Datasets:

gabrielaltay

/

pmcoa

like 0

==== Front Sci Total Environ Sci Total Environ The Science of the Total Environment 0048-9697 1879-1026 Elsevier B.V. S0048-9697(21)05694-1 10.1016/j.scitotenv.2021.150616 150616 Article Recovery of microbiological quality of long-term stagnant tap water in university buildings during the COVID-19 pandemic Ye Chengsong a Xian Xuanxuan a Bao Ruihan a Zhang Yiting ab Feng Mingbao a Lin Wenfang b Yu Xin a⁎ a College of the Environment & Ecology, Xiamen University, Xiamen 361102, PR China b Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, PR China ⁎ Corresponding author. 27 9 2021 1 2 2022 27 9 2021 806 150616150616 1 9 2021 19 9 2021 22 9 2021 © 2021 Elsevier B.V. All rights reserved. 2021 Elsevier B.V. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Stagnant water can cause water quality deterioration and, in particular, microbiological contaminations, posing potential health risks to occupants. University buildings were unoccupied with little water usage during the COVID-19 pandemic. It's an opportunity to study microbiological quality of long-term stagnant water (LTSW) in university buildings. The tap water samples were collected for three months from four types of campus buildings to monitor water quality and microbial risks after long-term stagnation. Specifically, the residual chlorine, turbidity, and iron/zinc were disqualified, and the heterotrophic plate counts (HPC) exceeded the Chinese national standard above 100 times. It took 4-54 days for these parameters to recover to the routine levels. Six species of pathogens were detected with high frequency and levels (101-105 copies/100 mL). Remarkably, L. pneumophilia occurred in 91% of samples with turbidity > 1 NTU. The absence of the culturable cells for these bacteria possibly implied their occurrence in a viable but non-culturable (VBNC) status. The bacterial community of the stagnant tap water differed significantly and reached a steady state in more than 50 days. Furthermore, a high concentration of endotoxin (>10 EU/mL) was found in LTSW, which was in accordance with the high proportion of dead bacteria. The results suggested that the increased microbiological risks require more attention and the countermeasures before the building reopens should be taken. Graphical abstract Unlabelled Image Keywords University buildings Long-term stagnation water Pathogens Viable but non-culturable state Endotoxin Editor: Ouyang Wei ==== Body pmc1 Introduction The stagnation of tap water is usually a problem of individual and occasional cases in residential community buildings. However, it may be recurrent incidents in public buildings like university buildings because of the periodical reduction of water consumption during winter and summer vacations, which can last for one to three months. All of these phenomena cannot compare to the situation during the COVID-19 pandemic. Hundreds of millions of university students and faculty members all over the world were absent from campus for several months. For instance, all of the campus in China was unoccupied from December 2019, and the universities did not reopen until May or September 2020. Stagnation has been associated with the deterioration of water quality at the distribution system level with the ubiquitous presence of harmful chemicals (Arnold and Edwards, 2012; Dias et al., 2017) or pathogens (Rhoads et al., 2016; Salehi et al., 2020). It is quite worthy to find out what impact the tap water stagnation with such a long time and large scale has brought to tap water quality, especially to microbiological quality, which is essential for water safety. A significant decline in the physicochemical and aesthetic quality of drinking water would occur if the tap water stagnates in the distribution pipelines (Caitlin et al., 2020; Salehi et al., 2020). In the absence of fresh tap water supplements, the chlorine residue in water decayed rapidly from 2.0 mg/L to 0.5 mg/L in 2 days, to 0 mg/L in 5 days (Ling et al., 2018). Moreover, the disinfectant residue was found to decay >140 times faster than in corresponding municipal water at highly stagnant taps (Rhoads et al., 2016). The dissolved oxygen (DO) in the water will also go down (Caitlin et al., 2020). Both of them cause water to shift from the oxidizing to reducing environment, which would lead to the breakage and detachment of pipe wall scaling and dissolving of heavy metals from the pipe materials as well as the long-term immersion works. Zhang et al. (2020) found that the turbidity of tap water could increase from 0.3 to 1.7 NTU in 48 h stagnation. The same stagnation duration could result in a Cu increase from below detected to 1370 and 1680 μg/L in kitchen and bathroom tap water, respectively (Zlatanovic et al., 2017). The increases of heavy metals including Fe, Mn, Cu, Ni, Cd were also observed in unlined cast pipe scales in a longer period of stagnation (132 h) in Zhengzhou, China (Li et al., 2020). Microbial growth during water stagnation is well documented (Lautenschlager et al., 2010; Ling et al., 2018). For example, Chen et al. (2020) showed that the colony of total bacteria increased to >500 CFU/mL after short stagnation in water purifiers. Moreover, the stagnation was highly related to the occurrence of waterborne pathogenic microorganisms (Schwake et al., 2016; Kinsey et al., 2017). Several studies have identified the growth of Legionella, Mycobacterium avium, Pseudomonas aeruginosa during stagnation, although the curve may plateau (Bédard et al., 2016; Cooper et al., 2008; Haig et al., 2018). For example, in a hospital water pipe network with a retention time of 3-6 days, a high level of Legionella co-occurred with the high pipe scale debris detachment and low residual chlorine (Schwake et al., 2016), which matched well with the occurrence of Legionnaires disease. However, all of the above-mentioned studies mainly focused on the analysis of water quality variation during short-term stagnation with a duration from several hours to several days (Zhang et al., 2020; Zlatanovic et al., 2017), which was not comparable with the stagnation during the current COVID-19 outbreak. Unfortunately, very limited information is available regarding this long-term stagnation on the tap water quality, especially on the microbiological parameters. In this study, a field study was performed in University buildings located in a southeast city in China to elucidate the potential microbial risks and recovery periods of LTSW-induced contamination. Stagnation water samples were collected for three months (May to August 2020) from four types of college buildings, i.e., laboratory, canteen, teaching building, and dormitory buildings, after nearly four months of stagnation (middle January to May 2020). The samples were analyzed based on the physicochemical parameters, HPC, Taqman-based qPCR, flow cytometry (FCM), endotoxin analysis, and high-throughput sequencing. This study aims to address the knowledge gap of LTSW-induced microbial risks and provide useful advice on the safety control of drinking water. 2 Materials and methods 2.1 Sampling sites and water consumption The studied facilities were three universities and one residential community in a city in Fujian province, Southeastern China. The campus water supply usages include everyday life (except cooking), cooking in the canteen, teaching, and lab work, etc. The detailed information of sampling site locations including frequency and water usage is shown in Table 1 . All water samples were taken from the end of the water supply (tap water from the sink). University A was the main sampling site, of which the information is shown in Fig. 1 . For university A, the winter vacation began in January 2020. Then, the sampling areas had seldom used tap water until graduate students returned to school in May because of the unexpected COVID-19 pandemic. In May 2020, the first month of sampling, the research and teaching activities were far from the normal levels since only a small part of students (about 1/12) returned to the campus (Fig. S1). The water consumption in the laboratory and teaching building of University A was almost zero. With the increased numbers of students returning to campus, the water supply was gradually restored to normal in August 2020. As additional sampling sites for both endotoxin and Legionella pneumophila indicators, the sinks of Universities B and C did not open during the sampling period. Their campus was closed for students and only opens for faculty and staff. By contrast, the residential community is located near University B, and the water supply in the community had not been interrupted and normal water consumption was maintained.Table 1 Description of sampling site locations including frequency and water usage. Water usage was based on system operator's observations and knowledge of building operations. Table 1Sampling site Description Size Sampling frequency Usage 1 University A, laboratory, sink, 13,000 students 2600 faculty members Two times per week then reduced to weekly Low-frequency usage building 2 University A, canteen, sink, 2nd floor Two times per week then reduced to weekly Low-frequency usage building 3 University A, teaching building, washroom sink, 2nd floor Two times per week then reduced to weekly Low-frequency usage building 4 University A, dormitory, sink, Two times per week then reduced to weekly Low-frequency usage building 5 University A, dormitory, sink, Once Low-frequency usage building 6 Community near university B, sink 1027 families Monthly, two times Continuously water consumption building 7 University B, teaching building, washroom sink, 1nd floor 15,000 students 2590 faculty members Monthly, two times Long stagnation building 8 University B, dormitory, sink, 1nd floor Monthly, two times Long stagnation building 9 University C, laboratory, washroom sink, 1nd floor 20,000 students 1100 faculty members Monthly, three times Long stagnation building 10 University C, teaching building, washroom sink, 1nd floor Monthly, three times Long stagnation building 11 University C, dormitory, sink, 1nd floor Monthly, three times Long stagnation building Fig. 1 Photographs of (a) sampling site information of the university A, (b) laboratory sampling site taps, (c) canteen sampling site taps, (d) teach building sampling site taps, and (e) dormitory sampling site taps. Fig. 1 2.2 Water sampling The study was mainly carried out in University A, and stagnation water sampling was performed over three months between May 2020 and August 2020. The water samples were taken in four types of public buildings in the university (Fig. 1). Besides, water samples were taken from University B, C and a residential community near University B, along with University A to determine endotoxin and L. pneumophila for the hazard assessment of long-term water retention. The residential community could be regarded as a negative control since its water supply was uninterrupted (Table 1). All stagnant water samples were collected from the tap at the sink. The water containers and tools were sterilized before sampling to eliminate the possible interference of bacterial contamination. Before sampling, the tap water will be flushed for 3 min. 10 L of tap water was collected at each sampling point, and the samples were transferred to the laboratory within 6 h (Guo et al., 2020). 2.3 Measurements of the physicochemical parameters of water samples The chlorine residue, DO, oxidation-reduction potential (ORP), turbidity, temperature, and pH were determined in the field. The residual chlorine was measured using the N,N-diethyl-p-phenylenediamine (DPD) chlorine analyzer (HACH, USA). All other parameters were determined in situ by a multi-parameter water quality analyzer (HACH, USA). The water samples were firstly filtered using a 0.45 μm polyethersulfone (PES) membrane (Millipore, USA) to remove the particles. Heavy metals were analyzed with inductively coupled plasma mass spectrometry (ICP-MS) (Agilent Technologies Inc. USA) after acidizing the water samples with nitric acid (5 mL/L) to pH < 2 (Sigma, USA). The total organic carbon (TOC) was detected by the TOC analyzer (Shimadzu, Japan). The nutrients, including the total dissolved nitrogen (TDN), total dissolved phosphorus (TDP), NO2 -, NO3 -, and NH4 +, were determined by automatic nutrients analyzer AA3 (San++ analyzer, Germany). Specifically, TDN and TDP were oxidized by 4% alkaline potassium persulfate before analysis. The concentrations of the above parameters were calculated according to the pre-conFig.d standard curves. 2.4 Microbial counting 1 mL of water samples was applied on nutrient agar (NA) (Hopebio, China) at 37 °C for 48 h to enumerate the total bacteria. The selective media for screening frequently-occurred pathogens was used to count P. aeruginosa (CN agar), E. coli (mTEC agar), Enterococcus faecalis (mEI agar), Shigella sp. (SS agar) (Hopebio, China), Salmonella sp. (BSA agar) (BD, USA) (USEPA, Method 1103.1: 2002, USEPA, Method 1600: 2006; Guo et al., 2020). In detail, the 100 mL water samples were concentrated by 0.45 μm PES membrane (Millipore, USA). Filtration membranes containing enriched bacteria were cultured on selected media. Two parallel groups were set for each sample. The specific preparation method and corresponding pathogen culture conditions are shown in Text S1. Concerning endotoxin, its concentration was determined using an Endotoxin test Limulus Kit (Bioendo Technology Co., Ltd., China) following the manufacturer's instructions. The ratio of dead and alive bacterial cells was determined using FCM (Millipore Guava EasyCyte, USA) and LIVE/DEAD BacLight bacterial viability kit (Invitrogen, Inc. USA) according to Lin et al. (2017). SYTO 9 and propidium iodide (PI) were added into water samples with a final concentration of 2 μM and 20 μM, respectively. 5000 cells for each sample were counted after being stained for 20 min in the dark under 488 nm of light irradiation at room temperature. 2.5 Taqman probe-based qPCR The water samples for qPCR and Illumina sequencing were concentrated by 0.22 μm PES membrane (Millipore, USA) and stored at -20 °C prior to DNA extraction. The genomic DNA was extracted using the FastDNA SPIN Kit (MP Biomedicals, USA) following the manufacturer's instructions. The Taqman-based probe was selected and designed (Table S1). Six representative waterborne pathogens (i.e., E. faecalis, P. aeruginosa, E. coli, Salmonella sp., L. pneumophila, Shigella sp.) were determined. The qPCR system with a final volume of 20 μL contained 10 μL of 2 × Taqman™ Gene Expression Master Mix (Thermo Fisher Scientific, USA), 0.05 μL of the probe (10 μM) (Sangon Biotech, China), 0.8 μL of each primer (10 μM), 2 μL of template DNA, and 6.35 μL of DNA-free water. The qPCR program consisted of a pro-denaturation step for 60 s at 95 °C, followed by 40 cycles of a denaturation step for 15 s at 95 °C, and an annealing step for 60 s at 60 °C using the ABI Q6 system (Life Technology, Singapore). Each target gene was run in triplicates. The standard curves were constructed from 10-fold serial dilutions of the plasmid standards that carry the target genes (Table S1). By comparison, the negative control used DNA-free water as the DNA template. 2.6 High-throughput sequencing (HTS) The sequencing analysis of water samples was performed on the Illumina NovaSeq platform (Illumina, USA). Briefly, quality controlled genomic DNA (1 ng/μL) was amplified with the bacteria-specific primers (338 F/806 R) containing a barcode. The PCR products were detected by electrophoresis with 2% agarose gel and recovered using the gel recovery Kit (Qiagen, Germany). TruSeq® DNA PCR-Free Sample Preparation Kit (Illumina, USA) was used to construct the library. After the library was qualified, NovaSeq6000 was used for sequencing (Novogene Science and Technology Co., Ltd., China). 2.7 Data analysis The data were plotted in Prism 8.0 and processed with R studio, SPSS 16.0, and Prism 8.0. Uparse v7.0.1001 (http://www.drive5.com/uparse/) was adopted for OTU clustering. Mothur was used for species annotation according to the SSUrRNA database (http://www.arb-silva.de/), and BugBase was used for functional annotation. Continuous sampling was selected to characterize the repetition and stability of the data and reduce the experimental error. 3 Results and discussion 3.1 Increased Zn and Fe of stagnant water samples In this study, eighteen physicochemical water quality parameters of the samples in University A were regularly measured, all of which could meet the tap water quality standard except the ones in Fig. 2 . As seen, the long-time water retention resulted in the deterioration of two metals (Zn and Fe), residual chlorine, and turbidity. The other measured physicochemical parameters of stagnant water samples were summarized and are illustrated in Table S2. Long-term water retention did not have a significant impact on these indicators.Fig. 2 The unqualified physicochemical parameters of different types of water samples collected during the stagnation. (a) Zinc, (b) iron, (c) residual chlorine, and (d) turbidity. Point “0” on the abscissa represents the starting point of the four-month water retention. The ordinate value of the dotted line is the national standard limit value. Fig. 2 After long-time stagnation, the leaching of both zinc (Zn2+) and iron (Fe2+/Fe3+) occurred in the plumbing system (Fig. 2). On the first day of the restarted water consumption, all of the samples suffered from the highest zinc and iron levels. The concentrations of zinc reached 7716.0, 6378.0, and 3082.0 μg/L in the samples from the laboratory, teaching building, and dormitory, respectively, which dramatically exceeded the Chinese national standard (i.e., 1000 μg/L). Similarly, the corresponding levels of iron were 1620.8, 700.4, and 716.9 μg/L, respectively, which were higher than the national standards of 300 μg/L. As the water usage got back to routine, the concentrations of zinc dropped rapidly and below the national standard within a week. Comparatively, the iron was restored even more quickly in three days. Compared with the other buildings, the concentrations of zinc and iron from canteen samples were lower than the national standard all the time, which may be attributed to the higher frequency and amount of water usage. Cast iron and galvanized steel are widely used in the stem pipes in water distribution systems. It should be the main reason that the concentrations of zinc and iron in the stagnant water were elevated. Besides the debris detachment due to the long-term immersion, previous studies have confirmed that the metal ions could leak from the corrosion layer through the electrochemical reactions (Clark et al., 2015; Lasheen et al., 2008; Li et al., 2020). However, the iron and zinc concentrations in our study were much higher than those reported in Li et al. (2020) for metal release in 132 h of stagnation pipes (Fe: 190-260 μg/L, Zn: 1-10 μg/L), which might be attributed to that long retention time during the pandemic enhanced the corrosion and aquatic chemistry process. In addition, even the zinc and iron concentrations in the tap water could go down to the routine levels shortly in several days, special attention should be paid that the long water stagnation would probably have created corrosion “hot spots” in the pipe walls, which could pose long-term adverse impacts (Masters et al., 2015). 3.2 Decayed chlorine residue and increased turbidity of stagnant water samples The qualified drinking water should contain residual chlorine to suppress bacterial regrowth. Because of the chlorine decay, the stagnant water contained much lower chlorine levels than the fresh one. In the first week of the water re-consumption, the residual chlorine in all samples was below 0.05 mg/L, except for the canteen water samples (i.e., 0.05-0.11 mg/L) (Fig. 2(c)). While the retained water was subsequently consumed, more and more freshwater was supplemented into the water supply system. The residual chlorine concentrations gradually increased to the standard requirement (0.05 mg/L) in a period from 10 days to 48 days (the laboratory samples). According to the data issued by the local administration in December 2019 (before the COVID-19 pandemic) (Water Quality Bulletin, 2019), the level of residual chlorine in pipes of this area was 0.48 mg/L that was much higher. Residual chlorine has been well recognized as the most important factor for microbial inhibition (Caitlin et al., 2020; Zhang et al., 2020). The low level of residual chlorine in the pipe network is certainly a concerning safety hazard. In this study, the recovery duration of residual chlorine (especially laboratory water samples) was dramatically longer than the short period of water retention (Zhang et al., 2019, Zhang et al., 2020), which suggested the occurrence of a much longer microbiological risk. Except for the laboratory samples, the long-term water retention appears to exert much fewer effects on turbidity (Fig. 2(d)). The average turbidity of the dormitory, teaching building, and canteen was 0.30 NTU, 0.20 NTU, and 0.23 NTU, respectively, similar to those (<0.5 NTU) before the pandemic in this area. However, the maximum turbidity of the laboratory samples was 4.94 NTU, which exceeded the standard concentration (1 NTU) by nearly four times. It took 24 days to get back to the normal level. The water quality with high turbidity could probably provide suitable conditions for microbial attachment and biofilm formation (Schwake et al., 2016). In fact, this study also found that the high turbidity was accompanied by the detection of L. pneumophilia, which was discussed in detail in Section 3.5. 3.3 Significant increase in culturable bacteria The influence of tap water supply recovery on the total culturable bacteria is reported in Fig. 3 . It was observed that water retention would result in increased HPC values, which was consistent with the findings of previous studies (Pepper et al., 2004). However, compared with short-term stagnation (Chen et al., 2020), much more culturable bacteria were detected in this work due to the much longer retention. According to the issued data by the local administration, the HPCs were usually zero CFU/mL. But the measured bacteria in all four types of samples exceeded the Chinese national standard (100 CFU/mL). Because the scientific research activities were at a standstill during the epidemic, the laboratory samples had the most serious bacterial contamination with significantly higher HPC concentrations of up to 1.5 × 104 CFU/mL, two more magnitude orders than the standard. The recovery period for laboratory tap water was the longest. The HPC took about eight weeks to fall to the routine levels (no detection). Concerning three other sampling sites, the water usage was a little higher due to the daily life of the small number of persons on the campus. All of them fluctuated sharply during the first sampling month with several samples exceeding the standards, respectively. They merely differed from the laboratory samples in the much shorter recovery durations of 4-5 weeks.Fig. 3 HPC and residual chlorine levels during stagnation. The black arrow indicates that the number of culturable bacteria would not drop to 0 CFU/mL immediately after the increase of residual chlorine (>0.5 mg/L). The dotted line is the time point when the number of culturable bacteria drops to 0 CFU/mL. There was a lag relationship between culturable bacteria and residual chlorine. Fig. 3 Microbial growth depends on different environmental factors, such as temperature, disinfectant residue, nutrients, and pipe network material, etc. The correlations between the HPC concentration and residual chlorine was thus analyzed by using Spearman, in which a significant correlation (P < 0.05) was found between HPC and residual chlorine in the laboratory samples (Fig. 3). This phenomenon was also documented in a recent study on the HPC growth in the drinking water system (Lin et al., 2020). It was interesting that the HPC value always dropped to zero CFU/mL after one week when the residual chlorine level reached up to the national standard (0.05 mg/L) (see the black arrows in Fig. 3). Since the residue chlorine is easier and real-time to be detected, this time lag might be used as an indicator ahead of time for the microbiological safety of the tap water with a long stagnation. 3.4 Safety risks of pathogenic microorganisms After long-time retention, the structures of the bacterial community in four different sampling sites were significantly different. The Principal Co-ordinates Analysis (PCoA) analysis showed that at the beginning of this study, four kinds of tap water samples were distributed in four different quadrants (Fig. S2), that is, their distribution was very dispersed. However, when the water supply resumed for about two months, microbial communities of the laboratory, canteen, and dormitory became uniformed. Among them, the community structure of canteen samples did not change significantly, which may be related to the constant tap water usage during the pandemic period. On the contrary, the teaching work always adopted the online mode, and only a small amount of flushing water was consumed, which might lead to the unstable results of microbial communities. This phenomenon suggested that the biological stability was gradually recovered in the tap water along with the water supply resumption. Besides the total bacteria, the absolute abundance of six typical waterborne pathogens in the retained water samples was determined in this study since they were directly related to human health (Fig. 4 ). To obtain the occurrence of pathogenic bacteria more accurately, the Taqman-based qPCR method with higher sensitivity and specificity was adopted. Based on the standard curves (Table S1), the minimum detection limits (MDLs) of all pathogens were about 10 copies/mL, except for Shigella sp. and L. pneumophila whose MDLs were 10-100 copies/mL. As presented in Fig. 4, all pathogens were detected. The pathogen with the highest detection level (1.95 × 105 copies/100 mL) was from L. pneumophila in the laboratory samples. Comparatively, the highest levels for Salmonella spp., Shigella sp., E. coli, and P. aeruginosa were 1.70, 7.08, 7.24, 1.62 × 103 copies/100 mL, respectively. The recovery for these pathogenic microorganisms, i.e., below their MDL values, was about 2-5 weeks except for L. pneumophila. The absolute abundance of pathogens in different water samples was also higher when water quality parameters deteriorate. Guo et al. (2020) found that the levels of these pathogenic bacteria in the effluent of full-scale drinking water treatment plants (DWTP) remained at 0-102 copies/100 mL. The relatively high detection levels and long recovery period suggested that the health risks from the pathogenic bacteria in the retained water were much higher and should receive more attention.Fig. 4 The absolute abundance of typical pathogens by using Taqman-qPCR. (a) L. pneumophilia, (b) Salmonella spp., (c) Shigella sp., (d) E. coli, (e) P. aeruginosa, and (f) E. faecalis. Unit: log10 copies/100 mL. Fig. 4 3.5 Detection of L. pneumophila and its relationship with residual chlorine and turbidity It is noteworthy to mention that L. pneumophila was continuously detected at high levels within three months during the recovery period of tap water supply in the laboratory samples (Fig. 4). Garrison et al. (2016) concluded that L. pneumophila was one of the most established causes of potable water-related disease outbreaks in the building plumbing systems. The outbreak of L. pneumophila was mainly connected with residual chlorine decay, iron release, and water stagnation. In fact, the decay of chlorine, which is the specific agent to inhibit microbial growth, was somewhat the result of the latter and could be accelerated by the latter (Lautenschlager et al., 2010; Ling et al., 2018). Likewise, Beer et al. (2015) and Shah et al. (2018) identified that depletion of residual disinfectant was the reason for Legionnaires disease outbreaks in public buildings. The above results suggested that chlorine maintenance should play a key role in the control of L. pneumophila in the stagnant water and its recovery. For example, shock disinfection within three weeks of planned occupancy was recommended for controlling remediation of Legionella colonization in the USA (ASHRAE Standards Committee, 2018). In addition, it is of great significance to make early warning of Legionella in the LTSW environment. A co-occurring phenomenon was found for the detection of L. pneumophila and the initial turbidity of LTSW samples in this study. In particular, L. pneumophila was detected in 91% of the water samples with high turbidity (>1 NTU) (Fig. 5 ). L. pneumophila preferred to live in biofilm or other microbial aggregates in pipe walls or other media surfaces (Garrison et al., 2016; Proctor et al., 2018). In stagnant water, the mild hydraulic conditions and quick chlorine decay would accelerate biofilm formation. High turbidity implied higher numbers of particular matters such as the scaling debris, which was advantageous for biofilm formation. So the high turbidity might be used as an indicator of early warning of L. pneumophila.Fig. 5 The abundance of L. pneumophila at different levels of turbidity. Data were collected from the samples taken in the first month of University A, and samples from University B and C. Percentage indicated the detection rate of L. pneumophila under specific water quality conditions. For example, 91% means that under the condition of turbidity greater than 1 NTU, the detection rate of L. pneumophila is 91% in water samples. Fig. 5 3.6 VBNC bacteria in the stagnant water In this study, the bacterial colonies with different morphology and colors were selected on the mediums, and a total of 86 strains were identified (Fig. 6(a)). It could be seen that Sphingomona was the dominant genera, accounting for 54.7% of all bacteria detected, followed by Methylorubrum at 13.3%. In addition, Acinetobacter, Aeromonas, and Pseudomonas were screened. Sphingomona is persistent and widely distributed in poor nutrition environments (e.g., mineral water or tap water) (Koskinen et al., 2000; Lee et al., 2001) and seldom present virulence or pathogenicity to human beings. However, a large amount of the bacterial cells might accumulate endotoxins when they were dead and decomposed, which would be discussed later. Methylorubrum, Pseudomonas, Acinetobacter, and Aeromonas were all reported as chlorine-resistant bacteria (Koskinen et al., 2000; Zhang et al., 2019; Zeng et al., 2020). Zhang et al. (2019) found that, with the increased secretion of extracellular polymers, Methylorubrum can form biofilms and thus resist the disinfectants. Overall, these pathogens were persistent in the LTSW environment, which require more attention. UV-based disinfection methods (e.g., UV/hydrogen peroxide and UV/peroxymonosulfate) were recommended for efficient inactivation of chlorine-resistant bacteria (CRB), therefore inhibiting the formation of biofilms (Zeng et al., 2020).Fig. 6 Bacterial community composition of (a) culturable bacteria, (b) pathogens-HTS, and (c) top 35 genera-HTS of the samples from university A. Values indicate the log10-transformed relative abundance of bacteria in each genus. Fig. 6 In this study, the top 35 genera in bacteria communities analyzed via HTS are listed in Fig. 6(c). Phreatobacter was the dominant genera, and its abundance was ranged from 17% to 81%, followed by Sphingomona with an abundance of 2%-17%. HTS results further confirmed that the proportion of pathogens in the total bacterial community was relatively low. The possible pathogens-HTS results were selected for the composition analysis of the pathogenic microorganisms (Fig. 6(b)). We found that results were much different from that culturable bacteria identification. In addition to the identified pathogenic bacteria in Fig. 6(a), 16S rRNA gene fragments of the E. coli, Helicobacter, Legionella, Mycobacterium, Staphylococcus, Streptomyces, and other suspected pathogens were sequenced by the HTS analysis, while the culturable ones were not detected in the selective medium (Fig. S3). Although the HTS results do not mean the existence of the active microbes, the risks of pathogenic bacteria should not be ignored, especially considering their entrance into the VBNC state. Legionella and Mycobacterium had the highest abundance, reaching 7.2 × 10-4 and 5.2 × 10-3, respectively. These results were consistent with the Taqman-based qPCR method (Fig. 4). Based on the above analysis, VBNC pathogens occurred very likely in the LTSW environment and could evade the HPC detection standards. Similarly, Kinsey et al. (2017) reported that P. aeruginosa outbreak in a neonatal intensive care unit (ICU) was related to water retention in hospitals, which deserves more attention in terms of their potential health risks. Felföldi et al. (2010) observed a higher detection of positive samples for Legionellae using the qPCR technique compared to the cultivation method. Since the culturing methods as HPC are not applicable in detecting VBNC bacteria, the real infectious risks of the LTSW environment might be inaccurately estimated in many cases. 3.7 High-level endotoxin in LTSW Endotoxin, composed of lipopolysaccharides, is a component of the cell wall of gram-negative (G-) bacteria. It is also called “pyrogen”, which could cause fever, microcirculation disorder, endotoxin shock, and disseminated intravascular coagulation, etc. (Anderson et al., 2002; Liao et al., 2010). The endotoxin was mainly released by the G- bacteria after death (Ren et al., 2019; Xue et al., 2019). It could be concluded that the bacterial biomass kept relatively high levels in the stagnant tap water. During the beginning days of this study, the 16s rRNA genes were at log 7-9 copies/L, and most of the cultural bacteria were at log 1-4 CFU/mL. Since the stagnant period was over 4 months, which obviously exceeded the bacterial growth cycles in most natural and artificial circumstances, it could be inferred that the bacterial biomass would be in a pseudo-steady state (Chen et al., 2020), i.e., the dead bacterial cell numbers should be equivalent to the newly-divided cells, during most time of the stagnant. Therefore, another problem for LTSW was the accumulation of endotoxin produced by the in-situ lysis of the bacteria. In this study, the levels of endotoxin in LTSW samples (the initial stage of water supply restoration) were analyzed (Fig. 7 (a)). The results showed that the endotoxin levels were all increased in LTSW compared with the control group (i.e., tap water of always used). t-Test showed that the results were significantly different (P < 0.05), except for the D2 sample. FCM results showed that the proportion of dead bacteria in the LTSW (69.0%-96.7%) was significantly higher than that in the tap water always used (53.4%) (Fig. 7(b)). A greater proportion of bacteria in the retained water samples were in a state of membrane damage or even breakage. This further confirmed that high contents of endotoxin were related to the percentage of dead bacteria. Traditional biological indicators cannot reflect the contamination level of endotoxin. The presence of endotoxin in the LTSW is worthy of attention owing to its environmental persistence and pathogenicity.Fig. 7 (a) Content of endotoxin and (b) the proportion of dead and alive bacteria in different types of stagnant water. Data were collected from samples 6-11 in Table 1. B: blank control, water sample free of endotoxin; C: control (sample 6), the tap water came from the residential area nearly university; L: laboratory (sample 9); T: teaching building (sample 7, 10); D: dormitory (sample 5, 8, 9). “*” indicates a significant difference between the sample and the control group (P < 0.05). Fig. 7 3.8 Prevention of water quality issues during the COVID-19 pandemic The University buildings impacted by COVID-19 had reduced or no water use for months. Our study confirmed that long-term water retention poses serious microbiological risks and thus prevention of water quality issues is essential. First of all, routine flushing is the most direct solution to pathogen control. Freshwater is regularly introduced to the pipeline network, and the stagnant environment cannot be formed, which helps prevent the problems. For the secondary water supply system of university buildings, attention should be paid especially to the cleaning of water tanks. It should be noted that recommissioning flushing could only reduce the levels of coliforms and heavy metals (Caitlin et al., 2020) but opportunistic pathogens can continue to grow (Hozalski et al., 2020), so it must be carried out with routine flushing during the COVID-19 pandemic. However, the frequency of routine flushing is difficult to determine. Factors such as plumbing design, the complexity of components, and the stored volume of water relative to water use need to be considered comprehensively. In the case of Legionella, Totaro et al. (2018) showed that effective control was achieved by maintaining a flushing frequency of 2 h. In addition, it is necessary to clean the water tank again for the university buildings with secondary water supply. It is important to maintain a disinfectant residual. By introducing a high level of disinfection for a short time, shock disinfection could effectively control pathogenic bacteria but must be weighed against the formation of disinfection by-products. Also, the water quality of University buildings should be monitored more frequently. Tap water with unqualified residual chlorine could be used for landscape irrigation, floor washing and other non-drinking purposes. If conditional, an automatic disinfectant device could be added to increase the delivery of disinfectant residual. 4 Conclusions The global outbreak of the COVID-19 has led to the ultra long-term stagnation of tap water in public buildings such as those in university campuses all over the world. It was of common sense that stagnation would result in the overall deterioration of water quality including both the chemical and microbiological aspects. However, the impacts from such an ultra long-term stagnation, i.e. several months, were still unclear. Especially what microbiological risks the stagnation would bring was of great interest due to the increasing concerns from the public under the context of the epidemic. It was expected the conclusions below could answer at least part of the questions.1) Long-term tap water retention resulted in the deterioration of water quality, while heavy metals (e.g., iron and zinc), turbidity, and chlorine were four key physicochemical parameters significantly exceeding the water quality guidelines, among which the latter two were closely connected to the microbial contamination. 2) Significant microbial growth occurred in the stagnant water, and the highest HPC of the samples reached two magnitude orders higher than the standards. It took 1-2 months to recover the bacterial levels to routine levels, which were much longer than the physicochemical parameters. 3) The microbiological risks in the LTSW were further confirmed by ubiquitous occurrence of six pathogenic species, among which L. pneumophilia had the highest detection frequency. However, these pathogens should probably be in the VBNC state due to the absence of the culturable ones. High turbidity (>1 NTU) might be an indicator for L. pneumophilia, suggested by their co-occurrence. 4) Endotoxin was a risk that has been overlooked in previous studies. A higher concentration of endotoxin (>10 EU/mL) in LTSW samples was detected, which resulted from the death of the high contents of the G- bacteria. 5) Routine flushing and shock disinfection were recommended as the possible microbiological risks control methods during the COVID-19 pandemic. CRediT authorship contribution statement Chengsong Ye: Conceptualization, Methodology, Investigation, Software, Writing-original draft. Xuanxuan Xian: Sampling, Methodology, Visualization. Ruihan Bao: Sampling, Methodology. Yiting Zhang: Methodology, Visualization, Software. Mingbao Feng: Conceptualization, Writing-review & editing. Wenfang Lin: Writing-review & editing. Xin Yu: Conceptualization, Writing-review & editing, supervision. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix A Supplementary data Supplementary material Image 1 Acknowledgments This research was supported by the 10.13039/501100001809 Natural Science Foundation of China (NSFC) (41861144023, and U2005206), 10.13039/100010166 Xiamen Municipal Bureau of Science and Technology (YDZX20203502000003), the 10.13039/501100003392 Natural Science Foundation of Fujian Province (2020J05090). Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.scitotenv.2021.150616. ==== Refs References Anderson W.B. Slawson R.M. Mayfied C.I. A review of drinking water-associated endotoxin, including potential routes of human exposure Can. J. Microbiol. 48 7 2002 567 587 12224557 Arnold R.B. Edwards M. Potential reversal and the effects of flow pattern on galvanic corrosion of lead Environ. Sci. Technol. 46 2012 10941 10947 22900550 ASHRAE Standards Committee ANSI/ASHRAE standard 188-2018. Legionellosis: risk management for building water systems. Atlanta, GA www.ashrae.org/technology 2018 Beer K.D. Gargano J.W. Roberts V.A. Hill V.R. Garrison L.E. Kutty P.K. Hilborn E.D. Wade T.J. Fullerton K.E. Yoder J.S. Surveillance for waterborne disease outbreaks associated with drinking water-United States, 2011–2012 Morb. Mortal. Wkly Rep. 64 31 2015 842 848 Bédard E. Prévost M. Déziel E. Pseudomonas aeruginosa in premise plumbing of large buildings Microbiol. Open 2016 1 20 Caitlin P. William J.R. Tim K. Considerations for Large Building Water Quality After Extended Stagnation 2020 10.31219/osf.io/qvj3b Chen X. Wang Y. Li W.Y. Zhao X.L. Lu Y.F. Yu Y.J. Chen S. Ding Z.B. Microbial contamination in distributed drinking water purifiers induced by water stagnation Environ. Res. 188 2020 109715 Clark B.N. Masters S.V. Edwards M.A. Lead release to drinking water from galvanized steel pipe coatings Environ. Eng. Sci. 32 8 2015 713 721 Cooper I.R. White J. Mahenthiralingam E. Hanlon G.W. Long-term persistence of a single Legionella pneumophila strain possessing the Mip gene in a municipal ahower despite repeated cycles of chlorination J. Hosp. Infect. 70 2008 154 159 18723253 Dias V.C.F. Besner M.C. Prévost M. Predicting water quality impact after district metered area implementation in a full-scale drinking water distribution system J. Am. Water Works Assoc. 109 2017 E363 E380 Felföldi T. Heéger Z. Vargha M. Marialigeti K. Detection of potentially pathogenic bacteria in the drinking water distribution system of a hospital in Hungary Clin. Microbiol. Infect. 16 1 2010 89 92 19519854 Garrison L.E. Kunz J.M. Cooley L.A. Moore M.R. Lucas C. Schrag S. Sarisky J. Whitney C.G. Vital Signs: Deficiencies in Environmental Control Identified in Outbreaks of Legionnaires' Disease-North America 2000–2014 16 2016 Centers for Disease Control and Prevention 3049 3058 Guo L.Z. Wan K. Zhu J.W. Ye C.S. Chabi K. Yu X. Detection and distribution of vbnc/viable pathogenic bacteria in full-scale drinking water treatment plants J. Hazard. Mater. 406 4 2020 124335 Haig S.J. Kotlarz N. Lipuma J.J. Raskin L. A high-throughput approach for identification of nontuberculous mycobacteria in drinking water reveals relationship between water age and Mycobacterium avium mBio 9 1 2018 e02354 Hozalski R.M. LaPara T.M. Zhao X.T. Kim T. Waak M.B. Burch T. McCarty M. Flushing of stagnant premise water systems after the COVID-19 shutdown can reduce infection risk by Legionella and Mycobacterium spp Environ. Sci. Technol. 54 24 2020 15912 15924 Lautenschlager K. Nico B. Wang W.W. Egli T. Hammes F. Overnight stagnation of drinking water in household taps induces microbial growth and changes in community composition Water Res. 44 17 2010 4868 4877 20696451 Kinsey C.B. Koirala S. Solomon B. Rosenberg J. Robinson B.F. Neri A. Halpin A.L. Arduino M.J. Moulton-Meissner H. Noble-Wang J. Chea N. Gould C.V. Pseudomonas aeruginosa outbreak in a neonatal intensive care unit attributed to hospital tap water Infect. Control Hosp. Epidemiol. 1–8 2017 Koskinen R. Ali-Vehmas T. Kampfer P. Laurikkala M. Tsitko I. Kostyal E. Atroshi F. Salkinoja-Salonen M. Characterization of Sphingomonas isolates from Finnish and Swedish drinking water distribution systems J. Appl. Microbiol. 89 2000 687 696 11054174 Lasheen M.R. Sharaby C.M. El-Kholy N.G. Elsherif I.Y. El-Wakeel S.T. Factors influencing lead and iron release from some Egyptian drinking water pipes J. Hazard Mater. 160 2008 675 680 18436372 Lee J.S. Shin Y.K. Yoon J.H. Takeuchi M. Pyun Y.R. Park Y.H. Sphingomonas aquatilis sp. nov. Sphingomonas koreensis sp. nov. and Sphingomonas taejonensis sp. nov. yellow-pigmented bacteria isolated from natural mineral water Int. J. Syst. Evol. Microbiol. 51 2001 1491 1498 11491350 Li M.J. Wang Y.H. Liu Z.W. Sha L. Korshin G.V. Chen Y.C. Metal-release potential from iron corrosion scales under stagnant and active flow, and varying water quality conditions Water Res. 175 2020 115675 Liao V.H.C. Chou W.C. Chio C.P. Ju Y.R. Liao C.M. A probabilistic approach to quantitatively assess the inhalation risk for airborne endotoxin in cotton textile workers J. Hazard. Mater. 177 1–3 2010 103 108 20036462 Lin H.R. Ye C.S. Chen S. Zhang S.H. Yu X. Viable but non-culturable E. coli induced by low level chlorination have higher persistence to antibiotics than their culturable counterparts Environ. Pollut. 230 2017 242 249 28662489 Lin W.F. Ye C.S. Guo L.Z. Hu D. Yu X. Analysis of microbial contamination of household water purifiers Appl. Microbiol. Biotechnol. 104 5 2020 Ling F.Q. Whitaker R. LeChevallier M.W. Liu W.T. Drinking water microbiome assembly induced by water stagnation ISME J. 12 2018 1520 1531 29588495 Masters S. Parks J. Atassi A. Edwards M.A. Distribution system water age can create premise plumbing corrosion hotspots Environ. Monit. Assess. 187 2015 559 26251058 Pepper I.L. Rusin P. Quintanar D.R. Haneya C. Josephsona K.L. Gerba C.P. Tracking the concentration of heterotrophic plate count bacteria from the source to the consumer's tap Int. J. Food Microbiol. 92 3 2004 289 295 15145587 Proctor C.R. Reimann M. Vriens B. Hammes F. Biofilms in shower hoses Water Res. 131 2018 274 286 29304381 Totaro M. Valentini P. Costa A.L. Giorgi S. Casini B. Baggiani A. Rate of legionella pneumophila coloniza- tion in hospital hot water network after time flow taps installation J. Hosp. Infect. 98 2018 60 63 28890285 Ren Y.R. Kong J.Y. Xue J.L. Shi X.J. Li H.J. Qiao J. Lu Y. Effects of ozonation on the activity of endotoxin and its inhalation toxicity in reclaimed water Water Res. 154 2019 153 161 30782557 Rhoads W.J. Pruden A. Edwards M.A. Survey of green building water systems reveals elevated water age and water quality concerns Environ. Sci. Water Res. Technol. 2 2016 164 173 Salehi M. Odimayomi T. Ra K. Ley C. Julien R. Nejadhashemi A.P. Hernandez-Suarez J.S. Mitchell J. Shah A.D. Whelton A. An investigation of spatial and temporal drinking water quality variation in green residential plumbing Build. Environ. 169 2020 106566 Schwake D. Garner E.D. Strom O.R. Pruden A. Edwards M.A. Legionella DNA markers in tap water coincident with a spike in Legionnaires' disease in Flint, MI Environ. Sci. Technol. Lett. 3 9 2016 311 315 Shah P.P. Barskey A.E. Binder A.M. Legionnaires' Disease Surveillance Summary Report, United States: 2014-1015 2018 Division of Bacterial Diseases, National Center for Immunization and Respiratory Diseases USEPA, Method 1103.1 Escherichia coli (E. coli) in Water by Membrane Filtration Using Membrane-Thermotolerant Escherichia coli Agar (mTEC) 2002 U.S. Environmental Protection Agency: Office of Water USEPA, Method 1600 Enterococci in Water by Membrane Filtration Using Membrane-Enterococcus Indoxyl-beta-D-glucoside Agar (mEI) 2006 U.S. Environmental Protection Agency: Office of Water Xiamen Municipal Garden Bureau Water Quality Bulletin in December 2019 2019 http://szyl.xm.gov.cn/js/csszgb/202002/t20200228_2425148.htm Xue J.L. Zhang J.S. Qiao J. Lu Y. Effects of chlorination and combined UV/Cl2 treatment on endotoxin activity and inhalation toxicity of lipopolysaccharide, gram-negative bacteria and reclaimed water Water Res. 155 2019 124 130 30836264 Zeng F. Cao S. Jin W. Zhou X. Ding W.Q. Renjie Tu R.J. Han S.F. Wang C.P. Jiang Q.J. Huang H. Ding F. Inactivation of chlorine-resistant bacterial spores in drinking water using UV irradiation, UV/Hydrogen peroxide and UV/Peroxymonosulfate: efficiency and mechanism J. Clean. Prod. 243 2020 118666 Zhang H.Y. Tian Y.M. Kang M.X. Chen C. Song Y.R. Li H. Effects of chlorination/chlorine dioxide disinfection on biofilm bacterial community and corrosion process in a reclaimed water distribution system Chemosphere 215 2019 62 73 30312918 Zhang M.L. Xu M.Y. Xu S.F. Zhang L.Y. Lin K.Z. Zhang L. Bai M. Zhang C. Zhou H. Response of the bacterial community and antibiotic resistance in overnight stagnant water from a municipal pipeline Int. J. Environ. Res. Public Health 17 6 2020 1995 32197379 Zlatanovic L. van der Hoek J.P. Vreeburg J.H.G. An experimental study on the influence of water stagnation and temperature change on water quality in a full-scale domestic drinking water system Water Res. 123 2017 761 772 28732329

33992139

PMC9752783

NO-CC CODE

2022-12-16 23:25:24

no

Lancet. 2021 May 13 15-21 May; 397(10287):1805-1806

latin-1

Lancet

10.1016/S0140-6736(21)00890-4

oa_other

==== Front Sci Total Environ Sci Total Environ The Science of the Total Environment 0048-9697 1879-1026 Elsevier B.V. S0048-9697(21)05535-2 10.1016/j.scitotenv.2021.150458 150458 Article Release of tens of thousands of microfibers from discarded face masks under simulated environmental conditions Wu Pengfei ab1 Li Jiangpeng ac1 Lu Xiao ac Tang Yuanyuan ac⁎ Cai Zongwei b⁎⁎ a State Environmental Protection Key Laboratory of Integrated Surface Water-Groundwater Pollution Control, School of Environmental Science and Engineering, Southern University of Science and Technology, Shenzhen 518055, PR China b State Key Laboratory of Environmental and Biological Analysis, Department of Chemistry, Hong Kong Baptist University, Hong Kong SAR, PR China c Guangdong Provincial Key Laboratory of Soil and Groundwater Pollution Control, School of Environmental Science and Engineering, Southern University of Science and Technology, Shenzhen 518055, PR China ⁎ Correspondence to: Y. Tang, State Environmental Protection Key Laboratory of Integrated Surface Water-Groundwater Pollution Control, School of Environmental Science and Engineering, Southern University of Science and Technology, Shenzhen 518055, PR China. ⁎⁎ Corresponding author. 1 These authors contributed equally to the work. 28 9 2021 1 2 2022 28 9 2021 806 150458150458 26 7 2021 13 9 2021 15 9 2021 © 2021 Elsevier B.V. All rights reserved. 2021 Elsevier B.V. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. While mechanical abrasion by water and sediment is a primary and critical step in weathering process, the upsurge of discarded face masks will undoubtedly become a potential source of micro-/nanofibers owing to the spread of novel coronavirus (COVID-19) pneumonia. However, effects of mechanical abrasion on discarded face masks have neither been seriously addressed nor understood. Therefore, we conducted a simulated experiment to explore abundance, size distribution and morphology of microfibers released from common, surgical and face filtering piece (FFP) masks after mechanical abrasion. Technologies such as Fourier transform infrared spectrometry, fluorescence microscopy, scanning electron microscopy, and confocal laser scanning microscopy were used. Results showed that the abundance of released microfibers followed order of surgical > common > FFP in both water and sediment environments, and the maximum abundance reached 272 ± 12.49 items per square centimeter of mask (items·cm−2) after sediment abrasion. Taking surgical mask for further investigation, the length of released fiber was observed to vary from 47.78 μm to 3.93 mm, and 72.41–89.58% of the total number of released microfibers fell in the range of 0.1–1 mm. However, microfibers with a very small length (1–100 μm) can occupy 0.09–13.59% of the total number of released fibers in sediment environment. The roughness of fiber surface after sediment abrasion was successively increased. Furthermore, the morphology analysis showed significant changes with countless cracks and many prominent protrusions on fiber surface after sediment abrasion. The cracks and protrusions may further accelerate mask decomposition, thereby potentially resulting in the adsorption of other contaminants and the release of self-containing chemicals. This study provides a valuable database of microfibers released from discarded face masks at the primary but critical stage, and further contributes knowledge on environmental impact of discarded personal protective equipment due to COVID-19. Graphical abstract Unlabelled Image Keywords Face masks Mechanical abrasion Microfiber release Confocal microscopy COVID-19 Editor: Jay Gan ==== Body pmc1 Introduction Novel coronavirus pneumonia (COVID-19) has been rapidly spreading worldwide since the beginning of 2020 (Murray et al., 2020). The World Health Organization (WHO) calls for strict interventions to prevent COVID-19 from further spread, if which, wearing a face mask is considered to be one of the best protective measures that can effectively block the spread of coronavirus by droplets from presymptomatic and asymptomatic individuals (Howard et al., 2020; Xu et al., 2020). This has led to a dramatic increase in global face mask production (up to 396.6%) by 2020 (Chua et al., 2020; Fadare and Okoffo, 2020). The daily capacity of masks in China has increased by 450%, from the 20 million at the beginning to 110 million by the end of February 2020 (Ren, 2020; Singh et al., 2020a). It was anticipated that even in the post-pandemic period, the compound annual growth rate (CAGR) of face masks could still be increased by 10.8% from 2020 to 2027 (Global Protective Face Masks Market (GPFMM), 2020). Such an extensive production of face masks will not only create a massive disruption of the upstream supply chain but also overwhelm the downstream waste management problems (Klemeš et al., 2020; Narendra et al., 2021; Singh et al., 2020b). If the used face masks are not properly managed, they will eventually enter and persist in the environment (Huang and Morawska, 2019). At present, a large amount of mask debris has been observed on the coast of Hong Kong, the United States, France, and Mainland China (Yeh, 2020), while dozens of masks have also been found floating on the waves of the Mediterranean Sea (Roberts et al., 2020). The abandoned face masks are likely to be fragmented when subjected to weathering effects and then release small secondary particles that can be recognized as microplastics (1 μm–5 mm) or nanoplastics (<1 μm) (Peeken et al., 2018; Wu et al., 2020a). In addition, some microfibers can also be formed and trapped in the face masks when the thermoplastic fiber is extruded and blown by high velocity and temperature airflow during the manufacturing processes (Hutten, 2007; Li et al., 2021). Currently, micro/nanofibers, as the primary constituent of micro/nanoplastics, have been considered as emerging contaminants with needle-like shapes (Kutralam-Muniasamy et al., 2020; Liu et al., 2019). Compared to micro/nanoplastics with other shapes, the micro/nanofiber can more easily penetrate the biological membrane, leading to the dysregulation of tissues and organs(de Sá et al., 2018; Galloway and Lewis, 2016; Wu et al., 2020b, Wu et al., 2020c) and biomagnification through the food chain from plankton to the human colon (de Sá et al., 2018; Ibrahim et al., 2021). Weathering has been regarded as the most critical process for the aging of carbonaceous polymers, including ultraviolet irradiation, temperature degradation, oxidative transformation and mechanical abrasion (Wu et al., 2019). Recent studies have shown that biotic and abiotic hydrolysis can gradually erose the surface of plastics, thereby reforming various reactive oxygen groups to accelerate the oxidative transformation of the micro/nanoplastic (Min et al., 2020). Nanoplastics, together with styrene monomers, were found to be released from coffee cups into nature after being subjected to ultraviolet irradiation (Lambert and Wagner, 2016). Mechanical abrasion caused by tidal currents, especially plastic wastes on beaches, has been widely recognized as the primary and critical process for the generation of micro/nanoplastics (Chubarenko et al., 2020; Song et al., 2017). The occurrence of the discarded face masks on the beaches indicates that mechanical abrasion by water and sediment may be critically involved in the weathering processes, especially during the formation of microfibers at an early stage. Several publications point out that the face mask as a potential source of microfibers can quickly intensify the already critical situation (Morgana et al., 2021; Saliu et al., 2021; Wang et al., 2021). The mechanical strength of masks decreased after ultraviolet irradiation (Wang et al., 2021), and then macro-, micro-, and even nanoplastics can be released from the face masks under different shear stress intensities (Morgana et al., 2021). However, the weathering effect of mechanical abrasion caused by the discarded face masks has not been seriously addressed and fully understood by the current reported studies. Particularly during the COVID-19 pandemic, the upsurge of face masks promoted the emergence of investigating such effects on the formation of micro/nanofibers at an early stage. Therefore, in this study, the aging of face masks towards micro/nanofibers under mechanical abrasion by water and sediment was investigated. Face masks of different types and brands were applied, and the increase in microfiber abundance was evaluated over the abrasion time in water and sediment environments. The abundance, size and color were also described to explain the possible release behavior of microfibers from the surgical face masks. Furthermore, the surface morphologies of the masks were characterized to clarify the variation of surface roughness during the aging process. This study provides a comprehensive investigation of the release of microfibers from the discarded face masks under mechanical abrasion in different environments, and further contributes towards the expansion of valuable database in current studies on environmental burden caused by the dramatic increase in personal protective equipment due to the COVID-19 pandemic. 2 Materials and methods 2.1 Raw materials and sample preparation As shown in Table S1 of the Supplementary Information (SI), three types of face masks generally used during the COVID-19 pandemic were selected for this research, including common masks, surgical masks, and face filtering piece masks (FFP; e.g. N95) (Greenhalgh et al., 2020). Three layers, namely the inner, middle and outer layers, consisted of the face masks mentioned above. To simulate a real environment, sediments were sampled from Soko Beaches, Hong Kong. The sediments were pretreated to remove the plastics according to the density separation and extraction reported by Wu et al. (2020b). The organic matter in the sediments was subsequently washed for three times with 18.2 MΩ deionized (DI) water (Millipore Co., USA) and combusted in a muffle furnace at 450 °C for 3 h. After cooling down, the particle size distribution (D50 = 562.25 μm) of the sediments was measured using a laser scattering technique (Mastersizer 3000, Malvern) (Fig. S1). Mechanical abrasion experiments were conducted in both the water and sediment environments. Each layer of the face mask was separated and cut into one square centimeter, and then transferred into a glass tube (Fisher Scientific, USA). The glass tubes were then filled with 15 mL of DI water or the saturated sands (6 g) with 10 mL of DI water with no head space, to represent the water or sediment environmental conditions, respectively. After that, all tubes were rotated end-over-end at 60 rpm at different time intervals from 0 to 240 h. After agitation, the suspension under water condition was filtered directly through a 1 μm GF/C glass fiber (Whatman, UK). Meanwhile, the microfibers in the sediment environment were suspended in 500 mL of NaI (Sigma-Aldrich, USA) solution (1.8 g·cm−3) by magnetic stirring for 4 h to separate microfibers. After settling for another 24 h, the solution was filtered through a glass fiber membrane, and then the membranes with extracted microfibers were transferred into the glass culture dish for oven-drying at 60 °C for 3 h for further analysis. All experiments were conducted in triplicates. For more detailed information, please refer to the extraction procedures reported in our previous study (Wu et al., 2020b). 2.2 Sample characterization and analysis Each layer of different mask types was measured using a Fourier transform infrared spectrometer (FT-IR) apparatus with attenuated total reflection (ATR) mode (Spotlight 200i, PerkinElmer, USA), and the spectra were recorded as six accumulations ranging from 400 to 4000 cm−1. According to the polymer library in the PerkinElmer database (around 16,000 reference spectra), PP polymer was used as the material for the three layers of the surgical masks, the middle and outer layer of the common masks, and the inner and outer layers of the N95 masks. The inner layers of the common masks and middle layer of the N95 masks were composed of high-density polyethylene (HDPE) (Table S1). Fluorescence microscopy (DM2500, Leica, Germany) with 10× magnification was applied to determine the abundance, size and color of the microfibers. The unit of the released microfibers was recorded as the number of microfibers per square centimeter of the face masks. The measurement of microfiber abundance was repeated six times using ImageJ software according to the protocol reported by Erni-Cassola et al. (2017). The comparison of microfiber abundance between each layer was investigated using Statistical Product and Service Solution 16.0 (SPSS Inc., USA) with one-way analysis of variance (ANOVA), and the statistical significance was set at a p-value < 0.05 (Barberán et al., 2012). The wear degree of the face masks was measured by confocal laser scanning microscopy (Nikon Eclipse Ti2, Japan) coupled with a Nikon Ti2-E inverted microscope platform. Small pieces of pristine and weathered face masks were cleaned and fixed with double-sided adhesive tape on. The samples were excited with the laser line at 491.2 nm and collected with a wavelength range of 494.0–531.0 nm. The scanning area was captured at 300 μm × 300 μm with a Z-step size spacing of 0.2 μm (50– 80 μm; controlled by Nikon NIS Elements AR software). The roughness was measured using ProflimOnline with surface analysis (Profilm Online, 2021). All the pristine and weathered face masks were fixed with carbon tape on the sample holder. The morphology with detailed element analysis was further analyzed using a scanning electron microscope (SEM; LEO1530, ZEISS, Germany) attached to an energy dispersive X-ray spectroscopy (EDS; ZEISS, Germany). The magnification of the observation was set at 0.2 k–30 k times by a secondary electron detector at 5 kV, while the mapping-mode EDS measurement was set at 20 kV. Prior to SEM-EDS, all samples were sputtered with a 3 nm layer of platinum using Leica coating system (Leica, Germany) to obtain a better conductivity. 2.3 Quality assurance and quality control Some quality control processes were adopted throughout the experiments. All containers were cleaned at least 3 times with DI water. A clean glass dish with a glass fiber membrane inside was placed on the experimental bench to collect airborne microfibers. Cotton lab-coat, nitrile gloves and cotton mask were worn to avoid cross-contamination during the extraction process. Procedural blanks with two replicates were also analyzed to cross-check microplastic contamination. Less than 2% of the microplastics were observed in each blank, indicating that the laboratory environment was clean enough to conduct microfiber experiments (Lin et al., 2018). 3 Results and discussion 3.1 Microfiber release from face masks with different types Fig. 1a illustrates the abundance of released microfibers from three types of face masks in water conditions; the total number was counted as 1909 items. Among them, the surgical face mask released the largest number of microfiber (mean ± SD: 272 ± 12.49 items·cm−2), comparing to the common face mask (165.7 ± 9.2 items·cm−2) and the FFP face mask (187.9 ± 9.45 items·cm−2). The abundance of the released microfiber varied among different layers of the three types of face masks, but the value can be ordered as middle layer > inner layer > outer layer for each type. The middle layer is mainly composed of 25 gsm (gram per square meter) of melt-blown fibers (0.5–10 μm) binding together insecurely through the melt blowing process (Table S1), while 20 gsm of the spunbonded non-woven fabrics were used for the inner and outer layers, resulting in the highest number of microfibers released from the middle layer (Fig. S2). For direct skin contact, the inner layer is relatively softer than the outer layer which is mainly designed according to the requirement for fluid repellency. Therefore, based on different functions, the inner layer with a velvet surface and loose structure is more likely to release microfibers in comparison with the outer layer (Du et al., 2020). This phenomenon was also observed in another study (Wang et al., 2021), which explained that the maximum load force of the middle layer (~3.5 N) is much lower than that of the outer and inner layers of the face masks. Thus, more microfibers can be released from the middle layer under the same weathering conditions.Fig. 1 The release of microfiber from different types and layers of the face masks. (a) Abundance and (b) size distribution of the microfibers released from the common, surgical and face filtering piece (FFP) masks. (c) The image, and (d, e) the abundance of the released microfiber from the outer, middle and inner layer of the surgical mask in both water and sediment environment. The time range for mechanical abrasion for the surgical mask is within the time range of 0–240 h, while the abundance of the microfiber in both water and sediment increased significantly in the first 96 h, and then gradually slowed down after approximately 168 h. On each time interval, the released microfiber abundance in sediment is higher than that in water (p-value < 0.05). Fig. 1 Mechanical abrasion can break the molecular chains of the polymer, resulting in a wide size range of the released microfibers. According to a previous study, microfibers can be classified as small (SMFs; 1–100 μm), medium (MMFs; 0.1–1 mm) and large (LMFs; 1–5 mm) by size (length) (Baldwin et al., 2016; Wu et al., 2020b). Therefore, based on this classification, the size distribution of the released microfibers in this study is summarized in Fig. 1b, which shows that the MMFs (0.1–1 mm) are predominant in every layer of all face masks, varying between 63.30 and 91.57%. The LMFs (1–5 mm) accounted for the second-highest proportion, ranging between 6.19% and 21.82%, followed by the SMFs (1.07–12.82%) with a size of 0.45–100 μm. It was noticed that some other studies also checked the microplastic release from the face masks, but with a larger size range (50–250 μm) caused by self-crosslinking under short-time UV weathering effects (Wang et al., 2021). Apart from the microfiber release from face masks, some other microplastics generated in the plastic product caused by physical effects in the water environment have also been reported. For example, the highest proportion of microplastics released from teabags is smaller than 0.15 mm (Hernandez et al., 2019). Du et al. (2020) also found that approximately 55–95.13% of microplastics with size <1 mm were released from take-out food containers after immersion in hot water. Meanwhile, Ranjan et al. (2021) reported a median value of 53.65 μm (~25,000 particles) from the paper cup after being contained in hot liquid (85–90 °C) for 15 min, which is smaller than that reported in our study. This discrepancy may be explained by the different shapes of the microplastic, and a longer size was observed for the fiber released from the face masks in comparison with the particles reported in other studies. 3.2 Microfiber release from surgical mask affected by brand, time, and environmental condition The surgical mask, as the most extensively used one among the three types of face masks (Howard et al., 2020), was found to have the most serious microfiber release, as shown in Fig. 1. Therefore, a surgical mask with five different brands was selected and intensively examined for the release behavior of the microfibers affected by abrasion time and environmental conditions. Table S2 shows similar releasing behavior of the microfiber from the five brand masks, with abundances of 272–297.5 items·cm−2 (P value < 0.05) and order as middle layer > inner layer > outer layer. The results indicated that the influence of the production process of different brands was negligible for the surgical mask. Therefore, the data obtained in this study can represent the general release level of microfibers from surgical face masks. Furthermore, the microfiber release from each layer of the surgical mask was evaluated, taking into consideration the environmental conditions (water or sediment) and time interval. Fig. 1c shows a photograph of the released fiber taken by a fluorescence microscope. It can be seen that the microfiber from the water environment is clean but that after sediment abrasion, it is entirely filled with sand particles. Fig. 1d–e further demonstrate the increasing tendency of the released microfibers in both water and sediment environments, that is the abundance increased significantly in the first 96 h, and then gradually slowed down after approximately 168 h. However, at each time interval, the abundance of the released microfiber in sediment was higher than that in water (P value < 0.05), with the discrepancy ranging from 4 items·cm−2 in the inner layer at 96 h to 31 items·cm−2 in the middle layer at 168 h. As shown in Fig. S3, a similar size distribution was observed for the microfibers obtained at each time interval. The MMFs (0.1–1 mm) accounted for the major component among the total amount (74.62–89.58% in water and 72.41–86.74% in sediment), followed by LMFs (7.64–22.54% in water and 9.71–23.2% in sediment) and SMFs (0–10.44% in water and 0.09–13.59% in sediment). After sediment abrasion, an obvious tendency was found, showing a decrease in the proportion of LMFs, but an obvious increase in SMFs over prolonged time for each layer of the surgical masks (Fig. S3). This phenomenon suggests that the SMFs can also be generated from the previously released larger microfibers in addition to directly originating from the raw surgical masks. In addition, as an important characteristic that can potentially reflect the origin of microfibers (Martí et al., 2020), the color was also determined and classified into six groups: transparent, blue, red, black, green, and brown. The transparent, translucent and white microfibers were classified as transparent because it is difficult to distinguish them due to diffuse reflection (Asamoah et al., 2019). As shown in Fig. 2 , the transparent microfiber accounted for 90.37% for the outer layer, 93.31% for the middle layer, and 91.80% for the inner layer in the water environment, and 91.26%, 93.85%, and 92.41% for the outer, middle and inner layers, respectively, after the mechanical processes in the sediment. The predominance of the transparent microfiber is reasonable as the face masks are mainly composed of colorless fibers. The rest are colored items consisting of blue, red, black, green, and brown, which are mainly attributed to the release of some impurities during the manufacturing processes of the face masks.Fig. 2 Color distribution of the released microfiber from the outer, middle and inner layer of the surgical mask in water and sediment environment. Transparent is the predominant color, accounting for 90.37–93.31% in water and 91.26–93.85% in sediment, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Fig. 2 3.3 Changes in morphology Changes in the surface texture are crucial considerations for evaluating the degree of weathering, especially after mechanical abrasion. In this study, two- or three-dimensional (2D or 3D) topographical images of the face masks before and after weathering processes were characterized using fluorescent and laser confocal microscopy, respectively (Fig. 3 ). The topography of the fibers from the raw mask was clearly observed, with the mean average diameter of 21.26 ± 6.08 μm (outer), 6.62 ± 1.58 μm (middle), and 18.58 ± 8.22 μm (inner). Subsequent to the abrasion processes, the topography changed dramatically (Fig. 3h–m); the average diameter of the fiber was changed to 21.75 ± 6.56 μm (outer), 7.81 ± 2.01 μm (middle), and 19.91 ± 9.36 μm (inner) in water condition and 21.21 ± 11.22 μm (outer), 7.52 ± 2.05 μm (middle), and 20.67 ± 11.21 μm (inner) in sediment condition, respectively. The obvious increase in the deviation of the fiber diameter of the face masks indicated that the roughness increased accordingly, owing to the mechanical abrasion. The surface roughness was then analyzed by measuring the average height of the peak above and below the test line drawn in the 2D topographical image (Fig. S4), with the calculation of arithmetic mean roughness (R a) and root mean square roughness (R q). The data in Table S3 shows that the R a and R q values are successively increased for all the mask fibers, including the pristine mask and the aged ones by water and sediment. Compared to water, the sediment caused an obviously larger increase in the fiber's roughness, because a higher Mohs hardness (~7) of the sediment can produce more energy during the abrasion to dissociate the C—C and C—H bonds of the molecular chain, and subsequently cause more severe mechanical abrasion (Ismail et al., 2020; Posch, 2011).Fig. 3 Florescent microcopy of the (a) outer, (b) middle and (c) inner layer of the pristine surgical face mask, and the confocal 3D micrographs (300 μm × 300 μm) of the outer, middle and inner layer of the pristine mask (d, e, f) and the mask after water (h, i, j) and sediment (k, l, m) abrasion. Fig. 3 The SEM-EDS analysis was further applied for surface characterization, which confirmed much stronger wearing caused by sediment abrasion (Fig. 4 ). Fig. 4(a–c) show the smooth surface of the virgin mask fiber but with some tiny cracks with a flaky surface, which may be self-carried during production (Han and He, 2021). After abrasion in water environment, the surface of the mask fiber was also relatively smooth (Fig. 4d–f), but with microplastics attached due to the breakage of the fiber. However, when the mask was aged in a sediment environment, significant changes with large numbers of cracks were detected on the fiber of the face masks (Fig. 4g–i). In addition to cracks, there are many prominent protrusions attached to the fiber surface, which were confirmed as sediment particles (Si oxides as the main component) from the EDS results (Fig. S5). The protrusions will further decrease the surface energy of the fiber, and accelerate microbial colonization when the microfiber is prolonged in the natural environment (Pan et al., 2019; Rummel et al., 2019). The appearance of the cracks and the attachment of the protrusions indicated that the fiber undergoes heavy friction by the sediment, which changes the surface properties of the fiber and therefore alters the ability of contaminant adsorption on the fiber surface (Liu et al., 2020). Such enhanced surface changes would accelerate the weathering process, the adsorption of other contaminants, and the release of self-containing chemicals such as formaldehyde and bronopol in the mask fibers (Aerts et al., 2020; Donovan and Skotnicki-Grant, 2007). The uptake of contaminants induces health risks through the following pathways: particle toxicity, chemical toxicity and pathogen/parasite vectors (Vethaak and Leslie, 2016; Wu et al., 2020c). Therefore, the emerging concern caused by the release of fibers from the discarded face masks is that the transport of the contaminants may become much easier and faster than the other plastic products (e.g. plastic bottles and bags) because of their smaller size, lower strength and more elastic surfaces of the fibers (Xu and Ren, 2021). Recent studies have documented that the particles, additives, and pathogens on polymer materials can cause potential toxicity to organisms (Rist et al., 2018). Microfibers with needle-like shapes can more easily penetrate the cell membrane together with plastic additives, inducing energy homeostasis, oxidative stress, circulatory systems, immune system dysregulation, and neurological dysfunction (de Sá et al., 2018; Galloway and Lewis, 2016; Rist et al., 2018; Wu et al., 2020c). Although no literature has reported the propagation of the coronavirus through plastics, it should be seriously considered that plastics can be vectors for the transmission of pathogens (e.g. Halofolliculina) or bacteria (e.g. Vibrio), causing the skeletal eroding band disease in coral reefs (Goldstein et al., 2014; Ziajahromi et al., 2017), particularly during the pandemic.Fig. 4 Surface morphology of the outer, middle and inner layer of the original surgical mask (a, b, c) and the mask after water (d, e, f) and sediment (g, h, i) abrasion. The fiber surface of the original face mask was relatively smooth. After water abrasion, the fiber surface was still smooth but with increased cracks. A significant difference was observed for the fiber surface after sediment abrasion, with numerous cracks and protrusions. Fig. 4 4 Conclusions This study provides exact data for the generation of microfibers and offers clear evidence for the severe destructive effects caused by mechanical abrasion. The results showed that the abundance of microfiber released from the surgical face masks can reach as high as 272.0 ± 12.49 items·cm−2. For a commonly used mask with dimensions of 20 cm × 10 cm, it can be estimated that after 240 h, approximately 54,400 ± 2498 items and 68,000 ± 4808 items can be released, attributed to water and sediment abrasion, respectively. More than 99% of microplastics are in the form of microfibers. Apart from the microfiber abundance, the dramatic increase in the surface roughness may decrease the surface energy and thereby accelerate microbial colonization, which can further accelerate the release of micro/nanofibers and other harmful chemicals during decomposition, embrittlement, and disintegration of the face masks. Therefore, this study can be of significant importance to understand the release behavior of microfibers from the face masks discarded in the natural environment, and to further contribute valuable information to the environmental transformation of plastics induced by weathering processes especially mechanical abrasion as the primary and critical step. CRediT authorship contribution statement Yuanyuan Tang and Zongwei Cai co-supervised the whole work with conceptualization and methodology. Pengfei Wu and Jiangpeng Li contributed equally to this work, by conducting the experiments, analyzing data and writing the draft manuscript. Xiao Lu helped with conducting experiments, analyzing data and revising the draft manuscript. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix A Supplementary data Supplementary material Image 1 Acknowledgment This work was supported financially by the 10.13039/501100001809 National Natural Science Foundation of China (NSFC) (41977329, 22106130), the 10.13039/501100003453 Natural Science Foundation of Guangdong Province (2021B1515020041), the Special Funds for the Cultivation of Guangdong College Students' Scientific and Technological Innovation (“Climbing Program” Special Funds pdjh2021c0038), the State Environmental Protection Key Laboratory of Integrated Surface Water-Groundwater Pollution Control, and the Guangdong Provincial Key Laboratory of Soil and Groundwater Pollution Control. We also thank Mr. Yingzhe She (Thermo Fisher Scientific, Guangzhou, China) for technical support. Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.scitotenv.2021.150458. ==== Refs References Aerts O. Dendooven E. Foubert K. Stappers S. Ulicki M. Lambert J. Surgical mask dermatitis caused by formaldehyde (releasers) during the COVID-19 pandemic Contact Dermatitis 88 2 2020 172 173 10.1111/cod.13626 Asamoah B.O. Kanyathare B. Roussey M. Peiponen K.E. A prototype of a portable optical sensor for the detection of transparent and translucent microplastics in freshwater Chemosphere 231 2019 161 167 10.1016/j.chemosphere.2019.05.114 31129396 Baldwin A.K. Corsi S.R. Mason S.A. Plastic debris in 29 Great Lakes tributaries: relations to watershed attributes and hydrology Environ. Sci. Technol. 50 2016 10377 10385 10.1021/acs.est.6b02917 27627676 Barberán A. Bates S.T. Casamayor E.O. Fierer N. Using network analysis to explore co-occurrence patterns in soil microbial communities ISME J. 6 2012 343 351 10.1038/ismej.2011.119 21900968 Chua M.H. Cheng W. Goh S.S. Kong J. Li B. Lim J.Y.C. Mao L. Wang S. Xue K. Yang L. Ye E. Zhang K. Cheong W.C.D. Tan Beng Hoon Li Z. Tan Ban Hock Loh X.J. Face masks in the new COVID-19 Normal: materials, testing, and perspectives Research 2020 2020 1 40 10.34133/2020/7286735 Chubarenko I. Efimova I. Bagaeva M. Bagaev A. Isachenko I. On mechanical fragmentation of single-use plastics in the sea swash zone with different types of bottom sediments: insights from laboratory experiments Mar. Pollut. Bull. 150 2020 10.1016/j.marpolbul.2019.110726 de Sá L.C. Oliveira M. Ribeiro F. Rocha T.L. Futter M.N. Studies of the effects of microplastics on aquatic organisms: what do we know and where should we focus our efforts in the future? Sci. Total Environ. 645 2018 1029 1039 10.1016/j.scitotenv.2018.07.207 30248828 Donovan J. Skotnicki-Grant S. Allergic contact dermatitis from formaldehyde textile resins in surgical uniforms and nonwoven textile masks Dermatitis 18 1 2007 40 44 10.2310/6620.2007.05003 17303043 Du F. Cai H. Zhang Q. Chen Q. Shi H. Microplastics in take-out food containers J. Hazard. Mater. 399 2020 122969 10.1016/j.jhazmat.2020.122969 Erni-Cassola G. Gibson M.I. Thompson R.C. Christie-Oleza J.A. Lost, but found with Nile red: a novel method for detecting and quantifying small microplastics (1 mm to 20 μm) in environmental samples Environ. Sci. Technol. 51 2017 13641 13648 10.1021/acs.est.7b04512 29112813 Fadare O.O. Okoffo E.D. Covid-19 face masks: a potential source of microplastic fibers in the environment Sci. Total Environ. 737 2020 140279 10.1016/j.scitotenv.2020.140279 Galloway T.S. Lewis C.N. Marine microplastics spell big problems for future generations Proc. Natl. Acad. Sci. U. S. A. 113 2016 2331 2333 10.1073/pnas.1600715113 26903632 Global Protective Face Masks Market (GPFMM) Global opportunity analysis and industry forecast (2019–2026): impact analysis of COVID-19 on protective face masks market [WWW Document]. URL https://www.researchdive.com/covid-19-insights/359/protective-face-masks-market 2020 Goldstein M.C. Carson H.S. Eriksen M. Relationship of diversity and habitat area in North Pacific plastic-associated rafting communities Mar. Biol. 161 2014 1441 1453 10.1007/s00227-014-2432-8 Greenhalgh T. Schmid M.B. Czypionka T. Bassler D. Gruer L. Face masks for the public during the covid-19 crisis BMJ 369 2020 1 4 10.1136/bmj.m1435 Han J. He S. Need for assessing the inhalation of micro(nano)plastic debris shed from masks, respirators, and home-made face coverings during the COVID-19 pandemic Environ. Pollut. 268 2021 115728 10.1016/j.envpol.2020.115728 Hernandez L.M. Xu E.G. Larsson H.C.E. Tahara R. Maisuria V.B. Tufenkji N. Plastic teabags release billions of microparticles and nanoparticles into tea Environ. Sci. Technol. 53 2019 12300 12310 10.1021/acs.est.9b02540 31552738 Howard J. Huang A. Li Z. Tufekcim Z. Zdimale V. van der Westhuizen H.-M. von Delfto A. Price A. Fridman L. Tang L.-H. Tang V. Watson G.L. Baxs C.E. Shaikh R. Questier F. Hernandez D. Chu L.F. Ramirez C.M. Rimoin A.W. Face masks against COVID-19: an evidence review Proc. Natl. Acad. Sci. U. S. A. 1–8 2020 10.20944/preprints202004.0203.v1 Huang W. Morawska L. Face masks could raise pollution risks Nat. Comment. 574 2019 29 30 10.1038/d41586-019-02938-1 Hutten I.M. Processes for nonwoven filter media Handb. Nonwoven Filter Media 195–244 2007 10.1016/b978-185617441-1/50020-2 Ibrahim Y.S. Tuan Anuar S. Azmi A.A. Wan Mohd Khalik W.M.A. Lehata S. Hamzah S.R. Ismail D. Ma Z.F. Dzulkarnaen A. Zakaria Z. Mustaffa N. Tuan Sharif S.E. Lee Y.Y. Detection of microplastics in human colectomy specimens JGH Open 5 2021 116 121 10.1002/jgh3.12457 33490620 Ismail M.Y. Patanen M. Kauppinen S. Kosonen H. Ristolainen M. Hall S.A. Liimatainen H. Surface analysis of tissue paper using laser scanning confocal microscopy and micro-computed topography Cellulose 27 2020 8989 9003 10.1007/s10570-020-03399-w Klemeš J.J. Fan Y.Van Tan R.R. Jiang P. Minimising the present and future plastic waste, energy and environmental footprints related to COVID-19 Renew. Sust. Energ. Rev. 127 2020 10.1016/j.rser.2020.109883 Kutralam-Muniasamy G. Pérez-Guevara F. Elizalde-Martínez I. Shruti V.C. An overview of recent advances in micro/nano beads and microfibers research: critical assessment and promoting the less known Sci. Total Environ. 740 2020 139991 10.1016/j.scitotenv.2020.139991 Lambert S. Wagner M. Characterisation of nanoplastics during the degradation of polystyrene Chemosphere 145 2016 265 268 10.1016/j.chemosphere.2015.11.078 26688263 Li L. Zhao X. Li Z. Song K. COVID-19: performance study of microplastic inhalation risk posed by wearing masks J. Hazard. Mater. 411 2021 1 9 10.1016/j.jhazmat.2020.124955 Lin L. Zuo L.Z. Peng J.P. Cai L.Q. Fok L. Yan Y. Li H.X. Xu X.R. Occurrence and distribution of microplastics in an urban river: a case study in the Pearl River along Guangzhou City, China Sci. Total Environ. 644 2018 375 381 10.1016/j.scitotenv.2018.06.327 29981986 Liu P. Qian L. Wang H. Zhan X. Lu K. Gu C. Gao S. New insights into the aging behavior of microplastics accelerated by advanced oxidation processes Environ. Sci. Technol. 53 2019 3579 3588 10.1021/acs.est.9b00493 30829479 Liu P. Lu K. Li J. Wu X. Qian L. Wang M. Gao S. Effect of aging on adsorption behavior of polystyrene microplastics for pharmaceuticals: adsorption mechanism and role of aging intermediates J. Hazard. Mater. 384 2020 121193 10.1016/j.jhazmat.2019.121193 Martí E. Martin C. Galli M. Echevarría F. Duarte C.M. Cózar A. The colors of the ocean plastics Environ. Sci. Technol. 54 2020 6594 6601 10.1021/acs.est.9b06400 32392043 Min K. Cuiffi J.D. Mathers R.T. Ranking environmental degradation trends of plastic marine debris based on physical properties and molecular structure Nat. Commun. 11 2020 10.1038/s41467-020-14538-z Morgana S. Casentini B. Amalfitano S. Uncovering the release of micro/nanoplastics from disposable face masks at times of COVID-19 J. Hazard. Mater. 419 2021 126507 10.1016/j.jhazmat.2021.126507 Murray O.M. Bisset J.M. Gilligan P.J. Hannan M.M. Murray J.G. Respirators and surgical facemasks for COVID-19: implications for MRI Clin. Radiol. 75 2020 405 407 32299617 Narendra S. A O.O. Yuanyuan T. Medical waste: current challenges and future opportunities for sustainable management Crit. Rev. Environ. Sci. Technol. 2021 1 23 10.1080/10643389.2021.1885325 Pan Z. Guo H. Chen H. Wang S. Sun X. Zou Q. Zhang Y. Lin H. Cai S. Huang J. Microplastics in the northwestern Pacific: abundance, distribution, and characteristics Sci. Total Environ. 650 2019 1913 1922 10.1016/j.scitotenv.2018.09.244 30286357 Peeken I. Primpke S. Beyer B. Gütermann J. Katlein C. Krumpen T. Bergmann M. Hehemann L. Gerdts G. Arctic Sea ice is an important temporal sink and means of transport for microplastic Nat. Commun. 9 2018 10.1038/s41467-018-03825-5 Posch W. Polyolefins Appl. Plast. Eng. Handb. 23–48 2011 10.1016/B978-1-4377-3514-7.10003-0 Ranjan V.P. Joseph A. Goel S. Microplastics and other harmful substances released from disposable paper cups into hot water J. Hazard. Mater. 404 2021 124118 10.1016/j.jhazmat.2020.124118 Ren Z. COVID-19: China boosts face mask production capacity by 450 per cent in a month, threatening a glut scenario [WWW Document] https://www.scmp.com/business/companies/article/3075289/china-boosts-face-mask-production-capacity-450-cent-month 2020 Rist S. Carney Almroth B. Hartmann N.B. Karlsson T.M. A critical perspective on early communications concerning human health aspects of microplastics Sci. Total Environ. 626 2018 720 726 10.1016/j.scitotenv.2018.01.092 29396337 Roberts K. Bowyer C. Kolstoe S. Fletcher S. Coronavirus face masks: an environmental disaster that might last generations [WWW Document] https://theconversation.com/coronavirus-face-masks-an-environmental-disaster-that-might-last-generations-144328 2020 Rummel C.D. Escher B.I. Sandblom O. Plassmann M.M. Arp H.P.H. Macleod M. Jahnke A. Effects of Leachates from UV-weathered microplastic in cell-based bioassays Environ. Sci. Technol. 53 2019 9214 9223 10.1021/acs.est.9b02400 31257880 Saliu F. Veronelli M. Raguso C. Barana D. Galli P. Lasagni M. The release process of microfibers: from surgical face masks into the marine environment Environ. Adv. 4 2021 100042 10.1016/j.envadv.2021.100042 Singh N. Tang Y. Ogunseitan O.A. Environmentally sustainable management of used personal protective equipment Environ. Sci. Technol. 10–12 2020 10.1021/acs.est.0c03022 Singh N. Tang Y. Zhang Z. Zheng C. COVID-19 Waste Management: Effective and Successful Measures in Wuhan, China 2020 Song Y.K. Hong S.H. Jang M. Han G.M. Jung S.W. Shim W.J. Combined effects of UV exposure duration and mechanical abrasion on microplastic fragmentation by polymer type Environ. Sci. Technol. 51 2017 4368 4376 10.1021/acs.est.6b06155 28249388 Vethaak A.D. Leslie H.A. Plastic debris is a human health issue Environ. Sci. Technol. 50 2016 6825 6826 10.1021/acs.est.6b02569 27331860 Wang Z. An C. Chen X. Lee K. Zhang B. Feng Q. Disposable masks release microplastics to the aqueous environment with exacerbation by natural weathering J. Hazard. Mater. 417 2021 126036 10.1016/j.jhazmat.2021.126036 Wu P. Cai Z. Jin H. Tang Y. Adsorption mechanisms of five bisphenol analogues on PVC microplastics Sci. Total Environ. 650 2019 671 678 10.1016/j.scitotenv.2018.09.049 30212696 Wu P. Tang Y. Cao G. Li J. Wang S. Chang X. Dang M. Jin H. Zheng C. Cai Z. Determination of environmental micro(nano)plastics by matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry Anal. Chem. 2020 10.1021/acs.analchem.0c01928 Wu P. Tang Y. Dang M. Wang S. Jin H. Liu Y. Jing H. Zheng C. Yi S. Cai Z. Spatial-temporal distribution of microplastics in surface water and sediments of Maozhou River within Guangdong-Hong Kong-Macao Greater Bay Area Sci. Total Environ. 717 2020 135187 10.1016/j.scitotenv.2019.135187 Wu P. Tang Y. Jin H. Song Y. Liu Y. Cai Z. Consequential fate of bisphenol-attached PVC microplastics in water and simulated intestinal fluids Environ. Sci. Ecotechnol. 2 2020 100027 10.1016/j.ese.2020.100027 Xu E.G. Ren Z.J. Preventing Masks From Becoming the Next Plastic Problem 15 2021 6 8 Xu H. Yan C. Fu Q. Xiao K. Yu Y. Han D. Wang W. Cheng J. Possible environmental effects on the spread of COVID-19 in China Sci. Total Environ. 731 2020 139211 10.1016/j.scitotenv.2020.139211 Yeh J. Where did 5,500 tonnes of discarded face masks end up? [WWW Document] https://www.greenpeace.org/international/story/44629/where-did-5500-tonnes-of-discarded-face-masks-end-up/ 2020 Ziajahromi S. Kumar A. Neale P.A. Leusch F.D.L. Impact of microplastic beads and fibers on waterflea (Ceriodaphnia dubia) survival, growth, and reproduction: implications of single and mixture exposures Environ. Sci. Technol. 51 2017 13397 13406 10.1021/acs.est.7b03574 29059522

34592275

PMC9752784

NO-CC CODE

2022-12-16 23:25:24

no

Sci Total Environ. 2022 Feb 1; 806:150458

utf-8

Sci Total Environ

10.1016/j.scitotenv.2021.150458

oa_other

==== Front Water Res Water Res Water Research 0043-1354 1879-2448 Elsevier Ltd. S0043-1354(21)00231-1 10.1016/j.watres.2021.117033 117033 Article An investigation into the leaching of micro and nano particles and chemical pollutants from disposable face masks - linked to the COVID-19 pandemic Sullivan G.L. a Delgado-Gallardo J. b Watson T.M. a Sarp S. b⁎ a SPECIFIC, College of Engineering, Swansea University, SA2 8PP, UK b SPEC, College of Engineering, Swansea University, SA2 8PP, UK ⁎ Corresponding author. 10 3 2021 15 5 2021 10 3 2021 196 117033117033 9 2 2021 5 3 2021 8 3 2021 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The production of disposable plastic face masks (DPFs) in China alone has reached to approximately 200 million a day, in a global effort to tackle the spread of the new SARS-CoV-2 virus. However, improper and unregulated disposals of these DPFs has been and will continue to intensify the plastic pollution problem we are already facing. This study focuses on the emission of pollutants from 7 DPF brands that were submerged in water to simulate environmental conditions if these DPFs were littered. The DPF leachates were filtered by inorganic membranes, and both particle-deposited organic membranes and the filtrates were characterized using techniques such as FTIR, SEM-EDX, Light Microscopy, ICP-MS and LC-MS. Micro and nano scale polymeric fibres, particles, siliceous fragments and leachable inorganic and organic chemicals were observed from all of the tested DPFs. Traces of concerning heavy metals (i.e. lead up to 6.79 µg/L) were detected in association with silicon containing fragments. ICP-MS also confirmed the presence of other leachable metals like cadmium (up to 1.92 µg/L), antimony (up to 393 µg/L) and copper (up to 4.17 µg/L). LC-MS analysis identified polar leachable organic species related to plastic additives and contaminants; polyamide-66 monomer and oligomers (nylon-66 synthesis), surfactant molecules, dye-like molecules and polyethylene glycol were all tentatively identified in the leachate. The toxicity of some of the chemicals found and the postulated risks of the rest of the present particles and molecules, raises the question of whether DPFs are safe to be used on a daily basis and what consequences are to be expected after their disposal into the environment. Graphical abstract Image, graphical abstract ==== Body pmc1 Introduction The SARS CoV-2 pandemic has caused an unprecedented rise of face mask wearing in order to mitigate the spread of this airborne viral disease (Wilson et al., 2020), as recommended by the World Health organisation (WHO)(Adyel, 2020; Aragaw, 2020; Fadare and Okoffo, 2020). This has seen an increase in mass production of face coverings to keep up with global demand. The WHO has estimated an additional 89 million disposable plastic face masks (DPFs) are required a month in order to protect health care workers alone (Aragaw, 2020; Fadare and Okoffo, 2020). Presently, China is the main manufacturer and distributer of DPFs, with an estimated production of 200 million DPFs per day (Aragaw, 2020), with up to a 1 billion DPFs being used globally on a monthly basis (Adyel, 2020).  This has seen a notable increase in DPF and clinical waste, irresponsibly discarded and disposed at landfill sites; with a large proportion inevitably ending up in the water course and the ocean (Fadare and Okoffo, 2020). The rise in DPFs waste is regarded as a new cause of pollution directly linked COVID-19 pandemic. However, there is still very little data on the effect of the pollution caused by DPFs on the wider environment (Adyel, 2020). Most DPFs are made entirely from plastic fibres (Aragaw, 2020; Jung et al., 2021), containing high percentages of polypropylene (PP) and polyethylene (PE), and other polymeric materials such as nylon and polystyrene (Aragaw, 2020; Fadare and Okoffo, 2020; Jung et al., 2021) with novelty face masks often coloured with dyes for customer appeal. The manufacturing of DPFs often involves electrospinning micro and nanofibres  of plastic, into three layers  (Aragaw, 2020; Fadare and Okoffo, 2020). In some manufacturing processes, silica nano particles maybe added as filler to enhance plastic properties such increasing mechanical strength and toughness of the material (Wu et al., 2005). There is a growing concern around the environmental fate of DPFs and whether they emit pollutants such as microscopic polymeric fibres (Aragaw, 2020; Fadare and Okoffo, 2020), micro-crystalline silica and other secondary pollutants, such as leachable chemicals (dyes, surfactants and glues). There is little information on the environment impacts of these pollutants and whether they may enter the food chain. Many of these chemical dyes are associated with heavy inorganic metals (Sungur and Gülmez, 2015) and are known to have an adverse impact on the environmental and potentially human health. Metals such as antinomy (Sb), copper (Cu) and chromium (Cr) are used as catalysts in dye manufacture and sometime residues are found in textiles materials as plastic additives (Hahladakis et al., 2018; Sungur and Gülmez, 2015). Some reactive dyes also form complexes with heavy metals, such as nickel (Ni), Cu and Cr. These chemicals may be associated with particles that can be transmitted from the face masks and enter the respiratory track via inhalation or  absorbed through skin contact; these chemicals may dissolve in moisture droplets of sweat and saliva, which may act as a medium for transport into the body. There are known hazards associated with these chemicals, especially heavy metals, ranging from mild allergic reactions, often from limited exposure, to more serious health issues from repeated exposure; such as renal disease, emphysema, cancer and often may be harmful to unborn children in pregnancy (Sungur and Gülmez, 2015; Tchounwou et al., 2012). Dye compounds themselves pose risk to environmental and public health (Lellis et al., 2019). As many of them are water soluble organic molecules and leachable, they can therefore enter the water course and food chain. Most dyes are also chromophores, competing for light in the environment and reducing the photosynthesis of aquatic plants, and thus disrupting the ecosystem (Lellis et al., 2019). Due to the polarity of these molecules, they are often difficult to remove by conventional water treatment methods, ending up in drinking water (Lellis et al., 2019). Most of these compounds have carcinogenic and mutagenic properties as they are, generally, highly aromatic. Furthermore, these dye molecules have the ability to enter cells  and intercalate with the cell's DNA, disrupting transcription processes of the cell (Lellis et al., 2019). Most dye compounds are deemed as persistent organic pollutants and have the potential to bioaccumulate in many species, including humans (Lellis et al., 2019). Not only chemical species but also the release of submicron particles from DPFs are cause of concerns. Recent studies of micro plastic (MPs)(>5 mm), and nano plastics (NP) (>1 µm), (Bianco and Passananti, 2020; Paul et al., 2020; Toussaint et al., 2019) which included polymeric fibres, showed that they have detrimental effects in animal models. MPs and NPs exhibit cytotoxic and genotoxic effects in terrestrial (including humans) and aquatic organisms (Bouwmeester et al., 2015). In recent articles, researchers have shown that some MPs and NPs can be adsorbed by the gut and then passed through the blood brain barrier (Kögel et al., 2020; Lusher et al., 2017) resulting in neurotoxic damage (Prüst et al., 2020). These particles often produce cellular damage by causing increased oxidative stress, as the particles are recognised by the organism as being foreign (Bhagat et al., 2020; Prüst et al., 2020). In one report, fibres were found to embed themselves easier in tissues rather than spherical particles. The result of which caused increased superoxide dismutase levels in gut tissue of zebra fish, an indication of oxidative stress and inflammation (Bhagat et al., 2020). Further to this, there is also evidence to suggest that NPs may be small enough to penetrate the cell wall, releasing persistent organic pollutants into tissues (Bhagat et al., 2020; Bouwmeester et al., 2015), or they can be activated NP particles, containing reactive sites on chain ends, which can inflict DNA damage. These effects can cause cell death, genotoxicity, or cancer formation. Additional to this, there is a potential human health concern around the presence of micro silica particles (SiMP) between 1 and 10 µm and nano silica particles (SiNP) <1 µm, which is often used in manufacturing process of plastics (Masuki et al., 2020; Murugadoss et al., 2017). There is strong evidence to support the human health implications of crystalline and amorphous silica if inhaled; this silica nano particles can cause lung irritation and silicosis (fibrosis of the lung) which  can develop into emphysema and lung cancer (Lin et al., 2006; Masuki et al., 2020; Murugadoss et al., 2017). SiMP and SiNP cause cell damage through oxidative stress, that leads to DNA damage, genotoxicity and cell death (Murugadoss et al., 2017).  SiMP and SiNP toxicity is also evident in other tissues, causing cancers in blood and bone tissue and causing neurotoxicity in brain tissues (Masuki et al., 2020; Murugadoss et al., 2017; You et al., 2018). However, SiMP and SiNP have lower environmental impact with mild toxic effects seen in aquatic organisms and terrestrial animals if ingested (Fruijtier-Pölloth, 2012). Nevertheless, the presence of these particles in DPFs water leachates, may be indicative of the ease of their discharge. This could raise further concerns regarding their potential release when wearing face coverings; could they be easily inhaled? The aim of this investigation is to identify the environmental impact of DPFs when disposed improperly, via formalizing a workflow procedure to characterize various pollutants emitted/leached from DPFs (Song et al., 2015; Sullivan et al., 2020) during simulated environmental conditions (face mask gently agitated in deionised water). 2 Materials and methods 2.1 Materials and instrumentation The following lists the consumables and instrumentation used in the present investigation: DPFs were purchased from several manufacturers and suppliers as listed in Table 1 and Fig. 1 , and contaminants were extracted using deionised water dispensed from Milli Q® type 1 dispenser, procedural blanks were also prepared using the same deionised water source. Whatman® Anodisc inorganic membranes of 0.1 μm pore size (Merck Group® UK) were used as membrane filters for particle deposition. A Glass vacuum manifold purchased from Sigma Aldrich®UK and vacuum pump Sparmax (The Airbrush Company Ltd, UK) were used to aid filtration of leachate. For microscopy analysis, a Zeiss Primotech light microscope (Carl Zeiss Ltd., Cambridge, UK) and TM3000 SEM electron microscope (Hitachi High-Technologies Corporation) were used for particle identification.Table 1 Shows the information associated with various DPFs analysed and supplier purchased from. Table 1Sample Description Brand Manufacturer Distributor Type Supplier Face Mask 1 Plain NA Huaxian Tiancheng Sanitary Material Co., Ltd NA NA Amazon Face Mask 2 Black PRO SFE Shanghai Careus Medical Product Co., Ltd Hunter Price International NA Poundland Face Mask 3 Plain children PMS International NA NA NA CKs store Face Mask 4 Novelty children (paw patrol) NA Guangzhou Quantum Laser Intelligent Equipment co., Ltd Sambro International Ltd. Type I Poundland Face Mask 5 Plain TLS Toiletry Sales Ltd Toiletry Sales Ltd Type IIR Sainsbury Face Mask 6 Plain CGM-3PLY Foshan Suncare Medical Products Co., Ltd Teucer (UK) Ltd NA Tesco Face Mask 7a Face Mask 7b Face Mask 7c Festive Red Festive blue Festive green NA Shanghai Ogo Medical Instrument Co., Ltd Sambro International Ltd. Type I Poundland Fig. 1 Shows images of the typical face masks used in this investigation. Image 1 corresponds to face mask 1; standard plain face mask (similar to face mask 3, 5 and 6), image 2; face mask 2 (black color), Image 4; face mask 4 (novelty kids), Image 7a- c; face masks 7a-c (Festive face masks). Fig 1 A Perkin Elmer® FTIR Frontier (UK) was used for chemical characterization of solid materials and a Perkin Elmer® ICP-MS NexIon 2000 for metal leachate identification and quantification. An Agilent® (UK) 1100 and  Dionex ultimate® (UK) 3000  was used for LCMS analysis of organic contaminates, which were tentatively identified using a Thermo LTQ Orbitrap XL  accurate Mass spectrometer. For chromatographic separation a reverse phase XBridge C18 column with dimensions: 3.5 µm x 2.1 mm x 150 mm and Guard column: XBridge C18 3.5 µm x 2.1 mm x 10 mm was used. 2.2 Methodology 2.2.1 Leaching and filtration of particles For particle analysis, 10 face masks for each batch were submerged in 1.5 L deionised water for 4 h and gently agitated by stirring every hour to ensure complete coverage and contact of DPFs with water. Post submersion, the eluent (leachate) was then filtered under vacuum through a 0.1 µm Al2O3 membrane filter. Once complete, the membranes were then transferred to glass petri dishes for drying in a drying cabinet for 2 h at 50 ° C. Procedural blanks were also run with each batch, this involved filtering 1.5 L of deionised water only, through the membranes and then drying for 2 h at 50 ° C. The filtration process was carried out using a glass vacuum filtration funnel and receiving flask with samples drawn through under negative pressure (Sullivan et al., 2020). 2.2.2 Microscopy Light microscopy of the membranes was used to determine coverage of particle contamination, this was done using Zeiss Primotech microscope (Carl Zeiss Ltd., Cambridge, UK) at 5 x, 10 x and 50 x magnification. For scanning electronic microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX) analysis, a Tabletop Microscope TM3000 was utilized (Hitachi High-Technologies Corporation), samples were mounted on carbon tape and placed in vacuums chamber prior. 2.2.3 FTIR characterization of DPFs and membrane Surface characterization was carried out using a Perkin Elmer FTIR Frontier for rapid solid sample analysis. Solid material placed over crystal housing and pressed down using the gage with a force of 30 (arbitrary units). Scan was between 650 and 4000 cm3, with 4 accumulations per second, to get an average spectrum. 2.2.4 ICP-MS elemental analysis Selected DPFs were placed in 250 mL deionised water and left submerged for 24 h typical of leachate analysis in environmental laboratories. The leachate was then subsampled and acidified using 1 mL of 1 M nitric acid and run on Perkin Elmer ICP-MS NexIon 2000 a standard procedure for elemental analysis. A procedural blank (deionised water passed through membrane) and reagent blank (deionised water only) was run with samples to check for background interference. Blanks were run after highest calibration and after every sample to check for carryover. Multielement calibration standards, representing all the analytes in the sample, were made up from PerkinElmer Pure single and multielement standards and diluted into 10% HNO3. Calibration range of 1 µg/L to 100 µg/L was preformed and standard curve required regression statistics above 0.9990 to be deemed acceptable, samples exceeding 100 µg/L were diluted accordingly to bring into dynamic range. Quality control samples prepared from separate batch of multielement standards and prepared at 50% of calibration range, acceptance criteria for accuracy and precision was deemed acceptable below 15%. Instrument parameters as follows: plasma gas flow set at 18 L/minute of argon, auxiliary gas flow set at 1.8 L /minute and nebuliser flow rate of set at 0.98 L/minute. Sample uptake was set at 300 µL/minute with 3 replicates per sample, the RF voltage was  applied at 1600 W, as determined by method optimization during method development (Hineman and Purcell-joiner, 2010). 2.2.5 LC-UV and LC-MS accurate mass of leachate A subsample of the leachate described in 2.1.4 was analysed for polar organic compounds by direct injection (5 µL) on LC-UV for initial sample contamination screening and LC-MS for compound identification. LC-UV was run using Agilent 1100 and a dionex ultimate 3000 for LCMS. A Waters XBridge C18 column: 3.5 µm x 2.1 mm x 150 mm and Guard column: XBridge C18 3.5 µm x 2.1 mm x 10 mm was used for analyte separation for both systems and used a Flow rate 200 µl/min for LC/UV; 150 µl/min for LC/MS. The composition of mobile phase A was 0.1% Formic acid; 2% (Acetonitrile) MeCN in (Water) H2O and mobile phase B 0.1% Formic acid in MeCN for both instrument set-ups. The elution gradient started at 2% B and increased to 90% at 32 min before returning to 2% B ending with a total run time of 37 min. For UV–VIS parameters the lamp was set at 254, 214 and 360 nm with bandwidth of 4, 16 and 32 nm respectively, compared to a reference cell set at 360 nm with bandwidth of 100 nm. The DAD spectrum operated at bandwidth of 190–700 with a refenced at bandwidth of 2 nm. For analysis: reagent blanks (LC grade water) and procedural blanks (deionised water used in sample preparation) were run before and after sample runs to check the background ions and potential carryover. A coumarin standard was also run as pre and post sample analysis to verify the system suitability i.e. retention time drift, analyte response and peak profile for coumarin. For MS parameters, the dionex ultimate 3000 was connected in series to a Thermo LTQ Orbitrap XL with API ion spray source;  sheath gas flow was set to 15 L/minute, auxiliary gas flow set at 2 L per minute, the probe was voltage was set at 4.3 kV with capillary and tube lens voltage of 43 and 150 V respectively. MS scan conditions was full mass profile mode m/z 200–1000 with resolution of 60,000. LCMS was tuned and calibrated using infusion at low flow rate of tune mix, containing a mixture of caffeine, MRFA and ultra Mark (Cal mix) prior to each analysis. A coumarin standard was also used to optimize parameters of LC flow rate. Reagent and procedural blanks were also run prior and post sample analysis to establish background of the system. 3 Results and discussion 3.1 Face mask and filtrate characterization: microscopy and FTIR analysis 9 separate batches of DPFs (from 7 brands) (Fig. 1) were tested for their leaching potential of dispersive pollutants in water. The membranes from the filtration process were then subjected to Microscopy (SEM-EDX and light microscopy) and FTIR, for characterization. Light microscopy was used for an initial indication of particle contamination (Fig. 2 ); all 9 batches of DPFs (Face mask 1 to 7c) emitted fibres believed to be PP and angular fragments that seemed to have crystalline appearance, suspected to be of siliceous composition. It appears that Novelty children face mask 4 and Novelty festive face masks 7a-c have significant fiber and particle contaminations. Light microscopy images of 7 a-c show many different coloured fibres, suggesting that some are stained with a dying agent and looks as though the membrane filters are also slightly tarnished. This is true for face mask 2, although in less abundance than 4, 7 a-c, the fibres are predominantly black (in line with appearance of original face mask). Face masks 1, 3, 5 and 6 appear to generate lower number of fibres, in comparison with 4, 7a-c, and generate a mixture of clear and blue fibres (similar appearance of the original face masks).Fig. 2 Light microscope images of the membrane filters post filtration of face masks1–7c taken and blank membrane at 5 x magnification. Notably face mask 4 (paw patrol) and 7a-c (festive novelty) have what appears significant fibres and particle contamination. Fig 2 FTIR analysis of the physical face masks confirms the manufacture of these masks are primarily out of PP, with some additional functionalisation seen of in the novelty face masks (4, 7 a b c), particularly on coloured sides. The white side presented spectrum indicative of PP; absorbance peaks 2950 cm−1 2917 cm−1are asymmetrical stretching of CH3 and CH2 respectively 2838 cm−1, stretching CH3, 1457 cm−1 and 1375 cm−1 are symmetrical stretching of CH3. The coloured side showed primarily an absorption band at 1712 cm−1, indicative of carboxylic acid or ketone functional (see Table 2 ). This could possibly arise from the side group of an acidic dye compound, polyamide used in synthesis of nylon −66 and sometimes found in oxidised PP (Charles et al., 2009; Coates, 2006; Jung et al., 2018). For plain face masks both blue and white side possessed peaks indicative of PP.Table 2 A table comparing the main FTIR absorbance peaks of the original physical DPF, with the filtered membrane supports. Peaks in red and blue font are peaks derived elsewhere other than PP. Table 2Image, table 2 For the membrane supports post filtration, FTIR was performed on all 9 batches; absorbance spectrum in line with main peaks of PP were identified on membrane supports from face mask 2, 3 and 7 a b c,  further supporting that membrane fibres derived from masks are PP in composition. For some face masks, less fibrous material was likely transferred to the support (i.e. face masks 1, 5 and 6 were plain typical medical type face masks as the detection of PP peaks where below the capabilities of the ATR-FTIR). Additional peaks were identified in novelty face masks coloured side (4, 7 a b c): peaks 1711 cm−1(carboxylic acid stretching), 1242 cm−1 (oxirane or sulfonic acid group) and 1095 cm−1(secondary alcohol stretch) (Bartošová et al., 2017; Coates, 2006). They were also often seen as the primary peaks on the membranes post filtration, and are often common groups in dyes such as eriochrome black and congo red (Bartošová et al., 2017). These membranes were often tarnished, suggesting that these peaks are more than likely derived from the inks or dyes used in the printing of the novelty graphics. More in-depth analysis of the membrane filters was preformed using SEM-EDX. It was confirmed that particle sized trapped on the support can be classified in micro particle (<1 mm) and nano (submicron particle size 0.1–1 µm) range with all face masks emitting significant amount of grain sized particles measured between 360 nm- 500 µm on SEM (Fig. 3 ), but likely to be even finer (limited by resolution of TM3000). Fibrous particles appeared to be in greater size range, often ranging from as little as 25 µm to several millimetres (2.5 mm). Fibres were found to be emitted from all face masks in this study (Masks 1–7c). Further to this EDX analysis using back scattered electrons suggested the elemental composition of particles. It was noted that fibrous particles had high percentage of carbon, most likely derived from polypropylene spun fibres. The majority of the grains contained high percentages of Si and oxygen and likely to be compositions of silica (Fig. 4 ). However, some grains analysed (thought to be silica) were often high in carbon and likely even finer fragments of plastics (Fig. 4).Fig. 3 SEM images of fibres and particles from face masks 1–7C and blank membrane filter. Elemental compositions of fibres were found to be mainly carbon, whilst the small angular fragments found in all face masks had high percentages of Silica and Oxygen. . Fig 3 Fig. 4 Shows SEMs images 1.S, 2.S and 3.S (face masks 3, 7c and 4) and a corresponding elemental map (1.M, 2.M and 3 M). 1.M coloured fuchsia representing carbon on a fiber and particle, 2.M is coloured green to represent silica on grains and 3.M is Purple indicating the presence of lead found on some of the grains. Fig 4 Additional to this, there was often presence of heavy metals associated with these particles, especially in the novelty face masks such as masks 2, 4, 7a, b, c. Some heavy metals were located on grain particles as seen in masks child's novelty face mask (4) shown in Fig. 5 and the festive face masks (7 a, b, c), and likely additives used in plastic manufacture. Heavy metals associated fibres are common chemical additives added during plastic manufacture such as Pb, Cd, Sb and Cu (Hahladakis et al., 2018).Fig. 5 Is an example of EDX data referring to the composition of a grain particle found in face mask 4 (paw patrol). (A) is the image generated by the SEM at x5000,  (B) shows the EDX data plotted graphically, whilst C is tabulated elemental composition data. Image D is false color map for elemental lead (Purple) and image E (blue) is an elemental map of Si the same grain. Fig 5 3.2 Leachable metals and organic compounds: ICP-MS and LC-MS analysis It was apparent from EDX analysis that heavy metals were associated mainly with coloured novelty face masks. It was therefore decided to analyze the coloured masks (face masks 2, 4, 7a, b, c) for heavy metal and organic compound leaching (suspected organic dyes). For comparative studies, a blank control using deionised water was run (same used in the leaching procedure), along with two plain standard face masks type I and II (mask 5, 6) to see if there was significance from novelty and plain face masks. In total, 8 DPFs were selected and placed in 250 mL deionised water and left submerged for 24 h. A procedural blank (deionised water taken through procedure) and reagent blank (deionised only) was also sampled to ascertain the quality of the background interference derived from the deioniser and glassware.  A subsample of the leachate was analysed on ICP-MS and LC-MS using parameters mentioned in section 2.1.4. For ICP-MS analysis of samples, a full external calibration was preformed to determine analyte concentration, and reagent and procedural blanks were used prior and post analysis of samples to assess potential carry over. All QC determinants for  Cd, Co, Cu, Pb, Sb, and Ti passed acceptance criteria and blank samples possessing concentration values below the analytical detection limits (See supportive evidence). Table 3 shows the leachable heavy metals from the DPFs determined by ICP-MS analysis of the sub sampled water, with concerning levels of Sb released from festive novelty face masks (7a,b,c) and ranging from 111–393 µg/L. Additional to this, all face masks appeared to release Cu with levels ranging from 0.85 µg/L (Mask 5) to highest levels of 4.17 µg/L (face mask 2). Copper is a known environmental pollutant which can induce toxic effects in a number of organisms including humans (Keller et al., 2017; Rehman et al., 2019). Lastly, leachable Pb was present in samples 2–7a, interestingly the highest value of 6.79 µg/L was associated with face mask 6, the plain face mask purchased from Tesco © (see Table 3 for more details). Lead is a serious cause of concern and has known carcinogenic and toxicological effect on organisms and has potential to bio-accumulative, even low exposures to lead can have adverse side effects to humans, such as neurological damage and detrimental to foetal development (Freedman et al., 1990; Gundacker and Hengstschläger, 2012; Tchounwou et al., 2012). Looking at the wider picture, the vast amount of DPFs generated due to the COVID-19 pandemic could easily see the cumulative release of these elements breaching current guideline limits, in addition to this, there is a potential worry for the face mask wearers, who are potentially being exposed to heavy metals with no time dependant exposure data available.Table 3 Lists some of the main heavy metals discovered in the DPF leachate (250 mL). Face masks 7a, b, c appears to have the highest release of Sb, whilst Cu is released from all masks. Table 3Sample Cd (µg/L) Co (µg/L) Cu (µg/L) Pb (µg/L) Sb (µg/L) Ti (µg/L) Procedural Blank N.D* N.D N.D N.D N.D N.D Face mask 2 (Leachate) N.D N.D 4.17 0.01 1.06 0.64 Face mask 4 (Leachate) 0.01 0.54 1.87 0.62 N.D 0.27 Face mask 4 (Leachate) repeat 0.04 0.59 1.22 0.89 N.D N.D Face mask 5 (Leachate) N.D N.D 0.85 0.75 3.07 N.D Face mask 6 (Leachate) 1.92 N.D 1.80 6.79 N.D N.D Face mask 7 a (Leachate) 0.53 N.D 2.06 1.62 111 N.D Face mask  7 b (Leachate) N.D N.D 2.31 N.D 393 0.12 Face mask  7 c (Leachate) N.D N.D 4.00 N.D 147 0.06 * Not Detected. In addition to heavy metal release, of the face masks analysed by LCMS (accurate mass), all emitted polar organic species of which, some tentatively identified as polyamide-66 ([M + H]+ mz= 227.1754. C12H23O2N2), polyamide-6 and various oligomers of polyamide (PA), typically associated with nylon (Fig. 6 ). Nylon is often used in the elasticated parts of the DPFs and may be used interwoven in various layers of the face covering. PA-66 and PA-6 monomers and oligomers were identified ([M + H]+ mz=340.2590 and 453.3421) (Tran and Doucette, 2006), in novelty face masks 4, 6, 7 a, b, c and in plain face mask 6 purchased from Tesco©. Medical type II mask (face mask 5) and black color face mask (6) showed little evidence of PA emission and is therefore likely to contain little, if any, nylon parts used in their fabrication. Of the compounds identified in the leachates, none of these species were identified in the procedural blank. The procedural blank was injected between samples to assess the potential background and carryover. The results of the blanks confirmed that there was no carryover between samples, and the background for each run was clear of the ions present in the samples. In some samples polyethylene glycol-like (PEG) derivatives were tentatively identified; in black color face mask (2) plain face masks (5, 6) and novelty face masks (7 a, b, c). PEG was not found in novelty children face mask 4 and in any of the procedural blank samples. PEG (C2nH4n+2On+1) is typically represented by a homologous series with a repeating mass difference of mz 22 and 44. PEG are typical contaminates associated with membrane filters and is often seen in LCMS as Multiple species on the chromatogram. However, due to limited PEG speciation in the leachate samples, and PEG was not present in all samples (Novelty face masks 4 and blank sample) it is therefore likely that PEG originated from DPFs. Fig. 6 Shows LC-MS data obtained from Leachate sample from face mask 4. Top A is the total ion chromatogram (TIC) and B-F are mass spectrums for the associated peaks of the TIC. Peak at Rt 7.73 has a mz of 249.1572 (B) was tentatively identified as Caprolactam. Mass spectrums C, D, E and F are therefore, likely to be oligomers of Caprolactam. More information regarding peak identity is found on table 4. Fig 6 Table 4 A table of LC-MS ions and there tentatively identified compound, preformed on the DPF leachates. Elemental formula takes into account of [M + H]+ or [M+Na]+ion formation. Table 4Image, table 4 Other molecules that were tentatively identified, but difficult to confirm was aromatic amines compounds (azo like), which were seen only at low level in a few samples (6, 7a). The peak corresponds to Rt 20.16 (supporting S3) also showed some absorption band on LC-UV (diode array in the HPLC stack of the LC-MS), further evidence that this is of a dye type compound. N-Undecyl-1-undecanamine was also a likely candidate found in face mask 5, which is a primary long chain surfactant type molecule, likely used in softening of PP during manufacture. MS/MS fragmentation data confirms N-Undecyl-1-undecanamine (mz = 326.3775) as a likely candidate, as it shows a loss of the decyl group a through probable homolytic cleavage, leaving a charged residing around a methyl-1-undecanamine fragment ion (mz = 186.2214). No other dye compounds were identified in other samples and it is likely that if any dye compounds were present in the leachate, then they were at significantly low levels and methodology would need further optimization in order to identify them, such as preconcentration and instrument parameter optimization. Further exploratory tests are required for more accurate identification, of components emitted from DPFs such as GCMS for volatiles analysis, LCMS/MS and NMR for structural elucidation of unknowns, but it is striking from preliminary data that these face masks are emitting organic compounds that may have adverse environmental fate and possibly have bioaccumulation properties (Hahladakis et al., 2018). 4 Conclusion There is a concerning amount of evidence that suggests that DPFs waste can potentially have a substantial environmental impact by releasing pollutants simply by exposing them to water. DPFs release small physical pollutants such as micro and nano size particles; mainly consistent with plastic fibres and silicate grains, which are well documented to have adverse effects on the environment and public health. In addition to the physical particles, harmful chemicals such as heavy metals (Pb, Cd and Sb), and organic pollutants are also readily released from. the DPFs when submerged in water. Many of these toxic pollutants have bio-accumulative properties when released into the environment and this research shows that DPFs could be one of the main sources of these environmental contaminants during and after the Covid-19 pandemic. It is, therefore, imperative that stricter regulations need to be enforced during manufacturing and disposal/recycling of DPFs to minimize the environmental impact of DPFs. Secondary to environmental concerns, there is a need to understand the impact of such particle leaching on public health, as all DPFs released micro/nano particles and heavy metals to the water during our investigation. One of the main concerns with these particles is that they were easily detached from face masks and leached into the water with no agitation, which suggests that these particles are mechanically unstable and readily available to be detached. Therefore, a full investigation is necessary to determine the quantities and potential impacts of these particles leaching into the environment, and the levels being inhaled by users during normal breathing.  This is a significant concern, especially for health care professionals, key workers, and children who are mandated to wear masks for large proportions of the working or school day (6–12 h). Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix Supplementary materials Image, application 1 Acknowledgements This work has been supported by Dr Gareth Davies and Ms. Kate Johns of Tata Steel Europe group for ICP-MS analysis and Dr. Ann Hunter for LCMS analysis at National Mass Spectrometry Facility. We would like to acknowledge the grant supports from 10.13039/501100000266 EPSRC (EP/R51312X/1; EP/N020863/1) and Swansea University Collage of Engineering. We would also like to thank to the Welsh Government Technical Advisory Group for their technical support and valuable feedback. Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.watres.2021.117033. ==== Refs References Adyel T.M. Accumulation of plastic waste during COVID-19 Science 369 2020 1314 1315 10.1126/SCIENCE.ABD9925 (80-. ). 32913095 Aragaw T.A. Surgical face masks as a potential source for microplastic pollution in the COVID-19 scenario Mar. Pollut. Bull. 2020 159 Bartošová, A., Blinová, L., Sirotiak, M., Michalíková, A., 2017. Usage of FTIR-ATR as non-destructive analysis of selected toxic dyes 25, 103–111. Bhagat J. Zang L. Nishimura N. Shimada Y. Zebrafish: an emerging model to study microplastic and nanoplastic toxicity Sci. Total Environ. 728 2020 138707 10.1016/j.scitotenv.2020.138707 Bianco A. Passananti M. Atmospheric micro and nanoplastics: an enormous microscopic problem Sustain 12 2020 10.3390/SU12187327 Bouwmeester H. Hollman P.C.H. Peters R.J.B. Potential health impact of environmentally released micro- and nanoplastics in the human food production chain: experiences from nanotoxicology Environ. Sci. Technol. 49 2015 8932 8947 10.1021/acs.est.5b01090 26130306 Charles J. Ramkumaar G.R. Azhagiri S. Gunasekaran S. FTIR and thermal studies on nylon-66 and 30% glass fibre reinforced nylon-66 E-J. Chem 6 2009 23 33 10.1155/2009/909017 Coates J. Interpretation of infrared spectra, a practical approach Encycl. Anal. Chem 2006 1 23 10.1002/9780470027318.a5606 Fadare O.O. Okoffo E.D. Covid-19 face masks: a potential source of microplastic fibers in the environment Sci. Total Environ. 2020 737 Freedman R. Olson L. Hoffer B.J. Toxic effects of lead on neuronal development and function Environ. Health Perspect. 89 1990 27 33 10.1289/ehp.908927 2088752 Fruijtier-Pölloth C. The toxicological mode of action and the safety of synthetic amorphous silica-a nanostructured material Toxicology 294 2012 61 79 10.1016/j.tox.2012.02.001 22349641 Gundacker C. Hengstschläger M. The role of the placenta in fetal exposure to heavy metals Wiener Medizinische Wochenschrift 162 2012 201 206 10.1007/s10354-012-0074-3 22717874 Hahladakis J.N. Velis C.A. Weber R. Iacovidou E. Purnell P. An overview of chemical additives present in plastics: migration, release, fate and environmental impact during their use, disposal and recycling J. Hazard. Mater. 344 2018 179 199 10.1016/j.jhazmat.2017.10.014 29035713 Hineman A. Purcell-joiner R. Digestion , testing , and validation of heavy metals in cannabis Perkin Elmer Appl. note 2010 1 5 Jung M.R. Horgen F.D. Orski S.V. Rodriguez C., V. Beers K.L. Balazs G.H. Jones T.T. Work T.M. Brignac K.C. Royer S.J. Hyrenbach K.D. Jensen B.A. Lynch J.M. Validation of ATR FT-IR to identify polymers of plastic marine debris, including those ingested by marine organisms Mar. Pollut. Bull. 127 2018 704 716 10.1016/j.marpolbul.2017.12.061 29475714 Jung S. Lee S. Dou X. Kwon E.E. Valorization of disposable COVID-19 mask through the thermo-chemical process Chem. Eng. J. 2021 126658 Keller A.A. Adeleye A.S. Conway J.R. Garner K.L. Zhao L. Cherr G.N. Hong J. Gardea-Torresdey J.L. Godwin H.A. Hanna S. Ji Z. Kaweeteerawat C. Lin S. Lenihan H.S. Miller R.J. Nel A.E. Peralta-Videa J.R. Walker S.L. Taylor A.A. Torres-Duarte C. Zink J.I. Zuverza-Mena N. Comparative environmental fate and toxicity of copper nanomaterials NanoImpact 7 2017 28 40 10.1016/j.impact.2017.05.003 Kögel T. Bjorøy Ø. Toto B. Bienfait A.M. Sanden M. Micro- and nanoplastic toxicity on aquatic life: determining factors Sci. Total Environ. 709 2020 136050 10.1016/j.scitotenv.2019.136050 Lellis B. Fávaro-Polonio C.Z. Pamphile J.A. Polonio J.C. Effects of textile dyes on health and the environment and bioremediation potential of living organisms Biotechnol. Res. Innov. 3 2019 275 290 10.1016/j.biori.2019.09.001 Lin W. Huang Y.wern Zhou X.D. Ma Y. In vitro toxicity of silica nanoparticles in human lung cancer cells Toxicol. Appl. Pharmacol. 217 2006 252 259 10.1016/j.taap.2006.10.004 17112558 Lusher A.L. Welden N.A. Sobral P. Cole M. Sampling, isolating and identifying microplastics ingested by fish and invertebrates Anal. Methods 9 2017 1346 1360 10.1039/c6ay02415g Masuki H. Isobe K. Kawabata H. Tsujino T. Yamaguchi S. Watanabe T. Sato A. Aizawa H. Mourão C.F. Kawase T. Acute cytotoxic effects of silica microparticles used for coating of plastic blood-collection tubes on human periosteal cells Odontology 108 2020 545 552 10.1007/s10266-020-00486-z 31997225 Murugadoss S. Lison D. Godderis L. Van Den Brule S. Mast J. Brassinne F. Sebaihi N. Hoet P.H. Toxicology of silica nanoparticles: an update Arch. Toxicol. 91 2017 2967 3010 10.1007/s00204-017-1993-y 28573455 Paul M.B. Stock V. Cara-Carmona J. Lisicki E. Shopova S. Fessard V. Braeuning A. Sieg H. Böhmert L. Micro- And nanoplastics-current state of knowledge with the focus on oral uptake and toxicity Nanoscale Adv 2 2020 4350 4367 10.1039/d0na00539h 36132901 Prüst M. Meijer J. Westerink R.H.S. The plastic brain: neurotoxicity of micro- and nanoplastics Part. Fibre Toxicol. 17 2020 1 16 10.1186/s12989-020-00358-y 31900181 Rehman M. Liu L. Wang Q. Saleem M.H. Bashir S. Ullah S. Peng D. Copper environmental toxicology, recent advances, and future outlook: a review Environ. Sci. Pollut. Res. 26 2019 18003 18016 10.1007/s11356-019-05073-6 Song Y.K. Hong S.H. Jang M. Han G.M. Rani M. Lee J. Shim W.J. A comparison of microscopic and spectroscopic identification methods for analysis of microplastics in environmental samples Mar. Pollut. Bull. 93 2015 202 209 10.1016/j.marpolbul.2015.01.015 25682567 Sullivan G.L. Gallardo J.D. Jones E.W. Hollliman P.J. Watson T.M. Sarp S. Detection of trace sub-micron (nano) plastics in water samples using pyrolysis-gas chromatography time of flight mass spectrometry (PY-GCToF) Chemosphere 249 2020 126179 10.1016/j.chemosphere.2020.126179 Sungur Ş. Gülmez F. Determination of metal contents of various fibers used in textile industry by MP-AES J. Spectrosc. 2015 10.1155/2015/640271 2015 Tchounwou P.B. Yedjou C.G. Patlolla A.K. Sutton D.J. Molecular, Clinical and Environmental Toxicology, Molecular, Clinical and Environmental Toxicology, Experientia Supplementum 2012 Springer Basel Basel 10.1007/978-3-7643-8340-4 Toussaint B. Raffael B. Angers-Loustau A. Gilliland D. Kestens V. Petrillo M. Rio-Echevarria I.M. Van den Eede G. Review of micro- and nanoplastic contamination in the food chain Food Addit. Contam. - Part A Chem. Anal. Control. Expo. Risk Assess. 36 2019 639 673 10.1080/19440049.2019.1583381 30985273 Tran J.C. Doucette A.A. Cyclic Polyamide oligomers extracted from nylon 66 membrane filter disks as a source of contamination in liquid chromatography/mass spectrometry J. Am. Soc. Mass Spectrom. 17 2006 652 656 10.1016/j.jasms.2006.01.008 16517177 Wilson N. Corbett S. Tovey E. Airborne transmission of covid-19 BMJ 370 2020 10 11 10.1136/bmj.m3206 Wu C.L. Zhang M.Q. Rong M.Z. Friedrich K. Silica nanoparticles filled polypropylene: effects of particle surface treatment, matrix ductility and particle species on mechanical performance of the composites Compos. Sci. Technol. 65 2005 635 645 10.1016/j.compscitech.2004.09.004 You R. Ho Y.S. Hung C.H.L. Liu Y. Huang C.X. Chan H.N. Ho S.L. Lui S.Y. Li H.W. Chang R.C.C. Silica nanoparticles induce neurodegeneration-like changes in behavior, neuropathology, and affect synapse through MAPK activation Part. Fibre Toxicol. 15 2018 1 18 10.1186/s12989-018-0263-3 29298690

33751973

PMC9752785

NO-CC CODE

2022-12-16 23:25:24

no

Water Res. 2021 May 15; 196:117033

utf-8

Water Res

10.1016/j.watres.2021.117033

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)00923-5 10.1016/S0140-6736(21)00923-5 Correspondence Working upstream Owen Mary a Westerhaus Michael bf Finnegan Amy d Surapaneni Laalitha c LaDuke Winona e a Center of American Indian and Minority Health, University of Minnesota, Duluth, MN, USA b Global Medicine, University of Minnesota, Twin Cities, MN, USA c Division of General Internal Medicine, University of Minnesota, Twin Cities, MN, USA d University of St Thomas, St Paul, MN, USA e Honor the Earth, MN, USA f Center for International Health, St Paul, MN 55104, USA 13 5 2021 15-21 May 2021 13 5 2021 397 10287 18031804 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcHealth-care workers often conceptualise addressing the social and structural determinants of health as working upstream.1 In response to the racial disparities of the COVID-19 pandemic and the Movement for Black Lives, health systems are acknowledging systemic racism, promoting implicit bias training, and screening for the social determinants of health. Although welcome, these changes will not achieve the social transformation necessary to eliminate health inequities. We must move even further upstream. Water Protectors, working upstream on the Mississippi River in northern Minnesota, USA, provide a model of what this work entails. They are a mix of Indigenous and environmental activists who are resisting construction of the Line 3 tar sands pipeline. Line 3 will be able to move up to 915 000 barrels of tar sands oil per day across hundreds of water bodies and wild rice beds, a nutritious grain integral to the Ojibwe people (also known as the Anishinaabe people) that does not grow anywhere else in the world. The pipeline will traverse sovereign treaty territory where Ojibwe people maintain the rights to hunt, fish, gather, and practise cultural traditions. Chronic disease, carcinogenic pollutants, and climate change are the possible downstream consequences of Line 3 on the health of humans, land, and water. 2, 3 In response, Water Protectors are acting to directly obstruct the flow of capital and challenge endless resource extraction. They lobby congress, protest in the streets, testify in courtrooms, implore divestment from banks funding pipelines, create camps on the basis of mutual aid, and physically impede construction. Health-care workers and health systems seeking to erase health inequities can learn a great deal from the Water Protectors. First, addressing the social determinants of health requires dismantling the upstream systems of power that structure society, such as racial capitalism and settler colonialism.4 Second, working upstream requires a collective, longitudinal pursuit of justice. The movement to resist Line 3 has been organising for 13 years, building restorative communities and germinating relationships of trust among multisectoral coalitions. Third, human and natural ecosystem health is inter-related and we must prioritise addressing climate change as essential health work. Fourth, just as some Water Protectors risk arrest or danger to their bodies, so too must health-care workers risk confrontation with power holders in our health-care and political systems. In our efforts to deeply engage with the social determinants of health, Water Protectors, who prevent the destruction of life and assert the sovereignty of Indigenous people, are an exemplar of truly upstream health and healing. © 2021 Michael Westerhaus 2021 WL is the executive director of Honor the Earth. All other authors declare no competing interests. ==== Refs References 1 McKinlay J The case for refocusing upstream: the political economy of illness Conrad P The sociology of health and illness: critical perspectives 7th edn. 2005 Worth Publishers New York, NY 551 564 2 Schulz LO Bennett PH Ravussin E Effects of traditional and western environments on prevalence of type 2 diabetes in Pima Indians in Mexico and the U.S. Diabetes Care 29 2006 1866 1871 16873794 3 King M Smith A Gracey M Indigenous health part 2: the underlying causes of the health gap Lancet 374 2009 76 85 19577696 4 Link BG Phelan J Social conditions as fundamental causes of disease J Health Soc Behav 35 1995 80 94

33992136

PMC9752786

NO-CC CODE

2022-12-16 23:25:24

no

Lancet. 2021 May 13 15-21 May; 397(10287):1803-1804

utf-8

Lancet

10.1016/S0140-6736(21)00923-5

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)01293-9 10.1016/S0140-6736(21)01293-9 Editorial We need a global conversation on the 2020 Olympic Games The Lancet 10 6 2021 12-18 June 2021 10 6 2021 397 10291 22252225 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcWith 6 weeks until the Tokyo 2020 Olympic Games begin, concerns over the safety of the Games amid the COVID-19 pandemic are intensifying. Public health experts have expressed strong reservations about how well the risks are being mitigated in articles and before parliamentary committees. The Olympic and Paralympic Games are now deeply unpopular in Japan: lengthy petitions have been signed, 10 000 volunteers have resigned, and several opinion polls have shown that most respondents thought the Games should be postponed or cancelled. The International Olympic Committee (IOC) has the ultimate power to decide on the Games but has huge economic and reputational incentives to continue, as does the Japanese Government. Both insist that the Games will go ahead safely. Writing on June 4, Kaori Yamaguchi, executive member of the Japanese Olympic Committee, suggested that the decision is a fait accompli. “We have been cornered into a situation where we cannot even stop now. We are damned if we do, and damned if we do not.” Is Yamaguchi right? Are we sliding into a dangerous Games? And shouldn't all who have a stake also have a voice? The Games are a global event, happening amid a global health crisis. Control and prevention of COVID-19, including vaccination, are highly variable worldwide. Although international spectators have been barred, 15 000 athletes from more than 200 countries will travel to Tokyo, as well as nearly 80 000 officials, journalists, and support staff. Their vaccination is not mandatory and mixing could risk avoidable transmission of SARS-CoV-2, including emerging viral variants, seeding fresh outbreaks when attendees return home. The Games might also adversely affect the COVID-19 situation within Japan, where case numbers are falling, but several regions remain under a state of emergency. Japan's vaccination roll-out has been sluggish, with 3·4% of the population immunised. As The Lancet goes to press, no final decision has been made on whether domestic fans can attend, but doctors are concerned about the pressure on health services. The Tokyo Medical Practitioners Association, representing around 6000 primary care doctors, has written to Prime Minister Yoshihide Suga calling for the Games to be halted, saying that Tokyo's hospitals “have their hands full and have almost no spare capacity”. The IOC has taken extraordinary measures to minimise the risk of transmission. The number of attendees has been halved. Participants have been given a stringent 60-page rule book for COVID-19, developed with WHO and public health experts: non-compliance could see expulsion from the Games. Two negative COVID-19 tests are required before departure for Japan as well as daily testing during the Games. Strict non-pharmaceutical interventions will be in place, and most events will take place outdoors. All possible places attendees might visit must be approved in advance, and use of public transport is barred. Historically, the overall number of travellers to a host country changes little during the Games. Narita airport in Tokyo is receiving close to 100 000 international passengers per month. The IOC might well also point to events such as the European Football Championship. Matches are being played, with spectators, in the UK, which had 30 724 infections in the week of May 31 and rising, compared with Japan, which had fewer than 19 000. Finally, the Olympic and Paralympic Games could encourage physical activity, provide a worldwide morale boost, and promote unity, showing how the global community can come together, after 18 traumatic and fractious months. But has the global community come together? All nations have an interest in the COVID-19 pandemic and the safety of the Games, yet discussions have largely rested with the IOC and the Japanese Government. Even Tokyo could not unilaterally cancel the Games without risking a multibillion dollar fine for breach of contract. The IOC, whose risk assessments are not publicly available, is the only one with the mandate to halt the Games. But the 2020 Games are not solely a sporting issue. Global health organisations have been largely silent on whether the Games should proceed. WHO refuses to be drawn on whether they should go ahead. The ECDC has told The Lancet it has not specifically performed or even discussed a risk evaluation for the Olympics. In 2016, amid Zika, US CDC director Tom Frieden, declared there was no public health reason to cancel or delay the Rio Games. The CDC has not responded to several requests from The Lancet to clarify its stance on Tokyo 2020. This silence is a deflection of responsibility. The risks of the Games, and how they are being managed, need wide scrutiny and approval. There needs to be a global conversation about the Games, and it needs to happen now. For more on reservations about the Games see BMJ 2021; 373: n962 For more on reservations given before parliamentary committees see https://www.japantimes.co.jp/news/2021/06/03/national/omi-olympics-comments/ For more on opinion polls see https://www.theguardian.com/sport/2021/may/10/tokyo-olympics-poll-shows-60-of-japanese-people-want-games-cancelled, and http://www.asahi.com/ajw/articles/14351670 For Yamaguchi's comments see https://english.kyodonews.net/news/2021/06/6c03987c339c-opinion-tokyo-olympics-have-no-meaning-if-dialogue-is-abandoned.html For more on the risks of transmission at the Games see N Engl J Med 2021; published online May 25. http://dx.doi.org/10.1056/NEJMp2108567 For more on the Tokyo Medical Practitioners Association see https://www.reuters.com/lifestyle/sports/tokyo-doctors-call-cancellation-olympic-games-due-covid-19-2021-05-18/ For the participant's rule book see https://stillmed.olympics.com/media/Document%20Library/IOC/Olympic-Games/Tokyo-2020/Playbooks/The-Playbook-Athletes-and-Officials-April-2021.pdf?_ga=2.78625682.749037116.1622712006-832070016.1622712006 For more on Tokyo's power to cancel the Games see https://www.smh.com.au/world/asia/can-tokyo-cancel-the-olympics-20210520-p57tnn.html © 2021 The Asahi Shimbun/Getty Images 2021

34119048

PMC9752787

NO-CC CODE

2022-12-16 23:25:24

no

Lancet. 2021 Jun 10 12-18 June; 397(10291):2225

utf-8

Lancet

10.1016/S0140-6736(21)01293-9

oa_other

==== Front J Environ Econ Manage J Environ Econ Manage Journal of Environmental Economics and Management 0095-0696 0095-0696 The Authors. Published by Elsevier Inc. S0095-0696(20)30140-6 10.1016/j.jeem.2020.102417 102417 Article Measuring environmental (dis)amenity value during a pandemic: Early evidence from Maryland Irwin Nicholas B. a Livy Mitchell R. b∗ a Department of Economics at the University of Nevada-Las Vegas. Contact Information: 4505 S. Maryland Pkwy, Las Vegas, NV 89154-6001, USA b Department of Economics at California State University, Fullerton. Contact Information: 800 N. State College Blvd. Fullerton, CA 92831-3599, USA ∗ Corresponding author. 11 1 2021 3 2021 11 1 2021 106 102417102417 23 7 2020 19 11 2020 30 12 2020 © 2021 The Authors 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. With the outbreak of COVID-19 and the implementation of stay-at-home (SAH) orders aimed to mitigate its spread, households became less mobile and sheltered in place. This behavior has potential implications for how households’ value environmental (dis)amenities, especially those that are underutilized during the pandemic. In this paper, we explore changes in the valuation of two prominent environmental (dis)amenities – major roadway and open space proximity – by households within the Baltimore metropolitan region. We find evidence that the housing price capitalization of immediate major roadway proximity changes due to the SAH order and associated policies that impact economic activity, suggesting a shift in household perceptions, while there is no evidence of open space valuation changes. These results may have significant implications for homeowner welfare if the altered capitalization of environmental (dis)amenities is temporary due to the SAH order. Keywords Environmental valuation Noise pollution Open space Hedonics COVID-19 ==== Body pmc1 Introduction The outbreak of COVID-19 and measures taken to decrease its spread have led to significant changes in household behaviors. Mobility has severely declined in response to protective policies, with highway travel decreasing at 53 percent across U.S. metropolitan areas (Fishbane and Tomer 2020; Engle et al., 2020). Given these behavioral adjustments and uncertainty surrounding their permanence, perceptions of environmental (dis)amenities may be impacted by COVID-19. We address these perceptions by examining the change in housing price capitalization of major roadway and open space proximity during the period following a COVID-19 stay-at-home (SAH) order. The results indicate that homebuyer perceptions of immediate major roadway features, but not open space, are impacted by the stay-at-home (SAH) order’s disruption to the larger economy and significant welfare effects may manifest if the impacts of the SAH order are transitory. This paper builds upon a broad environmental valuation literature, with a focus on the well-studied topics of major roadway and open space proximity. Roadway proximity is associated with the negative externalities of air and noise pollution, which are shown to negatively affect cardiovascular health, hypertension, lung development, and premature birth (WHO 2011; Gauderman et al., 2007; Currie and Walker 2011). Hedonic research finds a preference to avoid roadway noise (von Graevenitz, 2018). Quiros et al. (2013) examine the temporal dissipation of road-related pollutants from traffic reductions, finding ultrafine particle concentrations declined 80 percent during major highway closure, suggesting commuting changes have an abrupt impact on pollution. Focusing on open space, Acharya and Bennett (2001) find a positive impact on housing prices, but at a decreasing rate relative to open space levels. McConnell and Walls (2005) provide a literature overview noting a heterogeneous, small, and typically positive capitalization of public parks in nearby houses. We advance the literature by examining the impact of COVID-19 and its related mitigation policies on housing price capitalization of environmental (dis)amenities. We utilize housing sales and spatial (dis)amenity data from the Baltimore Metropolitan Statistical Area (MSA) between 2018 and 2020. Using a difference-in-differences approach, we regress housing prices on home attributes, spatial controls, and temporal measures to isolate the impact of COVID-19 and SAH polices on roadway and open space valuation. These were directly impacted by policies seeking to mitigate COVID-19, with partial closures of open spaces and decreased traffic on roadways. The estimates provide initial evidence that the negative capitalization for houses within 200-feet of major roadways is tempered after the implementation of the SAH, while larger distances were unaffected. In contrast, open space has no variation in capitalization stemming from the SAH order. The change in the capitalization of roadway proximity may have substantial impacts on household welfare if homebuyers were unaware of externalities associated with their locational choice. Assuming negative capitalization of these amenities is the correct state of the world, homebuyers created an aggregate property value wedge up to $1.6 million for those near major roadways relative to the estimated property value absent the SAH order. 2 Study area and data We use the Baltimore MSA as our study area, which is oft-utilized in valuation research due to the availability of transactions data from MDProperty View, a state database of property data (see Irwin et al. (2019) for a recent research compendium). The Baltimore MSA consists of six counties1 and the independent city of Baltimore. The MSA has a combined population of 2.7 million, 20th in the United States according to the 2010 Census. The impact of the COVID-19 pandemic in Maryland has been typical compared to the U.S., with over 115,000 cases, nearly 3,800 deaths,2 and a SAH order effective from March 30 through May 15 followed by a phased economic reopening. The reopening, while lifting the SAH order, severely restricted the types of businesses that operated and minimized large groups in a similar manner to the SAH, directly impacting roadway travel with implications for homebuyer open space preferences. Therefore, even though the official SAH order ended on May 15th, its effects continued through the end of our study period and, arguably, will continue until all restrictions on economic activity are lifted.3 We obtain arms-length single-family transactions from MDProperty View for the MSA covering January 2018 through July 2020. We clean the data to remove extreme outliers and observations missing key structural variables.4 In the upper left of Fig. 1 , we show the pattern of transactions by month and year for all of the houses in our sample. Generally, the pattern of sales increases through the spring, with a peak in June, before tapering off. 2020 was typical of previous years in January and February before sizable declines in sales from March through July due to the effects of COVID-19 and the SAH order. However, there were still 4,400 houses sold in the MSA after the SAH implementation. In the upper right and lower left panels of Fig. 1, the sales patterns for houses near major roadways and open space are similar to the overall sales trends. Finally, in the lower right panel of Fig. 1, we show the average (nominal) monthly sales price over our study period, which shows similarities across all years prior to the outbreak of COVID-19 and through the early portions of the SAH period. Table 1 reports the summary statistics for the whole sample and split by SAH. We note the statistics are very similar across groups, indicating the transactions composition was not significantly different during the post-SAH period. We show the spatial dispersion of transactions in Fig. 2 .5 Fig. 1 Housing sales by type and average sale prices. Fig. 1 Table 1 Summary statistics. Table 1 Full Sample (Jan. 2018–July 31, 2020) Pre-SAH (Jan. 2018–April 30, 2020) Post-SAH (April 30, 2020–July 31, 2020) Variable Mean Std. Dev. Min. Max. Mean Std. Dev. Mean Std. Dev. Sale price ($1000) $375.60 $213.81 $10 $2000 $375.27 $215.30 $380.03 $192.87 House size (100’s sqft.) 18.98 8.84 5.04 74.81 18.97 8.86 19.09 8.51 Parcel size (acres) 0.46 0.85 0.00 10.00 0.46 0.84 0.52 0.94 Year built 1978 27 1918 2020 1978 27 1977 26 Age 40.75 26.87 0.00 100 40.60 26.92 42.81 26.00 Stories 1.82 0.56 1.00 4.00 1.82 0.56 1.81 0.51 Basement (0/1) 0.70 0.46 0.00 1.00 0.70 0.46 0.73 0.44 Structural grade (1–9) 3.81 0.84 1.00 9.00 3.81 0.84 3.82 0.82 Vinyl siding (0/1) 0.64 0.48 0.00 1.00 0.64 0.48 0.63 0.48 Brick structure (0/1) 0.15 0.36 0.00 1.00 0.15 0.36 0.16 0.36 Distance to nearest major roadway (miles) 0.50 0.48 0.00 5.35 0.50 0.48 0.50 0.47 Distance to nearest open space (miles) 0.40 0.39 0.00 3.86 0.40 0.39 0.38 0.39 Number of observations 64,510 60,022 4,488 Fig. 2 Residential transactions in Baltimore MSA. Fig. 2 We link the transactions to a suite of additional data to aid our empirical estimation. We utilize the Maryland Tiger/Line shapefiles, which contains all primary (limited-access) and secondary (main arterial) roads and calculate the Euclidean distance from each house centroid to the nearest major roadway. Next, we obtain shapefiles for a variety of open space types from Maryland’s open GIS portal, including data on locally protected lands – parks and community space – in addition to data on state-managed and federally-owned lands which may be valuable open space for proximate homebuyers. We perform a similar distance calculation and report the distances in Table 1. The spatial difference in distance from each house to major roadways and open space provides the variation necessary for identifying temporal differences in the capitalization of these measures resulting from COVID-19 policies in our models. 3 Empirical implementation We implement a hedonic price model (Rosen, 1974) to investigate the impact of COVID-19 and the SAH polices on (dis)amenity capitalization. Following the discussion in Cropper et al. (1988), we implement the following log-linear regression to curtail estimation error:(1) ln(Pijt)=β0+βHXi+βAAi+βSSAHt+βASAi∗SAHt+δj+γt+eit where equation (1) estimates the relationship between housing prices, P, and X which is a vector of house attributes from Table 1, a distance cut-off parameter to the nearest roadway or open space type, A, an indicator variable for a house selling 30 days6 after the SAH order went into effect, SAH, and the interaction of the latter two variables. In equation (1), i, j, and t index individual houses, locations, and time, respectively. δj is a block-group fixed effect controlling for unobserved school quality, public services, and other spatially varying attributes that may bias our estimates. A control for temporal variation, γt, at the quarter-year level is included. Given the log-linear nature of equation (1), we apply Halvorsen and Palmquist’s (1980) indicator variable correction. The variable of interest in our model is the interaction of the amenity in question – roadway proximity or open space – and the SAH order. The interaction measures the difference in the valuation of roadway proximity and open space before and after the SAH order, which provides an estimate valuation change attributable to the SAH policy, i.e. a difference-in-difference. We define proximity in the following manner: for distance to major roadways, we create a dummy variable for houses within 200- and 400-feet of the roadway as this is the approximate distance whereby the decibel levels7 decrease substantially, conditional on noise barriers or geography (Caltrans, 2013). For open space, proximity is in terms of walkability, (400- and 800-feet) meaning houses will receive aesthetic open space value or are within walking distance of accessible space. In Fig. 3 , we provide residual plots of prices for transactions after controlling for time fixed-effects within our nearest distance bands for roadways and open space noting a clear difference in prices pre-SAH that shifts post-SAH in both groups. While this indicates some relationship between the SAH order and prices for proximate houses, our econometric model investigates this further and controls for a complete set of house and neighborhood confounders.Fig. 3 Residual figures for roadway and open space proximity. Fig. 3 An additional concern for identification is the amount of amenity proximity variation within block groups, but we note that for block groups with at least one observation within 400-feet of a major roadway, nearly 14 percent of observations are within this distance. For block groups containing at least one observation within 800-feet of open space less than 40 percent of observations are within this distance, implying sufficient proximity variation for model identification.8 4 Results We report the results of our first estimation – proximity to a major roadway – in Table 2 .9 In Column (1), we find a negative and significant value for houses located within 200-feet of a major roadway pre-SAH on the order of 3.4 percent – roughly $12,850 at mean price – which agrees with previous economic literature and indicates a negative externality associated with proximity to busy roads. Turning to the indicator variable for houses sold post-SAH, we find no evidence of any price effects for houses sold during this period, alleviating concern that these houses were atypical of pre-SAH sales. In the interaction term, we find that is it positive and significant but at a low level of significance (10 percent), indicating that there is a weak positive capitalization effect for the set of houses near major roadways post-SAH.Table 2 Regression results for changing capitalization of proximity to roadways and open space. Table 2 Dependent Variable: Log Sale Price (1) (2) (3) (2) Major road within 200 feet Major road within 400 feet Open space within 400 feet Open space within 800 feet Near roadway/open space −0.035∗∗∗ (0.009) −0.022∗∗∗ (0.006) −0.009∗ (0.005) −0.001 (0.006) Sold after SAH −0.015 (0.013) −0.014 (0.013) −0.013 (0.013) −0.013 (0.013) Near X Sold after SAH 0.043∗ (0.024) 0.002 (0.016) −0.001 (0.011) −0.003 (0.009) Property Characteristics Yes Yes Yes Yes Fixed Effects Block group Block group Block group Block group Temporal Fixed Effects Quarter-year Quarter-year Quarter-year Quarter-year R-Squared 0.76 0.76 0.76 0.76 Observations near roadway 1991 5780 9845 18,999 Observations sold after SAH 4488 4488 4488 4488 Observations in interaction group 128 384 783 1449 Total Observations 64,510 64,510 64,510 64,510 Notes: This table presents the estimation results of equation (1) for major roadway proximity (columns 1 & 2) and open space proximity (3 & 4). Near is defined as within the distance indicated in each column. Property characteristics are suppressed for space but reported in the Appendix. ∗∗∗p < .01, ∗∗p < .05, ∗p < .1. Robust standard errors shown in parenthesis are clustered at the indicated spatial fixed effects level. Although there is weak significance on this proximity variable, the more profound result is that houses sold pre-SAH have the expected negative capitalization, but houses sold during the post-SAH period no longer fully negatively capitalize this environmental externality into prices. We attribute this capitalization change to the fact significantly fewer cars were on the road during the SAH order, which provided false information about typical road noise and pollution levels. In turn, home-sellers that purchased these houses with a price discount from the externality received a higher price than they would have pre-SAH. We note that these results are based on a short temporal window and future research will be needed to unpack the persistence of this effect as the economy re-opens and traffic returns. To investigate the sensitivity of major roadway distance, we increase our proximity to 400-feet, more than doubling the number of affected houses and report the new results in Table 2, Column (2). Pre-SAH, the capitalization of roadway proximity is statistically significant but tempered compared to the 200-feet distance. With this expanded distance parameter, we find no evidence of a change in capitalization for the set of houses sold after the SAH. Our overall results provide clear and compelling evidence that the SAH order substantially tempered capitalization of major roadway proximity for those closest and most likely to be impacted by roadway externalities while having no effect at increased distances. We believe this outcome stems from the SAH’s effect on major roadway traffic due to the closing of non-essential businesses and schools despite the possible temporariness of the change.10 We turn next to estimating if open space valuation has similarly changed in response to the SAH order. In Table 2, Column (3), we report our results for houses that are within 400-feet of any open space and find that houses sold pre-SAH see a very small decrease in house prices of approximately one percent, which we attribute to the downside of being near open space such as park lights, event noise, and/or crowd congestion (McConnell and Walls, 2005; Anderson and West, 2006). However, we note that this estimate is significant only at the 10 percent level, indicating a weak capitalization effect. Turning to houses sold post-SAH, we find no evidence of any capitalization change for open space proximity in the nearest houses, indicating that the SAH impact on open space properties is indistinguishable from zero. Expanding our walkability to 800-feet in Table 2, Column (4), we find no pre-SAH capitalization for this set of houses, indicating that the observed negative effects of close proximity to open space may only manifest for the nearest of houses. We also continue to find a null result for houses sold post-SAH in this distance.11 From a policy perspective, these results have multiple important takeaways. First, homebuyers shopping for a house during the SAH period may respond differently to (dis)amenities than previous homebuyers, and in the case of roadway noise and pollution, counter to the literature. This has considerable aggregate welfare implications – up to $1.6 million near major roadways – due to the absence of compensation via decreased prices, i.e. if there is no offset for any noise or health effects as established in previous research. Second, the stability of the open space estimates demonstrates temporary changes do not impact perceptions for all (dis)amenities. Overall, the SAH order induced a price wedge between houses prices that account for the 200-feet roadway (dis)amenity externality and prices when externalities are temporarily absent, solely due to the SAH order. Together, the estimates suggest that disclosures or other education programs may be necessary if changes in (dis)amenities are expected to temporarily deviate from normal. This may mean that studying housing transactions during the entirety of the COVID-19 pandemic can reveal results that are unexpected and counter-intuitive; thus, the results would need important contextualization from researchers if the changes in the housing market are temporary. 5 Conclusion COVID-19 and its associated policies represent a substantial shift in societal and household behavior. With this research, we investigate the impact of SAH orders on housing price capitalization of major roadways and open space proximity to examine the change in valuation of these (dis)amenities expressed through the housing market. The results indicate that the negative capitalization associated with close (200-feet) roadway proximity was tempered by the SAH order and its associated impacts on economic activity, and the capitalization of larger roadway distances and open space were unaffected. While the negative externalities associated with immediate roadway proximity decreased during the SAH period, it is unclear how persistent these changes will be when the economy fully reopens, workers return to offices, and traffic increases. Policymakers may find it necessary to implement disclosure programs to educate homebuyers on these potential externalities or face constituents who may demand costly remediation activity. While this research has illuminated the immediate impact of COVID-19 on the valuation of spatial (dis)amenities, additional research is needed in tandem with new data. Untangling and better understanding the types of consumers on either end of these transactions and the permanence of the changes in capitalization is needed, as well as examining if similar effects manifest in other spatially provided (dis)amenities. Are these houses sold during the SAH under duress by homeowners who are worried about their economic future? Are homebuyers coming from inside or outside the MSA? Are there knock-on effects from expanded unemployment benefits on future housing markets? While these questions cannot be fully answered immediately, gaining understanding of the churn in the housing market and the extent to which this is causing MSA demographic shifts would be helpful to policymakers trying to respond to unprecedented economic conditions with the potential to create significant long-term damages both locally and nationally. Appendix Appendix Table 1 Complete regression results from roadway and open space proximity estimation (Table 2 in text) Appendix Table 1 Dependent Variable: Log Sale Price (1) (2) (3) (4) Major road w/i 200 feet Major road w/i 400 feet Open space w/i 400 ft. Open space w/i 800 ft. Near −0.035∗∗∗ (0.009) −0.022∗∗∗ (0.006) −0.009∗ (0.005) −0.001 (0.006) Sold after SAH −0.015 (0.013) −0.014 (0.013) −0.013 (0.013) −0.013 (0.013) Near X Sold after SAH 0.043∗ (0.024) 0.002 (0.016) −0.001 (0.011) −0.003 (0.009) Distance to highway/open space (miles) 0.041∗∗∗ (0.011) 0.038∗∗∗ (0.011) 0.034∗∗∗ (0.011) 0.037∗∗ (0.011) House size (100’s sqft) 0.043∗∗∗ (0.001) 0.043∗∗∗ (0.001) 0.043∗∗∗ (0.001) 0.043∗∗∗ (0.001) House size (100’s sqft) x Sold after SAH 0.001 (0.001) 0.001 (0.001) 0.001 (0.001) 0.001 (0.001) Acres 0.084∗∗∗ (0.008) 0.085∗∗∗ (0.008) 0.087∗∗∗ (0.008) 0.087∗∗∗ (0.008) Acres x Sold after SAH 0.000 (0.009) 0.000 (0.009) 0.000 (0.009) 0.000 (0.001) House size squared −0.000∗∗∗ (0.000) −0.000∗∗∗ (0.000) −0.000∗∗∗ (0.000) −0.000∗∗∗ (0.000) Acres squared −0.008∗∗∗ (0.001) −0.008∗∗∗ (0.001) −0.008∗∗∗ (0.001) −0.008∗∗∗ (0.001) Age −0.001 (0.001) −0.001 (0.001) −0.000 (0.001) −0.000 (0.001) Age squared 0.000∗ (0.000) −0.000∗ (0.000) −0.000∗∗ (0.000) −0.000∗∗ (0.000) Stories −0.07∗∗∗ (0.006) −0.07∗∗∗ (0.006) −0.07∗∗∗ (0.006) −0.07∗∗∗ (0.006) Basement 0.021∗∗∗ (0.006) 0.021∗∗∗ (0.006) 0.021∗∗∗ (0.006) 0.021∗∗∗ (0.006) Structural grade 0.108∗∗∗ (0.007) 0.108∗∗∗ (0.007) 0.109∗∗∗ (0.007) 0.109∗∗∗ (0.007) Vinyl siding 0.002 (0.007) 0.002 (0.007) 0.001 (0.007) 0.001 (0.007) Brick structure −0.045∗∗∗ (0.009) −0.045∗∗∗ (0.009) −0.047∗∗∗ (0.009) −0.047∗∗∗ (0.009) Constant 11.71∗∗∗ (0.047) 11.71∗∗∗ (0.047) 11.71∗∗∗ (0.046) 11.71∗∗∗ (0.047) Spatial fixed effects Block group Block group Block group Block group Temporal fixed effects Quarter-year Quarter-year Quarter-year Quarter-year Observations 64,510 64,510 64,510 64,510 R-squared 0.76 0.76 0.76 0.76 Notes: This table presents the complete estimation results of equation (1) for major roadway proximity (columns 1 & 2) and open space proximity (3 & 4). Near is defined as within the distance indicated in each column. ∗∗∗p < .01, ∗∗p < .05, ∗p < .1. Robust standard errors shown in parenthesis are clustered at the indicated spatial fixed effects level. Appendix Table 2 Estimation results from extending open space proximity and varying open space type Appendix Table 2 Dependent Variable: Log Sale Price (1) (2) (3) (4) Open space w/i 1600 ft. Explicit open space w/i 1600 ft. Park w/i 1600 ft. Easement w/i 1600 ft. Walkable to open space −0.004 (0.008) 0.008 (0.012) −0.002 (0.009) −0.008 (0.015) Sold after SAH −0.016 (0.014) −0.014 (0.013) −0.013 (0.013) −0.014 (0.013) Walkable to open space X Sold after SAH 0.003 (0.009) −0.003 (0.012) −0.006 (0.014) 0.000 (0.02) Distance to open space (miles) 0.034∗∗∗ (0.012) 0.039∗∗∗ (0.011) 0.037∗∗∗ (0.011) 0.037∗∗∗ (0.011) House size (sqft) 0.043∗∗∗ −0.001 0.043∗∗∗ (0.001) 0.043∗∗∗ (0.001) 0.043∗∗∗ (0.001) House size (sqft) x Sold after SAH 0.001 −0.001 0.001 (0.001) 0.001 (0.001) 0.001 (0.001) Acres 0.087∗∗∗ −0.008 0.087∗∗∗ (0.008) 0.087∗∗∗ (0.008) 0.087∗∗∗ (0.008) Acres x Sold after SAH 0.000 (0.009) 0.000 (0.009) 0.000 (0.009) 0.000 (0.009) House size squared −0.000∗∗∗ (0.000) −0.000∗∗∗ (0.000) −0.000∗∗∗ (0.000) −0.000∗∗∗ (0.000) Acres squared −0.008∗∗∗ (0.001) −0.008∗∗∗ (0.001) −0.008∗∗∗ (0.001) −0.008∗∗∗ (0.001) Age −0.000 (0.001) −0.000 (0.001) −0.000 (0.000) −0.000 (0.001) Age squared −0.000∗∗ (0.000) −0.000∗∗ (0.000) −0.000∗∗ (0.000) −0.000∗∗ (0.000) Stories −0.067∗∗∗ (0.006) −0.07∗∗∗ (0.006) −0.07∗∗∗ (0.006) −0.07∗∗∗ (0.006) Basement 0.021∗∗∗ (0.006) 0.021∗∗∗ (0.006) 0.021∗∗∗ (0.006) 0.021∗∗∗ (0.006) Structural grade 0.109∗∗∗ (0.007) 0.109∗∗∗ (0.007) 0.109∗∗∗ (0.007) 0.109∗∗∗ (0.007) Vinyl siding 0.001 (0.007) 0.001 (0.007) 0.001 (0.007) 0.001 (0.007) Brick structure −0.047∗∗∗ (0.009) −0.047∗∗∗ (0.009) −0.047∗∗∗ (0.009) −0.047∗∗∗ (0.009) Constant 11.71∗∗∗ (0.05) 11.71∗∗∗ (0.046) 11.71∗∗∗ (0.046) 11.71∗∗∗ (0.046) Spatial fixed effects Block group Block group Block group Block group Temporal fixed effects Quarter-year Quarter-year Quarter-year Quarter-year Observations 64,510 64,510 64,510 64,510 R-squared 0.76 0.76 0.76 0.76 Notes: This table presents the estimation results of extending open space proximity to 1600 feet (column 1) and then uses that proximity distance to explore differential capitalization of selected types of open space that previous literature suggests may capitalize into prices (2–4). ∗∗∗p < .01, ∗∗p < .05, ∗p < .1. Robust standard errors shown in parenthesis are clustered at the indicated spatial fixed effects level. BG is block group and Q-Y is quarter-year. Appendix Table 3 Roadway proximity results with 60-day closing Appendix Table 3 Dependent Variable: Log Sale Price (1) (2) Major road w/i 200 feet Major road w/i 400 feet Near major roadway −0.034∗∗∗ (0.009) −0.022∗∗∗ (0.006) Sold after SAH −0.002 (0.014) −0.001 (0.014) Near major roadway X Sold after SAH 0.033 (0.029) 0.002 (0.019) SAH Closing Restriction 60 days 60 days Property Characteristics Yes Yes Fixed Effects Block group Block group Temporal Fixed Effects Quarter-year Quarter-year R-Squared 0.76 0.76 Observations near roadway 1991 5780 Observations sold after SAH 3065 3065 Observations in interaction group 86 263 Total Observations 64,510 64,510 Notes: These estimation results provide a robustness check for the highway proximity results by modifying the SAH closing restriction from our original assumption of a 30-day closing period to a 60-day closing period for all housing transactions. ∗∗∗p < .01, ∗∗p < .05, ∗p < .1. Robust standard errors shown in parenthesis are clustered at the block group level. Appendix Table 4 Open space proximity results with 60-day closing Appendix Table 4 Dependent Variable: Log Sale Price (1) (2) (3) (4) (5) (6) Open space w/i 400 ft. Open space w/i 800 ft. Open space w/i 1600 ft. Explicit open space w/i 1600 ft. Park w/i 1600 ft. Easement w/i 1600 ft. Walkable to open space −0.009 (0.005) −0.000 (0.006) −0.004 (0.008) 0.008 (0.012) −0.003 (0.009) −0.008 (0.015) Sold after SAH 0.001 (0.014) 0.004 (0.015) −0.004 (0.016) −0.000 (0.014) 0.000 (0.014) −0.000 (0.014) Walkable to open space X Sold after SAH −0.005 (0.013) −0.011 (0.011) 0.005 (0.011) −0.006 (0.013) −0.005 (0.019) −0.004 (0.02) SAH Closing Restriction 60 60 60 60 60 60 Property Characteristics Yes Yes Yes Yes Yes Yes Fixed Effects Block group Block group Block group Block group Block group Block group Temporal Fixed Effects Quarter-year Quarter-year Quarter-year Quarter-year Quarter-year Quarter-year R-Squared 0.76 0.76 0.76 0.76 0.76 0.76 Observations near open space 9845 18,999 33,574 3586 6642 1720 Observations sold during SAH 3065 3065 3065 3065 3065 3065 Observations in interaction group 538 1003 1681 233 233 93 Observations 64,510 64,510 64,510 64,510 64,510 64,510 Notes: These estimation results provide a robustness check for the open space proximity results by modifying the SAH closing restriction from our original assumption of a 30-day closing period to a 60-day closing period for all housing transactions. ∗∗∗p < .01, ∗∗p < .05, ∗p < .1. Robust standard errors shown in parenthesis are clustered at the block group level. 1 Anne Arundel, Baltimore, Carroll, Harford, Howard, and Queen Anne’s. 2 As of September 20, 2020. 3 Results are qualitatively consistent across the sample period when disaggregated between the official SAH order and early stages of reopening, indicating the impact on valuation did not change from reopening. Therefore, we use the SAH label throughout for simplification. 4 One key structural variable missing is bathrooms, a prominent determinant of prices. MDProperty View no longer provides this datapoint as of their recent statewide data standardization, despite doing so in the past. 5 The sales volume for July shows an unexpected decline that runs counter to national evidence of increased summer sales. We believe this is due to MDProperty View data reporting delays from counties and we have used the full data as it is available. 6 We assume a standard 30-day closing window, allowing for the possible effects of COVID-19 on the valuation of these (dis)amenities to be included in homebuyer price formation. Our results are qualitatively similar if we relax this assumption to 60 days as shown in Appendix (Tables 3 and 4). 7 Previous research on roadway noise has utilized decibel level mapping to differentiate noise, which was not available. 8 The inclusion of tract fixed-effects leads to qualitatively similar results. 9 Complete regression tables in appendix. 10 It is possible that valuation changes are linked to household preferences for larger houses or more acres to compensate for SAH. We control for this by incorporating SAH interaction variables for house size and acres within the regressions. The coefficients are not statistically significant, indicating capitalization for these features was unaffected by the SAH order. Results reported in Appendix Table 1. 11 In Appendix Table 2, we further expand our walkability measure to 1600-feet and find similar results, i.e. no change in pre/post-SAH capitalization. We also explore the possibility that certain types of open space may be especially valued in the SAH period but again find no change in capitalization. ==== Refs References Acharya G. Bennett L.L. Valuing open space and land-use patterns in urban watersheds J. R. Estate Finance Econ. 22 2–3 2001 221 237 Anderson S.T. West S.E. Open space, residential property values, and spatial context Reg. Sci. Urban Econ. 36 6 2006 773 789 Currie J. Walker R. Traffic congestion and infant health: evidence from E-Z pass Am. Econ. J. Appl. Econ. 3 1 2011 65 90 Caltrans “Technical Noise Supplement to the Traffic Noise Analysis Protocol”. Division of Environmental Analysis, Environmental Engineering, Hazardous Waste, Air, Noise 2013 Paleontology Office Cropper M.L. Deck L.B. McConnell K. On the choice of functional form for hedonic price functions Rev. Econ. Stat. 70 4 1988 668 675 Engle S. Stromme J. Zhou A. Staying at Home: Mobility Effects of COVID-19 2020 Available at: SSRN: https://ssrn.com/abstract=3565703 Fishbane L. Adie T. Coronavirus Has Shown us a World without Traffic. Can We Sustain it? 2020 The Brookings Institute Gauderman W.J. Effect of exposure to traffic on lung development from 10 to 18 Years of age: a cohort study Lancet 369 9561 2007 571 577 17307103 Halvorsen R. Palmquist W.D. The interpretation of dummy variables in semilogarithmic equations Am. Econ. Rev. 70 3 1980 474 475 Irwin E.G. Grove J.M. Irwin N. Klaiber H.A. Towe C. Troy A. Effects of Disamenities and Amenities on Housing Markets and Locational Choices 2019 Yale University Press 92 110 McConnell V. Walls M. The Value of Open Space: Evidence from Studies of Non- Market Benefits 2005 Resources for the Future ” Washington, DC Quiros David C. Air quality impacts of a scheduled 36-H closure of a major highway Atmos. Environ. 67 2013 404 414 Rosen S. Hedonic prices and implicit markets: product differentiation in pure competition J. Polit. Econ. 82 1 1974 34 55 von Graevenitz K. The amenity cost of road noise J. Environ. Econ. Manag. 90 2018 1 22 WHO Burden of Disease from Environmental Noise. Technical Report 2011 The World Health Organization European Centre for Environment and Health

0

PMC9752788

NO-CC CODE

2022-12-16 23:25:24

no

J Environ Econ Manage. 2021 Mar 11; 106:102417

utf-8

J Environ Econ Manage

10.1016/j.jeem.2020.102417

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)01225-3 10.1016/S0140-6736(21)01225-3 Correspondence The fragility of abortion access in Europe: a public health crisis in the making Miani Céline a Razum Oliver a a School of Public Health, Department of Epidemiology and International Public Health, Bielefeld University, Bielefeld 33615, Germany 5 8 2021 7-13 August 2021 5 8 2021 398 10299 485485 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcPoland is rightly being criticised for suppressing abortion services.1 Since January, 2021, abortion is only legal if the pregnancy is directly life-threatening or the result of rape or incest. However, countries with allegedly more progressive policies have reasons to be self-critical as well. An example is Germany, considered a liberal country in terms of abortion law from an international perspective, since women can be granted an abortion on request for any reason, including socioeconomic reasons. Yet, abortion in Germany is technically a crime (albeit not punished up to 12 weeks from conception), and gynaecologists are losing court cases for stating on their websites that they provide abortion care in a supportive environment.2 Attacks on abortion rights and services are nourished by vocal conservative and religious forces whose agendas find support in a non-negligible share of the population. The number of doctors providing abortion services is declining,3 teaching of abortion techniques in medical schools is marginal,4 and a mandatory consultation before an abortion (in some regions done by religious organisations) and a so-called cooling-off period add barriers to access.5 As a result, some women from Germany (and other European countries) are seeking care in the Netherlands, as highlighted by the Europe Abortion Access Project. Women on Web, a non-governmental organisation, has recently recorded an increased demand for abortion pills in Germany.6 The COVID-19 pandemic has created further access challenges, in the form of reduced opening hours of clinics, fewer social infrastructures, and rise in domestic violence. Contrary to some other countries (eg, the UK and France),7 demands from reproductive health activists to modify medical abortion delivery (eg, through telecare or drug mailing) have remained unheard in Germany. Yet Germany is still seen as a safe haven for Polish women who are living in fear under one of the strictest abortion laws in Europe. Poland and Germany are only two examples of how fragile abortion access remains in Europe (in both constrained and more liberal societies) paving the way for a public health crisis. Denied or reduced access to abortion services has short-term and long-term health consequences, and disproportionately affects the most vulnerable groups in societies. Initiatives led by civil society (eg, Doctors for Choice, Women on Web) and crossborder care alone cannot compensate for the scarcity of governmental impetus, and cannot mitigate the threats to abortion rights coming from growing right-wing and anti-feminist movements in Europe. As a matter of health equity, abortion access needs to be sustainably guaranteed in practice, including beyond the allegedly permissive legislations. CM reports funding from Bielefeld University, Germany, for a postdoctoral researcher position. OR declares no competing interests. The funders had no role in the writing of or decision to submit this Correspondence. ==== Refs References 1 The Lancet Regional Health – Europe Anti-abortion laws—the antithesis of the fundamental rights of women Lancet Regional Health Europe 3 2021 100111 2 Deutsches Ärzteblatt OLG Frankfurt bestätigt §219a-Urteil gegen Hänel, Verfassungsbeschwerde angekündigt https://www.aerzteblatt.de/nachrichten/120309/OLG-Frankfurt-bestaetigt-219a-Urteil-gegen-Haenel-Verfassungsbeschwerde-angekuendigt Jan 19, 2021 3 Zeit Online Weniger Ärzte nehmen Schwangerschaftsabbrüche vor https://www.zeit.de/gesellschaft/zeitgeschehen/2018-08/schwangerschaftsabbrueche-statistisches-bundesamt-arztpraxen-kliniken Aug 28, 2018 4 Gamillscheg M Moderne Medizin https://www.zeit.de/campus/2019/01/medizinstudium-schwangerschaftsabbruch-papaya-werbeverbot-universitaet-bildung Dec 13, 2018 5 Bundeszentrale für gesundheitliche Aufklärung Der Beratungsschein https://www.familienplanung.de/schwangerschaftskonflikt/schwangerschaftsabbruch/der-beratungsschein/ March 20, 2020 6 Killinger K Günther S Gomperts R Atay H Endler M Why women choose abortion through telemedicine outside the formal health sector in Germany: a mixed-methods study BMJ Sex Reprod Health 2020 published online Nov 23. 10.1136/bmjsrh-2020-200789 7 Moreau C Shankar M Glasier A Cameron S Gemzell-Danielsson K Abortion regulation in Europe in the era of COVID-19: a spectrum of policy responses BMJ Sex Reprod Health 2020 published online Oct 22. 10.1136/bmjsrh-2020-200724

34364514

PMC9752818

NO-CC CODE

2022-12-16 23:25:25

no

Lancet. 2021 Aug 5 7-13 August; 398(10299):485

utf-8

Lancet

10.1016/S0140-6736(21)01225-3

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)01651-2 10.1016/S0140-6736(21)01651-2 Comment COVID-19: the turning point for gender equality Fisseha Senait a Sen Gita b Ghebreyesus Tedros Adhanom c Byanyima Winnie d Diniz Debora e Fore Henrietta H f Kanem Natalia g Karlsson Ulrika h Khosla Rajat i Laski Laura j Mired Dina k Mlambo-Ngcuka Phumzile l Mofokeng Tlaleng m Gupta Geeta Rao n Steiner Achim o Remme Michelle p Allotey Pascale p a Susan Thompson Buffet Foundation, Omaha, NE, USA b Public Health Foundation of India, Bangalore, India c World Health Organization, Geneva, Switzerland d Joint United Nations Programme on HIV/AIDS, Geneva, Switzerland e Brown University, Providence, RI, USA f UNICEF, New York, NY, USA g United Nations Population Fund, New York, NY, USA h Inter-Parliamentary Union, Stockholm, Sweden i Amnesty International, London, UK j Partnership for Maternal, Newborn & Child Health, New York, NY, USA k Union for International Cancer Control, Amman, Jordan l UN Women, New York, NY, USA m Office of the United Nations High Commissioner for Human Rights, Johannesburg, South Africa n United Nations Foundation, Washington, DC, USA o United Nations Development Programme, New York, NY, USA p United Nations University International Institute for Global Health, UKM Medical Centre, 56000 Cheras, Kuala Lumpur, Malaysia 20 7 2021 7-13 August 2021 20 7 2021 398 10299 471474 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcThe impacts of the COVID-19 pandemic have gone far beyond the disease itself. In addition to the increasing number of COVID-19 deaths,1 the pandemic has deepened social and economic inequalities.2 These indirect impacts have been compounded by pervasive gender inequalities, with profound consequences, especially for women, girls, and people of diverse gender identities.2 There has been an escalation in gender-based violence within households,3 increasing risk of child marriages and female genital mutilation,4 and an increased burden of unpaid care work,5 with impacts on mental health.6 Communities of people affected by HIV are, again, at the crossroads of injustice and targeted discrimination.7 Measures to control the pandemic have reduced access to essential health and social welfare services, including sexual and reproductive health services, reduced employment and labour force participation, and decimated many household incomes.8, 9, 10 Here again, women have borne the brunt of marginalisation, particularly those working in the informal sector. Intersectionality analyses have highlighted the inextricable effects of poverty, racial discrimination, harmful gender norms, and limited agency and opportunities for women, especially already marginalised women, even when they represent most of the front-line health workers.11 The diversion of funds from other health and development programmes into economic recovery means that the pandemic is further eroding health gains made over decades, stalling progress on tackling gender inequalities.9 There are, however, glimmers of hope. The gendered impacts of power, intersectionality, social, legal, and commercial determinants on health are foregrounded in public forums and can no longer be ignored. The pandemic has catalysed a need for concrete action on gender inequality. There is growing acknowledgment by governments that political leadership is required for key decisions about investments in health to ensure social protection and financial recovery, targeting of disadvantaged populations to ensure equity, and engaging with broader geopolitical challenges that impact on health.12 With commitments by governments to strengthen health systems and the health workforce, and to enhance the quality of care and self-managed care, there are opportunities to learn from previous efforts to address gender inequalities. Although there are still evidence gaps, our institutions, experts, and practitioners have decades of practice-based knowledge of strategies that work to drive gender equality impacts in health, including an understanding of the political and policy levers that are crucial drivers of change.13 For example, the progress made in advancing sexual and reproductive health and rights and tackling gender-based violence has consistently been realised through the strong leadership and engagement of feminist civil society and women's rights movements.14 The global health community is equipped now more than ever before to drive the gender equality agenda forward in pandemic responses and other health areas.15 There is considerable evidence on the technical solutions that promote gender equality,16 and many of these solutions provide transferable lessons. For example, gender-balanced community health worker teams can achieve increased service coverage.17 Group-based education interventions that tackle gendered power dynamics in relationships, communities, schools, and health-care settings can reduce intimate partner violence, HIV risk, and disrespectful maternity care.16, 18, 19 Indeed, four decades of the HIV response have revealed that approaches addressing discriminatory social and gender norms and power structures are effective in improving women's agency and sexual and reproductive health and rights.20 © 2021 Tommy Trenchard/Panos Pictures 2021 The UN and the Mexican and French Governments convened the Generation Equality Forum, with the most recent held in Paris on June 30 to July 2, 2021, to make concrete commitments to act on and resource gender equality and women's rights. Investments of US$40 billion were announced by governments and public sector institutions, UN entities, philanthropy, and the private sector.21 The UN is where the technical meets the political. It has a key leadership role in working with partners to bring evidence-based solutions together to promote healthy living and wellbeing for all as part of the 2030 Agenda for Sustainable Development. As representatives of UN agencies, the UN Special Rapporteur on the Right to Health, and civil society partners, we commit to leveraging the full power of our collective influence, access, and resources. We are seizing the opportunity to apply our collective knowledge and learning to focus efforts on strategies that have made change happen. First, we will reinforce and sustain our institutional capacity to deliver gender equality by increasing gender expertise in health, especially at senior levels. We will commit core financial resources to this agenda. Our organisations delivered results when our gender equality strategies were adequately resourced, and priority actions were central in our organisations' core programmes of work.13, 22 Second, we commit to obtaining sex-disaggregated data from our programmes and member states for priority health indicators. Data provide a powerful and empowering visibility to gender inequalities. Despite decades of guidance requiring health data to be sex-disaggregated, this basic requirement for informed decision making is still far from being the norm.23 The COVID-19 pandemic has shown that the health sector can do more to prioritise gender equality. Only 48% and 36% of 199 countries reported sex-disaggregated data on COVID-19 cases and deaths.24 Third, we will continue to leverage the expertise and capacity of feminist civil society to support the design, implementation, and monitoring of health policies, programmes, and community-centred solutions and hold our institutions accountable to our commitments.25 Innovative partnerships forged during the COVID-19 pandemic between governments, civil society, and the private sector have mitigated some disruptions in essential services, delivered essential commodities, supported survivors of violence, and brought justice to those unfairly persecuted.20 To drive structural change, voices from feminist civil society are needed at decision-making tables in clinics, communities, and our institutional governance structures. Finally, to control COVID-19 and mitigate its impacts, we need to tackle the structural determinants of gender inequality—eg, political participation and economic systems—and the intersections with other inequities. To do this effectively, we must join forces with social justice movements. Cross-movement activism had a pivotal role in the global HIV/AIDS response and ensuring access to HIV treatment in countries such as South Africa and Brazil.26 Therefore, our commitment to gender equality during this pandemic goes hand in hand with our commitment to global COVID-19 vaccine equity.27 Global vaccine inequity currently defines the global pandemic response. It is a moral and a public health failure that 75% of the 3·47 billion COVID-19 vaccine doses administered by July 12, 2021, were administered in only ten countries.28 A feminist response requires immediate sharing of knowledge, transfer of technological know-how, scale up of manufacturing, and the waiver of intellectual property protections for COVID-19 vaccines, as well as responding to gender-related barriers during vaccine deployment, including access for pregnant women. This approach will be essential if we are to have a fighting chance to prevent the erosion and reversal of hard-won health and gender equality gains. This online publication has been corrected. The corrected version first appeared at thelancet.com on August 5, 2021 We declare no competing interests. The views expressed in this Comment are personal and do not necessarily reflect the views of the UN. ==== Refs References 1 WHO WHO Coronavirus (COVID-19) Dashboard https://covid19.who.int/ July 12, 2021 2 UN The Sustainable Development Goals Report 2021 2021 United Nations Department of Economic and Social Affairs New York 3 Bourgault S Peterman A O'Donnell M Violence against women and children during COVID-19—one year on and 100 papers in 2021 Center for Global Development Washington, DC https://www.cgdev.org/sites/default/files/vawc-fourth-roundup.pdf 4 UNICEF COVID-19: a threat to progress against child marriage 2021 UNICEF Washington, DC https://data.unicef.org/resources/covid-19-a-threat-to-progress-against-child-marriage/ 5 Power K The COVID-19 pandemic has increased the care burden of women and families Sustainability Sci Pract Policy 16 2020 67 73 6 Almeida M Shrestha AD Stojanac D Miller LJ The impact of the COVID-19 pandemic on women's mental health Arch Womens Ment Health 23 2020 741 748 33263142 7 UNAIDS Rights in a pandemic—lockdowns, rights and lessons from HIV in the early response to COVID-19 2020 UNAIDS Geneva, Switzerland https://www.unaids.org/sites/default/files/media_asset/rights-in-a-pandemic_en.pdf 8 Azcona G Bhatt A Encarnacion J From insights to action: gender equality in the wake of COVID-19 2020 UN Women New York 9 WHO Second round of the national pulse survey on continuity of essential health services during the COVID-19 pandemic 2021 World Health Organization Geneva, Switzerland https://www.who.int/publications-detail-redirect/WHO-2019-nCoV-EHS-continuity-survey-2021.1 10 International Labour Organization ILO Monitor: COVID-19 and the world of work. Seventh edition https://www.ilo.org/wcmsp5/groups/public/@dgreports/@dcomm/documents/briefingnote/wcms_767028.pdf 2021 11 Morgan R Baker P Griffith DM Beyond a zero-sum game: how does the impact of COVID-19 vary by gender? Front Sociol 2021 published online June 15. 10.3389/fsoc.2021.650729 12 Kickbusch I Kirton J Health a political choice: act now, together 2020 http://repository.graduateinstitute.ch/record/298752/files/HPC-ActNowTogether.pdf 13 Lopes CA Allotey P Vijayasingham L Ghani F Ippolito AR Remme M What works in gender and health? Setting the agenda 2019 UNU-IIGH Kuala Lumpur, Malaysia 14 Htun M Weldon SL The civic origins of progressive policy change: combating violence against women in global perspective, 1975–2005 Am Political Sci Rev 106 2012 548 569 15 Wenham C Smith J Davies SE Women are most affected by pandemics— lessons from past outbreaks Nature 583 2020 194 198 32641809 16 Levy JK Darmstadt GL Ashby C Characteristics of successful programmes targeting gender inequality and restrictive gender norms for the health and wellbeing of children, adolescents, and young adults: a systematic review Lancet Glob Health 8 2020 e225 e236 31879212 17 Viswanathan K Hansen PM Rahman MH Can community health workers increase coverage of reproductive health services? J Epidemiol Community Health 66 2012 894 900 22068027 18 UNESCO Emerging evidence, lessons and practice in comprehensive sexuality education: a global review 2015 United Nations Educational, Scientific and Cultural Organization Paris 19 Grilo Diniz CS Rattner D Lucas d'Oliveira AFP de Aguiar JM Niy DY Disrespect and abuse in childbirth in Brazil: social activism, public policies and providers' training Reprod Health Matters 26 2018 19 35 30106349 20 UNAIDS Seizing the moment: tackling entrenched inequalities to end epidemics https://www.unaids.org/sites/default/files/media_asset/2020_global-aids-report_en.pdf 2020 21 UN Women Generation Equality Forum concludes in Paris with announcement of revolutionary commitments and global acceleration plan to advance gender equality by 2026. Generation Equality Forum https://forum.generationequality.org/news/generation-equality-forum-concludes-paris-announcement-revolutionary-commitments-and-global July 2, 2021 22 Gupta GR Oomman N Grown C Gender equality and gender norms: framing the opportunities for health Lancet 393 2019 2550 2562 31155276 23 WHO Addressing sex and gender in epidemic-prone infectious diseases 2007 World Health Organization Geneva, Switzerland 24 Global Health 50/50 The COVID-19 Sex-disaggregated Data Tracker: April update report https://globalhealth5050.org/wp-content/uploads/April-2021-Data-tracker-update.pdf 2021 25 Sen G Iyer A Chattopadhyay S Khosla R When accountability meets power: realizing sexual and reproductive health and rights Int J Equity Health 19 2020 111 32635915 26 Patterson D London L International law, human rights and HIV/AIDS Bull World Health Organ 80 2002 964 969 12571725 27 Georgieva K Ghebreyesus TA Malpass D Okonjo-Iweala N A new commitment for vaccine equity and defeating the pandemic June 1, 2021 International Monetary Fund. Views & Commentaries https://www.imf.org/en/News/Articles/2021/06/01/a-new-commitment-for-vaccine-equity-and-defeating-the-pandemic 28 Our World in Data COVID-19 vaccine doses administered https://ourworldindata.org/grapher/cumulative-covid-vaccinations 2021

34280381

PMC9752819

NO-CC CODE

2022-12-16 23:25:25

no

Lancet. 2021 Jul 20 7-13 August; 398(10299):471-474

utf-8

Lancet

10.1016/S0140-6736(21)01651-2

oa_other

==== Front Urology Urology Urology 0090-4295 1527-9995 Elsevier Inc. S0090-4295(21)00427-1 10.1016/j.urology.2021.05.019 Education #AUAMatch: The Impact of COVID-19 on Social Media Use in the Urology Residency Match Ho Patrick 1 Margolin Ezra 2 Sebesta Elisabeth 3 Small Alexander 4 Badalato Gina M. 2⁎ 1 Columbia University Vagelos College of Physicians and Surgeons, New York, NY 2 Department of Urology, Columbia University Irving Medical Center, New York, NY 3 Department of Urology, Vanderbilt University Medical Center, Nashville, TN 4 Department of Urology, Montefiore Medical Center, New York, NY ⁎ Address correspondence to: Gina M. Badalato, M.D., Department of Urology, Columbia University, 161 Fort Washington Avenue, 11th Floor, New York, NY 10032 23 5 2021 8 2021 23 5 2021 154 5056 2 3 2021 12 5 2021 © 2021 Elsevier Inc. All rights reserved. 2021 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Objectives To examine changes in Social Media (SoMe) use among urology residency applicants before and after the COVID-19 pandemic. Methods We distributed surveys to individuals who applied to our residency program for application cycles ending in 2018, 2019, and 2021. The surveys included questions about applicants’ SoMe use and perceptions of programs’ SoMe use during the application process, both before (2018/2019) and after (2021) the COVID-19 pandemic. The primary outcome was SoMe use for professional purposes. Results We received survey responses from 33% (162 of 496) and 29% (84 of 294) of applicants from the 2018/2019 and 2021 cohorts, respectively. There was a significant increase in professional SoMe use in the 2021 cohort (80%) compared with the 2018/2019 cohort (44%) (P < .001). In 2021 compared to 2018/2019, more applicants used SoMe to connect directly with residents (69% vs 34%, P < .001) and with faculty members (65% vs 15%, P < .001). Applicants in 2021 compared to 2018/2019 more often found SoMe to be useful for making decisions about applying to (33% vs 10%), interviewing at (26% vs 7%), and ranking programs (20% vs 9%) (all P < .05). Twitter was the most common platform for applicants to access program information, increasing from 38% to 71%. Conclusion The COVID-19 pandemic ushered in a period of unprecedented SoMe usage among urology applicants, who used it to learn about and connect with residency programs in new ways. The use of SoMe by residency programs has become an important component of trainee recruitment and is likely to continue in the future. ==== Body pmcSocial Media (SoMe) use has rapidly expanded within the urological community over the last decade. Urologists use SoMe platforms such as Facebook, Twitter, Instagram, and LinkedIn for networking, communicating at conferences, sharing research, and disseminating clinical information. Twitter is the most commonly used platform in professional settings amongst urologists, exemplified by a nearly fivefold increase in the volume of posts at the American Urological Association (AUA) annual meetings from 2013-2018.1 , 2 Sharing urological research over SoMe has become so prolific that article citations are now positively correlated with Twitter mentions, and researchers have developed predictive models of publication impact based on SoMe metrics.3 , 4 SoMe has also been useful to crowdsource advice for challenging cases and to share clinical practice guidelines.5 , 6 Additionally, SoMe has an important role in urological training and graduate medical education. Millennials (those born between the early-1980 and mid-1990) constitute the majority of current and upcoming urology trainees, and this generation has been defined by the ubiquity of technology, mobile devices, and social media in their lives. Multiple surveys have shown that SoMe has a significant influence on residents’ urological knowledge acquisition and is important for career development.7, 8, 9 Urology residency programs have recognized this importance, and the majority now have dedicated program Twitter accounts. Activity from these accounts has even been associated with departmental reputation scores.10 , 11 Despite the prevalence and importance of SoMe in urological training, little is known about its utility in recruiting future urology residents. Medical students leverage SoMe to learn about residency programs, demonstrate their interest, and engage directly with residents and faculty online. During the 2020-2021 application cycle, SoMe became particularly relevant to medical students, as the COVID-19 pandemic limited in-person interactions, mandated virtual interviews, and prohibited visiting sub-internships. Improving our understanding of SoMe use among applicants would be highly useful in future recruitment efforts. In this study, we examine the impact of the COVID-19 pandemic on SoMe use among urology residency applicants. MATERIALS AND METHODS We distributed surveys (Supplementary Material) via email to all individuals who applied to our urology residency program for application cycles ending in 2018, 2019, and 2021. The surveys reached 790 applicants out of 1398 who registered for the AUA Residency Match Program for those three cycles.12 The surveys were distributed after the conclusion of the AUA match results, and responses were anonymized to protect participants. Surveys were administered on the Qualtrics platform (Qualtrics International Inc, Provo, UT). The surveys queried applicants about basic demographic information, SoMe use during the residency application and interview processes, and attitudes toward residency programs’ SoMe use. In 2021, we added questions about changes to the residency application process related to the COVID-19 pandemic and about SoMe use in future application cycles. Incomplete survey responses were excluded. In order to determine the impact of the COVID-19 pandemic on SoMe use, we separated the survey responses into a pre-pandemic cohort (responses from 2018 and 2019) and a pandemic cohort (responses from 2021). The primary outcome was professional usage of SoMe. This was considered a positive outcome if the response to the question, “How would you describe the usage of your social media accounts?” indicated professional use only or a combination of professional and personal use. Descriptive statistics were summarized using medians and percentages. X2 tests and Wilcoxon rank-sum tests were used to compare categorical variables and ranked responses, respectively. A multivariable logistic regression model was constructed to evaluate the association between application cohort and professional SoMe use while controlling for demographic factors. The primary objective of the study was initially descriptive, so no power analysis was done at first (during the 2018 and 2019 distributions). However, during the COVID-19 pandemic, we amended the objective to analyze the impact of the pandemic on SoMe usage. At this point, we conducted a power analysis to determine the required sample size. A 2019 study examining the role of SoMe in the application process for anesthesia residency reported that 45% used SoMe to research programs during the application process.13 Assuming a similar pre-pandemic rate, a X2 test with a two-sided significance level of 0.05 required a sample size of 192 responses in order to detect a 20% increase in SoMe use with 80% power. Statistical analysis was completed using Stata 16 (StataCorp, College Station, TX). Sample size calculation was performed using G*Power 3.1 (Dusseldorf, Germany). Tests with a P value < .05 were considered statistically significant. This study was approved by the Columbia University Institutional Review Board. RESULTS Applicant Demographics A total of 496 and 294 applicants were sent surveys in the 2018/2019 and 2021 cohorts, respectively. One hundred sixty-two applicants (33%) responded in 2018/2019, and 84 applicants (29%) responded in 2021. The majority of respondents in both 2018/2019 and 2021 identified as male (65% and 55%, respectively, Table 1 ). The mean age was 27.0 (SD = 1.9) and 27.6 (SD = 2.1) for the 2018/2019 and 2021 cohorts, respectively.Table 1 Applicant demographics by cohort* Table 1 2018/2019 N = 162 2021 N = 84 P value Gender (%) .102  Male 106 (65) 46 (55)  Female 56 (35) 38 (45) Mean age (SD) 27.0 (1.9) 27.6 (2.1) .027 Ethnicity (%) <.001  African American; non–Hispanic black 11 (7) 3 (4)  Caucasian; non–Hispanic white 87 (54) 34 (40)  East Asian 18 (11) 16 (19)  Hispanic 7 (4) 18 (21)  Middle Eastern/North African 10 (6) 4 (5)  South Asian 16 (10) 3 (4)  Other 13 (8) 6 (7) Number of Applications (%) .042  <50 16 (10) 11 (13)  50-70 50 (31) 15 (18)  70-90 51 (31) 20 (24)  >90 45 (28) 38 (45) Number of Interviews (%) <.001  <5 6 (4) 27 (32)  5-10 16 (10) 12 (14)  10-15 55 (34) 11 (13)  15-20 60 (37) 23 (27)  >20 25 (15) 11 (13) ⁎ X2 tests used for hypothesis testing with categorical variables. Wilcoxon rank-sum tests used for continuous variables. Applicant SoMe Usage The percentage of respondents who had any social media accounts was 95% in both the 2021 cohort and the 2018/2019 cohort. The use of SoMe for professional purposes increased significantly in the 2021 cohort, compared with the 2018/2019 cohort (80% vs 44%, P < .001, Table 2 ). On multivariable logistic regression controlling for age and gender, applicants in the 2021 cohort were significantly more likely to use SoMe professionally, compared to applicants in the 2018/2019 cohort (OR 4.68, 95% CI 2.49-8.78, P < .001, Supplementary Table 1).Table 2 Applicant and program some use* Table 2 2018/2019 N = 162 2021 N = 84 P-value Applicant SoMe Usage Any SoMe accounts (%) 154 (95) 80 (95) .951 Professional SoMe use (%) 72 (44) 67 (80) <.001 SoMe connections (%)  With applicant(s) 118 (73) 64 (76) .570  With residents(s) 55 (34) 58 (69) <.001  With faculty 25 (15) 55 (65) <.001  With program coordinator(s) 7 (4) 31 (37) <.001 Privacy settings changed (%) 79 (49) 35 (42) .290 Posted on SoMe about interview process 40 (25) 29 (35) .104 Perceptions of Program SoMe Usage Percentage of programs with SoMe resources available (%) <.001  1%-25% 54 (33) 7 (8)  26%-50% 50 (31) 7 (8)  51%-75% 32 (20) 36 (43)  76%-100% 10 (6) 34 (40)  No response 16 (10) 0 (0) Programs’ SoMe resources were useful when deciding:  Whether to apply to a program (%) <.001   Yes 10 (6) 28 (33)   No 137 (85) 37 (44)   Maybe 15 (9) 19 (23)  Whether to interview at a program (%) <.001   Yes 12 (7) 22 (26)   No 137 (85) 45 (54)   Maybe 13 (8) 17 (20)  How to rank a program (%) .019   Yes 14 (9) 17 (20)   No 126 (78) 53 (63)   Maybe 22 (14) 14 (17) SoMe platform most likely to access to learn about a residency program (%) <.001  Twitter 61 (38) 60 (71)  Facebook 33 (20) 4 (5)  Instagram 22 (14) 9 (10)  YouTube channel 23 (14) 10 (12)  Other 7 (4) 2 (2)  No response 16 (10) 0 (0) ⁎ X2 tests used for hypothesis testing. Of the various SoMe platforms, there was significantly more frequent use of Twitter (P < .001) and LinkedIn (P = .036) in 2021 compared to 2018/2019 (Table 3 ). Half of applicant's pre-pandemic did not have a Twitter account, whereas in 2021 45% reported using Twitter at least once a day. There was a significant decrease in the frequency of Doximity use (P < .001), with 52% of 2018/2019 applicants using it at least once per month compared to 32% of the 2021 cohort. There was no significant difference in the frequency of Facebook or Instagram use.Table 3 Applicant some use by platform and frequency* Table 3 No account < 1/y At least 1x/y At least 1x/mo At least 1x/wk At least 1x/d P-value Facebook Use (%) 2018/2019  N = 162 18 (11) 2 (1) 6 (4) 10 (6) 29 (18) 97 (60) .081 2021  N = 84 7 (8) 2 (2) 8 (10) 5 (6) 24 (29) 38 (45) Twitter Use (%) 2018/2019  N = 162 81 (50) 7 (4) 15 (9) 16 (10) 13 (8) 30 (19) <.001 2021  N = 84 14 (17) 1 (1) 4 (5) 9 (11) 18 (21) 38 (45) LinkedIn Use (%) 2018/2019  N=162 74 (46) 18 (11) 37 (23) 28 (17) 5 (3) 0 (0) .036 2021  N = 84 28 (33) 9 (11) 23 (27) 18 (21) 5 (6) 1 (1) Instagram Use (%) 2018/2019  N = 162 40 (25) 4 (2) 6 (4) 5 (3) 25 (15) 82 (51) .150 2021  N = 84 12 (14) 0 (0) 3 (4) 7 (8) 14 (17) 48 (57) Doximity Use (%) 2018/2019  N = 162 40 (25) 11 (7) 26 (16) 67 (41) 18 (11) 0 (0) <.001 2021  N = 84 32 (38) 15 (18) 10 (12) 24 (29) 2 (2) 1 (1) ⁎ Wilcoxon rank-sum tests used for hypothesis testing The proportion of applicants who made SoMe connections (“friends” or “follows”) with other applicants did not significantly differ between the application cycles. However, the proportion of applicants who connected with residents more than doubled in 2021 compared to 2018/2019 (69% vs 34%, P < .001). Connections with faculty members increased by more than fourfold (65% vs 15%, P < .001), and connections with program coordinators increased by more than ninefold (37% vs 4%, P < .001). There was no significant difference in the applicants who changed their SoMe privacy settings in 2021 compared to pre-pandemic during the application season (49% vs 51%, P = .290). Of the 80 survey respondents in 2021 who reported using SoMe, 59 (74%) reported that application changes due to the COVID-19 pandemic (eg, virtual interviews, lack of away rotations) directly caused them to increase their SoMe use. Additionally, 32 (40%) reported posting original content about urology on SoMe, such as slides about a topic or a link to their own manuscript. Perceptions of Program SoMe Usage The median percentage of programs reported to have SoMe resources available increased significantly, from 26-50% in 2018/2019 to 51%-75% in 2021 (P < .001) (Table 2). The proportions of applicants who found SoMe to be useful when deciding whether to apply to, whether to interview at, and how to rank a particular program were also significantly higher in 2021 than in 2018/2019. Applicants in 2021 were most likely to access program information on Twitter (71%), compared to 38% of applicants in 2018/2019 (P < .001). The majority of applicants felt that program websites, colleagues, and attending physicians were the most important resources to gather program-specific information ( Fig. 1 ). However, the importance of SoMe in information gathering increased significantly, with 43% rating SoMe as “extremely important” or “very important” in 2021, compared with 9% in 2018/2019 (P < .001).Figure 1 Importance of residency resources for prospective applicants. *Represents P < .05, Wilcoxon rank-sum test. Figure 1 The posts on program SoMe pages that applicants found to be most useful were videos describing different aspects of the program or faculty (27%), followed by pictures of residents socializing (15%) (Supplementary Table 2). While 42% of applicants did not find any program SoMe practices to negatively influence their perceptions of a program, high frequency of posts and posting pictures/screenshots of applicants without their consent were the most commonly cited negative practices (15% and 17%, respectively). Impressions and Future Directions In the 2021 cohort, 39% of applicants felt that SoMe engagement had neither a positive nor a negative impact on their application and match prospects (Supplementary Table 3). The most commonly cited positive impact of SoMe was its ability to make applicants more visible to programs (32%). The most commonly cited negative impact of SoMe was feeling overshadowed by other applicants who were more active (55%). Looking to future application cycles, 21% of applicants would like to see more interaction between applicants and programs on SoMe, with 39% desiring a similar amount of interaction and 39% desiring less interaction. COMMENT Urology residency applicants are increasingly using SoMe for professional networking. The frequency of Twitter uses in particular underwent marked growth, with median usage increasing from “less than once a year” to “greater than once a week.” The percentage of applicants who used Twitter grew from 50%-83% between the 2018/2019 and 2021 cohorts. For United States. adults ages 18-29, Twitter use only grew from 40%-42% during 2018-2021.14 , 15 Facebook use decreased in this age group, while Instagram and LinkedIn use both increased slightly. More broadly, overall rates of SoMe use in this age group of the general U.S. population decreased from 88% (2018) to 84% (2021). Compared to prior cycles, the number of SoMe connections between applicants and residents, faculty, and program coordinators grew dramatically. As applicants gather information about training programs online, SoMe is an increasingly important resource, and applicants are incorporating information from SoMe as they make decisions about where to apply, where to interview, and how to rank programs. Social media habits among urology residency applicants, particularly during the 2020-2021 virtual application cycle, have not been well described. Our results showing increased professional SoMe uses are consistent with anecdotal experiences and other specialties seeing SoMe utilized to connect residency programs with prospective applicants. In a 2018 study of applicants applying for ophthalmology residency at Penn State University, almost half of respondents expressed that SoMe was helpful and desired an increase in program SoMe use to disseminate program information.16 In another study from Mayo Clinic surveying anesthesiology residency applicants for the 2017-2018 cycle, 45% of respondents used SoMe to research prospective programs.13 Finally, a study assessing applicants to an integrated Plastic and Reconstructive Surgery program in 2018 found that nearly three-fourths of respondents followed a residency SoMe account, with particular interest in resident life.17 The virtual application environment created by the COVID-19 pandemic heavily drove the rise in SoMe use, as reported by the vast majority of applicants in our survey. In a year when the in-person components of the application cycle were replaced with virtual open houses, virtual sub-internships, and virtual interviews, SoMe offered an easy and convenient platform for programs to share information, showcase their departments’ accomplishments, and connect with applicants.18, 19, 20 Other surgical subspecialties that traditionally required away rotations and evaluation letters from other institutions have similarly responded to the challenges of COVID-19. A 2020 study assessing otolaryngology resident recruitment during the pandemic found that over one-third of otolaryngology department and residency Twitter accounts were created in 2020, with the majority advertising virtual open house meet-and-greets.21 Urology applicants most commonly turned to departmental Twitter accounts to access updates and program information. Twitter also emerged as a platform for students to share their own original content with programs and fellow applicants, allowing the applicants to gain visibility in a virtual setting. Students were noted to participate in COVID-born virtual urology lecture series, such as the New York Section AUA Educational Multi-institutional Program for Instructing Residents (EMPIRE) and UCSF Urology Collaborative Online Video Didactics (COViD).22 They also flocked to participate in the “UroStream” online mentoring program that linked students and residents to write “Tweetorial” threads of posts on various urological topics. Applicants’ creativity and resilience were on full display like never before. While SoMe certainly helped connect applicants and residency programs in a cycle restricted by the pandemic, its heavily increased use also carries potential pitfalls that require further consideration. A recent study revealed that more than 11% of urology residents and fellows meet criteria for Social Media Disorder (SMD), a problem characterized by addiction to and compulsive use of SoMe.7 The majority of our survey respondents expressed feelings of being overshadowed by other applicants on SoMe, and as SoMe use becomes increasingly more important (or perceived as more important), the risk of disordered use may rise. Furthermore, while our survey revealed applicant perceptions of program SoMe use, the question of how programs perceive applicant SoMe use remains. As evident from the recent retraction of an article examining the prevalence of “unprofessional” SoMe use among young vascular surgeons, applicant SoMe use could be at risk for scrutiny and biased subjective judgments unrelated to professional competency.23 Should programs begin weighing applicant SoMe use in the evaluation of their candidacy, medical students should be encouraged to review and adopt SoMe professional guidelines. These include those published by BJU International, the European Association of Urology, and the American Urological Association.24, 25, 26 Finally, the rapid growth of SoMe connections between applicants and programs raises the important question of how SoMe use should be viewed in a residency match that emphasizes equity, exemplified by existing rules limiting post-interview communication. Our study has several limitations. Our survey responses represent 31% of our program's applicants and 18% of all 1,398 students who registered for the match during these three cycles. Thus, our results may not be generalizable to the entire applicant population. Our study had a slightly higher percentage of female respondents than the overall applicant pools for the respective time periods.27, 28, 29 However, gender was not found to be a significant predictor of professional SoMe use on multivariable logistic regression. Self-selection and response bias are potential concerns as well, though we attempted to limit response bias by ensuring anonymity and administering the survey following the match to alleviate any concerns that responses would affect the results. Additionally, we did not administer the survey in 2020, limiting our ability to measure the precise trends over time. Finally, while we report significant changes following a landmark event, we cannot definitively establish causation to link the COVID-19 pandemic to the rise in SoMe among residency applicants. Despite these limitations, this study offers important insights into the rising role of SoMe in the urology residency application process. This information can help programs better understand the evolving landscape of SoMe in order to optimize their online presence for their applicants. While the COVID-19 pandemic certainly created an unusual environment for residency applications, the extent to which this increased SoMe usage will continue in future application cycles has yet to be determined. Going forward, as applicants experience increasing pressure to use SoMe to connect with programs, it would be enlightening to further understand how applicants are using SoMe for their own promotion and whether programs find it to be a useful avenue to gather information about them. CONCLUSION Urology residency applicants are increasingly using SoMe to learn about and connect with residency programs. Twitter, in particular, has emerged as an essential resource for networking and information dissemination during the application process. The use of SoMe by residency programs has become an important component of trainee recruitment. Appendix SUPPLEMENTARY MATERIALS Image, application 1 Image, application 2 Image, application 3 Image, application 4 Declarations of interest: None Supplementary material associated with this article can be found in the online version at https://doi.org/10.1016/j.urology.2021.05.019. ==== Refs References 1 Loeb S Carrick T Frey C Titus T Increasing social media use in urology: 2017 american urological association survey Eur Urol Focus 6 2020 605 608 31351900 2 Matta R Doiron C Leveridge MJ The dramatic increase in social media in urology J Urol 192 2014 494 498 24576656 3 Hayon S Tripathi H Stormont IM Dunne MM Naslund MJ Siddiqui MM Twitter mentions and academic citations in the urologic literature Urology 123 2019 28 33 30278190 4 Sathianathen NJ Lane R 3rd Condon B Murphy DG Lawrentschuk N Weight CJ Early online attention can predict citation counts for urological publications: The #UroSoMe_score Eur Urol Focus 6 2020 458 462 31704280 5 Koo K Shee K Gormley EA Following the crowd: patterns of crowdsourcing on Twitter among urologists World J Urol 37 2019 567 572 30014160 6 Loeb S Roupret M Van Oort I N'dow J Gurp Mv Bloemberg J Darraugh J Ribal MJ Novel use of twitter to disseminate and evaluate adherence to clinical guidelines by the european association of urology BJU Int 119 2017 820 822 28170154 7 Dubin JM Greer AB Patel P Global survey of the roles and attitudes toward social media platforms amongst urology trainees Urology 147 2021 64 67 32950594 8 Rivas JG Socarras MR Patruno G Perceived role of social media in urologic knowledge acquisition among young urologists: a european survey Eur Urol Focus 4 2018 768 773 28753825 9 Salem J Borgmann H Baunacke M Widespread use of internet, applications, and social media in the professional life of urology residents Can Urol Assoc J 11 2017 E355 e366 29382458 10 Chandrasekar T Goldberg H Klaassen Z Wallis CJD Leong JY Liem S Teplitsky S Noorani R Loeb S Twitter and academic Urology in the United States and Canada: a comprehensive assessment of the Twitterverse in 2019 BJU Int 125 2020 173 181 31602782 11 Ciprut S Curnyn C Davuluri M Sternberg K Loeb S Twitter activity associated with U.S. news and world report reputation scores for urology departments Urology 108 2017 11 16 28669746 12 Urology and Specialty Matches 2021 https://www.auanet.org/education/auauniversity/for-residents/urology-and-specialty-matches PublishedAccessed February 15, 2021 13 Renew JR Ladlie B Gorlin A Long T The Impact of Social Media on Anesthesia Resident Recruitment J Educ Perioper Med 21 2019 E632 14 Social Media Use in 2018 2018 Pew Research Center Washington, DC 15 Social Media Use in 2021 2021 Pew Research Center Washington, DC 16 Goerlitz-Jessen M Behunin N Montijo M Wilkinson M Recruiting the digital-age applicant: the impact of ophthalmology residency program web presence on residency recruitment J Acad Ophthalmol 10 2018 e32 e37 17 Steele TN Galarza-Paez L Aguilo-Seara G David LR Social media impact in the Match: A survey of current trends in the United States Arch Plast Surg 48 2021 107 113 33503753 18 Jiang J Key P Deibert CM. Improving the Residency Program Virtual Open House Experience: A Survey of Urology Applicants Urology 146 2020 1 3 33049230 19 Kenigsberg AP Khouri RK Jr. Kuprasertkul A Wong D Ganesan V Lemack GE Urology residency applications in the COVID-19 Era Urology 143 2020 55 61 32562774 20 Margolin EJ Gordon RJ Anderson CB Badalato GM Reimagining the away rotation: A 4-week virtual subinternship in urology J Surg Educ 2021 21 DeAtkine AB Grayson JW Singh NP Nocera AP Rais-Bahrami S Greene BJ #ENT: otolaryngology residency programs create social media platforms to connect with applicants during covid-19 pandemic Ear Nose Throat J 2020 145561320983205 22 Smigelski M Movassaghi M Small A Urology virtual education programs during the covid-19 pandemic Curr Urol Rep 21 2020 50 33090272 23 Hardouin S Cheng TW Mitchell EL Raulli SJ Jones DW Siracuse JJ Farber A RETRACTED: Prevalence of unprofessional social media content among young vascular surgeons J Vasc Surg 72 2020 667 671 31882313 24 Social media best practices. american urological association. http://auanet.mediaroom.com/index.php?s=20294. Accessed April 2021. 25 Borgmann H Cooperberg M Murphy D Loeb S N'Dow J Ribal MJ Woo H Rouprêt M Winterbottom A Wijburg C Wirth M Catto J Kutikov A Online professionalism-2018 update of european association of urology (@uroweb) recommendations on the appropriate use of social media Eur Urol 74 2018 644 650 30177286 26 Murphy DG Loeb S Basto MY Challacombe B Trinh Q-D Leveridge M Morgan T Dasgupta P Bultitude M Engaging responsibly with social media: the BJUI guidelines BJU Int 114 2014 9 11 27 2019 Urology Residency Match Statistics. American Urological Association.https://www.auanet.org/documents/education/specialty-match/2019-Urology-Residency-Match-Statistics.pdf. Accessed April 27, 2021. 28 2018 Urology Residency Match Statistics. American Urological Association.https://www.auanet.org/Documents/education/specialty-match/2018-Urology-Residency-Match-Statistics.pdf. Accessed April 27, 2021. 29 2021 Urology Residency Match Statistics. American Urological Association.https://www.auanet.org/documents/education/specialty-match/2021-Urology-Residency-Match-Statistics.pdf. Accessed April 27, 2021.

34033828

PMC9752821

NO-CC CODE

2022-12-16 23:25:25

no

Urology. 2021 Aug 23; 154:50-56

utf-8

Urology

10.1016/j.urology.2021.05.019

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Published by Elsevier Ltd. S0140-6736(21)01747-5 10.1016/S0140-6736(21)01747-5 Department of Error Department of Error 5 8 2021 7-13 August 2021 5 8 2021 398 10299 490490 © 2021 Published by Elsevier Ltd. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcDanovaro-Holliday MC, Kretsinger K, Gacic-Dobo M. Measuring and ensuring routine childhood vaccination coverage. Lancet 2021; 398: 468–69—This Comment had incorrect copyright information. This correction has been made to the online version as of Aug 5, 2021, and the printed version is correct.

34364525

PMC9752822

NO-CC CODE

2022-12-16 23:25:25

no

Lancet. 2021 Aug 5 7-13 August; 398(10299):490

utf-8

Lancet

10.1016/S0140-6736(21)01747-5

oa_other

==== Front Urology Urology Urology 0090-4295 1527-9995 Published by Elsevier Inc. S0090-4295(21)00076-5 10.1016/j.urology.2020.12.043 Article AUTHOR REPLY Koven Alexander 13⁎ O'Kelly Fardod 13 El-Ghazzaoui Ali 13 Langer Jacob C. 23 Levin David 4 Diamond Aubie 15 Goldstein Yisroel 12 Reichman Ezriel 12 Koyle Martin A. 13 1 Division of Pediatric Urology, Hospital for Sick Children, Toronto, Ontario, Canada 2 Division of General and Thoracic Surgery, Hospital for Sick Children, Toronto, Ontario, Canada 3 Department of Surgery, University of Toronto, Toronto, Ontario, Canada 4 Department of Anesthesiology and Pain Medicine, Hospital for Sick Children, Toronto, Ontario, Canada 5 Abraham Diamond Medicine Professional Corporation ⁎ Address correspondence to: Alexander Koven, M.D., B.A.Sc., Division of Pediatric Urology, Hospital for Sick Children, Toronto, Ontario, Canada. 11 8 2021 8 2021 11 8 2021 154 248248 21 9 2020 13 12 2020 Copyright © 2021 Published by Elsevier Inc. 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcSince the COVID-19 pandemic began, our institution, like so many others, has been prioritizing cases and replacing in-person visits with virtual clinic platforms. In the operating room, family presence during induction and the presence of observers has also been restricted. These restrictions have led to challenges for the program described, but have also promoted creative thinking, innovative options and opportunities. In this context, we have modified the protocol, instead utilizing virtual communication with the family, who remain in the waiting area but are able to be involved with the ceremony portion after anesthesia has been safely induced and before the definitive procedure is performed. In addition, family and friends in other geographic locations may join in with no limitation to numbers, and without added risks such as contamination of sterile fields or distractions in the operating room.1 This preserves quality and safety, allows the ceremony to remain interactive and unchanged from that described, aside from family “in person” participation. This alternative may be applicable even after the pandemic abates to institutions where policy and procedures preclude “in-person” involvement with a cultural-based procedure such as this; thus, allowing a virtual family and patient focused alternative where none existed before. Alexander Koven M.D. Martin Koyle M.D., M.Sc. ==== Refs Reference 1 Lee, M., Koven, A., Chua, M., & Koyle, M. Incorporating modern technology with traditional ceremony of Brit Milah during the COVID-19 pandemic: virtual Brit Milah during COVID-19 pandemic. Can Urol Assoc J. Accepted for publication.

34389074

PMC9752823

NO-CC CODE

2022-12-16 23:25:25

no

Urology. 2021 Aug 11; 154:248

utf-8

Urology

10.1016/j.urology.2020.12.043

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Published by Elsevier Ltd. S0140-6736(21)01692-5 10.1016/S0140-6736(21)01692-5 Department of Error Department of Error 5 8 2021 7-13 August 2021 5 8 2021 398 10299 490490 © 2021 Published by Elsevier Ltd. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcFisseha S, Sen G, Ghebreyesus TA, et al. COVID-19: the turning point for gender equality. Lancet 2021; 398: 471–74—In this Comment, part of the fourth sentence of the first paragraph is corrected to read “increasing risk of child marriages and female genital mutilation” and the Mexican Government is now included at the start of the first sentence of the sixth paragraph to read “The UN and the Mexican and French Governments convened the Generation Equality Forum, with the most recent held in Paris…” These corrections have been made to the online version as of Aug 5, 2021, and the printed version is correct.

34364523

PMC9752824

NO-CC CODE

2022-12-16 23:25:25

no

Lancet. 2021 Aug 5 7-13 August; 398(10299):490

utf-8

Lancet

10.1016/S0140-6736(21)01692-5

oa_other

==== Front Urology Urology Urology 0090-4295 1527-9995 Elsevier Inc. S0090-4295(21)00392-7 10.1016/j.urology.2021.03.050 Infertility Video Visits are Practical for the Follow-up and Management of Established Male Infertility Patients Andino Juan a⁎ Zhu Alex a Chopra Zoey b Daignault-Newton Stephanie a Ellimoottil Chad ac Dupree James M. ac a Michigan Medicine Department of Urology, Ann Arbor, MI b University of Michigan Medical School, Ann Arbor, MI c Institute for Healthcare Policy and Innovation, Ann Arbor, MI ⁎ Address Correspondence to: Juan J Andino M.D., M.B.A., University of Michigan Department of Urology, A. Alfred Taubman Health Care Center - Room 3875, 1500 E. Medical Center Drive, SPC 5330, Ann Arbor, Michigan 48109-5330. 19 5 2021 8 2021 19 5 2021 154 158163 25 1 2021 31 3 2021 © 2021 Elsevier Inc. All rights reserved. 2021 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Objective To study the use of video visits for male infertility care prior to the COVID-19 pandemic Methods We reviewed video visits for male infertility patients completed at a tertiary academic center in southeast Michigan. These patients had follow-up after an initial in-person evaluation. We designed this retrospective case series to describe the diagnostic categories seen through telehealth, management steps completed during video visits, and to understand whether additional in-person care was required within 90 days of video visits. In addition, we estimated time and cost savings for patients attributed to video visits. Results Most men seen during video visits had an endocrinologic (29%) or anatomic (21%) cause for their infertility. 73% of video visits involved reviewing results; 30% included counseling regarding assistive reproductive technologies; and 25% of video visits resulted in prescribing hormonally active medications. The two patients (3%) who were seen in clinic after their video visit underwent a varicocelectomy in the interim. No patients required an unplanned in-person visit. From a patient perspective, video visits were estimated to save a median of 97 minutes (IQR 64-250) of travel per visit. Median cost savings per patient— by avoiding travel and taking time off work for a clinic visit—were estimated to range from $149 (half day off) to $252 (full day off). Conclusion Video visits for established male infertility patients were used to manage different causes of infertility while saving patients time and money. Telehealth for established patients did not trigger additional in-person evaluations. ==== Body pmcFollowing the coronavirus disease 2019 (COVID-19) pandemic, experts estimated that there were approximately 1 billion telehealth visits in the United States in 2020.1 Video visits—a form of telehealth using live, simultaneous audio and visual interactions to connect patients and providers—are not new. Historically, regulatory and reimbursement policies were cited as major barriers to wide-spread telehealth use.2 However, the declaration of a public health emergency in March 2020 resulted in a rapid expansion of telehealth services by relaxing regulations at the state and national levels.3, 4, 5 In particular, the Centers for Medicare & Medicaid temporarily changed regulatory requirements to allow more patients to engage in telehealth from their homes.6 , 7 Complementing these national policies, many state-specific changes have permitted Medicaid and privately insured patients to receive care from home and allowed providers to practice across state lines.8 The American Society of Reproductive Medicine (ASRM), Society for Male Reproduction and Urology, and the Society for the Study of Male Reproduction recommend clinicians use telehealth for reproductive consultations, to develop treatment plans, begin or continue evaluations, and educate patients.9 , 10 The COVID-19 public health emergency and associated telehealth regulatory changes have been extended through April of 2021.11 Infertility is defined as the inability to conceive after one year of unprotected intercourse and affects approximately 15% of US couples.12 Both the ASRM and National Institute for Healthcare Excellence (NICE) recommend male and female partners receive infertility evaluations.12 , 13 However, there are numerous barriers for accessing infertility care, including the geographic distribution of providers.14 Though video visits have been previously studied in general adult and pediatric urology populations,5 , 15, 16, 17, 18, 19 to date no studies have specifically evaluated the use of telehealth in male infertility care. We hypothesized that video visits for established patients served as substitutes to follow-up, in-person care. It is plausible that video visits for male infertility may effectively facilitate follow-up care while reducing financial and geographic constraints for patients. Conversely, the sensitive nature of infertility care as well as the importance of the physical exam may result in telehealth being used in addition to in-person visits. Since little is known about the role of telehealth in male infertility, we designed a retrospective case series to understand how video visits were used to provide male infertility care prior to the COVID-19 pandemic. Specifically, we sought to understand the etiologies for which established patients used telehealth for follow-up. We also reviewed how video visits were used in the evaluation and management of infertility. Finally, we estimated the financial benefits for patients using video visits by calculating travel costs and lost wages that would have been associated with in-person care. METHODS This case series is a retrospective review of outpatient video visits in the department of urology at a single academic institution from August 21, 2017 through March 17, 2020. This study was deemed exempt by the Institutional Review Board (HUM00141665). We ended the study when COVID-19 was declared a public health emergency. After March 17th, 2020 there were wide-spread changes to telehealth policies that would have introduced confounders to our study objectives.20 , 21 We included established patients seen for video visit follow-up of issues related to male infertility. All men had undergone a previous in-person examination with a urologist in the division of andrology. We excluded men younger than 18 years of age. Video visits were performed by a single urologist with andrology fellowship training. All video visits were performed using a HIPAA-compliant, video communication system integrated into the electronic health record (EHR). New patient video visits were not reimbursed or performed prior to March 17th, 2020 and were not included in this study. Chart review was conducted in the EHR to identify study variables, including age, gender, race and ethnicity, preferred language, referring provider, occupation, home zip code, and clinic location where an in-person visit would have taken place. Chart review was limited to data captured from Michigan Medicine and Mid-Michigan hospitals and affiliated outpatient clinics. Diagnostic Categories and Patient Management Our primary objective was to describe the landscape of male infertility diagnoses seen via video visits and what management was performed through telehealth. We first classified visits according to diagnostic category by evaluating the clinical history. Supplemental Figure 1 details all the individual diagnoses identified and how they were grouped into diagnostic categories. We then examined the management steps completed during the virtual encounters. Patient management categories included reviewing results, managing medications, referrals to other specialists, and counseling regarding sperm extraction, varicocelectomy, assistive reproductive technologies (ART), or cryopreservation. Visits often included multiple management steps, and these are broken down into more detail in supplemental figure 2. We extracted information about whether patients had clinic or emergency room visits for any urologic condition 90 days after their video visits within our health system, including Michigan Medicine and Mid-Michigan hospitals and affiliated outpatient clinics. Obtaining a semen analysis or DNA fragmentation index testing requires provided semen at infertility clinics but these interactions with the healthcare system were not categorized as in-person visits as they do not include interaction with a provider. Patient Time and Cost Savings Our secondary objective was to evaluate patient time saving and financial benefits from video visit utilization. Travel cost was estimated based on clinic visits with an andrologist and do not include travel for laboratory tests or semen analyses. We calculated round-trip driving distance and driving time using Google Maps࣪. We used each patients’ home addresses documented in the EHR and address of the outpatient urology clinic they would have visited in person. To control for effects of traffic variation in our calculations, we used Wednesday at 10AM as our index time. Total time saved from avoiding round trip travel was converted to ordinal categories ranging from less than one hour to greater than 9 hours of travel to depict the travel burden in this patient cohort. We calculated transportation costs by multiplying travel distance with the American Automobile Association's (AAA) cost per mile driving estimate of $0.59 per mile for 2017-2019.22 AAA's cost estimate includes the price of fuel, tires, maintenance, insurance, depreciation, license, and registration. We also estimated potential lost wages had the patient needed to take time off from work to attend an in-person appointment. Patient occupations were identified from chart. The provider who initially saw these patients always documents patient occupation as part of his note template. Salary data estimates were obtained from Glassdoor.com based on occupation documented in the EHR. We used the salaries from Glassdoor to project the potential lost wages from missing half or full days of work to attend an in-person visit. To analyze the financial impact that video visits have on patients of varying occupations, we categorized patients as being “blue collar” versus “white collar” employees using the International Standard Classification of Occupations (ISCO-08) from the International Labour Organization.23 Generally, “blue collar” workers engaged in outdoor, manual, agricultural, manufacturing, or service industries. “White collar” workers engaged in non-manual office work. We estimated total cost avoidance by summing the calculated transportation costs with the estimated wages lost had the patient had taken time off from work to attend an in-person appointment. The median cost savings of “blue collar” and “white collar” workers were calculated independently. Total median cost savings of all patients were also calculated, with no respect to “blue collar” or “white collar” identification. RESULTS Between August 21, 2017 and March 17, 2020, 70 infertility video visits were completed by 56 men. The median age of patients using video visits was 36 (range 20 to 56 years of age). Seventy six percent of patients self-identified as white and 96% identified their preferred language as English. Most patients were referred by their primary care provider (47%) or by their partner's reproductive endocrinologist (33%). There were a total of 49 unique occupations among the 56 men. 32% were blue collar workers and 68% were white collar workers. Blue collar workers had an estimated median annual income of $28,958 and white collar workers had a median estimated income of $61,240. Total median annual salary of our cohort was $51,331. See Table 1 for additional demographic data.Table 1 Baseline characteristics of established male infertility patients. Table 1No. visits 70 No. patients 56 No. of 1st time video visits 55 Age yrs, med (range) 36 (20-50) Ethnicity no. (%) White 53 (75.7%) Asian 9 (12.9%) Other 3 (4.3%) Black 2 (2.9%) Declined 2 (2.9%) Hispanic 1 (1.4%) Language no. (%) English 67 (95.7%) Albanian 2 (2.9%) Spanish 1 (1.4%) Referral no. % Primary care provider (PCP) 33 (47.1%) Reproductive endocrinology & Infertility (REI) 23 (32.9%) Urologist 8 (11.4%) Self 4 (5.7%) Obstetrician gynecologist 1 (1.4%) Endocrinologist 1 (1.4%) Occupation no. (%) Blue collar 18 (32%) White collar 38 (68%) Diagnostic Categories and Patient Management There was a broad array of male infertility diagnostic categories observed during video visits, including endocrinologic conditions (29%); anatomic causes of infertility (27%); idiopathic infertility (16%); concerns regarding medical treatments on fertility potential (9%); partners being evaluated by REI (9%); genetic abnormalities (7%); and low DNA integrity (3%). Video visit patients received a variety of interventions, as described in Figure 1 . The majority of men (73%) reviewed their test results with their provider during the video visits. Men also received counseling about ART (30%), changes in medications (25%), as well as counseling and indications for sperm extraction procedures (14%) and varicocelectomies (13%).Figure 1 Patient management completed during video visits. (Color version available online.) Figure 1 In the 90 days after video visits, there were only two in-person encounters (3%) within our health system, both of which were planned post-operative visits after varicocelectomy. Counseling regarding the impact of varicoceles on fertility as well as the risks, benefits and indications for surgery took place via a telehealth encounter once all infertility testing was completed. The remainder of the video visits did not result in additional in-person encounters. No patients required an unplanned office or emergency department visit in the three-month period after their telehealth follow-up. Patient Time and Cost Savings Video visits saved patients a median of 80 miles (interquartile range [IQR] 46-244) and 97 minutes (IQR 64-250) of round-trip travel time per visit. Patients travel time would have been between less than 1 hour for 21% of patients and greater than 9 hours for 3% of patients (Table 2 ). Median cost savings per patient from avoiding transportation to and from an in-person appointment was $47 (IQR $27-144).Table 2 Driving time avoided through the use of video visits. Table 2Round-trip Driving Times N (% of all patients) <1 hours 14 (21%) 1-3 hours 31 (44%) 3-5 hours 17 (24%) 5-7 hours 3 (4%) 7-9 hours 3 (4%) 9+ 2 (3%) When estimating lost wages, blue collar workers avoided a median loss of $58 (half day off) to $115 (full day off), and white collar workers avoided a median loss of $122 (half day off) to $244 (full day off), by not taking time off from work. Overall, patients across all occupations avoided a median loss of $102 (IQR $69 – 133) to $205 (IQR $137 – 266) in lost wages by not having to take a half or full day off from work, respectively (Table 3 ). In sum, we found that total potential cost avoidance per patient ranged from a median of $149 (IQR $96 – 277, half day off) to $252 (IQR $164-410, full day off).Table 3 Cost saving estimates from using video visits for follow-up. Table 3 Blue Collar White Collar All Patients Driving Cost (miles x $0.59) Median (IQR) Miles Driven 162 (94-280) 82 (60-250) 80 (46-244) Cost/Mile ($) 0.59 0.59 0.59 Total ($) 96 (55-165) 48 (35-148) 47 (27-144) Wages Saved ($) Median (IQR) Half Day 58 (56-77) 122 (94-145) 102 (69-133) Full Day 115(112-154) 244 (187-290) 205 (137-266) Total Savings ($) Median (IQR) Half Day 154 (111- 242) 168 (129-293) 149 (96-277) Full Day 211 (167-319) 292 (222-438) 252 (164-410) IQR, interquartile range. DISCUSSION In this pre-COVID case series, video visits were used to provide care for patients who had an initial in-person evaluation and were found to have a variety of different conditions impacting male infertility. From these visits, patients were able to review results, undergo medication changes and be counseled on a number of interventions for managing male infertility. We found that there were no unplanned clinic or emergency department visits 90 days after a video visit. Furthermore, these virtual encounters eliminated driving time and travel-related costs as well potentially preventing lost wages by reducing time off work. Collectively, these findings highlight that infertility video visits can serve as practical substitutes for in-person care for established patients. This is the first study to explore the application of telehealth for delivering male infertility care. Berg et al. described their institutional experience with telehealth use for male and female infertility care during the COVID-19 pandemic.24 Our findings corroborate their real-world experience and suggest that male patients with a variety of diagnoses can be provided telehealth follow-up care for male infertility. Importantly, we found that video visits seem to be used as substitutes for in-person care rather than as additive visits. The two patients who saw their urologist in the 90 days after a virtual encounter had surgery in the interim and opted for an in-person, post-operative visit. No patients required an unexpected or unplanned evaluation within 90 days after a video visit. The largest case-control study of telehealth use in urology compared 600 virtual visits to 600 clinic visits and found that less than 1% of patients required an unplanned, in-person evaluation in the 30 days after either an in-person or virtual appointment.25 Our study extends these results to the male infertility patient population over a longer period of follow-up when a physical exam was performed at the initial evaluation for infertility. In our study, patients seeking male infertility care avoided a median of 80 miles and 97 minutes of round-trip travel. Our results are consistent with previous publications regarding general adult and pediatric urology patients that estimated that video visits saved 82-95 miles of round-trip travel and 95-113 minutes of travel time.15 , 16 , 25 The potential to eliminate travel time is especially pertinent to the field of male infertility where significant geographic barriers to healthcare access exist. Twenty-nine states have five or fewer assisted reproductive technology (ART) centers, and 13 states do not have a single male reproductive urologist.14 However, traveling for care does not only burden patients who live far from infertility specialists. Urology patients in metropolitan cities report reduced travel burdens with the use of telehealth visits.26 Future work should evaluate whether the use of telehealth is leading to increased coordination with local providers and infertility clinics to minimize the burden of testing and whether at-home semen analysis kits27 are a reliable option for patients who opt to use video visits. Within the urologic literature, telehealth studies have estimated cost savings for patients from $48 to $150 by avoiding traveling for an appointment.10 , 12 Viers et al. found in their randomized study of post-operative prostatectomy visits that patients who had in-person follow-up reported having to miss a day of work, compared to no days missed by patients seen via video visit.15 Our study builds upon this previous work by estimating the scope of lost wages and is the first to estimate lost wages using patients’ occupations. After accounting for driving costs and lost wages, our patient cohort potentially avoided a median of $149 (half day off) to $252 (full day off) in costs by connecting with their provider through a video visit. Our calculated cost savings may underestimate financial benefits for patients since we could not calculate costs of parking, meals, childcare, lodging or other expenses incurred by travelling for an in-person visit. On the other hand, our calculated savings could overestimate the benefits seen during the COVID-19 pandemic. As more people are working from home, it may be easier for patients to attend doctors’ appointments without formally having to request time off work. Regardless, infertility care is already expensive28 so minimizing the financial impact of these appointments will be meaningful for patients. Future studies should examine real-world, patient reported cost savings to further understand the financial impact of telehealth. Our study has several limitations. First, this was a single institution study in an outpatient setting in the state of Michigan. These results are not generalizable to inpatient or emergency urological care, or to outpatient urology clinics in other states or countries where reimbursement policies may differ. Second, we evaluated established patients who were offered follow-up with video visits. We did not determine how many patients opted for an initial in-person visit over a telehealth encounter. Third, we did not measure patient satisfaction associated with our video visits. However, previous studies within the field of urology have shown that video visits have higher or equivalent levels of patient satisfaction rates as compared to in-person visits.15, 16, 17, 18, 19 , 29 Finally, this was a retrospective case series without a comparison group. This study was intended to be an initial descriptive analysis of how telehealth is being used to provide established infertility patients with follow-up care. These limitations notwithstanding, our finding have important implications for providers, patients, payors, and policymakers. Providers should be reassured that a broad spectrum of male infertility diagnostic categories can be followed-up using video visits without additional in-person evaluations. Our current study is possible because patients underwent an initial visit with an andrologist where a genitourinary exam was performed. The importance of the scrotal exam in male infertility remains paramount and how this can be integrated into virtual care models remains to be seen. For patients, this work highlights that video visits can reduce time spent driving to a clinic and can help avoid additional cost in seeking infertility care. For payors, telehealth for male infertility patients does not result in excess or inappropriate care as evidenced by the lack of unplanned clinic or emergency department visits in our health system within 90 days after video visits. Finally, for policymakers, this work can support advocacy efforts to ensure continued coverage and reimbursement of video visits. Given the changes in telehealth policy that have taken place due to the COVID-19 public health emergency, future research should help clarify how new patient video visit evaluations could impact access to male infertility specialists. Additionally, other forms of telehealth must be evaluated and reimbursed to ensure that patients have choices in how they receive infertility care. Early studies evaluating telemedicine use during the COVID-19 pandemic have found that age, race, ethnicity, and socioeconomic status impact whether patients use video or telephone visits to receive care.30 Relying on video visits, which require broadband internet and expensive hardware, as the only reimbursed form of telehealth could exacerbate health disparities. CONCLUSION Video visits for established male infertility patients were used to manage different causes of infertility while saving patients time and money. Telehealth for established patients did not trigger additional clinic or emergency evaluations and served as substitutes for in-person care. Appendix SUPPLEMENTARY MATERIALS Image, application 1 Image, application 2 Funding Support:1K08 HS027632-01grant support from the10.13039/100000133 Agency for Healthcare Research and Quality . Supplementary material associated with this article can be found in the online version at https://doi.org/10.1016/j.urology.2021.03.050. ==== Refs References 1 Coombs B. Telehealth visits could top 1 billion in 2020 amid the coronavirus crisis. Available at: https://www.cnbc.com/2020/04/03/telehealth-visits-could-top-1-billion-in-2020-amid-the-coronavirus-crisis.html. Accessed May 7, 2020. 2 Badalato GM Kaag M Lee R Vora A Burnett A AUA Telemedicine Workgroup the Role of Telemedicine in Urology: Contemporary Practice Patterns and Future Directions Urol Pract. March 7 2020 53 57 https://doi.org/10.1097/ UPJ.0000000000000094 3 Mehrotra A Ray K Brockmeyer DM Barnett ML Bender JA. Rapidly converting to “ virtual practices ”: outpatient care in the era of COVID 19 NEJM Catal 2020 10.1056/CAT.20.0091 4 IHPI Telehealth Division Caring at a distance: Telehealth and the COVID-19 pandemic IHPI News 2020 https://ihpi.umich.edu/news/caring-distance-telehealth-and-covid-19-pandemic Accessed June 5, 2020 5 Adam J. Gadzinski, Chad Ellimoottil, Anobel Y. Odisho, Kara L. Watts, John L. Gore. Telemedicine in urology: a crash course during the COVID-19 pandemic. Available at: https://www.urologytimes.com/coronavirus/telemedicine-urology-crash-course-during-covid-19-pandemic. 6 CMS Physicians and Other Clinicians : CMS Flexibilities to Fight COVID-19 2020 https://www.cms.gov/files/document/covid-19-physicians-and-practitioners.pdf 7 The White House Proclamation on Declaring a National Emergency Concerning the Novel Coronavirus Disease (COVID-19) Outbreak 2020 White House Proclamations Washington D.C., United States https://www.whitehouse.gov/presidential-actions/proclamation-declaring-national-emergency-concerning-novel-coronavirus-disease-covid-19-outbreak/ Accessed April 5, 2020 8 Center for Connected Health Policy. Covid-19 related state actions. Available at: https://www.cchpca.org/resources/covid-19-related-state-actions. 2020. Accessed April 5, 2020. 9 ASRM. Patient management and clinical recommendations during the coronavirus (COVID-19) pandemic. (2020): 356-57, https://doi.org/10.1097/01.ogx.0000666100.94243.bc. 10 ASRM. SMRU statement regarding male reproductive health and COVID-19, Accessed December 17, 2020, https://www.asrm.org/news-and-publications/covid-19/statements/smru-statement-regarding-male-reproductive-health-and-covid-19/. 11 Renewal of determination that a public health emergency exists,” Accessed January 24, 2020 https://www.phe.gov/emergency/news/healthactions/phe/Pages/covid19-07Jan2021.aspx. 12 American Society for Reproductive Medicine. (2020). Diagnosis and treatment of infertility in men: AUA/ASRM guideline [PDF]. https://www.asrm.org/globalassets/asrm/asrm-content/news-and-publications/practice-guidelines/for-non-members/diagnosis-and-treatment-of-infertility-in-men-aua-asrm.pdf 13 Fertility Problems: Assessment and Treatment 2017 National Institute for Health and Care Excellence (UK) London 14 Mehta A Nangia AK Dupree JM Smith JF. Limitations and barriers in access to care for male factor infertility Fertil Steril 105 2016 1128 1137 10.1016/j.fertnstert.2016.03.023 27054307 15 Viers BR Lightner DJ Rivera ME Efficiency, satisfaction, and costs for remote video visits following radical prostatectomy : a randomized controlled trial Eur Urol 68 2015 729 735 10.1016/j.eururo.2015.04.002 25900782 16 Finkelstein JB Cahill D Kurtz MP The use of telemedicine for the postoperative urological care of children: results of a pilot program J Urol 202 2019 159 163 10.1097/JU.0000000000000109 30707132 17 Andino JJ Castaneda PR Shah PK Ellimoottil C. The Impact of video visits on measures of clinical efficiecy and reimbursement Urol Pract 2020 10.1097/UPJ.0000000000000149 18 Chu S Boxer R Madison P Urologic care to remote clinics Urology 86 2015 255 261 10.1016/j.urology.2015.04.038 26168998 19 Thelen-Perry S Ved R Ellimoottil C. Evaluating the patient experience with urological video visits at an academic medical center mHealth 4 2018 10.21037/mhealth.2018.11.02 54-54 20 Centers for Medicare & Medicaid Services Medicare Telemedicine Health Care Provider Fact Sheet 2020 https://www.cms.gov/newsroom/fact-sheets/medicare-telemedicine-health-care-provider-fact-sheet Accessed April 7, 2020 21 Gadzinski AJ Ellimoottil C Odisho AY Watts KL Gore JL. Implementing telemedicine in response to the 2020 COVID-19 pandemic J Urol 203 2020 10.1097/JU.0000000000001033 22 AAA's Your Driving Costs | AAA Exchange, Accessed October 21, 2020, Available at: https://exchange.aaa.com/automotive/driving-costs/#.X5CnVNVKhhE. 23 ISCO - International Standard Classification of Occupations, Accessed November 27, 2020, Available at: https://www.ilo.org/public/english/bureau/stat/isco/isco08/. 24 Berg WT Goldstein M Melnick AP Rosenwaks Z. Clinical implications of telemedicine for providers and patients Fertil Steril 114 2020 1129 1134 10.1016/j.fertnstert.2020.10.048 33280717 25 Andino JJ Lingaya M-A Daignault-Newton S Shah PK Ellimoottil C. Video visits as a substitute for urological clinic visits Urology 2020 10.1016/j.urology.2020.05.080 26 Gettman M Rhee E Spitz A. Telemedicine in urology AUA White Pap 1 2016 3081 http://www.auanet.org/guidelines/telemedicine-in-urology 27 Samplaski MK Falk O Honig S Shin D Matthews W Smith JF. Development and validation of a novel mail-in semen analysis system and the correlation between one hour and delayed semen analysis testing Fertil Steril 2021 922 929 10.1016/j.fertnstert.2020.10.047 115 28 Dupree JM. Insurance coverage of male infertility: what should the standard be? Transl Androl Urol 7 2018 S310 S316 10.21037/tau.2018.04.25 30159237 29 Safir IJ Gabale S David SA Implementation of a tele-urology program for outpatient hematuria referrals: initial results and patient satisfaction Urology 97 2016 33 39 10.1016/j.urology.2016.04.066 27450940 30 Eberly LA Kallan MJ Julien HM Patient characteristics associated with telemedicine access for primary and specialty ambulatory care during the COVID-19 pandemic JAMA Netw Open 3 2020 e2031640 10.1001/jamanetworkopen.2020.31640

34022261

PMC9752825

NO-CC CODE

2022-12-16 23:25:25

no

Urology. 2021 Aug 19; 154:158-163

utf-8

Urology

10.1016/j.urology.2021.03.050

oa_other

==== Front Sci Total Environ Sci Total Environ The Science of the Total Environment 0048-9697 1879-1026 Elsevier B.V. S0048-9697(21)07253-3 10.1016/j.scitotenv.2021.152177 152177 Article Immunomodulatory activity of β-glucan polysaccharides isolated from different species of mushroom – A potential treatment for inflammatory lung conditions Murphy Emma J. ab1 Rezoagli Emanuele cde⁎1 Pogue Robert af Simonassi-Paiva Bianca a Abidin Ismin Izwani Zainol g Fehrenbach Gustavo Waltzer a O'Neil Emer a Major Ian g Laffey John G. cd Rowan Neil a a Bioscience Research Institute, Athlone Institute of Technology, Athlone, Ireland b Department of Graduate Studies, Limerick Institute of Technology, Limerick, Ireland c Lung Biology Group, Regenerative Medicine Institute at CURAM Centre for Medical Devices, School of Medicine, National University of Ireland Galway, Galway, Ireland d Anaesthesia and Intensive Care Medicine, University Hospital Galway, Galway, Ireland e Department of Medicine and Surgery, University of Milan-Bicocca, Monza, Italy f Post-Graduate Program in Genomic Sciences and Biotechnology, Catholic University of Brasilia, Brazil g Materials Research Institute, Athlone Institute of Technology, Athlone, Ireland ⁎ Corresponding author at: Department of Medicine and Surgery, University of Milan-Bicocca, Monza, Italy. 1 E.J. Murphy and E. Rezoagli equally contributed. 4 12 2021 25 2 2022 4 12 2021 809 152177152177 6 8 2021 29 11 2021 30 11 2021 © 2021 Elsevier B.V. All rights reserved. 2021 Elsevier B.V. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Acute respiratory distress syndrome (ARDS) is the most common form of acute severe hypoxemic respiratory failure in the critically ill with a hospital mortality of 40%. Alveolar inflammation is one of the hallmarks for this disease. β-Glucans are polysaccharides isolated from a variety of natural sources including mushrooms, with documented immune modulating properties. To investigate the immunomodulatory activity of β-glucans and their potential as a treatment for ARDS, we isolated and measured glucan-rich polysaccharides from seven species of mushrooms. We used three models of in-vitro injury in THP-1 macrophages, Peripheral blood mononuclear cells (CD14+) (PMBCs) isolated from healthy volunteers and lung epithelial cell lines. We observed variance between β-glucan content in extracts isolated from seven mushroom species. The extracts with the highest β-glucan content found was Lentinus edodes which contained 70% w/w and Hypsizygus tessellatus which contained 80% w/w with low levels of α-glucan. The extracts had the ability to induce secretion of up to 4000 pg/mL of the inflammatory cytokine IL-6, and up to 5000 pg/mL and 500 pg/mL of the anti-inflammatory cytokines IL-22 and IL-10, respectively, at a concentration of 1 mg/mL in THP-1 macrophages. In the presence of cytokine injury, IL-8 was reduced from 15,000 pg/mL to as low as 10,000 pg/mL in THP-1 macrophages. After insult with LPS, phagocytosis dropped from 70–90% to as low 10% in CD14+ PBMCs. After LPS insult CCL8 relative gene expression was reduced, and IL-10 relative gene expression increased from 50 to 250-fold in THP-1 macrophages. In lung epithelial cells, both A549 and BEAS-2B after IL-1β insult, IL-8 levels dropped from 10,000 pg/mL to as low as 6000 pg/mL. TNF-α levels dropped 10-fold from 100 pg/mL to just below 10 pg/mL. These results demonstrate the therapeutic potential of β-glucans in inflammatory lung conditions. Findings also advance bio-based research that connects green innovation with One Health applications for the betterment of society. Graphical abstract Unlabelled Image Keywords β-Glucans THP-1 macrophages Lung injury ARDS Medicinal mushrooms One-health Editor: Lotfi Aleya ==== Body pmc1 Introduction Acute respiratory distress syndrome (ARDS), is the most common form of acute severe hypoxemic respiratory failure in the critically ill (Rezoagli et al., 2017). The syndrome is defined by: acute onset of hypoxemia (PaO2:FiO2 ratio <300) and bilateral pulmonary opacities not explained by cardiac failure or fluid overload (Bellani et al., 2016). ARDS is a diffuse inflammatory reaction and can be characterised by an explosive acute inflammatory response in lung parenchyma (Crimi and Slutsky, 2004), impairing the principal function of gas exchange, which can lead to hypoxaemia. Treatment is mainly focused on clinical management as there remains no effective direct pharmacological therapy for this condition (Rezoagli et al., 2019). There is an urgent need for treatment as mortality and morbidity are unacceptably high at 40% (Horie et al., 2020). Furthermore, infection by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has led to further incidences of COVID-19-related ARDS, which is associated with 70% of fatal cases (Rezoagli et al., 2021; Li et al., 2020; Zhou et al., 2020). In both ARDS and in COVID-19-induced ARDS there is a marked increase in serum levels of inflammatory cytokines and chemokines, which is a major contributor to disease severity and ultimately death (G. Chen et al., 2020; X. Chen et al., 2020; Huang et al., 2020; Mehta et al., 2020; Qin et al., 2020). The pathophysiology of ARDS is associated with numerous target immune cells including macrophages. The lung microenvironment during injury determines the functional phenotype of macrophages, which can promote either wound healing or inflammation (Nanchal and Truwit, 2018). Thus, in developing potential therapeutics it is important to understand the potential effects of these therapeutics on lung tissue and on macrophages. An ideal treatment for this condition would aim at reducing the effects of the proinflammatory cascade and would seek to maximize the anti-inflammatory immunomodulatory response (Zambelli et al., 2021). β-Glucans are defined as complex polysaccharides that are found in an abundance of sources including fungi, yeast, grain, bacteria, and algae (Murphy et al., 2021). β-Glucans can be classified structurally as either 1,3 1,4-linked or 1,3 1,6-linked, which is dependent on their source (Cui et al., 2011; Pogue et al., 2021; Murphy et al., 2020). These molecules are widely marketed as biologically active molecules (bioactives) (Wang et al., 2017a). There are over 200 clinical trials registered for their use for a range of applications. There are also licenced drugs containing β-glucans on the market since 1980 in Japan, for the treatment of cancer (Novak and Vetvicka, 2008; Takeshita et al., 1991; Yang et al., 2019). β-Glucans as pharmaceutical agents have also been authorised in several countries, including the United States of America, Canada, Finland, Sweden, China and Korea (van Steenwijk et al., 2021). The diverse functional effects of these molecules include alteration of lipid and glucose metabolism, cholesterol reduction, obesity regulation and reduction of cardiovascular and diabetic risk, modulating the gut microbiome, altering lipid and glucose metabolism and beneficial effects on gastrointestinal conditions such as irritable bowel syndrome (Drozdowski et al., 2010; Maki et al., 2003; McRorie and McKeown, 2017; Sima et al., 2018; Tiwari and Cummins, 2011). β-Glucans, specifically from non-cereal sources, are widely documented for their immunomodulatory properties, with the ability to stimulate the immune response and initiate inflammatory properties, and to promote resistance to infections (Ooi and Liu, 2012). (Bohn and BeMiller, 1995). Mushroom-derived β-glucans are the most potent immune modulators (Borchers et al., 1999; Ooi and Liu, 2012; Lorenzen and Anke, 1998; Ooi and Liu, 1999; Tzianabos, 2000; Wasser and Weis, 1999). Moreover, they have demonstrated therapeutic effects in alleviating infective respiratory conditions (Fuller et al., 2012; Jesenak et al., 2013; Yamauchi et al., 2008). They have also been documented to reduce pro-inflammatory cytokines, increase anti-inflammatory cytokines, increased formation of antioxidants as well reduction of inflammatory cells in preclinical lung injury models (Bedirli et al., 2007; Jedinak et al., 2011; Johnson et al., 2009; Kofuji et al., 2012; Soltys and Quinn, 1999; Yamada et al., 2007). These beneficial effects can also be seen in clinical trials. When patients were administered β-glucans for the prevention of nosocomial pneumonia and sepsis, the treatment group compared to the control group had lower incidences of pneumonia as well as a lower mortality rate (De Felippe et al., 1993). We have previously investigated the effects of a commercial β-glucan and an in-house extract of β-glucans from the mushroom Lentinus edodes (Masterson et al., 2020; Murphy et al., 2020, Murphy et al., 2020, Murphy et al., 2020). Specifically, Murphy et al. (2019) showed that β-glucans from the same mushroom, one isolated by hot water extraction and one sourced commercially had different effects, namely reduction in inflammatory cytokines, reduction in phagocytic activity of macrophages after LPS insult and reduction of inflammatory response in in-vitro lung cells. Thus, to continue this work and understand the potential immunomodulatory properties of other mushroom β-glucans as a potential treatment for inflammatory lung conditions like ARDS we decided to replicate the assays and include additional test parameters. In the current study, we first extracted and measured β-glucans from seven species of mushroom to determine BRM variance among species by applying them to a monocytic cell line and an in-vitro lung injury model. Second, we isolated CD14+ monocytes from healthy volunteers and exposed the cells to the extracts, then measured phagocytic activity. Third, we simulated an injurious environment on two types of alveolar cell lines using IL-1β and measured cytokine expression. Finally, we extended this assay to a monocytic cell line, which was inflamed with different insults (LPS and cytomix). It has recently emerged that macrophages are reduced and equally as inflamed as lung cells during COVID-19 infection. Therefore, after injury we measured cytokine release, gene expression, and phagocytosis of these cells to determine immune-modulatory potential in an inflammatory micro-environment. 2 Materials & methods Commercial Lentinan (CLE) was sourced from Carbosynth (FL33321, Compton, Berkshire, UK). Fruiting bodies of mushrooms were kindly gifted by Garryhinch Wood Exotics Ltd. Garryhinch, Portarlington, Co Offaly, Ireland. The fruiting body of Agaricus blazeii was kindly gifted by Professor Leo van Griensven, Wageningen University, The Netherlands. Other species of mushroom included; Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazeii (A.B). 2.1 β-Glucan extraction To extract β-glucans from the fruiting bodies of the mushrooms, the method used previously by Murphy et al. (2019) was employed. Briefly – the fruiting bodies were washed and dried. The samples were blended into a fine powder. Roughly, 100 g of dried blended biomass was placed in 1 Litre of water and autoclaved. After autoclaving the polysaccharides were precipitated from supernatant using 100% Ethanol. Precipitates were dried and solubilised in PBS for analysis. 2.2 β-Glucan quantification Extracts were analysed for 1-3 1-6 β-glucan content using the Megazyme yeast and mushroom kit (K-YBGL) (Megazyme Ltd., Bray, Co. Wicklow, Ireland). Assays were carried out according to manufacturer's instructions. After milling, samples were placed in 12 M H2SO4 at −4 °C for 2 h to solubilize the β-glucans. Samples were then hydrolysed in 2 M H2SO4 at 100 °C for a further 2 h. Any remaining β-glucan fragments were quantitatively hydrolysed to glucose using a mixture of exo-1,3-β-glucanase and β-glucosidase which gives a measurement of total β-glucan content after substrate addition. The α-glucan content of the sample was determined by hydrolysing specifically to d-glucose and d-fructose. Glucose was measured with amyloglucosidase and invertase using a glucose oxidase peroxidase GOPOD reagent. β-Glucan content was determined by the difference between the two measurements. 2.3 Blood donor cell collection Blood sample collection was approved by the Athlone Institute of Technology Ethics Committee. Blood samples were obtained from healthy volunteers for isolation of immune cells. A total of 15 mL was collected from each donor. Individual cells were isolated from 5 mL aliquots of collected blood. Samples were magnetically labelled with whole blood microbeads (Miltenyi Biotec, Germany) to isolate cells based on specific surface molecules according to the manufacturer's instructions, using the autoMACS separator (Miltenyi Biotec). 2.4 Cell culture A549 cells (used at passage 90), BEAS-2B cells (used at passage 10), and THP-1 monocyte cells (used at passage 20), were obtained from the American Type Culture Collection (ATCC, Rockville, MD, USA). Cells were cultured in RPMI-1640 (Sigma-Aldrich, St. Louis, MO, USA), supplemented with 10% fetal calf serum (Sigma-Aldrich), 1% penicillin G (100 U/mL) and streptomycin (100 μg/mL) solution (Sigma-Aldrich), at 37 °C a 5% CO2 environment. For differentiation into macrophages, THP-1 monocyte cells were treated with phorbol 12-myristate 13-acetate (PMA) for differentiation into THP-1 macrophages. (Peprotech EC, London, UK), at a concentration of 100 ng/mL, for 48 h. 2.5 CD14+ PBMCs CD14+ cells were positively isolated based on the surface molecule CD14, which is primarily found on monocytes (Shin et al., 2019). Isolated cells were cultured in RPMI-1640 (Sigma-Aldrich, St. Louis, MO, USA), supplemented with 10% fetal calf serum (Sigma-Aldrich), 1% penicillin G (100 U/mL)/streptomycin (100 μg/mL) solution (Sigma-Aldrich) and 50 ng/mL of macrophage colony stimulating growth factor (MCSGF) (RnD Systems MN, USA) at 37 °C in a 5% CO2 environment. 2.6 Cell injury and β-glucan treatment All cell types were treated with 1 mg/mL of β-glucan in PBS based on other published work (Jung et al., 2007; Murphy et al., 2020, Murphy et al., 2020, Murphy et al., 2020; Sari et al., 2020; Sivinski et al., 2020). For injury assays THP-1 PMA differentiated macrophage cells were seeded at a density of 4 × 105 cells/well in 96-well plates, and 24 h later were injured with two different types of insult: either LPS (100 ng/mL) (Sigma) or cytomix (TNF-α, IFN-Υ, IL-1β), at 25 ng/mL (Immunotools), in RPMI supplemented with 1% penicillin/streptomycin. After 24 h cells were washed three times in PBS and treated with 1 mg/mL of extracts for a further 24 h before analysis. Pulmonary alveolar type II A549 cells were seeded at a density of 4 × 105 cells/well in 96 well plates. After 24 h cells were injured with 1 ng/mL of IL-1β (Peprotech, Rocky Hill, NJ) in RPMI supplemented with 1% penicillin/streptomycin. 2.7 Enzyme linked immunosorbent assay (ELISA) A human Duoset sandwich ELISA kit (RnD Systems MN, USA) was used to measure cytokine levels in the medium after β-glucan exposure. All ELISA assays were performed according to the manufacturer's instructions. Results were expressed either in pg/mL or in ng/mL. 2.8 Phagocytosis assays To determine Phagocytic activity, THP-1 macrophages (PMA differentiated) and CD14+ cells were seeded into 96-well plates at 4 × 105 cells/well. After 24 h, cells were either injured or treated with PBS. After a further 2 h, cells were treated with β-glucan extracts. After a further 24 h cells were washed with PBS and incubated with Alexa Fluor 488-conjugated E.coli (K-12 strain) Bioparticles (E13231; Life Technologies) for 2 h, after which cells were washed three times with PBS to remove residual particles before resuspension in FACS flow buffer and measured for fluorescent particles by flow cytometry (Miltenyi Biotec, Germany). 2.9 RNA extraction For RNA extraction from THP-1 macrophage cells, Media was removed, the cells were washed 3× with PBS, and RNA was extracted using the Purelink RNA Mini kit (Thermo-Fisher), according to the manufacturer's instructions. RNA was analysed using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, DE, USA) to determine RNA concentrations and A260/A280 ratios. 2.10 cDNA synthesis and real-time PCR cDNA was prepared for the replicate samples using the SensiFAST cDNA Synthesis kit (Bioline), according to the manufacturer's instructions. RNA input for all samples was normalized to the sample, so that 325.5 ng of total RNA were used in each reaction. Real-time quantitative PCR was performed using pre-designed TaqMan Gene Expression Assays for the respective genes, together with the TaqMan Gene Expression Master Mix (Thermo-Fisher). The transcripts examined were: TLR2, IL-10, CCL8, CLEC-7a and MCSGF. Reactions were carried out on the LightCycler 96 equipment (Roche), using the GAPDH transcript as endogenous control. Relative gene expression was calculated using the 2^-ddCq method. 2.11 Statistical analysis Continuous data were expressed as mean and standard error of the mean (SEM). Differences of continuous variables between species of mushrooms and PBS and injury (i.e. LPS or Cytomix or IL-1 β) were assessed by one-way analysis of variance for independent measures. Post-hoc comparisons were investigated by controlling the False Discovery Rate using the two-stage step-up method of Benjamini, Krieger and Yekutieli test. Statistical significance was considered with a p-value < 0.05 (two-sided). Statistical analyses were performed using STATA/MP 16.0 for Windows (StataCorp LLC, College Station, TX 77845, USA) and GraphPad Prism 8 for Windows (Version 8.0.2, FraphPad Software, Inc.). 3 Results 3.1 β-Glucans quantification β-glucans were extracted from seven species of mushrooms as previously described (Murphy et al., 2020, Murphy et al., 2020, Murphy et al., 2020). After extraction and isolation, the Megazyme 1,3 1,6 kit was used to determine the concentration of α- and β-glucans. There was variance between species as can be seen in the relative concentrations of β-glucan and α-glucan displayed in Fig. 1 . P.E appeared to have the purest concentration of β-glucan compared to other extracts. A.B and P.O appeared to have high levels of contaminating α-glucan present. H.T yielded the highest concentration of β-glucan with little contaminating α-glucan. H.T yielded the highest concentration of β-glucan with little contaminating α-glucan.Fig. 1 The percentage w/w α-glucan and β-glucan content in mushroom extracts using Megazyme. Commercial Lentinan (C.L.E.), Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazeii (A.B). Fig. 1 3.2 Effect of β-glucans on macrophages 3.2.1 The direct effect of β-glucans on THP-1 macrophages To understand the effects of β-glucan extracts on macrophages, THP-1 macrophages were treated with each of the extracts and cytokine secretion levels were measured by an ELISA. Results show (Fig. 2 ) that each extract had a different effect on the cytokine release profile from THP-1 macrophages. Extracts had the potential to increase secretion of both inflammatory cytokines (IL-6, IL-8, TNF-α) and anti-inflammatory cytokines (IL-10 and IL-22). CLE is a commercial source of β-glucan, and its extraction method is unknown; all other extracts were processed as described in the methods Section 2.1. Although the extracts have different effects on the cytokine secretion profile, the pattern was generally similar except for CLE. CLE induced lower secretion levels of IL-6 (Panel A), IL-22 (Panel C), and IL-10 (Panel E), compared to the other extracts and lower levels of IL-2 (Panel F) compared to the PBS control. The remaining extracts increased IL-6, TNF-α (Panel B) and IL-10 secretion and maintained IL-22 and IL-2 compared to PBS control. CLE and some of the extracts (L.E, C.L.E and P.C) appeared to increase the secretion of the chemokine IL-8 compared to control (Panel D).Fig. 2 The effect of the β-glucan extracts on cytokine expression in THP-1 macrophages (PMA differentiated) measured using ELISA. Panel A; IL-6, Panel B; TNF-α, Panel C; IL-22, Panel D; IL-8, Panel E; IL-10 Panel F; IL-2. p < 0.05 versus PBS. Cells were treated with 1 mg/mL of extracts for 24 h before cytokine analysis. Phosphate buffer saline (PBS), Commercial Lentinan (C.L.E.), Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazeii (A.B). Fig. 2 3.2.2 Effect of β-glucans on phagocytosis by THP-1 and CD14+ PBMC macrophages To compare the effects of β-glucan extracts on a macrophage cell line (THP-1) and on fresh PBMCs (CD14+), cells were treated with extracts, and after 24 h phagocytic activity was measured and displayed in Fig. 3 . Results varied: Panel A shows the effects on THP-1 macrophages; LPS significantly increased phagocytosis relative to untreated cells. Four extracts (P.M, P.C, P. E and H.T) were able to significantly reduce phagocytosis, with PM showing by far the greatest reduction. CD14+ PBMCs (Fig. 3, Panel B) showed varying responses to the extracts in terms of phagocytosis, as expected due to donor variability. The extract A.B significantly reduced the phagocytic activity, although the overall percentage phagocytosis was low in these cells. The extracts showed a tendency toward reduction in PBMCs, however, the results do not show significance, potentially due to donor variability.Fig. 3 The effects of the β-glucan extracts on percentage phagocytosis measured using flow cytometry analysis of uptake of Alexa Fluor 488-conjugated E.coli (K-12 strain) Bioparticles. THP-1 macrophages were treated with 1 mg/mL of extracts for 24 h before phagocytic analysis. Panel A; THP-1 macrophages; Panel B; CD14+ primary macrophages. Phosphate buffer saline (PBS), Commercial Lentinan (C.L.E.), Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazeii (A.B). Fig. 3 To gain a further understanding of the mechanism, expression levels of four genes were measured. Two genes were related to cytokine/chemokine response: IL-10 (Fig. 4 ; Panel B), and CCL8 (IL-8; Fig. 4, Panel D). Two genes corresponded to cell surface ligands associated with β-glucan recognition: toll-like receptor 2 TLR-2 (Fig. 4, Panel A), and dectin-1 (CLEC7a, Fig. 4, Panel C).Fig. 4 The effect of the β-glucan extracts on gene expression levels of THP-1 macrophages (PMA differentiated), relative to PBS-treated cells (expression level = 1.0). Panel A; TLR2, Panel B; IL-10, Panel C; CLEC7a, Panel D; CCL8. Differences in relative gene expression *p < 0.05 versus PBS; #p < 0.05 versus LPS. Cells were treated with 1 mg/mL of extracts for 24 h before analysis of gene expression levels. Phosphate buffer saline (PBS), Commercial Lentinan (C.L.E.), Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazeii (A.B). Fig. 4 Expression of TLR2 in THP-1 macrophages showed no significant difference in induction between LPS and the β-glucans. However, compared to PBS control, P.O induced expression of TLR2. All the extracts significantly inhibited the relative gene expression levels of CLEC7a compared to both PBS control and LPS in THP-1 macrophages. IL-10 expression showed no significant increase in gene expression compared to controls except for with P.M. The extracts did not significantly induce the expression of CCL8 compared to PBS controls and induction was significantly lower compared to LPS. To determine the effect of β-glucans on macrophages after injury, THP-1 macrophages were injured with Cytomix (IL-1β, TNF-α & IFN-γ), and then treated with β-glucan extracts (Fig. 5 ). ELISA assay results show the β-glucan extracts from L.E, P.E, H.T and A.B significantly increased the secretion of IL-6 after insult (Panel A). P.O and P.C did not have the same induction profile as when directly treated with injury (Fig. 2 Panel A), which induced ~2000 pg/mL secretory levels of IL-6. After insult with cytomix P.O and P.C induced ~1000 pg/mL secretory levels of IL-6. IL-8 secretion was increased after cytomix treatment alone (Fig. 5, Panel B). However, the strongest inducers of IL-8 secretion with direct treatment were L.E, C.L.E and P.C (Fig. 2 Panel D), all of which significantly reduced the secretion of this inflammatory chemokine after injury except for LE which was not significant. There was no significant effect of β-glucan treatments on TNF-α secretion after cytomix insult (Fig. 5, Panel C).Fig. 5 The effects of the β-glucan extracts on THP-1 macrophages (PMA differentiated) after cytokine insult measured using ELISA. Panel A; IL-6, Panel B; IL-8, Panel C; TNF-α. *p < 0.05 versus cytomix. Cells were treated with cytomix (TNF-α, IFN-Υ, IL-1β), at 25 ng/mL for 24 h, after which they were washed with PBS and treated with extracts (1 mg/mL) for 24 h before cytokine analysis. Commercial Lentinan (C.L.E.), Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazeii (A.B). Fig. 5 Both THP-1 macrophages and PBMCs were analysed for phagocytic activity (Fig. 6 ) after injury with LPS. The THP-1 macrophages after injury (Fig. 6, Panel A) had a very similar response to those with β-glucan extracts alone (Fig. 3 Panel A) compared with treatment after injury. When CD14+ cells were treated with β-glucan extracts, only A.B significantly reduced phagocytosis (Fig. 3 Panel B). However, when administered after LPS, all β-glucan extracts reduced percentage phagocytosis (Fig. 6, panel B). Panel C displays the phagocytosis percentage of THP-1 macrophages after cytomix insult; C.L.E, P.M, P.O, P.C and P.E reduced the phagocytic activity after insult and treatment.Fig. 6 The effects of the β-glucan extracts on THP-1 macrophages (PMA differentiated) after cytokine insult or LPS insult. Percentage phagocytosis measured using flow cytometry analysis of uptake of Alexa Fluor 488-conjugated E. coli (K-12 strain) Bioparticles. Panel A; THP-1 macrophages (PMA differentiated) treated with LPS, Panel B; CD14+ PBMCs treated with LPS. Panel C; THP-1 macrophages (PMA differentiated) treated with cytomix. *p < 0.05 versus Cytomix or LPS. Cells were treated with cytomix (TNF-α, IFN- Υ, IL-1β), at 25 ng/mL or LPS for 24 h after which they were washed with PBS and treated with extracts (1 mg/mL) for 24 h before phagocytic activity analysis. Commercial Lentinan (C.L.E.), Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazeii (A.B). Fig. 6 Inflammatory gene expression and anti-inflammatory gene expression markers were analysed after both types of injury, as displayed in Fig. 7 . The inflammatory marker CCL8 was measured after LPS injury (Panel A) and cytomix (Panel B). In the presence of LPS, CCL8 relative gene expression was significantly reduced after treatment with all extracts. In the presence of cytomix insult, L.E, P.O, P.C and H.T all increased the relative gene expression of CCL8. The anti-inflammatory marker IL-10 was measured after LPS injury (Panel C) and cytomix (Panel D). After LPS insult L.E, H.T and A.B all significantly increased the expression of IL-10 gene compared to injury alone. All the extracts increased the expression of IL-10 compared to injury alone.Fig. 7 The effects of the β-glucan extracts on THP-1 macrophages (PMA differentiated) after LPS (Panels A, C) or Cytomix (Panels B, D) injury relative to PBS-treated cells (expression level = 1.0). Panels A and B show IL-8 expression after β-glucan treatments, relative to insult; Panels C and D show IL-10 expression after β-glucan treatments, relative to insult. *p < 0.05 versus LPS /Cytomix. THP-1 macrophages (PMA differentiated) were treated with cytomix (TNF-α, IFN- Υ, IL-1β), at 25 ng/mL or LPS for 24 h after which they were washed with PBS and treated with extracts (1 mg/mL) for 24 h before gene expression analysis. Commercial Lentinan (C.L.E.), Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazei (A.B). Fig. 7 3.3 Effect of β-glucans in an in-vitro lung injury model In previous work carried out by this group, we established that β-glucan extracts (L.E) have the potential to reduce inflammation in alveolar A549 cell lines (Murphy et al., 2019). To further expand on this work, the same assays were repeated with another alveolar cell line BEAS-2B, using six new extracts. 3.3.1 A549 cells To determine the direct effects of the β-glucan extracts on lung cells, extracts were incubated with A549 cells for 24 h and supernatant was measured for cytokines. Results are displayed in Fig. 8 . Panel A shows the release of IL-6 after treatment; P.M, P.A, P.E, H.T and A.B all significantly induce the secretion of IL-6 compared to PBS control. Panel B shows the secretion of IL-8. All the extracts induced the secretion of IL-8 from A549 cells with respect to PBS control except for P.E. When A549 cells were treated with the extracts, TNF-α secretion was increased with L.E, P.O, P.E, H.T and A.B with respect to PBS control (Panel C).Fig. 8 The effect of the β-glucan extracts on cytokine expression in A549 cells measured using ELISA. Panel A; IL-6, Panel B; IL-8, Panel C; TNF-α. The effect of the β-glucan extracts on cytokine expression in BEAS-2B cells after IL-1β insult, measured using ELISA. Panel D; IL-6, Panel E; IL-8, Panel F; TNF-α. p < 0.05 versus PBS or IL-1β. Uninjured cells were treated with extracts (1 mg/mL) for 24 h before cytokine analysis. Cells were treated with IL-1β at 1 ng/mL for 24 h after which they were washed with PBS and treated with extracts (1 mg/mL) for 24 h before cytokine analysis. Phosphate buffer saline (PBS), Commercial Lentinan (C.L.E.), Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazei (A.B). Fig. 8 A549 cells were then treated with IL-1β to induce a cytokine injury. Subsequently the cells were treated with the β-glucans extracts and the cytokine analysis was repeated, as displayed in Fig. 8. Panel D shows that C.LE., P.M and P.A slightly reduced IL-6 secretion (though not significantly), after insult except for L.E which increases secretion with respect to injury alone. Panel E shows that after insult L.E, C.L.E, P.O, P.E and A.B reduce the section of IL-8, which is an opposite response to when cells are treated in the absence of injury (Fig. 8; Panel B). The extracts P.E, H.T and A.B significantly reduced TNF-α secretion after IL-1β insult (Panel F) which is again an opposite response to when the cells are treated alone with extracts (Panel C). 3.3.2 BEAS-2B To understand if the β-glucan extracts would have a similar effect in another lung epithelial cell line, BEAS-2B cells were treated with the β-glucans extracts. The supernatant was then measured for cytokine secretion. Results are displayed in Fig. 9 . All extracts induced the secretion of IL-6 (Panel A), and IL-8 (Panel B) with respect to PBS control. Panel C shows the secretion of TNF-α after treatment, C.L.E increased secretion but the other extracts had no effect.Fig. 9 The effect of the β-glucan extracts on cytokine expression in BEAS-2B cells measured using ELISA. Panel A; IL-6, Panel B; IL-8, Panel C; TNF-α. The effect of the β-glucan extracts on cytokine expression in BEAS-2B cells after IL-1β insult, measured using ELISA. Panel D; IL-6, Panel E; IL-8, Panel F; TNF-α. p < 0.05 versus PBS or IL-1β. Uninjured cells were treated with extracts (1 mg/mL) for 24 h before cytokine analysis. Cells were treated with IL-1β at 1 ng/mL for 24 h after which they were washed with PBS and treated with extracts (1 mg/mL) for 24 h before cytokine analysis. Phosphate buffer saline (PBS), Commercial Lentinan (C.L.E.), Lentinus edodes (L.E), Pholiota microspora (P.M), Pleurotus ostreatus (P.O), Pleurotus citrinopileatus (P.C), Pleurotus eryngii (P.E), Hypsizygus tessellatus (H.T) and Agaricus blazei (A.B). Fig. 9 Like the A549 cells, BEAS-2B cells were then treated with IL-1β to induce a cytokine injury; cells were then treated with the β-glucans extracts and the cytokine analysis was repeated, as displayed in Fig. 9. There was no effect on IL-6 secretion (Panel D). Panel E shows that after insult C.L.E, P.A, P.E, H.T and A.B reduced the section of IL-8 which is an opposite response to when cells are treated in the absence of injury (Fig. 9; Panel B). The extracts P.E, H.T and A.B significantly reduced TNF-α secretion after IL-1β insult (Fig. 9; Panel F) however these levels were the same in the absence of injury (Fig. 9; Panel C) suggesting expression levels are maintained in the presence of injury. 4 Discussion ARDS-associated lung injury develops a state which is marked by an increase in serum levels of inflammatory chemokines and cytokines; this is a major contributor to disease severity and ultimately death (G. Chen et al., 2020; X. Chen et al., 2020; Huang et al., 2020; Mehta et al., 2020; Qin et al., 2020). There is currently a vast literature relating to the immunomodulatory effects of β-glucan (Rao et al., 2020), and there is a huge level of enthusiasm regarding their therapeutic potential. Thus with this in mind, there are three aims to this work. Firstly, to understand if β-glucans extracted in the same way from different species of mushroom contained the same levels of β-glucan content and if these extracts had the same effect on a key player in cellular immunity and response – macrophages. The second aim was to determine if the samples could elicit a response or prime immune cells, and whether there would be a difference in priming effects. Our final aim was to determine whether there may be the potential for the extracts to be used in hyperinflammatory conditions such as ARDS. To understand this, two types of in-vitro models were used – injured macrophages and lung injury models. This approach will potentially facilitate a greater understanding into the biological variance of these compounds and realising their therapeutic potential. It is recognised in the literature that β-glucans from different sources can exert different biological effects, and that different extraction methods may help to optimize performance. Our experiments were performed using the same extraction method on all seven mushroom species. As such, we can compare the immunomodulatory effects of the β-glucan extracts by the same extraction procedure. Further studies may examine the effects of altering extraction parameters on β-glucan activity. Furthermore, future structural analyses would allow us to deepen the structure-function relationship. Further studies may examine the effects of altering extraction parameters on β-glucan activity. Furthermore, future structural analyses would allow us to deepen the structure-function relationship. The diverse mechanisms of action of β-glucans is unknown. There are differences in the effects of β-glucans that can be observed between similar preparations from the same species or source. The cellular pathways that are activated after recognition are also not fully understood. β-Glucans appear to be recognised as pathogen associated molecular patterns (PAMPs) and modulate immune function via this pathway (Brown and Gordon, 2005; Borchers et al., 1999). However, the exact mechanism by which β-glucans suppress inflammatory cytokines and induce anti-inflammatory cytokines are complex, and incompletely understood. With this in mind, previous work by this group, Murphy et al., 2019 investigated the differential effects of two β-glucan extracts in an in-vitro lung injury model and in an in-vivo model of pulmonary sepsis (Masterson et al., 2020). Once, determined fungal β-glucans had immune-modulatory effects in lung injury pre-clinical models, the next advancement is to highlight other potential fungal derived β-glucans with immune-modulatory activity. Once identified, future studies will investigate the structure-activity relationship to gain an understanding of how these molecules elicit their effects and the pathways associated with these effects. The Key findings of this work include - There is a variance in the levels of β-glucan between mushroom species extracted in the same way. This study found that Lentinus edodes and Hypsizygus tessellatus had the highest levels of β-glucan content when measured using the Megazyme assay. Most extracts had the ability to induce both pro and anti-inflammatory cytokines individually at a concentration of 1 mg/mL in THP-1 macrophages. In the presence of a paracrine insult of a cocktail of cytokines; IL-8 was reduced in THP-1 macrophages. Also observed was a reduction in phagocytosis in THP-1 macrophages and CD14+ macrophages in the presence and absence of injury. After LPS insult, CCL8 relative gene expression was reduced, and IL-10 gene expression was increased in THP-1 macrophages. In lung epithelial cells, the extracts had the ability to reduce two cytokines (IL-8 and TNF-α) which are heavily correlated to pathogenesis of inflammation in the presence of IL-1β. 4.1 β-Glucans quantification in mushroom species Hot water extracts were prepared from seven species of mushroom, and β-glucan content was determined. Results show that although extracts were isolated by the same method, each species yielded different levels of α- and β-glucans. Although there is some evidence to suggest that α-glucans can have immune-modulating properties (Masuda et al., 2017; Okamoto et al., 2007). There is substantially more evidence to suggest that the β-glucan molecule is the immune-stimulating compound found in mushrooms. These results highlight the variability between β-glucan contents in the different mushroom species. Two other studies using the same analysis procedure found variance among mushroom species (McCleary and Draga, 2016; Sari et al., 2017). Other studies have found that α- and starch glucans are usually of low abundance in cultivated mushrooms (Bak et al., 2014; Sari et al., 2017; Synytsya et al., 2008). 4.2 Effects of β-glucans on macrophages Macrophages have the potential to intensify inflammation or exhibit regulatory repair activity during injury (Wynn and Barron, 2010). As well as variance in content there is also evidential variance in response, which is most evident in Fig. 2 Panel A, measurement of IL-6. P.M and P.O have similar levels of β-glucan content (Fig. 1), yet P.M induced THP-1 macrophages to produce nearly double the amount of IL-6 in comparison to P.O, according to the ELISA assay. This could be correlated to the higher levels of α-glucan, in the P.O sample or to structural variances between β-glucans from different species. However, the high amount of α-glucan present in A.B sample does not hinder its activity in stimulating IL-6 secretion. Variance can also be seen in phagocytic activity (Fig. 3, Panel A), where some samples (P.M, P.C and H.T) reduced phagocytosis in THP-1 macrophages. A.B reduced phagocytic activity in the donor PBMCs. Other extracts had no effect on phagocytic index. The THP-1 macrophages were differentiated using PMA, and the PBMCs were differentiated using MCSGF. Thus, as they should have a high phagocytic potential in this assay, it is interesting that some of the samples appeared to reduce this. This ability is potentially useful especially in conditions where macrophages are hypersensitive, and phagocytosis is uncontrolled. Previous research has also shown that varied sources and structures lead to a varied biological response (Bohn and BeMiller, 1995; Bose et al., 2014; Demleitner et al., 1992; Driscoll et al., 2009; Goodridge et al., 2009; Volman et al., 2008; Wang et al., 2017b). As such, our results are in agreement with the literature in that β-glucan from different mushroom sources can induce varied responses. Further investigation into these correlations may identify optimized β-glucan sources for treatment of different pathological conditions. Dectin-1 is a type II membrane receptor, which is documented as one of the principal receptors for β-glucans (Baert et al., 2015). TLR 2, 4 and 6 co-bind to dectin-1 after β-glucan recognition (Guo et al., 2015), modulating and contributing to cell responses including the release of pro and anti-inflammatory cytokines and phagocytic activity (Kanjan et al., 2017). The results of the present study showed a low- to absent expression for the gene dectin-1 receptor (CLEC7a). This could be for two reasons; a limitation of this study was that the samples were taken at 24 h when the gene could be (temporarily) switched off. Secondly, dectin-1 does not recognise all β-glucans equally; studies have shown that dectin-1 reacts differently based on structural determinants such as side-branching and size of the molecule (Adams et al., 2008). No gene expression could also be correlated to inhibition of CLEC7a, which could be correlated to a negative feedback effect. This result warrants a further timeline study to understand this mechanism. Nonetheless these results demonstrate that the β-glucan samples are recognised by macrophages of a cell line lineage and from fresh PBMCs. This recognition can induce the secretion of both pro- and anti-inflammatory cytokines, reduce phagocytic activity, and alter gene expression levels reducing pro-inflammatory chemokines and increasing the secretion of the anti-inflammatory marker IL-10. Extracts increased secretion of both inflammatory cytokines (IL-6, IL-8, TNF-α) and anti-inflammatory cytokines (IL- 10 and IL-22). M1 macrophage polarization is associated with the secretion of pro-inflammatory cytokines: IL-1β, IL-6, and TNF-α (Bouhlel et al., 2007). M2 macrophage polarization is associated with the secretion of anti-inflammatory cytokines IL-10 (Arora et al., 2018; Wang et al., 2014). As the β-glucan extracts induce the secretion of both, it is possible that they stimulate the cells into a mixed population of M1/M2 macrophages. The commercial sample C.L.E had a different effect on the cells. L.E and C.L.E are isolated from the same mushroom species, again showing the great variances between β-glucan samples which can be dependent on cultivation, seasonal variation as well as extraction procedure. Taken together these results show the potential of β-glucans from mushrooms to behave as biological response modifiers. To understand the immunomodulatory effects of β-glucan in an inflammatory M1 phenotype-inducing environment two types of insult were used. Firstly, LPS which stimulates macrophages toward an M1 phenotype (Zheng et al., 2013) and secondly a cocktail of cytokines (cytomix) was used to stimulate an inflammatory environment (Farley et al., 2009). After insult β-glucan samples were added to determine if the effects of the insult could be tempered. After treatment with cytomix, some of the β-glucan extracts increased the secretion of IL-6. However, after insult, some of the extracts (P.O and P.C) induced less secretion of IL-6, compared with β-glucan alone, thus suggesting that the immune response is reduced in the presence of an injuring agent (cytomix) (Fig. 5 Panel a). One interesting finding in this study is the reduction of phagocytosis of PBMCs after LPS insult. All β-glucan extracts reduced the phagocytic index in the presence of LPS to just under half of the activity of positive controls. This result demonstrates the potential of β-glucans to modulate macrophage activity as these cells are from healthy volunteers. There is donor variation in these samples which is to be expected; future studies would investigate this effect in larger groups of healthy volunteers. The β-glucans also reduced phagocytosis after cytomix insult. Impressively, the β-glucan samples reduced IL-8 gene expression levels after LPS injury and increased the gene expression levels of IL-10. This demonstrates an intracellular shift from an inflammatory phenotype to an anti-inflammatory phenotype in the presence of LPS. Although IL-8 was not reduced in the presence of cytomix, IL-10 was increased, again demonstrating a shift to a more anti-inflammatory response. During SARS-CoV-2 macrophages communicate with target cells through chemokines and phagocytic signaling (Qi et al., 2020). Macrophages respond to initial infection as a result of the inflammatory cytokines secreted by type II alveolar cells which include IL-1β, IL-6 and TNF-α (Denney and Ho, 2018). When aiming to reduce the response of macrophages in inflammatory conditions, it is also important to target the alveolar cells at the centre of the injury. 4.3 Effect of β-glucans in an in-vitro lung injury model Cytokines and chemokines have an important role in immunity as well as in immune pathology as a dysregulated response has the potential to cause extensive tissue and organ damage, especially in the lungs (Pedersen and Ho, 2020). As SARS-Cov-2 infection is associated with the production of inflammatory cytokines we investigated the effects β-glucans would have in an inflammatory environment by measuring cytokine production after IL-1β insult on two types of alveolar cell lines; A549 and BEAS-2b. When A549 (Fig. 8) and BEAS-2B cells (Fig. 9) were treated with the β-glucan extracts, all inflammatory cytokines were elevated. However, in the presence of inflammatory insults, some of the inflammatory cytokines were reduced significantly. Fig. 8, Fig. 9 Panel E and F, shows that the extracts had the ability to reduce IL-8 and TNF-α in both A549 cells and BEAS-2B. In reducing the cytokine expression and inflammation of lung tissue the inflammatory process can be avoided and blood gas transfer potentially unaffected or minimally affected. After injury, when the invading pathogen is eliminated, large numbers of inflammatory monocytes and macrophages can be recruited to the distal alveolar space because of chemokine gradients, this can also exceed the total number of resident macrophages (Davies et al., 2013; Galli et al., 2011). As epithelial cells are the main source of anti-viral responses in the first 24–48 h window after infection, this is an important result. Important signals are transmitted to innate immune cells which are translated to adaptive immune responses (Geller and Yan, 2020). By firstly priming these cells with bioactives such as β-glucans to respond to infection, innate cells are recruited, and a memory is created for prevention of a secondary infection. More importantly, if the cells are primed, the hyper inflammatory reaction might not occur as cells are modulated by the β-glucans. As the infection is more lung-centred than multi-organ-centred (McGonagle et al., 2020). In-vitro lung epithelial cells represent a good model to determine potential targets. This study has shown that β-glucan extracts from mushrooms can reduce inflammatory responses in models of in-vitro lung injury. 5 Conclusions There is a growing awareness of therapies directed to modulate the immune response in many pathological contexts. Medicinal mushrooms, which contain the complex β-glucans sugars have been used to treat an array of conditions for centuries including inflammatory conditions. Previously, we have demonstrated that β-glucans from the same mushroom isolated by different methods have differential immune-modulation abilities in an in-vitro model as well as in an in-vivo preclinical model. Following on from this, the current work has demonstrated the potential of β-glucans as immunomodulators with dual functions, firstly as immune priming agents that may bolster the capacity of the body to maintain homeostasis in the face of infectious and other challenges, and secondly to temper the immune response following infection, thus helping to avoid the serious sequelae associated with immune hyper-inflammatory response in immune and epithelial cells in inflammatory lung conditions such as ARDS. Future work will investigate relationship between the structure of β-glucans and mechanistic effects at cell and molecular levels. The main findings of this research also strongly align with emergence of green innovation for OneHealth applications (Rowan and Galanakis, 2020). CRediT authorship contribution statement Emma Murphy: conceptualization, methodology, data curation, visualization, formal analysis, writing original draft, review and editing. Emanuele Rezoagli: conceptualization, methodology, data curation, visualization, formal analysis, writing original draft, review and editing. Robert Pogue: methodology, data curation, formal analysis, review and editing. Bianca Simonassi-Paiva: methodology, formal analysis, review and editing. Ismin Izwani Zainol Abidin: methodology, formal analysis, review and editing. Gustavo Waltzer Fehrenbach: methodology, formal analysis, review and editing. Emer O'Neil: methodology. John Laffey: supervision, funding acquisition, conceptualization, review and editing. Ian Major: methodology, supervision, review and editing. Neil Rowan: funding acquisition, project administration, conceptualization, supervision, methodology, review & editing. Declaration of competing interest The authors declare no conflict of interest. ==== Refs References Adams E.L. Differential high-affinity interaction of Dectin-1 with natural or synthetic glucans is dependent upon primary structure and is influenced by polymer chain length and side-chain branching J. Pharmacol. Exp. Ther. 325 1 2008 115 123 10.1124/jpet.107.133124 18171906 Arora S. Macrophages: their role, activation and polarization in pulmonary diseases Immunobiology 2018 Elsevier GmbH 383 396 10.1016/j.imbio.2017.11.001 Baert K. Sonck E. Cell type-specific differences in β-glucan recognition and signalling in porcine innate immune cells Dev. Comp. Immunol. 48 1 2015 192 203 10.1016/j.dci.2014.10.005 25453580 Bak W.C. Determination of glucan contents in the fruiting bodies and mycelia of lentinula edodes cultivars Mycobiology 2014 301 304 10.5941/MYCO.2014.42.3.301 25346611 Bedirli A. Beta-glucan attenuates inflammatory cytokine release and prevents acute lung injury in an experimental model of sepsis Shock 27 4 2007 397 401 10.1097/01.shk.0000245030.24235.f1 17414422 Bellani G. The LUNG SAFE study: a presentation of the prevalence of ARDS according to the Berlin Definition! Crit. Care 2016 10.1186/s13054-016-1443-x Bohn J.A. BeMiller J.N. (1→3)-β-d-glucans as biological response modifiers: a review of structure-functional activity relationships Carbohydr. Polym. 28 1 1995 3 14 10.1016/0144-8617(95)00076-3 Borchers A.T. Mushrooms, tumors, and immunity Exp. Biol. Med 1999 281 293 10.3181/00379727-221-44412 Bose N. Differential regulation of oxidative burst by distinct β-glucan- binding receptors and signaling pathways in human peripheral blood mononuclear cells Glycobiology 24 4 2014 379 391 10.1093/glycob/cwu005 24440830 Bouhlel M.A. PPARγ activation primes human monocytes into alternative M2 macrophages with anti-inflammatory properties Cell Metabolism 6 2 2007 137 143 10.1016/j.cmet.2007.06.010 Cell Press 17681149 Brown G.D. Gordon S. Immune recognition of fungal β-glucans Cell. Microbiol. 2005 471 479 10.1111/j.1462-5822.2005.00505.x John Wiley & Sons, Ltd 15760447 Chen G. Clinical and immunological features of severe and moderate coronavirus disease 2019 J. Clin. Investig. 130 5 2020 2620 2629 10.1172/JCI137244 32217835 Chen X. Macrophage polarization and its role in the pathogenesis of acute lung injury/acute respiratory distress syndrome 2020 Inflammation Research. Springer Science and Business Media Deutschland GmbH 883 895 10.1007/s00011-020-01378-2 Crimi E. Slutsky A.S. Inflammation and the acute respiratory distress syndrome Best Pract. Res. Clin. Anaesthesiol. 18 3 2004 477 492 10.1016/j.bpa.2003.12.007 15212340 Cui S.W. Wang Q. Zhang M. β-Glucans RSC Polymer Chemistry Series 1(3) 2011 Multidisciplinary Digital Publishing Institute 319 345 10.3390/encyclopedia1030064 Davies L.C. Tissue-resident macrophages Nature Immunology 2013 986 995 10.1038/ni.2705 24048120 Demleitner S. Kraus J. Franz G. Synthesis and antitumour activity of sulfoalkyl derivatives of curdlan and lichenan Carbohydr. Res. 226 2 1992 247 252 10.1016/0008-6215(92)84072-Z 1617688 De Felippe J. Infection prevention in patients with severe multiple trauma with the immunomodulator beta 1-3 polyglucose (glucan) Surg. Gynecol. Obstet. 177 4 1993 383 388 8211583 Denney L. Ho L.P. The role of respiratory epithelium in host defence against influenza virus infection Biomed. J. 2018 218 233 10.1016/j.bj.2018.08.004 30348265 Driscoll M. Therapeutic potential of various β-glucan sources in conjunction with anti-tumor monoclonal antibody in cancer therapy Cancer Biol. Ther. 8 3 2009 218 225 10.4161/cbt.8.3.7337 19106638 Drozdowski L.A. β-Glucan extracts inhibit the in vitro intestinal uptake of long-chain fatty acids and cholesterol and down-regulate genes involved in lipogenesis and lipid transport in rats J. Nutr. Biochem. 21 8 2010 695 701 10.1016/j.jnutbio.2009.04.003 19716281 Farley K.S. Wang L. Mehta S. Septic pulmonary microvascular endothelial cell injury: role of alveolar macrophage NADPH oxidase 296 3 2009 480 488 10.1152/AJPLUNG.90201.2008 https://doi.org/10.1152/ajplung.90201.2008. American Physiological Society Fuller R. Influence of yeast-derived 1,3/1,6 glucopolysaccharide on circulating cytokines and chemokines with respect to upper respiratory tract infections Nutrition 28 6 2012 665 669 10.1016/j.nut.2011.11.012 22465901 Galli S.J. Borregaard N. Wynn T.A. Phenotypic and functional plasticity of cells of innate immunity: macrophages, mast cells and neutrophils Nat. Immunol. 2011 1035 1044 10.1038/ni.2109 22012443 Geller A. Yan J. Could the induction of trained immunity by β-glucan serve as a defense against COVID-19? Front. Immunol. 11 2020 10.3389/fimmu.2020.01782 Goodridge H.S. Wolf A.J. Underhill D.M. Β-glucan recognition by the innate immune system Immunol. Rev. 230 1 2009 38 50 10.1111/j.1600-065X.2009.00793.x 19594628 Guo Y. Leptospiral lipopolysaccharide stimulates the expression of toll-like receptor 2 and cytokines in pig fibroblasts Anim. Sci. J. 86 2 2015 238 244 10.1111/asj.12254 25039909 Horie S. Emerging pharmacological therapies for ARDS: COVID-19 and beyond Intensive Care Med. 46 12 2020 2265 2283 10.1007/s00134-020-06141-z Springer Science and Business Media Deutschland GmbH 32654006 Huang C. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China Lancet 395 10223 2020 497 506 10.1016/S0140-6736(20)30183-5 31986264 Jedinak A. Anti-inflammatory activity of edible oyster mushroom is mediated through the inhibition of NF-κB and AP-1 signaling Nutr. J. 10 1 2011 10.1186/1475-2891-10-52 Jesenak M. Immunomodulatory effect of pleuran (β-glucan from Pleurotus ostreatus) in children with recurrent respiratory tract infections Int. Immunopharmacol. 15 2 2013 395 399 10.1016/j.intimp.2012.11.020 23261366 Johnson E. Effect of an extract based on the medicinal mushroom agaricus blazei murill on release of cytokines, chemokines and leukocyte growth factors in human blood ex vivo and in vivo Scand. J. Immunol. 69 3 2009 242 250 10.1111/j.1365-3083.2008.02218.x 19281536 Jung H.K. Production and physicochemical characterization of β-glucan produced by Paenibacillus polymyxa JB115 Biotechnol. Bioprocess Eng. 12 6 2007 713 719 Kanjan P. Immune effects of β-glucan are determined by combined effects on Dectin-1, TLR2, 4 and 5 J. Funct. Foods 37 2017 433 440 10.1016/j.jff.2017.07.061 Kofuji K. Antioxidant activity of β -glucan ISRN Pharmaceutics 2012 2012 1 5 10.5402/2012/125864 Li G. Coronavirus infections and immune responses J. Med. Virol. 2020 424 432 10.1002/jmv.25685 Lorenzen K. Anke T. Basidiomycetes as a source for new bioactive natural products Curr. Org. Chem. 2 1998 329 364 Maki K.C. Food products containing free tall oil-based phytosterols and oat β-glucan lower serum total and LDL cholesterol in hypercholesterolemic adults J. Nutr. 133 3 2003 808 813 10.1093/jn/133.3.808 12612157 Masterson C.H. Purified β-glucans from the Shiitake mushroom ameliorates antibiotic-resistant Klebsiella pneumoniae-induced pulmonary sepsis Lett. Appl. Microbiol. 71 4 2020 405 412 10.1111/lam.13358 Blackwell Publishing Ltd 32706908 Masuda Y. Antitumor activity of orally administered maitake α-glucan by stimulating antitumor immune response in murine tumor PLoS ONE 12 3 2017 10.1371/journal.pone.0173621 McCleary B.V. Draga A. Measurement of β-glucan in mushrooms and mycelial products J. AOAC Int. 99 2 2016 364 373 10.5740/jaoacint.15-0289 26957216 McGonagle D. The role of cytokines including Interleukin-6 in COVID-19 induced pneumonia and macrophage activation syndrome-like disease Autoimmun. Rev. 2020 10.1016/j.autrev.2020.102537 McRorie J.W. McKeown N.M. Understanding the physics of functional fibers in the gastrointestinal tract: an evidence-based approach to resolving enduring misconceptions about insoluble and soluble fiber J. Acad. Nutr. Diet. 117 2 2017 251 264 10.1016/j.jand.2016.09.021 27863994 Mehta P. COVID-19: consider cytokine storm syndromes and immunosuppression Lancet 2020 1033 1034 10.1016/S0140-6736(20)30628-0 32192578 Murphy E.J. Immunomodulation properties of a novel β-glucan extract from the mushroom lentinus edodes in an in-vitro lung injury model 2019 10.1164/ajrccm-conference.2019.199.1_meetingabstracts.a2114 pp. A2114–A2114 Murphy Emma J. Masterson C. β-glucan extracts from the same edible shiitake mushroom lentinus edodes produce differential in-vitro immunomodulatory and pulmonary cytoprotective effects — implications for coronavirus disease (COVID-19) immunotherapies Sci. Total Environ. 732 2020 10.1016/j.scitotenv.2020.139330 Murphy Emma J. β-glucan metabolic and immunomodulatory properties and potential for clinical application Journal of fungi (Basel, Switzerland). MDPI AG 6 4 2020 1 33 10.3390/jof6040356 Murphy Emma J. Rezoagli E. Β-glucan metabolic and immunomodulatory properties and potential for clinical application J. Fungi 6 4 2020 1 33 10.3390/jof6040356 Murphy Emma J. β-Glucans Encyclopedia 1 3 2021 831 847 10.3390/encyclopedia1030064 2021 Nanchal R.S. Truwit J.D. Recent advances in understanding and treating acute respiratory distress syndrome [version 1; referees: 2 approved] F1000Research 2018 10.12688/f1000research.15493.1 F1000 Research Ltd Novak M. Vetvicka V. β-glucans, history, and the present: immunomodulatory aspects and mechanisms of action J. Immunotoxicol. 2008 47 57 10.1080/15476910802019045 18382858 Okamoto S. Inflammatory immune responses by water-insoluble α-glucans J. Dent. Res. 86 3 2007 242 248 10.1177/154405910708600309 17314256 Ooi V.E.C. Liu F. A review of pharmacological activities of mushroom polysaccharides Int. J. Med. Mushrooms 1 3 1999 195 206 10.1615/intjmedmushrooms.v1.i3.10 Ooi V.E.C. Liu F. Immunomodulation and anti-cancer activity of polysaccharide-protein complexes Curr. Med. Chem. 7 7 2012 715 729 10.2174/0929867003374705 Pedersen S.F. Ho Y.C. SARS-CoV-2: a storm is raging J. Clin. Investig. 2020 2202 2205 10.1172/JCI137647 32217834 Pogue R. Exploiting immunomodulatory properties of β-glucans derived from natural products for improving health and sustainability in aquaculture-farmed organisms: concise review of existing knowledge, innovation and future opportunities Curr. Opin. Environ. Sci. Health 21 2021 Elsevier BV 100248 10.1016/j.coesh.2021.100248 Qi F. Single cell RNA sequencing of 13 human tissues identify cell types and receptors of human coronaviruses Biochem. Biophys. Res. Commun. 526 1 2020 135 140 10.1016/j.bbrc.2020.03.044 32199615 Qin C. Dysregulation of immune response in patients with coronavirus 2019 (COVID-19) in Wuhan, China Clin. Infect. Dis. 71 15 2020 762 768 10.1093/cid/ciaa248 32161940 Rao K.S. Role of immune dysregulation in increased mortality among a specific subset of COVID-19 patients and immune-enhancement strategies for combatting through nutritional supplements Front. Immunol. 2020 10.3389/fimmu.2020.01548 Rezoagli E. Fumagalli R. Bellani G. Ann. Transl. Med. 5 14 2017 Jul 282 10.21037/atm.2017.06.62 28828357 Rezoagli E. The safety and efficiency of addressing ards using stem cell therapies in clinical trials Stem Cell-Based Therapy for Lung Disease 2019 219 238 10.1007/978-3-030-29403-8_12 Rezoagli E. Development of a critical care response - experiences from Italy during the coronavirus disease 2019 pandemic Anesthesiol. Clin. 39 2 2021 265 284 10.1016/j.anclin.2021.02.003 34024430 Rowan N.J. Galanakis C.M. Unlocking challenges and opportunities presented by COVID-19 pandemic for cross-cutting disruption in agri-food and green deal innovations: Quo Vadis? Sci. Total Environ. 748 2020 10.1016/j.scitotenv.2020.141362 Sari M. Screening of beta-glucan contents in commercially cultivated and wild growing mushrooms Food Chem. 216 2017 45 51 10.1016/j.foodchem.2016.08.010 27596390 Sari M. The birch bracket medicinal mushroom, Fomitopsis betulina (Agaricomycetes)-bioactive source for beta-glucan fraction with tumor cell migration blocking ability Int. J. Med. Mushrooms 22 1 2020 1 13 10.1615/IntJMedMushrooms.2019033291 32463994 Shin J. Il Inflammasomes and autoimmune and rheumatic diseases: a comprehensive review J. Autoimmun. 2019 Sima P. Vannucci L. Vetvicka V. β-glucans and cholesterol (review) Int. J. Mol. Med. 2018 1799 1808 10.3892/ijmm.2018.3411 Spandidos Publications Sivinski S.E. Development of an in vitro macrophage screening system on the immunomodulating effects of feed components J. Anim. Sci. Biotechnol. 11 1 2020 10.1186/s40104-020-00497-4 Soltys J. Quinn M.T. Modulation of endotoxin- and enterotoxin-induced cytokine release by in vivo treatment with β-(1,6)-branched β-(1,3)-glucan Infect. Immun. 67 1 1999 244 252 10.1128/iai.67.1.244-252.1999 9864222 Synytsya A. Mushrooms of genus pleurotus as a source of dietary fibres and glucans for food supplements Czech J. Food Sci. 26 6 2008 441 446 10.17221/1361-cjfs Takeshita K. Diversity of complement activation by lentinan, an antitumor polysaccharide, in gastric cancer patients Nippon Geka Gakkai zasshi 92 1 1991 5 11 2014026 Tiwari U. Cummins E. Meta-analysis of the effect of β-glucan intake on blood cholesterol and glucose levels Nutrition 27 10 2011 1008 1016 10.1016/j.nut.2010.11.006 21470820 Tzianabos A.O. Polysaccharide immunomodulators as therapeutic agents: structural aspects and biologic function Clin. Microbiol. Rev. 2000 523 533 10.1128/CMR.13.4.523-533.2000 11023954 van Steenwijk H.P. Bast A. de Boer A. Immunomodulating effects of fungal beta-glucans: from traditional use to medicine Nutrients 13 4 2021 10.3390/nu13041333 Volman J.J. Ramakers J.D. Plat J. Dietary modulation of immune function by β-glucans Physiol. Behav. 94 2 2008 276 284 10.1016/j.physbeh.2007.11.045 18222501 Wang N. Liang H. Zen K. Molecular mechanisms that influence the macrophage M1-M2 polarization balance Front. Immunol. 2014 614 10.3389/fimmu.2014.00614 Frontiers Media S.A 25506346 Wang Q. Wang F. Xu Z. Ding Z. Bioactive mushroom polysaccharides: a review on monosaccharide composition, biosynthesis and regulation Molecules 22 6 2017 955 28608797 Wang Q. β-Glucans: relationships between modification, conformation and functional activities Molecules 22 2 2017 10.3390/molecules22020257 Wasser S.P. Weis A.L. Medicinal properties of substances occurring in higher basidiomycetes mushrooms: current perspectives (review) Int. J. Med. Mushrooms 1 1 1999 31 62 10.1615/intjmedmushrooms.v1.i1.30 Wynn T.A. Barron L. Macrophages: master regulators of inflammation and fibrosis Seminars in Liver Disease 2010 245 257 10.1055/s-0030-1255354 Yamada J. Alleviation of seasonal allergic symptoms with superfine β-1,3-glucan: a randomized study J. Allergy Clin. Immunol. 119 5 2007 1119 1126 10.1016/j.jaci.2007.02.005 17379290 Yamauchi E. Contribution of lung fibroblast migration in the fibrotic process of airway remodeling in asthma Allergol. Int. 57 1 2008 73 78 10.2332/allergolint.O-06-481 18209507 Yang D. Zhou Z. Zhang L. An overview of fungal glycan-based therapeutics Prog. Mol. Biol. Transl. Sci. 2019 135 163 10.1016/bs.pmbts.2019.02.001 Zheng X.F. Lipopolysaccharide-induced M2 to M1 macrophage transformation for IL-12p70 production is blocked by Candida albicans mediated up-regulation of EBI3 expression PloS one 8 5 2013 10.1371/JOURNAL.PONE.0063967 Zambelli V Hexarelin modulates lung mechanics, inflammation, and fibrosis in acute lung injury Drug Target Insights 2021 10.33393/dti.2021.2347 Zhou F. Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study Lancet 395 10229 2020 1054 1062 10.1016/S0140-6736(20)30566-3 32171076

34875322

PMC9752827

NO-CC CODE

2022-12-16 23:25:25

no

Sci Total Environ. 2022 Feb 25; 809:152177

utf-8

Sci Total Environ

10.1016/j.scitotenv.2021.152177

oa_other

==== Front J Urban Econ J Urban Econ Journal of Urban Economics 0094-1190 0094-1190 Elsevier Inc. S0094-1190(22)00003-1 10.1016/j.jue.2022.103426 103426 Article JUE insight: Demand for transportation and spatial pattern of economic activity during the pandemic☆ Chen Kong-Pin a Yang Jui-Chung b Yang Tzu-Ting ⁎a a Institute of Economics, Academia Sinica, Taiwan b Department of Economics, National Taiwan University, Taiwan ⁎ Corresponding author. 15 1 2022 1 2022 15 1 2022 127 103426103426 4 12 2020 20 12 2021 10 1 2022 © 2022 Elsevier Inc. All rights reserved. 2022 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Using traffic data from Taiwan for 2020, we quantify how the COVID-19 outbreak affected demand for public and private transportation. Despite there being no governmental restrictions, substantial shifts in travel modes were observed. During the peak of the pandemic in Taiwan within the study period (mid-March 2020), railway ridership declined by 40% to 60%, while highway traffic volume increased by 20%. Furthermore, railway ridership was well below pre-pandemic levels, though there were no locally transmitted cases in the eight-month period from mid-April to December. These changes in traffic patterns had implications for spatial patterns of economic activity: retail sales and nighttime luminosity data show that during the pandemic, economic activity shifted away from areas in the vicinity of major railway stations. Keywords COVID-19 Transportation mode Spatial pattern of economic activity ==== Body pmc1 Introduction COVID-19 has upended people’s lives around the world. Many recent studies have shown that human mobility and use of public transport fell dramatically following the onset of the pandemic (Engle, Stromme, Zhou, 2020, Fang, Wang, Yang, 2020, Monte, 2020, Cronin, Evans, 2021, Goolsbee, Syverson, 2021, Liu, Miller, Scheff, 2020, Xin, Shalaby, Feng, Zhao, 2021). However, it is unclear whether the observed changes in mobility and transport mode occurred because of voluntary changes in behaviors or because of enforced measures such as lockdowns and stay-at-home orders. Even without government intervention, rational individuals would still have curtailed their movements or changed the way they traveled in order to reduce their exposure to the virus. Understanding individuals’ efforts in the midst of a pandemic, especially in terms of mobility or mode of transportation, has important policy implications. On the one hand, people’s travel behavior is highly associated with the spread of COVID-19 (Li, Ma, 2021, Mangrum and Niekamp, 2022, Brinkman and Mangum, 2022). On the other hand, government regulations on travel can and could have resulted in huge economic and welfare costs. This raises questions as to whether these mobility restrictions are necessary or excessive. In addition, changes in human mobility are likely to affect commercial activities around transportation nodes during the pandemic. This paper studies transportation modes people used during the pandemic, without governmental restrictions, and its implications for spatial distribution of urban activity. In particular, we examine the effect of the COVID-19 outbreak on demand for public and private transportation in Taiwan. The experience of Taiwan during 2020 offers an ideal setting for this study because, except for a few minor regulations,1 no lockdown policy, stay-at-home order, or restrictions on mobility were imposed. Given this, the response of the general public in terms of mobility can be almost completely attributed to unrestricted choice of the people. We use a difference-in-differences design alongside 2018-2020 traffic administrative data of railway ridership and highway traffic volume to examine whether the utilization of public or private transport during 2020 had changed substantially compared to previous years (2018–2019). We then further investigate how changes in transportation mode affected the spatial pattern of economic activity. There are three key findings of this research. First, the number of railway passengers decreased immediately following the first COVID-19 case announcement. Moreover, during the period (mid-March 2020) when cases peaked in Taiwan during 2020, railway ridership dropped by more than 60% relative to the same weeks in prior years. As a matter of fact, COVID-induced decline in passenger flow persisted through the whole of 2020, despite Taiwan not having experienced any new local virus cases in the eight-month period from mid-April to the end of 2020. Furthermore, we use Google Trends data on COVID-19-related keywords to construct an index measuring the public perceptions of COVID-19 risk in Taiwan.2 Our results suggest that, on average, a 10% increase in the index of public perception of COVID-19 risk, equivalent to one additional coronavirus case, reduced the number of daily passengers by 1.6%. Second, in contrast to public transport, highway traffic flow did not change at the beginning of the COVID-19 outbreak but had increased by 20% when the number of new cases in Taiwan reached its peak during 2020. On average, highway traffic volume increased by 1.2% when the index of public perception of COVID-19 risk increased by 10%. Two effects influenced the demand for private transport during the pandemic. On the one hand, people avoided going out due to fear of contracting the coronavirus, so the demand for both public and private transport declined (the fear effect). On the other hand, individuals substituted private for public transportation when they needed to travel (the substitution effect), as the latter was deemed a far riskier mode of travel than the former. Our results indicate that the substitution effect dominated the fear effect. Finally, changes in transport mode have implications for the spatial pattern of economic activity. Since the pandemic substantially reduced passenger flow at railway stations, it shifted economic activity, measured by retail sales and nighttime luminosity, away from areas close to major railway stations. This paper contributes to three strands of extant literature. First, it complements the fast-growing body of work on impacts of the COVID-19 pandemic on individual mobility (Argente et al., 2022, Engle, Stromme, Zhou, 2020, Fang, Wang, Yang, 2020, Goolsbee, Syverson, 2021, Couture et al., 2022, Glaeser et al., 2022). In particular, we provide one of the first pieces of evidence indicating that individuals substituted private for public transport to reduce the risk of exposure to COVID-19. This finding is also related to the “prevalence response” in the literature on economic epidemiology (Ahituv, Hotz, Philipson, 1996, Gersovitz, Hammer, 2003, Lakdawalla, Sood, Goldman, 2006, Bennett, Chiang, Malani, 2015). Previous works on this issue have shown that people change their health-related behaviors when faced with an increase in disease risk. Our study contributes to this stream of literature by showing that people adjusted their mode of transport to reduce the risk of contracting an infectious disease. Moreover, our results indicate that people took proactive preventive actions even though the risk was very low.3 Second, our results are related to the literature on the relationship between public and private transport (Anderson, 2014, Chen, Whalley, 2012, Parry, Small, 2009, Nelson, Baglino, Harrington, Safirova, Lipman, 2007, Winston, Langer, 2006, Duranton, Turner, 2011). Several studies have shown how the provision of public transport affects traffic congestion (i.e., demand for private transport). This paper provides novel evidence on substitution between public and private transport, using an exogenous epidemic outbreak. Third, we also contribute to the literature on how risk perception affects spatial patterns of economic activity (Pope, 2008, Abadie, Dermisi, 2008, Manelici, 2017). Previous studies have found that fear of crime (Pope, 2008) or terror attacks (Abadie, Dermisi, 2008, Manelici, 2017) can affect housing prices and shift economic activity away from city centers (i.e., major railway stations). This paper offers new evidence showing that the risk of contracting an infectious disease could affect spatial distribution of economic activity by moving them away from areas close to crowded public places. This result is consistent with the recent evidence on COVID-induced reallocation of activities within and across US cities (Ramani, Bloom, 2021, Rosenthal et al., 2022). 2 Data and sample 2.1 Data This sectionbriefly introduces the administrative transportation data used to measure the demand for public transport (i.e., railway ridership) and private transport (i.e., highway traffic volume). Taiwan Railways (TR) is a 1065-kilometer railway network that services 21 of 22 counties via 241 stations. With annual journeys totaling more than 200 million kilometers, TR provides an extremely important form of transport in Taiwan. We collected daily passenger counts (entries plus exits) for each station from the government’s Open Data of Taiwan sharing platform.4 In addition to Taiwan Railways, another important transport mode is the national highway system. Currently, the 988.56-kilometer road network consists of nine lines in 20 of 22 counties. In our study, we focus on national highways where a toll is automatically collected by an electronic toll collection (ETC) system. While collecting fees, 327 toll reader devices also record vehicle speed, volume and other data. We collect data on traffic flow in five-minute intervals through each ETC station from the Freeway Bureau database.5 To maintain consistency with TR data, we aggregate five-minute traffic volume to a daily level. In addition, since we focus on private transport, data on bus and truck traffic are excluded from the sample, i.e., we use only private vehicle data. 2.2 Sample The sample is at the station-days level. The sample period is the first 24 weeks of 2018, 2019 and 2020.6 We only use TR stations and ETC stations that can be observed in the first 24 weeks of every year (i.e., a balanced panel). We also exclude TR stations located in Hualien and Taitung counties, where there is no highway. Among all TR stations, 180 satisfy the above criteria. In total, we have a sample size of 90,720 station-days for public transport. Similarly, 324 ETC stations fulfill balanced panel requirements, and we have a sample size of 162,648 station-days for private transport. 3 Empirical strategy and results Our identification strategy is the differences-in-differences (DID) design. Since the first COVID-19 case in Taiwan was reported on January 21st, 2020 (i.e., the 4th week of 2020), we use 2020 as the treated year and define the 1st to 3rd weeks and 4th to 24th weeks of the year as the pre-outbreak and post-outbreak periods, respectively. To control for seasonal patterns of the demand for public and private transport unrelated to the COVID-19 outbreak, we use 2018–2019 as untreated years, which helps construct the counterfactual trend of transportation patterns in 2020. 3.1 Effects of the COVID-19 outbreak on demand for public transport Since the impact of COVID-19 might have evolved over time, we need to trace the full dynamic trajectory of its effects. Therefore, following Chang etal. (2020) and Kleven etal. (2019), we implement a dynamic DID design by estimating the following regression:(1) Pidt=∑s≠−120βs·Y2020×Is+λt+ηw+θi+Xidtψ+εidt. Since we have daily numbers of passengers entering and exiting every TR station, estimation is implemented at the station-day level. Pidt represents outcomes of interest, namely, the log of the number of passengers exiting and entering station i on day d in year t. We include year fixed effects (λt) to capture the trend in demand for train travel over time. In addition, ηw denotes week of the year fixed effects. This helps to control for seasonal patterns in public transport demand over a year. To control for time-invariant confounding factors at the station level, we also include a full set of station fixed effects θi. Finally, Xidt is a set of covariates, including day-of-the-week fixed effects, various holiday dummies (e.g., Lunar New Year), daily temperature, daily rainfall, daily gasoline prices, and monthly population.7 Y2020 is a dummy variable for the treated year, which is set at one for the year 2020, and zero for 2018 and 2019 (untreated years). We denote the week in which the first COVID-19 case was reported with s=0, and then index all weeks relative to that week. The event time s runs from −3 to +20, since observations are from three weeks before the COVID-19 outbreak to 20 weeks after. Therefore, we use Is, whereby s=−3,−2,0,1,2…19,20, to denote the event time dummies. For example, I1 represents a dummy for the first week following the initial announcement of coronavirus cases. Since we use the week right before the outbreak as a baseline week, we omit the event time dummy at s=−1, i.e., the 3rd week of a year is used as the baseline period. The key variables used for identification in regression (1) are a set of event time dummies Is interacting with the dummy for the treated year Y2020. Coefficients of interest are βs, which measures the difference in demand for public transport between week s and the baseline week for 2020, relative to the difference in 2018 and 2019. Therefore, βs represents the COVID-19-induced change in demand for public transport, if the common trend assumption is valid. That is, in the absence of a COVID-19 outbreak, the time trend in railway ridership is assumed to be similar in both the treated and the untreated years. We examine this assumption by using data from the pre-outbreak period. To account for possible within-group error correlations, we use the multiway clustering approach proposed by Cameron etal. (2012) to calculate the standard errors clustered at both the date and the station levels. Fig. 1 a shows the results based on the TR data. The vertical axis of the figuredisplays the estimated βs and the corresponding 95% confidence intervals. Four key insights emerge from the figures. First, estimated coefficients at s=−3,−2 in the figuresare small and statistically insignificant, suggesting that trends in number of railway passengers in the treated year (i.e., 2020) and untreated years (i.e., 2018 and 2019) were similar before the COVID-19 outbreak. Therefore, the common trend assumption of our DID design is valid. Second, the TR ridership decreased by 25% within the first four weeks after the first COVID-19 case was announced, although there were only 22 new confirmed cases during this period.Fig. 1 The effect of the COVID-19 pandemic on transportation patterns. Notes: Sample period is the first 24 weeks of 2018-2020. The vertical axis of Fig.1 displays estimated βs in Eq. (1) and the corresponding 95% confidence level. The horizontal axis denotes weeks from the 4th week of a year. We define rush hours as 7am to 9am and 5pm to 9pm, and other times are defined as non-rush hours. Fig. 1 Third, the magnitude of the COVID-induced reduction was most pronounced at the peak of the pandemic in Taiwan during 2020 (i.e., mid-March 2020), with the number of passengers declining by more than 60%. Fourth, although negative effects of the COVID-19 outbreak gradually died away after Taiwan ceased to have any local COVID-19 cases starting from mid-April (i.e., April 12th, 2020), they did not recover to the pre-pandemic baseline. As a matter of fact, there were no new, locally transmitted cases in Taiwan for 253 consecutive days up to December 23rd, 2020. In Online AppendixA, we extend our sample period to the end of 2020 (i.e., the 48th week after the first case) and find that railway ridership was still 14% to 20% below pre-pandemic levels in December (see Fig.A1). This result is consistent with the survey evidence on the persistence of people’s behaviors reflecting the fear of virus infection. For example, according to a survey conducted by the National Taipei University of Nursing and Health Sciences in April,8 97.5% of Taiwanese people thought of the coronavirus as a serious disease, and over 90% of the interviewees correctly answered questions regarding how the virus spreads and what prevention measures were in place. Surveys conducted by YouGov (Smith, 2020) show that even at the end of 2020, approximately 60% of respondents said they were avoiding going to crowded public spaces (see Fig.A2). Interestingly, these numbers are comparable to the US, where the pandemic was still ongoing and more severe, implying that people would probably remain fearful of the coronavirus even after community transmission of COVID-19 is eliminated. So far, we have shown that the use of public transport declined substantially in response to the COVID-19 outbreak. To examine how people’s fear of infection affected demand for public transport, we use Google Trends data on search intensity of COVID-19-related keywords. Several medical studies (Ginsberg, Mohebbi, Patel, Brammer, Smolinski, Brilliant, 2009, McDonnell, Nelson, Schunk, 2012, Nuti, Wayda, Ranasinghe, Wang, Dreyer, Chen, Murugiah, 2014, Ayers, Leas, Johnson, Poliak, Althouse, Dredze, Nobles, 2020) suggest that Google Trends data on disease keywords can be a good proxy for the flu outbreak or fear of a flu pandemic. Following this idea, we sum up the search intensity of keywords “coronavirus” and “confirmed cases” to construct a measure for public perceptions of COVID-19 risk in Taiwan (hereafter, the COVID-19 Perception Index). Note that instead of showing the absolute search volume, Google Trends only provides a relative measure for the daily search volume, ranging from 0 to 100, where the numbers represent the search volume relative to the highest one. For example, the value of 100 is the peak popularity of a term, and a value of 50 means it is half as popular. Since we sum up two keywords, the maximum amount of our index is 200. In Online AppendixB, we examine the effect of new COVID-19 cases on the COVID-19 Perception Index. Fig.B1 shows that the evolution of new COVID-19 cases in Taiwan and the COVID-19 Perception Index have similar patterns. Our results suggest that one new COVID-19 case is associated with a 10% increase in the index. We then use the following regression to examine how the public perception of COVID-19 affects demand for public and private transport:(2) Pid=βCOV_PId+ηw+Xidψ+εid. Here, COV_PId is the log of the COVID-19 Perception Index on date d. The other notation is defined in the same way as in Eq. (1).9 In this specification, we use only 2020 data. The first three columns of Table 1 display the estimated coefficient of COV_PId for public transport. Panel A reports our main result, which shows that a 10% increase in the COVID-19 Perception Index is associated with a 1.6% decrease in number of daily passengers per TR station (Column (3)). We further conduct a subgroup analysis based on different pandemic periods defined in Online AppendixC: (1) Initial period (i.e., late January to mid-March, 1st to 8th weeks after the first COVID-19 case); (2) Peak period (i.e., mid-March to mid-April, 9th to 14th weeks after the first COVID-19 case); and (3) Recovery period (i.e., mid-April to June, 15th to 20th weeks after the first COVID-19 case). Panels B to D display results for the initial period, the peak period, and the recovery period, respectively. The results suggest that estimates in Panel A are mainly driven by the peak period when the COVID-19 Perception Index reached its peak in 2020 (see Panel C).Table 1 Effects of COVID-19 pandemic on the mode of transport. Table 1 Public transport (Railway) Private transport (Car) (1) (2) (3) (4) (5) (6) Panel A: 2020 COV_PI −0.133** −0.169** −0.164** 0.103*** 0.123*** 0.123*** (0.056) (0.073) (0.073) (0.033) (0.046) (0.046) Observations 30,240 54,108 Panel B: Initial period COV_PI −0.0649 −0.0873* −0.0865* 0.0740*** 0.0888*** 0.0958*** (0.040) (0.047) (0.047) (0.022) (0.031) (0.033) Observations 13,860 24,948 Panel C: Peak period COV_PI −0.709*** −0.745*** −0.728*** 0.548*** 0.579*** 0.589*** (0.235) (0.247) (0.242) (0.182) (0.188) (0.188) Observations 7,560 13,284 Panel D: Recovery period COV_PI 0.001 −0.004 −0.015 0.055* 0.060** 0.070** (0.031) (0.037) (0.046) (0.030) (0.026) (0.032) Observations 8,820 15,876 Basic controls √ √ √ √ √ √ Holiday FE √ √ √ √ Gasoline price √ √ Note: This table shows the estimated β (i.e. the coefficient on COV_PId) in Eq. (2). COV_PId is the log of the COVID-19 Perception Index as on d. The sample period in Panel A is the first 24 weeks of 2020. Note that the first confirmed COVID-19 case was announced on January 21st, 2020 (i.e., the fourth week of 2020). Panel B displays results for the initial period: the 1st to 8th weeks after the first COVID-19 case. Panel C displays results for the peak period: the 9th to 14th weeks after the first COVID-19 case. Panel D displays results for the recovery period: 15th to 20th weeks after the first COVID-19 case. Basic covariates includes week-of-the-year fixed effect, the day-of-week fixed effect, daily temperature, daily rainfall, and monthly county population. Note that daily temperature, daily rainfall, and monthly population are measured at the county level. Depending on where a TR station is located, we assign the corresponding county-level variables to that observation. Holiday FE includes a set of dummies for holidays, and election day, New Year’s Eve, New Year, Chinese New Year, Peace Memorial Day, Qing-Ming Festival, Labor Day, and the Dragon Boat Festival. Gasoline price includes daily gasoline prices at the national level. In order to account for possible within-group correlations among errors, we use the multiway clustering approach proposed by Cameron etal. (2012) to calculate standard errors clustered at both the date and station levels. Cluster-robust standard errors are reported in parentheses. *p<0.1**,p<0.05***,p<0.01 3.2 Effects of the COVID-19 outbreak on the demand for private transport In this section, we use ETC data to measure changes in demand for private transport. To compare it with demand for public transport, we use the log of daily number of cars passing through each ETC station as the outcome of interest, and as the same empirical specification as in Eqs. (1) and (2). The effect of the COVID-19 outbreak on the use of private transport is ambiguous. On the one hand, businesses could have shut down or shortened their working hours, since the pandemic had negative impacts on economic activity.10 Companies might also have adopted work-from-home policies to protect their employees from contracting COVID-19. According to an employee survey conducted by the 104 Job Bank, which is the largest human resource company in Taiwan, approximately 16% of employees worked from home during the pandemic in 2020.11 Moreover, people avoided exposure to the virus by postponing or canceling unnecessary outdoor activities. For all of these reasons, the COVID-19 pandemic reduced demand for both public and private transport. We call this the “fear effect. On the other hand, when people did go out, they adjusted their mode of transport by substituting private for public, as this could help maintain social distancing more easily. Thus, the “substitution effect can reduce demand for public transport but increase demand for private transport. Fig. 1 b displays the results for private transport. Again, the vertical axis of the figuredisplays the estimated βs and the corresponding 95% confidence intervals. There are three findings from the dynamic DID estimates. First, the COVID-19 outbreak had little impact on highway traffic volume at the beginning of the COVID-19 outbreak. The effects of COVID-19 on highway traffic turned out to be positive during the peak period of the 2020 pandemic in Taiwan (mid-March 2020). Most likely, at this time, passengers who would have ordinarily taken public transport were so concerned about the risk that they switched to private transport, so the substitution effect dominated the fear effect. Second, during rush hour, most trips are likely to be work-related and less discretionary. Since large numbers of people travel to work or go home after work during rush hours, the risk of contracting COVID-19 while using public transport is even greater. These facts suggest that the substitution effect could be stronger during rush hours than at non-peak times. Thus, we estimate Eq. (1) and report estimated βs by rush hour and non-rush hour, in Fig. 1c and d respectively. We define the rush hours as running from 7am to 9am and from 5pm to 9pm, while any other time is defined as “non-rush hour. Fig. 1c suggests that the rush hour traffic volume increased by approximately 25% when the 2020 pandemic in Taiwan was at its peak. In contrast, COVID-19 had little impact on the number of cars on national highways during non-rush hours (see Fig. 1d). Our results imply that people did shift to private vehicles when they had to go out during the pandemic. Third, the highway traffic flow increased by 17% to 28% during the period when Taiwan no longer reported any new local COVID-19 cases. Similar to public transport, we use Eq. (2) to estimate the effect of the public perception of COVID-19 risk on highway traffic. Estimated coefficients of COV_PId for private transport are reported in the last three columns of Table 1. Panel A shows the main result using 2020 ETC data. Our estimates suggest that a 10% increase in the COVID-19 Perception Index is associated with a 1.2% increase in the daily number of cars (see Column (6)). Combined with estimates in the first three columns of Panel A, our results suggest a strong substitution effect between public and private transport. Again, we conduct a subgroup analysis based on the same definition of the pandemic period as in Section 3.1 (see Panels B to D). Similar to public transport, the results suggest that our main estimate in Panel A is driven by the peak period, when the COVID-19 Perception Index rose quickly and attained its highest level in the period studied (see Panel C). 3.3 Impact of depressed public transit ridership on spatial patterns of urban activity So far, we have shown that the COVID-19 pandemic has induced a substantial decrease in railway ridership. Since most train stations, especially the major ones, are located in downtown areas, we posit that this decline in passenger flow during the pandemic may have negatively affected economic activity in urban areas close to main rail network nodes. In other words, the COVID-19 pandemic could have affected spatial patterns of business activities by shifting them away from areas close to major stations (i.e., city centers).12 Inspired by Ramani and Bloom (2021) and Rosenthal et al. (2022), we conduct two analyses to examine the above prediction, namely, between-district and within-district estimations. For the former, we examine whether the pandemic had a larger negative impact on retail sales in districts with major stations (i.e., urban areas) than in others. We use district-by-month-level retail transactional data for 2018 to 2020 and compare retail sales in districts with and without major TR stations, before and after the pandemic.13 For the latter, we further restrict the sample to districts with major stations and investigate the within-district reallocation of economic activities induced by the pandemic. For within-district estimation, given the difficulty in collecting data on business activities in small areas, following previous studies (Henderson, Storeygard, Weil, 2011, Chodorow-Reich, Gopinath, Mishra, Narayanan, 2020, Ch, Martin, Vargas, 2020), we exploit monthly nighttime lighting data from 2018 to 2020 as the proxy for local economic activity.14 The high spatial resolution of this nighttime lighting data allows for the comparison of luminosity within a 500-meter radius of a major station with that of 500 to 1000m away from the same station, before and after the pandemic.15 Fig. 2 shows the change in nighttime luminosity in areas surrounding Taipei Main Station, the busiest train station in Taiwan, as an example to illustrate how we use the luminosity data. We compare nighttime luminosity around this location in January 2019 (see Fig. 2a) and January 2020 (see Fig. 2c). The inner circles (outer circle) represent areas within a radius of 500m (500 to 1000m) from the railway station. Nighttime luminosity is measured by radiance values.16 A higher radiance value means a larger quantity of human-generated light in an area. Neither January 2019 nor January 2020 was affected by the COVID-19 pandemic and therefore we use the difference in nighttime luminosity in January as a baseline gap between 2019 and 2020. Fig. 2e indicates that nighttime luminosity was slightly brighter in January 2020 than in January 2019. Fig. 2b and d display similar graphs, using luminosity data in March 2019 and March 2020, respectively. In sharp contrast to Fig. 2e, we find that nighttime luminosity in March 2020 (i.e., the peak of the pandemic in Taiwan during the study period) was much darker than that in March 2019, especially within a 500-meter radius of Taipei Main Station (see Fig. 2f).Fig. 2 Nighttime luminosity in the area surrounding Taipei main station. Notes: This figuredisplays the geographic distribution of nighttime luminosity around Taipei Main Station. The inner circles (outer circle) represent the areas within a 500-meter (500-to 1000-meter) radius of Taipei Main Station. Fig.2a (2b) and 2c (2d) show nighttime luminosity in the area surrounding Taipei station in January (March) of 2019 and 2020, respectively. Fig.2 (2e) displays the difference in nighttime luminosity between January (March) 2020 and 2019. The nightlight luminosity is measured by radiance values. The unit of radiance value is nano watts per square centimeter per steradian (nW/cm2/sr). A higher value of radiance means a higher quantity of human-generated light in an area. Fig. 2 In the first instance, we estimate the following difference-in-differences model:(3) Ejmt=γY2020×Postm+λt+δm+θj+Xjmtψ+εjmt Ejmt represents the log of either retail sales or luminosity in district j in month m of year t.17 Y2020 is a dummy variable for the treated year, denoted by one if an observation is in 2020, and zero otherwise. Postm is a binary variable that takes the value one if an observation corresponds to the months between February and August (i.e., the post-outbreak period), and zero if the sample is observed in January (i.e., the pre-outbreak period). The year fixed effect λt controls for the general trend in local economic activity over time. The month-of-the-year fixed effect δm controls the seasonal patterns over a year. District fixed effects θj control for any time-invariant confounding factors at the district level. Finally, Xjmt refers to a set of covariates, including average temperature, average rainfall, number of households, population size, average house price, and number of real estate transactions. The key variable is the interaction term Y2020×Postm. Coefficient γ measures the difference in local economic activity (i.e., retail sales or nighttime lighting), before and after the COVID-19 outbreak in 2020, relative to the difference in the corresponding periods in 2018 and 2019. To identify the pandemic-induced reallocation of economic activity, we estimate Eq. (3) separately and compare the estimates of γ. For the between-districts analysis of retail sales, we estimate Eq. (3) by using districts with and without a major TR station. For the within-districts analysis of nighttime lights, we estimate the model using areas within 500m of major stations and those within 500 to 1000m away from the same station. Estimates are reported in Table 2 . Columns (1) to (3) show that during the pandemic, districts with major TR stations experienced a 14.5% decline in retail sales (see Panel A), while districts without major stations saw only an 11.8% decline (see Panel B). Using the luminosity data (nighttime lighting), we go one step further and examine the effects of COVID-19 on economic activities in areas surrounding major TR stations. Columns (4) to (6) suggest that nighttime luminosity within 500m of a major TR station (see Panel A, indicating a 16.7% decrease) experienced larger declines than areas slightly farther away (see Panel B, a 13.5% decrease).Table 2 Effects of COVID-19 pandemic on spatial patterns of urban activities. Table 2 Retail sales Nighttime light (1) (2) (3) (4) (5) (6) Panel A: Greater proximity to major TR stations Y2020×Post −153.153*** −173.173*** −145.145*** −174.174*** −202.202*** −167.167*** (0.013) (0.013) (0.013) (0.030) (0.036) (0.030) Observations 744 744 Panel B: Less proximity to major TR stations Y2020×Post −120.120*** −127.127*** −118.118*** −142.142*** −175.175*** −135.135*** (0.007) (0.008) (0.007) (0.025) (0.036) (0.025) Observations 5,100 744 Panel C: Triple-differences design Y2020×Post×Major −0317.0317** −035.035** −028.028* −032.032** −032.032** −032.032** (0.015) (0.015) (0.015) (0.014) (0.014) (0.014) Observations 5,844 1,488 Basic covariates √ √ √ √ √ √ District/County variables √ √ √ √ District FE √ √ Note: Panel A and Panel B show the estimated γ (i.e. the coefficient on Y2020×Post) in Eq. (3). Panel C shows the estimated γ1 (i.e. the coefficient on Y2020×Post×Major) in Eq. (4). Columns (1) to (3) show the results for between-townships analysis on retail sales. Columns (4) to (6) show the results for within-townships analysis on nighttime lights. Basic covariates for Panels A and B refers to year fixed effect and month fixed effect. Basic covariates for Panel C include a dummy variable for major stations Major, interaction terms Y2020×Majorj and Postm×Majorj, and year-by-month fixed effects. District/County variables include average temperatures, average rainfall, number of households, population size, average housing price, and number of real estate transactions. Note that average temperatures and average rainfall are measured at the county-by-month level. Number of households, population size, average housing price, and number of real estate transactions are measured at the district-by-year level. The real estate data were from administrative data on all house transactions in Taiwan provided by the Ministry of Interior (https://lvr.land.moi.gov.tw/). District FE includes a district fixed effect. In order to account for possible within-group correlations of the errors, we use the multiway clustering approach proposed by Cameron etal. (2012) to calculate the standard errors clustered at both the year-month and township level. Cluster-robust standard errors are reported in parentheses. *p<0.1**,p<0.05***,p<0.01 To summarize our findings, we consider the following triple-differences estimation.(4) Ejmt=γ0Majorj+γ1Y2020×Postm×Majorj+γ2Y2020×Majorj+γ3Postm×Majorj+λt×δm+θj+Xjmtψ+εjmt In this specification, we add a dummy variable Major, indicating districts with major TR stations (between-districts estimation) or areas within 500m of a major TR station (within-districts estimation). Therefore, we can control for the specific time trend and seasonality in areas close to major stations by including interaction terms Y2020×Majorj and Postm×Majorj. In addition, this empirical setting allows us to flexibly control for the time trend in economic conditions common in each district by including year-by-month fixed effects λt×δm. The key variable in the triple-differences design is Y2020×Postm×Majorj, which can capture the differential effect of the COVID-19 pandemic on economic outcomes in regions close to or far away from major TR stations. Estimates in Panel C of Table 2 show that retail sales in districts with major TR stations fell almost 2.8 percentage points relative to changes in other districts during the pandemic (see columns (1) to (3)). When using only districts with major rail nodes, we find that luminosity of nighttime lighting in areas surrounding major TR stations saw losses of approximately 3.2 percentage points compared to changes in areas slightly farther away from the same major nodes after the COVID-19 outbreak (see Columns (4) to (6)). There was extensive media coverage showing that hotels, theaters, and shopping malls, which are usually around public transit nodes, were either closed or had shortened their business hours during the pandemic.18 These stories are consistent with our findings related to the decline in nighttime luminosity in areas close to rail stations. Moreover, our result is consistent with the evidence provided by Rosenthal et al. (2022), whose results suggest that commercial rent premiums for properties close to rapid transit stations declined after the COVID-19 outbreak. To investigate the full dynamic trajectory of COVID-19s effects, we replace a dummy variable indicating the post-outbreak period Postm in Eq. (4) with event time dummies PostMonthm, where m=1,2,3,4,5,6,7,8. Note that we use January, the month just before the virus outbreak, as the baseline month and omit the event time dummy at m=1 (i.e., January). We estimate the following regression:(5) Ejmt=γ0Majorj+∑mαmY2020×Majorj×PostMonthm+γ2Y2020×Majorj+γ3Postm×Majorj+λt×δm+θj+Xjmtψ+εjmt The key coefficients αm measure the difference between economic outcomes for districts with and without major stations (area surrounding or slightly farther away from major nodes) in a given month, relative to the difference in the baseline month. Fig. 3 plots the estimated αm for effects on retail sales and nighttime luminosity, respectively. Fig. 3a suggests that compared to districts without major stations, those with major nodes experienced a relative fall in retail sales of approximately 2 to 6 percentage points, which was most pronounced in mid-March, the pandemic’s peak in Taiwan during 2020. Moreover, the retail sales gap gradually closed, but the point estimates did not return to pre-pandemic levels. This finding is consistent with evolution of the COVID-induced decline in TR ridership shown in Fig. 1a. A similar pattern can also be found in the within-district estimation, using nighttime luminosity as an outcome (see Fig. 3b).Fig. 3 Dynamic effects of COVID-19 pandemic on spatial patterns of economic activities. Notes: This figuredisplays the coefficients αm, which are the measure of difference in economic outcomes in a given month between the districts with and without major stations (areas surrounding and slightly farther away from major nodes) relative to the difference in the baseline month, in Eq. (4). The baseline month is January. Fig.3a shows the estimated αm for retail sales. Fig.3b shows the estimated αm for nighttime luminosity. Fig. 3 To sum up, our results clearly indicate that the pandemic could have induced movement of economic activity away from areas around major rail stations. Our finding is consistent with results found in recent studies using US data (Ramani, Bloom, 2021, Rosenthal et al., 2022), which suggests that COVID-19 reduced the value of living in city centers and led to reallocation of activities within or across cities. Given the low risk of contracting COVID-19 and the no-lockdown policy implemented in Taiwan, we believe our estimates could serve as a “lower bound for economic impacts of the decline in public transit ridership in other countries. 4 Conclusion Exploiting Taiwans unique experience and high-quality administrative data, we provide evidence that though there were no enforced restrictions on mobility during the pandemic, strong self-imposed restrictions existed. Specifically, our results indicate that the COVID-19 outbreak reduced the number of passengers taking a train journey by 40% to 60% at the peak of the pandemic in 2020. In contrast, highway traffic increased by 20% during the same period. This suggests that in the face of a pandemic, individuals not only curtailed mobility but also adjusted the mode of transport in order to reduce the risk of infection. Moreover, data of retail sales and nighttime luminosity show that this shift in transport modes is not only related to patterns of population mobility but also results in movement of economic activity away from areas around major rail stations. Since we also find that the decline in public transit ridership can persist even after a pandemic, our findings point towards some fruitful directions for future research. For example, it would be interesting to examine whether the pandemic would have a long-term or permanent impact on people’s mobility decisions or transport modes. In addition, future studies could investigate how this change affects spatial patterns of economic activity in a post-pandemic period. An interesting question is why people reacted so strongly and persistently to the pandemic in Taiwan, even though the risk was so low. Although we do not have direct evidence for this hypothesis, we speculate that the painful experience of SARS, which ravaged Taiwan (as well as China, Singapore, Hong Kong, Vietnam, South Korea and Canada) during 2002–2003, might have played a role in giving individuals in these areas a strong incentive to practice social distancing.19 However, since only a few regions experienced the SARS outbreak, this lesson might be difficult to carry over to other countries. Appendix A Supplementary materials Supplementary Data S1 Supplementary Raw Research Data. This is open data under the CC BY license http://creativecommons.org/licenses/by/4.0/ Supplementary Data S1 ☆ We thank Chung-Ying Lee, Yi-Cheng Kao, and participants of the COVID-19 study group at the Institute of Economics, Academia Sinica for their valuable comments. We also thank excellent assistance from Yung-Yu Tsai and Chih-Yun Wang. Kong-Pin Chen acknowledges financial support from the 10.13039/501100004663 Ministry of Science and Technology (MOST-109-2410-H-001-051-MY3). Jui-Chung Yang acknowledges financial support from the Ministry of Science and Technology (MOST-109-2410-H-002 -228 -MY2). Tzu-Ting Yang acknowledges financial support from the Ministry of Science and Technology (MOST-109-2628-H-001-001-MY2). 1 Several public events were canceled between March 25th and June 7th because the government had announced guidance suggesting that unnecessary public gatherings with more than 100 people indoors or 500 people outdoors should not be held. 2 We will discuss how to construct this index in Section 3.1 and Online Appendix. 3 Based on the accumulated number of COVID-19 cases reported as on October 28th, 2020, the incidence of COVID-19 per 1,000,000 population was approximately 23 in Taiwan and 26,960 in the US. 4 https://data.gov.tw/dataset/8792 5 https://www.freeway.gov.tw/ 6 Note that the definition of “week in this study follows the World Health Organization (WHO) definition, which always begins on a Sunday and ends on a Saturday, but does not necessarily start from January 1st. 7 Table A1 reports summary of statistics of variables used in Sections 3.1 and 3.2. 8 https://news.sina.com.tw/article/20200509/35111198.html, Date accessed Aug. 10th 2020 9 Similar to Eq. (1), we also use the multiway clustering approach proposed by Cameron etal. (2012) to calculate the standard errors clustered at both the date and the station levels. 10 Unemployment statistics from the Ministry of Labor, released in April 2020, indicate that the number of unemployed workers was 0.48 million, the highest since 2013. In addition, the number of employees working less than 35 hours per week was 0.40 million, higher by 0.21 million from 0.19 million in April 2019. 11 We obtained this statistic from the following news source: https://www.rti.org.tw/news/view/id/2101392. Several large companies implemented a work-from-home policy in 2020, as reported by newspapers. For example, “Approximately 3,245 employees in several financial firms were told to work from home for two weeks from April 6th (see https://www.taipeitimes.com/News/biz/archives/2020/04/07/2003734109). Taiwan Semiconductor Manufacturing Company (TSMC), having the largest semiconductor foundry in the world, implemented a work-from-home policy for employees not on production lines (see https://www.taiwannews.com.tw/en/news/3903344). 12 Table A2 lists 32 major rail stations, including four special-class stations and 28 first-class stations, in Taiwan and the corresponding location information. 13 We acquire transactional data of monthly retail sales at the district level from the open data platform offered by the Ministry of Finance (https://data.gov.tw/dataset/36862). 14 We obtain luminosity data on nighttime lighting from the National Oceanic and Atmospheric Administration (NOAA). 15 The advantage of this nighttime lighting data is its high spatial resolution (15 arc seconds, 0.5km×0.5km) and strong timeliness (monthly data). 16 The radiance value unit is nano watts per square centimeter per steradian (nW/cm2/sr). 17 Table A3 of the Online Appendixprovides summary statistics for these outcome variables. 18 For example, Eslite Mall at Taipei Rail Station closed one hour earlier (from 10:30pm to 9:30pm), since passenger flow decreased significantly during the pandemic (see https://udn.com/news/story/7934/4430775). Taipei 101 also shut down two hours earlier from April 2020 (see https://www.taipeitimes.com/News/biz/archives/2020/04/01/2003733744?fbclid=IwAR1zxqb14B4LA7v8tQmAnEn0IMUFH3gja_YiIb1hmnmjclQyjfTXsUfc7cQ). Another example is a five-star hotel close to Taichung station that decided not to open due to the pandemic (see https://news.ltn.com.tw/news/life/paper/1365882), and two theaters around Changhua rail station that closed after the COVID-19 outbreak (see https://www.taipeitimes.com/News/taiwan/archives/2020/04/27/2003735378?fbclid=IwAR0rHdJ6pgXAag1ue_AZx_R9Vg51WyJ4c6M7Y1qE0Pdv0ParQd58awA1QLI) 19 As of September 2021, these countries have relatively low COVID-19 incidence rates. For example, the total cases per 1,000,000 population is 40,380 in Canada but 121,520 in the US. Note that among these countries, Canada has the highest COVID incidence rate. Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jue.2022.103426. ==== Refs References Abadie A. Dermisi S. Is terrorism eroding agglomeration economies in central business districts? Lessons from the office real estate market in downtown Chicago J. Urban Econ. 64 2 2008 451 463 Ahituv A. Hotz V.J. Philipson T. The responsiveness of the demand for condoms to the local prevalence of AIDS J. Hum. Resour. 31 4 1996 869 897 Anderson M.L. Subways, strikes, and slowdowns: the impacts of public transit on traffic congestion Am. Econ. Rev. 104 9 2014 2763 2796 Argente D. Hsieh C.-T. Lee M. The cost of privacy: welfare effects of the disclosure of COVID-19 cases Rev. Econ. Stat. 104 1 2022 176 186 Ayers J.W. Leas E.C. Johnson D.C. Poliak A. Althouse B.M. Dredze M. Nobles A.L. Internet searches for acute anxiety during the early stages of the COVID-19 pandemic JAMA Intern. Med. 180 12 2020 1706 1707 32832984 Bennett D. Chiang C.-F. Malani A. Learning during a crisis: the SARS epidemic in Taiwan J. Dev. Econ. 112 2015 1 18 32287877 Brinkman, J., Mangum, K., 2022. JUE Insight: The geography of travel behavior in the early phase of the COVID-19 pandemic. J. Urban Econ., In Press. Cameron A.C. Gelbach J.B. Miller D.L. Robust inference with multiway clustering J. Bus. Econ. Stat. 29 2 2012 238 249 Ch R. Martin D.A. Vargas J.F. Measuring the size and growth of cities using nighttime light J. Urban Econ. 2020 103254 Chang H. Chang K. Fan E. The intended and unintended effects of drunk driving policies Oxf. Bull. Econ. Stat. 82 1 2020 23 49 Chen Y. Whalley A. Green infrastructure: the effects of urban rail transit on air quality Am. Econ. J. 4 1 2012 58 97 Chodorow-Reich G. Gopinath G. Mishra P. Narayanan A. Cash and the economy: evidence from India’s demonetization Q. J. Econ. 135 1 2020 57 103 Couture, V., Dingel, J.I., Green, A., Handbury, J., Williams, K.R., 2022. JUE Insight: Measuring movement and social contact with smartphone data: a real-time application to COVID-19. J. Urban Econ., In Press. Cronin C.J. Evans W.N. Total shutdowns, targeted restrictions, or individual responsibility: how to promote social distancing in the COVID-19 era? J. Health Econ. 79 2021 102497 34325285 Duranton G. Turner M.A. The fundamental law of road congestion: evidence from US cities Am. Econ. Rev. 101 6 2011 2616 2652 Engle S. Stromme J. Zhou A. Staying at home: mobility effects of COVID-19 COVID Econ. 4 2020 86 102 Fang H. Wang L. Yang Y. Human mobility restrictions and the spread of the novel coronavirus (2019-nCoV) in China J. Public Econ. 191 2020 104272 33518827 Gersovitz M. Hammer J.S. Infectious diseases, public policy, and the marriage of economics and epidemiology World Bank Res. Obs. 18 2 2003 129 157 Ginsberg J. Mohebbi M.H. Patel R.S. Brammer L. Smolinski M.S. Brilliant L. Detecting influenza epidemics using search engine query data Nature 457 7232 2009 1012 1014 19020500 Goolsbee A. Syverson C. Fear, lockdown, and diversion: comparing drivers of pandemic economic decline 2020 J. Public Econ. 193 2021 104311 33262548 Glaeser, E.L., Gorback, C., Redding, S.J., 2022. JUE Insight: How much does COVID-19 increase with mobility? Evidence from New York and four other U.S. cities. J. Urban Econ., In Press. Henderson V. Storeygard A. Weil D.N. A bright idea for measuring economic growth Am. Econ. Rev. 101 3 2011 194 199 25076786 Kleven H. Landais C. Søgaard J.E. Children and gender inequality: evidence from Denmark Am. Econ. J. 11 4 2019 181 209 Lakdawalla D. Sood N. Goldman D. HIV breakthroughs and risky sexual behavior Q. J. Econ. 121 3 2006 1063 1102 Li B. Ma L. Migration, transportation infrastructure, and the spatial transmission of COVID-19 in China J. Urban Econ. 2021 InPress Liu L. Miller H.J. Scheff J. The impacts of COVID-19 pandemic on public transit demand in the united states PLoS One 15 11 2020 e0242476 33206721 Manelici I. Terrorism and the value of proximity to public transportation: evidence from the 2005 London bombings J. Urban Econ. 102 2017 52 75 Mangrum D. Niekamp P. Urban, J. Econ. (Ed.), JUE Insight: College student travel contributed to local COVID-19 spread 2022 Press McDonnell W.M. Nelson D.S. Schunk J.E. Should we fear “flu fear” itself? Effects of H1N1 influenza fear on ED use Am. J. Emerg. Med. 30 2 2012 275 282 21208765 Monte F. Mobility zones Econ. Lett. 194 2020 109425 32834237 Nelson P. Baglino A. Harrington W. Safirova E. Lipman A. Transit in Washington, DC: current benefits and optimal level of provision J. Urban Econ. 62 2 2007 231 251 Nuti S.V. Wayda B. Ranasinghe I. Wang S. Dreyer R.P. Chen S.I. Murugiah K. The use of google trends in health care research: asystematic review PLoS One 9 10 2014 e109583 25337815 Parry I.W.H. Small K.A. Should urban transit subsidies be reduced? Am. Econ. Rev. 99 3 2009 700 724 Pope J.C. Fear of crime and housing prices: household reactions to sex offender registries J. Urban Econ. 64 3 2008 601 614 Ramani A. Bloom N. The donut effect of COVID-19 on cities NBER Working Paper 2021 Rosenthal S.S. Strange W.C. Urrego J.A. Urban, J. Econ. (Ed.), JUE insight: Are city centers losing their appeal? Commercial real estate, urban spatial structure, and COVID-19 2022 Press Smith M. International COVID-19 Tracker Update: 18 May Technical Report 2020 National Communications Commission Organization Winston C. Langer A. The effect of government highway spending on road users’ congestion costs J. Urban Econ. 60 3 2006 463 483 Xin M. Shalaby A. Feng S. Zhao H. Impacts of COVID-19 on urban rail transit ridership using the synthetic control method Transp. Policy 111 2021 1 16

0

PMC9752941

NO-CC CODE

2022-12-16 23:25:27

no

J Urban Econ. 2022 Jan 15; 127:103426

utf-8

J Urban Econ

10.1016/j.jue.2022.103426

oa_other

==== Front Atmos Ocean Opt Atmospheric and Oceanic Optics 1024-8560 2070-0393 Pleiades Publishing Moscow 4395 10.1134/S1024856022060136 Optics of Clusters, Aerosols, and Hydrosoles Cluster Composition of Anemophilous Plant Pollen Entering the Atmosphere Golovko V. V. golovko@kinetics.nsc.ru 1 Zueva G. A. 2 Kiseleva T. I. 12 1 grid.418912.7 0000 0000 9501 0228 Institute of Chemical Kinetics and Combustion, Siberian Branch, Russian Academy of Sciences, 630090 Novosibirsk, Russia 2 grid.465435.5 0000 0004 0487 2025 Central Siberian Botanical Garden, Siberian Branch, Russian Academy of Sciences, 630090 Novosibirsk, Russia 15 12 2022 2022 35 6 673679 23 3 2022 31 3 2022 11 4 2022 © Pleiades Publishing, Ltd. 2022, ISSN 1024-8560, Atmospheric and Oceanic Optics, 2022, Vol. 35, No. 6, pp. 673–679. © Pleiades Publishing, Ltd., 2022.Russian Text © The Author(s), 2022, published in Optika Atmosfery i Okeana. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Results of studying pollen emission into the atmosphere are presented for 26 anemophilous and three entomophilous plant species for which optional anemophily is possible. The percentage of clusters of two or more pollen grains of the total number of pollen particles entering the atmosphere is estimated. It is shown that such clusters were formed in significant quantities in all series of experiments. The percentage of pollen clusters reaches ∼71% of the total number of pollen particles. The percentage of pollen grains in the composition of the clusters reaches ∼94% of the total number of pollen grains. Keywords: pollen anemophilous plant atmospheric aerosol cluster issue-copyright-statement© Pleiades Publishing, Ltd. 2022 ==== Body pmcINTRODUCTION Pollen of anemophilous plants is an indispensable component of the coarse fraction of atmospheric aerosol. Its propagation in the atmosphere is a physical process necessary for seed reproduction of anemophilous plants, the main producers of extratropical land biomass. Pollen grains (PGs) cause outbreaks of allergic reactions in 30% of the population [1] and take part in the transfer of chemical elements in biocenoses. The high allergenicity of pollen has led to the creation of many aero-palynological monitoring stations throughout the world. They are often combined into regional and national networks where long-term observations are carried out by standard techniques. In 2016, there were 879 stations of pollen aerosol sampling in the world: 9 in Africa, 151 in America, 182 in Asia (143 in Japan), 525 in Europe, and 12 in Oceania [2]. In Russia, there are nine stations: in Moscow, St. Petersburg, Yekaterinburg, Krasnodar, Perm, Rostov-on-Don, Ryazan, Stavropol, and Tyumen (https://allergotop.com). Most often, pollen concentration in the atmosphere is determined using Hirst, Burkard, and Lanzoni slit aspiration traps (>600, or 70% of the stations). Features of pollen particle (PP) morphology (large size, the mean diameter of PG of anemophilous plants is 20–40 µm, drying deformation, and the presence of clusters) are caused by significant difficulties in the process of pollen aerosol (PA) entrapment. The sampling is nonisokinetic, PP settling on the walls of collecting devices is observed [3, 4], the sampling is accompanied by destruction of clusters [5], and the efficiency of trapping of >10-µm particles is unsatisfactory [6] (for example, ∼15% of cereal pollen is not trapped). The analysis of samples collected by impactors requires qualified executives and takes from three to ten days. The error of pollen concentration measurements in air is ∼30% [7]. The currently available technique of PA sampling did not provide continuous monitoring during the COVID-19 pandemic [8]. The quarantine and isolation of personnel prevented observations from being carried out. It is also necessary to note that collections of PA samples allow one to judge the pollen content in the atmosphere only in the vicinity of an observation site. Even in Europe (for example, in Bavaria) the number of pollen monitoring stations is insufficient for forecasting the pollen transfer in the atmosphere [2]. Currently, the AutoPollen program is operative in Europe. It is aimed at creating a prototype of a network of fully automatic pollen monitoring stations. To cover main bioclimatic zones of Europe, it is planned to additionally deploy 200–300 automatic stations allowing one to obtain data in the real time mode (several minutes after sampling). It is believed that timely acquisition of information about taxonomic affiliation, terms of entering, and concentration of PPs in the atmosphere will notably reduce the direct and indirect health costs associated with allergy, currently estimated between €50–150 billion/year [8]. The area of Russia, three times the area of Europe, with a much lower population density and an extremely uneven distribution of the population, objectively hinders the creation of a network of pollen monitoring stations. Models of pollen transport created allowing for vegetation features, pollen production of plants, and processes of PP propagation seem to be more promising for forecasting the pollen content in the atmosphere. The efficiency of the wind blowing PPs off the anther surface, the time of their stay in the air, distance of transport, and efficiency of capture by collecting surfaces depend on the sedimentation rate. The PGs of anemophilous plants entering the atmosphere dry out and change their size, shape, and density of the protoplast [9, 10]. In the cytoplasm of a PG vegetative cell, large air spaces can appear [11]. The PG sedimentation rate is most strongly affected by the formation of clusters of two and more PGs [12], the process which is still poorly studied. This work continues the studies [13–15] of the PA cluster composition during atmospheric pollen emission from 29 kinds of species of plants growing in the Central Siberian Botanical Garden (CSBG), Siberian Branch, Russian Academy of Sciences. MATERIALS AND METHODS All experiments were carried out in field conditions. We studied the cluster composition of PA entering the atmosphere from plants growing outdoors, both in natural populations and in collection areas of the CSBG. PPs were blown by wind from plant inflorescences to substrates covered with glycerin gelatin with addition of Coomassie blue (Fig. 1). The gust speed was measured by an anemometer and was 0.3–2.0 m/s. The substrates were arranged in the direction of the wind. The distance from the substrates was 20–25 cm, which made it possible to trap a sufficient amount of PPs and to avoid the contact of the substrates with inflorescences. Pollen of each species was sampled five times with intervals of several minutes. The exposure of the substrates lasted 1–2 s. Temperature and relative air humidity were simultaneously measured with a Center 311 device. The pollen particles (individual PGs and their clusters) were counted on ten transects at 10–40-fold magnification of the MBI-11U42 microscope lens. Fig. 1. Entrapment of Corylus colurna pollen at the time of its efflorescence from the anthers. Features of the PG morphology of anemophilous plants (the pollen is dry, with a thin and smooth outer shell (exine)) impede their coalescence; however, it is not clear how efficiently they interfere with cluster formation. It is supposed that individual PGs enter the atmosphere during the emission from anthers and the observed clusters are formed in the process of sedimentation directly on the microscope slide (Figs. 2 and 3). Fig. 2. Examples of PG clusters of Agropyron cristatum. Fig. 3. Density of the PG sediment of Poa pratensis on a substrate. The number of clusters of two and more PGs that would be formed on a substrate was estimated under the following assumptions: (i) sedimentation of PGs onto a substrate does not depend on sedimentation of other PGs and (ii) PGs are arranged in a layer in a forming cluster. With an increase in the number of PGs in clusters, their number on a substrate should decrease. If the mathematical expectation of the number of clusters consisting of an arbitrary number of PGs is less than unity, then similar (and larger) PPs are not formed on a substrate at a given number of PGs per unit area. A cluster is formed if the distance between the geometric centers of PGs does not exceed two radii. Thus, the mathematical expectation of the number of clusters including two or more PGs (N≥ 2) can be represented in the form1 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{N}_{{ \geqslant 2}}} = 4p{{N}_{{ \geqslant 1}}},$$\end{document} where N≥ 1 is the number of PGs in clusters of one and more PGs (in fact, the total number of PGs settled on the substrates);2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$p = {{{{S}_{{{\text{pg}}}}}{{N}_{{ \geqslant 1}}}} \mathord{\left/ {\vphantom {{{{S}_{{{\text{pg}}}}}{{N}_{{ \geqslant 1}}}} {{{S}_{{\text{T}}}}}}} \right. \kern-0em} {{{S}_{{\text{T}}}}}}$$\end{document} is the fraction of the substrate surface occupied by PGs (see Fig. 1), Spg is the average area of the PG projection, and ST is the area of the substrate examined. To estimate the area of individual PGs, photographs of ∼200 PGs of all examined plant species were taken in the Common Use Center for Microscopic Analysis of Biological Objects of the Siberian Branch, Russian Academy of Sciences. The areas of PG projections in the photographs were determined from image processing with the MapInfo Professional software. The mathematical expectation of the number of clusters of two PGs can be estimated by the relationship3 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{N}_{2}} = {{N}_{{ \geqslant 2}}} - {{N}_{{ \geqslant 3}}},$$\end{document} where N≥3 is the mathematical expectation of the number of clusters consisting of at least three PGs,4 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{N}_{{ \geqslant 3}}} = 7p{{N}_{{ \geqslant 2}}}.$$\end{document} In the general case, mathematical expectations of the number of clusters of j and more PGs can be represented in the form5 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{N}_{{ \geqslant j}}} = \left( {3(j - 1) + 4} \right)p{{N}_{{ \geqslant (j{\kern 1pt} - {\kern 1pt} 1)}}},$$\end{document} 6 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{N}_{j}} = {{N}_{{ \geqslant j}}} - {{N}_{{ \geqslant j{\kern 1pt} + {\kern 1pt} 1}}}.$$\end{document} The mathematical expectation of the number of individual PGs7 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{N}_{1}} = {{N}_{{ \geqslant 1}}} - 2{{N}_{2}} - 3{{N}_{3}} - \ldots - j{{N}_{j}}.$$\end{document} In the process of PG settling on a substrate, two alternative variants are possible: PGs appear on the substrate either individually or as components of clusters of two or more PGs. Therefore, the problem can be reduced to comparison of fractions of individual PGs of the total number of PGs settled on the substrates, i.e., to estimation of the significance of differences between fractions or percentages of the feature characterized by an alternative distribution. For this purpose, the Fisher criterion F with the φ-transformation (Fisher angular transformation) was used. It is intended for comparison of two samples according to the frequency of occurrence of an index a researcher is interested in:8 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$F = \frac{{{{{({{\varphi }_{1}} - {{\varphi }_{2}})}}^{2}}{{N}_{{a{\kern 1pt} \geqslant {\kern 1pt} 1}}}{{N}_{{b{\kern 1pt} \geqslant {\kern 1pt} 1}}}}}{{{{N}_{{a{\kern 1pt} \geqslant {\kern 1pt} 1}}}{{N}_{{b{\kern 1pt} \geqslant {\kern 1pt} 1}}}}} \sim {{F}_{{(a,{\kern 1pt} {\kern 1pt} d{{f}_{1}}{\kern 1pt} {\kern 1pt} d{{f}_{2}})}}},$$\end{document} where φ1 and φ2 are the transformed fractions and Na ≥ 1, Nb ≥ 1 are the sample volumes (the total numbers of PGs on the substrates). The value obtained was compared with the tabular one at a given level of significance and the number of the degrees of freedom df1 = 1; df2 = Na ≥ 1 + Nb ≥ 2 − 2. When the sample volumes number in hundreds and thousands, F = 3.8 at the significance level α = 0.05, 6.6 at α = 0.01, and 10.8 at α = 0.001. If the calculated values of F exceed the abovementioned values, the null hypothesis at a given level of significance should be rejected. RESULTS Emission of pollen entering the atmosphere directly from anthers of 29 plant species growing in the CSBG, including 26 anemophilous plant species: seven species of woody plants, 14 species of grasses, and five species of weeds (herbs) have been studied. The data about the presence of clusters in the pollen of the plants under study are shown in Table 1. Clusters were revealed in pollen samples of all plants under study. The percentage of clusters of two and more PGs reached 71% of the total number of PPs (Alnus hirsuta); the PG percentage in their composition was 93.7% (Populus alba) of the total number of trapped PGs. Table 1.   Presence of clusters in the plant pollen entering the atmosphere Specific name Observation date t, °C RH, % Number Percentage of clusters of two and more PGs, % PG percentage in clusters of two and more PGs, % PPs PGs Anemophilous plants Pinus mugo June 13, 2020 27.0 54.0 3195 5762 36.0 64.5 Carex cespitosa June 1, 2021 34.2 32.7 2072 2500 12.7 27.7 Agropyron cristatum June 23, 2020 26.8 52.7 4293 5143 13.4 27.7 Agrostis capillaris June 23, 2020 32.0 43.3 1351 2019 24.6 49.5 Cyperus papyrus June 3, 2021 22.4 62.8 494 607 12.3 28.7 Festuca rubra June 8, 2020 29.7 95.6 921 1259 21.0 42.2 Sanguisorba officinalis July 26, 2017 28.8 42.6 1637 1963 14.7 28.9 Corylus colurna May 3, 2021 14.2 30.7 4469 5356 13.5 27.8 Festuca ovina June 2, 2020 22.4 61.9 1352 1593 12.1 25.4 Poa pratensis June 8, 2020 26.6 49.7 396 566 16.4 41.5 Helictotrichon sempervirens June 1, 2021 35.9 20.8 422 685 25.4 54.0 Amaranthus caudatus Sept. 9, 2019 24.4 50.7 822 1572 35.4 66.2 Carex leporina June 1, 2021 31.7 41.8 670 914 19.1 40.7 Koeleria glauca June 8, 2020 28.8 51.3 1604 2206 18.6 40.8 Calamagrostis acutiflora Sept. 9, 2019 23.0 38.5 676 909 19.1 39.8 Fraxinus pennsylvanica May 15, 2021 23.6 19.7 4917 6195 15.7 33.1 Triticum aestivum July 26, 2019 24.2 72.7 514 730 18.5 42.6 Triticum durum July 26, 2019 29.6 65.0 428 659 26.9 52.5 Microbiota decussata Apr. 30, 2019 26.7 24.5 1539 2212 22.3 45.9 Festuca glauca June 8, 2020 25.7 65.4 1145 1645 17.6 42.6 Briza media June 23, 2020 37.0 46.9 584 1175 33.7 67.1 Alnus hirsuta Apr. 26, 2019 12.2 47.8 3805 7877 71.0 86.0 Chosenia arbutifolia May 8, 2019 22.2 37.4 380 462 15.5 30.5 Populus alba Apr. 26, 2017 17.6 53.2 2609 18321 55.4 93.7 Pennisetum alopecuroides Sept. 9, 2019 28.5 80.1 3394 4693 19.4 41.7 Avenella flexuosa June 8, 2020 26.6 51.7 3499 4722 16.9 38.4 Entomophilous plants for which wind pollination is possible Oxalis acetosella Aug. 9, 2018 39.5 54.0 487 1347 38.8 77.9 Tilia amurensis July 13, 2019 27.7 39.6 1669 2489 25.3 49.9 Tilia platyphyllos July 10, 2019 26.9 47.7 496 1240 40.5 76.2 The results of the statistical analysis of plant samples are presented in Table 2. Table 2.   Number (nj) and mathematical expectation (Nj) of PG clusters Specific name N ≥1 p, % Parameter Number j of PGs in the cluster composition F 1 2 3 4 5 6 7 8 9 ≥10 ≥20 ≥100 Anemophilous plants Pinus mugo 5762 0.98 n j 2046 566 268 140 73 37 19 8 4 28 6 69.0 N j 5293 211 14 1 Carex cespitosa 2500 0.13 n j 1808 185 38 21 9 5 1 3 2 31.8 N j 2416 13 Agropyron cristatum 5143 0.37 n j 3716 447 67 26 21 6 2 3 1 4 38.7 N j 4989 74 2 Agrostis capillaris 2019 0.15 n j 1019 178 68 44 16 14 3 4 3 2 39.2 N j 1995 12 Cyperus papyrus 607 0.09 n j 433 45 5 2 3 2 1 1 1 1 16.9 N j 603 2 Festuca rubra 1259 0.07 n j 728 135 34 4 4 6 3 2 0 5 31.5 N j 1251 4 Sanguisorba officinalis 1963 0.06 n j 1396 192 35 7 4 0 0 0 0 3 53.4 N j 1955 4 Corylus colurna 5356 0.14 n j 3865 430 120 31 11 8 2 0 0 1 1 46.5 N j 5296 30 Festuca ovina 1593 0.10 n j 1188 135 12 5 3 5 2 0 0 2 24.9 N j 1581 6 Poa pratensis 566 0.02 n j 331 25 17 8 5 4 2 0 1 3 21.6 N j 564 1 Helictotrichon sempervirens 685 0.05 n j 315 55 17 11 6 7 4 1 3 3 28.6 N j 683 1 Amaranthus caudatus 1572 0.04 n j 531 125 73 37 13 11 9 9 2 10 2 50.1 N j 1568 2 Carex leporina 914 0.05 n j 542 58 46 14 6 2 0 0 0 2 26.7 N j 910 2 Koeleria glauca 2206 0.12 n j 1305 153 83 28 12 7 4 7 2 3 39.7 N j 2186 10 Calamagrostis acutiflora 909 0.07 n j 547 75 26 16 6 4 0 2 26.3 N j 905 2 Fraxinus pennsylvanica 6307 0.18 n j 4253 505 145 52 30 8 5 4 4 10 54.5 N j 6214 46 1 Triticum aestivum 730 0.08 n j 419 51 18 6 7 5 2 3 2 1 23.7 N j 726 2 Triticum durum 659 0.07 n j 313 64 21 17 4 4 2 0 1 2 27.1 N j 655 2 Microbiota decussata 2212 0.18 n j 1196 213 71 23 14 10 1 1 0 6 4 41.3 N j 2179 16 Festuca glauca 1645 0.11 n j 944 130 18 14 7 8 7 3 5 5 4 35.5 N j 1631 7 Briza media 1175 0.10 n j 387 70 38 22 23 17 6 10 3 8 42.5 N j 1167 4 Alnus hirsuta 3897 0.28 n j 1103 322 128 86 46 30 18 25 14 20 12 75.6 N j 3807 43 1 Chosenia arbutifolia 462 0.01 n j 321 37 21 1 17.8 N j 462 Populus alba 18 321 0.63 n j 1163 318 194 93 82 60 43 47 29 270 248 62 208.0 N j 17373 443 19 1 Pennisetum alopecuroides 4693 0.09 n j 2734 403 118 64 21 21 10 6 6 8 3 54.2 N j 4659 16 Avenella flexuosa 4722 0.10 n j 2908 344 100 58 26 26 16 10 4 6 1 56.4 N j 4685 12 Entomophilous plants for which wind pollination is possible Oxalis acetosella 1347 0.03 n j 298 90 33 18 10 8 2 5 4 11 5 3 53.3 N j 1343 2 Tilia amurensis 2489 0.11 n j 1246 227 103 45 27 6 9 2 0 3 1 48.7 N j 2467 11 Tilia platyphyllos 1240 0.06 n j 295 73 42 22 13 7 4 11 8 13 8 49.4 N j 1234 3 The Fisher criterion varied from 16.9 (Cyperus papyrus) to 208 (Populus alba), which is obviously higher than its value even at a significance level of 0.001 (10.8). The null hypothesis about formation of clusters from individual PGs on the substrates should be rejected. Pollen of entomophilous plants significantly differed in its cluster composition (α = 0.001) from the PA produced by anemophilous plants. For example, comparison of percentages of individual PGs produced by Oxalis acetosella and Chosenia arbutifolia yields F = 42.3. CONCLUSIONS Results of studying the pollen entering the atmosphere for ∼1/12 of the total number (417) of species represented in the flora of Akademgorodok, Novosibirsk, corroborate preliminary conclusions about the character of the pollen emission [9–11]. The anemophilous plant pollen entering the atmosphere is not monodisperse. In addition to individual PGs, clusters containing two and more PGs enter the atmosphere. The percentage of such clusters of the total number of forming particles varies over a wide range and can be markedly different in different plant species. The following conclusions can be drawn from the work: (1) Morphological features of the structure of pollen grains of anemophilous plants do not prevent the formation of clusters in the process of pollen emission into the atmosphere. (2) Anemophilous plant pollen entering the atmosphere is not monodisperse but represented both by individual PGs and by clusters of two and more PGs. (3) The fraction of clusters of the total number of particles formed and the percentage of PGs in their composition vary over a wide range and can reach 71 and 93%, respectively. (4) Anemophilous plant pollen entering the atmosphere significantly differs from pollen of entomophilous plants in its cluster composition. CONFLICT OF INTEREST The authors declare that they have no conflicts of interest. Translated by A. Nikol’skii ==== Refs REFERENCES 1 Biedermann T. Winther L. Till S. J. Panzner P. Knulst A. Valovirta E. Birch pollen allergy in Europe Allergy 2019 74 1237 1248 30829410 2 J. T. M. Buters, C. Antunes, A. Galveias, K. C. Bergmann, M. Thibaudon, C. Galan, C. Schmidt-Weber, and J. Oteros, “Pollen and spore monitoring in the world,” Clin. Transl. Allergy 8 (9) (2018). 10.1186/s13601-018-0197-8 3 Crook B. Bioaerosols 1995 Boca Raton, FL Lewis Publishers 4 Crook B. Bioaerosols Handbook 1995 Boca Raton, FL Lewis Publishers 5 Fuks N. A. Aerosol Meckahincs 1955 6 Bohlmann S. Shang X. Giannakaki E. Filioglou M. Romakkaniemi S. Komppula M. Saarto A. Action and characterization of birch pollen in the atmosphere using a multiwavelength Raman polarization lidar and hirst-type pollen sampler in Finland Atmos. Chem. Phys. 2019 19 14559 14569 10.5194/acp-19-14559-2019 7 Beggs P. J. Davies J. M. Milic A. Haberl S. G. Johnston F. H. Jones P. J. Katelaris C. H. Newbigin E. Australian Airborne Pollen and Spore Monitoring Network Interim Standard and Protocols 2018 8 Tummon F. Arboledas L. A. Bonini M. Guinot B. Hicke M. Christophe J. Kendrovski V. McCairns W. Petermann E. Peuch V. H. Pfaar O. Sicard M. Sikoparija B. Clot B. The need for Pan-European automatic pollen and fungal spore monitoring: A stakeholder workshop position paper Clin. Transl. Allergy 2021 11 e12015 10.1002/clt2.12015 33934521 9 Raynor G. S. Ogden E. C. Haes J. V. Dispersion and deposition of ragweed pollen from experimental sources J. Appl. Meteorol. 1970 9 885 895 10.1175/1520-0450(1970)009<0885:DADORP>2.0.CO;2 10 Blackmore S. Barnes Y. S. Pollen and Spores. Form and Function 1986 London Academic Press 11 Harrington J. B. Kurt M. Ragweed pollen density Am. J. Bot. 1963 50 532 539 10.1002/j.1537-2197.1963.tb07226.x 12 Lacey J. Aggregation of spores and its effect on aerodynamic behavior Grana, No. 1991 30 437 445 10.1080/00173139109432005 13 Golovko V. V. Kutsenogii K. P. Istomin V. L. Agglomerate composition of pollen aerosol in the atmosphere of Novosibirsk Opt. Atmos. Okeana 2014 27 553 559 14 Golovko V. V. Belanova A. P. Zueva G. A. Study of the cluster composition of pollen particles entering the atmosphere during the bloom of anemophilic plants Opt. Atmos. Okeana 2019 32 476 481 10.15372/AOO20190610 15 Golovko V. V. Zueva G. A. Kiseleva T. I. Anemophilous plant pollen grains entering the atmosphere: Cluster composition Atmos. Ocean. Opt. 2021 34 483 490 10.1134/S1024856021050092

0

PMC9753010

NO-CC CODE

2022-12-16 23:25:27

no

Atmos Ocean Opt. 2022 Dec 15; 35(6):673-679

utf-8

oa_other

==== Front Atmos Ocean Opt Atmospheric and Oceanic Optics 1024-8560 2070-0393 Pleiades Publishing Moscow 4404 10.1134/S1024856022060276 Atmospheric Radiation, Optical Weather, and Climate Effect of Meteorological Conditions and Long-Range Air Mass Transport on Surface Aerosol Composition in Winter Moscow Vinogradova A. A. anvinograd@yandex.ru Gubanova D. P. gubanova@ifaran.ru Iordanskii M. A. Skorokhod A. I. grid.459329.0 0000 0004 0485 5946 Obukhov Institute of Atmospheric Physics, Russian Academy of Sciences, 119017 Moscow, Russia 15 12 2022 2022 35 6 758768 7 2 2022 23 3 2022 11 4 2022 © Pleiades Publishing, Ltd. 2022, ISSN 1024-8560, Atmospheric and Oceanic Optics, 2022, Vol. 35, No. 6, pp. 758–768. © Pleiades Publishing, Ltd., 2022.Russian Text © The Author(s), 2022, published in Optika Atmosfery i Okeana. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. We discuss the results from the complex experiment aimed at studying the composition and time variations in urban aerosol in the surface air at the center of Moscow based on daily data on PM10 and PM2.5 concentrations. In addition to these continuous observations every season (for 35–40 days), the total aerosol mass concentration (by gravimetric method) and 65 chemical elements in aerosol composition were measured daily. Winters 2019/2020 and 2020/2021 are considered. The aerosol composition is juxtaposed to the meteorological parameters in the surface atmosphere in Moscow, the direction of long-range air mass transport toward the Moscow region, and the distribution of dust in air over the European Russia (using the MERRA-2 reanalysis data). The detailed analysis of the aerosol elemental composition in Moscow made it possible to identify the elements of global/local spread, as well as of natural/anthropogenic genesis. Concentrations of all aerosol constituents in Moscow during winter did not exceed the corresponding daily average MPC values for the air of residential territories. It is shown that the accumulation of PM10 and PM2.5 in urban air was favored by calm weather conditions. The maximal levels of aerosol pollution were observed in December 2020 during southeasterly winds, when long-range atmospheric transport of admixtures to Moscow occurred from sources located in the southern regions of European Russia, the Caspian Depression, and western Kazakhstan. Keywords: surface aerosol mass concentration elemental composition meteorological conditions long-range transport air mass winter Moscow issue-copyright-statement© Pleiades Publishing, Ltd. 2022 ==== Body pmcINTRODUCTION Aerosol is a highly variable atmospheric constituent influencing both the environmental state and the climate in different natural zones and territories [1–3]. Sources (remote and local, natural and anthropogenic) supply to the atmosphere the primary aerosol, the composition of which differs in both particle sizes and chemical components. Moreover, the atmosphere contains not only aerosol, but also gas constituents, thus favoring the formation of secondary aerosol particles, as well as different physical and chemical conversions of aerosol leading to changes in its chemical composition and particle size distribution. The presence and the efficiency of these processes depend on the intensity of solar radiation, the meteorological parameters of the atmosphere, the variations in the emissions from local sources and in sinks of admixtures, the conditions of their long-range transport by air masses on specific days and in different seasons. Thus, the aerosol study in a big city is a very intricate problem, involving an analysis of many variables that describe the atmospheric properties and, if possible, the composition of emitted substance and locations of natural and anthropogenic sources of diverse atmospheric constituents. There are still not too many scientific publications devoted to studying urban aerosols [4]. Aerosol research has been most active in recent years, both worldwide [5] and in Moscow [6], under the conditions of the COVID-19 pandemic constraints that can, and should, influence the composition of urban air. However, the natural conditions in many countries are very diverse, so that general or, at least, similar tendencies of variations in aerosol parameters can still not be reliably identified against the background of their conventional “background” oscillations, characteristic for specific location and season [7]. Very important information for studying the atmospheric aerosol can be derived from its elemental composition that indirectly indicates the local/remote sources of atmospheric pollution, simplifying the determination of pathways of aerosol supply to an observation site [8]. The sources of atmospheric pollution are determined using certain tracers (concentrations of a number of chemical elements and their ratios), which are characteristic of the composition of a substance emitted into the atmosphere as a result of human activities or natural processes, such as different industries, motor vehicles, volcanoes, fires, dust storms, etc. [9–11]. In recent years, a growing number of works are devoted to studying the elemental composition of aerosols in different regions of Russia (such as, on desertified territories [10–13], remote regions of the Arctic [14, 15], Siberian cities [16, 17], etc. [18]). As regards the atmosphere of the Moscow region, the aerosol elemental composition was studied only episodically here, and the publications on the topic are few [7, 19–25]. In recent years, the staff of Moscow State University has carried out work on comparing the elemental composition of solid particles of soils and roadway dusts to that of the atmospheric aerosol in Moscow and the Moscow region [26–28]. In this work, we analyze the data on aerosol composition in the urban air of Moscow obtained at the Obukhov Institute of Atmospheric Physics, Russian Academy of Sciences (IAP RAS) (certain results from the analysis of these data were presented in [7, 25, 29, 30]). We consider winters 2019/2020 and 2020/2021, which are the seasons with the weakest aerosol pollution of the atmosphere over the two-year observation period. The emphasis is on the analysis of how the distribution of aerosol mass over particle sizes (mass composition) and elemental composition of aerosol vary when juxtaposed to the meteorological conditions and the direction of long-range air mass transport to the Moscow region. INITIAL DATA ACQUISITION AND ANALYSIS During the last two years (from October 2019 to October 2021), IAP RAS has conducted a complex experiment on studying the physicochemical characteristics of the surface aerosol. The observation site (approximate coordinates: 55.74° N; 37.62° E) is located in the IAP RAS yard in the central part of Moscow in a zone of dense urban development not far from moderately busy traffic arteries. Aerosol samples were taken daily at an altitude of ∼2 m above the ground; filters were replaced at 09:00 Moscow time. During the experiment, continuous time series of daily data on the aerosol particle size distribution function and number concentration (in the size range 0.25–10 μm) in the surface air layer were obtained for two years of observations. The well-known formula [1, section 8.1.1. P. 356, formula 8.8] was used to determine the particle mass distribution over sizes. Assuming that the dust and soot particles are the main components of urban aerosol [31], the particle density was assumed to be 1.8 g/cm3 in the calculations. The wetted particle density was assumed to be 1 g/cm3 under the conditions of high humidity. In addition, an intensive complex experiment was carried out for 35–40 days in each season; specifically, the daily samples collected by the aspiration method were used to determine the mass concentration of all filter-trapped aerosol particles (usually not exceeding 40 μm in size) and aerosol elemental composition. In the winter seasons considered here, these experiments were conducted from January 10 to February 14, 2020 and 2021. The aerosol sampling instrumentation comprises: — The aspiration aerosol samplers with analytical AFA-type filters made from Petryanov filtering sheet for a subsequent gravimetric and chemical analysis (for particles with sizes (diameters) D = 0.1 μm, the AFA filter skip factor is 0.1% [32]). — Six-stage impactors with hydrophobic filters for determining the aerosol mass and elemental composition in particle size ranges: < 0.5; 0.5–1.5; 1.5–2.5; 2.5–4.0; 4.0–6.5; and > 6.5 μm. — Laser aerosol spectrometers LAS-P for determining the number of aerosol particles with D = 0.15–2.0 μm every 5 min in 10 channels. — Optoelectronic aerosol spectrometers OEAS-05 for determining the number of aerosol particles with D = 0.2–10.0 μm every 5 min in 10 channels. Different aerosol parameters were determined and analyzed using the following methods and instruments: — Aerosol particle microphysical characteristics, by the methods of laser and electronic spectroscopy. — Aerosol mass concentration, by the gravimetry method or/and through numerical calculation applying data on the particle size distribution [1]. — Elemental composition, by the method of inductively coupled plasma atomic emission spectroscopy (ICP-AES), inductively coupled plasma mass spectrometry (ICP-MS), and through X-ray fluorescence (XRF) analysis [9–11, 33]. — Data statistics obtained by means of standard statistical software for automatic numerical processing and graphical display of information. Internet resources and databases used for analysis and interpretation of results included: — Meteorological parameters of the Balchug weather station closest to IAP RAS available at websites [34–36]. — Back trajectories of air mass transport to the observation site calculated using the HYSPLIT model available at the ARL NOAA website [37]. — Spatial distribution of dust in the surface air from MERRA-2 (version 2) reanalysis data obtained using satellite observations and real monitoring [38]. — Data from State Nature Organization Mosecomonitoring (MEM) network stations [39]. RESULTS AND DISCUSSION Snow cover isolates the underlying surface (soil and partially roads) from the atmosphere in Eurasian midlatitude winter in most continental regions. This strongly mitigates the effect of local natural sources of soil and dust aerosols on the composition of the surface atmosphere. Moreover, low temperatures slow down the formation of secondary aerosol particles from precursor gases, and also nearly block out the emissions of bioaerosols by vegetation. Thus, the surface urban aerosol during winter should predominantly be composed of anthropogenic components of local origin and natural/anthropogenic admixtures of global spread. Our data from the two-year continuous monitoring indicate that winter seasons are the periods when all aerosol components in the surface air of Moscow show minimal concentrations. It is noteworthy that winters may strongly differ in how strongly the weather conditions and long-range atmospheric transport of air masses influence the air composition. General Information on the Composition of Winter Aerosol Figure 1 shows the daily average mass concentrations of РМ10 and РМ2.5 and their variations in surface air at the center of Moscow in winters 2019/2020 and 2020/2021. The concentrations did not exceed the daily average maximum permissible concentration (MPC) (60 and 35 μg/m3 for РМ10 and РМ2.5, respectively) throughout the observation period. As can be seen from Fig. 1, the daily average mass concentrations of РМ10 and РМ2.5 synchronously vary for three months (the correlation coefficients are 0.84 and 0.75 in 2019–2020 and 2020–2021, indicating that the particles in these fractions have the same sources, or that these sources are nearly synchronously activated/deactivated). Fig. 1. Concentration of РМ10 (black squares) and РМ2.5 (open squares) in the surface air in Moscow during winter (December–February) seasons: (a) 2019/2020; (b) 2020/2021; dashed line boxes show the intensive monitoring periods. Comparison of daily mass concentrations of РМ10 and РМ2.5 obtained at the IAP RAS and at the nearest MEM network station (Spiridonovka) at the center of the city showed a good convergence of absolute values and variations in these characteristics. The correlation coefficients between the data from the IAP and Spiridonovka in each of these seasons are ∼0.8 for РМ10 and >0.9 for РМ2.5. The average mass concentrations of aerosol particles in different fractions and the total aerosol mass concentration М (D < 40 μm) are presented in Table 1. The most distinct differences between winters 2020/2021 and 2019/2020 are that the aerosol concentrations in air were much larger in the first half of December 2020; as well as that the mass concentrations of aerosols with sizes from 2.5 to 10 μm (the difference between the РМ10 and РМ2.5 fractions) were small in January–February 2021 (see Fig. 1). Anomalous values are bolded in Table 1. Table 1.   Average mass concentrations (μg/m3) ± SD of aerosol particles with different sizes in diverse time intervals during the winters under study Time interval D, μm М <2.5 2.5–10 <10 >10 Jan. 10–Feb. 14, 2020 4.7 ± 3.7 5.7 ± 2.7 10.4 ± 5.6 10.4 ± 6.7 20.6 ± 9.0 Jan. 10–Feb. 14, 2021 10.7 ± 6.8 2.5 ± 2.1 13.1 ± 8.0 18.4 ± 10.2 32.0 ± 12.6 3 months of 2019/2020 4.7 ± 2.8 6.7 ± 3.9 11.4 ± 5.9 3 months of 2020/2021 10.3 ± 5.9 6.5 ± 8.5 16.9 ± 11.9 Dec. 1–15, 2020 14.4 ± 3.5 23.1 ± 7.7 37.5 ± 10.6 Dec. 16, 2020–Mar. 3, 2021 9.5 ± 5.9 3.4 ± 3.4 12.9 ± 7.6 Meteorological Features of These Winters in Moscow and Associated Differences in Aerosol Particle Size Distribution Winters 2019/2020 and 2020/2021 in Moscow stood out in values of the main meteorological parameters: temperature T, relative air humidity U, and atmospheric pressure p. That was manifested primarily in anomalously high (positive) air temperatures (Table 2) during all three months of winter 2019/2020. The pressure and humidity in that period of time were slightly lower than in the next year, when all three indices were close to normal. The snow-covered period in winter 2019/2020 was just two months (December and January), given that the average (over 2005–2021) snow-covered period is longer than four months in Moscow. As a result, the snow-free soil was unfrozen and sufficiently dry to be a source of aerosol particles already in February. This latter could be why the concentration of aerosol particles with sizes from 2.5 to 10 μm was much larger in February 2020 than 2021 (see Table 1 and Fig. 1). Therefore, the size distribution of РМ10 from December 16, 2020, to early March 2021 should be more typical for winter aerosol in Moscow than the distribution in the preceding anomalous winter season. Table 2.   Monthly average meteorological indices of surface air in Moscow during the seasons under study Season Month T, °С p, mmHg U, % 2019/2020 December 1.3 747 82 January 0.7 748 79 February 0.5 744 75 2020/2021 December −3.9 760 81 January −5.0 749 83 February −8.3 751 75 The wind conditions in Moscow during these winters were characterized by quiet, weak winds; calm conditions were more frequent during winter 2020/2021 than during the preceding winter. However, wind roses in these seasons drastically differed (Figs. 2a and 2b): predominantly westerly winds in 2019/2020 versus a larger percentage of southeasterly winds in 2020/2021. Fig. 2. (a, c) Wind roses and (b, d) mass concentration roses of РМ10 and РМ2.5 particles in Moscow over three winter months: (a, b) in 2019/2020; and (c, d) in 2020/2021 along with the indices of calm conditions: (a, c) frequency, % and (b, d) average concentrations under calm conditions, μg/m3. These differences influenced the size distribution of aerosol mass for different wind directions. In particular, the data in Figs. 2c and 2d show that the mass concentrations of the РМ10 and РМ2.5 in the surface air in Moscow during winter are larger under southerly and easterly winds. Interestingly, precisely winds from southern, eastern, and southeastern rumbs were permanently recorded in Moscow from December 1 to 15, 2020, in the period of maximal aerosol concentrations (see Fig. 1b and Table 1). It is on these 15 days that the calm conditions were recorded in 40% of cases, which could also favor the increase in aerosol concentration in the surface layer when the outflow of pollutants from the city with air masses was weak. On the whole, the statistics for two winter seasons (without accounting for episode with increased aerosol concentration in the first half of December 2020) shows that, when calm conditions occur at a frequency of 12–18%, the mass concentrations of РМ10 and РМ2.5 increase by 4–11 and 7–13%, respectively. Long-Range Atmospheric Transport of Air Masses and Aerosol to Moscow during Winter During winter, the underlying surface is partially covered by snow, plants are defoliated, and air temperatures are often negative; therefore, the processes of removal of admixtures from the atmosphere and deposition onto the underlying surface are slowed down as compared to the warmer season, and aerosol is transported large distances away [40–42]. Air masses passing southward of Moscow, where the snow cover is either absent or very thin at this time, should contain natural (dust, soil) aerosols from the surface of these territories. Moreover, particles, having been formed on other territories and keeping the properties of the initial aerosol, can be carried to the atmosphere of these regions and then entrained into the flow travelling toward Moscow. These formation processes of aerosol composition in Moscow are quite probable considering the long lifetime of submicron and micron admixtures in winter air. Analysis of long-range air mass transport for each winter day in these seasons showed that air masses and aerosol arrive at Moscow from the south of European Russia, the Caspian Depression, and western Kazakhstan when winds blow from the southeast. This is important for identifying such days because standard meteorological data include data on wind direction. Figure 3 presents the distributions of the trajectories of air mass transport toward Moscow and dust trajectories in the surface atmosphere of Moscow (using MERRA-2 reanalysis data [38]) for two days typical for the interval from December 1 to 15, 2020, when aerosol content in the city was increased. The back trajectories of air mass motion to the IAP RAS observation site were calculated at the NOAA’s Air Resources Laboratory (ARL) website [37] (eight 72-hour trajectories per day at an interval of 3 h at an altitude of 100 m above the surface). Fig. 3. Maps of the dust distributions in the surface air and trajectories of air mass transport (black dots) toward Moscow (red star) on December (a) 2 and (b) 15, 2020. Winds in the city were from the southeastern rhombs in all 15 days. From Fig. 3 it can be seen that, for a three-day (duration of trajectories) period, air masses carrying sand and dust from the southeast of European Russia, Kalmykia, as well as from the Ryn-Peski Desert [43], situated in the Caspian Depression and in the west of Kazakhstan, could arrive to Moscow. On these territories, where rainfall is ∼230 mm per year, at negative wintertime air temperatures, the snow cover is very thin and dust storms occur quite often. It can be seen that, in early December 2020, the air masses could carry dust and sand from Kalmykia (northwestern Caspian coast), where dust storms developed on those days (Fig. 3a). We observed and recorded a similar situation (though with heavier aerosol pollution) in Moscow during fall 2020 [30]. By December 15, the trajectories of air mass transport to Moscow slightly shifted, and aerosols were transported from Caspian Depression and from more eastern territories of the Ryn-Peski Desert. On the other hand, the probability of air transport from those areas to the Moscow region in the winter season is estimated to be, on the whole, within 5% [44]. Thus, we recorded quite a Moscow-atypical event of a wintertime supply of arid aerosol from the southeastern regions of European Russia and western Kazakhstan. During those two winters, there were also other days, when aerosol transport to the region was from about the same areas as in the first half of December 2020. However, those were short-term episodes and, as such, could not significantly influence the aerosol concentration in Moscow. As was already noted above, the aerosol pollution in the city from December 16–17, 2020, to the end of the winter 2021 was about the same (see Table 1) as regards the average mass concentrations of the РМ10 and РМ2.5 particles and their variations. Seemingly, these parameters can be considered a conventional “background” level for winter air in Moscow. Aerosol Elemental Composition in Surface Air in Moscow during Winter During the experiment in Moscow, the aerosol elemental composition in each season was determined only during intensive monitoring, i.e., in the periods indicated in Fig. 1. From the figure it can be seen that the general aerosol indices in the periods from January 10 to February 14 little differed between 2020 and 2021. Out of all 65 chemical elements measured, for a more detailed analysis we chose 33 elements (Fig. 4), which are not only terrigenous and/or non-terrigenous elements, but also of global and/or local origin. Fig. 4. Average mass concentrations (μg/m3) and enrichment factors of chemical elements in surface aerosol in Moscow in January–February 2020 and 2021. Logarithmic scale; elements along horizontal scale line up in EF_2021 increasing order. Figure 4 shows the geochemical profiles of the average concentrations (С) and enrichment factors (EF) of chemical elements in the composition of surface aerosol in Moscow during winter 2020/2021. Evidently, these dependences are similar in different years. The enrichment factor was calculated using comparison to the average composition of the Earth’s crust (data were taken from [45]) according to the formula EF = (CX/CLa)aer/(СX/CLa)cr, where СX and CLa are the concentrations of the element Х and lanthanum La (a reference elements of predominantly terrigenous origin); and superscripts stand for aerosol (aer) or the Earth’s crust (cr). All elements considered are divided into predominantly terrigenous (EF < 10) and non-terrigenous (EF > 10). Based on this index, W, Cu, Mo, Hg, As, Zn, Sn, Pb, S, Cd, Sb, Se, and Bi in the surface aerosol in Moscow can be classified as non-terrigenous elements; this group is the same for both winters and has the same non-terrigenous elements as in the group identified in spring 2020 [7]. Elements with 1 < EF < 10 are frequently, and especially in the city, of mixed origin and associated with agricultural activities or with the soil components used in production (construction sites, motor vehicles, production of building materials, etc.). Analysis of variations in the concentrations of different elements made it possible to identify few groups, in which the concentrations of elements varied almost synchronously during winter months considered here. The pairwise correlation coefficients between elements of predominantly terrigenous origin (Mg, Al, P, Ca, Fe, Ba, Sr, Mn, Co, Hf, La, Th, U) were >0.8 in both winter seasons. The adjacent group (with the correlation coefficients from 0.7 to 0.8) contains elements of mixed or local origin: Cr, Cu, Mo, W, Sn, and Sb. In addition, we can single out the high correlations between variations in daily average concentrations for few groups of elements that we conventionally call the groups of sulfur (S, K, Cs, Se, Bi), iron (Fe, Mn, Cr, Co, Zn, Sn), and lead (Pb, Cd, Sn, Sb). Elements of global spread entered to the sulfur group; metals and metalloids of local or anthropogenic origin, to the iron group; and elements mainly associated with motor engine exhausts to the atmosphere, to the lead group [46]. On the whole, these regularities are valid for both winter seasons. The differences in elemental composition of surface aerosol between these winters in Moscow are manifested in changes in the concentrations of elements as functions of the wind direction (and, hence, the direction of long-range transport of air and pollutants to the Moscow region). Unfortunately, because of the vorticity of air flows, wind direction quite rarely indicates the possible remote territories from which air masses come. For instance, as can be seen from Fig. 3b, on December 15, 2020, the long-range transport of air masses was from southeastern regions for westerly winds in the city (these are possible, though rare, situations). The winds in the periods of intensive monitoring were distributed over directions in almost the same pattern as over the three months of the corresponding season (see Figs. 2a and 2b). The wind roses in these winters complemented each other with respect to directions; therefore, we considered two seasons together and plotted the so-called “concentration roses” of certain elements for an “undisturbed” average winter season in Moscow (the episode of long-range dust transport from southeastern regions in early December 2020 was not included in the intensive monitoring periods). These diagrams, averaged for two undisturbed winter periods (Figs. 5a–5d), show how the daily average concentration of each element in the surface air at the observation site changes with the wind direction in winter. It can be seen that the concentrations of such elements as S, Ca, Na, Mg, and P, on the whole, weakly depend on the wind direction in winter at the center of Moscow. On the contrary, the Pb, Cd, and Se concentrations in air at the center of Moscow during winter are higher under southerly and southeasterly winds. Of course, no distinction is made between effects from local and remote sources in such a consideration. Fig. 5. Distribution over wind directions for certain elements in surface air in Moscow during winter 2020/2021: (a–d) concentrations, μg/m3; (e–h) contributions to the average concentration. After taking into account the frequencies of winds from specific directions (wind rose) for the time interval under study (weighting coefficients of the wind direction), we can calculate the contributions to the average (over the winters considered) concentration of an element on those days when the wind blew in a corresponding direction. These “contribution roses” are shown in Figs. 5e–5h. It can be seen that more Ca, P, Mg, and Na is carried to the atmosphere of Moscow during winter with westerly winds, and more S, Pb, Cd, and Se, with easterly and southeasterly winds. CONCLUSIONS Winter is a season with minimal aerosol pollution in Moscow. The average total aerosol mass concentrations was 20.8 and 32.0 μg/m3 in winter seasons of 2019/2020 and 2020/2021, respectively. For comparison: the average (over 2020 and 2021) total aerosol mass concentrations were ∼57, 75, and 46 μm/m3 during spring, summer, and fall. The winter concentrations of all measured constituents in the surface air did not exceed the daily average MPC for residential areas during both winters. Winter 2019/2020 was anomalously warm, with the shortest-lasting snow cover in all years of observations, which possibly resulted in a higher concentration of micron-sized particles (as compared to the preceding winter). On the contrary, winter 2020/2021 was closer to normal in terms of the main meteorological parameters, although the wind rose was characterized by more frequent southeasterly winds. In the second half of December 2020 when winds were from the east and southeast, the concentration of aerosol (and especially of РМ10) was higher than average. The trajectory analysis of air mass transport toward Moscow, as well as the MERRA-2 reanalysis data on the spatial distribution of dust in the surface atmosphere showed that the aerosol sources in both cases were in the southern regions of European Russia, in Kalmykia, the northeastern Caspian region, and western Kazakhstan. It is shown that calm conditions favor the accumulation of the mass concentration of РМ10 and РМ2.5 up to maximal levels in urban air. In the absence of anomalous atmospheric pollutant transport, the frequency of calm conditions in Moscow during winter is 12–18%, and the mass concentrations of РМ10 and РМ2.5 increase by 4–11 and 7–13%, respectively. Study of variations in the enrichment factor (relative to the composition of the Earth’s crust) revealed the elements, similar for both winter seasons, of predominantly non-terrigenous origin: W, Cu, Mo, Hg, As, Zn, Sn, Pb, S, Cd, Sb, Se, and Bi. Correlation analysis of concentrations of different elements makes it possible to single out few more groups. These are terrigenous elements (Mg, Al, P, Ca, Fe, Ba, Sr, Mn, Co, Hf, La, Th, and U), elements of global spread (S, K, Cs, Se, and Bi), as well as two groups of elements of local anthropogenic origin (Fe, Mn, Cr, Co, Zn, and Sn) and (Pb, Cd, Sn, and Sb), seemingly associated with aerosol emissions from industrial plants and with motor engine exhausts. A joint analysis of wind direction and elemental composition of aerosol showed that westerly winds during winter carry to the Moscow atmosphere more Ca, P, Mg, and Na and other elements whose concentrations vary synchronously with the former. Easterly and southeasterly winds carry more S, Pb, Cd, and Se and, hence, more elements the concentrations of which vary synchronously with them. Thus, the study of the formation of aerosol pollution of the atmosphere of Moscow during winter, despite the low level of this pollution, makes it possible to identify important relationships between the absolute content and variations in different aerosol constituents and natural meteorological and synoptic conditions in specific winters. FUNDING This work was supported by the Russian Foundation for Basic Research (grant no. 19-05-50088). CONFLICT OF INTEREST The authors declare that they have no conflicts of interest. Translated by O. Bazhenov ==== Refs REFERENCES 1 Seinfeld J. H. Pandis S. N. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change 2006 New York Wiley 2 Ya K. Kondrat’ev L. S. Ivlev, and V. F. Krapivin, Atmospheric Aerosols: Properties, Generation Processes, and Effects. From Nano- to Global Scales 2007 St. Petersburg VVM 3 Ginzburg A. S. Gubanova D. P. Minashkin V. M. Influence of natural and anthropogenic aerosols on global and regional climate Russ. J. Gen. Chem. 2009 79 2062 2070 10.1134/S1070363209090382 4 Bezuglaya E. Yu. Smirnova I. V. Urban Air and Its Variations 2008 St. Petersburg Asterion 5 Sokhi R. S. Singh V. Querol X. Finardi S. Targino A. C. Andrade M. de Fatima, R. Pavlovic, R. M. Garland, J. Massague, Sh. Kong, A. Baklanov, L. Ren, O. Tarasova, G. Carmichael, V.-H. Peuch, V. Anand, G. Arbilla, K. Badali, G. Beig, L. C. Belalcazar, A. Bolignano, P. Brimblecombe, P. Camacho, A. Casallas, J.-P. Charland, J. Choi, E. Chourdakis, I. Coll, M. Collins, J. Cyrys, C. M. Silva, A. D. Di Giosa, A. Di Leo, C. Ferro, M. Gavidia-Calderon, A. Gayen, A. Ginzburg, F. Godefroy, Y. A. Gonzalez, M. Guevara-Luna, SkM. Haque, H. Havenga, D. Herod, U. Horrak, T. Hussein, S. Ibarra, M. Jaimes, M. Kaasik, R. Khaiwal, J. Kim, A. Kousa, J. Kukkonen, M. Kulmala, J. Kuula, N. La Violette, G. Lanzani, X. Liu, S. MacDougall, P.M. Manseau, G. Marchegiani, B. McDonald, S.V. Mishra, L.T. Molina, D. Mooibroek, S. Mor, N. Moussiopoulos, F. Murena, J. V Niemi., S. Noe, T. Nogueira, M. Norman, J. L. Perez-Camano, T. Petaja, S. Piketh, A. Rathod, K. Reid, A. Retama, O. Rivera, N. Y. Rojas, J. P Rojas-Quincho., R. San Jose, O. Sanchez, R. J. Seguel, S. Sillanpaa, Y. Su, N. Tapper, A. Terrazas, H. Timonen, D. Toscano, G. Tsegas, G. J. M. Velders, C. Vlachokostas, E. Schneidemesser von, R. Vpm, R. Yadav, R. Zalakeviciute, and M. Zavala, “A global observational analysis to understand changes in air quality during exceptionally low anthropogenic emission conditions Environ. Intern. 2021 157 106818 10.1016/j.envint.2021.106818 6 N. Ye. Chubarova, Ye. Ye. Androsova, and Ye. A. Lezina, “The dynamics of the atmospheric pollutants during the Covid-19 pandemic 2020 and their relationship with meteorological conditions in Moscow,” Geography, Environment, Sustainability 14 (5) (2021). 10.24057/2071-9388-2021-012 7 Gubanova D. P. Vinogradova A. A. Iordanskii M. A. Skorokhod A. I. Time variations in the composition of atmospheric aerosol in Moscow in spring 2020 Izv., Atmos. Ocean. Phys. 2021 57 297 309 10.1134/S0001433821030051 8 Rasmussen P. E. Long-range atmospheric transport of trace metals: The need for geoscience perspectives Environ. Geology 1998 33 96 108 10.1007/s002540050229 9 Salvador P. Artinano B. Pio C. Afonso J. Legrand M. Puxbaum H. Hammer S. Evaluation of aerosol sources at European high altitude background sites with trajectory statistical methods Atmos. Environ. 2010 44 2316 2329 10.1016/j.atmosenv.2010.03.042 10 Gubanova D. P. Kuderina T. M. Chkhetiani O. G. Iordanskii M. A. Obvintsev Yu. I. Artamonova M. S. Experimental studies of aerosols in the atmosphere of semiarid landscapes of Kalmykia 2. Landscape-geochemical composition of aerosol particles Geofiz. Protsessy Biosfera 2018 17 18 44 11 Trefilova A. V. Artamonova M. S. Kuderina T. M. Gubanova D. P. Davydov K. A. Iordanskii M. A. Grechko E. I. Minashkin V. M. The chemical composition and microphysical characteristics of aerosol over Moscow and its vicinity in June 2009 and during the fire peak of 2010 Geofiz. Protsessy Biosfera 2012 11 65 82 12 Abdullaev S. F. Maslov V. A. Nazarov B. I. Madvaliev U. Davlatshoev T. The elemental composition of soils and dust aerosol in the south-central part of Tajikistan Atmos. Ocean. Opt. 2015 28 347 358 10.1134/S1024856015040028 13 Artamonova M. S. Gubanova D. P. Iordanskii M. A. Lebedev V. A. Maksimenkov L. O. Minashkin V. M. Obvintsev Yu. I. Chkhetiani O. G. Variations of the aerosol concentration and chemical composition over the arid steppe zone of southern Russia in summer Geofiz. Protsessy Biosfera 2016 15 5 24 14 Shevchenko V. Lisitzin A. Vinogradova A. Stein R. Heavy metals in aerosols over the seas of the Russian Arctic Sci. Total Environ. 2003 306 11 25 10.1016/S0048-9697(02)00481-3 12699915 15 Starodymova D. P. Vinogradova A. A. Shevchenko V. P. Zakharova E. V. Sivonen V. V. Sivonen V. P. Elemental composition of near-ground aerosol near the northwestern coast of Kandalaksha Bay of the White Sea Atmos. Ocean. Opt. 2018 31 181 186 10.1134/S102485601802015X 16 Arshinov M. Yu. Belan B. D. Rasskazchikova T. M. Simonenkov D. V. Influence of the Tomsk city on the chemical and disperse composition of the surface aerosol Atmos. Ocean. Opt. 2008 21 421 425 17 Bortnikov V. Yu. Bukatyi V. I. Ryabinin I. V. Semenov G. A. “Miscophysical parameters and elemental composition of atmospheric aerosol in Barnaul in 2006–2008,” Izv. Altai Gos. Univ. 2009 61 106 110 18 V. N. Lukashin and A. N. Novigatsky, “Chemical composition of aerosols in the near-water surface atmospheric layer of the central Caspian Sea in the winter and autumn of 2005,” Oceanology 53 (6), 727–738 (2013.) 19 Volokh A. A. Zhuravleva M. G. Estimation of technogenic pollution of air in Moscow Izv. Akad. Nauk. Fiz. Atmos. Okeana 1994 30 182 188 20 Ogorodnikov B. I. Aerosol parameters in the atmospheric boundary layer over Moscow Izv. Akad. Nauk. Fiz. Atmos. Okeana 1996 32 163 171 21 Andronova A. V. Iordanskii M. A. Trefilova A. V. Lebedev V. A. Minashkin V. M. Obvintsev Yu. I. Artamonova M. S. Granberg I. G. Comparative analysis of pollution of the ground atmosphere layer in megacities using Moscow and Beijing as an example Geofiz. Protsessy Biosfera 2010 9 42 54 22 Gubanova D. P. Belikov I. B. Elansky N. F. Skorokhod A. I. Chubarova N. E. Variations in PM2.5 surface concentration in Moscow according to observations at MSU meteorological observatory Atmos. Ocean. Opt. 2018 31 290 299 10.1134/S1024856018030065 23 D. P. Gubanova, M. A. Iordanskii, P. P. Anikin, T. M. Kuderina, A. I. Skorokhod, and N. F. Elansky, “Elemental composition and mass concentration of near surface aerosols in Moscow region during unusual weather conditions in the fall 2019,” Proc. SPIE—Int. Soc. Opt. Eng. (2020). 10.1117/12.2575578 24 D. P. Gubanova, N. F. Elansky, A. I. Skorokhod, T. M. Kuderina, M. A. Iordansky, N. V. Sadovskaya, and P. P. Anikin, “Physical and chemical properties of atmospheric aerosols in Moscow and its suburb for climate assessments,” IOP Conf. Ser.: Earth Environ. Sci. 606 (2020). 10.1088/1755-1315/606/1/012019 25 Gubanova D. P. Iordanskii M. A. Kuderina T. M. Skorokhod A. I. Elansky N. F. Minashkin V. M. Elemental composition of aerosols in the near-surface air of Moscow: Seasonal changes in 2019 and 2020 Atmos. Ocean. Opt. 2021 34 475 482 10.1134/S1024856021050122 26 Vlasov D. V. Kasimov N. S. Kosheleva N. E. “Geochemistry of the road dust in the Eastern District of Moscow,” Vestn. Mosk Univ. Ser. 5. Geografiya, No 2015 1 23 33 27 Kosheleva N. E. Dorokhova M. F. Kuz’minskaya N. Yu. Ryzhov A. V. Kasimov N. S. “Impact of motor vehicles on the ecological state of soils in the Western District of Moscow,” Vestn. Mosk Univ. Ser. 5. Geografiya, No 2018 2 16 27 28 N. S. Kasimov, D. V. Vlasov, and N. E. Kosheleva, “Enrichment of road dust particles and adjacent environments with metals and metalloids in eastern Moscow,” Urban Climate 32, 100638 (2020). 10.1016/j.uclim.2020.100638 29 Gubanova D. P. Vinogradova A. A. Skorokhod A. I. Iordanskii M. A. Abnormal aerosol air pollution in Moscow near the local anthropogenic source in July 2021 Gidrometeorologicheskie Issledovaniya Prognozy, No. 2021 4 134 147 10.37162/2618-9631-2021-4-133-147 30 Gubanova D. Chkhetiani O. Vinogradova A. Skorokhod A. Iordanskii M. Atmospheric transport of dust aerosol from arid zones to the Moscow region during fall 2020 AIMS Geoscie. 2022 8 277 302 10.3934/geosci.2022017 31 D. A. Samulenkov, I. N. Mel’nikova, V. K. Donchenko, and M. V. Sapunov, “Lidar monitoring in the study of the air pollution,” Uchenye Zapiski Ros. Gos. Gidrometeorol. Univ., No. 48, 266–280 (2017). 32 AFA filters. https://niki-mlt.ru/article/filtry-afa-analiticheskie-aerozolnye.html. Cited December 7, 2021. 33 Karandashev V. K. Turanov A. N. Orlova T. A. Lezhnev A. E. Nosenko S. V. Zolotareva N. I. Moskvina I. R. Inductively coupled plasma mass spectrometry in the elemental analysis of environmental objects Zavodskaya Laboratoriya. Diagnostika Materialov 2007 73 12 22 34 http://rp5.ru. Cited December 7, 2021. 35 http://www.windy.com/ru. Cited December 7, 2021. 36 WeatherArchive. https://weatherarchive.ru/Pogoda/ Moscow. Cited December 7, 2021. 37 http://www.arl.noaa.gov. Cited December 7, 2021. 38 http://giovanni.gsfc.nasa.gov/giovanni. Cited December 7, 2021. 39 http://mosecom.mos.ru/. Cited December 7, 2021. 40 Bezuglaya E.Yu. Berlyand M.E. Climatic Characterization of the Conditions for Admixture Propagation in the Atmosphere 1983 Leningrad Gidrometeoizdat 41 Byzova N. L. Garger E. K. Ivanov V. N. Experimental Study of Atmospheric Diffusion and Calculation of Diffusion of Admixtures 1991 Leningrad Gidrometeoizdat 42 Vinogradova A. A. Distant evaluation of atmospheric pollution influence on the remote territories Geofiz. Protsessy Biosfera 2014 13 5 20 43 www.karatu.ru/pustyni-rossii/#i-5. Cited December 7, 2021. 44 Shukurov K. A. Shukurova L. M. Source regions of ammonium nitrate, ammonium sulfate, and natural silicates in the surface aerosols of Moscow oblast Izv., Atmos. Ocean. Phys. 2017 53 316 325 10.1134/S0001433817030136 45 V. V. Dobrovolsky, Terrestrial Biogeochemistry. Selected works. Vol. III (Nauchnyi mir, Moscow, 2009) [in Russian]. 46 N. S. Kasimov, Ecogeochemistry of Landscapes (IP Filimonov M.V., Moscow, 2013) [in Russian}.

0

PMC9753012

NO-CC CODE

2022-12-16 23:25:27

no

Atmos Ocean Opt. 2022 Dec 15; 35(6):758-768

utf-8

oa_other

==== Front Soc Indic Res Soc Indic Res Social Indicators Research 0303-8300 1573-0921 Springer Netherlands Dordrecht 3049 10.1007/s11205-022-03049-7 Original Research The Skills Wastage of Contract Court Clerks in China: Assessment and Countermeasures Sun Wen wensun@nju.edu.cn 123 Bai Menghan 676710769@qq.com 4 Gu Rui dg1805006@smail.nju.edu.cn 1 1 grid.41156.37 0000 0001 2314 964X Law School, Nanjing University, Nanjing, 210093 Jiangsu People’s Republic of China 2 grid.506697.b 0000 0004 1767 4896 The Johns Hopkins University Nanjing University Center for Chinese and American Studies, Nanjing, 210093 Jiangsu People’s Republic of China 3 grid.41156.37 0000 0001 2314 964X Yangtze River Delta Cultural Industry Development Research Institute of Nanjing University, Nanjing, 210093 Jiangsu People’s Republic of China 4 grid.41156.37 0000 0001 2314 964X Nanjing University Press, Nanjing University, Nanjing, 210093 Jiangsu People’s Republic of China 15 12 2022 118 8 12 2022 © The Author(s), under exclusive licence to Springer Nature B.V. 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Skills wastage among contract court clerks in China is becoming a concern for the court system, as it raises the turnover of essential personnel. This article explores how and why highly educated recruits (‘talents’) are attracted to, and often leave, this job through interviews with 79 newly recruited clerks of two courts of a provincial capital. The interviews reveal that underemployment—a common cause of dissatisfaction—reflects rigid recruitment policies, an inefficient talent allocation mechanism, and fundamental institutional changes to career progression in this specific occupation. To alleviate the problem, the data collected suggest encouraging employers to adjust their recruitment policies and provide more employment and career guidance. As many of the new recruits leave provincial towns to seek fortune in the capital, and this raises the incidence of under-employment among newly appointed court clerks through fiercer job competition, skills wastage among court clerks could also be reduced by narrowing the gap among cities to make talents evenly distributed across space. Keywords Court clerks Skill wastage Qualitative study Labour market Higher education Overeducation China http://dx.doi.org/10.13039/501100012456 National Social Science Fund of China 19BH151 Sun Wen ==== Body pmcIntroduction According to the theory of human capital, human resources are crucial to a country’s development and most countries pay much attention to talents’ cultivation (Schultz, 1971).1 However, along with the rapid development of higher education and an increasing supply of graduates (‘talents’), as China has experienced over the past three decades, a new phenomenon has emerged: namely, that well-educated and highly-skilled talents are employed in jobs in which their skills and knowledge cannot be fully utilized. This is known as “skills wastage”, “underutilization of skills”, “talents waste” or “brain waste” (Krahn et al, 2000; Liversage, 2009; Yoshida & Smith, 2005). Skills wastage has long been researched in high-income countries where it is now relatively well understood.2 It is noteworthy that this phenomenon is especially important for those with above average qualifications and skills, who are the most productive part of the labour force. The objective of increasing the level of education and skills in the population in an equitable and cost-effective way as a major policy to facilitate economic growth appears now inadequate and may entail the acceptance of some waste in the form of education-occupation mismatches.3 Thus, for policymakers, the emphasis now is not only promoting the popularization of higher education, but also paying more attention to the matching process in the labour market. In the case of court clerks in China, skills wastage has occurred since the reform of their personnel management system. Court expanded its recruitment of court clerks and open the job to non-local graduates, hence the proportion of migrants in court clerks largely increased. Court clerk is an attractive job and the competition for it is fierce, partly fuelled by internal migration. For instance, in Jiangsu Province, the statistics released by an exam training institution show that out of 5,400 people applying for a court clerk job only 795 could be recruited, a 1:7 ratio.4 Along with being a highly desirable job, court clerks experience a high turnover rate and clerks with bachelor or postgraduate degrees have a higher probability to resign from the job (Luo, 2017). Some of the hypotheses advanced to explain the observed turnover include a relatively low income, and poor prospect for career advancement (Luo, 2017; Pan, 2018). Clerks are an indispensible part in people’s courts, as they account for nearly 1/4 of court staff. More importantly, the work of clerks is the premise and basis for trial work. The high turnover rate of clerks has therefore a negative effect on the stability of court system and the efficiency of trial work (Marc, 1977). However, existing studies on court clerks tend to emphasise the deficiencies of the management system but do not apply a holistic approach, where the issue of skills wastage has more prominence. This paper aims at contributing to such analysis and discussion. In particular, we explore the skills wastage in China’s court system from both micro and macro viewpoints using interviews collected from a group of contract court clerks in two courts of a Chinese provincial capital. The interviews were conducted with newly recruited and former court clerks who resigned, as well as with other staffs of the two courts. The article presents the findings and its broader implications before concluding with a set of recommendations. Background China’s Push for Skills Creation The Chinese government realized the importance of talents for economic development and increased the education devotion after the reform and opening-up since the early 1980s, but especially in the 1990s. In 2002, the strategy of ‘Strengthening China through Human Resource Development’ had been established as one of the basic state policies. After its implementation, the higher education system took off. The number of colleges and universities and students enrolled by colleges and universities increased greatly, with the total number of students reaching 443 million—the largest in the world. The enrollment rate of higher education increased from 30% in 2012 to 57.8% in 2021, an increase of 27.8 percentage points.5 Compared with 2010, the number of universities educated people per 100,000 increased from 8930 to 15,467.6 The population with a bachelor’s degree is nowadays 218.36 million. However, even though the growth rate of highly-educated talents has been high relative to the grow rate of China’s GDP, there is still a shortage of talents in many regions and industries. Highly-educated talents are unevenly distributed in China. The proportion of people with college degree or above in the total population in Beijing, Nanjing and Shanghai are 41.98%, 35.23% and 33.87%, respectively. This is far above the national average (15.47%). The corresponding proportions in Guizhou, Guangxi and Qinghai are 10.95%, 10.81% and 1.49%. As a result, the employment competition in developed cities has prograssibely become fiercer, while employers in smaller centres cannot recruit graduates—widening the gap between large and small cities. The uneven distribution of talents also appears in different industries. Nearly half of the top 30 occupations in the ranking list of 100 occupations most short of talents published by the Ministry of Human Resources and Social Security are related to manufacturing, information transmission, software and information technology services and other industries, requiring a higher degree of specialization. However, talents majoring in Law, Accounting and Journalism and Communication are oversupplied. Higher educated talents have higher productivity, but the premise is that they are allocated to matching jobs (Shengde Lai, 1998). So the efficient functioning of the allocation mechanism is critical. In the planned economy era, the government exercised overall control over the use of labour resources. In particular, in the state-owned sector, the employment arrangement of employees is completely dependent on the administrative allocation mechanism. The number and scale of employment of enterprises were determined by the relevant government department. Labour could not freely flow among regions and employers. In nowadays’ market economy, the allocation mechanism maching labour demand and supply responds to the relative underlying forces. Labour can freely flow across regions and enterprises, but this has brought with it the uneven distribution of jobs and a labour market segmentation. While the vast majority of people with lower education or primary education are basically trapped in the secondary labor market, unable to transfer to the primary market, college graduates have more choices. However they tend to prefer staying in central or large cities even if this means having occupations that require less education than what they acquired. This in turn means that graduates compete in the labor market of people with lower education level, contributing to hidden unemployment. Skills Wastage in China China’s recent experience of educational mismatches in the labour market following the expansion of its higher education system has not been unnoticed (Wu and Li, 2021a, 2021b). Skills wastage has negative influences to both individuals and the country, as it leads to lower income and poor returns on education investments (Dollard & Winefield, 2002), as well as wastage of education resources and a loss of potential economic benefits that talents should have created (Wagner & Childs, 2006). Notwithstanding these premises, skills wastage in China differs in some aspects with that experienced in high-income countries. In more developed economies, overeducation often affects skilled immigrants, while in the case of China’s labour market overeducation is a domestic problem. Although China has high levels of internal migration, there is no evidence that non-locals are discriminated against in the process of employment, unlike immigrants in high income countries (Benjamin, 1994). However, it is undeniable that highly-educated people in China are more likely to migrate to larger cities to find high paying or more suitable jobs7 (Sunita, 2005) and their high proportion in larger cities makes them appear as if they are more likely to end up wasting their skills.8 It has been observed that there is a large number of talents migrating to developed cities for better job opportunities, salaries, welfare, and living conditions to name a few. The aggregation of talents in larger cities generates fiercer competition in the local job market and many talents are then unable to find jobs matching their skills and education. For example, according to the Advanced Research Institute of Shanghai University of Finance and Economics, the matching rate between the education level of employees and their jobs is 39.70%, and the incidence rate of overeducation is 24.40%.9 Under such circumstances, the inevitable conclusion is that some of them have to accept jobs with lower requirements and income, and lower rates of return on their education (He, 2009). The literature has not been silent on these outcomes. The first Chinese study on skills wastage dates back to 2006, when scholars studied the phenomenon from different perspectives. Yingming An (2006) holds that such wastage is the result of a poor allocation mechanisms that leads to talent waste. Massive numbers of newly formed graduates migrating to wealthier “upper tier”cities have become a salient feature of China’s recent rapid economic development. This movement is prominent among highly educated talents, and generates their uneven distribution among geographic areas. This in turn results in the paradox that the supply of talents in the most developed areas is excessive while in lesser urban centres is still insufficient. The resulting “overeducation” in the most populous urban centres is therefore viewed as a consequence of China’s higher education growing faster than the rest of the economy. However, despite its rise, many authors believe that overeducation is only a temporary phenomenon, because the proportion of college graduates in the total labor force is still low, and the expansion of higher education remains beneficial to China's economic development in the long run (Lu & Li, 2005; Zhou, 2008). In contrast, some authors view skills wastage as a kind of “recessive unemployment” caused by the labour market segmentation (Gong, 2010; Lai, 2001; Ma & Yue, 2011). This view holds that China's labor market is divided into primary and secondary labor markets across regions, sectors and occupations. Due to the high rigidity of wages and benefits of the primary labor market, college students compete and wait for suitable jobs in these industries, causing their temporary skills underuse. According to Yu and Chen (2006) this occurs as education fails to provide the signal function that enables employers to distinguish different levels of college quality and hence graduates’ potential productivity. This view is shared by other researchers who point out that China’s higher education system does not sufficiently take into account the needs of current professional employers, resulting in large numbers of graduates being unable to find jobs matching their majors and education degrees because of their inadequate training. The Labour Market of Contract Court Clerks Clerks are recuited through an open call since 2003. Before that, there was no public recruitment for clerks, as they were law graduates and assigned by their colleges. With China’s eocnomic growth, the increase of cases also expanded, and with it the demand for clerks. This pushed courts to reform the method of recruitment applied. Under the new recruitment policy, courts could set different recruitment requirements according to their actual and expected needs. At the beginning of the reform, many courts failed to recruit enough clerks due to setting requirements that were only met by law graduates. The courts relaxed this policy, resulting in higher numbers of applicants. However, significant differences in recruitment policies across cities have emerged. Recruitment requirements in large urban centres are more strongent than those of relatively smaller cities. For instance, recruitment announcements issued by the people’s courts in Shanxi, Gansu and Heilongjiang Provinces, require a college degree or above but no limits to the major in which the BA was acquired. In large urban areas, such as Shanghai, Shenzhen and Beijing, clerks are generally required to have a bachelor degree or above and are required to be law graduates, even though the work of clerks across the country bears little differences. Another significant change is that the restriction on applicant’s residence has been cancled, meaning that non-local graduates can apply for the job. Consequently, migrants have been a large part of court clerks in developed cities and cause some problems, such as higher turnover. Besides opening up admissions to a wider number of potential applicants, the management system of clerks was also reformed. From 1979 to 2003, court clerks were employed as "internal employees" and enjoyed the remuneration and promotion stipulated by the state.10 They followed the career path of court clerk acting judge judge. After 2003, a contract-based system for court clerks was introduced, and as a result court clerks were no longer internal employees of the court. The Measures for the Administration of the Clerks of the People's Court (for Trial Implementation) (hereinafter referred to as "Measures") formulated by the Organization Department of the Central Committee of the Communist Party of China (Organization Department, CCCPC), Ministry of Personnel and Supreme People's Court established contractual management for clerks and divided court clerks into ‘internal’ and ‘contract’ ones. According to the Measures, ‘contract’ court clerks belong to ancillary personnel who need to work under the guidance of judges. The newly reformed management system blocked their possibility to be promoted to assistance judge, curbing their promotion opportunities. Additionally, contract court clerks do not enjoy civil servant benefits, unlile internal employees, and so their income is barely higher than the local minimum wage. These reforms aimed to improve the utilization rate of human capital and the management level, but their result has been a higher turnover rate of contract court clerks. For instance in a certain Province, which we will not name, in the past three years, each court lost an average of 11 contract court clerks each year (Qi, et al., 2022). Most leaving clerks were previously recruited as contract court clerks (Luo, 2017). Method and Data Qualitative Method A qualitative approach is used as it suits investigating the experiences, feelings, opinions and social worlds of research objects (Fossey et al., 2002) and it provide a “deeper” understanding of social phenomena (Silverman, 2020). Court clerks with different educational backgrounds were motivated by various reasons to apply for the job. Hence, we decided to apply semi-structured interviews to analyse their experiences. This method is portrayed as an indispensable tool to uncover knowledge through interaction, conversations, and subjects from different life experiences. Moreover, the shared stories and life experience about other matters are interpreted to expand the knowledge into multiple platforms which could broaden our knowledge of the job (Chen, 2000). In order to have a comprehensive understanding of the relevant information about skills wastage, we recruited 87 participants who they were divided into four groups: specifically, 79 newly recruited clerks, 4 former clerks, 2 personnel department staffs and 2 judges. Different interview outlines have been prepared for different groups and adjustments to the interview protocols were made according to early experience and information provided by participants. We spent 20 to 40 mins on each interview over the course of two months. All interviews were conducted via videoconferencing. They were audio-recorded and transcribed, and only the primary researchers had access to the raw data. The research process was engaged to protect participants’ anonymity and to avoid any impression of coercion, as per the ethics’ approval. Participants were informed about the focus of the study and how we planned to use the data. Informed consent was obtained for all interviews discussed in this paper. Participants’ Characteristics All participants were from two courts in a provincial capital. Namely, the Provincial Higher People’s Court (PHPC) and the Municipal Intermediate People's Court (MIPC). Due to skills wastage generally occurring in the largest cities, we purposely selected the courts in this new first-tier city, located in Yangtze River Delta and attractive to university graduates. Additionally, there is a large number of migrant talents come to City A in recent years. The two courts have implemented the personnel management system reform of court clerks and established new recruitment policies based on a contractual management system. The two courts deal with issues at different types of issue, hence revealing the extent of skills wastage. Of the 79 newly recruited contract clerks participating in this study, 26 were recruited by the PHPC and the other 53 were recruited by the MIPC, as shown in Tables 1 and 2. They all have no prior experience in being a court clerk.Table 1 Information of the contract court clerks recruited by the PHPC: Number of recruits 26 Place of domicile City A in Province J 6 23.1% Other cities in Province J 9 34.6% Other cities outside Province J 11 42.3% Educational background Bachelor's degree 25 96.15% Master's degree 1 3.85% Major Bachelor of Laws 6 23.08% Bachelor of Linguistics 6 23.08% Bachelor of Arts 1 3.85% Bachelor of Economic Management 9 34.61% Bachelor of Other undergraduate majors (software, civil engineering, nursing) 3 11.53% Master of Sociology 1 3.85% Table 2 Information of the contract court clerks recruited by the MIPC Number of recruits 53 Place of domicile City A in Province J 9 16.98% Other cities in Province J 16 30.19% Other cities outside Province J 28 52.83% Educational background Degree obtained from a junior college 20 37.74% Bachelor's degree 32 60.38% Master's degree 1 1.89% Major Bachelor of Laws 19 35.85% Law degree obtained from a junior college 19 35.85% Bachelor of Arts 2 3.77% Bachelor of Economics and Management 7 13.21% Other undergraduate majors (software, civil engineering, nursing) 4 8.54% Other degrees obtained from a junior college 1 1.89% Master of Economics 1 1.89% The “place of domicile” of contract court clerks is based on their “parents’ place of domicile” Among the 79 clerks, 44 (19 junior college graduates and 25 undergraduates) studied law-related majors including law, judicial management-related majors, and security administration-related majors. The other 35 (1 junior college graduate and 34 undergraduates) studied non-law majors including art, economics and management, and even nursing, civil engineering. The proportion of new clerks with a bachelor degree or above is 74%. There are 2 postgraduates majoring in sociology and economics, respectively. As the PHPC only recruited clerks with bachelor degree or above, the 20 non-locals are all university graduates. There are 29 clerks with bachelor degree or above among the 44 non-locals in the MIPC. Among the 26 newly recruited clerks of the PHPC, 20 were non-locals, accounting for 77.9%. At the same time, 44 of the 53 clerks hired by the MIPC were from other cities, accounting for 83%. Interviews with the clerks focused on three topics: (1) the reasons why they choose that occupation; (2) how their major and education degree affected their work as clerks; and (3) whether they felt that their education was a suitable match for the requirements of the job and if not, why. We also recruited former clerks to gain a better understanding of their experience and feelings in the job and the exact reasons for resigning. Four former court clerks were interviewed. These were selected according to their education (must have a bachelor’s degree or above) and major (whether or not they majored in law). Many former clerks have changed their contact information and were not easy to find. We were fortunate that the four recruited ones not only could be contacted, but were also willing to share their experiences. Those former clerks include a balanced mix of undergraduates and postgraduate degree holders (2 each, respectively). Among these four clerks, only one person’s major is not law but a postgraduate major in finance. We also contacted personnel department staffs, who are familiar with information about court clerks and recruitment policies, to get their views on skills wastage. Two department staffs were interviewed. They are mainly responsible for recruiting and managing clerks. We also reached out to judges, as they work with court clerks and can assess clerks’ performance at work. We were particularly interested in understanding whether clerks’ education degrees and majors influence their work. We also sought judges’ views about whether they felt that clerks’ skills were wasted. We approached experienced judges who entered into the courts before 2003 as they have a deeper understanding of the changes brought about by the clerk management reform. Table 3 summarises some key detail about all the interviewees.Table 3 Details of interviewees Identity Majors Academic qualification Number Interview purposes In-service contract court clerks Law Bachelor 25 Learn about the reasons why they apply the job, whether their skills are utilized and whether they are satisfied with the job, and if not, the reasons Degree obtained from a junior college 19 Others Degree obtained from a junior college 1 Bachelor 32 Master 2 Resigned contract court clerks Law Bachelor 2 Learn about the reasons why they resigned the job Master 1 Others Master 1 Personnel department staffs 2 Learn about the information about the court clerks and the reasons for resignation Judges 2 The assessment of the contract court clerks system and the performance of contract court clerks Total amount 87 Data Collection and Analysis Interviews were anonymized and each participant was given a code number. We collected data through audio-recording and chart extraction. After transcribing the interviews, we closely examined the data collected and summarised participants’ views, knowledge and experiences. We also extracted insightful sentences and opinions from the data. The coding work started with existing research findings and was organized around two main topics: (1) situations in which clerks’ skills are wasted; (2) causes of such wastage. Then responses were coded under different themes. The themes emerged with respect to (2), which is the focus of the paper, can be categorized into 4 key areas:labour market: whether the labour market effectively matches supply and demand for talents; higher education: whether highly-educated graduates meet the quality and quantity requirements set by the labour market; recruitment policies: how do recruitment policies for clerks affect the quality of the recruited clerks; personal factors: how clerks’ personal initiative underpins their future career trajectories. Results Skills Wastage The collected data show that migrants constitute a high proportion of court clerks and they generally have higher education degrees in the 2 courts. As we have learned from the personnel department staffs, there are about 7–8 non-locals resigning the job every year since 2018, accounting for almost 72% of the total resignations. Among the reasons why they leave the job, the most significant ones are as follows: limited promotion channel, low salaries, instability and so on. Clerks who are not satisfied with the position look for other jobs or prepare for National Judicial Examination to change the current situation. Some return to less developed cities because of the high expenditure and the pressure of life in developed cities. The rising turnover of clerks reflects that there is a skills wastage among them. Due to migrants take a large part of court clerks, skill wastage is especially obvious among migrants. The information collected from the clerks indicates that most of the newly recruited clerks have a bachelor degree or above, and/or many of their majors are irrelevant to their duties. 37 newly recruited clerks believed that they could have been employed in jobs with higher education degree requirements and all of the 37 clerks are with bachelor’s degree or above. The 4 former clerks found better jobs after resigning: two became lawyers and the two others became civil servants. One of the former clerks said “I left the job largely for the reasons of low income and limited career development. It is proved that I can find a better job with higher income. The most important thing is that I can fully utilize my professional knowledge.” (fc1) This conclusion is shared by many current clerks who feel underemployed and therefore are not satisfied with their job. The lack of statisfaction is a major influence in seeking an alternative occupation. According to the Measures, the duties of clerk include: (1) handling routine work in the process of pre-trial preparation; (2) inspect the appearance of litigation participants at the court and announce court discipline; (3) take charge of the record work; (4) sorting, binding and filing material; (5) complete other routine work assigned by judges. Based on the above, the work of the clerk is complex, but not particularly needy of higher education workers. According to the Migration Policy Institute, the occupation ‘court clerk’ should be classified as a middle-skilled job that does not require a bachelor's degree.11 This requirement is at odds with the learning outcomes of law undergraduates, who are expected to be legal professionals with the ability to independently acquire and update knowledge related to laws; have basic skills to integrate professional theories and knowledge learned and apply them flexibly and comprehensively in their professional practice; employ creative thinking methods to perform scientific research and innovative and entrepreneurial practices; possess excellent skills in computer operation and foreign language. Following the requirements of China's unified qualification exam for legal professionals, only law graduates (but not junior college graduates) can participate in the selection process. Once they have obtained the legal professional qualification certificate, they can become lawyers or sit for the recruitment exam for internal assistant judges. However, these higher-educated talents are often competing for lower requirement jobs that traditionally target vocational college graduates. Jobs with tasks that are ancillary to those of judges. As one of the current clerks said “What I learned in college is useless in this job.” (c1) A comment shared by most of the interviewed clerks (including law and non-law graduates): “Even though my major is not law, I can get the skills that the job needs after the induction training.” (c16) Non-law clerks account for 44.3% of the two courts and none was fired at the end of the probation period, meaning that the job has little relationship with one’s major.12 Both of the interviewed judges held the view that what really matters is not legal professional knowledge, but work experience. It is as Dewey said that we should use our past experience to develop new and better experience for the future, and rationality should be prompted and tested in experience, and applied in every way possible to expand and enrich it through invention (Dewey, 2006). The more opportunities and time the court provides for skills training, the more quickly the "labour skills" of clerks get improved. Since 2016, the "Provincial Plan" of the location of the two courts has required that all court clerks must participate in an induction training and only those who pass an assessment can formally start their work. Moreover, before being promoted to higher posts, clerks must now undergo formal promotion training. Generally, after one-month training, the new recruits can deal with all the five types of "routine work" required by the Provincial Plan and they can even assist judges or judge assistants to proofread the judgment documents. Based on the interviews, clerks’ skills wastage mainly occurs in two situations: one is the clerk’s overeducation; the other one is that the clerk’s major being mismatched for the job. The two cases are acknowledged as common causes of skills wastage (McGuinness, 2006). Compared to the major mismatch (about 44.3% of clerks), the case of overeducation (about 74.68%) is more serious. The overeducated clerks devote more time and money on education, while the payback is the same as the one received by those with lower education. The result reflects evidence discussed in a report released by Shanghai University of Finance and Economics, whereby in for the same education level, the wages of overeducated employees are lower than those of correctly matched ones. In other words, the labour market discounts overeducation. Underuse of Court Clerks The interviews revealed that the four factors that affect the utilization of clerks’ skills are interrelated. From a macro-viewpoint, talents’ supply and demand and talents’ allocation mechanism are the three fundamental elements that influence the s employment outcome. Higher education policies determine the quality and quantity of talents provided by colleges and universities and thus affect talent supply. The efficiency of labour allocation mechanism determines whether these talents can be effectively allocated to matching vacancies. From a micro-viewpoint, employers as job providers have more power over employees: they can raise hiring requirements as a consequence of degree inflation, even though it may not lead to higher productive efficiency. Career choices also reflect an individual’s preferences and as such it is influenced by personal circumstances, including family obligations. These factors jointly condition if the job filled matches one’s education degree and skills.• Influencing factors of skills wastage: the higher education system For Chinese people, improving the level of education is a necessary way to increase incomes (Lai, 1998), and as to Chinese government, improving the level of national education is the key to economic development. However, with the popularization of higher education, employment difficulties of college students emerged, especially in large cities. Undergraduates are no longer a scarce resource in developed cities, let alone college students. “It seems that having a bachelor's degree is universal throughout China and it has been a necessity for job searching. When looking for a job, my degree has no advantage.” (coded under “overeducation”, c10) 58 (including both junior college graduate and university graduates) among the 79 newly recruited clerks mentioned this problem. This could be viewed as too many university graduates crowding out the local labour market and making competition for highly desirable jobs even fiercer. Under this circumstance, many people with ordinary educational background prefer to apply for jobs with lower competitive pressure (Li, 2017). Many interviewees also shared that “It is difficult for me to find a job matching my major.” (coded under “gap between education and the job market”, c28) The interviewees whose majors are art, education and civil engineering found that the job markets for their majors are limited, and can hardly find major-matching jobs. Nowadays, majors in philosophy, literature, history, science, agriculture, management and military science are basically in a state of “oversupply” in China’s labour market. Another theme that emerged is the quality of graduates. A staff member of personnel department said “Junior colleges set the major of legal affairs to cultivate professional clerks for courts, but the graduates lack practical experience and need to accept vocational training as others.” (p2) Junior colleges set the law major as a requirement to specifically cultivate professional clerks for courts, but the graduates show no special predispositions in the job. The two judges mentioned it in their interviews as “Compared to higher educated talents, the working ability of junior college graduates (with vocational education) is lower” (j1,2). Mo Rong (2022) believes that this contradiction is the main problem about the recent employment conditions of university students: the higher education sector in China expanded rapidly in a short period time but the quality (viewed as the set of notions acquired by students) of education has not kept pace with it, resulting in “high education degree, low working capability” (Wu, 2016).• Influencing factors of skills wastage: the labour market Like any commodity market, the labour market is subject to hysteresis and blindness that cannot efficiently adjust the relationship between supply and demand of new graduates. As a result, talents graduating from colleges and universities do not perfectly match employers’ demand. The interviews revealed a number of underlying factors at work. First, the degree of informatization is not sufficiently high as so economic transformation and industrial upgrading are hardly reflected in the labour market in a timely manner. This delayed signal may mislead investments in education especially for university students. “I didn’t know the employment situation before I applied for the major and it last until I began to look for a job.” (c34) Higher education lasts 3 to 4 years but demand for talents may change in the period of study, while colleges and universities programs cannot adapt for sudden changes. Secondly, China’s proportion of the third industry kept rising and this offered opportunities for new graduates. But in recent years, this growth has slowed, while talents with tertiary-relevant majors still kept increasing following major reforms in the higher education sector. Parts of the current mismatch reflects the adjusting industrial structure. Thirdly, the Chinese labour market’s segmentation by different occupations and regions contributes to an uneven spatial distribution of talents. A migrant said that“Most of my university classmates moved to Beijing and Shanghai which pushes me to find jobs in big cities.” (c20) Numerous talents congregate in several cities looking for jobs, but vacancies are limited and this raises job competition and the likelihood of mismatch when landing a job. Despite this, college graduates still chase jobs in China’s largest cities (Wang, 2010).• Influencing factors of skills wastage: recruitment policies According to the public recruitment information of the PHPC and the MIPC, the two courts do not impose restrictions based on graduates’ majors. Among the 63 courts in the Province under study, only 3 impose specific requirements. The absence of a restriction on majors gives graduates across different fields of study the possibility to apply for court clerk jobs, resulting in heightened competition. In 2021, the application-to-admission ratio of the PHPC and MIPC was 14:1 and 12:1, respectively. Therefore, employers can be picky, especially in the largest urban centres. “Although there is no clear requirement for academic qualifications, it actually has been taken into consideration.” (coded under “improved recruitment requirements”, p2) The Personnel department staff interviewed pointed out that the job competition has become fiercer and many talents with bachelor’s degree or even above apply for clerks’ jobs. Given the information asymmetry between employers and job seekers about the quality of candidates, it is difficult for employers to judge the productivity of job seekers in advance. For this reason, job applicants try to enhance and signal their work ability through their educational background, for instance by undertaking studies in prestigious universities.• Influencing factors of skills wastage: personal factors Personal preferences play an important role in career and locational preferences. According to “The 11th survey report on the best employers of Chinese College Students”, more than half of the college students expect to find jobs in government institutions (government, public institutions, state-owned enterprises).13 In the ranking of college students' pursuit of different career development goals, first is a job that is respected and recognized by society, followed by pay and stability and security. “In fact, there are some other jobs that are more compatible with my major, but working in a court is more decent and stable.” (coded under “preference for stable and decent work”, c35) 59 of the interviewed clerks mentioned this aspect (74.68%). Clerks with a university degree knew that their skills would be wasted in applying for a clerk position. Yet, they still chose to work in this job implicitly accepting what could be viewed as “voluntary skills wastage”. In Chinese tradition, working in government, public institutions and state-owned enterprises is desirable as it is stable and employees do not worry about losing their job or face pay cuts. This is valuable, especially in the period of economic downturn. “I just want to live in the city.” (coded under “motivation of living in big cities”, c15) Even though many highly-educated graduates migrate to large cities to find a better job or earn more, there are some migrants just motivated by enjoying a more varied life. Many belong to upper-middle class families. “My parents said that they hope me to find a stable job and they will give me financial support.” (c48) Living in the city is the prime consideration to these people and they do not need to worry about income. Under such circumstances, becoming a court clerk is attractive. The uneven flow of talents into larger regions and industries and the appeal exerted by government jobs, despite its limited recognition of individual productivity, contributes to raise overeducation but also reduce the amount of human capital flowing to productive sectors, hence hindering innovation. Discussion There is a skills wastage among court clerks, but according to the interviews, the most obvious characteristic of clerks is that it is a job related to the courts, a government position, and as such it attracts many graduate applicants. “My parents want me to work in government organs, but it is too difficult to pass the civil service exam. Most of my classmates also choose to prepare for the exam in the last year in university.” (c76) Influenced by parents, a large number of graduates take government jobs as “iron bowls”, with no possibility to be fired. Although some jobs require low skills or education, as long as they are provided by government, they will be especially popular. “Even though being court clerk is no more a pathway to become a civil servant, it is a relatively stable job compared to those in private enterprises. Under such a challenging employment situation, becoming court clerk is not a bad choice.” (c61) Faced with the increasingly stern situation of employment, contract posts in government organs are popular among university graduates. This does not mean that all jobs sought by talents will necessarily lead to skills wastage. As highlighted previously, a clerk is a middle-skilled job that most people can do properly through an induction training. Additionally, as the probability of skills wastage in large cities is higher than that in small cities, the recruitment policies of clerks vary a lot geographically. As the work content of court clerks is essentially the same, the education degrees of clerks varies greatly between large and small centres. Furthermore, “in recent years, the number of university graduates has increased sharply, but their personal qualities are uneven. Even though there are lots of law graduates applying for the job every year, many of them, especially junior college graduates, show no particular advantages relative others.” (p1). As a follow up, one of the judges told us “Universities and colleges pay too much attention to theories and ignore the importance of practice which leads to graduates being unadaptable to work. What is even worse is that many graduates haven’t participated in internships due to the coronavirus pandemic.” This comment highlights a disjunction between the education system and labour market demand, which an ensuing the waste of educational investments (Ma, 1995; An, 2007). The labour markets of some majors have become saturated or have no need, but colleges and universities still enroll a lot of students every year with no concern for the fact that these students will likely encounter skills wastage once they graduate. “I find that more and more graduates of non-law majors, such as art, economy and management, apply for the job. Of course, their universities are usually not good.” (p2) Another problem is that the curriculum design of junior colleges is unsuited for today’s employers’ needs, as the interviewed judges pointed out that college graduates are often incompetent for the job they carry out. As for the demand side, different industries have different absorptive abilities for college graduates, and this makes many college graduates of certain majors unable to find jobs that match their field of study. An under-researched but relevant aspect is that the first job after graduation generates an important signal that eases or prevents career development.“Seniors remind us that if our first job is in economically backward areas, it will be difficult to find jobs in big cities in the future. Therefore, most of us decide to move to big cities, even though the employment competition is fiercer.” (c33). Finally, notwithstanding the various reasons underpinning skills wastage presented so far, it should be noted that faced with an increasing supply of highly-educated people in some occupations, employers find it even more difficult and costly to screen and select suitable employees. Holding a degree appears to be no longer sufficient to carry out an effective screening. As stated by an interviewee, “we clearly know that the job doesn’t need talents with high education degrees, but if the requirement set is low, there will be even more graduates applying for the job and this increases our recruiting burden. The highest efficient way to recruit clerks is through education background.” (p2) but reliance on education enhances credentialism and the acquisition of degrees and qualifications without a real need for them.14 In other words, it incentivises further skills wastage. This is cogend for the court system, where some of the investments on recruitment and induction training is unnecessary. “The previous recruiting policy of court clerks seems to be more suitable when that courts could establish a long-term cooperation relationship with colleges and students could be cultivated with clear purpose according to the court’s needs.” (p1). Last but not least, from an individual perspective, when highly educated talents take stability as their primary consideration in seeking employment, the vitality and innovation of the whole society may be lost. Legal study is quite popular in China and the number of law graduates is increasing in recent years. Although the underutilized clerks are not satisfied with the job, the experience of being a court clerk still means a lot to career development. Some law graduates take it as a way to occupy legal resources, accumulate working experience and a transition to other better jobs. That explains despite of the rapid flow of court clerks, there are adequate supply of talents for it. Many graduates think that it is not necessary to choose a job matching their majors (Liu & Li, 2007), and universities seem unable to guide them. “I had no idea how to find jobs after the failure of civil service exam. I even didn’t know what court clerks are before I decided to apply for it. What I learned in university is only knowledge in books.” (c45) Universities hence share some responsibilities for the status quo. Conclusions Skills wastage has been more common than thought in China. This study focuses on a group of contract court clerks to explore how and why their skills are wasted. Clerk is a middle-skilled job that does not need high skills or knowledge but even though many junior college students are specifically trained for this job, more highly-educated graduates are recruited in the job every year. The reform of court clerks management system is aimed to establish a professional and stable team, but the result is opposite to it. Under the reformed policies, courts have recruited more migrant talents which accelerates the flow of court clerks. The high turnover supports that there is a skills wastage among clerks. Based on the information collected, the causes of such job mismatch reflect an imperfect signal between employers and universities about what needs are required to be suitably employed. Lags and lack of information result in the rise of credentialism, whereby employers in large cities raise the formal requirements for hiring while talents migrate there regardless of whether they will find a suitable job. Such blindness hinders the talents to find matching jobs. Personal preference plays a vital role in the skills wastage. Highly educated talents show a bias to developed cities and affected by traditional views, they prefer a job with high prestige and security. It is partly blamed to universities for that they do not take their due responsibilities to guide students in searching for jobs. The large scale of migration of graduates from small to larger city centres increases the employment pressure. It's hard to find satisfied work and they have to lower their job criteria which may results to the phenomenon of skills wastage. Therefore, the migration to developed cities is not a promise of a better future. Because skills wastage is harmful both to individual and society, it deserves more attention and understanding to be addressed through policy tools. For courts, it is suggested as following: (1) Widen the promotion channel for contract court clerks and allow those who have accumulated trial experience to be directly promoted to assistant judge through internal assessment; (2) Improve the salary and welfare of contract court clerks; (3) Establish cooperation between courts and junior colleges and only graduates specifically cultivated by the colleges are qualified to work as court clerks. To be more wildly applied to, suggested approaches include (1) reforming the education system to strangthen the link between universities and professional circles; (2) promoting economic and industrial structure upgrading to create more jobs for people with higher education; (3) encouraging employers to move away from the employment orientation of hiring graduated from "only famous schools", and establish a talent utilization mechanism targeted by actual job demand; (4) providing graduates with more employment guidance to help students make a more systematic and scientific plan for their career. With the help of professional ability tests, comprehensive evaluation should be carried out in combination with students' personality, ability and academic achievements, students can better know themselves and clarify their goals, career interests and potential. 1 The definition of talent in The National Medium- and Long-term Talent Development Plan (2010–2020) is as follows: "talent refers to people who have certain professional knowledge or skills, perform creative work and make contributions to society, and are workers with high ability and quality in human resources. In this paper, people with bachelor’s degree and above are referred to talents. 2 Freeman (1976) firstly argued that too many graduates relative to the number of jobs available would lead to several underutilized graduate workers. McGuinness’ (2006) literature survey found that the proportion of mismatched employees is quite common, affecting between 11 and 50 per cent of the employed labour force. 3 As the Australian National Engineering Taskforce (ANET) concluded, “The causes of the acute skills shortage and the impediments to resolving the capacity crisis are largely structural, and beyond the control of individual organisations: they are rooted in the nature and patterns of demand set by governments, the complex market structure…” (Wise et al., 2011). 4 2021 Statistics and analysis of the number of applicants for Jiangsu Court's recruitment of clerks, the competition ratio reaches 27:1 at the most, https://js.offcn.com/html/2021/04/255132.htm 5 See 240 million people received higher education in China, The People’s Daily, http://www.gov.cn/xinwen/2022-05/21/content_5691565.htm, last visited on September 25th, 2022. 6 National Bureau of Statistics, Main Data of the Seventh National Population Census, http://www.stats.gov.cn/tjsj/zxfb/202105/t20210510_1817176.html, visited on September 20th, 2020. 7 For instance, according to the figures released by Mycos (2021), the proportion of non-local university graduates in China’s largest (‘first-tier’) cities has been about 68% for during 2016–2020, and in newly-ranked first-tier cities it has risen from 32% in 2016 to 38% in 2020 (Mycos, 2021). 8 According to the Ranking talent attraction in Chinese cities: 2021 released by Zhopin Ltd., 56% of floating talents have bachelor's degree or above, higher than 47% of the total job seekers, so highly educated talents are more likely to seek jobs across cities. In 2021, the proportion of fresh graduates, masters and above who put their resumes in first tier cities respectively reaches to 20.7% and 30.0%, both higher than the proportion of floating talents flowing to first tier cities. Recent graduates, masters and above will prefer to gather in first tier and second tier cities, especially those with masters and above who prefer to gather in first tier cities. 9 See Annual Report on China's Macroeconomic Situation Analysis and Forecast (2020–2021), Advanced Research Institute of Shanghai University of Finance and Economics. 10 Internal employees refer to the employees subject to the staffing of public institutions "enjoy the salaries and benefits as prescribed by the state. 11 The Migration Policy Institute (MPI) assigns jobs to three skill levels: a. High-skilled jobs require at least a bachelor’s degree; b. Middle-skilled jobs require some postsecondary education or training (i.e., an associate’s degree or long-term on-the-job training or vocational training); and c. Low-skilled jobs require a high school degree or less, and little to moderate on-the-job training (Lofters, 2019). 12 Collins pointed out pointedly: “For most jobs, most skills, including the most advanced skills, are learned at work or through informal networks, and the education system is only trying to standardize the skills learned elsewhere.” The skills needed by clerks especially prove it. After China set up the post of assistant judge, court clerks have been mainly engaged in stylized and routine work. Due to the high reproducibility of the work, even non-law majors can perform well as a court clerk within half a year. With the exception of court transcription, which has high requirements for stenographic skills and the comprehension of legal language, the non-law majors are able to do a good job (Mo & Wang, 2018). 13 The 11th survey report on the best employers of Chinese College Students, http://www.docin.com/p-689668294.html 14 It has also been defined as "excessive reliance on credentials, especially academic degrees, in determining hiring or promotion policies." Credentialism occurs where the credentials for a job or a position are upgraded, even though there is no skill change that makes this increase necessary (Buton, 2019). Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References An Y. M. (2007). Analysis on the phenomenon of talent waste in China, People’s Forum, Z1. Baker B The Performance of Immigrants in the Canadian Labour Market Journal of Labour Economics 1994 12 369 405 10.1086/298349 Chen X.M. (2000). Qualitative Research in Social Sciences. Educational Science Press,.182–193. Dewey. (2006). The Transformation of Philosophy, in The Collected Works of Dewey, China Social Sciences Press, Beijing, 86, 114-115. Dollard MF Winefield AH Mental health: Overemployment, underemployment, unemployment and healthy jobs Australian e-Journal for the Advancement of Mental Health 2002 1 3 1 26 10.5172/jamh.1.3.170 Fossey E Harvey C McDermott F Davidson L Understanding and evaluating qualitative research Australian and New Zealand Journal of Psychiatry 2002 36 6 717 732 10.1046/j.1440-1614.2002.01100.x 12406114 Gong XY College Students' Difficulty in Employment: From the Perspective of Wage and Welfare Rigidity Chinese Talents 2010 22 169 170 He YH The Changes of the Rate of Return to Education: An Empirical Study Based on the Data of CHNS Chinese Journal of Population Science 2009 2 44 54 Krahn H Derwing T Mulder M Wilkinson L Educated and underemployed: Refugee integration into the Canadian labour market Journal of International Migration and Integration 2000 1 1 59 84 10.1007/s12134-000-1008-2 Lai DS Segmentation of Labor Market and Graduate Unemployment Journal of Beijing Normal University (humanities and Social Sciences Edition) 2001 4 69 77 Lai S Education. Labour Market and Income allocation Economic Research 1998 5 42 49 Li C. A. (2017). Degree Inflation Is Waste of Knowledge, Global Times, https://opinion.huanqiu.com/article/9CaKrnK5YRV, last visited on 17th Sptember, 2022. Li X. G. (2021a). High Academic Qualification but Low Requirements of Jobs: Did Our Education Mismatch the Job Requirements? GuangMing Daily, 25th March, 2021a, https://epaper.gmw.cn/gmrb/html/2021a-03/25/nw.D110000gmrb_20210325_5-02.htm , last visited on April 6, 2022. Li WF Research on Dynamic Management Mechanism of Organization Establishment Research of Administrative Science 2021 8 12 4 Liversage A Vital conjectures, shifting horizons: High skilled female immigrants looking for work Work, Employment and Society 2009 23 1 120 141 10.1177/0950017008099781 Lu HM Li CG Journal of South-Central University for Nationalities: Humanities and Social Science 2005 S1 68 69 Luo H. Q. (2017). The Study of Midland Court Clerks Turnover Influencing Factors on the Court of J city in J province, Master’s thesis of East China University of Politial Science and Law. Ma LP Yue CJ Research on the segment of labor market and the employment flow of college graduates Research in Educational Development 2011 3 1 7 McGuinness S Overeducation in the labour market Journal of Economic Surveys 2006 20 3 387 418 10.1111/j.0950-0804.2006.00284.x Pan L. (2018). Investigation and research on the professional status of court clerks in Guiyang, Master’s dissertation. Qi HG Zhao MF Liu SH Gao P Liu Z Evolution pattern and its driving forces of China's interprovincial migration of highly-educated talents from 2000 to 2015 Geographical Research 2022 41 2 456 479 Sunita D Ronald EL Brain drain from developing countries: how can brain drain be converted into wisdom gain? Journal of the Royal Society Medicine 2005 98 11 487 491 10.1177/014107680509801107 Wagner R Childs M Exclusionary narratives as barriers to the recognition of qualifications, skills and experience—A case of skilled migrants in Australia Studies in Continuing Education 2006 28 1 49 62 10.1080/01580370500525707 Wang BY Research on the employment difficulty of college students from the perspective of macroeconomic policy Education Exploration 2010 10 131 134 Wise, S., Schutz, H., Healy, J. and Fitzpatrick, D. (2011). Engineering Skills Capacity in the Road and Rail Industries, Workplace Research Centre, University of Sydney and the National Institute of Labour Studies, Flinders University, ANET Sydney, 102. Wu XG Higher education, elite formation and social stratification in contemporary China Chinese Journal of Sociology 2016 36 3 1 31 Yoshida Y Smith MR Training and the earning of immigrant males: Evidence from the Canadian workplace and employee survey Social Science Quarterly 2005 86 1218 1241 10.1111/j.0038-4941.2005.00343.x Zhou JG Social Transformation and Social Problems 2008 China, Gansu People Publishing Company 194 200

0

PMC9753013

NO-CC CODE

2022-12-16 23:25:27

no

Soc Indic Res. 2022 Dec 15;:1-18

utf-8

Soc Indic Res

10.1007/s11205-022-03049-7

oa_other

==== Front Atmos Ocean Opt Atmospheric and Oceanic Optics 1024-8560 2070-0393 Pleiades Publishing Moscow 4402 10.1134/S1024856022060252 Atmospheric Radiation, Optical Weather, and Climate Surface Ozone in the Atmosphere of Moscow during the COVID-19 Pandemic Stepanov E. V. eugenestepanov@yandex.ru 1 Andreev V. V. vvandreev@mail.ru 2 Konovaltseva L. V. konovaltseva-lv@rudn.ru 2 Kasoev S. G. sergey-kasoev@yandex.ru 1 1 grid.424964.9 0000 0004 0637 9699 Prokhorov General Physics Institute, Russian Academy of Sciences, 119991 Moscow, Russia 2 grid.77642.30 0000 0004 0645 517X Peoples’ Friendship University of Russia, 117198 Moscow, Russia 15 12 2022 2022 35 6 732740 12 5 2022 1 6 2022 15 6 2022 © Pleiades Publishing, Ltd. 2022, ISSN 1024-8560, Atmospheric and Oceanic Optics, 2022, Vol. 35, No. 6, pp. 732–740. © Pleiades Publishing, Ltd., 2022.Russian Text © The Author(s), 2022, published in Optika Atmosfery i Okeana. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. We present the results from monitoring surface ozone in the atmosphere of Moscow in 2020 and 2021 under lockdown conditions linked to the COVID-19 pandemic. These two years significantly differed in meteorological conditions and the level of anthropogenic environmental load. A level of surface O3 concentrations, relatively low for a megalopolis, was observed in Moscow in 2020. The annual average concentration was 28 μg/m3, and the annual maximal concentration was 185 μg/m3. That was due to relatively cool summer with the low content of pollutants in atmospheric air. Intense heat waves were observed in the megalopolis during summer 2021 under the conditions of a blocking anticyclone, when the daytime temperatures rose to 35°C. Combined with higher atmospheric air pollution, this resulted in unusually high O3 concentrations. The annual average concentration was 48 μg/m3, and the annual maximal concentration was 482 μg/m3. Keywords: atmosphere air surface ozone maximum permissible concentration surface layer troposphere ozone precursors COVID-19 pandemic lockdown issue-copyright-statement© Pleiades Publishing, Ltd. 2022 ==== Body pmcINTRODUCTION Surface ozone is an important chemical constituent of the Earth’s atmosphere, playing a significant role in forming its oxidation potential [1, 2]. In turn, oxidation of both organic and inorganic substances is one of the main components in the cycle of substances in nature [2]. Surface ozone strongly affects living organisms. The nonpolluted atmosphere contains a minor (as low as 30 μg/m3) background ozone amount. At these concentrations, ozone affects living organisms as a soft mutagenic and tonic factor, allowing them to adapt to environmental changes and keep evolving [3–8]. Introduced into the human body as an aqueous solution, ozone may have immunomodulatory, anti-inflammatory, antibacterial, antiviral, and antifungal effects [9]. At high concentrations observed during strong pollution of atmospheric air in big cities and industrial regions and exceeding accepted sanitary standards (daily average maximum permissible concentration (MPCd.a) is 30 μm/m3, and the one-time maximum permissible concentration (MPCm.o) is 160 μg/m3), ozone has pathogenic properties [10–14]. Being a strong oxidant, ozone is harmful to the respiratory organs and causes a systemic inflammation of blood circulating organs. Increased ozone content in the surface atmosphere is considered to cause not only a stronger morbidity of the respiratory organs, blood circulating organs, and nervous system, but also greater total mortality [15–17]. Increased surface ozone content also adversely affects the ecosystems, forests, individual plants, and productivity of certain agricultures [18, 19]. In the clean surface atmosphere, ozone is generated via the cycle of photochemical reactions with the participation of nitrogen oxides and solar UV radiation; conversely, ozone is destroyed after being chemically bound to nitrogen monoxide (NO) or subject to dry deposition [1, 2]. In the atmosphere polluted by the products of incomplete combustion (carbon monoxide, volatile hydrocarbons), the process of ozone binding to NO is slowed down, and the ozone generation rate may have been much larger than the rate of ozone destruction. In this case, ozone is accumulated and its concentration increases in the surface air layer. The ozone generation rate also strongly increases with rising temperature [1, 2]. The ozone content in the surface layer depends on air humidity, the intensity of wind-driven air-mass-mixing processes, destruction and sink upon the interaction with the Earth’s surface, vegetation, and soil. Increased surface ozone concentrations (SOCs) present the largest problem in big cities and industrially developed southern regions, such as North America, southern European countries, and China, where there are large anthropogenic environmental loads and a hot climate. In Russia and, in particular, in Moscow, high ozone concentrations were first recorded in the surface atmosphere in two recent decades [20–24]. First, this is because motor vehicles have rapidly grown in number, producing more exhaust emissions to the atmosphere; second, this is due to ongoing climate change, which resulted in regular heat waves with daytime temperatures of up to 35–40°С in Central Russia. In 2020 and 2021, a unique situation developed in the air basin of the Moscow region, making it possible to estimate the effect of both high temperatures and gaseous pollutants on the ozone generation rate and its accumulation in the surface atmosphere. The COVID-19 pandemic began in 2020; therefore, “high alert measures” and lockdown were imposed in Russia in March. As in the neighboring European countries and China, Russia rapidly reduced economic activities as well as the intensity of road and air traffic. A simultaneous compliance with the strict lockdown measures in many countries has led to a marked reduction of pollutant emissions to the atmosphere around the globe, which was recorded by many atmospheric air quality monitoring facilities [25–27]. It is noteworthy that the events with anomalous (both too high [33, 34] and too low [35–37]) ozone levels in the surface atmosphere were also reported [28–32]. After strict lockdown measures were imposed, we, too, recorded the effect of SOC reduction at the surface ozone monitoring station, located in background plain region in Central Russia in Vyatskiye Polyany, Kirov oblast [37, 38]. At this station in late March, i.e., just after the high-alert announcement in Russia, there was a jump-like three-fold reduction of SOC values, both monthly average nighttime minima and daytime maxima. The traditional springtime TOC maximum in April was weakly pronounced at this station. These results indicate not only a local, but also a global reduction of surface ozone. The year 2021 turned out to be unusual in terms of meteorological conditions. The pandemic-linked restrictions in Moscow were relaxed, so that pollutant concentrations in the atmosphere returned to the previous level. At the same time, spring and summer in Central Russia were warmer and dryer than usual, owing to the specific features of large-scale air circulation that created the conditions for both anomalously high temperatures and for pollutant accumulation in the surface air layer [39, 40]. Thus, the years 2020 and 2021 in the Moscow region strongly differed in meteorological conditions and in the level of anthropogenic load on the atmosphere. The purpose of this work is to compare the long-term SOC behaviors, recorded at the center of Moscow at the RUDN monitoring station in 2020 and 2021, and to clarify the role the temperature and the concentration of gas pollutants play in ozone generation under megalopolis conditions. INSTRUMENTS AND METHODS The data for analysis were acquired at the station for monitoring surface ozone, ozone precursors, and the main meteorological parameters; the station started to operate at the Peoples’ Friendship University of Russia with participation of Prokhorov General Physics Institute, Russian Academy of Sciences, in late 2019. The station is located at the center of Moscow within the Third Ring Road in Ordzhonikidze Street (55°42′37″ N, 37°36′78″ E, 149 m ASL). The station is surrounded by urban residential buildings, as well as by a few parks and boulevards. The nearest highways, which are the main sources of ozone precursors, are more than 1 km away from the station. There are no industrial structures nearby. In addition to measurements of the О3 concentration, the station also monitors NO, NO2, CO, CH4, and hydrocarbons and measures the mass concentrations of aerosol particles of different sizes and the main meteorological parameters. A 3-02P chemiluminescent gas analyzer, developed and manufactured by the Instrument-Making Enterprise OPTEC (St. Petersburg) and awarded the international certificate from U.S. Environmental Protection Agency [41, 42], is used for ozone measurements. The main metrological characteristics of the analyzer are: the dynamic range is 0–500 μg/m3, the sensitivity is 1 μg/m3, the error limit is 15%, the integration time is 1 min, and the recording rate is once a minute. To reduce the measurement error, the instrument is automatically calibrated every 10 min using a calibration gas mixture or “null gas.” The manufacturer tests and calibrates the instrument once a year using the first-rank working standard of the ozone molar fraction unit in ozone–air mixtures RE 154-1-33-2008, which is stored in the OPTEC instrument-making enterprise. The gas analyzer is operated as a part of automated measurement complex, ensuring data acquisition, storage, preliminary processing, and transfer, as well as data display and remote control. The air intake is carried out in the inner yard of the RUDN buildings at an altitude of ∼5 m via standard Teflon samplers. The measurements are carried out in the continuous long-term monitoring mode. The current parameters are measured once a minute and stored in the database at the measurement complex. Table 1.   Characteristics of SOC time series in Moscow in 2020 and 2021 Parameter Year 2020 2021 Annual average, μg/m3 28 48 60-minute maximum, μg/m3 185 482 Daily maximal 60-minute average, μg/m3 55 101 P80(1h) of annual time series of daily maximal 60-minute values, μg/m3 76 157 Period of concentrations in excess of MPCm.o, h 1 402 OZONE MONITORING IN THE SURFACE ATMOSPHERE We monitored the surface ozone continuously for 2020 and 2021. Figure 1 shows the time behavior of the hourly average SOCs recorded for these two years. We can clearly see two annual cycles of variations in the ozone concentration caused by the annual cyclicities of temperature, illumination, and daytime length. The surface ozone concentrations are minimal at low temperatures and during short days. The annual behavior of SOC is very typical for the atmosphere of a megalopolis. The growth of daytime temperatures during winter is accompanied by the SOC increase. The SOC reaches maxima in summer, in June or July. SOC starts gradually decreasing in August under the conditions of diminishing daily average temperatures and shortening daytime. It can be seen that the ozone content strongly varies in time during the year. The SOCs can vary from zero to maxima within quite short (a few hours) periods of time; therefore, the annual behavior look like “noisy” random fluctuations. Fig. 1. Time behavior of SOC at RUDN station, Moscow. The maximal SOCs in Moscow were no more than 180 μg/m3 in late June 2020; while in 2021, SOCs were anomalously high. A monotonic SOC growth has been apparent as early as March. In April, the daily average concentrations reached 100 μg/m3 and more, much larger than MPCd.a; and the daily maximal hourly concentration started regularly exceeding MPCm.o. The photochemical ozone generation in the surface atmosphere intensified in mid-March, so that the daily maximal concentrations exceeded 200 μg/m3 every day. The SOCs in the Moscow atmosphere were extremely high in June after passage of three heat waves. In that time, the weather was determined by a stable blocking anticyclone conducive to conditions, on the one hand, for anomalously high temperatures and low humidity and, on the other hand, for pollutant accumulation in the surface atmosphere [39, 40]. It is noteworthy that, for a few days, the daily maximal temperatures reached 35°С, while nighttime temperatures did not decrease to below 25°С. In daytime hours, it was dry, the relative humidity decreased to below 35%, the atmospheric pressure reached 758 mmHg, and wind blew from predominantly southwestern directions at a speed of up to 2 m/s. In nighttime hours, the wind changed the direction to southeast and calmed down. These conditions were ideal for intense photochemical ozone generation and ozone accumulation in daytime hours. For more than 40 days, from mid-June to mid-August, SOCs in excess of 160 μg/m3 were observed from 06:00 to 10:00 LT. The heat waves, lasting for 10–14 days, were followed by short periods of rainy weather, when the maximal daytime SOCs decreased to ∼100 μg/m3. A characteristic feature of the daily cycle of surface ozone in megalopolises during the spring–summer period is that the nighttime ozone concentrations were close to zero independent of how large the daytime concentrations were (Fig. 2). This trait, stemming from the relatively high NO content in the Moscow atmosphere at night, had been regular throughout summer 2021. Fig. 2. Variations in hourly average SOC in the periods of maximal levels in June–July of (a) 2020 and (b) 2021 at RUDN station. For comparison, Fig. 2, on an enlarged scale, shows ozone variations in those 2020 and 2021 periods, when the largest SOCs were recorded. Several most important features of the SOC time variations can be noted. First, we can clearly see the abovementioned wavelike character of daytime SOCs, alternating between high- and moderate-concentration periods. Figure 2a (2020) shows two waves of minor increase in SOC; and Fig. 2b (2021) shows two waves, which were most intense in that year, and during which SOC reached a maximum of ∼482 μg/m3. Second, we can clearly see the circadian rhythm of the time variations in the SOC, associated with alternating day and night. SOCs are very low (high) in nighttime (daytime) hours. Thus, the variations, “noisy” in character, become regularly periodic after a more careful inspection. The presence of circadian SOC rhythm makes it possible to average and accumulate the data over the day, which was found to be more informative than a simple sequential data smoothing. This approach is widely used to analyze both daily SOC variations and long-term SOC trends [1, 17, 20, 22–24, 43]. Figure 3 presents the daily variations in SOC and temperature after seasonal averaging. They are obtained after hourly summation and a subsequent normalization of the daily variations of these parameters in each season. It can be seen that, despite the fact that the temperature was, on average, ∼5°С lower during winter 2021 than in 2020, the wintertime daily SOC variations differ little between 2020 and 2021. This can be because the photochemical ozone generation is minor at low winter temperatures, so that this difference in the average temperatures does not markedly affect the daytime ozone production. Fig. 3. Daily variations in SOC and temperature in 2020 (open circles) and 2021 (closed circles) at the RUDN station in (a) winter, (b) spring, (c) summer, and (d) fall. The spring average temperature variations coincided in 2020 and 2021. At the same time, the amplitude of the daytime SOC maximum turned out to be almost a factor of two larger, and the daytime SOC increment (the difference between the daytime maximum and nighttime minimum) turned out to be almost a factor of three larger in spring 2021 than in 2020. That difference could be because air in the megalopolis was polluted by nitrogen oxides, carbon monoxide, and volatile organic compounds weaker in spring 2020 during the COVID-19 pandemic; as a consequence, less ozone was produced in daytime in 2020 than in 2021 under nearly identical temperature regimes. As is already mentioned above, summer 2021 was hot, with the summer average daily variations in the temperature lying ∼10°С higher than that for 2020. The difference in the summer variations in the daily SOCs between 2020 and 2021 is even more contrasting. Considering that the atmosphere was differently polluted in springs 2020 and 2021, the increase in SOC during summer 2021 can be attributed to the joint effect of higher temperatures and stronger pollution of Moscow air by ozone precursors. Despite the small difference in the averaged temperatures, the daily variations in SOC differ little between 2020 and 2021, which, as in the winter period, can be because the photochemical ozone production is weakly effective. The joint effect of increased temperature and ozone precursor concentrations on the SOC levels can be clearly seen on the distribution diagram of the hourly average SOCs as functions of the atmospheric air temperature (Fig. 4). This diagram accurately characterizes the difference in the temperature conditions between these years. The temperature varied in the ranges from −10 to +32°С in 2020 and from −20 to +35°С in 2021. It can be seen that SOC quite weakly depends on the temperature when the latter is below +10°С. A sharp exponential [1, 2] dependence becomes marked above +20°С. It is noteworthy that the SOC values recorded in 2020 are below those recorded in 2021 for all temperatures. The SOC variability ranges show a twofold difference up to +20°С. This difference rapidly increases above +20°С, until becoming threefold above +30°С. Fig. 4. Distribution of the hourly average SOCs for different atmospheric air temperatures in 2020 (black circles) and 2021 (gray circles) at the RUDN station. The differences in the distributions of SOC versus temperatures between 2020 and 2021, shown in the diagram, can also indicate lower air pollution of Moscow air in 2020. It can be seen that ozone concentration in the surface atmosphere exponentially increases when air is strongly polluted by ozone precursors at temperatures above +30°С. Figure 5 shows the monthly average NO, NO2, CO, and CH4 concentrations in the surface atmosphere of Moscow in 2020–2021 calculated from monitoring data. It can be seen that April 2020 stands out with a strong reduction of the concentrations of all these gases in air because of the beginning of the lockdown period. The measures were the strictest in April and May. Two months later, the atmospheric pollution level started gradually returning to the prelockdown and even higher values, possibly because industries and motor vehicles had adapted to pandemic conditions. In early 2021, we recorded the concentrations of both carbon-containing substances (СO and CH4) and NO2 the largest in that two-year period. The СО and NO2 contents in early 2021 were ∼30% larger than the average, and the СН4 content was ∼70% larger than the average. We turn attention to the specific features of the seasonal NO variations in the urban atmosphere, clearly manifested in the diagrams. The concentrations of this substance are minimal in those periods when ozone content is maximal. Fig. 5. Monthly average concentrations of the main surface ozone precursors CO, CH4, NO2, and NO in the atmosphere in 2020 (light-gray columns) and 2021 (dark-gray columns) at the RUDN station. DISCUSSION Our results show that the high-alert measures taken in Moscow in 2020 in response to the COVID-19 pandemic had led to strong changes in the concentrations of pollutants in the urban atmosphere. That was owing to the reduced total anthropogenic environmental load, because economic activities and the intensity of road and air traffic were significantly reduced in the city. The effect of the general removal of pollutants from Moscow air in this period is difficult to estimate quantitatively; however, the data from monitoring at the RUDN station clearly indicate that the local NO2, CO, and CH4 concentrations decreased by a factor of 1.5 in the surface atmosphere at the center of Moscow at the beginning of pandemic (Fig. 5). The first pandemic year differed little from the statistical average in meteorological conditions. Owing to the reduction of atmospheric air pollution throughout 2020, such a big megalopolises Moscow exhibited quite low concentrations of surface ozone, serving an integrated indicator of the total pollution of atmospheric air in that case [1, 2]. In particular, the concentrations in excess of MPCm.o were observed just once throughout the year. The annual maximum hourly average SOC was 158 μg/m3; the annual average SOC was 28 μg/m3; and the annual average daily maximal SOC was 55 μg/m3. The P80(1h) percentile in the annual distribution of daily maximal SOCs in 2020 had been 76 μg/m3. These values can be used subsequently as a target integral index of air quality in Moscow. The meteorological conditions in Moscow strongly differed between 2020 and 2021. Few heat waves with a maximal daytime temperature of +35°С were observed in summer 2021. (The strong difference between the years consists not only in higher temperature, but also in the heat waves in combination with the blocking anticyclone.) The weather in that period was determined by the blocking anticyclone that ensured rising temperature, air mass stagnation, weak inflow of clean air, and low relative humidity. The anthropogenic environmental load was also higher than in 2020 in view of the softer COVID-19 restrictions. Owing to the totality of these factors, the SOCs, larger than usual, were observed throughout 2021 starting from spring. In particular, the total period of time when MPCm.o was exceeded by the hourly average SOCs, was 402 h; the annual maximal hourly average SOC was 482 μg/m3; the annual average SOC was 48 μg/m3, the annual average daily maximal SOC was 101 μg/m3; and P80(1h) = 157 μg/m3. The data are summarized in Table 1 for convenience of comparing the parameters between 2020 and 2021. It should be noted the ozone concentration of ∼500 μg/m3 was observed last time in Moscow in August 2010, when the maximal daytime temperature exceeded 42°С [16, 23]. That summer, the urban atmosphere was burdened by smokes from forest fires in the Moscow region. These events are separated by 11 years, which is the time interval close to the solar activity cycle. Additional studies are required to confirm this relationship. CONCLUSIONS In this paper, we presented the results from surface ozone monitoring in the atmosphere of Moscow in 2020 and 2021, under the conditions of economic, motor-vehicle, and social activities reduced due to COVID-19 restrictions. The dynamics of the ozone content in the surface atmosphere is compared between these two years, which differed in both meteorological conditions and the level of the anthropogenic environmental load. The surface ozone concentrations, relatively low for the megalopolis, were observed throughout 2020 in Moscow. The annual average was 28 μg/m3, and the annual maximum was 185 μg/m3. This was because cool rainy weather during spring and summer was coupled with low pollutant content in atmospheric air after severe pandemic restrictions were implemented. During summer 2021, several intense heat waves were observed in the megalopolis under the conditions of a blocking anticyclone, when the maximal daytime temperatures reached 35°С. This, along with higher atmospheric air pollution, as compared to the preceding year, to produce extraordinarily high ozone concentrations. The annual average concentration was 48 μg/m3, and the annual maximal concentration was 482 μg/m3. CONFLICT OF INTEREST The authors declare that they have no conflicts of interest. Translated by O. Bazhenov ==== Refs REFERENCES 1 Belan B. D. Tropospheric Ozone 2010 Tomsk Publishing House of IAO SB RAS 2 Isidorov V. V. Organic Chemistry of the Atmosphere 2001 St. Petersburg Khimizdat 3 Carver J. H. The evolutionary role of atmospheric ozone J. Atmos. Sci. 1977 34 720 728 10.1175/1520-0469(1977)034<0720:TEROAO>2.0.CO;2 4 Walker J. C. G. The early history of oxygen and ozone in the atmosphere Pure Appl. Geophys. 1978 117 498 512 10.1007/BF00876630 5 Roshchina V. V. Roshchina V. D. Ozone and Plant Cell 2003 Dordrecht Kluwer 6 V. V. Roshchina and E. V. Mel’nikova, “Contribution of ozone and reactive oxygen species into the development of cell cyctems of plants,” in Proc. of the Intern. Conf. “Mitochondria, Cells, and Reactive Oxygen Species”, Ed. by V.P. Skulachev and V.P. Zinchenko (Biological Center of RAS, Pushchino, 2000), pp. 127–129 [in Russian]. 7 Roshchina V. V. Microbial Endocrinology. Interkingdom Signaling in Infectious Disease and Health 2010 New York, Berlin Springer 8 Zanardi I. Borrelli E. Valacchi G. Travagli V. Bocci V. Ozone: A multifaceted molecule with unexpected therapeutic activity Current Med. Chem. 2016 23 304 314 10.2174/0929867323666151221150420 9 Smith N. L. Wilson A. L. Gandhi J. Vatsia S. Ali-Khan S. Ozone therapy: An overview of pharmacodynamics, current research, and clinical utility Med. Gas Res. 2017 7 212 219 10.4103/2045-9912.215752 29152215 10 Zhang J. Wei Y. Fang Z. Ozone pollution: A major health hazard worldwide Front. Immunol. 2019 10 2518 10.3389/fimmu.2019.02518 31736954 11 Ebi K. L. McGregor G. Climate change, tropospheric ozone and particulate matter, and health impacts Environ. Health Perspect. 2008 116 1449 1455 10.1289/ehp.11463 19057695 12 Nuvolone D. Petri D. Voller F. The effects of ozone on human health Environ. Sci. Pollut. Res. 2018 25 8074 88 10.1007/s11356-017-9239-3 13 Koman P. D. Mancuso P. Ozone exposure, cardiopulmonary health, and obesity: A substantive review Chem. Res. Toxicol. 2017 30 1384 1395 10.1021/acs.chemrestox.7b00077 28574698 14 S. N. Kotel’nikov, “Main ways of interaction of ozone with living systems and eculiarities of the problem of surface ozone for Russia,” in Trudy IOFAN. Vol. 71. Problem of Tropospheric Ozone, Ed. by E.V. Stepanov (Nauka, Moscow, 2015), pp. 10–41 [in Russian]. 15 Turner M. C. Jerrett M. Pope C. A. Krewski D. Gapstur S. M. Diver W. R. Beckerman B. S. Marshall J. D. Su J. Crouse D. L. Burnett R. T. Long-term ozone exposure and mortality in a large prospective study Am. J. Respir. Crit. Care. Med. 2016 193 1134 1142 10.1164/rccm.201508-1633OC 26680605 16 S. N. Kotel’nikov and E. V. Stepanov, “Effect of surface ozone on the health of population,” in Trudy IOFAN. Vol. 71. Problem of Tropospheric Ozone, Ed. by E.V. Stepanov (Nauka, Moscow, 2015), pp. 72–93 [in Russian]. 17 Kotelnikov S. N. Stepanov E. V. Ivashkin V. T. Ground ozone concentrations and health of different age population groups in Moscow in summer 2010 Geofiz. Protsessy Biosfera 2019 18 26 38 18 Dingenen R. V. Dentener F. J. Raes F. Krol M. C. Emberson L. Cofala J. The global impact of ozone on agricultural crop yields under current and future air quality legislation Atmos. Environ. 2009 43 604 618 10.1016/j.atmosenv.2008.10.033 19 Felzer B. S. Cronin T. Reilly J. M. Melillo J. M. Wang X. Impacts of ozone on trees and crops Comptes Rendus Geosci. 2007 339 784 798 10.1016/j.crte.2007.08.008 20 Elansky N. F. Smirnova O. I. Ozone and nitrogen oxide concentrations in the surface air in Moscow Izv. Ross. Akad. Nauk. Fiz. Atmos. Okeana 1997 33 597 611 21 Zvyagintsev A. M. Kuznetsova I. N. Variability of surface ozone near Moscow: 10-year monitoring results Izv. Ross. Akad. Nauk. Fiz. Atmos. Okeana 2002 38 486 495 22 Zvyagintsev A. M. Kruchenitskii G. M. Belikov I. B. Elansky N. F. Skorokhod A. I. Egorov V. I. Nikolaeva A. N. Kuznetsova I. N. Obukhova Z. V. Positive anomalies in the surface ozone concentration in July-August 2002 over Moscow and its suburbs Izv., Atmos. Ocean. Phys. 2004 40 68 79 23 Elansky N. F. Mokhov I. I. Belikov I. B. Berezina E. V. Elokhov A. S. Ivanov V. A. Pankratova N. V. Postylyakov O. V. Safronov A. N. Skorokhod A. I. Shumskii R. A. Gaseous admixtures in the atmosphere over Moscow during the 2010 summer Izv., Atmos. Ocean. Phys. 2011 47 672 681 10.1134/S000143381106003X 24 Elansky N. F. Belikov I. B. Skorokhod A. I. Lokoshchenko M. A. Trifanova A. V. On contents of trace gases in the atmospheric surface layer over Moscow Izv., Atmos. Ocean. Phys. 2015 51 30 41 10.1134/S000143381501003X 25 Adam M. G. Tran P. T. M. Balasubramanian R. Air quality changes in cities during the COVID-19 lockdown: A critical review Atmos. Res. 2021 264 105823 10.1016/j.atmosres.2021.105823 34456403 26 A. Agarwal, A. Kaushik, S. Kumar, and R. Kumar, “Comparative study on air quality status in Indian and Chinese cities before and during the COVID-19 lockdown period,” Air Quality, Atmos. Health (2020). 27 Yang M. Chen L. Msigwa G. Tang K. H. D. Pow-Seng Yap P.‑S. Implications of COVID-19 on global environmental pollution and carbon emissions with strategies for sustainability in the COVID-19 era Sci. Total Environ. 2022 809 151657 10.1016/j.scitotenv.2021.151657 34793787 28 Mertens M. Jockel P. Matthes S. Nutzel M. Grewe V. Volker, and R. Sausen, “COVID-19 induced lower-tropospheric ozone changes Environ. Res. Lett. 2021 16 064005 10.1088/1748-9326/abf191 29 I. Bouarar, B. Gaubert, G. P. Brasseur, W. Steinbrecht, T. Doumbia, S. Tilmes, Y. Liu, T. Stavrakou, A. Deroubaix, S. Darras, C. Granier, F. Lacey, J.-F. Muller, X. Shi, N. Elguindi, and T. Wang, “Ozone anomalies in the free troposphere during the COVID-19 pandemic,” Geophys. Rev. Lett. 48 (16) (2021). 30 Peralta O. Ortinez-Alvarez A. Torres-Jardon R. Suarez-Lastra M. Castro T. Ruiz-Suarez L. G. Ozone over Mexico City during the COVID-19 pandemic Sci. Total Environ. 2021 20 143183 10.1016/j.scitotenv.2020.143183 31 Hashim B. M. Al-Naseri S. K. Al-Maliki A. Al-Ansari N. Impact of COVID-19 lockdown on NO2, O3, PM2.5, and PM10 concentrations and assessing air quality changes in Baghdad, Iraq Sci. Total Environ. 2021 754 141978 10.1016/j.scitotenv.2020.141978 32919315 32 He C. Hong S. Zhang L. Mu H. Xin A. Zhou Y. Liu J. Liu N. Su Y. Tian Ya. Ke B. Wang Y. Yang L. Global, continental, and national variation in PM2.5, O3, and NO2 concentrations during the early 2020 COVID-19 lockdown Atmos. Pollut. Res. 2021 12 136 145 10.1016/j.apr.2021.02.002 33584105 33 Siciliano B. Dantas G. Silva C. M. Arbilla G. Increased ozone levels during the COVID-19 lockdown: Analysis for the city of Rio De Janeiro, Brazil Sci. Total Environ. 2020 737 139765 10.1016/j.scitotenv.2020.139765 32480061 34 Wang L. Li M. Yu S. Chen X. Li Z. Jiang L. Xia Y. Li J. Liu W. Li P. Lichtfouse E. Rosenfeld D. Seinfeld J. H. Unexpected rise of ozone in urban and rural areas, and sulfur dioxide in rural areas during the coronavirus city lockdown in Hangzhou, China: Implications for air quality Environ. Chem. Lett. 2020 18 1713 1723 10.1007/s10311-020-01028-3 32837481 35 W. Steinbrecht, D. Kubistin, C. Plass-Dulmer, J. Davies, D. W. Tarasick, P. Gathen, H. Deckelmann, N. Jepsen, R. Kivi, N. Lyall, M. Palm, J. Notholt, B. Kois, P. Oelsner, M. Allaart, A. Piters, M. Gill, R. Van Malderen, A. W. Delcloo, R. Sussmann, E. Mahieu, C. Servais, G. Romanens, R. Stubi, G. Ancellet, S. Godin-Beekmann, S. Yamanouchi, K. Strong, B. Johnson, P. Cullis, I. Petropavlovskikh, J. W. Hannigan, J.-L. Hernandez, A. D. Rodriguez, T. Nakano, F. Chouza, T. Leblanc, C. Torres, O. Garcia, A. N. Rohling, M. Schneider, T. Blumenstock, M. Tully, C. Paton-Walsh, N. Jones, R. Querel, S. Strahan, R. M. Stauffer, A. M. Thompson, A. Inness, R. Engelen, K.-L. Chang, and O. R. Cooper, “COVID-19 crisis reduces free tropospheric ozone across the Northern Hemisphere,” Geophys. Rev. Lett. 48 (5) (2021). 36 Fu S. Guo M. Fan L. Deng Q. Han D. Wei Y. Luo J. Qin G. Cheng J. Ozone pollution mitigation in Guangxi (South China) driven by meteorology and anthropogenic emissions during the COVID-19 lockdown Environ. Pollut. 2021 272 115927 10.1016/j.envpol.2020.115927 33143981 37 Kotelnikov S. N. Stepanov E. V. Anomalous dynamics of tropospheric ozone in the spring of 2020 in Central Russia Bull. Lebedev Phys. Inst. 2021 48 92 96 10.3103/S1068335621030076 38 Andreev V. V. Arshinov M. Yu. Belan B. D. Davydov D. K. Elansky N. F. Zhamsueva G. S. Zayakhanov A. S. Ivlev G. A. Kozlov A. V. Kotel’nikov S. N. Kuznetsova I. N. Lapchenko V. A. Lezina E. A. Postylyakov O. V. Savkin D. E. Senik I. A. Stepanov E. V. Tolmachev G. N. Fofonov A. V. Chelibanovi I. V. Chelibanov V. P. Shirotov V. V. Surface ozone concentration over Russian territory in the first half of 2020 Atmos. Ocean. Opt. 2020 33 671 681 10.1134/S1024856020060184 39 https://meteoinfo.ru/. Cited April 12, 2022. 40 Andreev V. V. Arshinov M. Yu. Belan B. D. Belan S. B. Davydov D. K. Demin V. I. Dudorova N. V. Elansky N. F. Zhamsueva G. S. Zayakhanov A. S. Ivlev G. A. Kozlov A. V. Konovaltseva L. V. Kotel’nikov S. N. Kuznetsova I. N. Lapchenko V. A. Lezina E. A. Obolkin V. A. Postylyakov O. V. Potemkin V. L. Savkin D. E. Senik I. A. Stepanov E. V. Tolmachev G. N. Fofonov A. V. Khodzher T. V. Chelibanov I. V. Chelibanov V. P. Shirotov V. V. Shukurov K. A. Tropospheric ozone concentration in Russia in 2021 Opt. Atmos. Okeana 2022 35 559 571 10.15372/AOO20220706 41 Chelibanov V. P. Kotel’nikov S. N. Smirnov N. V. Yasenko E. A. Prospects of the use of PAE-8816 software-hardware complex in the design of global atmospheric air monitoring system Biosfera 2015 7 119 123 42 S. N. Kotel’nikov and E. V. Stepanov, “Tropospheric ozone monitoring in the air of megalopolises and weakly-urbanized, ” in Trudy IOFAN. Vol. 71. Problem of Tropospheric Ozone, Ed. by E.V. Stepanov (Nauka, Moscow, 2015) [in Russian]. 43 Kotelnikov S. N. Stepanov E. V. Positive trend of surface ozone in the north of the Privolzhskii Federal Region of the Russian Federation Bull. Lebedev Phys. Inst. 2019 45 24 28 10.3103/S1068335618010062

0

PMC9753014

NO-CC CODE

2022-12-16 23:25:27

no

Atmos Ocean Opt. 2022 Dec 15; 35(6):732-740

utf-8

oa_other

==== Front AIDS Behav AIDS Behav AIDS and Behavior 1090-7165 1573-3254 Springer US New York 3945 10.1007/s10461-022-03945-6 Substantive Review A Systematic Review of Self-Management Interventions Conducted Across Global Settings for Depressive Symptoms in Persons with HIV Yoo-Jeong Moka 1 Alvarez Gabriella 2 Khawly Gabriella 3 Voss Joachim 4 Wang Tongyao 45 Barroso Julie 6 http://orcid.org/0000-0003-2184-4045 Schnall Rebecca rb897@columbia.edu 23 1 grid.261112.7 0000 0001 2173 3359 Bouvé College of Health Sciences School of Nursing, Northeastern University, Boston, MA USA 2 grid.21729.3f 0000000419368729 School of Nursing, Columbia University, 560 West 168th Street, 10032 New York, NY USA 3 grid.21729.3f 0000000419368729 Mailman School of Public Health, Columbia University, New York, NY USA 4 grid.67105.35 0000 0001 2164 3847 Frances Payne Bolton School of Nursing, Case Western Reserve University, Cleveland, OH USA 5 grid.67105.35 0000 0001 2164 3847 Frances Payne Bolton School of Nursing, Sarah Cole Hirsh Institute for Evidence Based Practice at the Case Western Reserve University, Cleveland, OH USA 6 grid.152326.1 0000 0001 2264 7217 Vanderbilt University School of Nursing, Nashville, TN USA 15 12 2022 116 22 11 2022 © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Depressive symptoms can affect health outcomes in people living with HIV (PLWH) including adherence to treatment and disease prognosis. Self-management interventions targeting depressive symptoms have been effective in preventing these negative sequelae of depressive symptoms. The processes of self-management include learning skills related to living with the illness needs, accessing resources to manage the illness, and coping with the illness. A systematic literature review was conducted to appraise and synthesize the current evidence of self-management interventions targeting depressive symptoms in PLWH. Following the PRISMA guidelines, an electronic search of 4 databases was conducted. Original studies written in English that used a randomized controlled trial design to test the effect of self-management intervention on depressive symptoms were included. Studies were selected that were published on/before April 19, 2022, thus yielding 13 relevant articles. Risk of bias was assessed using the NIH Quality Assessment Tool for Controlled Intervention Studies and narrative synthesis was used to synthesize the results. 40 to 755 participants were included in the studies, with each using various measures to assess depressive symptoms pre-and post-intervention, and timepoints for assessing depressive symptoms post-intervention varied. While 12 studies showed a significant reduction in depressive symptoms post-intervention, only 4 studies that used individual coaching or technology showed lower depressive symptoms in intervention groups in comparison to the control groups. This review can be used to inform scale-up and dissemination of these interventions to improve depressive symptoms in PLWH. Resumen Los síntomas depresivos pueden afectar el estado de salud en personas que viven con el VIH (PLWH, por sus siglas en inglés), incluyendo la adherencia al tratamiento y el pronóstico de la enfermedad. Las intervenciones de autocuidado enfocadas a los síntomas depresivos han sido eficaces para prevenir estas secuelas negativas de los síntomas depresivos. Los procesos de autogestión incluyen habilidades de aprendizaje relacionadas con enfocarse a vivir con la enfermedad necesidades, acceder a recursos para manejar la enfermedad y afrontar a la enfermedad. Se realizó una revisión sistemática de la literatura para evaluar y sintetizar la evidencia actual de las intervenciones de autocuidado dirigidas a los síntomas depresivos en personas que viven con el VIH. Siguiendo las directrices PRISMA, se realizó una búsqueda electrónica en 4 bases de datos. Se incluyeron estudios originales escritos en inglés que utilizaron un diseño de ensayo controlado aleatorio para evaluar el efecto de la intervención de autocuidado sobre los síntomas depresivos. Se seleccionaron estudios que se publicaron el 19 de abril de 2022 o antes, obteniendo 13 artículos relevantes. El riesgo de sesgo se evaluó mediante la herramienta de evaluación de la calidad de los NIH para estudios de intervención controlados y se utilizó la síntesis narrativa para sintetizar los resultados. Se incluyeron de 40 a 755 participantes en los estudios. Los estudios utilizaron diversas medidas para evaluar los síntomas depresivos antes y después de la intervención, y los puntos temporales para evaluar los síntomas depresivos después de la intervención variaron. Mientras que 12 estudios mostraron una reducción significativa en los síntomas depresivos después de la intervención, solo 4 estudios que usaron entrenamiento individual o tecnología mostraron síntomas depresivos más bajos en los grupos de intervención en comparación con los grupos de control. Esta revisión se puede utilizar para informar la ampliación y difusión de estas intervenciones para mejorar los síntomas depresivos en las personas que viven con el VIH. Keywords Depressive symptoms HIV Interventions Self-management Systematic review Palabras Clave: Síntomas depresivos VIH Intervenciones Autogestión Revisión sistemática http://dx.doi.org/10.13039/100000056 National Institute of Nursing Research R01NR015737 K24NR018621 Schnall Rebecca ==== Body pmcIntroduction Depression is a common and serious medical illness that negatively affects behavior, thoughts, mood, and overall health outcomes. Persons suffering from depression can experience a constellation of symptoms including but not limited to: feeling sad, worthless, or guilty, loss of interest or pleasure in activities once enjoyed, difficulty thinking, concentrating or making decisions, trouble sleeping or sleeping too much, loss of energy or increased fatigue, changes in appetite, and thoughts of death or suicide [1]. Clinically, a person must have depressive symptoms most of the day, nearly every day, for at least 2 weeks to be diagnosed with depression [2]. While there is evidence that many people living with HIV (PLWH) may have a co-morbid diagnosis of depression with rates of depression two to four times higher in PLWH than the general population [3], a more recent study that focused on depressive symptoms reported by PLWH found that as many as 40% of PLWH experience depressive symptoms during their lifetime [4]. Depressive symptoms affect adherence to medication regimens and disease prognosis [5]. Prior research has revealed that PLWH who are experiencing depressive symptoms are at increased risk of non-adherence to antiretroviral therapy (ART) and causation is not known [6]). Depressive symptoms are linked to AIDS diagnosis and an elevated risk of mortality [7]. Thus, identifying effective interventions to manage depressive symptoms is an urgent priority. Self-management interventions can help people with chronic diseases achieve higher health-related quality of life, lower levels of depression, and better health behaviors and health outcomes [8–11]. The term self-management refers to “the individual’s ability to manage the symptoms, treatment, physical and psychological consequences, and lifestyle changes inherent in living with a chronic condition. Efficacious self-management encompasses the ability to monitor one’s condition and to affect the cognitive, behavioral, and emotional responses necessary to maintain a satisfactory quality of life” [12]. The processes of self-management include (a) learning skills [13, 14] focusing on the illness needs (e.g., decision-making, goal-setting, self-monitoring, problem-solving, emotional control, self-evaluation), (b) accessing resources (e.g., coordinating health care and social services, identifying psychological resources, communicating with health care providers overcoming social and environmental challenges being part of a spiritual community, seeking social support, and (c) coping with the illness (adjustment, processing emotions, and integrating illness into daily life). Additionally, self-management interventions focused on depressive symptoms in the general population of adults with a chronic condition experiencing depressive symptomology have been shown in a systematic review and meta-analysis to have a moderate effect on depressive symptoms [15]. This is especially timely given the increased burden on the healthcare system and dearth of healthcare professionals amidst the decrease in workforce post-COVID-19 [16]. Despite the evidence to support the efficacy of self-management interventions for improving depressive symptoms in general, no review has been conducted to assess and synthesize the evidence amongst PLWH. Our goal was to review the existing self-management interventions targeting depressive symptoms in PLWH and to assess the effectiveness of self-management interventions on reducing depressive symptoms. Methods This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (PRISMA-P) 2020 [16]. Inclusion/Exclusion Criteria We included peer-reviewed articles that focused on persons living with HIV ages 18 or older. The interventions were focused on the process of self-management of depression or depressive symptoms or HIV more generally. A study was deemed eligible for this review if the intervention focused on at least one process of self-management (needs, resources, and coping). Interventions that solely utilized antidepressants or psychotherapies, such as cognitive behavioral therapy (CBT), were excluded. The outcomes of interest were depressive symptoms or clinical depression. All included studies used a randomized control trial (RCT) design. Studies without results (i.e., protocol papers) were excluded. Search Strategy One author conducted a literature search of the following electronic databases for articles published from inception to April 19, 2022: PubMed, CINAHL, PsycInfo, and EMBASE. A range of Medical Subject Headings (MeSH), keywords, and indexed terms related to self-management, depressive symptoms, depression, and HIV were searched. Boolean operators and wildcards were utilized to define relationships between the terms. The reference lists of studies meeting the inclusion criteria were hand-searched to identify additional relevant studies. Literature search strategies and search results were tracked in a Microsoft Excel document. The search results from each database were exported into EndNote, whereby duplicates were removed. Articles were imported along with their full-texts into Covidence©, an online systematic review management platform that allows multiple users to review the publications and make decisions on whether to include or exclude the publication. Study Selection, Data Extraction, and Synthesis Two reviewers (MYJ, GA) independently reviewed the titles and abstracts, followed by full-text review over the course of 3 months. When either of the two reviewers considered a study potentially eligible or uncertain, its full-text was reviewed for further assessment. A third reviewer (RS) resolved any disagreements or discrepancies. One reviewer (GA) extracted the following data from included studies and synthesized the results using narrative synthesis [17]: study characteristics (sample size & setting/geographic location); intervention description (duration, frequency, dose, type, moderator); self-management process (needs, resources, coping); measurements for depressive symptoms; and any intervention effect on depressive symptoms. Quality Appraisal The National Institutes of Health National Heart, Lung, and Blood Institute’s (NIH/NHLBI) Quality Assessment Tool of Controlled Intervention Studies was used to assess the methodological quality of the included studies [18]. This tool was selected because of its comprehensiveness in assessing the internal validity of the study and its rigor of the methods of published studies based on certain study designs. The assessment tool for RCT design has 14 items where each item can be answered with a “Yes,” “No,” or “Other.” “Other” can have three responses (cannot determine, not applicable, or not reported). Two reviewers (MYJ, GA) independently rated the studies and any discrepancies between raters were resolved via inter-rater discussion or by involving a third reviewer (RS). Studies were not excluded based on the quality appraisal; rather, the quality appraisal was used to evaluate and discuss the rigor of available evidence. Results Figure 1 provides a flow chart of the literature search process. The database search identified 6,061 publications published up to April 19, 2022, resulting in a total of 3,272 unique titles after removing duplicates. We screened the full text of 27 publications against the inclusion and exclusion criteria. A total of 13 studies were included in this review and details of each study are described in Table 1. Fig. 1 PRISMA Diagram. (Note: PLWH = People living with HIV) Table 1 Study Characteristics Patient Population Sample Size & Setting Intervention Comparator Randomization Outcome: Measures of Depressive Symptoms and Administration Timing Setting Effect of Intervention on Depressive Symptoms Jones et al. (2010) N = 451, USA Intervention = 212 Control = 239 Based on the SMART/EST (Stress Management And Relaxation Training/Expressive-Supportive Therapy) Women’s Project Weekly 120-minute cognitive behavioral stress management /expressive supportive (CBSM+) group therapy sessions Individual information conditions (“low intensity experimental control”): 10-weekly individual 120-minute information sessions around stress management/relaxation training and coping with HIV/AIDS, equal in session length to intervention group Active control Blinding not mentioned in manuscript Randomized using a table of random numbers Beck Depression Inventory (BDI) 10 weeks Participants recruited from hospitals, community health centers/agencies, and via participant referrals in Miami-Dade County, FL, NYC, and New Jersey metropolitan area Specific care setting where intervention occurred not described in manuscript Interventions can be useful adjuncts for therapy and clinic waiting areas or adapted in clinical practice and mental health settings as short-term interventions Those in the CBSM + group who reported increases in self efficacy saw significant improvements in anxiety and depression at post-intervention and 12 months follow-up. Those in the control group did not report any significant changes in depression status. Laperriere et al. (2005) N = 154, USA Intervention = 80 Control = 74 Based on The SMART/EST Women’s Project (subgroup analysis) Weekly 120-minute CBSM + group therapy sessions led by doctoral and post-doctoral level psychologists Individual information conditions (“low intensity experimental control”): 10-weekly individual 120-minute information sessions around stress management/relaxation training and coping with HIV/AIDS, equal in session length to intervention group Active control Blinding not mentioned in manuscript Randomization method not mentioned in manuscript, but based on same project as above, therefore can infer randomization using a table of random numbers BDI 10 weeks Participants recruited from hospitals, community health centers/agencies, and via participant referrals in Miami-Dade County, FL, NYC, and New Jersey metropolitan area Specific care setting where intervention occurred not described in manuscript Interventions can be useful adjuncts for therapy and clinic waiting areas or adapted in clinical practice and mental health settings as short-term interventions Participants in both the intervention and control groups saw significant improvements in depressive symptoms at post-intervention. At 12 months follow up, participants in the intervention group scored an average BDI score of 7.4 which is below the scale’s range for depression categorization, while those in the control group scored an average of 10.2 (mild depression category). These results were significantly different from one another (p < 0.001). Hecht et al. (2018) N = 177, USA Intervention = 89 Control = 88 Weekly, in-person Mindfulness-Based Stress Reduction (MBSR) group sessions HIV disease self-management skills group (controlled for the effects of being in a group program): 8 weekly group sessions that covered variety of educational topics abut managing HIV infection) Active control Participants and staff were aware of group assignments. Personnel performing laboratory assays masked to group assignment Randomized using computer-generated randomization list with random assignment of participants in a 1:1 ratio using block sizes. A database manager generated randomization sequence with guidance from a study statistician. No other study staff had access to randomization sequence file BDI 8 weeks Specific care setting where intervention occurred not described in manuscript Recruitment conducted via broad advertising across United States Those in the intervention saw a statistically significant within group improvement in BDI scores for depressive symptoms from baseline to 3-months post-intervention (p < 0.05). BDI scores did not improve significantly within the control group from baseline to any point post-intervention. Guo et al. (2020) N = 300, China Intervention = 150 Control = 150 WeChat (internet)-based culturally adapted CBSM course Usual care waitlist control group: received a brochure on nutrition in addition to usual care for HIV treatment (i.e. ART as soon as feasible) Active control By nature of the trial design, the research staff nor the participants were blinded to allocation Randomized via computer-generated randomization list with block size of 4 using SAS Center for Epidemiologic Studies Depression Scale (CES-D) 3 months Specific care setting where intervention occurred not described in manuscript Recruited from the outpatient clinic of the only hospital designated for HIV treatment in Guangzhou, China. Those in the intervention group saw statistically significant improvements in depressive symptoms at all follow-up points compared to baseline (p < 0.001 at 3,6,& 9 months). Heckman & Carlson (2007) N = 299, USA Intervention (coping) = 108 Intervention (support) = 84 Control = 107 Teleconference coping or information support group led by a professional Usual care condition: no active intervention, but had access to all services provided by their ASO (i.e. case management, support groups, and assistance obtaining housing, employment, and child care) Active control Blinding not discussed in manuscript Randomization methods not mentioned in manuscript BDI and 20-item Hopkins Symptom Checklist (SCL-20) 8 weeks Intervention delivered via reserved conference calls using the participants’ standard telephones. Recruited through 27 different AIDS service organizations (ASOs) in Ohio, Pennsylvania, Virginia, New York, Wisconsin, West Virginia, Montana, Arizona, Indiana, and Alaska. Participants who received either form of the intervention in all conditions/subgroups did not report statistically significant improvements in depressive symptoms compared to the usual care group. Miles et al. (2003) N = 109, USA Intervention = 59 Control = 50 Biweekly emotional processing strategies and cognitive reframing sessions led by a registered nurse Usual care: health visits for primary care and specialty visits for HIV-related care. Most women got their care at the tertiary HIV clinics, where focus of care was on health problems rather than self-care symptom management. Active control Data collectors were blinded to the woman’s status. Randomized using a table of random numbers (performed by project manager who was part of intervention team) CES-D and Profile of Mood States (POMS) 3 months Intervention was conducted in the homes of the women by 3 nurses (2 African American nurses and one White nurse). Visits were arranged when home alone or with others who were aware of dx Recruited from two tertiary care university-based infectious disease clinics in the Southeast Those in the intervention group saw statistically significant reductions in feelings of Depression-Dejection (measured by POMS), from baseline to 1-month post-intervention (p < 0.003). Those in the control group saw no significant changes in feelings of Depression-Dejection at any follow-up point. The depressive symptoms measured with the CES-D scale did not change significantly in the intervention group at any time point. Inouye et al. (2001) N = 40, USA Intervention = 20 Control = 20 Individually administered self-management and coping skills training program Usual care: standard treatment of their primary care providers, an educational videotape about nutrition and health in relation to HIV/AIDS, wellness education, resource materials to take home. Invited to a shorter version of the protocol in a group setting after final session Active control Blinding not discussed in manuscript Randomization methods not mentioned in manuscript POMS 7 weeks Specific care setting where intervention occurred not described in manuscript Recruited from advertisements in newspapers, private physicians, hospital flyers, AIDS organization, and word-of-mouth communication. The intervention significantly improved participants’ mood as measured by the POMS, with a 27% improvement in Depression-Dejection scores for the intervention group from baseline (p < 0.05). Those in the control group did not show significant improvements in Depression-Dejections scores from baseline. van Luenen et al. (2018) N = 188, Netherlands Intervention = 97 Control = 91 Internet-based program adapted from a previously evaluated self-help booklet + individualized coaching session led by a master’s level graduate student in the field of psychology Control condition: put on a waiting list and received attention only via telephone calls from a coach. Invited to start the intervention after the second post-test; Wait List control; Participants, researchers, and coachers were not masked to the participant’s assigned treatment condition. Randomized using random number tables to generation randomization sequence with block sizes of 12, stratified by treatment center and sex, and concealed from the main researcher. Patient Health Questionnaire (PHQ-9) and CES-D 8 weeks All assessments were completed online via a secure website, with the exception of screening with the PHQ-9, which was completed via telephone conversation Recruited from 23 HIV treatment centers in the Netherlands A large proportion (64% - PHQ-9 and 67% - CES-D) of those in the intervention group had significant decreases in their depressive symptoms. These improvements were significantly different than those in the control group. Wantland et al. (2008) N = 755, Puerto Rico, USA, South Africa & Kenya Intervention = 426 Control = 349 Individually administered HIV/AIDS Symptom Management Manual Basic nutrition manual: of similar size as compared to intervention and at an eighth-grade reading level. Modified from the WHO HIV/AIDS Nutrition Guide; Active control Blinding not discussed in manuscript Randomization methods not discussed in manuscript Revised Sign and Symptom Checklist for Persons with HIV Disease (SSC-HIVrev) 2 months Specific care setting where intervention occurred not described in manuscript Recruited from clinics and HIV-focused community settings in Africa, Puerto Rico, and from 10 sites across the US. Those in the intervention group with comorbid depression showed a statistically significantly decrease mean symptom intensity post-intervention compared to those in the control group (p < 0.018). Schnall et al. (2018) N = 80, USA Intervention = 40 Control = 40 mHealth-app on HIV/AIDS Symptom Management Manual Control app without self-care strategies Placebo control Single-blinded study Randomized using an allocation sequence developed by the PI prior to the start of trial via a computerized random number generator. SSC-HIVrev 12 weeks Took place at study site, Columbia University School of Nursing. Research setting—care setting not applicable Recruited through flyers at a local HIV clinic and community-based organizations, and through e-mail invitations The intervention group participants showed a significantly greater improvement than the control group in depression (p = 0.001). Chen et al. (2018) N = 41, China Intervention = 21 Control = 20 Self-and family management intervention led by a nurse Usual care at the clinic as well as medical care as usual. Study sessions involved only assessment survey, no counseling; Placebo control Randomized using a computer generated number in recruitment sequence CES-D 1 month Intervention was conducted within clinic, and session times coordinated with patients’ clinic visits to reduce participant burden Recruited from Beijing Ditan Hospital and Shanghai Public Health Clinic Center (SPHCC) Those in the intervention group had a significant decrease in their depressive symptoms at both 1 and 3-months post-intervention, compared to the control group. Nyamathi et al. (2012) N = 68, India Intervention = 34 Control = 34 Six sessions of HIV-specific and life management skills led by a lay village Asha (health activist) Usual care: 6 group classes matched in terms of number and length of time to intervention. Included content from Ashas with the goal of monitoring barrier to ART adherence, inquire about side effects, and provide basic education. The women received basic nutrition supplementation to address malnourishment. Ashas did not assist with linkage to care. Did not fulfill the same supportive role as intervention Ashas. Placebo control Randomization methods not mentioned in manuscript CES-D 6 weeks Interventions involved visits to women performed by Ashas, assumed in private home setting Recruited from 2 high prevalence major HIV/AIDS villages or mandals in rural Andhra Pradesh were selected randomly from 16 major villages that were demographically alike in terms of HIV prevalence of 2% and served by a distinct Public Health Center. The intervention group had a much greater reduction in depressive symptom scores than those in the usual care group (B = 22.89, p = 0.001). Those with higher depressive symptom scores at baseline also showed greater reduction in symptoms than those with lower baseline depressive symptom scores. Eller et al. (2013) N = 222, Puerto Rico, USA, South Africa Intervention = 124 Control = 98 HIV/AIDS Symptom Management Manual led by a nurse Basic nutrition manual: of similar size as compared to intervention and at an eighth-grade reading level. Modified from the WHO HIV/AIDS Nutrition Guide. Reviewed for 30-minute sessions with a research nurse to mask treatment assignment Active control Additional blinding methods not mentioned in manuscript. Randomization methods not discussed in manuscript CES-D; Single item asking frequency, intensity, and impact of depressive symptoms One 30-minute session Specific care setting where intervention occurred not described in manuscript Recruited from clinics and HIV-focused community settings in Africa, Puerto Rico, and from 10 sites across the US. No significant differences between the intervention and control groups. Within the intervention group, there was significant reduction in depressive symptom frequency [F(2, 207) = 3.27, p = 0.05], intensity [F(2, 91) = 4.6, p = 0.01], and impact [F(2, 252) = 2.92, p = 0.05), at one month but not at two months. Study Characteristics The total number of participants ranged from 40 to 755 PLWH across all studies. One study focused specifically on low-income, racial/ethnic minority individuals living with HIV [19]. Four studies included only women living with HIV [19–22], one study included only women with HIV and their caregivers [23], and the remainder included both men and women living with HIV. The studies were conducted both in the United States (US) and internationally in China [23, 24], Netherlands [25], and India [22]. Two studies were from the same multi-site study comprised of sites in the US, Puerto Rico, Kenya, and South Africa [26, 27]. Two studies [20, 21] were from the same multi-site study based in the US using the Stress Management And Relaxation Training/Expressive-Supportive Therapy (The SMART/EST)- Women’s Project [28]. Participants were recruited from multiple settings including but not limited to community sites, AIDS service organizations, and HIV treatment centers/clinics. Quality of Included Studies Detailed results of the quality assessment can be found in Fig. 2; Table 2. No studies met all assessment items. Many of the studies lacked information on whether the assignment to the control/intervention group was concealed (n = 11; item 3), whether participants were blinded to the intervention group (n = 12; item 4), and whether individuals assessing the outcomes were blinded to the group assignments (n = 12; item 5). Nine studies had more than 20% drop out rates in the intervention group (item 7) and seven studies failed to report the adequacy of the sample size to detect a difference in the main outcome between groups with power (item 12). Fig. 2 Quality assessment of self-management intervention on depressive symptoms. (Quality of the selected observational study was assessed using the National Institutes of Health National Heart, Lung, and Blood Institute’s (NIH/NHLBI) Quality Assessment tool for Controlled Intervention Studies. Item (1) Was the study described as randomized, a randomized trial, a randomized clinical trial, or an RCT? Item (2) Was the method of randomization adequate (i.e., use of randomly generated assignment)? Item (3) Was the treatment allocation concealed (so that assignments could not be predicted)? Item (4) Were study participants and providers blinded to treatment group assignment? Item (5) Were the people assessing the outcomes blinded to the participants’ group assignments? Item (6) Were the groups similar at baseline on important characteristics that could affect outcomes (e.g., demographics, risk factors, co-morbid conditions)? Item (7) Was the overall drop-out rate from the study at endpoint 20% or lower of the number allocated to treatment? Item (8) Was the differential drop-out rate (between treatment groups) at endpoint 15% points or lower? Item (9) Was there high adherence to the intervention protocols for each treatment group? Item10. Were other interventions avoided or similar in the groups (e.g., similar background treatments) Item 11. Were outcomes assessed using valid and reliable measures, implemented consistently across all study participants? Item 12. Did the authors report that the sample size was sufficiently large to be able to detect a difference in the main outcome between groups with at least 80% power? Item 13. Were outcomes reported or subgroups analyzed prespecified (i.e., identified before analyses were conducted)? Item 14. Were all randomized participants analyzed in the group to which they were originally assigned, i.e., did they use an intention-to-treat analysis?; CD, cannot determine; NA, not applicable; NR, not reported.) Table 2 Results of the quality assessment of self-management intervention on depressive symptoms Jones et al. (2010) Laperriere et al. (2005) Hecht et al. (2018) Guo et al. (2020) Heckman & Carlson (2007) Miles et al. (2003) Inouye et al. (2001) van Luenen et al. (2018) Wantland et al. (2008) Schnall et al. (2018) Chen et al. (2018) Nyamathi et al. (2012) Eller et al. (2013) Item 1 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Item 2 Yes CD Yes Yes NR Yes NR Yes NR Yes Yes No Yes Item 3 CD No No No NR CD No Yes NR Yes NR No CD Item 4 No No No No NR Yes NR No NR No NR No No Item 5 No No Yes No NR CD NR No NR No NR CD CD Item 6 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Item 7 No No No No No No Yes No No Yes Yes Yes No Item 8 No No No Yes Yes No Yes Yes Yes Yes Yes Yes No Item 9 CD CD No Yes CD CD CD Yes Yes Yes Yes NR CD Item 10 Yes Yes No Yes No Yes Yes Yes Yes Yes Yes No No Item 11 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Item 12 No Yes Yes Yes No Yes No Yes Yes No No No No Item 13 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Item 14 Yes Yes Yes Yes Yes Yes CD Yes CD Yes Yes NR Yes CD = cannot determine; NR=not reported; Quality of the selected observational study was assessed using the National Institutes of Health National Heart, Lung, and Blood Institute’s (NIH/NHLBI) Quality Assessment tool for Controlled Intervention Studies. Item (1) Was the study described as randomized, a randomized trial, a randomized clinical trial, or an RCT? Item (2) Was the method of randomization adequate (i.e., use of randomly generated assignment)? Item (3) Was the treatment allocation concealed (so that assignments could not be predicted)? Item (4) Were study participants and providers blinded to treatment group assignment? Item (5) Were the people assessing the outcomes blinded to the participants’ group assignments? Item (6) Were the groups similar at baseline on important characteristics that could affect outcomes (e.g., demographics, risk factors, co-morbid conditions)? Item (7) Was the overall drop-out rate from the study at endpoint 20% or lower of the number allocated to treatment? Item (8) Was the differential drop-out rate (between treatment groups) at endpoint 15% points or lower? Item (9) Was there high adherence to the intervention protocols for each treatment group? Item10. Were other interventions avoided or similar in the groups (e.g., similar background treatments) Item 11. Were outcomes assessed using valid and reliable measures, implemented consistently across all study participants? Item 12. Did the authors report that the sample size was sufficiently large to be able to detect a difference in the main outcome between groups with at least 80% power? Item 13. Were outcomes reported or subgroups analyzed prespecified (i.e., identified before analyses were conducted)? Item 14. Were all randomized participants analyzed in the group to which they were originally assigned, i.e., did they use an intention-to-treat analysis? Measures of Depressive Symptoms Various measures were used to assess depressive symptoms across the studies, with Beck Depression Inventory (BDI) [20, 21, 29, 30] and Center for Epidemiologic Studies Depression Scale (CES-D) [19, 22–26] being the measures most widely used by the study investigators. Other measures that were used included the 90-item Hopkins Symptom Checklist (SCL-90-R) [30]; Profile of Mood States (POMS) Depression-Dejection subscale [19, 31]; and 64-item Revised Sign and Symptom Checklist for Persons with HIV Disease (SSC-HIVrev) [27, 32]. Four studies used two different methods for measuring depressive symptoms and CES-D was the measure that was consistently used along with another measurement. Among the four, three studies used reliable and validated scales for depressive symptoms: POMS [19]; Veterans Short Form-12 [23]; Patient Health Questionnaire (PHQ-9) [25]. One study used CES-D and a single-item Depressive Symptom Self-Report (DSSR) that asked whether a respondent has had depressive symptoms in the past week (Yes/No) with subsequent items asking for the frequency and the impact of such symptoms [26]. The internal reliability of these depression measures range from good to excellent with Cronbach α = 0.90 for BDI [30], Cronbach alphas ranging from 0.88 to 0.94 across several studies using CES-D [19, 22, 26], Cronbach α = 0.91 for SCL-90-R [30], Cronbach α = 0.89 for PHQ-9, [33] and Cronbach α = 0.92 for SSC-HIVrev [27]. Internal reliability for POMS was calculated across the six subscales (including Depression-Dejection) with Cronbach alphas ranging from 0.78 to 0.95 over time [19]. The Veterans Short Form-12 is an abbreviated version of the the 36-item Veterans Short Form (VSF-36), with the 36-item having Cronbach alphas ranging from 0.80 to 0.95 in veteran study populations [34]. Validity for BDI and CES-D is also high when comparing the BDI to the Hamilton Rating Scale for Depression (r = 0.73), [21] when comparing the CES-D to Lubin’s Depression Adjective Check (r = 0.70), [35] and when comparing the single item DSSR with CES-D (r = 0.48) [26]. The POMS Depression-Dejection subscale is correlated with BDI (r = 0.69), [36] PHQ-9 highly correlates to the Short Form 20 Mental Health subscale (r = 0.73), [33] and the SCL-90-R also has demonstrated validity within psychiatric populations. [37, 38] Assessment Time Points Time points for assessing depressive symptoms pre- or post-intervention varied across studies (Fig. 3). All studies assessed depressive symptoms at baseline before any intervention was given. Most studies (n = 8) assessed depressive symptoms right after the completion of the intervention, while the remaining studies relied on follow-up assessments between 1 and 12 months post-intervention [19; 22; 26; 27; 29; 32]. Only one study assessed depressive symptoms each week during the intervention period [4]. All but one [26] studies assessed the depressive symptoms using the same validated measure at each time point. Eller et al. (2013) assessed baseline depressive symptoms using the CES-D and a single item. However, in this study, the CES-D was not used in their post-intervention follow-up assessment and only the single item was used. Fig. 3 Time Points for Assessment of Depressive Symptoms Description of Control Groups Participants randomized to the control arm of the included studies received usual care. However, the definition of the usual care was not described in most studies. When explicitly written, usual care was an education on management of HIV or in general health. Usual care in the control groups was delivered via face-to-face or paper-based forms and sometimes deviated from the delivery method of the intervention. For example, a study by Guo et al. (2020) delivered intervention content via web-based platform during a period of 3 months, while the control group only received a single brochure on nutrition guide. Description of Self-Management Interventions Interventions were delivered through technology (web-based or smartphones) [24, 25, 30, 32], interactive group sessions with professionals [20, 21, 23, 29], individual session(s) with professionals or lay workers [19, 22, 26], or a combination of these different approaches [27, 31]. For interactive group sessions, the description about the context/setting where intervention was delivered was not captured for all studies. The duration of the intervention varied substantially across the studies, ranging from a single session [26] to 12 weeks [32] or 3 months [24]. Most of the studies did not explicitly report on the exact length and frequency of each session within an intervention. Based on our operationalization and categories of self-management interventions, there were twelve that were categorized under coping, ten that were categorized under needs, and one that was categorized under resources. Among twelve interventions that addressed coping, three specifically focused on coping skills [19, 30, 31] either on an individual basis or in a group setting. Only one study addressed resources that fall into self-management process, and this study utilized a lay worker to provide skills and guidance on where individuals can find additional support and resources [22]. Three studies used Cognitive Behavioral Stress Management (CBSM) [20, 21, 24], which is a short-term self-management approach focused on emotional/behavioral regulation [39]. This approach aligned with needs and coping under self-management process. Whereas two studies that used CBSM occurred in person as a group [20, 21] Guo et al. (2020) adapted the CBSM into a web-based program (i.e., WeChat) where it allowed participants in the intervention group to self-educate on the intervention content on a weekly basis. Three studies used the HIV/AIDS Symptom Management Manual, which is a paper-based symptom management manual with self-care strategies for 21 common HIV/AIDS symptoms developed by a team of researchers at the UCSF School of Nursing [27]. However, the mode of delivery of the intervention differed, where Wantland et al. (2008) used a paper-based format that was administered at their participants’ discretion (self-teach), Eller et al. (2013) used the paper-based format delivered by a nurse, and Schnall et al. (2018) used a mobile platform to deliver the intervention content. Individualized coaching sessions were used in two studies. van Luenen et al. (2018) adapted a self-help booklet into a web-based program, followed by an individualized coaching session based on motivational interviewing with a master’s level graduate student in Psychology. Similarly, in a study led by Nyamathi et al. (2012), participants in the intervention group received the Asha Life intervention and six individualized coaching sessions at individuals’ homes with a lay community health worker who offered lessons in HIV, nutrition, and life skills. Effects of Intervention on Depressive Symptoms Among the included studies, one study that utilized clinician-delivered teleconferencing in group settings did not show any effect of the intervention on depressive symptoms [30]. Of the remaining 12 studies with statistically significant reduction in depressive symptoms in the intervention group, four studies [22, 23, 25, 32] showed significant differences in depressive symptoms between intervention and its control groups. These four studies focused on individualized/personalized content either via personal coaching [22, 23, 25] or selecting content based on participants’ needs via mHealth format [32]. Discussion In our review, we identified 13 articles focused on self-management of depressive symptoms in PLWH using an RCT design. Many of the interventions showed initial efficacy in reduction of depressive symptoms, providing evidence for the use of self-management interventions for ameliorating depressive symptoms in PLWH. Comparatively, there is substantial literature documenting the efficacy of psychostimulants, conventional antidepressants (e.g. selective serotonin reuptake inhibitors), and dehydroepiandrosterone, for the treatment of depression in HIV [39]. Nonetheless, many of these studies lack rigor (i.e. comparative studies, no follow-up data) making it difficult to draw a firm consensus [39]. Therefore, there remains a need for self-management interventions, such as those detailed in this review, as well as consideration of psychotherapy which also has demonstrated efficacy for treatment of depression [39]. Differences in depressive symptoms were noted between the intervention and control groups depending on the method of content delivery. The use of individualized/personalized approaches leveraging technology or face-to-face time with a coach/professional [22, 23, 25, 32] yielded significant decreases in depressive symptoms compared to those conducted in a manualized format or a group setting. Personalized and regular engagement with the intervention content over time may have provided reinforcement of the behaviors, and therefore may have contributed to the significant reduction in depressive symptoms. Traditional manualized self-care interventions did not reduce depressive symptoms beyond the initial study period. These findings support the notion that knowledge of self-management strategies alone does not translate into behavior change [40]. Most studies emphasized coping as the primary form of self-management through their intervention contents. Moreover, the largest proportion of significant decreases in depressive symptoms came from studies centered around needs identification and coping skills development and may suggest the high efficacy of needs and coping-based self-management strategies for targeting depressive symptoms. This presumption aligns with current research which cites the effectiveness of HIV-related depression management strategies [41]. Our review identified only one study that used resource utilization for the management of depression symptoms. A study by Nyamathi et al. focused on resource-based self-management through the incorporation of linkages to community resources for depression management within the intervention group [22]. Apart from this study, we found no self-management intervention focusing on resource referrals amongst PLWH. However, the nature of referrals to other resources including professional mental health support is separate from the concept of “self-management” and may not be within the scope of this review. Given that only one study focused on resources, it is premature to conclude that self- management interventions incorporating resources are effective and therefore future studies should assess how to include resource-based strategies in self-management interventions. Notably, none of the studies presented findings related to sex differences in depressive symptoms. This is noteworthy since, in our own work, we noted significant differences in the symptom experience of women as compared to men living with HIV [41]. On the other hand, menopausal status in women living with HIV has not been shown to have a relationship with depressive symptoms [41] pointing to the complex sex-based differences in the symptom experience of PLWH. There were several limitations of this review. First, the standard of care/control arm was poorly described in most studies making it difficult to fully identify the intervention effects. Second, there was heterogeneity in the measurement of depressive symptoms across studies. Future studies would be strengthened by comparing the effectiveness of these interventions and using the same standardized measurement tools for evaluation. It is especially important to use comparable validated measures, studies in this review did not. Thirdly, there were risks of bias that were inherent in the studies due to lack of reporting on whether the study was able to conceal group assignments. This is due to the nature of the behavioral intervention where assignment to intervention/control groups can be difficult to conceal in comparison to drug trials. However, studies should mention strategies that they used to conceal assignments so that there is no risk of bias from the perspective of the participants. Additionally, many studies had high dropout rates in the intervention group, indicating that the strategies for retaining participants and designing recruitment to mitigate attrition are critical to further develop before study enrollment. Finally, many of the studies did not fully describe the duration and frequency of the intervention sessions. Despite these limitations, most of the included studies showed a positive reduction in depressive symptoms in PLWH who were in the intervention group. Given these findings and the need to address this disproportionately prevalent and burdensome issue of depressive symptoms among PLWH, future work should consider how to best disseminate these interventions as well as the best approach for their implementation in clinical and community settings. The findings in this review attest to the challenges in dissemination and implementation of interventions, despite positive findings from clinical trials [42]. Conclusion Given the increased rates of depressive symptoms in PLWH [43] and the negative health impacts of these symptoms, including treatment non-adherence and poor disease prognosis [44], there is a critical need for identifying efficacious interventions for addressing depressive symptoms in PLWH. Opportunities for an individualized approach and personalization through individual coaching or technology can hold promise for reducing depressive symptoms. This review synthesized the evidence of self-management interventions and can be used to inform scale-up and dissemination of interventions that effectively improve depressive symptoms in PLWH. Authors’ Contributions All authors on this paper have contributed to the conception and design of the study, drafted or have been involved in revising this manuscript, reviewed the final version of this manuscript before submission, and agree to be accountable for all aspects of the work. Using the CRediT taxonomy, the specific contributions of each author are as follows: Conceptualization: M. Yoo Jeong, J. Barroso, R. Schnall; Methodology: M. Yoo Jeong; Data Curation: M. Yoo Jeong, G. Alvarez; Formal Analysis: M. Yoo Jeong, G. Alvarez; Resources: G. Alvarez; Supervision: R. Schnall; Writing – original draft: M. Yoo Jeong, R. Schnall; Writing – review and editing: G. Alvarez, G. Khawly, J. Voss, T. Wang, J. Barroso. Funding Research reported in this publication was supported by the National Institute of Nursing Research through the National Institute of Health under award numbers: R01NR015737 and K24NR018621. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The findings and conclusions in this review are those of the authors and do not necessarily represent the official positions of the official views of the National Institutes of Health. Data Availability (data transparency) Not applicable. Code Availability (software application or custom code) Not applicable. Declarations Conflicts of interest/Competing Interests The authors report no real or perceived vested interests related to this article that could be construed as a conflict of interest. All authors declare that they have no conflicts of interest. Ethics Approval This article does not contain any studies with human participants or animals performed by any of the authors. Consent to Participate Not applicable. Consent for Publication Not applicable. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Association AP. Diagnostic and statistical manual of mental disorders. 2013. 2. Health NIoM. Depression 2018 [Available from: https://www.nimh.nih.gov/health/topics/depression. 3. Wang T Fu H Kaminga AC Li Z Guo G Chen L Prevalence of depression or depressive symptoms among people living with HIV/AIDS in China: a systematic review and meta-analysis BMC Psychiatry 2018 18 1 160 10.1186/s12888-018-1741-8 29855289 4. Schnall R Siegel K Jia H Olender S Hirshfield S Racial and socioeconomic disparities in the symptom reporting of persons living with HIV AIDS Care 2018 30 6 774 83 10.1080/09540121.2017.1417532 29353489 5. Mitzel LD Vanable PA Carey MP HIV-Related stigmatization and medication adherence: Indirect Effects of Disclosure concerns and Depression Stigma and health 2019 4 3 282 92 10.1037/sah0000144 31681850 6. Tao J Vermund SH Qian HZ Association between Depression and Antiretroviral Therapy Use among People living with HIV: a Meta-analysis AIDS Behav 2018 22 5 1542 50 10.1007/s10461-017-1776-8 28439754 7. So-Armah K Gupta SK Kundu S Stewart JC Goulet JL Butt AA Depression and all-cause mortality risk in HIV-infected and HIV-uninfected US veterans: a cohort study HIV Med 2019 20 5 317 29 10.1111/hiv.12726 30924577 8. Lin MY Liu MF Hsu LF Tsai PS Effects of self-management on chronic kidney disease: a meta-analysis Int J Nurs Stud 2017 74 128 37 10.1016/j.ijnurstu.2017.06.008 28689160 9. Murphy LA Harrington P Taylor SJ Teljeur C Smith SM Pinnock H Clinical-effectiveness of self-management interventions in chronic obstructive pulmonary disease: an overview of reviews Chronic Resp Dis 2017 14 3 276 88 10.1177/1479972316687208 10. Reynolds R Dennis S Hasan I Slewa J Chen W Tian D A systematic review of chronic disease management interventions in primary care BMC Fam Pract 2018 19 1 11 10.1186/s12875-017-0692-3 29316889 11. Schrijver J Lenferink A Brusse-Keizer M Zwerink M van der Valk PD van der Palen J Self-management interventions for people with chronic obstructive pulmonary disease Cochrane Database Syst Rev 2022 1 1 Cd002990 35001366 12. Barlow J Wright C Sheasby J Turner A Hainsworth J Self-management approaches for people with chronic conditions: a review Patient Educ Couns 2002 48 2 178 10.1016/S0738-3991(02)00032-0 13. Ryan P Sawin KJ The Individual and Family Self-Management Theory: background and perspectives on context, process, and outcomes Nurs Outlook 2009 57 4 217 25.e6 10.1016/j.outlook.2008.10.004 19631064 14. Schulman-Green D Jaser S Martin F Alonzo A Grey M McCorkle R Processes of self-management in chronic illness J Nurs scholarship: official publication Sigma Theta Tau Int Honor Soc Nurs 2012 44 2 136 44 10.1111/j.1547-5069.2012.01444.x 15. Ould Brahim L Lambert SD Feeley N Coumoundouros C Schaffler J McCusker J The effects of self-management interventions on depressive symptoms in adults with chronic physical disease(s) experiencing depressive symptomatology: a systematic review and meta-analysis BMC Psychiatry 2021 21 1 584 10.1186/s12888-021-03504-8 34800995 16. Page MJ McKenzie JE Bossuyt PM Boutron I Hoffmann TC Mulrow CD The PRISMA 2020 statement: an updated guideline for reporting systematic reviews J Clin Epidemiol 2021 134 178 89 10.1016/j.jclinepi.2021.03.001 33789819 17. Popay J, Roberts H, Sowden A, Petticrew M, Arai L, Rodgers M, et al. Guidance on the conduct of narrative synthesis in systematic reviews: a product from the ESRC Methods Programme2006. 18. National Institutes of Health/National Heart L, and Blood Institute. Study Quality Assessment Tools 2021 [Available from: https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools. 19. Miles MS Holditch-Davis D Eron J Black BP Pedersen C Harris DA An HIV self-care symptom management intervention for african american mothers Nurs Res 2003 52 6 350 60 10.1097/00006199-200311000-00002 14639081 20. Jones DL Ishii Owens M Lydston D Tobin JN Brondolo E Weiss SM Self-efficacy and distress in women with AIDS: the SMART/EST women’s project AIDS Care 2010 22 12 1499 508 10.1080/09540121.2010.484454 20845112 21. Laperriere A Ironson GH Antoni MH Pomm H Jones D Ishii M Decreased Depression up to one year following CBSM + intervention in Depressed Women with AIDS: the Smart/EST Women’s Project J Health Psychol 2005 10 2 223 31 10.1177/1359105305049772 15723892 22. Nyamathi A Salem BE Meyer V Ganguly KK Sinha S Ramakrishnan P Impact of an Asha intervention on depressive symptoms among rural women living with AIDS in India: comparison of the Asha-Life and Usual Care program AIDS Educ prevention: official publication Int Soc AIDS Educ 2012 24 3 280 93 10.1521/aeap.2012.24.3.280 23. Chen WT Shiu C Yang JP Wang K Zhang L Zhang J Quality of life in HIV-infected chinese women and their family caregivers: an intervention study AIDS Care 2018 30 12 1572 9 10.1080/09540121.2018.1510095 30111163 24. Guo Y Hong YA Cai W Li L Hao Y Qiao J Effect of a WeChat-Based intervention (Run4Love) on depressive symptoms among people living with HIV in China: a Randomized Controlled Trial J Med Internet Res 2020 22 2 e16715 10.2196/16715 32044751 25. van Luenen S Garnefski N Spinhoven P Kraaij V Guided internet-based intervention for people with HIV and depressive symptoms: a randomised controlled trial in the Netherlands The lancet HIV 2018 5 9 e488-e97 30135045 26. Eller LS Kirksey KM Nicholas PK Corless IB Holzemer WL Wantland DJ A randomized controlled trial of an HIV/AIDS Symptom Management Manual for depressive symptoms AIDS Care 2013 25 4 391 9 10.1080/09540121.2012.712662 22880943 27. Wantland DJ Holzemer WL Moezzi S Willard SS Arudo J Kirksey KM A randomized controlled trial testing the efficacy of an HIV/AIDS symptom management manual J Pain Symptom Manag 2008 36 3 235 46 10.1016/j.jpainsymman.2007.10.011 28. Weiss SM Tobin JN Antoni M Ironson G Ishii M Vaughn A Enhancing the health of women living with HIV: the SMART/EST Women’s Project Int J women’s health 2011 3 63 77 10.2147/IJWH.S5947 21445376 29. Hecht FM, Moskowitz JT, Moran P, Epel ES, Bacchetti P, Acree M, et al. A randomized, controlled trial of mindfulness-based stress reduction in HIV infection. Brain, behavior, and immunity. 2018;73:331–9. 30. Heckman TG Carlson B A randomized clinical trial of two telephone-delivered, mental health interventions for HIV-infected persons in rural areas of the United States AIDS Behav 2007 11 1 5 14 10.1007/s10461-006-9111-9 16705479 31. Inouye J Flannelly L Flannelly KJ The effectiveness of self-management training for individuals with HIV/AIDS J Association Nurses AIDS Care: JANAC 2001 12 5 71 82 10.1016/S1055-3290(06)60264-1 32. Schnall R Cho H Mangone A Pichon A Jia H Mobile Health Technology for improving Symptom Management in Low Income Persons living with HIV AIDS Behav 2018 22 10 3373 83 10.1007/s10461-017-2014-0 29299790 33. Kroenke K Spitzer RL Williams JB The PHQ-9: validity of a brief depression severity measure J Gen Intern Med 2001 16 9 606 13 10.1046/j.1525-1497.2001.016009606.x 11556941 34. Jones D Kazis L Lee A Rogers W Skinner K Cassar L Health status assessments using the Veterans SF-12 and SF-36: methods for evaluating otucomes in the Veterans Health Administration J Ambul Care Manage 2001 24 3 68 86 10.1097/00004479-200107000-00011 11433558 35. Radloff L The CES-D Scale A self-report Depression Scale for Research in the General Population Appl Psychol Meas 1977 1 3 385 401 10.1177/014662167700100306 36. Nyenhuis DL Yamamoto C Luchetta T Terrien A Parmentier A Adult and geriatric normative data and validation of the profile of mood states J Clin Psychol 1999 55 1 79 86 10.1002/(SICI)1097-4679(199901)55:1<79::AID-JCLP8>3.0.CO;2-7 10100834 37. Ardakani A Seghatoleslam T Habil H Jameei F Rashid R Zahirodin A Construct validity of Symptom Checklist-90-Revised (SCL-90-R) and General Health Questionnaire-28 (GHQ-28) in patients with drug addiction and diabetes, and Normal Population Iran J Public Health 2016 45 4 451 9 27252914 38. Derogatis LRCP Confirmation of the dimensional structure of the scl-90: a study in construct validation J Clin Psychol 1977 33 4 981 9 10.1002/1097-4679(197710)33:4<981::AID-JCLP2270330412>3.0.CO;2-0 39. Carlson LE Toivonen K Subnis U Integrative approaches to stress management Cancer J (Sudbury Mass) 2019 25 5 329 36 10.1097/PPO.0000000000000395 40. Arlinghaus KR Johnston CA Advocating for Behavior Change with Education Am J Lifestyle Med 2018 12 2 113 6 10.1177/1559827617745479 30283247 41. Olson B Vincent W Meyer JP Kershaw T Sikkema KJ Heckman TG Depressive symptoms, physical symptoms, and health-related quality of life among older adults with HIV Qual life research: Int J Qual life aspects Treat care rehabilitation 2019 28 12 3313 22 10.1007/s11136-019-02271-0 42. Fuller SM Koester KA Erguera XA Wilde Botta E von Beetzen F Steward WT The collaborative care model for HIV and depression: patient perspectives and experiences from a safety-net clinic in the United States SAGE open medicine 2019 7 2050312119842249 10.1177/2050312119842249 31044076 43. Choi SK Boyle E Cairney J Collins EJ Gardner S Bacon J Prevalence, recurrence, and incidence of current depressive symptoms among people living with HIV in Ontario, Canada: results from the Ontario HIV Treatment Network Cohort Study PLoS ONE 2016 11 11 e0165816 10.1371/journal.pone.0165816 27802346 44. Nanni MG Caruso R Mitchell AJ Meggiolaro E Grassi L Depression in HIV infected patients: a review Curr psychiatry Rep 2015 17 1 530 10.1007/s11920-014-0530-4 25413636

36520336

PMC9753016

NO-CC CODE

2022-12-16 23:25:27

no

AIDS Behav. 2022 Dec 15;:1-16

utf-8

AIDS Behav

10.1007/s10461-022-03945-6

oa_other

==== Front SN Soc Sci SN Soc Sci Sn Social Sciences 2662-9283 Springer International Publishing Cham 550 10.1007/s43545-022-00550-1 Review Paper A missing theoretical element of online higher education student attrition, retention, and progress: a systematic literature review http://orcid.org/0000-0001-7298-6444 Rotar Olga olga.y.rotar@gmail.com grid.410682.9 0000 0004 0578 2005 Centre for Institutional Studies, National Research University Higher School of Economics, Moscow, Russia 15 12 2022 2022 2 12 27811 3 2021 18 10 2022 © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Online learning technologies have facilitated higher education in many ways, making it more flexible and available for learners with multiple life and work responsibilities. Yet information regarding graduation rates suggests that the vast majority of online learners drop out. By systematically analysing 30 empirical studies published between 2009 and 2020, this paper aims to highlight factors critical for online students’ attrition, retention, or progress, focusing on the adult student population. Four groups of factors influencing adult students’ online learning were identified: (a) student factors, (b) course factors, (c) social factors, and (d) support factors. These four groups are analysed and discussed in light of selected theoretical models on student attrition, retention, and progress. The results show that student support remains a missing element in these models. Finally, recommendations based on the study findings are offered. Keywords Attrition Retention Student progress Adult learning Distance education Online higher education Systematic literature review http://dx.doi.org/10.13039/501100000307 British Federation of Women Graduates GA-00764 Rotar Olga issue-copyright-statement© Springer Nature Switzerland AG 2022 ==== Body pmcIntroduction Online education is an attractive option for students with multiple responsibilities due to its flexible structure, lower costs, and opportunity to learn as suits the individual (Ilgaz and Gülbahar 2015; Ladell-Thomas 2012; Muljana and Luo 2019). However, these benefits are not always realised, and the majority of online learners do not graduate (Woodley and Simpson 2014). Although precise figures on dropout rates in online higher education are not available, the literature suggests that graduation rates in online programmes are much lower compared to those in a traditional setting. Simpson (2013) and Woodley and Simpson (2014) pointed out that in the UK, graduation rates from online programmes vary from 0.5 to 20%. The University of Phoenix in the US and the University of South Africa have 5% and 6% graduation rates, respectively (Woodley and Simpson 2014). The lower graduation rates from online programmes indicate what Simpson (2013) calls a “distance education deficit” (p. 105). Research on the dropout phenomenon is ongoing, and the issue remains an “elephant in the room” (Woodley and Simpson 2014, p. 462). Many studies have examined online learners’ experiences and perceptions to understand what contributes to their learning progress and success (see Hart 2012; Park and Choi 2009; Simpson 2004). Scholars emphasise the diversity of online students in regard to their backgrounds, personal characteristics and skills, and the complexity of factors that influence their online learning experience and behaviour. This paper brings together empirical evidence of students’ learning experiences in online higher education and discusses the result of a systematic literature review in relation to selected theoretical models on student attrition, retention, and progress. While online education provides multiple benefits for learners (see Coomley and Stephenson’s 2001 meta-analysis) and offers unprecedented opportunities for students to learn from where they are and at their own pace, that opportunity comes with high risk. Failure to complete the first online course may lead students to experience lower self-confidence or self-esteem, and discourage them from registering for other online courses (Moore and Kearsley 1996). Dropout experience can cause social isolation, economic loss (Rumberger 1987), and marginalisation (Sosu and Pheunpha 2019). An examination of theoretical models of student attrition, retention, and progress through the discussion of new empirical evidence can reveal weaknesses of these models and shed light on the problem of online students dropping out. To do this, this study addressed the following research questions:What factors affect adult student attrition, retention, or progress in online higher education? Which of these factors are underrepresented in the theoretical models selected for the analysis? The scope of the article is limited in two ways. Firstly, it looks at the adult student cohort in online higher education. The focus on adult students is justified by the great proportion of mature students enrolled in pre-COVID-19 online programmes (Pozdnyakova and Pozdnyakov 2017). Secondly, this paper focused on factors of student retention, attrition, and progress, rather than dropout factors, through the analysis of adults’ experiences of online learning. Theoretical background There have been numerous attempts to systematically explain the processes of student learning and decision making through theoretical models of attrition, retention, and progress. Among the most recognised, there are the models of Spady (1970), Tinto (1975), Bean and Metzner (1985), Kember (1995), and Rovai (2003), and Falcone (2011). Two of the early theoretical models developed by Spady (1970) and Tinto (1975) draw on Durkheim’s theory of suicide (Durkheim 1951). Durkheim argued that suicide is a result of the individual’s “malintegration” into society due to the dissonance of values or an “insufficient collective affiliation” (Tinto 1975, p. 91). Both Spady (1970) and Tinto (1975) found an analogy between Durkheim’s concept of suicide and student dropout. Spady’s (1970) Model of Undergraduate Dropout Process contains the following elements: students’ background, normative congruence, academic potential, friendship support, grade performance, intellectual development, social integration, satisfaction, and institutional commitment. This model suggests that the decision to drop out depends on a student’s lack of successful integration into the life of the educational institution, which is determined by social and academic factors. Tinto’s (1975) Model of Dropout Behaviour synthesised research on student attrition and Durkheim’s study, and depicted student learning as a process of social and academic interactions moving towards student integration. The model presumes that students’ backgrounds and personal characteristics determine their ability to integrate into the learning environment, interact with others, and this affects their social and academic outcomes (Eaton and Bean 1995). Tinto distinguished two types of student withdrawal or dropout: voluntary and forced. The difficulty with the application of Tinto’s model to the analysis of student dropout in online education is in that he considers forced withdrawal to be a result of “insufficient levels of academic performance (poor grades)” or “the breaking of established rules concerning proper social and academic behaviour” (p. 92). What he does not include in the model are the external factors that are likely to be faced by the online student population, primarily represented by adult students. Addressing this limitation, Bean and Metzner (1985) proposed a Model of Non-traditional Undergraduate Student Attrition, which stressed the influence of the external environment, e.g. financial or familial difficulties, professional workload, adult student socialisation, persistence, and level of goal commitment. This model contains the following elements: student background, academic characteristics, environmental factors, and academic and psychological outcomes. Due to the focus on the non-traditional student cohort, Bean and Metzner accounted for environmental factors, suggesting that barriers associated with the external environment can influence adult student integration. Another comprehensive theoretical framework of adult student progress in distance education was developed by Kember (1995). He drew on Tinto’s work, his own research, and an extensive literature review to theoretically explain the connections between the factors presented in the model. Pointing to the specific context of Tinto’s model, Kember argued that adult students often have additional family and work responsibilities, and face different barriers in their studies. Similarly to Tinto, Kember distinguished two types of integration—social and academic. Social integration is the ability to integrate learning with other life and work responsibilities, and academic integration is associated with integration to the programme and the relationship between the educational institution and the learner. To represent adults’ competing demands, Kember added individual characteristics into the model, such as gender, prior work and learning experience, and family status. He argued that these greatly impact retention or attrition. Kember’s model has been tested in several quantitative studies within different national settings (Woodley et al. 2001). Although the quantitative tests run by Kember showed the model is reliable (Kember 1995), Woodley et al. (2001) pointed out weaknesses in Kember’s inventory instrument. They argued that the individual items used in the instrument did not measure the intended concept and concluded that the model could not adequately explain adult student progress in distance education. Another model, the Composite Persistence Model, was proposed by Rovai (2003). He synthesised the models of Tinto, and Bean and Metzner. Rovai’s model consists of four elements: student characteristics, student skills, external factors, and internal factors. Finally, a detailed conceptual model was developed by Falcone (2011), where additional elements were added: self-efficacy (or habitus); social, economic, cultural, and other forms of capital; and different levels of belongings to communities within and beyond the educational institution. The latter element shapes students’ goals and educational and social commitments. All these elements influence students’ learning experiences, the perceptions of their academic fit, and behaviour regarding learning processes and progress. These theoretical models provide a useful conceptual framework to discuss factors identified through the systematic literature review, rather than limiting the analysis by the direct application of a particular theoretical model. Approach to the literature review The purpose of this article is to identify factors or elements that influence student learning experiences and analyse them against the considered theoretical models. To achieve this aim, I reviewed existing studies that reported empirical research results from 2009 to 2020. The Scopus database was used to search for relevant studies. The variety of key words and word combinations, such as “adult student”, “non-traditional student”, “online higher education”, “online education”, and “distance education” were used as search terms. Only studies published in the English language were reviewed. Additional studies were identified through a “snowball” method by using reference lists of the selected articles (Webster and Watson 2002). Initially, I identified 144 studies. From these, I excluded the following studies: (a) conducted in a K-12 setting; (b) were not empirical, e.g. opinion or conceptual papers; (c) doctoral theses; (d) conference papers; and (e) papers, publications in magazines, and reports which were not peer reviewed. Consequently, I selected 30 empirical studies on students’ experiences in online higher education that had been published in peer-reviewed journals. In the quantitative studies, only factors that were suggested as statistically significant have been considered. In the qualitative studies, factors that were critical for online students’ experience were included in the analysis. To distinguish and group factors, I employed the Constant Comparative Method for the analysis (Lincoln and Guba 1985). Below, I provide an explanation of the steps of the analysis. First, a large number of factors were selected from the reviewed studies according to their specific features. The label of each factor has been either borrowed from the study where it was reported or was created to reflect main features of the factor. The second step involved the comparison of the factors. The goal of the comparison was to distinguish conceptual similarities and differences between them (Boeije 2002). From the initial number of factors, I selected one and allocated it to the first category (e.g. individual characteristics). Then, I chose another factor and compared it to the first one, to decide whether this factor can be added to the first category or represents a new theme. In this way, all identified factors have been evaluated until I had 15 categories which were further combined into four groups (see Table 1).Table 1 Factors suggested as being critical for adult students’ online learning Factor Sub-factor Factor attributes Student factors Individual characteristics Individual characteristics (Xu and Jaggars 2014; Martin and Bolliger 2018) Academic background GPA (Willging and Johnson 2009; Knestrick et al. 2016; Cochran et al. 2014) Relevant experiences No experience of successful completion of any previous online courses (Hachey et al. 2012) Prior experience of withdrawal (Cochran et al. 2014) Student factors Skills Self-regulation skills (Lee et al. 2013; Geduld 2014) Self-efficacy (Geduld 2014; Backs 2017; Reilly et al. 2012; Harnett et al. 2011; Joo et al. 2015; Cox 2018) Self-regulation, and self-discipline (Lee et al. 2013) Competency in using information communication technologies (Beqiri et al. 2009; Pena and Yeung 2010; Cole et al. 2014) Time management (Cox 2018; Holder 2007; Ilgaz and Gülbahar 2015) Students’ expectations Students’ expectations about the difficulty of the course (Bourdeaux and Schoenack 2016; Pierrakeas et al. 2004) Expectations of the instructor’s feedback (Gaytan 2015; Martin and Bolliger 2018) Psychological attributes Persistence (Park and Choi 2009; Yang et al. 2017) Satisfaction with the course (Chyung et al. 1998; Noel-Levitz 2011) Students’ motivation (Kim and Frick 2011; Zaborova et al. 2017) Locus of control (Lee et al. 2013) Course factors Course design Course design (Li et al. 2017; Rienties and Toetenel 2016) Course flexibility The flexibility of online learning (Sorensen and Donovan 2017) Online modality (Wladis et al. 2014) Integration of learning with working experience (Kahu 2013) Relevancy of the course Relevancy of the course for personal needs (Yang et al. 2017) Higher importance of career development or personal development goals (Stoessel et al. 2015) Interest in and the utility of the programme for the student’s professional career (Yang et al. 2017) or personal development (Stoessel et al. 2015; Knestrick et al. 2016) Social factors Interactions Successful online interactions and relationships with other students (Baxter 2012; Burns 2013) Collaborative learning activities (Nistor and Neubauer 2010) Lack of interaction (Cole et al. 2012) Othering (Phirangee and Malec 2017) Online interactions (Phirangee and Malec 2017; Kuo and Belland 2016; Cole et al. 2014; Kuo et al. 2014) Engagement The amount of time spent on communication activities (Rienties and Toetenel 2016) Engagement (Martin and Bolliger 2018; Banna et al. 2015; Britt et al. 2015; Meyer 2014; Backs, 2017; Wlodkowski 2008; Chametzky 2013; Stone and O’Shea 2019) Connectedness Connectedness (Boyle et al. 2010; Johnson 2014) Teacher connection (Stone and O’Shea 2019) Social presence Social presence (Richardson et al. 2017) Support factors Institutional support Proactive support (Simpson 2013; Russo-Gleicher 2013) Tutors’ support and guidance (Brown and Wilsom 2016) Targeted, promoted, appropriate, and easily available support (Stone and O’Shea 2019) Instruction and feedback (Gaytan 2015) Embedded within the curriculum support (Stone 2017) External support Support from family and at the workplace (Park and Choi 2009; Pierrakeas et al. 2004; Lee et al. 2013) Results The review of the literature resulted in the differentiation of four groups of factors that have been suggested as critical for students’ online learning experience and contributed to either students’ attrition, retention, or progress. The four groups of factors are: student factors, course factors, social factors, and support factors. Table 1 provides a description of attributes of the identified categories of factors and their composite sub-factors. In the following part of the paper, I explain the role of each group of factors and its composite sub-factors on adult students’ online learning as they have been described in the empirical studies. I then analyse these factors in their relation to the considered theoretical models. Student factors Individual characteristics Evidence of the influence of students’ individual characteristics on their attrition, retention, or learning progress is mixed. For instance, Park and Choi (2009) found no significant difference between the students’ individual characteristics and their learning behaviour. They also concluded that gender, age, previous education and work experience have no significant impact on students’ attrition. Xu and Jaggars (2014) and Cochran et al. (2014) suggest students’ gender is a significant predictor of online students’ retention. Xu and Jaggars (2014) found that persistence and learning outcomes vary significantly among students of a different gender, race as well as level of academic preparation. Martin and Bolliger (2018) also analysed how gender, age, and previous online learning experience influence students’ perceptions of engagement strategies, which were associated with learning progress. In regard to gender, they found that the use of additional online resources for learning was more important for female than for male students. Regarding age, it was more important for younger students to receive regular updates or email reminders from the instructor than it was for older students. Similarly, Knestrick et al. (2016) named age among four important variables that can explain students’ leaving or withdrawal, reporting that students who are over 40 years old are twice more likely to leave their study programme before graduation. Surprisingly, another study suggested that students aged 50 years and over are at a lower risk of attrition due to their learning goal orientation towards personal development, and greater value of the opportunity for personal growth and development through learning (Stoessel et al. 2015). Overall, there is no consensus on the effect of the individual characteristics, academic background on the students’ attrition, retention, or progress. Yet, the student characteristics element is presented in all theoretical models selected for the analysis in this paper. Academic background The effect of students’ academic background on their learning is also not clear as the results of the analysed empirical studies are contradictory (see Willging and Johnson 2009; Knestrick et al. 2016; Cochran et al. 2014). Xu and Jaggars (2014) and Cochran et al. (2014) suggested that academic preparation and prior academic performance are significant predictors of online students’ retention. Knestrick et al. (2016) found that four variables, namely grade point average (GPA), specialty of the programme, student status (full time or part time), and age (younger than or over 40 years old) explain 27% of the absence or withdrawal from study. They also found that the number of earned credits is a significant predictor of absence and withdrawal. Interestingly, Willging and Johnson (2009) found that dropped out students tend to have a higher GPA, and students with the high employment status, e.g. director or manager, are less likely to discontinue their study. Although this study was conducted in the context of a single online program, these conclusions suggest that the effect of academic background on the learning progress in the context of online education should be further explored. Relevant experiences Past research demonstrated that students with little or no online learning experience are at greater risk of attrition (Cochran et al. 2014; Xu and Jaggars 2014). Xu and Jaggars (2014) argued that previous experience of dropout from online courses is a significant predictor of online students’ retention. They found that students with weaker academic backgrounds have “significantly stronger negative coefficients for online learning compared with their peers, in terms of both course persistence and course grade” (Xu and Jaggars 2014, p. 23). Hachey et al. (2012) revealed that students who had not successfully completed any previous online courses had low retention rates compared to those who successfully completed prior online classes, suggesting that previously unsuccessful online learners require additional support. Yet, Willging and Johnson (2009) found that students who completed their first two courses are more at risk of dropout. Li et al. (2017), in their investigation of learning experiences of new and continuing students, found that there is a dramatic difference in how online learning environments are experienced between the two groups. Specifically, continuing learners expressed much lower satisfaction (70% less than new students) if their learning was not aligned with their wider professional development aims. An overall recommendation from the research is that there is a need to identify and support online learners at risk, namely freshmen, those with lower GPA, and those with prior experience of attrition. Student skills There is substantial evidence for the effect of student skills on student attrition, retention, and online learning progress. This ranges from the ability to effectively allocate time and make realistic timetables, to academic self-efficacy, self-regulation, and self-discipline. Findings from qualitative studies suggest that competency in using information communication technologies is related to a greater satisfaction (Beqiri et al. 2009; Pena and Yeung 2010), which influences academic progress (Cole et al. 2014). Lai (2011) determined that the readiness for self-direction is a critical element of mature students’ learning progress. Hashim et al. (2015) argued that adult students, despite their assumed self-direction, require the same level of guidance and motivation as their younger peers (Hashim et al. 2015). Geduld (2014) showed that higher achieving students are more self-regulated, whereas those students who are lacking self-regulation are at a greater risk of attrition. Research has also shown the impact of time management skills on students’ academic progress (Cox 2018; Ilgaz and Gülbahar 2015). Students who are able to effectively allocate time and set up realistic timetables are more likely to be satisfied with their studies and successfully progress in an online course (Ilgaz and Gülbahar 2015). The lack of awareness of the effort required for online learning was the main cause of attrition for time-poor adult students (Romero and Barbera 2011). Another important concept that has been emphasised in the literature in relation to online learning is academic self-efficacy—a student’s confidence in their ability to perform the tasks successfully (Bandura 1997; Geduld 2014). Research indicates that a lack of self-efficacy negatively affects students’ academic progress (Shen et al. 2013), and may cause emotional stress, feelings of isolation (Betts 2009) and frustration (Artino and Stephens 2009). In a similar way, Backs (2017) pointed out the substantial negative impact of low levels of self-efficacy, arguing that students with a lack of self-efficacy for learning in an online environment are at risk of disengagement and attrition and, emphasising the communication with instructors and peers as an important support strategy. Psychological attributes Persistence, satisfaction with the course, motivation, locus of control and love of learning are important contributing factors for online student retention and learning progress. Persistence proved to be a strong predictor of students’ retention in online education (Yang et al. 2017). Yang et al. (2017) categorise persistence attributes into individual and programme factors. They emphasised the relevance of the course as an important programme attribute. Learning satisfaction is also an important indicator of online learning progress (Cole et al. 2014; Ilgaz and Gülbahar 2015; Lee 2014), although not all research supports this finding. For instance, Rienties and Toetenel (2016) did not find a significant relationship between satisfaction and online student retention, arguing that online learning is not necessarily a pleasant experience. Bourdeaux and Schoenack (2016) found that adult students’ satisfaction can be negatively affected when tutors do not meet their expectations, use pedagogical tools poorly, do not provide instructional clarity or show a lack of respect. Feedback from instructors was an important expectation of online learners (Gaytan 2015; Martin and Bolliger 2018). Specifically, students expect timely, meaningful, and clear and comprehensive feedback from their tutors so that they can improve their academic performance and make stable progress. Online students appreciate a development of more personal relationships with their tutors (Martin and Bolliger 2018). Another important psychological attribute is motivation (Kim and Frick 2011). Such factors as external commitments, lower tuition fees, an opportunity and flexibility to combine work and study, influence students’ motivation for learning, and, as a result, their retention (Zaborova et al. 2017). Harnett et al. (2011) argued that the level of students’ motivation for learning depends on a combination of factors, both internal and external. Among the internal factors for online student progress they identified interaction with instructors and tasks that are interesting, relevant, and applicable. Important external factors are family and work-related commitments. Locus of control, or a student’s perception of the causes or control over their learning, is another psychological factor that is critical for online student progress (Lee and Choi 2011). Based on the analysis of differences between successful and dropout students, Lee et al. (2013) found that successful online learners have a high level of locus of control, a greater feeling of responsibility for their learning, and are more self-regulated. Course factors Course design In a study on the influence of online programme design on the learning progress, Lee and Rha (2009) reported two main findings: learners who participated in a structured course expressed their satisfaction with the structure of the course, whereas learners who participated in the interactive course were more satisfied with interpersonal communication. Li et al. (2017) compared the learning experiences of a large sample of students (99,976 continuing students and 16,670 new students) and concluded that such design elements as assessment, learning materials, workload, and focus on career development strongly correlate with student retention. Another large study by Rienties and Toetenel (2016) also looked at the impact of the online course design on students’ (n = 111,256) satisfaction, learning behaviour and academic progress. The results provided insights for online educators on the importance of the social side of learning that goes in parallel with cognitive development. As Ladell-Thomas (2012) emphasised, structured content with diverse social activities and authentic tasks is an important expectation of online students. Course flexibility The flexibility of online learning mode is commonly presented as a benefit. Adult learners who choose such study have greater opportunities to combine their learning and other responsibilities (Kahu 2013). Nevertheless, the flexibility feature of online learning may be misinterpreted as learning that requires less commitment. Based on the responses from 396 dropout students, Sorensen and Donovan (2017) found that those who valued flexibility of online learning, specifically as an opportunity to work following an individualised schedule, were more likely to discontinue their study due to the difficulty to combine learning with other commitments. These findings suggest that online learning might not be suitable for those individuals who misjudge the concept of flexibility and struggle to juggle multiple responsibilities. Instead, more guided instruction and support may be needed for these learners. A similar insight on the need for more structure and guidance has been offered by Farrell et al. (2016). They argued that to enhance participation in online learning, students should be provided with comprehensive information regarding the programme schedule and required commitments. Relevance of the course The relevancy of the course has been highlighted as a critical factor for student progress in the analysed studies. Yang et al. (2017) found that the course relevance for students’ professional or personal needs have a significant impact on their persistence and progress. The importance of personal development as a result of learning has been also associated with student academic progress (Stoessel et al. 2015; Knestrick et al. 2016). As Stoessel et al. (2015) concluded, the alignment of learning with career development or personal development goals lowers the risk of online student attrition. Social factors Engagement A higher level of engagement enhances satisfaction with and motivation for learning, eliminates feelings of isolation, and positively impacts academic progress (Martin and Bolliger 2018). Banna et al. (2015), Britt (2015) and Meyer (2014) also emphasised that greater engagement leads to better academic progress due to the cognitive commitment and effort required for students’ cognitive development. Banna et al. (2015) stressed that while the quality of the content of learning materials played the main role in the past, for the successful self-directed learning engagement is more crucial. Online interactions Previous research relates academic progress in online higher education to factors that increase feelings of disconnection and isolation, including the lack of interaction between students (Phirangee and Malec 2017; Kuo and Belland 2016). Cole et al. (2014) determined that the rarity of interactions negatively affects learning. Similarly, Kuo and Belland (2016) identified that learner–content interactions and learner–instructor interactions were significant predictors of student satisfaction and retention in an online course. Martin and Bolliger (2018) found that students, particularly those who liked to work on collaborative group activities or assignments, valued interactions with peers and reported enjoying being involved in group discussions. Despite the importance of online interactions, there is evidence that online learners may face challenges to maintaining interactions in their courses, and do not always adapt well to the constructivist learning activities often used in online learning environments (Backs 2017). Connectedness Belongingness to the community, which implies a connection to a group or an institution, is critical for decreasing attrition rates (Boyle et al. 2010; Rovai 2003). Closely linked to the sense of belonging to the community is connectedness (Hart et al. 2011; Shackelford and Maxwell 2012). In the study conducted by Boyle et al. (2010), students reported little sense of connection and belonging to the learning community, and, as a result, dissatisfaction with learning and a lack of progress. Johnson (2014) also found that in an online environment that facilitating connectedness is critical for student retention as in an online learning environment there is a risk for a student to feel disconnected (Johnson 2014). Past research divided the concept of ‘connectedness’ into three themes: continuity (i.e. course tutor meeting with students at each study day); structure (university regulations, dates, and deadlines); and a ‘human touch’ (genuineness, caring, and commitment to students) (Carnwell et al. 2001) and suggest that if these aspects of connectedness are realised, disconnectedness and student attrition can be significantly reduced. Social presence It is argued that instructors should aim to foster social presence in order to support student retention and facilitate their learning progress. Richardson et al. (2017) suggest that social presence has an influence on students’ motivation and participation in online learning environment and may accurately predict student satisfaction. Moreover, Richardson et al. (2017) found that social presence may influence students’ online learning progress. A study conducted by Arbaugh (2014) also confirmed that social presence, in this study measured as a learning behaviour associated with an active perception of others, can carefully predict student satisfaction with learning, and consequently, their academic progress. Support factors Institutional support Research suggests that institutional support plays a critical role in ensuring student retention and progress (Simpson 2013). Stone (2017) and Stone and O’Shea (2019) stress the importance of the learning support that is “embedded within the curriculum as much as possible, hence delivering it where and when it is most needed” (Stone 2017, p.10). This may include support with academic skills, technological and personal services embedded into the course design. In other words, Stone (2017) advocates the inclusion of support elements into the content of the discipline, “integrated within the classroom task, and usually within the assessment task”. (Stone 2017, p. 10). Brown and Wilson (2016) found that less proactive students rely strongly on a study handbook, and guidance and support from the instructor in order to develop adequate skills and successfully progress in their learning. Russo-Gleicher (2013) proposed that instructors can contribute to students’ retention by merely monitoring and redirecting students to appropriate support services. This is in line with Jones (2010) who argues that academic caring is important for learners who study online and with Farrell et al. (2016) who state that online participation can be enhanced if the learners are provided with adequate information, guidance, and schedule. External support Scholars within the field of online and adult education often claim that although students are drawn to online learning for flexibility and convenience, some of them struggle to balance multiple priorities and require external support (see for instance Park and Choi 2009; Sorensen and Donovan 2017). However, Lee et al. (2013) did not find correlation between the completion of the online programme and existing support from family or an employer. The authors explain this contradiction by a possible influence of other variables that have been included in their analysis. Thus, external support may positively impact online student retention and progress, but its form and effect on students’ learning should be better explored (Simpson 2003). Synthesis of the literature This section provides a synthesis of the reviewed literature. Past research generates findings on multiple reasons for online student attrition, retention, and progress which are associated with student-related, program-related, social, and support factors. Although considered factors are discussed individually, this is done for the simplicity of analysis. The contradictory results of the past research promote the idea that these factors are interrelated and intertwined in their influence on online student learning. Some scholars found that student factors influence their preferences towards the course structure (Rienties et al. 2012; Ladell-Thomas 2012). The relation between student factors (e.g. student skills, academic background, experience in learning at a distance) and course factors is particularly noticeable within the misconception about the taken-for-granted flexibility of online learning. Although online students have an opportunity to study at any time and from any geographical location, they still must comply with course requirements and assessment deadlines. Furthermore, despite the proffered flexibility of online education, more structured courses suggested to improve student retention (Sorensen and Donovan 2017; Farrell et al. 2016). Not only the structure of the course, but the quality and relevance of the content determines the influence of social factors. For instance, the alignment of online learning objectives with career development or personal development needs of the learner may enhance student engagement and social presence (Stoessel et al. 2015). Therefore, the barriers associated with student factors can be eliminated by paying closer attention to course factors. When discussing support factors in relation to the other factor categories, an overall recommendation from scholars is a targeted approach of students at risk, meaning those with low GPA and a higher likelihood to withdraw. The discussion about the need for support often runs in parallel with consideration for a particular factor in the reviewed studies. Indeed, online students may require support at different stages of their learning (Rotar 2021) since their retention and progress depends on a combination of both internal and external factors. As Stone (2017), and Stone and O’Shea (2019) emphasise, learning support should be “embedded within the curriculum” (Stone 2017, p. 10), so it can address a wide range of attrition, retention, and progress determinants. Furthermore, the assumption of online students having good self-regulating skills should be questioned (Geduld 2014), contributing to the development of an educational institutions’ responsibility for student retention and progress. The assumptions about the self-directedness and self-motivation of adult online learners can lead to accelerating attrition rates in online education. Social factors also overlap with student factors on the issues of learning community, connectedness, social presence, and engagement. Although communication with instructors and peers is highlighted as an important support strategy (Rotar 2021), Lee and Rha (2009) pointed out that different online learners may have varying needs in the frequency of online interactions. Furthermore, due to the diversity of online students, a more personalised approach (Martin and Bolliger 2018) and a continuous adjustment of teaching practices are needed to address the negative impact of student factors on learning (Gaytan 2015). Thus, the consideration of social factors is only relevant when discussed in relation to the needs of an individual student. For example, Geduld (2014) states that only students with low self-regulation require additional support, since those who are able to realistically allocate time are more likely to be successful. Similarly, Lee and Rha (2009) argue that not all online students feel satisfied with group participation and would prefer individual work. Discussion Existing theoretical models suggest critical elements of students’ retention (Tinto 1975), attrition (Bean and Metzner 1985), persistence (Rovai 2003; Falcone 2011) or progress (Kember 1995) in either traditional or a distance learning setting. These well-known models emphasise the importance of academic and social integration for student retention and successful progress in learning. They suggest that unsatisfactory integration of the student into the social life of the educational institution, or an incompatibility with the learning demands are major causes for a student’s decision to withdraw from a course of study. Each model provides a comprehensive list of elements that should be considered when applying a model for analysis of student learning and/or a decision-making process. However, the review of empirical studies revealed new factors that may be associated with student attrition and are critical for students’ retention and academic progress in an online environment which are not explicit in the considered theoretical models. The systematic literature review identified four groups of factors that may influence learners in an online learning environment. Among them are student factors, course factors, social factors, and support factors. The revision of these factors in relation to the existing theoretical models of student attrition, retention, and progress revealed a lack of attention to the importance of student support in the previously formulated models and a weak emphasis on the influence of course factors, particularly the relevance of the course for students’ professional and personal development. All considered models incorporated students’ personal characteristics as an important element. The examined literature also suggests that the consideration of individual differences in understanding the students’ online learning experience is important and should not be neglected due to the peculiarity of the online student population. The influence of the individual differences with regard to age, gender, and previous educational background, personal characteristics, circumstances, commitments, and so on should be better examined in relation to their effect on students’ online learning experience due to the contradictory results in the empirical research. In all well-known models, the role of course factors is not fully explained. For instance, such a sub-factor as relevance of the course for professional development and student career seems to be missing in the considered theoretical models, despite its importance highlighted in the analysed empirical studies. More importantly, the criticality of support factors is not explicitly mentioned in the reviewed theoretical model, showing a significant gap in understanding the influence of student support, or its lack, for students’ attrition, retention, and progress. Yet, many scholars suggest that institutional support and external support can predict students’ persistence in online learning (Park and Choi 2009; Perry et al. 2008; Pierrakeas et al. 2004). A discussion of the need for additional, proactive support emerged from the fact that the majority of online students are adults employed either part or full time, and who possess additional family responsibilities. As a result of their busy lifestyles, online distance learners have fewer opportunities than their younger peers to interact directly with available institutional support services or have less immediate contact with their tutors. Prominent researchers in the field of online education Simpson (2013) and Woodley and Simpson (2014) provide examples of the successful interventions practices and argue that proactive institutional support is one of the effective ways to reduce online students’ attrition and improve retention rates. Nevertheless, this element is missing or not explicitly explained in the theoretical models of students’ attrition, retention, and progress evaluated in this paper. Such discrepancy may be explained by the fact that, according to Woodley and Simpson (2014), online students tend to blame themselves for their failure and underestimate the role of institutional support for their learning progress and success. Growing research in the area of online and adult learning indicates that for online students predominantly represented by the adult population of different ages and levels of commitments (Street 2010; Buck 2016) a supportive study environment and availability of support services are among the most significant factors of their successful learning (La Padula 2003; Buch 2016; Simpson 2013). Rather than limiting educational opportunities for those students, educational institutions should be ready to provide them with proactive support. Furthermore, when designing an online programme, online students’ needs and barriers for progress should be examined rather than assumed, and the role of the external support should be better investigated. Conclusion This study reviewed previous research on online student attrition, retention, and progress through the lens of theoretical models developed by Tinto (1975), Bean and Metzner (1985), Kember (1995), Rovai (2003), and Falcone (2011). The analysis demonstrates that factors which have been proposed to predict or explain student attrition, retention, or progress can be broadly split into four categories: student factors, course factors, social factors, and support factors. In discussing these factors, there is a common agreement that if the appropriate support is offered, some internal and external influences are likely to be mitigated. The three groups of factors, namely student factors, course factors and social factors, are apparent in the considered theoretical models. The significance of these factors is supported by many studies which employed a variety of research methods. However, the last group of factors, specifically support factors, seems to be neglected as this review revealed that this factor is not presented in the considered models. The importance of student support is emphasised in the analysed studies in various ways. For instance, when discussing student skills, scholars point out that scaffolding is one way to address the issue of student attrition. The psychological barriers to online learning can also be eliminated by support interventions. Furthermore, since this paper focused on the adult student population, the role of support proved to be even more critical for adults’ success due to the presence of multiple commitments in their lives. Online students with multiple responsibilities have various constraints in their learning. While they can successfully manage most of their challenges, student support plays a significant role in affecting (positively or negatively) their learning experience. Although Kember’s model (1995) suggests that a supportive environment and encouragement are necessary for successful social integration, support interventions for creating such an environment are not fully investigated. Further research is needed to integrate support factors into the theoretical models of online students’ attrition, retention, and progress. The course factors are also not emphasised in the considered models. This excludes the possibility of addressing the issues of students’ misconception around the flexibility of online learning. Furthermore, a lack of awareness of the importance of course and social factors dismiss the need for pedagogical and instructional design training of online educators who should regularly evaluate and develop technological, communication, and facilitation skills. The flexible nature of online education enables adult students, with a variety of commitments, to integrate more successfully academically, professionally, socially, and psychologically to their learning if they are adequately supported. The absence of support factors in the considered theoretical models requires a re-examination of how we address the problem of student attrition. A popular approach is to assume the influence of a great variety of factors on students’ learning experience yet neglect the role of student support. In existing models, it appears to be an unexamined assumption that individual learners are fully responsible for their successful adaptation into the academic and social life of their educational institution. Based on the identified limitations, the following recommendations can be provided:Online student populations vary significantly in regard to the student factors. Due to this heterogeneity, a more personalised approach, and a continuous adjustment of teaching practices during the teaching and learning process are needed to address the negative impact of student factors on learning. A proffered flexibility of online learning may create misconceptions around the potential academic and time commitments. Thus, to address the negative influence of the course factors an adequate information about the course structure and course requirements should be provided to the prospective students prior to enrolment. A development of technological and pedagogical skills of educators proved to eliminate the negative effect of social and course factors on student learning. Professional development programmes that cover pedagogical and instructional design training should be offered to educators. Student support is only relevant and can address a negative effect of a particular factor if offered on time and at the right stage of learning. Therefore, it is recommended to embed support interventions within different stages of the learning cycle to address the wide variety of internal and external factors. A revision of existing models to include elements of course factors and support factors into their structure is recommended. Inclusion of support and course factors in the theoretical models can be a way forward for addressing the problem of online student attrition and increase retention rates in online classes and programs. To conclude, current research appears to be moving away from understanding of how to better support online students to a more extensive examination of attrition, retention, and progress factors that constitute different theoretical models. The results of this systematic literature review highlight a fundamental problem of neglecting the role of institutional and external support on student learning in an online learning environment. Given the lack of attention to the element of support in theoretical models, further research needs to explore the importance of proactive institutional and external support in ensuring online students’ success. Limitations The aim of this paper was to shed light on the structural limitations of the considered theoretical models of students’ attrition, retention, and progress. This paper, however, did not intend to examine recommendations developed alongside or on the basis of these models. Following this approach, the results suggest that student support is a missing structural element in the models’ architecture. However, the results of the review cannot be generalised for all student groups due to the specific interest of this paper in the adult student population. Finally, it is important to note that the number of the revised studies is limited due to the application of the inclusion and exclusion criteria to maintain the focus of the study. Abbreviations GPA Grade point average OU Open university Acknowledgements I would like to thank Professor Don Passey for his suggestions to improve the final version of this manuscript. Author contributions Not applicable. Funding This research was supported by the FfWG of the British Federation of Women Graduates, Ref: GA-00764. Availability of data and material Data for the analysis were extracted from publications identified through the Scopus databases and can be accessed online. Declarations Conflict of interest No competing interests have been identified. Ethical approval Not applicable. Informed consent Not applicable. ==== Refs References Arbaugh JB System, scholar or students? Which most influences online MBA course effectiveness? J Comput Assist Learn 2014 30 4 349 362 10.1111/jcal.12048 Artino AR Jr Stephens JM Academic motivation and self-regulation: a comparative analysis of undergraduate and graduate students learning online Internet High Educ 2009 12 3–4 146 151 10.1016/j.iheduc.2009.02.001 Backs A (2017) Promoting online learner self-efficacy through instructional strategies and course supports (Doctoral dissertation, Capella University) Bandura A (1977) Social Learning Theory. Englewood Cliffs, NJ: Prentice Hall Bandura A Guide for constructing self-efficacy scales Self Effic Beliefs Adolesc 2006 5 1 307 337 Banna J Lin M-FG Stewart M Fialkowski MK Interaction matters: strategies to promote engaged learning in an online introductory nutrition course J Online Learn Teach 2015 11 2 249 261 27441032 Baxter J (2012) Who am I and what keeps me going? Profiling the distance learning student in higher education. International Review of Research in Open and Distributed learning, 13(4), 107–129. Bean JP Metzner BS A conceptual model of nontraditional undergraduate student attrition Rev Educ Res 1985 55 4 485 540 10.3102/00346543055004485 Beqiri MS Chase NM Bishka A Online course delivery: an empirical investigation of factors affecting student satisfaction J Educ Bus 2009 85 2 95 100 10.1080/08832320903258527 Betts K (2009) Lost in translation: Importance of effective communication in online education. Online J of Distance Learning Administration, 12(2):1–14 Boeije H (2002) A purposeful approach to the constant comparative method in the analysis of qualitative interviews. Qual Quant 36(4):391–409 Botha J Coetzee M The influence of biographical factors on adult learner self-directedness in an open distance learning environment Int Rev Res Open Distance Learn 2016 17 4 242 263 10.19173/irrodl.v17i4.2345 Bourdeaux R Schoenack L Adult student expectations and experiences in an online learning environment J Contin High Educ 2016 64 3 152 161 10.1080/07377363.2016.1229072 Boyle F, Kwon J, Ross C, Simpson O (2010) Student–student mentoring for retention and engagement in distance education. Open Learning: The Journal of Open, Distance and e-Learning, 25(2):115–130 Britt M Goon D Timmerman M How to better engage online students with online strategies Coll Stud J 2015 49 3 399 404 Brown CJ, Wilson CB (2016) One university making a difference in graduate education: Caring in the online learning environment. J of Holistic Nursing 34(4):402–407 Buck S (2016) In their own voices: Study habits of distance education students. J of Library & Information Services in Distance Learning, 10(3–4):137–173 Burns M (2013) Staying or leaving? Designing for persistence in an online educator training programme in Indonesia. Open Learning: The J of Open, Distance and e-Learning, 28(2):141–152 Carnoy M, Rabling BJ, Castano-Munoz J, Montoliu JMD, Sancho-Vinuesa T (2012) Who attends and completes virtual universities: the case of the open university of Catalonia (UOC). High Educ 63(1):53–82 Carnwell R, Moreland N, Helm R (2001) Co-opting learners: addressing their learning support needs through a learning support needs questionnaire. part two. a comparison of the learning support needs of campus and distance based community nurse students. Res Post-Compulsory Educ 6(1):51–66 Chametzky B (2013) Offsetting the affective filter: A classic grounded theory study of post-secondary online foreign language learners (Doctoral dissertation, Northcentral University) Cheng B Wang M Yang SJH Peng J Acceptance of competency-based workplace e-learning systems: effects of individual and peer learning support Comput Educ 2011 57 1 1317 1333 10.1016/j.compedu.2011.01.018 Chyung Y, Winiecki DJ, Fenner JA (1998) A case study: Increase enrollment by reducing dropout rates in adult distance education. (ERIC Document Reproduction Service No. ED 422 848) Cochran JD Campbell SM Baker HM Leeds EM The role of student characteristics in predicting retention in online courses Res High Educ 2014 55 1 27 48 10.1007/s11162-013-9305-8 Cole MT Shelley DJ Swartz LB Online instruction, e-learning, and student satisfaction: a three year study Int Rev Res Open Distrib Learn 2014 15 6 111 131 Coleman JA, Furnborough C (2010) Learner characteristics and learning outcomes on a distance Spanish course for beginners. Sys 38(1):14–29 Coomey M, Stephenson J (2001) Online learning: it is all about dialogue, involvement, support and control-according to the research. Teaching and learning online: pedagogies for new technologies, pp 37–52 Cornelius S Gordon C Ackland A Towards flexible learning for adult learners in professional contexts: an activity-focused course design Interact Learn Environ 2011 19 4 381 393 10.1080/10494820903298258 Cox A (2018) The online alone together paradox as a context for incidental and informal learning. New Directions for Adult and Continuing Education, 2018(159):71–89 Durkheim E (1951) Suicide: a study in sociology (J. A. Spaulding, & G. Simpson, Trans.). Free Press, Glencoe (Original work published 1897) Dwyer D Thompson DE Thompson CK Adult learners' perceptions of an undergraduate HRD degree completion programme: reasons for entering, attitudes towards programme and impact of programme Assess Eval High Educ 2013 38 2 131 141 10.1080/02602938.2011.608906 Eaton SB Bean JP An approach/avoidance behavioral model of college student attrition Res High Educ 1995 36 6 617 645 10.1007/BF02208248 Falcone T (2011) Toward a new model of persistence in higher education. In: ERIC, Paper presented at the Annual Meeting of the Association for the Study of Higher Education. Charlotte, NC, Nov 19, 2011. https://eric.ed.gov/?id=ED530661. Accessed 30 May 2020 Farrell B Ward N Jennings B Jones C Jorgenson D Gubbels-Smith A Kennie N Participation in online continuing education Int J Pharm Pract 2016 24 1 60 71 10.1111/ijpp.12202 26173959 Geduld B Self-Directedness in open distance learning: twists and turns Int J Instr Technol Distance Learn 2014 11 9 23 Gravani MN (2015) Adult learning in a distance education context: theoretical and methodological challenges. Int J Lifelong Educ 34(2):172–193 Gaytan J (2015) Comparing faculty and student perceptions regarding factors that affect student retention in online education. American J of Distance Edu 29(1):56–66 Hachey AC Wladis CW Conway KM Is the second time the charm? Investigating trends in online re-enrollment, retention and success J Educ Online 2012 9 1 1 25 Hart C (2012) Factors associated with student persistence in an online program of study: a review of the literature. J Interact Online Learn 11(1):19–42 Hartnett M, St George A, Dron J (2011) Examining motivation in online distance learning environments: Complex, multifaceted, and situation-dependent. Int Review of Research in Open and Distributed Learn 12(6):20–38 Hashim R Ahmad H Abdullah CZ Assessing the attitudes of distance learners toward the use of ICT in education Turk Online J Distance Educ 2010 11 2 125 134 Hashim R Ahmad H Ahmad N Evaluation of andragogy: cognitive engagement of distance learners Int J Technol Enhanc Learn 2011 3 3 322 331 10.1504/IJTEL.2011.040228 Hashim KF, Tan FB, Rashid A (2015) Adult learners' intention to adopt mobile learning: A motivational perspective. Br J Educ Technol 46(2):381–390 Ilgaz H Gülbahar Y A snapshot of online learners: E-readiness, e-satisfaction and expectations Int Rev Res Open Dist Learn 2015 16 2 171 187 Jan YI Kao C Huang C Chang K-W Exploring adult learners' preferences toward online learning environments: the role of internet self-efficacy and attitudes Anthropologist 2013 16 3 487 494 10.1080/09720073.2013.11891374 Johnson S Applying the seven principles of good practice: technology as a lever-in an online research course J Interact Online Learn 2014 13 2 41 50 Jones, B. D. (2010). An examination of motivation model components in face-to-face and online instruction. Electronic J of Research in Educ Psychology, 8(3):915–44 Joo YJ, Oh E, Kim SM (2015) Motivation, instructional design, flow, and academic achievement at a Korean online university: A structural equation modeling study. J of Comput in Higher Educ 27(1):28–46 Kahu ER (2013) Framing student engagement in higher education. Studies in Higher Educa 38(5):758–773. Kim KJ, Frick TW (2011) Changes in student motivation during online learning. J of Educ Comput Research 44(1):1–23 Kember D Open learning courses for adults: a model of student progress 1995 Englewood Cliffs Educational Technology Publications Knestrick JM Wilkinson MR Pellathy TP Lange-Kessler J Katz R Compton P Predictors of retention of students in an online nurse practitioner program J Nurse Practitioners 2016 12 9 635 640 10.1016/j.nurpra.2016.06.011 Kuo YC, Walker AE, Schroder KE, Belland BR (2014) Interaction, Internet self- efficacy, and selfregulated learning as predictors of student satisfaction in online education courses. The Int and Higher Educ 20:35–50 Kuo Y Belland BR An exploratory study of adult learners’ perceptions of online learning: minority students in continuing education Educ Technol Res Dev 2016 64 4 661 680 10.1007/s11423-016-9442-9 Ladell-Thomas J Do-it-yourself information literacy: self-directed learning at a distance J Libr Inf Serv Dist Learn 2012 6 3–4 376 386 10.1080/1533290X.2012.705168 Lai HJ (2011) The influence of adult learners' self-directed learning readiness and network literacy on online learning effectiveness: A study of civil servants in Taiwan. J of Educ Technol and Soc 14(2):98–106. LaPadula M (2003) A comprehensive look at online student support services for distance learners. The American J of Distance Educ 17(2):119–128 Lee J An exploratory study of effective online learning: assessing satisfaction levels of graduate students of mathematics education associated with human and design factors of an online course Int Rev Res Open Distrib Learn 2014 15 1 111 132 Lee Y, Choi J (2011) A review of online course dropout research: Implications for practice and future research. Educ Technol Res Develop 59(5):593–618 Lee Y Choi J A structural equation model of predictors of online learning retention Internet High Educ 2013 16 36 42 10.1016/j.iheduc.2012.01.005 Lee HJ Rha I Influence of structure and interaction on student achievement and satisfaction in web-based distance learning J Educ Technol Soc 2009 12 4 372 382 Lee Y Choi J Kim T Discriminating factors between completers of and dropouts from online learning courses Br J Edu Technol 2013 44 2 328 337 10.1111/j.1467-8535.2012.01306.x Li N, Marsh V, Rienties B, Whitelock D (2017) Online learning experiences of new versus continuing learners: a large-scale replication study. Assessment & Eval in High Educa 42(4):657–672 Li N Marsh V Rienties B Whitelock D Online learning experiences of new versus continuing learners: a large-scale replication study Assess Eval High Educ 2017 42 4 657 672 10.1080/02602938.2016.1176989 Lincoln YS Guba EG Naturalistic inquiry 1985 Newbury Park Sage Martin F, Bolliger DU (2018) Engagement matters: Student perceptions on the importance of engagement strategies in the online learning environment. Online Learning 22(1):205–222 Meyer KA (2014) Student engagement in online learning: What works and why. ASHE higher education report, 40(6):1–114 Meyers CA Bagnall RG The challenges of undergraduate online learning experienced by older workers in career transition Int J Lifelong Educ 2017 36 4 442 457 10.1080/02601370.2016.1276107 Moore MG Kearsley G Distance education: a systems view 1996 Belmont Wadsworth Publishing Muljana PS Luo T Factors contributing to student retention in online learning and recommended strategies for improvement: a systematic literature review J Inf Technol Educ Res 2019 18 19 59 Nistor N, Neubauer K (2010) From participation to dropout: Quantitative participation patterns in online university courses. Comput and Educ 55(2):663–672 Noel-Levitz (2011) National online learners priorities report. Coralville, IA: Author Park JH Choi HJ Factors influencing adult learners' decision to drop out or persist in online learning J Educ Technol Soc 2009 12 4 207 217 Park CL Boman J Care WD Edwards M Perry B Persistence and attrition: what is being measured? J Coll Stud Retent Res Theory Pract 2008 10 2 223 233 10.2190/CS.10.2.g Pena M Yeung A Satisfaction with online learning: does students’ computer competence matter? Int J Technol Knowl Soc 2010 6 97 108 10.18848/1832-3669/CGP/v06i05/56148 Phirangee K, Malec A (2017) Othering in online learning: an examination of social presence, identity, and sense of community. Distance Educa 38(2):160–172 Pierrakeas C Xenos M Panagiotakopoulos C Vergidis D A comparative study of dropout rates and causes for two different distance education courses Int Rev Res Open Distrib Learn 2004 5 2 1 15 Pozdnyakova O Pozdnyakov A Adult students’ problems in the distance learning Procedia Eng 2017 178 243 248 10.1016/j.proeng.2017.01.105 Reilly JR, Gallagher-Lepak S, Killion C (2012) Me and my computer: emotional factors in online learning. Nursing Educ perspectives 33(2):100–105 Remedios R, Richardson JT (2013) Achievement goals and approaches to studying: evidence from adult learners in distance education. Distance Educ 34(3):271–289 Richardson JC Maeda Y Lv J Caskurlu S Social presence in relation to students' satisfaction and learning in the online environment: a meta-analysis Comput Hum Behav 2017 71 402 417 10.1016/j.chb.2017.02.001 Rienties B, Kaper W, Struyven K, Tempelaar DT, Van Gastel L, Vrancken S, Virgailaite-Meckauskaite E (2012) A review of the role of Information Communication Technology and course design in transitional education practices. Interactive Learning Environments 20(6):563–581 Rienties B, Toetenel L (2016) The impact of learning design on student behaviour, satisfaction and performance: A cross-institutional comparison across 151 modules. Comput in Human Behavior 60:333–341 Rienties B Toetenel L The impact of learning design on student behaviour, satisfaction and performance: a cross-institutional comparison across 151 modules Comput Hum Behav 2016 60 333 341 10.1016/j.chb.2016.02.074 Romero M, Barbera E (2011) Quality of learners time and learning performance beyond quantitative time-ontask. Int Review of Research in Open and Distributed Learning 12(5):125–137 Rotar O (2021) Online student support: a framework for embedding support interventions into the online learning cycle (Preprint 10.31234/osf.io/s4kr6) Rovai AP (2003) In search of higher persistence rates in distance education online programs. The internet and Higher Educ 6(1):1–16 Rovai AP In search of higher persistence rates in distance education online programs Internet High Educ 2003 6 1 1 16 10.1016/S1096-7516(02)00158-6 Rumberger RW High school dropouts: a review of issues and evidence Rev Educ Res 1987 57 2 101 121 10.3102/00346543057002101 Russo-Gleicher R Qualitative insights into faculty use of student support services with online students at risk: implications for student retention J Educ Online 2013 10 1 1 32 Shackelford JL Maxwell M Sense of community in graduate online education: contribution of learner to learner interaction Int Rev Res Open Distrib Learn 2012 13 4 228 249 Shen D Cho MH Tsai CL Marra R Unpacking online learning experiences: online learning self-efficacy and learning satisfaction Internet High Educ 2013 19 10 17 10.1016/j.iheduc.2013.04.001 Simpson O (2004) Access, retention and course choice in online, open and distance learning. 3rd Eden Research Workshop, Oldenburg, Germany. http://www.eurodl.org/materials/contrib/2004/Ormond_Simpson Simpson O (2003) Student retention in online, open and distance learning. London: Kogan Page Simpson O Supporting students for success in online and distance education 2013 New York Routledge Sorensen C, Donovan J (2017) An examination of factors that impact the retention of online students at a for profit university. Online Learning 21(3):206–221 Sosu EM Pheunpha P Trajectory of university dropout: investigating the cumulative effect of academic vulnerability and proximity to family support Front Educ 2019 4 6 10.3389/feduc.2019.00006 Spady WG Dropouts from higher education: an interdisciplinary review and synthesis Interchange 1970 1 1 64 85 10.1007/BF02214313 Stoessel K Ihme TA Barbarino ML Fisseler B Stürmer S Sociodemographic diversity and distance education: who drops out from academic programs and why? Res High Educ 2015 56 3 228 246 10.1007/s11162-014-9343-x Stone C (2017) Opportunity through online learning: improving student access, participation and success in higher education. NCSEHE 2016 Equity Fellowship Final Report. Retrieved from the National Centre for Student Equity in Higher Education website https://www.ncsehe.edu.au/publications/opportunity-online-learning-improving-student-accessparticipation-success-higher-education/ Stone C O’Shea S Older, online and first: recommendations for retention and success Australas J Educ Technol 2019 35 1 57 69 10.14742/ajet.3913 Street H (2010) Factors influencing a learner’s decision to drop-out or persist in higher education distance learning. Online J of Distance Learning Administration 13(4):1–5 Tinto V Dropout from higher education: a theoretical synthesis of recent research Rev Educ Res 1975 45 1 89 125 10.3102/00346543045001089 Webster J, Watson RT (2002) Analysing the past to prepare for the future: writing a literature review. MIS quarterly, pp xiii–xxiii Willging PA Johnson SD Factors that influence students' decision to dropout of online courses J Asynchronous Learn Netw 2009 13 3 115 127 Wladis C, Hachey AC, Conway K (2014) An investigation of course-level factors as predictors of online STEM course outcomes. Comput and Educ 77:145–150 Wlodkowski R (2008) Enhancing adult motivation to learn: A comprehensive guide for teaching all adults. San Francisco, CA: Jossey-Bass Woodley A de Lange P Tanewski G Student progress in distance education: Kember's model re-visited Open Learn J Open Distance e-Learn 2001 16 2 113 131 10.1080/02680510123105 Woodley A, Simpson O (2014) The elephant in the room.&nbsp;Online distance education: towards a research agenda, pp 459–485 Xu D Jaggars SS Performance gaps between online and face-to-face courses: differences across types of students and academic subject areas J High Educ 2014 85 5 633 659 10.1353/jhe.2014.0028 Yang D Baldwin S Snelson C Persistence factors revealed: students’ reflections on completing a fully online program Distance Educ 2017 38 1 23 36 10.1080/01587919.2017.1299561 Zaborova EN, Glazkova IG, Markova TL (2017) Distance learning: Students’ perspective. Sociological Studies 2(2):131–139

0

PMC9753023

NO-CC CODE

2022-12-16 23:26:19

no

SN Soc Sci. 2022 Dec 15; 2(12):278

utf-8

SN Soc Sci

10.1007/s43545-022-00550-1

oa_other

==== Front Prev Med Prev Med Preventive Medicine 0091-7435 1096-0260 Elsevier Inc. S0091-7435(21)00047-5 10.1016/j.ypmed.2021.106463 106463 Short Communication Disparities in the distribution of COVID-19 testing sites in black and Latino areas in new York City Grigsby-Toussaint Diana S. a⁎ Shin Jong Cheol b Jones Antwan cd a Department of Behavioral and Social Sciences, Department of Epidemiology, Brown University School of Public Health, United States of America b Department of Behavioral and Social Sciences, Brown University School of Public Health, United States of America c Department of Sociology, George Washington University, United States of America d Department of Epidemiology, George Washington University, United States of America ⁎ Corresponding author at: Department of Behavioral and Social Sciences, Department of Epidemiology, School of Public Health, Brown University, United States of America. 26 2 2021 6 2021 26 2 2021 147 106463106463 19 6 2020 7 2 2021 20 2 2021 © 2021 Elsevier Inc. All rights reserved. 2021 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. In New York City (NYC), there are disproportionately more cases and deaths from COVID-19 for Blacks and Latinos compared to Whites. Using data from the NYC coronavirus data repository and the 2018 American Community Survey 5-year census estimates, we examined the distribution of testing sites across NYC areas (zip code tabulation areas) by race in May 2020. ArcGIS was used to create majority race zip code-level maps showing the distribution of testing sites on May 1, 2020 and May 17, 2020 in NYC. t-tests were used to determine whether significant differences existed in the number of testing sites by the majority race of zip codes. Between May 1, 2020 and May 17, 2020, testing sites in majority Black areas increased by more than 240% from nine to 31, and more than 90% from 16 to 31 in majority Latino areas. Black (M = 1257.7) and Latino (M = 1662.3) areas had significantly more COVID-19 cases (p < 0.05) compared to White areas. Nonetheless, White (n = 70; 38.9%) areas had most of the 180 testing sites on May 17, 2020, compared to Black (n = 31;17.2%) and Latino (n = 31;17.2%) areas. Due to the socio-economic and underlying health conditions that may place Blacks and Latinos at high risk for COVID-19, it is imperative that access to testing is improved for vulnerable groups. Keywords Disparities Black African-American Hispanic Latino Access Testing COVID-19 ==== Body pmc1 Introduction The novel severe acute respiratory coronavirus 2 (SARS-Cov-2) has infected more than 1.5 million Americans as of May 23, 2020, leading to approximately 95,000 deaths due to COVID-19, the disease caused by the virus.(Gorbalenya et al., 2020; Johns Hopkins Coronavirus Resource Center, 2020) New York City (NYC), initially considered the epicenter of COVID-19 in the US, has been hit especially hard by the virus. As of May 21, 2020, 192,840 cases of COVID-19 have been identified in NYC, along with 16,233 confirmed deaths.(Health, 2020) Non-Hispanic Black (Black) and Latino New Yorkers, however, have experienced the greatest burden of COVID-19. Recent estimates indicate 1470 cases of COVID-19 per 100,000 Blacks, 1307 cases per 100,000 Latinos, compared to 934 cases per 100,000 non-Hispanic Whites (White) in NYC.(Health, 2020) Moreover, Black (208 per 100,000) and Latino (217 per 100,000) New Yorkers are dying at much higher rates of COVID-19 compared to White New Yorkers (104 per 100,000).(Health, 2020) Anecdotal evidence highlights racial and ethnic disparities in access to testing that is critical for identifying cases and preventing the spread of COVID-19 among these groups.(Farmer, 2020) However, no studies at the time that this research was performed had explored the distribution of testing sites within predominately Black and Latino areas in NYC. Moreover, few studies have accounted for the potential impact of the introduction of the Paycheck Protection Program and Health Care Enhancement Act (PPP) on the availability of testing options.(Probasaco, 2020) The PPP was signed on April 24, 2020 to provide $25 billion for increased COVID-19 testing across the US.(Probasaco, 2020) Visualizing the distribution of testing sites is critical for determining how to target efforts addressing testing access for vulnerable COVID-19 populations such as Blacks and Latinos. We hypothesize that compared to White areas, Black and Latino areas will have fewer COVID-19 testing sites, despite having a higher percentage of positive tests. 2 Materials and methods Four primary data sources were used to complete this analysis. The addresses for testing sites were identified from the New York state government testing website which is run in collaboration with Castlight Health, a health navigation company.(New York State: Find a Test site near you, 2020) Castlight Health was selected as the source for NYC testing sites due to standardized data collection methods (e.g., consistent definition of what constitutes a testing site across New York, and uniform verification procedures for identifying testing sites). In addition, it is the sole resource provided by the New York state government for locating testing sites.(New York State: Find a Test site near you, 2020) We analyze 2020 data on two dates: May 1 to reflect the introduction of the PPP in April 2020 and May 17, which reflects the most recent data at the time of writing.(New York State: Find a Test site near you, 2020) Testing sites were geocoded across 177 zip code tabulation areas (zip codes hereafter) in NYC as the smallest unit of analysis with publicly available COVID-19 data. We then extracted demographic data from the 2018 American Community Survey 5-year estimates collected by the Census to determine the concentration of Blacks and Latinos across those zip codes.(Tiger/Line with Selected Demographic and Economic Data, 2020) Specifically, we designated areas as majority White, Black, or Latino, if 50% or more of the residents were identified as belonging to at least one group. In addition, if there was no plurality group, we indicated these areas as “no majority.” We also further segmented the “no majority” areas into majority White-Asian, White-Black, White-Hispanic, Black-Asian, Black-Hispanic and Asian-Hispanic if the combination of the two racial and ethnic groups accounted for 50% or more of the residents in a zip code. Lastly, to map COVID-19 cases by zip code, we calculated the percentage of residents that tested positive using the NYC COVID-19 city-wide data portal, which is updated on a daily basis.(Health, 2020) We used ArcGIS version 10.8.1 (ESRI)(ArcGIS Desktop, 2019) to create 1) majority race zip code-level maps showing the distribution of testing sites on May 1, 2020 and May 17, 2020 in NYC and 2) majority race zip code-level maps showing the percentage of residents identified as having COVID-19 based on the cumulative number of individuals tested on May 1, 2020 and May 17, 2020. Independent samples t-tests were used to determine whether significant differences existed in the number of testing sites by the majority race of zip codes. 3 Results Despite increased reports of higher numbers of cases and deaths among Blacks and Latinos due to COVID-19,(Yancy, 2020) clear disparities exist in the distribution of COVID-19 testing sites in NYC. On May 1, 2020 there were a total of 126 testing sites across NYC with 63 (50%) in majority White zip codes, compared to 9 (7.1%) in majority Black zip codes, and 16 (12.7%) in majority Latino zip codes (Fig. 1 ). This represented 2.5 sites per 100,000 White residents, 0.6 sites per 100,000 Black residents, and 1.1 sites per 100,000 Latino residents (Table 1 ). Moreover, of the total number of tests performed, 35.8% of residents in predominately White areas were found to be positive, while a much higher percentage of Blacks (46.8%) and Hispanics (48.8%) were found to have positive tests for COVID-19(Fig. 1). The differences in testing sites and the number of positive tests by majority race of areas were found to be statistically significant (Table 1).Fig. 1 Distribution of COVID-19 testing sites and cases identified in New York City areas by majority race/ethnicity, May 2020 Fig. 1 Table 1 COVID-19 testing sites and cases by majority race of zip code in New York City, May 2020. Table 1 White majority Black majority Asian majority Hispanic majority No majority Black vs white Hispanic vs white n (%) n (%) n (%) n (%) n (%) Test sites, 5/1 (n = 126) 63 (50%) 9 (7.1%) 1 (0.8%) 16 (12.7%) 37 (29.4%) Test sites, 5/17 (n = 180) 70 (38.9%) 31 (17.2%) 2 (1.1%) 31 (17.2%) 46 (25.6%) Number of zip code (n = 177) 63 (35.6%) 27 (13.6%) 3 (1.7%) 27 (13.6%) 57 (32.2%) M (SD) M (SD) M (SD) M (SD) M (SD) t-test t-test Average number of testing sites, 5/1 1 (1) 0.3 (0.48) 0.3 (0.58) 0.6 (0.57) 0.6 (0.79) −3.687 ⁎⁎⁎ −1.979 Average number of COVID cases, 5/1 638.7 (577.78) 1143.3 (633.97) 802.3 (337.72) 1486.7 (741.15) 937.9 (561.97) 3.552 ⁎⁎⁎ 5.847 ⁎⁎⁎ Testing sites per 100 K, 5/1 2.5 (2.94) 0.6 (0.85) 0.4 (0.71) 1.1 (1.12) 1.5 (2.06) −3.447 ⁎⁎⁎ −2.538 ⁎⁎⁎ Cases/tested, 5/1 (%) 35.8 (6.79) 46.8 (3.8) 45.8 (6.15) 48.8 (4.36) 45.2 (4.61) 7.833 ⁎⁎⁎ 9.108 ⁎⁎⁎ Average number of testing sites, 5/17 1.1 (1.15) 1.1 (1.23) 0.7 (1.15) 1.1 (1.03) 0.8 (0.95) 0.133 0.151 Average COVID cases, 5/17 703.4 (645.01) 1257.7 (695.61) 910.7 (403.11) 1662.3 (828.2) 1032.5 (622.67) 2.583 ⁎ 4.409 ⁎⁎⁎ Testing sites per 100 K, 5/17 2.8 (3.28) 2.2 (2.38) 0.8 (1.43) 2.1 (2.22) 1.8 (2.2) −0.995 −1.13 Cases/tested, 5/17 (%) 27.1 (6.18) 38 (4.53) 36.9 (5.37) 39.2 (4.1) 36.8 (4.53) 9.289 ⁎⁎⁎ 9.329 ⁎⁎⁎ Note. M = mean, SD = standard deviation. ⁎ p < 0.05. ⁎⁎ p < 0.01. ⁎⁎⁎ p < 0.001. By May 17, 2020, testing sites in Black areas increased more than 240% from 9 to 31, and more than 90% from 16 to 31 in Latino areas (Fig. 1). There were no significant differences found in the average number of test sites across zip codes categorized by majority race. The number of testing sites increased to 2.2 per 100,000 Blacks, and to 2.1 per 100,000 Latinos on May 17. The percentage of residents testing positive relative to the number of tests administered also decreased to 38 per 100,0000 Blacks, and 39.2 per 100,000 for Latinos (Table 1). Nonetheless, more testing sites (70 (38.9%) of the 180 sites) in NYC were located in predominately White areas (Table 1) compared to any other subgroup. Moreover, the number of test sites (2.8 per 100,000 residents) was highest in White areas and the test positivity rate (27.1%) was more than 10% lower than the test positivity rates in Black (38%) and Latino (39.2%)areas. When we examined geographic unitsge using combinations of two racial and ethnic groups, we found the greatest disparity for the distribution of testing sites with combinations including Black race (Fig. 1). Majority Black-Asian (n = 0), Black-Hispanic (n = 9), and White-Black (n = 4) areas had far fewer testing sites on May 17, 2020, compared to White-Asian (n = 12), White-Hispanic (n = 9) and Asian-Hispanic areas (n = 12) (Fig. 1). 4 Discussion In this paper, we mapped the distribution of testing sites in NYC in majority Black, White, and Hispanic areas. Initially, we found a significant disparity in the availability of testing sites for Black and Hispanic neighborhoods compared to White areas (Table 1). This disparity is consistent with the literature showing the marginalization of majority Black areas due to economic disenfranchisement and limited health-promoting attributes.(Poteat et al., 2020; Williams and Collins, 2001) Notably, efforts to increase testing availability between May 1, 2020 and May 17, 2020 (after the passage of the PPP) to decrease the disparity in the placement of testing sites seem to have made an impact. We did not find significant differences in the availability of testing sites between White and Black or Hispanic areas on May 17, 2020 (Table 1). New York City and state government should continue to make efforts to target and increase the placement of COVID-19 testing sites in Black and Latino areas, since this is likely to mitigate COVID-19 fatalities among these vulnerable groups. 5 Conclusions Our study represents one of the first studies to map the geographic distribution of COVID-19 testing sites for predominately Black and Latino areas in NYC. While this analysis could not be performed at a smaller geographic unit (such as a census tract) due to the lack of COVID-19 data at smaller units of geography, our study is one of the first to map the distribution of COVID-19 testing sites by race in NYC. However, there is some limited generalizability of our findings, due to both the unique size and racial/ethnic make-up of a large city such as NYC. As more testing sites are being continually added, this research suggests further monitoring and exploration of race, testing access, and COVID-19 issues in NYC, and comparable cities impacted by the pandemic. Credit author statement DGT, AJ and JCS conceived the study. DGT was responsible for project administration, including overseeing the data analysis and data curation for the paper, as well as writing the original draft of the paper. JCS was responsible for data analysis. All authors reviewed, commented and edited later drafts of the paper, and approved the final version. CRediT authorship contribution statement Professor Grigsby-Toussaint: Conceptualization, data curation, methodology, analysis, project administration and writing, review and editing. Professor Antwan Jones: Conceptualization, methodology, writing, review and editing. Dr. Jong Cheol Shin: Conceptualization, methodology, analysis, writing, review, and editing. Declaration of competing interest None. ==== Refs References ESRI ArcGIS Desktop Release 10.7.1 2019 Environmental Systems Research Institute Redlands, CA Farmer B. The coronavirus doesn’t discriminate, but U.S. health care showing familiar biases April 2, 2020 NPR https://www.npr.org/sections/health-shots/0220/04/02/825730141/the-coronavirsu-doesnt-discriminate-but-u-s-healthcare-showing-familiar-biases Accessed April 2020 Gorbalenya A.E. Baker S.C. Baric R.S. de Groot R.J. Drosten C. Gulyaeva A.A. The species severe acute respiratory syndrome-related coronavirus: classifying 2019-nCOV and naming it SARS-CoV-2 Nat. Microbiol. 5 4 2020 536 544 32123347 NYC Health COVID-19 Data https://www1.nyc.gov/site/doh/covid/covid-19-data.page 2020 accessed May 2020 Johns Hopkins Coronavirus Resource Center https://coronavirus.jhu.edu/ 2020 Accessed May 2020 New York State Find a Test site near you https://coronavirus.health.ny.gov/find-test-site-near-you 2020 Accessed May 2020 Poteat T. Millett G.A. Nelson L. Beyrer C. Understanding COVID-19 risks and vulnerabilities among black communities in America: the lethal force of syndemics Ann. Epidemiol. 47 2020 1 3 32419765 Probasaco J. The Paycheck Protection Program and Health Care Enhancement Act for small businesses, hospitals, and health care workers https://www.investopedia.com/paycheck-protection-program-and-health-care-enhancement-act-4843094 2020 Accessed April 2020 Tiger/Line with Selected Demographic and Economic Data U.S Census Bureau https://www.census.gov/geographies/mapping-files/time-series/geo/tiger-data.html 2020 Accessed May 2020 Williams D.R. Collins C. Racial residential segregation: a fundamental cause of racial disparities in health Public Health Rep. 116 5 2001 404 416 12042604 Yancy C.W. COVID-19 and African Americans JAMA. 2020 10.1001/jama.2020.6548

33647352

PMC9753026

NO-CC CODE

2022-12-16 23:25:27

no

Prev Med. 2021 Jun 26; 147:106463

utf-8

Prev Med

10.1016/j.ypmed.2021.106463

oa_other

==== Front Appl Entomol Zool Appl Entomol Zool Applied Entomology and Zoology 0003-6862 1347-605X Springer Japan Tokyo 810 10.1007/s13355-022-00810-9 Review Cockroaches: a potential source of novel bioactive molecule(s) for the benefit of human health Siddiqui Ruqaiyyah 13 Elmashak Yara 1 http://orcid.org/0000-0001-7667-8553 Khan Naveed Ahmed naveed5438@gmail.com 23 1 grid.411365.4 0000 0001 2218 0143 College of Arts and Sciences, American University of Sharjah, University City, 26666 Sharjah, United Arab Emirates 2 grid.412789.1 0000 0004 4686 5317 Department of Clinical Sciences, College of Medicine, University of Sharjah, University City, Sharjah, United Arab Emirates 3 grid.508740.e 0000 0004 5936 1556 Department of Medical Biology, Faculty of Medicine, Istinye University, 34010 Istanbul, Turkey 15 12 2022 111 20 9 2022 30 11 2022 © The Author(s), under exclusive licence to The Japanese Society of Applied Entomology and Zoology 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Cockroaches are one of the hardiest insects that have survived on this planet for millions of years. They thrive in unhygienic environments, are able to survive without food for up to 30 days, without air for around 45 min and being submerged under water for 30 min. Cockroaches are omnivorous and feed on a variety of foods, including cellulose and plastic, to name a few. It is intriguing that cockroaches are able to endure and flourish under conditions that are harmful to Homo sapiens. Given the importance of the gut microbiome on its’ host physiology, we postulate that the cockroach gut microbiome and/or its metabolites, may be contributing to their “hardiness”, which should be utilized for the discovery of biologically active molecules for the benefit of human health. Herein, we discuss the biology, diet/habitat of cockroaches, composition of gut microbiome, cellular senescence, and resistance to infectious diseases and cancer. Furthermore, current knowledge of the genome and epigenome of these remarkable species is considered. Being one of the most successful and diverse insects, as well as their extensive use in traditional and Chinese medicine, the lysates/extracts and gut microbial metabolites of cockroaches may offer a worthy resource for novel bioactive molecule(s) of therapeutic potential for the benefit of human health and may be potentially used as probiotics. Keywords Cockroaches Gut microbiome Microbial metabolites Probiotics ==== Body pmcIntroduction Cockroaches are a fascinating and ancient species, and are hemimetabolous insects of the class Insecta, order Blattodea, which also includes termites, with their ancestors originating from the Carboniferous period, emerging approximately 300–350 million years ago (Tinker and Ottesen 2021; Wang et al. 2017; Zhao et al. 2017). To date, around 4700 species of cockroaches are known; however, it is speculated that at least twice this number is yet to be ascertained (Beccaloni 2014). Cockroaches partake in an important role in terrestrial ecosystems, through breakdown of organic materials and release of nutrients from recycling dead plants, dead animals, and it has been suggested that extinct cockroaches (Blattulidae) were probably involved in the clean-up of dinosaur excrements (Vršanský et al. 2013). On the contrary, Homo sapiens are merely one species in the midst of millions of others and are a relatively new addition to Earth (Harari 2014). In comparison, cockroaches have been able to adapt, evolve and survive successfully over millions of years, indicating that we should learn from these species. Of note, cockroaches have been utilized in traditional Chinese medicine. For example, extracts of the American cockroach (Periplaneta americana) which is usually considered a pest, has been utilised in traditional Chinese medicine (Lu et al. 2021; Xin et al. 2015). These extracts have been used to treat aches and pains, inflammation and even chronic heart failure for hundreds of years (Ma et al. 2018; Xin et al. 2015). Ethanol extract of P. americana, known as “Kangfuxin”, has been used to treat skin and mucosal injuries since the 1980s and has the approval of the Chinese Food and Drug Administration (Wang et al. 2020). Cockroaches are one of the “hardiest” insects and are able to survive without food for up to a month, without air for around 45 min and being submerged under water for 30 min (Lee et al. 2012; Wharton and Wharton 1959). Moreover, they can endure high doses of radiation: 15 times higher than humans (Lee et al. 2011; Wharton and Wharton 1959). Accordingly, it is plausible to suggest that cockroaches represent a rich and diverse source of novel molecules with biological activity originating either in their tissues or their gut microbiome contributing to their overall health and resilience (Akbar et al. 2018; Ali et al. 2017; Moshaeb et al. 2018). Several studies indicate the significance of the gut microbiome and its role in the overall health and immunity of the host (Malard et al. 2021; Siddiqui et al. 2021a, b, c; Zheng et al. 2020). Cockroaches represent an ideal organism for studying gut microbiome composition as well as host microbial interactions, as they have evolved to host a multifaceted gut microbiome comprising several species of microbes (Tinker and Otteson 2021). Here, we describe the cockroach biology, diet and habitat, gut microbiome composition, in relation to cellular senescence, the genome and epigenome, as well as resistance of cockroaches to infectious diseases and cancer. Preliminary studies investigating the anti-cancer and anti-bacterial effects of cockroach gut bacterial metabolites and the molecules elucidated are also discussed (Fig. 1). Due to their hardiness, diversity, and successful survival throughout history, we suggest that roaches need to be studied as they are an excellent resource that we can use to extract novel bioactive molecules that can act as antimicrobials, antibacterial, or anticancer drugs for the benefit of human health.Fig. 1 Novel molecules from immune/gut microbiota of cockroaches Biology and classification Life cycle of cockroaches may contribute to maximizing cockroaches’ reproductive fitness by reproducing a high number of eggs in order to compensate for constantly living under a high risk of death by external factors. After reaching sexual maturity, a female cockroach can reproduce an egg capsule every 9 days which contains around 14 embryos (Perrott and Miller 2010). After the juvenile cockroaches complete their molting stages, they will reach sexual maturity in 6 to 12 months and have a lifespan of up to a year and half (Perrott and Miller 2010) (Fig. 2). However, there are slight differences among different species. For instance, differences can be observed between the German and the Brown-Banded cockroaches. While the female German cockroaches are known to carry their egg cases until hatching, the female Brown-Banded cockroaches attach their egg cases to discrete locations (Malinoski 1999). The developmental period between egg to adulthood is around 95–276 days for the Brown-Banded cockroaches but only 55–68 days for the German cockroaches (Malinoski 1999). Some cockroaches can undergo parthenogenesis to reproduce and others can store sperm to reproduce at a later stage (Guzman et al. 2020). Cockroaches have incomplete metamorphosis, which refers to the nymphs being generally similar to adults, but with underdeveloped genitalia and wings. Some cockroaches have been able to survive in the laboratory for up to 4 years (Daly et al. 1978).Fig. 2 The life cycle of a cockroach through the three stages including (i) egg stage, (ii) nymph or juvenile cockroach, and (iii) adult cockroach In the phylogenetic tree, Blattodea is divided into Corydioidea, Blaberoidea, and Blattoidea. Corydioidea includes sister groups called Nocticolidae and Corydiidae (Wang et al. 2017). Ectobius and Blaberidae are the sister groups of Blaberoidea (Evangelista et al. 2019; Wang et al. 2017). Blattoidea, a family of cockroaches and termites, includes Blattidae, Tryonicidae, Lamproblattidae, Anaplectidae, Cryptocercidae and Isoptera (Wang et al. 2017). Irrespective of the number of differences among the species, including behavioural and phenotypical differences, cockroaches are considered as monophyletic due to the results gathered by DNA sequencing of mitochondrial rRNA genes (Kambhampati 1995; Li et al. 2018). Recently used classifications are Princis, McKittrick, and Grandcolas (Roth 2003). The Princis classification is mostly based on differences in external structures of the cockroaches which gives rise to “4 suborders, 28 families, and 21 subfamilies” (Roth 2003) while McKittrick classification, the most acknowledged and accepted type, is based on the morphological characters of the male and female genitalia as well as their musculature, proventriculus, and oviposition behaviour (Kambhampati 1995; Roth 2003). The Grandcolas classification system “redescribed 6 families and sub-families” (Roth 2003) after analysing 221 genera and these classifications mostly coincide with McKittrick’s system but with a few exceptions. The genome and epigenome of cockroaches As previously mentioned, cockroaches are known to be able to survive in tough environments, variable climates, and have flexible feeding habits. These properties of cockroaches may have arisen from the epigenetic processes which allow them to regulate gene expression while maintaining the same DNA sequences without modifications and these processes may include DNA methylation as well as histone modifications (Qi et al. 2019; Villalba et al. 2021). Therefore, the outcome of epigenetic mechanisms may be key factors that contribute towards the cockroaches’ ability to adapt to several climates and thermal stress (Villalba et al. 2021). In addition, studies on several species showed that epigenetics may also contribute to the generation of new heritable variations in the phenotypic traits of the organism (Bird 2007; Villalba et al. 2021). Another interesting hallmark about cockroaches is that their genome ranks as the second largest of all insect species, and is sized at about 3.38 Gb (Li et al. 2018), while containing a relatively lower number of species-specific genes, but the highest number of multicopy universal genes, whereby studies revealed a number of 13,555 genes in P. Americana in comparison to 7708 genes which was the highest number of universal genes in other insects (Li et al. 2018). The cockroaches’ large genome size is also a result of containing repetitive elements which constitute 60% of the roaches’ entire genome (Li et al. 2018). Additionally, the genome of P. Americana shares 90% of their genome with the rest of the Blattodea family and, therefore, contains some similar features to other insect species such as the exon number per gene; however, it differs from other insects when it comes to the length of introns which, too, is the second largest of all insect species numbered 3 Kb (Li et al. 2018). Remarkably, the American cockroaches’ genome shows around 80% homology with termites, Z. nevadensis and M. natalensis, which is more than the 75% homology with German cockroaches and this suggests that the American cockroaches are more related to termites than German cockroaches representing an evolutionarily significant finding (Li et al. 2018). Moreover, the 80% sequence identity between American cockroaches and termites are found in 29 pathophysiological pathways such as development and immunity while the 75% sequence identity with the German cockroaches are found in only 6 pathways including signal transduction (Li et al. 2018). Furthermore, in order to be able to survive and thrive in environments inhabited by humans, cockroaches contain a very distinctive genome which includes numerous genes regulating detoxification of xenobiotics through the function of P450s as well as the encoding of 154 olfactory receptors and 522 gustatory receptors (GRs) which far surpasses that of any other insect where 329 of these GRs are bitter receptors which benefit the roaches by allowing them to tolerate bitter food found and expanding the diversity of their diets (Li et al. 2018). Diet and habitat Cockroaches are omnivorous and generally feed on a variety of foods (McPherson et al. 2021; Schal et al. 1984; Lauprasert 2006). Notably, female cockroaches require more protein than males due to their high investment in egg development, while males preferred food with a high carbohydrate content (Lauprasert 2006). Cockroaches are also classified as omnivores due to having a varied diet which include any type of food eaten by humans or animals as well as their wastes such as fecal matter (Roth and Willis 1957), hair, hygienic products found in bathrooms, or even glue (Potter 2010). They also exhibit cannibalism during starvation (Guzman et al. 2020). As roaches eat items that are disposed of by humans, cockroaches have the ability to feed on certain types of plastic including plastic insulation and leather (Mogbo et al. 2013). These plastics vary in thickness but are still torn apart by the roaches’ strong mandibles (Weihmann et al. 2015). Cockroaches are able to digest cellulose and can utilise a diet of crystallisine cellulose (Slaytor 1992). The ability of cockroaches to thrive in unhygienic environments possibly carrying infectious agents is one of the key factors that contribute to the cockroaches’ importance in the medical community (Akbar et al. 2019; Li et al. 2018; Koehler et al. 1999; Malinoski 1999; Roth and Willis 1957). Their habitats vary greatly as they can survive in forest canopy beds, arid land, water surfaces, etc. (Bell et al. 2007). With human interactions, cockroaches are mostly found in kitchens as a result of the presence of grease and also in bathrooms due to the readily available source of water, cracks in walls, large openings such as vents or pipes, as well as sewers (Koehler et al. 1999; Malinoski 1999). Cockroaches can also adapt and survive in a wide range of climates including tropical as well as polar temperatures by using glycerol as an anti-freeze and this results in different sizes of cockroaches (Channel, 2007). The differences in adaptations to temperatures also means that cockroaches have different tolerances to humidity. For example, while German, Brown-Banded, American, and Oriental cockroaches all prefer warm temperatures, only Brown-Banded and American cockroaches can be found in dry habitats, whereas German and Oriental cockroaches prefer moist environments (Malinoski 1999). Gut bacteria and its role Cockroaches are routinely associated with microbes in the environments in which they reside (Ali et al. 2017). They are associated with bacteria which are primary symbionts residing within specialized cells in the fat body and are involved in nitrogen metabolism, as well as facultative secondary symbionts that circulate in the gut (Domínguez-Santos et al. 2021; Guzman and Vilcinskas 2020). Gut microbiome of cockroaches has a major role in digestion, metabolism, absorption, immunity against pathogens, growth, behavior, etc. (Chen et al. 2020, Jandhyala et al. 2015; Tinker and Ottesen 2016). A number of studies have demonstrated the importance of gut microbiome and its role in the overall health and immunity of the host (Malard et al. 2021; Siddiqui et al. 2021a, b, c; Sieksmeyer 2021; Zheng et al. 2020). Cockroaches represent an ideal organism for studying gut microbiome composition as well as host microbial interactions, as they have evolved to host a multifaceted gut microbiome comprising hundreds of unique species of microbes, similar to a variety of omnivorous animals including humans and mice (Tinker and Otteson 2021). In a recent study, a weak but significant phylosymbiotic signature was detected, suggesting that cockroach phylogeny may have a role in structuring the gut microbiome over shorter evolutionary distances and possibly extended periods of evolutionary time (Tinker and Otteson 2020). The developmental stages in lifecycle of cockroaches contribute to differences between species of microorganisms found in the gut microbiome of cockroaches (Chen et al. 2020). Nevertheless, there are certain bacteria that dominate throughout all three developmental stages of roaches indicating that some microorganisms found in the gut microbiome are inherited from mother to eggs which include Bacteroidetes, Firmicutes and Proteobacteria as these bacteria aid in the process of digestion, absorption and defence against pathogens (Chen et al. 2020; Tinker and Ottesen 2016). However, the diversity of the gut microbiome in eggs remain lower than that found in nymph or adult stages where major difference is the presence of the Blattabacterium and Lactobacillus which are also inherited from the mother (Chen et al. 2020). These dominant bacteria of the eggs are specifically beneficial since Blattabacterium has a significant role in nitrogen cycling as well as absorption and synthesis of important nutrients while Lactobacillus acts as a probiotic and protects the eggs from pathogens (Chen et al. 2020). On the other hand, Desulfovibrio and Parabacteroides are the dominant microorganisms in the nymph stages and Bacteroides, Dysgonomonas, Porphyromonadaceae and Alistipes dominate the adult stage of the cockroaches (Chen et al. 2020). The gut microbiome of cockroaches does not only depend on which developmental stage they are moving through as it also depends on their diet. To begin with, it is important to understand that the gut microbiome of cockroaches differs from other insects and humans. This is mainly due to the diet habits of the species whereby cockroaches, which are considered as a typically gregarious or social species, feed on a big assortment of foods and have a varied diet while other insects, such as bees, who feed on a smaller array of food and are said to have more specialized diets (Tinker and Ottesen 2016). The varied diet of cockroaches enables this species to contain a highly diverse and complex gut microbiome which is comprised of a great number of microorganisms. This also means that roaches have evolved methods to maintain a stable gut microbiome regardless of any changes that may occur to their diet which is advantageous in their daily lives (Chen et al. 2020). Some of these advantages contribute to their ability to survive in various habitats as well as adapt to unfavourable conditions and as a result P. americana is observed to be found across continents (Tinker and Ottesen 2016). In addition, there is an obvious distinction in the abundance of the microorganisms depending on the division of the gut. For instance, the hindgut of the cockroaches contains the biggest and most diverse portion of the gut microbiome (Jahnes et al. 2021; Tinker and Ottesen, 2016). Therefore, the hindgut microbiome plays a significant role in the digestion of the materials that pass through the foregut and midgut as well as the promotion of social behaviour through the production of pheromones (Chen et al. 2020; Tinker and Ottesen 2016). A recent study was conducted to compare gut microbial diversity between laboratory P. americana as well wild-caught populations of P. Americana and Periplaneta fuliginosa, prior to, during, and following a 2-week period in the laboratory environment (Tinker and Ottesen 2021). The data revealed that gut microbial changes were apparent, based on the species and environment and laboratory-based and wild-captured cockroaches from the same species depicted distinct gut microbiomes. Interestingly, being based in a laboratory environment led to decreased microbiome diversity for both species of wild-caught insects, suggesting that cockroaches could be used as a model to study changes in gut bacterial diversity as a result of various changes (Tinker and Otteson 2021). In another report, the hindgut microbiome of Blattella germanica cockroaches was analyzed for gut bacteria, fungi, archaea and viruses (Domínguez-Santos et al. 2021). In agreement with prior studies, the most abundant core genera were Bacteroides (Bacteroidetes), Desulfovibrio (Proteobacteria), Fusobacterium (Fusobacteria) and Clostridium (Firmicutes) and are known to participate in protein and polysaccharide digestion, protection versus pathogens and nitrogen fixation (Domínguez-Santos et al. 2021). In addition, 70 families of archaea were also detected, indicating a potential role of these microorganisms on cockroach physiology. The most abundant species in adults and nymphs were from the families: Methanobacteriaceae, Methanosarcinaceae, and Methanomassiliicoccaceae, which are methanogenic archaea that may be involved the hindgut nitrogen–carbon balance by nitrogen fixation (Domínguez-Santos et al. 2021). German cockroach, Blatella germanica, is thought to be a vector of several enteric bacterial pathogens, including E. coli, among livestock and humans (Ray et al. 2020). In this study, B. germanica were orally infected with E. coli. The results revealed that E. coli is mostly cleared within 48 h, whereas one strain may persist in a majority of cockroaches for longer than 3 days with limited impact on cockroach longevity. The study also revealed that some strains of E. coli were greater in cockroach nymphs than adults. Interestingly, clearance of E. coli was significantly reduced in gnotobiotic cockroaches that were reared in the absence of environmental bacteria. This suggested a possible protective role for the microbiome versus bacterial pathogens (Ray et al. 2020). Anti-microbial activity Since roaches live in unhygienic and insanitary environments and niches, cockroaches have adapted, as a result of external stimuli, in ways by which they can protect themselves from exposure to contaminants or from microbial infections (Akbar et al. 2019; Ali et al. 2017; Latifi et al. 2015). This means that cockroaches may contain defence mechanisms and their “gut microbiota produce molecules to thwart invading pathogens” (Ali et al. 2017). Their ability to do this comes from the lectin proteins which identify the foreign or harmful bacteria and stimulate the innate immunity response against pathogens (Latifi et al. 2015). Another significant factor that may contribute to the cockroach immunity is the complex passageway found in their cavity which consists of antimicrobials that destroy the pathogens before reaching the haemocoel (Balasubramanian et al. 2017). Therefore, the cockroaches’ ability to be resistant to superbugs and other pathogens, due to the presence of lectin and antimicrobials in their cavities, indicates that their anti-bacterial properties need to be studied further in hopes of a new medical breakthrough such as the discovery of new antibiotics that could be useful for humanity. It is not an uncommon practice for insects including cockroaches to be used therapeutically against some diseases such as malaria as well as asthma (Balasubramanian et al. 2017) since there are currently 50 anti-bacterial molecules being used in the medical field today that have been extracted from insects (Latifi et al. 2015). Furthermore, cockroaches have been observed to have anti-bacterial activities against Gram- positive and -negative bacteria as well as anti-amoebic properties (Akbar et al. 2018). In a study, some of the bacteria found in the gut microbiome of cockroaches were isolated which included Serratia marcescens, Escherichia coli, Klebsiella sp., Bacillus sp., and Streptococcus sp. This allowed for the development of conditioned media comprising the gut bacterial metabolites and the antibacterial activities of these metabolites were determined of various against pathogenic bacteria (Akbar et al. 2018). The gut microbial metabolites converse with the immune system and modulate immune responses, and play a profound role in cellular signaling, inflammation and interaction with the immune cells (Belkaid and Hand 2014; Kau et al. 2011). These gut bacterial metabolites were found to have bactericidal and inhibitory effects on methicillin-resistant Staphylococcus aureus (MRSA), Streptococcus pyogenes, Bacillus cereus, E. coli K1, P. aeruginosa, K. pneumoniae, Salmonella enterica and S. marcescens (Akbar et al. 2018). However, the extraordinary results did not stop there but continued when the conditioned media also indicated 40–60% amoebicidal effects when tested against the free-living amoebae A. castellanii (Akbar et al. 2018). In addition, other studies have also shown that roaches have antimicrobial properties against methicillin-susceptible Staphylococcus aureus (MSSA) (Billah et al. 2015), M. luteus (Basseri et al. 2016) Staphylococcus aureus and Bacillus subtilis (Mahboub et al. 2021). Another study conducted in Saudi Arabia on Blattella vaga extracted Bacillus licheniformis, Bacillus subtilis and Kocuria rosea from the cockroach’s gut microbiome and evaluated their anti-microbial properties against Salmonella enterica, MRSA, Streptococcus mutans and Candida albicans which are all considered as drug-resistant pathogens (Alkhalifah 2021). Their analysis showed that Bacillus subtilis does not demonstrate any antimicrobial activity while Bacillus licheniformis demonstrated inhibitory effects against Candida albicans and Kocuria rosea also showed antimicrobial properties against MRSA and Streptococcus mutans (Alkhalifah 2021). These studies and their results indicate that gut bacterial metabolites of cockroaches may be developed into effective antibiotics or probiotics to be utilized as a treatment or prevention against drug-resistant pathogens, which may solve a major public health crisis that we are facing today. Nevertheless, the gut bacteria of cockroaches are not the only factor contributing to their antibacterial properties as cockroach brain extracts as well as hemolymph have demonstrated significant bactericidal activities against some pathogenic bacteria (Ali et al. 2017). Aspirated hemolymph demonstrated 35% antibacterial activity against MRSA and 20% against E. Coli K1 while brain lysates showed 90% antibacterial activity when tested on MRSA as well as E. Coli K1 (Ali et al. 2017). While the results clearly indicate that cockroach brain lysates are superior to hemolymph as antibacterial compounds, hemolymph have also exhibited potent anti-viral and anti-tumour properties, antimicrobial activities against parasitic worm embryos, aided in the treatment of several conditions and diseases such as diabetes (Ali et al. 2017). For instance, a study isolated the hemolymph of cockroaches and tested it against P. aeruginosa, P. mirabilis, S. aureus, E. coli and Salmonella typhi using a zone of inhibition test which exhibited antimicrobial properties against all 5 bacterial strains but with different degrees (Balasubramanian et al. 2017). At the same time, brain lysates have too exhibited positive outcomes in the treatment of viruses and tumours while also acting as anti-diabetic and anti-inflammatory compounds (Ali et al. 2017). The remarkable therapeutic effects of the cockroach hemolymph and brain lysates arise from the structure of their functional groups as well as the presence of other active compounds such as flavanones (Ali et al. 2017). Cockroach anti-viral properties are significant as they act as their natural defense mechanism since roaches can be infected with viruses, which may cause behavioural changes, just as much as being biological vectors (Sieksmeyer 2021). Studies have shown that cockroaches’ anti-viral activity and antimicrobial peptides are effective against herpes simplex virus (Ali et al. 2017; Wang et al. 2011). Other studies have also demonstrated that cockroaches contain compounds found in their brain and hemolymph lysates which demonstrate anti-viral activity against Influenza A virus, Parainfluenza virus, HIV-1, Norovirus, retroviruses and several other types of viruses (Ali et al. 2017). Some studies have attributed the cockroaches’ anti-viral properties as well as other anti-microbial activities to 1, 2, 4-triazole compounds found in their hemolymph (Balasubramanian et al. 2017). This is understandable since 1, 2, 4-triazole compounds are found and used in several commercial drugs found in the market today including anti-viral and anti-fungal agents such as ribavirin and fluconazole, respectively (Balasubramanian et al. 2017). Other than hemolymph, gut and brain extracts, antimicrobial properties in cockroaches can also originate from chitosan which is a polysaccharide as well as an antimicrobial agent found in the roaches’ exoskeleton (Balasubramanian et al. 2017; Mahboub et al. 2021). Interestingly, the activities and effects of chitosan found in the cockroaches depend upon the roaches’ species such that the antimicrobial properties associated with chitosan differs between the American cockroach and the German cockroach (Basseri et al. 2019). Chitosan can act as an antibacterial agent against Gram-positive bacteria, such as S. aureus and B. subtilis, where the minimum inhibitory concentration (MIC) was found to be 2000 μg/ml (Mahboub et al. 2021). Chitosan is also effective against Gram-negative bacteria where MIC of chitosan against E. coli was recorded to be 1000 μg/ml and MIC against Salmonella typhimurim was 2000 μg/ml (Mahboub et al. 2021). Yet, studies have shown different degrees of antibacterial activities in roaches and this can be a result of differences in nutrition and severity of the infections as well as other factors (Latifi et al. 2015). Irrespective of its extensive antimicrobial properties, chitosan also exhibits anti-fungal activities against A. flavus and A. albican by inhibiting their growth (Basseri et al. 2019), but studies conducted by Mahboub et al. 2021 found that chitosan did not exhibit any anti-fungal activity against Candida albicans which agreed with research by Basseri et al. (2019) but was contradictory to other studies which revealed anti-fungal capabilities (Alburquenque et al. 2010). However, cockroaches contain another peptide, Periplanetasin-2, which demonstrates non-hemolytic anti-fungal properties (Yun et al. 2017). When Periplanetasin-2 was extracted and tested against mitochondria of Candida albicans, the results showed the stimulation of oxidative stress and lipid peroxidation resulted in inducing apoptosis that was brought about by the externalization of phosphatidylserine, DNA fragmentation, membrane depolarization and an increase in the calcium level and mitochondrial glutathione, but a decrease in cytosolic glutathione (Yun et al. 2017). These results of the antimicrobial activity of cockroaches functioning as broad-spectrum antimicrobial peptides (Li et al. 2018) are all well and good but useless if they have potential harmful effects on humans and cannot be used. Hence, the Roche cytotoxicity detection kit was used to study the effects on human cells when infected with MRSA or E. coli K1 which resulted in 70% cytotoxicity (Ali et al. 2017). On the other hand, the detection kit demonstrated a minimal amount of cytotoxicity and cell damage when the infected human cells were treated with cockroach lysates (Ali et al. 2017). Therefore, this revealed that cockroach lysates contain antibacterial molecules, less than 10 kDa in molecular mass (Ali et al. 2017), that can treat multi-drug resistant infections and are safe to be used on human cells. Anti-cancer activity Cancer is established as one of the leading causes of mortality and morbidity and was ranked in 2020 as one of the top 3 causes of death, worldwide along with heart disease and COVID-19 (Ahmad et al. 2021). Western medicine has relied on chemotherapy, radiotherapy, and stem-cell therapy as well as other similar options for cancer treatment (Soopramanien et al. 2019; Zhao et al. 2017). However, due to drug resistance and several complications as well as harmful effects of the currently routinely used types of cancer treatment (Zhao et al. 2017), the medical community has been trying to innovate in hopes of improving healthcare quality and better patient care by discovering and developing novel anti-cancer agents and treatments without harming or compromising or suppressing the patient’s immune system (Mahboub et al. 2021). Therefore, in some parts of the world, cancer patients have resorted to Traditional Chinese Medicine (TCM) instead of Western medicine due to its lower cytotoxicity (Zhao et al. 2017). In particular, cancer patients benefit from TCM which uses plant derivatives along with insect secretions, including cockroaches’, which have proven to have anticancer properties (Seabrooks and Hu 2017; Wang et al. 2011). Aside from their anticancer properties, cockroach extracts also contain wound-healing activities and are also used in TCM (Zhu et al. 2018). The effects of these extracts vary from treating blood stasis to burns, tissue repair, or wounds from the first day of treatment (Li et al. 2018; Zhu et al. 2018). A plausible explanation to the cockroach derivatives ability as inhibitors of tumour progression, maybe due to the roaches’ ability to withstand radiation as well as their living conditions in tough environments filled with pollutants such as chemicals, traces of heavy metals and infectious microorganisms (Soopramanien et al. 2021; Wang et al. 2011). Thus, traditional medicine introduced dried worms and adult cockroaches as potential treatments for several diseases due to their pharmacological effects such as blood pressure stabilization, detoxification, immunity enhancement and promotion of diuresis when necessary (Zhao et al. 2017). For example, cockroaches contain antimicrobial peptides that can be used in treating and stimulating liver recovery after hepatitis B infection (Zhao et al. 2017) and also in the treatment of Newcastle disease (Wang et al. 2011). Roaches are also involved in traditional medicine for numerous other health conditions which include heart diseases, asthma, digestive conditions, ulcers, burns and most importantly cancer (Seabrooks and Hu 2017; Wang et al. 2011) as a by-product of the interplay and effect of several components such as unsaturated fatty acids, ester, cyclic peptides, human essential and semi-essential amino acids, pheromones, polysaccharides and, once again, chitosan (Zhao et al. 2017). And since drug resistance is one of the major threats on public health including cancer treatments, it is important to note that one of the anti-cancer properties of cockroaches relates to their extracts targeting of multidrug resistance proteins and breast cancer resistance proteins which reverse the effects of drug resistance on cells and improve the possibility of successful cancer treatments (Zhao et al. 2017). As mentioned above, cockroach brain and hemolymph have many therapeutic effects including anti-tumour activity against various forms of cancers such as ovarian, breast, lung and prostate cancers (Ali et al. 2017). In addition, chitosan, which is responsible for the roaches’ antibacterial properties, is similarly responsible for resulting in the cockroaches containing anticancer agents due to its role of acetylation and its molecular weight (Mahboub et al. 2021). Studies on this polysaccharide have shown to be an effective natural treatment for hepatoblastoma and breast cancer (Mahboub et al. 2021; Seabrooks and Hu 2017). Furthermore, there are several other researchers that have carried out studies on the anticancer properties of chitosan. For instance, it was shown that chitosan derived from shrimp can be an effective treatment for human bladder cell carcinoma (Younes et al. 2014). In another study, positive results against hepatocellular carcinoma cells using chitosan extracted from cockroaches were observed (Azuma et al. 2014). Furthermore, it was noted that chitosan is also effective against laryngeal cancer and human embryo rhabdomyosarcoma cells (Ganesan et al. 2020). Research has also revealed that chitosan found in cockroaches may be used against chronic myeloid leukemia by inducing G2/M phase arrest as well as acting as an EGFR-inhibiting agent (Seabrooks and Hu 2017). Nevertheless, chitosan is not the only effective anticancer agent found in cockroaches since their gut microbial metabolites may also contribute to having anti-cancer molecules as a result of their need to protect themselves from diseases by the microorganism rich environments that they dwell in (Soopramanien et al. 2021). For instance, Staphylococcus xylosus extracted from the cockroach gut microbiome exhibited anti-cancer effects by reducing the growth of a human prostate cancer cell line (PC-3) (Soopramanien et al. 2019). Additionally, as with the action of chitosan on myeloid leukemia, cockroaches have the ability to prompt cell cycle arrest (Zhao et al. 2017). Some examples of the effects of cockroach extracts include the reduced growth of Lewis lung carcinoma due to blocking the G0/G1 phase of the cell cycle as well as reduced growth of endometrial cancer cells due to overexpression of p53 and reduced expression of C-erbB-2 as a result of cell cycle arrest. Cockroach extracts have been observed to also induce apoptosis in cancer cells such as in human hepatoma cells as well as in leukaemia cells which were brought about by the action of perplanetasin-5 (Kim et al. 2021; Zhao et al. 2017). Moreover, polypeptide extracts from cockroaches can reduce the tumour microvessel density and the expression of vascular endothelial growth factor and, therefore, have anti-angiogenic effects and impede tumour growth (Zhao et al. 2017). Therefore, it is clear that cockroaches may provide humanity with novel pharmacological drugs, such as anti-tumour agents, which would significantly help the medical community, but more research is still required for to develop a better understanding of the exact modes of action of the cockroaches’ anti-cancer compounds and the appropriate isolation of their chemical constituents to result in successful tumour suppression in humans (Soopramanien et al. 2019; Zhao et al. 2017). Cellular senescence in cockroaches While ageing is a continuous pathophysiological process, the mechanisms underlying this phenomenon are complex and vary between species. Cockroaches are typically a gregarious and social species and so may be useful model organisms to investigate the evolution of cellular senescence (Kramer et al. 2021; Tinker and Otteson 2020). For instance, one of these mechanisms is the accumulation of cells that have undergone cellular senescence in body tissues (Collado et al. 2007) which refers to a permanent but stable arrest in the cell cycle due to the action of stressors which result in the limitation of cellular replication or cell proliferation (Ben-Porath and Weinberg 2005; Herranz and Gil 2018). Therefore, cellular senescence refers to the decline or deterioration of organisms as they age which may also be due to the build-up of physiological or oxidative damage causing what are commonly known as age-associated diseases (de Verges and Nehring 2016; Herranz and Gil 2018; Hseih and Hsu 2011; Lucas and Keller 2014). In human somatic cells, one of the obvious stressors which lead to cellular senescence include the overexpression of oncoproteins or extensive cellular damage (Ben-Porath and Weinberg 2005). Another stressor which triggers senescence in humans is related to telomere shortening as a result of the inactivity of telomerase and this acts as our natural protection mechanism against cancer as this prevents the proliferation of cells (Greider 1998; Hornsby 2007). Therefore, in species where telomerase activity is organ-specific as that in humans (Sasaki and Fujiwara 2000), the ageing process is slowed down, attributable to cellular senescence reducing tumour growth (Hornsby 2007). In contrast, high non-tissue specific telomerase activity was found in cockroaches’ germ as well as somatic cells including fat bodies, neural tissue and muscles (Sasaki and Fujiwara 2000). This increased telomerase activity contributes to their shorter lifespan, in comparison to humans, and this results in the cockroaches portraying higher cell proliferation (Sasaki and Fujiwara 2000). On the other hand, research on cellular senescence in honeybees is quite extensive (de Verges and Nehring 2016; Lucas and Keller 2014). Similar to humans, older honeybees demonstrate signs of ageing by becoming weaker and show a decline in their learning curve as well as jelly production and immune system activity (de Verges and Nehring 2016). Honeybees are considered as model organisms for cellular senescence research since honeybees are social insects which exhibit different phenotypes and behaviours during their lifetime depending on their role in the caste (Kramer et al. 2021). For instance, there is a clear difference in the lifespan of honeybees which are categorised as workers or foragers and reproductive individuals as well as the queen (Kramer et al. 2021). Workers are known to live a shorter life due to extrinsic as well as physiological damages which include protein carbonylation, possibly causing brain protein carbonylation damage, indicating the consequences of oxidative stress in relation to cell senescence (Kramer et al. 2021; Seehuus et al. 2006). The oxidative stress hypothesis states that cellular senescence is mainly due to the harmful effects of the accumulation of reactive oxygen species (ROS), and it is suggested that that senescence of cells in honeybees is also highly correlated with their social role (Hseih and Hsu 2011). Nevertheless, the oxidative stress hypothesis cannot be generalized to all social insects as it is considered as an inconsistent link when predicting cellular senescence in some species (Kramer et al. 2021) and this calls for using other age-related molecules or processes to determine senescence in honeybees. For example, biochemical assays of trophocytes and fat cells in honeybees are considered as good predictors of cellular senescence and illustrate that the age of honeybees is not entirely unrelated to their senescence. Old trophocytes and fat cells express higher levels of lipid peroxidation, protein oxidation, senescence-associated β galactosidase (SA-β-Gal) and lipofuscin granules where all act as indicators of ageing or cellular senescence (Hseih and Hsu 2011). However, unlike cockroaches and humans, the telomerase activity in trophocytes and fat cells of honeybees is not influenced by age and is not associated with cellular senescence as there are no significant differences in telomere length of newly emerged and old worker honeybees (Hseih and Hsu 2011). Another age-related molecule found in honeybees is vitellogenin (Vg) as these proteins are found in higher quantities in queens than workers which is one of the reasons pertaining to the significantly longer lifespan of queens and this is due to Vg acting as an anti-oxidant by providing protection against oxidative stress (Kramer et al. 2021). Additionally, Vg levels decrease during flying which is mostly performed by worker honeybees during foraging and this results in an increased risk of damages due oxidative stress, hence, relating social role and cellular senescence once again (Kramer et al. 2021). However, very few studies investigating cellular senescence in cockroaches have been accomplished and these are warranted, given the hardiness of these species. Conclusions and future perspectives Infectious diseases, cancer, and ageing are amongst the biological challenges affecting human health; thus understanding the precise mechanisms of senescence and how disease etiologies come about in model organisms such as cockroaches, locusts and other interesting species such as crocodiles is of immense value (Siddiqui et al. 2021b; Siddiqui et al. 2021c). Insects such as cockroaches have been on this earth for millions of years and are one of the hardiest insects, able to survive without food for up to a month, without air for around 45 min and being submerged under water for 30 min (Lee et al. 2012; Wharton and Wharton 1959). Moreover, they can endure high doses of radiation: 15 times higher than humans (Lee et al. 2012; Wharton and Wharton 1959). Accordingly, it is possible that cockroaches represent a rich and diverse source of novel molecules with biological activity originating either in their tissues or their gut microbiome playing a role in contributing to the overall health and resilience. Furthermore, cockroaches may be a tractable model for research on senescence in natural populations. Till date, various studies have been conducted to determine the gut microbial composition of cockroaches, however limited work has been accomplished to understand their gut bacterial metabolites and their potential as novel and active biological molecules for use in human health, and this warrants further study, although preliminary studies have isolated and identified potentially innovative as well as some already known molecules from cockroach gut bacterial metabolites which depict inhibition of cell metabolic activity or viability reduction, and cell survival inhibition in cancerous cell lines (Akbar et al. 2018; Ali et al. 2017; Soopramanien et al. 2021). Future studies on the microbiome and associated metabolites might allow for identification of novel therapeutic leads for clinical and pre-clinical investigations in invertebrate and animal models of ageing and/or disease. Another alternative strategy may be the direct implantation of select and biologically active gut microbiome species (portraying senloytic, anti-cancer/ anti- microbial effects) into mammalian models of disease or ageing. Although these notions may seem improbable, other studies such as the discovery of insulin was made when it was found that an aqueous pancreatic extract was able to normalize diabetes in a dog. Previously, insulin for clinical use was normally obtained from cows and pigs (Crasto et al. 2016). In another study, obesity was elevated with an increase in Firmicutes in a mouse model (Turnbaugh et al. 2008). Such studies could be emulated to utilize the unique microbiome of cockroaches for the benefit of Homo sapiens with in vivo work and clinical trials in the prospective years. Being one of the most successful and diverse insects, as well as their extensive use in traditional and Chinese medicine, the lysates and gut microbiome of cockroaches may offer a worthy resource for novel bioactive molecules of therapeutic potential. Funding RS and NAK are funded by the Air Force Office of Scientific Research (AFOSR), grant number: FA 8655-20-1-7004. Declarations Conflict of interest No conflict of interest exists. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References Ahmad FB Cisewski JA Miniño A Anderson RN Provisional mortality data—United States, 2020 Morb Mortal Wkly Rep 2021 70 519 10.15585/mmwr.mm7014e1 Akbar N Siddiqui R Iqbal M Sagathevan K Khan NA Gut bacteria of cockroaches are a potential source of antibacterial compound (s) Lett Appl Microbiol 2018 66 416 426 10.1111/lam.12867 29457249 Akbar N Siddiqui R Sagathevan KA Khan NA Gut bacteria of animals/pests living in polluted environments are a potential source of antibacterials Appl Microbiol Biotechnol 2019 103 3955 3964 10.1007/s00253-019-09783-2 30941460 Alburquenque C Bucarey SA Neira-Carrillo A Urzúa B Hermosilla G Tapia CV Antifungal activity of low molecular weight chitosan against clinical isolates of Candida spp Med Mycol J 2010 48 1018 1023 10.3109/13693786.2010.486412 Ali SM Siddiqui R Ong SK Shah MR Anwar A Heard PJ Khan NA Identification and characterization of antibacterial compound (s) of cockroaches (Periplaneta americana) Appl Microbiol Biotechnol 2017 101 253 286 10.1007/s00253-016-7872-2 27743045 Alkhalifah DH Evaluation of antimicrobial activity of bacterial symbionts isolated from wild field cockroach Blattella vaga from Saudi Arabia Saudi J Biol Sci 2021 28 6239 6244 10.1016/j.sjbs.2021.06.092 34759743 Azuma K, Ifuku S, Osaki T, Okamoto Y, Minami S (2014) Preparation and biomedical applications of chitin and chitosan nanofibers. J Biomed Nanotechnol 10(10):891–2920 Balasubramanian S Priya K Revathi I Revathi A Venkatesh P Gunasekaran G Screening of antibacterial activity and biochemical assay from haemolymph of cockroach Blatta orientalis (Linnaeus, 1758) J Entomol Zool Stud 2017 5 753 758 Basseri HR Dadi-Khoeni A Bakhtiari R Abolhassani M Hajihosseini-Baghdadabadi R Isolation and purification of an antibacterial protein from immune induced haemolymph of American cockroach Periplaneta Americana J Arthropod-Borne Dis 2016 10 519 28032104 Basseri H Bakhtiyari R Hashemi SJ Baniardelani M Shahraki H Hosainpour L Antibacterial/antifungal activity of extracted chitosan from American cockroach (Dictyoptera: Blattidae) and German cockroach (Blattodea: Blattellidae) J Med Entomol 2019 56 1208 1214 10.1093/jme/tjz082 31139829 Beccaloni GW (2014) Cockroach species file online.Publisher. http://cockroach.speciesfile.org/ Accessed 28 Nov 2021 Belkaid Y Hand TW Role of the microbiota in immunity and inflammation Cell 2014 157 121 141 10.1016/j.cell.2014.03.011 24679531 Bell WJ Roth LM Nalepa CA Cockroaches: ecology, behavior, and natural history 2007 Baltimore JHU Press Ben-Porath I Weinberg RA The signals and pathways activating cellular senescence Int J Biochem Cell Biol 2005 37 961 976 10.1016/j.biocel.2004.10.013 15743671 Billah MK Pesewu GA Otu H Adetokunbo M In vitro antibacterial activities of cockroach extracts against selected bacterial pathogens Am J Res Commun 2015 3 78 94 Bird A Perceptions of epigenetics Nature 2007 447 396 10.1038/nature05913 17522671 Channel D Animal planet the most extreme bugs 2007 John Wiley & Sons Chen Z, Yang B, Ou P, Jin X (2020) Differences in the Diversity and Structure of the Gut Microbiome in Different Life Stages of the American Cockroach (Periplaneta Americana). Collado M Blasco MA Serrano M Cellular senescence in cancer and aging Cell 2007 130 223 233 10.1016/j.cell.2007.07.003 17662938 Crasto W Jarvis J Davies MJ Handbook of insulin therapies 2016 Springer Daly HV Doyen JT Ehrlich PR Introduction to insect biology and diversity 1978 McGraw-Hill Book Company De Verges J Nehring V A critical look at proximate causes of social insect senescence: damage accumulation or hyperfunction? Curr Opin Insect Sci 2016 16 69 75 10.1016/j.cois.2016.05.003 27720053 Domínguez-Santos R Pérez-Cobas AE Cuti P Pérez-Brocal V García-Ferris C Moya A Latorre A Gil R Interkingdom gut microbiome and resistome of the cockroach Blattella germanica mSystems 2021 6 e01213 e1220 10.1128/mSystems.01213-20 33975971 Evangelista DA Wipfler B Béthoux O Donath A Fujita M Kohli MK Simon S An integrative phylogenomic approach illuminates the evolutionary history of cockroaches and termites (Blattodea) Proc R Soc B 2019 286 1895 20182076 10.1098/rspb.2018.2076 Ganesan AR Guru MS Balasubramanian B Mohan K Liu WC Arasu MV Al-Dhabi NA Duraipandiyan V Ignacimuthu S Sudhakar MP Seedevi P Biopolymer from edible marine invertebrates: a potential functional food J King Saud Univ - Sci 2020 32 1772 1777 10.1016/j.jksus.2020.01.015 Greider CW Telomeres and senescence: the history, the experiment, the future Curr Biol 1998 8 R178 R181 10.1016/S0960-9822(98)70105-8 9501064 Guzman J Vilcinskas A Bacteria associated with cockroaches: health risk or biotechnological opportunity? Appl Microbiol Biotechnol 2020 104 10369 10387 10.1007/s00253-020-10973-6 33128616 Harari YN Sapiens: a brief history of humankind 2014 New York Random House Herranz N Gil J Mechanisms and functions of cellular senescence J Clin Investig 2018 128 1238 1246 10.1172/JCI95148 29608137 Hornsby PJ Telomerase and the aging process Exp Gerontol 2007 42 575 581 10.1016/j.exger.2007.03.007 17482404 Hsieh YS Hsu CY Honeybee trophocytes and fat cells as target cells for cellular senescence studies Exp Gerontol 2011 46 233 240 10.1016/j.exger.2010.10.007 20965244 Jahnes BC, Poudel K, Staats AM, Sabree ZL (2021) Microbial colonization promotes model cockroach gut tissue growth and development. J Insect Physiol 133:104274. Jandhyala SM, Talukdar R, Subramanyam C, Vuyyuru H, Sasikala M, Reddy DN (2015) Role of the normal gut microbiota. World j gastroenterol: WJG 21(29):8787. Kambhampati S A phylogeny of cockroaches and related insects based on DNA sequence of mitochondrial ribosomal RNA genes Proc Natl Acad Sci 1995 92 2017 2020 10.1073/pnas.92.6.2017 7534409 Kau AL Ahern PP Griffin NW Goodman AL Gordon JI Human nutrition, the gut microbiome and the immune system Nature 2011 474 327 336 10.1038/nature10213 21677749 Kim IW Choi RY Lee JH Seo M Lee HJ Kim MA Kim SH Kim I Hwang JS Anticancer activity of periplanetasin-5, an antimicrobial peptide from the cockroach Periplaneta americana J Microbiol Biotechnol 2021 31 1343 1349 10.4014/jmb.2104.04040 34409948 Koehler PG Ol F Branscome D Cockroaches and their management 1999 Florida University of Florida Cooperative Extension Service, Institute of Food and Agriculture Sciences, EDIS Kramer BH Nehring V Buttstedt A Heinze J Korb J Libbrecht R Meusemann K Paxton RJ Séguret A Schaub F Bernadou A Oxidative stress and senescence in social insects: a significant but inconsistent link? Philos Trans R Soc B 2021 376 20190732 10.1098/rstb.2019.0732 Latifi M, Alikhani MY, Salehzadeh A, Nazari M, Bandani AR, Zahirnia AH (2015) The antibacterial effect of American cockroach hemolymph on the nosocomial pathogenic bacteria.SID 0–0. Lauprasert P Sitthicharoenchai D Thirakhupt K Pradatsudarasar AO Food preference and feeding behavior of the German cockroach, Blattella germanica (Linnaeus) J Sci Res Chula Univ 2006 31 121 126 Lee S Siddiqui R Khan NA Animals living in polluted environments are potential source of antimicrobials against infectious agents Pathog Glob Health 2012 106 218 223 10.1179/2047773212Y.0000000033 23265422 Li S Zhu S Jia Q Yuan D Ren C Li K Liu S Cui Y Zhao H Cao Y Fang G The genomic and functional landscapes of developmental plasticity in the American cockroach Nat Commun 2018 9 1 1 29317637 Lu K Zhou J Deng J Li Y Wu C Bao J Periplaneta americana oligosaccharides exert anti-inflammatory activity through immunoregulation and modulation of gut microbiota in acute colitis mice model Molecules 2021 26 1718 10.3390/molecules26061718 33808686 Lucas ER Keller L Ageing and somatic maintenance in social insects Curr Opin Insect Sci 2014 5 31 36 10.1016/j.cois.2014.09.009 32846739 Ma X Hu Y Li X Zheng X Wang Y Zhang J Fu C Geng F Periplaneta americana ameliorates dextran sulfate sodium-induced ulcerative colitis in rats by Keap1/Nrf-2 activation, intestinal barrier function, and gut microbiota regulation Front Pharmacol 2018 9 944 10.3389/fphar.2018.00944 30186174 Mahboub MT Hassan MI Bream AS Mohamed AF Abdel-Samad MR Evaluation of the antibacterial and antifungal activities of chitosan prepared from the American cockroach, (Periplaneta americana) Egypt Acad J Biol 2021 13 39 46 Malard F Dore J Gaugler B Mohty M Introduction to host microbiome symbiosis in health and disease Mucosal Immunol 2021 14 547 554 10.1038/s41385-020-00365-4 33299088 Malinoski MK Cockroaches 1999 University of Maryland, College Park University of Maryland Extension McPherson S Wada-Katsumata A Hatano E Silverman J Schal C Comparison of diet preferences of laboratory-reared and apartment-collected german cockroaches J Econ Entomol 2021 114 2189 2197 10.1093/jee/toab139 34260722 Mogbo TC, Okeke JJ, Nwosu MC, Ijuh II (2013) Assessment of ecofriendly methods of eliminating cockroaches. Adv Biosci Biotechnol 1(2):81–85 Mosaheb MU Khan NA Siddiqui R Cockroaches, locusts, and envenomating arthropods: a promising source of antimicrobials Iran J Basic Med Sci 2018 21 873 30524685 Perrott RC Miller DM American cockroach 2010 Virginia Virginia Cooperative Extension (VCE) Publications Potter M Cockroach elimination in homes & apartments 2010 Lexington University of Kentucky College of Agriculture Qi C Xu CJ Koppelman GH The role of epigenetics in the development of childhood asthma Expert Rev Clin Immunol 2019 15 1287 1302 10.1080/1744666X.2020.1686977 31674254 Ray R Potts R Pietri JE The persistence of Escherichia coli infection in German cockroaches (Blattodea: Blattellidae) varies between host developmental stages and is influenced by the gut microbiota J Med Entomol 2020 57 1964 1971 10.1093/jme/tjaa108 32516418 Roth LM Systematics and phylogeny of cockroaches (Dictyoptera: Blattaria) Orient inSects 2003 37 1 86 10.1080/00305316.2003.10417344 Roth LM Willis ER The medical and veterinary importance of cockroaches SMC 1957 134 1 147 Sasaki T Fujiwara H Detection and distribution patterns of telomerase activity in insects Eur J Biochem 2000 267 3025 3031 10.1046/j.1432-1033.2000.01323.x 10806402 Schal C Gautier JY Bell WJ Behavioural Ecology of Cockroaches Biol 1984 59 209 254 Seabrooks L Hu L Insects: an underrepresented resource for the discovery of biologically active natural products Acta Pharm Sin B 2017 7 409 426 10.1016/j.apsb.2017.05.001 28752026 Seehuus SC Krekling T Amdam GV Cellular senescence in honey bee brain is largely independent of chronological age Exp Gerontol 2006 41 1117 1125 10.1016/j.exger.2006.08.004 17052880 Siddiqui R Qaisar R Goswami N Khan NA Elmoselhi A Effect of microgravity environment on gut microbiome and angiogenesis Life 2021 11 1008 10.3390/life11101008 34685381 Siddiqui R Cruz Soares N Khan NA Crocodile gut microbiome is a potential source of novel bioactive molecules ACS Pharmacol Transl Sci 2021 4 1260 1261 10.1021/acsptsci.1c00108 34151219 Siddiqui R Muhammad JS Khan NA Locust as an in vivo model ACS Chem Neurosci 2021 12 1469 1471 10.1021/acschemneuro.1c00190 33877824 Sieksmeyer T (2021) Host-microbiome-pathogen interactions in cockroaches (Doctoral dissertation). https://refubium.fu-berlin.de/bitstream/handle/fub188/29499/Dissertation_Thorben_Sieksmeyer.pdf?sequence=4&ampisAllowed=y Slaytor M Cellulose digestion in termites and cockroaches: what role do symbionts play? Comp Biochem Physiol Part B 1992 103 775 784 10.1016/0305-0491(92)90194-V Soopramanien M Mungroo MR Sagathevan KA Khan NA Siddiqui R Invertebrates living in polluted environments are potential source of novel anticancer agents Marmara Pharm J 2019 23 1079 1089 Soopramanien M Khan NA Siddiqui R Gut microbiota of animals living in polluted environments are a potential resource of anticancer molecules J Appl Microbiol 2021 131 1039 1055 10.1111/jam.14981 33368930 Tinker KA Ottesen EA The core gut microbiome of the American cockroach, Periplaneta americana, is stable and resilient to dietary shifts Appl Environ Microbiol 2016 82 6603 6610 10.1128/AEM.01837-16 27590811 Tinker KA Ottesen EA Phylosymbiosis across deeply diverging lineages of omnivorous cockroaches (order Blattodea) Appl Environ Microbiol 2020 86 e02513 e2519 10.1128/AEM.02513-19 31953337 Tinker KA Ottesen EA Differences in gut microbiome composition between sympatric wild and allopatric laboratory populations of omnivorous cockroaches Front Microbiol 2021 12 703785 10.3389/fmicb.2021.703785 34394050 Turnbaugh PJ Backhed F Fulton L Gordon JI Marked alterations in the distal gut microbiome linked to diet-induced obesity CELL HOST MICROBE 2008 3 213 10.1016/j.chom.2008.02.015 18407065 Villalba de la Pena M Piskobulu V Murgatroyd C Hager R DNA methylation patterns respond to thermal stress in the viviparous cockroach Diploptera punctata Epigenetics 2021 16 313 326 10.1080/15592294.2020.1795603 32713247 Vršanský P van de Kamp T Azar D Prokin A Vidlička LU Vagovič P Cockroaches probably cleaned up after dinosaurs PLoS ONE 2013 8 e80560 10.1371/journal.pone.0080560 24324610 Wang XY He ZC Song LY Spencer S Yang LX Peng F Liu GM Hu MH Li HB Wu XM Zeng S Chemotherapeutic effects of bioassay-guided extracts of the American cockroach, Periplaneta americana Integr Cancer Ther 2011 10 NP12 23 10.1177/1534735411413467 21733985 Wang Z Shi Y Qiu Z Che Y Lo N Reconstructing the phylogeny of Blattodea: robust support for interfamilial relationships and major clades Sci Rep 2017 7 1 8 28127051 Wang T Liao Q Wu Y Wang X Fu C Geng F Qu Y Zhang J A composite hydrogel loading natural polysaccharides derived from Periplaneta americana herbal residue for diabetic wound healing Int J Biol Macromol 2020 164 3846 3857 10.1016/j.ijbiomac.2020.08.156 32841667 Weihmann T Reinhardt L Weißing K Siebert T Wipfler B Fast and powerful: biomechanics and bite forces of the mandibles in the American cockroach Periplaneta americana PLoS ONE 2015 10 e0141226 10.1371/journal.pone.0141226 26559671 Wharton DR Wharton ML The effect of radiation on the longevity of the cockroach, Periplaneta americana, as affected by dose, age, sex and food intake Radiat Res 1959 11 600 615 10.2307/3570814 13844254 Xin L Yu-Ping TA Rui LI Yi JI Jian-Ming GU Jin-Long ZH Shao-Xiong DI Xiang-Zhi LI Ru-Rong LI Jin-Ao DU Antipyretic and anti-inflammatory activities of Thais luteostoma extracts and underlying mechanisms Chin J Nat Med 2015 13 192 198 25835363 Younes I Hajji S Frachet V Rinaudo M Jellouli K Nasri M Chitin extraction from shrimp shell using enzymatic treatment. Antitumor, antioxidant and antimicrobial activities of chitosan Int J Biol Macromol 2014 69 489 498 10.1016/j.ijbiomac.2014.06.013 24950313 Yun J Hwang JS Lee DG The antifungal activity of the peptide, periplanetasin-2, derived from American cockroach Periplaneta americana Biochem J 2017 474 3027 3043 10.1042/BCJ20170461 28733329 Zhao Y Yang A Tu P Hu Z Anti-tumor effects of the American cockroach, Periplaneta americana Chin Med 2017 12 1 6 10.1186/s13020-017-0149-6 28053654 Zheng D Liwinski T Elinav E Interaction between microbiota and immunity in health and disease Cell Res 2020 30 492 506 10.1038/s41422-020-0332-7 32433595 Zhu JJ, Yao S, Guo X, Yue BS, Ma XY, Li J (2018) Bioactivity-guided screening of wound-healing active constituents from American cockroach (Periplaneta americana). Molecules 23(1):101.

0

PMC9753028

NO-CC CODE

2022-12-16 23:25:27

no

Appl Entomol Zool. 2022 Dec 15;:1-11

utf-8

Appl Entomol Zool

10.1007/s13355-022-00810-9

oa_other

==== Front Clin Neuroradiol Clin Neuroradiol Clinical Neuroradiology 1869-1439 1869-1447 Springer Berlin Heidelberg Berlin/Heidelberg 1238 10.1007/s00062-022-01238-y Original Article Clinical and Neuroimaging Characteristics of Ischemic Stroke in Rhino-Orbito-Cerebral Mucormycosis Associated with COVID-19 Najafi Mohammad Amin 1 Zandifar Alireza 2 Ramezani Neda 3 Paydari Hanie 1 Kheradmand Mohsen 1 Ansari Behnaz 1 Najafi Mohammad Reza 1 http://orcid.org/0000-0002-8654-4548 Hajiahmadi Somayeh som.hajiahmadi1372@gmail.com 45 Khorvash Fariborz 1 Saadatnia Mohammad 1 Vossough Arastoo 2 1 grid.411036.1 0000 0001 1498 685X Isfahan Neurosciences Research Center, Department of Neurology, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran 2 grid.239552.a 0000 0001 0680 8770 Department of Radiology, Children’s Hospital of Philadelphia, Philadelphia, PA USA 3 grid.411036.1 0000 0001 1498 685X Isfahan Neurosciences Research Center, Isfahan University of Medical Sciences, Isfahan, Iran 4 grid.411036.1 0000 0001 1498 685X Department of Radiology, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran 5 grid.411036.1 0000 0001 1498 685X Division of Neuroradiology, Department of Radiology, Alzahra Hospital, Isfahan University of Medical Sciences, Hezar Jarib Street, Isfahan, Iran 15 12 2022 19 12 6 2022 7 11 2022 © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Purpose The aim of this study was to compare clinical, neuroimaging, and laboratory features of rhino-orbito-cerebral mucormycosis (ROCM) in COVID-19 patients with and without ischemic stroke complications. Methods This observational study was conducted between August and December 2021 and 48 patients who had confirmed ROCM due to COVID-19, according to neuroimaging and histopathology/mycology evidence were included. Brain, orbit and paranasal sinus imaging was performed in all included patients. Data pertaining to clinical, neuroimaging, and laboratory characteristics and risk factors were collected and compared between patients with and without ischemic stroke complications. Results Of the patients 17 were diagnosed with ischemic stroke. Watershed infarction was the most common pattern (N = 13, 76.4%). Prevalence of conventional risk factors of stroke showed no significant differences between groups (patients with stroke vs. without stroke). Cavernous sinus (p = 0.001, odds ratio, OR = 12.8, 95% confidence interval, CI: 2.3–72) and ICA (p < 0.001, OR = 16.31, 95%CI: 2.91–91.14) involvement was more common in patients with stroke. Internal carotid artery (ICA) size (on the affected side) in patients with ischemic stroke was significantly smaller than in patients without stroke (median = 2.4 mm, interquartile range, IQR: 1.3–4 vs. 3.8 mm, IQR: 3.2–4.3, p = 0.004). Superior ophthalmic vein (SOV) size (on the affected side) in patients with stroke was significantly larger than patients without stroke (2.2 mm, IQR: 1.5–2.5 vs. 1.45 mm IQR: 1.1–1.8, p = 0.019). Involvement of the ethmoid and frontal sinuses were higher in patients with stroke (p = 0.007, OR = 1.85, 95% CI: 1.37–2.49 and p = 0.011, OR = 5, 95% CI: 1.4–18.2, respectively). Patients with stroke had higher D‑dimer levels, WBC counts, neutrophil/lymphocyte ratios, and BUN/Cr ratio (all p < 0.05). Conclusion Stroke-related ROCM was not associated with conventional ischemic stroke risk factors. Neuroimaging investigations including qualitative and quantitative parameters of cavernous sinus, ICA and SOV are useful to better understand the mechanism of stroke-related ROCM in COVID-19 patients. Keywords COVID-19 associated mucormycosis Mucormycosis associated stroke Brain MRI Cerebrovascular involvement Internal carotid artery Cavernous sinus ==== Body pmcIntroduction Mucormycosis, caused by fungi of the order Mucorales, can manifest as an angioinvasive fungal infection [1]. Some conditions such as overt diabetes mellitus, hematologic malignancies, and organ transplantation predispose patients to mucormycosis [2]. Different clinical presentations have been described, including pulmonary, cutaneous, gastrointestinal, rhinocerebral, and disseminated forms [3, 4]. An outbreak of mucormycosis has been described during the coronavirus disease 2019 (COVID-19) pandemic, caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in India. There was a more than fivefold rise in hospital admissions due to invasive fungal infection during the COVID-19 pandemic compared to the previous 2 years [1, 5, 6]. There is no sufficient evidence on the exact incidence and risk factors for cerebrovascular events in mucormycosis patients. Two recent studies have reported that 11.8% and 14.8% of patients with COVID-19 associated mucormycosis (CAM) had cerebrovascular events, including different types of strokes [7–9]. Kulkarni et al. reported that ischemic stroke is the most common neurological manifestation [8]. Prior studies that assessed the imaging characteristics of rhino-orbito-cerebral mucormycosis (RCOM), had found paranasal sinus infection, orbital infection, arterial thrombosis, cavernous sinus thrombosis, cerebral hemorrhage, and mycotic aneurysms are among the findings in this group of patients [10]. Although some studies have reported on the clinical characteristics and epidemiology of neurological complications of CAM, there is a paucity of studies on the neuroimaging and laboratory findings in these patients. In this study, we aimed to compare clinical, neuroimaging, and laboratory features of ROCM in the setting of CAM between patients with and without ischemic stroke complications, to find the potential risk factors associated with stroke in these patients. Method Study Design This single-center, observational study included a total of 48 adult patients who were admitted to the largest medical center, exclusively designated for the care of COVID-19 patients in our country. They all had pertinent clinical and neuroimaging findings with histopathology/mycology confirmed ROCM due to CAM. The study period was between August 2021 and December 2021, during the fourth wave of COVID-19 outbreak in the country. Subjects A COVID-19 infection was diagnosed based on SARS-CoV‑2 reverse transcription-polymerase chain reaction (RT-PCR) SARS-CoV‑2 viral ribonucleic acid (RNA) or positive rapid antigen test (Arvin Biohealth Co., Tehran, Iran) on oropharyngeal and nasopharyngeal swab specimens. The initial diagnosis of mucormycosis was made by clinical and imaging evaluations of patients who presented with suspicious symptoms including new onset ptosis, facial numbness, extraocular movement restriction, and diminution of vision. The initial diagnosis was confirmed based on histopathology and mycology evaluation of biopsied tissues including staining with potassium hydroxide (KOH) or calcofluor stain of obtained tissue and/or fungal growth in cultures. CAM was defined based on previous literature as mucormycosis occurrence in patients with confirmed COVID-19 diagnosis within the prior 12 weeks [11]. ROCM was defined as histopathology confirmed mucormycosis infection in the samples from nose, paranasal sinuses, orbit, and/or intracranial structures. All patients underwent brain computed tomography (CT) with or without magnetic resonance imaging (MRI) of the brain, orbits, and paranasal sinuses. The MRI protocol in patients with possible diagnosis of ROCM included T1 weighted image (T1WI), T2WI, fluid attenuation inversion recovery (FLAIR), T2WI with fat suppression, gradient recalled echo (GRE), diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC), magnetic resonance angiography, magnetic resonance venography, and postcontrast T1WI sequences, except in cases with MR contraindications. All neuroimaging studies were evaluated by a radiologist (S.H. with more than 10 years of experience) and one neurologist (M.S. with more than 20 years of experience and neuroradiology fellowship). In order to determine the potential risk factors associated with stroke in CAM patients, cases were divided into two groups based on whether they had schemic stroke or not. The etiology of ischemic stroke was further established utilizing additional vascular imaging modalities (e.g., cervical and transcranial color Doppler ultrasound, CT angiography, MR angiography), cardiac evaluations (including echocardiography and electrocardiogram). Variables Demographic and clinical historical data were acquired from the patients directly or the patients’ relatives upon enrolment. We collected conventional medical risk factors for stroke including hypertension (HTN), diabetes mellitus (DM), ischemic heart disease (IHD), hyperlipidemia (HLP), heart failure (HF) and atrial fibrillation (AF). We also recorded the detailed neurological examination data at the time of admission, the time interval between diagnosis of COVID-19 and mucormycosis, and patient survival outcome. Imaging features including pattern of stroke (watershed/non-watershed), laterality, and vascular territory as well as imaging features related to of the cavernous sinus, internal carotid artery (ICA), superior ophthalmic vein (SOV), cerebral parenchyma, paranasal sinuses, soft tissue spaces surrounding the paranasal sinuses, and orbits were evaluated for all patients. Cavernous sinus involvement was considered positive as lateral wall bulging (Fig. 1b and 2b), increased signal intensity on spin echo MRI, postcontrast filling defect (Fig. 1b), associated dural enhancement (Fig. 1c), and asymmetric increase in cavernous sinus size measurement (Fig. 1d and 2c). Cavernous sinus size was measured as maximum width of cavernous sinus on the axial plane (Fig. 1d and 2c). ICA involvement was considered as asymmetrically increased signal intensity in the ICA and its adventitia (Fig. 1e,f and 2d), and asymmetric decrease in size of ICA lumen (Fig. 1f and 2e). ICA size was calculated based on the flow void diameter of the vertical C3/C4 segments of the ICA perpendicular to the axial plane just as it enters the cavernous sinus (Fig. 2e). SOV involvement was delineated by asymmetric increase in size of SOV lumen, increased signal intensity in the SOV, and enhancement of adventitia of SOV. SOV diameter was measured perpendicular to the SOV on coronal images on the closest slice to the rear of the globe. Proptosis was defined as a globe protrusion of more than 21 mm anterior to the interzygomatic line at the level of the lens on axial images (Fig. 2f; [12]). Ischemic stroke subtypes were categorized according to TOAST criteria classification: 1) large artery atherosclerosis, 2) cardioembolism, 3) small vessel occlusion, 4) stroke of other determined etiology, and 5) stroke of undetermined etiology [13].Fig. 1 A patient in his late 30s with history of COVID-19 infection three weeks before the admission for hypoesthesia in the left forehead and cheek for two days. It was accompanied by ptosis, proptosis, extraocular movement limitation, and diminished vision. Axial diffusion-weighted imaging (a) demonstrates unilateral ischemic stroke in deep (internal) and cortical (external) watershed zones. Axial postcontrast MRI showing a filling defect (yellow arrow) along with bulging of the lateral wall of the left cavernous sinus (red arrow) (b), dural enhancement in nearby left middle cranial fossa (yellow arrow) (c) and increased cavernous sinus width to 11.1 mm (yellow arrow) (d). Axial T2-weighted fat suppressed MRI showing increased signal intensity in petrosal (e) and proximal cavernous (f) segments of the left internal carotid artery (ICA) (red arrows). The normal right ICA is shown with yellow arrows (e,f) Fig. 2 A patient in his early 30s with history of COVID-19 infection four weeks before the admission for hypoesthesia of the left forehead and cheek for two days. It was accompanied by ptosis, proptosis, frozen eyes, and diminution of vision. Histological evaluation confirmed the diagnosis rhino-orbito-cerebral mycosis. Axial diffusion-weighted imaging (a) demonstrates unilateral watershed zone ischemic stroke. Post-contrast brain MRI shows a filling defect and bulging of the lateral wall of the left cavernous sinus (yellow arrow) (b) with the cavernous sinus width increased to 10.4 mm (c). Axial fat suppressed T2-weighted MRI shows subtle increased signal intensity in and around the petrosal segment of the left internal carotid artery (ICA) (red arrow) with the unaffected right ICA shown with a yellow arrow (d). ICA dimeters in the affected (3.6 mm) and non-affected (5.1 mm) sides are shown (e). Axial fat suppressed T2-weighted MRI showing proptosis of left eye (26 mm protrusion, anterior to the interzygomatic line at the level of the lens) (f) and abnormal signal affecting the periantral (perimaxillary) fat on the affected side (yellow arrow) (g) Statistical Analysis Statistical analysis was conducted using IBM SPSS Statistics for Windows, Version 26.0 (Armonk, NY, USA). Quantitative results are presented as median (interquartile range). Categorical variables were reported as counts and percentages, as appropriate, χ2-tests were used to compare categorical variables and odds ratios (OR, 95% confidence interval, CI) were calculated for dichotomous variables. Mann-Whitney U test and Wilcoxon signed-rank test were applied to compare continuous variables between two groups and two related samples, respectively. A P value < 0.05 was considered as the level of statistical significance. Results Demographics and Clinical Characteristics Over the study period, 75 patients were admitted to our hospital with the clinical suspicion of ROCM and recent history of COVID-19 and 59 patients had neuroimaging findings suggestive of possible ROCM; however, histopathology confirmed the diagnosis in only 50 patients. Of the patients two were excluded due to the history of COVID-19 preceding the 12-week period before clinical onset of ROCM (Fig. 3) and 48 patients (34 males, and 14 females) were included in the analysis with a median age of 58.5 years (IQR: 46–65.5 years) (Table 1). Of the 48 patients, acute ischemic stroke (AIS) was diagnosed in 17 (35.4%). The median interval time between the first symptoms of COVID-19 and first symptoms of ROCM was 15 days (IQR: 9.75–23.2 days) and there was no statistically significant difference between patients with and without AIS (15 days, IQR: 5–25 days vs. 15 days, IQR: 12.5–22.5, P = 0.472). Prevalence of conventional risk factors of stroke, i.e. HTN, DM, IHD, HLP and AF, showed no statistically significant differences between the two groups. Neurological examination of patients at the time of admission also showed no significant differences in terms of cavernous sinus related symptoms; however, other symptoms including hemiparesis, hemi-hypoesthesia and dysarthria were more common among patients with ischemic stroke (all p < 0.05). Ten patients died during hospitalization, which includes 41.1% (N = 7) of patients with ischemic stroke and 9.6% (N = 3) of patients without ischemic stroke (OR = 6.53, 95% CI: 1.4–30.2, P = 0.001) (Table 1).Fig. 3 The study workflow diagram for the selection of patients with rhino-orbito-cerebral mucormycosis (RCOM) Table 1 Demographics, risk factors, and neurological examination at the time of admission Total (n = 48) AIS (n = 17) Without AIS (n = 31) p Value OR (95% CI) Age (years)a 58.5 (46–65.5) 56 (41.5–65) 59 (49–66) 0.353 n/a Gender Male 34 13 21 0.525 1.5 (0.4–5.9) Female 14 4 10 Conventional stroke risk factors HTN 18 6 12 0.75 0.8 (0.2–2.8) DM 32 9 23 0.135 0.39 (0.11–1.4) IHD 9 2 7 0.359 0.45 (0.08–2.5) HLP 5 1 4 0.426 0.4 (0.04–3.97) AF 1 1 0 0.174 n/a Duration between mucormycosis and COVID-19 diagnoses (days)a 15 (9.75–23.2) 15 (5–25) 15 (12.5–22.5) 0.472 n/a Neurological examination in admit Ptosis 26 11 15 0.278 1.95 (0.58–6.61) Facial numbness 20 5 15 0.202 0.44 (0.12–1.56) Extra ocular movement restriction 29 12 17 0.286 1.97 (0.56–6.98) Diminution of vision 33 14 19 0.132 2.94 (0.7–12.45) Impaired pupillary reflex 28 12 16 0.202 2.25 (0.64–7.92) Limb weakness 6 5 1 0.009* 12.5 (1.3–118.4) Limb hypoesthesia 4 4 0 0.005* 3.38 (2.14–5.34) Dysarthria 6 5 1 0.009* 12.5 (1.3–118.4) Headache 38 15 23 0.252 2.61 (0.48–14) Death in hospital 10 7 3 0.001* 6.53 (1.4–30.2) HTN hypertension, DM diabetes mellitus, IHD ischemic heart disease, HLP hyperlipidemia, AF atrial fibrillation, n/a not applicable, OR odds ratio, CI 95% confidence interval, AIS acute ischemic stroke *p value < 0.05 aQuantitative data reported as Median (interquartile range) Neuroimaging Findings All included patients underwent MRI, 17 patients had ischemic stroke and 5 patients had intraparenchymal abscesses, including in the cerebellum (N = 2), temporal lobe (N = 2), and frontal lobe (N = 1). The cavernous sinus was involved in 10 patients (20.8%), 8 of those developed AIS and 2 did not have any infarcts (OR = 12.8, 95% CI: 2.3–72, p = 0.001). There were no differences between patients with and without stroke in terms of bulging of cavernous sinus lateral wall, cavernous signal intensity in spin echo sequence (flow void vs. increased signal intensity), cavernous sinus filling defect, and dural enhancement. There were also no differences between patients with and without ischemic stroke with respect to the width of the cavernous sinus (Table 2).Table 2 Imaging findings of rhino-orbito-cerebral mucormycosis (ROCM) Total (N = 48) AIS (N = 17) Without AIS (N = 31) p Value OR (95% CI) Cavernous sinus features A. Cavernous sinus involvement 10 8 2 0.001* 12.8 (2.3–72) A.1. Cavernous sinus lateral wall bulging 3 2 1 0.490 0.33 (0.01–8.1) A.2. Cavernous signal intensity on spin echo MRI 0.301 n/a  Flow void 3 3 0  Increased signal intensity 7 5 2 A.3. Cavernous sinus filling defect 2 2 0 0.429 0.75 (0.5–1.1) A.4. Nearby abnormal dural enhancement 6 4 2 0.197 1.5 (0.8–2.6) B. Cavernous sinus size (mm)a 6.6 (6.1–8.3) 6.4 (5.3–7.3) 6.8 (6.3–8.6) 0.178 n/a ICA features A. ICA involvement 11 9 2 0.001* 16.3 (2.9–91.1) A.1. ICA signal intensity on spin echo MRI 0.011 * n/a  Flow void 3 1 2  Increased signal intensity 8 8 0 A.2. ICA enhancement 4 3 1 0.809 1.5 (0.05–40.6) B. ICA size (mm)a 3.6 (2.5–4.2) 2.4 (1.3–4) 3.8 (3.2–4.3) 0.004 n/a Superior ophthalmic vein features A. SOV involvement 8 5 3 0.079 3.9 (0.8–19) A.1. SOV signal intensity in spin echo MRI 0.035 n/a  Flow void 2 0 2  Increased signal intensity 6 5 1 A.2. SOV enhancement 5 3 2 0.439 n/a B. SOV size (mm)a 1.6 (1.2–2.1) 2.2 (1.5–2.5) 1.45 (1.1–1.8) 0.019 n/a Intraorbital involvement 27 13 14 0.079 3.25 (0.8–12.4) Proptosis 16 7 9 0.43 1.6 (0.47–5.6) Paranasal sinus involvement Maxillary 43 15 28 0.821 0.8 (0.1–5.3) Frontal 22 12 10 0.011 5 (1.4–18.2) Sphenoid 34 15 19 0.049 4.7 (0.91–24.5) Ethmoid 37 17 20 0.007* 1.85 (1.37–2.49) Nasal cavity 35 14 21 0.351 2 (0.46–8.7) Periantral (perimaxillary) fat signal abnormality 36 10 26 0.03* 0.22 (0.05–0.91) Superior orbital fissure involvement 9 5 4 0.329 2.1 (0.46–9.8) AIS acute ischemic stroke, OR odds ratio, CI 95% confidence interval, n/a not applicable, ICA Internal carotid artery, SOV Superior ophthalmic vein aQuantitative data reported as Median (interquartile range), p value The ICA was involved in 11 patients (22.9%), 9 of those patients developed ischemic stroke (OR = 16.31, 95% CI: 2.91–91.14, p < 0.001). ICA signal intensity on spin echo sequence showed preservation of the ICA flow void on the affected side in 3 patients and increased signal intensity in 8 patients (all in those patients with ischemic stroke) (p = 0.011). ICA diameter (on the affected side) in patients with ischemic stroke was significantly smaller than patients without stroke (2.4 mm, IQR: 1.3–4 vs. 3.8 mm, IQR: 3.2–4.3, p = 0.004) (Table 2). SOV was involved in 8 patients (16.6%), 5 patients had ischemic stroke and 3 did not (p = 0.079). SOV signal intensity in spin echo sequence showed flow void on the affected side in 2 patients (both in patients without ischemic stroke) and increased signal intensity in 6 patients (5 had ischemic stroke and 1 in the other group) (p = 0.035). SOV size (on the affected side) in patients with ischemic stroke was significantly larger than patients without ischemic stroke (2.2, IQR: 1.5–2.5 vs. 1.45, IQR: 1.1–1.8, p = 0.019) (Table 2). Imaging of the paranasal sinuses revealed maxillary sinus, ethmoid sinus, sphenoid sinus, and frontal sinus involvement in 43 (89.5%), 37 (77%), 34 (70%) and 22 (45%) patients, respectively. Involvement of ethmoid (OR = 1.85, 95% CI: 1.37–2.49, p = 0.007) and frontal (OR = 5, 95% CI: 1.4–18.2, p = 0.011) sinuses were significantly higher in patients with stroke (p = 0.009 and 0.041, respectively). Involvement of other sinuses showed no significant differences between patients with stroke and without stroke. Involvement of superior orbital fissure showed no significant differences between patients with and without stroke (p = 0.329). Presence of periantral fat signal abnormalities was significantly higher in patients without stroke (OR = 0.022, 95% CI: 0.05–0.91, p = 0.03) (Table 2). Stroke Characteristics Using the TOAST classification of stroke, among the patients with AIS, most were categorized as other determined etiology (N = 13, 76.4%) and 5 patients (29.4%) were classified as undetermined etiology (negative results after standard evaluation). Regarding the laterality of AIS, 13 patients (76.4%) had unilateral cerebral hemisphere stroke and 4 patients (23.5%) had bilateral stroke. In terms of the affected territory, anterior circulation was involved in 2 patients (11.7%) (2 anterior cerebral artery territory), posterior circulation was involved in 2 patients (10.5%), and 13 patients (76.4%) had watershed (border zone) infarcts. Of those with a watershed infarct pattern, the involvement was that of deep (internal) watershed in 8 patients, cortical (external) watershed in 1 patient, and mixed internal and external watershed infarcts in 4 patients. Laboratory Data Patients with AIS had a higher D‑dimer level (2100 ng/mL, IQR: 1200–3100 vs. 850 ng/mL, IQR: 510–2074], p = 0.031), white blood cell count (15,100 per µL, IQR: 10,900–16,875 vs. 11,100 per µL, IQR: 7900–14,900, p = 0.023), and neutrophil to lymphocyte ratio (NLR) (10.1, IQR: 7.4–14 vs. 4.6, IQR: 2.5–9.7, p = 0.01). Lymphocyte percentage was significantly lower in patients with AIS (8.45%, IQR: 6.25–11.2 vs. 16.8%, IQR: 8.8–27.2, p = 0.01). Moreover, BUN/Cr ratios were significantly higher in patients with AIS (21.4, IQR: 15.8–23 vs. 14.7, IQR: 10.9–19.1, p = 0.048). Other laboratory values showed no significant differences between patients with and without ischemic stroke (Table 3).Table 3 Baseline laboratory findings of patients at the time of hospital admission Biomarkera Total (n = 48) AIS (n = 17) Without AIS (n = 31) p Value Erythrocyte sedimentation rate, mm/hour 56 (35–80) 74 (49–85.5) 53 (28.5–72.5) 0.82 C‑reactive protein, mg/L 62.5 (4.5–80.75) 78 (50–84) 13 (2–80) 0.148 D‑Dimer, ng/mL 1350 (520–2500) 2100 (1200–3100) 850 (510–2074) 0.031* White blood cell count, /mm3 11,800 (8400–15,700) 15,100 (10,900–16,875) 11,100 (7900–14,900) 0.023* Lymphocyte percentage 11 (7.4–21.2) 8.45 (6.25–11.2) 16.8 (8.8–27.2) 0.01* Neutrophil to lymphocyte ratio 7.54 (3.43–11.7) 10.1 (7.4–14) 4.6 (2.5–9.7) 0.01* Hemoglobin, g/dl 12.1 (10.9–13.2) 12.7 (11.1–13.2) 11.8 (10.9–13.5) 0.87 Platelet count, /mm3 258 × 10 3 (176 × 103 − 329 × 10 3) 253.5 × 10 3 (186.7 × 10 3 − 314 × 10 3) 263 × 10 3 (167 × 10 3 − 358 × 10 3) 0.472 Alanine aminotransferase, U/L 32 (21–47) 32 (21.7–39) 28.5 (18.2–53.7) 0.88 Aspartate aminotransferase, U/L 24 (16–33) 21.5 (15.2–27.2) 27 (18.5–35.2) 0.2 Blood urea nitrogen (BUN), mg/dl 17 (12–26) 22.5 (15.2–27) 15 (11–24) 0.074 Creatinine (Cr), mg/dl 1 (0.9–1.15) 1 (0.8–1.1) 1 (0.9–1.2) 0.49 BUN/Cr ratio 16 (11.5–22.1) 21.4 (15.8–23) 14.7 (10.9–19.1) 0.048* Blood glucose, mg/dl 228 (150–313) 243 (185–371) 203 (126.5–279) 0.16 AIS Arterial Ischemic Stroke *p value < 0.05 aQuantitative data reported as Median (interquartile range) Discussion Rhino-orbito-cerebral mucormycosis (RCOM) is a life-threatening condition which needs timely attention to prevent severe complications. These infections are often difficult to diagnose. Knowledge of the clinical, neuroimaging, and laboratory features of ROCM in patients with CAM and the possible risk factors associated with stroke in these patients can be helpful for better management of these patients. Lack of significant differences in conventional risk factors of ischemic stroke between patients with/without ischemic stroke suggests that the core mechanism of stroke in ROCM is different from stroke in other typical settings. Stroke in ROCM can be caused by invasion of vessels, local spread to nearby arteries, and hematogenous spread [14]. ROCM is an invasive disease with predilection for the vessels which pass through the nasal mucosa and paranasal sinuses. It can reach the cavernous sinuses either directly via the walls of the sphenoid sinus or via the pterygopalatine fossa [15]. Mucormycosis extension into the cavernous sinus with resultant septic thrombophlebitis and subsequent reactive vasospastic narrowing or frank invasion of ICA may result in occlusion or significant narrowing of the ICA, which in turn can cause ischemic stroke [6, 14]. Involvement of the basilar artery by further fungal invasion of the skull base may cause stroke in the posterior circulation [16]. Watershed infarct was the most common pattern of ischemic stroke among our patients (73%). Watershed infarction can be defined as the lack of blood supply in junctional areas between vascular territories of the major cerebral arteries and is most commonly seen in patients with significant arterial stenosis or prolonged severe hypotension. In our study, among 17 patients with AIS, 9 of 17 had ICA involvement and watershed infarct was the pattern of AIS in all of these 9 patients. In other patients without evidence of ICA involvement, half of them also had watershed infarct. These findings suggest that structural narrowing of carotid from the outside is not the only mechanism for stroke in RCOM patients and other mechanisms of AIS such as microscopic invasion of ICA could be considered in ROCM patients as well [9]. To best of our knowledge, there is no study that details the structural qualitative and quantitative imaging characteristics of cavernous sinus, ICA, SOV, paranasal sinuses and orbit in patients with AIS in the setting of ROCM associated with COVID-19. Although cavernous sinus involvement is more common among patients with AIS, other imaging features including size of cavernous sinus showed no significant differences between patients with and without AIS. These findings suggested that the key factor for ischemic stroke is the effect of cavernous sinus involvement on the ICA and that the mere size or pattern of cavernous sinus involvement may not have the determining role in ischemic stroke. Among the various venous connections of the cavernous sinus, the superior and inferior ophthalmic veins are important afferent tributaries that drain into the cavernous sinus. Presence of cavernous sinus thrombosis can impair the drainage of blood from the SOV, leading to engorgement or thrombosis of the SOV, thus it can be considered as an indirect sign of cavernous sinus involvement [12]. In line with this fact, our study showed that diameter of SOV in patients with stroke is larger than patients without stroke. Also, increased signal intensity of SOV on spin echo MRI sequence was more prevalent in patients with stroke. This finding can be due to blood stasis and possible thrombosis in SOV that may be sign of more extensive involvement of cavernous sinus that result in stroke [12, 17]. The cavernous segment of the ICA is an important component that can be significantly affected in the presence of cavernous sinus lesions. Involvement of ICA could be due to external compression (secondary to cavernous sinus involvement), intraluminal obstruction, and angioinvasion of fungus [12]. In our study, radiological evidence of ICA involvement showed significant differences regarding ICA lumen diameter between patients with and without stroke. Accordingly, comparing ICA diameter of involved side vs. uninvolved side in each group revealed a significant difference only in patients with stroke, which signify the importance of ICA evaluation in these patients. An arterial clot can alter signal intensity that would be different from the normal dark flow void (because of fast flowing blood) [18]. Thus, increased signal intensity of ICA on spin echo MRI (lack of flow void) can be indicative of slow flow or an ICA occlusion [18]. Similarly, our results showed that increased signal intensity of ICA on MRI is more common among patients with AIS. Regarding paranasal sinus involvement, our study showed that the ethmoid and frontal sinuses are the most commonly affected sinuses, which is in accordance with the study of Dubey et al. [9]. In addition, we presume that the higher involvement of frontal and ethmoid sinuses among stroke patients could be because of more extensive fungal spread or the proximity of the ethmoid sinus to ICA [19]. A hyperinflammatory state from cytokine storm followed by a prothrombotic state can have complications and cause both venous and arterial thromboembolisms. Thus, this mechanism could also be considered as an additional reason for higher incidence of ischemic stroke in RCOM patients [20]. In accordance with this hypothesis, in our study we found significant differences between patients with and without AIS in terms of laboratory findings that indicate systemic inflammation [21, 22], including higher level of D‑dimer levels, higher white blood cell counts, higher neutrophil to lymphocyte ratios (NLR), lower lymphocyte percentages, and higher BUN/creatinine ratios. We acknowledge that this study has some limitations that can be addressed in future studies. First, there was a lack of uniform follow-up imaging in performing longitudinal analysis of the imaging findings over the time to find possible prognostic factors of ischemic stroke in ROCM patients associated with CAM. Second, a brain MRI was not performed for a small number of our patients and only brain CT had been performed. Third, we were not able to formally evaluate the severity of COVID-19 manifestations as a possible factor affecting patient prognosis. Nevertheless, we believe that this paper contributes to the early literature and raises awareness of this potentially devastating disorder. Conclusion Ischemic stroke is a common consequence of rhino-orbito-cerebral mucormycosis (ROCM) associated with COVID-19 infection. Stroke in the setting of ROCM is not associated with conventional ischemic stroke risk factors. Neuroimaging features including qualitative and quantitative evaluation of the cavernous sinus, internal carotid artery (ICA) and superior ophthalmic vein (SOV) are useful in better understanding the mechanism and risk profile of ischemic stroke in patients with ROCM. The most common pattern of stroke in these patients is watershed infarction. Laboratory findings including higher levels of inflammatory biomarkers can also be associated with stroke in this group of patients. Meticulous and timely evaluation of COVID-19 associated mucormycosis (CAM) patients in terms of neuraxis involvement by having a high index of suspicion and appropriate neuroimaging is crucial to prevent the potentially debilitating consequences. Funding No funding was received for conducting this study. Author Contribution MAN analyzed the data, monitored data collection, and drafted and revised the paper. AZ analyzed the data, monitored data collection, and drafted and revised the paper. NR designed data collection tools, monitored data collection, and drafted and revised the paper. HP prepared the statistical analysis plan, monitored data collection, and revised the draft paper. MK initiated the collaborative project, designed data collection tools, monitored data collection, analyzed the data, and drafted and revised the paper. MRN monitored data collection and revised the draft paper. SH designed data collection tools, monitored data collection, wrote the statistical analysis plan, analyzed the data, and drafted and revised the paper. FKH monitored data collection and revised the draft paper. MS monitored data collection and revised the draft paper. AV supervised the data collection, analyzed the data, and revised the draft paper. Declarations Conflict of interest M.A. Najafi, A. Zandifar, N. Ramezani, H. Paidari, M. Kheradmand, B. Ansari, M.R. Najafi, S. Hajiahmadi, F. Khorvash, M. Saadatnia and A. Vossough declare that they have no competing interests. Ethical standards This study was performed in line with the principles of the Declaration of Helsinki. This study was approved by the Institutional Review Board of the Isfahan University of Medical Sciences (IR.ARI.MUI.REC.1400.025). Verbal consent to participate in this study was obtained from all patients or their surrogates by one of the co-investigators. The authors Mohammad Amin Najafi and Alireza Zandifar contributed equally to the manuscript. ==== Refs References 1. Prakash H Chakrabarti A Global epidemiology of mucormycosis J Fungi 2019 5 1 26 10.3390/jof5010026 2. Ahmadikia K Hashemi SJ Khodavaisy S Getso MI Alijani N Badali H Mirhendi H Salehi M Tabari A Mohammadi Ardehali M Kord M Roilides E Rezaie S The double-edged sword of systemic corticosteroid therapy in viral pneumonia: a case report and comparative review of influenza-associated mucormycosis versus COVID-19 associated mucormycosis Mycoses 2021 64 8 798 808 10.1111/myc.13256 33590551 3. Roden MM Zaoutis TE Buchanan WL Knudsen TA Sarkisova TA Schaufele RL Sein M Sein T Chiou CC Chu JH Kontoyiannis DP Walsh TJ Epidemiology and outcome of zygomycosis: a&nbsp;review of 929 reported cases Clin Infect Dis 2005 41 5 634 653 10.1086/432579 16080086 4. Skiada A Pavleas I Drogari-Apiranthitou M Epidemiology and diagnosis of mucormycosis: an update J Fungi 2020 6 4 265 10.3390/jof6040265 5. Singh AK Singh R Joshi SR Misra A Mucormycosis in COVID-19: a&nbsp;systematic review of cases reported worldwide and in India Diabetes Metab Syndr Clin Res Rev 2021 15 4 102146 10.1016/j.dsx.2021.05.019 6. Srivastava PM Vishnu VY Pandit AK Mucormycosis epidemic and stroke in India during the COVID-19 pandemic Stroke 2021 52 10 e622 e623 34525840 7. Kulkarni R Pujari S Gupta D Advani S Soni A Duberkar D Dhonde P Batra D Bilala S Agrawal P Aurangabadkar K Rhino-orbito-cerebral mycosis and COVID-19: From bad to worse? Ann Indian Acad Neurol 2022 25 1 68 10.4103/aian.aian_463_21 35342244 8. Kulkarni R Pujari SS Gupta D Ojha P Dhamne M Bolegave V Dhonde P Soni A Adwani S Diwan A Duberkar D Cerebrovascular involvement in mucormycosis in COVID-19 pandemic J Stroke Cerebrovasc Dis 2022 31 2 106231 10.1016/j.jstrokecerebrovasdis.2021.106231 34890962 9. Dubey S Mukherjee D Sarkar P Mukhopadhyay P Barman D Bandopadhyay M Pandit A Sengupta A Das S Ghosh S Adhikari S COVID-19 associated rhino-orbital-cerebral mucormycosis: an observational study from Eastern India, with special emphasis on neurological spectrum Diabetes Metab Syndr Clin Res Rev 2021 15 5 102267 10.1016/j.dsx.2021.102267 10. Mazzai L Anglani M Giraudo C Martucci M Cester G Causin F Imaging features of rhinocerebral mucormycosis: from onset to vascular complications Acta Radiol 2022 63 2 232 244 10.1177/0284185120988828 33615823 11. Hoenigl M Seidel D Carvalho A Rudramurthy SM Arastehfar A Gangneux JP Nasir N Bonifaz A Araiza J Klimko N Serris A Lagrou K Meis JF Cornely OA Perfect JR White PL Chakrabarti A ECMM and ISHAM collaborators. The emergence of COVID-19 associated mucormycosis: a review of cases from 18 countries Lancet Microbe 2022 3 e543 52 10.1016/S2666-5247(21)00237-8 35098179 12. Bhatia H Kaur R Bedi R MR imaging of cavernous sinus thrombosis European J Radiol Open 2020 7 100226 10.1016/j.ejro.2020.100226 32154332 13. Adams HP Jr Bendixen BH Kappelle LJ Biller J Love BB Gordon DL Marsh EE 3rd Classification of subtype of acute ischemic stroke. Definitions for use in a&nbsp;multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment Stroke 1993 24 1 35 41 10.1161/01.STR.24.1.35 7678184 14. Chikley A Ben-Ami R Kontoyiannis DP Mucormycosis of the central nervous system J Fungi 2019 5 3 59 10.3390/jof5030059 15. Hosseini SM Borghei P Rhinocerebral mucormycosis: pathways of spread Eur Arch Otorhinolaryngol 2005 262 11 932 938 10.1007/s00405-005-0919-0 15891927 16. Little JS Cheng MP Hsu L Corrales CE Marty FM Invasive Fungal Carotiditis: A Rare Manifestation of Cranial Invasive Fungal Disease: Case Series and Systematic Review of the Literature Open Forum Infect Dis 2019 6 ofz392 10.1093/ofid/ofz392 31660355 17. Sotoudeh H Shafaat O Aboueldahab N Vaphiades M Sotoudeh E Bernstock J Superior ophthalmic vein thrombosis: What radiologist and clinician must know? Eur J Radiol Open 2019 6 258 264 10.1016/j.ejro.2019.07.002 31641683 18. Makkat S Vandevenne JE Verswijvel G Ijsewijn T Grieten M Palmers Y De Schepper AM Parizel PM Signs of acute stroke seen on fluid-attenuated inversion recovery MR imaging Am J Roentgenol 2002 179 1 237 243 10.2214/ajr.179.1.1790237 12076944 19. Lee WH Kim JW Lim JS Kong IG Choi HG Chronic rhinosinusitis increases the risk of hemorrhagic and ischemic stroke: A longitudinal follow-up study using a&nbsp;national sample cohort PLoS ONE 2018 13 3 e0193886 10.1371/journal.pone.0193886 29494700 20. Klein DE Libman R Kirsch C Arora R Cerebral venous thrombosis: atypical presentation of COVID-19 in the young J Stroke Cerebrovasc Dis 2020 29 8 104989 10.1016/j.jstrokecerebrovasdis.2020.104989 32689590 21. Ok F Erdogan O Durmus E Carkci S Canik A Predictive values of blood urea nitrogen/creatinine ratio and other routine blood parameters on disease severity and survival of COVID-19&nbsp;patients J Med Virol 2021 93 2 786 793 10.1002/jmv.26300 32662893 22. Liu Q Wang Y Zhao X Wang L Liu F Wang T Ye D Lv Y Diagnostic performance of a blood urea nitrogen to creatinine ratio-based nomogram for predicting in-hospital mortality in COVID-19 patients Risk Manag Healthc Policy 2021 14 117 10.2147/RMHP.S278365 33469395

36520187

PMC9753029

NO-CC CODE

2022-12-16 23:25:27

no

Clin Neuroradiol. 2022 Dec 15;:1-9

utf-8

Clin Neuroradiol

10.1007/s00062-022-01238-y

oa_other

==== Front Atmos Ocean Opt Atmospheric and Oceanic Optics 1024-8560 2070-0393 Pleiades Publishing Moscow 4403 10.1134/S1024856022060033 Atmospheric Radiation, Optical Weather, and Climate Tropospheric Ozone Concentration on the Territory of Russia in 2021 Andreev V. V. vvandreev@mail.ru 1 Arshinov M. Yu. 2 Belan B. D. bbd@iao.ru 2 Belan S. B. 2 Davydov D. K. 2 Demin V. I. 3 Dudorova N. V. 2 Elansky N. F. 4 Zhamsueva G. S. 5 Zayakhanov A. S. 5 Ivlev G. A. 2 Kozlov A. V. 2 Konovaltseva L. V. 1 Kotel’nikov S. N. 6 Kuznetsova I. N. 7 Lapchenko V. A. 8 Lezina E. A. 9 Obolkin V. A. 10 Postylyakov O. V. 4 Potemkin V. L. 10 Savkin D. E. 2 Senik I. A. 4 Stepanov E. V. 6 Tolmachev G. N. 2 Fofonov A. V. 2 Khodzher T. V. 10 Chelibanov I. V. 11 Chelibanov V. P. 11 Shirotov V. V. 12 Shukurov K. A. 4 1 grid.77642.30 0000 0004 0645 517X Peoples’ Friendship University of Russia, 117198 Moscow, Russia 2 grid.435125.4 0000 0004 0638 2644 V. E. Zuev Institute of Atmospheric Optics, Siberian Branch, Russian Academy of Sciences, 634055 Tomsk, Russia 3 grid.467115.0 0000 0004 0577 251X Polar Geophysical Institute, Russian Academy of Sciences, 184209 Apatity, Russia 4 grid.459329.0 0000 0004 0485 5946 Obukhov Institute of Atmospheric Physics, Russian Academy of Sciences, 119017 Moscow, Russia 5 grid.415877.8 0000 0001 2254 1834 Institute of Physical Material Science, Siberian Branch, Russian Academy of Sciences, 670047 Ulan-Ude, Russia 6 grid.424964.9 0000 0004 0637 9699 Prokhorov General Physics Institute, Russian Academy of Sciences, 111933 Moscow, Russia 7 Hydrometeorological Center of Russia, 123242 Moscow, Russia 8 grid.4886.2 0000 0001 2192 9124 Vyazemsky Karadag Scientific Station – Nature Reserve of Russian Academy of Sciences – Branch of Kovalevsky Institute of Biology of Southern Seas of Russian Academy of Sciences, 298188 Feodosia, Russia 9 State Nature Organization Mosecomonitoring, 119019 Moscow, Russia 10 grid.425246.3 0000 0004 0440 2197 Limnological Institute, Siberian Branch, Russian Academy of Sciences, 664033 Irkutsk, Russia 11 Instrument-Making Enterprise OPTEC, 199178 St. Petersburg, Russia 12 Typhoon Scientific and Production Association, 249038 Obninsk, Kaluga oblast Russia 15 12 2022 2022 35 6 741757 7 4 2022 28 4 2022 16 5 2022 © Pleiades Publishing, Ltd. 2022, ISSN 1024-8560, Atmospheric and Oceanic Optics, 2022, Vol. 35, No. 6, pp. 741–757. © Pleiades Publishing, Ltd., 2022.Russian Text © The Author(s), 2022, published in Optika Atmosfery i Okeana. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Ozone is one of the most toxic admixtures in the troposphere. Therefore, it is among the main pollutants and its concentration is monitored. This work represents an overview of continuous measurements of the ozone content in the troposphere on the territory of Russia throughout 2021 carried out on an initiative of scientific and educational institutions at 17 stations in different Russian regions. The monitoring results showed that the daily average ozone concentration exceeded the MPCd.a level during a major part of the year at all observation sites, and by a factor of two or even three at a number of stations. At six stations, concentrations in excess of the maximum permissible one-time concentration MPCm.o were recorded. This requires a more comprehensive analysis of the composition and concentration of ozone precurcors and the development of measures to reduce their emission into the atmosphere. Keywords: atmosphere air concentration ozone maximum permissible concentration surface layer troposphere issue-copyright-statement© Pleiades Publishing, Ltd. 2022 ==== Body pmcINTRODUCTION The urgency of studying the tropospheric ozone stems from its physicochemical properties [1–5] and its effect on biological entities and environmental structures after its content increases in air. As biological and medical studies have shown, ozone in the troposphere is a virulent poison exhibiting, in addition to the general toxic effect, such properties as mutagenic and carcinogenic potentialities and radiomimetic effect (an action on blood similar to ionizing radiation) [2–4]. Based on [6], 30-min inhalation of ozone at the concentration 0.8 mg/L is equivalent to a 100-R dose. Ozone is even more toxic than such a well-known poison as hydrocyanic acid. Therefore, it is classified as a class one hazardous substance in regulatory documentation [7]. In high concentrations, ozone strongly inhibits the activity of plant life. Plant response to the increased ozone concentration is reduced productivity and even death in some cases. Calculations of American scientists [8] have shown that the economic losses from reduced crop productivity are from 1.9 to 3.3 trillion dollars yearly in the United States. Analogous losses for a part of southeastern Asia are 68 billion dollars [9]. In addition to crop productivity reduction, ozone also decreases the uptake of carbon dioxide by vegetation, which can lead to enhancement of the Earth’s radiative forcing [10, 11]. In addition to the effects mentioned above, ozone is the strongest oxidant capable of destroying rubber and caoutchouc and oxidizing many metals, even from the platinum group [12–16]. With its long (from few days to few months) lifetime in the atmosphere and strong solar radiation absorption, tropospheric ozone plays an important role in the greenhouse effect. Estimates [17] indicate that it contributes more than 8% of the total air heating due to absorption of solar radiation by greenhouse gases. Later estimates show that this contribution may be even larger. This variety of possible adverse consequences from an increasing concentration of tropospheric ozone for both human beings and the environment call for closer attention to the trends in variations in its content in surface air. This gas is considered the number one air pollutant in all developed countries. The authors of work [18] mentioned that there were over 10 thousand stations for monitoring ozone and its precursors in Europe as early as 2003. Most important is that the information is made available to the population and used in decision making by governing bodies. The United States and Europe have already succeeded in reducing ozone concentrations in air. For instance, based on data from 119 stations in Great Britain, the measures taken resulted in a reduction of the concentration of surface ozone from 1980 to 2019 by a factor of 2–6 depending on the region [19]. The United States managed to reduce the emissions of ozone precursors by a factor of two [20, 21]. China undertook similar efforts; however, a significant reduction of the emissions of ozone precursors could be achieved only in certain industries [22–24]. The former Soviet Union and present-day Russia did not pay due attention to monitoring and measures for reducing the ozone content. Rosgydromet, entrusted with the responsibility for monitoring ozone content, is proceeding with the technological modernization of the observational network, and so far is measuring surface ozone in just a few large and industrial cities. The two biggest megalopolises in Russia, St. Petersburg and Moscow, have competitive systems for monitoring surface ozone and other pollutants. An ecological monitoring network of State Nature Organization Mosecomonitoring has been operated in Moscow since 2002, which is a specially authorized governmental ecological monitoring organization in Moscow [25]. The surface ozone concentration is monitored at 17 automatic air pollution control stations (AAPCS) hourly and around-the-clock. The 20‑min averages are stored in a database. The Mosecomonitoring network stations carry out measurements using gas analyzers of three types based on ultraviolet photometry: Casella Monitor ME 9810B, Environnement S.A. O3 42M, and HORIBA Ltd. APXA-370 model APOA-370, and a OPSIS AB AR500 analyzer, based on differential optical absorption spectroscopy (DOAS). The instruments are included in the State Register of Measuring Instruments and certified by the State Meteorological Service. Analytical materials on the state of the environment in Moscow are annually reported [26]. However, data on the content of surface ozone on the territory of Russia are still not provided in state reports [27, 28]. In the rest of Russia, the ozone observations are carried out on an initiative basis, mainly by scientific or higher-education institutions. The purpose of this review is to inform the scientific community about the ozone content in the surface air layer in 2021, and about the causes for its variations and the compliance of ozone concentrations recorded at different monitoring sites to national hygienic standards [7]. In this review, we used the data obtained by the coauthors at 17 sites in Russia, differing by their geographic and climatic characteristics, as well as by the anthropogenic load on the environment. The spatiotemporal variations in the surface ozone on the territory of Moscow are analyzed using averaged measurements at Mosecomonitoring AAPCS of two types: seven urban AAPCSs and four traffic AAPCSs (https://mosecom.mos.ru/vozdux/); it is noteworthy that the maximal ozone concentrations are represented by highest hourly average concentrations recorded at all AAPCSs. It should be noted that, like previous half-year reviews in 2020 [29, 30], 2021 coincided with the period of coronavirus pandemic and, as such, can reflect the lockdown results. The meta-analysis carried out in [31–33] using monitoring at tens of stations around the globe showed that reduction of emissions of the main admixtures was usually accompanied by the growth of ozone concentration in the surface air layer. Interestingly, ozone concentrations decreased in the free troposphere during the pandemic [34, 35]. There was no aim of elucidating the lockdown consequences in the review, because this requires the data for previous years unavailable at a number of stations. Here, the changes in the ozone concentrations in the free troposphere are verified using results from aircraft sensing. 1 NEW STATIONS AND INSTRUMENTS USED The total set of stations and the instrumentation installed at these stations, as well as the operational modes and calibrations, were listed in previous reviews [29, 30]. In 2021, we resumed measurements at stations Slyudyanka and Tarusa, opened new OPTEC stations in Karelia and Boyarsky settlement in Buryatia. In this section, we describe these stations. The atmospheric monitoring station Listvyanka (51°50′48″ N, 104°53′58″ E, 670 m ASL) is located in Irkutsk oblast on the southwestern coast of Lake Baikal, in a region of the source of the Angara River, at the top of a coastal hill (200 m above the lake level) on the territory of the Astrophysical Observatory of the Institute of Solar-Terrestrial Physics, Siberian Branch, Russian Academy of Sciences. The nearest settlement Listvyanka is 2 km off the station on the coast of the lake. The location of the station at the top of the hill makes it possible to escape the effect of local sources of atmospheric pollution (settlement, motor vehicles) and to track the regional transports of pollutants, primarily from the direction of Irkutsk, Angarsk, and Shelekhov. The complex of automatic gas analyzers year-round monitors the presence in the atmosphere of different admixtures, including ozone, a new optical ozone analyzer F-105 (OPTEC, St. Petersburg, Russia) was installed at the station in February 2021. Since 2001, the station had become a participant of the International Program “Acid Deposition Monitoring Network in East Asia” (EANET); and the data from the station are planned to be made available on the Internet. Under the auspices of the Peoples’ Friendship University of Russia (RUDN) and the Prokhorov General Physics Institute, Russian Academy of Sciences (GPI RAS), Russian Academy of Sciences, an automatic station for monitoring surface ozone and the main meteorological parameters started operating in summer 2021 in Tarusa, Kaluga oblast. The station is located on the territory of Tarusa branch of GPI RAS (54°43′36″ N, 37°10′40″ E, 128 m ASL) situated at the center of the city in the residential building zone 350 m away from the coast of the Oka River. Tarusa is 110 km south of Moscow on the high bank of Oka bend, surrounded by pine forests in the north. Its population is slightly more than nine thousand residents. Tarusa and its surroundings have no industrial plants and, as such, are considered as one of the resort regions in the far Moscow oblast. The main local sources of anthropogenic pollution of the atmosphere are motor vehicles and municipal services. The nearest busy motorway Serpukhov–Kaluga is 1.5 km away from the center of Tarusa. The distances to the nearest bigger cities are: ∼30 km to Serpukhov, ∼70 km to Kaluga, and ∼80 km to Tula. The monitoring station is equipped with a chemiluminescent ozone analyzer 3.02P-A (OPTEC) with a sensitivity of ∼1 μg/m3. Sampling is carried out via Teflon pipes at an altitude of 5 m above the Earth’s surface. The measurements are performed in the continuous long-term monitoring mode. Current values of the parameters measured are recorded once a minute, followed by their averaging over 20 min and storing the result in the database. The atmospheric monitoring station OPTEC-Karelia is in the settlement Voloma (63°44′41″ N, 31°56′33″ E, 185 m ASL) in the northern part of the Republic of Karelia, about 80 km away from the border with Finland (Russian Far North). The settlement is in a depression surrounded by hills up to 100 m in height. Outside the settlement there are large forest massifs as well as hundreds of small lakes. The average air temperature is −28°С in winter and +25°С in summer. The lowest observed temperature reaches −50°С in winter and the highest temperature is +40°С in summer. The wind direction at the location of the station is predominantly S–W and the wind speed is 2–3 m/s. The annually average pressure does not exceed 738–740 mmHg. The average height of snow cover is 1.2–1.5 m. There are no big industrial plants in the region; the nearest, Segezha Pulp and Paper Mill (PPM) is 110 km away, and the Kostomuksha ore mining and processing enterprise is 114 km away. Quite rarely, a bad smell from cellulose processing at the Segezha PPM is detected when the wind is strong. There is small industrial plant for wood production and drying in the settlement. Temperature inversions in the atmosphere are observed mainly in winter owing to the specific orography at the location of the settlement and predominant type of anticyclonic weather. This leads to intense accumulation of pollutants (СО and CO2) in the surface air layer. For instance, the average concentrations are 200 μg/m3 for CO and 400 mg/m3 for CO2 during summer, and 800 μg/m3 for СО and 1200 mg/m3 for CO2 during winter. The accumulation of the carbon oxides in the surface air layer is likely due to the specific properties of the fuel used in the settlement: wood-burning stoves are used in the local boiler and in private houses. The OPTEC-Karelia station has operated since May 2021 in the pilot mode; it comprises the channels of measuring the concentrations of О3 (ozone), СО (carbon monoxide), CO2 (carbon dioxide), and 1Δg(O2) singlet oxygen. The surface concentrations of ozone and singlet oxygen are measured using domestic solid-state chemiluminescent analyzers mod. 3.02P-A and 102-А, respectively. The limiting values of the main error of the 3.02P-A analyzer measurements are ±20% for the range of 0–30 μg/m3 and ±20% for the range of 30–50 μg/m3. The limiting values of the main error of the 102-A analyzer measurements are ±20% for the range 0–10 μg/m3 and ±20% for the range 0–200 μg/m3. Boyarsky station is located on the southeastern coast of Lake Baikal, 160 km away from Ulan-Ude. This region is characterized by large temperature contrasts between the lake and adjoining territory, intensifying due to the closed position of Lake Baikal, surrounded in all directions by mountain ridges. The temperature gradient between the lake depression and adjoining dry hollows, reaching 20°С and more, is one of the main factors of formation and development of intrahollow circulation and its propagation into the lake basin, often favoring accumulation of atmospheric pollutants. Boyarsk village can be considered to experience weak anthropogenic impact: certain effects can be due to small industrial centers (Babushkin (22 km), Kamensk settlement (50 km), Selenginsk settlement (60 km), and others). A mixed forest (birch, pine, and cedar) lies in the immediate vicinity of the station. The concentration of surface ozone was measured using a 3.02 P-A chemiluminescent gas analyzer. The instrument was calibrated using a Mod. 8500 Monitor Labs calibrator. The observations at Boyarsky station were episodic, in the period of expeditionary works (April 13–18 and July 21–August 20, 2021). 2 MEASUREMENT RESULTS 2.1 Annual Average Data Data on annual average ozone concentrations in the surface air layer at all stations that conducted measurements in 2021 are presented in Fig. 1. Fig. 1. Annual average ozone concentrations at Russian stations. On the one hand, Fig. 1 shows that the annual average ozone concentration is higher than 30 μg/m3 at all stations. This is even larger than the daily average maximum permissible concentration (MPCd.a), indicating that the norm is exceeded throughout the year [7]. On the other hand, with respect to the annual average concentration, the stations line up in quite an intricate order. The concentrations are the largest at Kislovodsk high-mountain station (KHMS), in Karadag, and in Listvyanka, i.e., at locations far removed from anthropogenic sources of ozone precursors. The concentrations are the smallest in Obninsk, its station subject to urban conditions, where ozone can be destroyed in emissions from plants and in vehicular exhausts. Both background (Vyatskiye Polyany and OPTEC-Karelia), and urban (RUDN) and suburban (Tropospheric Ozone Research (TOR)) stations become the members of the group with medium values of ozone concentrations. Stations in the group with minimal values line up in a similar order. Here, again, there are urban (OPTEC-N, -P, and -PR, Ulan-Ude, and Obninsk), suburban (Tarusa), and background (Fonovaya Observatory) stations. It also follows from Fig. 1 that there is no longitudinal or latitudinal dependence of the annual averages, possibly due to the contribution from local sources of ozone precursors and anthropogenic factors. Also, there can be a contribution from long-term interannual variations in ozone concentration, when the annual average concentration can vary by as much as a factor of four [36]. A separate study is required to answer this question. 2.2 Annual Behavior of Ozone Concentration Monthly average data are used to consider the variations in ozone concentration at stations that operated throughout the year (Fig. 2). Fig. 2. Annual behavior of ozone concentration at Russian stations based on monthly average data. From Fig. 2, it can be seen that the main concentration maximum in annual behavior is observed in spring at seven stations (TOR, Fonovaya, Listvyanka, Obninsk, Apatity, Vyatskiye Polyany, and Ulan-Ude), usually classified as background or suburban. The springtime concentration maximum was also recorded at a number of other stations [37–39]. Based on 20‑year monitoring, the springtime maximum in Tomsk was observed in 88.5% of cases [40]. Its specific feature is that it does not coincide in time with the maximum of incoming solar radiation. Considering that ozone is photochemically produced, this is not explainable yet [41]. Based on 11-year monitoring, the springtime maximum in Ulan-Ude was observed in 27.3% of cases [42]. Considering that organic gases may account for more than a half of the initial volume of ozone precursors in background regions [43, 44], a more probable process seems to be that associated with the springtime intensification of vegetation activity of plants delivering organic gases, i.e., ozone precursors [45, 46]. For stations located in St. Petersburg and Moscow (OPTEC-N, OPTEC-P, and RUDN), the main maximum of ozone concentration in 2021 was observed in July under the conditions of a blocking anticyclone and anomalously hot dry weather, associated with the large-scale circulation. The climatic annual maximum of surface ozone in midlatitudes and, in particular, in Tomsk and Moscow, is usually observed in spring (April–May). This was recently illustrated through the analysis of data from a continuous monitoring in 2005–2020 in [47]. In megalopolises in the southern latitudes, the ozone maximum occurs in July because of photochemical ozone production from anthropogenic emissions [48, 49]. There are two concentration maxima at the OPTEC-PR station. The first (not primary) maximum is recorded in the spring–summer period, and the second maximum is in December. This fact is difficult to explain. Possibly, it is associated with any local sources of ozone precursor gases, because no concentration increases were recorded in that period at the other St. Petersburg stations. An extraordinary annual behavior is observed at KHMS (Fig. 2). The increased monthly average values are usually observed at KHMS in spring (March–May) and summer (July–August) and do not coincide in time with the maximum of the sunshine duration [50]. This regular effect has been observed and confirmed since the beginning of measurements of surface ozone at KHMS in 1989. The springtime local maximum is also manifested at other high-mountain stations such as Jungfraujoch (JFJ), where the stratosphere-troposphere exchange and mountain-valley circulation also influence the ozone regime, in addition to photochemical processes. The ozone concentration at KHMS is minimal in fall–winter. The absolute hourly average maximum in 2021 (140 μg/m3) was on July 19; and on August 13 and 18, the hourly averages reached 120 μg/m3. Under high-mountain conditions, that high ozone concentrations at KHMS can be associated with the stratospheric intrusions to the free troposphere, followed by mixing in the zone of orographic disturbances and, in particular, during foehn formation [51]. These events are generally short-term, lasting from one to few hours. The increased concentrations can also be associated with ozone production in polluted air during long-range transport. The trajectory analysis of air masses coming to KHMS was carried out to consider the contribution of long-range transport to the extreme values observed. The method for calculating the 7-day back trajectories was described in [30]. The 2021 measurements contain gaps, for technical reasons; therefore, fewer trajectories than in 2020 (∼17 000) were simulated. In contrast to urban conditions, where decreases in ozone concentration down to very low values signify strong pollution by nitrogen oxides, the anomalously low ozone concentrations at KHMS, which is located in a clean terrain and above the atmospheric boundary layer, are associated not so much with the long-range transport, but rather with dry deposition onto the Earth’s surface. This process is most active in a slowly moving air mass, where the contact of the analyzed air with the vegetation-covered surface is longest. The effect of pollutants transported from lower atmospheric levels from the direction of Kislovodsk (750–850 m ASL) on days with conditions favoring the development of the mountain-valley circulation, was shown in [50] to lead, on the contrary, to an increased daytime maximum, though not by a significant amount. Moreover, fogs favor the reduction of ozone concentration. Therefore, we excluded from analysis the back trajectories for days with a high (larger than 80%) humidity at the trajectory end point at KHMS. As a result, ∼13 000 trajectories remained in the dataset. This dataset was processed to select two sets of trajectories corresponding to extreme negative and extreme positive ozone anomalies, respectively, in the first and last deciles of the ozone anomaly distribution function, calculated with respect to the second-order polynomial fit. For extreme values of the ozone concentration anomalies of both signs, we retrieved the fields of the probability P (%) of air particle transport to KHMS in spatial grid cells 1° × 1° in size (Fig. 3). Fig. 3. Transport probability of air particles (elementary air masses) associated with 10% of the lowest (upper panel) and 10% of the highest (lower panel) anomalies of surface ozone concentrations at KHMS in 2021 over different territories. On the whole, our results are consistent with the 2020 observations: extremely high anomalies of the surface ozone at KHMS in 2020 were associated with the southerly air transport and extremely low ozone was due to northwesterly transport. The extremely high anomalies in 2021 as compared to 2020 had an extra contribution from the southeastern transport direction: air particles associated with extremely high ozone anomalies in 2021 most probably moved not only over Turkey, as they did in 2020, but also over Azerbaijan, the Southern Caspian Sea, Turkmenistan, and Uzbekistan. Trajectories associated with extremely low surface concentrations in 2021, as in 2020, most probably passed over Krasnodar Krai, the Azov Sea, and the Ukrainian Azov region. That is, in the structure of the set of trajectories, we can distinctly single out two clusters: one associated with air masses coming from the Middle East, and the other (eastern) has a structure close to zonal. The trajectories of both types pass through the regions of intense oil and gas production and processing. If we consider the seasonal distribution of extreme ozone values, it is seen that the maximal concentrations are predominant in the spring–summer season, when there is a stable easterly transport associated with the Middle Asian anticyclone. Under the conditions of high temperatures and solar irradiance, the oxidation of volatile organic compounds in the plume from plants of the oil and gas industry leads to active ozone production and a stable increase in ozone concentration at KHMS. 2.3 Dynamics of Daily Average Concentrations One of the normalized characteristics of ozone content in air is the daily average ozone concentration, which should not exceed 30 μg/m3 [7]. The data on this characteristic are presented in Fig. 4. Fig. 4. Daily average ozone concentrations at Russian stations in 2021; horizontal lines indicate 1MPC (red), 2MPC (green), and 3MPC (lilac). It can be seen that the daily average MPC of ozone is exceeded at all observation sites for a major part of the year. We give below the numerical information on this. The MPC is exceeded two- and even three-fold at a number of stations. At urban stations, there are some days or short periods, usually during fall and winter, when the daily average ozone concentration decreases to zero. This seems to be due to low photochemical formation rate of ozone in these periods and its destruction by pollutants from automobile exhausts. 2.4 Daily Maximum Concentrations Still another normalized characteristic is the hourly maximal ozone concentration. Based on [7], it should not exceed 160 μg/m3. These data are summarized in Fig. 5. Fig. 5. Maximal hourly average ozone concentrations; red horizontal lines indicate MPCm.o. From Fig. 5, it follows that the one-time maximum MPC was not exceeded at 9 out of 14 stations in 2021. Four stations recorded a onefold excess over MPCm.o. A threefold excess was recorded in Moscow at RUDN station, the pollution of which can be considered to be as strong as photochemical smog. Thus, the regime of surface ozone in the Moscow region in 2021 differed from previous years in the following. First, in the concentrations largest on record and not observed before. Second, in the number of ozone episodes with characteristics exceeding Russian standards. Third, in the development of a summertime maximum that became the primary maximum in the annual behavior of surface ozone [52–54]. This markedly differs from data presented in previous reviews [29, 30]. In 2020, the weather in Moscow, which was rainier and colder than usual, was accompanied by decreased ozone concentrations, a poorly defined springtime maximum, and the absence of a summertime maximum. The specific features of large-scale circulation brought about a positive anomaly of air temperature in the warm period in April–August 2021 in the central regions of Russia. Data available at https://meteoinfo.ru/ indicate that the weather was 1–2° warmer than usual in spring and 3–4° warmer in summer. In summer months, the temperature of surface air rose above +30° on 31 days, 15 of which had temperatures as high as 33–36°. Hot dry weather favors intense photochemical ozone production [55–57]. It was the anomalous weather conditions that created prerequisites for the occurrence of high levels of surface ozone during summer. The springtime maximum of surface ozone in Moscow was observed in April–May; the hourly concentrations of surface ozone at Mosecomonitoring AAPCS increased up to 130–150 μg/m3 on separate days. A prolonged springtime maximum in 2021 was determined by the specific features of its formation. The summer ozone episodes have no analogs in the entire history of the regular ozone observations in Moscow [47]. In 2002 and 2010, in periods when smoke from forest fires influenced the ozone level, the high concentrations developed in late July–early August under the impact of long-range transport of ozone and its precursors from remote sources (see, e.g., [54]). In 2021, the main factor of anomalous growth of the surface ozone concentration had been an intense photochemical ozone production in air, polluted by local sources, under the conditions of atmospheric circulation weakened in blocking anticyclones. The longest ozone episode with extreme surface concentrations was observed in the second half of June under the conditions of the highest UV irradiance of the year. The two-week heat wave in June (since June 14, the air temperature rose above +25° at afternoon hours) was accompanied by severe radiation inversions and weak transport in the lower atmospheric layers; it reached its peak in the last week of the month. As shown in Fig. 6, the surface ozone concentration at certain AAPCSs exceeded the Russian standard (MPCm.o) by a factor of 1.2–1.6 for eight days; and on the days with extremely high pollution, the excesses were a factor of 1.8 (July 13) and a factor of 2.2 (June 23). Fig. 6. Surface ozone concentration at certain Mosecomonitoring stations in June 2021. Using trajectory analysis, we found that ozone concentrations grew to extremely high levels in the air mass that circulated at the center of anticyclone over the Moscow agglomeration, thus favoring air loading by pollutants. The diurnal dynamics were maximal in this period. After the nighttime destruction/sink of ozone to 10–20 μg m−3, the concentrations rapidly grew up to 100–150 μg/m−3 and higher in the morning hours (from 08:00 to 12:00 LT). This process was accompanied by a rapid depletion of nitrogen oxides in air. At the same time, we note that the maximum increase in the surface ozone concentration in the June 23 episode coincided in time with the highest NO2 level characteristic of photochemical smogs [58–60]. The plumes of polluted air with a high ozone content propagated long distances away from Moscow; calculations using SILAM chemical transport model (https://www.ventusky.com/) indicated that ozone-rich air masses moved to the neighboring regions; in particular, in the episode with the maximal ozone level in Moscow on June 23 and 24, the plume of anthropogenic ozone was carried northeast of Moscow, i.e., toward Ivanovo, Vladimir, and partly Nizhny Novgorod oblasts (Fig. 7). Fig. 7. Surface ozone concentration on (a) June 23 and (b) 24, 2021, calculated from SILAM chemical transport model. In ozone episodes, the surface ozone field inside the megalopolis was characterized by high inhomogeneity: the difference in the maximal concentrations between urban and roadside stations reached 80–100 μg/m3 on certain days (Fig. 6). A prolonged ozone episode ended on June 28 after passage of a cold atmospheric front and change of air masses, as well as an inflow of clean air from the Baltic region. The next ozone episode occurred in Moscow in July. It emerged against the background of a new wave of 30-degree heat on July 7–18 and was interrupted only on July 11 due to a short-term air inflow from the north. As in the June episode, the maximal increase in the level of ozone in the surface air was observed in the period from 15:00 to 17:00 LT, when air was maximally heated, the humidity was ∼30%, and the weather in the region was calm in the lower atmospheric layers with a slowly moving anticyclone above. As was shown in [61], low air humidity also favors an increase in ozone concentration. It is important that the NO2 concentrations at night maximally increased (to 100–120 μg/m3) precisely on July 8 and 13, thereby ensuring a high chemical activity in the morning hours for a daylight ozone buildup, and for ozone destruction in evening. We can also note that the level of surface ozone turned out to be ∼50 μg/m3 lower on July 13 at a temperature of +35° than it was on July 8 at a temperature of +32°, for all other atmospheric parameters remaining almost the same. The July ozone episode ended on July 19 owing to change of synoptic process and arrival of a clean air mass from the Baltic region. In August, Moscow experienced another three short-term waves of 30-degree heat. However, no episodes of air pollution by ozone and its precursors occurred because of the absence of stagnant synoptic situations in those cases and a decrease in the level of UV radiation; only few urban stations recorded an atypical increase in ozone up to 0.8–0.9 MPCm.o. The maximal surface ozone concentrations (SOCs) in 2021 at the (SBEM) Karadag background ecological monitoring station were observed on May 8 and August 6, on clear-sky wind-free days (150 and 141 μg/m3, respectively); and minimal SOCs, on December 17 (6 μg/m3), when the humidity was higher than 90%. In the summer period, the daily maximum ozone concentrations were observed under southerly and southeasterly winds, signifying the transport toward the AAPCS location from the direction of the sea. For the first time in the observation period since 2006, a morning ozone maximum was recorded on May 8. Three SOC peaks were noted on that day: the first at 04:00 LT, the second at 08:00 LT, and the third at 20:00 LT (131, 150, and 116 μg/m3, respectively). Presumably, the nighttime peaks are associated with the stratospheric ozone source, as well as with intense vertical mixing between the surface layer and the free troposphere. Table 1 summarizes absolute maxima of ozone concentrations for each of the stations considered here. Table 1.   Absolute maxima of hourly averaged ozone concentrations in 2021 at Russian stations Station Concentration, μg/m3 Moscow, RUDN 490 Mosecomonitoring (northwest) 358 Obninsk 253 OPTEC-P 193 OPTEC-N 172 Vyatskiye Polyany 169 Boyarsky 151 Karadag 150 KHMS 134 TOR 129 Fonovaya 127 OPTEC-Karelia 126 OPTEC-PR 116 Ulan-Ude 116 Listvyanka 115 Apatity 104 Tarusa 92 3 OZONE DISTRIBUTION IN THE TROPOSPHERE From July 1997 to the present, V.E. Zuev Institute of Atmospheric Optics, Siberian Branch, Russian Academy of Sciences (IAO SB RAS), has carried out monthly flights on the Optik aircraft laboratory to determine the vertical distributions of the gaseous and aerosol compositions of the atmosphere. The flights were first performed on an An-30 aircraft [62], and then on a Tu-134 aircraft [63]. The aircraft laboratory flies over the region of Karakan pine forest 100 km southwest of Novosibirsk to eliminate the urban contribution. The aircraft takes off at noon, when there is the maximal photochemical ozone production lasting for 2 h. The altitude range is from 0 to 7 km. Not all flights were performed in 2021 due to the coronavirus pandemic. Since, as discussed before [64, 65], ozone measurements under nonbackground conditions are problematic, three ozonometers are simultaneously operated onboard the aircraft: a chemiluminescent ozone analyzer 3.02P and two ultraviolet Thermo Environmental Instruments (TEI) model 49C UV ozone analyzers (United States). The ozonometers are pre-flight calibrated using ozone generator GS-2. The measurements of the vertical ozone distribution, shown in Fig. 8, show that no ozone is produced in the atmospheric boundary layer during the cold period (March), when the Earth’s surface is covered by snow. Ozone is produced only in May. Thus, О3 was mainly transported from the stratosphere. Note that the vertical distribution in the middle troposphere was near-neutral indicating that the ozone flux was not very intense. Fig. 8. Vertical distribution of ozone concentration over Western Siberia in 2021. The measurements in Fig. 8 strongly differ from long-term sounding results, which were summarized previously [66] for this same region. It was noted that there was almost constant photochemical ozone production in the surface or boundary layer of the atmosphere, which in 2021 was recorded as late as April. Based on long-term measurements, Fig. 9 was plotted to identify and analyze the trends of variations in ozone concentration in the troposphere over Western Siberia noted in [34, 35]. This figure presents the annual average ozone concentrations at different altitudes. Fig. 9. Variations in ozone concentrations at different altitudes over Western Siberia in 2011–2021. From Fig. 9 it can be seen that the ozone concentration in the period of the coronavirus pandemic varies in opposite directions at the altitudes of the lower and upper troposphere. The variations are within the limits of long-term variability of the ozone concentration at these altitudes. 4 COMPLIANCE WITH HYGIENIC STANDARDS The Russian Federation accepted the following standards regarding ozone concentration in the surface air layer [7]: 30 μg/m3 for the daily average maximum permissible concentration (MPCd.a), 160 μg/m3 for the one-time maximum permissible concentration (MPCm.o), and 100 μg/m3 for a duration of 8 h for the maximum permissible concentration of harmful substance in the air of a work zone (MPCw.z). Table 2 summarizes the cases where the abovementioned MPCs were exceeded. Table 2.   Events of ozone concentrations above MPC in the surface air layer on the territory of Russia in the second half of 2021 Station MPCd.a, 30 μg/m3 MPCw.z, 100 μg/m3 MPCm.o, 160 μg/m3 1MPC, days/% 2MPC, days/% 3MPC, days/% OPTEC-PR 179/53.4 12/3.6 0/0 0 0 OPTEC-P 158/51.3 37/12.0 3/1.0 2 8 OPTEC-N 191/57.5 37/11.1 3/0.9 0 2 OPTEC-Karelia 117/90.0 30/23.1 0 1 0 SBEM Karadag 305/97.1 171/54.5 41/13.1               23 0 Obninsk 228/70.8 9/2.8 0/0 0 1 RUDN (Moscow) 212/58.9 115/31.9 55/15.3              145             402 KHMS 239/99.2 205/85.1 19/7.9 2 0 Vyatskiye Polyany 326/89.3 149/40.8 17/4.6 4 1 TOR station 306/84.8 51/14.1 0 0 0 Fonovaya 288/81.1 89/25.1 7/2.0 3 0 Listvyanka 327/98.2 157/48.0 2/0.6 0 0 Apatity 301/81.8 98/27.6 0 0 0 Tarusa 36/23.5 5/3.3 0 0 0 Ulan-Ude 191/57.2 36/10.8 0 0 0 Boyarsky 1/1.4 26/38.2 7/10.3 7 0 From Table 2 it can be seen that MPCd.a could be exceeded in all regions where ozone was monitored. If KHMS is disregarded as a peculiar station, the frequency of occurrence of the daily average concentrations 30 μg/m3 and larger is within 23.5–97.1%. The concentrations 60 μg/m3 (2MPC) and larger also occur in all regions at a frequency ranging from 3.6 to 54.5%. Concentrations above MPCw.z are recorded in a number of regions. The MPCm.o is exceeded in five regions. It should be noted that two stations operated for a part of 2021; otherwise, the lower limits of the frequency of occurrence would be even higher. Of special note is the RUDN station, at which we recorded 145 periods longer than 8 h, when the concentration exceeded 100 μg/m3, and 402 cases with the hourly concentrations of 160 μg/m3 and larger. All features taken together, a classical photochemical smog persisted in Moscow for few days at afternoon hours. We note that the State Nature Organization Mosecomonitor regularly provides the information on ozone content in the surface air online at https://mosecom.mos.ru/vozdux/. However, residents know little about how hazardous ozone is for human health and what they should do during the events of hazardous ozone concentrations. More diseases, reported in numerous mass media, were merely attributed to high air temperatures in those periods of time. CONCLUSIONS Our review shows that the ozone concentration in the surface air layer exceeded the national hygienic standards at all sites on the territory of Russia in 2021. This motivates the more comprehensive analysis of ozone precursors and development of measures for reducing their emission to the atmosphere. It is also evident that the data in this review are mosaic in character. There are no data available for many big regions in the country. Therefore, there should be more cities and background regions where ozone would be regularly monitored. The smog situation recorded during summer 2021 in Moscow can recur at any time under the conditions of the warming climate. This indicates that the system for warning the population about dangerous pollution of atmospheric air should be updated to include prognostic data on ozone concentration calculated using statistical and numerical models. ACKNOWLEDGMENTS The authors thank the Department of Nature Management and Environmental Protection of Moscow and, personally, E.G. Semutnikova for creation of the state-of-the-art competitive monitoring system, and for its development and support in meeting modern demands. FUNDING The review was prepared based on data acquired with the use of the infrastructure of V.E. Zuev Institute of Atmospheric Optics, Siberian Branch, Russian Academy of Sciences, including the Common Use Center “Atmosphere”, created and operated within the State order no. 121031500342-0 under partial support of the Ministry of Science and Higher Education of the Russian Federation (agreement no. 075-15-2021-661), infrastructure of the Institute of Physical Material Science, Siberian Branch, Russian Academy of Sciences created and operated within the State order no. 121032500027-3, infrastructure of Karadag Scientific Station—Nature Reserve of Russian Academy of Sciences—Branch of the Institute of Biology of Southern Seas, Russian Academy of Sciences, within the State order no. 121032300023-7, infrastructure of the Institute of Atmospheric Physics, Russian Academy of Sciences, within the State order no. 129-2022-0012, and infrastructure of the Limnological Institute, Siberian Branch, Russian Academy of Sciences, within the State order no. 0279-2021-0014. CONFLICT OF INTEREST The authors declare that they have no conflicts of interest. Translated by O. Bazhenov ==== Refs REFERENCES 1 Lunin V. V. Popovich M. P. Tkachenko S. N. Physical Chemistry of Ozone 1998 Moscow MSU 2 Perov S. P. Khrgian A. Kh. Modern Problems of Atmospheric Ozone 1980 Leningrad Gidrometeoizdat 3 Razumovskii S. V. Zaikov G. E. Ozone and Its Reactions with Organic Compounds (Kinetics and Mechanics) 1974 Moscow Nauka 4 Harmful Chemicals Inorganic Compounds of the V–VIII groups 1989 Leningrad Khimiya 5 Belan B. D. Tropospheric Ozone 2010 Tomsk Publishing House of IAO SB RAS 6 Lazarev N. V. Handbook for Chemists, Engineers, and Doctors. Harmful Substances in Industry 1971 Leningrad Khimiya 7 SanPiN 1.2.3685-21 Hygienic Standards and Requirements for Ensuring the Safety and (or) Security of Environmental Factors for Humans. Introduced January 28, 2021 (Standart inform, Moscow, 2021) [in Russian]. 8 Mauzerall D. L. Wang X. “Protecting agricultural crops from the effects of tropospheric ozone exposure: Reconciling science and standard setting in the United States, Europa, and Asia,” Ann. Rev Energy Environ, No. 2001 26 237 268 9 “Air pollution takes a big bite out of Asia’s grain crops,” Nature 601 (7894), 487 (2022). 10 Wang B. Shugart H. H. Lerdau M. T. Sensitivity of global greenhouse gas budgets to tropospheric ozone pollution mediated by the biosphere Environ. Res. Lett. 2017 12 084001 10.1088/1748-9326/aa7885 11 Anav A. De Marco A. Collalti A. Emberson L. Feng Z. Lombardozzi D. Sicard P. Verbeke T. Viovy N. Vitale M. Paoletti E. Legislative and functional aspects of different metrics used for ozone risk assessment to forests Environ. Pollut. 2022 295 118690 10.1016/j.envpol.2021.118690 34921939 12 Altshuller A. P. Wartburg A. F. The interaction of ozone with plastic metallic materials in a dynamic flow system Intern. J. Air Water Pollut. 1961 4 70 78 13 Daubendick R. L. Calvert J. G. The reaction of ozone with perfluorinated polyolefins Environ. Lett. 1974 6 253 272 10.1080/00139307409437369 4835568 14 Screpani A. DeMarco A. Corrosion on cultural heritage buildings in Italy: A role for ozone? Environ. Pollut. 2009 157 1513 1520 10.1016/j.envpol.2008.09.046 19026474 15 Coleman B. K. Destaillats H. Hodgson A. T. Nazaroff W. W. Ozone consumption and volatile byproduct formation from surface reactions with aircraft cabin materials and clothing fabrics Atmos. Environ. 2008 42 642 654 10.1016/j.atmosenv.2007.10.001 16 Rice R. G. Century 21—pregnant with ozone Ozone Sci. Engeen. 2002 24 1 15 10.1080/01919510208901590 17 M. J. Rowlinson, A. Rap, D. S. Hamilton, R. J. Pope, S. Hantson, S. R. Arnold, J. O. Kaplan, A. Arneth, M. P. Chipperfield, P. M. Forster, and L. Nieradzik, “Tropospheric ozone radiative forcing uncertainty due to pre-industrial fire and biogenic emissions,” Atmos. Chem. Phys. 20 (18), 10 937–10 951. 18 Air Pollution by Ozone in Europe in Summer 2004. Overview of Exceedances of EC Ozone Threshold Values during April–September 2004. Technical Report No. 3 (EEA, 2005). 19 Gouldsbrough L. Hossaini R. Eastoe E. Young P. Y. A temperature dependent extreme value analysis of UK surface ozone, 1980–2019 Atmos. Environ. 2020 273 118975 10.1016/j.atmosenv.2022.118975 20 Langford A. O. Senff C. J. IIR. J. Alvarez, K. C. Aikin, S. Baidar, T. A. Bonin, W. A. Brewer, J. Brioude, S. S. Brown, J. D. Burley, D. J. Caputi, S. A. Conley, P. D. Cullis, Z. C. J. Decker, S. Evan, G. Kirgis, M. Lin, M. Pagowski, J. Peischl, I. Petropavlovskikh, R. B. Pierce, T. B. Ryerson, S. P. Sandberg, C. W. Sterling, A. M. Weickmann, and L. Zhang, “The Fires, Asian, and Stratospheric Transport-Las Vegas Ozone Study (FAST-LVOS) Atmos. Chem. Phys. 2022 22 1707 1737 10.5194/acp-22-1707-2022 21 S.-W. Kim, B. C. McDonald, S. Seo, K.-M. Kim, and M. Trainer, “Understanding the paths of surface ozone abatement in the Los Angeles Basin,” J. Geophys. Res.: Atmos. 127 (4) (2022). 22 Cao J. Qiu X. Liu Y. Yan X. Gao J. Peng L. Identifying the dominant driver of elevated surface ozone concentration in North China Plain during summertime 2012–2017 Environ. Pollut. 2022 300 118912 10.1016/j.envpol.2022.118912 35092729 23 Wu K. Wang Y. Qiao Y. Liu Y. Wang S. Yang X. Wang H. Lu Y. Zhang X. Lei Y. Drivers of 2013–2020 ozone trends in the Sichuan Basin, China: Impacts of meteorology and precursor emission changes Environ. Pollut. 2022 300 118914 10.1016/j.envpol.2022.118914 35124125 24 Gao J. Li Y. Xie Z. Hue B. Wang L. Bao F. Fan S. The impact of the aerosol reduction on the worsening ozone pollution over the Beijing–Tianjin–Hebei region via influencing photolysis rates Sci. Total Environ. 2022 821 153197 10.1016/j.scitotenv.2022.153197 35063532 25 https://mosecom.mos.ru. Cited March 24, 2022. 26 www.mos.ru/eco/documents/doklady/view/. Cited March 24, 2022. 27 Review of the State and Pollution of the Environment in the Russian Federation for 2020 (Rosgidromet, Moscow, 2021) [in Russian]. 28 Overview of the Background State of the Environment on the Territory of the CIS Countries for 2019 (Yu.A. Israel Institute of Global Climate and Ecology; Rosgidromet, Moscow, 2020) [in Russian]. 29 Andreev V. V. Arshinov M. Yu. Belan B. D. Davydov D. K. Elanskii N. F. Zhamsueva G. S. Zayakhanov A. S. Ivlev G. A. Kozlov A. V. Kotel’nikov S. N. Kuznetsova I. N. Lapchenko V. A. Lezina E. A. Postylyakov O. V. Savkin D. E. Senik I. A. Stepanov E. V. Tolmachev G. N. Fofonov A. V. Chelibanov I. V. Chelibanov V. P. Tikhonov N. A. Zakharova S. A. Davydova M. A. Surface ozone concentration over Russian territory in the first half of 2020 Atmos. Ocean. Opt. 2020 33 671 681 10.1134/S1024856020060184 30 Andreev V. V. Arshinov M. Yu. Belan B. D. Belan S. B. Davydov D. K. Demin V. I. Elanskii N. F. Zhamsueva G. S. Zayakhanov A. S. Ivlev G. A. Kozlov A. V. Kotel’nikov S. N. Kuznetsova I. N. Lapchenko V. A. Lezina E. A. Postylyakov O. V. Savkin D. E. Senik I. A. Stepanov E. V. Tolmachev G. N. Fofonov A. V. Chelibanov I. V. Chelibanov V. P. Shirotov V. V. Shukurov K. A. Surface ozone concentration in Russia in the second half of 2020 Atmos. Ocean. Opt. 2021 34 347 356 10.1134/S1024856021040035 31 Mertens M. Jockel P. Matthes S. Nutzel M. Grewe V. Volker, and R. Sausen, “COVID-19 induced lower-tropospheric ozone changes Environ. Res. Lett. 2021 16 064005 10.1088/1748-9326/abf191 32 Adam M. G. Tran P. T. M. Balasubramanian R. Air quality changes in cities during the COVID-19 lockdown: A critical review Atmos. Res 2021 264 105823 10.1016/j.atmosres.2021.105823 34456403 33 Yang M. Chen L. Msigwa G. Tang K. H. D. Yap Pow-Seng P.‑S. Implications of COVID-19 on global environmental pollution and carbon emissions with strategies for sustainability in the COVID-19 era Sci. Total Environ. 2022 809 151657 10.1016/j.scitotenv.2021.151657 34793787 34 W. Steinbrecht, D. Kubistin, C. Plass-Dulmer, J. Davies, D. W. Tarasick, P. Gathen, H. Deckelmann, N. Jepsen, R. Kivi, N. Lyall, M. Palm, J. Notholt, B. Kois, P. Oelsner, M. Allaart, A. Piters, M. Gill, R. Van Malderen, A. W. Delcloo, R. Sussmann, E. Mahieu, C. Servais, G. Romanens, R. Stubi, G. Ancellet, S. Godin-Beekmann, S. Yamanouchi, K. Strong, B. Johnson, P. Cullis, I. Petropavlovskikh, J. W. Hannigan, J.-L. Hernandez, A. D. Rodriguez, T. Nakano, F. Chouza, T. Leblanc, C. Torres, O. Garcia, A. N. Rohling, M. Schneider, T. Blumenstock, M. Tully, C. Paton-Walsh, N. Jones, R. Querel, S. Strahan, R. M. Stauffer, A. M. Thompson, A. Inness, R. Engelen, K.-L. Chang, and O. R. Cooper, “COVID-19 crisis reduces free tropospheric ozone across the Northern Hemisphere,” Geophys. Rev. Lett. 48 (5), e2020GL091987 (2021). 35 I. Bouarar, B. Gaubert, G. P. Brasseur, W. Steinbrecht, T. Doumbia, S. Tilmes, Y. Liu, T. Stavrakou, A. Deroubaix, S. Darras, C. Granier, F. Lacey, J.-F. Muller, X. Shi, N. Elguindi, and T. Wang, “Ozone anomalies in the free troposphere during the COVID-19 pandemic,” Geophys. Rev. Lett. 48 (16), e2021GL094204 (2021). 36 Antokhin P. N. Arshinov M. Yu. Belan B. D. Belan S. B. Sklyadneva T. K. Tolmachev G. N. Many-year variability of ozone and aerosol near Tomsk and justification of the ten-year prediction of their yearly average concentrations Opt. Atmos. Okeana 2010 23 772 776 37 U. Feister and W. Warmbt, “Long-term surface ozone increase at Arcona,” in Proc. Quadrennial Ozone Symposium, Greece, (1980), p. 782–787. 38 Volz A. Kley D. Evaluation of the montsouris series of ozone measurements made in the nineteenth century Nature 1988 332 240 242 10.1038/332240a0 39 Cartalis C. Varotsos C. Surface ozone in Athens, Greece, at the beginning and at the end of the twentieth century Atmos. Environ. 1994 28 3 8 10.1016/1352-2310(94)90018-3 40 Belan B. D. Savkin D. E. Tolmachev G. N. Study of the relationship between snow cover and ozone concentration in the surface air layer Opt. Atmos. Okeana 2018 31 665 669 41 Monks P. S. A review of the observations and origins of the spring ozone maximum Atmos. Environ. 2000 34 3545 3561 10.1016/S1352-2310(00)00129-1 42 Zayakhanov A. S. Zhamsueva G. S. Tsydypov V. V. Bal’zhanov T. S. Results of surface ozone monitoring in the atmosphere over Ulan-Ude Russ. Meteorol. Hydrol. 2013 38 846 852 10.3103/S1068373913120066 43 Isidorov V. A. Volatile Plant Excretions: Composition, Emission Rate, and Ecological Role 1994 St. Petersburg Alga 44 Curci G. Beekmann M. Vautard R. Smiatek G. Steinbrecher R. Theloke J. Friedrich R. Modelling study of the impact of isoprene and terpene biogenic emissions on european ozone levels Atmos. Environ. 2009 43 1444 1455 10.1016/j.atmosenv.2008.02.070 45 Collins W. J. Sitch S. Boucher O. How vegetation impacts affect climate metrics for ozone precursors J. Geophys. Res. 2010 115 D23308 10.1029/2010JD014187 46 Sadiq M. Tai A. P. K. Lombardozzi D. Martin M. V. Effects of ozone-vegetation coupling on surface ozone air quality via biogeochemical and meteorological feedbacks Atmos. Chem. Phys. 2017 17 3055 3066 10.5194/acp-17-3055-2017 47 N. F. Elansky, A. V. Shilkin, N. A. Ponomarev, P. V. Zakharova, M. D. Kachko, and T. I. Polyakov, “Spatio-temporal variations in pollution of the Moscow air basin and pollutant emissions,” Izv. Ross. Akad. Naul. Fiz. Atmos. Okeana 58 (1), 92–108 (2022.) 48 Tan Z. Ma X. Lu K. Jiang M. Zou Q. Wang H. Zeng L. Zhang Y. Direct evidence of local photochemical production driven ozone episode in Beijing: A case study Sci. Total Environ. 2021 800 148868 10.1016/j.scitotenv.2021.148868 34384967 49 Liu X. Guo H. Zeng L. Lyu X. Wang Y. Zeren Y. Yang J. Zhang L. Zhao S. Li Y. Zhang G. Photochemical ozone pollution in five Chinese megacities in summer 2018 Sci. Total Environ. 2021 801 149603 10.1016/j.scitotenv.2021.149603 34416603 50 Senik I. A. Elansky N. F. Belikov I. B. Lisitsyna L. V. Galaktionov V. V. Kortunova Z. V. Main patterns of the temporal variability of surface ozone in the region of the town of Kislovodsk at 870 and 2070 m above sea level Izv., Atmos. Ocean. Phys. 2005 41 67 79 51 N. P. Shakina, A. R. Ivanova, N. F. Elansky, and T. A. Markova, “Transcontinental observations of surface ozone concentration in the TROICA experiments: 2. The effect of the stratosphere-troposphere exchange,” Izv. Atmos. Ocean. Phys. 37 (2001). 52 Borisov D. V. Shalygina I. Yu. Lezina E. A. Studying seasonal and daily variability of surface ozone Gidrometeorol. Issledovaniya Prognozy 2020 377 122 135 53 Ivanova N. S. Kuznetsova I. N. Shalygina I. Yu. Lezina E. A. Ozone content over the Russian Federation in 2020 Russ. Meteorol. Hydrol. 2021 46 129 137 10.3103/S1068373921020084 54 Zvyagintsev A. M. Belikov I. B. Egorov V. I. Elansky N. F. Kruchenitskii G. M. Kuznetsova I. N. Nikolaev A. N. Obukhova Z. V. Skorokhod A. I. Positive anomalies in the surface ozone concentration in July-August 2002 over Moscow and its suburbs Izv., Atmos. Ocean. Phys. 2004 40 68 79 55 Belan B. D. Savkin D. E. Tolmachev G. N. Generation of ozone in the surface air layer versus air temperature Atmos. Ocean. Opt. 2018 31 187 196 10.1134/S1024856018020045 56 Porter W. C. Heald C. L. The mechanisms and meteorological drivers of the summertime ozone-temperature relationship Atmos. Chem. Phys. 2019 19 13367 13381 10.5194/acp-19-13367-2019 57 Ulpiani G. Ranzi G. Santamouris M. Local synergies and antagonisms between meteorological factors and air pollution: A 15-year comprehensive study in the Sydney region Sci. Total Environ. 2021 788 147783 10.1016/j.scitotenv.2021.147783 34029820 58 Platikanov S. Terrado M. Pay M. T. Soret A. Tauler R. Understanding temporal and spatial changes of O3 or NO2 concentrations combining multivariate data analysis methods and air quality transport models Sci. Total Environ. 2022 806 150923 10.1016/j.scitotenv.2021.150923 34653450 59 Hossain M. S. Frey H. C. Louie P. K. K. Lau A. K. H. Combined effects of increased O3 and reduced NO2 concentrations on short-term air pollution health risks in Hong Kong Environ. Pollut. 2021 270 116280 10.1016/j.envpol.2020.116280 33360064 60 Liu C. Shi K. A review on methodology in O3‑NOx-VOC sensitivity study Environ. Pollut. 2021 291 118149 10.1016/j.envpol.2021.118249 61 Belan B. D. Savkin D. E. The role of air humidity in variations of near-surface ozone concentration Atmos. Ocean. Opt. 2019 32 586 589 10.1134/S1024856019050038 62 Zuev V. E. Belan B. D. Kabanov D. M. Kovalevskii V. K. Luk’yanov O. Yu. Meleshkin V. E. Mikushev M. K. Panchenko M. V. Penner I. E. Pokrovskii E. V. Sakerin S. M. Terpugova S. A. Tumakov A. G. Shamanaev V. S. Shcherbatov A. I. The “OPTIK-E” AN-30 aircraft-laboratory for ecological investigations Atmos. Ocean. Opt. 1992 5 658 663 63 Anokhin G. G. Antokhin P. N. Arshinov M. Yu. Barsuk V. E. Belan B. D. Belan S. B. Davydov D. K. Ivlev G. A. Kozlov A. V. Kozlov V. S. Morozov M. V. Panchenko M. V. Penner I. E. Pestunov D. A. Sikov G. P. Simonenkov D. V. Sinitsyn D. S. Tolmachev G. N. Filippov D. V. Fofonov A. V. Chernov D. G. Shamanaev V. S. Shmargunov V. P. OPTIK Tu-134 aicraft laboratory Opt. Atmos. Okeana 2011 24 805 816 64 Arshinov M. Yu. Belan B. D. Krasnov O. A. Kovalevskii V. K. Pirogov V. A. Plotnikov A. P. Tolmachev G. N. Fofonov A. V. Comparison of ultraviolet and chemiluminescent ozonometers Atmos. Ocean. Opt. 2002 15 656 658 65 Dunlea E. J. Herndon S. C. Nelson D. D. Volkamer R. M. Lamb B. K. Allwine E. J. Grutter M. Ramos Villegas C. R. Marquez C. Blanco S. Cardenas B. Kolb C. E. Molina L. T. Molina M. J. Technical note: Evaluation of standard ultraviolet absorption ozone monitors in a polluted urban environment Atmos. Chem. Phys. 2006 6 3163 3180 10.5194/acp-6-3163-2006 66 Belan B. D. Tolmachev G. N. Fofonov A. V. Vertical ozone distribution in troposphere above south regions of West Siberia Atmos. Ocean. Opt. 2011 24 181 187 10.1134/S1024856011020059

0

PMC9753031

NO-CC CODE

2022-12-16 23:25:27

no

Atmos Ocean Opt. 2022 Dec 15; 35(6):741-757

utf-8

oa_other

==== Front J Plant Res J Plant Res Journal of Plant Research 0918-9440 1618-0860 Springer Nature Singapore Singapore 1426 10.1007/s10265-022-01426-4 Regular Paper – Physiology/biochemistry/molecular and Cellular Biology Metabolomics and transcriptomics provide insights into the flavonoid biosynthesis pathway in the roots of developing Aster tataricus Jia Kaixuan 12 Zhang Xiaoling 12 Meng Yijiang 23 Liu Shuqi 12 Liu Xiaoqing 12 Yang Taixin 12 Wen Chunxiu 4 Liu Lingdi nkyliulingdi@126.com 4 Ge Shujun gshj@hebau.edu.cn 12 1 grid.274504.0 0000 0001 2291 4530 College of Agronomy, West Campus of Hebei Agricultural University, Lianchi District, Baoding, 071000 Hebei China 2 grid.419897.a 0000 0004 0369 313X Key Laboratory of Crop Germplasm Resources Research and Utilization in North China, Ministry of Education, Baoding, 071000 China 3 grid.274504.0 0000 0001 2291 4530 College of Life Science, Hebei Agricultural University, Baoding, 071000 China 4 grid.464364.7 0000 0004 1808 3262 Institute of Cash Crops, Medicinal Plant Research Center West of Hebei Academy of Agriculture and Forestry Sciences, Nongke Road, Xiyuan Street, Xinhua District, Shijiazhuang, 050000 Hebei China 15 12 2022 118 13 7 2022 9 11 2022 © The Author(s) under exclusive licence to The Botanical Society of Japan 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Aster tataricus (L.) is an important medicinal plant in China. Its roots are rich in flavonoids, the main medicinal components. However, the molecular basis of flavonoid biosynthesis in the roots of A. tataricus remains unclear. In this study, the content of total flavonoid of A. tataricus roots at different developmental stages was measured first, and the results showed that the content of total flavonoid gradually decreased from September to November, which may be caused by the stagnation of A. tataricus growth due to the decrease in temperature after September. Then, an integrated analysis of transcriptome and metabolome was conducted on five developing stages of A. tataricus roots to identify flavonoid compositions and potential genes involved in flavonoid biosynthesis. A total of 80 flavonoid metabolites, of which 75% were flavonols and flavonoids, were identified in metabolomic analyses, among which isorhamnetin, kaempferol, quercetin, and myricetin were the main skeletons of these flavonoids. Cluster analysis divided these 80 flavonoids into 3 clusters. The compounds in cluster I mainly accumulated in S1, S3, and S5. In cluster II, the relative content of the flavonoid metabolites showed an upward trend from S2 to S4. In cluster III, the flavonoids decreased from S1 to S5. A total of 129 structural genes, including 43 PAL, 23 4CL, 9 C4H, 4 CHS, 18 CHI, 3 F3H, 5 F3’H, 1 F3′5′H, 21 FLS, and 2 FSII, and 65 transcription factors, including 22 AP2/ERF, 7 bHLH, 5 bZIP, 8 MYB, 11 NAC, and 12 WRKY, showed significant correlation with total flavonoid content. Eighteen genes (7 4CL, 5 C4H, 2 CHI, 1 F3H, and 3 FLS) and 30 genes (5 PAL, 9 4CL, 1 C4H, 2 CHI, 1 F3H, 1 DFR, 7 3AT, 1 BZ1, and 3 UGT79B1) were identified as key structural genes for kaempferol and anthocyanins biosynthesis, respectively. Our study provides valuable information for understanding the mechanism of flavonoid biosynthesis in A. tataricus root. Supplementary Information The online version contains supplementary material available at 10.1007/s10265-022-01426-4. Keywords Aster tataricus Differentially expressed genes Flavonoids Metabolite Transcriptome Key Research and Development Project of Hebei Province: Scientific and Technological Innovation Team of Modern Seed Industry of Traditional Chinese Medicine21326312D Ge Shujun Hebei Chinese medicine industry Technical systemHBCT2018060201 Ge Shujun China Agriculture Research System of MOF and MARACARS-21 Ge Shujun ==== Body pmcIntroduction Aster tataricus (L.) (known as ‘Ziwan' in Chinese) is an important medicinal plant in the Compositae family. In China, A. tataricus is mainly distributed in Anguo City, Hebei Provinces, the genuine producing area of aster, and Bozhou City, Anhui Province. In addition, it is also cultivated in Japan and North Korea. Its dried root and rhizome (Asteris Radix et Rhizoma, AR) have been widely used as a traditional medicine to treat cough, inflammation, and asthma for more than 2,000 years (Chen et al. 2020a; Yu et al. 2015). Also, AR can be used in the treatment of novel coronavirus pneumonia (COVID-19) because of the effect of moistening lung for arresting cough, resolving phlegm, and relieving asthma (Chen et al. 2020b; Ren et al. 2021). The roots of A. tataricus are rich in chemical components, including phenolic acids, organic acids, flavonoids, terpenoids, and coumarin (Sun et al. 2018; Zhao et al. 2015). Flavonoids, such as quercetin and kaempferol, are important active ingredients in A. tataricus roots. Chen et al. identified 31 flavonoids from A. tataricus by UHPLC-Q-TOF-MS (Chen et al. 2019). Sun et al. extracted 31 flavonoids from rhizomes of A. tataricus by UHPLC-Q-TOF-MS and showed quercetin may have antidepressant effects (Sun et al. 2018). (2R,2′R)-7-O-methyl–2,3,2′′,3′′–tetrahydrorobustaflavone, a new flavonoid isolated from A. tataricus, can significantly inhibit A594 cancer cells proliferation (Chen et al. 2022). As the largest group of secondary metabolites in plants, flavonoids not only contribute to diverse human health benefits, but also provide plants with a variety of biological functions, such as diseases treatment for humans and stress resistance for plants (Ng et al. 2003; Treutter 2005; Wang et al. 2020; Wen et al. 2021). Flavonols may facilitate maize (Zea mays) seedling drought tolerance by scavenging H2O2 and stomatal closure (Li et al. 2021a). Quiroz et al. reported that flavonoids formononetin and genistein protect plants by eliciting an antifeedant effect on H. obscurus (Quiroz et al. 2017). In addition, anthocyanins are flavonoids that impart bright color to plant tissues (Olivas-Aguirre et al. 2016). However, the current research on A. tataricus mainly concentrates on the pharmacological and chemical components, and very little research focuses on the molecular mechanism of accumulation of flavonoids (Cheng and Shao 1993; Su et al. 2019; Yu et al. 2015). Therefore, it is important to dissect the biosynthesis pathway of flavonoid in the roots of A. tataricus, which will be helpful for guiding the quality breeding and increasing the content of flavonoids. In recent years, with the development of UPLC-MS/MS and sequencing technology, the combination of metabolomics and transcriptomics has been widely used to investigate the metabolites and reveal the biosynthesis pathway of metabolites in plants (Li et al. 2019). In this study, we integrated the analysis of metabolomics and transcriptomics to investigate flavonoid biosynthesis in developing roots of A. tataricus at 5 different stages. The purpose of our study was to examine the species of flavonoids, and analyze the differentially expressed genes involved in flavonoid biosynthesis in the roots of A. tataricus during development. Our results revealed the accumulation pattern of flavonoids during root development of A. tataricus, identified the key genes involved in the flavonoid biosynthesis pathway, and provided valuable information for the further study of flavonoids in A. tataricus. Materials and methods Plant materials The A. tataricus ‘Qiziwan’, which is a landrace grown in Anguo, China, were grown at the germplasm resource center at Chinese Medicine Capital Expo Park, Anguo, China, in 2019 and subjected to normal field management during the growth periods. A. tataricus grow rapidly from August to September. The growth of A. tataricus became slow after September, and the aerial parts of A. tataricus gradually began to dry up. Only the underground part remains in December. We speculate that the content of flavonoids in A. tataricus roots is not constant from August to December. Therefore, roots were harvested on the 15th of August (S1), September (S2), October (S3), November (S4), and December (S5), respectively (Fig. 1a). For each sample, the mixed roots were collected from five independent plants and had three independent biological replicates. All materials were frozen in liquid nitrogen and stored at – 80 ℃ until further use.Fig. 1 Phenotypes, total content of flavonoid, and number of flavonoid metabolites of A. tataricus “Qiziwan” at different developmental stages. S1, S2, S3, S4, and S5 represent A. tataricus were harvested on the 15th of August, September, October, November and December, respectively. a Phenotypes of A. tataricus “Qiziwan” at different developmental stages. b Changes in total content of flavonoid during A. tataricus root development. Error bars represent ± standard deviation (n = 3). Statistical analyses were performed by Duncan’s multiple range tests, P < 0.05. (c) Number of flavonoid metabolites detected in A. tataricus roots in each categorie. The number above the column represents the number of the categories Measurement of total flavonoid content Total flavonoid measurement was carried out by the aluminum nitrate colorimetric method (Hossain and Rahman 2011). The freeze–dried roots were crushed using a mixer mill (MM 400, Retsch) with a zirconia bead for 1.5 min at a frequency of 30 Hz. Approximately 1.0 g root powder was weighed and extracted 15 min at 30 ℃ with 15 mL 80% methanol aqueous solution to obtain the crude extract. 300 μL crude extract was mixed with 90 μL of 5% NaNO2 solution and 1.5 mL deionized water. Then 180 μL 10% Al(NO3)3 solution was added after 6 min of incubation, and the mixture was incubated for another 5 min. Subsequently, 600 μL of 1 mol L−1 NaOH solution was added and the final volume of the mixture solution was 3 mL. Then the absorbance was measured at a wavelength of 510 nm by the microplate reader. Rutin was used as a standard solution to prepare a calibration curve, and the results were expressed as rutin equivalent on a dry weight basis. Metabolite extraction and profiling analysis Metabolite extraction and profiling analyses were performed by Metware Biotechnology Co., Ltd. (Wuhan, China). The experimental procedure was done following the company's standard instruction (Chen et al. 2013; Dong et al. 2014, 2019). In brief, 100 mg root powder was weighed and extracted overnight at 4 ℃ with 0.6 mL 70% methanol aqueous solution. After centrifugation at 10,000 g for 10 min, the supernatant was filtered through a 0.22 μm microporous membrane for UPLC-MS/MS analysis. Chromatographic separation was carried out on a Waters ACQUITY UPLC HSS T3 system (Shim-pack UFLC SHIMADZU CBM30A), which was equipped with an Agilent SB-C18 column (1.8 µm, 2.1 mm × 100 mm), at 40 ℃. The injection volume was 4 μL at a flow rate of 0.35 mL min−1. The mobile phase was ultrapure water (with 0.1% formic acid): acetonitrile. The elution gradient was as follows: 0 min, 95:5 water/acetonitrile (v/v); 9.0 min, 5:95 water/acetonitrile; 10.0 min, 5:95 water/acetonitrile; 11.1 min, 95:5 water/acetonitrile; and 14.0 min, 95:5 water acetonitrile. The effluent was alternatively connected to an ESI–triple quadrupole–linear ion trap (QTRAP)-MS. Metabolites were detected using an Applied Biosystems 4500 QTRAP LC/MS/MS system equipped with linear ion trap (LIT) and triple quadrupole (QQQ) scans. This system was controlled by Analyst 1.6.3 software. The ESI source operation parameters were as follows: ion source, turbo spray; source temperature, 550 °C; ion spray voltage (IS), 5,500 V; curtain gas (CUR) 25.0 psi; and collision–activated dissociation (CAD), high. Ten and 100 μmol L−1 polypropylene glycol solutions were used in the QQQ and LIT modes, respectively, for the instrument tuning and mass calibration. QQQ scans were acquired as multiple reaction monitoring (MRM) experiments with collision gas (nitrogen) set to medium. Declustering potential (DP) and collision energy (CE) measurements for individual MRM transitions were completed with further DP and CE optimization. A specific set of MRM transitions was monitored for each period according to the metabolites eluted within the period (Chen et al. 2013). Mass spectral data analysis was conducted with the software Analyst 1.6.3. The flavonoids were qualitatively and quantitatively analyzed by blasting with the local database. Principal component analysis (PCA) and partial least-squares discriminant analysis (OPLS-DA) analysis were carried out for identifying differentially expressed flavonoids. Significantly different metabolites between groups were determined by the variable importance in projection (VIP) ≥ 1 and fold change ≥ 2 or ≤ 0.5. RNA extraction, quantification and sequencing Total RNA was extracted from frozen roots using the OminiPlant RNA Kit (Dnase) (cwbio, Jiangsu, China). RNA integrity and contamination was monitored by 2% agarose gel. RNA purity was detected using a NanoDrop one spectrophotometer (Thermo Scientific, Shanghai, China). The concentration of RNA was accurately measured using a Qubit 2.0 Fluorometer. The quality of RNA was detected using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Poly(A) mRNA was enriched from total RNA by Oligo(dT) magnetic beads, followed by fragmentation of poly(A) mRNA into short fragments. The first-strand cDNA was generated by reverse transcription using the fragment RNA as a template and using random hexamers primers. Second-strand cDNA was synthesized using DNA polymerase I. Double-stranded cDNA was then purified using AMPure XP beads. End–repaired and addition of a poly(A) tail of the purified double-stranded cDNA, and a sequencing adapter was ligated. Fragments of suitable size were isolated. The cDNA library was obtained by PCR enrichment. The cDNA library was sequenced on an Illumina HiSeq™ 2500 platform. RNA sequencing data analysis High–quality reads are the basis for accurate subsequent analysis. To acquire high–quality reads, strict quality control of the data was carried out. Clean reads were obtained from raw data by removing reads with adapters, low-quality reads, and reads containing too many N (the N content exceeds 10% of the number of bases in the read). The transcriptome was obtained by assembling clean reads using Trinity software. The longest cluster sequence obtained after corset hierarchical clustering was used as unigenes for subsequent analysis. The unigenes sequence were aligned with the KEGG, NR, Swiss-Prot, GO, COG/KOG, and Trembl databases using BLAST software. After predicting the amino acid sequence of unigenes, the HMMER software was used to compare with the Pfam database to obtain the annotation information of unigenes. Gene expression levels were estimated using bowtie 2 in RSEM software. Differentially expressed genes (DEGs) between different sample groups were analyzed using DESeq 2 (Love et al., 2014; Varet et al., 2016). The false discovery rate (FDR) was obtained by multiple hypothesis testing for hypothesis testing probability (P-value) using the Benjamini–Hochberg method. Significantly different genes between groups were determined by |log2 Fold Change|≥ 1 and FDR < 0.05. The KEGG pathway and GO term of the DEGs were obtained from the Kyoto Encyclopedia of Genes and Genomes websites (KEGG, https://www.genome.jp/kegg) and the Gene Ontology websites (GO, http://www.geneontology.org/), respectively (Gene Ontology Consortium 2015; Kanehisa et al. 2016). We conducted a hypergeometric test to find the KEGG pathways or GO terms that are significantly enriched in DEGs compared to the whole genome background. The DEGs were aligned to the KOG database using Blast software to obtain the KOG annotation. Joint analysis of transcriptome and metabolome The DEGs and differentially accumulated metabolites (DAMs) of flavonoid biosynthesis–related pathways in each comparison group (S1 vs S2, S1 vs S3, S1 vs S4, S1 vs S5, S2 vs S3, S2 vs S4, S2 vs S5, S3 vs S4, S3 vs S5, and S4 vs S5) were screened according to the relative contents of flavonoid metabolites and gene expression values in the roots of A. tataricus at different developmental stages. First, the correlation of DEGs and DAMs was obtained by calculating their Pearson correlation coefficients. Pearson’s correlation coefficient ≥ 0.8 was considered to be significantly correlated. Then, DEGs and DAMs, which with significant correlations, were mapped to the KEGG pathway database to gather information about their shared pathways to better comprehend their interaction. Weighed gene co-expression network analysis In order to further examine the genes that are strongly connected to the traits (total flavonoid content, kaempferol content, 5 anthocyanins content, and the variation trends of 3 clusters of flavonoid metabolites) of A. tataricus, the R software WGCNA package was used to generate a weighted gene co-expression network (Ghazalpour et al. 2006). The co-expression modules were obtained by using the one–step network construction function with following parameters: maxBlockSize: 20,000, soft–threshold power: 13, TOMtype: unsigned, mergeCutHeight: 0.25, minModuleSize: 300, and other parameters: default. The correlation between the modules and the traits was further obtained by calculating the eigenvectors of each module. qRT-PCR analysis Twenty DEGs involved in flavonoid biosynthesis obtained in the RNA-seq data were selected for confirmation by qRT-PCR. The cDNA, which was used as a template to measure gene expression level, was obtained by reverse transcription of the total RNA of A. tataricus root according to the TRUEscript RT MaseterMix (OneStep gDNA Removal) (Aidlab, Beijing, China). The A. tataricus actin gene was taken as a reference in all experiments to normalize gene expression level. The comparative 2−ΔΔCT method was used to quantify the gene expression level as described previously (Schmittgen and Livak 2008). The primers, including the flavonoid biosynthesis related genes and the A. tataricus actin gene (internal control), used in qRT-PCR are listed in Table S1. Statistical analysis Statistical analysis was performed using Excel 2016 software (Microsoft Office, USA). Data are presented as means ± standard deviations (SD). Differences between samples were determined by one–way analysis of variance (ANOVA) and the levels of statistical significance were calculated by the least significant difference (P < 0.05). Results Dynamic changes of total flavonoid content during the root development in A. tataricus To investigate the accumulating trends of flavonoids in the roots of A. tataricus, the total content of flavonoid was measured at five developmental stages. As shown in Fig. 1b, the total content of flavonoids in roots at S1 and S2 stages was about 12.95 mg g−1 of dry weight, while at S3 and S4, it was significantly reduced to 10.85 mg g−1 and 8.94 mg g−1, respectively, and it was 9.10 mg g−1 at S5, which had no significant difference from S4. These results showed that the total flavonoid content in roots of A. tataricus began to decrease at S2 until S4. Flavonoids showing differential accumulated during the root development in A. tataricus Furtherly, we analyzed the metabolome data obtained by UPLC/ESI–Q TRAP-MS/MS. A total of 80 flavonoid metabolites, including 31 flavonols, 29 flavonoids, 6 isoflavones, 5 dihydroflavones, 5 anthocyanins, 2 flavonoid carbonosides, 1 dihydroflavonol, and 1 flavanol, were identified at these five stages (Tables 1, S2, Fig. 1c). 70% of the flavonoids detected here are flavonoid glycosides, among which the most prevalent three glycosides were glucoside (37), galactoside (6), and hexoside (6) (Table S3).Table 1 Classification and relative content of the top flavonoids determined during A. tataricus roots development Compounds Class S1 S2 S3 S4 S5 Quercetin-3-O-α-l-rhamnopyranoside Flavonoid 2.94 × 107 1.17 × 107 1.92 × 107 3.06 × 107 2.36 × 107 Luteolin 3'-O-β-d-glucoside Flavonoid 2.89 × 107 1.16 × 107 1.87 × 107 3.03 × 107 2.32 × 107 Kaempferol-3-O-glucoside (Astragalin) Flavonols 2.87 × 107 1.12 × 107 1.89 × 107 3.03 × 107 2.31 × 107 Luteolin-4'-O-β-d-glucoside Flavonoid 2.85 × 107 1.15 × 107 1.85 × 107 3.02 × 107 2.35 × 107 Kaempferol-3-O-(6´´-acetyl)-glucoside Flavonols 1.50 × 107 7.95 × 106 1.36 × 107 2.06 × 107 1.02 × 107 Quercetin-3-O-(6´´-O-malonyl)-galactoside Flavonols 1.45 × 107 8.15 × 106 2.10 × 107 3.09 × 107 2.53 × 107 Kaempferol-3-O-(6´´-malonyl)-glucoside Flavonols 1.07 × 107 5.22 × 106 9.50 × 106 1.56 × 107 8.80 × 106 Isorhamnetin O-malonylglucoside Flavonoid 8.85 × 106 2.57 × 106 6.40 × 106 6.51 × 106 6.95 × 106 Quercetin-3-O-(6´´-O-acetyl)-galactoside Flavonols 7.21 × 106 4.18 × 106 9.20 × 106 1.59 × 107 1.16 × 107 6-Hydroxykaempferol-7-O-glucoside Flavonols 5.56 × 106 1.87 × 106 4.88 × 106 1.27 × 107 7.26 × 106 Myricetin-O-rhamnoside Flavonoid 5.52 × 106 1.68 × 106 4.66 × 106 1.21 × 107 7.01 × 106 isohyperoside Flavonols 5.50 × 106 1.81 × 106 4.49 × 106 1.23 × 107 6.99 × 106 mlyricetin 3-O-B-d-glucopyranoside Flavonoid 4.96 × 106 1.66 × 106 4.31 × 106 1.11 × 107 6.21 × 106 Cyanidin-3-O-(6´´-Malonylglucoside) Anthocyanins 4.11 × 106 3.67 × 106 5.09 × 107 1.72 × 108 5.73 × 107 Trifolin Flavonols 3.69 × 106 1.25 × 106 2.44 × 106 4.71 × 106 3.63 × 106 Kaempferol-7-O-glucoside Flavonols 3.57 × 106 1.31 × 106 2.35 × 106 4.73 × 106 3.65 × 106 Hyperin Flavonols 3.54 × 106 1.23 × 106 3.12 × 106 8.38 × 106 4.66 × 106 Isorhamnetin-acetyl hexoside Flavonols 2.84 × 106 8.76 × 105 2.40 × 106 3.14 × 106 3.21 × 106 Diosmetin-7-O-galactoside Flavonoid 2.38 × 106 1.29 × 106 3.56 × 106 1.51 × 106 2.22 × 106 Chrysoeriol-O-malonylhexoside Flavonoid 2.19 × 106 1.04 × 106 2.87 × 106 9.78 × 105 1.68 × 106 Table 2 Summary statistics of functional annotation of unigenes in developing roots of A. tataricus Database Number of genes Percentage (%) KEGG 226,372 44.8 NR 286,755 56.75 SwissProt 191,339 37.86 Trembl 283,896 56.18 KOG 174,830 34.6 GO 243,280 48.14 Pfam 197,740 39.13 Annotated in at least one Database 294,773 58.33 Total Unigenes 505,334 100 To gain insight into the variance of metabolites in different developmental stages of A. tataricus roots, DAMs were identified using the identification criterion of VIP ≥ 1 and fold change ≥ 2 or fold change ≤ 0.5 between pairwise comparisons (S1 vs S2, S2 vs S3, S3 vs S4, and S4 vs S5). Compared with S1, S2 had 4 flavonoids increased (1 flavonoid, 1 flavonol, 1 dihydroflavone, and 1 flavonoid carbonoside) and 38 flavonoids decreased (16 flavonoid, 19 flavonols, 2 isoflavones, and 1 dihydroflavone) (Fig. 2a). In the S2 and S3 comparison group, 27 differentially accumulated flavonoids were screened, with 20 upregulated (10 flavonoid, 6 flavonols, 2 isoflavones, and 2 anthocyanins) and 7 downregulated (1 flavonoid, 3 isoflavones, 2 dihydroflavone, and 1 flavonoid carbonoside) (Fig. 2b). The number of differentially accumulated flavonoids between S3 and S4 was 24, of which 17 (4 flavonoid, 10 flavonols, 1 dihydroflavone, and 2 anthocyanins) increased and 7 (5 flavonoid, 1 flavonol, and 1 dihydroflavonol) decreased (Fig. 2c). In the comparison group S4 and S5, 13 flavonoids increased and 11 decreased. The increased flavonoids include 5 flavonoid, 4 flavonols, 2 isoflavones, 1 dihydroflavone, and 1 dihydroflavonol, and the decreased flavonoids include 1 flavonoid, 5 flavonols, and 5 anthocyanins (Fig. 2d). It is interesting that although the total flavonoid content of S1 and S2 was not significantly different, in the comparison group S1 vs S2, the number of differential flavonoid compounds is the largest. It is worth noting that all the 80 detected flavonoids showed differentially accumulation during A. tataricus root development (Table S2).Fig. 2 Clustering heat map of differential accumulation of flavonoid metabolites in different comparison groups. a–d Represent the S1 vs S2, S2 vs S3, S3 vs S4, and S4 vs S5 comparison groups, respectively. The content of each metabolite is represented by a different color. Red indicates high content of metabolite, while low content metabolite is shown in green. S1, S2, S3, S4, and S5 represent aster harvested in August, September, October, November and December, respectively To understand the trends of metabolites in different developmental stages of A. tataricus roots, 80 flavonoids were clustered into 3 clusters based on the accumulation patterns of different metabolites using a K-means clustering algorithm (Fig. 3). The relative content of flavonoids in cluster I showed a trend of decline, rise, decline, and rise again during the five growth periods. In cluster II, the relative content of flavonoid metabolites showed an upward trend from S2 to S4, however, S1 to S2 and S4 to S5 showed a downward trend. In cluster III, the flavonoids kept decreasing from S1 to S5 (Fig. 3b). Notably, 5 anthocyanins, including 2 pelargonidin derivatives (Pelargonidin-3-O-(6′′-acetylglucoside) and Pelargonidin-3-O-(6′′-malonylglucoside)) and 3 cyanidin derivatives (Cyanidin-3-O-(3′′,6′′-dimalonylglucoside), Cyanidin-3-O-(6′′-Malonylglucoside), and Cyanidin-3-O-glucoside (Kuromanin)), were detected in our study, and they all belong to cluster II. The contents of these five anthocyanins were higher in the later stages of A. tataricus development (S4 and S5), especially in the S4 stage, when most of the A. tataricus roots were purple–red (Fig. 1a and Table S2). Pelargonidin derivatives have been reported to change the flower to vivid red, and cyanidin derivatives can turn fruit red (Nakatsuka et al. 2007; Olivas–Aguirre et al. 2016). Therefore, we speculate that the accumulation of these five anthocyanins is related to the reddening of A. tataricus roots.Fig. 3 Cluster analysis of all detected flavonoid metabolites in A. tataricus roots of five stages based on K-means clustering method. S1, S2, S3, S4, and S5 represent A. tataricus were harvested on the 15th of August, September, October, November and December, respectively. a Clustering heat map of all detected flavonoid metabolites in A. tataricus roots of five stages based on K-means clustering algorithm. Each column represents a sample and each row represents a metabolite. The content of each metabolite is represented by a different color. Red indicates high content of metabolite, while low content metabolite is shown in green. b Line chart of flavonoid accumulation patterns for three clusters Analysis of differential expressed genes during root development in A. tataricus To understand the potential molecular biosynthesis pathway of flavonoid in the developmental roots of A. tataricus. Fifteen cDNA libraries were constructed and subjected to high–throughput RNA-seq analysis. As a result, a total of 125.56 G clean reads and 837,144,592 base pairs were obtained from 15 independent samples of A. tataricus roots (Table S4). These clean reads were further assembled into 631,853 transcripts with a mean length of 798 bp and N50 length of 1,204 bp, and 505,334 unigenes with a mean length of 929 bp and N50 length of 1,307 bp using Trinity software (Table S5). To annotate the assembled unigenes, all unigenes were blasted with 7 publicly available nucleotide and protein databases, including KEGG, NR, Swiss-Prot, Trembl, KOG, Go, and Pfam. The results showed that 226,372 (44.8%), 286,755 (56.75%), 191,339 (37.86%), 283,896 (56.18%), 240,483 (47.59%), 163,286 (32.31%), and 197,740 (39.13%) of unigenes were annotated in KEGG, NR, Swiss-Prot, Trembl, KOG, Go, and Pfam database, respectively (Table 2). With the criteria of |Log2Fold Change |≥ 1 and FDR < 0.05 in the transcriptome of A. tataricus roots, 23,012 (12,743 up- and 10,269 down-regulation), 32,792 (13,044 up- and 19,748 down-regulated), 9,124 (3,486 up- and 5,638 down-regulated), and 23,402 (12,417 up- and 10,985 down-regulated) DEGs were identified in S1 vs S2, S2 vs S3, S3 vs S4, and S4 vs S5, respectively (Table S6 and Fig. 4).Fig. 4 Numbers of DEGs in different comparison groups. S1, S2, S3, S4, and S5 represent A. tataricus were harvested on the 15th of August, September, October, November and December, respectively To understand the biological function of the DEGs, GO term enrichment was analyzed. 11,186 (S1 vs S2) and 16,452 (S2 vs S3) DEGs were divided into 58 functional groups, including 18 cellular component categories, 13 molecular function categories, and 27 biological process categories (Fig. 5a, b, Table S7). 4,669 DEGs in S3 vs S4 were divided into 56 functional groups, including 18 cellular component categories, 12 molecular function categories, and 26 biological process categories (Fig. 5c, Table S7). In S4 vs S5, 11,808 DEGs were divided into 18 cellular component categories, 11 molecular function categories, and 28 biological process categories, a total of 57 functional groups. (Fig. 5d, Table S7). In the category of cellular component, cell part, cell, and organelle were most prevalent. Within the molecular function category, the most common terms were catalytic activity, binding, and transporter activity. Among biological process category, metabolic processes, cellular processes, and response to stimulus were the greatest abundance terms.Fig. 5 GO enrichment of DEGs identified in different comparison groups. a–d Represent the S1 vs S2, S2 vs S3, S3 vs S4, and S4 vs S5 comparison groups, respectively. S1, S2, S3, S4, and S5 represent A. tataricus were harvested on the 15th of August, September, October, November and December, respectively To further identify the metabolic pathways of the DEGs, they were annotated with KEGG. 7,187 (S1 vs S2), 9,930 (S2 vs S3), 2,278 (S3 vs S4), and 6,963 (S4 vs S5) DEGs were assigned to 143, 143, 141, and 141 KEGG pathways, respectively (Table S8). Among these, 32, 45, 27, and 39 pathways were significantly enriched with a p-value < 0.05, respectively (Table S9). Notably, phenylpropanoid biosynthesis (ko00940) and flavonoid biosynthesis (ko00941), which are involved in the flavonoid biosynthesis, were significantly enriched in both of the four comparison groups (Fig. 6, Table S9). In addition, the enrichment pathways of the four comparison groups (S1 vs S2, S2 vs S3, S3 vs S4, and S4 vs S5) could be further divided into six categories: metabolism, genetic information processing, cellular processes, environmental information processing, organismal systems, and human diseases. In the six categories, the metabolic category contained the most pathways in all four comparison groups.Fig. 6 Significantly enriched KEGG pathways (P < 0.05) from DEGs in different comparison groups. a–d Represent the S1 vs S2, S2 vs S3, S3 vs S4, and S4 vs S5 comparison groups, respectively. The enriched phenylpropanoid biosynthesis pathway and flavonoid biosynthesis pathway are labeled by the red frame. S1, S2, S3, S4, and S5 represent A. tataricus were harvested on the 15th of August, September, October, November and December, respectively Combined transcriptome and metabolome analysis revealed the biosynthesis of flavonoid in the roots of A. tataricus DEGs that encode enzymes related to flavonoid biosynthesis were screened out based on the richen KEGG pathways and gene functional annotation, among which 90 structural genes, including 35 PAL, 16 4CL, 11 C4H, 1 CHS, 10 CHI, 1 F3H, 2 F3′H, and 14 FLS, showed significant correlation either with total flavonoid content or 23 individual flavonoid (r > 0.8) (Fig. 7, Table S10). Kaempferol, a flavonoid, displays several pharmacological activities, such as anti-inflammatory, antioxidant, and antitumor (Imran et al. 2019). Ng et al. isolated kaempferol from A. tataricus, and studies have shown that it has outstanding antioxidant activity (Ng et al. 2003). Our correlation analysis showed that the expression levels of 1 C4H, 1 CHI, and 1 FLS gene were highly correlated with the kaempferol content, especially the FLS (Table S10). Therefore, we speculate that these three genes are the key genes regulating the synthesis of kaempferol in the roots of A. tataricus.Fig. 7 Correlation coefficient heatmap of flavonoid biosynthesis related pathway genes and flavonoids in A. tataricus roots. Each row represents a gene and each column represents a flavonoid. Red indicates positive correlation, while negative correlation is shown in green. PAL phenylalanine ammonia-lyase, 4CL 4 coumarate CoA ligase, C4H cinnamate-4-hydroxylase, CHI chalcone isomerase, CHS chalcone synthase, F3H flavanone 3-hydroxylase, F3´H flavonoid 3´-hydroxylase, FLS flavonol synthase TFs, such as AP2/ERF, bHLH, bZIP, MYB, NAC, and WRKY, have been reported to be involved in flavonoid synthesis (Nabavi et al. 2020; Hichri et al. 2011). DEGs that belonged to the above six classes of TFs were identified in our transcriptome data, and the correlation between the changes of total flavonoid content and the 6 TFs in the five different developmental stages of A. tataricus roots was further analyzed. As a result, 65 TFs, including 22 AP2/ERF, 7 bHLH, 5 bZIP, 8 MYB, 11 NAC, and 12 WRKY, showed significant correlation with total flavonoid content (r > 0.8) (Table S11). These TFs might contribute to flavonoid metabolites in the roots of A. tataricus. Weighted gene co-expression network analysis to identify differential genes related to flavonoid synthesis To get a comprehensive understanding of the relationship between A. tataricus root samples in 5 different developmental stages, samples were clustered based on fragments per kilobase of transcript per million fragments mapped (FPKM) of 83,398 DEGs. The cluster dendrogram showed that samples were divided into 2 main clusters with 3 subclusters (Fig. 8a). Cluster I included S1 and S2. S4 and S5 were grouped together and belong to cluster II along with S3. The results suggested that the gene expression levels during root development in A. tataricus changed greatly from S2 to S4, especially from S2 to S3. A weighted gene co-expression network analysis (WGCNA) was further performed to divide all EDGs into 12 distinct modules, labeled with different colors, in which genes in the same modules had high correlation coefficients (Fig. 8b). Furthermore, the total flavonoid content, kaempferol content, 5 anthocyanins contents, and the variation trends of 3 clusters (cluster I, cluster II, and cluster III) of flavonoid metabolites as traits data for the module–trait relationship analysis. Three modules, including brown, turquoise, and purple, had substantial positive link with total flavonoid content and cluster III. Green and yellow modules were significantly associated with 5 anthocyanins contents and cluster II. Blue, pink, and greenyellow modules were correlated with cluster I. Kaempferol content had highest correlation with blue module (Fig. 8c). As a result, these eight modules would be chosen as interesting modules for further investigation.Fig. 8 Identification of WGCNA modules. a Cluster dendrogram of 15 samples based on FPKM of 83,398 DEGs. b Hierarchical clustering tree. c Module–trait relationship analysis. The value inside each box represents Pearson’s correlation coefficient between the module with trait, and the number in each parentheses represents p-value. The color scale on the right represents the degree of correlation between modules and trait and the red represent high correlation To further identify the metabolic pathways of the DEGs in these 8 modules, they were annotated with KEGG. The results showed that the DEGs in these 8 modules were all enriched in flavonoid biosynthesis–related pathways (Fig. 9). One hundred and twenty-seven genes, including 43 PAL, 23 4CL, 9 C4H, 4 CHS, 18 CHI, 2 F3H, 5 F3'H, 1 F3′5'H, 20 FLS, and 2 FSII, that related to flavonoid biosynthesis were selected from brown, turquoise, and purple modules (Table S12). They may be related to the biosynthesis of flavonoids in cluster III and the accumulation of total flavonoids. In green and yellow modules, 30 genes, including 5 PAL, 9 4CL, 1 C4H, 2 CHI, 1 F3H, 1 DFR, 7 3AT, 1 BZ1, and 3 UGT79B1, were screened (Table S12). These genes may be involved in the biosynthesis of flavonoids in cluster II (including 5 anthocyanins) (Fig. 10). There were 24 genes (10 4CL, 6 C4H, 2 CHI, 3 F3H, and 3 FLS) in blue, pink, and greenyellow modules that may promote the biosynthesis of flavonoids in cluster I (Table S12). The blue module contains 7 4CL, 5 C4H, 2 CHI, 1 F3H, and 3 FLS, which may be the key genes regulating the biosynthesis of kaempferol (Table S12, Fig. 10).Fig. 9 KEGG histograms of different gene modules enriched in flavonoid biosynthesis–related pathways. The number outside the right square brackets of the column represents the ratio of the number of genes annotated to the pathway to the number of annotated genes. The numbers in the right square brackets of the columns indicate the number of genes annotated to the pathway Fig. 10 Biosynthetic pathway of flavonoids during root development in A. tataricus. The histogram displays the levels of kaempferol and 5 anthocyanins (The histogram of flavonoids are expressed as relative content). Error bars represent ± standard deviation (n = 3). Heatmap showing the expression patterns of the candidate structural genes involved in the regulation of flavonoid compounds during the root development in A. tataricus. The expression of each gene is represented by a different color. Red indicates high content of metabolite, while low content metabolite is shown in green. PAL phenylalanine ammonia-lyase, C4H cinnamic acid 4-hydroxylase, 4CL 4 coumarate CoA ligase, CHI chalcone isomerase, F3H flavanone 3-hydroxylase, FLS flavonol synthase, DFR dihydroflavonol 4-reductase, BZ1 anthocyanidin 3-O-glucosyltransferase, 3AT anthocyanidin 3-O-glucoside 6´´-O-acyltransferase, UGT79B1 anthocyanidin 3-O-glucoside 2´´´-O-xylosyltransferase. S1, S2, S3, S4, and S5 represent A. tataricus were harvested on the 15th of August, September, October, November and December, respectively Confirmation of DEGs relates to flavonoid biosynthesis using qRT-PCR To confirm the credibility of the transcriptome information, we further selected 12 structural genes and 8 TFs to validate their expression by qRT-PCR. As shown in Fig. 11, expression level of all of the selected genes displayed high consistency with the RNA-seq data.Fig. 11 qRT-PCR verified the DEGs related to flavonoid biosynthesis DEGs during aster root development. Relative expression levels of qRT-PCR were calculated using actin as a standard. Pearson correlation coefficients were calculated by comparing qRT-PCR and RNA-seq data for each gene across all samples. Error bars represent ± standard deviation (n = 3). PAL phenylalanine ammonia-lyase, C4H cinnamate-4-hydroxylase, FLS flavonol synthase. S1, S2, S3, S4, and S5 represent A. tataricus were harvested on the 15th of August, September, October, November and December, respectively Discussion A. tataricus has been used as a traditional medicinal herb in China for more than 2,000 years due to its medicinal properties, such as anti-oxidation, anti-inflammatory, and anti-cancer, for humans (Du et al. 2017; Li et al. 2021b; Su et al. 2019; Wang et al. 2020). Li et al. reported that the anti-lung cancer function of A. tataricus is closely related to flavonoids such as quercetin, kaempferol, isorhamnetin, and luteolin (Li et al. 2021b). The (2R,2′′R)-7-O-methyl–2,3,2′′,3′′–tetrahydrorobustaflavone, a flavonoid isolated from A. tataricus, remarkably inhibited the proliferation of A549 cancer cells (Chen et al. 2022). However, research about the molecular mechanism of flavonoid biosynthesis in A. tataricus has not been reported. In our study, the total flavonoid content of five stages of A. tataricus roots was quantified, and an integrated transcriptomic and metabolite profile analysis was performed to understand the flavonoid biosynthesis in the developing roots of A. tataricus. Quantification of total flavonoids showed that it continued to decrease from S2 to S4. However, there was no significant difference in the total flavonoid content of S1–S2 and S4–S5 (Fig. 1b). It is obvious that the aerial parts of A. tataricus grow rapidly from S1 to S2. The growth of the aerial part of A. tataricus became slow after S2, and gradually began to dry up until S4 (Fig. 1a). These results indicated that the content of total flavonoid in the roots of A. tataricus changed with the growth period. As the plant grows and develops, the total content of flavonoid in the plants is not constant. Xie et al. reported that the content of total flavonoids in Dryopteris erythrosora varied in different seasons (Xie et al. 2015). Huang et al. reported that the content of total flavonoid in the roots of Abrus cantoniensis was the highest in October, and decreased after October (Huang et al. 2006). The content of total flavonoid in Tartary buckwheat seeds showed a pattern of first increasing and then decreasing (Li et al. 2019). In addition, the accumulation of secondary metabolites is also affected by temperature. Alhaithloul et al. reported that low temperature stress promoted the content of total flavonoid in tomato (Solanum lycopersicum) seedlings, while the content of total flavonoid decreased under high temperature stress (Alhaithloul et al. 2021). Under low temperature (4 ℃) or high temperature (40 ℃) stress, the content of total flavonoid in basil (Ocimum basilicum cv. 'Genovese') leaves was higher than that under normal temperature (25 ℃) (Jakovljević et al. 2021). In our study, the content of total flavonoid in A. tataricus roots of S4 (the average daily maximum temperature is about 11 °C) was lower than that of A. tataricus roots of S2 (the daily average maximum temperature is about 27 °C). We speculate that the decrease in total flavonoid content in A. tataricus roots may be caused by the stagnation of A. tataricus growth after S2 due to the decrease in temperature. Through metabolite profiling analysis, a total of 80 flavonoids were identified in the roots of A. tataricus. All detected flavonoids had differential accumulation during A. tataricus root development (Table S2). Of these flavonoids, flavonols (31) and flavonoid (29) are the major flavonoid compounds (Fig. 1c, Table S2). Furthermore, 70% of these 80 flavonoids belong to glycosides (Table S3). The skeletons of most flavonoids in A. tataricus root were isorhamnetin, kaempferol, quecetin, and myricetin. It was reported that flavonoids such as isorhamnetin and kaempferol have the effects of cardiovascular and cerebrovascular protection, anti-tumor, anti-inflammatory, anti-oxidation, organ protection, and prevention of obesity (Gong et al. 2020; Imran et al. 2019). Liu et al. reported that isorhamnetin protected against liver fibrosis by reducing ECM formation and autophagy via inhibition of TGF–β1–mediated Smad3 and p38 MAPK signaling pathways (Liu et al. 2019). Liu et al. reported that kaempferol improves osteoporosis by downregulating miR-10a-3p and upregulating CXCL12 (Liu et al. 2021a). Quercetin might decrease the susceptibility of neutrophils to pro–inflammatory factors to achieve anti-inflammatory effects (Liu et al. 2005). So far, there are few studies on the types and accumulation patterns of flavonoids in A. tataricus roots. This study gives us a more comprehensive understanding of flavonoids in A. tataricus roots. Our study provides some guidance for future studies on the pharmacological effects of A. tataricus. Anthocyanins are secondary metabolites in plants of the flavonoid family. They are responsible for the vigorous colors of various botanic organs and are also substantial dietary compounds (Qin et al. 2010; Rodriguez-Saona and Wrolstad 2001; Salamone et al. 2012). Pelargonidin and cyanidin derivatives are the important pigments in bright red fruits (Jaakola 2013). Shen et al. reported that glycosidic pelargonidin is absent in white F. nilgerrensis fruits compared to red fruits (Shen et al. 2020). Miyazawa et al. reported that cyanidin 3-O-glucoside confers a red hue to fruits (Miyazawa et al. 1999). In our data, two pelargonidin derivatives and three cyanidin derivatives were detected in the roots of A. tataricus. The contents of these five anthocyanins were highest and most of the A. tataricus roots were purple-red in the S4 stage, when the temperature was lower compared to previous developmental stages. It has been demonstrated that low temperature induces anthocyanin synthesis in various species (Chalker-Scott 1999; Choi et al. 2009). Therefore, we speculate that the accumulation of anthocyanins induced by low temperature caused the reddening of A. tataricus roots. It is well known that the structural genes that encode enzymes are involved in the biosynthesis of flavonoids. In our study, based on KEGG enrichment analysis and gene functional annotation, 182 DEGs that encode enzymes associated with flavonoid biosynthesis, including PAL, 4CL, C4H, CHS, CHI, F3H, F3′H, F3′5′H, FLS, FSII, DFR, 3AT, BZ1, and UGT79B1, were identified by WGCNA and calculating Pearson correlation coefficients between genes and flavonoids. Of these DEGs, 129 DEGs (43 PAL, 23 4CL, 9 C4H, 4 CHS, 18 CHI, 3 F3H, 5 F3′H, 1 F3′5’H, 21 FLS, and 2 FSII) showed significant correlation with total flavonoid content. Gam et al. reported that the expression levels of CHI and FLS genes in Anoectochilus roxburghii are related to the total flavonoid content (Gam et al. 2020). Specific flavonoids biosynthesis is preceded by a general phenylalanine pathway involving PAL, C4H, and 4CL (Dong and Lin 2021). It is reported that PAL, C4H, and 4CL genes are related to the synthesis of flavonoids (Li et al 2015; Liu et al. 2006; Singh et al. 2009). Cheng et al. showed that PAL activity has linked to the concentration of anthocyanins in strawberry (Fragaria × ananassa) fruit (Cheng and Breen 1991). The p-coumaroyl–CoA produced in the general phenylalanine pathway is catalyzed by chalcone synthase (CHS) to generate chalcone. Through the catalysis of CHI, F3H, and FLS, chalcone generates dihydroflavone, dihydroflavonol, and flavonols sequentially (Liu et al. 2021b). CHI, the first reported enzyme involved in flavonoid biosynthetic pathway, is a key enzyme in flavonoid biosynthesis (McKhann et al. 1998). CHI over–expression could increase flavonol accumulation in Arabidopsis, and stimulate the accumulation of apigenin in Astragalus trigonus (Elatabi et al. 2021; Jiang et al. 2015). Flavonol synthase catalyzes the production of flavonols from dihydroflavonols, the flavonols is a subclass of the flavonoids. Overexpression of FLS gene in chrysanthemum morifolium can increase the content of flavanols in tobacco (Wang et al. 2021). The above studies showed that CHI and FLS genes can promote the accumulation of flavonols. Our study found 2 CHI and 10 FLS genes were highly associated with flavonols, such as kaempferol and isorhamnetin. These results indicated that these genes are the key genes involved in the regulation of flavonol synthesis in A. tataricus. In addition to the structural genes, many classes of regulatory genes, such as AP2/ERF, MYB–bHLH–WD40 complexes, bZIP, MYB, NAC, and WRKY have been identified to play a role in flavonoid biosynthesis in higher plants (Morishita et al. 2009; Terrier et al. 2009; Wang et al. 2018; Xu et al. 2015; Zhang et al. 2022; Zhao et al. 2021). The AP2/ERF transcription factors modulate the accumulation of flavonoid by regulating CHI in citrus (Zhao et al. 2021). In apple (Malus pumila Mill.), bZIP44 promotes anthocyanin in response to ABA by enhancing the binding of MYB1 to the promoters of downstream target genes (An et al. 2018). Overexpression of the Arabidopsis NAC078 transcription factor results in a significant increase in the transcriptional levels of genes related to flavonoid biosynthesis and the levels of anthocyanins in Arabidopsis under high–light (Morishita et al. 2009). Potato (Solanum tuberosum) StWRKY13 can promote anthocyanin biosynthesis by activating the transcription of flavonoid biosynthesis–related genes such as StCHS, StF3H, and StDFR in potato tubers (Zhang et al. 2021). In our study, based on transcriptome data, we found that 65 important TFs including AP2/ERFs, bHLH, bZIP, MYB, NACs, and WRKY showed higher correlation values (r > 0.8) with total flavonoid content (Table S11). These differentially expressed TFs might be candidate regulators of flavonoid synthesis in A. tataricus roots. In summary, the total flavonoid content was different in roots of A. tataricus at different developmental stages. Based on metabolome and transcriptome data, we revealed the flavonoid biosynthesis metabolic pathway in A. tataricus root. A total of 83,398 DEGs and 80 flavonoid metabolites were identified with differential accumulation during A. tataricus root development. The skeletons of most flavonoids in A. tataricus roots were isorhamnetin, kaempferol, quecetin, and myricetin. Classification of 80 flavonoids into three subgroups with different accumulation patterns base on a K-means clustering algorithm. The genes involved in flavonoid biosynthesis were identified by combined analysis of transcriptome and metabolome data of the roots of A. tataricus. Our results provide valuable information on understanding flavonoid compositions and accumulation patterns and the candidate genes involved in the flavonoid biosynthesis pathways in A. tataricus. Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (XLSX 4080 KB) Acknowledgements We thank Metware company for transcriptome and metabolome detection. Author contributions KJ, XZ, LL and SG conceived and designed the research. YM and SL prepared the experimental materials. XL, TY, and CW performed some experiments. KJ and XZ analyzed the data and wrote the manuscript. SG provided intellectual input and revised the manuscript. All authors read and approved the final manuscript. Funding This work was supported by the Key Research and Development Project of Hebei Province: Scientific and Technological Innovation Team of Modern Seed Industry of Traditional Chinese Medicine (21326312D), Hebei Chinese medicine industry Technical system (HBCT2018060201), and China Agriculture Research System of MOF and MARA (CARS-21). Availability of data and materials The whole set of raw data can be found in the national center for biotechnology information (NCBI) SRA database (accession number PRJNA874188). Declarations Conflict of interest The authors declare that they have no conflict of interest. Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Kaixuan Jia, Xiaoling Zhang have contributed equally to this work. ==== Refs References Alhaithloul HSA Galal FH Seufi AM Effect of extreme temperature changes on phenolic, flavonoid contents and antioxidant activity of tomato seedlings (Solanum lycopersicum L.) PeerJ 2021 9 e11193 10.7717/peerj.11193 34026345 An JP Yao JF Xu RR You CX Wang XF Hao YJ Apple bZIP transcription factor MdbZIP44 regulates abscisic acid-promoted anthocyanin accumulation Plant Cell Environ 2018 41 2678 2692 10.1111/pce.13393 29940702 Chalker-Scott L Environmental significance of anthocyanins in plant stress responses Photochem Photobiol 1999 70 1 9 10.1111/j.1751-1097.1999.tb01944.x Chen W Gong L Guo Z Wang W Zhang H Liu X A novel integrated method for large-scale detection, identification, and quantification of widely targeted metabolites: application in the study of rice metabolomics Mol Plant 2013 6 1769 1780 10.1093/mp/sst080 23702596 Chen YJ Wu H Wei ZY Chen LM Dong JJ Liu J Identification of chemical constituents in Aster tataricus by UHPLC-Q-TOF-MS Acta Pharm Sin 2019 54 1645 1654 10.16438/j.0513-4870.2019-0214 Chen Y Wu H Li Y Liu J Jia Z Xu W Aster tataricus attenuates asthma efficiently by simultaneously inhibiting tracheal ring contraction and inflammation Biomed Pharmacother 2020 130 110616 10.1016/j.biopha.2020.110616 32784051 Chen Z Zhang H Wang L Cai Q Li J Cao H Three cases of coronavirus disease 2019 (COVID-19) report J Tradit Chin Med 2020 61 745 748 10.13288/j.11-2166/r.2020.09.003 Chen T Yang P Chen HJ Huang B A new biflavonoids from Aster tataricus induced non–apoptotic cell death in A549 cells Nat Prod Res 2022 36 1409 1415 10.1080/14786419.2021.1882456 33615932 Cheng GW Breen PJ Activity of phenylalanine ammonia-lyase (PAL) and concentrations of anthocyanins and phenolics in developing strawberry fruit J Am Soc Hortic Sci 1991 116 865 869 10.21273/JASHS.116.5.865 Cheng D Shao Y Terpenoid glycosides from the roots of Aster tataricus Phytochemistry 1993 35 173 176 10.1016/S0031-9422(00)90528-4 Choi S Kwon YR Hossain MA Hong SW Lee B Lee H A mutation in ELA1, an age–dependent negative regulator of PAP1/MYB75, causes UV–and cold stress–tolerance in Arabidopsis thaliana seedlings Plant Sci 2009 176 678 686 10.1016/j.plantsci.2009.02.010 Dong NQ Lin HX Contribution of phenylpropanoid metabolism to plant development and plant–environment interactions J Integr Plant Biol 2021 63 80 209 10.1111/jipb.13054 Dong X Chen W Wang W Zhang H Liu X Luo J Comprehensive profiling and natural variation of flavonoids in rice J Integr Plant Biol 2014 56 76 886 10.1111/jipb.12204 Dong T Han R Yu J Zhu M Zhang Y Gong Y Anthocyanins accumulation and molecular analysis of correlated genes by metabolome and transcriptome in green and purple asparaguses (Asparagus officinalis, L.) Food Chem 2019 271 8 28 10.1016/j.foodchem.2018.07.120 Du H hang M, Yao K, Hu Z, Protective effect of Aster tataricus extract on retinal damage on the virtue of its antioxidant and anti-inflammatory effect in diabetic rat Biomed Pharmacother 2017 89 617 622 10.1016/j.biopha.2017.01.179 28262614 Elarabi NI Abdelhadi AA Sief-Eldein AG Ismail IA Abdallah NA Overexpression of chalcone isomerase A gene in Astragalus trigonus for stimulating apigenin Sci Rep 2021 11 1 9 10.1038/s41598-021-03704-y 33414495 Gam DT Khoi PH Ngoc PB Linh LK Hung NK Anh PTL LED lights promote growth and flavonoid accumulation of Anoectochilus roxburghii and are linked to the enhanced expression of several related genes Plants 2020 9 1344 10.3390/plants9101344 33053736 Gene Ontology Consortium Gene ontology consortium: going forward Nucleic Acids Res 2015 43 D1049 D1056 10.1093/nar/gku1179 25428369 Ghazalpour A Doss S Zhang B Wang S Plaisier C Castellanos R Integrating genetic and network analysis to characterize genes related to mouse weight PLoS Genet 2006 8 e130 10.1371/journal.pgen.0020130 Gong G Guan YY Zhang ZL Rahman K Wang SJ Zhou S Isorhamnetin: A review of pharmacological effects Biomed Pharmacother 2020 128 110301 10.1016/j.biopha.2020.110301 32502837 Hichri I Barrieu F Bogs J Kappel C Delrot S Lauvergeat V Recent advances in the transcriptional regulation of the flavonoid biosynthetic pathway J Exp Bot 2011 62 2465 2483 10.1093/jxb/erq442 21278228 Hossain MA Rahman SM Total phenolics, flavonoids and antioxidant activity of tropical fruit pineapple Food Res Int 2011 44 672 676 10.1016/j.foodres.2010.11.036 Huang R Yu Y Hu Y Sheng X Determination of total flavonoids in Abrus cantoniensis and its dynamic changes Zhongguo Zhongyao Zazhi 2006 17 1428 1431 10.3321/j.issn:1001-5302.2006.17.009 Imran M Salehi B Sharifi-Rad J Aslam Gondal T Saeed F Imran A Kaempferol: A key emphasis to its anticancer potential Molecules 2019 24 2277 10.3390/molecules24122277 31248102 Jaakola L New insights into the regulation of anthocyanin biosynthesis in fruits Trends Plant Sci 2013 18 477 483 10.1016/j.tplants.2013.06.003 23870661 Jakovljević D Momčilović J Bojović B Stanković M The short–term metabolic modulation of basil (Ocimum basilicum L. cv. ‘Genovese’) after exposure to cold or heat Plants (basel) 2021 10 590 10.3390/plants10030590 33801088 Jiang W Yin Q Wu R Zheng G Liu J Dixon RA Role of a chalcone isomerase–like protein in flavonoid biosynthesis in Arabidopsis thaliana J Exp Bot 2015 66 7165 7179 10.1093/jxb/erv413 26347569 Kanehisa M Sato Y Kawashima M Furumichi M Tanabe M KEGG as a reference resource for gene and protein annotation Nucleic Acids Res 2016 44 D457 D462 10.1093/nar/gkv1070 26476454 Li Y Kim JI Pysh L Chapple C Four isoforms of Arabidopsis 4-coumarate: CoA ligase have overlapping yet distinct roles in phenylpropanoid metabolism Plant Physiol 2015 169 2409 2421 10.1104/pp.15.00838 26491147 Li H Lv Q Ma C Qu J Cai F Deng J Metabolite profiling and transcriptome analyses provide insights into the flavonoid biosynthesis in the developing seed of tartary buckwheat (Fagopyrum tataricum) J Agric Food Chem 2019 67 11262 11276 10.1021/acs.jafc.9b03135 31509416 Li BZ Fan RN Sun GL Sun T Fan YT Bai SL Flavonoids improve drought tolerance of maize seedlings by regulating the homeostasis of reactive oxygen species Plant Soil 2021 461 389 405 10.1007/s11104-020-04814-8 Li MZ Yang Z Wu Q Sun DX Chen G Tang DX Investigation of anti-lung cancer mechanism of herbal pair of semen Lepidii seu Descurainias and Radix Asteris base on network pharmacology J Oncol Chin Med 2021 3 13 20+7 10.19811/j.cnki.ISSN2096-6628.2021.01.003 Liu JJ Song CW Yue Y Duan CG Yang J He T Quercetin inhibits LPS–induced delay in spontaneous apoptosis and activation of neutrophils Inflamm Res 2005 54 500 507 10.1007/s00011-005-1385-2 16389571 Liu R Xu S Li J Hu Y Lin Z Expression profile of a PAL gene from Astragalus membranaceus var. Mongholicus and its crucial role in flux into flavonoid biosynthesis Plant Cell Rep 2006 25 705 710 10.1007/s00299-005-0072-7 16456646 Liu N Feng J Lu X Yao Z Liu Q Lv Y Isorhamnetin inhibits liver fibrosis by reducing autophagy and inhibiting extracellular matrix formation via the TGF–β1/Smad3 and TGF–β1/p38 MAPK pathways Mediators Inflamm 2019 10.1155/2019/6175091 Liu H Yi X Tu S Cheng C Luo J Kaempferol promotes BMSC osteogenic differentiation and improves osteoporosis by downregulating miR-10a-3p and upregulating CXCL12 Mol Cell Endocrinol 2021 520 111074 10.1016/j.mce.2020.111074 33157164 Liu W Feng Y Yu S Fan Z Li X Li J The flavonoid biosynthesis network in plants Int J Mol Sci 2021 22 12824 10.3390/ijms222312824 34884627 Love MI Huber W Anders S Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2 Genome Biol 2014 15 1 21 10.1186/s13059-014-0550-8 McKhann HI Paiva NL Dixon RA Hirsch AM Expression of genes for enzymes of the flavonoid biosynthetic pathway in the early stages of the Rhizobium–legume symbiosis Adv Exp Med Biol 1998 439 45 54 10.1007/978-1-4615-5335-9_4 9781294 Miyazawa T Nakagawa K Kudo M Muraishi K Someya K Direct intestinal absorption of red fruit anthocyanins, cyanidin-3-glucoside and cyanidin–3,5–diglucoside, into rats and humans J Agric Food Chem 1999 47 1083 1091 10.1021/jf9809582 10552420 Morishita T Kojima Y Maruta T Nishizawa-Yokoi A Yabuta Y Shigeoka S Arabidopsis NAC transcription factor, ANAC078, regulates flavonoid biosynthesis under high–light Plant Cell Physiol 2009 50 2210 2222 10.1093/pcp/pcp159 19887540 Nabavi SM Šamec D Tomczyk M Milella L Russo D Habtemariam S Flavonoid biosynthetic pathways in plants: Versatile targets for metabolic engineering Biotechnol Adv 2020 38 107316 10.1016/j.biotechadv.2018.11.005 30458225 Nakatsuka T Abe Y Kakizaki Y Yamamura S Nishihara M Production of red–flowered plants by genetic engineering of multiple flavonoid biosynthetic genes Plant Cell Rep 2007 26 1951 1959 10.1007/s00299-007-0401-0 17639403 Ng TB Liu F Lu Y Cheng CHK Wang Z Antioxidant activity of compounds from the medicinal herb Aster tataricus Comp Biochem Physiol Part C Toxicol Pharmacol 2003 136 109 115 10.1016/s1532-0456(03)00170-4 Olivas-Aguirre FJ Rodrigo-García J Martínez-Ruiz ND Cárdenas-Robles AI Mendoza-Díaz SO Álvarez-Parrilla E Cyanidin-3-O-glucoside: physical-chemistry, foodomics and health effects Molecules 2016 21 1264 10.3390/molecules21091264 27657039 Qin C Li Y Niu W Ding Y Zhang R Shang X Analysis and characterisation of anthocyanins in mulberry fruit Czech J Food Sci 2010 28 117 126 10.17221/228/2008-cjfs Quiroz A Mendez L Mutis A Hormazabal E Ortega F Birkett MA Parra L Antifeedant activity of red clover root isoflavonoids on Hylastinus obscurus J Soil Sci Plant Nutr 2017 17 231 239 10.4067/S0718-95162017005000018 Ren W Ma Y Wang R Liang P Sun Q Pu Q Research advance on Qingfei Paidu Decoction in prescription principle, mechanism analysis and clinical application Front Pharmacol 2021 11 589714 10.3389/fphar.2020.589714 33584265 Rodriguez-Saona LE Wrolstad RE Extraction, isolation, and purification of anthocyanins Curr Protoc Food Anal Chem 2001 1 F.1.1.1 F1.1.11 10.1002/0471142913.faf0101s00 Salamone F Li Volti G Titta L Puzzo L Barbagallo L La Delia F Moro orange juice prevents fatty liver in mice World J Gastroenterol 2012 18 3862 3868 10.3748/wjg.v18.i29.3862 22876038 Schmittgen TD Livak KJ Analyzing real-time PCR data by the comparative CT method Nat Protoc 2008 3 1101 1108 10.1038/nprot.2008.73 18546601 Shen JC Shao WL Du ZJ lu HF Li JM Integrated metabolomic and transcriptomic analyses reveal differences in the biosynthetic pathway of anthocyanins in Fragaria nilgerrensis and Fragaria pentaphylla Sci Hortic 2020 271 109476 10.1016/j.scienta.2020.109476 Singh K Kumar S Rani A Gulati A Ahuja PS Phenylalanine ammonia-lyase (PAL) and cinnamate 4-hydroxylase (C4H) and catechins (flavan-3-ols) accumulation in tea Funct Integr Genom 2009 9 125 134 10.1007/s10142-008-0092-9 Su XD Jang HJ Li HX Kim YH Yang SY Identification of potential inflammatory inhibitors from Aster tataricus Bioorg Chem 2019 92 103208 10.1016/j.bioorg.2019.103208 31473471 Sun Y Li L Liao M Su M Wan C Zhang L A systematic data acquisition and mining strategy for chemical profiling of Aster tataricus rhizoma (Ziwan) by UHPLC-Q-TOF-MS and the corresponding anti-depressive activity screening J Pharm Biomed Anal 2018 154 216 226 10.1016/j.jpba.2018.03.022 29573734 Terrier N Torregrosa L Ageorges A Vialet S Verries C Cheynier V Ectopic expression of VvMybPA2 promotes proanthocyanidin biosynthesis in grapevine and suggests additional targets in the pathway Plant Physiol 2009 149 1028 1041 10.1104/pp.108.131862 19098092 Treutter D Significance of flavonoids in plant resistance and enhancement of their biosynthesis Plant Biol 2005 7 581 591 10.1055/s-2005-873009 16388461 Varet H Brillet-Guéguen L Coppée JY Dillies MA SARTools: a DESeq2–and EdgeR–based R pipeline for comprehensive differential analysis of RNA-Seq data PLoS One 2016 11 e0157022 10.1371/journal.pone.0157022 27280887 Wang N Liu W Zhang T Jiang S Xu H Wang Y Transcriptomic analysis of red–fleshed apples reveals the novel role of MdWRKY11 in flavonoid and anthocyanin biosynthesis J Agric Food Chem 2018 66 7076 7086 10.1021/acs.jafc.8b01273 29909630 Wang X Fan L Yin H Zhou Y Tang X Fei X Protective effect of Aster tataricus extract on NLRP3-mediated pyroptosis of bladder urothelial cells J Cell Mol Med 2020 24 13336 13345 10.1111/jcmm.15952 33030301 Wang Y Zhou LJ Wang Y Liu S Geng Z Song A Functional identification of a flavone synthase and a flavonol synthase genes affecting flower color formation in Chrysanthemum morifolium Plant Physiol Biochem 2021 166 1109 1120 10.1016/j.plaphy.2021.07.019 34328869 Wen K Fang X Yang J Yao Y Nandakumar KS Salem ML Recent research on flavonoids and their biomedical applications Curr Med Chem 2021 28 1042 1066 10.2174/0929867327666200713184138 32660393 Xie Y Zheng Y Dai X Wang Q Cao J Xiao J Seasonal dynamics of total flavonoid contents and antioxidant activity of Dryopteris erythrosora Food Chem 2015 186 113 118 10.1016/j.foodchem.2014.05.024 25976799 Xu W Dubos C Lepiniec L Transcriptional control of flavonoid biosynthesis by MYB–bHLH–WDR complexes Trends Plant Sci 2015 20 176 185 10.1016/j.tplants.2014.12.001 25577424 Yu P Cheng S Xiang J Yu B Zhang M Zhang C Expectorant, antitussive, anti-inflammatory activities and compositional analysis of Aster tataricus J Ethnopharmacol 2015 164 328 333 10.1016/j.jep.2015.02.036 25701752 Zhang H Zhang Z Zhao Y Guo D Zhao X Gao W StWRKY13 promotes anthocyanin biosynthesis in potato (Solanum tuberosum) tubers Funct Plant Biol 2021 49 102 114 10.1071/FP21109 34794538 Zhang F Huang J Guo H Yang C Li Y Shen S OsRLCK160 contributes to flavonoid accumulation and UV-B tolerance by regulating OsbZIP48 in rice Sci China Life Sci 2022 10.1007/s11427-021-2036-5 Zhao DX Hu BQ Zhang M Zhang CF Xu XH Simultaneous separation and determination of phenolic acids, pentapeptides, and triterpenoid saponins in the root of Aster tataricus by high-performance liquid chromatography coupled with electrospray ionization quadrupole time-of-flight mass spectrometry J Sep Sci 2015 38 571 575 10.1002/jssc.201401008 25491750 Zhao C Liu X Gong Q Cao J Shen W Yin X Three AP2/ERF family members modulate flavonoid synthesis by regulating type IV chalcone isomerase in citrus Plant Biotechnol J 2021 19 671 688 10.1111/pbi.13494 33089636

36520245

PMC9753034

NO-CC CODE

2022-12-16 23:25:27

no

J Plant Res. 2022 Dec 15;:1-18

utf-8

J Plant Res

10.1007/s10265-022-01426-4

oa_other

==== Front Intensive Care Med Intensive Care Med Intensive Care Medicine 0342-4642 1432-1238 Springer Berlin Heidelberg Berlin/Heidelberg 6950 10.1007/s00134-022-06950-4 Editorial Intensive Care Medicine facing the future Citerio Giuseppe giuseppe.citerio@unimib.it grid.7563.7 0000 0001 2174 1754 School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy 15 12 2022 14 © Springer-Verlag GmbH Germany, part of Springer Nature 2022 This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. ==== Body pmcAfter the tsunami caused by the coronavirus disease 2019 (COVID-19) pandemic [1], Intensive Care Medicine (ICM) has returned to “normal” life, as most medical journals did. We had anticipated this in an editorial published during the COVID-19 waves [2]. The impressive impact factor released in 2022, which surpassed 41, documented that during the storm ICM kept its line and attracted, in a timely manner, manuscripts that have been heavily quoted. We strived to control the speed and the quality of the journal with a small and efficient team of committed Editors and an incredible editorial office. In addition to this exceptional result, the number of articles downloaded has been skyrocketing, surpassing 10 million per year, and our presence on social media remains massive. The number of manuscripts submitted to ICM has returned to the levels prior to the COVID-19 explosion, and the Journal remains focused on clinical intensive care medicine. The new year offers an opportunity to stop and think about our past and future. The first thought goes to our community, composed of a vast readership, devoted reviewers (listed at the end of this editorial), brilliant authors, and tireless editors to whom we gratefully express our warmest thanks. ICM is your journal, and its success is our success. ICM is always keen to receive the best clinical research and contributions on all aspects of intensive care medicine, from large, randomized clinical trials to personal reflections on being an ICU patient, family, or staff member. We remain focused on keeping the competitive speed of the peer-review and publication processes and on ensuring the quality of the articles we publish, including commissioned pieces enriched by nice, custom-made, illustrations. ICM is proud to be the scientific voice of the European Society of Intensive Care Medicine (ESICM) while keeping its independent editorial line. With ESICM the journal started to develop trustworthy guidelines with the involvement of worldwide clinical experts and ESICM methodologists aimed at helping clinicians at the bedside. In 2023, updated guidelines on the management of acute respiratory distress syndrome (ARDS), fluid management and antibiotic stewardship are expected to be released. ICM is also opening a new topical collection, “My green ICU”, on the climate responsibilities in intensive care medicine, aimed to promote the reduction of energy use, life cycle assessments, ICU recycling, avoidance of futility and “less is more” in the technological ICU environment for a climate protective campaign. The connection with ESICM has never been stronger. From January 2023, ICM, until now co-owned by ESICM and the European Society of Paediatric and Neonatal Intensive Care (ESPNIC), will be owned entirely by ESICM as ESPNIC launches a new open-access Journal, ICM Paediatric and Neonatal (ICMpn), guided by Dick Tibboel. The ICM journal family is thus growing with the new ICM Paediatric and Neonatal and ICM Experimental as “sister” journals. At the end of 2024, the mandate of the Editor -in-Chief (EiC) will expire. Serving the Journal in the latest years and watching its continuous growth during the last decade has been an incredible experience. To continue along this path, a call for a new Editor-in-Chief will be launched to find a visionary, committed Editor who will guide the Journal in the future. Last, but not least, we would like to thank the reviewers who supported ICM throughout 2022 and took the necessary time and effort to review the submitted manuscripts. We sincerely appreciate all their valuable comments and suggestions, which helped ICM in improving the quality of the manuscripts. Top Reviewers (in alphabetical order; more than 15 reviews in 2022) Thomas Bein, Warwick Butt, Davide Chiumello, Audrey De Jong, Hans Flaatten, Luciano Gattinoni, Boris Jung, Marc Leone, Daniele Poole, Paola Rebora, Chiara Robba. Reviewers (in alphabetical order) Yassir Aarab, Aoife Abbey, Hafiz Abdul-Aziz, Osama Abou Arab, Darryl Abrams, Gareth Lewis Ackland, Stefan Acosta, Neill Adhikari, Rachel S. Agbeko, Nadia Aissaoui, Hafid Ait-Oufella, Evangelia Akoumianaki, Guillermo Albaiceta, Jan-Willem Alffenaar, Waleed Alhazzani, Marlon Aliberti, Andre Amaral, Lar Wiuff Andersen, Nina Christine Andersen-Ranberg, Anders Aneman, Eduardo Angles-Cano, Federico Angriman, Derek C. Angus, Ghada Ankawi, Djillali Annane, Sardar Ansari, Matthew H. Anstey, David Antcliffe, Christian Apfelbacher, Andrew Argent, Jean-Michel Arnal, Nitin Arora, Antonio Artigas, Nishkantha Arulkumaran, Karim Asehnoune, Pierre Asfar, Selena Au, Cécile Aubron, Gerard, Audibert, Luciano Cesar Pontes Azevedo, Elie Azoulay, Rafael Badenes, Sean M. Bagshaw, Heather Baid, John Kenneth Baillie, Sébastien Bailly, Jan Bakker, Bruno Guedes Baldi, Jonathan Ball, Arun Kumar Baranwal, Ryan P. Barbaro, Carmen Sílvia Valente Barbas, Ferran Barbé, François Barbier, Matthew Barhight, Tatiana Barichello, Michael David Barnett, Erin F. Barreto, Jason Batten, Philippe R. Bauer, Michael Bauer, Danielle Bear, Jacqueline Becker, Jeremy R. Beitler, Max Bell, Giacomo Bellani, Giuseppe Bellelli, Francois Beloncle, Julie Sarah Benbenishty, Dominique Benoit, Mette M. Berger, Tobias Bergler, Jesus F. Bermejo-Martin, Lorenzo Berra, Laurent Bertoletti, Bruno Adler Maccagnan Pinheiro Besen, Daniele Guerino Biasucci, O. Joseph Bienvenu, Shailesh Bihari, Thomas Billyard, Federico Bilotta, Julian Bion, Brittany Bissell, Laurent Bitker, Erin Blakeney, Lluis Blanch, Frank Bloos, Stijn Blot, Yelena Bodien, Bernd W. Boettiger, Joachim Boldt, Bernardo Bollen Pinto, Ezio Bonanomi, Timothy Bonnici, Lieuwe D. Bos, Somnath Bose, Josee Bouchard, Belaid Bouhemad, Carole Boulanger, Arnaud Bourdin, Mohamed Boussarsar, Salah Boussen, Lionel Bouvet, Hendrik Bracht, Sally Brady, Bruna Brandao Barreto, Richard Branson, Susan L. Bratton, Jeffrey Brent, Josef Briegel, Laurent J. Brochard, Daniel Brodie, Karim Brohi, Roy Brower, Kate L. Brown, Nicolas J. Bruder, Niccolo Buetti, Aidan JC Burrell, Lisa D. Burry, Katharina Busl, Pietro Caironi, Conceição Calhau, Kristin Canavera, Mathieu Capdevila, Gilles Capellier, Maurizia Capuzzo, Joe Carcillo, Pablo Cardinal-Fernández, Anselmo Caricato, Alain Cariou, Jean Carlet, Ana Paula de Carvalho Panzeri Carlotti, Eric Carlton, Michael Carter, Agostinho Carvalho, Michael Paul Casaer, Michael Casaer, Maurizio Cecconi, Maurizio Cereda, Abhimanyu Chandel, Gerald Chanques, Sanjay Chawla, Yih-Sharng Chen, Jonathan Chun-Hei Cheung, Michelle Chew, Minesh Chotalia, Kenneth B. Christopher, Raphael Cinotti, Giuseppe Citerio, Jan Claassen, William Clark, Sara Clohisey, Davide Colombo, Kirsten Colpaert, Alain Combes, Jean-Michel Constantin, Andrew Conway Morris, Deborah J. Cook, Rodrigo Cornejo, Francesco Corradi, Andrea Cortegiani, Howard Corwin, Andrea Costamagna, Jean-Maxime Côté, Rémi Coudroy, Francis Couturaud, Thomas Craven, Jose Melo Cristino, J. Randall Curtis, Felipe Dal Pizzol, Heidi Dalton, Neha Subhash Dangayach, Michael Darmon, Camille Daste, Dieter Frans Dauwe, Marcelo Gama de Abreu, Daniel De Backer, Liesbet De Bus, Dylan W. de Lange, Daniele De Luca, Gennaro De Pascale, Nicolas de Prost, Silvia De Rosa, Greet De Vlieger, Steven Deem, Akash Deep, Freda DeKeyser Ganz, Lorenzo Del Sorbo, Anthony Delaney, Julien Demiselle, Alexandre Demoule, Mark Dennis, Pieter Depuydt, Lars Desmet, Allan Detsky, Shelly Dev, John Devlin, Mark Devonald, Rajat Dhar, Joanna Colleen Dionne, Michael N. Diringer, Ilija Djordjevic, Annemarie Beth Docherty, Thomas Donaldson, Dirk Donker, martin dres, Jun Duan, Julie Dubourg, Guillaume Dumas, Jacques Duranteau, Mark Edwards, Ingrid Egerod, Moritoki Egi, Stephan Ehrmann, Sharon Einav, Maarten Eisma, Paul Elbers, Jonathan Elmer, E. Wesley Ely, Christian Emsden, Zoltan H. Endre, Andres Esteban, Angel Estella, Stine Estrup, Lisbeth Evered, Beverley Anne Ewens, Eddy Fan, Vito Fanelli, Jennifer Fang, David Faraoni, Chris Farmer, Aarne Feldheiser, Niall D. Ferguson, Fatima Carneiro Fernandes, Rafael Fernandez, Shannon M. Fernando, Luis Ferreira Moita, Wojciech Filipiak, Matthew Fish, Richard Fisher, Christoph Fisser, Ana Bruschy Fonseca, Lui Forni, Giuseppe Foti, Robert Fowler, Bruno Francois, Corinne Frere, Laura Galarza, Stefania Galimberti, Adriank Gallardo, Allan Garland, José Garnacho-Montero, Marc Garnier, Nicolas Gaspard, Stephane Gaudry, Etienne Gayat, Diego Gazzolo, Thomas Geeraerts, Dimitrios Georgopoulos, Hayley B. Gershengorn, Maddalena Giannella, Alberto Giannini, Timothy D. Girard, massimo girardis, Armand Girbes, Thomas Godet, Alberto Goffi, Ewan Christopher Goligher, João Gonçalves-Pereira, Anthony C. Gordon, Johannes Grand, Wilson Grandin, Asger Granfeldt, Cristina Granja, Giacomo Grasselli, Francesca Graziano, Massimiliano Greco, David Greer, Cesare Gregoretti, Domenico Luca Grieco, Antoine Grillon, David Grimaldi, Anne-Marie Guerguerian, Claude Guérin, Bertrand Guidet, Antoine Guillon, Jan Gunst, Alejandra Gutierrez, Bas Haak, Ryan W. Haines, David Hajage, Sophie Rym Hamada, Nadjib Hammoudi, Rashan Haniffa, Fraser Hanks, Michael Oscar Harhay, David A. Harrison, Anatole Harrois, Laura Hawryluck, Gregory W. J. Hawryluk, Guillaume Hekimian, Raimund Helbok, Julie Helms, Claude Hemphill, Jeroen Hermanides, Greet Hermans, Gonzalo Hernandez, Glenn Hernandez, Margaret Herridge, Christian von Heymann, Michael Joerg Hiesmayr, Julie Highfield, Stephanie Hill, Peter Buhl Hjortrup, Carol Hodgson, Cornelia Hoedemaekers, Martin Hoenigl, Edward Hoffer, Magdalena Hoffmann, Mathias Johan Holmberg, Patrick M. Honore, Janneke Horn, Megan Hosey, Eric Hoste, Sami Hraiech, May Hua, James Hurley, Faeq Husain-Syed, Sam Hutchings, Robert Hyzy, Gaetano Iapichino, Rosenfeld J, Matthieu Jabaudon, Samir Jaber, Emma Jackson, Matthias Jaeger, Arthur James, Paul S. Jansson, Marc G. Jeschke, Olivier Joannes-Boyau, Michael Joannidis, Tiffanie Jones, Mercedes Jourdain, Mathieu Jozwiak, Nicole P. Juffermans, Christian Jung, Pierre Kalfon, Andre C. Kalil, Constantinos Kanaris, Sandra Kane-Gill, Christian Karagiannidis, Narayan Karunanithy, Kianoush Kashani, Nancy Kentish-Barnes, Thomas Kerforne, Jozef Kesecioglu, Michelle Kho, Marion Kibler, Antoine Kimmoun, Detlef Kindgen-Milles, Derek J.B. Kleinveld, Michael Klompas, Kada Klouche, Stefan Kluge, Martin Kneyber, Matthieu Komorowski, Matthieu Komorowski, Rafal Kopanczyk, Katarzyna Laura Kotfis, Jay L. Koyner, John Kress, Sapna Kudchadkar, Michael Alexander Kuiper, Pedro Kurtz, Erwan L'Her, Norbert Lameire, Giulia Lamiani, Francois Lamontagne, Christian Lanckohr, Giovanni Landoni, Theis Lange, Thomas Langer, Stephen Lapinski, Romaric Larcher, Jean Baptiste Lascarrou, Sigismond Lasocki, Jos M. Latour, Nicola Latronico, Kevin B. Laupland, Alexandra Laurent, Clément Le Bihan, David Leaf, Clement Lebihan, Thomas Lecompte, Pierre-Louis Leger, Matthieu Legrand, Stephane Legriel, François Lellouche, virginie lemiale, François Lersy, Jerrold H. Levy, Bruno Levy, Mitchell Levy, Jie Li, Chunsheng Li, Daniel Lichtenstein, Lindsay Lief, Lowell Ling, Jeffrey Lipman, T. Lisboa, Vincent Liu, Suzana Margareth Lobo, Olivier Lortholary, Alawi Luetz, Roman Lukaszewski, Nuttha Lumlertgul, Tommaso Lupia, Charles-Edouard Luyt, Andrew I. R. Maas, Graeme MacLaren, Robert MacLaren, Salvatore Maurizio Maggiore, Ula Mahadeva, Yazine Mahjoub, Kathryn Maitland, Manu L.N.G. Malbrain, Maximilian Valentin Malfertheiner, John Marini, Eric Mariotte, Hugo Marques, John Marshall, Thomas Reed Martin, Greg S. Martin, Ignacio Martin-Loeches, Gennaro Martucci, Paul Masi, David Maslove, Gerald Matchett, Michael A. Matthay, Tommaso Mauri, Eric Maury, Stephan Mayer, Paul H. Mayo, Florian Mayr, Manuel Mayr, Monty Mazer, Danny McAuley, Forbes McGain, David Mcilroy, Bairbre McNicholas, Joanne McPeake, Gianfranco Umberto Meduri, Melanie Meersch, Bruno Megarbane, Sangeeta Mehta, Puja Mehta, Nilesh Mehta, Andreas Meiser, Armand Mekontso Dessap, Livia Melro, Colin Melville, David Menon, Spyros D. Mentzelopoulos, Sebastiano Mercadante, Hamid Merdji, Antonio Messina, Victoria Metaxa, Nuala J. Meyer, Geert Meyfroidt, Frédéric Michard, Olivier Mimoz, Eduardo Mireles-Cabodevila, Giovanni Mistraletti, Nicolas Molinari, Morten Hylander Møller, Celine Monard, Patricia Moniz, Guillaume Monneret, Xavier Monnet, Antoine Monsel, Lynne Moore, Alessandro Morandi, Rui Moreno, marc moss, Damien Motavasseli, Anna Motos, Paul Mouncey, Pierre-Henri Moury, Thomas Muders, Laurent Muller, Laveena Munshi, Patrick T. Murray, Raghavan Murugan, John Muscedere, Sheila Nainan Myatra, Marek Nalos, Antonio Paulo Nassar Jr, Mai-Anh Nay, Edward Needham, Sarah Nelson, Thanh Neville, Virginia Newcombe, Javier A. Neyra, Tri-Long Nguyen, Alistair Nichol, Nathan Dean Nielsen, Jacob Ninan, Jacob Ninan, Akira Nishisaki, Vandack Nobre, Saad Nseir, Peter Nydahl, Mauro Oddo, Yewande Odeyemi, Theresa Olasveengen, Olusegun Olusanya, Marlies Ostermann, Mehdi Oualha, Valerie Page, Paulo Paixao, Paul M. M. Palevsky, Pratik Pandharipande, Marios Papadopoulos, Laurent Papazian, Ken Kuljit Singh Parhar, Marcelo Park, Soojin Park, Bhakti Patel, Brijesh Patel, Christopher Patriquin, Jean-Francois Payen, Didier Payen, Sandra Peake, Marcus Peck, Paolo Pelosi, Frederic Pene, Joris Pensier, Roger Peredes, Bruno Pereira, Gavin Perkins, Anders Perner, Mark J. Peters, Matteo Petrosino, Carmen Andrea Andrea Pfortmueller, Barbara Janet Philips, Michael Piagnerelli, Peter Pickkers, David Pilcher, Marc Pineton De Chambrun, Michael R. Pinsky, Dominique Piquette, Lise Piquilloud, Romain Pirracchio, Gina Piscitello, Gaël Piton, Simone Piva, Julien Poissy, Kees Polderman, Garyphallia Poulakou, Pedro Póvoa, Jonah Powell-Tuck, Geraldine Sarah Power, Jean-Charles Preiser, Hallie Prescott, Susanna Price, Alessandro Protti, Catherine Proulx, Kathleen Puntillo, Zudin Puthucheary, Kathryn A. Puxty, Edward Qian, Jean-Pierre Quenot, Michael Quintel, Alejandro A. Rabinstein, Padmanabhan Ramnarayan, Vito Marco Ranieri, Otavio T. Ranzani, Samiran Ray, Jean Reignier, Annika Reintam Blaser, Annika Reintam Blaser, Dinis Reis Miranda, Jordi Rello, Steve Reynolds, Emanuele Rezoagli, Gabriela Ribeiro, Jean-Damien Ricard, Jean Christophe Marie Richard, Deborah Ridout, Jordi Riera, Reimer Riessen, Cassia Righy, Thomas Julien Rimmele, Giuseppe Ristagno, Jason Roberts, Oriol Roca, Angelo Rocha, Bram Rochwerg, Camila Rodrigues, Julie Rogan, Angela Rogers, Antoine Roquilly, Regis Goulart Rosa, Louise Rose, Peter Rosenberger, Andrea Rossetti, Anahita Rouze, Hadrien Roze, Sacha Rozencwajg, Kristina E. Rudd, Kurt Ruetzle, James A. Russell, Lene Russell, Vincenzo Russotto, Peter Sackey, Jorge Salluh, Claudio Sandroni, Filippo Sanfilippo, Fabio Sangalli, Catherine S. Sassoon, Maximilian Sebastian Schaefer, Stefan J. Schaller, Miet Schetz, Matthieu Schmidt, Mathieu Schmidt, Katharina R.L. Schmitt, Ronny Schnabel, Antoine G. Schneider, Brendon Scicluna, Mypinder Sekhon, Rodrigo Serafim, Ary Serpa Neto, Shahzad Shaefi, Manu Shankar-Hari, Sameer Sharif, Tarek Sharshar, Yahya Shehabi, Kiran Shekar, Jodi Sherman, Benjamin Shickel, Lori Shutter, Edward Siew, Stein Silva, Jonathan A. Silversides, Jon Silversides, Mervyn Singer, Pierre Singer, Benjamin Singer, Pratik Sinha, Alessandro Sionis, Fredrik Sjovall, Michael Sklar, Markus B. Skrifvars, Yoanna Skrobik, Michel Slama, Michael Smith, Jasmeet Soar, Ryann Sohaney, Romain Sonneville, Vicente Souza-Dantas, Savino Spadaro, Claudia D. Spies, Peter E. Spronk, Vijay Srinivasan, Nattachai Srisawat, Thomas Staudinger, Irene Steinberg, Scott Stephens, Nino Stocchetti, Alexander Supady, Timothy Sweeney, Daniel Sweeney, Alexis Tabah, Fabio Silvio Taccone, Jean-Marc Tadié, Daniel Talmor, Kenichi Tanaka, Leandro Utino Taniguchi, Guido Tavazzi, Jean-Louis Teboul, Armando Teixeira-Pinto, Irene Telias, Ruben Thanacoody, Holger Thiele, Thomas Thiele, Arnaud W. Thille, Dick Tibboel, Shane M. Tibby, Jean-François Timsit, Joseph Tonna, Antoni Torres, Pieter Tuinman, Chris Turner, Roman Ullrich, Takeshi Unoki, Charles-Hervé Vacheron, Emily Anne Vail, Andreas Valentin, Mark van den Boogaard, Iwan van der Horst, Mathieu van der Jagt, Sean van Diepen, Pouline van Oort, Niels Van Regenmortel, Trevor Van Schooneveld, Rosanna Vaschetto, Bala Venkatesh, Peter Verhamme, James Versalovic, Antoine Vieillard-Baron, Elizabeth Vigilanti, Philippe Vignon, Anitha Vijayan, Jean-Louis Vincent, Emmanuel Vivier, Alexander P.J. Vlaar, Guillaume Voiriot, Giovanni Volpicelli, Alain Vuylsteke, Jan J. De Waele, Mats Wallin, Timothy S. Walsh, Yize Wan, Jingping Wang, Jim Watchorn, Joost Wauters, Bjoern Weiss, Emmanuel Weiss, Nicolas Weiss, Julia Wendon, Stephen Whebell, Tony Whitehouse, Simon Whiteley, Eveline Wiegers, M. Elizabeth Wilcox, Dominic Wilkinson, David R. Williamson, Wolfram Windisch, Xavier Wittebole, Stefan Wolf, Adrian Wong, Hannah Wozniak, Hermann Wrigge, Vincent Wu, Hannah Wunsch, Ryo Yamamoto, Christopher J. Yarnell, Takeshi Yoshida, Paul Young, Jean-Ralph Zahar, Fernando Godinho Zampieri, Alberto Zanella, Alexander Zarbock, Tobias Zellner, Uwe Zeymer, Laurent Zieleskiewicz. Declarations Conflicts of interest The author declares that he has no conflict of interests. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Citerio G Bakker J Brochard L Critical care journals during the COVID-19 pandemic: challenges and responsibilities Intensive Care Med 2020 10.1007/s00134-020-06155-7 2. Citerio G Citerio G Jaber S And once the storm is over… ICM will remain the intensivist’s beacon Intensive Care Med 2021 10.1007/s00134-021-06402-5

36520174

PMC9753062

NO-CC CODE

2022-12-16 23:25:28

no

Intensive Care Med. 2022 Dec 15;:1-4

utf-8

Intensive Care Med

10.1007/s00134-022-06950-4

oa_other

==== Front Russ. Engin. Res. Russian Engineering Research 1068-798X 1934-8088 Pleiades Publishing Moscow 3221 10.3103/S1068798X22100070 Article Digital PR and Communication with B2B Enterprises in the Digital Era Bulantseva L. V. k501@mai.ru grid.17758.3c 0000000088920127 Moscow Aviation Institute, Moscow, Russia 15 12 2022 2022 42 10 10891092 21 1 2022 21 1 2022 21 1 2022 © Allerton Press, Inc. 2022, ISSN 1068-798X, Russian Engineering Research, 2022, Vol. 42, No. 10, pp. 1089–1092. © Allerton Press, Inc., 2022.Russian Text © The Author(s), 2022, published in STIN, 2022, No. 8, pp. 56–60. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. There is an evident need for the rapid introduction of digital public relations (PR) at Russian enterprises, so as to raise brand awareness. The essential features of digital PR as a social medium are discussed; the basic concepts are defined. Particular attention is paid to the use of digital PR at B2B (business-to-business) enterprises, as reflected in published articles and authoritative expert opinion. Keywords: digital public relations (DPR) benefits systems communication channels tools technologies trends B2B enterprises issue-copyright-statement© Allerton Press, Inc. 2022 ==== Body pmcThe global coronavirus epidemic has been challenging for national economies, populations, and businesses, as reflected generally in communications media and specifically in public relations (PR). In enforced social isolation, with increased reliance on information channels such as the Internet, PR departments have been exploring new ways of reaching potential clients. In these changed conditions, adaptation of PR communications has been costly. To be effective, PR specialists have been forced to use the tools of digital PR (DPR), also known as electronic PR (e-PR) and online PR. In Russia, experiences with digital PR date back to the beginning of this century, but widespread use only began in 2020, with the pandemic constraints. At that point, traditional PR approaches still predominated, although many specialists were aware of specific digital PR tools and some had the necessary skills. At first, a very limited set of digital PR resources was employed. Today, we may speak of a new conceptual approach based on new devices and specific digital PR channels, tools, and technologies for communication with the target audiences in different economic sectors. There is a clear trend to digitization: transition to digital communications and the digital representation and transmission of data. Soon these trends will bring PR to a new level of activity, as confirmed by the research described in what follows, which is based on content analysis. Content analysis permits the identification of trends on the basis of facts and figures. It is enlightening to consider expert opinion and media such as Рressfeed.ru, B2B‑journal.ru, Likeni.ru, Bi-school.ru, Sostav.ru, Mediabitch.ru, and Cossa.ru. BASIC CONCEPTS Digital PR is a set of public relations practices utilizing online media, search engines, social networks, and other web technologies [5]. The World Wide Web (www) is a system of interconnected hypertext documents (websites) located on computers and connected to the global Internet. Essentially, digital PR is a new type of communication in the current digital era, when the main link between the enterprises and its customers and suppliers is the following chain: information–digital devices–tools and technologies–online channels–knowledge–opinions. Like traditional PR, digital PR has an indirect effect on company profits. A good reputation affects customers’ decision to purchase a particular product or service. Hence, sales and profits may be increased by disseminating knowledge regarding the company, its leaders, and its brand and thereby inspiring the confidence and loyalty of the target groups. The goal is the same as in traditional PR: by distributing truthful information, to create a favorable social, political, and psychological environment benefiting specific economic actors (companies). We may regard digital PR as traditional PR in a new format, incorporating digital technologies in the quest for the same goals. The capabilities of these technologies offer new possibilities for successful outcomes. 1. They are timely. Real-time delivery of information is possible. 2. They are multichannel. Individual online channels are integrated into a single system for communication with the target audience. 3. They are interactive. They permit dialog and information exchange between all the participants. (Most Internet users are very willing to share opinions as long as the information is real and has the ability to go viral and reach new audiences.) 4. They are viral. Rapid spread of information through the Internet is possible, for example, through links, through social networks, and through instant message (IM) services. In Russia in 2021, the most popular IM services were WhatsApp (2 billion users), Facebook Messenger (1.3 billion users), WeChat (1.2 billion users), QQ (648 million users), Telegram (500 million users), Snapchat (433 million users), Discord (300 million users), Viber (260 million users), Line (250 million users), and Skype (50 million users). The audience is enlarged by chain reaction: for example, free redistribution of information arriving through various online channels. 5. They generate loyalty. With online familiarity, the audience’s confidence in the company and brand increases. The audience receives interesting, useful, and truthful content. A company today must go where the audience is. 6. They are popular. As the company’s content is cited on third-party sites (social media, blogs, etc.), search engines rank that material higher in the feed and, once again, it is shown to more users. 7. They are personalized. Users are sent information that is tailored to their interests, which intensifies their response. 8. They build lasting relationships. If the content provided is interesting and of high quality, users will continue to pay attention. 9. They are measurable. It is possible to track practically all information regarding the company, including media mentions of its leaders and brands. Their tone may be noted. Comparison with the other characteristics of a particular site is possible by means of digital services and tools, quantitative metrics, and special indices. The following resources are available in digital PR: devices for PR communications; digital PR communication channels; digital PR tools; digital PR technologies; and communications specialists with evolving competencies. We now consider those resources. 1. The basic devices employed are TVs and radios, computers, notebooks, tablets, and smart phones, as well as other devices that may store information and distribute it digitally. 2. Digital PR communication channels encompass the capabilities of digital devices available for the communication, transmission, and exchange of information with target groups, such as digital TV and radio; online media; websites, social networks, influencers (bloggers and opinion leaders); video hosting; search engines; review portals; messaging services; and forums (special sites or areas of a site or portal organized for the exchange of views on a specific topic). Video hosting is a service for viewing and introducing videos in the browser through a special player. Users employ such services to view content for free, without downloading it to their device. Authors use such services to promote their brand, attract an audience, and monetize their channel. The most popular web hosting services currently are YouTube; RuTube; Dailymotion, Vzaar, Video@Mail.Ru; Yandex.Video, Myvi.ru, and Toxicbun. 3. Digital PR tools are messages addressed to a target audience for the transmission of text, visual, and audio information. The messages are adapted to the digital format of the specific communication channel, by means of specific methods and algorithms. The basic digital PR tools are messages (posts) placed on the company’s site, in social media, in forums, in blogs, and on video hosting services. Depending on the communication channel, they may contain text, images, links, graphics, or audio and video materials. They may provide news or generate discussion. They may contain updates on future product releases, promotions, or contests. Posts may provide competitions to attract new customers or may contain stories, business cards, portfolios, surveys, or memes (entertaining digital images). They may include quizzes (for example, on the company’s history), instructions, or reviews. The PR department may organize online events (conferences, presentations, webinars, live broadcasts, or digital art viewings. The tools for working with online media include the same list of informational materials as for traditional PR, taking account of the specifics of the digital format. Some important tools are press and video releases, backgrounders, biographies of the organization’s leaders, feature articles, case histories, signed articles (byliners), roundup articles, and interviews. Others include official documents, expert commentaries, opinion pieces, and information bulletins. 4. Digital PR technologies are sequences of procedures and working methods tailored to Internet services and automated systems so that communications with the targeted audience can be optimized and efficiently managed. Examples include the following. (This list does not include technologies for digital marketing and advertising.) 4.1. Technology for deriving information regarding online users of the company’s website and web analytics based on services such as Google Analytics, Yandex.Metrika, and Rating.Mail.ru. 4.2. Search technologies for the target media and specific journalists writing on specific themes relevant to the company’s activity, on the basis of services such as World-newspapers.com, HARO (Help A Reporter Out), Nutcall:PREX, Pressfeed; HackPack, Katalog SMI Yandex, and Deadline.Media. 4.3. Technology for monitoring and analysis of references to brand and enterprise names in open sources (media, blogs, social networks, specialized forums, review sites, etc.) using automated systems such as Angry Scan, Babkee, Brand Analytics, IQBuzz, SCAN, SemanticForce, YouScan, Katyusha, and Medialogy. 4.4. Technology for managing the reputation of the company’s brand and image, with the goal of neutralizing indifferent and negative reactions from the target social groups and generating positive responses, by means of SEO (search engine optimization), ORM (online reputation management), and SERM (search engine reputation management) systems. These are of primary importance, since the goal of corporate communications is to create a positive impression, thereby ensuring the company’s success in the future. SEO (search engine optimization) systems optimize the site for Yandex, Google, and other search engines so as to ensure that materials regarding the company and its products appear at the top of the search response. ORM (online reputation management) permits expansion of the positive material regarding the brand and company that is available in the Internet SEO. It facilitates identification and neutralization of negative comments; and locates complaints and claims against the company, so that they can be eliminated. SERM (search engine reputation management) addresses similar problems in the Yandex, Google, and other search engines. Pages and sites with negative comments regarding the company and its products are moved to remote positions in the list of search responses, and positive comments are moved forward. As a result, Internet searches provide predominantly positive information. This technology relies on well produced and organized PR, techniques for search engine promotion, and coordination with ORM technology. 4.5. Technology for content creation using various services, including free services. Content is any significant information on an information resource, such as a website. Video is created using the following popular services: Supa, Clideo, Coverr, Pexels, and Online Video Cutter. Pictures and photos are produced by means of Canva, Picture Plus, Stock Up, Gratisography, and Kaboompics. Text may be produced by means of Google-doc, Text, Advego, and Glavred. To download and save content, we may use Savefrom, Joxi, and other services. 5. To master the details of social communications using digital PR, specialists with new competencies are required. They must not only be familiar with the tools and technologies already noted but also be adept in related areas such as SMM (social media marketing), which provides assistance to clients, buyers, and business partners; SEM (search engine marketing), which helps users find a website that is appropriate to their query; and web analytics, which analyzes the effectiveness of methods of promotion and assesses their results, thereby providing a measure of the company’s online advertising for its product. Other necessary specialists include a copywriter who creates biographies for social networks and apps; a manager of expert groups; and opinion leaders in certain communities (community managers). In the near future, PR specialists will need to be retrained so that they are up to speed with the available digital technologies. DIGITAL PR IN THE B2B SECTOR AND FUTURE TRENDS Today, economic competition is expressed in terms of intangibles: brands and brand loyalty. The positive image and reputation of enterprises in the B2B (business-to-business) sector are created and popularized by both traditional and digital PR. That provides additional resources in the ongoing global information wars. A significant advantage of digital PR is the quick and effective dissemination of information. That permits prompt response to competitors’ attacks. By means of digital PR, a B2B enterprise will have a global presence and can influence opinion with timely and targeted materials for groups of distinct social and geographic types. B2B enterprises generate complex high-tech products. In assessing company prospectuses, buyers need detailed information regarding the product’s characteristics, so as to make an informed choice. For this purpose, they use online resources: for example, corporate websites, search engines, review portals, social networks, and so on. However, as already noted, these are channels of information distribution in digital PR. Communications and PR specialists use them to convey messages aimed at shaping the opinions of specific target groups. In 2019, Fresh Russian Communications provided data regarding digital communications in the B2B departments of various commercial organizations (in its report B2B Communication Vector 2020). We may itemize the following digital PR trends in the B2B sector on the basis of those findings [2, 3] and other specialist data [4–9]. 1. Problems in digital B2B communications include inadequate budgets, because marketing is favored; extreme communication for the attention of the target audience, since digital PR resources are now generally available; limited use of digital PR resources, for lack of skilled professionals (the available staff may be deficient not only in technical knowledge but in creativity); and problems with contractors, especially in failure to meet deadlines for outsourced projects and poor product quality. 2. The most popular communication channels in the B2B sector are corporate websites; and social networks (in order of popularity, Facebook, Instagram, Vkontakt, and YouTube video hosting). While there is still interest in Instagram, which regularly adds new features and functions, attention has been shifting to TikTok and similar channels. It is promising for B2B applications. 3. The future belongs to digital PR. Therefore, most companies should prioritize digital developments, so as to more effectively strengthen the corporate reputation and expand the online presence of the brand, company, and management. 4. The following are the short-terms trends for digital PR in the B2B sector. 4.1. Continued active use of text and voice formats to reach the target audience. 4.2. Messaging that is accessible, concise, comprehensible, relevant, and valuable. 4.3. Primary use of social networks for information transmission. 4.4. Use of services such as Telegram, WhatsApp, and Viber, since most Internet users prefer brief messages that appear as onscreen notifications. 4.5. Increasing use of video, which is the most popular format among Internet users. 4.6. Parallel use of online and offline formats in PR. Offline events include conferences, excursions, presentations, and exhibits. 4.7. Continued popularity of webinars and live broadcasts, especially if contagion-related social distancing continues or recurs. 4.8. An emphasis on online formats in media work. 4.9. The appearance of new online approaches to reaching target audiences. 4.10. Expanded use of influencers and bloggers, within a strategic framework. 4.11. Large companies’ growing use of Yandex.Zen, a relatively new platform for brand promotion with considerable potential. 4.12. Growing demand for PR specialists fluent in digital media. Specialists will need up-to-date digital skills in order to prosper in a competitive labor market. CONCLUSIONS 1. Digital PR supplements traditional techniques. Today, information technologies must be constantly upgraded if an enterprise is to communicate success fully with potential audiences. 2. The development of digital PR will incorporate not only existing resources for reaching the target audience but also valuable new tools. That will expand companies’ capabilities for reaching audiences and shaping public perceptions of their brands and management. The speed of PR communications and responses is constantly increasing, and communication prospects are rapidly expanding in the current era of digital information. Translated by B. Gilbert ==== Refs REFERENCES 1 Gavra, D., Digital PR of the territory. On the question of concepts, Korporativ. Imidzhel., 2011. http://www.ci-journal.ru/article/601/digital-pr-territorii. Accessed July 25, 2021. 2 Alekseeva, K.M., Digital PR in B2B: Reputation and sales, B2B J., 2020. https://b2b-journal.ru/article/digital-pr-v-b2b-reputacziya-i-prodazhi. Accessed August 24, 2021. 3 Alekseeva, K.M., The research of professional activity specificity in B2B-communication, Gumanitar. Nauki. Vestn. Finans. Univ., 2020, vol. 10, no. 4, pp. 131–136. 10.26794/2226-7867-2020-10-4-131-136 https://cyberleninka.ru/article/n/issledovanie-spetsifiki-professionalnoy-deyatelnosti-v-sfere-b2b-kommunikatsiy. Accessed August 24, 2021. 4 PR trends in 2021, 2020. https://news.pressfeed.ru/pr-trends-2021/. Accessed August 25, 2021. 5 Trends 2021. New normality: Is there any PR after the pandemic, 2021. https://vc.ru/marketing/239492-trendy-2021-novaya-normalnost-est-li-pr-posle-pandemii/. Accessed August 19, 2021. 6 PR people are not afraid of the digital future: The Pro-Vision survey showed the main trends of the profession, 2021. https://www.sostav.ru/publication/piarshchikov-ne-pugaet-tsifrovoe-budushchee-opros-pro-vision-pokazal-glavnye-trendy-v-professii-49616.html. Accessed August 23, 2021. 7 Baikov, E.A., Application of digital marketing technologies to achieve strategic competitive advantage, in Strategicheskoe upravlenie razvitiem tsifrovoi ekonomiki na osnove umnykh tekhnologii (Strategic Management of Digital Economy Development Based on Smart Technologies), Babkin, A.V., Ed., St. Petersburg, 2021, pp. 621–643. 8 Il’in, G., What’s wrong with you? Common PR mistakes in digital, 2019. https://www.likeni.ru/analytics/chto-s-toboy-ne-tak-rasprostranennye-oshibki-pr-v-digital/. Accessed September 27, 2021. 9 Kalacheva, A., PR Tech or free digital tools for PR specialists, 2021. https://bi-school.ru/pr-tech-ili-besplatnye-tsifrovye-instrumenty-dlya-pr-spetsialistov/. Accessed July 30, 2021.

0

PMC9753068

NO-CC CODE

2022-12-16 23:25:28

no

Russ. Engin. Res.. 2022 Dec 15; 42(10):1089-1092

utf-8

oa_other

==== Front Lat Stud Lat Stud Latino Studies 1476-3435 1476-3443 Palgrave Macmillan UK London 390 10.1057/s41276-022-00390-3 Original Article Cascading disasters: The impact of hurricane Maria and Covid-19 on post-disaster Puerto Rican migrants’ adaptation and integration in Florida Desastres en cascada: El impacto del huracán María y COVID-19 en la adaptación e integración en Florida de los migrantes posdesastre de Puerto RicoAranda Elizabeth earanda@usf.edu 1Elizabeth Aranda is Professor of Sociology at the University of South Florida. A native of Puerto Rico, she has dedicated herself to documenting the lived experience of migration and to share (im)migrants’ stories through her research, teaching, and service. Her research addresses migrants’ emotional well-being and how they adapt to challenges posed by racial and ethnic inequalities and legal status. She is author of Emotional Bridges to Puerto Rico and co-author of Making a Life in Multiethnic Miami. Blackwell Rebecca rblackwell@usf.edu 12Rebecca Blackwell is a Social Science Researcher at the University of South Florida. She has a multicultural personal background and a multidisciplinary education in the areas of linguistics, Latin American and Caribbean studies, and sociology. Her research in migration, human rights, and health and illness explores ways in which social communication, social psychology, and emotions intersect in the perception, contestation, and reproduction of inequalities in society. Escue Melanie mescue@usf.edu 1Melanie Escue (she, her, hers) is a doctoral candidate in the Department of Sociology at The University of South Florida. Her research interests include undocumented migration, Puerto Rican studies, post-disaster migration, and emotional well-being. At the heart of her research, teaching, and service, she strives to draw attention to the diverse backgrounds, experiences, and needs of im(migrants) in the United States. Currently, her work explores the emotional well-being of undocumented young adults as they navigate transitions to adulthood in the United States. Rosa Alessandra amrosa1@usf.edu 1Alessandra Rosa (she, her, ella) is a sociocultural anthropologist, professor, researcher, activist, public speaker, and consultant. Currently, she is a visiting assistant professor of instruction in the Department of Sociology at the University of South Florida (USF). Her areas of expertise include social movements, digital activism, education, Latin America & Caribbean studies with a focus on Puerto Rican studies, women & gender studies, post-disaster migration studies, emotional well-being, and media discourse analysis. As a transnational feminist scholar, she has dedicated her teaching, research, and service to fostering diversity, equity, and justice. 1 grid.170693.a 0000 0001 2353 285X University of South Florida, Tampa, USA 2 St. Petersburg, USA 15 12 2022 124 20 9 2022 © The Author(s), under exclusive licence to Springer Nature Limited 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Based on data from 103 surveys of Puerto Rican migrants living in Florida and 54 in-depth interviews with a subgroup of them, we examine how Puerto Ricans who left the archipelago after Hurricane Maria have navigated settlement in their new homes. In this article, we observed and classified our participants’ descriptions of how they managed opportunities and challenges regarding education, employment, and social relations, the traditional benchmarks for the assessment of societal integration. We also observed how our participants described Covid-19’s interaction with these benchmarks. We found that our participants have experienced a series of cascading disasters since 2017—namely, Hurricane Maria, the earthquakes that affected Puerto Rico starting in late 2019, the humanitarian crises that followed both disasters, and now the global pandemic. These disasters, compounded with migration, have resulted in a process of adaptation to Florida in which social and labor-market integration and the ability to nurture social ties have been significantly diminished. Resumen Basándonos en los datos de 103 encuestas con migrantes puertorriqueños residentes en Florida y cincuenta y cuatro entrevistas a fondo con un subgrupo de estas personas, examinamos cómo los puertorriqueños que abandonaron el archipiélago después del huracán María han lidiado con la adaptación a sus nuevos hogares. En este artículo observamos y clasificamos las descripciones de los participantes: cómo manejaron las oportunidades y los retos relacionados con la educación, el empleo y las relaciones sociales, que son los puntos de referencia tradicionales para evaluar la integración social. Observamos también cómo los participantes describían la interacción de la pandemia de COVID-19 con estos puntos de referencia. Encontramos que estas personas han experimentado una serie de desastres en cascada desde 2017, específicamente el huracán María, los terremotos que comenzaron a afectar a Puerto Rico a finales de 2019, las crisis humanitarias que siguieron a ambos desastres y ahora la pandemia global. Estos desastres, agravados por la migración, han tenido como resultado un proceso de adaptación a la Florida en el cual ha disminuido significativamente la integración social y laboral y la capacidad de nutrir los vínculos sociales. Keywords Cascading disasters Hurricane Maria Post-disaster migration Adaptation Covid-19 Puerto Rico Florida Palabras clave Desastres en cascada Huracán María migración posdesastre Adaptación COVID-19 Puerto Rico Florida ==== Body pmcSince the early 2000s, scholars of U.S. migration have examined how Latinos/as are integrating into new destinations (Massey and Capoferro 2008), particularly in what has been called the “new south” (Winders 2005). The area, which roughly comprises a region from North Carolina in the north to Florida in the south and west to Arkansas, is called “new” because of the growth in Latino/a presence beginning in the mid-1980s and early 1990s (Kochhar et al. 2005). Integration outcomes differ depending on the contexts of reception (Portes and Rumbaut 1996), and research indicates that Latinos/as who migrate to the new south, particularly those who arrive with greater human capital such as higher education, English fluency, and legal status, gain greater access to opportunities within their workplaces and communities (Marrow 2011). Puerto Ricans would be among the groups in the new south that fit this upward-mobility profile, given their citizenship status in the United States and their overall human capital (Marrow 2011; Schleef 2009). Therefore, we might expect them to have very positive integration outcomes. In this paper, we argue that measures of integration from traditional migration studies have been dominated by rational choice theories and economic and political analyses (Hollifield 2020) that neglect the inescapable emotional processes that complement the instrumental practices of adaptation for migrants (Aranda 2007). Despite what we know about Puerto Rican communities stateside (Acosta-Belen and Santiago 2018) and in Florida specifically (Duany 2011; Silver 2020), researchers know little about the adaptation of the most recent migrants who fled Hurricane Maria’s catastrophic effects. Moreover, the ways in which Covid-19 has shaped the challenges migrants confront as they adapt and integrate into a new society are yet to be explored from the perspective of migration studies, and the context in which post-disaster migrants have arrived raises questions regarding their trajectories of incorporation, particularly considering existing differences among groups in the Puerto Rican community’s integration outcomes. This analysis is an effort to begin to fill in this gap. We thus ask, what do the adaptation outcomes look like for climate refugees, and how has the Covid-19 pandemic that began to affect the U.S. in 2020 shaped these incorporation processes? Moreover, how does the combination of previous and new stressors affect post-disaster migrants’ adaptation and integration outcomes? Our primary data source consists of 54 in-depth interviews, a subset of 103 Puerto Rican migrants whom we had initially surveyed in the context of a larger research project that investigates how Puerto Ricans are navigating the stressors of relocation to Central Florida after the traumatic events of Hurricanes Irma and Maria in 2017, while also confronting the Covid-19 pandemic. Moreover, some of our interviewees were also affected by the series of earthquakes that took place in the archipelago in December 2019, since they worried about their kin still living there. We combine a number of theoretical frameworks to gain understanding of the nuances in their adaptation process. First, we provide a brief summary of the history of Puerto Rican migration. Next, we lay out our theoretical and conceptual framework, focusing on disaster risk, coloniality, and emotions in migration. Then we describe our method of analysis, present our findings, and discuss their implications. Brief background of Puerto Rican migration There is a long history of Puerto Rican migration to the continental U.S. Since 1898, Puerto Rico has been a U.S. colony, under the classification of an unincorporated territory. In 1917, Puerto Ricans were granted a collective citizenship provision under the Jones-Shafroth Act, but it was not until 1940 that the U.S. Congress passed the Nationality Act granting birthright citizenship to any person born in Puerto Rico (Venator-Santiago 2017). As U.S. citizens, Puerto Ricans have been able to move between the archipelago and the fifty U.S. states, in what has been described as “circular migration” (Duany 2002). More recent research has shown that migration for Puerto Ricans is viewed as a flexible survival strategy allowing them to capitalize on their ability to move in search of a better quality of life throughout the life course (Duany 2011; Aranda 2007). Out-migration from Puerto Rico to the continental U.S. has grown consistently over the years, especially more so after the catastrophic effects of natural hazards. In Florida, the Puerto Rican population was recently estimated to be 1,190,891 (Social Explorer n.d.-a)—surpassing the population in New York and representing the largest concentration of Puerto Ricans in the continental U.S. We focus on the Florida region that encompasses cities in the Interstate-4 corridor, which crosses the center of the state from east to west. This area is a popular gateway for this group since more than a decade ago (Velez and Burgos 2010) and is home to four of the five largest communities in the state (Duany 2015). Our sample draws from the Tampa Bay area, where approximately 22 percent of the nearly 56,000 Puerto Ricans who relocated to Florida after Hurricane Maria settled (Gamarra 2018). In September 2017, Hurricanes Irma (a category 5 storm) and Maria (a category 4 storm) impacted Puerto Rico within two weeks of each other. Though Hurricane Irma passed near Puerto Rico, it left in its wake power outages and water-service interruptions for several days (Rand Corporation n.d.). Then Hurricane Maria directly struck Puerto Rico and left a death toll of approximately 2975 people (George Washington University 2020), wiped out the archipelago’s power grid, and affected its communication infrastructure. Residents lacked access to medication (Melin et al. 2018), fresh food, and potable water for an extended time period and had limited access to health services given that most roads were impassable (Rand Corporation n.d.). In the aftermath of Hurricane Maria, the Federal Emergency Management Agency (FEMA) mismanaged the distribution of aid to Puerto Rico. Commodities shipped to aid Puerto Rico sat in FEMA’s custody for approximately forty-eight days, and life-sustaining commodities like water and food experienced shipping delays of seventy-one and fifty-nine days, respectively (U.S. Department of Homeland Security 2020). Moreover, the local government did not adequately track the supplies it received from FEMA, and 40% of Puerto Rico’s municipalities (twelve out of thirty) had problems with receiving expired food (U.S. Department of Homeland Security 2020). As a result of these conditions, migration from Puerto Rico increased. Some scholars have described this type of migration movement as “displacement” (Vargas-Ramos and Venator-Santiago 2019), while others call these migrants “climate refugees” (Duany 2021). We use both the terms “post-disaster migrants” and “climate refugees” when we discuss our sample. Just over half of these climate refugees from Puerto Rico moved to the U.S. South, and most moved to Florida. Within this state, Orange County (Orlando metropolitan area) received the greatest number of evacuees, followed by Osceola (also in Orlando), Miami-Dade (Miami area), Polk (Lakeland and surrounding communities), and Hillsborough (Tampa area) Counties (Hinojosa et al. 2018). It is estimated that between 159,415 and 176,603 Puerto Ricans left the archipelago in the year after the hurricanes (Hinojosa and Meléndez 2018). These numbers represent a significant increase compared to previous years. Moreover, migrants continued to arrive after the numerous earthquakes that began in December 2019, lasting through the spring of 2020 (Sanchez 2020). Table 1 (see Appendix) illustrates the Florida counties with the largest shares of Puerto Ricans and shows how the population is significantly higher in the counties along the I-4 corridor (Orange, Osceola, Hillsborough, and Polk Counties). Though these data come from the ACS 2019 five-year estimates, they show the counties where Puerto Ricans have been concentrating both before and after the hurricanes. Figure 1 illustrates the geographic location of these counties and the proportion of Puerto Ricans in each.Fig. 1 Percent of Puerto Ricans, by Florida County. Source: SocialExplorer.com. (nd, b) ACS 2019 (5-year estimates) Although Puerto Rican migrants have citizenship status in the United States (Grosfoguel 2003), as colonial migrants, they have undergone “decades of discrimination in labor and housing markets” (Silver and Vélez 2017, p. 99), which has challenged their adaptation and successful integration. Moreover, in the United States, there are strong cultural biases against migrants who speak Spanish or have Spanish heritage, for they can be perceived as noncitizens, inherently undocumented, poor, uneducated, criminally inclined, refusing to learn proper English, and, hence, outsiders (Parsons 2011). Preliminary studies suggest that Florida and specifically parts of Central Florida present an advantageous context of reception (Velez and Burgos 2010), yet questions remain as to which populations benefit most from this context. While many Puerto Rican families have found a better quality of life in Florida than in Puerto Rico (Duany 2011; Duany and Matos-Rodríguez 2006), evidence suggests that there may be bifurcated integration patterns, separating professional and working-class Puerto Ricans (Delerme 2013). Disaster risk, coloniality, and emotions in migration We incorporate an analytical framework from disaster risk theory, an area of human geography that addresses the interaction of social and biophysical systems (McGowran and Donovan 2021) and theorizes that society’s vulnerabilities, resilience, and adaptation processes in the face of natural hazards are structurally affected by socio-political factors that can sometimes lead to cascading disasters (Pescaroli and Alexander 2015; Thomas et al. 2020). Moreover, given the colonial history of Puerto Rico, we also consider theories that explore disparities in the relationship between the continental U.S. and the Puerto Rican archipelago (Bonilla 2020), since they are the receiving society and the society of origin of our participants. We build on works that explore emotions in the migration experience and expose underlying factors affecting the self-perception of well-being after relocation (Aranda 2007) and works that explore the relationship between pre-migration trauma and migration stressors (Li 2016; Kilic et al. 2006). Disaster risk theory Thomas et al. (2020) draw from the fields of geography, political science, and health and behavioral science to develop a people-centered conceptual model—referred to as “CHASMS” (Cascading Hazards to disAsters that are Socially constructed eMerging out of Social vulnerability)—for the analysis of the preexisting social and structural forces that lead to inequitable outcomes. They build on the following definition of “cascading disasters” from the field of human geography:Cascading disasters are extreme events, in which cascading effects increase in progression over time and generate unexpected secondary events of strong impact. These tend to be at least as serious as the original event, and to contribute significantly to the overall duration of the disaster’s effects. These subsequent and unanticipated crises can be exacerbated by the failure of physical structures, and the social functions that depend on them, including critical facilities, or by the inadequacy of disaster mitigation strategies, such as evacuation procedures, land use planning and emergency management strategies. Cascading disasters tend to highlight unresolved vulnerabilities in human society. (Pescaroli and Alexander 2015, p. 65) The CHASMS model belongs to a body of literature that is currently underscoring the need to incorporate social sciences knowledge into research on disaster risk reduction. Its underlying premise is that in order to produce knowledge that can effectively assess hazards and risks and understand the futures that are being constructed in the present, various types of knowledge are necessary, including knowledge about situated power dynamics and cultural beliefs, alongside knowledge of structural vulnerabilities and possible geophysical forces (McGowran and Donovan 2021). The CHASMS model proposes that the vulnerability and resilience of a society in the face of cascading disasters (e.g., a prolonged humanitarian crisis generated by lack of electricity, food, or water after a natural hazard) may be further impacted by cascading hazards that occur in the same area before recovery is possible. The impact of Covid-19 in Puerto Rico before recovery from the hurricanes is an example of multiple hazards worsening or prolonging existing crises. An argument of this model is that in some cases the term “resilience” may be wrongly used to describe “survival.” These works also suggests that social vulnerability and resilience are interrelated and that the concept of “resilience can deflect attention away from enduring vulnerabilities and the exhaustion of resilience in the face of multiple or cascading disasters” (Thomas et al. 2020, p. 3). These propositions aim to complement more traditional apolitical and technical approaches to the study of disasters and disaster risk management by proposing the analysis of preexisting inequalities, with an emphasis on the role of structural factors in the generation of a cascade, which “emerges out of social, political, cultural, and economic systems that shape community and individual risk at multiple temporal and spatial scales” (Thomas et al. 2020, p. 3). These works have important points in common with scholarship on what has been called “disaster capitalism” (Klein and Smith 2008; Schuller and Maldonado 2016), which proposes that neoliberal actors profit from disasters by implementing measures on populations who accept them only because they are weakened by the hardship they are facing during the disaster. Coloniality Bonilla (2020) has argued that the management of resources and treatment of the people in the aftermath of Hurricane Maria revealed a historical “racialized neglect” on the part of the United States. The underlying assumption in her work (see also Bonilla and LeBrón 2019) is that the colonial occupation of Puerto Rico has been marked by the same form of racialized governance that is present in the continental U.S. This perspective, applied as a framework for the analysis of disasters stateside, is also found in works about Hurricane Katrina in New Orleans (Meyers and Hunt 2014). In Puerto Rico’s case, Bonilla (2020) argues that the hurricanes dismantled the notion that U.S. development had elevated the Puerto Rican standard of living. Racialized neglect is the social abandonment of a dispossessed people, which renders them vulnerable to the natural hazards of climate change. Hurricane Maria became catastrophic because of the archipelago’s vulnerability, making the colonial dispossession not only palpable but also reconfirmed by the lack of an effective emergency response on the part of the federal government (Bonilla 2020), but also the insular government (Rodríguez-Díaz 2018). Together, the disaster risk theory concept of cascading disasters and the notion of a racialized governance provide a nuanced framework for the analysis of post-disaster migration in a colonial context. We apply these perspectives to examine our participants’ migration outcomes because we wish to further understand some of the ways in which colonially induced inequalities, including local governance failures affected by the extralocal political situation, may be taken for granted. These factors may be impacting trajectories of incorporation. Our goal is to gain insight into the multiple scales of time, space, and social systems at play in our participants’ migration experience. Emotions in migration research To study migrant integration in the U.S., scholars have typically measured “how different or similar to other Americans are immigrants and their children in terms of socioeconomic standing, residential segregation, language use, and intermarriage?” (Waters and Jiménez 2005, p. 106). Moreover, what scholars call “straight-line” assimilation theories argue that successful integration often leads to upward social mobility (Gordon 1964; Portes and Rumbaut 1996). Existing theoretical understandings of “adaptation” and “integration” frequently associate these two concepts with processes of “acculturation,” or “the adoption of the host society’s mainstream values, attitudes, sentiments, behaviors, and practices” (Riosmena et al. 2015, p. 444), and “assimilation,” or the decline in the migrant’s ethnic characteristics (Alba and Nee 2009). By contrast, our analysis of the recent lived experiences of Puerto Rican migrants has been able to demonstrate, through the migrants’ perceptions of well-being, that ethnic traits and identity may remain the same or even increase during the incorporation process. For post-disaster migrants, research has shown how stressors encountered in the new home shape social and health outcomes. Some scholars refer to these stressors as “secondary trauma,” referring to “events that prolong the disaster experience” (Gil 2007, p. 615), or “secondary stressors” (Kessler et al. 2012; Li 2016). Secondary stressors are the challenges that operate in conjunction with disasters and that have a negative effect (Kessler et al. 2012, p. 36), such as relocation, changes in career, or poor health related to the original trauma, as well as the overall slow pace of the practical recovery process. Research has shown that post-disaster migration can lead to worse outcomes in mental health, and secondary stressors can compromise well-being upon resettlement (Li 2016). For instance, Kilic et al. (2006) found that post-disaster relocation was associated with poor mental health outcomes; they attribute this to disrupted social networks resulting from migration. Moreover, studies have shown that migrants who have been exposed to pre-migration trauma are more prone to experiencing acculturative stress upon resettlement, associated with the loss of roots; employment and language barriers; difficulties finding housing, health care, and education; social isolation; racism; and guilt for leaving loved ones behind (Li 2016). The traumatic losses due to major disasters can produce an expanded disruption and loss of the sense of community (Erikson 1976), and “the pile-up effect of multiple losses, dislocations, and adaptation challenges can be overwhelming” (Walsh 2007, p. 216). We draw from this literature to analyze the patterns that we found in our sample of post-disaster migrants, related to their experiences of secondary stressors during their adaptation processes in the context of the global pandemic. Data and methods Epistemology We combine naturalist and interpretivist methodologies, creating a confluence of epistemologies in order to better understand the experiences of a group of Puerto Ricans who migrated to Florida after Hurricane Maria. Our analysis involves two sources of data: a survey conducted with 103 individuals and in-depth interviews with a subgroup of fifty-four individuals. While the survey data provide us a glimpse into the impact of the pandemic on migration outcomes, the interviews allow us to unpack these experiences. Survey We partnered with a nonprofit organization, Mujeres Restauradas por Dios (MRD), a faith-based local agency established by Nancy Hernandez in November 2013 to serve victims of intimate partner violence and human trafficking. By January 2014, MRD was continuing its work supported by the Tampa Underground Network (UG), a Christian organization that supports more than 100 small mission communities. Once Hurricane Maria climate refugees began to arrive in Tampa, the organization expanded its services to provide emergency disaster relief services for this population. MRD received funding from the Tampa Bay Disaster Relief and Recovery Fund through the Community Foundation of Tampa Bay to continue assisting the Puerto Rican families that were arriving. By establishing a one-stop hub, MRD facilitated access to local services such as job placement, housing, social services coordination (e.g., school enrollment, health care, food assistance), emotional support, and as a faith-based organization, spiritual guidance. We obtained MRD’s list of clients affected by Hurricane Maria (over 1000) and contacted them through phone to administer a survey that took about forty-five minutes to complete. Survey questions were adapted from a survey of Hurricane Katrina survivors conducted by the Hurricane Community Advisory Group, at the Harvard Medical School (Kessler 2010). Our cooperation rate was 63% (AAPOR 2016, Cooperation Rate 1). Of those surveyed, a subset of the sample was interviewed virtually (i.e., over Zoom, Facetime, or WhatsApp). All surveys were conducted in Spanish by native Spanish speakers. Table 2 illustrates the main demographic characteristics of the sample, and Table 3 illustrates some of their experiences post-migration, such as their employment status during the Covid-19 pandemic, changes in their financial stability, and changes in their social ties (see Appendix). It is important to mention that this sample is not representative of Puerto Ricans who moved stateside; given their financial and other needs (for which they sought MRD’s services), it is a disadvantaged sample, especially considering all that they lost due to the hurricanes. Most survey participants were women (82%), the average age was forty-seven, and most identified as having white or light skin (88%). Thirty-five percent were married, 43% were single, and most had children (85%). Regarding education, 21% had a high school diploma or less, and 36% had a bachelor’s or graduate degree. Most of those surveyed did not know English very well or not at all (60%), and the average number of months they reported living in the Tampa region was twenty-eight months. Regarding changes in their relationships, the average number of loose and close ties diminished when comparing their ties pre-hurricane to those they had post-migration.1 Of those who reported that their employment changed due to the pandemic (just under one-third of the sample), one in three were laid off, and one in five permanently lost their jobs or had their hours reduced. In addition to the precarity in employment, when asked about their financial difficulties, the percentage who reported very difficult financial problems before the hurricane (20%) declined after migration (15%) and subsequently increased during Covid-19 (27%). Given the character of the sample (e.g., predominantly women, somewhat older in age, and disadvantaged when it comes to English fluency), it is possible that these characteristics introduced certain biases in the data and led to some of the outcomes we report below. Thus, while some of the issues we discuss may be pertinent to other Puerto Rican climate refugees, we cannot say that they are representative of the Puerto Rican post-disaster experience. In-depth interviews Several months after the survey, we contacted respondents to gauge their interest in participating in interviews, and 54 individuals agreed. Interviews lasted approximately two to three hours. The themes in the interviews included their experiences of the hurricanes; their reasons for migrating to Florida; their experiences finding jobs, housing, and social services; and their perceived physical and mental health at various times throughout the migration journey. We also probed how they were handling the social restrictions due to the pandemic. All interviews were conducted in Spanish. Data analysis For the qualitative data, a constructivist grounded theory approach was adopted, which allowed theory to be developed from data in an iterative process (Charmaz 2014). This approach involved an interplay between inductive and deductive reasoning; an important aspect of this perspective involves employing the existing literature and interweaving it throughout the research process. Grounded theory carries the potential of developing new theories that not only are rooted in participants’ accounts but also can be set into the context of existing theories (McGhee et al. 2007). All interviews were transcribed verbatim and entered into MAXQDA, a qualitative analysis software. Data were coded based on conceptual themes developed following the interviews. For example, one code that was used was “adaptation,” and subcodes of “adaptation” included what the source of the challenge was (e.g., problems with the language, social ties, etc.). Through the patterns in the qualitative data, specifically those related to the code “adaptation,” we identified social ties, education, and employment as among the major challenges to adaptation that were most affected by Covid-19. Thus, in the findings, we focus our analysis on these three challenges, especially since they are also among the most often used benchmarks to measure integration (Waters and Jiménez 2005). Methodological considerations The research team consisted of five women with diverse backgrounds: two are Puerto Rican (the PI and Co-PI); two are Latin American immigrants, from Venezuela and Colombia; and one is a white woman. The interviews were conducted by the two Puerto Rican women and the Venezuelan woman, whose positionalities helped to gain the trust of the population studied. The PI and Co-PI are Puerto Rican women who were raised in Puerto Rico, which facilitated the connection with the organization we worked with (headed by a Puerto Rican woman). In addition, these two researchers had family who experienced the hurricanes and connected with participants in this regard. Findings Many of our participants moved to the continental U.S. because they lived through dire conditions in the aftermath of Hurricane Maria. In some of the examples below, participants describe the prolonged experience of enduring in a state of emergency. They report a lack of federal response to the devastation of the hurricane, which constitutes an example of the racialized neglect that Bonilla discusses (2020). The pandemic then occurred before they were able to overcome previous crises and compromised their path to adaptation and integration. The pandemic is a hazard event that piled onto previous disasters caused by preexisting inequalities, which is consistent with the CHASMS model (Thomas et al. 2020). The pile-up effect (Walsh 2007) of Covid-19 has also been exacerbated by the process of migration, which with its linguistic, cultural, economic, and individual challenges operated as a secondary stressor (Kessler et al. 2012; Li 2016) prolonging the trauma (Gil 2007, p. 615) of the experience of the hurricanes. Participants described suffering hunger in the aftermath of the hurricane. Fabiola,2 who is twenty years old, has an associate’s degree, and works in hospitality and leisure, said, “I tell you, it was eating once a day, going to bed hungry.” They described losing their jobs, as did Laura, a 49-year-old mother of two who attained a bachelor’s degree and currently works as a clerk at Walmart. She explained, “I didn’t have a job. They closed the school where I worked. And I said to myself, ‘Where am I going to go? What do I do? What can I do?’” Some lost their homes, like Aurora, who is a 40-year-old housewife, high school graduate, and mother of one child. She reported having been denied all forms of assistance by FEMA: “I don’t know what is the problem with those people. They denied us all the types of help they supposedly offered.” Aurora struggled to describe the traumatic experience of the hurricane:Hurricane Maria marks you, because nobody is prepared for something like that to happen. It was something so strong. Our house fell apart. … God had mercy that the door of my house didn’t explode, because it was something where the winds, … the water got into my house. It was a horrible thing. It was horrible, horrible, horrible. It was something that no one could understand unless they live it. With the pandemic, the sense of helplessness that many of our participants felt from the hurricane reawakened. Many of them lost their jobs again or experienced reduced working hours, with their children forced into stressful remote-learning situations complicated by language barriers in the family. Some participants experienced physical health issues related to isolation and lack of resources to pay for medicines, and some lost access to medical services. Many reported stress related to the worry over the health of family members left behind in Puerto Rico. We identified in the data three salient stressors that challenged adaptation and integration: the absence of social ties, educational challenges, and the loss of employment and financial stability. Social ties Ironically, what is needed to foster resilience after disaster-induced trauma is what is compromised upon migration—social support. As Walsh (2007, p. 220) states, “The comfort and security provided by warm, caring relationships is especially critical in withstanding trauma events, which induce social and personal uprooting, family disruption, separation and loss, mental and physical suffering, and vast social change.” When participants described the most difficult aspects of moving to the continental U.S., an overwhelming majority of them mentioned that leaving families behind was the hardest, showing a sign of acculturative stress (Li 2016). Some could not avoid crying while talking about this aspect. Aurora, for instance, said, “The most difficult part for me was having to leave my family.” Migration also had compounding effects beyond missing family and kin relationships. For Aurora, adapting to life in the continental U.S. has been complicated by the language barrier, which has made communication with those who can provide her with assistance seem impossible: “I cannot go to Catholic Charity because of Covid, and I cannot speak English with the American on the phone.” Her English classes were also interrupted by the pandemic lockdown, halting her acculturation process. Despite these challenges, individuals employ their creativity to overcome loneliness. Aurora used her love of plants to make herself spend time outdoors: “With this problem of Covid, I go outside. I love plants, so I plant things and take my chair outside and sit there in the afternoons when the sun goes down, and if there is someone, I say hello.” Participants used to look forward to visiting or receiving kin from Puerto Rico, particularly if their stateside networks were underdeveloped, but Covid compromised their ability to travel. Such was the case of Yanira, a 28-year-old housewife and high school graduate whose youngest son is on the autism spectrum and who felt a great deal of anxiety because of the confinement and loss of her son’s therapies but also because of their inability to travel due to the pandemic. Yanira told us,We had plans to go to Puerto Rico in March, but I wasn’t going to expose the children to being in a plane. … I said to my grandparents, “I’m sorry. I know I’ve not seen you for two years, but for now we cannot go there.” So that was one of the hardest things for us during the beginning of Covid. Visits such as these are important, since research has shown them to be a mechanism that migrants use to uplift their emotional well-being, which is affected by the loss of face-to-face contact with kin upon migration (Aranda 2007). Some participants described their anguish in the isolation of the pandemic. Social-distance restrictions on outings were particularly difficult when migrants did not have communities of support; perhaps these outings filled the void that the loss of social ties left. Nando, a 34-year-old father of two who attained an associate’s degree and works in manufacturing, discussed how pandemic restrictions on the family’s mobility exacerbated the effects of having lost the connection to a support network:Well, in a way it affected us. We are always a family that doesn’t follow a routine. We like to go out, to experiment. … In that sense, the lockdown affected us. So, not being able to keep discovering or traveling, going out and sharing … Because here it’s not like in Puerto Rico. It’s more lonely here. … People live in their own world. Laura, mentioned earlier, discussed how the pandemic interrupted a process by which she was making friends and developing relationships, which is part of social integration: “I had to change my routine, change my lifestyle. I can’t go to places anymore, see friends. I don’t have much contact with friends. It’s like a process was interrupted.” One way that Laura manages to maintain social ties is to keep up with her friends who are in Puerto Rico; this includes her family too: “I stay in touch with my friends there … [through] Facebook. They’ve come to see me here … before the pandemic. … I have few friends, but I see my family as friends too.” Like other immigrants who are part of transnational kinship networks, Puerto Ricans stay in touch through phone and other social media to increase feelings of connectedness with these networks, particularly during the pandemic, when mobility was restricted. Confinement seems to have been extremely challenging for many of our participants when the trauma that they had previously experienced because of the hurricane is considered. Natalia, a thirty-four-year-old mother of two who earned a graduate degree and works in administrative services, made an association between the confinement of the pandemic and what she lived after the hurricane: “The [pandemic] confinement gives me a lot of anxiety because I was locked up for a year due to the hurricane. There was no light. There was nothing to do. It was confinement all the time.” Here, we see that emotional responses to the pandemic are conflated with those experienced after Hurricane Maria. The pandemic brings back memories of Maria’s aftermath. Among the things that brought Natalia some relief was that she connected with a friend from Puerto Rico who introduced her to a Latino church that resembled the church she attended in the archipelago. She continued to attend this church despite the pandemic. And as much as the stay-at-home order was hard on her, she thinks the experience would have been more difficult had she be living in Puerto Rico. Despite having to relive the trauma of the hurricane, she shared that she was generally happier living in Tampa than in Puerto Rico. Overall, the loss of face-to-face social ties after migration was worsened by the restrictions associated with Covid. Moreover, how Puerto Ricans cope with these losses (e.g., outings, visits to Puerto Rico) is also affected by the pandemic as the process of social integration stalls. Education The change to remote learning proved to be particularly difficult for our participants. For some, online education represented a hazard to their children’s physical health. Lydia, a 36-year-old health worker who had an associate’s degree, described the hardship and chaos her family endured because she and her partner had to continue working outside their home and her children stayed home taking virtual classes without supervision. Both of her children struggled with anxiety and had sleeping and eating issues. Lydia described how the pandemic affected her family:It was chaos for me. This is a moment when you most need your family, who are not around. You also have the panic that something can happen to your parents, who are in Puerto Rico alone, and you cannot do anything. You need them here with you helping you. … You end up having a lot of feelings, depression, anxiety. It’s very difficult. … You have two children without supervision, who do not know what to do, for whom the online process was very complicated. It was chaos. … They were constantly awake, all night sometimes. … They gained a lot of weight. They ate a lot out of anxiety and without control, and I was not at home to say, “Eat this or that.” … My thirteen-year-old boy was used to going out a lot, after he left school. He could not go out. It was very difficult, because we have to be locked up. For him, it was the most difficult, and he was the one who gained the most weight. … When I took him to the doctor, … he had gained fifty pounds. He came out pre-diabetic in the lab work. Lydia and her family came to Florida because they lost their belongings, their home was damaged, and they lost their jobs after the hurricane. She told us, “He [her partner] wasn’t working, and neither was I. … We were living like homeless people who get into houses that are like this, all abandoned, all broken.” After migrating to Florida, her first job was in a cookie factory, which hurt her hands and where she earned $9 per hour. She currently works for $10.99 an hour as a home health aide (for which she took a course in Florida) but has no benefits. Trying to keep this job was the reason her children were unattended. She bemoaned the absence of family to help take care of them, and the effects of the transition to remote learning affected the whole family’s physical and mental health, not to mention its effects on her children’s education itself. Despite these challenges, Lydia expressed not wanting to return to Puerto Rico, which she describes as an abandoned site, a neglected place with deep social consequences such as high crime: “Puerto Rico has become a no man’s land. … Crime is very high. … I don’t want my children to stay in that environment. Here they may have problems with bullying, discrimination, and things like that, but those are manageable problems.” Lydia’s narrative about enduring discrimination in the context of running from violence speaks to the survival strategies that people affected by compounding crisis are forced to develop. For many Puerto Rican migrants, language is a barrier to integration, and some participants felt that they could not assist their children with their schoolwork for this reason. This is the case for Yulayda, a 27-year-old housewife who has an associate’s degree and has four children. When asked if she speaks English, she said, “I can barely understand. I can understand what you want to say to me, if you tell me slowly and with basic words”:At school they are already in a program called ESL, which helps them with the transition, and it has affected their grammar and reading, because there they had their teacher, who took them aside for a certain time as if it were a “Title 1” [school program for disadvantaged children], as one says. They would take them aside and deal with what they need. Well, at home they don’t have that. Similarly, Misty, a 31-year-old mother of two and a high school graduate, described her experience during Covid: “It is affecting me with my girls here because of home schooling. It’s like going backwards again. Not having the language, you get lost a lot.” Misty shared that she knew several Puerto Ricans who had moved to Florida after the hurricane; they were acquaintances, including a childhood friend, whom she could rely on for help and companionship. They occasionally got together, but Covid interrupted that. She said, “We haven’t seen each other in ages, because all this happened. … But at least there is the phone.” Misty referred to a time when one of the acquaintances in this network helped her with one of her children’s homework assignments, illustrating the strength of relying on coethnic ties, albeit loose ties, as a strategy to navigate adaptation. Other participants felt that they could not continue with their own educational plans. Manuel, a 38-year-old father of two who earned an associate’s degree and works for a moving company, saw his studies interrupted in Puerto Rico due to the hurricane and the process of migration as well, since he could not continue in Tampa what he had started studying in Puerto Rico:I’ve thought about it. I’ve always thought about finishing it. … I don’t have long before finishing my bachelor’s. I only have a few credits left. … I was in the middle of studying to finish my bachelor’s, and then the hurricane happened. I had to leave. There were no classes. There was no way to get to Bayamón. … I couldn’t continue. … Right now, with Covid-19, it’s so difficult to make money that I prefer to be working as much as I can than studying. I have the studies. I have half. I would like to finish one day. Under normal circumstances, continuing his education would have involved navigating the system of transferring credits and deciphering what coursework completed in Puerto Rico would count for a degree stateside. But when considered alongside Covid-19, Manuel determined that it was best to earn as much money as he could. In this regard, though he wishes to someday continue his studies, his strategy to maintain a sense of well-being for the present time is to work and make the money he needs to subsist. This desire to support oneself and one’s family is seen across the board with our participants. As we see in the next section, however, Covid-19 interrupted financial stability and labor-market integration for some our participants. Employment and financial stability Finding employment upon migration was challenging enough; however, the pandemic presented new challenges to those who had found jobs. These challenges led to greater feelings of economic insecurity. Berta, a 45-year-old certified nurse who earned a bachelor’s degree, came to the continental U.S. because she could not continue her cancer treatment in Puerto Rico after the hurricane, during which she lost her home as well:When Maria happened, I had to come here. … My doctors there lost their offices, and back in Puerto Rico, everything was chaos. Everybody lost everything. I lost the house. My brother lost his house, … even the doctors, God bless. So, the American Cancer Society was the one that brought me here to Florida. After the American Cancer Society helped her move to Florida to continue her treatment, she stayed on disability only until she found a job at a call center. Berta described a form of endurance and resilience that Bonilla (2020) would refer to as “neoliberal resilience,” in which individuals take on roles that should be adopted by governing authorities. She could not work because of the effect of the chemotherapy, so she told us,You know what I had to do? I had to go to the news … to see if someone saw my story because I was desperate. … And so a Mexican man came. He called me when he saw the news, and he lent me a mobile home. … He thought I was going to die, … but he didn’t know that Puerto Rican women are very strong. If these challenges were not enough, Berta felt a very strong sense of vulnerability during the pandemic. She described her experience during this time as very difficult because for some time she had no work or income beyond her disability payment, which was not enough to live on:Well, it has been difficult as … I couldn’t work because of the situation. When the pandemic started, I was working in a call center for medical plans stuff. … In the place where I worked … we were more than 120 people in a single room. Then they had to start sending us all home, because, obviously, people started to get sick. … I spent more than six weeks without work, without payment, and without working because I was new. I had started in January and still had no benefits. Adding to the hardship of unemployment was the fact that Berta’s sister and mother and her sister’s boyfriend contracted Covid-19, and her sister’s boyfriend died. The pandemic also affected Nando’s employment. His hours were cut, and the new schedule that his employer proposed was unfeasible. He persevered by transitioning to another job in a factory:I was working, but in that job that I had, they cut my hours because of the pandemic, and then I had to stay home. I was working as … cleaning with pressure washer machines at the mall. They ended up cutting my hours, and … well, for a long time I was without … like four months passed. Then they called, but the schedule they wanted to give me wasn’t possible for me, so we couldn’t – I couldn’t go back there. Some months after I lost the job I had, I ended up working in another job, where things aren’t going so great. I work in a lid-packaging factory. Someone has to do it. Both Berta and Nando had to change course in terms of their employment due to Covid-19. These changes represent secondary stressors that they contended with after migration. These disruptions that affected their trajectories to societal incorporation were dealt with as challenges to adaptation that they took in stride. However, for some, the challenges surrounding each of these standalone disasters combine so that the pain presented by any of them spills over onto the others. Lily’s case illustrates this process. Lily, who is 52 years old and has a bachelor’s degree and whose husband has cancer, is very worried for their lives. She is even scared of going out for a walk. Her nephew’s wife died in Puerto Rico from Covid-19, and her nephew and children had the virus too. After explaining this, she mentioned that thinking about the hurricanes, especially on anniversaries, was like reliving them again, and she compared it to September 11. She mentioned how everything changed after Hurricane Maria and described the compounding effect of both the pandemic and hurricanes on lack of job security and uncertainty about moving forward in life. She was especially worried about her son, who was still living under a blue tarp as a roof in Puerto Rico, three years after the hurricane:In Puerto Rico, my nephew’s wife died from Covid, and my nephew has Covid, his daughter, her children too. … And they are locked up there. … They are locked up. [After a pause, she says, crying, that it is hard to talk about the hurricane.] You relive those moments, and it feels like it was happening right now. … It’s like a movie, seeing it all over again. … I say that it’s like 9/11, when that tragedy happened. It is very difficult. And it’s been three years now, and to think that my son still has that roof, that he hasn’t been able to do much. Now with this pandemic, the jobs are very limited. It is not easy. One begins to think, after Hurricane Maria, everything has been a total change. Lily’s case is a prime example of the accumulation of cascading disasters. As she discusses the effects of Covid-19, she transitions to the emotional response she has when thinking about the hurricane, comparing it to 9/11. Disaster upon disaster has left her conflating the pain she feels from both, exacerbated by the job insecurity that persists and worries about her family’s well-being. It’s important to note that there were things in Lily’s life that brought her joy. Above all, her greatest accomplishment was her family and the family unity they enjoyed in Puerto Rico. She recently had a month-long visit from two of her grandchildren, and she lit up telling us about that time. She also was grateful for technology: “I’m grateful for technology because I can see my grandchildren every day. They love me, and I love them.” Moreover, she and her husband were intentional about finding ways to connect with others. For example, prior to the pandemic, they attended English classes twice a week—one at the Catholic Church they belonged to and the other at a local Center for Hispanics. Though Covid-19 brought that to a halt, they still received calls from a few acquaintances they made who checked in on them, including the head of a local Puerto Rican nonprofit organization that assisted them. Lily indicated that these calls helped ease the isolation. In lieu of these activities, Lily and her husband would go to a local track and walk when there were not many people there, and they would do things like wash the car outdoors to get outside. Lily also had a sewing machine that kept her occupied. Speaking about looking for things that would entertain her mind, she said, “With my sewing machine, I’m always inventing stuff, the seams of pants for people I know. I keep myself busy. Even my dentist, … I made all the medical caps for the employees, and they are all happy.” In addition to that, she shared that she likes to cook and bake, which she was gearing up for in anticipation of Christmas. Despite the hardships, she believed that going through the hurricane was much more harrowing than the pandemic, as with the latter, she felt she could take precautions, whereas during the hurricane, she indicated that at one point she gave up trying to get the water out of their house and just sat down to pray until it ended. Thus, through everyday coping strategies, including her religion, hobbies, and connecting with her family, Lily continued with life. In sum, our participants have undergone various traumatic experiences since Hurricanes Irma and Maria that have affected how they have adapted to U.S. society and their pathways to societal incorporation. For some, the latest experience with the pandemic has interrupted their abilities to adapt and integrate and, at a broader level, has affected their overall well-being. However, as we also see, Puerto Ricans adapt to the challenges presented by either changing jobs if they can or, as in Berta’s case, appealing to the news or, as in Lily’s case, connecting with their family and trying to stay occupied. Moreover, the increase in the Puerto Rican population in the Tampa area also has helped, because migrants often turn to coethnics for help and social support. Thus, as Puerto Ricans adapt to their new communities, their influx in those same communities is helping newcomers as well. Though the coping strategies we have illustrated throughout speak to Puerto Ricans’ resiliency in overcoming hardship, we draw on Bonilla’s (2020) analysis of individual resilience to argue that these individual strategies should not be leveraged by other entities (e.g., governments) to abandon the needs of a population or to engage in further racialized neglect. Just because Puerto Ricans can show they are resilient does not mean they should be left to suffer on their own, when in fact, many of the sources of these challenges are structural in nature. Conclusion The Covid-19 pandemic has not only represented widespread illness, deaths, and hospitalizations; for Puerto Rican post-disaster migrants, it also has interrupted social and labor-market integration through the loss of work, educational challenges, and obstacles to their ability to nurture social ties. For some, these effects have piled onto accumulated social challenges that reach back prior to Hurricane Maria. The climate refugees who participated in our study left Puerto Rico because they were living under dire circumstances. Our analysis demonstrates that the emotional fallout from experiencing these hurricane-related conditions continued to affect them as the pandemic evoked those memories again. As Puerto Ricans attempted to rebuild their lives post-migration, our findings show that the accumulation of stressors that prolong trauma (Gil 2007, p. 615), which exacerbate the vulnerability of this population, may compromise their ability to recover. At the root of these disasters lies the preexisting inequalities experienced by the Puerto Rican people as a result of a “racialized neglect” on the part of the U.S. federal government (Bonilla 2020). Moreover, the systemic, geopolitical, and racialized governance that caused the cascading disasters after the 2017 hurricanes in Puerto Rico continue to affect our participants after migration. When the new hazard of Covid took place, our participants were not yet recovered from the multiple losses that led to relocation. Our findings reveal themes connected to integration theories, and these themes show that specific patterns of unattended vulnerabilities are affecting the adaptation and resilience of this community in the face of hazards. The coping strategies that they deploy to ease the burden of cascading disasters and promote resilience are also significant findings. Together our findings suggest that more research is needed on actions that can be taken to enhance mitigation and prevention of crises in the specific areas of education, housing, job security, language acquisition, and social ties. However, we believe it is imperative that disaster risk research seriously consider the colonial status and geopolitical relations between vulnerable places and the racialized systems of governance that control them so that proposed disaster mitigation strategies address past and current vulnerabilities in the context of colonial conditions. Finally, future research should also consider the experiences not just of those who have come to the continental U.S. but also of those who remained in Puerto Rico after the hurricanes and those who migrated stateside yet returned to Puerto Rico due to the challenges they encountered. Migration scholarship in particular should attend to the emotional impact of migration in a post-disaster context and prioritize the stories of climate refugees and their accompanying emotions, as this can enhance our understanding of the contexts that promote integration and what factors compel some to return to Puerto Rico. This could lead to community-level interventions in both societies that can be developed to assist in climate refugees’ adaptation. Understanding these factors, which surface through an analysis of emotions, can help identify the sociocultural conditions and systemic roots of migration and formulate policy recommendations to help Puerto Ricans upon resettlement. Appendix See Tables 1, 2 and 3.Table 1 Number and percent of Puerto Ricans, by select Florida Counties County No. of Puerto Ricans % of County pop Broward County 88,416 4.6 Hillsborough County 118,467 8.3 Miami-Dade County 97,755 3.6 Orange County 199,936 14.8 Osceola County 113,258 32.2 Polk County 63,890 9.3 Source: SocialExplorer.com ACS 2019 (5-Year Estimates) Table 2 Main demographic characteristics (N = 103) Variable Description Mean SD Gender Female 0.82 0.39 Male 0.18 0.39 Respondent’s age In years 46.74 14.75 Respondent’s skin color White/light-skinned 0.88 0.32 Black/dark-skinned 0.12 0.32 Marital status Married 0.35 0.48 Widowed 0.06 0.24 Separated/divorced 0.15 0.35 Single 0.43 0.50 Children Has children 0.85 0.36 Does not have children 0.15 0.36 Educationa Less than HS diploma 0.07 0.25 HS diploma 0.14 0.34 Technical/Vocational school 0.15 0.35 Some college 0.14 0.34 AA degree 0.16 0.36 BA/BS degree 0.23 0.42 Graduate/professional 0.13 0.33 Social class Middle class 0.26 0.44 Working class 0.53 0.50 Lower class 0.21 0.41 English fluencya Very well 0.16 0.36 Well 0.25 0.44 Not very well 0.47 0.50 Not at all 0.13 0.33 Time in US In months 28.24 5.77 Source: Puerto Rican Post-Disaster Migration Project aDue to rounding the total percentage exceeds 100% Table 3 Financial insecurity, social ties, and changes in employment Pre-Hurricanes Pre-COVID During COVID f % f % f % Financial insecurity  Very difficult 21 20.39 16 15.53 28 27.18  Somewhat difficult 22 21.36 26 25.24 34 33.01  Not difficult 60 58.25 61 59.22 41 39.81 M SD M SD Social ties  Loose ties 7 10.48 1.80 2.42  Close ties 5.73 7.68 2.04 2.31 F % Change in employment  Got a job 3 10.71  Temporarily fired 9 32.14  Permanently dismissed 2 7.14  Reduction in hours 4 14.29  Othera 10 35.71 aSource Puerto Rican Post-Disaster Migration Project, Survey Data; 1 company closed, unpaid leave, resigned Acknowledgements We would like to thank Maritza Novoa and Nancy Hernández for their collaboration on this project. We would also like to acknowledge Andrew Katz for his copyediting assistance and Elizabeth Vaquera for assistance with data analysis. Funding for this work has been supported by the National Science Foundation, Grant Number 1918241. 1 Loose ties were measured by asking respondents about the number of people they could ask for favors such as saving their mail if they went on a trip, and close ties by asking about number of friends or family they could share private feelings with. 2 All names of participants are pseudonyms that we use to protect their identity. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References AAPOR (American Association for Public Opinion Research) Standard definitions: Final dispositions of case codes and outcome rates for surveys 2016 9 Alexandria AAPOR Acosta-Belen E Santiago C Puerto Ricans in the United States: A contemporary portrait 2018 Boulder Lynne Rienner Alba RD Nee V Remaking the American mainstream: Assimilation and contemporary immigration 2009 Cambridge, MA Harvard University Press Aranda E Emotional bridges to Puerto Rico: Migration, return migration, and the struggles of incorporation 2007 Lanham, MD Rowman and Littlefield Aranda E Rivera FI Puerto Rican Families in Central Florida: Prejudice, discrimination, and their implications for successful integration Women, Gender, and Families of Color 2016 4 1 57 85 10.5406/womgenfamcol.4.1.0057 Bonilla Y The coloniality of disaster: Race, empire, and the temporal logics of emergency in Puerto Rico, USA Political Geography 2020 78 102181 10.1016/j.polgeo.2020.102181 Bonilla Y LeBrón M Aftershocks of disaster: Puerto Rico before and after the Storm 2019 Chicago Haymarket Books Charmaz K Constructing grounded theory 2014 Thousand Oaks, CA Sage Delerme S The Latinization of Orlando: Language, whiteness, and the politics of place Centro: Journal of the Center for Puerto Rican Studies 2013 25 2 60 95 Duany J Mobile livelihoods: The sociocultural practices of circular migrants between Puerto Rico and the United States International Migration Review 2002 36 2 355 388 10.1111/j.1747-7379.2002.tb00085.x Duany J Blurred borders: Transnational migration between the Hispanic Caribbean and the United States 2011 Chapel Hill University of North Carolina Press Duany, J. 2015. Florirricans en Crecimiento. El Nuevo Dia, 14 October. Duany, J. 2021. Rethinking Puerto Rican Migrants as Climate Refugees. Paper presented at the Bregando with Disasters: Post Hurricane Maria Realities and Resiliencies Symposium, Contemporary Art Museum, University of South Florida. Tampa: 2 October. Duany J Matos-Rodríguez F Puerto Ricans in Orlando and Central Florida Center for Puerto Rican Studies Policy Report 2006 1 1 1 39 Erikson KT Loss of communality at Buffalo Creek American Journal of Psychiatry 1976 133 3 302 305 10.1176/ajp.133.3.302 1259041 Gamarra, E. 2018. Nuestro Futuro Is in Your Hands: Puerto Ricans in Florida. Paper presented at Florida International University, Miami: 2 July. George Washington University Ascertainment of the estimated excess mortality from Hurricane María in Puerto Rico 2020 Washington, DC Milken Institute School of Public Health Gil DA Secondary trauma or secondary disaster? Insights from Hurricane Katrina Sociological Spectrum 2007 27 6 613 632 10.1080/02732170701574941 Gordon MM Assimilation in American life: The role of race, religion, and national origins 1964 New York Oxford University Press Grosfoguel R Colonial subjects: Puerto Ricans in a global perspective 2003 Berkeley University of California Press Hinojosa J Meléndez E Puerto Rican Exodus: One Year since Hurricane Maria Center for Puerto Rican Studies Research Brief 2018 5 1 8 Hinojosa J Román N Meléndez E Puerto Rican Post-maria relocation by States Center for Puerto Rican Studies Research Brief 2018 1 1 15 Hollifield JF Is migration a unique field of study in social sciences? A response to Levy, Pisarevskaya, and Scholten Comparative Migration Studies 2020 8 1 1 9 10.1186/s40878-020-00192-3 Kessler, R. C. 2010. Hurricane Katrina Community Advisory Group Study [United States]. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 10 June. Kessler RC McLaughlin KA Koenen KC Petukhova M Hill ED The importance of secondary trauma exposure for post-disaster mental disorder Epidemiology and Psychiatric Sciences 2012 21 1 35 45 10.1017/S2045796011000758 22670411 Kilic C Aydin I Taskintuna N Ozcurumez G Kurt G Eren E Lale T Ozel S Zileli L Predictors of psychological distress in survivors of the 1999 earthquakes in Turkey: Effects of relocation after the disaster Acta Psychiatrica Scandinavica 2006 114 3 194 202 10.1111/j.1600-0447.2006.00786.x 16889590 Klein N Smith N The shock doctrine: A discussion Environment and Planning D: Society and Space 2008 26 4 582 595 10.1068/d2604ks Kochhar, R., R. Suro, and S. Tafoya. 2005. The New Latino South: The Context and Consequences of Rapid Population Growth. Pew Hispanic Center, 26 July. Li M Pre-migration trauma and post-migration stressors for Asian and Latino American immigrants: Transnational stress proliferation Social Indicators Research 2016 129 1 47 59 10.1007/s11205-015-1090-7 Marrow H New destination dreaming: Immigration, race, and legal status in the Rural American South 2011 Stanford, CA Stanford University Press Massey D Capoferro C Massey Douglas The geographic diversification of American immigration new faces in new places: The changing geography of American immigration 2008 New York Russell Sage Foundation 25 50 McGhee G Marland GR Atkinson J Grounded Theory Research: Literature Reviewing and Reflexivity Journal of Advanced Nursing 2007 60 3 334 342 10.1111/j.1365-2648.2007.04436.x 17908129 McGowran P Donovan A Assemblage theory and disaster risk management Progress in Human Geography 2021 45 6 1601 1624 10.1177/03091325211003328 Melin K Maldonado WT López-Candales A Lessons learned from Hurricane Maria: Pharmacists’ perspective Annals of Pharmacotherapy 2018 52 5 493 494 10.1177/1060028017751691 29310451 Meyers T Hunt R The other global south The Lancet 2014 384 9958 1921 1922 10.1016/S0140-6736(14)62270-4 Parsons H Language and migration to the United States Annual Review of Anthropology 2011 40 227 240 10.1146/annurev-anthro-081309-145634 Pescaroli G Alexander D A definition of cascading disasters and cascading effects: Going beyond the ‘toppling dominos’ metaphor Global Risk Forum Davos Planet Risk 2015 3 1 58 67 Portes A Rumbaut RG Immigrant America 1996 Berkeley University of California Press Rand Corporation Hurricanes Irma and Maria: Impact and aftermath 2020 Homeland Security Research Division Homeland Security Operational Analysis Center, Rand Corporation Riosmena F Everett BG Rogers RG Dennis JA Negative acculturation and nothing more? Cumulative disadvantage and mortality during the immigrant adaptation process among Latinos in the United States International Migration Review 2015 49 2 443 478 10.1111/imre.12102 28104925 Rodríguez-Díaz CE María in Puerto Rico: Natural disaster in a colonial Archipelago American Journal of Public Health 2018 108 1 30 32 10.2105/AJPH.2017.304198 29211541 Sanchez, R. 2020. For Puerto Rico's Earthquake Families, Trauma Continues Thousands of Miles from Home. CNN, 29 January. Schleef DJ Latinos in Dixie: Class and assimilation in Richmond, Virginia 2009 Albany, NY SUNY Press Schuller M Maldonado J Disaster capitalism Annals of Anthropological Practice 2016 40 1 61 72 10.1111/napa.12088 Silver P Sunbelt Diaspora: Race, Class, and Latino Politics in Puerto Rican Orlando 2020 Austin University of Texas Press Silver P Vélez W ‘Let me go check out Florida’: Rethinking Puerto Rican Diaspora Centro: Journal of the Center for Puerto Rican Studies 2017 29 3 98 125 Social Explorer. n.d.-a. ACS 2019 (1-Year Estimates) Hispanic or Latino: Puerto Rican [Map]. https://www.socialexplorer.com/data/ACS2019/documentation/ Social Explorer. n.d.-b. ACS 2019 (5-Year Estimates) Hispanic or Latino: Puerto Rican [Map]. https://www.socialexplorer.com/data/ACS2019/documentation/ Thomas D Jang S Scandlyn J The CHASMS conceptual model of cascading disasters and social vulnerability: The COVID-19 case example International Journal of Disaster Risk Reduction 2020 51 101828 10.1016/j.ijdrr.2020.101828 32895627 U.S. Department of Homeland Security FEMA mismanaged the commodity distribution process in response to hurricanes Irma and Maria 2020 Washington, DC Office of the Inspector General, Homeland Security Vargas-Ramos C Venator-Santiago CR Displacement and migration in the age of climate change: Holyoke’s response to Hurricane Maria for Massachusetts’s Municipal Vulnerability Program 2019 Boston, MA Commonwealth of Massachusetts Velez W Burgos G The impact of housing segregation and structural factors on the socioeconomic performance of Puerto Ricans in the United States Centro: Journal of the Center for Puerto Rican Studies 2010 22 1 174 197 Venator-Santiago CR Mapping the contours of the history of the extension of U.S. Citizenship to Puerto Rico, 1898–Present Centro: Journal of the Center for Puerto Rican Studies 2017 29 1 38 Walsh F Traumatic loss and major disasters: Strengthening family and community resilience Family Process 2007 46 2 207 227 10.1111/j.1545-5300.2007.00205.x 17593886 Waters MC Jiménez TR Assessing immigrant assimilation: New empirical and theoretical challenges American Review of Sociology 2005 31 105 125 10.1146/annurev.soc.29.010202.100026 Winders J Changing politics of race and region: Latino migration to the US south Progress in Human Geography 2005 29 6 683 699 10.1191/0309132505ph577oa

0

PMC9753070

NO-CC CODE

2022-12-16 23:25:28

no

Lat Stud. 2022 Dec 15;:1-24

utf-8

Lat Stud

10.1057/s41276-022-00390-3

oa_other

==== Front AIDS Behav AIDS Behav AIDS and Behavior 1090-7165 1573-3254 Springer US New York 3943 10.1007/s10461-022-03943-8 Substantive Review Determinants of Pre-exposure Prophylaxis (PrEP) Implementation in Transgender Populations: A Qualitative Scoping Review Zamantakis Alithia 2 Li Dennis H. 123 Benbow Nanette 13 Smith Justin D. 5 http://orcid.org/0000-0001-9222-5116 Mustanski Brian brian@northwestern.edu 1234 1 grid.16753.36 0000 0001 2299 3507 Department of Psychiatry, Northwestern University Feinberg School of Medicine, Chicago, IL USA 2 grid.16753.36 0000 0001 2299 3507 Institute for Sexual and Gender Minority Health and Wellbeing, Northwestern University, Chicago, IL USA 3 grid.16753.36 0000 0001 2299 3507 Center for Prevention Implementation Methodology, Northwestern University Feinberg School of Medicine, Chicago, IL USA 4 grid.16753.36 0000 0001 2299 3507 Department of Medical Social Sciences, Northwestern University Feinberg School of Medicine, Chicago, IL USA 5 grid.223827.e 0000 0001 2193 0096 Department of Population Health Sciences, Spencer Fox Eccles School of Medicine at the University of Utah, Salt Lake City, UT USA 15 12 2022 119 21 11 2022 © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. We conducted a scoping review of contextual factors impeding uptake and adherence to pre-exposure prophylaxis in transgender communities as an in-depth analysis of the transgender population within a previously published systematic review. Using a machine learning screening process, title and abstract screening, and full-text review, the initial systematic review identified 353 articles for analysis. These articles were peer-reviewed, implementation-related studies of PrEP in the U.S. published after 2000. Twenty-two articles were identified in this search as transgender related. An additional eleven articles were identified through citations of these twenty-two articles, resulting in thirty-three articles in the current analysis. These thirty-three articles were qualitatively coded in NVivo using adapted constructs from the Consolidated Framework for Implementation Research as individual codes. Codes were thematically assessed. We point to barriers of implementing PrEP, including lack of intentional dissemination efforts and patience assistance, structural factors, including sex work, racism, and access to gender affirming health care, and lack of provider training. Finally, over 60% of articles lumped cisgender men who have sex with men with trans women. Such articles included sub-samples of transgender individuals that were not representative. We point to areas of growth for the field in this regard. Resumen En este revisión de alcance, examinamos los factores contextuales que impiden la adopción y el cumplimiento de la profilaxis previa a la exposición en las comunidades transgénero. Este revisión sistemática se formó a partir de una revisión sistemática más grande. Utilizando un proceso de selección de aprendizaje automático, filtración de los titulus y examines, y revision del texto complete, el primer revisión sistemática identificó 353 artículos por el analisis. Estes artículos fueron estudios revisados por pares, relacionados con la implementación de la PrEP en los EE.UU. publicados despues de 2000. Veintidós artículos se identificaron en esta b?squeda como relacionados con personas transgénero. Se identificaron once artículos adicionales a través de citas de estos veintidós artículos, lo que resultó en treinta y tres artículos en el análisis actual. Estos treinta y tres artículos fueron codificados cualitativamente en NVivo utilizando construcciones adaptadas del Marco Consolidado para la Investigación de Implementación (CFIR) como códigos individuales. Los códigos fueron evaluados temáticamente. Señalamos las barreras de la implementación de la PrEP, como la falta de esfuerzos intencionales de difusión y asistencia al paciente, las barreras estructurales como el trabajo sexual, el racism, y el acceso a la salud de afirmación de género, y la falta del entrenamiento de los doctores. Finalmente, más de sesenta por ciento de los artículos tuvieron submuestras de personas transgénero que no eran representativas. Se?alamos áreas de crecimiento para el campo en este sentido. Keywords HIV/AIDS Pre-exposure prophylaxis Transgender health Implementation science Determinants of implementation Scoping review Qualitative analysis http://dx.doi.org/10.13039/100015691 Division of Microbiology and Infectious Diseases, National Institute of Allergy and Infectious Diseases P30 AI117943 ==== Body pmcPre-exposure prophylaxis (PrEP) is an oral pill taken once daily, or, as of 2022, a bimonthly injection, to prevent the acquisition of HIV. There are two options for oral PrEP—Truvada (emtricitabine/tenofovir disoproxil fumarate) and Descovy (emtricitabine/tenofovir alafenamide)—and one option for long acting injectable (LAI) PrEP—Apretude (cabrotegravir). When taken daily or injected bimonthly, PrEP has been found to reduce likelihood of HIV acquisition by 99%, with 60% effectiveness in real world trials (Jourdain, de Gage, Desplas, & Dray-Spira, 2022; Raphael J Landovitz et al., 2021; Mayer et al., 2020). PrEP has been found effective at reducing rates of HIV seroconversion for transgender women, transgender men, and nonbinary people (Grant et al., 2020), in addition to cisgender populations (i.e., those who identify with the sex they were assigned at birth). Provider prescription of PrEP has increased, with 25% of those indicated for PrEP use by the Centers for Disease Control and Prevention (CDC) criteria (e.g. cisgender men who have sex with men [cisgender MSM], transgender women, and Black/Latinx individuals, among others) having a prescription for PrEP in 2020 compared to only about 3% in 2015 (Centers for Disease Control and Prevention, 2020). However, large disparities in PrEP uptake and reach exist across minoritized populations, including Black, Latinx, transgender, and/or youth populations (Kuehn, 2018; Poteat & Radix, 2020; Spinelli and Buchbinder, 2019). A recent study details that these disparities may have widened between 2015 and 2020 (Kamitani, Johnson, Wichser, Adegbite, & Mullins, 2020). The first National HIV Behavioral Surveillance System’s Transgender Cycle found 92% of HIV-negative transgender women surveyed were aware of PrEP, but only 32% reported using it (Centers for Disease Control and Prevention, 2021). Recent research found that while 84.1% of transgender men who have sex with men (transgender MSM; i.e. men who were assigned female at birth and who have sex with other men) surveyed had heard of PrEP, only 28% of all participants reported using PrEP (Reisner, Moore, Asquith, Pardee, & Mayer, 2021). Implementation science enables analysis of the gap between knowledge and uptake of PrEP that can inform strategies for effectively bringing this intervention to this population with unmet need. While myriad evidence-based innovations (EBIs; e.g., PrEP, motivational interviewing, housing first) have been identified for HIV prevention and treatment, in addition to other illnesses and infections, researchers have identified a gap, referred to as the implementation gap, of seventeen years between identification of an EBI and its translation into real-world settings (Morris, Wooding, & Grant, 2011). That is, if the EBI is ever implemented in real-world settings at all. Implementation science is the study of how to increase uptake of EBIs (referred to as “innovations”), promote adherence and/or sustainment, and, thus, improve quality and effectiveness (Bauer, Damschroder, Hagedorn, Smith, & Kilbourne, 2015). The U.S. plan to end the HIV epidemic requires speedy translation of findings into practice, ensuring that the right innovations are utilized to their greatest effect by highly vulnerable populations (US Department of Health and Human Services, 2022). Implementation science examines what changes are needed in policy, practice, procedure, or behavior to actualize that goal in real time. Researchers know that PrEP is efficacious and effective and have identified disparities in uptake and adherence. Implementation science takes the reigns from there to identify how to then overcome barriers leading to disparities. Within implementation science, barriers, as well as facilitators, to implementing, adapting, modifying, and/or improving innovations are referred to as determinants of implementation; that is, they enable and/or hinder adoption, acceptability, and reach of an innovation, among other outcomes (Wensing et al., 2011). A predominant framework for conceptualizing determinants is the Consolidated Framework for Implementation Research (CFIR). CFIR has five domains, or categories of determinants, each broken down by sub-constructs (Damschroder et al., 2009). CFIR domains include: innovation characteristics (e.g. innovation source, evidence), inner setting (i.e. factors of implementation internal to an organization—may be identified by studies in delivery settings but may also include perceptions of patients), outer setting (i.e. sociopolitical, structural, and environmental factors), characteristics of individuals (including providers and patients), and process determinants (e.g. how engaging patients in implementation may shape the outcome of delivery or uptake). Building on a larger systematic review (Li et al., 2022) of determinants of PrEP implementation for all populations, this manuscript focuses on transgender populations. While the larger review provides breadth, this manuscript contributes depth by focusing on a specific population and qualitatively exploring how factors hinder and enable implementation, how researchers identify determinants, and how specific constructs are experienced by transgender people and those that provide care for them. However, most articles included focus on transgender women, with few including transgender men and/or nonbinary people. While we included four studies that explicitly examined determinants of PrEP implementation for transgender men and one that did so for nonbinary transfeminine individuals, given the scope of the literature that we review, this manuscript is more centrally focused on determinants of implementation for transgender women, except when explicitly detailed otherwise. In this review, we ask: What barriers and facilitators have been identified as determinants of PrEP implementation for transgender populations, and how are these barriers and facilitators experienced by transgender people? What existing gaps in the literature on potential barriers and facilitators need greater analysis? How do study designs, samples, and methods hinder or enable identification of determinants of PrEP implementation for transgender populations? Methods The authors conducted this scoping review as an in-depth, qualitative analysis of the transgender population within a previously published systematic review of PrEP for all populations (Li et al., 2022). Between November 2020 and January 2022, our team conducted a database search of studies related to implementation of PrEP. The full protocol of the initial systematic review has been described elsewhere (Li et al., 2022). Through a multi-step, computerized, and manual-screening process, 50,128 unique records were identified through a database search of Ovid MEDLINE, PsycINFO, and Web of Science. The number of unique records was then reduced through deduplication, title and abstract screening, and full-text review (see Fig. 1 for the PRISMA-ScR and full details of exclusion by number). The initial systematic review included 353 articles for analysis. Authors of the initial systematic review coded studies using CFIR constructs as codes (Damschroder et al., 2009), as well as by common implementation settings (e.g., HIV/AIDS health centers, LGBT health centers, pharmacies) and priority populations (e.g., Latinx communities, transgender communities). The first author (az) conducted full-text review of each of the twenty-three articles coded through this process as “transgender population” and found that one article simply mentioned transgender populations but did not include them in their study or analysis. This article was excluded, reducing the number of articles to twenty-two. An additional eleven articles were identified from the citations of these twenty-two articles as potentially relevant to this scoping review. These eleven articles were not identified in the initial review, as many were published after the initial computerized search or were published in more niche journals that were not uploaded that our initial search did not identify. After screening, these eleven articles were included into the scoping review, resulting in a total of 33 articles. These thirty-three articles were uploaded to NVivo (Version 1.6.1) and qualitatively coded using adapted CFIR constructs (Damschroder et al., 2009; Li et al., 2022) as individual codes. Additional codes included study methods; purpose of the study; whether analyses and results lumped transgender women and cisgender MSM and/or lumped transgender men and cisgender women (i.e., women assigned female at birth); and whether multi-population analyses had large enough transgender sub-populations for substantive analysis. We added two additional codes as outer setting barriers, “cissexism” and “racism,” and used these two sociopolitical concepts as codes. Segments coded as cissexism and/or racism included findings from studies that demonstrated the presence of interpersonal, institutional, and/or systemic discrimination, and oppression that shaped the health outcomes and experiences of transgender patients and/or patients of color. This coding scheme was intentionally chosen in order to more explicitly attend to health inequities experienced by Black transgender women and other transgender populations as structural and systemic (Shelton, Adsul, & Oh, 2021). Codes were thematically assessed by the first author (az), examining not only the prevalence of analysis of determinants but also how determinants were measured, what authors found, and how they analyzed the data. Axial coding was also performed to examine differences across year of publication, intersecting population(s), and types of method(s) used by researchers. Analysis of the codes “cissexism” and “racism” occurred intersectionally, with an attentiveness to the ways in which the two sociopolitical forces differentially shape health and implementation outcomes. We used intersectional theory (Bowleg, 2012; Crenshaw 2018) to elucidate the interlocking nature of various axes of power (e.g. race, gender, sexual orientation, etc.). We conducted a search for existing reviews of determinants of PrEP implementation in transgender populations. We did not identify any existing published reviews of PrEP implementation that focused on transgender populations; however, we identified several ongoing reviews by Matos et al. (PROSPERO ID: CRD42021239360), Peixoto et al. (PROSPERO ID: CRD42020154059), Canoy, Hannes, and Thapa (PROSPERO ID: CRD42018089956), Algarin et al. (PROSPERO ID: CRD42019130858) and Garcia, Rehman, and Mbuagbaw (PROSPERO ID: CRD42022300631). These reviews examine factors associated with PrEP in transgender populations. Our scoping review differs, in that, we sought to qualitatively explore factors leading to and/or hindering implementation of PrEP solely for transgender people. Further, our review goes beyond identifying determinants of implementation and analyzes how determinants are hindering and/or enabling uptake and adherence. To the best of our knowledge, no published or registered ongoing review examines determinants of PrEP implementation along the PrEP cascade for transgender people of all genders and races, as a group separate from cisgender people, using implementation science frameworks. Results Delivery Setting Four studies examined determinants of PrEP implementation in non-specialty primary care (Brooks et al. 2019; Chan et al. 2019; Rhodes et al. 2020; Wu et al. 2020), one in pharmacies (Havens et al., 2019), and five in multiple settings (Carter et al. Jr 2022; Cohen et al. 2015; Hoenigl et al., 2019; Rael et al. 2018; Sevelius, Keatley, Calma, & Arnold, 2016). The remaining twenty-three studies examined determinants broadly, without attention to specific delivery settings. The single study focused on PrEP delivery in pharmacies examined the acceptability and feasibility of implementing PrEP in pharmacies and identified determinants in CFIR’s inner and outer settings and characteristics of individuals domains (Havens et al., 2019). This study described delays in communication between pharmacists and medical providers and pharmacist discomfort with discussing sexual histories as barriers to PrEP delivery that could be enhanced to facilitate better delivery in the future. Studies examining delivery in non-specialty primary care or multiple settings identified determinants in each of the CFIR domains but with little attention to process determinants. Study Methods There was little variation in identification of barriers and facilitators across study methods (see Table 1). Most articles utilized either qualitative methods or quantitative methods, with only three utilizing a mixed-methods approach (Rael et al. 2018; Theodore et al. 2020; Zarwell et al., 2021). Across study methods, outer setting determinants, characteristics of individuals (patients), and innovation characteristics were the most analyzed CFIR domains. However, qualitative articles were more likely to identify process determinants than quantitative or mixed-methods articles (n = 5, 2, & 0 respectively). Qualitative articles identified the facilitative role that engaging with consumers/patients plays in implementing PrEP in transgender communities (Brooks et al. 2019; Carter et al. Jr 2022; Galindo et al. 2012; Klein and Golub 2019; Reback et al. 2019). Table 1 Number of measured determinants by CFIR 2.0 construct and study method among n = 33 articles CFIR 2.0 Domain Qualitative Articles Quantitative Articles Mixed Methods Articles Innovation Characteristics 7 6 2 Outer Setting 9 9 1 Inner Setting 3 2 0 Characteristics of Individuals 9 8 2 Process 5 2 0 Years of Publication Over 72% (n = 24) of articles were published between 2018 and 2022. One article was identified from 2012 (Galindo et al. 2012), one from 2013 (Golub et al. 2013), one from 2015 (Cohen et al. 2015; Liu et al. 2016), one from 2016 (Liu et al. 2016), three from 2017 (Eaton et al. 2017; Landovitz et al. 2017; Wood et al. 2017), and two from 2018 (Lalley-Chareczko et al. 2018; Rael et al. 2018). The single article from 2012 examined barriers and facilitators that we coded in each of the five CFIR domains through focus groups with transgender women (Galindo et al. 2012). Of the nine articles prior to 2018, only one focused on barriers and facilitators of PrEP implementation specifically for transgender communities (Sevelius, Keatley, et al., 2016). The other seven each categorized transgender women with cisgender MSM in their study design and focused on cisgender MSM more so than their smaller sub-samples of transgender women. Articles published 2018 and after largely identified barriers in the outer setting, barriers and facilitators vis-à-vis characteristics of individuals, and barriers regarding innovation characteristics. Study Priority Populations Twenty of the thirty-three articles (60.6%) examined transgender women and cisgender MSM together. Only nine (27.3%) attended to determinants for Black transgender communities, and eight (24.2%) identified determinants for Latinx transgender communities. Only three identified determinants for transgender people who use injectable drugs (PWID) (Chan et al. 2019; Havens et al., 2019; Wu et al. 2020). While there was little variation in determinants identified across priority populations, articles that included transgender PWIDs in their studies largely examined outer setting determinants, with none identifying process or innovation characteristic determinants. Innovation Characteristics Fifteen articles (45.5%) measured and identified determinants regarding innovation characteristics (see Table 2). Another five mentioned determinants identified by other scholars, for example, noting identified disparities and issues of medical mistrust when introducing the study, but these articles did not study innovation characteristics as determinants of implementation within their own research. Innovation characteristics refer to core and adaptable pieces of an innovation (Damschroder et al., 2009). Study authors utilized focus group, interview, and/or survey data to assess patients’ and—to a lesser extent—providers’ perceptions of PrEP’s complexity, and cost. Four articles identified “acceptable” and “unacceptable” side effects of PrEP: using semi-structured interviews with six transgender women, Galindo et al. (2012) highlighted that participants deemed mild (e.g., headaches, nausea) and/or short-term side effects as acceptable and long-term and more serious side-effects (e.g., bone damage) as unacceptable. However, Rael et al. (2018) found that participants from four focus groups (n = 18 participants) were more seriously concerned about stomach pain and nausea than Galindo et al.’s sample. Rael et al. (2020; 2021) were also the only study authors to focus on determinants of injectable PrEP implementation; using focus group data with a sample of eighteen in 2020 and interviews with a sample of fifteen in 2021, they found participant concerns lay in where the injection occurs. Participants felt that the gluteal muscle was unacceptable due to fears of disfiguration or scarring in the area, as well as potential incompatibility of gluteal injections due to prior or future silicone injections. However, these same participants displayed excitement at the idea of injectable PrEP and felt that it was more compatible with their routine injections of estradiol than oral PrEP. Table 2 Number and proportion of articles measuring v. mentioning determinants by CFIR 2.0 construct among n = 33 articles CFIR 2.0 Domain & Construct Mentioning Proportion of Articles Only Mentioning (%) Measuring Proportion of Articles Measuring (%) Innovation Characteristics Innovation source 4 12.1 12 36.4 Evidence strength & quality 7 21.2 7 21.2 Relative advantage 2 6.1 1 3.0 Adaptability 1 3.0 0 0 Trialability 0 3.0 1 3.0 Complexity 4 12.1 8 24.2 Design quality and packaging 6 18.2 7 21.2 Cost 1 3.0 8 24.2 Other intervention characteristic 2 6.1 8 24.2 Subtotal 5 15.2 15 45.5 Outer Setting Mass disruptions 0 0 0 0 Population needs & resources 12 36.4 11 33.3 Community characteristics 7 21.2 16 48.5 Partnerships & connections 2 6.1 2 6.1 Market forces 0 0 0 0 External policy & incentives 2 6.1 3 9.1 Other outer setting characteristic 1 3.0 1 3.0 Subtotal 8 24.2 19 57.6 Inner Setting Structural characteristics 0 0 0 0 Networks & communications 0 0 3 9.1 Culture/climate 0 0 2 6.1 Tension for change 0 0 0 0 Compatibility 0 0 1 3.0 Relative priority 0 0 0 0 Organizational incentives & rewards 0 0 0 0 Goals & feedback 0 0 0 0 Leadership engagement 1 3.0 0 0 Available resources 1 3.0 3 9.1 Access to knowledge and information 0 0 3 9.1 Other inner setting characteristic 0 0 1 3.0 Subtotal 2 6.1 5 15.2 Characteristics of Individuals—Patients Knowledge & beliefs about the intervention 7 21.2 10 30.3 Self-efficacy 0 0 8 24.2 Individual stage of change 0 0 5 15.2 Individual identification with organization 0 0 0 0 Other personal attributes 2 6.1 10 30.3 Other individual characteristic 2 6.1 17 51.5 Subtotal 3 9.1 19 57.6 Characteristics of Individuals–Providers Knowledge & beliefs about the intervention 0 0 3 9.1 Self-efficacy 1 3.0 0 0 Individual stage of change 1 3.0 1 3.0 Individual identification with organization 0 0 0 0 Other personal attributes 0 0 1 3.0 Other individual characteristic 0 0 6 18.2 Subtotal 1 3.0 7 21.2 Process Planning 0 0 0 0 Engaging 1 3.0 6 18.2 Executing 0 0 2 6.1 Reflecting & evaluating 0 0 0 0 Other process 0 0 1 3.0 Subtotal 1 3.0 7 21.2 As a CFIR construct, complexity refers to the perceived difficulty of an intervention (Greenhalgh et al. 2004). Studies that attended to the complexity of PrEP described a recurring question of whether a daily pill is acceptable to transgender patients. While articles noted that the use of a daily pill could be integrated into other pill regimens (including hormone replacement therapy [HRT]), they also note that there are many reasons that may be undesirable for transgender people. For example, Rael et al. (2018) described three patient concerns regarding oral PrEP. Patients worried about the size of PrEP and difficulty of swallowing larger pills, as well as a lack of desire to add another daily pill to an already complex hormone regimen that may include taking estradiol in the morning and/or evening along with one or two androgen blockers and, perhaps, progesterone. Further, participants explained that patients who are homeless, impoverished, or otherwise socioeconomically marginalized may have higher priorities than remembering to take PrEP each day. Finally, the addition of recurring visits for lab tests can be cumbersome for some transgender individuals who are unable to fit another doctor appointment into their routine, particularly for those in rural areas who must travel to access gender-affirming care, let alone PrEP-care. While injectable or long-acting PrEP, including Apretude (cabotegravir), may solve the problem regarding pill size and daily use, Rael et al. found that “overwhelmingly, participants felt that visits with their healthcare provider to administer injectable PrEP were cumbersome and inconvenient” (2020, p. 1455). Eight articles identified participant concerns regarding the costs of PrEP. A lack of knowledge of PrEP among patients across studies also resulted in a lack of awareness that there are payment assistance programs, including that of Gilead, the manufacturer of both Truvada and Descovy. Further, Klein and Golub (2019) detailed participants’ desires for “active assistance” vis-à-vis cost support. Klein and Golub conceptualized “active assistance” as “navigation services that include help with obtaining/maintaining insurance and other payment options” (2019, p. 266) through patient navigation and by facilitating systemic change in cost and access to PrEP. Interestingly, Rael et al.’s (2018) participants were the sole sample to not identify cost as a significant barrier to PrEP. The authors utilized focus groups with eighteen participants, and nearly the entire sample had prior knowledge of PrEP. Instead of concerns with cost, they were more concerned with pill side effects (long-term and short-term) and pill size. However, this may be in part due to the location of the study, with participants solely from New York City, where, as Rael et al. mention, “the Department of Health, public health clinics, and other community-based organizations proactively outreach to city residents to enroll qualifying individuals in healthcare plans, including Medicaid” (2018, p. 12). Study authors also noted patient desires for stronger evidence regarding the efficacy of PrEP for transgender women, as well as transgender people as a whole. Additionally, study authors noted a need for studies of potential cross-interactions between HRT and PrEP. This note came from studies of transgender women participants, thus only focusing on patient concerns of cross-interactions of estradiol and PrEP but not testosterone HRT. Regarding the innovation source, or who/what organization is promoting the innovation, twelve articles utilized focus group and interview data to identify a lack of trust in medical providers and researchers by transgender participants, and two articles found similar results through quantitative analyses. For example, 16.7% of Restar et al.’s (2018) sample (n = 230 participants) cited mistrust with providers and researchers as reasons for not wanting to take PrEP. Regression analysis detailed that access to trans-affirming providers significantly increased acceptability of PrEP amongst transgender participants. Finally, seven articles (21.2%) examined design quality and packaging of PrEP, including marketing, as a determinant of implementation in transgender communities. Such articles highlighted findings from qualitative research that the marketing for PrEP has resulted in transgender women feeling as though PrEP is not for them. This is due to an overwhelming marketing of PrEP towards cisgender MSM in advertisements and by providers. A transfeminine participant from Klein et al.’s (2019) focus group research described an experience of a provider handing them a flyer for PrEP. The flyer asked if the reader of the flyer was a cisgender MSM and/or a transgender woman. This direct lumping of cisgender MSM and transgender women resulted in the participant declining PrEP. Indeed, 90% of Klein et al.’s sample (n = 30 participants) felt that poor uptake of PrEP among transgender populations is due to poor marketing. Participants across studies also felt that even when marketing focuses on transgender women, the models/actresses “pass” as cisgender, or in other words are not “noticeably” transgender, and that such marketing often does not include transgender women of all body sizes, transgender women who engage in survival sex work, non-monogamous transgender women, and/or other diverse representations of transgender women. Thus, participants desired marketing that addresses transgender patients’ fears regarding cross-interactions between HRT and PrEP, as well as cost assistance program information. Sevelius et al. (2016), for example, detailed that empowering, gender-affirming, and sex-positive marketing can serve to facilitate uptake and acceptability of PrEP for transgender populations . Outer Setting Nearly all articles (n = 27; 81.8%) mentioned determinants of PrEP implementation in the outer setting for transgender communities and patients. Fewer articles (n = 19; 57.6%) measured determinants in the outer setting. No article examined market forces (e.g., pressure to be competitive by providing additional or unique services/programming) or mass disruptions (e.g., inability to manufacture or transport PrEP due to local, state, or national emergencies). Only a small number looked at the role of external policy and incentives or partnerships and connections as barriers and/or facilitators (n = 5 and n = 4 respectively). Eleven articles (33.3%) identified disparities in prevalence as determinants of implementation. All twenty-three noted that transgender women have a higher risk of HIV infection compared to cisgender women, with higher lifetime risks of HIV acquisition for Black and/or Latina transgender women than for white transgender women, transgender women engaged in sex work, and transgender women who inject drugs. These articles highlighted structural factors, such as incarceration, homelessness, and poverty, are barriers to preventing HIV acquisition and increasing PrEP uptake. However, some also found that, when presented with information about PrEP and disparities in HIV rates and PrEP uptake, transgender women were more likely to desire PrEP and were upset that they did not previously know about PrEP. For example, transgender women sex workers identified support to initiate PrEP from other transgender women engaged in sex work (Brooks et al. 2019; Sevelius, Keatley, et al., 2016). In these cases, PrEP was viewed as a mechanism to protect themselves and their clientele. Wilson et al. (2019) additionally elucidated that transgender women’s HIV acquisition rates are different and more varied than cisgender MSM, noting that the repetitive grouping of transgender women and cisgender MSM in PrEP-related implementation research may be preventing more robust analysis and more accurate PrEP guidelines and specific supports for transgender women. Sixteen articles (48.5%) also identified community characteristics as barriers in PrEP implementation. Community characteristics include social determinants of health (SDOH), the built environment, and discrimination. These included disparities in homelessness, poverty, incarceration, engaging in survival sex work, unemployment, and intimate partner violence, as well as a lack of access to gender-affirming health care, health insurance, and familial support, and lower rates of educational attainment. Articles detailed that these disparities are informed by experiences of discrimination, as well as structural level factors (e.g., policy), and as such, we coded many of these disparities as part of a larger structure of cissexism and racism. Brooks et al. (2019), for example, discussed the intersecting roles of cissexism and racism for Black and Latina transgender women who reported many incidents of cisgender women of color’s viewing all Black and Latina transgender women as HIV-positive, vectors of HIV, and “whores” who take no health precautions. The interview participants of Brooks et al. explained that when cisgender women see their PrEP, they use it as further evidence that Black and Latina transgender women are HIV-positive and are keeping it secret from others. Mehrota et al. (2018) surveyed individuals who participated in the original iPrex trials that found PrEP to be efficacious. They found a significant relationship between social integration and PrEP uptake; thus, the greater connection to an LGBT community, the greater probability that an individual (i.e., cisgender MSM or transgender woman) would be likely to initiate and maintain PrEP use. Access to LGBT community, though, is not universal across the US, and Latina transgender women reported to Rhodes et al. (2020) that their migration to the US came with disappointment, as they expected greater social acceptance. Spaces created to provide services for women—often targeting cisgender women—were identified as spaces that are not yet trans-inclusive enough to enhance transgender women’s health care access. Further, the presence of LGBT and trans-inclusive agencies/organizations does not guarantee that transgender individuals are able to access such spaces. Due to fears of deportation and criminalization, undocumented transgender women, transgender women engaged in sex work, and transgender women who use drugs and/or street hormones may avoid even inclusive agencies. Using baseline survey data from a behavioral intervention for transgender women, Restar et al. (2018) found, in part, that creating partnerships between PrEP prescribers and gender-affirming care serves a facilitative role to PrEP uptake. Finally, Klein and Golub’s (2019) transgender women and nonbinary interview participants reported frustration with researchers and implementers who continue to group transgender women and nonbinary individuals with cisgender MSM, with such grouping serving as a barrier to PrEP uptake. Inner Setting Only five articles (15.2%) identified inner setting determinants of PrEP implementation. The studies that identified inner setting determinants utilized qualitative methods—primarily in-depth interviews and focus groups—and only noted the inner setting briefly. Instead, the focus was often on outer setting determinants or characteristics of individuals, and inner setting determinants were mentioned as a small portion of the article. Four articles elucidated the facilitative role of available resources, noting that staff capacity building, technical assistance, agency capacity to provide transportation to patients, and the ability to integrate gender-affirming care into PrEP services or vice versa may each increase PrEP uptake among transgender patients (Carter et al. Jr 2022; McMahan et al. 2020; Sevelius, Keatley, et al., 2016). Article authors also point to the importance of leadership buy-in for the development of an equity-centered culture within organizations to facilitate patient-centered approaches to care. For example, Carter et al. (2022) evaluated three health departments funded by a CDC health equity cluster grant. Using mixed methods from the three health departments’ projects, the authors cited a need for CBO staff to have access to ongoing training that provides information and skills needed to provide PrEP to transgender patients. Characteristics of Individuals Twenty articles (60.6%) identified individual-level patient and provider barriers and facilitators, with more attention to patient determinants (n = 19; 67.6% of all articles) than provider (n = 7; 21.2% of all articles). At the provider-level, study authors detailed participant reports of their providers’ telling them to not disclose PrEP use to others. While these providers encouraged their patients to keep their PrEP use a secret to mitigate stigma, interview participants felt it made them more cognizant of the stigma placed on transgender women’s use of PrEP. Other interview participants wished that their providers would grow in comfort and capacity to navigate discussions of sexual assault, sex work, and other experiences transgender women experience at a disproportionate rate. Finally, Sevelius, Keatley, Calma, and Arnold (2016) argued from interviews with transgender women that “providers and clinics that serve MSM are not necessarily equipped to recruit, retain, and provide care to transgender women” (2016, p. 1073). They explained that providers and implementers often expect that they can map cisgender MSM interventions, strategies, and guidelines onto transgender women due to a shared sex assignment at birth, but unique barriers, needs, and lived experiences of transgender women require providers and clinicians to expand their knowledge and toolsets. At the patient-level, the largest barrier noted was a lack of knowledge and awareness about PrEP. Additionally, though transgender individuals may have had knowledge of PrEP, they did not always believe that it was a necessary intervention for them. This, in part, was due to the marketing of PrEP to cisgender MSM, resulting in transgender women feeling that PrEP is only for cisgender MSM. This was also due to perceptions that an individual needs to be engaging in a high rate of unprotected sex to be PrEP indicated. Articles also identified barriers in self-efficacy. For example, McMahan et al. (2020) found that transgender women who used methamphetamine sought daily reminders to take their PrEP as their drug use made it difficult to remember. Brooks et al. (2019) conducted focus groups with Black and Latina transgender women and found low self-efficacy vis-à-vis condom use because of interrelationship power dynamics that made it difficult for Black and Latina transgender women to negotiate the use of a condom. Their participants reported fears of even lower confidence in ability to demand condom usage if their partners knew they were on PrEP. Many participants across studies did not feel confident in their ability, or desire, to take a pill every day, and some did not want to have to start with oral PrEP to access injectable PrEP (Rael et al. 2020), a step that was required in injectable PrEP efficacy trials. Additional barriers lay in Black and Latina transgender women’s high rates of mental illness, making daily adherence to oral PrEP difficult; lower educational attainment, which limits health literacy; and prioritization of hormone replacement therapy. One barrier arose for those who were on HRT: HRT can require numerous appointments for doctor check-ups and routine blood work, and transgender women on HRT did not feel they had the time or capacity to have to add on additional frequent appointments for PrEP (Sevelius, Keatley, et al., 2016). Lastly, Brooks et al. (2019) detailed experiences of “anticipated stigma” for heterosexual, transgender women. One transgender Latina participant explained to researchers: Most of the sexual partners that I have sex with are heterosexual men. So…when you tell them PrEP, they’re like, ‘‘What’s that?’’ And then I don’t want to go explain…because then they’re going to start thinking, “Oh, you’re HIV-positive.” (p. 192) A lack of in PrEP (and general HIV) awareness by cisgender, heterosexual men can make such conversations difficult for transgender, heterosexual women. Facilitators for PrEP initiation and maintenance included gender affirmation, with transgender women who had socially or medically transitioned being more likely to engage in protective behaviors including PrEP use; familial and peer acceptance of their identity; and engagement in sex work. Transgender women who engaged in sex work experienced encouragement and celebration from fellow sex workers for initiating PrEP. Further, transgender women engaging in sex work felt that PrEP offered them the possibility of gaining some control over their bodies that they often do not have otherwise due to pressure from clients to forgo condoms and higher pay for not using condoms. Similarly, transgender women in relationships also felt that PrEP offered them a way to protect themselves, as they may feel powerless to negotiate condom use or bring up sexual histories when their dating pool is already constrained. Rael et al. (2021) identified self-efficacy in self-injections of PrEP as a source of empowerment; thus, patients being able to inject Apretude themselves rather than having to visit their provider may facilitate PrEP uptake. Transgender women who lacked familial support but hoped for it in the future felt that PrEP offered them a way to survive and remain “healthy” for when their family comes around. Finally, while not a facilitator of PrEP use for transgender women, one article identified transgender women’s use of PrEP as a potential facilitator of uptake by cisgender, heterosexual men who could gain greater awareness of PrEP’s utility and efficacy through their transgender women partners. Process Only seven articles (21.2%) identified process determinants. The most noted (50% of coded “process” references) were regarding engaging patients and customers. Six out of seven articles identified facilitators to increase PrEP uptake by engaging patients through numerous methods. Carter et al. (2022) found that one health department’s use of patient storytelling of experiences of cissexism and racism increased staff empathy and served to potentially reduce medical mistrust. They also found that another health department’s engagement in a workshop, Undoing Racism, enhanced staff motivation to work with patients to overcome barriers to PrEP implementation. Galindo et al. (2012) noted the role of community mobilization strategies in increasing PrEP uptake by engaging with consumers more broadly and working with the community to create buy-in. In large, Carter et al. (2022), Galindo et al. (2012), and other articles that identified similar determinants explicated a need to engage with transgender communities, actively listen to transgender patients, and learn from transgender people. Transgender participants in numerous studies provided feedback on marketing to aid researchers in better reaching transgender communities, and transgender communities may be able to provide feedback and assistance on PrEP implementation in other regards. Rhodes et al. (2020) analyzed the implementation of HOLA, an implementation strategy to increase PrEP uptake and other sexual health practices, among transgender Latinas. HOLA relied on the development of formally appointed implementation leaders and identification of key stakeholders to provide peer navigation assistance to transgender Latinas. Finally, Brooks et al. (2019) noted a need to adapt PrEP implementation for Black and Latina transgender women through interviews with medical providers. They highlighted that providers need to adapt strategies to the unique needs of Black and Latina transgender women and that “supportive services [specific to them] should be embedded within the delivery of PrEP to [Black and Latina transgender women]” (p. 194). Such supportive services specific to Black and Latina transgender women might include tapping into the familial and peer networks of Black and Latina transgender women that, when supportive of their gender identity, have been found to enhance PrEP uptake and levels of resilience and protective factors (Brooks et al. 2019). Lumping Cisgender MSM & Transgender Women and Cisgender Women & Transgender Men Twenty-one articles (63.6%) identified determinants from studies that either analyzed cisgender MSM and transgender women together (often referring to this lumped category of populations as MSMTW) or cisgender women and transgender men together. Ten articles (30.3%) focused solely on transgender people. Three of these (Rael et al. 2018, 2020, 2021) were from the same primary author, and only one (3% of all articles) included transgender men (Zarwell et al., 2021). The majority of articles lumping cisgender MSM and transgender women together did not detail that the MSM included are cisgender. Rather, this was deduced from acknowledgment within study methods that all participants had to be assigned male at birth and were either cisgender MSM or transgender women. Often, studies that included cisgender men and women and transgender men and women combined transgender men and transgender women as “transgender people” due to the small sample size of transgender people compared to the much larger sample size of cisgender people. For example, Wu et al. (2020) included a sample of 412 participants, only 20 of whom were transgender. Results were not broken out between sub-samples, except for one regression comparing “low-level PrEP retention” by gender and sexual orientation. There were not significant findings regarding the transgender sub-sample, though, most likely due to the sample size. Only one study explicitly sought to analyze determinants of PrEP implementation among transgender women and nonbinary people (Klein and Golub 2019). While three articles explicitly attended to determinants for transgender men (Theodore et al., 2020; Westmoreland et al., 2020; Zarwell et al., 2021) one of these three lumped together cisgender women with transgender men, briefly mentioning that transgender men are rarely included in studies of PrEP implementation before continuing to largely focus on cisgender women (Theodore et al., 2020). More quantitative articles lumped cisgender MSM and transgender women than did qualitative articles (thirteen and seven, respectively; see Table 3). This lumping of cisgender MSM and transgender women made it difficult to tease apart the results of each study that pertain to transgender women and whether such studies are representative of transgender women. This is, in part, due to the small sub-samples of transgender women that were ultimately overwhelmed by much larger sub-samples of cisgender MSM. For example, Chan et al. (2019) included two transgender women in a sample of 282 participants, the rest of whom were cisgender MSM. While the study identified numerous barriers to PrEP retention and uptake among cisgender MSM, it did not do so, nor could it, for transgender women. Cohen et al. (2015) included fourteen transgender women in a sample of 1,069 participants. While articles like ones by Colson et al. (2020) found associations between education, depression, and time in follow-up and PrEP adherence—and this seems plausible to be true for transgender women, as well—it cannot be known due to only ten transgender women being included in a study of 204 total participants, all of whom else were cisgender MSM . Table 3 Number of articles lumping cisgender MSM and transgender women and cisgender women and transgender men, as well as number of articles that had too few transgender participants for sub-population analysis by study method of n = 21 articles Axial Code Qualitative Articles Quantitative Articles Mixed-Method Article Lumping cisgender MSM and transgender women 7 13 0 Lumping cisgender women and transgender men 0 0 1 Too Few Transgender Participants to Break Out Results2 2 10 0 1 Less than 10% of the total sample. Havens et al. (2019), one of the few to include transgender men, had a total sample of sixty, which included two transgender men. They noted, though, that the majority, 91.7% (55 out of 60) were men. Another three were cisgender women, and transgender men were not included in their categorization of “men.” Due to the small sample size of transgender people, article titles and abstracts often mentioned transgender women, transgender people, and, sometimes, transgender men, only for the small handful of transgender participants included to be referenced in the literature review and then again at the end of the study as a limitation. Further, Rael et al. noted in their article that, at the time, “We found that with the exception of one study, all research on [transgender women and oral PrEP] grouped transgender women with gay and bisexual men” (2018, p. 2). As seen in Fig. 2, there is an association between the lumping of cisgender MSM and transgender women and a lack of a large enough transgender sub-sample to breakout the results. Thus, including cisgender MSM and transgender women in a single study may not serve to identify determinants of PrEP implementation for transgender women. Finally, more studies lumping cisgender MSM and transgender women and cisgender women and transgender men have been published in the past four years than the earlier four-year period of 2012–2016 (see Table 4). Fig. 1 PRISMA-ScR flowchart for PrEP implementation in transgender populations Table 4 Number of articles lumping cisgender MSM and transgender women and cisgender women and transgender men, as well as number of articles with unsubstantiated findings for transgender sub-samples by year of n = 21 articles Axial Code 2012 2013 2015 2016 2017 2018 2019 2020 2021 2022 Total Lumping cisgender MSM and transgender women 1 1 2 0 3 2 5 4 0 1 20 Lumping cisgender women and transgender men 0 0 0 0 0 0 0 1 0 0 1 Too Few Transgender Participants to Break Out Results1 1 1 2 0 1 0 3 4 0 0 12 1 Less than 10% of the total sample. Discussion While more studies lumping cisgender MSM and transgender women, as well as those lumping cisgender women and transgender men, have been published in recent years, this may in part be due to an increase in attention to transgender health within the field of HIV research in the U.S. (Siskind et al., 2016; Wansom, Guadamuz, & Vasan, 2016). In part, the increase in trans-focused studies after 2018 may have been influenced by coinciding institutional factors. The National Institutes of Health (NIH) officially formed the Sexual & Gender Minority Research Office in 2015, the CDC expanded their National HIV Behavioral Surveillance System to focus on transgender women as a distinct population for the first time in 2019, and NIH released its first R01 focused on sex and gender with particular attention to transgender populations in the Fall of 2020. Though insufficient, the growth in this area of research has provided formative knowledge regarding determinants of PrEP implementation for transgender individuals. The findings point to barriers to overcome for expanding PrEP implementation in transgender communities and facilitators that can be utilized to do so. Such barriers include a need to actively and intentionally expand awareness of PrEP availability, side effects, safety to take with HRT, and cost assistance programs. Additionally, whatever form education and awareness-building take, transgender individuals should be engaged in the development of these. Further, marketing of PrEP at large needs to better engage transgender individuals in the design process. In doing so, PrEP marketing can better target the needs and experiences of transgender people. Additionally, scholars highlight structural factors impeding transgender individuals’ access to PrEP uptake and, especially, adherence. These include engagement in sex work, histories of incarceration, access to gender affirming health care and health insurance, educational attainment, and systemic racism. Structural barriers will require structural strategies. System-level barriers include the need for providers to increase their awareness and knowledge of transgender experiences and transgender health needs, as well as for non-HIV-specific primary care providers and pharmacists to obtain training and resources to prescribe PrEP and mitigate transportation barriers, such as reducing the number of required appointments. Gaps in the literature remain, though, preventing scholars and practitioners from a clear understanding of how best to implement PrEP in transgender communities. Few articles attended to process determinants. Our research team similarly found few articles attending to process determinants in the initial systematic review that provided the basis for this manuscript (Li et al., 2022). This is most likely due to the overlap between implementation strategies and process determinants. Implementation strategies are “methods or techniques used to enhance the adoption, implementation, and sustainability of a clinical program or practice” (Proctor et al., 2013). Process determinants include engaging customers, clients, and patients, executing interventions and implementation strategies, planning (e.g., building local capacity, tailoring interventions to localized communities), and reflecting and evaluating. Each of these are actions undertaken to overcome barriers or utilize facilitators to implementing an innovation. Thus, there may be literature on implementation strategies for PrEP with transgender communities that contain process determinants that were not included within our review. As such, our research team is currently conducting a systematic review of PrEP implementation strategies. Other scholars, as well, should attend to such strategies for barriers and facilitators we have identified in this manuscript. An example includes the facilitative role of peer navigators. Community role models or peer navigators with shared lived experiences can help increase knowledge about PrEP and model its use by individuals having various rates and kinds of sex (Klein and Golub, 2019; McMahan et al., 2020; Waltz et al., 2019) developed a tool matching CFIR determinants with implementation strategies from the Expert Recommendations for Implementation Change compilation. Using this tool can be helpful in matching barriers with the strategies to overcome them. For example, the need to adapt behavioral interventions for increasing PrEP uptake and adherence can be facilitated through capturing and sharing local knowledge (Waltz et al., 2019). Integrating gender affirmation surgery and PrEP may also serve to increase uptake by overcoming transportation burdens and time constraints, as well as integrating PrEP care into health care that is outwardly inclusive and aware of transgender individuals’ health needs (Sevelius, Deutsch, et al., 2016). Developing ongoing training and education and hiring transgender women can also help to cultivate organizations, clinics, and healthcare facilities with climates that are open and affirmative of transgender individuals and thus increase sustained engagement with transgender patients (Auerbach et al., 2021). Connecting already trialed implementation strategies, such as these with specified determinants, may facilitate PrEP uptake and adherence. Domestic implementation science, transgender health, and applied transgender studies are each nascent, albeit burgeoning, fields. As such, gaps identified in this manuscript identify key opportunities for future PrEP implementation research with these key populations. There are currently numerous studies identifying individual-level determinants for transgender women (e.g., self-efficacy), community and society-level—or outer setting—determinants (e.g., cissexism, racism, lack of access to gender-affirming healthcare), and innovation-level determinants (e.g., saturation of marketing with cisgender MSM and a lack of marketing specifically to transgender people). We identify four areas of needed growth: expanding research on determinants for transgender men and nonbinary people, analyzing inner setting determinants, rethinking the continued lumping of transgender women and cisgender MSM, and better identifying facilitators of implementation. Finally, we identify what implementation science may offer the field of transgender health writ large. First, much of the literature on determinants of PrEP implementation for transgender people focuses on transgender women. In our search, we identified only four articles that explicitly included transgender men in their study samples and only one explicitly including nonbinary people. Two studies including transgender men (Theodore et al., 2020; Westmoreland et al., 2020) had too small of sub-samples (3.9% and 0.3% of the total sample respectively) to analyze transgender men’s barriers and facilitators to PrEP implementation. While Descovy is not yet indicated for use by transgender men and nonbinary people assigned female at birth (AFAB), Truvada and Apretude are. And while HIV prevalence rates are not fully known for transgender MSM and nonbinary people, transgender MSM and nonbinary people have been found to engage in high-risk sexual behavior (Herbst et al., 2008; Reisner et al., 2016). Thus, as HIV implementation science continues to grow and develop, researchers must ensure that they expand their focus beyond focusing primarily on transgender women. If future effectiveness trials are conducted with Descovy and other forms of PrEP for transgender men and nonbinary people, we recommend the use of hybrid approaches involving the collection of data on barriers and facilitators to PrEP within these populations (Curran et al. 2013). Additionally, it may be useful to examine determinants of implementation at the partner-level. For example, Gamarel and Golub (2015) found that partner-level determinants influenced PrEP adoption for cisgender MSM, with participants who reported condomless anal sex outside their primary relationship reporting higher likelihood to initiate PrEP use. As noted earlier, transgender women have reported difficulties negotiating PrEP use with cisgender, heterosexual men partners due to stigma; fear; and lack of awareness among cisgender, heterosexual men. Thus, it may be important to analyze determinants of implementation among cisgender, heterosexual men and to match those determinants with strategies to overcome barriers at the individual- and partner-levels. Second, implementation researchers should undertake more studies analyzing inner setting barriers and facilitators of PrEP implementation for transgender people. For example, our scoping review details that gender-affirming health care may be an ideal location to increase PrEP uptake. Thus, studies like those examining inner setting determinants for implementing PrEP within pharmacies and primary care settings (Hopkins et al., 2021; Zhao et al., 2021) are needed for gender-affirming health care. Do gender-affirming care providers have the education and awareness to prescribe PrEP and maintain patients in care? Are there policy, protocol, or guideline factors to consider in the outer setting to undertake such implementation, particularly in the wake of increasing attacks on transgender health care across the US (Conron et al., 2022)? Do gender-affirming care environments have the capacity to implement PrEP? These and more questions are worth examining. Similar analyses can be undertaken examining inner setting determinants for implementing PrEP for transgender women in women-centered care, as well as within trans-specific and trans-led nonprofit organizations. Third, implementation science (and public health at large) needs to analyze determinants for implementing PrEP for transgender women separate from studies on cisgender MSM. The historical categorization of transgender women as MSM has shifted to current studies analyzing cisgender MSM and transgender women. However, differences in HIV rates and acquisition between cisgender MSM and transgender women (Poteat, German, & Flynn, 2016) and the continued lumping of cisgender MSM and transgender women has made it difficult to assess and review literature on PrEP implementation for our review, as well as others, such as Baldwin, Light, and Allison’s (2020) narrative review of literature on PrEP for cisgender and transgender women. While there are additional factors to consider in making such decisions regarding the lumping of transgender women and cisgender MSM (e.g., funding), this recurring practice is not contributing to the literature for transgender women as much as it may appear from a glance at titles and abstracts in a PubMed search. An alternative approach would be to include transgender MSM with cisgender MSM and transgender women with cisgender women. Such studies would categorize by gender rather than sex assigned at birth and would better represent gendered dynamics of individuals’ social ecologies that shape their health behaviors and experiences. Examples of such research can be found in some implementation strategies and interventions trials, including Walters et al.’s (2021) implementation of a PrEP intervention for cisgender and transgender women and Auerbach et al.’s (2021) interviews with cisgender and transgender women on developing HIV care for all women. It must be noted, though, that limited sub-samples of transgender individuals within samples lumping cisgender MSM and transgender women is not unique to implementation science. Indeed, a review of the inclusion of transgender women in PrEP trials found that transgender women comprised just 0.2% of total trial enrollments (Escudero et al., 2015). In order to attend to transgender populations’ needs separate from cisgender MSM and cisgender women, researchers can identify strategies used by other researchers in transgender health to obtain larger sample sizes. For example, Kronk et al. (2022) detail a two-step gender and sex identification system within electronic health records that could be utilized for data reporting and analysis. Other scholars have utilized internet samples to reach “hard to reach” sexual and gender minority populations for studies of cannabis-use comorbidities (Gonzalez, Gallego, & Bockting, 2017), mental health and resilience (Bockting, Miner, Romine, Hamilton, & Coleman, 2013), and other analyses (Mathy, Schillace, Coleman, & Berquist, 2002). Researchers can utilize alternative measures for sex and gender in survey research to better capture cisgender, transgender, and nonbinary identification (Puckett, Brown, Dunn, Mustanski, & Newcomb, 2020; Saperstein & Westbrook, 2021). It is also possible that fostering an increased presence of Black and/or Latinx transgender researchers in implementation science as faculty and lead researchers could lead to shifts in how research is undertaken and planned in formative phases (Everhart et al., 2022). Further, creating equitable partnerships with community members by engaging in community-based participatory research can mitigate missteps on the part of researchers (Israel et al., 1998). Finally, it is important that scholars begin to mark analyses of cisgender MSM as such (Brekhus, 1998). In our review, few studies explicitly stated that their analyses were of cisgender MSM despite the MSM included all being AMAB. Implementation scientists and researchers need to additionally utilize and continue developing and adapting implementation frameworks centering health equity (Baumann and Cabassa, 2020; Brownson et al., 2021). Implementation scientists and researchers develop strategies to mitigate barriers to implementing innovations in everyday life; however, if health equity is not central to such work, including explicit attention to issues of systemic racism, capitalism, cissexism, and ableism, among other axes of oppression, then huge opportunities to increase access to health care and health services will be missed. Institutionalizing health equity within implementation science requires that researchers adequately study and acquire the knowledge and skills necessary to engage in such work (e.g. developing community-engaged recruitment and research skills, engaging in conversation through citation with scholars who have developed a foundation for health equity research, and committing to ongoing racial justice praxis) (Lett et al., 2022). To do so, implementation scientists and researchers can utilize the framework of Critical Race Public Health Praxis, which builds on Critical Race Theory to better capture race as a structural (rather than individual or “biological”) factor and requires that researchers address racism in research context, conceptualization, and measurement and knowledge production (Ford and Airhihenbuwa, 2010). Finally, scholars can look to existing adaptations of implementation frameworks, including Allen et al.’s (2021) racism-conscious adaptation of CFIR and Woodward et al.’s (2019) development of a health equity implementation framework, which was constructed from integration and adaptation of the integrated-promoting action on research implementation in health services (i-PARIHS) framework (Harvey and Kitson, 2016) and the health care disparities framework (Kilbourne et al., 2006). As Shelton et al. (2021) note, “To be a part of the solution in helping to achieve racial/ethnic justice, [implementation science] needs to ground the field in extant scholarship on health equity and racism, and reframe a foundational focus on social justice, equity, and real-world impact.” While many barriers to implementation (e.g., stigma, SDOH, medical mistrust) are well identified, fewer facilitators of implementation are identified and assessed. Here, implementation researchers could look to the work of transnational implementation science, as well as intersectional HIV research in the US. Wilson et al. (2019) examined barriers and facilitators to PrEP among 339 Brazilian transgender women. They highlighted how transgender women’s engagement in sex work served as a source of “PrEP empowerment,” a potential facilitator of implementation rather than simply a barrier as it almost entirely was in our review. Additionally, researchers can expand work on facilitative factors of PrEP implementation by expanding attention to transgender individuals who currently are on PrEP, with particular attention to determinants of sustained use at a prevention effective level. While not a PrEP- or trans-focused study, Dācus, Voisin, and Barker (2017) performed interviews with HIV-negative, Black, cisgender MSM to better understand how they stayed HIV-negative. They identified individual, social, and community-level resiliency characteristics that serve as facilitators. Xavier Hall et al. (2022) qualitatively analyzed daily diaries of YMSM who were PrEP adherent to understand what facilitated their adherence. They identified strategies for adherence, including checking in with friends, having pill counters, and counseling that adapts strategies to individuals prior to initiating PrEP. Johns et al. (2018; 2021) similarly identified individual-, social-, and community-level protective factors for transgender youths’ overall health and wellbeing, including parental and familial support, having romantic and sexual relationships, transgender visibility, and self-advocacy. Implementation researchers could similarly examine whether such attributes and characteristics facilitate PrEP implementation for transgender people. Finally, transgender health can gain from implementation science. Transgender people, particularly Black transgender individuals, experience barriers to most, if not all, types of care across the US. Use of CFIR can guide evaluations of barriers and facilitators to implementing new centers of care, as scholars have recently done for the creation of a rural transgender health center (Tinc, Wolf-Gould, Wolf-Gould, & Gadomski, 2020). As clinical advances are made in the field of transgender health, hybrid study designs trialing effectiveness and implementation can be utilized to ensure advances are put into practice with the greatest reach from the start. Implementation strategies can be developed to enhance health equity for transgender populations vis-à-vis COVID-19 prevention (Teixeira da Silva et al., 2021), eating disorder treatment (Duffy, Henkel, & Earnshaw, 2016; Rosenvinge & Pettersen, 2015), and even employment interventions to address structural health inequities (Thompson et al. 2022). Further, it is possible that leveraging implementation science in transgender health and vice versa could provide benefits to both fields and transgender populations at large. Limitations Our review has some limitations that should be considered. First, studies included in our review largely focused on transgender women with only four including transgender men and one including nonbinary individuals. Second, we did not differentiate determinant coding by stages of the PrEP cascade, and it will be useful for researchers to do so in the future to assess barriers and facilitators that may require more attention in awareness compared to uptake and/or adherence. Third, our review does not include studies based outside the US. There is a wealth of non-domestic implementation research that attends to the barriers and facilitators of PrEP implementation for transgender populations, and it may be beneficial for future researchers to examine what non-US implementation research on PrEP for transgender individuals has to offer domestic researchers and implementers. Fourth, there were few articles coded as identifying determinants within the process domain of CFIR; process constructs are similar to implementation strategies, which would not have been captured in the initial systematic review identification and selection process focused on determinants. Fifth, our review only consisted of published scholarship in peer-reviewed journals. We did not seek to include the grey literature of unpublished studies. Finally, only the first author (az) performed coding and substantive analysis for this manuscript. We did not assess coding reliability; however, az did routinely assess similarities and differences between her coding and coding in the initial systematic review as a form of reliability checking. Conclusion Utilizing CFIR to identify barriers and facilitators to implementing PrEP, we found numerous barriers identified regarding individual characteristics for transgender women; innovation characteristics like cost, marketing, and evidence quality; and outer setting factors, including structural cissexism/racism and other SDOHs. We recommend that implementation researchers expand analyses on inner setting determinants of barriers and facilitators for implementing PrEP for transgender populations in general and specialty health care settings, individual-level barriers and facilitators at the deliverer level, and individual-level barriers and facilitators at the recipient level for transgender men and nonbinary people. Our review also identifies a need for researchers to increase sample sizes of transgender women in multi-population analyses, as well as to attend to the unique barriers, facilitators, and needs of transgender individuals alone (i.e., separate from cisgender MSM and cisgender women). The next step for researchers is to match trans-specific determinants of PrEP implementation with strategies to overcome barriers, enhance facilitators of implementation, and mitigate health inequities. Fig. 2 Number of articles lumping cisgender MSM and transgender women and cisgender women and transgender men, as well as number of articles with unsubstantiated findings for transgender sub-samples by year of n = 33 articles. (1TW refers to transgender women, CW refers to cisgender women, and TM refers to transgender men. 2 Less than 10% of the total sample.) Authors’ Contributions The first author conducted the identification of articles for the scoping review, coded the data, analyzed, and drafted the manuscript. The second author provided leadership in the initial systematic review from which this scoping review was developed. The second, third, fourth, and fifth authors contributed to writing and editing of the manuscript. Funding This work was supported by a supplement grant to the Third Coast Center for AIDS Research, an NIH-funded center (P30 AI117943; PIs: Mustanski & D’Aquila; Supplement PIs: Mustanski & Benbow), and a training grant from the National Institute of Mental Health (T32MH130325; PI: Newcomb). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The sponsors had no involvement in the conduct of the research or the preparation of the article. Data Availability Individuals interested in obtaining data from this review can contact the corresponding author. A list of all articles included within the analysis is also included at the end of the manuscript. Code Availability Not applicable. Declarations Conflict of interest The authors declare that they have no conflicts of interest. Ethics Approval Not applicable. Consent to Participate Not applicable. Consent for publication Not applicable. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References Allen M, Wilhelm A, Ortega LE, Pergament S, Bates N, Cunningham B. (2021). Applying a race(ism)conscious adaptation of the CFIR framework to understand implementation of a school-based equity-oriented intervention. Ethn Dis, 31. doi:10.18865/ed.31.S1.375. Auerbach JD Moran L Weber S Watson C Keatley J Sevelius J Implementation strategies for creating inclusive, all-women HIV care environments: perspectives from trans and cis women Women’s Health Issues 2021 31 4 332 40 10.1016/j.whi.2021.04.001 33941451 Baldwin A Light B Allison WE Pre-exposure prophylaxis (PrEP) for HIV infection in cisgender and transgender women in the US: a narrative review of the literature Arch Sex Behav 2021 50 4 1713 28 10.1007/s10508-020-01903-8 34075504 Bauer MS Damschroder L Hagedorn H Smith J Kilbourne AM An introduction to implementation science for the non-specialist BMC Psychol 2015 3 1 1 12 10.1186/s40359-015-0089-9 25678981 Baumann AA, Cabassa LJ. (2020). Reframing implementation science to address inequities in healthcare delivery. Implement Sci, 20(190). doi:10.1186/s12913-020-4975-3. Bockting WO Miner MH Romine RES Hamilton A Coleman E Stigma, mental health, and resilience in an online sample of the US transgender population Am J Public Health Nations Health 2013 103 5 943 51 10.2105/ajph.2013.301241 Bowleg LJ The problem with the phrase women and minorities: Intersectionality—an important theoretical framework for public health Am J Public Health 2012 102 7 1267 73 10.2105/AJPH.2012.300750 22594719 Brekhus W A sociology of the unmarked: redirecting our focus Sociol Theory 1998 16 1 34 51 10.1111/0735-2751.00041 Brooks RA, Cabral A, Nieto O, Fehrenbacher A, Landrian A. (2019). Experiences of pre-exposure prophylaxis stigma, social support, and information dissemination among Black and Latina transgender women who are using pre-exposure prophylaxis. Transgender Health, 4. doi:10.1089/trgh.2019.0014. Brownson RC, Kumanyika SK, Kreuter MW, Haire-Joshu D. (2021). Implementation science should give higher priority to health equity. Implement Sci, 16(28). doi:10.1186/s13012-021-01097-0. Carter JW Jr, Salabarría-Pena Y, Fields EL, Robinson WT. (2022). Evaluating for health equity among a cluster of health departments implementing PrEP services. Eval Program Plan, 90. doi:10.1016/j.evalprogplan.2021.101981. Centers for Disease Control and Prevention. (2020). PrEP for HIV prevention in the US. Retrieved from https://www.cdc.gov/nchhstp/newsroom/fact-sheets/hiv/PrEP-for-hiv-prevention-in-the-US-factsheet.html#:~:text=Preliminary%20CDC%20data%20show%20only%20about%2016%25%20(38%2C454)%20of,among%20people%20based%20on%20sex. Centers for Disease Control and Prevention. (2021). HIV infection, risk, prevention, and testing behaviors among transgender women–national HIV behavioral surveillance, 7 U.S. cities, 2019–2020. Retrieved from https://www.cdc.gov/hiv/pdf/library/reports/surveillance/cdc-hiv-surveillance-special-report-number-27.pdf. Chan PA, Patel RR, Mena L, Marshall BD, Rose J, Coats S, Nunn C, A. (2019). Long-term retention in pre-exposure prophylaxis care among men who have sex with men and transgender women in the United States. Journal of the International AIDS Society, 22(e25385). doi:http://onlinelibrary.wiley.com/doi/10.1002/jia | 10.1002/jia2.25385. Cohen SE, Vittinghoff E, Bacon O, Doblecki-Lewis S, Postle BS, Feaster DJ,. . Liu AY. (2015). High interest in pre-exposure prophylaxis among men who have sex with men at risk for HIV-infection: baseline data from the US PrEP demonstration project. J Acquir immune Defic syndrome, 68(4). doi:10.1097/QAI.0000000000000479. Colson PW, Franks J, Wu Y, Winterhalter FS, Knox J, Ortega H,. . Hirsch-Moverman Y. (2020). Adherence to pre–exposure prophylaxis in black men who have sex with men and transgender women in a community setting in Harlem, NY. AIDS Behav, 24. doi:10.1007/s10461-020-02901-6. Conron KJ, O’Neill KK, Vasquez LA, Mallory C. (2022). Prohibiting gender-affirming medical care for youth. Retrieved from https://williamsinstitute.law.ucla.edu/publications/bans-trans-youth-health-care/. Crenshaw K. Demarginalizing the intersection of race and sex: a black feminist critique of antidiscrimination doctrine, feminist theory, and antiracist politics [1989]. In: Feminist legal theory. Routledge; 2018. pp. 57–80. Curran GM, Bauer M, Mittman B, Pyne JM, Stetler C. (2013). Effectiveness-implementation hybrid designs. Med Care, 50(3). doi:10.1097/MLR.0b013e3182408812. Dācus J-d Voisin DR Barker J “Proud i am negative”: maintaining HIV-seronegativity among Black MSM in New York City Men and Masculinities 2018 21 2 276 90 10.1177/1097184x17696174 Damschroder LJ Aron DC Keith RE Kirsh SR Alexander JA Lowery JC Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science Implement Sci 2009 4 1 1 15 10.1186/1748-5908-4-50 19123945 Duffy ME Henkel KE Earnshaw VA Transgender clients’ experiences of eating disorder treatment J LGBT Issues Couns 2016 10 3 136 49 10.1080/15538605.2016.1177806 35356098 Eaton LA, Kalichman SC, Price D, Finneran S, Allen A, Maksut J. (2017). Stigma and conspiracy beliefs related to pre-exposure prophylaxis (PrEP) and interest in using PrEP among black and white men and transgender women who have sex with men. AIDS Behav, 21(5). doi:10.1007/s10461-017-1690-0. Escudero DJ Kerr T Operario D Socías ME Sued O Marshall BD Inclusion of trans women in pre-exposure prophylaxis trials: a review J AIDS Care 2015 27 5 637 41 10.1080/09540121.2014.986051 Everhart AR, Boska H, Sinai-Glazer H, Wilson-Yang JQ, Butler Burke N, LeBlanc G,. . Marshall Z. (2022). ‘I’m not interested in research; i’m interested in services’: how to better health and social services for transgender women living with and affected by HIV. Soc Sci Med, 292. doi:10.1016/j.socscimed.2021.114610. Ford CL, Airhihenbuwa CO. (2010). The public health critical race methodology: Praxis for antiracism research. Soc Sci Med, 71(8). doi:10.1016/j.socscimed.2010.07.030. Galindo GR, Walker JJ, Hazelton P, Lane T, Steward WT, Morin SF, Arnold EA. (2012). Community member perspectives from transgender women and men who have sex with men on pre-exposure prophylaxis as an HIV prevention strategy: implications for implementation. Implementation science, 7. doi:http://www.implementationscience.com/content/7/1/116. Gamarel KE Golub SA Intimacy motivations and pre-exposure prophylaxis (PrEP) adoption intentions among HIV-negative men who have sex with men (MSM) in romantic relationships Ann Behav Med 2015 49 2 177 86 10.1007/s12160-014-9646-3 25124457 Golub SA, Gamarel KE, Rendina HJ, Surace A, Lelutiu-Weinberger CL. (2013). From efficacy to effectiveness: facilitators and barriers to PrEP acceptability and motivations for adherence among MSM and transgender women in New York City. AIDS Patient Care STDs, 27(4). doi:10.1089/apc.2012.0419. Gonzalez CA Gallego JD Bockting WO Demographic characteristics, components of sexuality and gender, and minority stress and their associations to excessive alcohol, cannabis, and illicit (noncannabis) drug use among a large sample of transgender people in the United States J Prim Prev 2017 38 4 419 45 10.1007/s10935-017-0469-4 28405831 Grant RM Pellegrini M Defechereux PA Anderson PL Yu M Glidden DV Deutsch MB Sex hormone therapy and Tenofovir Diphosphate concentration in dried blood spots: primary results of the interactions between antiretrovirals and transgender hormones study J Clin Infect Dis 2020 73 7 e2117 23 10.1093/cid/ciaa1160 Greenhalgh T, Robert G, Macfarlane F, Bate P, Kyriakidou O. (2004). Diffusion of innovations in service organizations: systematic review and recommendations. Milbank Quarterly, 82. Harvey G, Kitson A. (2016). PARIHS revisited: from heuristic to integrated framework for the successful implementation of knowledge into practice. Implement Sci, 11(33). doi:10.1186/s13012-016-0398-2. Havens JP Scarsi KK Sayles H Kelpser DG Swindells S BAres SH Acceptability and feasibility of pharmacist-led human immunodeficiency virus pre-exposure prophylaxis program in the midwestern United States Open Forum Infectious Diseases 2019 10.1093/ofid/ofz365 Herbst JH Jacobs ED Finlayson TJ McKleroy VS Neumann MS Crepaz N Estimating HIV prevalence and risk behaviors of transgender persons in the United States: a systematic review J AIDS Behav 2008 12 1 1 17 10.1007/s10461-007-9299-3 Hoenigl M Hassan A Moore DJ Anderson PL Corado K Dubé MP 2019 California Collaborative Treatment Group 10.1097/QAI.0000000000002003 Hopkins R Josma D Morris J Klepser DG Young HN Crawford ND Support and perceived barriers to implementing pre-exposure prophylaxis screening and dispensing in pharmacies: examining concordance between pharmacy technicians and pharmacists J Am Pharm Association (American Pharm Association) 2021 61 1 115 20 Israel BA, Schulz AJ, Parker EA, Becker AB. (1998). Review of community-based research: assessing partnership approaches to improve public health. Annu Rev Public Health, 19(1). doi:10.1146/annurev.publhealth.19.1.173. Johns M Beltran O Armstrong HL Jayne PE Barrios LC Protective factors among transgender and gender variant youth: a systematic review by socioecological level J Prev Res 2018 39 3 263 301 Johns M, Andrzejewski zamantakis,a, Boyce J, Rasberry L, C. N., & Jayne PE. (2021). Minority stress, coping, and transgender youth in schools—results from the resilience and transgender youth study. J Sch Health, 91(11). doi:10.1111/josh.13086. Jourdain H de Gage SB Desplas D Dray-Spira R Real-world effectiveness of pre-exposure prophylaxis in men at high risk of HIV infection in France: a nested case-control study The Lancet Public Health 2022 7 6 e529 36 10.1016/S2468-2667(22)00106-2 35660214 Kamitani E Johnson WD Wichser ME Adegbite AH Mullins MM Growth in proportion and disparities of HIV PrEP use among key populations identified in the United States national goals: systematic review & meta-analysis of published surveys J Acquir immune Defic syndrome 2020 84 4 379 86 10.1097/QAI.0000000000002345 Kilbourne AM, Switzer G, Hyman K, Crowley-Matoka M, Fine MJ. (2006). Advancing health disparities research within the health care system: a conceptual framework. Am J Public Health, 96(12). doi:10.2105/AJPH.2005.077628. Klein A, Golub SA. (2019). Increasing access to pre-exposure prophylaxis among transgender women and transfeminine nonbinary individuals. AIDS Patient Care STDs, 33(6). doi:10.1089/apc.2019.0049. Kronk CA, Everhart AR, Ashley F, Thompson HM, Schall TE, Goetz TG, … Karnoski R. (2022). Transgender data collection in the electronic health record: Current concepts and issues. Journal of the American Medical Informatics Association, 29(2), 271–284. Kuehn BJJ PrEP disparities JAMA Netw 2018 320 22 2304 4 10.1001/jama.2018.18947 Lalley-Chareczko L, Clark D, Conyngham C, Zuppa A, Moorthy G, Mounzer K, Koenig H. (2018). Delivery of TDF/FTC for pre-exposure prophylaxis to prevent HIV-1 acquisition in young adult men who have sex with men and transgender women of color using a urine adherence assay. Journal of acquired immune deficiency syndrome, 79. Landovitz RJ, Beymer M, Kofron R, Amico KR, Psaros C, Bushman L,. . Jordan WC. (2017). Plasma tenofovir-levels to support adherence to TDF/FTC preexposure prophylaxis for HIV prevention in MSM in Los Angeles, California. J Acquir immune Defic syndrome, 76(5). doi:10.1097/QAI.0000000000001538. Landovitz RJ Donnell D Clement ME Hanscom B Cottle L Coelho L Frank I Cabotegravir for HIV prevention in cisgender men and transgender women N Engl J Med 2021 385 7 595 608 10.1056/NEJMoa2101016 34379922 Lett E, Adekunle D, McMurray P, Asabor EN, Irie W, Simon MA,. . McLemore MR. (2022). Health equity tourism: ravaging the justice landscape. Health Policy, 46(17). doi:10.1007/s10916-022-01803-5. Li DH, Benbow N, Keiser B, Mongrella M, Ortiz K, Villamar J,. . Smith JD. (2022). Determinants of implementation for HIV pre-exposure prophylaxis based on an updated Consolidated Framework for implementation research: a systematic review. J Acquir immune Defic syndrome, 90. doi:10.1097/QAI.0000000000002984. Liu AY, Cohen SE, Vittinghoff E, Anderson PL, Doblecki-Lewis S, Bacon O,. . Kolber MA. (2016). Pre-exposure prophylaxis for HIV infection integrated with municipal- and community-based sexual health services. JAMA Intern Med, 176(1). doi:10.1001/jamainternmed.2015.4683. Mathy RM Schillace M Coleman SM Berquist BE Methodological rigor with internet samples: new ways to reach underrepresented populations CyberPsychology & Behavior 2002 5 3 253 66 10.1089/109493102760147259 12123248 Mayer KH Molina J-M Thompson MA Anderson PL Mounzer KC De Wet JJ Ruane P Emtricitabine and tenofovir alafenamide vs emtricitabine and tenofovir disoproxil fumarate for HIV pre-exposure prophylaxis (DISCOVER): primary results from a randomised, double-blind, multicentre, active-controlled, phase 3, non-inferiority trial The Lancet 2020 396 10246 239 54 10.1016/S0140-6736(20)31065-5 McMahan VM, Frank N, Buckler S, Violette LR, Baeten JM, Banta-Green CJ,. . Stekler JD. (2020). Protocol development for HMU! (HIV prevention for methamphetamine users), a study of peer navigation and text messaging to promote pre-exposure prophylaxis adherence and persistence among people who use methamphetamine: Qualitative focus group and interview study. JMIR Formative Research, 4(9). doi:10.2196/18118. Mehrota ML, Amico KRM, Glidden V, Defechereux DV, Guanira P, J. V., & Grant RM. (2018). The role of social relationships in PrEP uptake and use among transgender women and men who have sex with men. AIDS Behav, 22. doi:10.1007/s10461-018-2151-0. Morris ZS Wooding S Grant J The answer is 17 years, what is the question: understanding time lags in translational research J R Soc Med 2011 104 12 510 20 10.1258/jrsm.2011.110180 22179294 Poteat T German D Flynn C The conflation of gender and sex: gaps and opportunities in HIV data among transgender women and MSM J Global Public Health 2016 11 7–8 835 48 10.1080/17441692.2015.1134615 Poteat T Radix A HIV antiretroviral treatment and pre-exposure prophylaxis in transgender individuals Drugs 2020 80 10 965 72 10.1007/s40265-020-01313-z 32451925 Proctor EK, Powell BJ, McMillen JC. (2013). Implementation strategies: recommendations for specifying and reporting. Implementation science, 8. Puckett JA Brown NC Dunn T Mustanski B Newcomb ME Perspectives from transgender and gender diverse people on how to ask about gender LGBT Health 2020 7 6 305 11 10.1089/lgbt.2019.0295 32598211 Rael CT, Lopez–Ríos J, McKenna SA, Das D, Dolezal C, Abascal E,. . Bockting WO. (2021). Transgender women’s barriers, facilitators, and preferences on tailored injection delivery strategies to administer long–acting injectable cabotegravir (CAB–LA) for HIV pre–exposure prophylaxis (PrEP). AIDS Behav, 25. doi:10.1007/s10461-021-03357-y. Rael CT, Martinez M, Giguere R, Bockting WO, MacCrate C, Mellman W,. . Carballo-Diéguez A. (2018). Barriers and facilitators to oral PrEP use among transgender women in New York City. AIDS Behav, 22(11). doi:10.1007/s10461-018-2102-9. Rael CT, Martinez M, Giguere R, Bockting WO, MacCrate C, Mellman W,. . Hope TJ. (2020). Transgender women’s concerns and preferences on potential future long–acting biomedical HIV prevention strategies: the case of injections and implanted medication delivery devices (IMDDs). AIDS Behav, 42. doi:10.1007/s10461-019-02703-5. Reback CJ, Clark KA, Rünger D, Fehrenbacher AE. (2019). A promising PrEP navigation intervention for transgender women and men who have sex with men experiencing multiple syndemic health disparities. J Community Health, 44. doi:10.1007/s10900-019-00705-x. Reisner SL Hughto JMW Pardee DJ Kuhns L Garofalo R Mimiaga MJ LifeSkills for men (LS4M): pilot evaluation of a gender-affirmative HIV and STI prevention intervention for young adult transgender men who have sex with men J Urb Health 2016 93 1 189 205 10.1007/s11524-015-0011-z Reisner SL Moore CS Asquith A Pardee DJ Mayer KH The pre-exposure prophylaxis cascade in at-risk transgender men who have sex with men in the United States LGBT Health 2021 8 2 116 24 10.1089/lgbt.2020.0232 33567245 Restar AJ, Kuhns L, Reisner SL, Ogunbajo A, Garofalo R, Mimiaga MJ. (2018). Acceptability of antiretroviral pre-exposure prophylaxis from a cohort of sexually experienced young transgender women in two U.S. cities. 22(11). doi:10.1007/s10461-018-2127-0. Rhodes DD, Alonzo J, Mann-Jackson L, Song EY, Tanner AE, Garcia M,. . Reboussin BA. (2020). A peer navigation intervention to prevent HIVamong mixed immigrant status Latinx GBMSM and transgender women in the United States: outcomes, perspectives and implications for PrEP uptake. Health Educ Res, 35(3). doi:10.1093/her/cyaa010. Rosenvinge JH Pettersen G Epidemiology of eating disorders part III: Social epidemiology and case definitions revisited Adv Eat Disorders 2015 3 3 320 36 10.1080/21662630.2015.1022197 Saperstein A Westbrook L Categorical and gradational: alternative survey measures of sex and gender Eur J Politics Gend 2021 4 1 11 30 10.1332/251510820X15995647280686 Sevelius JM, Deutsch MB, Grant R. (2016). The future of PrEP among transgender women: the critical role of gender affirmation in research and clinical practices. J Int AIDS Soc, 19(7). doi:10.7448/IAS.19.7.21105. Sevelius JM Keatley J Calma N Arnold E ‘I am not a man’: Trans-specific barriers and facilitators to PrEP acceptability among transgender women Glob Public Health 2016 11 7 8 10.1080/17441692.2016.1154085 Shelton RC Adsul P Oh A Recommendations for addressing structural racism in implementation science: a call to the field Ethn Dis 2021 31 Suppl 1 357 10.18865/ed.31.S1.357 34045837 Shelton RC Adsul P Oh A Moise N Griffith DM Application of an antiracism lens in the field of implementation science (IS): recommendations for reframing implementation research with a focus on justice and racial equity Implement Res Pract 2021 2 26334895211049482 Siskind RL Andrasik M Karuna ST Broder GB Collins C Liu A Renzullo PO Engaging transgender people in NIH-funded HIV/AIDS clinical trials research J Acquir immune Defic syndrome 2016 72 Suppl 3 243 10.1097/QAI.0000000000001085 Spinelli MA, Buchbinder SP. (2019). Pre-exposure prophylaxis persistence Is a critical issue in PrEP implementation. Clinical Infectious Diseases, 71(3), 583–585. Teixeira da Silva D Biello K Lin WY Valente PK Mayer KH Hightow-Weidman L Bauermeister JA COVID-19 vaccine acceptance among an online sample of sexual and gender minority men and transgender women Vaccines 2021 9 3 204 10.3390/vaccines9030204 33804530 Theodore DA, Zucker J, Carnevale C, Grant W, Adan M, Borsa A. . Sobieszczyk ME. (2020). Pre-exposure prophylaxis use among predominantly african american and hispanic women at risk for HIV acquisition in New York City. J Assoc Nurses AIDS Care, 31(1). doi:10.1097/JNC.0000000000000147. Thompson HM, Clement AM, Ortiz R, Preston TM, Quantrell ALW, Reback C,. . Milburn N. (2022). Community engagement to improve access to healthcare: a comparative case study to advance implementation science for transgender health equity. Tinc PJ Wolf-Gould C Wolf-Gould C Gadomski A Longitudinal use of the consolidated framework for implementation research to evaluate the creation of a rural center of excellence in transgender health Int J Environ Res Public Health 2020 17 23 9047 10.3390/ijerph17239047 33291621 US Department of Health and Human Services. (2022). What is Ending the HIV Epidemic in the U.S.? About Ending the HIV Epidemic in the U.S Retrieved from https://www.hiv.gov/federal-response/ending-the-hiv-epidemic/overview. Walters SM Platt J Anakaraonye A Golub SA Cunningham CO Norton BL Blackstock OJ Considerations for the design of pre-exposure prophylaxis (PrEP) interventions for women: Lessons learned from the implementation of a novel PrEP intervention J AIDS Behav 2021 25 12 3987 99 10.1007/s10461-021-03353-2 Waltz TJ, Powell BJ, Fernández ME, Abadie B, Damschroder LJ. (2019). Choosing implementation strategies to address contextual barriers: Diversity in recommendations and future directions. Implementation science, 14(42). Wansom T Guadamuz TE Vasan S Transgender populations and HIV: unique risks, challenges and opportunities J Virus Eradication 2016 2 2 87 93 10.1016/S2055-6640(20)30475-1 Wensing M, Oxman A, Baker R. (2011). Tailored implementation for chronic diseases (TICD): A project protocol. Implementation science, 6(103). doi:10.1186/1748-5908-6-103. Westmoreland DA, D’Angelo A, Nash D, Grov C. (2020). Individual and partner characteristics associated with intentions to use PrEP among partnered men, trans men, and trans women in sero-concordant and -discordant relationships in the United States. AIDS Education and Prevention, 32(5). Wilson EC Jalil EM Castro C Martinez Fernandez N Kamel L Grinsztejn B Barriers and facilitators to PrEP for transwomen in Brazil Glob Public Health 2019 14 2 300 8 10.1080/17441692.2018.1505933 30084305 Wood SM, Lee S, Barg FK, Castillo M, Dowshen N. (2017). Young transgender women’s attitudes toward HIV preexposure prophylaxis. J Adolesc Health, 60(5). doi:10.1016/j.jadohealth.2016.12.004. Woodward EN, Matthieu MM, Uchendu USR, S, & Kirchner JE. (2019). The health equity implementation framework: proposal and preliminary study of hepatitis C virus treatment. Implement Sci, 14(26). doi:10.1186/s13012-019-0861-y. Wu L, Schumacher C, Chandran A, Fields EL, Price A, Greenbaum A,. . the IMPACT Partner Collaborative. (2020). Patterns of PrEP retention among HIV pre-exposure prophylaxis users in Baltimore City, Maryland. Journal of acquired immune deficiency syndrome, 85. Xavier Hall CD, Bundy C, Foran JE, Carrillo H, Newcomb ME, Mustanski B. (2022). Identifying strategies for improving pre-exposure prophylaxis adherence: Perspectives from a sample of mostly adherent Young Men Who Have Sex with Men. (In Press). Zarwell M John SA Westmoreland D Mirzayi C Pantalone DW Golub S Grov C PrEP uptake and discontinuation among a US National sample of transgender men and women AIDS Behav 2021 25 4 1063 71 10.1007/s10461-020-03064-0 33057893 Zhao A, Dangerfield DT, Nunn A, Patel R, Farley JE, Ugoji CC, Dean LT. (2021). Pharmacy-based interventions to increase use of HIV pre-exposure prophylaxis in the United States: a scoping review. Journal of AIDS Behavior, 1–16.

36520334

PMC9753072

NO-CC CODE

2022-12-16 23:25:28

no

AIDS Behav. 2022 Dec 15;:1-19

utf-8

AIDS Behav

10.1007/s10461-022-03943-8

oa_other

==== Front Biomass Convers Biorefin Biomass Convers Biorefin Biomass Conversion and Biorefinery 2190-6815 2190-6823 Springer Berlin Heidelberg Berlin/Heidelberg 3641 10.1007/s13399-022-03641-4 Original Article A novel pH-sensitive antibacterial bilayer film for intelligent packaging Li Huiru Liu Guozhao Ye Kairu He Wanping Wei Hongyuan Dang Leping dangleping@tju.edu.cn grid.33763.32 0000 0004 1761 2484 School of Chemical Engineering and Technology, Tianjin University, Tianjin, 300072 People’s Republic of China 15 12 2022 114 5 8 2022 1 12 2022 5 12 2022 © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Intelligent single-layer packaging is widely used in food monitoring and storage. However, most single-layer intelligent packaging has poor mechanical strength and water barrier properties. In this study, a bilayer intelligent detector film based on polyvinyl alcohol-chitosan (PVA-CS)/nano-ZnO/sodium alginate (SA) combined with anthocyanin extract (cyanidin chloride) was prepared using a layer-by-layer solution casting assembly technique. The effects of different levels of anthocyanin extracts on the physical and functional properties of the films, including microstructure, mechanical property, barrier property, pH sensitivity, and antibacterial property, were investigated. The results show that the bilayers exhibit excellent physical properties, lower water vapor permeability, better light transmission and UV-blocking properties, a broader pH sensitivity (ΔE > 10), and good antibacterial activity. In short, the bilayer films studied are superior to the single-layer films in terms of their packaging potential for products with low moisture content, offering new directions for active intelligent packaging and biodegradable materials for the food industry. Graphical Abstract Keywords Composite bilayer films pH sensitivity Antibacterial Intelligent packaging ==== Body pmcIntroduction During storage, transportation, and distribution, food products will be damaged by environmental and other factors such as microorganisms and reducing freshness [1]. Moreover, food waste has been a further concern due to the spread of COVID-19 during the last 2 years. Intelligent packaging is aimed at protecting and controlling food products [2]. In particular, increased safety, quality, and information are provided by their ability to perform functions such as detection, tracking, and communication [3]. As a result, intelligent packaging films are widely used to detect food quality and extend food storage. Currently, the development of colorimetric pH indicator films for application in intelligent packaging to monitor the freshness of food has attracted people’s attention. This indicator film changes color in response to the food’s pH change. Thus, consumers can quickly distinguish whether food is spoiled without opening the package, which may reduce the amount of food wasted [4]. Anthocyanins are a class of water-soluble natural pigments widely found in plants, producing blue, purple, red, and intermediate tones in plant tissues [5]. The colors of anthocyanins are sensitive to pH changes because of their structural transformations [6, 7]. Therefore, incorporating anthocyanins into food packaging films can effectively monitor varying pH resulting from food decay. Meanwhile, it is necessary to have good antibacterial properties for a well-behaved intelligent packaging film. So far, metal nanoparticles including zinc oxide nanoparticles, copper oxide nanoparticles, and silver nanoparticles have been applied to a wide range of packaging films, among which nano-ZnO has attracted more attention due to its good antibacterial properties and non-toxicity [7–9]. Therefore, nano-ZnO was added to the PVA-CS blended film as an enhanced antimicrobial agent to extend the food preservation time while detecting the food quality. The traditional plastic-based food packaging films are derived from petroleum, which is not only non-degradable but has serious environmental concerns [10]. Recently, biomaterials have received increasing attention as intelligent packaging materials for food storage and quality monitoring. Among them, chitosan (CS) is derived from a non-toxic, renewable resource and has antimicrobial and antioxidant properties that prevent oxidative deterioration and prolong the shelf life of food products. In addition, it is widely used in many applications for its excellent biocompatibility and biodegradability [11]. Polyvinyl alcohol (PVA) is a non-toxic, biocompatible, biodegradable polymer with remarkable film-forming ability and barrier off-odors, which can be used in food packaging [12]. Pereira, Arruda, and Stefani reported that the mechanical properties of pure chitosan films are shallow, while the blending films from PVA and CS had enhanced physical properties and high antioxidant activities [13, 14]. Also, sodium alginate (SA) is a biocompatible, biodegradable, and non-toxic polysaccharide with sufficient fibrous chemistry to form gels in the presence of divalent cations and is a biopolymer that creates high-quality films [15]. Unlike single-layer films, bilayers could exhibit the uniqueness that single-layer films have even better. Ebrahimzadeh et al. prepared electrospun chitosan-polyvinyl alcohol bilayers containing essential oils by hybrid casting and electrostatic spinning methods, which exhibited better antimicrobial and physical properties than chitosan monolayers [16]. Moradi et al. developed plasma-improved chitosan/polyethylene bilayers containing summer savory essential oil, which also demonstrated reduced water vapor permeability (WVP) and increased mechanical strength of the bilayers compared to chitosan films [17]. However, sodium alginate is a hydrophilic group and cross-linking with multivalent cations is usually required to improve water resistance. Guo et al. reported that a high selectivity of Ca2+ was found due to its ability to form dense membrane networks by cross-linking only with G blocks [18]. Also, alginate’s high affinity and strong ionic bonding allow Ca2+ to obtain better film properties than Zn 2+, Mn 2+, and Mg 2+ when cross-linked. On the other hand, due to the electrostatic interaction between -NH3+ on chitosan and -COO− on sodium alginate forming a white insoluble polyelectrolyte polymer, a stable polyelectrolyte multilayer film is formed by a layer-by-layer casting method and this multilayer film can achieve even better performance [19]. To improve the performance of single-layer films, sodium alginate and chitosan solutions could be used to form multilayer composite films. Nevertheless, there are few reports on the ability to perform pH-responsive bilayer membranes for real-time monitoring of food products. Therefore, the present study aims to prepare a bilayer film that could provide real-time monitoring of food quality and improve the shortcomings of single-layer films as intelligent packaging for food by extending the shelf life of the food and enhancing the physical properties of water resistance. In this work, we have developed a pH-sensitive antibacterial bilayer film. The inner SA layer and the outer PVA-CS layer are prepared by a solution casting method and cross-linked by calcium ions. The bilayer films are further enhanced for food storage by adding nano-ZnO as an antimicrobial agent. The performance optimization of the new bilayer films is comprehensively evaluated by comparing the nano-bilayer films with single-layer films. It is demonstrated to have broad applications in food preservation indication and intelligent bionic devices, especially for low moisture substances. Experimental section Materials Polyvinyl alcohol (PVA, 1799, degree of alcoholics: 98–99%) and chitosan (degree of deacetylation 85%) were purchased from Tianjin Heowns Biochem, Ltd. Sodium alginate (AR grade), cyanidin chloride (purity: 25+%), nano-ZnO (purity: 99.9%, particle size: 30 ± 10 nm), glycerol (purity: 99.5%), and 1% acetic acid aqueous solution were purchased from Tianjin Kmart Chemical Technology Co., Ltd. and CaCl2 (0.5%) solution was purchased from Nanjing SenbeiJia Biotechnology Co., Ltd. Preparation of films 1.2 g of chitosan (CS) was dissolved in 110 mL of 1% aqueous acetic acid solution with glycerol (20% w/w based on chitosan) added as a plasticizer, and 0.3 g of nano-ZnO as an antibacterial agent. The solution was then stirred magnetically at room temperature until the chitosan powder was completely dissolved and the nano-ZnO was uniformly dispersed. In addition, 5.5 g of polyvinyl alcohol (PVA) with 1.1 g (20% w/w based on PVA) of glycerol added was dissolved in 110 ml before stirring at 95 ± 2 ℃, and cooled to room temperature. Subsequently, the anthocyanins were mixed at 0%, 5%, 10%, and 20% (w/w based on chitosan) content with the chitosan solution and PVA solution by magnetic stirring and vacuuming to remove air bubbles to obtain the outer film-forming solution. The inner film-forming solution was obtained by dissolving 2.2 g of sodium alginate (SA) in 220 mL of distilled water and adding 0.44 g of glycerol (20% w/w based on chitosan), followed by magnetic stirring at room temperature to obtain a clarified solution. Meanwhile, 20 g of the outer film-forming solution and 25 g of the inner film-forming solution were cast in separate plastic petri dishes (d = 35 mm) and dried at 55 °C. Then, the films were quickly wetted (t < 30 s) with 1 mL CaCl2 (0.5%, w/v) solution and dried at 55℃ to obtain the corresponding films as a control group. The single-layer films obtained from the inner film-forming solution were named SA, and the single-layer films obtained from the outer film-forming solution were named CP-1, CP-2, CP-3, and CP-4, respectively. The preparation of the bilayers was based on the multilayer solution casting method of Zhuang et al. [19]. Firstly, 22 g of the inner film-forming SA solution was cast on a medium plastic petri dish (d = 35 mm) and dried at 55℃ for 15 h. Subsequently, 10 g of the outer film-forming solution was decanted again into this petri dish and dried in an oven at 55℃ for 7 h. The resultant bilayer film was quickly wetted (t < 30 s) with 1 mL CaCl2 (0.5%, w/v) solution and dried at 55℃ for 2 h. And they were named SA-CP-1, SA-CP-2, SA-CP-3, and SA-CP-4. The dried films are stored at room temperature, and 50% relative humidity for at least 48 h before testing. Structural characterization Fourier transform infrared (FT-IR) spectroscopy FTIR spectra were obtained by scanning films over a range of 400–4000 cm−1 using a spectrometer (Nicolet IS 10, America). X-ray diffraction The phase structure of the films was measured by X-ray diffraction (SmartLab, Japan), with Cu produced at 40 kV and 200 mA radiation, and scanned at a 5°/min rate over a diffraction angle range of 5–60° (2θ) [20]. Scanning electron microscope The surface and cross sections of the films were observed using SEM (Regulus 8100, S-4800, Hitachi, Japan) with an accelerating voltage of 3.00 kV. Physical properties measurement Thickness and mechanical properties The thickness of the film was measured by selecting five random points using a digital micrometer. The tensile strength (TS) and elongation at break (EB) of films were measured by a universal mechanical testing machine (CTM6103, MTS, USA) with an initial grip separation of 40 mm and a strain rate of 60 mm/min. The films were cut into strips (60 mm × 6 mm), and the samples were taken five times in parallel [21]. Moisture content and swelling degree The film samples (d = 35 mm) (M0, g) were placed in an oven set at 70 °C for 24 h to evaporate the water contained in the film and weighed (M1, g). The dried films were then immersed in 50 ml of distilled water at room temperature for 24 h. The excess water was wiped off the film’s surface with filter paper and weighed (M2, g). Afterwards, they are dried again in an oven at 70 °C for 24 h and weighed (M3, g) [20]. The moisture content (MC (%)) and swelling degree (Sw (%)) were calculated by the following equations:MC%=M0-M1/M0×100 SW%=M2-M3/M1×100 Water vapor permeability (WVP) The film water vapor permeability was determined according to the method of Yong et al. with modifications [22]. Six grams g of anhydrous silica gels was poured into a test tube (d = 20 mm), and the mouth of the tube was sealed with a film. Then, the tubes were placed in a desiccator at a temperature maintained at 20 °C, containing distilled water (100% relative humidity), and weighed only once a day for 6 days. WVP was calculated as follows:WVP=W×xt×A×ΔP where W is the change in weight of the tube (g), x is the film thickness(mm), t is the time interval (s), ΔP is the water vapor pressure difference across the film at 20 °C(KPa), and A is the area of m2. Contact angle test The water contact angle of the film surface is measured using a contact angle tester (Do nano, China). The film sample was cut into a rectangle of 4 cm × 1 cm. A drop of ultrapure water (10 μL) was dropped onto the surface of the film samples fixed to the platform. The contact angle of the drop was recorded after 8 s of water drop contact. All films were selected at 5 different locations for measurement and the average value was reported. Light transmittance and opacity [9] The light transmittance of the films was determined by scanning the film samples in the 200–800 nm range using a UV-1601 spectrophotometer (Beifen Riley, China). The opacity of the film is calculated as follows:Opacity=Abs600x where Abs600 and x are the film absorbance at 600 nm and the film thickness (mm), respectively. Functional properties of food packing pH sensitivity According to Liu et al., the pH-sensitivity of this anthocyanin was assessed in the method with several modifications [23]. In simple terms, 5 mg of anthocyanins were dissolved separately in the 50 ml buffer of different pH for 20 min. The UV–Vis spectra of the extracted liquids could then be recorded on a UV–Vis spectrophotometer (TU-1900, Pu-Analysis, China) through a sweep of 450–700 nm. To verify the pH-sensitivity of the films, the films (d = 35 mm) were dipped into the different buffer solutions (pH = 2–12) for 20 min. The chromatic parameters of the films were recorded with a colorimeter (CR-10, Konica Minolta, Japan) and photographed. The results were expressed in terms of the parameters L*, a*, b*, and ΔE [24]. ΔE was calculated as follows:ΔE=L∗-L02+a∗-a02+b∗-b02 where L0, a0, and b0 are the original grey value of films. L* = lightness, a* = red to green, and b* = yellow to blue. Ammonia sensitivity Film samples (2 cm × 2 cm) were placed on top of the inner Petri dishes (d = 9 cm) containing 15 mL of 0.1 mg/mL ammonia solution at room temperature for 300 min. The color changes of the film samples were recorded at different time intervals (5, 10, 20, 30, 40, 50, 60, 180, 300 min). Antibacterial activity The antibacterial capacity of the film samples was determined by judging the inhibition circle (mm) against Staphylococcus aureus and Escherichia coli using the agar diffusion test described by Abral et al. with some modifications [25]. Bacterial colonies were counted based on the 0.5 McFarland standard. By comparing the turbidity between the 0.5 McFarland standard and the microbial culture, the microbial suspension cell density was approximately 1.5 × 108 CFU/mL. Firstly, 100 μL of the microbial suspension was evenly spread on the Luria–Bertani agar. Then, using forceps, 10 mm film discs, sterilized under a UV lamp for 15 min, were placed in an orderly fashion on the microbial suspension to be tested. Finally, all the media were incubated in a biochemical incubator at 37 °C for 24 h before recording their inhibition circle diameter (mm). Three sets of parallel experiments were set. Statistical analysis All films were measured in parallel greater than three times, and results were expressed as average ± standard deviation (SD). All statistics were handled using SPSS software (25.0, SPSS Inc., Chicago, IL, USA), and Duncan’s multiple range test (p < 0.05) was used to compare differences among the data. Results and discussion Molecular bonds analysis FTIR spectra of all films are shown in Fig. 1. The single-layer films exhibit similar peak areas, and both the SA and bilayer films also exhibit similar peak areas. Single-layer films which showed bands at 3302–3286 cm−1, 2939 cm−1, 1569 cm−1, 1412 cm−1, and 1045 cm−1 were attributed to O–H stretching, C-H asymmetric stretching, amide II groups, CH-CH2 bending in the basic carbon skeleton, and C-O stretching, respectively [23]. In addition, single-layer films presented at 1328 cm−1 (amide III), 1142 cm−1 (C–O–C stretching), 921 cm−1 (C-O stretching), and 847 cm−1 (C-H vibrational stretching) [26]. The bond at 3302 cm−1 (O–H stretching) in the bilayer containing the SA layer shifted towards lower wave numbers, while the bands at 1594 cm−1, 1406 cm−1, and 1022 cm−1 became sharper and broader, probably due to the presence of the SA layer and electrostatic interactions between the SA and CP layers [19]. Notably, with the addition of anthocyanin (cyanidin chloride), the waveband of both films at 3302 cm−1 (O–H stretching) shifted to a lower wave number near 3290 cm−1, as a result of the hydrogen bonds formed between the components of the film, altering the physical and chemical interactions between the aromatic rings of the anthocyanins and polysaccharides [9].Fig. 1 FT-IR spectra of SA and films with different contents Phase structure analysis The XRD patterns of CP-1, CP-2, CP-3, CP-4, SA-CP-1, SA-CP-2, SA-CP-3, and SA-CP-4 films are shown in Fig. 2. Both SA-CP and CP films showed broad peaks in the 2θ range of 19 to 20°. The intensity of the peaks was significantly lower in the SA-CP film than in the CP film, which was attributed to the lower content of the CP mixture in the SA-CP film. In terms of peak intensity, it is found that the peak intensity at 2θ = 19.7° decreases with the addition of anthocyanin, while it increases with the increase of anthocyanin content. This phenomenon may result from the formation of new hydrogen bonds between the anthocyanins and the two polymers (PVA, CS), thus disrupting the original interactions between the polymeric substrates and promoting the spatial reconfiguration of the polymer chains, which was supported by FTIR spectrograms, and secondly, the plasticization and electrostatic interactions between anthocyanins and all components [27, 28]. However, as the anthocyanin content increases, this could weaken the interaction between the two interpretations, thereby allowing an increase in the crystallinity of the polymer matrix.Fig. 2 XRD spectra of films with different contents Cross-sectional micromorphologies observation Figure 3 exhibited the surface and cross-sectional micromorphology of the films. For the surfaces (Fig. 3A), both the single-layers and bilayers are smooth, dense, and homogeneous, indicating a good compatibility between CS, nano-ZnO, PVA, and anthocyanin extracts. However, the smoothness of the bilayers is lower than the single-layer films, which may be caused by the uneven drying of the bilayers during the layer-by-layer casting process. It can be seen that all the films show relatively continuous, dense, and non-cracking cross sections (Fig. 3B). Moreover, all the single layers demonstrated a homogenous and smooth cross section without voids, indicating good compatibility between CS, nano-ZnO, PVA, cyanidin chloride, and glycerol. SA-CP-2 films are slightly rough in cross sections with small agglomerates, which could be caused by the cross-linking reaction of Ca2+ with sodium alginate to form insoluble calcium alginate. In the SA-CP-3 film, a stratification was observed between the top and bottom layers, which may result from the SA layer’s low moisture content prior to casting the CP layer, preventing the two layers from being well compatible [29]. Besides, the results show that the other cross sections do not clearly exhibit a bilayer structure; instead, the outer layer exhibits a continuous and dense microstructure tightly bound to the inner layer. The ability of CP and SA to form a stable bilayer is further demonstrated. It stems from the fact that CP and SA could coalesce to form polyelectrolyte complexes by electrostatic gravity, while Ca2+ chelated with the guluronic acid blocks of the SA chain, creating a physically cross-linked SA/Ca2+ network to enhance their electrostatic binding [19]. These results were also confirmed in the FTIR analysis.Fig. 3 SEM images of surface (A) and cross sections (B). SA (a), CP-1(b), CP-2 (c), CP-3 (d), and CP-4 (e) are single-layer films and SA-CP-1 (f), SA-CP-2 (g), SA-CP-3 (h), and SA-CP-4 (i) are bilayer films Thickness and mechanical properties measurement The thickness and mechanical properties of the films are presented in Table 1. There was no significant difference between the thickness of the SA-CP and CP films. Among them, the thickness of the SA layer (0.035 ± 0.003 mm) is the thinnest. The thicknesses of single-layer films are thicker than bilayers. Since the SA films obtained by casting and drying equal amounts of solution are thicker than the CP film, the bilayers with inner of a SA layer and the outer a CP layer are thus less thick than the CP films.Table 1 Summary of the physical properties of the films Films Thickness (mm) Tensile strength (MPa) Elongation at break (%) Moisture content (%) WVP (g mm/m2day Kpa) Opacity (mm−1) SA 0.035 ± 0.003d 60.55 ± 1.91a 8.33 ± 0.11f 17.30 ± 0.42a 3.64 ± 0.34e 0.05 ± 0.01e SA-CP-1 0.055 ± 0.004c 22.94 ± 4.46de 20.25 ± 1.05e 11.31 ± 0.46bc 5.69 ± 0.92d 0.05 ± 0.04e SA-CP-2 0.068 ± 0.009b 31.01 ± 2.81cde 27.90 ± 0.76de 12.05 ± 0.13b 7.41 ± 0.83c 0.35 ± 0.05d SA-CP-3 0.063 ± 0.004bc 20.89 ± 0.68e 34.75 ± 0.12d 12.13 ± 0.26b 6.74 ± 0.70c 0.61 ± 0.02c SA-CP-4 0.065 ± 0.012bc 26.07 ± 6.45cde 24.59 ± 0.15e 11.86 ± 0.42b 6.96 ± 0.43c 1.02 ± 0.05b CP-1 0.124 ± 0.009a 32.96 ± 4.02bcd 116.45 ± 7.52a 11.35 ± 0.73bc 13.34 ± 0.73a 0.06 ± 0.02e CP-2 0.116 ± 0.004a 33.08 ± 6.05bcd 109.51 ± 7.67ab 10.90 ± 0.44cd 12.64 ± 0.57ab 0.55 ± 0.09c CP-3 0.114 ± 0.019a 35.56 ± 7.11bc 104.59 ± 10.72bc 10.14 ± 0.53de 12.31 ± 0.39b 1.07 ± 0.14b CP-4 0.125 ± 0.011a 43.13 ± 8.24b 99.10 ± 8.84c 9.82 ± 0.51e 13.23 ± 0.65a 2.26 ± 0.06a All data are shown as mean ± standard deviation (SD). Different letters indicate differences at the p < 0.05 level are significant. Elongation at break (EB) and tensile strength (TS) reflect the flexibility and mechanical resistance of a food packaging material, respectively [30]. CP films have a higher elongation at break (EB) than SA-CP and SA films, while CP films have higher tensile strength (TS) than SA-CP films but are lower than SA films. Assis et al. observed that an increase in the polymer network’s randomization could diminish the generation of structural consistency, resulting in a lower value of TS [31]. With the anthocyanin incorporation, the EB of the CP films decreased, and the TS increased, reflecting the lower flexibility of the chitosan/PVA films and the enhanced mechanical resistance of the films by adding anthocyanin extract. One is that the interaction between the chains of film components is hindered by anthocyanins, thereby reducing flexibility [30]. Another is hydrogen bonds could be formed between anthocyanins and the film matrix and between OH− in the SA layer and NH3+ in the chitosan layer (Fig. 4), thus increasing the mechanical resistance [6]. The difference in mechanical properties is insignificant for SA-CP films containing less anthocyanin than CP films. Nevertheless, in terms of mechanical properties, the tensile strength of the bilayer films in this study was not as high as that of the single-layer films up to 32.34 ± 3.5 Mpa, but the tensile strength of the prepared films was 103% higher compared to the data reported by Liu et al. at 15.90 ± 2.86 MPa [23]. In addition, the elongation at the break of the bilayer films was similar to the reported value of 39.50 ± 5.51%, while the single-layer films were much higher than this value. Consequently, the mechanical properties were all improved and could meet the needs of food packaging.Fig. 4 Schematic diagram of ionic interactions between chitosan and alginate in a bilayer Moisture content, swelling, and WVP According to Table 1, SA film (7.30 ± 0.42%) has the highest moisture content of all films, and the water solubility is close to 100% due to its hydrophilicity. Other than this, SA-CP films exhibit similar water contents. In the case of CP films, for CP-1 without anthocyanin addition, the moisture content was higher than for CP films with anthocyanin addition. This phenomenon was related to forming molecular hydrogen bonds between CS/PVA bound to the anthocyanin and replaced a part of the CP interacting with moisture. In previous studies, it was also confirmed that the addition of anthocyanin extracts reduced the water content of the films [32, 33]. Swelling ability reflects the film’s hydrophilicity as an essential film property [34]. The swelling of all films is shown in Fig. 5a.The results indicate that the swelling of the SA-CP films is much more significant than that of the CP films, which could be related to the strong hydrophilic of the SA layer in the bilayer. The presence of an SA layer in the bilayer, which is strongly hydrophilic, and the interaction between the SA layer and the water molecules during swelling increase molecular spacing and volume.Fig. 5 Swelling index (a) and water contact angle (b) of films [Different letters indicate significant differences. (p < 0.5)] The WVP of food packaging film is essential in reducing the ability to transfer moisture from food to the environment and is a vital reference for comparing moisture barrier performance [30]. The result of WVP for SA-CP, CP, and SA films is shown in Table 1. The WVP values of SA-CP films and SA films were significantly weaker than those of CP films among all the films tested. This study found no significant change in WVP values after adding anthocyanins to the CP/PVA solution. Generally, the WVP of laminated films depends mainly on the WVP value of each layer of the film, and SA has a lower WVP because of the thinness of its film [35]. The lower WVP of the bilayers may be related to the bilayers having fewer free hydroxyl groups compared to the single-layer films. The free hydroxyl group could facilitate the attachment of water molecules down the films and through the films by enhancing the interaction between the water molecules and the polysaccharide chains. For bilayers, hydrogen bonding between the bilayers and electrostatic interactions between them reduced the adsorption of hydrophilic groups to water, making them more water resistant. In terms of the addition of anthocyanins, no change in the corresponding groups was induced, so the water resistance was not significantly altered. The hydrophobicity of all films was determined by the water contact angle and the results are shown in Fig. 5b. The water contact angle of the SA film was 41.77° (< 65°), indicating the hydrophilic nature of the sodium alginate film, indicated also by the swelling properties. As the anthocyanin content increased, the water contact angle increased from 77.85 to 84.39° for the singles and from 86.29 to 96.00° for the bilayers. This might be attributable to the reduction of surface-free energy through enhanced hydrogen bonding interactions between anthocyanin, chitosan, and polyvinyl alcohol molecules, resulting in the increased hydrophobicity of the films [36]. Compared to the singles, the water contact angle of the bilayers was smaller and the bottom area increased, which could be attributed to the fact that the inner layer of the bilayer was an SA film making it less hydrophobic. Light transmittance The UV–visible barrier capability is an essential factor in preventing UV–vis light-induced food oxidation to extend food storage [37, 38], and also an integral element in assessing whether the appearance could be used for practical packaging applications. Figure 6 presents the light transmission spectra of all films. It can be seen that SA, SA-CP-1, and CP-1 represented in black, pink, and navy, respectively, have a light transmission more excellent than 80% and thus are transparent films. The following curves represent the single-layer films with the addition of anthocyanins and the bilayer films. The light transmission of the films tends to decrease with the accumulation of anthocyanins, which indicates that the films with the addition of anthocyanins have better UV–visible light barrier properties [32]. It was also discovered that all films incorporating anthocyanins had a significantly lower light transmission than all films without anthocyanins, which indicated that films containing anthocyanins were expected to have better UV–visible light barrier properties.Fig. 6 Light transmittance of films In addition, the films exhibited a gradual decrease in light transmission as the anthocyanin content increased. As a result, the light transmission of the SA-CP films was higher than that of the CP films for the added anthocyanins, which was attributed to the SA-CP containing fewer anthocyanins than the CP at the same level of incorporation. According to Table 1, the opacity of the bilayers is lower than that of the monolayers, and the CP-4 film has the highest opacity (2.26 ± 0.06a), indicating that sufficient anthocyanins could effectively diminish the light transmission of the films [9]. According to Table 1, all films in this study have low opacity values and could be used for food packaging, but double-layer films have better transparency and UV–visible light barrier properties. pH sensitivity evaluation Color response analysis of anthocyanin extract Anthocyanins can cause structural changes at a wide range of pH values, resulting in visible color changes [39]. According to Fig. 7, the color range of anthocyanin solutions was peach > pink > purple > celadon > yellow from acidic buffered solutions to basic buffered solutions. An increased pH from 2 to 6 decreased the intensity of the anthocyanin color until a minimum value of 6, after which an increase in pH increased the intensity of the color. This phenomenon is ascribed to the presence of anthocyanins as flavylium cation at pH values of approximately 3 or lower, producing structural shifts in carbinol pseudobase, quinone base, and quinone anion as the pH increases [26]. With the structural changes in anthocyanin, the maximum absorption peak shifted from 510 to 570 nm.Fig. 7 Color variations of cyanidin chloride extracts and UV–vis spectra in different buffer solutions (pH 2–12) pH sensitivity of films In this study, we used the changing color of the films to evaluate pH sensitivity, a crucial indicator for intelligent packaging, and the results are displayed in Fig. 8. The results indicate that both SA-CP and CP films were highly pH-sensitive. The film color appears pH = 2–3 in pink, pH = 4–6 in purple, pH = 7–11 in blue, and pH = 12 in green. In particular, the color of the film changes related to the anthocyanin content, and the color intensity increases with increasing anthocyanins. According to Table 2, ΔE values can be obtained for all films of CP and SA-CP. The color change (ΔE > 3) could be easily observed with the naked eye [34]. In this work, the ΔE values for both SA-CP and CP films are more significant than 3, which is to say, all films could be visually detected over a wide range of pH values. The ΔE values in this study are much greater than the reported, so the bilayer films in this study have better-naked eye detection performance [23]. In contrast, CP-4 exhibited a slightly smaller ΔE (p < 0.5) under basic conditions at the same pH, while the ΔE of the SA-CP films was slightly more significant than the CP films, indicating that the bilayer had a more pronounced visually observable color change. In addition, CP or SA-CP films (Fig. 8) and cyanidin chloride solution (Fig. 7) exhibited different color changes in the same pH buffer solution. The change is attributed to the strong electrostatic interaction between cyanidin chloride and chitosan, leading to a structural modification of the anthocyanin, and thus affecting the discoloring ability [40].Fig. 8 Visual color change of CP-2, CP-3, CP-4, SA-CP-2, SA-CP-3, and SA-CP-4 films at pH 2–12 Table 2 The ΔE values of films after being soaked into pH 2 to 12 buffer solutions for 20 min pH values SA-CP-2 SA-CP-3 SA-CP-4 CP-2 CP-3 CP-4 pH = 2 32.68 ± 1.56a 35.16 ± 0.26a 37.73 ± 0.63a 35.89 ± 0.77a 44.91 ± 0.96a 39.79 ± 1.10a pH = 3 31.73 ± 1.58a 30.36 ± 0.75b 31.36 ± 0.54b 31.18 ± 0.80b 42.69 ± 0.84a 35.08 ± 1.33b pH = 4 24.64 ± 1.59b 29.57 ± 1.41b 26.60 ± 0.64c 27.73 ± 0.84c 31.57 ± 1.53b 36.14 ± 0.97b pH = 5 18.76 ± 1.73c 26.69 ± 1.40c 22.76 ± 0.58d 23.43 ± 0.81d 26.61 ± 1.26c 24.14 ± 1.11c pH = 6 19.23 ± 1.59c 15.61 ± 0.44d 16.05 ± 0.56fg 23.01 ± 1.22d 23.80 ± 1.91c 22.51 ± 0.84c pH = 7 14.68 ± 1.71d 8.15 ± 1.01f 14.58 ± 0.75h 14.39 ± 0.49f 5.19 ± 0.97f 6.84 ± 0.74e pH = 8 19.22 ± 2.36c 7.85 ± 0.98f 12.19 ± 0.71i 14.45 ± 0.90f 10.60 ± 1.72e 3.74 ± 0.57f pH = 9 12.52 ± 2.25d 15.33 ± 2.25d 15.08 ± 0.67gh 8.55 ± 0.43h 7.36 ± 1.62f 3.98 ± 1.28f pH = 10 12.88 ± 2.20d 8.28 ± 1.59f 15.44 ± 0.61gh 7.13 ± 0.43i 13.89 ± 2.17d 3.02 ± 1.46f pH = 11 12.40 ± 2.14d 10.83 ± 1.68e 19.40 ± 0.63e 10.48 ± 0.68g 16.71 ± 2.90d 17.08 ± 1.03d pH = 12 13.58 ± 1.49d 26.42 ± 0.36c 16.90 ± 0.08f 17.61 ± 0.99e 24.81 ± 2.10c 6.57 ± 1.40e All data are shown as mean ± standard deviation (SD). Different letters indicate differences at p < 0.05 level are significant. Ammonia sensitivity In the action of bacteria and enzymes, the proteins in animal foods are gradually broken down into peptides and amino acids. Furthermore, the protein is degraded into low molecular compounds containing nitrogen, which is the source of the spoilage smell of animal food [41]. Thus, indicator films are essential for the detection of ammonia sensitivity. In this study, ammonia was used to test the pH sensitivity of the films. Ammonia could diffuse into the films and then be hydrolyzed to hydroxide ions, creating an alkaline environment in the films. As shown in Fig. 9, all films except CP-4 showed significant differences from the initial film after 10 min. CP-4 film did not show significant changes visible to the naked eye at 60 min, probably due to the high anthocyanin content of the film resulting in a darker color that was not easily observed. At 180 min and 300 min, the films showed significant changes. The color gradually changes from purple to grey-blue to yellow-green as the time increases. The bilayers had a large difference in ΔE, suggesting that the bilayers may have good ammonia sensitivity.Fig. 9 The colorimetric response of bilayer films SA-CP-2 (a), SA-CP-3 (b), SA-CP-4 (c), and single-layer films CP-2 (d), CP-3 (e), CP-4 (f) for 0.1 mg/mL ammonia for 60 min; the colorimetric response of all films for 0.1 mg/mL ammonia for 180 min and 300 min (g) Antibacterial activity To extend food products’ storage life, the films’ antibacterial properties play a vital role. E. coli and S. aureus are the most common Gram-negative and Gram-positive bacteria, respectively. They are common foodborne pathogenic microorganisms and a common source of disease infection in humans. The antibacterial activity of the samples was evaluated by the diameter of the inhibition circle for two bacteria, E. coli (Gram-negative) and S. aureus (Gram-positive). As shown in Fig. 10 and Table 3, the results revealed that the SA film did not inhibit two bacteria, while the other films all had varying degrees of bacteria inhibition. Meanwhile, the maximum inhibition circle diameters were 21.17 ± 1.04 mm and 13.05 ± 0.56 mm for S. aureus and E. coli, respectively, indicating that the remaining eight films had more effective antibacterial capacity against Gram-positive bacteria. The antimicrobial activity of nano-ZnO had been demonstrated in several previous studies, both by the release of Zn2+ from the film matrix and by the production of hydrogen peroxide on the surface of the ZnO nanoparticles [42, 43]. Furthermore, according to Fig. 10, the bilayers we prepared were as effective as the single layers in inhibiting bacteria. At the same time, the data for the single layers demonstrate a tendency for the diameter of inhibition to rise with increasing amounts of ZnO added. Therefore, in the subsequent work, we can adjust the ZnO content in the bilayers according to the requirements of specific food storage conditions for bacterial inhibition, thus regulating the performance of the composite films by design.Fig. 10 Inhibitory zone of the fabricated packaging films (SA is sodium alginate film, addition of different anthocyanins to the bilayers is SA-CP-1 to SA-CP-4, and the single-layer films are CP-1 to CP-4.) Table 3 Antibacterial activity of the films Films Diameter of inhibition zone (mm) Staphylococcus aureus ( +) Escherichia coli ( −) SA 0h 0e SA-CP-1 13.98 ± 0.39g 10.59 ± 0.24d SA-CP-2 15.75 ± 0.22f 11.81 ± 0.20c SA-CP-3 16.38 ± 0.14ef 11.87 ± 0.47c SA-CP-4 16.72 ± 0.28de 12.86 ± 0.16a CP-1 17.23 ± 0.25d 12.63 ± 0.16ab CP-2 19.47 ± 0.45b 12.56 ± 0.49ab CP-3 18.23 ± 0.59c 12.08 ± 0.13c CP-4 21.17 ± 1.04a 13.05 ± 0.56a All data are shown as mean ± standard deviation (SD). Different letters indicate differences at the p < 0.05 level are significant. Conclusions In this study, a novel homogeneous and flexible bilayer film is developed by a calcium ion cross-linking method, which is designed for the food intelligent packing. The results show that the bilayers exhibit a dense and homogeneous monolayer structure and excellent physical properties. With the addition of anthocyanins, the films show a decrease in elongation at break, an increase in tensile strength, an enhancement in UV–visible light barrier properties, a decrease in water vapor permeability, and an increase in swelling properties. Under the same conditions of addition, the bilayer films offer better light transmission and excellent UV-blocking properties. Meanwhile, the bilayer films have antimicrobial properties, which prolong the storage life of the food. Last, the bilayers demonstrated a broader pH sensitivity (ΔE > 10) in different pH buffer solutions, which is critical for intelligent packaging that effectively monitors the freshness of food products in real-time. Ultimately, this study presented the overall performance of the bilayer pH-responsive films, providing a new direction for biodegradable active intelligent packaging. Author contribution Huiru Li: investigation, methodology, formal analysis, data curation, visualization, writing—original draft. Guozhao Liu: writing—review and editing. Kairu Ye: writing—review and editing. Wanping He: writing—review and editing. Hongyuan Wei: writing—review and editing. Leping Dang: writing—review and editing, supervision, funding acquisition, project administration. Data availability Not applicable. Declarations Ethical approval Not applicable. Competing interests The authors declare no competing interests. Highlights • Stable bilayers could be formed by electrostatic interactions between chitosan and sodium alginate. • Preparation of chitosan/polyvinyl alcohol/sodium alginate bilayers for intelligent packaging. • Bilayer films exhibit superior water vapor barrier and pH sensitivity compared to single-layer films. Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Zhang J Huang X Shi J Liu L Zhang X Zou X Xiao J Zhai X Zhang D Li Y Shen T A visual bi-layer indicator based on roselle anthocyanins with high hydrophobic property for monitoring griskin freshness Food Chem 2021 355 129573 10.1016/j.foodchem.2021.129573 33799267 2. Pereira PF Andrade CT Optimized pH-responsive film based on a eutectic mixture-plasticized chitosan Carbohyd Polym 2017 165 238 246 10.1016/j.carbpol.2017.02.047 3. JH Han, CH Ho, ET (2005) Rodrigues, Intelligent packaging, Innovations in Food Packaging 138–155 10.1016/B978-012311632-1/50041-3 4. Li Y Wu K Wang B Li X Colorimetric indicator based on purple tomato anthocyanins and chitosan for application in intelligent packaging Int J Biol Macromol 2021 174 370 376 10.1016/j.ijbiomac.2021.01.182 33539953 5. He J Giusti MM Anthocyanins: natural colorants with health-promoting properties Annu Rev Food Sci T 2010 1 1 163 187 10.1146/annurev.food.080708.100754 6. Wang X Yong H Gao L Li L Jin M Liu J Preparation and characterization of antioxidant and pH-sensitive films based on chitosan and black soybean seed coat extract Food Hydrocolloid 2019 89 56 66 10.1016/j.foodhyd.2018.10.019 7. Zhang X Liu Y Yong H Qin Y Liu J Liu J Development of multifunctional food packaging films based on chitosan, TiO2 nanoparticles and anthocyanin-rich black plum peel extract Food Hydrocolloid 2019 94 80 92 10.1016/j.foodhyd.2019.03.009 8. Qin Y Liu Y Yuan L Yong H Liu J Preparation and characterization of antioxidant, antimicrobial and pH-sensitive films based on chitosan, silver nanoparticles and purple corn extract Food Hydrocolloid 2019 96 102 111 10.1016/j.foodhyd.2019.05.017 9. Sun J Jiang H Wu H Tong C Pang J Wu C Multifunctional bionanocomposite films based on konjac glucomannan/chitosan with nano-ZnO and mulberry anthocyanin extract for active food packaging Food Hydrocolloid 2020 107 105942 10.1016/j.foodhyd.2020.105942 10. Atarés L Chiralt A Essential oils as additives in biodegradable films and coatings for active food packaging Trends Food Sci Tech 2016 48 51 62 10.1016/j.tifs.2015.12.001 11 Ali A Ahmed S A review on chitosan and its nanocomposites in drug delivery Int J Biol Macromol 2018 109 273 286 10.1016/j.ijbiomac.2017.12.078 29248555 12. Li X Yang X Wang Z Liu Y Guo J Zhu Y Shao J Li J Wang L Wang K Antibacterial, antioxidant and biocompatible nanosized quercetin-PVA xerogel films for wound dressing Colloid Surface B 2022 209 112175 10.1016/j.colsurfb.2021.112175 13. Hajji S Chaker A Jridi M Maalej H Jellouli K Boufi S Nasri M Structural analysis, and antioxidant and antibacterial properties of chitosan-poly (vinyl alcohol) biodegradable films Environ Sci Pollut R 2016 23 15 15310 20 10.1007/s11356-016-6699-9 14. Pereira VA de Arruda INQ Stefani R Active chitosan/PVA films with anthocyanins from Brassica oleraceae (Red Cabbage) as Time-Temperature Indicators for application in intelligent food packaging Food Hydrocolloid 2015 43 180 188 10.1016/j.foodhyd.2014.05.014 15. Afshar M Dini G Vaezifar S Mehdikhani M Movahedi B Preparation and characterization of sodium alginate/polyvinyl alcohol hydrogel containing drug-loaded chitosan nanoparticles as a drug delivery system J Drug Deliv Sci Tec 2020 56 101530 10.1016/j.jddst.2020.101530 16. Ebrahimzadeh S Bari MR Hamishehkar H Kafil HS Lim LT Essential oils-loaded electrospun chitosan-poly(vinyl alcohol) nonwovens laminated on chitosan film as bilayer bioactive edible films LWT 2021 144 111217 10.1016/j.lwt.2021.111217 17. Moradi E Moosavi MH Hosseini SM Mirmoghtadaie L Moslehishad M Khani MR Jannatyha N Shojaee-Aliabadi S Prolonging shelf life of chicken breast fillets by using plasma-improved chitosan/low density polyethylene bilayer film containing summer savory essential oil Int J Biol Macromol 2020 156 321 328 10.1016/j.ijbiomac.2020.03.226 32234439 18. Liling G Di Z Jiachao X Xin G Xiaoting F Qing Z Effects of ionic crosslinking on physical and mechanical properties of alginate mulching films Carbohyd Polym 2016 136 259 265 10.1016/j.carbpol.2015.09.034 19. Zhuang C Jiang Y Zhong Y Zhao Y Deng Y Yue J Wang D Jiao S Gao H Chen H Mu H Development and characterization of nano-bilayer films composed of polyvinyl alcohol, chitosan and alginate Food Control 2018 86 191 199 10.1016/j.foodcont.2017.11.024 20 Ferreira ARV Torres CAV Freitas F Sevrin C Grandfils C Reis MAM Alves VD Coelhoso IM Development and characterization of bilayer films of FucoPol and chitosan Carbohyd Polym 2016 147 8 15 10.1016/j.carbpol.2016.03.089 21. Liu J Liu S Wu Q Gu Y Kan J Jin C Effect of protocatechuic acid incorporation on the physical, mechanical, structural and antioxidant properties of chitosan film Food Hydrocolloid 2017 73 90 100 10.1016/j.foodhyd.2017.06.035 22. Yong H Wang X Zhang X Liu Y Qin Y Liu J Effects of anthocyanin-rich purple and black eggplant extracts on the physical, antioxidant and pH-sensitive properties of chitosan film Food Hydrocolloid 2019 94 93 104 10.1016/j.foodhyd.2019.03.012 23. Liu J Huang J Ying Y Hu L Hu Y pH-sensitive and antibacterial films developed by incorporating anthocyanins extracted from purple potato or roselle into chitosan/polyvinyl alcohol/nano-ZnO matrix: comparative study Int J Biol Macromol 2021 178 104 112 10.1016/j.ijbiomac.2021.02.115 33609585 24. Ge Y Li Y Bai Y Yuan C Wu C Hu Y Intelligent gelatin/oxidized chitin nanocrystals nanocomposite films containing black rice bran anthocyanins for fish freshness monitorings Int J Biol Macromol 2020 155 1296 1306 10.1016/j.ijbiomac.2019.11.101 31751683 25. Abral H Ariksa J Mahardika M Handayani D Aminah I Sandrawati N Pratama AB Fajri N Sapuan SM Ilyas RA Transparent and antimicrobial cellulose film from ginger nanofiber Food Hydrocolloid 2020 98 105266 10.1016/j.foodhyd.2019.105266 26. Lian Z Zhang Y Zhao Y Nano-TiO2 particles and high hydrostatic pressure treatment for improving functionality of polyvinyl alcohol and chitosan composite films and nano-TiO2 migration from film matrix in food simulants Innov Food Sci Emerg 2016 33 145 153 10.1016/j.ifset.2015.10.008 27. Sun J Du Y Ma J Li Y Wang L Lu Y Zou J Pang J Wu C Transparent bionanocomposite films based on konjac glucomannan, chitosan, and TEMPO-oxidized chitin nanocrystals with enhanced mechanical and barrier properties Int J Biol Macromol 2019 138 866 873 10.1016/j.ijbiomac.2019.07.170 31356935 28 Liang T Sun G Cao L Li J Wang L A pH and NH3 sensing intelligent film based on Artemisia sphaerocephala Krasch. gum and red cabbage anthocyanins anchored by carboxymethyl cellulose sodium added as a host complex Food Hydrocolloid 2019 87 858 868 10.1016/j.foodhyd.2018.08.028 29. Zhao Y Qiu J Xu J Gao X Fu X Effects of crosslinking modes on the film forming properties of kelp mulching films Algal Res 2017 26 74 83 10.1016/j.algal.2017.07.006 30. Liu Y Qin Y Bai R Zhang X Yuan L Liu J Preparation of pH-sensitive and antioxidant packaging films based on κ-carrageenan and mulberry polyphenolic extract Int J Biol Macromol 2019 134 993 1001 10.1016/j.ijbiomac.2019.05.175 31129204 31 Assis RQ Lopes SM Costa TMH Flôres SH Rios AdO Active biodegradable cassava starch films incorporated lycopene nanocapsules Ind Crops Prod 2017 109 818 827 10.1016/j.indcrop.2017.09.043 32. Yong H Liu J Qin Y Bai R Zhang X Liu J Antioxidant and pH-sensitive films developed by incorporating purple and black rice extracts into chitosan matrix Int J Biol Macromol 2019 137 307 316 10.1016/j.ijbiomac.2019.07.009 31276717 33. Musso YS Salgado PR Mauri AN Smart gelatin films prepared using red cabbage (Brassica oleracea L.) extracts as solvent Food Hydrocolloid 2019 89 674 681 10.1016/j.foodhyd.2018.11.036 34. Kurek M Garofulić IE Bakić MT Ščetar M Uzelac VD Galić K Development and evaluation of a novel antioxidant and pH indicator film based on chitosan and food waste sources of antioxidants Food Hydrocolloid 2018 84 238 246 10.1016/j.foodhyd.2018.05.050 35. Yong Cho S Park JW Rhee C Properties of laminated films from whey powder and sodium caseinate mixtures and zein layers LWT-Food Sci Technol 2002 35 2 135 139 10.1006/fstl.2001.0826 36. Dong H Ling Z Zhang X Zhang X Ramaswamy S Xu F Smart colorimetric sensing films with high mechanical strength and hydrophobic properties for visual monitoring of shrimp and pork freshness Sensors Actuators B: Chem 2020 309 127752 10.1016/j.snb.2020.127752 37 Cazón P Vázquez M Velazquez G Cellulose-glycerol-polyvinyl alcohol composite films for food packaging: evaluation of water adsorption, mechanical properties, light-barrier properties and transparency Carbohyd Polym 2018 195 432 443 10.1016/j.carbpol.2018.04.120 38. Zarandona I Puertas AI Dueñas MT Guerrero P de la Caba K Assessment of active chitosan films incorporated with gallic acid Food Hydrocolloid 2020 101 105486 10.1016/j.foodhyd.2019.105486 39. Wrolstad RE Durst RW Lee J Tracking color and pigment changes in anthocyanin products Trends Food Sci Technol 2005 16 9 423 428 10.1016/j.tifs.2005.03.019 40. Yong H Liu J Kan J Liu J Active/intelligent packaging films developed by immobilizing anthocyanins from purple sweetpotato and purple cabbage in locust bean gum, chitosan and κ-carrageenan-based matrices Int J Biol Macromol 2022 211 238 248 10.1016/j.ijbiomac.2022.05.046 35561863 41. Naghdi S Rezaei M Abdollahi M A starch-based pH-sensing and ammonia detector film containing betacyanin of paperflower for application in intelligent packaging of fish Int J Biol Macromol 2021 191 161 170 10.1016/j.ijbiomac.2021.09.045 34536478 42. Eskandarabadi SM Mahmoudian M Farah KR Abdali A Nozad E Enayati M Active intelligent packaging film based on ethylene vinyl acetate nanocomposite containing extracted anthocyanin, rosemary extract and ZnO/Fe-MMT nanoparticles Food Packaging Shelf 2019 22 100389 10.1016/j.fpsl.2019.100389 43 Pirsa S Shamusi T Intelligent and active packaging of chicken thigh meat by conducting nano structure cellulose-polypyrrole-ZnO film Mat Sci Eng C-Mater 2019 102 798 809 10.1016/j.msec.2019.02.021

0

PMC9753073

NO-CC CODE

2022-12-16 23:25:28

no

Biomass Convers Biorefin. 2022 Dec 15;:1-14

utf-8

Biomass Convers Biorefin

10.1007/s13399-022-03641-4

oa_other

==== Front AIDS Behav AIDS Behav AIDS and Behavior 1090-7165 1573-3254 Springer US New York 3956 10.1007/s10461-022-03956-3 Original Paper Development of an Economic and Relationship-Strengthening Intervention for Alcohol Drinkers Living with HIV in Malawi http://orcid.org/0000-0002-0609-5077 Conroy Amy A. amy.conroy@ucsf.edu 1 Tebbetts Scott 1 Darbes Lynae A. 2 Hahn Judith A. 3 Neilands Torsten B. 1 McKenna Stacey A. 4 Mulauzi Nancy 5 Mkandawire James 1 Ssewamala Fred M. 6 1 grid.266102.1 0000 0001 2297 6811 Center for AIDS Prevention Studies, Division of Prevention Sciences, Department of Medicine, University of California San Francisco, 550 16th Street, 3rd Floor, San Francisco, CA USA 2 grid.214458.e 0000000086837370 Department of Health Behavior and Biological Sciences, Center for Sexuality and Health Disparities, School of Nursing, University of Michigan, Ann Arbor, MI USA 3 grid.266102.1 0000 0001 2297 6811 Department of Medicine, University of California San Francisco, San Francisco, CA USA 4 Stacey McKenna, LLC, Fort Collins, CO USA 5 Invest in Knowledge, Zomba, Malawi 6 grid.4367.6 0000 0001 2355 7002 Brown School, Washington University, St. Louis, MO USA 15 12 2022 116 6 12 2022 © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Heavy alcohol use among people with HIV in sub-Saharan Africa is driven by household economics such as poverty and unemployment and has negative impacts on couple relationships. Multilevel interventions have the potential to reduce alcohol use and improve relationship outcomes by addressing the web of co-occurring economic, social, and dyadic factors. This objective of this study was to develop an economic and relationship-strengthening intervention for couples in Malawi, consisting of matched savings accounts with financial literacy training and a couples counseling component to build relationship skills. Informed by the ADAPT-ITT framework, we collected multiple rounds of focus group data with key stakeholders and couples to gain input on the concept, session content, and procedures, held team meetings with field staff and an international team of researchers to tailor the intervention to couples in Malawi, and refined the intervention manual and components. The results describe a rigorous adaptation process based on the eight steps of ADAPT-ITT, insights gained from formative data and modifications made, and a description of the final intervention to be evaluated in a pilot randomized clinical trial. The economic and relationship-strengthening intervention shows great promise of being feasible, acceptable, and efficacious for couples affected by HIV and heavy alcohol use in Malawi. Keywords Economic strengthening Couples HIV/AIDS Antiretroviral therapy Sub-Saharan Africa NIAAA027983 Conroy Amy A. ==== Body pmcIntroduction Heavy alcohol use is a major health threat for people with HIV (PWH), described as “adding fuel to the fire” of the HIV epidemic [1]. Rates of heavy drinking are alarmingly high among PWH and may be almost twice that of the general population [2]. Alcohol use directly impacts ART adherence and HIV clinical outcomes, and contributes to malnutrition, liver disease, and HIV disease progression [3–7]. In sub-Saharan Africa, individuals who use alcohol report some of the highest per capita alcohol consumption rates in the world [8], despite that the majority of adults abstain from alcohol use. Economic, Social, and Dyadic Correlates of Alcohol Use in Sub-Saharan Africa The main drivers and consequences of heavy alcohol use in sub-Saharan Africa are economic, social, and dyadic. At the economic level, the daily stress of poverty is inextricably linked with heavy alcohol use [9, 10]. Population-based data from South Africa showed that people living in poverty with few assets have a higher odds of heavy alcohol use than those with some or many assets [11]. Unemployment, boredom, and coping with a stressful, low-wage job have been documented as reasons for heavy alcohol use [12]. At the social level, barriers to reducing alcohol use include peer pressure and the social benefits of drinking; alcohol use is an active part of social life in many African settings, allowing men to form social bonds and express masculinity [12–15]. On the other hand, positive forms of social pressure from relatives, peers, and healthcare providers is a strong motivator to reduce alcohol use [15, 16]. At the dyadic level, primary partners can influence each other’s drinking behaviors, and alcohol use can be both a cause and consequence of relationship distress [17]. In South Africa, heavy alcohol use was positively associated with poor communication and mistrust in couples [18]. In Malawi, a qualitative study found that men living with HIV who drink alcohol experienced challenges with ART adherence; however, a more striking finding was that women living with HIV, who often did not drink themselves, also faced adherence issues as a result of their husband’s drinking [16]. Women described how alcohol-related violence, food insecurity, and a lack of partner support played a role in missing doses of ART [16]. Moreover, alcohol use can indirectly affect ART adherence and overall health by damaging couple relationships needed for support, survival, and well-being [12, 16]. Alcohol use is a trigger of intimate partner violence (IPV), which is one of the strongest predictors of adherence [19]. While alcohol use can negatively impact the health of both partners, partners can also play an important health-promoting role. In South Africa, partners can mitigate the deleterious effects of alcohol use by helping PWH manage alcohol use and maintain adherence to ART even while drinking [20]. Interventions at the Economic, Dyadic, and Social Levels: Need for Multilevel Approach Multilevel approaches are needed to address alcohol use among PWH in sub-Saharan Africa. The majority of behavioral interventions have focused on individuals who use alcohol, using approaches such as cognitive behavioral therapy or motivational interviewing [21]. A recent meta-analysis based on 21 studies found that such interventions were successful at reducing alcohol use, and had additional benefits on adherence [22]. However, only four of the studies took place in sub-Saharan Africa and the results of these studies were mixed. One of the most promising interventions was a cognitive behavioral therapy intervention in Kenya, which showed significant reductions in alcohol use in a pilot study [23] and full efficacy trial [24]. The intervention used paraprofessionals to deliver cognitive behavioral therapy and emphasized locally-salient motivations for reducing drinking, including economic reasons [25]. To complement individual-level interventions, there is a strong need for approaches that modify the broader socio-economic and dyadic context of drinking. Economic-strengthening interventions could have meaningful impacts on alcohol use, especially with couples. Savings-based interventions combined with financial literacy training (FLT) may provide a sustainable option by breaking the cycle of poverty through investments, liquid assets, and lifelong financial knowledge [26–28]. In Uganda, the Suubi intervention (also called Bridges) provided incentivized savings accounts combined with FLT to facilitate savings and investments. The intervention had positive impacts on mental health, family cohesion, and sexual risk-taking among adolescents at risk for HIV, and virologic suppression among adolescents living with HIV [29–31]. Given the economic determinants and consequences of alcohol use in sub-Saharan Africa, interventions like Suubi could adapted to target heavy alcohol use, such as by making the FLT curriculum more relevant to people who drink alcohol. In addition, a focus on couples within saving-based interventions may be particularly effective and synergistic when intervening on the couple relationship and household economics together. Most economic strengthening interventions, including those with micro-savings, have focused solely on empowering women rather than couples [32–34]. By working with both partners together, men may be less likely to view women’s financial activities as a threat to their masculinity and provider roles [35, 36] and respond with control and dominance [32]. Moreover, providing both partners with FLT could reinforce material learned and encourage partners to engage in shared decision-making and take joint responsibility for financial goals. Relationship-strengthening interventions have been effective at reducing alcohol use and addressing HIV-related behaviors. In the US, behavioral couples therapy was more effective than an individual-based approach by addressing the relationship dynamics that intersect with alcohol use [37]. Other studies found that when comparing individuals who drink alcohol heavily, those in the behavioral couples therapy arm reported lower rates of IPV and higher relationship functioning than those in an individual-based arm [38, 39]. In South Africa, one of the few interventions that addressed alcohol use in couples was an HIV prevention intervention known as Couples Health Co-Op [40]. The intervention reinforced couples’ relationships with skill-building exercises around communication and sex. Men in the couples arm were less likely to report heavy alcohol use as compared to the control arm of male-only groups [41]. Another couple-based intervention developed in South Africa, called Uthando Lwethu, which consisted of activities to improve relationship dynamics and build problem-solving and communication skills, was effective at increasing uptake of couples HIV testing and counseling [42, 43]. Conceptual Framework: Integrating Dyadic, Economic, and Social Theory We posit that gaining relationship skills will help couples work together on financial goals and reduce alcohol use, while increasing savings and financial stability will alleviate stress on couples, encourage planning for the future, and reduce drinking, thereby enhancing couple functioning. In this study, we developed a theoretically-based intervention called Mlambe, named after the mlambe (Baobab) tree in Malawi which is a symbol of strength and life, by adapting and integrating two efficacious interventions focused on strengthening household economics and relationship dynamics. Economic-strengthening activities were based on the Suubi intervention in Uganda, which consisted of incentivized savings accounts and FLT [30, 44, 45]. Relationship-strengthening activities were based on the Uthando Lwethu intervention in South Africa, which consisted of relationship education and skills [42, 43]. Mlambe will intervene at the dyadic, economic, and social levels of the social-ecological model [46] to impact alcohol use and HIV clinical outcomes (Fig. 1). At the dyadic level, improving relationship dynamics and communication patterns can help couples move from a self-centered to a relationship-centered orientation in which couples interpret health issues as having significance for the relationship. This in turn enables couples to support each other in positive ways to carry out their shared vision and work collaboratively around health issues. At the economic level, asset theory posits that assets such as education, savings accounts, and an income-generating activity (IGA) can change not only household economic status, but behaviors, attitudes, and hope for the future [47, 48]. Because alcohol use is often used to cope with economic stressors and feelings of hopelessness, asset-building could address some of the underlying reasons for drinking. The accumulation of assets may also reduce economic stressors and its negative impact on couple relationships [49–53], creating new opportunities for couples to engage in healthy behaviors together. At the social level, social learning theory argues that new behaviors can be acquired by observing and reproducing the behaviors of others and through social reinforcement [54]. Thus, by including individuals who use alcohol together in a health-promoting and supportive environment, there will be opportunities to learn the best strategies from others, to develop new friendships, and to receive peer support around alcohol use. Fig. 1 Conceptual Framework for the Mlambe Intervention 1 Study Purpose The purpose of this study was to describe the process of adapting and synthesizing two interventions into a combined economic and relationship-strengthening intervention to reduce alcohol use by employing the ADAPT-ITT framework [55]. Specifically, we describe how we adapted the interventions to address heavy alcohol use, tailored the study procedures and content to the Malawi context, and made modifications to ensure the intervention was relevant for couples. We add to the literature by outlining how to combine interventions operating at different levels of the socio-ecological model. Relationship-focused research grounded in psychology does not typically intervene upon structural or economic factors driving poor couple relationships [52, 53]. Economic approaches, on the other hand, often ignore couple interactions, thereby missing opportunities to harness the power of relationships to overcome and cope with economic constraints at the household level. The ADAPT-ITT framework was originally designed to adapt a single, evidence-based intervention (EBI) for a new purpose, target population, or setting [55], and has been successfully used to adapt HIV-focused behavioral interventions for adolescents, men who have sex with men, and other at-risk populations [56, 57]. A few more recent studies have used ADAPT-ITT to combine interventions based on two or more EBIs [58, 59], but this is a rare application of ADAPT-ITT and there is still little guidance of the process of integrating EBIs. We build on this work by considering additional adaptation elements for combining multiple EBIs such as spacing and length of sessions to accommodate a larger number of sessions from two interventions; cost considerations to counteract greater participant burden across many visits; the ordering of the two intervention components and how to blend intervention content; and skills and training of facilitators needed to deliver two different interventions. Methods and Results Process for Intervention Development Similar to other intervention development studies for alcohol use [60], we carried out all 8 steps of the ADAPT-ITT 8 steps: Assessment, Decision, Administration, Production, Topic Experts, Integration, Training, and Testing [55]. Table 1 provides a description of each ADAPT-ITT step and a summary of how we implemented each step. Table 1 Development of the Mlambe Intervention using the ADAPT-ITT Framework ADAPT-ITT Step Question Answered by Step Methodology Used 1. Assessment Who is the target population and why is alcohol a problem? • Collected and evaluated mixed-methods data with target population of couples who drink alcohol in Malawi. 2. Decision What evidence-based intervention (EBI) is going to be selected and is it going to be adopted or adapted? • Identified need for combined economic and relationship-strengthening intervention. • Reviewed literature for EBIs, consulted experts, and held team meetings. • Made decision to adapt Suubi for economic component and Uthando Lwethu for relationship component. 3. Administration What in the EBI needs to be adapted and how should it be adapted? • Presented intervention concept to academic experts and key stakeholders for feedback. • Explored elements of the interventions to adapt. 4. Production How do you produce a draft of the intervention and document adaptations? • Generated draft of Mlambe manual and curriculum for combined intervention based on existing manuals and materials. Tailored the content for contextual issues in Malawi (literacy issues, common words/phrases) and to specifically address alcohol use. 5. Topic experts Who can help to adapt the EBI? • Conducted focus group discussions with couples and key stakeholders for input. 6. Integration What is going to be included in the adapted EBI that is to be piloted? • Analyzed focus group discussion (FGD) data to further refine the manual and intervention design. 7. Training Who needs to be trained and how? • Generated facilitator training manual and conduct trainings with facilitators. 8. Testing Was the adaptation successful and how did it improve alcohol outcomes? • Developed study procedures and data collection instruments for pilot randomized controlled trial to follow. Assessment Step: Defining the Target Population and Intervention Needs Evidence to support the need for Mlambe came from a four-year, mixed-methods investigation of couples living with HIV in Malawi called the Umodzi M’Banja (UMB) project. Couples were recruited from HIV clinics in the Zomba district and were eligible if married/cohabitating, over age 18, and had at least one partner living with HIV on ART who had disclosed their HIV status; the sample has been described elsewhere [61]. The UMB study showed a strong link between alcohol use, relationship dynamics, and ART adherence using survey data collected with 211 couples [16, 20, 62–64]. First, we found that participants who drank alcohol had a lower odds of self-reported ART adherence (AOR = 0.38; p < 0.05). Second, higher levels of relationship unity (AOR = 2.11), satisfaction (AOR = 3.80), and partner social support (AOR = 1.12) were associated with higher adherence (all p < 0.05). Third, participants who reported higher physical (AOR = 0.72; p < 0.05) and sexual IPV (AOR = 0.72; p < 0.05) had lower odds of adherence. Fourth, drinking alcohol was positively associated with higher levels of physical (β = 0.80; p < 0.001) and emotional IPV (β = 0.48; p < 0.01). Within UMB, we enrolled 23 couples who had a positive Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) screen [65, 66] (score of ≥3 for men and ≥4 for women) in order to identify multilevel barriers and facilitators of alcohol use, and potential intervention options. The sample has been described elsewhere [67]. Men were the primary drinker with few women reporting alcohol use. Wives tried to persuade their partners to reduce their alcohol intake, but were often unsuccessful, citing issues with communication (e.g., not listening, arguing). Effective couple collaboration around alcohol use was constrained by negative peer influence and men’s desire for friendship to cope with economic stressors. Women were primarily concerned about the expense of alcohol and described how alcohol prevented families from meeting basic needs and investing in the future. A major theme among men was the desire for peer groups or an economic intervention to reduce alcohol use. Based on participants’ needs and preferences, we concluded that an intervention should include efforts to improve couple communication around alcohol, economic-strengthening activities for couples, and alcohol support groups [67]. Decision Step: Choosing an Intervention to Adapt We decided to adapt two EBIs, Suubi and Uthando Lwethu, based on our review of the literature, consultation with experts, and formative work pointing to the need for an economic and relationship-strengthening intervention [16, 68]. Through this process, we learned of the drawbacks of microcredit or finance interventions in Africa (e.g., stress of repaying a loan; increased economic vulnerability when defaulting on a loan) [32, 35, 36], and decided to focus on savings instead of microloans or credit. This led us to choose Suubi, a savings-based intervention grounded in asset theory, for the economic component. For the relationship-strengthening component, we decided to adapt Uthando Lwethu. Although designed to address HIV testing rather than alcohol use, we chose this intervention because of its grounding in dyadic theory and personalized focus on couple communication and problem-solving skills—which were identified as intervention targets in our formative work [16, 67]. The combined intervention would consist of the following components: (1) an incentivized joint savings account at a national bank; (2) FLT covering topics such as banking, saving, budgeting, and debt management; (3) financial goal training and linkage with community-based extension workers to help support IGAs focused on agriculture and livestock raising; and (4) relationship skills education plus one-on-one counseling sessions to gain communication and problem-solving skills to help couples develop plans for reducing alcohol use and improving family finances. Administration Step: Identifying Content to be Adapted We held four focus group discussions (FGDs) with thirty key stakeholders to obtain reactions to the intervention concept and identify areas to adapt. Two FGDs were conducted with couples from urban and rural areas (N = 16) who participated in the UMB study. Similar to earlier stages, couples were married, aged 18 or older, had at least one partner with a positive AUDIT-C screen who was living with HIV and on ART. The study PI conducted an FGD with providers in English and a trained research assistant conducted the other FGDs in Chichewa. FGDs lasted between 1 and 2 h and were held in private rooms at clinics or community-based venues. Two FGDs were conducted with HIV care providers (e.g., nurses, clinical officers, clinic volunteers) and traditional leaders such as village chiefs (N = 14), identified through our professional networks. After summarizing the proposed activities based on Suubi and Uthando Lwethu, stakeholders were asked questions around the couple-based approach, whether session topics would be useful for couples and why, and what additional content should be included or modified. The investigative team also wanted feedback on whether the banking approach used in Suubi would work in rural Malawi given long distances to banks and whether alternatives like mobile money (i.e., mobile phone banking using airtime as a currency) would be more acceptable. Finally, the team wanted to explore whether the matched incentive should be done as a group, versus at the individual level, to capitalize on peer support to reduce alcohol use. See Table 2 for other topics. Audio files were transcribed, translated into English, and coded by the study PI using deductive codes derived from the interview guide and inductive codes that emerged from the data. Table 2 What Needs to be Adapted? Stakeholder Findings from Administration Step of ADAPT-ITT Key Questions Asked Perceptions from Key Stakeholders Conclusions and Modifications needed What are the challenges around saving, banking, starting a business, and creating a budget? • People cannot save because of poverty. • People need a business to save. • People need capital (e.g., loans) to improve their households. • Villagers do not know how to make a budget. Address myths that people cannot save into financial literacy sessions. Retain sessions on savings, budgeting, and debts. Will couples be able to work together on saving money and starting a business? • Men and women often do not communicate around household purchases; men “waste” money because they do not know household needs but sometimes women use the money for frivolous purchases. This causes conflict. Learning how to create a household budget as a couple will resolve conflict from one member making decisions independently. What are the beliefs around banks, mobile money, microfinance, and loans that we need to address? • Banks are not trustworthy; money is taken out for fees that people do not understand. There is no point in saving with banks. • It is hard for the average person to open a bank account. • Banks are for wealthy people. • Village bank loans for businesses can help, but a husband could take the money and “spend it on concubines and beer”. • Mobile money banking makes money too accessible which could be used for alcohol versus depositing money in bank account. Address banking myths in financial sessions. Involve bank representatives. Retain focus on formal banking approach. Should we offer the matched incentive as a group-based match or couple-based match? • Match should be based on what each couple can save, not the group: “everyone should reap what they sow” Use couple-based (not group-based) matched incentive How to ensure sobriety at sessions? • If couples are informed about rules to attend sober, they will comply. • There could be penalties for coming drunk to sessions. Establish ground rules around sobriety. Should men and women be separated or stay as a couple? • “There should be no secrets between husbands and wives” and thus they should attend sessions together. • Families should budget together so burden does not fall on women. • Couples should be educated together on alcohol and finances so that they are on the same page. Do not separate men and women; retain focus on couples. Will couples be open to learning communication skills? • Arguments stem from couples not agreeing or respecting each other. • Counselling should be done with each couple one-on-one so couples’ privacy is respected and couples feel open to participate together. Retain couples counseling on communication skills. Should be one-on-one with each couple, not in a group. How can we ensure privacy and that people feel comfortable with participating? • “HIV is no longer shameful as it was. There are worse diseases.” • People will feel free talking about alcohol in a group. Group format for other sessions is acceptable. Overall, there was strong support that the intervention would be feasible and acceptable for the target population (see Table 2). Most importantly, participants confirmed the appropriateness of a couple-based approach. When asked if partners should be together or separated for the intervention sessions, the consensus (across all FGDs) was that couples should be intervened upon together. The groups provided several reasons for why a couple-based approach is preferred: because the couple is married/family, to help them agree how to spend money, to prevent one partner from misusing financial resources, to reinforce financial lessons, and to reduce the burden on one partner for making financial decisions. For example, one healthcare provider described how women should be more involved in financial decision-making through a couple-based approach: “Men have the financial mastery in most families. But I believe that this way [the couple approach] can encourage the wife or partner to start taking part in the finances in the home. Because we are trying to make them work together.” All other focus groups had similar views about the importance of involving both partners together on finances. The sessions on relationship skills were deemed equally valuable by all groups, which were believed to help couples avoid disagreements, “enhance love in families by encouraging tolerance of each other”, and to communicate their opinions to one another. One female participant in an FGD with couples said: “This is helpful because if we say family, it means the two, being one flesh. It is also good to know what your spouse likes than to tell your husband what you would like them to do for you. You should agree as husband and wife. This suggestion [to include a session on relationship skills] is good and it may have good results.” The FLT sessions on savings, budgeting, and debt management were also deemed valuable overall, and several suggestions were made to address local beliefs such as the belief that banks were only for rich people and to obtain buy-in from participants. Mobile money (i.e., storing money in the form of airtime instead of in a formal bank account) was viewed as more convenient, but also perceived as problematic for people who use alcohol since cash could be easily accessed. All four groups believed that the savings match should be at the couple/family level, not group level, so that couples individually benefit based on what they contribute. They agreed that most sessions could be group-based, while others such as the couple counseling sessions should be done individually with each couple for privacy reasons. Overall, the findings confirmed our broader approach and provided input on areas under debate with the design (see Table 2). Production Step: Drafting the Intervention Manual Through an iterative process led by the PI, the study team in the US created an outline of intervention sessions, compiled materials from the interventions to adapt, and generated a first draft of the Mlambe manual. The manual included a module on the control arm, which consisted of brief alcohol advice lasting 10–15 min modelled off WHO recommendations [69]. Brief alcohol advice would be delivered immediately following a randomization ceremony with each couple individually using alcohol messages personalized to their AUDIT-C score. For the intervention, we decided on a total of 10 monthly sessions to allow couples sufficient time to save for their IGA over 10 months. We included four sessions on FLT from Suubi (e.g., savings and banking, budgeting, debt management) with a fifth session to provide couples with goal-specific training with community-based extension workers on their selected IGA. We combined these five sessions with three sessions from Uthando Lwethu (i.e., one group session on general relationship dynamics and two couples counseling sessions on communication and problem-solving skills). In Uthando Lwethu, couples are asked to choose an HIV-related topic to practice their communication and relationship skills and we adapted the practice exercises to focus on issues related to alcohol use, savings, and ART adherence. We also added an introductory session to explain the Mlambe intervention and introduce couples to banking so that they had basic knowledge and skills to get started using their bank accounts immediately. Because we posited that couples could save by reducing spending on alcohol, we added a session to understand the harms of alcohol use on health (e.g., including HIV) and relationships, and develop strategies to reduce based on the Indashyikiwa intervention from Rwanda with couples [70]. See Table 3 for an outline of sessions. With the exception of two counseling sessions, all sessions were conducted in groups based on the recruitment block at randomization. Couples would attend sessions with the same cohort across the 10-month period, unless making up sessions in which case they could join another group. Table 3 Outline of Mlambe Intervention Sessions Number Session Title Adapted from Main Activities 1 Introduction to Mlambe, savings and banking, and mobile money Suubi • Building group rapport • Introduction to Mlambe and matched savings • Introduction to bank services and mobile money • Establishing couple banking agreement 2 Understanding and reducing alcohol use Indashyikirwa • Understanding alcohol harms and impacts on relationships • Strategies for reducing alcohol use 3 Relationship dynamics and power Uthando Lwethu Indashyikirwa • Positive aspects of relationships • Appreciating our partners • Expressing heartfelt support • Types of power, balancing economic power 4 Couples counseling session 1: Introduction to communication and problem-solving skills Uthando Lwethu • Learning the Initiator-Receiver technique • Exploring expectations in the relationship • Introduction to problem-solving and goal setting 5 Couples counseling session 2: Working together on alcohol issues Uthando Lwethu • Practicing Initiator-Receiver technique • Plan going forward • Goal setting 6 Bank services Suubi • Perceptions of banks, dispelling myths • Benefits of using a bank • Planning financial goals 7 Budgeting and spending wisely Suubi • Identifying expenses and sources of income • Developing a financial plan • Creating a budget • Identifying ways to cut spending • Planning financial goals 8 Savings, asset building, and asset accumulation Suubi • Introduction to savings • Building skills to save • Planning financial goals 9 Debt management Suubi • Steps in borrowing money • Managing debt and responsible borrowing • Planning financial goals 10 Goal-specific financial support Suubi • Introduction of Mlambe to extension workers • Overview of extension workers areas of expertise • One-on-one meetings with couples and extension worker to plan for income-generating activity The ordering of sessions was determined as follows. We hypothesized that couple communication skills would be needed first so that couples could work collaboratively together on finances and alcohol reduction from the start. Couples would receive an overview of banking in session 1 and alcohol use in session 2, then worked on building their relationships in sessions 3–5 and received FLT in sessions 6–9. During the financial sessions, couples would start planning out the type of IGA that their family will invest in, such as raising cows, pigs, or growing produce to sell at the market. In the last session, community-based extension workers (i.e., government employees who work in communities with training on livestock raising and agriculture) would be invited to learn about the Mlambe study and provide an overview of their knowledge and services to the group. Couples would then be matched with extension workers based on where they live and the type of IGA chosen. Thereafter, the extension workers would meet with couples in their communities to provide further education and assistance on starting their IGA based on their financial goals and circumstances (e.g., advantages and disadvantages of a starting a piggery business over raising goats). At the first session attended by bank representatives, couples open a joint bank account and then save money every month by making deposits into their bank account. Couples are eligible for a 1:1 match up to $10 per month for each Malawi Kwacha saved. The matched contribution is contingent upon authorized withdrawals for education, business, or medical expenses, which were verified with receipts. Otherwise, unauthorized expenses were deducted from a couple’s matched contribution for that month. To encourage attendance at sessions, couples would need to attend at least 7 out of 10 sessions to receive the matched component at the end of the study but could make up sessions if missed. At the end of each month, bank statements are requested and reviewed and the matched contribution is calculated, tracked in a ledger, and then allocated in separate project bank account containing the matched contribution for all participants. At the end of the 10-month intervention period, participants coordinate with the project staff to make purchases for their family IGA using their individual savings plus matched component. Couples continue working with their extension workers to learn about the most appropriate IGA for their circumstances and financial goals until they are ready to commit to an IGA and invest their savings, which can be done up until the 15-month follow-up visit. After outlining activities for each session and generating an initial draft of the manual, we tailored the content for couples and incorporated alcohol themes into the activities and examples. For example, in the counseling sessions, practice exercises were revised such that couples could apply their communication skills to issues either alcohol use, savings, or HIV. In the FLT sessions, illustrations and examples were tailored to couples who drink alcohol. For example, in the banking session, there is a story of a woman who instead of putting her money into a bank, kept it in a hiding place at home and then finds out her husband spent the money on alcohol. In the budgeting session, we added an exercise to calculate how much money would be saved per month if alcohol use was reduced (see Fig. 2). We focused adaptations on spurring change around alcohol use rather than HIV-related behaviors because other alcohol interventions have been less successful at addressing alcohol use when intervening on multiple behaviors at once [22] and we believed that improved HIV treatment behaviors would follow from reduced alcohol use. We also believed that addressing HIV behaviors in addition to alcohol use and savings could be overly burdensome and further lengthen an already long intervention. Moreover, in our prior work, participants, especially wives, expressed greater concerns around alcohol use than HIV [68]. Fig. 2 Example of Adapted Financial Literacy Content to Address Alcohol Use in Malawi 2 Topic Experts Step: Obtaining Input on the Intervention To gain further input on the intervention, we held six FGDs with six couples (2 FGDs) and 22 key stakeholders (4 FDGs) consisting of HIV care providers, religious and community leaders, bank, microfinance, and mobile money representatives, alcohol vendors, for a total of 34 participants. We included alcohol vendors who work in bars, bottle shops, brew beer, or make Kachasu (a locally made spirit) because they have important insight into whether the intervention concept would be acceptable, feasible, and appropriate given their direct observations of drinking patterns and behaviors, and the harms of drinking through their regular interactions with patrons. Couples were identified from the UMB study using the same eligibility criteria as above. Key stakeholders were recruited through our professional networks. A trained research assistant conducted the FGDs, which lasted 60–90 min and were conducted in private rooms at clinics or community-based venues. The FGD guide was divided into 12 sections on core features of the intervention (e.g., matched savings accounts), planned intervention sessions from Suubi (e.g., bank services, budgeting, savings, asset building, and asset accumulation) and Uthando Lwethu (e.g., relationship dynamics and power, one-on-one couples counseling sessions), and practical aspects of intervention delivery (e.g., how to tailor materials to low literacy, how to increase attendance and participation). Participants were presented with an overview of each session and then asked a series of questions about what they liked/disliked, what was confusing or unclear, and suggestions for improvements. FGDs were audio-recorded, translated from Chichewa into English, and transcribed into electronic format. FGD transcripts were coded line-by-line by one of the study authors with qualitative expertise. Using an iterative process, themes were identified from the coded transcripts around the financial, relationship, and alcohol components of the intervention. Three patterns emerged within the themes around likes, concerns, and recommendations. Within these categories, the main themes were selected based on frequency in which they came up across and within the different groups and by their intensity (i.e., not necessarily common, but important or unexpected) [71]. We then looked for agreements and disagreements within and across the different types of stakeholders and weighted the significance of stakeholder recommendations based on their relevant experience of a topic (e.g., bankers would know more about banking than care providers). Overall, FGD participants expressed many positive sentiments for the intervention. Similar to the first FGDs, and all groups highlighted their appreciation for the couple-based approach (e.g., doing activities together will reinforce the material and help them be more transparent around spending money) (see Table 4). One of the FGDs with couples noted that involving husbands in the FLT sessions could be a challenge since this usually falls within the women’s domain, but they believed it was very beneficial to work with both partners to help avoid quarrels over money. Most groups really liked the financial session on budgeting and found it to be highly relevant for most families, and also believed the session on loans and debt management was important given the many high-interest loans in these communities that could bring financial hardship. Participants were very supportive of the matched savings approach and thought this would be a strong incentive to save and reduce alcohol, although some worried about whether participants would borrow from family or friends to qualify for the match. There were differences in opinions regarding the acceptability of formal banks, with some groups mentioning that it was time-consuming, confusing/difficult to navigate, unexpected fees, and transport could be expensive; while others noted that banks were more secure than village savings and loan programs and other methods to save money in homes where it could be stolen or used to purchase beer. It was noted that with proper education and training, the challenges of using banks for rural families could be overcome. Table 4 Who Can Help to Adapt the Intervention? Stakeholder Findings from Topic Experts Step of ADAPT-ITT Stakeholder Likes Concerns Recommendations Modifications made Alcohol vendors • Sessions as a couple • Financial literacy and matched savings • Exercises to reduce drinking • Communication skills could be hard to learn • Privacy/safety of funds • Banks are time-consuming, challenges with deposits and withdrawals • Include protections for a joint bank account so that one partner cannot steal • Include advice not to bring much money to the beer halls to help curb drinking • Emphasize confidentiality in group sessions • Address issues around transport to bank 1. Added couple financial agreement on bank withdrawals to first session 2. Tailored alcohol content to provide tips from alcohol vendors 3. Added confidential agreement to first session 4. Included mobile banking agents in rural areas to reduce transport costs 5. Mobile money providers are now invited to first session to provide education on services 6. Meal is provided at each group session 7. To address difficulties travelling to banks, we included village bank agents to assist couples in making deposits and withdrawals 8. Facilitators will have formal education and authority to build trust (e.g., teacher or counselor). 9. Included training on bank loans HIV care providers • Teaching people how to save and bank independently • Exercises and vignettes around alcohol use • Counseling sessions and relationship activities • People will borrow to get the match • Difficulties accessing bank services in rural areas • Sessions could be too long • Include mobile money providers in financial education • Discourage borrowing from friends to get match • Address food insecurity • Address issues around transport to bank Religious and community leaders • Match will be a strong incentive • Debt management and loan content • Understanding harms of alcohol • Focus on positive aspect of relationships • Working with couples and involve husbands as saving as often falls on wife • Concerns about whether people can save in rural areas • People will borrow to get match • Transport to banks is an issue • Banking fees are confusing, missing money from accounts • Mobile money may be preferred over banks • Recommend bank accounts require dual signatures for withdrawals • Include mobile money representatives • Facilitators should be experts on financial issues • Emphasize confidentiality in group sessions Financial representatives • Saving as a couple promotes transparency • Financial skills will give partners tools to solve problems together • Will promote a savings culture • Budget session is important for identifying monthly expenses • Groups will help couples learn from each other • Concerns about whether people can save in rural areas • Confidentiality in group sessions • People may not be able to save and will borrow from friends • Mobile banking agents can help • Bank representations should do financial sessions • Divorce rates are high; protections are needed so one partner does not take all the savings and leave • Encourage people to save and not take out high interest loans or borrow from friends Couples • Saving will bring couples closer • Difficult to navigate banking system so this will help • Match incentives • Exercises on alcohol reduction • Focus on positive aspects of relationship • Will help with school fees, a big challenge • Good to learn financial skills together (especially men) • People can learn from other in groups • Couple communication exercises should not focus on one partner’s wrongdoings • Financial literacy sessions could be difficult to understand • Men may struggle more, since budgeting often falls on wife • Some people fear the banks • Will need assistance with opening accounts and banking • Facilitators should be kind and non-judgmental • Include mobile money providers • Need to promote confidentiality groups • Include a meal at sessions Participants had several suggestions of ways to improve the intervention: (1) incorporating mobile money providers to round out the FLT; (2) ensuring group sessions were private and information was not shared outside the group; (3) putting protections in place so that one partner could not withdraw all the savings and divorce the other or spend the money on alcohol; (4) addressing the issue regarding transportation to town centers where banks are located; (5) ensuring that intervention facilitators were knowledgeable, respectful, and non-judgmental; and (6) providing a meal at sessions to increase attention and motivation for attendance (see Table 4 for more details). Integration Step: Incorporating Input to Refine the Intervention Based on the topic expert findings, we made the following modifications to the intervention. Rather than enforce strict rules around withdrawals, we empowered couples to negotiate the terms of banking and added a couple financial agreement on bank withdrawals that is negotiated between both partners at the first session to determine who can make bank withdrawals and under what conditions. We involved mobile money providers in the financial sessions as recommended by the FGD participants so that participants can learn about these services and could use mobile money as a channel (but not to replace the formal banking system) to make deposits and withdrawals into bank accounts. Our team’s prior work in Uganda with Suubi found that linking individuals into the formal banking system with bank accounts is important not only for ensuring the money is safe and can earn interest, but also is a step towards accessing a range of banking products that can benefit people. Thus, we maintained our original approach. We incorporated the use of local bank agents (employed by the formal bank) who could assist clients with banking in their villages to avoid transport to town banks. We also added a confidentiality agreement for all participants to pledge at the first session to protect the information of other group members as much as possible, given that we could not guarantee confidentiality in a group format. Given the long distances traveled by couples to attend the sessions, high levels of food insecurity, and to encourage participation over multiple sessions, we provided a full meal at each group sessions with a lunch activity relevant to the session material on a given day. Finally, we hired facilitators with counseling and/or educational backgrounds who were able to explain complex information on FLT to participants with low education levels (see Table 4). Rather than employ community extension workers to deliver the intervention as done in Suubi, it was noted by FGD participants that facilitators should have the education and skills to gain the respect of participants and be viewed as models of financial success (i.e., community workers were often poor themselves). We also believed that higher education of the facilitators was important to be able to train facilitator to lead sessions requiring different skill sets (e.g., financial literacy and couples counselling). Training Step: Training Staff to Deliver the Intervention Intervention facilitators, data collectors, and the research manager were trained over a six-week period by the US study investigators and research coordinator on the research objectives, study procedures, delivery of the intervention sessions and control arm, general facilitation skills, and couples counseling skills. Because the training started in May 2021 when international travel was not possible due to COVID-19 pandemic, we developed a training curriculum that could be delivered remotely using a triad of complementary videos, independent readings, and practical activities to reinforce content. Instructional videos were created in Loom® for each session and could be re-watched as needed by clicking on a web link. Because we were adapting two interventions from Uganda and South Africa, we also held conference calls between the Mlambe investigators and the other study teams in both countries to better understand intervention delivery, challenges on the ground, and implementation procedures, which were recorded and stored, transferred into a link using Loom®, and then used to train the facilitators. Daily conference calls were held between the US research team and the Malawi team to reinforce training material, answer questions on content learned each day, and provide further guidance. Facilitators also conducted mock sessions and a mock randomization ceremony with other study staff to gain practical skills and recorded short videos for the investigators to review and provide further training. While the economic sessions were highly structured and more straightforward to deliver, the couples counseling sessions required more training, especially for the Initiator-Receiver technique, and thus it took several rounds of videos with feedback until the team agreed the facilitators were proficient in their couples counseling skills. Audio recordings were taken of all intervention sessions; however, given the multiple hours of the sessions and cost and time associated with translation and transcription of large audio files, and because most sessions were highly scripted and structured, we opted to assess intervention fidelity using checklists after each session to confirm that all activities were completed. Weekly calls were also held between the facilitators, project managers, and investigative team in which facilitators presented on each session and explained how activities were received by participants, which helped to ensure the comprehension of the manual, spark conversations around field challenges, and ensure the intervention was being delivered as intended. The research manager attended most intervention sessions to further ensure fidelity to the intervention manual and to assist with session activities and help couples with literacy challenges. Once the investigative team was able to travel again, the study PI, co-investigator, and the Suubi team from Uganda provided additional training on the FLT components and attended intervention sessions to observe operations and provide additional feedback such as giving stretching breaks, encouraging couples to consume refreshments instead of saving for family members, arranging the chairs in the room to foster better participation, and encouraging more engagement among couples. Pilot Study Step: Testing and Evaluating the Intervention For the final ADAPT-ITT step, we developed the standard operating procedures (SOPs) for the pilot clinical trial to assess the feasibility and acceptability of the intervention. SOPs covered all study procedures such as informed consent process, ethical procedures, adverse event reporting, monitoring for intimate partner violence, COVID-19 risk reduction, recruitment and enrollment, randomization process, survey data collection, use of REDCap, tracking bank savings and matched savings component, and blood collection and laboratory testing. SOPs were developed through an iterative process between the study investigators and research managers in the US and our implementing partner in Malawi, and were further refined as questions came up during the training in Malawi and during initial rollout of the study. During the roll-out and pilot testing of the intervention, several implementation challenges came up. It was discovered that couples needed a national identification card to establish a bank account which some did not have. Thus, study procedures were modified such that upon enrollment, the facilitators helped couples obtain documentation needed to apply for ID cards in time for the first session. There were also some challenges with making deposits using village banking agents in which couples deposited the money into the wrong bank account, which led to mistrust and confusion when account balances did not reflect the deposits. The facilitators had to troubleshoot banking issues with couples and work with the bank to correct transaction errors and then re-train couples on entering their account information properly. In another scenario, a husband did not make the deposit as he told his wife and spent the money on other purchases that they did not agree on, which angered the wife when she discovered his lie. Thus, the husband had violated their financial agreement. The facilitators worked with the couple on the misunderstanding, reminding them to use the Initiator-Receiver technique to communicate collaboratively on how deposits should be made going forward. Thus, the communicate techniques became a tool that was used throughout the intervention when disagreements arose between partners, not just during the counselling sessions. In a few cases, the husband sold the phone provided by the study which angered the wife who was committed to the intervention. Couples were not given a second phone in this case and the facilitators counselled the couple to decide whether they wanted to continue in the study (which they ultimately did). Finally, while we had hoped that the group sessions would encourage social cohesion and social support, learning in Malawi is often very top-down and instructive and thus we had to re-train the facilitators to provide FLT in a more engaging way and by encouraging couples to participate in group exercises and encourage more active learning. We provided a lunch after the sessions which we had hoped could also be a time for couples to get to know each other, however, given the extreme poverty in this setting, couples often did not eat the meal and took it home with them to share with family members. Facilitators had to encourage couples to eat the meal so that they could be attentive and reap the full benefits of the FLT. Discussion With the growing number of intervention development and pilot studies conducted prior to a full-scale efficacy trial, the process of adapting an intervention is rarely documented and even less is known about how to adapt multiple EBIs into a combined intervention. Through an iterative process, we collected multiple rounds of key stakeholder data, as well as input from the investigative team and field team in Malawi, to obtain feedback on the initial concept, intervention components and sessions, and intervention procedures. This rigorous process produced a structured intervention manual, suite of SOPs, and training plan with videos and a format that can be delivered remotely, in preparation for a future trial of Mlambe to establish efficacy. This research provides a practical illustration of the types of issues to consider when combining multiple interventions and intervening at multiple levels of the socio-ecological framework. Economic and relationship-strengthening interventions that have worked elsewhere, even within sub-Saharan Africa, must be carefully tailored to the local setting before scale-up. This was important given that Malawi is resource-poor country in sub-Saharan Africa, and even in comparison to other African settings. Our formative work required that we ensure that the same banking, savings, and income-generation elements used in the parent intervention from Uganda could be implemented in rural Malawi. We found that mobile phones and use of local banking agents in the rural areas were essential for couples to be able to make bank account deposits given that transport to the physical banks can be cost-prohibitive. There were additional contextual issues that had to be considered, particularly when working with couples, such as the high rates of divorce in this setting and concerns that one partner could withdraw savings from the joint banking account and exit the marriage or spend the money on alcohol. Thus, we incorporated a couple financial agreement and taught couples the skills to negotiate the terms of their bank accounts as a preventative measure. When developing multilevel interventions using approaches from different academic fields, in our case, economic development and relationship science, it is important to consider the synergies between the two approaches and additions that may be needed to offset the strengths and weakness of each. During the implementation phase, there were instances of couples who violated their financial agreements or made financial decisions without involving the other partner, or had other disagreements, and thus the facilitators needed to remind couples of how to resolve their issues using learned communication skills. This reinforced our decision to augment FLT with relationship skills and to have the couple communication sessions earlier on so that couples had the tools to communicate when financial disagreements transpired. Other lessons learned when combining EBIs include the importance of hiring facilitators with diverse skill sets that are trusted sources of information or with credentials that allow for training on multiple areas such as counseling and FLT. The adaptation process demonstrated that while some modifications were necessary, the underlying interventions showed promise of being transferrable and implementable in an entirely different African setting and population, even in rural settings like Malawi where access for formal banks can be challenging. While we do not report on participants’ socio-economic status for our pilot study because it is currently ongoing, we expect based on our prior studies that couples will have lower education levels, high food insecurity, and low employment rates [16, 61, 68, 72]. Other modifications would be needed for populations with higher income and employment levels such as, increasing the matched contribution (i.e., higher than $10 USD per month) or changing the types of investments allowed at the end of the study if participants already have assets. Because our participants are underemployed, it was feasible to schedule and conduct group sessions; thus, couples who are formally employed, may not be able to attend sessions during business hours and may have less flexibility, which could require one-on-one sessions. Finally, we worked with married couples living together in the same household which made it more feasible for couples to support each other around alcohol, attend sessions together, and conduct banking and business activities together. Couples from settings with low marriage or cohabitation rates could face more difficulties with coordination of intervention activities and attending sessions and may require more support. Strengths and Limitations The strength of our approach was that we used an established adaptation framework and conducted a rigorous study that gathered local input at all stages of the process. While our data suggest that our initial approach may be feasible and acceptable, it was critical to conduct this formative research with a different study population, health behaviors (i.e., heavy alcohol use), and socio-cultural context, as compared to the original interventions. An advantage of using this process is that it maximizes the likelihood that the final intervention will be successful and will allow us to focus our attention on demonstrating efficacy versus implementation in the next steps. No study is without weaknesses and our study is no exception. As with any method, there is the potential for social desirability bias in which respondents, and even field staff, provide more socially acceptable responses when asked for input on the intervention concept and details of the sessions. Couples and key stakeholders were generally supportive and optimistic of the approach to intervene with both members as a couple but could have been providing more favorable responses to the study investigators. This concern is mitigated, however, by multiple rounds of focus groups, both of which revealed the same theme: intervening with couples has many different benefits and would be well-received and beneficial for families. Our finding is consistent with reports from other studies that participants find couple-based approaches to be enjoyable and valuable experiences for couples to connect [40, 73, 74]. We were also limited by the COVID-19 pandemic in our ability to train staff in-person and oversee study implementation through more frequent field site visits, which could have established better rapport within the team. However, the PI and field staff in Malawi had a long history of collaboration through prior studies and the research manager in Malawi was highly skilled in overseeing study activities and maintained regular communication with the US-based team. The team was in constant communication with daily video calls followed by weekly or monthly calls as activities progressed. This learning experience also resulted in a training curriculum that could be delivered remotely during upcoming COVID-19 surges which may prohibit travel. Conclusion Using a proven adaptation model, we synthesized and adapted two efficacious interventions through a collaborative process that actively involved members of the target population, diverse key stakeholders, implementers, and our colleagues in the US, Malawi, Uganda, and South Africa. We learned that it takes a global village to deliver a behavioral intervention at multiple levels and to adapt what worked in other African countries to Malawi. Although the ultimate success of our adaptation will be determined after the trial, we believe this process resulted in a culturally-informed intervention that shows great promise of addressing important multilevel determinants of heavy alcohol use and reducing the harms of alcohol on people with HIV. Our approach is innovative by combining programming from the fields of economics and relationship science to strengthen couple relationships from the ground up, symbolic of the Mlambe tree in Malawi. The process described in this paper provides a practical example for global health researchers working in resource-poor settings around the world on how to develop and implement a complex, multi-component intervention that is feasible and acceptable for a target population. Author Contributions AC led the conceptualization and design of this study, led the analysis, and drafted this manuscript. LD, JH, TN, JM, and FS conceptualized the study and edited the manuscript. SM assisted with the analysis and edited the manuscript. ST, NM, and JM assisted with data collection and edited the manuscript. All authors contributed to the interpretation of findings and approved this manuscript. Funding This study was funded by the National Institutes of Health under grants R34-AA027983. Data Availability Not available due to the potential to identify participants. Code Availability Not applicable. Declarations Conflict of interest There are no conflicts of interest for any of the study authors. Ethical Approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This study was approved by the UCSF Human Research Protection Program (HRPP) and the National Health Sciences Research Committee (NHSRC) in Malawi. Informed Consent Informed consent was obtained from all individual participants included in the study. Consent for publication All authors approve the publication of this manuscript. 1 ESA = Economic Strengthening Activities; RSA = Relationship Strengthening Activities 2 Notes: Kachasu is a locally-made distilled spirit of high alcohol content. MK = Malawi Kwacha Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Hahn JA Woolf-King SE Muyindike W Adding fuel to the fire: alcohol’s effect on the HIV epidemic in Sub-Saharan Africa Curr HIV/AIDS Rep 2011 8 3 172 10.1007/s11904-011-0088-2 21713433 2. Schneider M Chersich M Temmerman M Degomme O Parry CD The impact of alcohol on HIV prevention and treatment for South Africans in primary healthcare Curationis 2014 37 1 1 8 10.4102/curationis.v37i1.1137 3. Hahn JA Samet JH Alcohol and HIV disease progression: weighing the evidence Curr HIV/AIDS Rep 2010 7 4 226 33 10.1007/s11904-010-0060-6 20814765 4. Williams EC, Hahn JA, Saitz R, et al. Alcohol use and human immunodeficiency virus (HIV) infection: current knowledge, implications, and future directions. Alcoholism: Clinical and Experimental Research. 2016;40(10):2056-72. 5. Hendershot CS Stoner SA Pantalone DW Simoni JM Alcohol use and antiretroviral adherence: review and meta-analysis J Acquir Immune Defic Syndr 2009 52 2 180 202 10.1097/QAI.0b013e3181b18b6e 19668086 6. Salmon-Ceron D Lewden C Morlat P Liver disease as a major cause of death among HIV infected patients: role of hepatitis C and B viruses and alcohol J Hepatol 2005 42 6 799 805 10.1016/j.jhep.2005.01.022 15973779 7. Braithwaite RS Bryant KJ Influence of alcohol consumption on adherence to and toxicity of antiretroviral therapy and survival Alcohol Res Health 2010 33 3 280 23584069 8. World Health Organization Global status report on alcohol and health 2014 Geneva World Health Organization 9. Kalichman SC Simbayi LC Kagee A Associations of poverty, substance use, and HIV transmission risk behaviors in three south african communities Soc Sci Med 2006 62 7 1641 9 10.1016/j.socscimed.2005.08.021 16213078 10. Dunkle KL Jewkes RK Brown HC Transactional sex among women in Soweto, South Africa: prevalence, risk factors and association with HIV infection Soc Sci Med 2004 59 1581 92 10.1016/j.socscimed.2004.02.003 15279917 11. Parry CD, Plüddemann A, Steyn K, et al Alcohol use in South Africa: findings from the first Demographic and Health Survey (1998). J Stud Alcohol. 2005;66(1):91 – 7. 12. Setlalentoa B Pisa P Thekisho G Ryke E Loots Du T The social aspects of alcohol misuse/abuse in South Africa South Afr J Clin Nutr 2010 23 sup2 11 5 10.1080/16070658.2010.11734296 13. Morojele NK Kachieng’a MA Mokoko E Alcohol use and sexual behaviour among risky drinkers and bar and shebeen patrons in Gauteng province, South Africa Soc Sci Med 2006 62 1 217 27 10.1016/j.socscimed.2005.05.031 16054281 14. Woolf-King SE Maisto SA Alcohol use and high-risk sexual behavior in Sub-Saharan Africa: a narrative review Arch Sex Behav 2011 40 1 17 42 10.1007/s10508-009-9516-4 19705274 15. Sundararajan R Wyatt MA Woolf-King S Qualitative study of changes in alcohol use among HIV-infected adults entering care and treatment for HIV/AIDS in rural southwest Uganda AIDS Behav 2015 19 4 732 41 10.1007/s10461-014-0918-5 25323678 16. Conroy AA McKenna SA Ruark A Couple interdependence impacts alcohol use and adherence to antiretroviral therapy in Malawi AIDS Behav 2019 23 1 201 10 10.1007/s10461-018-2275-2 30218319 17. Rodriguez LM Neighbors C Knee CR Problematic alcohol use and marital distress: an interdependence theory perspective Addict Res Theory 2014 22 4 294 312 10.3109/16066359.2013.841890 18. Woolf-King SE, Conroy AA, Fritz K, et al. Alcohol use and relationship quality among south african couples. Subst Use Misuse. 2018:1–10. 19. Hatcher AM Smout EM Turan JM Christofides N Stocki H Intimate partner violence and engagement in HIV care and treatment among women: a systematic review and meta-analysis AIDS 2015 29 16 2183 94 10.1097/QAD.0000000000000842 26353027 20. Conroy AA McKenna SA Leddy A “If she is Drunk, I don’t want her to take it”: Partner Beliefs and Influence on Use of Alcohol and antiretroviral therapy in south african couples AIDS Behav 2017 21 7 1885 91 10.1007/s10461-017-1697-6 28150121 21. Brown JL DeMartini KS Sales JM Swartzendruber AL DiClemente RJ Interventions to reduce alcohol use among HIV-infected individuals: a review and critique of the literature Curr HIV/AIDS Rep 2013 10 4 356 70 10.1007/s11904-013-0174-8 23990322 22. Scott-Sheldon LA Carey KB Johnson BT Carey MP Team MR Behavioral interventions targeting alcohol use among people living with HIV/AIDS: a systematic review and meta-analysis AIDS Behav 2017 21 2 126 43 10.1007/s10461-017-1886-3 28831609 23. Papas RK Sidle JE Gakinya BN Treatment outcomes of a stage 1 cognitive–behavioral trial to reduce alcohol use among human immunodeficiency virus-infected out‐patients in western Kenya Addiction 2011 106 12 2156 66 10.1111/j.1360-0443.2011.03518.x 21631622 24. Papas RK Gakinya BN Mwaniki MM A randomized clinical trial of a group cognitive–behavioral therapy to reduce alcohol use among human immunodeficiency virus-infected outpatients in western Kenya Addiction 2021 116 2 305 18 10.1111/add.15112 32422685 25. Papas RK Sidle JE Martino S Systematic cultural adaptation of cognitive-behavioral therapy to reduce alcohol use among HIV-infected outpatients in western Kenya AIDS Behav 2010 14 3 669 78 10.1007/s10461-009-9647-6 19967441 26. Swann M Economic strengthening for retention in HIV care and adherence to antiretroviral therapy: a review of the evidence AIDS Care 2018 30 sup3 85 98 10.1080/09540121.2018.1476665 29985055 27. Swann M Economic strengthening for HIV prevention and risk reduction: a review of the evidence AIDS Care 2018 30 sup3 37 84 10.1080/09540121.2018.1479029 30235937 28. Allen H Panetta D Savings groups: what are they 2010 Washington DC SEEP Network 2 29. Ssewamala FM Han C-K Neilands TB Asset ownership and health and mental health functioning among AIDS-orphaned adolescents: findings from a randomized clinical trial in rural Uganda Soc Sci Med 2009 69 2 191 8 10.1016/j.socscimed.2009.05.019 19520472 30. Bermudez LG, Ssewamala FM, Neilands TB, et al. Does economic strengthening improve viral suppression among adolescents living with HIV? Results from a Cluster Randomized Trial in Uganda. AIDS Behav. 2018:1–10. 31. Ksoll C Lilleør HB Lønborg JH Rasmussen OD Impact of Village Savings and Loan Associations: evidence from a cluster randomized trial J Dev Econ 2016 120 70 85 10.1016/j.jdeveco.2015.12.003 32. MacPherson E Sadalaki J Nyongopa V Exploring the complexity of microfinance and HIV in fishing communities on the shores of Lake Malawi Rev Afr Polit Econ 2015 42 145 414 36 10.1080/03056244.2015.1064369 33. Leatherman S, Dunford C, Metcalfe M et al, editors. Integrating Microfinance and Health benefits, Challenges and reflections for moving Forward. Global Microcredit Summit commissioned Workshop paper; 2011. 34. Dworkin SL Blankenship K Microfinance and HIV/AIDS Prevention: assessing its Promise and Limitations AIDS Behav 2009 13 462 9 10.1007/s10461-009-9532-3 19294500 35. Slegh H Barker G Kimonyo A Ndolimana P Bannerman M ‘I can do women’s work’: reflections on engaging men as allies in women’s economic empowerment in Rwanda Gend Dev 2013 21 1 15 30 10.1080/13552074.2013.767495 36. Dunbar MS Maternowska MC Kang M-SJ Findings from SHAZ!: a feasibility study of a microcredit and life-skills HIV prevention intervention to reduce risk among adolescent female orphans in Zimbabwe J Prev Interv Community 2010 38 2 147 61 10.1080/10852351003640849 20391061 37. Klostermann KC Fals-Stewart W Intimate partner violence and alcohol use: exploring the role of drinking in partner violence and its implications for intervention Aggress Violent Beh 2006 11 6 587 97 10.1016/j.avb.2005.08.008 38. Fals-Stewart W Birchler GR Kelley ML Learning sobriety together: a randomized clinical trial examining behavioral couples therapy with alcoholic female patients J Consult Clin Psychol 2006 74 3 579 10.1037/0022-006X.74.3.579 16822114 39. Fals-Stewart W Clinton-Sherrod M Treating intimate partner violence among substance-abusing dyads: the effect of couples therapy Prof Psychology: Res Pract 2009 40 3 257 10.1037/a0012708 40. Wechsberg WM, El-Bassel N, Carney T, et al. Adapting an evidence-based HIV behavioral intervention for South African couples. Substance abuse treatment, prevention, and policy. 2015;10(1):6. 41. Wechsberg WM Zule WA El-Bassel N The male factor: outcomes from a cluster randomized field experiment with a couples-based HIV prevention intervention in a south african township Drug Alcohol Depend 2016 161 307 15 10.1016/j.drugalcdep.2016.02.017 26946991 42. Darbes LA McGrath NM Hosegood V Results of a couples-based randomized controlled trial aimed to increase testing for HIV JAIDS J Acquir Immune Defic Syndr 2019 80 4 404 13 10.1097/QAI.0000000000001948 30730356 43. Darbes LA van Rooyen H Hosegood V Uthando Lwethu (‘our love’): a protocol for a couples-based intervention to increase testing for HIV: a randomized controlled trial in rural KwaZulu-Natal, South Africa Trials 2014 15 1 64 10.1186/1745-6215-15-64 24552199 44. Ssewamala FM Bermudez LG Neilands TB Suubi4Her: a study protocol to examine the impact and cost associated with a combination intervention to prevent HIV risk behavior and improve mental health functioning among adolescent girls in Uganda BMC Public Health 2018 18 1 693 10.1186/s12889-018-5604-5 29871619 45. Ssewamala FM Neilands TB Waldfogel J Ismayilova L The impact of a comprehensive microfinance intervention on depression levels of AIDS-orphaned children in Uganda J Adolesc Health 2012 50 4 346 52 10.1016/j.jadohealth.2011.08.008 22443837 46. Brofenbrenner U The Ecology of Human Development Cambridge 1979 MA Harvard University Press 47. Sherraden M Stakeholding Notes on a theory of welfare based on assets Soc Serv Rev 1990 64 4 580 601 10.1086/603797 48. Sherraden M. Can the poor save?: saving and asset building in individual development accounts. Routledge; 2017. 49. Montgomery BE Rompalo A Hughes J Violence against women in selected areas of the United States Am J Public Health 2015 105 10 2156 66 10.2105/AJPH.2014.302430 25790408 50. Goodman LA Smyth KF Borges AM Singer R When crises collide how intimate Partner violence and poverty intersect to shape women’s Mental Health and Coping? Trauma Violence & Abuse 2009 10 4 306 29 10.1177/1524838009339754 51. de Cássia Ribeiro-Silva R Fiaccone RL Barreto ML The association between intimate partner domestic violence and the food security status of poor families in Brazil Public Health Nutr 2016 19 07 1305 11 10.1017/S1368980015002694 26411418 52. Diamond-Smith N, Conroy AA, Tsai AC, Nekkanti M, Weiser SD. Food insecurity and intimate partner violence among married women in Nepal. Journal of Global Health. 2018;In Press. 53. Conroy AA Cohen M Frongillo E Food insecurity and violence in a prospective cohort of U.S. women at risk or living with HIV PLoS ONE 2019 14 3 e0213365 10.1371/journal.pone.0213365 30840700 54. Bandura A, Walters RH. Social learning theory:. Prentice-hall Englewood Cliffs, NJ; 1977. 55. Wingood G DiClemente R The ADAPT-ITT Model A novel method of adapting evidence-based HIV interventions JAIDS 2008 47 S1 40-S6 56. Latham TP Sales JM Boyce LS Application of ADAPT-ITT: adapting an evidence-based HIV prevention intervention for incarcerated african american adolescent females Health Promot Pract 2010 11 3_suppl 53S 60S 10.1177/1524839910361433 20488969 57. Sullivan PS Stephenson R Grazter B Adaptation of the african couples HIV testing and counseling model for men who have sex with men in the United States: an application of the ADAPT-ITT framework Springerplus 2014 3 1 1 13 24422185 58. Embleton L Di Ruggiero E Odep Okal E Adapting an evidence-based gender, livelihoods, and HIV prevention intervention with street-connected young people in Eldoret, Kenya Glob Public Health 2019 14 12 1703 17 10.1080/17441692.2019.1625940 31162989 59. Shrestha R Altice F Karki P Copenhaver M Developing an integrated, brief biobehavioral HIV prevention intervention for high-risk drug users in treatment: the process and outcome of formative research Front Immunol 2017 8 561 10.3389/fimmu.2017.00561 28553295 60. Woolf-King SE, Firkey M, Foley JD, et al. Development of a telephone-delivered Acceptance and Commitment Therapy intervention for people living with HIV who are hazardous drinkers. AIDS Behav. 2022:1–16. 61. Conroy AA, McKenna S, Ruark A, et al. Relationship Dynamics are Associated with Self-Reported adherence but not an Objective Adherence measure in Malawi. AIDS Behav. 2022:1–12. 62. Conroy A, Tuthill E, Neilands T, Johnson M, Mkandawire J. A dyadic investigation of relationship dynamics and depression in HIV-infected couples from Malawi. Poster presented at the International AIDS Conference 23–27 July 2018. Amsterdam, Netherlands. 63. Conroy A, Ruark A, Darbes L, et al. Relationship dynamics and dyadic coordination are associated with ART adherence among married couples from Malawi. Paper presented at the Conference on HIV Treatment and Prevention Adherence. Miami2018, June. 64. Conroy AA, Leddy A, Neilands TB, Mulauzi N. Mkandawire JIntimate partner violence impacts adherence to ART: Evidence from married couples in Malawi Poster presented at the Conference on HIV Treatment and Prevention Adherence; 8–10 June 2018; Miami2018. 65. Bush K Kivlahan D McDonell M Fihn S Bradley K The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking Arch Intern Med 1998 158 1789 95 10.1001/archinte.158.16.1789 9738608 66. Bradley KA DeBenedetti AF Volk RJ AUDIT-C as a brief screen for alcohol misuse in primary care Alcoholism: Clin Experimental Res 2007 31 7 1208 17 10.1111/j.1530-0277.2007.00403.x 67. Conroy AA, Ruark A, McKenna SA, et al. The unaddressed needs of alcohol-using couples on antiretroviral therapy in Malawi: formative research on multilevel interventions. Under Review. 68. Conroy AA, Ruark A, McKenna SA, et al. The unaddressed needs of Alcohol-Using couples on antiretroviral therapy in Malawi: Formative Research on Multilevel Interventions. AIDS Behav. 2019:1–13. 69. Babor TF, Higgins-Biddle JC, Organization WH. Brief intervention for hazardous and harmful drinking: a manual for use in primary care. 2001. 70. Dunkle K Stern E Chatterji S Heise L Effective prevention of intimate partner violence through couples training: a randomised controlled trial of Indashyikirwa in Rwanda BMJ global health 2020 5 12 e002439 10.1136/bmjgh-2020-002439 71. Foss SK Rhetorical criticism: exploration and practice 2009 Long Grove Waveland Press 72. Conroy AA McKenna SA Comfort ML Marital infidelity, food insecurity, and couple instability: a web of challenges for dyadic coordination around antiretroviral therapy Soc Sci Med 2018 214 110 7 10.1016/j.socscimed.2018.08.006 30172140 73. Darbes LA, Van Rooyen H, Hosegood V, et al. Uthando Lwethu (‘our love’): a protocol for a couples-based intervention to increase testing for HIV: a randomized controlled trial in rural KwaZulu-Natal, South Africa. Trials. 2014;15(64). 74. Pettifor A MacPhail C Nguyen N Feasibility and acceptability of Project Connect: a couples-based HIV-risk reduction intervention among young couples in Johannesburg, South Africa AIDS Care 2013 26 4 476 85 10.1080/09540121.2013.841827 24116954

36520335

PMC9753077

NO-CC CODE

2022-12-16 23:25:29

no

AIDS Behav. 2022 Dec 15;:1-16

utf-8

AIDS Behav

10.1007/s10461-022-03956-3

oa_other

==== Front Sci China Chem Sci China Chem Science China. Chemistry 1674-7291 1869-1870 Science China Press Beijing 1408 10.1007/s11426-022-1408-5 Invited Reviews Structural and functional imaging of brains Liu Zhichao 1 Zhu Ying 2 Zhang Liming 1 Jiang Weiping 3 Liu Yawei 4 Tang Qiaowei 2 Cai Xiaoqing 2 Li Jiang 2 Wang Lihua 2 Tao Changlu 7 Yin Xianzhen 8 Li Xiaowei 9 Hou Shangguo 10 Jiang Dawei 11 Liu Kai 5 Zhou Xin xinzhou@wipm.ac.cn 3 Zhang Hongjie hjzhang2019@mail.tsinghua.edu.cn 45 Liu Maili ml.liu@wipm.ac.cn 3 Fan Chunhai fanchunhai@sjtu.edu.cn 6 Tian Yang ytian@chem.ecnu.edu.cn 1 1 grid.22069.3f 0000 0004 0369 6365 Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai, 200241 China 2 grid.410726.6 0000 0004 1797 8419 Interdisciplinary Research Center, Shanghai Synchrotron Radiation Facility, Zhangjiang Laboratory, Shanghai Advanced Research Institute, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Shanghai, 201210 China 3 grid.462167.0 0000 0004 1769 327X State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, Chinese Academy of Sciences, Wuhan National Laboratory for Optoelectronics, Wuhan, 430071 China 4 grid.9227.e 0000000119573309 State Key Laboratory of Rare Earth Resource Utilization, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, 130022 China 5 grid.12527.33 0000 0001 0662 3178 Department of Chemistry, Tsinghua University, Beijing, 100084 China 6 grid.16821.3c 0000 0004 0368 8293 School of Chemistry and Chemical Engineering, Frontiers Science Center for Transformative Molecules, Institute of Translational Medicine, Shanghai Jiao Tong University, Shanghai, 200240 China 7 grid.9227.e 0000000119573309 Interdisciplinary Center for Brain Information, Brain Cognition and Brain Disease Institute, Shenzhen-Hong Kong Institute of Brain Science-Shenzhen Fundamental Research Institutions, Faculty of Life and Health Sciences, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, 518055 China 8 Lingang Laboratory, Shanghai, 201602 China 9 grid.16821.3c 0000 0004 0368 8293 School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, 200240 China 10 grid.510951.9 0000 0004 7775 6738 Institute of Systems and Physical Biology, Shenzhen Bay Laboratory, Shenzhen, 518055 China 11 grid.33199.31 0000 0004 0368 7223 Department of Nuclear Medicine, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430022 China 9 12 2022 143 27 7 2022 28 9 2022 © Science China Press and Springer-Verlag GmbH Germany, part of Springer Nature 2022 This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic. Analyzing the complex structures and functions of brain is the key issue to understanding the physiological and pathological processes. Although neuronal morphology and local distribution of neurons/blood vessels in the brain have been known, the subcellular structures of cells remain challenging, especially in the live brain. In addition, the complicated brain functions involve numerous functional molecules, but the concentrations, distributions and interactions of these molecules in the brain are still poorly understood. In this review, frontier techniques available for multiscale structure imaging from organelles to the whole brain are first overviewed, including magnetic resonance imaging (MRI), computed tomography (CT), positron emission tomography (PET), serial-section electron microscopy (ssEM), light microscopy (LM) and synchrotron-based X-ray microscopy (XRM). Specially, XRM for three-dimensional (3D) imaging of large-scale brain tissue with high resolution and fast imaging speed is highlighted. Additionally, the development of elegant methods for acquisition of brain functions from electrical/chemical signals in the brain is outlined. In particular, the new electrophysiology technologies for neural recordings at the single-neuron level and in the brain are also summarized. We also focus on the construction of electrochemical probes based on dual-recognition strategy and surface/interface chemistry for determination of chemical species in the brain with high selectivity and long-term stability, as well as electrochemophysiological microarray for simultaneously recording of electrochemical and electrophysiological signals in the brain. Moreover, the recent development of brain MRI probes with high contrast-to-noise ratio (CNR) and sensitivity based on hyperpolarized techniques and multi-nuclear chemistry is introduced. Furthermore, multiple optical probes and instruments, especially the optophysiological Raman probes and fiber Raman photometry, for imaging and biosensing in live brain are emphasized. Finally, a brief perspective on existing challenges and further research development is provided. Keywords brain structure brain function brain chemistry chemical signal biosensing and bioimaging ==== Body pmcAcknowledgements This work was supported by the National Natural Science Foundation of China (22004037 for Liu Z; 22022410 and 82050005 for Zhu Y; 22022402 and 21974051 for Zhang L; 21635003 and 21811540027 for Tian Y; 22125701 and 21834007 for Liu K; 22020102003 for Zhang H; 91859206 and 21921004 for Zhou X), the Innovation Program of Shanghai Municipal Education Commission (201701070005E00020 for Tian Y), the Research Funds of Happiness Flower ECNU (2020JK2103 for Tian Y), the Shanghai Municipal Science and Technology Commission (19JC1410300 for Fan C), the Youth Innovation Promotion Association of CAS (2016236 for Zhu Y), and the National Key Research and Development Project of China (2018YFA0704000 for Zhou X). Conflict of interest The authors declare no conflict of interest. These authors contributed equally to this work. ==== Refs References 1 Wang Y Tan H Hui X Biomed Res Int 2018 2018 7848901 29805977 2 Uddin MJ Zidorn C Nat Prod Bioprosp 2020 10 377 410 10.1007/s13659-020-00269-7 3 Cabrera DeBuc D Somfai GM Arthur E Kostic M Oropesa S Mendoza Santiesteban C Front Physiol 2018 9 1721 10.3389/fphys.2018.01721 30574092 4 Shipp S Curr Biol 2007 17 R443 R449 10.1016/j.cub.2007.03.044 17580069 5 Wang Q Ding SL Li Y Royall J Feng D Lesnar P Graddis N Naeemi M Facer B Ho A Dolbeare T Blanchard B Dee N Wakeman W Hirokawa KE Szafer A Sunkin SM Oh SW Bernard A Phillips JW Hawrylycz M Koch C Zeng H Harris JA Ng L Cell 2020 181 936 953 10.1016/j.cell.2020.04.007 32386544 6 Carlén M Science 2017 358 478 482 10.1126/science.aan8868 29074767 7 Rasman BG Forbes PA Peters RM Ortiz O Franks I Inglis JT Chua R Blouin JS eLife 2021 10 e65085 10.7554/eLife.65085 34374648 8 Chang RB Strochlic DE Williams EK Umans BD Liberles SD Cell 2015 161 622 633 10.1016/j.cell.2015.03.022 25892222 9 Denoth-Lippuner A Jessberger S Nat Rev Neurosci 2021 22 223 236 10.1038/s41583-021-00433-z 33633402 10 Hudetz AG Humphries CJ Binder JR Front Syst Neurosci 2014 8 234 10.3389/fnsys.2014.00234 25565989 11 Rings T Mazarei M Akhshi A Geier C Tabar MRR Lehnertz K Sci Rep 2019 9 1744 10.1038/s41598-018-38372-y 30741977 12 Osmanlıoğlu Y Tunç B Parker D Elliott MA Baum GL Ciric R Satterthwaite TD Gur RE Gur RC Verma R NeuroImage 2019 199 93 104 10.1016/j.neuroimage.2019.05.064 31141738 13 Fu Z Sui J Turner JA Du Y Assaf M Pearlson GD Calhoun VD Hum Brain Mapp 2021 42 80 94 10.1002/hbm.25205 32965740 14 Zhao L Wang E Zhang X Karama S Khundrakpam B Zhang H Guan M Wang M Cheng J Shi D Evans AC Li Y Biomed Res Int 2015 2015 1 23 15 Tang Q Li J Zhu Y Wang L Tian Y Fan C Hu J Chem J Chinese U 2022 43 20220379 16 Konkel L Environ Health Perspect 2018 126 112001 10.1289/EHP2268 30457876 17 Pugh JR Raman IM Trends Neuroscis 2009 32 170 177 10.1016/j.tins.2008.12.001 18 van Veenendaal TM Backes WH Tse DHY Scheenen TWJ Klomp DW Hofman PAM Rouhl RPW Vlooswijk MCG Aldenkamp AP Jansen JFA NeuroImage-Clin 2018 19 47 55 10.1016/j.nicl.2018.04.006 30035001 19 Onakomaiya MM Porter DM Oberlander JG Henderson LP Hormones Behav 2014 66 283 297 10.1016/j.yhbeh.2014.04.008 20 Petushkova AI Zamyatnin AA Jr. Biomolecules 2020 10 650 10.3390/biom10040650 32340246 21 Wang H Jordan VC Ramsay IA Sojoodi M Fuchs BC Tanabe KK Caravan P Gale EM J Am Chem Soc 2019 141 5916 5925 10.1021/jacs.9b00603 30874437 22 Kavalali ET Nat Rev Neurosci 2015 16 5 16 10.1038/nrn3875 25524119 23 Reiner A Levitz J Neuron 2018 98 1080 1098 10.1016/j.neuron.2018.05.018 29953871 24 Grace AA Nat Rev Neurosci 2016 17 524 532 10.1038/nrn.2016.57 27256556 25 Mumtaz F Khan MI Zubair M Dehpour AR Biomed Pharmacother 2018 105 1205 1222 10.1016/j.biopha.2018.05.086 30021357 26 Barnham KJ Bush AI Curr Opin Chem Biol 2008 12 222 228 10.1016/j.cbpa.2008.02.019 18342639 27 Zhang X Liang M Qin W Wan B Yu C Ming D Front Hum Neurosci 2020 14 244 10.3389/fnhum.2020.00244 32792927 28 Tereshchenko S Kasparov E Behaval Sci 2019 9 62 10.3390/bs9060062 29 Shahbazi A Kinnison J Vescovi R Du M Hill R Joesch M Takeno M Zeng H da Costa NM Grutzendler J Kasthuri N Scheirer WJ Sci Rep 2018 8 14247 10.1038/s41598-018-32628-3 30250218 30 Lee TJ Lewallen CF Bumbarger DJ Yunker PJ Reid RC Forest CR J Colloid Interface Sci 2018 531 352 359 10.1016/j.jcis.2018.07.068 30041112 31 Bourne JN Harris KM Curr Opin Neurobiol 2012 22 372 382 10.1016/j.conb.2011.10.019 22088391 32 Chen Y Akin O Nern A Tsui CYK Pecot MY Zipursky SL Neuron 2014 81 280 293 10.1016/j.neuron.2013.12.021 24462095 33 Beyene AG Yang SJ Landry MP J Vacuum Sci Tech A 2019 37 040802 10.1116/1.5051047 34 Tsai D Yuste R Shepard KL IEEE Trans Biomed Circuits Syst 2018 12 13 23 10.1109/TBCAS.2017.2750484 29377795 35 Wang B Wen X Cao Y Huang S Lam HA Liu T Chung PS Monbouquette HG Chiou PY Maidment NT Lab Chip 2020 20 1390 1397 10.1039/D0LC00021C 32211718 36 Steinmetz NA Koch C Harris KD Carandini M Curr Opin Neurobiol 2018 50 92 100 10.1016/j.conb.2018.01.009 29444488 37 Eftekhari M Assadi M Kazemi M Saghari M Esfahani AF Sichani BF Gholamrezanezhad A Beiki D BMC Med Phys 2005 5 6 10.1186/1471-2385-5-6 38 Mullins RJ Diehl TC Chia CW Kapogiannis D Front Aging Neurosci 2017 9 118 10.3389/fnagi.2017.00118 28515688 39 Joshi VS Sheet PS Cullin N Kreth J Koley D Anal Chem 2017 89 11044 11052 10.1021/acs.analchem.7b03050 28920437 40 Wang R Zhang H Liu Q Liu F Han X Liu X Li K Xiao G Albert J Lu X Guo T Nat Commun 2022 13 547 10.1038/s41467-022-28267-y 35087063 41 Liu YL Wang XY Xu JQ Xiao C Liu YH Zhang XW Liu JT Huang WH Chem Sci 2015 6 1853 1858 10.1039/C4SC03123G 28706641 42 Dai Y Liu Z Bai Y Chen Z Qin J Feng F RSC Adv 2018 8 42246 42252 10.1039/C8RA09298B 35558776 43 Mamede AP Santos IP Batista de Carvalho ALM Figueiredo P Silva MC Tavares MV Marques MPM Batista de Carvalho LAE Cancers 2021 13 5336 10.3390/cancers13215336 34771500 44 Jeong S Widengren J Lee JC Nanomaterials 2022 12 21 10.3390/nano12010021 45 Braunschweig AB Schmucker AL Wei WD Mirkin CA Chem Phys Lett 2010 486 89 98 10.1016/j.cplett.2010.01.009 20396668 46 Jia HR Zhu YX Xu KF Pan GY Liu X Qiao Y Wu FG Chem Sci 2019 10 4062 4068 10.1039/C8SC04884C 31015947 47 Chen Y Li X Biomacromolecules 2011 12 4367 4372 10.1021/bm201350d 22040128 48 Ahrens MB Orger MB Robson DN Li JM Keller PJ Nat Methods 2013 10 413 420 10.1038/nmeth.2434 23524393 49 Mahou P Vermot J Beaurepaire E Supatto W Nat Methods 2014 11 600 601 10.1038/nmeth.2963 24874570 50 Streich L Boffi JC Wang L Alhalaseh K Barbieri M Rehm R Deivasigamani S Gross CT Agarwal A Prevedel R Nat Methods 2021 18 1253 1258 10.1038/s41592-021-01257-6 34594033 51 Liu J Liu Z Wang W Tian Y Angew Chem Int Ed 2021 60 21351 21359 10.1002/anie.202106193 52 Wang W Zhao F Li M Zhang C Shao Y Tian Y Angew Chem Int Ed 2019 58 5256 5260 10.1002/anie.201814286 53 Liu Z Zhang Z Liu Y Mei Y Feng E Liu Y Zheng T Chen J Zhang S Tian Y Angew Chem Int Ed 2022 61 e2111630 54 Zhu Y Zhang J Li A Zhang Y Fan C Curr Opin Chem Biol 2017 39 11 16 10.1016/j.cbpa.2017.04.016 28521258 55 Zhu Y Earnest T Huang Q Cai X Wang Z Wu Z Fan C Adv Mater 2014 26 7889 7895 10.1002/adma.201304281 24687860 56 Fan C Zhao Z Synchrotron Radiation in Materials Science: Light Sources, Techniques, and Applications 2018 Weinheim John Wiley & Sons 57 Qiang Y Artoni P Seo KJ Culaclii S Hogan V Zhao X Zhong Y Han X Wang PM Lo YK Li Y Patel HA Huang Y Sambangi A Chu JSV Liu W Fagiolini M Fang H Sci Adv 2018 4 eaat0626 10.1126/sciadv.aat0626 30191176 58 Lin ZC Xie C Osakada Y Cui Y Cui B Nat Commun 2014 5 3206 10.1038/ncomms4206 24487777 59 Zhang L Tian Y Acc Chem Res 2018 51 688 696 10.1021/acs.accounts.7b00543 29485847 60 Liu Z Tian Y Sci China Chem 2021 64 915 931 10.1007/s11426-020-9961-3 61 Abbott J Ye T Krenek K Gertner RS Ban S Kim Y Qin L Wu W Park H Ham D Nat Biomed Eng 2020 4 232 241 10.1038/s41551-019-0455-7 31548592 62 Zhao Y You SS Zhang A Lee JH Huang J Lieber CM Nat Nanotechnol 2019 14 783 790 10.1038/s41565-019-0478-y 31263191 63 Chai X Zhou X Zhu A Zhang L Qin Y Shi G Tian Y Angew Chem Int Ed 2013 52 8129 8133 10.1002/anie.201302958 64 Dong H Zhou Q Zhang L Tian Y Angew Chem Int Ed 2019 58 13948 13953 10.1002/anie.201907210 65 Liu L Zhao F Liu W Zhu T Zhang JZH Chen C Dai Z Peng H Huang JL Hu Q Bu W Tian Y Angew Chem Int Ed 2017 56 10471 10475 10.1002/anie.201705615 66 Liu W Dong H Zhang L Tian Y Angew Chem Int Ed 2017 56 16328 16332 10.1002/anie.201710863 67 Luo Y Zhang L Liu W Yu Y Tian Y Angew Chem Int Ed 2015 54 14053 14056 10.1002/anie.201508635 68 Zhang C Liu Z Zhang L Zhu A Liao F Wan J Zhou J Tian Y Angew Chem Int Ed 2020 59 20499 20507 10.1002/anie.202006318 69 Huang S Zhang L Dai L Wang Y Tian Y Anal Chem 2021 93 5570 5576 10.1021/acs.analchem.1c00151 33757286 70 Li S Zhu A Zhu T Zhang JZH Tian Y Anal Chem 2017 89 6656 6662 10.1021/acs.analchem.7b00881 28560874 71 Zhang L Han Y Zhao F Shi G Tian Y Anal Chem 2015 87 2931 2936 10.1021/ac504448m 25630826 72 Zhou J Zhang L Tian Y Anal Chem 2016 88 2113 2118 10.1021/acs.analchem.5b03634 26768309 73 Zhao F Shi GY Tian Y Chin J Anal Chem 2019 47 347 354 10.1016/S1872-2040(19)61146-1 74 Liu Y Liu Z Zhao F Tian Y Angew Chem Int Ed 2021 60 14429 14437 10.1002/anie.202102833 75 Zhao F Liu Y Dong H Feng S Shi G Lin L Tian Y Angew Chem Int Ed 2020 59 10426 10430 10.1002/anie.202002417 76 Li H Zhang Z Zhong J Ruan W Han Y Sun X Ye C Zhou X NMR Biomed 2016 29 220 225 10.1002/nbm.3465 26915791 77 Zhou X Sun Y Mazzanti M Henninger N Mansour J Fisher M Albert M NMR Biomed 2011 24 170 175 10.1002/nbm.1568 20821723 78 Zhou X Mazzanti ML Chen JJ Tzeng YS Mansour JK Gereige JD Venkatesh AK Sun Y Mulkern RV Albert MS NMR Biomed 2008 21 217 225 10.1002/nbm.1184 17557274 79 Zhou YP Mei MJ Wang XZ Huang SN Chen L Zhang M Li XY Qin HB Dong X Cheng S Wen L Yang B An XF He AD Zhang B Zeng WB Li XJ Lu Y Li HC Li H Zou WG Redwood AJ Rayner S Cheng H McVoy MA Tang Q Britt WJ Zhou X Jiang X Luo MH JCI Insight 2022 7 e152551 10.1172/jci.insight.152551 35014624 80 Lauterbur PC Nature 1973 242 190 191 10.1038/242190a0 81 Gonçalves FG Serai SD Zuccoli G Top Magn Reson Imag 2018 27 387 393 10.1097/RMR.0000000000000189 82 Hashemi RH Bradley WG Lisanti CJ Mri: The Basics 2012 Philadelphia Lippincott Williams & Wilkins 83 Nitz WR Reimer P Eur Radiol 1999 9 1032 1046 10.1007/s003300050789 10415233 84 Stanisz GJ Odrobina EE Pun J Escaravage M Graham SJ Bronskill MJ Henkelman RM Magn Reson Med 2005 54 507 512 10.1002/mrm.20605 16086319 85 Spijkerman JM Petersen ET Hendrikse J Luijten P Zwanenburg JJM Magn Reson Mater Phy 2018 31 415 424 10.1007/s10334-017-0659-3 86 De Coene B Hajnal JV Gatehouse P Longmore DB White SJ Oatridge A Pennock J Young I Bydder G Am J Neuroradiol 1992 13 1555 1564 1332459 87 Hagiwara A Warntjes M Hori M Andica C Nakazawa M Kumamaru KK Abe O Aoki S Investig Radiol 2017 52 647 657 10.1097/RLI.0000000000000365 28257339 88 Ji S Yang D Lee J Choi SH Kim H Kang KM Magn Reson Imag 2022 55 1013 1025 10.1002/jmri.27440 89 Blystad I Warntjes JBM Smedby O Landtblom AM Lundberg P Larsson EM Acta Radiol 2012 53 1158 1163 10.1258/ar.2012.120195 23024181 90 Tanenbaum LN Tsiouris AJ Johnson AN Naidich TP DeLano MC Melhem ER Quarterman P Parameswaran SX Shankaranarayanan A Goyen M Field AS AJNR Am J Neuroradiol 2017 38 1103 1110 10.3174/ajnr.A5227 28450439 91 Bobman SA Riederer SJ Lee JN Suddarth SA Wang HZ Drayer BP MacFall JR Am J Neuroradiol 1985 6 265 269 2984911 92 Scheffler K Hennig J Magn Reson Med 2001 45 720 723 10.1002/mrm.1097 11284003 93 Gulani V Schmitt P Griswold MA Webb AG Jakob PM Investig Radiol 2004 39 767 774 10.1097/00004424-200412000-00008 15550838 94 Ehses P Seiberlich N Ma D Breuer FA Jakob PM Griswold MA Gulani V Magn Reson Med 2013 69 71 81 10.1002/mrm.24225 22378141 95 Pfister J Blaimer M Kullmann WH Bartsch AJ Jakob PM Breuer FA Magn Reson Med 2019 81 3488 3502 10.1002/mrm.27657 30687949 96 Warntjes JBM Leinhard OD West J Lundberg P Magn Reson Med 2008 60 320 329 10.1002/mrm.21635 18666127 97 Schmidbauer V Geisl G Diogo M Weber M Goeral K Klebermass-Schrehof K Berger A Prayer D Kasprian G Eur Radiol 2019 29 7063 7072 10.1007/s00330-019-06325-2 31286188 98 Di Giuliano F Minosse S Picchi E Marfia GA Da Ros V Muto M Muto M Pistolese CA Laghi A Garaci F Floris R Magn Reson Mater Phy 2020 33 549 557 10.1007/s10334-019-00804-9 99 Cheng CC Preiswerk F Hoge WS Kuo TH Madore B Magn Reson Med 2019 81 1699 1713 10.1002/mrm.27525 30320945 100 Cheng CC Preiswerk F Madore B Magn Reson Med 2020 83 2310 2321 10.1002/mrm.28077 31755588 101 Ma D Gulani V Seiberlich N Liu K Sunshine JL Duerk JL Griswold MA Nature 2013 495 187 192 10.1038/nature11971 23486058 102 Panda A Mehta BB Coppo S Jiang Y Ma D Seiberlich N Griswold MA Gulani V Curr Opin Biomed Eng 2017 3 56 66 10.1016/j.cobme.2017.11.001 29868647 103 Hsieh JJL Svalbe I J Med Radiat Sci 2020 67 333 344 10.1002/jmrs.413 32596957 104 Mazor G Weizman L Tal A Eldar YC Med Phys 2018 45 4066 4084 10.1002/mp.13078 105 Zhao B Setsompop K Ye H Cauley SF Wald LL IEEE Trans Med Imag 2016 35 1812 1823 10.1109/TMI.2016.2531640 106 Buonincontri G Biagi L Retico A Cecchi P Cosottini M Gallagher FA Gómez PA Graves MJ McLean MA Riemer F Schulte RF Tosetti M Zaccagna F Kaggie JD NeuroImage 2019 195 362 372 10.1016/j.neuroimage.2019.03.047 30923028 107 Tang BL Biochem Biophysl Res Commun 2006 345 911 916 10.1016/j.bbrc.2006.05.040 108 Vollmer T Signorovitch J Huynh L Galebach P Kelley C Di-Bernardo A Sasane R J Neurol Sci 2015 357 8 18 10.1016/j.jns.2015.07.014 26238166 109 Sanai SA Saini V Benedict RH Zivadinov R Teter BE Ramanathan M Weinstock-Guttman B Mult Scler 2016 22 717 725 10.1177/1352458516634871 26895718 110 Giorgio A De Stefano N J Magn Reson Imag 2013 37 1 14 10.1002/jmri.23671 111 Reiter K Nielson KA Durgerian S Woodard JL Smith JC Seidenberg M Kelly DA Rao SM J Alzheimer Dis 2017 55 1363 1377 10.3233/JAD-160504 112 Ashburner J Friston KJ Neuroimage 2000 11 805 821 10.1006/nimg.2000.0582 10860804 113 Wetzel SG Johnson G Tan AGS Cha S Knopp EA Lee VS Thomasson D Rofsky NM Am J Neuroradiol 1962 23 995 1002 114 Whitwell JL J Neurosci 2009 29 9661 9664 10.1523/JNEUROSCI.2160-09.2009 19657018 115 Ashburner J NeuroImage 2012 62 791 800 10.1016/j.neuroimage.2011.10.025 22023741 116 Ashburner J Magn Reson Imag 2009 27 1163 1174 10.1016/j.mri.2009.01.006 117 Jenkinson M Beckmann CF Behrens TEJ Woolrich MW Smith SM NeuroImage 2012 62 782 790 10.1016/j.neuroimage.2011.09.015 21979382 118 Avants BB Tustison N Song G Insight J 2009 2 1 35 119 Fischl B Neuroimage 2012 62 774 781 10.1016/j.neuroimage.2012.01.021 22248573 120 Manjón JV Coupé P Front Neuroinf 2016 10 30 10.3389/fninf.2016.00030 121 Ramzan F Khan MUG Iqbal S Saba T Rehman A IEEE Access 2020 8 103697 103709 10.1109/ACCESS.2020.2998901 122 Wang L Xie C Zeng N IEEE Access 2019 7 39670 39679 10.1109/ACCESS.2019.2906890 123 Bernal J Kushibar K Asfaw DS Valverde S Oliver A Martí R Lladó X Artif Intell Med 2019 95 64 81 10.1016/j.artmed.2018.08.008 30195984 124 Risacher SL Anderson WH Charil A Castelluccio PF Shcherbinin S Saykin AJ Schwarz AJ Neurology 2017 89 2176 2186 10.1212/WNL.0000000000004670 29070667 125 Tsagkas C Magon S Gaetano L Pezold S Naegelin Y Amann M Stippich C Cattin P Wuerfel J Bieri O Sprenger T Kappos L Parmar K Neurology 2018 91 e349 e358 10.1212/WNL.0000000000005853 29950437 126 Qin Y Wu J Chen T Li J Zhang G Wu D Zhou Y Zheng N Cai A Ning Q Manyande A Xu F Wang J Zhu W J Clin Investigation 2021 131 147329 10.1172/JCI147329 127 Chen S Lan Y Li H Xia L Ye C Lou X Zhou X Biomedicines 2022 10 781 10.3390/biomedicines10040781 35453531 128 Xue D Liu Y Jin L Wang Y Cui F Liu J Li X Zhang S Zhao Y Yin N Niu R Wang D Zhang H Sci China Chem 2021 64 558 564 10.1007/s11426-020-9920-x 129 He W Ai K Lu L Sci China Chem 2015 58 753 760 10.1007/s11426-015-5351-8 130 Peng XJ He HP Liu Q She K Zhang BQ Wang HS Tang HT Pan YM Sci China Chem 2021 64 753 760 10.1007/s11426-020-9958-6 131 Regenhardt RW González RG He J Lev MH Singhal AB Radiology 2022 302 400 407 10.1148/radiol.2021210455 34726532 132 Rava RA Snyder KV Mokin M Waqas M Allman AB Senko JL Podgorsak AR Shiraz Bhurwani MM Hoi Y Siddiqui AH Davies JM Levy EI Ionita CN AJNR Am J Neuroradiol 2020 41 206 212 10.3174/ajnr.A6395 31948951 133 Konno Y Hiraka T Kanoto M Sato T Tsunoda M Ishizawa T Matsuda A Makino N Pancreatology 2020 20 1406 1412 10.1016/j.pan.2020.08.010 32888809 134 He J Yang Y Wang Y Zeng D Bian Z Zhang H Sun J Xu Z Ma J IEEE Trans Med Imag 2018 38 371 382 10.1109/TMI.2018.2865202 135 Siegler JE Rosenberg J Cristancho D Olsen A Pulst-Korenberg J Raab L Cucchiara B Messé SR Int J Stroke 2020 15 299 307 10.1177/1747493019869702 31409213 136 Li Y Sci China Chem 2013 56 1682 1692 10.1007/s11426-013-5004-8 137 Liu CY Jiang SD Sci China Ser B-Chem 2009 52 2195 2199 10.1007/s11426-009-0195-8 138 Maurer GD Brucker DP Stoffels G Filipski K Filss CP Mottaghy FM Galldiks N Steinbach JP Hattingen E Langen KJ J Nucl Med 2020 61 505 511 10.2967/jnumed.119.234757 31519802 139 Kim D Chun JH Yi JH Ko HY Chung JI Lee M Park YM Nam MH Kim J Kim SY Park Y Moon JH Kang SG Chang JH Lee CJ Kim SH Yun M Clin Nucl Med 2022 47 863 868 10.1097/RLU.0000000000004341 35868002 140 Li A Gong H Zhang B Wang Q Yan C Wu J Liu Q Zeng S Luo Q Science 2010 330 1404 1408 10.1126/science.1191776 21051596 141 Ding W Li A Wu J Yang Z Meng Y Wang S Gong H J Microsc 2013 251 168 177 10.1111/jmi.12058 23781931 142 Wu J He Y Yang Z Guo C Luo Q Zhou W Chen S Li A Xiong B Jiang T Gong H NeuroImage 2014 87 199 208 10.1016/j.neuroimage.2013.10.036 24185025 143 Zhang X Yin X Zhang J Li A Gong H Luo Q Zhang H Gao Z Jiang H Natl Sci Rev 2019 6 1223 1238 10.1093/nsr/nwz124 34692000 144 Xiong B Li A Lou Y Chen S Long B Peng J Yang Z Xu T Yang X Li X Jiang T Luo Q Gong H Front Neuroanat 2017 11 128 10.3389/fnana.2017.00128 29311856 145 Yin X Zhang X Zhang J Yang W Sun X Zhang H Gao Z Jiang H Front Neurosci 2022 16 870520 10.3389/fnins.2022.870520 35516801 146 Xue S Gong H Jiang T Luo W Meng Y Liu Q Chen S Li A PLoS ONE 2014 9 e88067 10.1371/journal.pone.0088067 24498247 147 Li Y Li N Yu X Huang K Zheng T Cheng X Zeng S Liu X Sci Rep 2018 8 1086 10.1038/s41598-018-19373-3 29348413 148 Huisken J Swoger J Del Bene F Wittbrodt J Stelzer EHK Science 2004 305 1007 1009 10.1126/science.1100035 15310904 149 Keller PJ Schmidt AD Wittbrodt J Stelzer EHK Science 2008 322 1065 1069 10.1126/science.1162493 18845710 150 Stelzer EHK Nat Methods 2015 12 23 26 10.1038/nmeth.3219 25549266 151 Power RM Huisken J Nat Methods 2017 14 360 373 10.1038/nmeth.4224 28362435 152 Planchon TA Gao L Milkie DE Davidson MW Galbraith JA Galbraith CG Betzig E Nat Methods 2011 8 417 423 10.1038/nmeth.1586 21378978 153 Chen BC Legant WR Wang K Shao L Milkie DE Davidson MW Janetopoulos C Wu XS Hammer JA III Liu Z English BP Mimori-Kiyosue Y Romero DP Ritter AT Lippincott-Schwartz J Fritz-Laylin L Mullins RD Mitchell DM Bembenek JN Reymann AC Böhme R Grill SW Wang JT Seydoux G Tulu US Kiehart DP Betzig E Science 2014 346 1257998 10.1126/science.1257998 25342811 154 Vaadia RD Li W Voleti V Singhania A Hillman EMC Grueber WB Curr Biol 2019 29 935 944.e4 10.1016/j.cub.2019.01.060 30853438 155 Ueda HR Ertürk A Chung K Gradinaru V Chédotal A Tomancak P Keller PJ Nat Rev Neurosci 2020 21 298 10.1038/s41583-020-0291-5 156 Zhao S Todorov MI Cai R -Maskari RA Steinke H Kemter E Mai H Rong Z Warmer M Stanic K Schoppe O Paetzold JC Gesierich B Wong MN Huber TB Duering M Bruns OT Menze B Lipfert J Puelles VG Wolf E Bechmann I Ertürk A Cell 2020 180 796 812.e19 10.1016/j.cell.2020.01.030 32059778 157 Dodt HU Leischner U Schierloh A Jährling N Mauch CP Deininger K Deussing JM Eder M Zieglgänsberger W Becker K Nat Methods 2007 4 331 336 10.1038/nmeth1036 17384643 158 Murakami TC Mano T Saikawa S Horiguchi SA Shigeta D Baba K Sekiya H Shimizu Y Tanaka KF Kiyonari H Iino M Mochizuki H Tainaka K Ueda HR Nat Neurosci 2018 21 625 637 10.1038/s41593-018-0109-1 29507408 159 Renier N Adams EL Kirst C Wu Z Azevedo R Kohl J Autry AE Kadiri L Umadevi Venkataraju K Zhou Y Wang VX Tang CY Olsen O Dulac C Osten P Tessier-Lavigne M Cell 2016 165 1789 1802 10.1016/j.cell.2016.05.007 27238021 160 Park YG Sohn CH Chen R McCue M Yun DH Drummond GT Ku T Evans NB Oak HC Trieu W Choi H Jin X Lilascharoen V Wang J Truttmann MC Qi HW Ploegh HL Golub TR Chen SC Frosch MP Kulik HJ Lim BK Chung K Nat Biotechnol 2019 37 73 83 10.1038/nbt.4281 161 Chen F Tillberg PW Boyden ES Science 2015 347 543 548 10.1126/science.1260088 25592419 162 Ku T Swaney J Park JY Albanese A Murray E Cho JH Park YG Mangena V Chen J Chung K Nat Biotechnol 2016 34 973 981 10.1038/nbt.3641 27454740 163 Gao R Asano SM Upadhyayula S Pisarev I Milkie DE Liu TL Singh V Graves A Huynh GH Zhao Y Bogovic J Colonell J Ott CM Zugates C Tappan S Rodriguez A Mosaliganti KR Sheu SH Pasolli HA Pang S Xu CS Megason SG Hess H Lippincott-Schwartz J Hantman A Rubin GM Kirchhausen T Saalfeld S Aso Y Boyden ES Betzig E Science 2019 363 eaau8302 10.1126/science.aau8302 30655415 164 Voigt FF Kirschenbaum D Platonova E Pagès S Campbell RAA Kastli R Schaettin M Egolf L van der Bourg A Bethge P Haenraets K Frézel N Topilko T Perin P Hillier D Hildebrand S Schueth A Roebroeck A Roska B Stoeckli ET Pizzala R Renier N Zeilhofer HU Karayannis T Ziegler U Batti L Holtmaat A Lüscher C Aguzzi A Helmchen F Nat Methods 2019 16 1105 1108 10.1038/s41592-019-0554-0 31527839 165 Todorov MI Paetzold JC Schoppe O Tetteh G Shit S Efremov V Todorov-Völgyi K Düring M Dichgans M Piraud M Menze B Ertürk A Nat Methods 2020 17 442 449 10.1038/s41592-020-0792-1 32161395 166 Hillman EMC Voleti V Li W Yu H Annu Rev Neurosci 2019 42 295 313 10.1146/annurev-neuro-070918-050357 31283896 167 Mostany R Miquelajauregui A Shtrahman M Portera-Cailliau C Verveer P Two-photon excitation microscopy and its applications in neuroscience Advanced Fluorescence Microscopy. Methods in Molecular Biology 1975 New York Humana Press 168 Truong TV Supatto W Koos DS Choi JM Fraser SE Nat Methods 2011 8 757 760 10.1038/nmeth.1652 21765409 169 Wolf S Supatto W Debrégeas G Mahou P Kruglik SG Sintes JM Beaurepaire E Candelier R Nat Methods 2015 12 379 380 10.1038/nmeth.3371 25924070 170 Zong W Zhao J Chen X Lin Y Ren H Zhang Y Fan M Zhou Z Cheng H Sun Y Chen L Cell Res 2015 25 254 257 10.1038/cr.2014.124 25257466 171 Lin PY Hwang SPL Lee CH Chen BC J Biomed Opt 2021 26 116503 34796706 172 Takanezawa S Saitou T Imamura T Nat Commun 2021 12 1 5 10.1038/s41467-021-23249-y 33397941 173 de Vito G Turrini L Müllenbroich C Ricci P Sancataldo G Mazzamuto G Tiso N Sacconi L Fanelli D Silvestri L Vanzi F Pavone FS Biomed Opt Express 2022 13 1516 1536 10.1364/BOE.434146 35414999 174 Gasparoli FM Escobet-Montalbán A Early J Bruce GD Dholakia K OSA Continuum 2020 3 2935 2942 10.1364/OSAC.405369 175 Voleti V Patel KB Li W Perez Campos C Bharadwaj S Yu H Ford C Casper MJ Yan RW Liang W Wen C Kimura KD Targoff KL Hillman EMC Nat Methods 2019 16 1054 1062 10.1038/s41592-019-0579-4 31562489 176 Yang B Chen X Wang Y Feng S Pessino V Stuurman N Cho NH Cheng KW Lord SJ Xu L Xie D Mullins RD Leonetti MD Huang B Nat Methods 2019 16 501 504 10.1038/s41592-019-0401-3 31061492 177 Sapoznik E Chang BJ Huh J Ju RJ Azarova EV Pohlkamp T Welf ES Broadbent D Carisey AF Stehbens SJ Lee KM Marín A Hanker AB Schmidt JC Arteaga CL Yang B Kobayashi Y Tata PR Kruithoff R Doubrovinski K Shepherd DP Millett-Sikking A York AG Dean KM Fiolka RP eLife 2020 9 e57681 10.7554/eLife.57681 33179596 178 Patel KB Liang W Casper MJ Voleti V Li W Yagielski AJ Zhao HT Perez Campos C Lee GS Liu JM Philipone E Yoon AJ Olive KP Coley SM Hillman EMC Nat Biomed Eng 2022 6 569 583 10.1038/s41551-022-00849-7 35347275 179 Yang B Lange M Millett-Sikking A Zhao X Bragantini J Vijay-Kumar S Kamb M Gómez-Sjöberg R Solak AC Wang W Kobayashi H McCarroll MN Whitehead LW Fiolka RP Kornberg TB York AG Royer LA Nat Methods 2022 19 461 469 10.1038/s41592-022-01417-2 35314838 180 Malkinson G Mahou P Chaudan É Gacoin T Sonay AY Pantazis P Beaurepaire E Supatto W ACS Photonics 2020 7 1036 1049 10.1021/acsphotonics.9b01749 33335947 181 Zhou B Shi B Jin D Liu X Nat Nanotech 2015 10 924 936 10.1038/nnano.2015.251 182 Richardson DS Lichtman JW Cell 2015 162 246 257 10.1016/j.cell.2015.06.067 26186186 183 Royer LA Lemon WC Chhetri RK Wan Y Coleman M Myers EW Keller PJ Nat Biotechnol 2016 34 1267 1278 10.1038/nbt.3708 27798562 184 Ji N Nat Methods 2017 14 374 380 10.1038/nmeth.4218 28362438 185 Li C Moatti A Zhang X Troy Ghashghaei H Greenabum A Biomed Opt Express 2021 12 5214 5226 10.1364/BOE.427099 34513252 186 Palay S Exp Cell Res (Suppl) 1958 5 275 293 187 Estable C Reissig M De Robertis E Exp Cell Res 1954 6 255 262 10.1016/0014-4827(54)90172-X 13173478 188 Robertson JD Exp Biol Med 1953 82 219 223 10.3181/00379727-82-20071 189 Liao M Cao E Julius D Cheng Y Nature 2013 504 107 112 10.1038/nature12822 24305160 190 Zhang Y Ye F Zhang T Lv S Zhou L Du D Lin H Guo F Luo C Zhu S Nature 2021 596 301 305 10.1038/s41586-021-03769-9 34321660 191 Liu YT Tao CL Lau PM Zhou ZH Bi GQ Curr Opin Struct Biol 2019 54 152 160 10.1016/j.sbi.2019.02.012 30904821 192 Liu YT Tao CL Zhang X Xia W Shi DQ Qi L Xu C Sun R Li XW Lau PM Zhou ZH Bi GQ Nat Neurosci 2020 23 1589 1596 10.1038/s41593-020-00729-w 33139942 193 Huang S Xu P Tan Y You C Zhang Y Jiang Y Xu HE Cell Res 2021 31 1036 1038 10.1038/s41422-021-00527-4 34239069 194 Fitzpatrick AWP Falcon B He S Murzin AG Murshudov G Garringer HJ Crowther RA Ghetti B Goedert M Scheres SHW Nature 2017 547 185 190 10.1038/nature23002 28678775 195 Nakane T Kotecha A Sente A McMullan G Masiulis S Brown PMGE Grigoras IT Malinauskaite L Malinauskas T Miehling J Uchański T Yu L Karia D Pechnikova EV de Jong E Keizer J Bischoff M McCormack J Tiemeijer P Hardwick SW Chirgadze DY Murshudov G Aricescu AR Scheres SHW Nature 2020 587 152 156 10.1038/s41586-020-2829-0 33087931 196 Trinkaus VA Riera-Tur I Martínez-Sánchez A Bäuerlein FJB Guo Q Arzberger T Baumeister W Dudanova I Hipp MS Hartl FU Fernández-Busnadiego R Nat Commun 2021 12 1 10.1038/s41467-021-22108-0 33397941 197 Guo Q Lehmer C Martínez-Sánchez A Rudack T Beck F Hartmann H Pérez-Berlanga M Frottin F Hipp MS Hartl FU Edbauer D Baumeister W Fernández-Busnadiego R Cell 2018 172 696 705.e12 10.1016/j.cell.2017.12.030 29398115 198 Tao CL Liu YT Zhou ZH Lau PM Bi GQ Front Neuroanat 2018 12 48 10.3389/fnana.2018.00048 29942253 199 Tao CL Liu YT Sun R Zhang B Qi L Shivakoti S Tian CL Zhang P Lau PM Zhou ZH Bi GQ J Neurosci 2018 38 1493 1510 10.1523/JNEUROSCI.1548-17.2017 29311144 200 Radhakrishnan A Li X Grushin K Krishnakumar SS Liu J Rothman JE Proc Natl Acad Sci USA 2021 118 e2024029118 10.1073/pnas.2024029118 33468631 201 Sun R Liu YT Tao CL Qi L Lau PM Zhou ZH Bi GQ Biophys Rep 2019 5 111 122 10.1007/s41048-019-0092-4 202 Schaffer M Mahamid J Engel BD Laugks T Baumeister W Plitzko JM J Struct Biol 2017 197 73 82 10.1016/j.jsb.2016.07.010 27444390 203 White JG Southgate E Thomson JN Brenner S Phil Trans R Soc Lond B 1986 314 1 340 10.1098/rstb.1986.0056 22462104 204 Shapson-Coe A, Januszewski M, Berger DR, Pope A, Wu Y, Blakely T, Schalek RL, Li PH, Wang S, Maitin-Shepard J. BioRxiv, 2021, doi: 10.1101/2021.05.29.446289 205 Scholl B Thomas CI Ryan MA Kamasawa N Fitzpatrick D Nature 2021 590 111 114 10.1038/s41586-020-03044-3 33328635 206 Henstridge M Pfeiffer C Wang D Boltasseva A Shalaev VM Grbic A Merlin R Science 2018 362 439 442 10.1126/science.aat5915 30361369 207 Hwu Y Margaritondo G Chiang A-S BMC Biol 2017 15 1 9 10.1186/s12915-017-0461-8 28100223 208 Töpperwien M van der Meer F Stadelmann C Salditt T Proc Natl Acad Sci USA 2018 115 6940 6945 10.1073/pnas.1801678115 29915047 209 Foxley S Sampathkumar V De Andrade V Trinkle S Sorokina A Norwood K La Riviere P Kasthuri N NeuroImage 2021 238 118250 10.1016/j.neuroimage.2021.118250 34116154 210 Withers PJ Bouman C Carmignato S Cnudde V Grimaldi D Hagen CK Maire E Manley M Du Plessis A Stock SR Nat Rev Methods Primers 2021 1 1 2 10.1038/s43586-021-00015-4 211 Sanchez-Cano C Alvarez-Puebla RA Abendroth JM Beck T Blick R Cao Y Caruso F Chakraborty I Chapman HN Chen C Cohen BE Conceição ALC Cormode DP Cui D Dawson KA Falkenberg G Fan C Feliu N Gao M Gargioni E Glüer CC Grüner F Hassan M Hu Y Huang Y Huber S Huse N Kang Y Khademhosseini A Keller TF Körnig C Kotov NA Koziej D Liang XJ Liu B Liu S Liu Y Liu Z Liz-Marzán LM Ma X Machicote A Maison W Mancuso AP Megahed S Nickel B Otto F Palencia C Pascarelli S Pearson A Peñate-Medina O Qi B Rädler J Richardson JJ Rosenhahn A Rothkamm K Rübhausen M Sanyal MK Schaak RE Schlemmer HP Schmidt M Schmutzler O Schotten T Schulz F Sood AK Spiers KM Staufer T Stemer DM Stierle A Sun X Tsakanova G Weiss PS Weller H Westermeier F Xu M Yan H Zeng Y Zhao Y Zhao Y Zhu D Zhu Y Parak WJ ACS Nano 2021 15 3754 3807 10.1021/acsnano.0c09563 33650433 212 Andreasen NC Science 1988 239 1381 1388 10.1126/science.3279509 3279509 213 Lv J Di Biase M Cash RFH Cocchi L Cropley VL Klauser P Tian Y Bayer J Schmaal L Cetin-Karayumak S Rathi Y Pasternak O Bousman C Pantelis C Calamante F Zalesky A Mol Psychiatry 2021 26 3512 3523 10.1038/s41380-020-00882-5 32963336 214 Bosch C Ackels T Pacureanu A Zhang Y Peddie CJ Berning M Rzepka N Zdora MC Whiteley I Storm M Bonnin A Rau C Margrie T Collinson L Schaefer AT Nat Commun 2022 13 1 6 34983933 215 Kong H Li M Xu Q Zhou H Yu F Wang Q Am J Nucl Med Mol Imaging 2021 11 529 536 35003890 216 Brooks J Everett J Lermyte F Tjhin VT Banerjee S O’Connor PB Morris CM Sadler PJ Telling ND Collingwood JF Angew Chem Int Ed 2020 59 11984 11991 10.1002/anie.202000239 217 Kong H Zhang J Li J Wang J Shin HJ Tai R Yan Q Xia K Hu J Wang L Zhu Y Fan C Natl Sci Rev 2020 7 1218 1227 10.1093/nsr/nwaa055 34692146 218 Si-Mohamed S Tatard-Leitman V Laugerette A Sigovan M Pfeiffer D Rummeny EJ Coulon P Yagil Y Douek P Boussel L Noël PB Sci Rep 2019 9 1 8 10.1038/s41598-019-44821-z 30626917 219 McCollough CH Leng S Yu L Fletcher JG Radiology 2015 276 637 653 10.1148/radiol.2015142631 26302388 220 Fratini M Bukreeva I Campi G Brun F Tromba G Modregger P Bucci D Battaglia G Spanò R Mastrogiacomo M Requardt H Giove F Bravin A Cedola A Sci Rep 2015 5 8514 10.1038/srep08514 25686728 221 Hu J Li P Yin X Wu T Cao Y Yang Z Jiang L Hu S Lu H J Synchrotron Rad 2017 24 482 489 10.1107/S1600577517000121 222 Cedola A Bravin A Bukreeva I Fratini M Pacureanu A Mittone A Massimi L Cloetens P Coan P Campi G Spanò R Brun F Grigoryev V Petrosino V Venturi C Mastrogiacomo M Kerlero de Rosbo N Uccelli A Sci Rep 2017 7 1 10.1038/s41598-017-06251-7 28127051 223 Bukreeva I Campi G Fratini M Spanò R Bucci D Battaglia G Giove F Bravin A Uccelli A Venturi C Mastrogiacomo M Cedola A Sci Rep 2017 7 1 10.1038/srep41054 28127051 224 Böhm T Joseph K Kirsch M Moroni R Hilger A Osenberg M Manke I Johnston M Stieglitz T Hofmann UG Haas CA Thiele S Sci Rep 2019 9 7646 10.1038/s41598-019-42544-9 31113972 225 Khimchenko A Schulz G Thalmann P Müller B APL Bioeng 2018 2 016106 10.1063/1.5022184 31069291 226 Bikis C Rodgers G Deyhle H Thalmann P Hipp A Beckmann F Weitkamp T Theocharis S Rau C Schulz G Müller B Appl Phys Lett 2019 114 083702 10.1063/1.5085302 227 Hubert M Pacureanu A Guilloud C Yang Y da Silva JC Laurencin J Lefebvre-Joud F Cloetens P Appl Phys Lett 2018 112 203704 10.1063/1.5026462 228 Khimchenko A Bikis C Pacureanu A Hieber SE Thalmann P Deyhle H Schweighauser G Hench J Frank S Müller-Gerbl M Schulz G Cloetens P Müller B Adv Sci 2018 5 1700694 10.1002/advs.201700694 229 Khimchenko A Pacureanu A Bikis C Hieber SE Thalmann P Deyhle H Schweighauser G Hench J Frank S Muller-Gerbl M Schulz G Cloetens P Muller B Microsc Microanal 2018 24 354 355 10.1017/S143192761801406X 230 Kuan AT Phelps JS Thomas LA Nguyen TM Han J Chen CL Azevedo AW Tuthill JC Funke J Cloetens P Pacureanu A Lee WCA Nat Neurosci 2020 23 1637 1643 10.1038/s41593-020-0704-9 32929244 231 Hill AV Nature 1929 123 9 11 10.1038/123009a0 232 Hubel DH Wiesel TN J Physiol 1959 148 574 591 10.1113/jphysiol.1959.sp006308 14403679 233 Hamill OP Marty A Neher E Sakmann B Sigworth FJ Pflügers Archiv 1981 391 85 100 10.1007/BF00656997 6270629 234 Zhang C Liu JQ Tian HC Kang XY Du JC Rui YF Yang B Yang CS Microsyst Technol 2015 21 139 145 10.1007/s00542-013-2017-3 235 Wang K Chung-Chiun Liu K Durand DM IEEE Trans Biomed Eng 2009 56 6 14 10.1109/TBME.2008.926691 19224713 236 Brüggemann D Wolfrum B Maybeck V Mourzina Y Jansen M Offenhäusser A Nanotechnology 2011 22 265104 10.1088/0957-4484/22/26/265104 21586820 237 Kim JH Kang G Nam Y Choi YK Nanotechnology 2010 21 085303 10.1088/0957-4484/21/8/085303 238 Khodagholy D Doublet T Gurfinkel M Quilichini P Ismailova E Leleux P Herve T Sanaur S Bernard C Malliaras GG Adv Mater 2011 23 H268 H272 10.1002/adma.201102378 21826747 239 Blau A Murr A Wolff S Sernagor E Medini P Iurilli G Ziegler C Benfenati F Biomaterials 2011 32 1778 1786 10.1016/j.biomaterials.2010.11.014 21145588 240 Ansaldo A Castagnola E Maggiolini E Fadiga L Ricci D ACS Nano 2011 5 2206 2214 10.1021/nn103445d 21341752 241 Keefer EW Botterman BR Romero MI Rossi AF Gross GW Nat Nanotech 2008 3 434 439 10.1038/nnano.2008.174 242 Chung JE Joo HR Fan JL Liu DF Barnett AH Chen S Geaghan-Breiner C Karlsson MP Karlsson M Lee KY Liang H Magland JF Pebbles JA Tooker AC Greengard LF Tolosa VM Frank LM Neuron 2019 101 21 31.e5 10.1016/j.neuron.2018.11.002 30502044 243 Dipalo M Rastogi SK Matino L Garg R Bliley J Iachetta G Melle G Shrestha R Shen S Santoro F Feinberg AW Barbaglia A Cohen-Karni T De Angelis F Sci Adv 2021 7 eabd5175 10.1126/sciadv.abd5175 33827809 244 Khodagholy D Doublet T Quilichini P Gurfinkel M Leleux P Ghestem A Ismailova E Hervé T Sanaur S Bernard C Malliaras GG Nat Commun 2013 4 1575 10.1038/ncomms2573 23481383 245 Salatino JW Ludwig KA Kozai TDY Purcell EK Nat Biomed Eng 2017 1 862 877 10.1038/s41551-017-0154-1 30505625 246 Feiner R Dvir T Nat Rev Mater 2018 3 17076 10.1038/natrevmats.2017.76 247 Musk E J Med Internet Res 2019 21 e16194 10.2196/16194 31642810 248 Khodagholy D Gelinas JN Thesen T Doyle W Devinsky O Malliaras GG Buzsáki G Nat Neurosci 2015 18 310 315 10.1038/nn.3905 25531570 249 Fu TM Hong G Zhou T Schuhmann TG Viveros RD Lieber CM Nat Methods 2016 13 875 882 10.1038/nmeth.3969 27571550 250 Lee W Kim D Matsuhisa N Nagase M Sekino M Malliaras GG Yokota T Someya T Proc Natl Acad Sci USA 2017 114 10554 10559 10.1073/pnas.1703886114 28923928 251 Lee J Ozden I Song YK Nurmikko AV Nat Methods 2015 12 1157 1162 10.1038/nmeth.3620 26457862 252 Park S Guo Y Jia X Choe HK Grena B Kang J Park J Lu C Canales A Chen R Yim YS Choi GB Fink Y Anikeeva P Nat Neurosci 2017 20 612 619 10.1038/nn.4510 28218915 253 Zhang Y Zheng N Cao Y Wang F Wang P Ma Y Lu B Hou G Fang Z Liang Z Yue M Li Y Chen Y Fu J Wu J Xie T Feng X Sci Adv 2019 5 eaaw1066 10.1126/sciadv.aaw1066 31086809 254 Wu F Stark E Ku PC Wise KD Buzsáki G Yoon E Neuron 2015 88 1136 1148 10.1016/j.neuron.2015.10.032 26627311 255 Minev IR Musienko P Hirsch A Barraud Q Wenger N Moraud EM Gandar J Capogrosso M Milekovic T Asboth L Torres RF Vachicouras N Liu Q Pavlova N Duis S Larmagnac A Vörös J Micera S Suo Z Courtine G Lacour SP Science 2015 347 159 163 10.1126/science.1260318 25574019 256 Shin H Son Y Chae U Kim J Choi N Lee HJ Woo J Cho Y Yang SH Lee CJ Cho IJ Nat Commun 2019 10 3777 10.1038/s41467-019-11628-5 31439845 257 Hodgkin AL Huxley AF Nature 1939 144 710 711 10.1038/144710a0 258 Cole KS Curtis HJ J Gen Physiol 1939 22 649 670 10.1085/jgp.22.5.649 19873125 259 Cole K Arch Sci Physiol 1949 3 253 258 260 Marmont G J Cell Comp Physiol 1949 34 351 382 10.1002/jcp.1030340303 15406358 261 Hodgkin AL Huxley AF J Physiol 1952 116 449 472 10.1113/jphysiol.1952.sp004717 14946713 262 Neher E Sakmann B Nature 1976 260 799 802 10.1038/260799a0 1083489 263 Peier AM Moqrich A Hergarden AC Reeve AJ Andersson DA Story GM Earley TJ Dragoni I McIntyre P Bevan S Patapoutian A Cell 2002 108 705 715 10.1016/S0092-8674(02)00652-9 11893340 264 Noma A Nature 1983 305 147 148 10.1038/305147a0 6310409 265 Nowak L Bregestovski P Ascher P Herbet A Prochiantz A Nature 1984 307 462 465 10.1038/307462a0 6320006 266 Song E Lee YK Li R Li J Jin X Yu KJ Xie Z Fang H Zhong Y Du H Zhang J Fang G Kim Y Yoon Y Alam MA Mei Y Huang Y Rogers JA Adv Funct Mater 2018 28 1702284 10.1002/adfm.201702284 267 Buzsáki G Anastassiou CA Koch C Nat Rev Neurosci 2012 13 407 420 10.1038/nrn3241 22595786 268 Cohen MX Trends Neuroscis 2017 40 208 218 10.1016/j.tins.2017.02.004 269 Pravdich-Neminsky V Zbl Physiol 1913 27 951 960 270 Berger H Archiv für Psychiatrie und Nervenkrankheiten 1929 87 527 570 10.1007/BF01797193 271 Adrian ED Matthews BHC Brain 1934 57 355 385 10.1093/brain/57.4.355 272 Adrian ED Yamagiwa K Brain 1935 58 323 351 10.1093/brain/58.3.323 273 Jasper H Electroen Clin Neuro 1958 10 370 375 10.1016/0013-4694(58)90053-1 274 Petrov Y Nador J Hughes C Tran S Yavuzcetin O Sridhar S Neuroimage 2014 90 140 145 10.1016/j.neuroimage.2013.12.041 24398333 275 Tucker DM Electroencephalogr Clin Neurophysiol 1993 87 154 163 10.1016/0013-4694(93)90121-B 7691542 276 Hämäläinen M Hari R Ilmoniemi RJ Knuutila J Lounasmaa OV Rev Mod Phys 1993 65 413 497 10.1103/RevModPhys.65.413 277 He B Sohrabpour A Brown E Liu Z Annu Rev Biomed Eng 2018 20 171 196 10.1146/annurev-bioeng-062117-120853 29494213 278 Dehghani N Bédard C Cash SS Halgren E Destexhe A J Comput Neurosci 2010 29 405 421 10.1007/s10827-010-0263-2 20697790 279 Ahlfors SP Magnetoencephalography: From Signals to Dynamic Cortical Networks 2019 2nd Ed Cham Springer 259 278 280 Stefan H Nakasato N Papanicolaou AC Handb Clin Neurol 2012 107 347 358 10.1016/B978-0-444-52898-8.00021-5 22938981 281 Palmini ALF Epileptic Disorders 2006 8 S10 S15 17012068 282 Kuruvilla A Flink R Seizure 2003 12 577 584 10.1016/S1059-1311(03)00095-5 14630497 283 Bazhenov M Lonjers P Skorheim S Bedard C Destexhe A Phil Trans R Soc A 2011 369 3802 3819 10.1098/rsta.2011.0119 21893529 284 Hong G Lieber CM Nat Rev Neurosci 2019 20 330 345 10.1038/s41583-019-0140-6 30833706 285 Rossant C Kadir SN Goodman DFM Schulman J Hunter MLD Saleem AB Grosmark A Belluscio M Denfield GH Ecker AS Tolias AS Solomon S Buzsaki G Carandini M Harris KD Nat Neurosci 2016 19 634 641 10.1038/nn.4268 26974951 286 Mishanina TV Libiad M Banerjee R Nat Chem Biol 2015 11 457 464 10.1038/nchembio.1834 26083070 287 Kimura H Proc Jpn Acad Ser B-Phys Biol Sci 2015 91 131 159 10.2183/pjab.91.131 288 Greiner R Pálinkás Z Bäsell K Becher D Antelmann H Nagy P Dick TP Antioxid Redox Signal 2013 19 1749 1765 10.1089/ars.2012.5041 23646934 289 Yadav PK Martinov M Vitvitsky V Seravalli J Wedmann R Filipovic MR Banerjee R J Am Chem Soc 2016 138 289 299 10.1021/jacs.5b10494 26667407 290 Tan BH Wong PTH Bian JS Neurochem Int 2010 56 3 10 10.1016/j.neuint.2009.08.008 19703504 291 Elrod JW Calvert JW Morrison J Doeller JE Kraus DW Tao L Jiao X Scalia R Kiss L Szabo C Kimura H Chow CW Lefer DJ Proc Natl Acad Sci USA 2007 104 15560 15565 10.1073/pnas.0705891104 17878306 292 Hu B Kong F Gao X Jiang L Li X Gao W Xu K Tang B Angew Chem Int Ed 2018 57 5306 5309 10.1002/anie.201712921 293 Gao P Wang M Wan X Liu X Pan W Li N Tang B Chem Commun 2020 56 14267 14270 10.1039/D0CC05527A 294 Soto RJ Hall JR Brown MD Taylor JB Schoenfisch MH Anal Chem 2017 89 276 299 10.1021/acs.analchem.6b04251 28105839 295 Liu X Zhang M Xiao T Hao J Li R Mao L Anal Chem 2016 88 7238 7244 10.1021/acs.analchem.6b01476 27327860 296 Liu X Xiao T Wu F Shen MY Zhang M Yu HH Mao L Angew Chem Int Ed 2017 56 11802 11806 10.1002/anie.201705900 297 Li X Ren L Dunevall J Ye D White HS Edwards MA Ewing AG ACS Nano 2018 12 3010 3019 10.1021/acsnano.8b00781 29513514 298 Wang X Chan HN Desbois N Gros CP Bolze F Li Y Li HW Wong MS ACS Appl Mater Interfaces 2021 13 18525 18532 10.1021/acsami.1c01585 33852279 299 Robert P Vives V Grindel AL Kremer S Bierry G Louin G Ballet S Corot C Radiology 2020 294 117 126 10.1148/radiol.2019182953 31660804 300 Wahsner J Gale EM Rodríguez-Rodríguez A Caravan P Chem Rev 2018 119 957 1057 10.1021/acs.chemrev.8b00363 30350585 301 D’Arceuil H Coimbra A Triano P Dougherty M Mello J Moseley M Glover G Lansberg M Blankenberg F NeuroImage 2013 83 200 209 10.1016/j.neuroimage.2013.06.066 23831413 302 Iv M Ng NN Nair S Zhang Y Lavezo J Cheshier SH Holdsworth SJ Moseley ME Rosenberg J Grant GA Yeom KW Radiology 2020 297 438 446 10.1148/radiol.2020200378 32930651 303 Calvete MJF Pinto SMA Pereira MM Geraldes CFGC Coord Chem Rev 2017 333 82 107 10.1016/j.ccr.2016.11.011 304 Lenora CU Carniato F Shen Y Latif Z Haacke EM Martin PD Botta M Allen MJ Chem Eur J 2017 23 15404 15414 10.1002/chem.201702158 28707809 305 Morelli JN Runge VM Williams JM Beissner RS Tweedle M Investig Radiol 2011 46 169 177 10.1097/RLI.0b013e3181f7a0b0 21150792 306 Sillerud LO Solberg NO Chamberlain R Orlando RA Heidrich JE Brown DC Brady CI Vander Jagt TA Garwood M Vander Jagt DL J Alzheimer Dis 2013 34 349 365 10.3233/JAD-121171 307 Cheng Y Cheng A Jia Y Yang L Ning Y Xu L Zhong Y Zhuang Z Guan J Zhang X Lin Y Zhou T Fan X Li J Liu P Yan G Wu R ACS Appl Mater Interfaces 2021 13 56909 56922 10.1021/acsami.1c16530 34807583 308 Barandov A Bartelle BB Williamson CG Loucks ES Lippard SJ Jasanoff A Nat Commun 2019 10 1 9 10.1038/s41467-019-08558-7 30602773 309 Novotna B Herynek V Rossner P Turnovcova K Jendelova P Int J Nanomed 2017 12 4519 4526 10.2147/IJN.S133156 310 Lee T Cai LX Lelyveld VS Hai A Jasanoff A Science 2014 344 533 535 10.1126/science.1249380 24786083 311 Grist JT McLean MA Riemer F Schulte RF Deen SS Zaccagna F Woitek R Daniels CJ Kaggie JD Matys T Patterson I Slough R Gill AB Chhabra A Eichenberger R Laurent MC Comment A Gillard JH Coles AJ Tyler DJ Wilkinson I Basu B Lomas DJ Graves MJ Brindle KM Gallagher FA NeuroImage 2019 189 171 179 10.1016/j.neuroimage.2019.01.027 30639333 312 Lee CY Soliman H Geraghty BJ Chen AP Connelly KA Endre R Perks WJ Heyn C Black SE Cunningham CH Neuroimage 2020 204 116202 10.1016/j.neuroimage.2019.116202 31557546 313 Yuan Y Guo Q Zhang X Jiang W Ye C Zhou X J Mater Chem B 2020 8 5014 5018 10.1039/D0TB00484G 32301463 314 Li H Zhao X Wang Y Lou X Chen S Deng H Shi L Xie J Tang D Zhao J Bouchard LS Xia L Zhou X Sci Adv 2021 7 eabc8180 10.1126/sciadv.abc8180 33219111 315 Witte C Martos V Rose HM Reinke S Klippel S Schröder L Hackenberger CPR Angew Chem Int Ed 2015 54 2806 2810 10.1002/anie.201410573 316 Yanagisawa D Ibrahim NF Taguchi H Morikawa S Tomiyama T Tooyama I Molecules 2021 26 1362 10.3390/molecules26051362 33806326 317 Li S Jiang W Yuan Y Sui M Yang Y Huang L Jiang L Liu M Chen S Zhou X ACS Appl Mater Interfaces 2020 12 57290 57301 10.1021/acsami.0c13865 33231083 318 Walker-Samuel S Ramasawmy R Torrealdea F Rega M Rajkumar V Johnson SP Richardson S Gonçalves M Parkes HG Arstad E Thomas DL Pedley RB Lythgoe MF Golay X Nat Med 2013 19 1067 1072 10.1038/nm.3252 23832090 319 Longo DL Sun PZ Consolino L Michelotti FC Uggeri F Aime S J Am Chem Soc 2014 136 14333 14336 10.1021/ja5059313 25238643 320 Branzoli F Pontoizeau C Tchara L Di Stefano AL Kamoun A Deelchand DK Valabrègue R Lehéricy S Sanson M Ottolenghi C Marjańska M Neuro-Oncology 2019 21 765 774 10.1093/neuonc/noz031 30726924 321 Okada S Bartelle BB Li N Breton-Provencher V Lee JJ Rodriguez E Melican J Sur M Jasanoff A Nat Nanotech 2018 13 473 477 10.1038/s41565-018-0092-4 322 Choi C Ganji SK DeBerardinis RJ Hatanpaa KJ Rakheja D Kovacs Z Yang XL Mashimo T Raisanen JM Marin-Valencia I Pascual JM Madden CJ Mickey BE Malloy CR Bachoo RM Maher EA Nat Med 2012 18 624 629 10.1038/nm.2682 22281806 323 Andronesi OC Kim GS Gerstner E Batchelor T Tzika AA Fantin VR Vander Heiden MG Sorensen AG Sci Transl Med 2012 4 116ra114 10.1126/scitranslmed.3002693 324 Branzoli F Deelchand DK Sanson M Lehéricy S Marjańska M Magn Reson Med 2019 82 1259 1265 10.1002/mrm.27810 31131476 325 Cuccarini V Antelmi L Pollo B Paterra R Calatozzolo C Nigri A DiMeco F Eoli M Finocchiaro G Brenna G Tramacere I Bruzzone MG Anghileri E Neurol Sci 2020 41 347 355 10.1007/s10072-019-04087-9 31650436 326 Lei H Ugurbil K Chen W Proc Natl Acad Sci USA 2003 100 14409 14414 10.1073/pnas.2332656100 14612566 327 Rijpma A van der Graaf M Meulenbroek O Olde Rikkert MGM Heerschap A Neuroimage-Clin 2018 18 254 261 10.1016/j.nicl.2018.01.031 29876246 328 Rango M Bonifati C Bresolin N J Cereb Blood Flow Metab 2006 26 283 290 10.1038/sj.jcbfm.9600192 16094320 329 Levine SR Helpern JA Welch KM Vande Linde AM Sawaya KL Brown EE Ramadan NM Deveshwar RK Ordidge RJ Radiology 1992 185 537 544 10.1148/radiology.185.2.1410369 1410369 330 Le Page LM Guglielmetti C Taglang C Chaumeil MM Trends Neuroscis 2020 43 343 354 10.1016/j.tins.2020.03.006 331 Hesketh RL Brindle KM Curr Opin Chem Biol 2018 45 187 194 10.1016/j.cbpa.2018.03.004 30007214 332 Miloushev VZ Granlund KL Boltyanskiy R Lyashchenko SK DeAngelis LM Mellinghoff IK Brennan CW Tabar V Yang TJ Holodny AI Sosa RE Guo YWW Chen AP Tropp J Robb F Keshari KR Cancer Res 2018 78 3755 3760 10.1158/0008-5472.CAN-18-0221 29769199 333 Park I Larson PEZ Gordon JW Carvajal L Chen HY Bok R Van Criekinge M Ferrone M Slater JB Xu D Kurhanewicz J Vigneron DB Chang S Nelson SJ Magn Reson Med 2018 80 864 873 10.1002/mrm.27077 29322616 334 Mammoli D Gordon J Autry A Larson PEZ Li Y Chen HY Chung B Shin P Van Criekinge M Carvajal L Slater JB Bok R Crane J Xu D Chang S Vigneron DB IEEE Trans Med Imag 2019 39 320 327 10.1109/TMI.2019.2926437 335 Autry AW Gordon JW Chen HY LaFontaine M Bok R Van Criekinge M Slater JB Carvajal L Villanueva-Meyer JE Chang SM Clarke JL Lupo JM Xu D Larson PEZ Vigneron DB Li Y Neuroimage-Clin 2020 27 102323 10.1016/j.nicl.2020.102323 32623139 336 Flogel U Ding Z Hardung H Jander S Reichmann G Jacoby C Schubert R Schrader J Circulation 2008 118 140 148 10.1161/CIRCULATIONAHA.107.737890 18574049 337 Chirizzi C De Battista D Tirotta I Metrangolo P Comi G Bombelli FB Chaabane L Radiology 2019 291 351 357 10.1148/radiol.2019181073 30888930 338 Strauss WL Unis AS Cowan C Dawson G Dager SR Am J Psychiat. 2002 159 755 760 10.1176/appi.ajp.159.5.755 11986128 339 Wenger KJ Richter C Burger MC Urban H Kaulfuss S Harter PN Sreeramulu S Schwalbe H Steinbach JP Hattingen E Bähr O Pilatus U Cancers 2020 12 3175 10.3390/cancers12113175 33138036 340 Prinz C Starke L Millward JM Fillmer A Delgado PR Waiczies H Pohlmann A Rothe M Nazaré M Paul F Niendorf T Waiczies S Theranostics 2021 11 2490 2504 10.7150/thno.47130 33456555 341 Kwong KK Belliveau JW Chesler DA Goldberg IE Weisskoff RM Poncelet BP Kennedy DN Hoppel BE Cohen MS Turner R Cheng HM Brady TJ Rosen BR Proc Natl Acad Sci USA 1992 89 5675 5679 10.1073/pnas.89.12.5675 1608978 342 Ogawa S Tank DW Menon R Ellermann JM Kim SG Merkle H Ugurbil K Proc Natl Acad Sci USA 1992 89 5951 5955 10.1073/pnas.89.13.5951 1631079 343 Bandettini PA Wong EC Hinks RS Tikofsky RS Hyde JS Magn Reson Med 1992 25 390 397 10.1002/mrm.1910250220 1614324 344 Qiu D Zaharchuk G Christen T Ni WW Moseley ME Neuro-image 2012 62 1726 1731 22584230 345 Baumgartner R Cho W Coimbra A Chen C Wang Z Struyk A Venketasubramanian N Low M Gargano C Zhao F Williams D Reese T Seah S Feng D Apreleva S Petersen E Evelhoch JL J Magn Reson Imag 2017 46 124 133 10.1002/jmri.25499 346 Li N Jasanoff A Nature 2020 580 239 244 10.1038/s41586-020-2158-3 32269346 347 Bednařík P Tkáč I Giove F DiNuzzo M Deelchand DK Emir UE Eberly LE Mangia S J Cereb Blood Flow Metab 2015 35 601 610 10.1038/jcbfm.2014.233 25564236 348 Luo Y Kim EH Flask CA Clark HA ACS Nano 2018 12 5761 5773 10.1021/acsnano.8b01640 29851460 349 Savić T Gambino G Bokharaie VS Noori HR Logothetis NK Angelovski G Proc Natl Acad Sci USA 2019 116 20666 20671 10.1073/pnas.1908503116 31548425 350 Lee T Zhang X Dhar S Faas H Lippard SJ Jasanoff A Chem Biol 2010 17 665 673 10.1016/j.chembiol.2010.05.009 20609416 351 Ghosh S Li N Schwalm M Bartelle BB Xie T Daher JI Singh UD Xie K DiNapoli N Evans NB Chung K Jasanoff A Nat Neurosci 2022 25 390 398 10.1038/s41593-022-01014-8 35241803 352 Li M, Liu Z, Wu Y, Zheng N, Liu X, Cai A, Zheng D, Zhu J, Wu J, Xu L. Mol Psychiatry, 2022, doi: 10.1038/s41380-022-01540-0 353 Davis KA Nanga RPR Das S Chen SH Hadar PN Pollard JR Lucas TH Shinohara RT Litt B Hariharan H Elliott MA Detre JA Reddy R Sci Transl Med 2015 7 309ra161 354 Sun PZ Cheung JS Wang E Lo EH J Cereb Blood Flow Metab 2011 31 1743 1750 10.1038/jcbfm.2011.23 21386856 355 Zhang X Yuan Y Li S Zeng Q Guo Q Liu N Yang M Yang Y Liu M McMahon MT Zhou X Magn Reson Med 2019 82 577 585 10.1002/mrm.27753 30968442 356 Tiwari V Daoud EV Hatanpaa KJ Gao A Zhang S An Z Ganji SK Raisanen JM Lewis CM Askari P Baxter J Levy M Dimitrov I Thomas BP Pinho MC Madden CJ Pan E Patel TR DeBerardinis RJ Sherry AD Mickey BE Malloy CR Maher EA Choi C Neuro-Oncology 2020 22 1018 1029 10.1093/neuonc/noaa034 32055850 357 Rao MR Stewart NJ Griffiths PD Norquay G Wild JM Radiology 2018 286 659 665 10.1148/radiol.2017162881 28858563 358 Zeng Q Guo Q Yuan Y Yang Y Zhang B Ren L Zhang X Luo Q Liu M Bouchard LS Zhou X Anal Chem 2017 89 2288 2295 10.1021/acs.analchem.6b03742 28192930 359 Zhang B Guo Q Luo Q Zhang X Zeng QB Zhao L Yuan Y Jiang W Yang Y Liu M Ye C Zhou X Chem Commun 2018 54 13654 13657 10.1039/C8CC07822J 360 Fernando PUAI Shepelytskyi Y Cesana PT Wade A Grynko V Mendieta AM Seveney LE Brown JD Hane FT Albert MS De-Boef B ACS Omega 2020 5 27783 27788 10.1021/acsomega.0c02565 33163761 361 Shapiro MG Ramirez RM Sperling LJ Sun G Sun J Pines A Schaffer DV Bajaj VS Nat Chem 2014 6 629 634 10.1038/nchem.1934 24950334 362 Yang Y Zhang Y Wang B Guo Q Yuan Y Jiang W Shi L Yang M Chen S Lou X Zhou X Chem Sci 2021 12 4300 4308 10.1039/D0SC06969H 34163694 363 Zeng Q Bie B Guo Q Yuan Y Han Q Han X Chen M Zhang X Yang Y Liu M Liu P Deng H Zhou X Proc Natl Acad Sci USA 2020 117 17558 17563 10.1073/pnas.2004121117 32661173 364 Jayapaul J Komulainen S Zhivonitko VV Mareš J Giri C Rissanen K Lantto P Telkki VV Schröder L Nat Commun 2022 13 1 10 10.1038/s41467-022-29249-w 34983933 365 Liu Z Pei H Zhang L Tian Y ACS Nano 2018 12 12357 12368 10.1021/acsnano.8b06322 30418752 366 Zhang Z Liu Z Tian Y iScience 2020 23 101344 10.1016/j.isci.2020.101344 32688287 367 Zhu A Qu Q Shao X Kong B Tian Y Angew Chem Int Ed 2012 51 7185 7189 10.1002/anie.201109089 368 Liu Z Jing X Zhang S Tian Y Anal Chem 2019 91 2488 2497 10.1021/acs.analchem.8b05360 30618245 369 Liu Z Wang S Li W Tian Y Anal Chem 2018 90 2816 2825 10.1021/acs.analchem.7b04934 29376642 370 Liang M Liu Z Zhang Z Mei Y Tian Y Chem Sci 2022 13 4303 4312 10.1039/D2SC00326K 35509464 371 Wu Z Liu M Liu Z Tian Y J Am Chem Soc 2020 142 7532 7541 10.1021/jacs.0c00771 32233469 372 Mei Y Zhang QW Gu Q Liu Z He X Tian Y J Am Chem Soc 2022 144 2351 2359 10.1021/jacs.1c12959 35099950 373 Gong Z Liu Z Zhang Z Mei Y Tian Y CCS Chem 2022 4 2020 2030 10.31635/ccschem.021.202101038 374 Kong B Zhu A Ding C Zhao X Li B Tian Y Adv Mater 2012 24 5844 5848 10.1002/adma.201202599 22933395 375 Lavis LD Raines RT ACS Chem Biol 2008 3 142 155 10.1021/cb700248m 18355003 376 Li B Zhao M Feng L Dou C Ding S Zhou G Lu L Zhang H Chen F Li X Li G Zhao S Jiang C Wang Y Zhao D Cheng Y Zhang F Nat Commun 2020 11 3102 10.1038/s41467-020-16924-z 32555157 377 Wu J Shao C Ye X Di X Li D Zhao H Zhang B Chen G Liu HK Qian Y ACS Sens 2021 6 863 870 10.1021/acssensors.0c01914 33438997 378 Liu S Chen C Li Y Zhang H Liu J Wang R Wong STH Lam JWY Ding D Tang BZ Adv Funct Mater 2020 30 1908125 10.1002/adfm.201908125 379 Samanta S Huang M Li S Yang Z He Y Gu Z Zhang J Zhang D Liu L Qu J Theranostics 2021 11 2137 2148 10.7150/thno.53780 33500716 380 Liu Y Liu J Chen D Wang X Zhang Z Yang Y Jiang L Qi W Ye Z He S Liu Q Xi L Zou Y Wu C Angew Chem Int Ed 2020 59 21049 21057 10.1002/anie.202007886 381 Zhang Z Fang X Liu Z Liu H Chen D He S Zheng J Yang B Qin W Zhang X Wu C Angew Chem Int Ed 2020 59 3691 3698 10.1002/anie.201914397 382 Liu Y Lu Y Chen G Wang Q Biotechnol J 2020 15 2000086 10.1002/biot.202000086 383 Yan P Acker CD Zhou WL Lee P Bollensdorff C Negrean A Lotti J Sacconi L Antic SD Kohl P Mansvelder HD Pavone FS Loew LM Proc Natl Acad Sci USA 2012 109 20443 20448 10.1073/pnas.1214850109 23169660 384 Treger JS Priest MF Iezzi R Bezanilla F Biophysl J 2014 107 L09 L12 10.1016/j.bpj.2014.07.054 385 Zebibula A Alifu N Xia L Sun C Yu X Xue D Liu L Li G Qian J Adv Funct Mater 2018 28 1703451 10.1002/adfm.201703451 386 Li C Li W Liu H Zhang Y Chen G Li Z Wang Q Angew Chem Int Ed 2020 59 247 252 10.1002/anie.201911803 387 Chen G Zhang Y Peng Z Huang D Li C Wang Q Nano Res 2019 12 1321 1326 10.1007/s12274-019-2283-1 388 Ast J Arvaniti A Fine NHF Nasteska D Ashford FB Stamataki Z Koszegi Z Bacon A Jones BJ Lucey MA Sasaki S Brierley DI Hastoy B Tomas A D’Agostino G Reimann F Lynn FC Reissaus CA Linnemann AK D’Este E Calebiro D Trapp S Johnsson K Podewin T Broichhagen J Hodson DJ Nat Commun 2020 11 467 10.1038/s41467-020-14309-w 31980626 389 Campbell BC Nabel EM Murdock MH Lao-Peregrin C Tsoulfas P Blackmore MG Lee FS Liston C Morishita H Petsko GA Proc Natl Acad Sci USA 2020 117 30710 30721 10.1073/pnas.2000942117 33208539 390 Qian Y Piatkevich KD McLarney B Abdelfattah AS Mehta S Murdock MH Gottschalk S Molina RS Zhang W Chen Y Wu J Drobizhev M Hughes TE Zhang J Schreiter ER Shoham S Razansky D Boyden ES Campbell RE Nat Methods 2019 16 171 174 10.1038/s41592-018-0294-6 30664778 391 Akemann W Mutoh H Perron A Rossier J Knöpfel T Nat Methods 2010 7 643 649 10.1038/nmeth.1479 20622860 392 Jin L Han Z Platisa J Wooltorton JRA Cohen LB Pieribone VA Neuron 2012 75 779 785 10.1016/j.neuron.2012.06.040 22958819 393 Liang T Wang Q Li Z Wang P Wu J Zuo M Liu Z Adv Funct Mater 2020 30 1910765 10.1002/adfm.201910765 394 Zhong Y Ma Z Zhu S Yue J Zhang M Antaris AL Yuan J Cui R Wan H Zhou Y Wang W Huang NF Luo J Hu Z Dai H Nat Commun 2017 8 737 10.1038/s41467-017-00917-6 28963467 395 Wang T Wang S Liu Z He Z Yu P Zhao M Zhang H Lu L Wang Z Wang Z Zhang W Fan Y Sun C Zhao D Liu W Bünzli JCG Zhang F Nat Mater 2021 20 1571 1578 10.1038/s41563-021-01063-7 34326504 396 Liu MH Zhang Z Yang YC Chan YH Angew Chem Int Ed 2021 60 983 989 10.1002/anie.202011914 397 Qi J Alifu N Zebibula A Wei P Lam JWY Peng HQ Kwok RTK Qian J Tang BZ Nano Today 2020 34 100893 10.1016/j.nantod.2020.100893 398 Scott BB Thiberge SY Guo C Tervo DGR Brody CD Karpova AY Tank DW Neuron 2018 100 1045 1058.e5 10.1016/j.neuron.2018.09.050 30482694 399 Yu W Guo B Zhang H Zhou J Yu X Zhu L Xue D Liu W Sun X Qian J Sci Bull 2019 64 410 416 10.1016/j.scib.2019.02.019 400 Yu X Feng Z Cai Z Jiang M Xue D Zhu L Zhang Y Liu J Que B Yang W Xi W Zhang D Qian J Li G J Mater Chem B 2019 7 6623 6629 10.1039/C9TB01381D 31591622 401 Sych Y Chernysheva M Sumanovski LT Helmchen F Nat Methods 2019 16 553 560 10.1038/s41592-019-0400-4 31086339 402 Takezaki M Kawakami R Onishi S Suzuki Y Kawamata J Imamura T Hadano S Watanabe S Niko Y Adv Funct Mater 2021 31 2010698 10.1002/adfm.202010698 403 Pisano F Pisanello M Lee SJ Lee J Maglie E Balena A Sileo L Spagnolo B Bianco M Hyun M De Vittorio M Sabatini BL Pisanello F Nat Methods 2019 16 1185 1192 10.1038/s41592-019-0581-x 31591577 404 Nguyen T Kim M Gwak J Lee JJ Choi KY Lee KH Kim JG J Biophotonics 2019 12 e201800298 10.1002/jbio.201800298 30963713 405 Husain SF Yu R Tang TB Tam WW Tran B Quek TT Hwang SH Chang CW Ho CS Ho RC Sci Rep 2020 10 1 9 10.1038/s41598-020-66784-2 31913322 406 Feng E Zheng T He X Chen J Tian Y Sci Adv 2018 4 eaau3494 10.1126/sciadv.aau3494 30406203 407 Zhou Y Gu Q Qiu T He X Chen J Qi R Huang R Zheng T Tian Y Angew Chem Int Ed 2021 60 26260 26267 10.1002/anie.202112367 408 Feng E Tian Y Chem Res Chin Univ 2021 37 989 1007 10.1007/s40242-021-1263-7

0

PMC9753096

NO-CC CODE

2022-12-16 23:25:29

no

Sci China Chem. 2022 Dec 9;:1-43

utf-8

Sci China Chem

10.1007/s11426-022-1408-5

oa_other

==== Front Urol Clin North Am Urol Clin North Am The Urologic Clinics of North America 0094-0143 1558-318X Elsevier Inc. S0094-0143(21)00006-9 10.1016/j.ucl.2021.01.006 Article Telemedicine in Urology The Socioeconomic Impact Kirshenbaum Eric MD a∗ Rhee Eugene Y. MD, MBA bc Gettman Matthew MD d Spitz Aaron MD e a Uropartners, Suite 312, 1475 E Belvidere Rd, Grayslakle, IL 60030, USA b Kaiser Permanente Urology, 4405 Vandever Ave, San Diego, CA 92120, USA c Urology, Permanente Federation d Mayo Clinic Department of Urology, 200 First Street SW, Rochester, MN 55905, USA e Orange County Urology, 23961 Calle De La Magdalena, Laguna Hills, CA 92653, USA ∗ Corresponding author. 12 3 2021 5 2021 12 3 2021 48 2 215222 © 2021 Elsevier Inc. All rights reserved. 2021 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The emergence of the COVID-19 pandemic and subsequent public health emergency (PHE) have propelled telemedicine several years into the future. With the rapid adoption of this technology came socioeconomic inequities as minority communities disproportionately have yet to adopt telemedicine. Telemedicine offers solutions to patient access issues that have plagued urology, helping address physician shortages in rural areas and expanding the reach of urologists. The Centers for Medicare & Medicaid Services have adopted changes to expand coverage for telemedicine services. The expectation is that telemedicine will continue to be a mainstay in the health care system with gradual expansion in utilization. Keywords Telemedicine Telehealth Telesurgery Urology Disparities Socioeconomic ==== Body pmcKey points • The emergence of the COVID-19 public health emergency has propelled telemedicine into the future by alleviating many of the barriers that telehealth adopters faced. • The Centers for Medicare & Medicaid Services have made several changes that allow practitioners to continue to utilize telemedicine through a variety of platforms. • With the adoption of telemedicine came socioeconomic disparities in care and access. It is crucial to ensure equal access to this emerging technology. Introduction In the year 2000, Reed Hastings, the founder and chief executive officer of a new start-up company based on movie rentals by mail, approached Blockbuster, a giant in the movie industry, to merge and lead Blockbuster’s online brand. Blockbuster, rigid in its philosophy, refused to adapt with the online revolution and eventually filed for bankruptcy 10 years later. On the other hand, Reed Hastings grew his company to a current valuation of $125 billion and today Netflix leads the digital entertainment industry. This ability to adapt with changing times propelled Netflix into a leadership position in the industry. In many ways, telemedicine has the same trajectory as Netflix. An industry revolution for health care is transforming the ways in which health care is delivered. As usually happens, the adoption of disruptive technology varies greatly, and this may be exacerbated in certain socioeconomic demographics. Although advocates of telemedicine have been frustrated by the historical lack of action in expanding telehealth services, the emergence of the COVID-19 public health emergency (PHE) virtually overnight has accelerated the adoption of telemedicine several years forward. This article reviews the recent changes in Centers for Medicare & Medicaid Services (CMS) rules regarding telemedicine and postulates its future impact; additionally, it focuses on the socioeconomic impact of telemedicine adoption on urology patients. Background Urologists will need to deliver telemedical care strategically in a scalable fashion that does not further the socioeconomic gap that already exists in health care. Telemedicine and telehealth are terms that describe the interactive exchange of health care information electronically between patients, providers, and consultants for the purpose of education, evaluation, decision making, and treatment. These interactions include text, audio, video, and audio-video communication. This may be live (synchronous) or as store-and-forward (asynchronous) interactions. Platforms are increasing and include personal computers, pads, tablets, smartphones, watches, wireless wearable sensors, and other emerging technologies. Under the official pronouncement by the federal government declaring COVID-19 a PHE, waivers were issued that eased many telemedicine restrictions in a matter of weeks that had challenged the progress of telemedicine for years in the pre–COVID-19 era. Limitations on in-person visits due to safety concerns coupled with this liberalization of telemedicine requirements fast-tracked physicians across most specialties to integrate telemedicine into their practices. The early adoption of telemedicine by a minority of urologists helped onboard the vast majority of the nation’s remaining urologists. Workforce shortage Telemedicine may provide a high-value solution for the workforce and access shortages in urology that have a disproportionate impact on communities from lower socioeconomic strata. The urology workforce will continue to be stressed with a progressive imbalance in available urologists to patients. It is estimated that by 2030, 20% of the population will be age 65 or older with increasing surgical needs. By 2030, urology will face a 32% (3884 urologists) shortage for greater than 350 million US citizens. According to the 2018 American Urological Association (AUA) census data, 30% of urologists are greater than 65 years old, a clear concerning sign of a progressive urologist shortage.1 Urology is the second oldest specialty after thoracic surgery, confronting the profession with the prospect of losing a quarter of its workforce in a short period of time.1 , 2 Telemedicine can improve workforce efficiency and provide competitive advantages. Telemedicine can allow urologists to care for patients in larger geographic areas, addressing the scarcity of urologists in rural areas; 62.2% of US counties have 0 urologists, limiting patient access to necessary care. Even large urology groups may be challenged to meet the demands of their contracted population in urban and suburban markets but may use telemedicine to meet these challenges. Well after the COVID-19 PHE, telemedicine can allow urologists to leverage their subspecialty expertise across a larger population, allowing them to reach patients in need of their services whom they otherwise would not be able to encounter in person. Additionally, groups that adopt telemedicine for their patients’ convenience will meet rising patient expectations for telemedical access, because many patients will have been introduced to telemedicine during the PHE who otherwise would not have. Traditional sectors of society already have incorporated and transformed a customer-centric focus, including travel, retail sales, and banking. Medicine will as well. Pre-COVID-19 telehealth coverage Prior to the PHE, health care reform, with its emphasis on value, was bringing more attention to the prospect of telemedicine. In many states, private payers were mandated to cover telemedicine services, often on par with office encounters. Medicaid covered telemedicine in almost all states. The Department of Veterans Affairs (VA) had been a true trailblazer in the world of telehealth. In 2019, the VA reported more than 900,000 veterans utilized its telehealth services (235% jump from previous year). Furthermore, the VA announced a 17% increase in televisits, delivering greater than 2.6 million telehealth episodes in 2019.3 Medicare had been very restrictive prior to the PHE, limiting coverage to remote geographies or certain chronic care scenarios but allowing some liberalization through alternative payment models, such as alternative care organizations or bundled payments. At the start of 2020, traditional Medicare and Medicare Advantage plans expanded telehealth coverage, limiting restrictions on patient location requirements and expanding coverage for more diagnosis.4 As of March 1, 2020, during the PHE, traditional Medicare has radically liberalized access to telemedicine across effectively all locations and all conditions. The extent to which this will remain after the PHE has concluded is unknown, but even with a significant retraction, urologists can be well positioned to participate in alternative payment models that would compensate for telemedical services. Types of telemedical services Video Visits A video visit is a live face-to-face electronic audio-video interaction between a provider and patient. Prior to the PHE, the combination of both audio and video was required by almost all payers for reimbursement. Telephone encounters are reimbursed by Medicare during the PHE but the extent to which that will remain is uncertain. Other payers may or may not provide reimbursement for telephone calls. In spite of the limitations of an on-screen encounter, video visits are successfully providing alternatives to traditional visits in a variety of settings, including clinic, office, urgent care, hospital, and skilled nursing facilities. Online Digital Evaluation Services Patients may access portions of their electronic medical record and communicate with their urology care team. Value-added services also include online appointment scheduling, form submission such as history intake questionnaires, and online bill payment. Additionally, a practice can provide patient alerts for preventative services as well as definitions of conditions and guidelines and reminders for care, such as urology cancer rechecks. These services typically were not billable but they brought value to patients and helped satisfy government mandated meaningful use criteria for electronic health records.5 During the PHE, traditional Medicare is reimbursing certain components of these services, known as virtual check-ins and digital evaluation services, whereby patients may communicate with their providers via text, e-mail, or chart portals, to determine the need for an in office visit or to efficiently address concerns not requiring an imminent office visit. eConsults eConsults allow a urologist to asynchronously answer another provider’s focused questions about the diagnosis or management of a specific patient. The urologist reviews the supporting material from the electronic medical record and provides a formal response to the focused question. Consultations most likely are solicited from large academic centers but also can be solicited from large urology groups. eConsults also can be synchronous (involving real-time interactions similar to video visits) and in this context also are known as video consults. In contrast to video visits, a patient typically is not present during a video consult. eConsults are performed using electronic software and hardware specifically for electronic consultations. For example, a portal called WebSphere (Cisco, San Jose, CA) has been used for eConsults. The response is documented in the medical record of the provider completing the eConsult. eConsults are convenient for requesting providers and patients alike because they provide timely access to specialty expertise without requiring the patient to actually travel to visit with another urologist. Thus, eConsults allow patient visits to be completed in a shorter time frame and streamline the delivery of care. Scheduling also is more flexible with eConsults, especially asynchronous consults that can be completed outside of regular business hours. By using asynchronous eConsults particularly for more straightforward urologic problems, it is anticipated that more time would be spared for synchronous eConsults or traditional face-to-face consults on more complex patients during regular business hours. The concepts used in synchronous eConsults also have been applied to virtual tumor boards. Studies assessing the benefit of eConsults uniformly have demonstrated a reduction in the number of required in-person consults, ranging from 62% to 92%, optimizing patient time and physician efficiency.6 With virtual tumor boards, detailed case presentations are made in the same way as traditional tumor boards. Radiologic studies and histologic findings are reviewed ahead of time with a radiologist and pathologist, respectively. Using this approach, a second opinion for the patient is obtained and the mechanism is a powerful means to keep high standards of clinical care and performance within the control of a private large group practice. Tele-intraoperative Consultation Although adoption is still in its infancy, the intraoperative consultation is perhaps the most innovative and valuable application of telemedicine. The urologist offers electronic consultations regarding intraoperative findings with other surgeons remotely. Tele-intraoperative consultations promise to enhance productivity and surgical quality as this matures. Hung and colleagues7 categorized currently available telesurgical technologies into 4 categories: verbal guidance, telestration, guidance with teleassist, and telesurgery. Verbal guidance is simply a 2-way communication between surgeon and consultant with video monitoring, the benefit being it is easily implemented with low cost and minimal bandwidth requirements. Telestration allows a consultant or mentor to telestrate in 2-dimensions or 3-dimensions, aiding the surgeon through various aspects of a procedure. TIMS Consultant (TIMS Medical, Chelmsford, Massachusetts), an interactive video broadcast network within hospitals, such as Kaiser Permanente, provides high-resolution, live video streams anywhere remotely, including compliant smartphones and laptop computers. Remote multiparty interactivity and collaboration are provided through integrated audio conferencing and live telestrations. Teleassists allow for the mentor or consultant to reach into the patient remotely and physically aid in a procedure. Lastly, telesurgery allows for the entire surgery to be completed remotely.7 Sterbis and colleagues8 performed 4 porcine radical nephrectomies utilizing the daVinci robot (Intuitive Surgical: Sunnyvale, CA) from greater than 1300 miles away. Telementoring and Teleproctoring A urologist can serve as a mentor and/or proctor during a telesurgical procedure creating telementoring and teleproctoring services.9, 10, 11, 12 This has broad implications with licensure and credentialing in the practice of urology. Telementoring and teleproctoring can address cumbersome and impractical barriers posed by physical proctoring. Hinat and colleagues,13 in 1 of the first reported series of telementoring and teleproctoring in robotics, used a telementoring system to promote surgical techniques associated with robotic-assisted radical prostatectomies. The group demonstrated proper function and acceptable latency with no differences in surgical outcomes, incline operative times, complication rates, early continence status, and positive margin rates between the telementoring and direct mentoring groups. Telesimulation and Telesurgical Rehearsal Simulators now are being used to teach minimally invasive surgical techniques. A network of simulators can assist teaching and evaluating novice surgeons and those who desire improvement. Simulators can standardize teaching and as well allow for interactive proctoring during the simulated procedures. A surgical dress rehearsal may be possible before the actual operation. Currently, urology patient–specific simulations are in development that could be integrated into established training programs rapidly and easily. Telemedicine: Systems and Procedures As with traditional care delivery, telemedicine can be enhanced by following standard operating procedures. General and specialty medical societies currently are generating guidelines for telemedicine that include technical instructions and ethical considerations. Guidelines also increase the delivery of high-quality and safe patient care, key themes needed for telemedicine to be supported by legislators and payers. Prior to the PHE, telemedicine was required to be delivered using secure Internet-based videoconferencing technologies. During the PHE, encrypted platforms still are encouraged but not required by Medicare when their unavailability creates a barrier to access. Non–Health Insurance Portability and Accountability Act (HIPAA) compliant platforms, such as FaceTime, are allowable but public facing platforms, such as Facebook Live, are not. Furthermore, during the PHE, telephone encounters may be conducted for reimbursement on par with video encounters. Nonetheless, after the PHE, telephone-only encounters may or may not continue to be reimbursed and HIPAA requirements for secure connections likely will return. It is advisable to select a secure video platform for a long-term telemedical strategy. The network used for telemedicine typically is a secure virtual private network, with software that typically is licensed to a host institution. Videoconferencing with encryption software can be downloaded to connect with patients directly in their own homes or other noninstitutional settings. Internet-based platforms also can be used as an alternative portal for urologists and patients seeking telemedicine services. Telemedical encounters also may be initiated by patients via direct-to-consumer telemedical companies, such as Hims and Roman, which specialize in select male health concerns. For practices considering implementation of video visits, it is useful to consider the urologic diagnoses for which video visits will be most effective. The experience with many urologists during the PHE is that most, if not all, diagnoses can be managed in part or in whole with telemedicine. Telemedicine need not be an either/or proposition, such that a visit that is conducted telemedically does not preclude a prompt follow-up in-person visit when deemed necessary. Informed consent is established prior to the start of the video visit. The consent typically is conducted in real time following laws within a patient’s jurisdiction. The provider should document the consent in the medical record. The consent should include a discussion about the structure and timing of services, record keeping, scheduling, privacy, risks, confidentiality, mandatory reporting, and billing. Confidentiality and the limits of confidentiality in electronic communication also should be discussed. It also is important that the issue of video recordings be discussed. Specifics regarding technical failure of the video visit, protocols for contact between sessions, and conditions upon which the video visits will be terminated in lieu of a traditional visit also need to be established. Video visits need to be carried out in an appropriate environment for both the provider and the patient to maximize privacy. Video cameras and lighting should be optimized for both the patient and provider during a video visit. If a patient attempts to carry out a video visit in a public space, the provider should recommend that the consultation be delayed until a suitable private space is identified. The consultation should start with identity verification of both the provider and patient. In many instances, a host clinic may perform the verification prior to starting the video visit. The location of the provider and the patient also should be established during the video visit. Contact information for both the provider and the patient should be verified during the video visit. Lastly, the expectations regarding the video visit and any subsequent visits should be discussed. The urologist must make an entry in the medical record in a fashion similar to that for traditional visits once the video visit is complete. The medical record entry should include an assessment and plan, patient information, contact information, history, informed consent, and information regarding fees and billing. As part of the documentation, it also is important to note that the patient was seen using telemedicine technologies. Regarding connectivity, audio-video telemedicine services can be provided through personal computers or mobile devices that use Internet-based videoconferencing software programs. A bandwidth of 384 kilobits per second or higher in both the downlink and uplink directions is recommended.2 Because different technologies provide different video quality results at the same bandwidth, each endpoint should use a bandwidth sufficient to achieve at least a minimum of 640 pixels × 360 pixels resolution at 30 frames per second. Each party should use the most reliable connection to the Internet during the video visit. Next, it is important to verify the patient has the required hardware and software capabilities for a video visit and sufficient broadband connectivity. The patient should be provided contact information for technical support in case troubleshooting is required. Rather than have the patient manage the requirements of the video visit, another option is for the patient to report to a telemedicine center where the hardware, software, and connectivity are provided and standardized for maximum reliability. The increasing availability of 5G networks throughout the country will have a significant impact on telehealth availability as adequate information technology infrastructure will be available to remote patients and clinicians; 5G wireless ecosystems will continue to grow, given both regional and national initiatives from network and wireless providers. Compared with 4G, 5G can be expected to be 100-times faster, with 25-times lower lag times and 1 million devices supported in 1 square mile. The 5G systems will allow for reliable, faster connections, resulting in high-quality video connections and data transfer.14 Efforts should be taken to make audio and video transmission secure by using point-to-point encryption that meets recognized standards. Currently, Federal Information Processing Standard Publication 140-2, is the US Government security standard used to accredit software encryption and lists encryption types, such as advanced encryption standard, as providing acceptable levels of security. When patients or providers use a mobile device, special attention should be paid to the relative privacy of information being communicated over such technology. Mobile devices should require a passcode and should be configured to have an inactivity timeout function not exceeding 15 minutes. Gaps to access With the emergence of telemedicine in urology came socioeconomic inequalities in care. Many patients do not possess the basic technology required for a robust telehealth visit. In the initial PHE rules, telephone encounters were not included creating an access gap for those who did not have adequate Internet, smartphones, or computers. On March 31, 2020, CMS allowed for telephone services to be covered during the PHE (Current Procedural Terminology code 99441-99443), including creating parity between telephone and televideo visits. This bridged a significant access gap for those unable to perform video visits. Prior to the PHE, coverage and reimbursement for telephone calls were severely limited. G2021 Healthcare Common Procedure Coding System (HCPCS) code (brief communication technology-based service) seldom was used and had limited reimbursement. Although telephone coverage expanded access, investigators are assessing its adequacy as a telemedical platform. The utilization of telephone calls was studied by Safir and colleagues,15 who compared telephone with face-to-face encounters for hematuria consults in the VA population. They found access improved from 72 days to 12 days, although overall satisfaction with visit was higher in the face-to-face visit cohort (92% VS 84%). The launch of telemedicine in the Bronx, New York during the PHE offers a glimpse into its utilization and limits in a socioeconomically disadvantaged population. Montefiore Medical Center, one of the largest and diverse hospital systems in the country, transitioned almost overnight to 95% telehealth visits. This was in a population where approximately 50% of households did not have adequate Internet access, 20% preferred non-English language, 44% relied on Medicaid, and 40% lived in poverty. The group found 88% satisfaction with video visits and 81% for telephone-only encounters; 67% of patients felt they received similar care as in office visit whereas 79% would choose a telemedicine encounter. Similar to previous research, clear travel time and clinic wait time advantages were seen for telehealth. Lastly, there was a preference for phone visits compared with video visits, a majority citing technologic limitations as the reason telephone is preferred.16 This small study highlights the need for continued expansion of telephone-only encounters. Although CMS has recognized telephone services as telehealth by adding them to Medicare telehealth services, the degree of coverage beyond the PHE has yet to be seen. CMS understood the importance of this technology because it allowed for social distancing and expanded access for underserved populations. As the new 2021 CMS rule currently stands, telephone-only encounters will not continue after the PHE. CMS, however, did introduce G2252 HCPCS code, which is a virtual check-in for established patients with incremental times to account for differing time spent on the phone. This will lead to a significant cut in relative value unit per time spent with patients on phone. In addition, this cannot lead to an in-person visit within 7 days and may not result from a recent visit. New 2021 Centers for Medicare & Medicaid Services rules Several changes in the 2021 CMS rules will have a long-lasting impact on providers and patients utilizing telemedicine. The most significant change in the 2021 rules is the emphasis on medical decision making (MDM) when determining levels of service. No longer are particular history and physical examination components required; rather, providers determine level of service on MDM alone. The history and examination are left to the discretion of the provider to perform what is medically appropriate. By eliminating the physical examination requirements, providers can focus on MDM and be reimbursed for their work based on the complexity of the patient. This is important especially when utilizing telemedicine because the ability to perform a comprehensive physical examination is limited. By way of example, a new patient with metastatic prostate cancer with multiple complicating factors can be billed at the same level as an in-person encounter because CMS no longer requires the detailed physical examination previously required. Furthermore, for time-based billing, the total time now includes previsit and postvisit preparation and coordination. For many providers, this better represents the amount of time it takes to care of patients. Although the telehealth visit itself may take only 10 minutes, CMS now reimburses for time spent reviewing imaging, records, and coordinating care at the conclusion of the visit. Evaluation and Management Changes During the PHE, the list of covered telehealth services was expanded significantly. Some of these are set to expire after the PHE whereas others were added to the permanent covered services. A list of services covered through 2021 is in Table 1 .Table 1 Current procedural terminology codes for services covered through 2021 Rest home visits 99336-99337 Home visits 99349-99350 Therapy services 97161-97168, 97110, 97112, 97116, 97535, 97750, 97755, 97760, 97761, 92521-92524, 92507 Critical care codes 99469, 99472, 99476, 99478-99480, 99291-99292 Discharge codes 99315-99316, 99238-99239, Observation management 99217, 99224-99226 Emergency department 99281-99285 For the time being, many of the restrictions associated with originating sites will revert back to pre-PHE rules when the PHE expires. This means that services covered on the PHE list, including evaluation and management codes, will be restricted to those in rural areas (health shortage regions) and at approved facilities. Legislation surrounding the originating site is expected to change because there is a strong push to permanently eliminate originating site restrictions. One area of particular interest is the status of remote patient monitoring and its reimbursement. This has the potential to expand the diagnostic reach to patients in more rural areas of the country bridging significant geographic disparities in care. The final rule states that remote patient monitoring can be used for established patients but must be a Food and Drug Administration–approved device with certain data collection requirements. Advanced Practice Providers The role for advanced practice providers (APPs) has expanded exponentially over the past decade. There is a projected shortage of physicians, ranging from 46,000 to 121,000 by 2032. Urology is likely to be plagued greatly by this shortage given it is the second oldest surgical specialty, with more than 18% of its workforce greater than 65 years old.2 Although increasing the role of APPs potentially can address this workforce shortage, adequate supervision is imperative. Technology-based supervision has the potential to expand access with an increasing APP workforce while simultaneously providing adequate and efficient supervision. During the PHE, CMS has allowed for supervision of APPs utilizing audio and visual technology. There are new CMS rules surrounding direct supervision of APPs. Direct supervision traditionally meant that a physician must be in the same office and immediately available to assist. During the PHE, direct supervision is able to be accomplished through combined audio-video communication. Audio alone does not satisfy the requirement. This allows APPs to perform procedures and see patients without the overseeing doctor being physically in the office. This has been extended until December 31, 2021, or the end of the PHE, whichever is later. Disparity in telemedicine care With the accelerated rise in telehealth adoption, many healthcare professionals are concerned that this has resulted in an unequal distribution of health care resources, further widening racial and socioeconomic disparities. At its inception, telemedicine was meant to expand the health care reach to underserved populations and those living in rural areas. With telehealth companies focusing on well-resourced patients in order to expand their market presence, the concern is that those for whom telehealth initially was intended will be left behind. Furthermore, relying on telehealth algorithms for care risks magnifying disparities because underrepresented populations often are not included in algorithmic data. Few studies have assessed the socioeconomic impact of the rapid changes in telehealth that have occurred under the PHE. Weber and colleagues,17 a group from one of the largest health care systems in New York City, evaluated telehealth utilization for COVID-19-related care at the height of the initial COVID-19 wave. They found that compared with whites, blacks and Hispanics were more likely to go in person to emergency rooms and in-office visits rather than utilizing telehealth services. Similarly, patients greater than 65 years old utilized in-person emergency room and office visits at higher rates than their younger cohort. Similarly, Eberly and colleagues18 reviewed records of 150,000 patients who scheduled telemedicine visits at the beginning of the pandemic (March 2020–May 2020). They found that 54% of patients followed through with their visit. Furthermore, they found disparities in utilization for age, race, and income. They found that when comparing utilization of video visits versus telephone encounters there was a lower video utilization in women, blacks, Hispanics, and low-income families.17 , 18 Some reasons for this inequality include technology barriers for older patients, language barriers, Internet constraints in lower-income patients, and lack of adaptations for those with disabilities (ie, visually and hearing impaired). Because there may be benefits unique to video visits as providers are able to examine patients visually and pick up on various visual cues, it is important for physicians, hospitals, politicians, and insurers to ensure equal representation in all telehealth services. Contemporary (public health emergency) utilization The emergence of COVID-19 and the subsequent PHE propelled telemedicine into the future. Urologists were quick to adopt telemedicine to facilitate social distancing, continue care for their patients, and keep practices financially viable. Urology saw a dramatic acceleration of utilization, with 71.5% of urologists stating they participated in telemedicine during the PHE according to the 2020 AUA annual census data. The most common telemedicine topics were benign prostatic hyperplasia, elevated prostate-specific antigen, erectile dysfunction, stone disease, and voiding dysfunction. Approximately half of urologists provided telemedicine encounters for new patients and 77% provided encounters for established patients. According to census data, of those urologists participating in telemedicine, urologists receive compensation primarily for video visits (93.9%) and telephone calls (77%) whereas fewer than 11% reported receiving compensation for eConsults, video visits with other providers, and text messages. Summary Just as Reed Hastings adapted to the changing landscape of movie rentals, so has the field of urology adapted with the rapid emergence of telehealth. Telemedicine appears positioned to be a mainstay of health care systems beyond the PHE and urologists are well positioned to pioneer the expansion of services. In addition to meeting the current demands for social distancing arising from the pandemic, telemedicine specifically may help address the needs of underserved communities by addressing workforce shortages. It also will prove instrumental to increasing clinical and surgical productivity, improving patient access to care, and facilitating data quality reporting. It is crucial to focus on those who benefit most from this new technology (ie, underserved populations) and ensure equal access. As the initial limitations have been radically removed and as new solutions are developed both in the technology and regulatory sides, it is possible that telemedicine will become completely integrated into urologic training and health care delivery to fulfill its promise of access and quality urologic care. Disclosure The authors have nothing to disclose. ==== Refs References 1 Jennifer M. Ortman and Christine E. Guarneri: United States population Projections: 2000 to 2050 2009 United States Census Bureau Available at: https://www.census.gov/content/dam/Census/library/working-papers/2009/demo/us-pop-proj-2000-2050/analytical-document09.pdf 2 Nuewahl: AAMC projection through 2025 2012 AAMC Available at: https://www.aamc.org/media/45976/download 3 Anon: VA reports significant increase in Veteran use of telehealth services 2019 United States Department of Veterans Affairs Available at: https://www.va.gov/opa/pressrel/pressrelease.cfm?id=5365#:~:text=WASHINGTON%20%E2%80%93%20The%20U.S.%20Department%20of,telehealth%20care%20in%20FY%202019 4 Anon: Telehealth Coverage. Medicare.gov. Available at: https://www.medicare.gov/coverage/telehealth, Accessed January 21, 2020. 5 Anon: Website. Meaningful use definition & objectives Available at: HealthIT.gov 2015 HealthIT.gov https://www.healthit.gov/providers-professionals/meaningful-use-definition-objectives Accessed September 13, 2016, Accessed January 21, 2021 6 Modi P.K. Portney D. Hollenbeck B.K. Engaging telehealth to drive value-based urology Curr Opin Urol 28 2018 342 347 29697472 7 Hung A.J. Chen J. Shah A. Telementoring and Telesurgery for Minimally Invasive Procedures J Urol 199 2018 355 369 28655529 8 Sterbis J.R. Hanly E.J. Herman B.C. Transcontinental Telesurgical Nephrectomy Using the da Vinci Robot in a Porcine Model Urology 71 2008 971 973 18295861 9 Sathiyakumar V. Apfeld J.C. Obremskey W.T. Prospective randomized controlled trial using telemedicine for follow-ups in an orthopedic trauma population J Orthop Trauma 29 2015 e139 e145 24983434 10 Hwa K. Wren S.M. Telehealth follow-up in lieu of postoperative clinic visit for ambulatory surgery: results of a pilot program JAMA Surg 148 2013 823 827 23842982 11 Matimba A. Woodward R. Tambo E. Tele-ophthalmology: Opportunities for improving diabetes eye care in resource- and specialist-limited Sub-Saharan African countries J Telemed Telecare 22 2016 311 316 26407990 12 Canon S. Shera A. Patel A. A pilot study of telemedicine for post-operative urological care in children J Telemed Telecare 20 2014 427 430 25316038 13 Hinata N. Miyake H. Kurahashi T. Novel telementoring system for robot-assisted radical prostatectomy: impact on the learning curve Urology 83 2014 1088 1092 24642077 14 Editorial Team: 5G vs. 4G - A Side-by-Side Comparison Available at: https://datamakespossible.westerndigital.com/5g-vs-4g-side-by-side-comparison/ 2019 Accessed January 21, 2021 15 Safir I.J. Gabale S. David S.A. Implementation of a Tele-urology Program for Outpatient Hematuria Referrals: Initial Results and Patient Satisfaction Urology 97 2016 33 39 27450940 16 Watts K.L. Abraham N. “Virtually perfect” for some but perhaps not for all: launching telemedicine in the bronx during the COVID-19 pandemic J Urol 204 2020 903 904 32519903 17 Weber E. Miller S.J. Astha V. Characteristics of telehealth users in NYC for COVID-related care during the coronavirus pandemic J Am Med Inform Assoc 27 2020 1949 1954 32866249 18 Eberly L.A. Khatana S.A.M. Nathan A.S. Telemedicine outpatient cardiovascular care during the COVID-19 pandemic: bridging or opening the digital divide? Circulation 142 2020 510 512 32510987

33795055

PMC9753100

NO-CC CODE

2022-12-16 23:25:29

no

Urol Clin North Am. 2021 May 12; 48(2):215-222

utf-8

Urol Clin North Am

10.1016/j.ucl.2021.01.006

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Published by Elsevier Ltd. S0140-6736(21)01850-X 10.1016/S0140-6736(21)01850-X Editorial The NHS: the many challenges for leadership The Lancet 12 8 2021 14-20 August 2021 12 8 2021 398 10300 559559 © 2021 Published by Elsevier Ltd. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcAmanda Pritchard has her work cut out. The new Chief Executive of NHS England has taken charge of an organisation reeling from the pressures of the past 18 months, with huge backlogs in care, an exhausted workforce, and the COVID-19 pandemic not yet over. Reforms to the health service are making their way through parliament in the teeth of opposition from union leaders who say the timing is wrong. Pritchard is respected and seemingly well liked. “I am realistic. I am also optimistic”, she said on taking up her new role on Aug 1. Just what are the current realities of the NHS that she faces? The immediate outlook is bleak. The Health Secretary Sajid Javid has warned that waiting lists could potentially reach 13 million patients this year. This figure not only reflects the effects of huge disruptions to care, but also—in a country of 67 million people—suggests the parlous state of population health in the UK. Modelling done by the Institute for Fiscal Studies suggests that even in the best case scenario, waiting lists will rise to 9 million people next year, as people who avoided or did not seek care during 2020–21 return to the system. And that's if the NHS's capacity is substantially expanded above prepandemic levels, at a cost of at least £2 billion per year. Much will hinge on the Chancellor's autumn spending review, and the Government's ability to control COVID-19. The Commonwealth Fund, in its influential Mirror, Mirror 2021 report, demonstrates the consequences of more deep rooted challenges in the NHS. The study, which compares the health systems of 11 high-income countries, shows the UK overtaken by Australia, Norway, and the Netherlands, after topping previous iterations of the study in 2014 and 2017 (the USA was bottom on almost every measure). Difficulties in accessing care is one reason. The UK also ranked ninth for health outcomes. And the country has fallen behind in the coordination of care and in equity. Social care is in a precarious financial state and suffers from a gaping disconnect from the health service and COVID-19 has ruthlessly demonstrated the harms of this disconnect. The Health and Care Bill, currently at committee stage, seeks to undo some of the fragmentation wrought by the Lansley reforms of 2012—lessening the emphasis on competition and bringing local authorities and the NHS closer together. Experts have expressed muted enthusiasm for its aims, but disquiet over the details. Pritchard will have to direct its implementation by an already stretched workforce. Overall, these findings will be of little surprise to readers familiar with health in the UK. COVID-19 might have exacerbated these issues, but they have long affected the health service, driven by consistent underfunding, damaging reform, and a neglect of prevention and the social determinants of health. The danger is that they will worsen. The response must be a renewed, long-term vision for the NHS. Not a drastic overhaul, but sustained and progressive evolution towards a health system that is designed and funded to meet the changing physical and mental health needs of an ageing population with increasing multimorbidities. One that attends to workforce welfare and staffing shortfalls amid low pay and the challenges of Brexit. One that strengthens the long tradition of clinical research within the health system. One that is proactive in dealing with the climate emergency—the NHS, as the country's largest employer and a major carbon emitter, ought to have a large stake in the health agenda of COP26. And one that works as part of a strong broad public sector to address the social determinants of health. The separation of prevention and treatment, combined with a policy of austerity, has enabled a steady decline in public health funding, resulting in poor health outcomes and rising inequalities. The LSE–Lancet Commission on the Future of the NHS, published in May, articulates such a vision, and provides the template for putting it into effect. The NHS now needs the likes of Pritchard to have the courage and determination to make this plan a reality. Her key challenge will be eliciting the proper support from politicians. The NHS has many strengths. The provision of care, free to all, is fundamental to life in the UK. It is admired around the world and much loved at home. COVID-19 has put acute pressure on the NHS. But it has also shown—for those who did not already know—the centrality of a well functioning, equitable, and resilient health system to the wellbeing of individuals, communities, and the entire country. Might Pritchard use this impetus and political capital as an opportunity to bring about the progress the NHS needs? A realist might think it unlikely under the current Government. Proving otherwise will take more than optimism. © 2021 Yui Mok/PA Images/Alamy Stock Photo 2021

34391487

PMC9753118

NO-CC CODE

2022-12-16 23:25:30

no

Lancet. 2021 Aug 12 14-20 August; 398(10300):559

utf-8

Lancet

10.1016/S0140-6736(21)01850-X

oa_other

==== Front Appl Geogr Appl Geogr Applied Geography (Sevenoaks, England) 0143-6228 0143-6228 The Authors. Published by Elsevier Ltd. S0143-6228(21)00140-5 10.1016/j.apgeog.2021.102524 102524 Article The city turned off: Urban dynamics during the COVID-19 pandemic based on mobile phone data Romanillos Gustavo a∗ García-Palomares Juan Carlos a Moya-Gómez Borja b Gutiérrez Javier a Torres Javier c López Mario c Cantú-Ros Oliva G. c Herranz Ricardo c a tGIS, Department of Geography, Universidad Complutense de Madrid, Spain b TRANSyT, Universidad Politécnica de Madrid, Spain c Nommon Solutions and Technologies, Spain ∗ Corresponding author. tGIS, Department of Geography, Universidad Complutense de Madrid, C/ Profesor Aranguren, s/n, Ciudad Universitaria, 28040, Madrid, Spain. 28 7 2021 9 2021 28 7 2021 134 102524102524 10 10 2020 17 6 2021 23 7 2021 © 2021 The Authors 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Due to the rapid expansion of the COVID-19 pandemic, many countries ordained lockdowns, establishing different restrictions on people's mobility. Exploring to what extent these measures have been effective is critical in order to better respond to similar future scenarios. This article uses anonymous mobile phone data to study the impact of the Spanish lockdown on the daily dynamics of the Madrid metropolitan area (Spain). The analysis has been carried out for a reference week prior to the lockdown and during several weeks of the lockdown in which different restrictions were in place. During these weeks, population distribution is compared during the day and at night and presence profiles are obtained throughout the day for each type of land use. In addition, a spatial multiple regression analysis is carried out to determine the impact of the different land uses on the local population. The results in the reference week, pre-COVID-19, show how the population in activity areas increases in each time slot on a specific day and how in residential areas it decreases. However, during the lockdown, activity areas cease to attract population during the day and the residential areas therefore no longer show a decrease. Only basic essential commercial activities, or others that require the presence of workers (industrial or logistics) maintain some activity during lockdown. Keywords COVID-19 Lockdown Spatio-temporal demographics Mobile phone data Land use Urban geography Madrid (Spain) ==== Body pmc1 Introduction Different pandemics have altered the rhythm of cities, even in recent years (Hanson, 2006, p. 232). However, none have impacted modern living as much as the current COVID-19 pandemic. This pandemic has suddenly changed the way in which citizens interact, move or make use of different urban activities. The change has been radical. In particular in the early phases of the pandemic, with the adoption of the most severe measures and the lockdown, which has led to the closure of most activities and changes in habits when carrying out the most basic activities. Without any warning, cities were forced to slow down, reduce and even stop much of their activity for months. Knowing how the pandemic has transformed urban dynamics and what the patterns of these dynamics are in the phases of lockdown and subsequent restrictions is essential for decision-making, establishing new measures or evaluating their effectiveness in preventing and controlling the spread of the pandemic and in understanding the city's resilience to these measures to contain severe outbreaks. Big Data obtained from geolocated devices provides valuable spatial and temporal information to evaluate measures implemented to prevent and control of the spread of the pandemic (Zhou et al., 2020). In particular, due to the heterogeneity and large size of the sample as well as the high temporal granularity, anonymous mobile phone records constitute an excellent source of Big Data for the analysis of the distribution of the population throughout the day. Each user's activity records allow the reconstruction of their spatio-temporal trajectories, differentiating between the time they remain in one place and that taken to move between places (trips). This information is crucial for the analysis and modelling of the spread of the disease. The possibilities offered by new geolocation technologies to study population mobility and the possible spread of contagious diseases are well known (Sirkeci & Yucesahin, 2020; Ferretti et al., 2020). Mobile phone data had only been rarely used in epidemiological research, but their enormous potential has been demonstrated during the COVID-19 pandemic. Despite the short time that has elapsed since the beginning of the pandemic and the restricted access to these data, researchers and governments have started to collaborate with private companies, particularly mobile network operators and location intelligence companies, to estimate the effectiveness of the control measures in a number of countries, including Austria, Belgium, Chile, China, Germany, France, Italy, Spain, United Kingdom, and the United States (Oliver et al., 2020a). Oliver et al. (2020b) have reviewed how mobile phone data can help to tackle the COVID-19 pandemic. In their work, they classify the investigations that have used mobile data according to the type of questions they are trying to answer or according to the actual time at which they appear during the pandemic. In general, studies using mobile phone data have focused, on the one hand, on the analysis of the population's mobility patterns during the different phases of the pandemic. These studies evaluate the follow-up and impact of the lockdown in different countries according to the number of people who stopped traveling during this period and, therefore, remain at home (see Badr et al., 2020; Gao et al., 2020; Lai et al., 2020; Kraemer et al., 2020; Pepe et al., 2020; Pullano et al., 2020). In some cases, sociodemographic differences have been further examined in the monitoring of the measures (Bushman et al., 2020) or in the analysis of the temporal changes in human mobility behavior, social contact rates, and their correlations with the transmissibility of COVID-19 (Paez et al., 2020; Yabe et al., 2020). On the other hand, another group of works has used mobile phone data to analyze the spread patterns of COVID-19 and build predictive models on the expansion of the pandemic. Mazzoli et al. (2020) or Sun et al. (2020, p. 13860) have used the mobility data obtained from mobile telephones to analyze and model how the epidemic has spread in Spain and the United States, while Peixoto et al. (2020) use mobility data to model future scenarios in the possible expansion of the pandemic in Brazil or Aleta et al. (2020) to model possible second-wave scenarios once restrictions have been lifted. The objective of this paper is to study the impact of the pandemic on the dynamics of the city throughout the day and its spatial relationship with land uses, an aspect that the authors believe has not as yet been discussed in depth. The presence of the population in each area of the metropolitan area of Madrid (Spain) is calculated throughout the day using information from mobile phones. This daily distribution of the population is analyzed for a typical week, taking as reference the period between February 14 and 20, 2020, and is compared with the daily distributions in the weeks of confinement decreed due to the state of emergency in Spain. During this lockdown period, we also analyze the effects of the different phases, where measures to restrict the mobility of the population and the opening of the different activities have been tightened or eased. In order to analyze urban dynamics, mobile phone data were crossed with the distribution of land uses within each transport zone. Typical hourly activity profiles have been obtained for each land use and multiple regression models (OLS and spatial models) have been calculated for four major moments in time (morning, afternoon, evening and night). The methodology implemented is similar to that used in García-Palomares et al. (2018), where urban dynamics were analyzed through Twitter activity in each area of Madrid. However, in our case, it is applied to the lockdown and containment scenarios in response to the COVID-19 pandemic. Furthermore, spatial regressions have been used to improve the quality of the models and mitigate the problems of spatial autocorrelation in the distribution of the residuals. The analyses carried out show the level of activity throughout the day that each type of land use has maintained according to the degree of restrictions imposed. This paper contributes to the literature in several ways. First, unlike most previous studies, a highly detailed spatial scale is used. This article does not focus on the impact of the measures in the study area as a whole, but rather analyzes the distribution of their impact according to the type of land use within each zone of the study area. Second, this article uses high temporal detail. Normally previous works analyze the impacts of lockdown on total daily mobility. Here we have analyzed how the type of measures imposed are reflected in the changes in the hourly distribution of the population present in each zone, week by week. Third, in this article data on the hourly distribution of the population are crossed with the distribution of land uses, to analyze the impact of the measures on the temporal dynamics of the different urban activities. Only Google's COVID-19 Community Mobility Report1 performs a similar analysis, but with aggregations in large territorial units (regions and countries). However, here we have gone further. Mobile phone data have been used to reconstruct the trips between transport zones (whose area is usually of a few m2) and crossed with the information on the distribution of land uses of the Cadastre within these transport zones. The crossing of information between mobile phone and land use data has allowed us to carry out regression analyses to take advantage of the high level of detail of the land use data, in addition to temporal distribution profiles for each use. Profiles have been constructed based on the dominant land use in each area, since mobile phone data do not have sufficient spatial resolution to determine the land use by each resident present at any given time of the day. We know the size of the population present in each transport zone in each time zone, but not their exact location (land use) within each transport zone, since most of these zones are used for different purposes. However, the regression analysis allows us to obtain an accurate calculation of the weight of each form of land use by the population present in the transport zone and during each week of the lockdown. The selection of the Madrid metropolitan area as a case study is also of special interest, given the high impact that the disease has had. Spain has been one of the countries most affected by the pandemic, with rates of confirmed cases and deaths among the highest in the world: more than 270,000 total cases at the end of July 2020 (Johns Hopkins University, 2020), which means almost 6000 cases per million inhabitants. Madrid has been the most affected metropolitan area. In addition, it was one of the first affected areas in Europe to establish a lockdown and has also witnessed various phases in the application of the measures. These measures were strictly respected by citizens during the weeks of study, partially due to a remarkable level of police surveillance across the city. Because of this, the study allows us to evaluate the impact of different types of measures and serves as a reference in the evaluation of the same. Although this paper does not study the relationship between the mobility restrictions and the control of the pandemic spread, it is important to highlight the overall effectiveness of the severe lockdown measures to contain the unexpected explosive situation of the first wave. The Madrid Region reached a peak of almost 3500 new COVID-19 cases per day during the last week of March, while the reported new cases in mid May were around 300 per day (Consejería de Sanidad de la ). The remainder of this paper is structured as follows. Section 2 describes the study case, including the data and the methodology used. Section 3 describes results, and Section 4 contains the conclusions. 2 Case study, data and methodology 2.1 Study area and phases The selected study area covers the municipalities of the Morphological Urban Area (MUA) (ESPON, 2014) of Madrid that are located within the Region of Madrid. With an extension of 202,478.46 Ha, the study area enables us to analyze Madrid's behavior on a metropolitan scale, and study in detail what happens in each of the 1062 transport zones into which it is divided (Fig. 1, Fig. 2 ). Just over 5.7 million people reside in the metropolitan area of Madrid according to the 2019 census, and its population increases to almost 5.9 million people in the morning hours due to the balance of people commuting to and from outside the metropolitan area.Fig. 1 Land Use and Distribution of population (in the reference week at night) in the study area. Fig. 1 Fig. 2 Layout of the video-visualization representing the variation of population density according to time slot for the reference week (W0) and the second week of the lockdown (W2). Fig. 2 Regarding the time frame, the research analyzes the impact of the COVID-19 pandemic on the distribution of the population in the study area over 6 weeks (March 23 - May 10, except Easter). In these weeks, the Government of Spain had activated the State of Alarm prior to the adoption of the Transition Plan to the New Normal. They were the weeks of greatest restrictions, with various measures to regulate activities in the different phases. Additionally, the analysis extends to the week of February 14–20, 2020, taken as a reference (W0), representing the distribution of the population in a normal week, prior to the pandemic. The weekly analysis allows us to study the impact of the different measures decreed by the government on mobility and the degree of confinement of the population. To understand the results obtained, the phases of the lockdown decreed by the Government of Spain and the most important measures established in each of them (Table 1 ) must be defined. Table 2 shows the dates of the study weeks, relating them to the phases and measures indicated in Table 1.Table 1 Phases, dates and measures adopted by the Spanish Government during the lockdown. Table 1Phase Dates (2020) Summary of measures Declaration of a State of Alarm 14–28 March ● Suspension of face-to-face classes in all learning centers. ● Prohibition to circulate in the streets, except for: Buying food or medicine, going to health centers, going to or coming from the workplace, going to banks or insurance companies, taking care of the elderly or children. ● Recommendation of teleworking (most companies where it was possible adopted this recommendation) ● Closure of most premises, shops and businesses. Exceptions: Food stores, pharmacies, medical centers, gas stations, and others. ● Closure of museums, libraries and leisure or sports centers. ● The public transport service is maintained, with exceptional measures depending on the specific service. Extension of State of Alarm 1 29 March - 12 April This is the phase with the greatest restriction of activities. The measures adopted during the State of Alarm also include:● Suspension of non-essential face-to-face work activity. Fundamentally, the following are considered essential activities: health, food and fuel distribution, public maintenance services, cleaning and waste collection, state security, postal services, funeral services and the media. Extension of State of Alarm 2 13–26 April Measures relating to those defined in the previous phase:● People are allowed to return to their workplaces for non-essential activities where teleworking measures cannot be implemented. ● Circulation of private vehicles is allowed to carry out the permitted activities. Extension of State of Alarm 3 27 April - 10 May Measures relating to those defined in the previous phase:● Children under 14 years of age may go out with someone 1 h a day but must not go further than 1 km from home. ● From 2 May: Those defined by the Transition Plan to the New Normal. Transition Plan to the New Normality: De-escalation - Phase 0 2–10 May On April 28, a 4-phase Transition Plan to the New Normal was established. On May 2, Madrid enters Phase 0 of the plan, allowing:● Departure for minors, individual non-contact sports activities and walks, once a day and at regulated hours. ● Opening of establishments by appointment for individual customer service. Table 2 Study weeks, dates and correspondence with the State of Alarm phases. Table 2Study weeks Dates Corresponding phase W0 14–20 February Reference week. Normal activity prior to COVID-19, before the state of alarm W1 23–29 March Second week after the Declaration of the State of Alarm W2 30 March - 5 April First week of Extension of State of Alarm 1 W3 13–19 April First week of Extension of State of Alarm 2 W4 20–26 April Second week of Extension of State of Alarm 2 W5 27 April - 3 May First week of Extension of State of Alarm 3 W6 4–10 May Second week of Extension of State of Alarm 3 and First week of Transition Plan to the New Normality: Phase 0 The results for the reference week (W0) are shown in all analyses. For reasons of space, sometimes only the results for weeks W1, W2, W4 and W6 are shown, which are a good reflection of behavior in the different phases of the State of Alarm. In other cases, the comparison is made between the reference week (W0) and the week with the greatest restrictions (W2). 2.2 Data sources and data preprocessing 2.2.1 Data sources The data sets on which this study is based are described below:1. Mobile phone records. The data used for the extraction of mobility indicators consists of a set of anonymized mobile phone records corresponding to the defined weeks of study, obtained through a collaboration agreement with one of the three main Mobile Network Operators (MNOs) in Spain, with a market share of more than 20 %. The homogeneous penetration of the MNO in virtually all socioeconomic groups of the population, together with the size of the sample, grants a good representativeness of the whole Spanish population. The records include Call Detail Records (CDRs), produced every time a mobile phone interacts with the network through a voice call, a text message or an Internet data connection, as well as passive events coming from network probes. Among other information, each record contains an anonymized identifier of the user, a timestamp and the position of the tower to which the device is connected at that particular moment. This provides an indication of the geographical position of the user at certain moments along the day. The registers do not provide the exact location of the users. This typically provides an accuracy of dozens or hundreds of meters in urban environments, and up to a few kilometres in rural areas, where the mobile network is less dense. The temporal resolution of the records depends on the frequency of use of the mobile device; most users typically generate a register at least every 15–20 min. 2. Land Use data. Land use data provided by the Directorate General for Cadastre in Spain (Cadastre), by built entity of the study area. The databases define the surface area [m2] of each type of land use. These data are updated every 6 months and the data set used corresponds to the update of January 24, 2020. Fig. 1 represents the transport zones of the study area according to this classification of predominant land uses (see section 2.4). 3. Population Data. Census data for 2019 at the census section level, obtained from the National Institute of Statistics. The data has been aggregated at the transport zone level, and it has been used as the sampling frame for expanding the sample of the MNO customers. Fig. 1 shows the population distribution in the study area according to the Register. 4. Territorial boundaries. The demarcation of the Morphological Urban Area (MUA) of Madrid has been obtained from the ESPON DATABASE project. Only the municipalities belonging to the Region of Madrid have been considered for this study. The transport zones defined in Madrid have been obtained from the Open Data Portal of the Consorcio de Transportes de la Comunidad de Madrid.2 5. Data on State of Alarm phases and measures. They come from the Royal Decree of the Ministry of the Presidency of the Government of Spain published in the Official State Gazette. 2.2.2 Phone data preparation The extraction of activity and mobility information from mobile phone records consists of the main following sub-processes:1. Data pre-processing and cleansing: mobile phone registers are pre-processed to ease their storage and management. An integrity analysis is also performed to filter out errors in the raw data, in order to ensure the quality of the results. It is important to detect and fix these errors to avoid the incorrect detection of trips. 2. Sample selection: an effective sample is built by selecting only those users with enough mobile phone activity such that it is possible to reconstruct their mobility and activity patterns with an adequate level of accuracy and reliability. In addition, the data have been processed to discriminate the trips made on a recurring basis by potential transport professionals, i.e. those who make more than 6 trips a day and travel over 50 km, than those made by travellers. 3. Extraction of activity-travel diaries: an “activity” is defined as an interaction or set of interactions with the environment that takes place in the same location and motivates an individual to move there. A “trip” is defined as a sequence of one or more displacements (“stages” or “legs”) between two consecutive activities. This way, a trip has a main purpose determined by the activity at origin and/or the destination. Different criteria based on stay times, itineraries and longitudinal behavioural patterns are used to identify activities, trips, intermediate stops subordinate to the trip and the different stages or legs of a trip. The result of this process is the sequence of activities and trips performed by each user in the sample for the period of study. The information associated to each activity includes its location, the start and end times, and the type of activity: home, work, study, other frequent activities, non-frequent activities (e.g., based on the analysis of the user's longitudinal behavioural patterns during several weeks/months, the place of residence of each user is identified as the place where the user sleeps more often). Once activity diaries are extracted at an antenna level, a layer of land use information is used to refine the estimation of the user position inside antenna coverage areas. Users are assigned to different areas served by the same antenna through a probabilistic method that takes into account the type of land use (residential, commercial, industrial, etc.). The information associated to each trip includes its origin and destination (i.e., the locations of the preceding and subsequent activities), the start and end times (i.e., the end time of the preceding activity and the start time of the subsequent activity), and the location and the start/end times of the intermediate stops (if any). 4. Expansion of the sample to the total population: in order to extract meaningful mobility indicators, the sample is expanded to the total population of Spain. This expansion is performed at transport zone level. The expansion factor is calculated as the ratio between the number of residents of the district according to the census information and the sample of users with their home location at the given district. This procedure allows the correction of any possible spatial heterogeneity of the MNO's market share. 5. OD matrices generation: In the present study the expanded activity travel diaries extracted from mobile phone records were used to build OD matrices with origin and/or destination in the Region of Madrid. The matrices were segmented by day and start time of the trip, considering 24-time segments. The zoning used for trips aggregation consisted of the 1259 transport zones defined for the Community of Madrid, in addition to 51 external zones that refer to the rest of the provinces of Spain. These data have enabled us to calculate the presence of the population each hour of the day in the 1062 zones of the study area. 2.3 Analysis of the spatial distribution of population according to time slot Population distribution in the study area varies throughout the day as a consequence of the different activities carried out. For analysis purposes, the number of people present in each transport zone for each hour and week was estimated from the O-D matrices. The following criteria were considered:1. A single matrix of hourly trips per week was obtained, in which the average number of trips for each O-D pair is the average of the trips made between Monday and Thursday of that week. 2. It was considered that the number of people in the census corresponds to the number of people present in each transport zone at 02:00, when the lowest number of trips generated in W0 in the study area is observed. 3. The number of people present in each transport zone per hour was estimated as those indicated in the census (situation 02:00 h) plus the sum of average weekly trips of people attracted to that transport zone, between 02:00 and the corresponding time, minus the sum of average trips for the week generated in that transport zone for the same time period. If negative populations are obtained, they become zero. The process takes into account the entire day. An initial exploration of the data was carried out through video-visualization, which represents the evolution of the population density [people/km2] for each time slot in the reference week (W0) and the week of greatest restrictions (W2) (see video available in the supplementary material). The two weeks are shown simultaneously and according to the same symbology, so that a visual comparison can be made. The video also contains an animated graph showing the weekly evolution of the population in each type of urban area according to the basic classification of predominant land use. Secondly, bivariate Ordinary Least Squares (OLS) analyses were performed in order to compare the different population distributions according to time slots for each of the study weeks: Morning (08:00 to 14:00), Afternoon (14:00 to 19:00), Evening (19:00 to 22:00) and Night (22:00 to 00:00). The coefficient of determination indicates the degree of overlap between population distributions, while the regression residual maps show where differences (positive or negative) between time slot distributions emerge. This analysis was focused on differences between the reference time slot (night) and the rest of the time slots for each of the study weeks. These differences are expected to be especially high in the reference week W0 (people move within the city without restrictions) and particularly low during the week of strictest home confinement W2 (most people stay at home). The methodology is based on comparing the presence of people in the different weeks of analysis, according to time slots, for the average of the trips made between Monday and Thursday of that week. The different weeks with mobility restrictions are compared to the week of reference and then, also among themselves, allowing us to evaluate not only the impact of the measures in relation to the normal mobility of the city, but also to the mobility which results from the application of the different measures. The methodology guarantees the comparability of the different weeks and time slots, taking into account that we estimate the presence of people following the same methodology and based on the same dataset. 2.4 Temporal profiles according to predominant land use Temporal profiles of population presence according to predominant land use were calculated for each time slot and study week. The total number of people present in a zone was assigned to the predominant land use in the zone, and then the total number of people according to land use was added up for each time slot in order to obtain the specific temporal profile of each land use in each study week. In order to perform this temporal analysis, the percentage of built-up area pertaining to each land use in each transport zone was calculated based on cadastral data. Firstly, three main types of transport zones were distinguished: residential (when more than 66.6 % of built-up area in the zone is residential), activity (when more than the 66.6 % is non-residential, e.g. offices, industry, retail or education) and mixed residential (all other cases). Secondly, activity (non-residential) areas were classified in 10 types: offices, industry, retail, health, education, culture, entertainment, large transport terminals, parks and others. Fig. 1 shows the predominant land use in each of the transport zones and Table 3 the built-up area per land use category and number of zones as predominant land use.Table 3 Summary statistics of the built-up area per land use category and number of zones as predominant land use. Table 3Land use Built-up area [hm2] Zones as predominant land use Residential 28,872 553 Mixed residential 171 Activity 24,077 338 Total 52,859 1062 Activity Built-up area [hm2] Zones as predominant land use Retail 3100 34 Culture 1441 6 Educational 333 30 Entertainment 181 8 Industrial 4,49 107 Offices 3137 41 Park and sport 4597 29 Health care 825 31 Large transport terminals 54 4 Other 5918 48 Total activity 24,077 338 This analysis is also based on the comparison of the presence of people in the different areas of the city characterized by their main land use, during the different weeks of study. Taking into account the high reliability and accuracy of cadastral data and considering that we estimate the presence of people following the same criteria and based on the same dataset, the methodology followed guarantees the comparability of the different scenarios analyzed. 2.5 Multiple regression analyses With the aim of exploring and quantifying the impact of the different land uses on urban dynamics during COVID-19 pandemic, we have performed Multiple regression analyses. This exploration allows us to analyze the relationship between land use and people presence, overcoming the problem of land use mix within each transport zone. The dependent variable in each model was the people's presence in every transport zone per major time slot (Morning, Afternoon, Evening and Night) and study weeks. The independent variables were the amount of build-up square hectometers of each type of land use in each of the transport zones, based on cadastral data. Distance to the city center was included in the models as a control variable in order to consider the spatial component. Non-significant variables were removed from the models (blank spaces in Table 6, Table 7, Table 8 in section 3.2). In the first step, we performed Ordinary Least Squares (OLS) regressions, and then we analyzed the results of the model and the effect of the spatial dependence using Lagrange Multiplier (lag and error) and Moran Index. The results of Robust Lagrange Multiplier (error) revealed that spatial error was significant at the level of 5 %, while Robust Lagrange Multiplier (lag) values were significant only in one case. Therefore, we eventually performed Spatial Error Models (SEM). We selected the spatial relationship of queen contiguity in SEM because we obtained better fits with this spatial relationship than with longer distances. Here, we present the OLS results for the reference week (W0) and SEM results for all weeks. Statistical analyses were performed using GEODA software. The coefficients of the independent variables indicate the number of additional people present as each independent variable (land use) increases by one unit. The changes in the coefficients throughout the day show the time slots in which each type of land use is most active. If the comparison is made between weeks for the same time slot, the differences in the coefficients express the extent to which the restrictions adopted in each phase of confinement affect each type of land use. The coefficients of land uses with activities that require the physical presence of workers (such as industrial) are expected to experience less variations than those used for other purposes with activities involving teleworking (for example, offices). 3 Results 3.1 Spatial distribution of population according to time slot The visual analysis of the variation in spatial distribution of the population according to time slots in weeks W0 and W2 through video-visualization shows a very clear picture of the impact of the measures restricting mobility and performance of activities established with the decree of the State of Alarm. During the lockdown week (W2), the population variations with respect to night-time distribution are minimal, which is also shown in the animated graph that represents the evolution of the population in each type of urban area according to the basic classification of predominant land use: residential, mixed residential and activity. However, a more detailed visual inspection reveals more significant changes in specific areas of the city, where activity registers a particularly sharp decline (for example in educational, financial or office areas) or where it remains at outstanding levels (some areas of logistics or health). Animated maps locate and identify some of these areas of interest for comparative reading of the results. Fig. 2 shows a screenshot of the video-visualization, which is attached as supplementary material in this paper. The weekly bivariate analysis between large time slots complements the first visual approximation and allows us to obtain a numerical indicator for comparing the different scenarios. Taking night-time as the base period, the differences in distributions of the population throughout the day can be analyzed from bivariate correlations (Table 4 ). Madrid and its metropolitan area have a high mix of land uses, meaning that the coefficients of determination are high in all cases. The biggest differences are between night-time (residence) and morning (activities). On the contrary, between Night-time - Afternoon and, especially, between Night-time - Evening the correlations are very high, because many people have already returned to their areas of residence. The confinement situation makes the correlations between night-time and the rest of the time slots practically equal to 1. Night is reproduced during the day. However, despite how restrictive the measures have been in Madrid since the beginning of the pandemic, the different phases are reflected in the morning and night-time correlations with very high values in the most active closing week (W2) but slightly lower in subsequent weeks (W4 and W6).Table 4 Relationships in the distribution of population according to time slot (R2). Table 4Week Morning 08:00 to 14:00 Afternoon 14:00 to 19:00 Evening 19:00 to 22:00 Night 22:00 to 00:00 W0 Night (22:00 to 24:00) 0.711*** 0.814*** 0.978*** 1 W1 Night 0.987*** 0.996*** 0.999*** 1 W2 Night 0.994*** 0.998*** 0.999*** 1 W4 Night 0.984*** 0.996*** 0.999*** 1 W6 Night 0.976*** 0.993*** 0.998*** 1 ***P Value < 0.001. The mapping of the correlation residuals between night-time and morning in weeks W0 and W2 (Fig. 3 ) shows a very different spatial behavior. In a normal situation (W0), the morning activity spaces become highly active (positive residual in yellow), such as office areas (Points of interest 1 and 8) and mixed areas of the center, industrial areas (Points 4, 7 or 9), large facilities, university campuses (Point 2) or hospitals, as well as transport terminals, such as railway stations or the airport (Point 3). Whereas residential areas have high negative residues (blue color).Fig. 3 Residuals in the bivariate correlations of the distribution of population at night and in the morning. Fig. 3 During the week of greatest restrictions (W2) the intensity of the residuals is very low. Some equipment areas are shut off (for example, Ciudad Universitaria - Point 2) and the intensity of activity is significantly reduced in the central office spaces (Points 1 and 8). On the other hand, some industrial spaces on the periphery now show the greatest deviations (Points 4, 7 and 9), together with strategic logistics facilities, such as Mercamadrid (Point 5). Mercamadrid is the largest wholesale market in Spain, and presents an even greater deviation than in the reference week (W0), which is related to the fact that supermarkets increased sales during the first weeks of the state of alarm. Finally, attention should be drawn to the activity detected in specific points of the city, such as the Feria de Madrid-IFEMA (6), which was converted into the largest emergency hospital in Madrid during the State of Alarm. On the contrary, residential areas tend to lose a large part of their population during the hours of activity during the reference week (W0), especially those with the highest density located to the south and northwest of the central districts of Madrid, but these losses have decreased substantially until there is practically no difference between night-time and morning in the week with the highest restrictions (W2). Although this paper is not focused on evaluating the impact of mobility restrictions on the different socioeconomic populations groups, it is possible to draw some basic conclusions from the obtained results: We can infer that certain professionals were more affected by the restrictions than others. For instance, on the one hand, we can conclude that office workers and education professionals and staff mainly stayed at home teleworking, since, as we previously stated, the intensity of activity was significantly reduced in the central office spaces (Points 1 and 8) and education areas (2) during the week of greatest restrictions (W2). And, on the other hand, we can infer that people working in industry and logistics were the group of professionals less affected by the mobility restrictions, of course along with the health professionals, considering that industrial spaces and strategic logistics facilities show the greatest deviations (Points 4, 7 and 9), along with hospitals (6). 3.2 Temporal profiles according to predominant land use Population distribution according to land use and time slots for the reference week (W0) is shown in Table 5 and Fig. 4 a. Most of the population can be found in residential areas during all time slots. Although residential use is dominant in these transport zones, other activities, mainly commercial, services and equipment can also be found. Many of these areas therefore maintain a high population presence also during working hours (morning and afternoon). During the night time residential areas concentrate 74 % of the population. In the morning, the presence of the population in residential areas falls, but they still concentrate 65 % of the total population. This mix is even more marked in mixed residential transport zones, where the presence of the population increases during these daytime hours. But the zones that show higher fluctuation of population along the day are those classified as activity areas. These areas show a decreasing population trend from morning to night time. The population of the activity areas during the morning doubles the population of these during the night time. This population, however, represents only 14 % of the total population in the morning time and 7.5 % at night. According to the Census data a 7.2 % of the population of Madrid (more than 400,000) is resident of these areas.Table 5 Population distribution in the reference week and use of the predominant one. Table 5Uses Morning 08:00 to 14:00 Afternoon 14:00 to 19:00 Evening 19:00 to 22:00 Night 22:00 to 00:00 Census (02:00)** Residential 3,848,781 3,932,847 4,136,932 4,244,711 4,272,127 Mixed residential 1,251,975 1,221,602 1,135,494 1,082,954 1,061,866 Activities 825,004 725,800 525,987 430,393 415,866 Total 5,891,478 5,847,032 5,767,322 5,727,775 5,719,651 Activities Morning Afternoon Evening Night Census Commercial 62,930 65,727 56,917 40,180 37,027 Cultural 11,214 9309 6011 4534 4054 Educational 66,949 54,536 32,305 24,064 24,298 Shows 18,055 18,415 22,477 17,416 13,194 Industrial 277,921 241,765 172,052 147,700 144,688 Offices 155,203 130,749 69,290 48,758 44,691 Parks and sports 59,288 52,842 39,716 33,269 33,036 Health 71,861 62,169 49,901 43,528 42,760 Transports 18,458 12,957 5638 2325 3314 Others 48,844 44,114 40,589 38,336 38,597 Total 825,004 725,800 525,987 430,393 415,866 Fig. 4 Population distribution profiles throughout the day during the study weeks (W1–W6), including the reference week W0 (a) and without it (b). Fig. 4 Among the activities present, industrial and office areas are the ones with the largest population density. Some of the areas where industrial land is dominant also have residences, where almost 150,000 people live in these areas. These areas also receive a high number of people during working hours (morning and afternoon). During the night time industrial areas concentrate 2.5 % of the total population and office areas the 0,8 %, but in the morning time the presence of the population in these areas grows to represent 4.7 % and 2.5 %, respectively. These distributions with a higher population density in the morning and afternoon are repeated in educational, health or cultural facilities, while commercial activities or parks and sports areas have a significant population density in the evening. Temporal profiles according to large types of land use and their behavior during the weeks of the pandemic are very different from the reference week (Fig. 4, Fig. 5 ). In week W0, the departure of the population from residential areas does not compensate for arrivals, and these spaces lose more than 425,000 people during the morning hours and more than 340,000 in the afternoon. The areas of activity show an opposite profile, with very important gains in the morning (almost 410,000 people) until reaching the peak at 11 a.m. and falling during the afternoon (+310,000). The mixed areas have an intermediate situation, with positive balances both in the morning and in the afternoon, but with less intensity and also a smaller decline in the afternoon. However, during the pandemic, the three curves have tended to flatten, significantly reducing both the negative balances of residential spaces and the positive balances of mixed and activity spaces. Data also show the evolution of the different phases, with a flatter curve in the first and especially the week with the greatest restrictions (W2). With the closure of non-essential activities, imbalances were reduced by up to 85 % in residential spaces in the morning in week W2 (they lost only 63,000 people), by 90 % in the afternoon (with losses of only 34,000) and 100 % at night. These balances are reproduced inversely in the mixed and activity spaces. With the easing of the restrictions, profile curves and balances have been recovering, although activity spaces had a positive balance that represented only 28 % of the usual (W0) in the last week (W6).Fig. 5 Changes in the resident population in the census according to large land uses during the different study weeks (W0 – W6). Fig. 5 The temporal distribution of the population in the activity areas shows different profiles and different behaviors during the study weeks according to the characteristics of their activity (Fig. 6 ). During the reference week (W0), the curves of office or industrial activities are very similar, however their behavior is different during the pandemic. Office workers have been able to implement teleworking to a greater extent, so that the presence of the population in these spaces has been reduced very significantly, to the point of practically flattening the profiles. However, industrial activities require the increased presence of workers, so profiles maintain a steeper curve. Another important factor is that in the industrial space there is a greater difference between weekly profiles (W2) and the rest of the weeks of the pandemic as a result of the closure of non-essential industrial activity during this period.Fig. 6 Population distribution throughout the day in the different activity areas during the study weeks (W0–W6). Fig. 6 Among the rest of the activities, we see how large transport infrastructures (airport and railway stations), educational or health areas are activated especially in the morning, while commercial areas are activated somewhat more in the afternoon, and leisure and entertainment at night-time. However, all areas have significantly flattened their curves, especially in activities that have completely closed, such as education and entertainment, and to a lesser extent commerce, transport and health (Fig. 6). 3.3 Links between land use and population distribution The results of the multiple regression models obtained for the different time slots show the close link between the spatial distribution of land use and the population in all time slots. The adjusted coefficients of determination were very similar and close to 0.7 in all time slots and weeks, and all have significant F-statistic values at the 0.000 level. No collinearity problems appeared in any of the models between the explanatory variables (VIF values less than 2 and tolerance greater than 0.6 in all cases). However, the models showed problems of spatial dependence. The distribution of the residuals was spatially autocorrelated (Moran's I error is statistically significant at the 0.05 level). Robust Lagrange Multiplier (error) was significant at the level of 5 %, while Robust Lagrange Multiplier (lag) values were significant in just one case. Based on the OLS models obtained for the population distributions in a reference week W0, all variables have the expected signs (Table 6 ). During the morning, the residential and activity areas have positive signs (they tend to concentrate the population) and the distance to the center of Madrid has a negative sign (greater intensity of use in central spaces). The highest coefficients correspond to variables such as transport, culture, education and commerce, where there is very intensive use of the land (large population density per unit area), while they were much lower in activities such as offices and, above all, industrial land. Throughout the day the coefficients of residential use are created, and the highest values are registered at night. Meanwhile, the coefficients decrease between morning and afternoon in all activities, except commercial ones, with higher values in evening. Evening and night-time variables such as offices and parks have negative signs, and educational and health uses at night are also negative. This has to do with the high mix of uses in the city and the weight of mixed residential areas, so that in these mixed-use zones the presence of activities reduces the potential presence of the population in these areas. Finally, the coefficient of the distance to the city center variable decreases throughout the day and ceases to be significant at night, showing the center-periphery gradient of activity, since the weight of residential use is greater in the periphery than in the center.Table 6 Regression models (OLS) in the reference week (W0). Table 6 Morning Afternoon Evening Night B Beta B Beta B Beta B Beta (Constant) 1663.2** 1643.8** 1834,2** 1864,2** Residential [Ha] 110.6** 0.697 113.7** 0.725 118,1** 0,744 119,8** 0,737 Offices [Ha] 77.8** 0.122 47.3** 0.90 −46,5** −0,087 −78,8** −0,144 Industrial [Ha] 19.2** 0.069 11.3** 0.041 Commercial [Ha] 111.6** 0.117 125.8** 0133 125,4** 0,131 83,9** 0,086 Educational [Ha] 125.2** 0.083 74.2** 0.50 −70,0** −0,045 Cultural [Ha] 143.9** 0.075 104.1** 0.055 Transports [Ha] 218.1** 0.058 139.6** 0.037 Health [Ha] −49,9* −0,029 Parks and sports [Ha] −2,5* −0,029 −4,0* −0,044 Others [Ha] 80.2** 0.122 71.9** 0.111 69,9** 0,106 69,8** 0,104 Distance [km] −29.7** −0.072 −22.8** −0.056 −12,9* −0,031 No. Observations 1062 1062 1062 1062 R2 0.674 0.690 0.694 0.686 Adj. R2 0.671 0.688 0.692 0.684 AIC 19397.3 19318.1 19324.8 19403.5 Moran's I (error) 6.845** 8.045** 14.334** 17,654** Lagrange Mult. (lag) 24.533** 21.027** 45.903** 68.473** Robust LM (lag) 4.409** 1.175 0.029 0.960 LM (error) 41.527** 57.974** 190.118** 296.986** Robust LM (error) 21.403** 38.122** 144.245** 229.473** LM (SARMA) 45.936** 59.149** 190.150** 297.946** **Significant at the 0.05 level; * Significant at the 0.10 level. The spatial error models in the reference week (W0) during morning and afternoon are very similar to the OLS, with similar equations and slightly better fits (Table 7 ). However, the R2 values increase in the evening and night. The coefficients of residential, industrial and commercial uses have values and trends similar to the results obtained in the OLS models. Nevertheless, during the evening, only land uses with positive signs are maintained and the distance to the center is no longer significant. At night, offices have negative sign, but a much lower coefficient than in the OLS model. The LAMBDA is significant at the level of 5 % in all models and grows from morning to night, showing a greater spatial dependence at night, when the spatial distribution of population is mainly explained by residential use.Table 7 Spatial Error models (SEMs) in the reference week (W0). Table 7 Morning Afternoon Evening Night B Std.Error B Std.Error B Std.Error B Std.Error (Constant) 1771.0** 188.8 1742.9** 189.4 1701.2** 147.5 1751.7** 163.9 Residential [Ha] 111.3** 3.5 114.5** 3.4 119.0** 2.8 121.0** 2.98 Offices [Ha] 76.1** 11.3 52.2** 10.8 −31.6** 10.0 Industrial [Ha] 21.6** 5.3 14.1** 5.0 Commercial [Ha] 104.4** 17.6 114.5** 16.8 99.8** 15.8 74.5** 15.9 Educational [Ha] 119.7** 27.8 75.4** 26.7 Cultural [Ha] 132.3** 35.5 100.6** 33.8 Transports [Ha] 214.2** 64.3 145.5** 61.1 Health [Ha] Parks and sports [Ha] Others [Ha] 61.3** 15.3 51.3** 14.6 36.8** 12.4 43.0** 13.7 Distance [km] −31.1** 11.96 −24.2** 10.2 LAMBDA 0.283** 0.071 0.472** 0.06 0.513** 0.037 0.577** 0.03 No. Observations 1062 1062 1062 1062 R2 0.688 0.710 0.743 0.756 AIC 19362.4 19268.4 19184 19200 **Significant at the 0.05 level; * Significant at the 0.10 level. The Spatial Error Models (SEMs) obtained for the weeks of the State of Alarm show similar coefficients and signs in all time slots, but with some nuances (Table 8, Table 9 ). In general, with the population confined to their homes and numerous activities closed, the equations tend to reproduce the situation occurring during the night of the reference week in all time slots. The residential land has similar coefficients in all weeks and all time slots, with a similar value to the one registered at night in week W0. Business activities, which have maintained a basic level, also show positive signs every week. However, their values are lower than those of night-time in a typical week, even in the morning, and they decrease throughout the day. Educational and park land, which have suffered a major closure, do not enter in the models. Offices have signs and values similar to the night of the reference week. Only industrial land has maintained positive signs in the morning. Finally, the variable distance to the center is not significant in any of the time slots.Table 8 Spatial Error models during the weeks of confinement (morning and afternoon). Table 8 Morning Afternoon B – W1 B – W2 B – W4 B – W6 B – W1 B – W2 B – W4 B – W6 (Constant) 1709.8** 1742.3** 1694.8** 1678.5** 1767.1** 1770.2** 1759.7** 1748.5** Residential [Ha] 120.8** 120.3** 121.0** 120.9** 120.4** 120.5** 120.5** 120.8** Offices [Ha] −24.9** −28.4** −25.5** −23.6** −24.8** −28.1** −24.9** −23.9** Industrial [Ha] 15.1** 9.0* 15.8** 18.0** Commercial [Ha] 73.6** 72.3** 74.5** 76.2** 72.2** 70.8** 72.7** 74.6** Educational [Ha] Cultural [Ha] Transports [Ha] Health [Ha] Parks [Ha] Other [Ha] 41.3** 42.6** 42.7** 44.3** 40.0** 41.2** 40.9** 40.8** Distance [km] LAMBDA 0.542** 0.568** 0.530** 0.518** 0.568** 0.579** 0.565** 0.554** No. Observations 1062 1062 1062 1062 1062 1062 1062 1062 R2 0.751 0.752 0.753 0.754 0.753 0.754 0.754 0.757 AIC 19146 19173 19132 19115 19178 19186 19171 19152 **Significant at the 0.05 level; * Significant at the 0.10 level. Table 9 Regression models (Spatial Error) during the weeks of confinement (evening and night). Table 9 Evening Night B – W1 B – W2 B – W4 B – W6 B – W1 B – W2 B – W4 B – W6 (Constant) 1768.1** 1768.1** 1762.7** 1766** 1761** 1763.3** 1757** 1762.3** Residential [Ha] 120.9** 121.1** 121.0** 121.0** 121.5** 121.5** 121.5** 121.6** Offices [Ha] −30.8** −31.4** −30.8** −30.9** −32.9** −32.6** −32.6** −33.6** Industrial [Ha] Commercial [Ha] 69.8** 68.9** 69.9** 70.7** 68.2** 68.1** 68.3** 67.1** Educational [Ha] Cultural [Ha] Transports [Ha] Health [Ha] Parks [Ha] Other [Ha] 41.8** 41.5** 41.8** 41.3** 41.9** 40.9** 41.8** 42.3** Distance [km] LAMBDA 0.587** 0.591** 0.588** 0.585** 0.597** 0.599** 0.598** 0.597** No. Observations 1062 1062 1062 1062 1062 1062 1062 1062 R2 0.754 0.756 0.755 0.757 0.756 0.757 0.757 0.757 AIC 19203 19199 19201 19186 19213 19234 19210 19208 **Significant at the 0.05 level; * Significant at the 0.10 level. The different phases followed in confinement also introduce significant nuances in the equations. In the week with the greatest restrictions (W2), the negative coefficients of activities such as offices are higher, and they decrease as the restrictions are lifted. Industrial activity, largely closed also during that W2 week, has a much lower coefficient during this week, while recovering in the weeks with the least restrictions (W4 and W6). 4 Conclusions The expansion of the COVID-19 pandemic has led to a radical change in urban dynamics and the distribution of the population in relation to land uses throughout the day. The measures imposed to control the spread of the virus imply the total or partial closure of many urban activities, with direct repercussions on people's activity, their mobility and their distribution in the city. Knowing the keys to urban dynamics during lockdown phases and the restrictions imposed is essential for their management and a crucial element for containment against possible outbreaks or second waves. From this study, it is possible to draw some conclusions and contribute to this knowledge. First, this paper shows that mobile phone data provides information with great potential for analyzing the impact of the measures taken regarding urban dynamics and the intensity of recovery in the different areas of the city after the lifting of restrictions. In this paper we have taken advantage of the high level of temporal detail of mobile data and have crossed them with information on land use with a high level of spatial and thematic disaggregation in order to determine how the restrictions imposed change the temporal profile of city use, with different impacts according to the types of activities present in some areas or others. Second, this research provides evidence of the different impact of the restrictions implemented on the city dynamics during the weeks of analysis, over the course of a typical day. More specifically, a first visual and dynamic analysis, through video-visualization, allowed us to explore the variation in population in the different urban areas throughout the day, comparing the reference week prior to the lockdown (W0) with that with the greatest restrictions on mobility (W2), showing very significant changes. A second analysis, based on the study of bivariate correlations of population distributions between large time slots, allowed us to obtain a numerical indicator to globally compare the impact of the lockdown measures in the different study weeks. The city of Madrid presents a high mix of land use, so that even in a reference week (W0) the correlations between the strips are very high. However, while in week W0 the correlations between morning and night decreased, due to the differences between residential and activity spaces, correlations were practically 1 during the lockdown, showing a similar distribution of the population in all time bands at night. The city turned off. The mapping of the residuals of these correlations showed that the few active zones during the morning hours were mainly logistics and industrial areas (positive residues). Third, from the results obtained in this paper, it is possible to describe the different impact of the measures on the diverse areas of the city, characterized by its main land use (and therefore, also to evaluate the impact on the activities related to the different land uses). Through hourly population distribution profiles for the dominant land use in each zone, the study provides evidence of a ra°dical change with respect to the reference week (W0), especially in the weeks of greatest restrictions (W2). These profiles are simplified, since they consider only the dominant use, when in most areas there are several uses of the land. However, the results explicitly showed the drastic reduction in population in activity spaces in the morning and afternoon, while residential spaces conserve the population in those time bands. All profile curves tended to flatten significantly, but once again the activities related to industry, commerce or health maintained more active profiles, compared to very subdued educational, leisure or office areas. Finally, this research provides evidence of the close link between the presence of people and the spatial distribution of the different land uses in all time slots of the different weeks of study. Unlike the profiles, the multiple regression analysis allowed us to consider the influence of the different land uses of each zone by the population and not to work only with the dominant use. In the reference week (W0) the coefficients of the independent variables showed the expected signs, positive in the variables of land use and negative in the distance to the center. In that week (W0), the variation of the coefficients throughout the day was consistent: increasing for residential and commercial land (except at night), and decreasing for the rest of the activities. Some land uses, such as offices, education, health or parks had negative signs at night, as a result of the high mix of uses in the city and the weight of mixed residential areas. In these mixed-use areas, the presence of activities reduces the potential presence of the population in these mostly residential areas. During the lockdown weeks, the equations tend to reproduce the night-time situation of the reference week (W0) for any period of the day. Only basic activities, such as commerce, have been active in all time frames of the day, but now with higher coefficients in the morning. Industry has also maintained some activity, but only in the morning and in the weeks of less restriction. Meanwhile, activities such as offices, education or parks showed negative signs in all time slots and in all weeks of the pandemic. This research does not explore the expected relationship between mobility restrictions and the pandemic spread. Although this is a crucial topic, it is out of the scope of the paper. This study aims at providing useful information for pandemic management and post-recovery planning, focusing on the impact of mobility restrictions across the city. In the first place, it enables us to improve our knowledge of urban dynamics during each of the confinement phases and the degree of restricted mobility of the population. Changes in population density according to mobility restrictions help to assess the level of follow-up of the measures. Second, it helps us to determine in which spaces and activities a greater presence of the population is concentrated during the weeks with restrictions and those when the restrictions are lifted. This is of interest to identify the areas of the city, the activities and the population groups associated with them, which remain functional and, consequently, pose a greater risk of virus transmission. In addition, once the restrictions are lifted, these analyses performed are able to show the pace of the city's recovery and the different recovery speeds of each urban activity. CRediT authorship contribution statement Gustavo Romanillos: Conceptualization, Methodology, Visualization, Writing – review & editing. Juan Carlos García-Palomares: Conceptualization, Methodology, Visualization, Writing – review & editing. Borja Moya-Gómez: Conceptualization, Methodology, Visualization, Writing – review & editing. Javier Gutiérrez: Conceptualization, Methodology, Writing – review & editing, Supervision. Javier Torres: Data curation, Conceptualization, Methodology, Writing – review & editing. Mario López: Data curation, Conceptualization, Methodology, Writing – review & editing. Oliva G. Cantú-Ros: Data curation, Conceptualization, Methodology, Writing – review & editing. Ricardo Herranz: Data curation, Conceptualization, Methodology, Writing – review & editing. Appendix A Supplementary data The following are the Supplementary data to this article:Multimedia component 1 Multimedia component 1 Multimedia component 2 Multimedia component 2 Acknowledgements The authors gratefully acknowledge funding from the 10.13039/501100004837 Spanish Ministry of Science and Innovation and the European Regional Development Fund (Project DynMobility, RTI2018-098402-B-I00) as well as the Comunidad de Madrid (Project INNJOBMAD-CM - H2019/HUM-5761). Borja Moya-Gómez would also like to thank Juan de la Cierva–2018 Training Aids for Labour Contracts plan [Grant: FJC2018-036613-I] of the 10.13039/501100004837 Spanish Ministry of Science and Innovation . The authors thank Dr. Marcin Stępniak for downloading the catastral data. 1 https://www.google.com/covid19/mobility/(accessed 07.09.20). Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.apgeog.2021.102524. 2 https://data-crtm.opendata.arcgis.com. (accessed 05.06.20). ==== Refs References Aleta A. Martín-Corral D. y Piontti A.P. Ajelli M. Litvinova M. Chinazzi M. Pentland A. Modelling the impact of testing, contact tracing and household quarantine on second waves of COVID-19 Nature Human Behaviour 2020 1 8 Badr H.S. Du H. Marshall M. Dong E. Squire M.M. Gardner L.M. Association between mobility patterns and COVID-19 transmission in the USA: A mathematical modelling study The lancet infectious diseases 2020 10.1016/S1473-3099(20)30553-3 Bushman K. Pelechrinis K. Labrinidis A. Effectiveness and compliance to social distancing during COVID-19. arXiv preprint arXiv:2006.12720 2020 Consejería de Sanidad de la Comunidad de Madrid Informe de la evolución de la infección por SARS-CoV-2 en la Comunidad de Madrid. Mayo 2020. Madrid Retrieved from https://www.comunidad.madrid/sites/default/files/doc/sanidad/200601_cam_covid19.pdf 2020 ESPON ESPON 2013 database dictionary of spatial unites [WWW Document]. URL http://database.espon.eu/db2/jsf/DicoSpatialUnits/DicoSpatialUnits_onehtml/index.html 2014 accessed 5.20.20 Ferretti L. Wymant C. Kendall M. Zhao L. Nurtay A. Abeler-Dörner L. Fraser C. Quantifying SARS-CoV-2 transmission suggests epidemic control with digital contact tracing Science 368 6491 2020 Gao S. Rao J. Kang Y. Liang Y. Kruse J. Mapping county-level mobility pattern changes in the United States in response to COVID-19 SIGSPATIAL Special 12 1 2020 16 26 García-Palomares J.C. Salas-Olmedo M.H. Moya-Gómez B. Condeço-Melhorado A. Gutiérrez J. City dynamics through Twitter: Relationships between land use and spatiotemporal demographics Cities 72 B 2018 310 319 10.1016/j.cities.2017.09.007 Hanson S. Imagine Journal of Transport Geography 3 14 2006 232 233 Kraemer M.U. Yang C.H. Gutierrez B. Wu C.H. Klein B. Pigott D.M. Brownstein J.S. The effect of human mobility and control measures on the COVID-19 epidemic in China Science 368 6490 2020 493 497 32213647 Lai S. Ruktanonchai N.W. Zhou L. Prosper O. Luo W. Floyd J.R. Yu H. Effect of non-pharmaceutical interventions for containing the COVID-19 outbreak in China medRxiv 2020 10.1101/2020.03.03.20029843 Mazzoli M. Mateo D. Hernando A. Meloni S. Ramasco J.J. Effects of mobility and multi-seeding on the propagation of the COVID-19 in Spain 2020 medRxiv Oliver N. Lepri B. Sterly H. Lambiotte R. Deletaille S. De Nadai M. Colizza V. Mobile phone data for informing public health actions across the COVID-19 pandemic life cycle Science Advances 6 23 2020 eabc0764 10.1126/sciadv.abc0764 Oliver N. Letouzé E. Sterly H. Mobile phone data and COVID-19: Missing an opportunity? 2020 arXiv:2003 12347 Paez A. Lopez F.A. Menezes T. Cavalcanti R. Pitta M.G.D.R. A spatio‐temporal analysis of the environmental correlates of COVID‐19 incidence in Spain Geographical Analysis 2020 https://covid19.elsevierpure.com/en/publications/a-spatio-temporal-analysis-of-the-environmental-correlates-of-cov Peixoto P.S. Marcondes D. Peixoto C. Oliva S.M. Modeling future spread of infections via mobile geolocation data and population dynamics. An application to COVID-19 in Brazil PLoS One 15 7 2020 e0235732 Pepe E. Bajardi P. Gauvin L. Privitera F. Lake B. Cattuto C. Tizzoni M. COVID-19 outbreak response: A first assessment of mobility changes in Italy following national lockdown medRxiv 2020 10.1101/2020.03.22.20039933 Pullano G. Valdano E. Scarpa N. Rubrichi S. Colizza V. Population mobility reductions during COVID-19 epidemic in France under lockdown 2020 medRxiv Sirkeci I. Yucesahin M.M. Coronavirus and migration: Analysis of human mobility and the spread of covid-19 Migration Letters 17 2 2020 379 398 Sun Q. Pan Y. Zhou W. Xiong C. Zhang L. Quantifying the influence of inter-county mobility patterns on the COVID-19 outbreak in the United States. arXiv preprint arXiv:2006 2020 Yabe T. Tsubouchi K. Fujiwara N. Wada T. Sekimoto Y. Ukkusuri S.V. Non-compulsory measures sufficiently reduced human mobility in Japan during the COVID-19 epidemic. arXiv preprint arXiv:2005 2020 09423 Zhou C. Su F. Pei T. Zhang A. Du Y. Luo B. Song C. COVID-19: Challenges to GIS with big data. Geography and Sustainability 2020 Further reading Aloi A. Alonso B. Benavente J. Cordera R. Echániz E. González F. Perrucci L. Effects of the COVID-19 lockdown on urban mobility: Empirical evidence from the city of santander (Spain) Sustainability 12 9 2020 3870 Bonato P. Cintia P. Fabbri F. Mobile phone data analytics against the COVID-19 epidemics in Italy: Flow diversity and local job markets during the national lockdown https://arxiv.org/abs/2004.11278 2020 Walters C.E. Meslé M.M. Hall I.M. Modelling the global spread of diseases: A review of current practice and capability Epidemics 25 2018 1 8 29853411 Zhu X. Zhang A. Xu S. Spatially explicit modeling of 2019-nCoV epidemic trend based on mobile phone data in mainland China https://www.medrxiv.org/content/10.1101/2020.02.09.20021360v2 2020

0

PMC9753119

NO-CC CODE

2022-12-16 23:25:30

no

Appl Geogr. 2021 Sep 28; 134:102524

utf-8

Appl Geogr

10.1016/j.apgeog.2021.102524

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)01693-7 10.1016/S0140-6736(21)01693-7 Correspondence Very rare thrombosis with thrombocytopenia after second AZD1222 dose: a global safety database analysis Bhuyan Prakash a Medin Jennie c da Silva Hugo Gomes e Yadavalli Madhavi b Shankar Nirmal Kumar f Mullerova Hana e Arnold Matthew e Nord Magnus d a Global Clinical Development Late R&I, AstraZeneca, Gaithersburg, MD, USA b Global Patient Safety BioPharmaceuticals, AstraZeneca, Gaithersburg, MD, USA c Global Medical BioPharmaceuticals, AstraZeneca, Gothenberg SE431 83, Sweden d Global Patient Safety BioPharmaceuticals, AstraZeneca, Gothenberg SE431 83, Sweden e Global Medical BioPharmaceuticals, AstraZeneca, Cambridge, UK f Global Patient Safety BioPharmaceuticals, AstraZeneca, Bangalore, India 27 7 2021 14-20 August 2021 27 7 2021 398 10300 577578 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcSince COVID-19 vaccine roll-out, very rare cases of thrombosis with thrombocytopenia syndrome (TTS), which has been referred to as vaccine-induced immune thrombotic thrombocytopenia, have been reported. Here we describe case details of TTS identified in the AstraZeneca global safety database, which captures all spontaneously reported adverse events from real-world use of its medicines and vaccines worldwide. All cases of TTS occurring within 14 days of intramuscular administration of first or second AZD1222 (ChAdOx nCoV-19) dose up to April 30, 2021, were included. In alignment with the Brighton Collaboration definition of TTS,1 TTS cases were searched using standardised Medical Dictionary for Regulatory Activities (version 23.1) queries “embolic and thrombotic events”, “hematopoietic thrombocytopenia”, and high-level term “thrombocytopenias”. The understanding of the role of anti-platelet factor 4 (PF4) antibodies in TTS is still evolving, so in line with the Brighton criteria, all cases meeting the above definition, irrespective of anti-PF4 antibodies, were included. This research was led and funded by AstraZeneca. At data cutoff, 13 cases of TTS were identified after the second AZD1222 dose, occurring 1–13 days post-vaccination (appendix p 1; no cases were observed outside the 14-day window). The reported events included eight individuals with pulmonary embolism, co-occurring with cerebral venous sinus thrombosis (CVST) in two individuals; one individual with CVST occurring alone; one individual with deep vein thrombosis; one individual with thrombotic stroke; and two individuals with unspecified embolisms. The 13 vaccinees reporting TTS were aged 45–85 years (one age unknown); eight were female (a lower proportion than in initial TTS reports2). Medical history was available for 11 vaccinees; one had previous pulmonary embolism, one had thrombocytopenia, three had cancer, and one had COVID-19. Other medication details were available for seven vaccinees and included cancer treatments, antihypertensives, anticoagulants, and statins. Anti-PF4 test results were available for three of 13 cases, all of which were negative (appendix p 1). At data cutoff, six vaccinees were reported as “not recovered”, three were “recovering”, three died, and one “recovered with sequalae”. Based on weekly data from the European Centre for Disease Prevention and Control and the UK Department of Business Energy and Industrial Strategy, as of April 25, 2021, approximately 5·62 million people were estimated to have received the second AZD1222 dose in the EU/EEA and in the UK (93·5% administered in the UK). Based on this exposure level, the estimated rate of TTS within 14 days of the second AZD1222 dose was 2·3 per million vaccinees. By comparison, within the same timeframes, the estimated rate of TTS within 14 days of the first dose was 8·1 per million vaccinees, based on 399 cases of TTS identified after the first AZD1222 dose and approximately 49·23 million first doses administered, with 45·2% of doses administered in the UK and 54·8% in the EU/EEA. We estimated background TTS rates using two analysis methods with the US Truven MarketScan Commercial Claims and Encounters database from Jan 1, 2019, to Dec 31, 2019. The very low rate of TTS reported following a second AZD1222 dose is within preliminary estimates of the background range in an unvaccinated population pre-COVID-19 (appendix p 4).3 Limitations of this safety database analysis include a reliance on health-care provider-reported and vaccinee-reported data, which might result in event under-reporting. Furthermore, heightened media attention might have led to event misclassification. Therefore, to provide a cautious estimate for the event rate, data used for the number of doses administered was limited to the EU/EEA and the UK, while all cases reported globally were included. As common in post-market reporting, limited information was provided in many cases, including medical history and concomitant medication. Additionally, comparing incidence rates, including background rates, can be challenging due to dependence on data sources, event definitions, data collection method, timeframe, and the patient population. The TTS background event rate reported here was derived from a large US population with health insurance in 2019, prior to the emergence of COVID-19, a disease which itself has been associated with thrombotic events.4 Although post-marketing surveillance reporting does not enable full characterisation and contextualisation of each case, overall results support the continued administration of AZD1222 in a two-dose schedule, as indicated.5 This is particularly relevant in light of recent data demonstrating the efficacy of two doses of AZD1222 against SARS-CoV-2 variants of concern, including protection against hospitalisation with the Delta variant.6, 7 © 2021 Dinendra Haria/SOPA Images/LightRocket/Getty Images 2021 We are employees of AstraZeneca and might have stock or options, or both. Details of contributions are shown in the appendix. Supplementary Material Supplementary appendix ==== Refs References 1 Brighton Collaboration Updated Proposed Brighton Collaboration process for developing a standard case definition for study of new clinical syndrome X, as applied to Thrombosis with Thrombocytopenia Syndrome (TTS) https://brightoncollaboration.us/wp-content/uploads/2021/05/TTS-Interim-Case-Definition-v10.16.3-May-23-2021.pdf May 18, 2021 2 Greinacher A Thiele T Warkentin TE Thrombotic thrombocytopenia after ChAdOx1 nCov-19 vaccination New Engl J Med 384 2021 2092 2101 33835769 3 Burn E Li X Kostka K Background rates of five thrombosis with thrombocytopenia syndromes of special interest for COVID-19 vaccine safety surveillance: incidence between 2017 and 2019 and patient profiles from 20·6 million people in six European countries medRxiv 2021 published online May 13. 10.1101/2021.05.12.21257083 (preprint). 4 Eslamifar Z Behzadifard M Soleimani M Behzadifard S Coagulation abnormalities in SARS-CoV-2 infection: overexpression tissue factor Thromb J 18 2020 38 33323111 5 AstraZeneca Vaxzevria, COVID-19 Vaccine (ChAdOx1-S [recombinant]). Summary of Product Characteristics https://www.ema.europa.eu/en/documents/product-information/vaxzevria-previously-covid-19-vaccine-astrazeneca-epar-product-information_en.pdf June, 2021 6 Emary KRW Golubchik T Aley PK Efficacy of ChAdOx1 nCoV-19 (AZD1222) vaccine against SARS-CoV-2 variant of concern 202012/01 (B.1.1.7): an exploratory analysis of a randomised controlled trial Lancet 397 2021 1351 1362 33798499 7 Bernal JL Andrews N Gower C Effectiveness of the Pfizer-BioNTech and Oxford-AstraZeneca vaccines on covid-19 related symptoms, hospital admissions, and mortality in older adults in England: test negative case-control study BMJ 373 2021 n1088

34329583

PMC9753120

NO-CC CODE

2022-12-16 23:25:30

no

Lancet. 2021 Jul 27 14-20 August; 398(10300):577-578

utf-8

Lancet

10.1016/S0140-6736(21)01693-7

oa_other

==== Front Appl Geogr Appl Geogr Applied Geography (Sevenoaks, England) 0143-6228 0143-6228 Elsevier Ltd. S0143-6228(21)00142-9 10.1016/j.apgeog.2021.102526 102526 Article Why the Navajo Nation was hit so hard by coronavirus: Understanding the disproportionate impact of the COVID-19 pandemic Wang Haoying Department of Business and Technology Management, New Mexico Tech, 801 Leroy Pl, Socorro, NM, 87801, USA 29 7 2021 9 2021 29 7 2021 134 102526102526 9 12 2020 23 7 2021 26 7 2021 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The COVID-19 (SARS-CoV-2) pandemic of 2019–2020 has incurred astonishing social and economic costs in the United States (US) and worldwide. Native American reservations, representing a unique geography, have been hit much harder than other parts of the country. This study seeks to understand the reasons for the disproportionate impact of the pandemic on Native American communities by focusing on the Navajo Nation – the largest Native American reservation in the US. I first reviewed the historical pandemics experienced by Native Americans. Guided by the literature review, an institutional analysis focusing on the Navajo Nation suggests a lack of both institutional resilience and healthcare preparation. The analysis further identified four factors that could help explain the Navajo's slow response to the COVID-19 pandemic: prevalence of underlying chronic health conditions, lack of institutional resilience, the relationship between the federal government and tribal governments, and lack of social trust. Relevant policy implications are discussed. For instance, to better prepare Native American communities for shocking events like the COVID-19 pandemic in the future, policymaking should integrate informal institutions to build efficient formal institutions for self-governance. Promoting public health education and establishing collaborations between Native and non-Native communities are also necessary long-run strategies. Keywords Coronavirus COVID-19 pandemic Formal institutions Informal institutions Resilience Navajo nation ==== Body pmc1 Introduction Since March 2020, the coronavirus SARS-CoV-2 (COVID-19) has quickly spread across the United States (US) and the world. By the end of summer 2020, several regions had been hit particularly hard as the pandemic expanded, including Washington, New York, California, and Florida in terms of the total positive cases reported. By early December 2020, many states across the country saw an escalating number of positive cases. The total number of coronavirus cases in the US reached 15 million, with a death toll of close to 300,000. As several vaccines became available from the beginning of 2021, both the daily new positive cases and deaths have been declining as the vaccination rolls out across the country. As of this writing (July 2021), the COVID-19 pandemic has not been under effective control due to new coronavirus variants (e.g., Delta variant) and stagnating vaccination rates (Bekiempis, 2021). Both the total positive cases and deaths doubled from the December 2020 statistics. Based on some preliminary estimates, the total economic cost of the COVID-19 pandemic can be as high as US$16 trillion, which is more than three-quarters of the US annual GDP (Cutler & Summers, 2020). According to a more recent study by Chen et al. (2021), published in late May 2021, the total burden of the pandemic will be between US$17 and 94 trillion over the next decade in the US. Behind all the shocking numbers and estimates, there seem to be very few questions asked about one unique geography – the Native American reservations. According to the 2010 US Decennial Census, 5.2 million people (1.7 % of the US population) self-identified as American Indian or Alaska Native scattered across over 500 federally recognized tribes. Steckel and Prince (2001) once commented in their economic history study: “… … Fortunately for European-Americans, they rarely had to test their system of production and distribution against the kinds of demographic disasters faced by Plains tribes.” The COVID-19 pandemic has put the system under a tough test. In spite of their low-density population and remoteness, Native Americans have been hit much harder by the COVID-19 pandemic than the US white population, according to a US CDC (Centers for Disease Control and Prevention) report (Hatcher et al., 2020). Based on data collected from 23 states, the cumulative COVID-19 incidence rate among American Indians (and Alaska Natives) was 3.5 times that among non-Hispanic white persons. According to the statistics compiled by Johns Hopkins University (https://coronavirus.jhu.edu/us-map), McKinley County (New Mexico) had 5515 positive cases per 100,000 residents as of July 14, 2020. It was the highest per-capita cases in the country. By the first week of December 2020, the number nearly doubled. Neighboring Navajo County and Apache County in Arizona were also among the top 10 counties nationwide in terms of per-capita cases. All three counties are mostly located in the Navajo Nation (with the Hopi reservation in the middle) – the largest Native American reservation in the US. The reservation covers approximately 70,000 square km2 (more than the states of Connecticut, Massachusetts, New Hampshire, and Rhode Island combined), but only with a population of 156,823 living on the reservation as of the 2010 US Census.1 As Fig. 1 shows, the reported positive COVID-19 cases were already more than 10 % of the Navajo Nation residents by the end of November 2020, which is significantly higher than the US national average (about 4 % at the same time). Based on our common knowledge of epidemic diseases, a low-density population is more resistant to disease spread. So why has the Navajo Nation been hit so hard by the pandemic? The existing research focuses mainly on documenting the epidemic facts and studying the general population (e.g., Dyer, 2020; Laurencin & McClinton, 2020, van Dorn et al., 2020). No in-depth analysis has explored the pandemic situation and its causes and impacts in Native American communities. This study seeks to understand the disproportionate impact of the COVID-19 pandemic by looking into the Navajo Nation and its healthcare system, institutions, and related factors.Fig. 1 The reported COVID-19 positive cases on the Navajo Nation from March 2020 to November 2020 (the first two waves of the pandemic) Data source: Compiled from the data reported by the Navajo Nation Department of Health; https://www.ndoh.navajo-nsn.gov/COVID-19/Data, accessed on Jul 23, 2021. Note: The highest daily case count was reached on Nov 21, 2020 (401 cases), based on data up to Jul 21, 2021. Fig.Q3: Please check Figure captions 1 The literature has found that health disparities between indigenous and non-indigenous people persist worldwide (e.g., Leeuw et al., 2012). In this study, I ask the specific question of what is causing the severity of the COVID-19 pandemic in the Navajo Nation. Is it the lack of institutional resilience and healthcare resources, or other factors?2 The Navajo Nation sits around the four-corner region and is located mainly in the state of New Mexico and Arizona's remote areas (Fig. 2 ). There is no metropolitan area inside the reservation. As Fig. 2 shows, several small to medium-size cities are around the reservation. Large metropolitan areas like Albuquerque and Phoenix are hundreds of miles away. Its location and geographical layout give rise to the first hypothesis that the inadequacy of healthcare resources has resulted in slow responses to the pandemic. As early as April 2020, the Navajo Nation instituted the country's most extensive and restrictive lockdown orders, but underfunded infrastructure and lack of access to basic needs have worked against the efforts (Baek, 2020). It leads to questions about the unique legal relationship between the US federal government and Native American tribes. The federal-tribal relationship is a complicated result of history, culture, and traditions. This suggests another hypothesis that the lack of institutional resilience explains the situation, which requires us to examine the institutional environment (formal and informal institutions and their interactions, see footnote 5 for definitions of formal and informal institutions) of the Navajo Nation. Or simply, people are not following public health recommendations due to social norms? In the following sections, I first review the history of public health pandemics with a focus on Native American tribes. I then analyze the institutional environment of the Navajo Nation and explore factors behind the severity of the pandemic. Methodologically, I synthesize a multi-disciplinary literature review, anecdotal evidence, and conversations with Navajo scholars.3 The choice of the methodology considers the limitations of collecting individual/household-level data on the reservation. I conclude the paper by discussing the policy implications of my findings and highlighting the suggestions for moving forward. This study contributes to the broad literature on indigenous geography (Butzer, 1992; Larsen & Johnson, 2012).Fig. 2 The map of Navajo Nation agencies and major nearby cities. Data sources: GIS shapefiles are provided by the Navajo Nation Land Department. All city locations indicate the coordinates of the city hall or city council based on Google Maps. Fig. 2 2 Background and literature review The COVID-19 pandemic is not the first pandemic that the Navajo Nation and other Native American tribes have faced. And it is unlikely to be the last one. If we look through the history of pandemic shocks by exploring the literature, the pandemics that Native Americans have survived since the first contact with Europeans can be broadly categorized into two: internally emerged pandemics and exogenously imported pandemics. The difference between the two is not always clear. In general, internally emerged pandemics are more likely to be chronic pandemics, such as opioid addiction and obesity. Of course, certain external factors can influence the spread and severity of these chronic pandemics. Exogenously imported pandemics, on the other hand, tend to be more swift, such as the diphtheria pandemic of 1801 in California, the Influenza pandemics, and the COVID-19 pandemic. Among Native American communities, the most consequential chronic pandemics include drug disorder (opioid pandemic), alcohol problems, obesity, diabetes, and hypertension (e.g., Broussard et al., 1995; Frank et al., 2000; Jernigan et al., 2020; McLaughlin, 2010; Wiedman, 2012). Earlier studies also suggest that HIV/AIDS has been another pandemic faced by some Native American communities (Campbell, 1989; Hamill & Dickey, 2005; Nebelkopf & King, 2003). These chronic pandemics can take a toll on the labor force on top of those already sluggish tribal economies. It also puts additional stress on the underfunded tribal healthcare system. The reasons for the prevalence of these chronic health problems in Native American communities range from biological reasons to cultural & socio-economic environments (Wiedman, 2012). Several previous studies have explored the biological reasons (e.g., Broussard et al., 1995; Frank et al., 2000; Wiedman, 2012). For instance, one hypothesis related to obesity is that Native Americans have a genetic predisposition to obesity when exposed to the modern environment of high calorie food and low body energy consumption (Broussard et al., 1995). Other research has suggested alternatives to the biological response hypothesis. For instance, studies have shown that physical activities play a significant role in preventing obesity among Native Americans (e.g., Esparza et al., 2000; Stevens et al., 2004). However, socio-economic factors are believed to be more relevant to the prevalence of chronic conditions. First, lagging socio-economic conditions link to “diseases of poverty” (or infections of poverty in some studies). Hotez (2008) summarized some of the common diseases of poverty among tribes in the US West, including Navajo. Second, Native American tribes, including the Navajo Nation, face much higher transaction costs in economic development than their non-Native counterparts (Yonk et al., 2017; Lofthouse, 2019). One of the consequences is the underinvestment in healthcare and lack of employer-sponsored health insurance.4 Meanwhile, the complicated relationship between the federal government and tribal governments leads to underfunded public health facilities and other infrastructure necessary to maintain and improve the quality of life (Tipps 2018). Feir and Akee (2019)'s study of the First Nations in Canada suggests that the nutrition and health conditions of Native Indians have rarely changed in the last several decades. To prevent and eradicate these internally emerged chronic pandemics, informal and formal institutions need to work integratively.5 The traditional Native American culture and belief systems, which show a great amount of heterogeneity across hundreds of federally recognized tribes, value peoplehood and tend to be receptive to health interventions (Broussard et al., 1995; Plough et al., 2011; Lerma, 2014).6 Castillo (1999) found that Native Americans’ adherence to their traditional belief patterns has been robust. They did not simply abandon their belief systems and become Christian but have maintained their cultural diversity, for instance, through traditional storytelling (Grandbois & Sanders, 2012). It suggests that Native Americans have built-in resilience in terms of informal institutions and culture.7 Institutional resilience is critical in healthcare systems and recovering from a public health pandemic (Carthey et al., 2001; Wu et al., 2020). When designing public health interventions, institutional resilience should be channeled and utilized instead of being suppressed or ignored (Dufrene et al, 1992). One example is the Wellness Courts program adopted by several tribal groups across the US (Tipps et al., 2018). It has been a successful institutional innovation of the healthcare systems for some tribal groups. Meanwhile, a thorough understanding of these chronic pandemics is essential. As we will further explore in the following sections, an important reason why some Native American communities were hit hard by exogenously imported pandemics is the prevalence of underlying conditions. The current COVID-19 pandemic suggests that this is not a problem unique to Native Americans. People with underlying conditions are more vulnerable to coronavirus. However, it is worth noting that Native American communities usually have a significantly higher prevalence of chronic diseases (La Ruche et al., 2009; McLaughlin, 2010), which makes them more vulnerable to exogenously imported pandemics like H1N1 virus (influenza) and coronavirus (COVID-19). Most Native Americans with chronic illnesses are a collateral result of the prevailing chronic pandemics. The exogenously imported pandemics in Native American tribes date back to the sixteenth to the seventeenth century. Historically, especially in the pre-Columbian era, most Native Americans had a nutrition advantage (Steckel & Prince, 2001). However, this nutrition advantage has gradually disappeared since the first contact with Europeans. From the sixteenth through the twentieth century, according to Campbell (1989), Native Americans had experienced more than several detrimental epidemics such as smallpox, rubella, influenza, malaria, yellow fever, and cholera. All these diseases appeared after the first contact with Europeans and followed the expansion of the frontier. During the first contact period that lasted over two centuries, many diseases with common cures in Europe became exogenously imported pandemics as the frontier expanded across the American continent. Castillo (1999) presented an in-depth study about the diphtheria pandemic of 1801 among Tongva and Chumash tribes in Southern California. The traditional belief systems and social norms of these tribes played a critical role in their fight against the pandemic. Pre-Columbian historical evidence also suggests that Native Americans are vulnerable to infectious diseases – something that can easily be imported by outsiders (Martin & Goodman, 2002). It is worth noting that such a vulnerability is not unique to American Indians. It has been observed globally as the global transport networks expanded in the past several centuries (Tatem et al., 2006). Another exogenously imported pandemic that significantly affected Native Americans was the 1918–1920 Influenza pandemic. According to Dahal et al. (2018), there was a significant decline in the number of births occurring 9–11 months after peak pandemic mortality in Northern Arizona counties where Native Americans, including the Navajo people, have lived for centuries. A similar case, but well-documented, was the recent Influenza H1N1 pandemic in 2009. The literature identified two key factors that affected the pandemic interventions: (1) The prevalence of underlying health conditions; (2) the cultural, social, and economic barriers to adoption of pandemic interventions (e.g., Hutchins et al., 2009; La Ruche et al., 2009). A US CDC report indicated that American Indians/Alaska Natives had an H1N1 mortality rate four times higher than all other racial/ethnic populations combined (US CDC 2009). From the literature on historical pandemics that were exogenously imported into Native American communities, we can seek a better understanding of the disproportionate impact of the COVID-19 pandemic on them. For instance, one thing is clear – pandemic vulnerability directly relates to the institutional environment of Native American reservations and their institutional resilience (Groom et al., 2009; La Ruche et al., 2009; McLafferty, 2010; Tipps et al., 2018; Wiedman, 2012). This study focuses on the Navajo Nation – the largest Native American reservation in North America. In particular, I analyze its institutional environment and related factors to understand how it presents both challenges and opportunities for Navajo people when facing exogenously imported pandemics like the COVID-19. The analysis considers both dimensions of its institutional environment: informal institutions and formal institutions (and their interactions). The goal is to unpack the current public health situation under the COVID-19 pandemic and shed light on how to better prepare for similar shocking events in the future. 3 The institutional environment of the Navajo Nation 3.1 Informal institutions Before Europeans arrived, Native American tribes lived mainly under the governance of informal institutions. Even though there were certain types of formal institutions established by some tribes (e.g., private property rights, see Miller (2018)), the formal institutions only started playing active roles after the treaty-making between sovereign Indian tribes and the US government.8 Informal institutions have been playing essential roles in the social-economic life of Native Americans, which in turn shapes their public health conditions (Plough et al., 2011). Navajo's informal institutions are notable in social norms and networks. They are passed on to generations mainly through storytelling and ceremonies. Storytelling is an effective and critical way for Navajos to sustain cultural history, language, social customs, and knowledge systems (Iseke, 2013; Sage, 2019). Their informal institutions have two key characteristics: Peoplehood and connection to nature. Holm et al. (2003) argued that peoplehood in Native American culture explains why colonialism failed to destroy the Diné (Navajo People)'s traditional institutions of governance. A qualitative comparative analysis of Navajo history by Lerma (2014) suggests that between the first contact and 1923, the impact of colonial strategies on the elimination of traditional Diné institutions was minimal. Feir and Gillezeau (2018)'s study of unemployment of Native Americans during the Great Recession also suggests that relying on peoplehood is an important social trait of the Navajo people. Language is a good example to contextualize the importance of peoplehood for Navajo. Language is one of the four building elements of peoplehood (Holm et al., 2003). A Navajo member is expected to speak their own language. In practice, not every Navajo member can speak the Navajo language (also known as Diné Bizaad). But it is a shared expectation, which fits the conceptual framework of informal institutions by Helmke and Levitsky (2004). The Navajo language is also the fundamental medium of storytelling that sustains Navajo's cultural history and social norms. The connection to nature is a common element of Native Americans' culture and belief systems. It was documented in detail by Castillo (1999) when discussing the Tongva and Chumash tribes in Southern California. Even nowadays, the strength of connecting to nature is reflected in the Navajo Nation's Climate Adaptation Plan (Navajo Nation Department of Fish and Wildlife, 2018). Such a robust integrant of their informal institutions is also revealed in their attitude towards energy and natural resources development. For example, Necefer et al. (2015) found that Navajo people attach significant importance to environmental preservation, not only for the sustainability of future generations but also for the viability of their culture and identity that have supported environment and natural resource stewardship for centuries. However, their informal institutions have faced challenges. For instance, the lack of social trust with the surrounding non-Native communities has been an invisible barrier. Social trust is not necessarily in the realm of informal institutions, but informal institutions affect social trust.9 As an indispensable part of social capital, social trust builds on social norms and expected behaviors (Jensen & Svendsen, 2016). The recent Days Inn incident between Alamo Navajo Reservation (a non-contiguous branch of the Navajo Nation) and the City of Socorro in New Mexico suggests that the lack of social trust with the non-Native communities may have slowed them down in managing the COVID-19 pandemic.10 Such challenges to Navajo's informal institutions have also increased the transaction costs in other aspects of their socio-economic life. For example, Cattaneo and Feir (2019) showed that Native Americans face an average interest rate of nearly two percentage points above the average loan for non-Native Americans in the mortgage loan market. Such an elevated risk premium not only reflects discrimination in the financial markets but also suggests a deeply rooted lack of social trust between Native American communities and the non-Native communities. Another challenge to the Navajo informal institutions is that many public health recommendations from the federal agencies and local organizations (e.g., large healthcare providers) during the pandemic are not customized. For example, a generic 6 or 12-feet social distancing recommendation may not be consistent with some of the important Navajo social norms like attending seasonal tribal ceremonies. A more customized and balanced (between disease prevention and social norms) social distancing recommendation could prevent people from ignoring the universal social distancing recommendation from the CDC. A related example concerns the face covering recommendation. A generic message like ‘get your mask’ or ‘wear your mask’ does not work for many Navajo residents who already lack public health education. Because it takes them several hours of driving to go to the nearest chain store like Walmart to get supplies, and many do not have internet access to place an online order (and this already assumes that they have a credit/debit card). A customized message like ‘make your own mask and here is how’ is much more effective. 3.2 Formal institutions Historically, Navajo people have struggled in building effective formal institutions relative to their success in sustaining strong informal institutions. One way to understand this is by looking at the Navajo Nation's path to promoting economic growth. Despite their unique culture and way of life, Navajo people have been seeking changes, if not revolutions, to be part of the modern economy as American Indian history scholar Peter Iverson depicts in his book:“At the time of the 1970 [Navajo Nation chairman] race, [Peter] MacDonald represented a new face, with the promise of a new way of doing things. Well known throughout the reservation because of his work with the Office of Navajo Economic Opportunity, he obtained the support both of more traditional Navajos and of younger Navajos looking for a chairman who understood the old ways and yet could meet the Anglo world on more equal terms.” - from <The Navajo Nation> (Iverson, 1981; Page xv) During this process of seeking change, the role of formal institutions becomes more and more important. In the past several decades, the local governance institutions of the Navajo Nation have become increasingly relevant to its economic development and the improvement of its social safety net. It became even clear after the federal government released much responsibility for reservation governance from the Bureau of Indian Affairs (BIA) to the tribal governments in the 1980s (Dippel, 2014). The process of rebuilding and enhancing self-governance has proved difficult, which affects economic development and has consequences for Navajo's healthcare system today (Spirling, 2012; Dippel, 2014). According to the literature on Native American economic development and entrepreneurship, some of the main weaknesses of Native Americans' formal institutions include ambiguity in property rights, heavy reliance on federal grants and aids, dual bureaucracy, tax complexity, financial discrimination, etc. (Yonk et al., 2017; Lofthouse, 2019; Cattaneo & Feir, 2019). The weaknesses in formal institutions that contribute directly to the poor public health conditions are dual bureaucracy and limited access to financial resources and services. Dual bureaucracy refers to circumstances when the federal and tribal officials, who both have broad authority and discretion regarding tribal public policies, attempt to manage resources that are under their dual jurisdiction (Lofthouse, 2019). A consequence of dual bureaucracy is that, after the 1980s federal-tribal relation reform, some power and responsibilities were given back to the Navajo Nation while necessary formal institutions and infrastructure were not there for the tribal government to make good decisions. Some of the changes concerned Navajo's healthcare system (e.g., S.2728 - Indian Health Care Amendments of 1980). According to the treaties between Native American tribes and the federal government, the federal government is obligated to provide healthcare services to Native Americans on the reservation as part of the federal trust (Tipps et al., 2018). The federal healthcare services for Native Americans are managed by the Indian Health Service (IHS), established in 1955. However, as Tipps et al. (2018) pointed out, the system has been underfunded in recent years. Some IHS clinics operated on the reservation have suffered from inadequate staffing, out-of-date facilities, failure to implement new programs, etc. Some well-funded reservations, for example, those who have a relatively small population but with significant casino and tourism revenues, can supplement the federal programs or build their own healthcare infrastructure. But this is usually not the case for large reservations like the Navajo Nation. According to recent national news amid the COVID-19 pandemic (e.g., Baek, 2020), 30 % of the Navajo Nation residents do not have access to running water almost 200 years after its invention. It makes critical public health measures, such as frequent hand-washing, difficult to implement on the reservation during a pandemic. Since the 1980s, there has been a decline in the role of the BIA in Native American communities. Meanwhile, the persistent social divisions lead to fractional politics within some reservations. According to Dippel (2014), a critical institutional and historical reason is the forced coexistence during the formation of Native American reservations in the 19th century. As Dippel (2014) argued, when different tribes or bands (even with a shared cultural identity) were forced to live together, it often did not imply coherent shared governance. The forced coexistence has significantly increased reported cases of internal conflicts and corruption in the policy processes (Dippel, 2014). It is essentially a consequence of the lack of resilience in formal institutions that were never fully functional as expected. Such a consequence becomes more pronounced when tribal self-governance began to matter more after the 1980s (Dippel, 2014). To some extent, it explains the inadequate preparation observed through several recent pandemics, including the COVID-19 pandemic. In the case of the Navajo Nation, there were no forced coexisting bands. For administrative and constitution-related reasons, however, the Navajo Indians were organized into over 100 chapters and five agencies (Fig. 2). They are represented by delegates in the Navajo Nation Council – the legislative body of the Navajo Nation government. All the Navajo chapters share the same culture with little difference in traditions and social norms (e.g., ceremonies). It potentially explains Navajo's strength in informal institutions. We will explore the impact of the exogenously introduced governance structure on its response to the COVID-19 pandemic in the next section. As discussed before, the intricate relationship between the federal government and tribal governments has resulted in sluggish economic development and scanty employment opportunities. It limits many Native American residents’ ability to supplement their healthcare through employer-sponsored or private insurance. It also keeps potential financial firms away from providing lending and other financial services on or near reservations. The fact that many Native American residents cannot use their real estate property as collateral makes the situation potentially worse. It means that they have limited or only expensive lending opportunities to use as leverage to help with family healthcare. Such a disadvantage also applies to situations like entrepreneurship development and human capital accumulation (e.g., education and professional training). Of course, one potential explanation for the restriction on land titling is the concern that privatization may lead to concentrations of economic power and a worse outcome for tribal members (Holm et al, 2003). But the debate around tribal land privatization does indicate the level of challenge faced by Native Americans. Over time, the challenge has contributed to several chronic pandemics across Native American communities, which we reviewed in the previous section. No matter whether it is the dual bureaucracy or the limited financial resources at their disposal, they all leave many Native Americans unprepared for shocking events like the COVID-19 pandemic, even though research has shown that Native Americans are receptive to health interventions (Dufrene et al., 1992; Broussard et al., 1995; Plough et al., 2011). 4 The COVID-19 pandemic: why the Navajo Nation was hit so hard? Based on the literature review and the institutional analysis of the Navajo Nation and other Native American reservations in general, we can identify four key factors that could explain why the Navajo Nation was hit hard by the coronavirus from the onset of the pandemic through December 2020 (the first two waves). They are (1) prevalence of underlying chronic health conditions; (2) lack of institutional resilience to external changes and psychosocial stresses; (3) complicated relationship between the federal government and tribal governments; (4) lack of social trust. They are interrelated, but each reflects a distinct aspect of the challenges that the Navajo people have faced. Here we explore them separately in the context of the COVID-19 pandemic. 4.1 Prevalence of underlying health conditions As we reviewed in the background section, underlying chronic conditions such as drinking problems, obesity, diabetes, and hypertension have frequently exposed Native Americans to abrupt public health risks. The Navajo Nation is not immune to these problems. For example, Dabelea et al. (2009) showed that diabetes has been a critical health problem for Navajo youth. Infectious diseases are another high-risk health burden among communities in the Navajo Nation (Sutcliffe et al., 2019). Such chronic health conditions have already put some Navajo families under financial and psychosocial stresses before an exogenously imported pandemic hits. During the COVID-19 pandemic, affected individuals and families rely mainly on the underfunded IHS clinics to provide treatment and care, at least during the first several months. However, the IHS is not an insurance program, and it does not cover care from external providers. When the tribe-supported supplemental healthcare services cannot meet the gap, the affected individuals and families were trapped by the healthcare system. Despite the extremely low population density, the infection rates in terms of per-capita cases in several Navajo Nation counties were among the highest in the country. 4.2 Lack of institutional resilience Formal and informal institutions, if operating effectively, can reinforce each other. The interaction and integration of formal and informal institutions have been found as a principal driver of social stability and economic development (Pejovich, 1999). It would be the ideal scenario and socioeconomic path for the Navajo Nation. The reality, on the other hand, has been that despite the traditionally robust informal institutions, the Navajo Nation has not been able to build efficient formal institutions that can advance its social-economic life to a level comparable to the non-Native world.11 Historically, the Navajo Nation relied on informal institutions to guide tribal affairs while depending on the federal government to provide essential services and infrastructure. After the federal government released some of the power back to the tribes, a vacuum of formal institutions appeared. Suddenly, many tribal governments found themselves weak in self-governance institutions. Tribes have to catch up in institutions and infrastructure development to take the new responsibilities. It has resulted in a lack of institutional resilience when facing abrupt changes, especially exogenously imported shocks. Although the Navajo Nation is a very adapting group as Tolan (1989) depicted, the robustness of its informal institutions alone cannot revamp through the collision between the Native world and the non-Native world during a period of rapid social and economic changes. As a passive reaction to the lack of formal institutional resilience, the Navajo Nation has developed a strong dependence on federal support and its natural resources endowment since World War II. According to the Navajo Nation Division of Economic Development, coal, oil, and uranium have been the foundation of the Navajo economy since the 1920s.12 The issue with such a resource-dependent economy is path dependence, similar to that of the dependence on the federal budget. The income and welfare of the Navajo people are directly subject to the fluctuation of the commodity markets. At the reservation level, this limits the supplementary investment in healthcare services. At the individual and household level, frequent disruptions of income affect consumption and health, as well as the ability to prepare for abrupt changes such as a pandemic. 4.3 The relationship between the federal government and tribal governments Dippel (2014) shows that reservations that combined multiple tribal bands in the 1800s on average are 30 % poorer today. The result does not apply directly to the Navajo Nation. However, it has implications for Navajo Nation's governance structure. As previously mentioned, the Navajo Nation was ‘required’ to organize into over 100 chapters and five agencies for administrative and constitution-related reasons. Each chapter is a sub-governmental entity within the Navajo Nation delegated to address local issues such as land and health of its respective chapter residents (Navajo Nation Government, 2013). Since there are no different bands within the Navajo Nation, the question is that does the Navajo Nation needs this many administrative units to effectively self-govern. In practice, such a widely divided government entity weakens the Navajo Nation's bargaining position during negotiations with the federal government. It may also reduce the efficiency of decision-making processes concerning internal tribal affairs, including public health education and disease prevention. But the internal organization is not the only factor that prevents the Navajo Nation from working with the federal government effectively. Even if we treat the Navajo Nation as a coherent government entity, it still faces two issues when working with the federal government: (1) A disadvantage in the negotiation because of its small size; (2) the cultural differences. The first issue is easy to see. Although being the largest reservation in the US, the total Navajo population was 332,129 in 2010, according to the 2010 US Census. Only close to half of them lived on the reservation, and others lived in nearby off-reservation towns and other areas (Navajo Nation Government, 2013). The second issue is more implicit, and it often seems esoteric to the non-Native world. However, cultural differences can have significant influences on the Navajo Nation's pursuit of economic prosperity under self-governance. As a 1989 Los Angeles Times article (Jones, 1989) titled: “MacDonald Faces Cultural Conflicts: Embattled Navajo Leader Tries to Straddle 2 Worlds,” tribal leaderships often struggle for a balance between strong traditional social norms (e.g., taking care of families and relatives, or nepotism) and a set of well-established rules to follow (formal institutions). Given more than 500 federally recognized American Indian and Alaska Native tribes in the US, the federal support for Native Americans is often stretched even without any budget cuts. As far as healthcare services and support are concerned, what happens often is a delayed distribution of funds and/or insufficient funds (Newton, 2020). Such situations can worsen the spread of underlying chronic conditions and, in the case of a pandemic, exacerbate the challenges faced by Native American communities. 4.4 Lack of social trust Social trust concerns both shared values and shared expectations within a community. Based on Helmke and Levisky (2004)'s conceptual framework, social trust overlaps culture and informal institutions. Lack of social trust can explain some of the challenges faced by the Navajo people. Internally, social trust affects the efficiency of group decision-making processes. Native Americans have a history of making decisions as a group under the dominant institutional environment of traditions and norms (Dufrene et al, 1992). It was also one of the potential reasons that the Navajo people survived the Long Walk between 1863 and 1866 (Denetdale, 2007). As its social & economic environment changes and formal institutions begin to matter, the Navajo Nation has suffered from failures of group decision-making. Although sharing the same culture, group interests overshadow social ties and trust as socio-economic conditions change. It has affected their efficiency of responding to exogenous shocks. One example is the allocation of federal funds. Despite the frequent insufficient federal support, how to timely and effectively allocate federal funds is a common obstacle that the Navajo decision-makers have to deal with. During the 2020 COVID-19 pandemic, the Navajo Nation received $716 million from the federal government through the CARES Act. However, there were different opinions within the tribal government regarding how to spend the fund (Smith, 2020), which potentially delayed the use of the fund and the relief of Navajo residents affected by the pandemic. If the relief fund can be used effectively and assistance programs can be implemented in a timely manner, the risk exposure to the COVID-19 pandemic would have been greatly reduced. Some tribal leaders believed that part of the relief fund should be allocated to improve essential infrastructures such as electricity, water systems, bathroom additions, and telecommunications access (Smith, 2020), which could generate long-term public health benefits and economic benefits. Externally, social trust is particularly relevant to individual healthcare and public health in the Navajo Nation. Similar to many federally recognized tribes, the Navajo Nation relies on the IHS clinics for primary healthcare. However, a typical issue is the instability of the federally sponsored healthcare system. For instance, doctors who work at the clinics are usually on a short rotation, ranging from several months to a few years. There is little incentive for both doctors and patients to build trust, connection, and stable communication mechanisms. A derivative consequence of the lack of communication and hence social trust is the lack of customized education on nutrition and public health. The latter is critical for pandemic preparation and disease prevention. Understaffing has been another challenge faced by the Navajo healthcare system (Albino et al., 2017; Kim, 2000). 5 Discussion This study reviewed the background and the literature on previous public health pandemics that Native Americans had faced. In the context of the COVID-19 pandemic, I conducted an institutional analysis of the social & economic environment of the Navajo Nation, incorporating conversations with Navajo scholars, to answer the question of why the Navajo Nation was hit so hard in the first two waves of the pandemic. The policy implications of the analysis can be explored from several aspects. At the informal institutions level, first, we should realize the values and social norms embedded in the Navajo social-economic life and culture. Such values and social norms are prevailing across Native American communities (Nebelkopf & King, 2003; Miller, 2018), although with variations. The design of federal assistance programs and their implementation should take into account these values and social norms to alleviate poverty and widespread underlying health conditions, which will help Native American communities to prepare for future pandemics in the long run. The same suggestion applies to any cooperative effort between the federal government and tribal governments. Ignoring the existence and value of Native American cultural history and social norms can only make policy implementation less effective and exacerbate any inconsistencies between federal policy and self-governance. In this regard, several programs proposed in the public health arena have proved successful, such as group counseling and wellness courts (Dufrene et al, 1992; Tipps et al., 2018). These programs usually consider a Native individual's cultural identity and traditional values carefully. Besides, giving attention to indigenous rights is an effective way to improve social trust in healthcare systems (Nelson & Wilson, 2021). Another essential aspect of enhancing Navajo's informal institutions is to build social trust and communication with the non-Native world, especially with the non-Native communities where some Navajo lives. As the 2010 US Census suggests, 10 % of Navajo people live in border towns surrounding the Navajo Nation (Navajo Nation Government, 2013). The social gap between a reservation and its border towns is not necessarily the smallest, but the geographical proximity is something that can be leveraged. Historically, during their decades and centuries of encountering and exchange with the non-Native world, Native American tribes have been able to prosper for most of the time. Setbacks did happen. For example, the incrimination of the unprecedented four-term chairman Peter MacDonald, who was once a popular public figure in the country, was a major setback in the Navajo Nation leadership in 1989 and led to the restructuring of the tribal government in 1991. Between 1989 and 2003, the Navajo Nation had six chairmen, and none of them served more than one term. But a short setback like this should not stop the effort to collaborate with non-Native communities, both nearby and far away. On another note, tribal governments mainly dealt with the federal government in the past. To build social trust and communication with the non-Native communities, it is also worthwhile to collaborate with local state and municipal governments. The recent Days Inn incident in New Mexico (Ihrig, 2020) highlights the importance of such local-level collaborations. At the formal institutions level, policymakers and tribal leaders should realize the importance of formal institutions and their relationships with culture and informal institutions. First and foremost, the established formal institutions should match the tribal culture. It is one of the four key factors for successful economic development identified by the Harvard Project on American Indian Economic Development (HPAIED, Cornell and Kalt (2000)). For instance, this is essentially the direction of the 1989 Navajo Nation Council reform (i.e., Navajo Nation Local Governance Act; for more discussion, see Morris (1991)). The new Act gave some local Navajo chapters the ability to make their own decisions. Another example is clear property rights. Dippel et al. (2020) showed that fee simple title (a form of freehold ownership, which gives the owner absolute property rights) adds between $4000 and $15,000 in value to an acre of land using allotment data from Native American reservations. The estimate does not necessarily reflect any potential value from the future improvement of lending opportunities with better collateral (the land), as well as any health effect of increased willingness to invest in home improvement. This monetary estimate illustrates the value of formal institutions. It is worth noting that the idea of titling reservation lands has had both successful cases (e.g., Alaska Native village corporations, see Dayo and Kofinas (2009)) and questionable ones (e.g., Hernando de Soto's design, see Otto (2009)). One implication from these mixed outcomes is that it is important to use informal institutions to overcome the shortfalls of formal institutions. More importantly, building efficient formal institutions is a critical step towards self-governance. Without strong self-governance, the resident health and economic livelihood on the reservation are always subject to the fluctuations of federal grants and funds. Federal support often does not go to residents on the reservation directly. Complaints tend to link such a way of fund distribution to corruption among tribal officials. Strong self-governance can reduce reliance on federal support and create employment opportunities for tribe members, which eradicates the root cause of corruption related to federal funds. It is consistent with Howitt (2012) that “sustainable Indigenous futures cannot arise from policy interventions that rely on creating wealth for state and corporate appropriation and assume enough of this wealth can be redistributed to local Indigenous communities to constitute ‘development.’” In addition, strong self-governance allows tribal governments to strategically utilize federal support. Historically, a lot of federal funds allocated to Native American communities were used for consumptive purposes. With strong self-governance and a strategic relationship with the federal government, much of the federal funds could be diverted to production purposes, such as investment in telecommunication infrastructure (e.g., broadband) and supporting small and medium enterprise development. Another essential component of building efficient formal institutions is to invest in the education system. As far as pandemic prevention is concerned, promoting public health education in Native American communities is critical. It is an indispensable vehicle for building health equity and eliminating health disparities. Over the last several decades, there has been a trend toward transferring public health programs from the IHS to tribal governments. Noren et al. (1998) argued that for the tribe-sponsored programs to be effective and sustainable, they need to be operated by skilled healthcare professionals and managers to adapt to the changes in the healthcare environment. Promoting public health education can help to address the shortage of nurses and physicians on reservations while managing the burden on their healthcare system in the long run. It is also an effective way to integrate informal institutions and formal institutions. For instance, Dignan et al. (1994) found that the unique cultures and the diversity of Native Americans are critical factors in developing successful health education programs for their communities. It can also help address the high turnover rates of doctors and nurses at the IHS clinics and grow social trust in the healthcare system. It is worth noting that it may be necessary for tribal governments to work with local state and municipal governments to develop successful public health education programs. First, the already-stretched federal funds and limited tribal resources make it difficult for tribal governments to invest in the needed education infrastructure from scratch. Collaborating with local state and municipal governments may allow them to share the already existing infrastructure. A recent example is a cooperative agreement between Navajo Technical University (a tribally controlled university) and a local public research university in New Mexico to address the drinking water quality issue on the reservation (Davis, 2020). Similar collaborations can be expanded to public health education. Second, working with local state governments on education programs will also help address the poor higher education situation for Native Americans. In that regard, Native American students have been struggling (Pruitt & Flores, 2020). Overall, it needs to stress that with policies and initiatives integrating both dimensions of tribal institutions, such as the suggested above, institutional resilience can be forged to prepare Native American reservations like the Navajo Nation for pandemics in the future. 6 Concluding remarks Based on reviewing the pandemic history of Native Americans and analyzing the institutional environment of the largest Native American reservation – the Navajo Nation, this study aims at understanding the disproportionate impact of the COVID-19 pandemic on Native American communities. I use the Navajo Nation as a case study to seek answers for the slow response to the pandemic among some Native American communities, especially during the beginning stage. Exploring its formal and informal institutions provides us a way to identify the critical factors behind the severity of the pandemic on the reservation compared to the non-Native world. The analysis indicates that there is a lack of both institutional resilience and healthcare preparation. They are essentially two intertwined aspects of the same challenge that the Navajo Nation and other Native American reservations have faced. Specifically, four factors can help explain the severity of the pandemic: prevalence of underlying chronic health conditions, lack of institutional resilience, the relationship between the federal government and tribal governments, and lack of social trust. They provide implications for the way to move forward. The Navajo Nation traditionally has robust informal institutions but weak formal institutions, especially institutions necessary for effective self-governance. Taking that into consideration, the study offers several policy suggestions. First, policymaking should balance the traditional culture and social norms of a Native American tribe and the public health needs instead of ignoring or suppressing them. Second, integrating informal institutions to build efficient formal institutions is an essential step towards self-governance. Besides, as a component of building strong self-governance, tribal economic development policy should increase investment in its education system. In particular, promoting public health education could help Native American communities to better prepare for pandemics in the future. In the long run, to avoid institutional lock-in or institution vacuum, local-level collaboration and cooperation should be promoted to enhance social trust and communication between Native and non-Native communities. One way is through strategically integrating Native American culture and regional economic development. It requires the federal government to work with tribal governments to reduce business transaction costs and improve financial opportunities for entrepreneurial activities. The momentum of economic growth will help to build resilience to, for example, chronic conditions and psychosocial stresses in the long run. Furthermore, addressing policy vacuum and policy ambiguity is as necessary as crafting new policies and agreements. For instance, according to van Dorn et al. (2020), “the IHS support does not cover care [for COVID-19] from external providers. Although there is a provision of the CARES Act stimulus bill that is intended to cover those costs, it is unclear how effective it would be if someone covered by the IHS is transferred to a non-IHS facility.” Hopefully, such policy ambiguity due to dual bureaucracy can be resolved soon in the wake of the COVID-19 pandemic. CRediT authorship contribution statement Haoying Wang: Conceptualization, Formal analysis, Literature Review, Visualization, Analysis, Writing – review & editing. Acknowledgements The author is grateful for the comments and suggestions made by editors Giulia Urso, Luca Storti, Neil Reid, and two anonymous reviewers. The author would also like to thank Historian Dr. Scott C. Zeman (deceased Sep 2020, a student of American Indian Historian Peter Iverson) and Navajo scholar Dr. Anne Gray (Lukachukai, AZ) for their suggestions. 1 According to the 2010 US Census on Navajo, the total Navajo population is 332,129 if including Navajos who live off the reservation. 2 Resilience is commonly defined as a system or society's ability to adapt to changes and its capacity to recover from shocks and maintain function (e.g., Holling, 1973; Rose, 2007). 3 Following the convention of human geography, the author identifies himself as non-indigenous. The paper hence presents a non-Native perspective on the issues studied. 4 In general, if a tribe member has employer-sponsored health insurance, it serves as the primary coverage. The HIS (Indian Health Service)-provided health services serve as secondary coverage. 5 This study defines informal and formal institutions following the literature of development studies (including development economics) and political science (e.g., see Helmke and Levitsky (2004) and Steer and Sen (2010)). Formal institutions include laws and regulations, the court system, written contracts. Informal institutions cover unwritten rules of behavior, social norms and rules with shared expectations, and social networks. 6 Peoplehood is a conceptual framework proposed by Holm et al. (2003) to transform American Indian studies. It is defined as an interconnected system of language, sacred history (storytelling tradition), place (territory), and ceremony. Peoplehood is different from informal institutions, but it can be considered part of Navajo culture according to Helmke and Levitsky (2004)'s framework. Note that peoplehood is different from the concept of citizenship in indigenous geography. The latter is a more fragmented concept (Staeheli, 2011). 7 See footnote 2 for the definition of resilience. 8 There were approximately 368 Indian treaties that had been ratified from 1778 to 1868 between the US Congress and Native Americans (Gover, 2014). 9 Some literature does consider social trust as a dimension of informal institutions (e.g., Lu et al., 2018). 10 According to Ihrig (2020), in November 2020, the City of Socorro Mayor Ravi Bhasker placed barricades outside the local Days Inn after finding out that the motel was being used to house COVID-19-positive patients and others who are in quarantine from the nearby Alamo Navajo Reservation. The incident caused debates and small protests among local residents. 11 In a 2020 interview with CBS News, the Navajo Nation president Jonathan Nez highlighted such a social and economic goal of the Navajo people. The documentary that contains the interview is available in Baek (2020). 12 https://navajoadvantage.com/pages/natrlrs.htm, accessed on Nov 10, 2020. ==== Refs References Albino J. Batliner T.S. Tiwari T. Preventing caries in American Indian children: Lost battle or new hope? JDR .Clin.Transl. Res. 2 4 2017 406 409 Baek G. Navajo Nation residents face coronavirus without running water https://www.cbsnews.com/news/coronavirus-navajo-nation-running-water-cbsn-originals/ 2020 accessed on Bekiempis V. Delta variant rapidly gaining ground in US west as vaccination rates stagnate https://www.theguardian.com/world/2021/jul/06/delta-variant-infections-us-west-vaccines 2021 accessed on Broussard B.A. Sugarman J.R. Bachman‐Carter K. Booth K. Stephenson L. Toward comprehensive obesity prevention programs in Native American communities Obesity Research 3 S2 1995 289s 297s 8581789 Butzer K.W. The Americas before and after 1492: An introduction to current geographical research Annals of the Association of American Geographers 82 3 1992 345 368 Campbell G.R. The changing dimension of native American health: A critical understanding of contemporary native American health issues American Indian Culture & Research Journal 13 3–4 1989 1 20 Carthey J. de Leval M. Reason J. Institutional resilience in healthcare systems BMJ Quality and Safety 10 1 2001 29 32 Castillo E.D. Blood came from their mouths: Tongva and Chumash responses to the pandemic of 1801 American Indian Culture & Research Journal 23 3 1999 47 61 Cattaneo L. Feir D. The higher price of mortgage financing for Native Americans Working Paper #1906 2019 Center for Indian Country Development Minneapolis, MN Chen S. Prettner K. Kuhn M. Bloom D.E. The economic burden of COVID-19 in the United States: Estimates and projections under an infection-based herd immunity approach J. Econ. Ageing 20 2021 100328 34123719 Cornell S. Kalt J.P. Where's the glue? Institutional and cultural foundations of American Indian economic development The Journal of Socio-Economics 29 5 2000 443 470 Cutler D.M. Summers L.H. The COVID-19 pandemic and the $16 trillion virus Journal of the American Medical Association 324 15 2020 1495 1496 33044484 Dabelea D. DeGroat J. Sorrelman C. Glass M. Percy C.A. Avery C. Diabetes in Navajo youth: Prevalence, incidence, and clinical characteristics: The SEARCH for diabetes in youth study Diabetes Care 32 S2 2009 S141 S147 19246579 Dahal S. Mizumoto K. Bolin B. Viboud C. Chowell G. Natality decline and spatial variation in excess death rates during the 1918–1920 influenza pandemic in Arizona, United States American Journal of Epidemiology 187 12 2018 2577 2584 30508194 Davis T. Universities to treat water on reservation 2020 https://www.abqjournal.com/1499673/universities-to-treat-water-on-reservation.html accessed Dayo D. Kofinas G. Institutional innovation in less than ideal conditions: Management of commons by an Alaska native village corporation International Journal of the Commons 4 1 2009 142 159 Denetdale J. The long walk: The forced Navajo exile 2007 Chelsea House Publications Broomall, PA Dignan M. Sharp P. Blinson K. Michielutte R. Konen J. Bell R. Lane C. Development of a cervical cancer education program for native American women in North Carolina Journal of Cancer Education 9 4 1994 235 242 Dippel C. Forced coexistence and economic development: Evidence from native American reservations Econometrica 82 6 2014 2131 2165 Dippel C. Frye D. Leonard B. Property rights without transfer rights: A study of Indian land allotment 2020 National Bureau of Economic Research NBER Working Paper #w27479 Dufrene P.M. Coleman V.D. Gainor K.A. Counseling native Americans: Guidelines for group process Journal for Specialists in Group Work 17 4 1992 229 234 Dyer O. Covid-19: Black people and other minorities are hardest hit in US BMJ 369 2020 m1483 32291262 Esparza J. Fox C. Harper I.T. Bennett P.H. Schulz L.O. Valencia M.E. Ravussin E. Daily energy expenditure in Mexican and USA pima Indians: Low physical activity as a possible cause of obesity International Journal of Obesity 24 1 2000 55 59 10702751 Feir D. Akee R. First Peoples lost: Determining the state of status First Nations mortality in Canada using administrative data Canadian Journal of Economics 52 2 2019 490 525 Feir D. Gillezeau R. Return to the homeland? The impact of the great recession on employment outcomes and labor mobility for native Americans J. Econ.Race.Pol. 1 2–3 2018 60 74 Frank J.W. Moore R.S. Ames G.M. Historical and cultural roots of drinking problems among American Indians American Journal of Public Health 90 3 2000 344 351 10705850 Gover K. Nation to nation: Treaties between the United States and American Indian Nations Am. Indian Art Mag. 15 2 2014 37 39 Grandbois D.M. Sanders G.F. Resilience and stereotyping: The experiences of Native American elders Journal of Transcultural Nursing 23 4 2012 389 396 22786944 Groom A.V. Jim C. LaRoque M. Mason C. McLaughlin J. Neel L. Pandemic influenza preparedness and vulnerable populations in tribal communities American Journal of Public Health 99 S2 2009 S271 S278 19461107 Hamill S. Dickey M. Cultural competence: What is needed in working with Native Americans with HIV/AIDS? Journal of the Association of Nurses in AIDS Care 16 4 2005 64 69 Hatcher S.M. Agnew-Brune C. Anderson M. Zambrano L.D. Rose C.E. COVID-19 among American Indian and Alaska native persons - 23 states, january 31–july 3, 2020 Morbidity and Mortality Weekly Report 69 34 2020 1166 1169 32853193 Holling C.S. Resilience and stability of ecological systems Annual Review of Ecology and Systematics 4 1 1973 1 23 Holm T. Pearson J.D. Chavis B. Peoplehood: A model for the extension of sovereignty in American Indian studies Wicazo Sa Review 18 1 2003 7 24 Hotez P.J. Neglected infections of poverty in the United States of America PLoS Neglected Tropical Diseases 2 6 2008 e256 18575621 Howitt R. Sustainable indigenous futures in remote indigenous areas: Relationships, processes and failed state approaches Geojournal 77 6 2012 817 828 Hutchins S.S. Fiscella K. Levine R.S. Ompad D.C. McDonald M. Protection of racial/ethnic minority populations during an influenza pandemic American Journal of Public Health 99 S2 2009 S261 S270 19797739 Ihrig C. Socorro mayor decries ‘peculiar situation Albuq.J.Nova 17 2020 2020 Iseke J. Indigenous storytelling as research International Review of Qualitative Research 6 4 2013 559 577 Iverson P. The Navajo nation 1981 University of New Mexico Press Albuquerque, NM Jensen S. Svendsen G.T. Social trust, safety and the choice of tourist destination Business and Management Horizons 4 1 2016 1 9 Jernigan V.B.B. D'Amico E.J. Duran B. Buchwald D. Multilevel and community-level interventions with Native Americans: Challenges and opportunities Prevention Science 21 1 2020 65 73 Jones T. MacDonald faces cultural conflicts: Embattled Navajo leader tries to straddle 2 worlds https://www.latimes.com/archives/la-xpm-1989-03-13-mn-448-story.html 1989 accessed on Kim C. Recruitment and retention in the Navajo area Indian health service Western Journal of Medicine 173 4 2000 240 243 11017982 La Ruche G. Tarantola A. Barboza P. Vaillant L. Gueguen J. The 2009 pandemic H1N1 influenza and indigenous populations of the Americas and the Pacific Euro Surveillance 14 42 2009 19366 19883543 Larsen S.C. Johnson J.T. In between worlds: Place, experience, and research in indigenous geography Journal of Cultural Geography 29 1 2012 1 13 Laurencin C.T. McClinton A. The COVID-19 pandemic: A call to action to identify and address racial and ethnic disparities J. Racial.Ethn. Health.Disparities 7 2020 398 402 32306369 Leeuw S.D. Maurice S. Holyk T. Greenwood M. Adam W. With reserves: Colonial geographies and First Nations health Annals of the Association of American Geographers 102 5 2012 904 911 Lerma M. Shocks to the Navajo (Diné) political system: Resiliency of traditional Diné institutions in the face of colonial interaction (Contact to 1923) Indig.Policy J. 25 1 2014 1 20 Lofthouse J.K. Institutions and economic development on Native American lands Independent Review 24 2 2019 227 248 Lu J.W. Song Y. Shan M. Social trust in subnational regions and foreign subsidiary performance: Evidence from foreign investments in China Journal of International Business Studies 49 6 2018 761 773 Martin D.L. Goodman A.H. Health conditions before columbus: Paleopathology of native North Americans Western Journal of Medicine 176 1 2002 65 68 11788545 McLafferty S. Placing pandemics: Geographical dimensions of vulnerability and spread Eurasian Geography and Economics 51 2 2010 143 161 McLaughlin S. Traditions and diabetes prevention: A healthy path for native Americans Diabetes Spectrum 23 4 2010 272 277 Miller R.J. Sovereign resilience: Reviving private-sector economic institutions in Indian country BYU Law Review 2018 1331 Morris C. Navajo nation council reforms American Indian Law Review 16 2 1991 613 617 Navajo Nation Department of Fish and Wildlife Climate adaptation plan for the Navajo nation 2018 Navajo Nation Window Rock, AZ Navajo Nation Government Navajo population profile - 2010 US Census https://www.nec.navajo-nsn.gov/Portals/0/Reports/NN2010PopulationProfile.pdf 2013 accessed on Nebelkopf E. King J. A holistic system of care for Native Americans in an urban environment Journal of Psychoactive Drugs 35 1 2003 43 52 12733757 Necefer L. Wong-Parodi G. Jaramillo P. Small M.J. Energy development and native Americans: Values and beliefs about energy from the Navajo nation Energy Res.Soc. Sci. 7 2015 1 11 Nelson S.E. Wilson K. Rights and health versus rights to health: Bringing Indigenous Peoples’ legal rights into the spaces of health care services Political Geography 85 2021 102311 Newton C. Why has Navajo Nation been hit so hard by the coronavirus? https://www.aljazeera.com/news/2020/5/27/why-has-navajo-nation-been-hit-so-hard-by-the-coronavirus 2020 accessed on Noren J. Kindig D. Sprenger A. Challenges to Native American health care Public Health Reports 113 1 1998 22 33 9885525 Otto J.M. Rule of law promotion, land tenure and poverty alleviation: Questioning the assumptions of Hernando de Soto Hague J.Rule. Law 1 1 2009 173 194 Pejovich S. The effects of the interaction of formal and informal institutions on social stability and economic development Journal of Markets and Morality 2 2 1999 164 181 Plough A. Bristow B. Fielding J. Caldwell S. Khan S. Pandemics and health equity: Lessons learned from the H1N1 response in Los Angeles county Journal of Public Health Management and Practice 17 1 2011 20 27 21135657 Pruitt L.R. Flores D.E. Access to higher education in New Mexico: Racial, ethnic and geographic disparities 10.2139/ssrn.3677575 2020 Rose A. Economic resilience to natural and man-made disasters: Multidisciplinary origins and contextual dimensions Environmental Hazards 7 4 2007 383 398 Sage F. Storytelling is indigenous science AGU Fall Meeting Abstracts 2019 2019 U32A-03 Smith N.L. Nez defends stance for funding Navajo Nation's COVID-19 financial relief program https://www.daily-times.com/story/news/local/navajo-nation/2020/12/04/navajo-nation-president-defends-stance-covid-19-financial-relief-program/3810767001/ 2020 accessed Spirling A. US treaty making with American Indians: Institutional change and relative power, 1784–1911 American Journal of Political Science 56 1 2012 84 97 Staeheli L.A. Political geography: Where's citizenship? Progress in Human Geography 35 3 2011 393 400 Steckel R.H. Prince J.M. Tallest in the world: Native Americans of the great Plains in the nineteenth century The American Economic Review 91 1 2001 287 294 Stevens J. Suchindran C. Ring K. Baggett C.D. Jobe J.B. Story M. Caballero B. Physical activity as a predictor of body composition in American Indian children Obesity Research 12 12 2004 1974 1980 15687399 Sutcliffe C.G. Grant L.R. Reid A. Douglass G.K. Weatherholtz R.C. The burden of Staphylococcus aureus among Native Americans on the Navajo Nation PloS One 14 3 2019 e0213207 Tatem A.J. Rogers D.J. Hay S.I. Global transport networks and infectious disease spread Advances in Parasitology 62 2006 293 343 16647974 Tipps R.T. Buzzard G.T. McDougall J.A. The opioid epidemic in Indian Country Journal of Law Medicine & Ethics 46 2 2018 422 436 Tolan S. Showdown at window rock N. Y. Times Mag.November 26 1989 1989 Us Cdc Deaths related to 2009 pandemic influenza A(H1N1) among American Indian/Alaska Natives - 12 states Morbidity and Mortality Weekly Report 58 48 2009 1341 1344 20010508 van Dorn A. Cooney R.E. Sabin M.L. COVID-19 exacerbating inequalities in the US Lancet 395 10232 2020 1243 32305087 Wiedman D. Native American embodiment of the chronicities of modernity: Reservation food, diabetes, and the metabolic syndrome among the Kiowa, Comanche, and Apache Medical Anthropology Quarterly 26 4 2012 595 612 23361887 Wu A.W. Connors C. Everly G.S. Jr. COVID-19: Peer support and crisis communication strategies to promote institutional resilience Annals of Internal Medicine 172 12 2020 822 823 32251512 Yonk R.M. Hoffer S. Stein D. Disincentives to business development on the Navajo nation Journal of Developmental Entrepreneurship 22 2017 1750012 02

0

PMC9753121

NO-CC CODE

2022-12-16 23:25:30

no

Appl Geogr. 2021 Sep 29; 134:102526

utf-8

Appl Geogr

10.1016/j.apgeog.2021.102526

oa_other

==== Front J Res Pers J Res Pers Journal of Research in Personality 0092-6566 0092-6566 Elsevier Inc. S0092-6566(21)00012-X 10.1016/j.jrp.2021.104075 104075 Article Stockpiling during the COVID-19 pandemic as a real-life social dilemma: A person-situation perspective Fischer Moritz a⁎ Twardawski Mathias a Steindorf Lena b Thielmann Isabel c a Ludwig-Maximilians-Universität München, Munich, Germany b Universität Heidelberg, Heidelberg, Germany c Universität Koblenz-Landau, Landau, Germany ⁎ Corresponding author at: Ludwig-Maximilians-Universität München, Department of Psychology, Leopoldstrasse 13, 80802 Munich, Germany. 10 2 2021 4 2021 10 2 2021 91 104075104075 9 9 2020 1 2 2021 2 2 2021 © 2021 Elsevier Inc. All rights reserved. 2021 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Prior research using economic games has shown that personality drives cooperation in social dilemmas. In this study, we tested the generalizability of these findings in a real-life social dilemma during the COVID-19 pandemic, namely stockpiling in the presence of low versus high resource scarcity. Honesty-Humility was negatively related to stockpiling intentions and justifiability of stockpiling. Moreover, we found a positive albeit weaker effect of Emotionality on stockpiling intentions. Victim Sensitivity was mostly positively associated with stockpiling intentions. None of the personality traits interacted with resource scarcity to predict stockpiling. Our findings replicate established associations between personality and cooperation in a real-life social dilemma, and suggest that the characteristics of interdependent situations during a pandemic additionally afford the expression of Emotionality. Keywords Social dilemma Stockpiling Personality HEXACO Justice sensitivity ==== Body pmc1 Introduction The COVID-19 pandemic reached Europe and North America during the early months of 2020. At that time, many people started to stockpile medical supplies and other goods, arguably to prepare for quarantines, national shut-downs, and anticipated disruptions of supply chains. In Germany, for instance, the purchase of soap more than doubled in the last week of February 2020, and purchase of sanitizer even increased sevenfold in comparison to the same week in 2019 (Statistisches Bundesamt, 2020). Sales surges were also reported for durable foods such as flour, yeast, or pasta (Statistisches Bundesamt, 2020). Although having an adequate stock of these goods can be individually beneficial in times of crisis, stockpiling may pose negative consequences on others: Some stores may experience short-term supply shortages as a consequence of individual stockpiling, and other customers may eventually be left empty-handed. In the present article, we argue that stockpiling during a pandemic can be conceptualized as a social dilemma in a naturalistic setting, and that insights from standardized economic games modeling social dilemmas can be used to predict stockpiling. To test this claim, we recruited a large sample of German adults during the emerging COVID-19 pandemic and investigated to what extent well-known predictors of cooperation in game-based social dilemmas – such as broad personality traits (i.e., HEXACO dimensions), more specific traits (i.e., Victim Sensitivity), and context factors (i.e., resource scarcity) – account for stockpiling intentions and justifiability of stockpiling behavior. Social dilemmas are “situations in which a non-cooperative course of action is (at times) tempting for each individual in that it yields superior (often short-term) outcomes for self, and if all pursue this non-cooperative course of action, all are (often in the longer-term) worse off than if all had cooperated” (Van Lange, Joireman, Parks, & Van Dijk, 2013, p. 126). Various types of social dilemmas exist that differ structurally, for example with regard to the number of individuals involved (i.e., dyads versus groups), the duration of interaction between individuals (e.g., one-shot versus repeated interaction), and whether “cooperation” refers to contribution to or non-consumption of a common resource (Van Lange et al., 2013). Stockpiling during a pandemic arguably constitutes a real-life social dilemma that involves a group of individuals (e.g., customers of a supermarket) engaging in repeated interactions (e.g., multiple purchases over the course of a pandemic) and in which cooperation is reflected in refraining from overconsuming a common resource (e.g., groceries or sanitary products). In fact, the situational structure of the stockpiling dilemma is best described by the “tragedy of the commons” or a so-called commons dilemma (Hardin, 1968). In a commons dilemma, members of a group decide how much to harvest from a common resource. The amount each individual consumes is no longer available for other group members. Usually, the interaction (i.e., individuals harvesting from a common resource) is repeated over a certain amount of time, with the resource remaining after each interaction recovering at a certain reproduction rate. The interaction ends once the common resource is fully depleted, that is, once individuals’ consumption exceeded the reproduction rate over a certain amount of time. During the COVID-19 pandemic, individuals likewise decided how much to consume from a common resource such as groceries or sanitary products. The common resource would suffice for most or even all individuals if a sufficient number of individuals cooperated by refraining from overconsumption. However, if too many individuals decide to defect by purchasing more than they actually need, the common resource will become depleted eventually. Empirical studies in behavioral economics and psychology typically use standardized economic games which provide a precise and parsimonious approach to model social dilemmas and to investigate cooperation in controlled experimental settings (Murnighan & Wang, 2016). However, the real-life social dilemma of stockpiling during a pandemic is arguably more complex than highly controlled game-like situations. In a commons dilemma in form of an economic game, for example, the resource remaining after each round usually replenishes at a certain – and often known – reproduction rate. Thus, resource quantity in a subsequent round depends on both the remaining resource from the previous round and the reproduction rate. In the stockpiling case, however, the reproduction rate is represented by the amount a supermarket restocks and thus, is not necessarily linked to the resource quantity that was left on a given day. Another aspect that may be specific about the stockpiling dilemma is that consumption has short-term monetary costs – in the sense that buying stocks of groceries or sanitary products is costly – whereas consumption in a game-based commons dilemma is non-costly in short-term but exclusively beneficial to the individual. In addition, stockpiling during a pandemic is characterized by more uncertainty and insecurity than non-cooperation within a respective economic game. This is because in economic games, individuals are usually fully informed about the structure of the situation. For example, in a game-based commons dilemma, individuals usually know about the reproduction rate and they are also aware of the consequences their own and the interaction partners’ actions have. By contrast, in the stockpiling case, information about reproduction (i.e., whether or not a store restocks overnight) is essentially unavailable, and it is therefore also unknown to individuals whether their (over)consumption will affect their own and others’ outcomes in the future. Uncertainty and insecurity about the structure of the situation are unique features of this real-life social dilemma that are typically not (or at least not comparably) apparent in a game-based commons dilemma. Consequently, stockpiling is likely to serve an emotional function in the sense that having a stock of the respective goods may reduce stress, anxiety, and feelings of uncertainty. Taken together, the stockpiling dilemma has conceptual similarities with a game-based commons dilemma but also incorporates some differences. Therefore, it is unclear whether insights on cooperative behavior from game-based social dilemmas can be transferred to stockpiling during the COVID-19 pandemic. Addressing this issue, the aim of the present research is to examine whether personality traits and context factors that have been shown to predict cooperation in economic games also relate to stockpiling in real-life. To this end, we assessed two forms of responses to the stockpiling dilemma, namely (i) how strongly individuals intend to stockpile (i.e., stockpiling intentions) and (ii) how justifiable they evaluate stockpiling in general (i.e., justifiability of stockpiling behavior). We differentiated these constructs to reflect the conceptual distinction between behavioral intentions to stockpiling and judgements of stockpiling behavior, respectively (for a similar approach, see Dammeyer, 2020). This distinction was also implied by the fact that some personality traits should be specifically linked to stockpiling intentions (i.e., Emotionality and Victim Sensitivity) whereas other should be specifically linked to justifiability of stockpiling behavior (i.e., Agreeableness), and even others should be associated with both (i.e., Honesty-Humility). Prior research using game-based social dilemmas to study cooperation has consistently shown that individuals differ in the extent to which they cooperate and that personality traits can account for such interindividual variation (for a meta-analysis, see Thielmann, Spadaro et al., 2020). Importantly, personality traits are more or less relevant for cooperation in certain situations, depending on the situational structure of the respective social dilemma. In other words, social dilemmas entail situational properties (i.e., affordances) that enable the expression of certain traits in behavior. Specifically, it has been proposed that social dilemmas may primarily involve four such affordances, namely possibility for exploitation, possibility for reciprocity, dependence on others under uncertainty, and temporal conflict between short- and long-term interests (Thielmann, Spadaro et al., 2020). In the following section, we discuss which of these affordances are particularly relevant for the stockpiling dilemma and employ this affordance perspective to derive hypotheses regarding the impact of personality and context factors on stockpiling. In addition to these four key affordances, Thielmann, Spadaro et al., (2020) proposed several subaffordances of the exploitation and reciprocity affordances that specifically allow the expression of certain social motives (e.g., altruism, fairness, greed etc.) in behavior. However, given that additional consideration of these subaffordances would have resulted in the exact same hypotheses, we will not refer to these subaffordances further in what follows. First, the stockpiling dilemma involves a possibility for exploitation because individuals can increase their outcomes without fearing retaliation by their interaction partners. That is, individuals can stockpile certain goods without expecting sanctions from other customers or authorities. In such situations, Honesty-Humility from the HEXACO model should become relevant. Specifically, Honesty-Humility denotes the “tendency to be fair and genuine in dealing with others, in the sense of cooperating with others even when one might exploit them without suffering retaliation” (Ashton & Lee, 2007, p. 156). Thus, Honesty-Humility can be understood to represent active cooperativeness that drives cooperative behavior particularly in situations providing a possibility for exploitation (Hilbig et al., 2013, Hilbig et al., 2018) – a proposition that has also been meta-analytically confirmed (Thielmann, Spadaro et al., 2020). Correspondingly, recent studies also found Honesty-Humility to be negatively associated with stockpiling intentions (Columbus, 2020). A large meta-analytic investigation across five samples in Denmark and Germany also found a significant yet very small correlation of r = −0.04 between Honesty-Humility and (self-reported) hoarding behavior during the COVID-19 pandemic (Zettler, Schild, et al., 2020). Note, however, that this correlation was non-significant (on the rather conservative significance level of p < .010) when considering each sample in isolation (Zettler, Schild, et al., 2020) and that another study also found a non-significant relationship between Honesty-Humility and stockpiling of toilet paper (Garbe, Rau, & Toppe, 2020). Nonetheless, based on the theoretical conceptualization of Honesty-Humility as a tendency of non-exploitation towards others (Ashton & Lee, 2007), we predicted a negative association of Honesty-Humility with stockpiling intentions as well as with justifiability of stockpiling behavior. Second, stockpiling entails a possibility for reciprocity, in the sense that individuals can respond to others’ behavior. For instance, one indirect proxy for other customers’ stockpiling is the availability of resources in the supermarket: If certain resources are scarce during the weekly shopping trip, the impression that other people stockpile becomes more salient. If these resources are, however, available in abundance, people might conclude that others do, apparently, refrain from stockpiling. Indeed, this context factor seems particularly relevant as implied by research showing that a key predictor of cooperation in interdependent situations is the expectation that others will behave cooperatively (for meta-analyses, see Balliet and Van Lange, 2013, Pletzer et al., 2018). By contrast, individuals expecting that others might take advantage of those who cooperate (e.g., through stockpiling) should behave uncooperatively, too, to avoid being exploited. We therefore manipulated resource scarcity (high versus low) to induce beliefs about the (un)cooperativeness of others. Our prediction was that high (as compared to low) resource scarcity would lead to more stockpiling intentions and stronger justifiability of stockpiling behavior. In terms of personality traits, the situational affordance of reciprocity should be particularly associated with HEXACO Agreeableness which represents “the tendency to be forgiving and tolerant of others, in the sense of cooperating with others even when one might be suffering exploitation by them” (Ashton & Lee, 2007, p. 156). Thus, Agreeableness captures reactive forms of cooperativeness (Hilbig et al., 2013), manifesting itself in more cooperation in situations in which one can reciprocate another’s uncooperative behavior (Hilbig et al., 2016, Thielmann et al., 2020). Therefore, Agreeableness is most likely to be expressed in lenient attitudes towards stockpiling behavior, and we consequently hypothesized a positive association of Agreeableness with justifiability of stockpiling behavior. In turn, because Agreeableness is particularly likely to play out if others behave uncooperatively, we also predicted an interaction of Agreeableness with resource scarcity on justifiability of stockpiling behavior, in the sense that the positive association of Agreeableness should be stronger under high (versus low) resource scarcity. Third, stockpiling during the COVID-19 pandemic entails dependence under uncertainty. That is, individuals do not have full control over their individual outcomes (i.e., the type and amount of goods they can purchase) and often learn about others’ behavior (i.e., whether or not others stockpile) only after having made their own decision. For example, customers individually have to decide how much to purchase without knowing how other customers will behave in the future. In such situations, beliefs about others (un)cooperative behavior may guide individual decisions (Thielmann, Spadaro et al., 2020). A personality trait that should be afforded by the situational characteristic of dependence under uncertainty is Victim Sensitivity (Schmitt, Neumann, & Montada, 1995). This construct captures a stable tendency to perceive situations as unjust to oneself, and to show strong cognitive (e.g., rumination), emotional (e.g., anger), and behavioral reactions (e.g., punitiveness) to this perceived injustice (Schmitt, Gollwitzer, Maes, & Arbach, 2005). Victim Sensitivity often predicts less prosocial behavior (Fetchenhauer and Huang, 2004, Gollwitzer et al., 2005), but victim sensitive individuals are not per se uncooperative or selfish. Rather, their willingness to cooperate is particularly low when they fear being exploited (Baumert et al., 2020). The notion that Victim Sensitivity involves a fear of being exploited also lies at the heart of the sensitivity to mean intentions (SeMi) model (Gollwitzer and Rothmund, 2009, Gollwitzer et al., 2013). This model builds on the assumption that victim sensitive individuals particularly attend to cues of others’ untrustworthiness in socially uncertain situations. In the presence of such cues, a so-called “suspicious mindset” is activated. As a consequence, victim sensitive individuals revoke cooperation to preemptively avoid being exploited by others (Gollwitzer et al., 2013). Based on this evidence and the SeMi model, we expected Victim Sensitivity to interact with resource scarcity – a situational cue indicating others’ untrustworthiness – to predict stockpiling. Specifically, we hypothesized that Victim Sensitivity will be positively related to stockpiling intentions under high resource scarcity, but less so under low resource scarcity. In addition to being sensitive to injustice from a victim’s perspective, people can also be sensitive to injustice as neutral observers (“Observer Sensitivity”), beneficiaries (“Beneficiary Sensitivity”), or perpetrators (“Perpetrator Sensitivity”). We did not derive any hypotheses about the effects of these perspectives of Justice Sensitivity on our dependent variables. However, we measured them for the sake of completeness and controlled for their shared variance in some of our analyses. The hypotheses we derived so far are based on three situational affordances (i.e., possibility for exploitation, possibility for reciprocity, and dependence under uncertainty) that have been identified to be present in social dilemmas (Thielmann, Spadaro et al., 2020). As detailed above, the real-life dilemma of stockpiling is, however, more complex than a game-based social dilemma and should therefore involve additional situational affordances. More specifically, stockpiling may help to cope with anxiety, stress, and feelings of uncertainty elicited by the pandemic and thereby serve an emotional function (e.g., by assuring to have a personal stock of the goods one needs). Thus, we argue that insecurity should be another situational affordance involved in the stockpiling situation. This situational affordance has been identified in the so-called situation, trait and outcome activation (STOA) model and should particularly allow the expression of HEXACO Emotionality (De Vries, Tybur, Pollet, & van Vugt, 2016). According to Ashton and Lee (2007), “[…] Emotionality represents tendencies relevant to the construct of kin altruism (Hamilton, 1964), including not only empathic concern and emotional attachment toward close others (who tend to be one’s kin) but also the harm-avoidant and help-seeking behaviors that are associated with investment in kin” (p. 156). This suggests that individuals high in Emotionality might additionally view stockpiling as a means to care for one’s family and close others if goods are bought for or shared with them. Supporting this reasoning, a recent study showed a positive relation between HEXACO Emotionality and stockpiling of toilet paper (Garbe et al., 2020). Moreover, another study found that Big Five Neuroticism – which has close conceptual links to Emotionality (Ashton, Lee, & De Vries, 2014) – was positively correlated with self-reported stockpiling during the COVID-19 pandemic (Dammeyer, 2020). By contrast, however, other studies found non-significant relationships of Big Five Neuroticism (Aschwanden et al., 2021) and/or HEXACO Emotionality (Zettler, Schild, et al., 2020) with self-reported hoarding behavior. Based on the affordance perspective, we hypothesized a positive relation of Emotionality with stockpiling intentions. In addition, we predicted an interaction effect of Emotionality and resource scarcity on stockpiling intentions in the sense that the positive association of Emotionality should be stronger under high (versus low) resource scarcity. The goal of the present study was to test whether associations of personality traits with prosocial behavior from economic games can be transferred to stockpiling during the COVID-19 pandemic as a real-life social dilemma. We conducted an online experiment during the beginning of the COVID-19 pandemic in Germany. At this time, reports about stockpiling were highly prevalent in the media and clearly observable in many stores. We manipulated resource scarcity (high versus low) in a fictional shopping scenario and examined the extent to which personality (i.e., HEXACO dimensions and Victim Sensitivity) predicts participants’ self-reported stockpiling intentions and justifiability of stockpiling behavior.1 2 Method We report how we determined our sample size, all data exclusions, all manipulations, and all measures in the study (Simmons, Nelson, & Simonsohn, 2012). The study materials, dataset, and analysis script are available on the Open Science Framework (OSF; https://osf.io/wbxnr/). 2.1 Context Our study was online for 14 days between March 23 and April 5, 2020. This period was characterized by strict governmental regulations in response to the rapid spread of COVID-19 in Germany. Universities, schools, and kindergartens in most German states were closed after March 13. Gatherings of more than two people who were not living together were prohibited as of March 22. Likewise, restaurants and bars were closed as of March 22 and were only allowed to offer takeaway food. Supermarkets and pharmacies were open throughout but some products (e.g., disinfectants, face masks, toilet paper, or flour) were sold out regularly, presumably as a consequence of stockpiling (for example, see Diemand, Jansen, Müssgens, Piller, & Preuss, 2020). Stockpiling tendencies decreased again around mid-April 2020 (Statistisches Bundesamt, 2020). In addition, prices for certain products (e.g. face masks, hand sanitizer) might have increased over the course of the pandemic, particularly in online reselling shops. By contrast, prices for groceries in German supermarkets did not increase significantly during that time. For the sake of comparability between the stockpiling dilemma and game-based commons dilemmas, we wanted to rule the possibility out that participants stockpiled simply because they anticipated further price increases (e.g., for face masks), and thus, focused in the present study on stockpiling of groceries in supermarkets. 2.2 Measures 2.2.1 HEXACO dimensions We used the German version (Moshagen, Hilbig, & Zettler, 2014) of the 60-item HEXACO Personality Inventory Revised (Ashton & Lee, 2009) to measure the six HEXACO dimensions. Items were presented on one survey page, in randomized order. Participants rated their agreement with each item on a 5-point response scale, ranging from 1 = strongly disagree to 5 = strongly agree. Internal consistencies were satisfactory for all dimensions (see Table 1 ). We also embedded an instructed attention check (i.e., “Please select ‘strong agreement’ here (this serves to check your attention).”) in the HEXACO items.Table 1 Correlations and internal consistencies for main study variables. Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Honesty-Humility 0.76 [0.73, 0.79] 2. Emotionality −0.03 0.80 [−0.10, 0.04] [0.78, 0.82] 3. Extraversion 0.06 −0.21*** 0.78 [−0.01, 0.13] [−0.28, −0.14] [0.76, 0.81] 4. Agreeableness 0.20*** −0.12*** 0.07 0.75 [0.14, 0.27] [−0.19, −0.05] [−0.00, 0.14] [0.72, 0.78] 5. Conscientiousness 0.06 −0.01 0.07 −0.07* 0.77 [−0.01, 0.13] [−0.08, 0.06] [−0.00, 0.14] [−0.14, −0.00] [0.74, 0.79] 6. Openness to Experience 0.11** −0.07 0.20*** 0.01 0.06 0.71 [0.04, 0.17] [−0.14, 0.00] [0.13, 0.27] [−0.06, 0.08] [−0.01, 0.13] [0.68, 0.74] 7. Victim Sensitivity −0.31*** 0.23*** −0.24*** −0.22*** −0.01 −0.18*** 0.75 [−0.37, −0.24] [0.16, 0.29] [−0.30, −0.17] [−0.28, −0.15] [−0.08, 0.06] [−0.24, −0.11] [0.71, 0.80] 8. Observer Sensitivity 0.10** 0.19*** 0.01 −0.00 0.05 0.13*** 0.26*** 0.69 [0.03, 0.17] [0.12, 0.26] [−0.06, 0.08] [−0.07, 0.07] [−0.02, 0.12] [0.06, 0.20] [0.19, 0.32] [0.63, 0.74] 9. Beneficiary Sensitivity 0.20*** 0.30*** −0.10** 0.03 0.03 0.06 0.19*** 0.40*** 0.75 [0.13, 0.27] [0.24, 0.36] [−0.17, −0.03] [−0.04, 0.10] [−0.04, 0.10] [−0.01, 0.13] [0.12, 0.26] [0.34, 0.46] [0.72, 0.79] 10. Perpetrator Sensitivity 0.43*** 0.18*** 0.01 0.14*** 0.07 0.12*** 0.01 0.29*** 0.34*** 0.76 [0.37, 0.48] [0.11, 0.25] [−0.06, 0.08] [0.07, 0.21] [−0.00, 0.14] [0.05, 0.19] [−0.06, 0.08] [0.23, 0.36] [0.28, 0.40] [0.71, 0.81] 11. Resource scarcity: High (versus low) −0.01 −0.06 0.05 0.03 0.04 0.00 −0.01 −0.05 0.01 0.02 – [−0.08, 0.06] [−0.13, 0.00] [−0.02, 0.11] [−0.04, 0.10] [−0.03, 0.11] [−0.07, 0.07] [−0.08, 0.06] [−0.12, 0.02] [−0.06, 0.08] [−0.05, 0.09] [−, −] 12. Stockpiling intentions −0.21*** 0.10** −0.06 −0.04 −0.05 −0.04 0.14*** 0.03 −0.09* −0.08* 0.07 0.86 [−0.27, −0.14] [0.03, 0.17] [−0.13, 0.01] [−0.11, 0.03] [−0.12, 0.02] [−0.11, 0.03] [0.07, 0.20] [−0.04, 0.10] [−0.15, −0.02] [−0.15, −0.01] [−0.00, 0.14] [0.83, 0.89] 13. Justifiability of stockpiling −0.17*** 0.05 −0.07* 0.04 −0.06 −0.06 0.14*** −0.03 −0.11** −0.08* −0.03 0.64*** 0.82 [−0.23, −0.10] [−0.02, 0.12] [−0.14, −0.00] [−0.03, 0.10] [−0.13, 0.01] [−0.13, 0.01] [0.07, 0.20] [−0.10, 0.04] [−0.18, −0.04] [−0.15, −0.01] [−0.10, 0.04] [0.60, 0.68] [0.79, 0.85] Note. Estimates in the lower part are Pearson’s zero-order correlations. Resource scarcity was experimentally manipulated (low vs. high) and thus, these are point biserial correlations. Resource scarcity is coded as −1 = low and 1 = high. Estimates on the diagonal (in italics) are omega (ωTotal) for measures with three or more items (i.e., the HEXACO dimensions and justifiability of stockpiling) as recommended by Dunn, Baguley, and Brunsden (2014). For two-item measures (i.e., Justice Sensitivity perspectives and stockpiling intentions), estimates on the diagonal are Spearman-Brown coefficients (Eisinga, Te Grotenhuis, & Pelzer, 2013). Numbers in brackets represent lower and upper bounds of 95% confidence intervals. Confidence intervals for internal consistency measures were bootstrapped with 5000 iterations. * p < .05. ** p < .01. *** p < .001. 2.2.2 Justice Sensitivity We applied the Justice Sensitivity Inventory in form of its German short version comprising eight items (Baumert et al., 2014). Participants completed the Victim Sensitivity items first before continuing with the Observer, Beneficiary, and the Perpetrator Sensitivity items in that order. The order of the two items within each subscale was randomized. Participants rated each item on a 5-point response scale, ranging from 1 = strongly disagree to 5 = strongly agree. We followed the recommendation by Eisinga, Te Grotenhuis, & Pelzer, 2013 and report the Spearman-Brown coefficients as internal consistency estimates for all two-item scales. This procedure yielded an estimate of ρ = 0.75 for Victim Sensitivity. 2.2.3 Dependent variables We used self-developed measures to assess our dependent variables (DVs). Stockpiling intentions were assessed using two items reading “I would buy more products than I normally need for a week” and “I would only buy as many products as I normally need for a week” (the latter was reverse-coded; ρ = 0.86). Justifiability of stockpiling behavior was measured using four self-developed items. More specifically, the items were “In such a situation, it is… […perfectly fine /…absolutely appropriate /…egoistic /…irresponsible] to buy more than one actually needs” (the latter two reverse-coded; ωTotal = 0.82). All six items pertaining to our dependent variables were rated on a 6-point response scale, ranging from 1 = strongly disagree to 6 = strongly agree. Items were presented in random order. 2.2.4 Measures for exploratory purposes We also measured additional variables for exploratory purposes. Specially, we assessed Need for Cognition with a short four-item scale (Beißert, Köhler, Rempel, & Beierlein, 2015; ωTotal = 0.51), negative attitudes towards others’ stockpiling behavior (10 self-developed items; ωTotal = 0.89), negative emotions towards stockpiling (seven items adapted from the German version of the Positive and Negative Affect Scale; Krohne, Egloff, Kohlmann, & Tausch, 1996; ωTotal = 0.77), acceptability of a policy that limits the purchase quantity to household standards (seven self-developed items; ωTotal = 0.87), trust in authorities during the pandemic (four self-developed items; ωTotal = 0.73), and factual knowledge about COVID-19 (10 dichotomous “true or false” items). These measures are not relevant for the present research question and are therefore not further discussed. All item wordings and the dataset are available on the OSF. Moreover, we provide means, standard deviations, and correlations between all variables assessed in the study in the Supplementary Online Materials (Table S1 and Table S2). 2.3 Procedure Participants first provided informed consent and demographic information. Next, they filled out the HEXACO-60 and the Justice Sensitivity Inventory.2 We randomized the order of the two respective survey pages. We then introduced our experimental manipulation of resource scarcity. Participants were asked to imagine going to the supermarket for the weekly shopping during the ongoing COVID-19 pandemic, being presented with the following passage (translated from German): “Imagine going to the supermarket to do your weekly shopping during the ongoing corona pandemic these days. You encounter the following situation: Most of the shelves are quite full [empty].” This text passage was accompanied by a picture of either filled shelves (low resource scarcity) or emptied shelves (high resource scarcity). We used five different pictures per condition presenting different goods such as groceries or sanitary products to prevent any systematic effects caused by a specific type of product. Each participant was presented with one randomly selected picture. The content of pictures (i.e., goods) was the same across conditions (i.e., the stimulus set contained one picture with a filled pasta shelf and one picture with an emptied pasta shelf etc.).3 Subsequently, we measured our dependent variables which were presented on the same page as the experimental manipulation. Importantly, instructions emphasized that participants should imagine that they would usually go to the supermarket once a week and that they would thus return again the week thereafter. This instruction was implemented to ensure that stockpiling can be validly interpreted as non-cooperative rather than reflecting a prosocial act (e.g., an attempt to go to the supermarket less often to reduce the risk of infecting others). Next, participants responded to two exploratory control items designed to assess the plausibility of the scenario (“The situation in the supermarket does not seem plausible to me during the current corona pandemic”, reverse-coded) and perceptions of resource scarcity in real-life (“When I go shopping these days (since the beginning of the corona pandemic), most supermarket shelves are pretty empty.”; both on a 6-point response scale ranging from 1 = strongly disagree to 6 = strongly agree). They also completed a manipulation check asking the extent to which the supermarket shelves depicted on the picture were 1 = very empty to 6 = very filled.4 On this survey page, we also asked whether participants had completed the survey attentively and whether they thought their responses were reliable using a so-called “use me” item. Finally, we raffled four vouchers each worth 25€ as incentive for participation and informed participants about the purpose of our study. 2.4 Participants We recruited participants online via a large mailing list, social media, and the PsyWeb panel (https://psyweb.uni-muenster.de). Given the very dynamic situation with regard to the spread of COVID-19, we decided to base our sampling strategy on pragmatic constraints and collected data from as many participants as possible within a pre-defined timeframe (i.e., until April 5, 2020). A total of 941 participants started the study of which 811 completed it. As pre-registered, we excluded participants who failed the instructed attention check that was embedded in the HEXACO-60 (n = 13). Further, one participant indicated on the “use me” item that their responses were not reliable and, thus, was excluded from the analyses.5 Our final sample consisted of N = 797 participants. The participants’ age ranged from 18 to 84 years (M = 35.83, SD = 15.99) and the sample was predominantly female (597 women, 197 men, 3 other). Participants in the final sample were almost equally distributed across conditions (i.e., n = 400 in the low resource scarcity condition and n = 397 in the high resource scarcity condition), suggesting that exclusions due to drop-out or non-attentive responding did not systematically vary across experimental conditions (see Zhou & Fishbach, 2016). Only 1% (n = 10) of participants were currently or had previously been infected with the virus, and 7% (n = 53) had a (currently or previously) infected close friend or family member. Moreover, 7% (n = 55) were quarantined and 25% (n = 198) had a close friend or family member who was quarantined during the time of participating in the study. To identify the effect sizes we were able to detect with our sample size, we conducted two sensitivity power analyses using G*Power (Faul, Erdfelder, Buchner, & Lang, 2009), both with a power of 1- β = 0.90 and a significance level of α = 0.05. First, we estimated the effect size we could uncover for one-sided t-tests (for investigating mean differences in our dependent variables between conditions). This yielded a very small effect size of d = 0.10. Second, we estimated the minimum effect size of a single regression coefficient (increase in R2) we were able to detect in a multiple regression including three predictors. This analysis suggested that our sample size allowed detecting small effects of f 2 = 0.01, corresponding to an R2 = 0.01 or a Pearson’s correlation of r = 0.10. We considered this effect size appropriate for our purposes because it is comparable to the size of meta-analytic estimates for the (uncorrected) correlations of Honesty-Humility and Agreeableness with cooperation in game-based commons dilemmas (r = 0.14 and r = 0.10, respectively; Thielmann, Spadaro et al., 2020). We did not find empirical benchmarks for the respective effects of Emotionality, Victim Sensitivity, resource scarcity, nor for the interaction effects. 3 Results Analyses were conducted in R (R Core Team, 2020), mainly using the packages tidyverse (Wickham et al., 2019), sjPlot (Lüdecke, 2018), MBESS (Kelley, 2017) and apaTables (Stanley, 2018). Correlations between the main study variables are displayed in Table 1. 3.1 Manipulation check We first checked the effectiveness of our manipulation by comparing the extent to which individuals rated the depicted shelves as empty or full between experimental conditions. Participants in the low resource scarcity condition rated the supermarket shelves as less empty (M = 5.12, SD = 1.11) than participants in the high resource scarcity condition (M = 1.77, SD = 0.91), yielding a large effect, d = 3.29, 95% CI [3.08, 3.51]. Levene’s test for homogeneity of variances indicated that variances of this item were significantly different between conditions, F(1, 795) = 8.05, p = .005. Consequently, we conducted Welch’s t-test (one-sided) to test for mean differences between conditions, and found a significant effect, t(767.98) = 46.54, p < .001. This suggests that our experimental manipulation was successful in the sense that resource scarcity was perceived as higher in the high as compared to the low resource scarcity condition. 3.2 Resource scarcity First, we investigated whether experimentally induced resource scarcity influenced participants’ stockpiling intentions and their perceptions of justifiability of stockpiling. Descriptively, stockpiling intentions were greater on average when resource scarcity was high (M = 2.70, SD = 1.32) rather than low (M = 2.52, SD = 1.33). In line with our hypothesis, a one-sided t-test yielded a significant result, t(795) = −1.93, p = .027, although the effect size was only small, d = −0.14, 95% CI [−0.28, 0.00].6 The descriptive patterns for justifiability of stockpiling behavior indicated that participants judged stockpiling as slightly less justifiable when resource scarcity was high (M = 2.33, SD = 1.03) rather than low (M = 2.39, SD = 1.10). This mean difference was non-significant in a one-sided t-test, t(795) = 0.76, p = .776, and indeed negligible in size, d = 0.05, 95% CI [−0.09, 0.19]. Thus, contrary to our hypothesis, resource scarcity did not affect participants’ justifiability of stockpiling behavior. 3.3 Personality We continued with investigating the effects of the personality variables under scrutiny (i.e., Honesty-Humility, Agreeableness, Emotionality, and Victim Sensitivity) on stockpiling intentions and justifiability of stockpiling behavior. To this end, we computed several multiple linear regressions, separately for each personality trait and for both DVs. Specifically, we entered one personality trait (mean centered), resource scarcity (effect-coded with −1 = low and 1 = high), and their interaction as predictors and either stockpiling intentions or justifiability of stockpiling behavior as criterion. Fig. 1 and Table 2 show the results for stockpiling intentions; Fig. 2 and Table 3 show the results for justifiability of stockpiling behavior.Fig. 1 Prediction of stockpiling intentions by personality under high and low resource scarcity. Personality dimensions were mean centered. Circles represent observations in the low resource scarcity condition, rectangles represent observations in the high resource scarcity condition. Table 2 Linear Regression Analyses for Stockpiling Intentions. Honesty-Humility Emotionality Agreeableness Victim Sensitivity Predictor B SE 95% CI p Predictor B SE 95% CI p Predictor B SE 95% CI p Predictor B SE 95% CI p LB UB LB UB LB UB LB UB Intercept 2.61 0.05 2.52 2.70 <0.001 Intercept 2.61 0.05 2.52 2.70 <0.001 Intercept 2.61 0.05 2.52 2.70 <0.001 Intercept 2.61 0.05 2.52 2.70 <0.001 RS 0.09 0.05 0.00 0.18 0.058 RS 0.10 0.05 0.01 0.19 0.034 RS 0.09 0.05 0.00 0.18 0.049 RS 0.09 0.05 0.00 0.18 0.047 HH −0.44 0.08 −0.59 −0.30 <0.001 EM 0.20 0.07 0.06 0.34 0.006 AG −0.11 0.09 −0.27 0.06 0.225 VS 0.20 0.05 0.10 0.29 <0.001 HH * RS −0.09 0.08 −0.24 0.06 0.237 EM * RS 0.10 0.07 −0.05 0.24 0.181 AG * RS −0.16 0.09 −0.33 0.01 0.060 VS * RS 0.01 0.05 −0.09 0.10 0.917 F 13.46 <0.001 F 4.68 0.003 F 2.85 0.037 F 6.44 <0.001 R2 0.05 0.02 0.08 R2 0.02 0.00 0.04 R2 0.01 0.00 0.02 R2 0.02 0.00 0.04 Note. RS = resource scarcity (effect-coded, −1 = low, and 1 = high); HH = Honesty-Humility; EM = Emotionality; AG = Agreeableness; VS = Victim Sensitivity; CI = confidence interval; LB = lower bound of CI; UB = upper bound of CI. HH, EM, AG, and VS were mean centered. Degrees of freedom are ν1 = 3 and ν2 = 793 for all four models. Fig. 2 Prediction of justifiability of stockpiling behavior by personality under high and low resource scarcity. Personality dimensions were mean centered. Circles represent observations in the low resource scarcity condition, rectangles represent observations in the high resource scarcity condition. Table 3 Linear Regression Analyses for Justifiability of Stockpiling Behavior. Honesty-Humility Emotionality Agreeableness Victim Sensitivity Predictor B SE 95% CI p Predictor B SE 95% CI p Predictor B SE 95% CI p Predictor B SE 95% CI p LB UB LB UB LB UB LB UB Intercept 2.36 0.04 2.29 2.44 <0.001 Intercept 2.36 0.04 2.29 2.44 <0.001 Intercept 2.36 0.04 2.29 2.44 <0.001 Intercept 2.36 0.04 2.29 2.44 <0.001 RS −0.03 0.04 −0.10 0.04 0.407 RS −0.03 0.04 −0.10 0.05 0.496 RS −0.03 0.04 −0.10 0.04 0.429 RS −0.03 0.04 −0.10 0.05 0.471 HH −0.29 0.06 −0.41 −0.17 <0.001 EM 0.07 0.06 −0.04 0.19 0.231 AG 0.07 0.07 −0.04 0.11 0.336 VS 0.15 0.04 0.07 0.23 <0.001 HH * RS −0.09 0.06 −0.21 0.03 0.134 EM * RS 0.06 0.06 −0.05 0.18 0.295 AG * RS −0.08 0.07 −0.12 0.03 0.251 VS * RS 0.01 0.04 −0.07 0.09 0.812 F 8.59 <0.001 F 1.13 0.338 F 0.98 0.400 F 5.13 0.002 R2 0.03 0.01 0.06 R2 0.00 0.00 0.01 R2 0.00 0.00 0.01 R2 0.02 0.00 0.04 Note. RS = resource scarcity (effect-coded, −1 = low, and 1 = high); HH = Honesty-Humility; EM = Emotionality; AG = Agreeableness; VS = Victim Sensitivity; CI = confidence interval; LB = lower bound of CI; UB = upper bound of CI. HH, EM, AG, and VS were mean centered. Degrees of freedom are ν1 = 3 and ν2 = 793 for all four models. For Honesty-Humility, we hypothesized negative main effects across conditions on stockpiling intentions as well as on justifiability of stockpiling behavior. Descriptively, zero-order correlations with stockpiling intentions (r = −0.21) and justifiability of stockpiling behavior (r = −0.17) were in line with our hypotheses. Correspondingly, Honesty-Humility was a significant negative predictor of stockpiling intentions, B = −0.44, p < .001, and justifiability of stockpiling behavior, B = −0.29, p < .001 in the regression analyses. The interaction effects of Honesty-Humility with resource scarcity were non-significant in both models, B = −0.09, p = .237, and B = −0.09, p = .134, respectively (see Table 2, Table 3). For Agreeableness, we predicted a positive relation to justifiability of stockpiling behavior and an interaction with resource scarcity. The zero-order correlation with justifiability of stockpiling behavior was close to zero, r = 0.04. Accordingly, and contrary to our predictions, we found neither a main effect of Agreeableness, B = 0.07, p = .336, nor an interaction with resource scarcity, B = −0.08, p = .251 in predicting justifiability of stockpiling behavior (see Table 3). Note that Agreeableness was also unrelated to stockpiling intentions, r = −0.04. For Emotionality, we hypothesized a positive main effect and an interaction with resource scarcity on stockpiling intentions. The direction of the respective zero-order correlation was indeed positive, albeit small, r = 0.10. Correspondingly, the regression analysis yielded a significant and positive main effect of Emotionality on stockpiling intentions, B = 0.20, p = .006, thereby confirming our first prediction. Contrary to expectations, however, the interaction with resource scarcity did not reach statistical significance, B = 0.10, p = .181 (see Table 2). Also, note that Emotionality was unrelated to justifiability of stockpiling behavior, r = 0.05. For Victim Sensitivity, we predicted an interaction with resource scarcity on stockpiling intentions in the sense that Victim Sensitivity should be positively related to stockpiling intentions under high resource scarcity, but less so under low resource scarcity. Across conditions, there was a positive zero-order correlation of Victim Sensitivity with stockpiling intentions, r = 0.14. Contrary to our hypothesis, there was no interaction with resource scarcity, B = 0.01, p = .917. Interestingly, the main effect of Victim Sensitivity was positive and significant, B = 0.20, p < .001 (see Table 2), indicating that individuals with high Victim Sensitivity expressed more intentions to stockpile in general across both levels of resource scarcity. Of note, Victim Sensitivity was also positively correlated with justifiability of stockpiling behavior, r = 0.14. We also aimed at testing whether the reported effects were robust when taking shared variances with other personality traits into account. For this reason, we entered all six HEXACO dimensions and all four Justice Sensitivity perspectives into the same regression model, separately for both DVs. We also included resource scarcity as a main effect, but no interaction terms with personality given that we did not find any significant interaction effects in the previous analyses. As summarized in Fig. 3 and consistent with our hypotheses, the multiple regression for stockpiling intentions revealed significant effects of resource scarcity, B = 0.11, SE(B) = 0.05, 95% CI [0.02, 0.20], p = .021, Honesty-Humility, B = -0.35, SE(B) = 0.10, 95% CI [−0.52, −0.17], p < .001, and Emotionality, B = 0.22, SE(B) = 0.08, 95% CI [0.06, 0.37], p = .006. Beyond predictions, we also found a significant effect for Beneficiary Sensitivity, B = −0.18, SE(B) = 0.06, 95% CI [−0.30, −0.07], p = .002. The effect of Victim Sensitivity was non-significant in this analysis, B = 0.10, SE(B) = 0.06, 95% CI [−0.01, 0.22], p = .076, which is not compatible with the analysis in which Victim Sensitivity was included as single personality predictor. None of the other personality traits were significantly related to stockpiling intentions in this analysis, B’s < 0.12, all p’s > 0.116. Overall, this regression model explained 7% of the variance in stockpiling intentions, F(11, 785) = 5.70, p < .001, R2 = 0.07, 95% CI [0.04, 0.11].Fig. 3 Multiple regression results for resource scarcity, HEXACO dimensions, and Justice Sensitivity perspectives predicting stockpiling intentions. Estimates are unstandardized regression coefficients. Resource scarcity is coded as −1 = low and 1 = high. The intercept is omitted in this figure. Personality dimensions were mean centered. Error bars represent 95% confidence intervals. * p < .05, ** p < .01, *** p < .001. Fig. 4 summarizes the results from the multiple regression for justifiability of stockpiling behavior. As predicted, there were significant effects of Honesty-Humility, B = −0.20, SE(B) = 0.07, 95% CI [−0.34, −0.05], p = .007, and Agreeableness, B = 0.19, SE(B) = 0.07, 95% CI [0.05, 0.33], p = .007. Of note, the significant effect of Agreeableness in this multiple regression is somewhat at odds with the regression results when Agreeableness was included as single personality predictor. We further delve into this finding in the exploratory analyses. Moreover, the regression model yielded significant effects of Victim Sensitivity, B = 0.14, SE(B) = 0.05, 95% CI [0.05, 0.23], p = .002, and Beneficiary Sensitivity, B = −0.16, SE(B) = 0.05, 95% CI [−0.25; −0.06], p = .001, both of which were not hypothesized. The effects of all other predictors – including resource scarcity – were non-significant, with all B’s < 0.10, all p’s > 0.118. Together, these predictors explained 6% of the variance in justifiability of stockpiling behavior, F(11, 785) = 4.84, p < .001, R2 = 0.06, 95% CI [0.03, 0.10].Fig. 4 Multiple regression results for resource scarcity, HEXACO dimensions, and Justice Sensitivity perspectives predicting justifiability of stockpiling behavior. Estimates are unstandardized regression coefficients. Resource scarcity is coded as −1 = low and 1 = high. The intercept is omitted in this figure. Personality dimensions were mean centered. Error bars represent 95% confidence intervals. * p < .05, ** p < .01. 3.4 Exploratory analyses We followed four lines of exploratory analyses: First, we elucidated why our manipulation of resource scarcity had only weak effects on the DVs. Second, we explored why Agreeableness was only related to justifiability of stockpiling behavior when controlling for the shared variance with other personality traits. Third, we investigated whether gender differences had an influence on stockpiling intentions and justifiability of stockpiling behavior, as well as controlled for gender in some of our analyses to account for the unequal gender distribution in our sample. Fourth, we examined the amount of variability in the DVs that was accounted for by different pictures within experimental conditions. One possible explanation for the small effects of the resource scarcity manipulation on the DVs is that individuals generally perceived supermarket shelves as relatively empty at the time the study was conducted. If this were true, asking participants to imagine encountering full supermarket shelves might have been perceived as unrealistic; instead, imagining a shopping scenario might have automatically elicited an imagination of resource scarcity. Indeed, inspection of the two control questions included in the survey supported this idea: Participants judged the situation in the low resource scarcity condition as less plausible (M = 4.68, SD = 1.44) than participants did in the high resource scarcity condition (M = 5.18, SD = 1.26). Levene’s test for homogeneity of variances suggested that these variances were heterogenous, F(1, 795) = 8.04, p = .005. We therefore applied Welch’s t-test (one-sided) and found this mean difference to be significant, t(781.41) = −5.18, p < 0.01, d = −0.37, 95% CI [−0.51, −0.23]. Correspondingly, and irrespective of condition, participants also indicated that supermarket shelves were relatively empty at the time (M = 3.78, SD = 1.53; on a response scale ranging from 1 to 6). These findings suggest that our manipulation of resource scarcity might have been too subtle to systematically affect the perception that others stockpile and thereby exploit those who cooperate. Interestingly, the perception that the supermarket shelves were relatively empty at the time of the survey was influenced by personality: Victim Sensitivity was positively associated with perceiving supermarket shelves as empty, r = 0.08, whereas Honesty-Humility was negatively associated with such perceptions, r = −0.12. This is compatible with evidence showing that victim sensitive individuals often expect malevolence (Gollwitzer et al., 2013), whereas individuals high in Honesty-Humility often expect benevolence in others (Thielmann and Hilbig, 2014, Thielmann, Hilbig, & Zettler, 2020). We next turned to the effect of Agreeableness on justifiability of stockpiling behavior. Recall that we hypothesized that Agreeableness should be positively related to justifiability of stockpiling behavior (i.e., a main effect) and that it should interact with resource scarcity in the sense that the positive association of Agreeableness should be stronger under high (versus low) resource scarcity. The linear regression analyses including only Agreeableness as trait predictor supported neither of these hypotheses (see above and Table 2). By contrast, Agreeableness did positively predict justifiability of stockpiling when the shared variance with other personality dimensions was statistically controlled for (see Fig. 4). A potential explanation for this inconsistency is the relatively high correlation between Agreeableness and Honesty-Humility in our data (r = 0.20; see also Moshagen, Thielmann, Hilbig, & Zettler, 2019, for corresponding meta-analytic estimates). A recent meta-analysis demonstrated that controlling for the shared variance of these two dimensions in particular allows a more unique mapping of Agreeableness and Honesty-Humility on theoretically-related outcomes (Zettler, Thielmann, Spadaro, & Balliet, 2020). Following this reasoning, we extended the pre-registered regression model including Agreeableness, resource scarcity, and their interaction as predictors of justifiability of stockpiling by a main effect of Honesty-Humility. In this model, the effect of Agreeableness was indeed positive and significant, B = 0.14, SE(B) = 0.07, 95% CI [>0.00, 0.28], p = .048. The interaction between Agreeableness and resource scarcity remained non-significant, B = −0.07, SE(B) = 0.07, 95% CI [−0.20, 0.07], p = .334. Taken together, higher levels of Agreeableness were associated with higher justifiability of stockpiling behavior when controlling for the shared variance with Honesty-Humility, but there was no evidence for an interaction with resource scarcity whatsoever. Our sample consisted of substantially more women (n = 597) than men (n = 197). Interestingly, females descriptively reported lower levels of stockpiling intentions than men in both the low resource scarcity (Mfemale = 2.47, SDfemale = 1.29 versus Mmale = 2.70, SDmale = 1.44) and the high resource scarcity condition (Mfemale = 2.64, SDfemale = 1.31 versus Mmale = 2.86, SDmale = 1.35) but both mean differences were non-significant in two-sided t-tests, t(395) = −1.45, p = .148, d = −0.17, 95% CI [−0.41, 0.06] and t(395) = −1.53, p = .128, d = −0.17, 95% CI [−0.40, 0.05], respectively. Similarly, females judged stockpiling as descriptively less justifiable than men in both the low resource scarcity (Mfemale = 2.34, SDfemale = 1.09 versus Mmale = 2.57, SDmale = 1.33) as well as in the high resource scarcity condition (Mfemale = 2.27, SDfemale = 0.96 versus Mmale = 2.51, SDmale = 1.19) but these trends were again non-significant in two-sided tests, t(395) = −1.76, p = .080, d = −0.21, 95% CI [−0.45, 0.03] and Welch’s t(159.67) = −1.85, p = .067, d = −0.23, 95% CI [−0.45, −0.01], respectively.7 Nevertheless, these descriptive trends (but not the non-significant inferential results) are in line with meta-analytic evidence showing that women tend to cooperate more than men in resource dilemmas (Balliet, Li, Macfarlan, & Van Vugt, 2011). Due to these descriptive gender differences, we also considered it important to replicate our main confirmatory findings when statistically controlling for gender. More specifically, we extended the linear regressions models with single personality predictors (i.e., the analyses reported in Table 2 and Table 3) by a main effect of gender. The results for these models are reported in the Supplementary Online Materials (Table S3 and Table S4). In essence, being male was associated with greater stockpiling intentions and greater justifiability of stockpiling behavior in all models, except for the model in which stockpiling intentions were regressed on Honesty-Humility, B = 0.20, p = .069. All main effects of personality on both DVs replicated the results from the confirmatory analyses, with the exception that Emotionality now also positively predicted justifiability of stockpiling behavior, B = 0.14, p = .027. Again, there was no evidence for significant interactions between the personality variables and resource scarcity. Lastly, we tested whether variance in our DVs caused by different stimuli (i.e., 10 different pictures of empty vs. full supermarket shelves) within conditions affected the results. Specifically, in our main analyses we did not consider the different stimuli used to manipulate resource scarcity. To address this issue, we extended the pre-registered regression analyses to linear mixed models by adding random intercepts on the stimulus level (lme4 package; Bates, Mächler, Bolker, & Walker, 2015). The intraclass correlations (ICCs) for the stockpiling intentions models were consistently small (all ICCs < 0.01). For justifiability of stockpiling, all ICCs were estimated to be zero, reflecting a singular fit of these models. Together, this indicates that – if at all – only negligible amounts of variance in the DVs can be attributed to the different stimuli used. Consequently, the multiple regression analyses neglecting differences between stimuli are appropriate for the present data structure. 4 Discussion The present study tested whether well-known predictors of cooperation in game-based social dilemmas can likewise account for stockpiling during the COVID-19 pandemic – a real-life social dilemma. We recruited a large German sample (N = 797) during a time when stockpiling occurred frequently and tested the impact of the HEXACO dimensions and Victim Sensitivity on stockpiling intentions and justifiability of stockpiling behavior under high and low resource scarcity. Hypotheses were based on an affordance perspective (Thielmann, Spadaro et al., 2020) allowing to derive specific hypotheses about which traits should be relevant in the stockpiling situation. In line with our predictions, Honesty-Humility yielded small to medium-sized (Funder & Ozer, 2019) negative relations with stockpiling intentions and justifiability of stockpiling behavior.8 This closely replicates recent findings by Columbus (2020) who also found Honesty-Humility to negatively predict stockpiling in the past and intentions to do so in the future in a UK sample. By contrast, Garbe et al. (2020) found no evidence for the role of Honesty-Humility for multiple indicators of toilet paper stockpiling during COVID-19 in a sample of participants from 22 European and North American countries. A potential explanation for this inconsistency – besides differences between samples – is that Garbe and colleagues only used a 24-item short scale to measure the HEXACO dimensions (De Vries, 2013) whereas both Columbus and we used longer scales. Then again, Zettler, Schild, et al. (2020) used both short and longer scales to assess Honesty-Humility across five samples in Denmark and Germany, showing no systematic differences in relations with self-reported hoarding behavior as a function of scale length. Thus, we suspect that the described discrepancies between findings may also simply be due to sampling error or other systematic differences in samples and methods between studies. In any case, the finding that Honesty-Humility negatively predicts stockpiling is compatible with its theoretical conceptualization (Ashton & Lee, 2007) and prior research on the link between Honesty-Humility and cooperation in game-based social dilemmas (Thielmann, Spadaro et al., 2020). Agreeableness, in turn, was positively associated with justifiability of stockpiling, yielding a small effect size but (unexpectedly) only when controlling for the shared variance with other trait predictors – most prominently Honesty-Humility. Nonetheless, the finding that the unique effect of Agreeableness becomes apparent once accounting for the shared variance with Honesty-Humility is in line with meta-analytic evidence on the unique predictive ability of Agreeableness for theoretically-relevant outcomes (Zettler, Thielmann et al., 2020). That said, these findings should be replicated in the future, as the corresponding analyses were only exploratory in nature. Moreover, it should be noted that Agreeableness did not interact with resource scarcity to predict justifiability of stockpiling behavior, as would have been expected based on the theoretical conceptualization of Agreeableness reflecting forgivingness versus retaliation. Nonetheless, the absence of this effect should also be interpreted with caution given the limited effectiveness of our resource scarcity manipulation. Further, our results showed a small positive relation between Emotionality and stockpiling intentions. However, we found no evidence for an interaction between Emotionality and resource scarcity in predicting stockpiling intentions. The finding that Emotionality plays a role for cooperation in the stockpiling dilemma is particularly noteworthy because Emotionality is usually unrelated to cooperation in game-based social dilemmas (Thielmann, Spadaro et al., 2020). As reasoned above, situations like a pandemic may elicit feelings of stress, anxiety, and uncertainty and thus afford the expression of Emotionality in uncooperative behavior such as stockpiling. More precisely, individuals high in Emotionality may use stockpiling as a means to reduce fear and stress, which is compatible with the positive associations of Emotionality with anxiety-related outcomes (Zettler, Thielmann et al., 2020). As such, this finding corroborates the notion that emotions may indeed “create social or moral forces that game matrices have not captured” (Murnighan & Wang, 2016, p. 91). Future research is needed to examine whether Emotionality is only relevant in certain social dilemmas in naturalistic settings, for example by directly manipulating whether or not (non-)cooperation can reduce feelings of anxiety. Building on the affordance-based reasoning above, it is conceivable that Emotionality is unrelated to cooperation in less emotionally-charged settings, but negatively related to cooperation when the outcome helps to cope with anxiety or uncertainty. Finally, Victim Sensitivity was positively associated with stockpiling intentions. However, in contrast to our hypotheses, this association was apparent irrespective of whether resource scarcity was high or low. The theoretical rationale for the hypothesized interaction was based on the SeMi model (Gollwitzer et al., 2013) which proposes that “victim-sensitive individuals react more sensitively to cues of untrustworthiness; thus, the presence of such cues reduces their willingness to cooperate in social dilemmas” (p. 418). However, the operationalization of untrustworthiness cues in form of our resource scarcity manipulation may have been ineffective in eliciting meaningful differences between conditions. The pandemic was a “strong situation” in itself and was likely to suffice for eliciting a suspicious mindset (i.e., expectations that others stockpile) among many individuals (i.e., both people high and low on Victim Sensitivity). This could explain why Victim Sensitivity was positively linked to stockpiling intentions in general, irrespective of resource scarcity. 4.1 Limitations and directions for future research Although our results provide important insights into the personality determinants of stockpiling during a pandemic, some limitations of the present research ought to be acknowledged. First, the experimental manipulation of resource scarcity only had a small effect on stockpiling intentions and no effect at all on justifiability of stockpiling behavior. Moreover, none of the focal personality traits interacted with resource scarcity to predict the DVs, suggesting that this manipulation was too subtle and thus ineffective in triggering differential effects of personality in the two experimental conditions. Moreover, we argued that empty supermarket shelves may indicate others’ engagement in stockpiling. However, this is not the only plausible interpretation. Empty supermarket shelves may also reflect environmental influences, such as interruptions in delivery of certain products. Arguably, the attribution of resource scarcity as caused by other people versus environmental influences may also impact individuals’ willingness to cooperate. Evidence suggests that individuals tend to harvest more from a scarce resource in the commons dilemma when scarcity is caused by group members rather than environmental influences (Rutte, Wilke, & Messick, 1987). However, given that we did not measure participants’ construal of the situation, we cannot empirically control for potential differences in this attribution. Future research addressing this issue is desired. The theoretical rationale of the present study is based on the notion that stockpiling during a pandemic represents a social dilemma that has conceptual overlap with the commons dilemma. However, whether individuals indeed perceived the situation in a structurally similar way as a (game-based) commons dilemma is an open question. In this regard, it would be informative to investigate whether participants were aware that stockpiling may reduce the availability of certain products for other customers. Future research is needed to examine when and how real-life situations are perceived and interpreted in terms of social dilemmas as modelled in the lab (Columbus, Molho, Righetti, & Balliet, 2020), and to what extent such perceptions may affect the influence of personality on behavior (Columbus, Thielmann, & Balliet, 2019). Moreover, the situational affordances we ascribed to the situation of stockpiling were based on theoretical considerations and were not empirically assessed. For example, we argued that stockpiling represents an opportunity for exploitation in the sense that customers can increase their individual outcome through stockpiling without fearing punishment from other customers or authorities – because no formal sanctioning systems for stockpiling behavior were present. Stockpiling is, however, publicly observable, and thus some customers may have anticipated informal sanctions for stockpiling, for example in form of public criticisms by other customers. Whether or not participants anticipated sanctions for non-cooperation (i.e., stockpiling) is relevant because previous research suggests that the effects of certain personality traits on cooperation in social dilemmas are more pronounced in the absence (vs. presence) of punishment (Hilbig, Zettler, & Heydasch, 2012). Another limitation is the correlational nature of the study, with the exception of our resource scarcity manipulation. Thus, the study does not allow any causal claims regarding the link between personality and stockpiling. Moreover, our DVs were self-reports and therefore potentially prone to socially desirable responding. Observing actual stockpiling behavior (rather than measuring stockpiling intentions with self-report scales) would have provided a stronger test of our hypotheses. Lastly, our results are specific to real-life social dilemmas during a pandemic and may thus not directly transfer to other real-life social dilemmas. More specifically, we expect our results to generalize to situations in which (i) others’ stockpiling of basic goods such as groceries or sanitary products is visible (e.g., through media reports or empty supermarket shelves) and (ii) stockpiling is to some extent driven by expecting that these goods may be unavailable in the near future. We have no reason to believe that the results depend on other characteristics of the participants, materials, or context (cf. Simons, Shoda, & Lindsay, 2017). 5 Conclusion Most previous research investigated individual and situational antecedents of cooperation in social dilemmas using economic games. The present study demonstrates that particularly the association of Honesty-Humility with cooperation generalizes to the real-life social dilemma of stockpiling, thereby speaking to the robustness of this effect and the validity of economic games as a research paradigm. By contrast, our study also reveals that economic games may underestimate or neglect the importance of other personality traits that become particularly relevant in the presence of specific situational affordances, such as the impact of Emotionality for stockpiling during a pandemic. We hope that our research encourages future work along these lines. 6 Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. 7 Open practices Study materials, data set, and analysis scripts can be accessed via the following link: https://osf.io/wbxnr/. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix A Supplementary material The following are the Supplementary data to this article:Supplementary data 1 Acknowledgements The authors are grateful to the panel PsyWeb (https://psyweb.uni-muenster.de/) for their immense support with the recruitment of participants. 1 Our hypotheses, methods, and analytic plan were pre-registered (https://aspredicted.org/gu4bn.pdf). The situation regarding the spread of COVID-19 and the resulting stockpiling tendencies changed rapidly around the time of data collection. We therefore wanted to start data collection as quickly as possible and thus, launched the survey before we had completed the pre-registration. More specifically, we started data collection on 23rd of March, completed the pre-registration on 26th of March, and the study was online until 5th of April 2020. We did not inspect nor analyze any data before data collection was completed. We obtained ethics approval for an umbrella project on the psychological effects of the COVID-19 pandemic from the local ethics committee at the University of Koblenz-Landau’s psychology department. 2 At this stage, we also assessed the Need for Cognition scale which was presented on the same survey page right after the Justice Sensitivity Inventory. 3 Unfortunately, the license for these picture does not permit sharing the original files. More information is available from the first author upon request. 4 This item was included to assess the effectiveness of our manipulation and not to screen for inattentive responding. Thus, no participants were excluded based on this item (see “Participants” section and pre-registration for our exclusion criteria). 5 As preregistered, we also screened the data set for non-serious participation but did not find evidence of “clicking through”. Thus, no additional participants were excluded based on this screening. 6 We used one-sided t-tests because our pre-registered hypotheses were that high (vs. low) resource scarcity would lead to more stockpiling intentions and stronger justifiability of stockpiling behavior (i.e., directed hypotheses). We therefore consider one-sided tests appropriate. Note, however, that a two-sided test for the effect of resource scarcity on stockpiling intentions would have been non-significant (p = .054). 7 We used Welch’s t-test to compare means in justifiability of stockpiling behavior between female and male participants in the high resource scarcity condition because Levene’s test indicted that these variances were heterogenous, F(1,395) = 5.93, p = .015. Note, however, that this mean difference would be significant when assuming equal variances instead, t(395) = −2.04, p = .042, and that this explains why the 95% confidence interval for Cohen’s d does not include zero despite Welch’s t-test being non-significant. 8 The evaluations of effect sizes in this section are based on the zero-order correlations (see Table 1). We followed the recommendation of Funder & Ozer (2019) to interpret r ≥ |0.05| as a very small, r ≥ |0.10| as a small, r ≥ |0.20| as a medium, and r ≥ |0.30| as a large effect in psychological research. Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.jrp.2021.104075. ==== Refs References Aschwanden D. Strickhouser J.E. Sesker A.A. Lee J.H. Luchetti M. Stephan Y. …Terracciano A. Psychological and behavioural responses to coronavirus disease 2019: The role of personality European Journal of Personality 35 1 2021 51 66 10.1002/per.2281 Ashton M.C. Lee K. Empirical, theoretical, and practical advantages of the HEXACO model of personality structure Personality and Social Psychology Review 11 2 2007 150 166 10.1177/1088868306294907 18453460 Ashton M.C. Lee K. The HEXACO–60: A short measure of the major dimensions of personality Journal of Personality Assessment 91 4 2009 340 345 10.1080/00223890902935878 20017063 Ashton M.C. Lee K. De Vries R.E. The HEXACO honesty-humility, agreeableness, and emotionality factors: A review of research and theory Personality and Social Psychology Review 18 2 2014 139 152 10.1177/1088868314523838 24577101 Balliet D. Li N.P. Macfarlan S.J. Van Vugt M. Sex differences in cooperation: A meta-analytic review of social dilemmas Psychological Bulletin 137 6 2011 881 909 10.1037/a0025354 21910518 Balliet D. Van Lange P.A.M. Trust, conflict, and cooperation: A meta-analysis Psychological Bulletin 139 5 2013 1090 1112 10.1037/a0030939 23231532 Bates D. Mächler M. Bolker B. Walker S. Fitting linear mixed-effects models using lme4 Journal of Statistical Software 67 1 2015 1 48 10.18637/jss.v067.i01 Baumert A. Beierlein C. Schmitt M. Kemper C.J. Kovaleva A. Liebig S. Rammstedt B. Measuring four perspectives of justice sensitivity with two items each Journal of Personality Assessment 96 3 2014 380 390 10.1080/00223891.2013.836526 24066854 Baumert A. Maltese S. Reis D. MacLeod C. Tan-Mansukhani R. Galang A.J.R. …Schmitt M. A cross-cultural study of justice sensitivity and its consequences for cooperation Social Psychological and Personality Science 11 7 2020 899 907 10.1177/1948550619896895 Beißert H. Köhler M. Rempel M. Beierlein C. Eine deutschsprachige Kurzskala zur Messung des Konstrukts Need for Cognition [A German-language short scale for measuring the construct Need for Cognition] GESIS-Working Papers 32 2015 Columbus, S. (2020). Honesty-Humility, beliefs, and prosocial behaviour: A test on stockpiling during the COVID-19 pandemic. Retrieved on January 19, 2021, from 10.31234/osf.io/8e62v. Columbus, S., Molho, C., Righetti, F., & Balliet, D. (2020). Interdependence and cooperation in daily life. Journal of Personality and Social Psychology. Advance online publication. 10.1037/pspi0000253. Columbus S. Thielmann I. Balliet D. Situational affordances for prosocial behaviour: On the interaction between Honesty-Humility and (perceived) interdependence European Journal of Personality 33 6 2019 655 673 10.1002/per.2224 Dammeyer J. An explorative study of the individual differences associated with consumer stockpiling during the early stages of the 2020 Coronavirus outbreak in Europe Personality and Individual Differences 2020 10.1016/j.paid.2020.110263 De Vries R.E. The 24-item brief HEXACO inventory (BHI) Journal of Research in Personality 47 6 2013 871 880 10.1016/j.jrp.2013.09.003 De Vries R.E. Tybur J.M. Pollet T.V. van Vugt M. Evolution, situational affordances, and the HEXACO model of personality Evolution and Human Behavior 37 5 2016 407 421 10.1016/j.evolhumbehav.2016.04.001 Diemand S. Jansen J. Müssgens C. Piller T. Preuss S. Hamstern was das Zeug hält [Hoarding as much as possbile] Frankfurter Allgemeine Zeitung 2020 Dunn T.J. Baguley T. Brunsden V. From alpha to omega: A practical solution to the pervasive problem of internal consistency estimation British Journal of Psychology 105 3 2014 399 412 10.1111/bjop.12046 24844115 Eisinga R. Te Grotenhuis M. Pelzer B. The reliability of a two-item scale: Pearson, Cronbach, or Spearman-Brown? International Journal of Public Health 58 4 2013 637 642 10.1007/s00038-012-0416-3 23089674 Faul F. Erdfelder E. Buchner A. Lang A.-G. Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses Behavior Research Methods 41 2009 1149 1160 10.3758/brm.41.4.1149 19897823 Fetchenhauer D. Huang X. Justice sensitivity and distributive decisions in experimental games Personality and Individual Differences 36 5 2004 1015 1029 10.1016/s0191-8869(03)00197-1 Funder D.C. Ozer D.J. Evaluating effect size in psychological research: Sense and nonsense Advances in Methods and Practices in Psychological Science 2 2 2019 156 168 10.1177/2515245919847202 Garbe L. Rau R. Toppe T. Influence of perceived threat of Covid-19 and HEXACO personality traits on toilet paper stockpiling PLoS ONE 15 6 2020 e0234232 10.1371/journal.pone.0234232 Gollwitzer M. Rothmund T. When the need to trust results in unethical behavior: The sensitivity to mean intentions (SeMI) model De Cremer D. Psychological perspectives on ethical behavior and decision making 2009 Information Age Charlotte, NC 135 152 Gollwitzer M. Rothmund T. Süssenbach P. The Sensitivity to Mean Intentions (SeMI) model: Basic assumptions, recent findings, and potential avenues for future research Social and Personality Psychology Compass 7 7 2013 415 426 10.1111/spc3.12041 Gollwitzer M. Schmitt M. Schalke R. Maes J. Baer A. Asymmetrical effects of justice sensitivity perspectives on prosocial and antisocial behavior Social Justice Research 18 2 2005 183 201 10.1007/s11211-005-7368-1 Hardin G. The tragedy of the commons Science 162 3859 1968 1243 1248 10.1126/science.162.3859.1243 Hilbig B.E. Kieslich P.J. Henninger F. Thielmann I. Zettler I. Lead us (not) into temptation: Testing the motivational mechanisms linking Honesty-Humility to cooperation European Journal of Personality 32 2 2018 116 127 10.1002/per.2149 Hilbig B.E. Thielmann I. Klein S.A. Henninger F. The two faces of cooperation: On the unique role of HEXACO Agreeableness for forgiveness versus retaliation Journal of Research in Personality 64 2016 69 78 10.1016/j.jrp.2016.08.004 Hilbig B.E. Zettler I. Heydasch T. Personality, punishment and public goods: Strategic shifts towards cooperation as a matter of dispositional honesty–humility European Journal of Personality 26 3 2012 245 254 10.1002/per.830 Hilbig B.E. Zettler I. Leist F. Heydasch T. It takes two: Honesty-Humility and Agreeableness differentially predict active versus reactive cooperation Personality and Individual Differences 54 5 2013 598 603 10.1016/j.paid.2012.11.008 Kelley, K. (2017). MBESS (Version 4.0.0 and higher) [computer software and manual]. Retrieved from http://cran.r-project.org. Krohne H.W. Egloff B. Kohlmann C.W. Tausch A. Untersuchungen mit einer deutschen Version der “Positive and negative Affect Schedule” (PANAS) [Investigations with a German version of the ‘‘Positive and negative Affect Schedule’’ (PANAS)] Diagnostica 42 1996 139 156 Lüdecke, D. (2018). sjPlot: Data visualization for statistics in social science. Retrieved from https://CRAN.R-project.org/package=sjPlot. Moshagen M. Hilbig B.E. Zettler I. Faktorenstruktur, psychometrische Eigenschaften und Messinvarianz der deutschsprachigen Version des 60-item HEXACO Persönlichkeitsinventars [Factor structure, psychometric properties, and measurement invariance of the German-language version of the 60-item HEXACO personality inventory] Diagnostica 60 2 2014 86 97 10.1026/0012-1924/a000112 Moshagen M. Thielmann I. Hilbig B.E. Zettler I. Meta-analytic investigations of the HEXACO Personality Inventory(-Revised): Reliability generalization, self-observer agreement, intercorrelations, and relations to demographic variables Zeitschrift für Psychologie 227 2019 186 194 10.1027/2151-2604/a000377 Murnighan J.K. Wang L. The social world as an experimental game Organizational Behavior and Human Decision Processes 136 2016 80 94 10.1016/j.obhdp.2016.02.003 Pletzer J.L. Balliet D. Joireman J. Kuhlman D.M. Voelpel S.C. Van Lange P.A. Social value orientation, expectations, and cooperation in social dilemmas: A meta-analysis European Journal of Personality 32 1 2018 62 83 10.1002/per.2139 R Core Team R: A language and environment for statistical computing 2020 R Foundation for Statistical Computing Vienna, Austria Retrieved from https://www.R-project.org/ Rutte C.G. Wilke H.A. Messick D.M. Scarcity or abundance caused by people or the environment as determinants of behavior in the resource dilemma Journal of Experimental Social Psychology 23 3 1987 208 216 10.1016/0022-1031(87)90032-1 Schmitt M. Gollwitzer M. Maes J. Arbach D. Justice sensitivity: Assessment and location in the personality space European Journal of Psychological Assessment 21 3 2005 202 211 Schmitt M. Neumann R. Montada L. Dispositional sensitivity to befallen injustice Social Justice Research 8 4 1995 385 407 10.1007/bf02334713 Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2012). A 21 word solution. Retrieved from: 10.2139/ssrn.2160588. Simons D.J. Shoda Y. Lindsay D.S. Constraints on generality (COG): A proposed addition to all empirical papers Perspectives on Psychological Science 12 6 2017 1123 1128 10.1177/1745691617708630 28853993 Stanley, D. (2018). apaTables: Create American Psychological Association (APA) style tables. Retrieved from https://CRAN.R-project.org/package=apaTables. Statistisches Bundesamt (2020). Corona-Krise: Experimentelle Daten zeigen aktuelles Kaufverhalten [Corona crisis: Experimental data show current buying behavior]. Retrieved on 23 May from https://www.destatis.de/DE/Presse/Pressemitteilungen/2020/05/PD20_178_61.html;jsessionid=1ECD957F54DDDE696E8C7D1256534DA2.internet8711. Thielmann I. Hilbig B.E. Trust in me, trust in you: A social projection account of the link between personality, cooperativeness, and trustworthiness expectations Journal of Research in Personality 50 2014 61 65 10.1016/j.jrp.2014.03.006 Thielmann I. Hilbig B.E. Zettler I. Seeing me, seeing you: Testing competing accounts of assumed similarity in personality judgments Journal of Personality and Social Psychology 118 1 2020 172 198 10.1037/pspp0000222 30321044 Thielmann I. Spadaro G. Balliet D. Personality and prosocial behavior: A theoretical framework and meta-analysis Psychological Bulletin 146 1 2020 30 90 10.1037/bul0000217 31841013 Van Lange P.A.M. Joireman J.A. Parks C.D. Van Dijk E. The psychology of social dilemmas: A review Organizational Behavior and Human Decision Processes 120 2 2013 125 141 10.1016/j.obhdp.2012.11.003 Wickham H. Averick M. Bryan J. Chang W. McGowan L.D. Francois R. …Yutani H. Welcome to the tidyverse Journal of Open Source Software 4 43 2019 1686 10.21105/joss.01686 Zettler I. Schild C. Lilleholt L. Kroencke L. Utesch T. Moshagen M. …Geukes K. The role of personality in COVID-19 related perceptions, evaluations, and behaviors: Findings across five samples, nine traits, and 17 criteria PsyArXiv 2020 10.31234/osf.io/pkm2a Zettler I. Thielmann I. Hilbig B.E. Moshagen M. The nomological net of the HEXACO model of personality: A large-scale meta-analytic investigation Perspectives on Psychological Science 15 3 2020 723 760 10.1177/1745691619895036 32324493 Zhou H. Fishbach A. The pitfall of experimenting on the web: How unattended selective attrition leads to surprising (yet false) research conclusions Journal of Personality and Social Psychology 111 4 2016 493 504 10.1037/pspa0000056 27295328

0

PMC9753122

NO-CC CODE

2022-12-16 23:25:30

no

J Res Pers. 2021 Apr 10; 91:104075

utf-8

J Res Pers

10.1016/j.jrp.2021.104075

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier S0140-6736(21)01851-1 10.1016/S0140-6736(21)01851-1 World Report Haiti's health woes intensify Daniels Joe Parkin 12 8 2021 14-20 August 2021 12 8 2021 398 10300 567567 . 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcAssassination, after months of violence, fuel shortages, and protests, has led to instability in Haiti, causing interruptions to normal care and the COVID-19 response. Joe Parkin Daniels reports. When Haiti's president, Jovenel Moïse, was assassinated in his home outside the capital, Port-au-Prince, in the early hours of July 7, 2021, the country was already in crisis. Gang violence, fuel shortages, and street protests had been intensifying for more than a year in much of Haiti, the western hemisphere's poorest nation. Not a single dose of any vaccine against COVID-19 had been administered when allegedly Colombian mercenaries rained down gunfire on Moïse. Natural disasters and political turmoil regularly rattle Haiti, and medical workers say that the current instability has only exacerbated long-existing health challenges. Widespread gang violence, often politically motivated, continues to roil Port-au-Prince and other cities, with sporadic roadblocks of rubble and burning tyres, complicating the movement of health professionals and patients alike. National programmes to prevent and treat HIV/AIDS, tuberculosis, and malaria have been interrupted, and ambulances are often stuck in blockades. Non-governmental organisations (NGOs)—which often make up the shortfall of the underfunded health ministry—are struggling to operate. Doctors Without Borders (MSF) announced on Aug 2 that it had permanently closed its clinic in Martissant neighbourhood in the capital amid shootouts between rival gangs. The hospital offered free medical care to 300 000 people, and had been hit by stray bullets in early June. “MSF continues to call on armed actors in Haiti to respect the safety of health personnel, patients, equipment and medical facilities”, the NGO said in a statement following the announcement of the hospital closure. “Vehicles and ambulances must also be able to circulate safely”. Another NGO, Partners In Health—which operates primarily across Haiti's remote and impoverished Central Plateau and Artibonite regions—also expressed concern about how freely its workers can circulate. “We are very worried about the situation”, Kenia Vissieres, a physician working with Partners In Health, told The Lancet. Vissieres added that the security crisis has compounded the problems caused by Haiti's crumbling or non-existent health infrastructure. “The general hospital, which is Port-au-Prince's primary health-care facility, is still under construction, and rural centres are not fit to receive many patients. Facilities across the country must be renovated as they are the basis of a strong national health-care system.” GHESKIO—another health centre near Martissant originally founded to tackle Haiti's HIV/AIDS epidemic—remains open but has seen daily patient visits decrease from 2134 in 2019, to 1535 today, amid fears of being caught in escalating violence, according to workers there. Health professionals now carry out visits into nearby slums to administer health care rather than have patients make risky journeys. On July 14, 1 week after Moïse's assassination, Haiti received its first batch of 500 000 COVID-19 vaccines from the USA via the COVAX programme. Although COVID-19 has troubled Haiti, where sanitation and infection prevention measures are difficult to implement, fewer than 600 deaths have been reported. This situation could change, however, as new strains reach the country. Since May, a third wave driven by the Gamma variant has caused more deaths than the previous two waves combined. The vaccination campaign continues, although it faces acute challenges, including safe transport within the country, keeping vaccines refrigerated amid rolling power outages, as well as convincing people to get inoculated. Preliminary results of a UNICEF-supported study carried out in June by the University of Haiti found that only 22% of Haitians would get vaccinated. Jean William Pape, director of GHESKIO and professor of medicine at Weill Cornell Medical College, said that one reason for vaccine hesitancy is that many vulnerable people do not recognise the severity of COVID-19. “For AIDS and cholera epidemics, it was different; they saw people dying around them—their parents and friends—but this is not the case for the COVID-19 pandemic”, Pape said. “In a recent survey carried out by Safitek, 85% of people did not know anyone who had COVID-19”. Health workers are at least able to celebrate the near eradication of cholera. According to the country's Ministry of Public Health and Population, 819 000 people showed symptoms of the disease from October, 2010, to December, 2019, initially brought to the country by aid workers following the catastrophic earthquake in January, 2010. “We’re happy to report that with the addition of contact tracing and interventions, and WASH [water, sanitation, and hygiene] we have had not a single cholera case in 2 years”, Pape said.

34391488

PMC9753123

NO-CC CODE

2022-12-16 23:25:30

no

Lancet. 2021 Aug 12 14-20 August; 398(10300):567

utf-8

Lancet

10.1016/S0140-6736(21)01851-1

oa_other

==== Front Appl Geogr Appl Geogr Applied Geography (Sevenoaks, England) 0143-6228 0143-6228 Elsevier Ltd. S0143-6228(21)00120-X 10.1016/j.apgeog.2021.102504 102504 Article Abrupt changes, institutional reactions, and adaptive behaviors: An exploratory study of COVID-19 and related events' impacts on Hong Kong's metro riders Zhou Jiangping ∗ Wu Jiangyue Ma Hanxi Department of Urban Planning and Design, Faculty of Architecture, The University of Hong Kong, Pok Fu Lam, Hong Kong SAR, China ∗ Corresponding author. 10 7 2021 9 2021 10 7 2021 134 102504102504 18 12 2020 11 6 2021 28 6 2021 © 2021 Elsevier Ltd. All rights reserved. 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Abrupt socioeconomic changes have become increasingly commonplace. In face of these, both institutions and individuals must adapt. Against the backdrop of the COVID-19 pandemic, suddenness, scale, and impacts of which are unprecedented as compared to its counterparts in history, we first propose transferable measures and methods that can be used to quantify and geovisualize COVID-19 and subsequent events' impacts on metro riders' travel behaviors. Then we operationalize and implement those measures and methods with empirical data from Hong Kong, a metropolis heavily reliant on transit/metro services. We map out where those impacts were the largest and explores its correlates. We exploit the best publicly available data to assemble probable explanatory variables and to examine quantitatively whether those variables are correlated to the impacts and if so, to what degree. We find that both macro- and meso-level external/internal events following the COVID-19 outbreak significantly influenced of metro riders' behaviors. The numbers of public rental housing residents, public and medical facilities, students' school locations, residents’ occupation, and household income significantly predict the impacts. Also, the impacts differ across social groups and locales with different built-environment attributes. This means that to effectively manage those impacts, locale- and group-sensitive interventions are warranted. Keywords COVID-19 Adaptive travel behaviors Measurement Geovisualization Predictor ==== Body pmc1 Introduction In past few decades, abrupt changes caused by economic and financial crises, social protests, terrorist attacks, and outbreak of pandemics/epidemics have become increasingly commonplace across countries and regions. Taking Hong Kong as an example, it has endured and/or is still enduring at least the following abrupt and shocking events and/or their derivatives: the 2003 SARS outbreak, the 2007/08 Asian Financial Crisis, the 2007 to 2010 Global Financial Crisis, the 2014 Umbrella Movement, the Anti-Extradition Upheavals (2019-present), and the COVID 19 pandemic (2020-present). In face of these, local institutions and individuals have undertaken many (unprecedented) countermeasures, which have profoundly affected their respective operations or behaviors. As shown in the ensuing survey of existing literature, different authors have used various indicators and methods to measure and geovisualize travel behaviors of different social groups and market segments across different spatiotemporal scales. However, only recently have academics done so in extreme situations where there are large-scale health-related abrupt and shocking events such as COVID-19. Therefore, we still know little about which indicators and methods would be more relevant in those situations and how we should adapt existing or develop new indicators and methods to better identify and epitomize wide-ranging and lasting impacts of extreme situations. In the existing literature, few also have taken advantage of non-traditional data such as smartcard data, which contain continuous and rich travel information about a much larger sample of transit riders at the individual level than traditional data such as surveys or interviews. The information provided by non-traditional data includes but is not limited to: transit stops and stations that these riders have been to, ingress/boarding and egress/alighting times, type/amount of fares paid, route selected, distance travelled, and frequency of travel. In principle, by linking the information to other publicly available data such as census, land use, and transit network, one can efficiently:(a) operationalize indicators and/or visuals to quantify/compare travel behaviors of millions of transit riders by different spatiotemporal units of analysis before, amid, and after an abrupt event (e.g., COVID-19), (b) fit quantitative models to predict impacts of the event on those behaviors and identify and/or verify their respective influencing factors, and (c) based on the above indicators, visuals, and models, explore socioeconomic implications of corresponding (spatiotemporal) variations across subgroups of travelers. In short, in the existing study, little has been done how to select, prioritize, and operationalize indicators and/or visuals based on both non-traditional and traditional data concerning adaptive travel behaviors and their variations across space and social groups given the occurrence of an abrupt event, especially COVID-19. This motivates us to (a) propose/adapt indicators and methods that are suitable for capturing impacts of COVID-19 and (b) demonstrate the relevance and usefulness of the proposed/adapted indicators by conducting an empirical study of Hong Kong's metro riders' adaptive behaviors at the metro station level and their predictors amid COVID-19. Hong Kong is a typical high-density city where public transit carries the lion's share (90%+) of passenger trips. If a shocking/abrupt event like COVID-19 and its derivatives influence the local society and economy, corresponding impacts can be more or less unfolded by looking at transit riders' behaviors and related changes. Most notably, the significance and meanings of different locations in the city are to a large extent (re)shaped by its extensive and expanded transit network and services and the reliance and usage of residents on them (c.f., Bertolini, 1996; Cervero and Murakami, 2009). Against the backdrop of COVID-19, the above significance and meanings can greatly change when fewer people go out and/or use transit and/or when transit services are adjusted or suspended. For instance, when lots of white-collar workers rely more or more on virtual meetings, the local central business district is no longer a place for intensive face-to-face interactions as it did prior to the outbreak of COVID-19. Thus, transit riders' behaviors and related changes amid COVID-19 provide a lens though which one can identify the impacts of a shocking/abrupt event on different riders across the space. When extra efforts are undertaken to quantify the sociodemographic attributes of the riders and how they are related to those impacts, we can also unravel how those impacts vary across different socioeconomic groups. Specifically, we attempt to examine in this study:(a) how residents, metro riders in particular, across different locales in Hong Kong have responded to the COVID-19 pandemic and subsequent or concurrent macro- and meso-level events, e.g., the Wuhan/Hubei lockdown and the local government's working from home policy, which are unprecedented institutional reactions to a shocking/abrupt event? (b) how different responses have reshaped activity and mobility patterns of metro riders, who constitute 37% of all the local transit patronage (Transport & Housing Bureau, 2017)? (c) Which built environment, metro network, socioeconomic, and spatial factors can explain the spatiotemporal variations in mobility patterns of metro riders? (d) In light of the above, what kind of (transferable) policy implications we can identify? In our study, we exploit both non-traditional (big) data, especially Octopus (smartcard) data and traditional data (e.g., censuses, surveys, and land use maps) to investigate the change in metro riders' travel pattern ex ante and ex post one remote/local event amid COVID-19. New or adapted indicators and visuals are introduced to quantify the pattern and its change at the metro station level. We hypothesize that metro station area (built environment) characteristics, metro station's network features, sociodemographic characteristics, and increased health risk among people due to the outbreak of COVID-19 would all significantly predict the change. Fitting regression models between the change and its probable influencing factors and comparing the pattern ex ante and ex post an event would allow us to unravel the possible determinants of the pattern and its change. For instance, whether and to what degree the percentage of blue-collar workers around a metro station would predict the number/distribution of destinations among riders from that station after the introduction of the local working from home policy? Also, whether and to what degree a metro station's centrality in the local metro network is correlated to the temporal concentration of incoming or outgoing trips of that station? The remainder of the paper is organized as follows. The next section (Section 2) reviews existing literature on travel behaviors and their changes, synthesizing (a) which indicators and/or methods have been and can be used to measure and geovisualize adaptive behaviors among transit/metro riders and (b) research gaps one can fill in the existing literature. Section 3 introduces measures and methods proposed by the authors in light of the literature review to quantify or geovisualize (adaptive) behaviors among transit/metro riders pre- and post-a typical COVID-19-related event. Section 4 presents an empirical study in the context of Hong Kong. In this section, we illustrate how to measure and geovisualize the (adaptive) behaviors and related changes using real-world data. We also quantify which explanatory variable(s) would significantly influence metro riders’ travel behaviors and related changes at the metro station area level. Section 5 concludes. 2 Changes in travel patterns/behaviors 2.1 Changes in travel patterns/behaviors without abrupt events Even without mega abrupt events such as COVID-19, there still can be observable changes in travel patterns and behaviors. Most studies prior to COVID-19 focus on these changes and how to quantify and geovisualize them. In the relatively stable pre-COVID-19 world, Zhao et al. (2018) illustrated there exists “abrupt, substantial and persistent changes” in travel behaviors, which can be quantified in three dimensions: the frequency of travel, time of travel, and origins/destinations. In face of abnormal natural events, e.g., extreme weather, De Palma and Rochat (1999) and Sabir et al. (2010) showed that trips can be re-scheduled, re-routed, and shifted to another mode. Corcoran and Tao (2017) adapted a method called “flow-comap” by Tao et al. (2014) to geovisualize the weather-transit usage relationship at the station-to-station level by hour of a day. Flow-comap can be used to identify and geovisualize probable trajectories of transit riders between origin-destination pairs. It is comparable to “desired lines” in travel demand modeling, which are the aggregated flows along the shortest path between centroids of two units of analysis, e.g., traffic analysis zone (Caliper, 2020). At the route level, transit users can stick to the same route or choose multiple routes for any given pair of transit stations or stops. Kim et al. (2017) developed a stickiness index to quantify this, which they defined as the range of preferences of transit users in route selection for a given period. They believe that between the same pair of start and end stations/stops, there can be users who prefer the same route or few routes and those do not. At the network/system levels, there could be “central places” for transit users, where there are the largest number of incoming or outgoing trips in few locales in a city or region across hours of a day or days of week. The number and distribution of central places, the origins and destinations of the trips to and from those places and associated desired lines can be used to simplify and characterize transit flows of a transit network across different temporal units. Using empirical data from Brisbane, Australia, Wei and Zhou (2016) showed the number and distribution of central places of local transit riders and their respective trip origins using a standard deviational ellipse. To identify those central places, they adapted the head-tail algorithm by Jiang (2013) which “partitions all of the data values around the mean into two parts and continues the process iteratively for the values (above the mean) in the head until the head part values are no longer heavy-tailed distributed”. To map out the desired lines between those central places and trip origins, they assumed that transit riders always chose the shortest path between any pair of transit stations or stops. Over a longer period and across spatial granularities, travel patterns of all travelers or a subset of them can remain stable or change significantly, which could have important socioeconomic and public policy implications (e.g., whether a subgroup of the population in a community consistently suffer from long commute). Hu et al. (2017) paid special attention to workers' commuting time and distance variability and stability, using three sets of the US Census data between 1990 and 2010. As those data already aggregated commuters’ origins and destinations by census tract, estimating accurate distance can be a challenge because of aggregation error and scale effect. They thus proposed and implemented a Monte Carlo simulation of individual trips to address the challenge. Their analyses show that lowest-wage workers suffered from poor mobility. Their workplaces were in few isolated locations and their home or workplace tracts had few transport options. 2.2 Changes in travel patterns/behaviors amid abrupt events Abrupt events are not exceptions in the world we live in. Prior to COVID-19, such events occurred too and notably impacted travel patterns/behaviors. In the existing scholarship, academics have examined the travel-related impacts of abrupt events such as the 911 terrorist attacks in the US (Blunk et al., 2006; Ito & Lee, 2005) and the Occupy Central Movement (OCM) in Hong Kong (Loo & Leung, 2017). Their studies reconfirm that “large-scale disruptions caused not just by natural hazards but also human beings” (Loo & Leung, 2017, p. 100). In addition, the impacts of the abruption, including those on travel patterns/behaviors are multidimensional and across different spatiotemporal scales. Often, new indicators and methods are needed to quantify and ascertain those impacts. To better quantify, manage, and mitigate impacts of events such as OCM, for instance, Loo and Leung (2017) proposed Key Resilience Performance indicators for the local transport system. Blunk et al. (2006) exploited the counter-factual forecast to ascertain whether the US domestic airline travel demand had restored to the average of normal years after the 911 terrorist attacks. Outside academia, there have also been many (online) narratives concerning travel impacts of social abruptions caused by abrupt civil unrests in Paris, London, and Hong Kong (e.g., Securewest International, 2021; The Major, 2019). The outbreak of COVID-19 has engendered a series of publications on impacts of the pandemic on people's travel and activities. These publications can be quickly identified through Google Scholar using key words such as COVID-19 and travel or mobility. They can be roughly categorized into two streams based on the input data used, corresponding indicators formulated, and what kind of information those indicators can convey and disclose. Stream 1 uses data from traditional sources such as surveys, which can better inform us about variations in the impacts across different social groups, modes of travel, and trip purposes. Beck and Hensher (2020), for instance, used survey responses to examine how the Australian government's various COVID-19 containment measures influenced travel activity patterns of Australians. Their study shows what kind of households and activities and which mode of travel were affected more: Younger households still made significantly more trips than other households; Public transport suffered the most—people have a high degree of trepidation with this mode after COVID-19. In Istanbul, Shakibaei et al. (2021) conducted longitudinal and cross-sectional surveys after COVID-19 and found that trips of all purposes were significantly suppressed. Again, they found that public transport experienced a decline in both (short-term) attractiveness and patronage. Stream 2 employs data from non-traditional sources such as social media, mobile phone location data, traffic control cameras, and public transport ITS, which often enables quantification and mapping of the impacts across more spatiotemporal units of analysis. Huang et al. (2020) used 580 million tweets to see how the single-day distance and the cross-day distance varied across countries at the global level and across states in the US. By exploiting mobile phone location data, Hara and Yamaguchi (2021) compared the travel behaviors before and after the official deceleration of COVID-19 outbreak in Japan. They found that the total number of trips decreased significantly, and population density index derived from the mobile phone location data was down 20%. Their findings are in general in lines with those reported in other studies such as Jenelius and Cebecauer (2020) in Sweden and Teixeira and Lopes (2020) in New York. Jenelius and Cebecauer (2020) used a combination of traditional ticket validation and (non-traditional) automatic passenger counting sensor data. Teixeira and Lopes (2020) used data from local bike-sharing and subway databases. The bike-sharing database contained both bike usage and a little sociodemographic information concerning shared bike users at the individual level whereas the subway database provided only rider counts by station for each 4 h of a day. 2.3 Influencing factors of changes in travel patterns/behaviors In existing studies, scholars have been investigating the explanatory factors that influence travel patterns/behaviors and their changes for decades. At some risk of oversimplifying the reality, there are approximately five groups of the explanatory factors that have been identified or extensively examined:(1) socio-demographic characteristics at the community level (Hanson, 1981; Pas, 1984; Goodchild et al., 1984), (2) land-use pattern or built-environment (Hansen, 1959; Boarnet & Sarmiento, 1998; Crane & Crepeau, 1998), (3) direct or indirect policies and transport strategies (Cervero, 1996; Meyer, 1999; Crane, 2000), (4) transport infrastructure and availability (Meyer, 1999; Handy, 2005), and (5) individual attributes and preferences (Hensher, 1994; Vogt, 1976). In the existing studies, however, only few authors investigate how commuters alter their travel demand and preferred modes when facing some unplanned events, such as terrorist attacks (Prager et al., 2011; Rubin et al., 2007; Ito & Lee, 2005.) and earthquakes (Gray et al., 1990; Yashinsky, 1999). Given its sudden outbreak, COVID-19, related events, and their impacts on travel patterns have not been well studied. Emerging (and limited) studies, however, have indicated that the impacts can be widespread, significant, and long lasting. In Budapest, Hungary, trips of all modes of transport decreased significantly amid COVID-19, with public transit suffering from the greatest decline (Bucsky, 2020). In Hong Kong, MTR saw only 637 million passengers in the first half of 2020, down 38% from 2019. The decline resulted in HK$400 million net loss to MTR, a very rare occurrence since the 1970s (Yau, 2020). In summary, the existing literature reviewed above indicates that changes in travel behaviors can occur regardless of there is an abrupt event like COVID-19 or not. Indicators and methods have been formulated to capture those changes. Most of those indicators and methods, however, are not tailored to measure and geovisualize the specific impacts of COVID-19 on travel behaviors. Little has been done to see which indicators and methods would be more appropriate than others to deal with those impacts. To the best knowledge of the authors, multiday smartcard data and its combination with other public available data such as census and point of interest data have also not been exploited to:(1) measure and geovisualize the impacts of COVID-19 on travel behaviors at the metro station level, (2) identify the spatial variation of those impacts, (3) see whether and how sociodemographic and network attributes (e.g., centrality) of a metro station can predict the impacts, and (4) whether and to what degree increased health risk among people because of the outbreak of COVID-19 would amplify the impacts. There are research gaps that one can fill. In this study, we attempt to adapt and operationalize a few existing indicators and methods using empirical data to fill some of those gaps. 3 Measures and methods proposed and used The outbreak of an abrupt shock like COVID-19 can have multidimensional impacts on different people in across locales. For a transit-reliant city like Hong Kong, many of those impacts can be identified should we formulate appropriate indicators and develop/adapt appropriate methods to measure or visualize the transit riders' travel patterns before and after a COVID-19 related event. In this study, we argue that the total numbers of trips, the average distance/duration of trips, spatiotemporal distribution of trips’ departure/arrival times/destinations by metro station can be used to quantify some impacts of those events on transit/metro riders. For instance, after the Wuhan/Hubei lockdown, fewer people would go out and thus fewer transit trips would be observed. For those who still travelled, they might reduce their respective travel distances/durations to minimize the healthy risk posed by COVID-19. Riders would also do their best to avoid peak hours and congested destinations to keep reasonable social distancing. Specifically, we propose the following six sets of indicators (or visuals) to measure and geovisualize different aspects of metro riders’ (adaptive) behaviors pre- and post-a special COVID-19 related event. We hypothesize that such event would affect the values of all those indicators and the variations in the values can be explained by sociodemographic, built environment, metro network, and spatial factors. For instance, metro services became less popular when a COVID-19-related event occurred thus most metro stations saw fewer incoming and outgoing trips. But there could be variations across stations and across rider groups (e.g., high-income vs. low-income and youth vs. elderly). Or in other words, the same COVID-19 related event can affect different locales and/or people differently. More details about the proposed indicators or methods and their relevance are as follows. Incoming (A) and outgoing (P) trips by station per hour. These trips reflect how popular a station is to riders from other stations and how many riders from a station are enticed to other stations. We hypothesize that where there are more essential facilities (e.g., groceries and medical services) in or around a station, the station would be more popular amid COVID-19 and thus residents around the station would make fewer outgoing trips whereas the number of incoming trips would not significantly decrease. A and P are calculated using the following equations:(1) Ai=∑j=1n−1tjiKi (2) Pi=∑i=1n−1tijKj where.tji is the incoming trips from Station j to Station i, j<>i and n is the total number of stops/stations in the local transit/metro system; tij is the outgoing trips from Station i to Station j. ki,j is the total hours of operation at Station i or j, ki and kj can be the same or be different. Average trip length (L) and average trip duration (D) for either A (or P). The following formulas are used to calculate L and D:(3) Li=∑j=1n−1(tji∗lji)Ai (4) Di=∑j=1n−1(tji∗dji)Ai where.lji is the travel distance between Station j to Station i; dij is the travel time between Station i and j. Given the outbreak of COVID-19, all stations should expect changes in both L and D as most riders would reduce the number, length, and duration of trips, if possible (c.f., Beck & Hensher, 2020). The temporary and spatial “stickiness” indices, which measure to what degree riders from Station i stick to few destinations (Sd) (or few periods (Sp)) and which are inspired by Simpson (1949) and Kim et al. (2017). For both Sd and Sp, they can adapt the formula by Kim et al. (2017) to calculate. For instance, Sdi, Sd for Station i is calculated as:(5) Sdi=∑j=1q(tijAi)2 where, q is the total number of stations (destinations) other than Station i that have non-zero trips from Station i. S has a value between 0 and 1. The larger S the fewer destinations, origins, or time slots are involved. If we talk about riders from a station, the spatial and temporal stickiness indices measure these riders' capacity to spread out or compress their outgoing trips into a larger or smaller set of destinations or time slots. We assume that those who can visit few other stations/locales and/or travel across more (off-peak) time slots would be safer amid COVID-19. The stickiness indices can also measure the similar capacity of incoming riders of a station, e.g., how widely distributed are their origins and how widely their arrival times can distribute across x-min time slots. The standard deviational ellipse is used to measure the degree of concentration in the spatial pattern of origins (or destinations) of incoming trips (or outgoing trips) by station. If there are few origins toward the periphery (a spatial normal distribution), a one standard deviation ellipse will cover approximately 68% of the origins and two standard deviations will contain approximately 95% of the origins (ESRI, 2018). The larger the ellipse, the more widely across the space a station draws its incoming riders from or sending its outgoing riders to. In this study, destinations of outgoing distinct riders are chosen to produce the two-standard deviational ellipses by station, weighted by the number of distinct riders of each destination on a day. In addition, the values of the major and minor axes (Amax and Amin) of are extracted as extra variables for us to quantify and characterize the spatial patterns of those destinations. They reflect the two major directions where most origins or destinations are located. Given the above, the size of the two-standard deviational ellipses and values of Amax and Amin allow us to quantify, geovisualize and compare attractiveness of (1) metro stations amid COVID-19 and (2) of a particular metro station before and after a special COVID-19-related event. 4 Case study area Hong Kong, a city with 7.5 million residents and 1100+ square kilometer land area is chosen as our study site (see Fig. 1 ).Fig. 1 Hong Kong in China and MTR network*. *Notes: unless otherwise stated, all figures in the text were made by the authors. Fig. 1 As of 2020, Hong Kong has a metro system (called Mass Transit Railway [MTR] locally) consisting of both heavy rail and light rail. On a typical weekday pre COVID-19, MTR carries over 4 million trips, which account for about 37% of the transit trips by the local residents and visitors (Transport & Housing Bureau, 2017). Like other Asian metropolises such as Seoul, Tokyo, and Singapore, the COVID-19 pandemic and related events' impacts on human mobility patterns in Hong Kong can thus be encapsulated by the changes in local public transit/metro usage. Furthermore, those patterns and changes can be quantified with the measures and methods proposed above. In our empirical study, besides the quantification based on Octopus (smartcard) data on four special days, we have also assembled local traditional data such as censuses and surveys to quantify probable influencing factors of the patterns and changes. The four special days are: January 17 (Friday), 22 (Wednesday), 24 (Friday), and 29 (Wednesday), 2020. Two of the four days represent the next day right after a macro or meso-level COVID-19 related event, i.e., January 24, 2020, a day after the Wuhan/Hubei lockdown and January 29, 2020, a day after Hong Kong Government announced its “working from home” mandate. Two comparable days one week ahead of these two days were selected: January 17 and 22, 2020. The situations of these two days serve as baselines for us to fathom the impacts of the COVID-19 related event(s) on travel behaviors and patterns of local metro riders. These baselines were chosen because we assume that travel patterns of the same weekday in two consecutive weeks were the most comparable. The further the baselines went back in time, the more factors could have influenced travel patterns, e.g., New Year's Day break, Chinese New Year celebration, seasonality, and socioeconomic changes. Comparing the patterns/visuals of January 27 and 24 2020 (Comparison 1) allows us to detect probable impacts of the Wuhan/Hubei lockdown event on the local metro usage whereas comparing the patterns/visuals of January 22 and 29, 2020 (Comparison 2) enables us to estimate probable impacts of both the Wuhan/Hubei lockdown and the “working from home” policy events (See Fig. 2 ).Fig. 2 Two comparison visualization. Fig. 2 The two events were selected because the Wuhan/Hubei lockdown occurred remotely, and the local government did not enforce any restrictions on local travel. Therefore, if there were any changes in local travel patterns, it was voluntary behaviors of residents. In contrast, the “working from home” policy was introduced by the local government and most public servants were affected. Thus, we can see to what degree restrictions on public servants' travel plus some subsequent voluntary residents’ reactions can affect the local travel pattern. More detail about how we operationalize the patterns/changes and their probable influencing factors is given as follows. 4.1 Patterns and changes: dependent variables We use a matrix-based method to effectively extract information from the raw Octopus data (a few samples are shown in Table 1 ) on the four days for us to obtain six sets of indicators or to get ready to implement the methods mentioned above, which allow us the measure travel patterns and their changes on the four days and between two comparable days. There are multiple types of Octopus data, this study focuses on ADL, i.e., adult card only as they are more likely to be (essential) workers and have less discretion regarding whether, when, and where to ride metro. On a typical day, 90% of metro riders are also adult riders.Table 1 Samples of octopus data. Table 1Card ID (Encrypted) Date Time Swipe typea Station ID 903537064 1/17/2020 9:00 ENT 19 903537064 1/17/2020 9:23 EXIT 29 … … … … … 903537065 1/17/2020 19:10 ENT 13 903537065 1/17/2020 19:23 EXIT 21 a ENT: Swiping into a station, EXIT: Swiping out of a station. Specifically, we first create matrices that contain four types of information about incoming or outgoing trips by station on each of the four days and average travel distance and time of those trips (See Table 2 ). In those matrices, each cell's row index (Ri) represents the origin of trips whereas column index (Cj) is the destination of the trip, for example, in the trip matrix, the value in cell (1, 2) represents the number of outgoing trips, i.e., 22, from Station 1 to Station 2.Table 2 Sample matrix of incoming and outgoing trips on a special day. Table 2Ri/Cj 1 2 … j 1 0 22 … … 2 12 0 … … … … … … … i … … … … Then we create a long table that contains the number of incoming or outgoing trips by station and by 15-min interval on each of the four days (see Table 3 ).Table 3 Number of incoming trips by station and by 15-min interval. Table 3Station ID Time period Number of trips 1 5:30–5:45 110 1 5:45–6:00 200 … … … 1 0:15–0:30 45 … … … With the above matrices and tables, we then calculate the numbers of incoming and outgoing trips by metro station per hour, the average travel distance and duration of those trips, standard deviational ellipses concerning those trips' origins or destinations, and stickiness indices measuring spatiotemporal preferences of those trips. Table 4 presents the descriptive statistics of indicators of all the 90 stations whereas Fig. 3 illustrates the standard deviational ellipse concerning outgoing trips’ destinations for Station LOHAS Park on January 24 and 17.Table 4 Descriptive statistics of indicators (dependent variables) of all the stations. Table 4Date Incoming trips Outgoing trips Aa L (km) D (mins) Sd Sp P L (km) D (mins) Sd Sp Jan 17 2296.7 (1406.3) 7.10 (2.68) 24.0 (4.1) 0.048 (0.030) 0.018 (0.002) 2296.7 (1293.8) 7.12 (2.66) 23.9 (4.2) 0.046 (0.026) 0.019 (0.004) Jan 22 2141.5 (1302.4) 7.08 (2.70) 24.2 (4.2) 0.046 (0.026) 0.018 (0.003) 2141.5 (1204.4) 7.08 (2.70) 24.2 (4.3) 0.045 (0.025) 0.019 (0.003) Jan 24 1735.7 (1026.6) 7.08 (2.73) 23.5 (4.3) 0.048 (0.028) 0.017 (0.003) 1735.7 (964.2) 7.10 (2.73) 23.5 (4.2) 0.048 (0.027) 0.017 (0.002) Jan 29 1048.3 (616.4) 7.09 (2.74) 23.5 (4.1) 0.051 (0.035) 0.017 (0.002) 1048.3 (568.1) 7.09 (2.73) 23.3 (4.7) 0.049 (0.026) 0.017 (0.002) a Standard deviation is in the parentheses and the total number of stations is 90. Fig. 3 Standard deviational ellipses for LOHAS park station on Jan17 and Jan 24. Fig. 3 Table 4 shows that after the two events—the Wuhan/Hubei lockdown and Hong Kong's Working from Home policy— the following results can be seen:(1) fewer incoming and outgoing trips by metro station can be observed, meaning that some riders did perceive the health risk posed by COVID-19 and had reduced their trip frequencies; (2) For those continued traveling, their average travel distance seemed to be constant before and after either of the two special days whereas their average travel time slightly decreased; (3) For those continued traveling, they seemed to go to slightly fewer destinations and spread out their departure times into more 15-min intervals on a day despite that frequencies of MTR's train services largely remained unchanged; (4) given (1) to (3), it can also be concluded that those stopping traveling tended to have longer average travel time than those continued traveling. Fig. 3 indicates that after the Wuhan/Hubei lockdown event, the standard deviational ellipse for 95% of the destinations of outgoing trips from Station LOHAS Park shrunk significantly. The shrinkage can be seen in the differences in the centroid, size of the ellipse, and the values and directions of both major and minor axes. Of all the metro stations, Station LOHAS Park was the one experiencing the most changes in these regards. In Fig. 4a, Fig. 4b, Fig. 4c, Fig. 4d (a) to 4(d) , we map out how stickiness index for outgoing trips changed between January 17 and 24 and between January 22 and 29. Fig. 4a, Fig. 4b(a) and (b) show the possible impacts of the Wuhan/Hubei lockdown on the temporal and spatial index changes whereas Fig. 4c, Fig. 4d(c) and (d) indicate the possible impacts of Hong Kong's working from home policy (possibly the continuous impacts of the Wuhan/Hubei lockdown as well) on the changes. Again, the index measures whether and to what degree riders from one metro station go to few other stations or travel in few time segments.Fig. 4a Temporal Stickiness Index (Sp) Change after the Wuhan/Hubei lockdown. Fig. 4a Fig. 4b Spatial Stickiness Index (Sd) Change after the Wuhan/Hubei lockdown. Fig. 4b Fig. 4c Temporal stickiness index (Sp) change after the working from home mandate. Fig. 4c Fig. 4d Spatial stickiness index (Sd) change after the working from home mandate. Fig. 4d To better geovisualize the spatial variation, we map out the index by tertiary planning unit (TPU), which is a geographic reference system demarcated by the Planning Department of the Hong Kong Government. In all the figures, blue, transparent, or red colors in each TPU represent different types of changes where blue represents negative change, transparent means little or no change, and red is positive change. Interestingly, if we tentatively overlook the sign of the change, the TPUs that experienced the most change in the four figures resemble. This means that those TPUs were most influenced by COVID-19 related events occurring in a relatively short period (e.g., two weeks in our case) might not totally be random. It can be the same set of communities that saw the most impacts in the weeks. Furthermore, those TPUs contain one checking point, theme park, and/or terminal station tended to be consistent members of this set, e.g., West Kowloon (high-speed railway station and (checking point), Lok Ma Chau (checking point/terminal station), Lo Hu (checking point/terminal station), Disneyland (theme park), Wong Chuk Hang (Ocean Park), Tung Chung (terminal station), and LOHAS park (terminal station). Similarly, the local central business districts (e.g., Central and Tsim Sha Tsui) also suffered significant impacts. Several other TPUs also experienced significant changes, e.g., Tai Shui Hang, Sha Tin, and Heng Fa Chuen. Without first-hand data about these polygons/communities, we have no clues to know what happened there. But one thing is certain based on Fig. 4a, Fig. 4b, Fig. 4c, Fig. 4d(a)–(d): the impacts of macro- and meso-level external/internal events would not have homogeneous impacts on different TPUs. 4.2 Influencing factors of the patterns and changes: independent variables To investigate whether and to what degree MTR riders’ travel patterns and related changes are affected by different explanatory factors, we compiled local publicly available data. The data include censuses, land use maps, and transit network files (.shp files) that can be fed into Geographic Information System. Based on these data, we formulate indicators to different explanatory factors. In light of the literature review and local data availability, we formulate four sets of variables by TPU to measure the predictors for the change in MTR riders’ travel patterns between a special day and the baseline. Using the TPU boundaries and attribute tables and other data available to us, we created four sets of variables:(1) socio-demographic characteristics from the 2011 Hong Kong Census Data; (2) built-environment factors depicted by the 2017 point of interest (POI) data of Hong Kong; (3) the centrality degree of each station in the MTR network based on the January 2020 Octopus (smartcard) data and MTR network. shp files; (4) the changing rate of the total incoming and outgoing trips by MTR station. Sets 1 variables are rather conventional according to the existing literature. They include all factors that can explain variations in impacts of COVID-19 and subsequent event(s) on travel patterns of metro riders. These factors include education, age, employment status, occupation, household attributes, commuting distance, and income. Existing studies have indicated, for instance, younger households would still make significantly more trips amid COVID-19 (e.g., Beck & Hensher, 2020). Concerning Set 2 variables, six categories of POIs are separately considered: recreational, transportation, public facility, medical service, commercial, and others. The first five are singled out as we subjectively regard them as essential functions that people must turn to regardless of there exists a pandemic or not. But we are unsure what would be more important among them amid COVID-19. In addition, we use Simpson index to measure diversity of different categories of POIs. We hypothesize that diversity in a locale would reduce outgoing trips of residents therein whereas increase incoming trips of residents elsewhere. But in this study, we focus on outgoing trips only. Set 3 variables are inspired by (Zhou et al., 2019). Specifically, the average travel time based on Octopus data from an MTR station to all other stations is calculated to reflect the global centrality of that station. The shorter this time is, the higher centrality degree. Besides, the number of total population residing in 15 min’ metro ride to an MTR station are considered as an extra measurement of the regional centrality of a station. We hypothesize that lots of facilities and opportunities would be around stations with high centrality and therefore people would still need to travel to those stations despite of the outbreak of COVID-19 and increased health risk subsequently. Set 4 variables--the rates of change in average daily trips are included because they partially reflect how incoming and outgoing metro riders by metro station perceived and reacted to the health risk of COVID-19 and related news about in the two weeks with the Wuhan/Hubei lockdown and “working from home” in presence as compared to the baseline we subjectively chosen: the first working week of 2020 after the New Year's Day break and the last week when there was still little in local media about the pandemic. We hypothesize that the more sensitive riders to and from a metro station to the health risk of COVID-19 and related news the more likely that riders' travel patterns of that station would change, e.g., fewer incoming and outgoing trips, trips staggering into more x-min intervals, and trips to fewer other stations. Table 5 presents descriptive statistics of all the variables that we formulate.Table 5 Descriptive statistics of the independent variables. Table 5 Variables Mean(stdev) Socio-demographics Education Primary and below 19367.42 (18131.27) Secondary/Sixth Form 30770.21 (28804.82) Post-secondary 15787.09 (13016.58) Demographic Total population 65924.72 (58252.43) Age under 15 7539.12 (7101.31) Age 15-24 8175.94 (8422.27) Age 25-44 20534.33 (17653.44) Age 45-64 20925.63 (18960.56) Age over 65 8749.69 (7522.86) Age median 42.35 (2.50) Employment Working population 33132.94 (29122.51) Persons not in working population 32791.78 (29439.06) Employees 29669.61 (26422.38) Employers 1376.08 (1126.73) Self-employed 1963.38 (1712.55) Unpaid family worker 123.88 (110.29) Income Median income 14124.11 (5841.84) Monthly income under $10,000 13057.18 (12391.24) Monthly income $10,000-$20,000 11268.17 (10830.95) Monthly income $20,000-$40,000 6142.21 (5273.49) Monthly income over $40,000 2665.39 (2490.57) Occupation Managers and administrators 3100.67 (2557.91) Professionals/Associate professionals 8871.81 (7646.49) Clerical support, service and sales workers 10802.56 (10311.76) Craft and related workers 4151.92 (4655.40) Elementary occupations and agricultural workers 6205.99 (5447.83) Household Total households 22235.94 (19124.99) Nuclear family households 14968.48 (13548.32) Relative households 3170.73 (2816.48) Other households 4096.73 (3196.52) Household size 1-3 14918.08 (12429.83) Household size 4 4766.12 (4612.29) Household size 5 or up 2551.74 (2353.17) Average household size 2.82 (0.30) Median monthly income 28606.11 (21380.90) Household monthly income under $10,000 5274.68 (5459.49) Household monthly income $10,000 to $20,000 5304.36 (5325.44) Household monthly income $20,000 to $40,000 6696.14 (6210.51) Household monthly income over $40,000 4960.77 (4216.29) Housing type Public rental housing 19676.19 (29904.47) Subsidized home ownership housing 13989.80 (23208.75) Private permanent housing 30318.41 (22258.68) Non-domestic housing 218.51 (748.08) Temporary housing 160.81 (281.56) Population in domestic households 64363.72 (57229.47) Population in non-domestic households 1561.00 (1230.34) Commuter Workers work in same district 5466.37 (4508.11) Workers work in different district 27666.58 (25973.09) Students studying in same district 6096.36 (6048.43) Students studying in different district 4771.27 (5255.56) Total number of students 10867.62 (10775.39) Build environment Number of POIs Recreation facilities 31.90 (55.96) Commercial facilities 69.42 (123.13) Public facilities 31.30 (33.51) Medical facilities 69.37 (113.12) Transport facilities 7.71 (13.08) Related index Total number of all facilities 209.70 (299.45) Simpson diversity 0.70 (0.11) Centrality degree of a station Average travel time to other stations 40.11 (8.86) Population within 15 min' travel 41179.85 (26804.46) Perceived health risk Wuhan/Hubei Lockdowna Incoming riders 1.04 (0.05) Outgoing riders 1.05 (0.06) “Working from Home”b Incoming riders 0.72 (0.09) Outgoing riders 0.72 (0.08) a Rate of change as compared to the baseline—average daily trips between 6 and 10 Jan 2020. b Rate of Change as compared to the baseline. 4.3 Regression models: what affected changes in travel patterns of MTR users Given that little has been done in the existing studies concerning what affects travel patterns in a pandemic, we assume that there exists linear relationship between travel pattern (including its changes) and explanatory factors. Our overall hypothesis is that both events would see metro riders going to fewer stations and stagger their departure times into more 15-min intervals after either of the events, i.e., at least some riders would feel higher health risks because of COVID-19 and related events and therefore adapted their travel behaviors. In terms of size and direction of the two events’ impacts, the above-mentioned four sets of independent variables would influence them. Then we adopted a two-step method (see Fig. 5 ) to fit a series of ordinary least square (OLS) regression models using the assembled the dependent and independent variables (see Table 4, Table 5) we assembled as input.Fig. 5 The two-step method to fit regression. Fig. 5 Our models’ results are presented in Table 6 , followed by our interpretations of them. The most significant results are highlighted in bold.Table 6 Stickiness index changes after remote/local events and their predictors. Table 6 Dependent variable a (Time stickiness index change rate between Jan 17 and 24a) Dependent variable b (Destination stickiness index change rate between Jan 17 and 24a) Dependent variable c (Time stickiness index change rate between Jan 22 and 29a) Dependent variable d (Destination stickiness index change rate between Jan 22 and 29a) Independent variablesb Socio-demographics Public rental housing −0.257 (0.008) *** −0.189 (0.067) * Temporary housing 0.133 (0.148) 0.163 (0.138) Subsidized home ownership housing −0.136 (0.447) Population in non-domestic households Non-domestic housing −0.173 (0.124) Age over 65 0.354 (0.100) Monthly income 20-40k Monthly income over 40k Household income over 40k 0.567 (0.000) *** Income median 0.138 (0.219) Craft and related occupation −0.453 (0.039) ** Students studying in different districts 0.468 (0.009) *** Built environment Public facilities 0.461 (0.000) *** Medical facilities 0.311 (0.009) *** Centrality degree of a station Population within 15 min' travel 0.130 (0.162) 0.092 (0.491) Average travel time to other stations 0.189 (0.241) Perceived health risk Outgoing riders 0.169 (0.072) * Model results F 10.265 7.291 2.339 12.686 R2 0.328 0.203 0.146 0.374 Notes: Standardized coefficients are in listed on the table only when the independent variables are considered in the model. *** Indicators significance at the 99% level. ** Indicators significance at the 95% level. *Indicators significance at the 90% level. a Descriptive statistics of the stickiness indices on the selected dates are shown in Italic Table 4. b Only variables selected by Automatic linear model and certified as non-collinear are listed. Table 6 shows how the temporal and spatial stickiness index for outgoing trips changed at the metro station level is associated with different predictors at the TPU level after a macro-level and external event: The Wuhan/Hubei lockdown on January 23, 2020 and a meso-level and internal event: The Hong Kong Government's “working from home” policy released on January 28, 2020. By definition, a reduction in the index means that metro riders from one station would go to more other stations (or travel in more x-min intervals) whereas an increase indicates that those riders would go to fewer other stations (or travel in fewer x-min intervals) (see Equation (5)). Predictors for Wuhan/Hubei lockdown's impacts on temporal distribution of outgoing trips by 15-min interval: Not surprisingly, the Wuhan/Hubei lockdown seemed to have enticed more outgoing riders to stagger their departure times into more 15-min intervals, where the time stickiness index changed from 0.019 on January 17 to 0.017 on January 24—a 10% change. In our regression model results, the numbers of residents in public rental housing and public facilities can significantly predict the change at the 99% confidence level. The larger the number of public rental housing residents, the larger reduction in the stickiness index—meaning that outgoing trips would stagger into more 15-min intervals. The more public facilities there were around a metro station, the larger growth in the stickiness index—meaning that outgoing trips from that station would stagger into fewer 15-min intervals. Predictors for Wuhan/Hubei lockdown's impacts on destination distribution of outgoing trips by metro station: According to Table 4, the Wuhan/Hubei lockdown seemed to have slightly prevented a metro station's riders from traveling to other metro stations ex post the outbreak of COVID-19. At the network level, metro riders on average went to slightly fewer destinations (mean Sd = 0.048) on January 24 than on January 17 (mean Sd = 0.046). In terms of predictors for the change rate in the destination stickiness index, only the number of students studying in different districts is statistically significant at the 99% level. The larger the number, the larger the positive change in the Sd value, meaning that trips from a metro stations would go to fewer other stations. This indicates that many students used to travel frequently by MTR prior to the Wuhan/Hubei lockdown. Predictors for “Working from home” impacts on temporal distribution of outgoing trips by 15-min interval: As expected, the mandate had reduced the temporal stickiness index between January 22 and 29, which means that riders from all the stations on average would stagger their departure times into more 15-min intervals. In terms of which predictors that can significantly predict the size of the change, it is the number of residents who worked in craft and related occupations (at the 95% level) and the number of medical facilities (at the 99% level). Specifically, the more residents who worked in craft and related occupations, the bigger the negative change, meaning the more residents who worked in craft and related occupations around a metro station, the more 15-min intervals that riders from that station would travel after the introduction of the “working from home” mandate. In other words, those workers in craft and related occupations were more likely to have discretion to stagger their departure time from an MTR station whereas other workers such as public servants might not have the same degree of discretion. In addition, the more medical facilities around a metro station the fewer 15-min intervals that riders from that station would travel after the introduction of the mandate. Predictors for “Working from home” impacts on destination distribution of outgoing trips by metro station: Hong Kong Government's working from home mandate did reduce the average number of destinations that metro riders from different stations would go to, where the stickiness changed from 0.045 on January 22 to 0.048 on January 29 (see Table 4). In the regression model, the number of households with HK$40,000 or more income (i.e., local middle-class or above households) can positively predict the change rate in the destination stickiness index, meaning that the more well-to-do households living around a metro station the fewer destinations that riders from that station would visit after the mandate was enforced. 5 Discussion and conclusions COVID-19 and related events significantly influence urban dynamics. Local travel patterns and transit demand are important proxies one can use to measure those dynamics. Amid COVID-19, changes in travel patterns and transit demand vary across locales and people. To date, the empirical information afforded by COVID-19, related events, and subsequent institutional changes and adaptive behaviors have largely been underexploited because of challenges in (a) acquiring such information prior to the availability of innovative sensors and corresponding big data and (b) deriving useful, complete, and comprehensive information from big data alone. In addition, we are still unsure what kind of measures and methods can be used to efficiently analyze those data. In this study, we hypothesize that transit riders would adapt their travel behaviors given the occurrence of macro- and meso-level COVID-19 related events. We identify and propose feasible measures and methods that can be used to quantify transit riders' adaptive behaviors and their change amid COVID-19. Based on empirical data from Hong Kong, which consist of both big and traditional (small) data, we illustrate how those measures and methods can be operationalized and implemented. In light of the existing studies and local data availability, we also examine how those adaptive behaviors were affected by four sets of sociodemographic, built environment, metro network, and spatial factors/variables after the Wuhan/Hubei lockdown and Hong Kong's working from home mandate, two unprecedented institutional reactions to an abrupt shock. Based on all the above, some generic and transferable conclusions can be drawn: First, both local and remote institutional reactions to abrupt shocks like COVID-19 can significantly influence urban dynamics, which can be measured by indictors concerning transit travel patterns across locales and people in a transit-dependent city like Hong Kong. Our above empirical studies in Hong Kong indicate that metro riders’ adaptation in departure time distribution and outgoing destination choice amid COVID-19 can be significantly predicted by sociodemographic attributes such as the percentage of residents in public housing, the numbers of students studying in different districts, workers in craft and related occupations, and well-to-do households at the metro station area level. Ideally, people/riders should travel less often, shorten their trip distance and duration, stagger their departure times, and visit fewer destinations to reduce the health risk posed by COVID-19. But not all the people can succeed in doing so as reflected in our regression model results. For instance, in Hong Kong that wealthy households or communities tended to reduce more destinations than other households when the perceived health risk was high. Before, amid and after COVID-19, the magnitude and even sign of the impacts of sociodemographic attributes could vary. The number of students studying in different districts, for instance, can mean trips to more destinations from a metro station before COVID-19. But the same number is correlated to trips to fewer destinations amid COVID-19. Our results also indicate that certain built environment factors would influence the metro riders' adaptation capacity, e.g., the numbers of medical facilities and public housing. All the above mean that public policy making and built environment planning should account for the heterogenous impacts of both sociodemographic and built environment factors on people's adaptation capacity. How to identify and manage all the impacts systematically become a new topic for policymakers. Second, given the suddenness and scale of COVID-19, little has been done on how the pandemic and related events influence transit riders' travel patterns in the context where there is a high reliance on transit. In existing technical reports or guidelines on travel and transit demand modeling and transit services planning and operations, policy scenarios during and after pandemics such as COVID-19 have also rarely been considered. There is a need for us to quantify transit riders’ adaptive behaviors against the backdrop of COVID-19. Some of these behaviors might not exist prior to COVID-19, for instance, staggering trips into more time slots to avoid health risk. It is even better if we can have data concerning predictors of those behaviors. Those efforts can help us better identify vulnerable transit riders, frequent/choice transit riders, and probable transit services gaps where local transit operators can fill. Third, both macro-level (external) and meso-level (internal) events’ impacts can be measured and detected conveniently based on indicators or tools such as the numbers of incoming/outgoing trips, average distance and duration of these trips, stickiness indices of the origins or destinations of these trips, and 95% standard ellipses of the origins or destinations. Some of these indicators and tools (e.g., the 95% standard ellipses) have long been in existence but others (e.g., stickiness indices) must be invented to consider the impacts of an abrupt shock like COVID-19. It is meaningful to inventory all these indicators and tools and their respective relevance and utilities so that they better serve policy analysts and decisionmakers. Four, in addition to those metro stations or subareas serving cross-border passengers, tourists, and exurban riders around the terminal stations, other stations and subareas can be profoundly influenced by both macro- and meso-level internal/external events. In Hong Kong, for instance, we found that Tai Shui Hang and Heng Fa Chuen can belong to the latter, where riders tended to visit fewer other stations after those events. Due to first-hand data constraints, we still do not know what happened in those stations and their adjacencies. But if we assume that mobility is a necessary condition for sufficient options and decent quality of life, then riders who had to significantly reduce more of their destinations than others in the same city when there were COVID-19 events might suffer from fewer choices and decline in quality of life (c.f., Adey, 2017). Despite the above progresses, our study can still be improved and enhanced in at least the following aspects. First, once there is extra financial and community support, we should conduct surveys among riders to supplement the data used in this study. The Octopus data used, for instance, only inform us where and when riders travel to and from. They do not tell us who they were, why they travelled, and how they perceived their trips. Second, due to time, data, and budgetary constraints, this study only examined impacts of two special events on metro riders' travel patterns. In the future, impacts of more events, especially micro-level events, e.g., revealed confirmed cases and associated venues and upper customer limit for restaurants, on transit riders' travel patterns should be investigated. Finally, this study does not control adjustments of bus services when fitting the models to quantify or geovisualize impacts of different predictors on metro riders’ travel pattern changes, e.g., the stickiness index change. It also does not investigate how those adjustments might have impacted certain metro/bus riders or certain locales or metro/bus corridors more than others. In the future, more work should be done in these aspects. Funding Funded by a grant from 10.13039/501100012479 General Research Fund of Hong Kong (Project number: 17603220). Authors statement Jiangping Zhou, PhD, corresponding author, designed the research, secured funding and data to carry out the research, performed data analysis, wrote and revised the manuscript; Jiangyue Wu, PhD candidate and Hanxi Ma, PhD student, performed data analysis and reviewed the manuscript draft. ==== Refs References Adey P. Mobility 2017 Routledge London, UK Beck M.J. Hensher D.A. Insights into the impact of COVID-19 on household travel and activities in Australia – the early days of easing restrictions Transport Policy 2020 10.1016/j.tranpol.2020.08.004 Bertolini L. Nodes and places: Complexities of railway station redevelopment European Planning Studies 1996 10.1080/09654319608720349 Blunk S.S. Clark D.E. McGibany J.M. Evaluating the long-run impacts of the 9/11 terrorist attacks on US domestic airline travel Applied Economics 2006 10.1080/00036840500367930 Boarnet M.G. Sarmiento S. Can land-use policy really affect travel behaviour? A study of the link between non-work travel and land-use characteristics Urban Studies 1998 10.1080/0042098984538 Bucsky P. Modal share changes due to COVID-19: The case of Budapest Transportation Research Interdisciplinary Perspectives 2020 10.1016/j.trip.2020.100141 Caliper What are desire lines? Available at https://www.caliper.com/glossary/what-are-desire-lines.htm 2020 Cervero, R., & Murakami, J. Rai and Property Development in Hong Kong: Experiences and Extensions. Urban Studies. 10.1177/0042098009339431. Cervero R. Radisch C. Travel choices in pedestrian versus automobile oriented neighborhoods Transport Policy 1996 10.1016/0967-070X(96)00016-9 Corcoran J. Tao S. Mapping spatial patterns of bus usage under varying local temperature conditions Journal f Maps 2017 10.1080/17445647.2017.1378933 Crane R. The influence of urban form on travel: An interpretive review Journal of Planning Literature 2000 10.1177/08854120022092890 Crane R. Crepeau R. Does neighborhood design influence travel? : A behavioral analysis of travel diary and GIS data Transportation Research Part D 1998 10.1016/S1361-9209(98)00001-7 De Palma A. Rochat D. Understanding individual travel decisions: Results from a commuters survey in Geneva Transportation 1999 2018 http://doi:10.1023/A:1005121605832ESRI ESRI Directional distribution (standard deviational ellipse) Available at https://tinyurl.com/y4tomuz9 2018 Goodchild M.F. Janelle D.G. The city around the clock: Space-time patterns of urban ecological structure (halifax, Canada) 1984 Environment & Planning A 10.1068/a160807 Gray G. Roberts J. Markowitz J. Aftershock: Dealing with the highway Crisis after the loma prieta earthquake 1990 TR News Handy S. Cao X. Mokhtarian P. Correlation or causality between the built environment and travel behavior? Evidence from northern California Transportation Research Part D 2005 10.1016/j.trd.2005.05.002 Hansen W.G. How accessibility shapes land use Journal of the American Planning Association 1959 10.1080/01944365908978307 Hanson S. Hanson P. The travel-activity patterns of urban residents: Dimensions and relationships to sociodemographic characteristics 1981 Economic Geography 10.2307/144213 Hara Y. Yamaguchi H. Japanese travel behavior trends and change under COVID-19 state-of-emergency declaration: Nationwide observation by mobile phone location data Transportation Research Interdisciplinary Perspectives 2021 10.1016/j.trip.2020.100288 Hensher D.A. Stated preference analysis of travel choices: The state of practice Transportation 1994 10.1007/BF01098788 Huang X. Li Z. Jiang Y. Li X. Porter D. Twitter reveals human mobility dynamics during the COVID-19 pandemic PloS One 2020 10.1371/journal.pone.0241957 Hu Y. Wang F. Wilmot C.G. Commuting variability by wage groups in Baton Rouge, 1990–2010 Papers in Applied Geography 2017 10.1080/23754931.2016.1248577 Ito H. Lee D. Assessing the impact of the September 11 terrorist attacks on U.S. airline demand Journal of the Economics of Business 2005 10.1016/j.jeconbus.2004.06.003 Jenelius E. Cebecauer M. Impacts of COVID-19 on public transport ridership in Sweden: Analysis of ticket validations, sales and passenger counts Transportation Research Interdisciplinary Perspectives 2020 10.1016/j.trip.2020.100242 Jiang B. Head/tail breaks: A new classification scheme for data with a heavy-tailed distribution The Professional Geographer 2013 10.1080/00330124.2012.700499 Kim J. Corcoran J. Papamanolis M. Route choice stickiness of public transport passengers: Measuring habitual bus ridership behaviour using smart card data Transportation Research Part C 2017 10.1016/j.trc.2017.08.005 Loo B.P.Y. Leung K.Y.K. Transport resilience: The Occupy central movement in Hong Kong from another perspective Transportation Research Part A 2017 10.1016/j.tra.2017.09.003 Major T.he Measurement of traffic and pollution impacts from protests Available at https://www.london.gov.uk/questions/2019/19677#a-174956 2019 Meyer M.D. Demand management as an element of transportation policy: Using carrots and sticks to influence travel behavior Transportation Research Part A 1999 10.1016/S0965-8564(99)00008-7 Pas E.I. The effect of selected sociodemographic characteristics on daily travel-activity behavior Environment & Planning A 1984 10.1068/a160571 Prager F. Beeler Asay G.R. Lee B. Winterfeldt V.D. Exploring reductions in London Underground passenger journeys following the July 2005 bombings Risk Analysis 2011 10.1111/j.1539-6924.2010.01555.x Rubin G.J. Brewin C.R. Greenberg N. Hughes J.H. Simpson J. Wessely S. Enduring consequences of terrorism: 7-month follow-up survey of reactions to the bombings in London on 7 july 2005 British Journal of Psychiatry 2007 10.1192/bjp.bp.106.029785 Sabir M. Ommeren J. Koetse M.J. Rietveld P. Impact of weather on daily travel demand Proceedings of the tinbergen institute discussion paper 2010 VU University Amsterdam http://www.fietsberaad.nl/library/repository/bestanden/Sabir%20et%20al%20(2010a).Pdf Securewest International How civil unrest affects travel Available at https://www.securewest.com/how-civil-unrest-affects-travel/ 2020 Shakibaei S. de Jong G. Alpkokin P. Rashidi T.H. Impact of the COVID-19 pandemic on travel behavior in istanbul: A panel data analysis Sustainable Cities and Society 2021 10.1016/j.scs.2020.102619 Simpson E.H. Measurement of diversity Nature 1949 10.1038/163688a0 Tao S. Rohde D. Corcoran J. Examining the spatial-temporal dynamics of bus passenger travel behaviour using smart card data and the flow-comap Journal of Transport Geography 2014 10.1016/j.jtrangeo.2014.08.006 Teixeira J.F. Lopes M. The link between bike sharing and subway use during the COVID-19 pandemic: The case-study of New York's Citi Bike Transportation Research Interdisciplinary Perspectives 2020 10.1016/j.trip.2020.100166 Transport Housing Bureau Public transport strategy study 2017 Transport and Housing Bureau Hong Kong Vogt J.W. Wandering: Youth and travel behaviour Annals of Tourism Research 1976 10.1016/0160-7383(76)90051-7 Wei M. Zhou J. Central places of transit riders: A visual story of Brisbane, Australia Regional Studies, Regional Science 2016 10.1080/21681376.2016.1262277 Yashinsky M. Repairing highways after earthquakes 1999 Technical Council on Lifeline Earthquake Engineering Monograph Yau C. Hong Kong's MTR faces HK$400 million net loss for first half of 2020 amid coronavirus pandemic, protest fallout. South China Morning Post Available at https://www.scmp.com/news/hong-kong/transport/article/3092205/hong-kongs-mtr-faces-hk400-million-net-loss-first-half 2020 Zhao Z. Koutsopoulos H.N. Zhao J. Detecting pattern changes in individual travel behavior: A bayesian approach Transportation Research B 2018 10.1016/j.trb.2018.03.017 Zhou J. Yang Y. Gu P. Yin L. Zhang F. Zhang F. Li D. Can TODness improve (expected) performances of TODs? An exploration facilitated by non-traditional data Transportation Research Part D 2019 10.1016/j.trd.2019.07.008

0

PMC9753124

NO-CC CODE

2022-12-16 23:25:30

no

Appl Geogr. 2021 Sep 10; 134:102504

utf-8

Appl Geogr

10.1016/j.apgeog.2021.102504

oa_other

==== Front Appl Geogr Appl Geogr Applied Geography (Sevenoaks, England) 0143-6228 0143-6228 Elsevier Ltd. S0143-6228(21)00122-3 10.1016/j.apgeog.2021.102506 102506 Article How will COVID-19 impact Australia's future population? A scenario approach Charles-Edwards Elin a∗ Wilson Tom b Bernard Aude a Wohland Pia a a The University of Queensland, St Lucia, Brisbane, 4072, Australia b The University of Melbourne, Parkville, Melbourne, 3010, Australia ∗ Corresponding author. 6 7 2021 9 2021 6 7 2021 134 102506102506 14 7 2020 11 5 2021 3 7 2021 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The impact of COVID-19 has been massive and unprecedented, affecting almost every aspect of our daily lives. This paper attempts to quantify the impact of COVID-19 on the future size, composition and distribution of Australia's population by projecting a range of scenarios. Drawing on the academic literature, historical data and informed by expert judgement, four scenarios representing possible future courses of economic and demographic recovery are formulated. Results suggest that Australia's population could be 6 per cent lower by 2040 in a Longer scenario than in the No Pandemic scenario, primarily due to a huge reduction in international migration. Impacts on population ageing will be less severe, leading to a one percentage point increase in the proportion of the population aged 65 and over by 2040. Differential impacts will be felt across Australian States and Territories, with the biggest absolute and relative reductions in growth occurring in the most populous states, Victoria and New South Wales. Given the ongoing nature of the crisis at the time of writing, there remains significant uncertainty surrounding the plausibility of the proposed scenarios. Ongoing monitoring of the demographic impacts of COVID-19 are important to ensure appropriate planning and recovery in the years ahead. Keywords Population scenarios COVID-19 Migration Projections ==== Body pmc1 Introduction Since its emergence in Wuhan, China in late 2019, Severe Acute Respiratory Syndrome Coronavirus (SARs CoV-2) and its associated disease (COVID-19) has spread rapidly, impacting more than 219 countries and territories (Worldometers, 2021). The effects of COVID-19 have been pervasive, impacting all economic, social and political domains, and causing 3.2 million deaths globally by the beginning of May 2021 (CSSE, 2020). COVID-19 is expected to affect the size, composition and distribution of national populations due to disruptions in long-term trends in mortality (Banerjee et al., 2020; Ferguson et al., 2020; Goldstein & Lee, 2020; Trias-Llimós & Bilal, 2020), fertility (Luppi, Arpino, & Rosina, 2020b; Stone, 2020a) and migration (Balbo et al., 2020). There is significant uncertainty with respect to the duration of the crisis, with many countries experiencing a third wave of infections (Henley, 2021) at the time of writing. There is additional uncertainty tied to the timing and equity of the global vaccination rollout (Science Alert, 2021). Mortality and migration are likely to be most affected components of demographic change in the short-term term, but there is significant global variation in the magnitude of these impacts (Kontis et al., 2020). As of the beginning of May 2021, Australia had experienced fewer than 1000 deaths due to COVID-19 (Australian Government Department of Health, 2021c). The disruption to the global migration system (Zambrano, Bueermann, & Sullivan, 2020), however, is expected to substantially impact Australia's future population size and composition, exacerbated by Australia's ongoing international border closure (Australian Government Department of Health, 2021b). The distribution of Australia's population is likely to be affected by internal travel restrictions and lockdowns and the economic fallout which places downward pressure on internal migration rates (Levy, 2017). In concert, these demographic forces will produce very different futures than those forecast prior to the emergence of COVID-19. Population projections are a fundamental tool for economic, social and regional planning. Assumptions underpinning projections are generally trend-based, derived from mathematical modelling and/or informed by the expert judgement of demographers. COVID-19 represents an unprecedented disruption to demographic behaviour, rendering existing projections unusable, and mathematical modelling of trends difficult. While there is a clear need to update projections for this new COVID-19 world, uncertainty as to the duration and scale of impacts as well as a lack of contemporaneous data, makes standard population forecasts unfeasible. The best option available to demographers is the development of scenarios which model a range of plausible population futures. This is the approach taken in this paper. The paper presents a range of population scenarios to capture the potential impact of COVID-19 on the size, composition and distribution of Australia's population. We begin with some background on Australia's pre-COVID-19 population dynamics, before providing an overview of the COVID-19 restrictions expected to impact the components of demographic change in Australia. In Section 3, we describe the projection model and the data inputs. In Section 4, we outline the approach adopted to formulate our scenarios, including a survey of expert opinion, a targeted literature review and an assessment of historical equivalences. In Section 5, we report the results of our projections for Australia and its states and territories from 2019 out to 2040. In Section 6, we provide some discussion of our findings and recommendations for the on-going monitoring of population trends in the post-COVID-19 world. 2 Background 2.1 Trends in Australian population growth Prior to the emergence of COVID-19, Australia had one of the fastest growing populations in the Organisation for Economic Co-operation and Development (OECD), with an annual growth rate of 1.5 per cent in 2018, behind only Iceland, Luxembourg and Israel (World Bank, 2020). As of the June 30, 2019, Australia had an estimated resident population of 25.4 million people. The population had increased by more than 380,000 from the year prior, with approximately two-thirds of this growth due to Net Overseas Migration (NOM). NOM has been the main contributor to Australia's population growth since the mid-2000s (Fig. 1 ), averaging 218,310 per annum between 2010 and 2019, equivalent to 59% of annual growth (ABS, 2020b). To be counted in NOM, a person must have spent 12 months in Australia during any 16 month period. As a consequence, NOM includes individuals on permanent and temporary visas, as well as returning and departing Australian citizens. In recent years, temporary migrants, most notably international students, have been the largest single contributor, with an average contribution of 105,843 to NOM over the financial years 2017–18 and 2018–19, out of a total of 247,057 (ABS, 2020b). By contrast, natural increase (the excess of births over deaths) contributed roughly 154,455 people per annum to Australia's population between 2008 and 2018 (ABS, 2019a).Fig. 1 Components of population growth, Australia 2000–2018 Fig. 1Source: ABS, 2019a. There have been differential impacts of population growth across Australian States and Territories driven in large part by the concentration of immigrants in the capital cities of Sydney and Melbourne, as well as an internal migration system that relocates population away from New South Wales and into Queensland and Victoria (Charles-Edwards, Bell, Cooper, & Bernard, 2018). Spatial differentials in fertility (ABS, 2019c) and mortality (ABS, 2019d) are also observed. In 2018, the highest fertility was observed in the Northern Territory (TFR = 2.03) while the lowest was observed in the Australian Capital Territory (TFR = 1.55) (ABS, 2019c). In 2016–18, the Northern Territory recorded the lowest life expectancy in Australia of 75.5 for males and 80.2 for females. Victoria recorded the highest life expectancy of 81.7 for males and 85.3 for females (ABS, 2019d). 2.2 Australia's responses to COVID-19 and its impacts on population components 2.2.1 Net overseas migration (NOM) Australia's first recorded case of COVID-19 was a passenger who flew into Melbourne on January 19, 2020 from Guangzhou, China. The Australian government responded by denying entry to anyone who had left or transited mainland China in the previous 14 days from February 1, 2020. An exception was made for Australian citizens and permanent residents who were placed in mandatory 14-day quarantine following their return. This travel ban coincided with one of the busiest months for international student arrivals, which fell by 34 per cent compared to the same month in the previous year (ABS, 2020a). Travel bans were later imposed for Iran (1st March), South Korea (5th March) and Italy (11th March), before a general travel ban was introduced on the 20th March, which banned travel to Australia for non-citizens and non-residents, and also restricted Australians from travelling overseas. The outcome of these unprecedented international border restrictions was a 99.3 per cent drop in overseas arrivals in April 2020 compared to the previous year (ABS, 2020a). In January 2021, international arrivals remained 99 per cent lower than in January 2020 (ABS, 2021b). At the time of writing international borders remain closed and are unlikely to open until the Australian adult population has been vaccinated, which was originally estimated to occur around October 2021 (Zagon, 2021), but has experienced a setback due to issues with the AstraZeneca COVID-19 vaccine (Remeikis, 2021). 2.2.2 Net interstate migration (NIM) Following the closure of the international border, Australian States and Territories introduced restrictions on interstate travel. Tasmania was the first to introduce border controls on March 19, 2020, followed by the Northern Territory, Western Australia and South Australia on 24th March, and then Queensland on 26th March. Controls varied between States and Territories with most involving a 14-day compulsory quarantine (Tasmania, Northern Territory, South Australia). Queensland and Western Australia enacted a ban on all non-resident visitors with exceptions for essential workers, on compassionate grounds, and new permanent residents. The two most populous states, New South Wales and Victoria, along with the Australian Capital Territory, initially kept their borders open, but did introduce strict travel restrictions for non-essential travel. In June 2020, Victoria experienced a surge of COVID-19 cases leading to an extended lockdown in Greater Melbourne. The lockdown ended in October 2020, after 112 days, making it one of the longest lockdowns in the world (BBC News, 2020). This led to the closure of the New South Wales border for the first time in 100 years (Khalil, 2020) as well as a travel bans from Victoria to other States and Territories. Since the end of 2020, there have been a few temporary State border closures in response to localised outbreaks and snap lockdowns, with Brisbane for example experiencing a three-day full lockdown in March 2021. These are expected to continue until the population is fully vaccinated. While these restrictions limit travel temporarily, with the exception of Western Australia, they do not prohibit moving to take up permanent residence in a new State or Territory. They do however restrict the ability for prospective migrants to take short-term trips for job interviews and other activities, which may serve as a precursor to migration. 2.2.3 Mortality and fertility As of the beginning of May 2021, Australia had recorded about 30,000 cases of COVID-19 and about 900 deaths (Australian Government Department of Health, 2021c). Due to Australia's track record in limiting community transmission through widespread testing and contact tracing, community lockdowns as well strict quarantine measures (Stanaway, Irwig, Teixeira-Pinto, & Bell, 2020), it is not expected that COVID-19 will have a population-level impact on mortality in Australia. The commencement of Australia's vaccination program in February 2021 (Australian Government Department of Health, 2021d) provides further reason to believe that significant excess mortality may be avoided. However, at the time of writing the Federal Government had abandoned its original target of all eligible Australians being offered their first COVID-19 vaccine by October 2021. Evidence of demographic response to shocks and economic recessions (Matysiak, Sobotka, & Vignoli, 2020) suggest that the pandemic is likely to negatively affect fertility. Indeed, research from Europe shows that COVID-19 has already affected fertility planning there, leading to a postponement of planned births (Aassve, Cavalli, Mencarini, Plach, & Bacci, 2020; Luppi et al., 2020b). While limited, evidence from Australia is mixed. Recent modelling of long-term fertility trend suggests that COVID-19 will exert a depressing effect on Australian fertility (McDonald, 2020), although there is evidence of a slight year-on-year increase in access to prenatal services between March and June 2020 (Moaven & Brown, 2021). 3 Projection model Population projections were created by a program which incorporates cohort-component models using directional inward and outward migration flows at national and State/Territory geographical scales. The projection system is based on a movement population accounts framework (Rees, 1984) and uses a series of linked bi-regional models in place of data-hungry fully multiregional models to handle internal migration (Wilson & Bell, 2004). A minor adjustment is made to ensure that internal in-migration summed over all subnational geographical areas exactly matches total internal out-migration. The program outputs projected populations by sex and single years of age from age 0 to 110, together with projected demographic components of change with the same age-sex detail, for Australia and each of the States and Territories for each year of the projection horizon. Overseas migration is modelled using immigration numbers and emigration rates, with an optional constraint to annual NOM assumptions, which was applied for this study. Interstate migration is projected via bi-regional in- and out-migration rates with an optional constraint to annual NIM numbers, which was also applied in this case. The heart of the projection system comprises a set of population accounting equations. At the State and Territory scale, the population aged a+1 at time t+1 is calculated as:(1) Ps,a+1i(t+1)=Ps,ai(t)−Ds,a→a+1i+Is,a→a+1i−Es,a→a+1i+IMs,a→a+1i−OMs,a→a+1i where P refers to population, i a State/Territory, t a point in time, s sex, a age group, a→a+1 the period-cohort aged a at time t and a+1 at time t+1, D deaths, I immigration, E emigration, IM interstate in-migration, and OM interstate out-migration. All demographic components in equation (1) cover the t to t+1 projection interval, but explicit labelling is omitted to avoid clutter. For newly-born infants, the start-of-interval population consists of the number of babies born during the t to t+1 projection interval. Each of the terms in equation (1) is obtained by multiplying occurrence/exposure demographic rates by populations at risk. A separate national scale projection is also calculated; State and Territory projections are made consistent with the national projections by constraining projected State/Territory births, deaths and overseas migration terms to those at the national scale each year.1 4 Projection assumptions Future trends in the demographic components of change are generally determined by mathematical modelling and/or the expert judgment of demographers (ONS & Statistics Canada, 2014). Given the unprecedented nature of COVID-19, we adopted a multi-strand approach to inform our assumptions. First, we undertook a review of the academic literature on demographic responses to shocks including, natural disasters, economic recessions and pandemics. The second component was a review of historical data in Australia to understand the impact of past shocks on the various components of demographic change, such as the recessions of 1982–83 and 1990–91 and the global financial crisis (2007–09) as well as contemporary guidance on the likely economic impact of COVID-19 (RBA, 2020). The third dimension was a survey of Australian demographers to elicit their views on the likely impact of COVID-19 on international and internal migration. The survey sought to gauge opinion on future levels of NOM by visa category as well as NIM for the financial years out to 2022–23, however the response rate was too low (5 responses out of 19 invitations) to use these to set assumptions directly. Reponses were instead used to inform our judgement on likely scenarios. 4.1 Overseas migration Overseas migration is the most volatile component of population change both in Australia and globally. Overseas migration, particularly labour migration flows, are pro-cyclical, declining in periods of recession (Tilly, 2011). Evidence suggests, however, that declines in overseas migration tend to be smaller than expected and are followed by a relatively quick return to ‘normal’ (Beets & Willekens, 2009; Dobson, Latham, & Salt, 2009). Labour migration can be affected by a reduction in demand for labour at the destination as well as the introduction of reactive migration policies during periods of recession, reducing immigration (Martin, 2009). At the same time, massive return migration of temporary workers tends not to occur, reducing the level of emigration, particularly from advanced economies (Dobson et al., 2009). Family migration is less sensitive to economic conditions, as are humanitarian flows (Papademetriou & Terrazas, 2009), however, resettlement processes may be slowed due to policy in receiving countries (Castles & Vezzoli, 2009). Student inflows are also sensitive to economic conditions and financial resources at origins and labour market opportunities at the destination, as many students support their studies through part-time work (Beets & Willekens, 2009; Papademetriou & Terrazas, 2009). In Australia, NOM has fluctuated in line with the economic cycle, with substantial declines in the years following major economic recessions (Fig. 2 ). Since 1970 there have been only three technical economic recessions in Australia, defined as two consecutive quarters of negative real GDP growth. This makes it difficult to statistically model the impact of recessions on NOM due to a lack of data points. This is exacerbated by changes in the definition of NOM which occurred in 2006 breaking the data series (ABS, 2012). Another factor complicating the modelling of NOM is the shift in Australia's migration program away from permanent settlement to temporary migration which accelerated in the first decade of this century (Hugo, Khoo, McDonald, & Voigt-Graf, 2003). Temporary migrants now account for approximately two-thirds of NOM, with international students, working holiday makers and New Zealand citizens, and short-term working visas the largest contributors. The shift to temporary migration adds a layer of complexity to NOM forecasting: while the number of permanent migration places is set annually by the federal government, temporary migration is demand driven (Wilson, 2017) and thus falls outsize the direct control of the Australian Migration and Humanitarian Programmes. Temporary flows are differentially influenced by a range of Australian and international factors. For overseas students, factors such as the economic conditions at their country of origin as well as perceived quality of Australian universities, quality of life in Australia and pathways to permanency play a major role (Rafi & Lewis, 2013). For flows of New Zealand migrants, relative labour market conditions in Australia and New Zealand are a major driver.Fig. 2 Net overseas migration, 1971–2019. Note: Since 1970 there have been only three technical economic recessions in Australia, defined as two consecutive quarters of negative real GDP growth. The Global Financial Crisis of 2007–09 led to an economic slowdown but not a recession. Fig. 2Source: ABS, 2019b. The survey of expert demographers asked participants for their views on the likely impact of COVID-19 on overseas migration to and from Australia for the financial years 2020–21, 2021–22, and 2022–23. Questions were asked about the possible future of immigration, emigration and NOM in 7 key visa/citizenship categories:• permanent residence visa holders (part of the Migration program) • permanent residence visa holders (part of Humanitarian Program) • international students • working holiday makers • Temporary Work Skilled visa holders • Australian citizens • New Zealand citizens. For each category we asked experts to estimate possible immigration, emigration and NOM numbers, and provide a brief summary of reasons for their responses. Undoubtedly this was a very challenging request to make, and many questions were answered with a range of values, and others were left unanswered. The overall finding was that a major drop in NOM was expected for the next 1–3 years by the survey participants. A number of respondents emphasised the importance of international student migration to future NOM. If students were able to return relatively soon (under some sort of safe passage arrangement for example), the impacts on NOM would be reduced. The latest government advice suggests that large scale return of international students is unlikely to occur before 2022 (Hare, 2021). The most recent estimates indicate that in the year to June 2020, NOM declined by 19.4 per cent (ABS, 2021a), reflecting the early effects of COVID-19. 4.2 Internal migration Australia is one of the most migratory nations in the world (Bell et al., 2015), with nearly 40 per cent of the population changing address every five years. However, historical data shows significant volatility, with drops in the annual interstate migration rate that coincide with economic downturns (Fig. 3 ). These were all followed by significant rebounds in interstate migration, with the exception of the global financial crisis of 2007–09, which did not lead to a recession in Australia. Internal migration is well-known to be pro-cyclical (Saks & Wozniak, 2011). The level of internal migration varies in tandem with the business cycle as migrants respond to the labour and housing markets. Evidence from the United States suggest that the cyclicality of internal migration is not driven by variations in the geographic dispersion of economic opportunities but a reduction in the net benefits of migrating overall. Empirically this is manifested in the negative association of internal migration rates with unemployment (van der Gaag & van Wissen, 2008) and regional inequalities (Alvarez, Bernard, & Lieske, 2021). In Australia, the unemployment reached 7.1 per cent in May 2020 (ABS, 2020e), a 2 percentage point increase in just three months and the highest level since October 2001. By March 2021, the unemployment rate had decreased to 5.6 per cent, a better performance than the Reserve Bank of Australia forecast (RBA, 2020), although the effect of the withdraw of COVID-19 employment support packages by the Federal Government at the end of March 2021 is yet to be reflected employment statistics.Fig. 3 Year-on-year change in interstate migration rates (%). Fig. 3Source:ABS, 2019b. These short-term variations have occurred in the context of a long-term decline in internal migration that started in the mid-1990s (Bell, Wilson, Charles-Edwards, & Ueffing, 2017). While explanations have long revolved around changes in the composition of population, recent evidence suggests that the effect of population ageing has been fully counteracted by an increase in the share of more migratory groups, lone-households, renters and tertiary-educated individuals in particular (Kalemba, Bernard, Charles-Edwards, & Corcoran, 2020). This means that the downward trend in internal migration is the result of a secular, behavioural change in migration behaviour, possibly caused by a substitution with tele-working (Cooke & Shuttleworth, 2017) and long-distance commuting (Brown, Champion, Coombes, & Wymer, 2015). A trend analysis of reason-specific migration intensities has showed that all reasons for migrating interstate in Australia have declined over time (Bernard et al., 2020), which supports the idea of ‘rootedness' or increased place attachment (Cooke, 2011). Alternatively, some individuals may simply be ‘stuck in place’ because they do not have the means to migrate in a context of stagnating wages (Foster, 2016). Some of these factors will continue to play out in the short-term and are likely to exert an additional downward pressure on interstate migration, even when state borders re-open. A possible offset to the expected decline in internal migration, at least in the short term, is an increased rate of return migration. Large scale return migration has been observed in India and parts of Latin America, as urban migrant workers return to rural villages (World Bank, 2020). In Australia, there is some evidence of return to the parental home among young people, however, the intensity and pattern of these movements are unclear. One study suggests the up to 331,000 young adults have returned to their parental home during COVID-19, however this is a relatively small sample and includes overseas returnees (Hassan, 2020). It is also unclear whether these represent permanent or temporary returns, with four per cent of Australian's reporting that one person had temporarily stayed in their household since the first of March due to COVID-19 (ABS, 2020c). The increase in work from home arrangements is another potential disrupter to the internal migration patterns. A number of media reports suggest that work from home arrangements may lead a resurgence in counter-urban flows (Goldlust, 2020), altering the pattern of internal migration in Australia. There is limited evidence for this at present. The survey of Australian demographers elicited opinion on the likely impact of COVID-19 on internal migration in Australia for the next three financial years. Respondents were uniform in predicting a large decline in interstate migration in the year 2020–21, with drops ranging from 50 to 80 per cent of flows from the previous financial year, but largely recovering to pre COVID-19 levels by 2022–23. Associated commentary suggested that while the long-term outlook was good, the pace and pattern over the coming years was dependent upon economic recovery and the relative economic performance of Australian States and Territories. The scenarios proposed by experts in the survey are extreme by historical standard. The largest ever decrease year-on-year in interstate migration in Australia was −12.4 per cent following the economic recession of the early 1980s. The declines following the economic recession of the early 1990s and the Global Financial Crisis were smaller, at −7 per cent and −6 per cent respectively. Given that the closure of internal borders did not prohibit internal migration but only prevented temporary interstate movement, a more moderate set of assumptions were adopted. This is supported by the most recent data released by the ABS which suggested a 10 per cent drop in the level of interstate migration in the third quarter of 2020 compared to the previous quarter (ABS, 2021c). 4.3 Fertility Fertility in Australia has been below replacement since the mid-1970s and has been steadily declining for the last decade (ABS, 2019a). In 2019 the Total Fertility Rate reached a new low of 1.66. However, this is still higher than many other Western industrialised countries (Gray & Evans, 2018). This trend is assumed to continue, with no recovery of fertility rates assumed in the near future (ABS, 2018). The COVID-19 pandemic will most likely further depress fertility in the short-term. Evidence of the pandemic's effect on fertility in Australia is currently limited due to the lag between conception and childbirth, and the lag between birth occurrences and the publication of statistics. When COVID-19 first struck Europe and lockdown measures were put in place, news speculated about a baby boom (Burke, 2020; Fordham, 2020; Graham-McLay, 2020). By now these thoughts have changed and more commentators expect a downturn in fertility (Bell, 2020; Murray, 2020). However, only one publication to date on the impact of COVID-19 on fertility has been published. It is based on a European Survey, considering fertility intentions. This study finds an overall intention to either delay or to forego having children due to COVID-19 (Luppi et al., 2020b). To formulate assumptions on fertility for this study, we researched the impact on fertility induced by pandemics and economic downturns and also considered the impact of a change in the composition of the population due to a reduction in international migration. Research shows various pandemics lead to a clear decline in fertility (Stone, 2020b). The decline in fertility rates triggered by pandemics is often the result of high mortality, mortality of potential parents as well as unborn children, or as in the case of the Zika virus due to pregnancy avoidance. In Australia, low case numbers and deaths due to COVID 19 mean that mortality is unlikely to influence fertility. Declines in fertility due to economic downturns are also well documented (Sobotka, Skirbekk, & Philipov, 2011). In Australia, national and international lockdown regulations led to a considerable number of people losing their jobs, reduced working hours or being stood down (ABS, 2020d). Even though the Australian government put some measures in place – free child care, a job-keeper allowance, lower hurdles for job seeker allowance (ABS, 2020d) – to ease the impact of COVID-19 restriction on the economy, most of these measures are temporary. The reduction of immigration to Australia is expected to have a limited impact as women born abroad and in Australia have similar fertility rates (ABS, 2019a). With a reduction in international migration the overall number of births in Australia will be affected due to smaller childbearing-age populations, though the influence of overseas migration on fertility rates is thought to be small. For this reason, the reduction in international migration is not considered in our fertility assumptions. 4.4 Mortality At the time of writing (May 2021), about 900 deaths had been attributed to COVID-19 (Australian Department of Australian Government Department of Health, 2021c). Recent analysis of Australian deaths data shows a small increase excess deaths in Australia during the first half of 2020, however it is not clear whether these are due to population growth or are in fact COVID-related (Bennett, 2020). Excess deaths are relatively few based on current data. The impacts of COVID-19 on mortality may reach beyond deaths directly attributable to COVID-19, including delays in elective surgery, suicide due to financial uncertainty and domestic violence. The available evidence at the time of writing suggests that the impact on mortality in Australia is likely to be minimal. 4.5 Headline assumptions For this study we developed scenarios based on the assumed demographic and economic recovery process from the COVID-19 pandemic in Australia. As of May 2021, many restrictions imposed across States and Territories have been lifted and Australia's vaccination program has commenced. However, recovery may not be straightforward. This is highlighted Australia's second wave of COVID-19 in June 2020 following the virtual elimination of community transmission following the first wave (Milne et al., 2020). According to the recovery process we assume the pandemic to have different levels of impact on population behaviour affecting demographic indicators. We set out five scenarios – one reference scenario assuming No Pandemic and four scenarios each considering a different speed of recovery and extent of impact. Table 1 summarises the key ‘headline’ demographic indicators of these four scenarios the national scale as well as the interstate migration assumptions.Table 1 Summary of headline projection assumptions for each scenario. Table 1 Scenario Shorter Moderate Spatial Shift Longer Reference: No Pandemic Fertility TFR dips to 1.65 in 2020–21 before recuperating to 1.75 in the mid-2020s and settling at a long-run value of 1.70 TFR falls to 1.55 in 2020–21 before recuperating to 1.75 in the late 2020s and settling at a long-run value of 1.70 TFR drops to 1.45 in 2020–21 and takes a decade to recover to long-run value of 1.70 Same as for Shorter Impact scenario Overseas migration NOM drops to 0 in 2020–21 and recovers by 2023–24 to the long-run value of 210,000 per annum NOM falls to −50,000 for 2020–21, increasing to a long-run value of 210,000 per annum by 2024–25 NOM falls to −100,000 for 2020–21 and recovers over many years reaching the long-run value of 210,000 per annum by 2028–29 NOM is 210,000 per annum from 2020–21 onwards Interstate migration Decline of 5% in 2020–21 then an increase in 2021–22 to +5% over 2018-19 levels then trend to 10-year average in 3 years Decline of 10% in 2020–21 then 5% in 2021–22 then return to 10-year average in 3 years Decline in 2020–21 pro-rated based on third quarter 2020 migration estimatesa then uniform 5% decline in 2021–22 then returning to 10-year trend in 3 years. Decline of 15% in 2020–21 then 25% in 2021–22 then return to 10-year trend after 5 years. NIM returning to 10-year average 5 years after jump off Mortality Life expectancy at birth continues long-run increases, reaching 88.8 years for females and 86.3 years for males by 2040–41. Same assumptions used for all scenarios a This means greater net losses for New South Wales and Victoria and corresponding gains for Queensland and Western Australia than the Moderate scenario. In the Moderate scenario the pandemic reduces fertility to a TFR of 1.55 in the 2020–21 financial year before a recovery over the next few years, followed by a minor increase which represents a modest tempo effect of previously delayed births. In all scenarios the assumed long-run TFR is 1.70, which is in line with current TFR assumptions of the official ABS population projections. In the Shorter impact scenario the economic impact of the pandemic is less severe than originally envisaged and quite short, and the TFR hardly falls before recovering. This situation corresponds to a better economic recovery than anticipated. In the Longer impact scenario, the serious and long-lasting economic consequences of the pandemic drives fertility sharply down to 1.45 before a slow recovery over many years. Net Overseas Migration in the Moderate scenario falls substantially to −50,000 in the 2020-21 year before recovering over the next few years and maintaining 210,000 per annum in the long-run. The Shorter scenario assumes international migration resumes quickly in late 2020 and early 2021, resulting in NOM of 0 in 2020–21 before quickly increasing over the following year This situation is conceivable in the event of mass vaccination by the end of 2021 followed by the re-opening of international borders. In contrast, the Longer scenario sees NOM plummet to −100,000 and take eight years to increase to the long-run level of 210,000 per annum, as a result of a slower economic recovery and protracted international border restrictions. In all scenarios life expectancy at birth is assumed to continue its long-run upward trajectory, and no mortality differentials between scenarios are included. At the State and Territory scale, both fertility and life expectancy at birth assumptions are tied to the national level assumptions. State/Territory TFRs are assumed to follow the national TFR assumption plus or minus average differentials of the last decade. Similarly, State/Territory life expectancies are tied to national life expectancy assumptions plus or minus average differentials of the last decade. National projections of immigration and emigration totals are distributed to States and Territories using immigration and emigration distributions of the last decade. Interstate migration assumptions in the Moderate, Shorter and Longer scenarios area limited to changes in the overall level of migration, with the pattern and age structure to remain unchanged. In the Shorter scenarios, the level of interstate migration is assumed to drop 5 per cent in 2020–2021, before rebounding +5 per cent on 2021–22. This assumes that, while some moves were postponed during the initial response to COVID-19, they are recovered in the following year. The Longer scenario has NIM dropping by 15 per cent in 2020–2021 and then by a further 25 per cent in 2020–21. This assumes that a significant level of foregone moves, due to deteriorating economic conditions. The decline in the Longer scenario is considerably smaller than proposed by the experts. A judgement was made to moderate the decline due to a number of factors: the first is the potential for a significant number returns to the parental home; the second is that many moves already in train would likely have occurred. The scale of the decline in the Longer scenario is still more than twice the maximum decline in Australian interstate migration Australia since the 1970s. A fourth scenario (Spatial Shift) was developed based on a single quarter of interstate migration data released by the ABS in February 2021 and reflects observed changes in pattern of interstate migration flows (ABS, 2021c), with larger net losses for New South Wales and Victoria and corresponding gains for Queensland and Western Australia than the Moderate scenario. 5 Results 5.1 National-level results In this section we present scenario results of the most important population measures from the projections. For each scenario we analyse total populations, population growth rates and population ageing for Australia and for States and Territories. Fig. 4 shows the total projected population of Australia from 2020 to 2040 for the five scenarios as well as annual average growth rates. The reference No Pandemic scenario has Australia's population reaching 27.6 million by 2025, 29.4 million by 2030 and 33.1 million by 2040. Totals from the No Pandemic scenario are lower than the ABS's current Series B projection which reaches 33.6 million by 2040. Based on the modelled scenarios, COVID-19 is expected to have a measurable and persistent impact on Australia's population. Under the Shorter, Moderate and Longer scenarios, Australia's population by 2040 is, respectively, 0.46 million, 1 million and 1.90 million lower than in the No Pandemic scenario (Table 2 ). The impacts of COVID-19 are felt most strongly in the short term with annual population growth dropping to 0.55 per cent in the Shorter scenario, 0.28 per cent in the Moderate scenario, and just 0.01 per cent in the Longer scenario. For historical context, Australia's annual population growth last dropped below the projected rates in 1916 at which time growth was negative (ABS, 2019b). The impacts on Australia’ growth rates are persistent for the Moderate and Longer scenarios, with annual growth not returning to No Pandemic levels until the late 2020s in the late 2030s for the Longer scenario. The boost in annual growth rates in the Moderate scenario in 2027–28 is due to the return to pre-COVID levels of migration and fertility.Fig. 4 Total population and average annual growth rates (%), Shorter, Moderate, Longer and No pandemic Scenarios, Australia, 2019–2040 Fig. 4 Table 2 Total population by scenario and difference with the No Pandemic scenario, selected years. Table 2Scenario 2020 2025 2030 2040 Total population Difference with the no pandemic scenario Total population Difference with the no pandemic scenario Total population Difference with the no pandemic scenario Total population Difference with the no pandemic scenario No pandemic 25,732,381 na 27,569,219 na 29,436,029 na 33,089,100 na Shorter 25,697,295 −35,087 27,251,897 −317,322 29,073,714 −362,315 32,627,382 −461,718 Moderate 25,697,295 −35,087 26,782,228 −786,991 28,568,497 −867,532 32,040,911 −1,048,188 Longer 25,697,295 −35,086 26,288,548 −1,280,671 27,840,447 −1,595,582 31,190,083 −1,899,017 5.2 Components of change The projected demographic components of change reveal which processes drive population change. These are illustrated for each scenario in Fig. 5 .Fig. 5 Components of population growth, Australia, 2019–2040. Note: The Spatial Shift scenario is not presented here separately as the assumptions on mortality, fertility and NOM are the same as for the Moderate scenario and have on a national level the same outcomes. Fig. 5 COVID-19 has the potential to exert a profound impact on the number of births in Australia over the next 30 years. Compared to the reference scenario, close to 860,000 fewer babies born if the Longer scenario comes to pass, about 420,000 less if we assume a Moderate recovery period and close to 180,000 fewer under a Short recovery scenario. Even though in each scenario the TFR converges to 1.7, the resulting number of births does not because of differences in the number of females of reproductive age. In the last projection period of 2040–41, there are still about 25,000 fewer babies born in the Longer scenario, about 15,000 fewer in the Moderate scenario and about 8,500 fewer in the Shorter scenario compared to the reference scenario. By affecting the size of successive birth cohorts, the demographic reach of COVID-19 is expected be long-lasting. There is little difference in the number of deaths between scenarios. This is largely due to the use of identical mortality assumptions across the scenarios. The annual number of deaths in each scenario increases from about 170,00 to about 230,000 over the projection horizon because of the rapid growth of population at the oldest ages where death rates are highest. Even though NOM is assumed to converge back to long-term trends after a maximum of 10 years in all scenarios, there is a substantial variation in total NOM across the projection horizon. In the Longer Scenario, overall NOM is over a million people lower compared to No Pandemic. The difference reduces in line with the speed of recovery. In the Moderate scenario, NOM is about 640,000 fewer and in the Shorter scenario is 280,000 fewer compared to the No Pandemic scenario, which highlights the strength of the demographic effect of COVID-19 through international border closures. Our results suggest the impact of COVID-19 on population ageing to be relatively modest at the national level. This reflects the fact that NOM impacts projections more than any other component of change. While migration does exert some influence on age structure, the magnitude is less than for fertility. Under the No Pandemic scenario the proportion of Australians aged 65 and over is expected to increase from 15.9 per cent in 2020 to 20.0 per cent by 2040, this is only one percentage point less in 2040 than in the Longer scenario. 5.3 State variation Due to differences in international and internal migration levels and pattern, the demographic impacts of COVID-19 will vary across Australian States and Territories. Fig. 6 illustrates the total population growth for States and Territories, while Fig. 7 reports differences between the Shorter, Moderate and Longer scenarios and the reference scenario. Most States and Territories will experience less population growth due to the pandemic out to 2040. The largest absolute impacts will be felt in the two most populous states, New South Wales and Victoria, reflecting the importance of NOM to growth in these two states, followed by Queensland and Western Australia. If the pandemic has the longer impact assumed in our scenarios, then the population of New South Wales will be 642,416 fewer at the end of the projection horizon compared to if the pandemic had not taken place, while in Victoria the population is 638,451 lower. For Queensland, the population is 266,175 lower under the Longer scenario, while in Western Australia it is 199,745.Fig. 6 Population totals, Shorter, Moderate, Longer and No pandemic Scenarios, Australian States and Territories, 2019–2040. Note: Population sizes differ so much between States and Territories that different scales for the y-axes needed to be used. Fig. 6 Fig. 7 Difference in population totals between Shorter, Moderate, Longer and No pandemic Scenarios, Australian States and Territories, 2019–2040. Fig. 7 In contrast, the absolute impacts of the Shorter scenario range from a difference of −178,313 in New South Wales relative to the reference scenario to a small positive value of +1,247 in the Northern Territory. In relative terms, the largest impact out to 2040 in the Longer Scenario is in Victoria, which will have a population 7.5 per cent smaller than in the No Pandemic scenario followed by New South Wales (6.7 per cent smaller). The relative impact will be much smaller in the Northern Territory and Tasmania, with the Longer scenario leading to populations 3.8 per cent and 3.5 per cent smaller respectively than under the No Pandemic scenario. The impact of COVID-19 on the distribution of Australia's population by State and Territory is relatively small under the current scenarios, with the Longer scenario leading a slight reduction in the relative share of Victoria's population in 2040 and a slight gain in Queensland (Table 3 ). This does however have a number of potential impacts including the relative distribution of tax revenue and the apportionment of seats in the Australian House of Representatives (Corr, 2016).Table 3 Distribution of population by States and Territories, 2019 and 2040, No pandemic, Shorter, Moderate, Spatial shift and Longer Scenarios (per cent). Table 3 2020 2040 No pandemic Shorter Moderate Spatial shift Longer New South Wales 31.8% 31.3% 31.2% 31.1% 31.1% 31.1% Victoria 26.1% 27.8% 27.5% 27.5% 27.4% 27.4% Queensland 20.1% 20.0% 20.2% 20.3% 20.3% 20.3% South Australia 6.9% 6.1% 6.1% 6.1% 6.1% 6.2% Western Australia 10.3% 10.5% 10.5% 10.5% 10.6% 10.5% Tasmania 2.1% 1.8% 1.8% 1.8% 1.8% 1.8% Northern Territory 1.0% 0.8% 0.9% 0.9% 0.9% 0.9% Australian Capital Territory 1.7% 1.7% 1.7% 1.7% 1.7% 1.7% Total 100% 100% 100% 100% 100% 100% The population in all states and territories will age under each of the scenarios (Fig. 8 ). Compared to the reference scenario, the population of Western Australia will age more under the Longer scenario, with the percentage of the population aged 65 and over 1.5 percentage points higher by 2040 than in the No Pandemic scenario. The impact on ageing is smallest in the Australian Capital Territory (0.86 percentage point higher). This reflects the relatively small contribution of net overseas migration (the component of change most affected by the pandemic) to population change in that jurisdiction.Fig. 8 Population aged 65+ (%), Shorter, Moderate, Longer and No pandemic Scenarios, Australia, 2019–2040. Fig. 8 6 Discussion and conclusion The impacts of COVID-19 have been substantial and pervasive and will ultimately be reflected in the demography of national and sub-national populations. While in early stages, much of the focus of demographers has been on mortality and morbidity, attention is now turning to other components of demographic change, and ultimately the cumulative impact of COVID -19 on population size, structure and distributions. Population projections are the obvious and appropriate tool to scope these impacts. In normal times, projections are informed by recent data, extrapolative models of fertility, mortality and migration, as well as expert views as to likely demographic futures. Creating projections in a period of unprecedented global upheaval, with limited data and the potential for long lasting changes in demographic behaviour, is fraught. However there is a clear need to try and quantify the potential impact, which we attempt here via a number of scenarios. To formulate our assumptions, we synthesised academic research on the impact of external shocks on the various components of demographic change, used past data to establish meaningful limits and a survey of experts to garner a range of views. These were informed by the economic forecasts produced by the Reserve Bank of Australia and framed as Shorter, Moderate, Spatial Shift and Longer. Ultimately, the scenarios represent our informed view as to plausible futures for Australia's population and thus will be subject to error, especially in view of the pandemic still ongoing, no officially specified end in sight of an end to international travel restrictions and Australia's vaccination program still in early stages at the time of writing. However, we have to start planning for life after the pandemic now and decision makers need data to start this process sooner rather than later. The modelled scenarios suggest the potential for significant and long-lasting impacts on the size of Australia's population, but a relatively modest impact on population ageing (see Wilson, Temple, & Charles-Edwards, 2021 for a fuller discussion). The scale of the expected impact differs across Australia's States and Territories, with New South Wales and Victoria experiencing the greatest decrease in population in absolute terms, while Victoria is most impacted in relative terms. The impact on the spatial distribution of the Australia's population across state and territories is minor under the current scenarios. If shifts in internal migration patterns emerge due to differences in the relative economic recovery of states and territories or other yet unknown factors, greater differences in spatial patterns may occur. Broader questions emerge from this exercise with respect to how population projections are created and used in response to shocks and crises. In this study we focused on the impact of COVID-19. However, over the same period Australia was subject to other environmental, political and economic forces likely to impact the components of demographic change. These include Australia's extreme bushfire season of 2019–2020 (Bradstock et al., 2021), potentially impacting internal migration flows, and rising tensions with China which could affect international student flows in the coming years (Foster, 2021). There is clear value in developing more complex scenarios incorporating a fuller range of influences, although this falls outside the scope of the current study (Amran, 2019; Rees et al., 2010). Population projections inevitably contain a degree of error, which propagates as the projection timeline increases and the size of the area decreases. Probabilistic approaches have increasingly been used to quantify the degree of uncertainty in demographic futures, however, when uncertainty is so high bounds may be less useful. Scenario approaches may provide more targeted ranges, however, they still are only as good as their input data, which has the potential to shift rapidly in response to policies such as large scale border closures. Perhaps a more fruitful approach would be a dynamic projection system which produces regular and high frequency outputs. This requires significant investment in both statistical infrastructure but also in academic research into the nowcasting of demographic events. There are significant opportunities in this space, as recent research on digital demography has shown (Cesare, Lee, McCormick, Spiro, & Zagheni, 2018), however there is some way to go before leading indicators of demographic change are identified and validated. This would allow for a shift in the conceptualization of projections from looking backwards to look forwards, to a contemporary picture of plausible population futures. Author statements Elin Charles-Edwards: Conceptualization, Data curation; Formal analysis, Writing - Original Draft; Writing - Review & Editing; Supervision. Tom Wilson: Conceptualization, Data curation; Formal analysis, Writing - Original Draft; Writing - Review & Editing. Aude Bernard: Conceptualization, Writing - Original Draft; Writing - Review & Editing. Pia Wohland: Conceptualization, Writing - Original Draft; Writing - Review & Editing. 1 The impact of this top-down constraining is relatively minor. By the end of the projection horizon, constrained State and Territory Total Fertility Rates differed by 0.01–0.07 from the input data, with the exception of the Northern Territory where the difference was 0.12. For mortality, constrained life expectancy at birth for both males and females never differed by more than 0.1 years for any State or Territory. ==== Refs References Aassve A. Cavalli N. Mencarini L. Plach S. Bacci M.L. The COVID-19 pandemic and human fertility Birth trends in response to the pandemic will vary according to socioeconomic conditions Science 369 6502 2020 370 371 10.1126/science.abc9520 32703862 ABS 3412.0 - migration, Australia, 2010-11, explanatory notes Australian Buerau of statistics, Canberra 2012 https://www.abs.gov.au/ausstats/abs@.nsf/lookup/3412.0explanatory+notes12010-11 ABS 3222.0 - population projections, Australia, 2017 (base) - 2066 2018 https://www.abs.gov.au/AUSSTATS/abs@.nsf/Lookup/3222.0Main+Features12017%20(base)%20-%202066?OpenDocument ABS 3101.0 Australian demographic statistics 2019 Australian Bureau of Statistics Canberra https://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/3101.0Jun%202019 ABS 3105.0.65.001 Australian historical population statistics 2019 Australian Bureau of Statistics Canberra https://www.abs.gov.au/statistics/people/population/historical-population/latest-release ABS 3301.0 births, Australia, 2018 2019 Australian Bureau of Statistics Canberra ABS 3302.0.55.001 life Tables, states, territories and Australia, 2016-2018 2019 Australian Bureau of Statistics Canberra ABS 3401.0 overseas arrivals and Departures, Australia, Feb 2020 2020 Australian Bureau of Statistics Canberra https://www.abs.gov.au/statistics/industry/tourism-and-transport/overseas-arrivals-and-departures-australia/latest-release ABS 3412.0 migration, Australia, 2018-19 2020 Australian Bureau of Statistics Canberra ABS 4940.0 household impacts of COVID-19 survey, 12-15 may 2020 2020 Australian Bureau of Statistics Canberra https://www.abs.gov.au/statistics/people/people-and-communities/household-impacts-covid-19-survey/latest-release ABS 6202.0 - labour force, Australia, may 2020 2020 https://www.abs.gov.au/AUSSTATS/abs@.nsf/Lookup/6202.0Main+Features1May%202020?OpenDocument ABS 6202.0 labour force 2020 Australian Bureau of Statistics Canberra ABS Migration, Australia A. B. o. Statistics 2021 Canberra https://www.abs.gov.au/statistics/people/population/migration-australia/latest-release#net-overseas-migration ABS Statistics A.B.o. Overseas arrivals and Departures, Australia Vol. 2021 2021 Canberra https://www.abs.gov.au/statistics/industry/tourism-and-transport/overseas-arrivals-and-departures-australia/latest-release ABS Regional internal migration estimates, provisional -Provisional statistics on internal migration at state/territory and capital city levels Vol. 2021 2021 https://www.abs.gov.au/statistics/people/population/regional-internal-migration-estimates-provisional/latest-release Alvarez M. Bernard A. Lieske S.N. Understanding internal migration trends in OECD countries Population, Space and Place 2021 e2451 10.1002/psp.2451 n/a (n/a), Amran G.B.A. Conte A. Gailey N. Ghio D. Grapsa E. Jensen K. W L. Demographic scenarios for the EU - migration, population and education 2019 Publication Office of the European Union Luxembourg 10.2760/590301 Australian Government Department of Health Coronavirus (COVID-19) advice for international travellers Vol. 2021 2021 https://www.health.gov.au/news/health-alerts/novel-coronavirus-2019-ncov-health-alert/coronavirus-covid-19-restrictions/coronavirus-covid-19-advice-for-international-travellers Australian Government Department of Health Coronavirus (COVID-19) current situation and case numbers Vol. 2021 2021 https://www.health.gov.au/news/health-alerts/novel-coronavirus-2019-ncov-health-alert/coronavirus-covid-19-current-situation-and-case-numbers Australian Government Department of Health First COVID-19 vaccinations in Australia 2021 https://www.health.gov.au/news/first-covid-19-vaccinations-in-australia Balbo N. Kashnitsky I. Melegaro A. Meslé F. Mills M.C. de Valk H. Demography and the coronavirus pandemic Population and policy brief Vol. 25 2020 Population Europe Banerjee A. Pasea L. Harris S. Gonzalez-Izquierdo A. Torralbo A. Shallcross L. Estimating excess 1-year mortality associated with the COVID-19 pandemic according to underlying conditions and age: A population-based cohort study The Lancet 395 10238 2020 1715 1725 10.1016/S0140-6736(20)30854-0 BBC News Covid in Australia: Melbourne to exit 112-day lockdown BBC 2021 2020 https://www.bbc.com/news/world-australia-54686812 Beets G. Willekens F. The global economic crisis and international migration: An uncertain outlook Vienna Yearbook of Population Research 7 2009 19 37 www.jstor.org/stable/23025522 Bell T. No baby boom in the time of corona The Guardian 2020 https://www.theguardian.com/commentisfree/2020/jun/14/no-baby-boom-in-the-time-of-corona Bell M. Charles-Edwards E. Ueffing P. Stillwell J. Kupiszewski M. Kupiszewska D. Internal migration and development: Comparing migration intensities around the world 41 1 2015 33 58 10.1111/j.1728-4457.2015.00025.x Bell M. Wilson T. Charles-Edwards E. Ueffing P. Australia: The long-run decline in internal migration intensities Champion T. Cooke T.J. Shuttleworth I. Internal migration in the developed world: Are we becoming less mobile? 2017 Routledge Abingdon, United Kingdom 147 172 10.4324/9781315589282-7 Bennett C. Have there been uncounted coronavirus deaths in Australia? We can't say for sure, but the latest ABS data holds clues in the Conversation 2020 https://theconversation.com/have-there-been-uncounted-coronavirus-deaths-in-australia-we-cant-say-for-sure-but-the-latest-abs-data-holds-clues-141363 Bernard A. Charles-Ewdards E. Alvarez M. Wohland P. Loginova J. Kalemba S. Anticipating the impact of COVID-19 on international migraiton 2020 Centre for Populaiton, Federal Department of Treasury Canberra, Australia Bradstock R. Clarke H. Collins L. Clarke M. Nolan R.H. Penman T. A staggering 1.8 million hectares burned in ‘high-severity’ fires during Australia's Black Summer The Conversation 2021 https://theconversation.com/a-staggering-1-8-million-hectares-burned-in-high-severity-fires-during-australias-black-summer-157883 Brown D.L. Champion T. Coombes M. Wymer C. The Migration-commuting nexus in rural England. A longitudinal analysis Journal of Rural Studies 41 2015 118 128 10.1016/j.jrurstud.2015.06.005 Burke D. Coronavirus: 2020 baby boom predicted as world goes into lockdown Worksop Guardian 2020 https://www.worksopguardian.co.uk/health/coronavirus/coronavirus-2020-baby-boom-predicted-world-goes-lockdown-2503886 Castles S. Vezzoli S. The global economic crisis and migration: Temporary interruption or structural change? Paradigms 2 2009 68 75 Cesare N. Lee H. McCormick T. Spiro E. Zagheni E. Promises and pitfalls of using digital traces for demographic research Demography 55 5 2018 1979 1999 10.1007/s13524-018-0715-2 30276667 Charles-Edwards E. Bell M. Cooper J. Bernard A. Population shift: Understanding internal migration in Australia 2071.0 - Census of population and housing: Reflecting Australia - Stories from the Census, 2016 2018 Australian Bureau of Statistics Canberra Cooke T.J. It is not just the economy: Declining migration and the rise of secular rootedness 17 3 2011 193 203 10.1002/psp.670 Cooke T.J. Shuttleworth I. Migration and the Internet Migration Letters 14 3 2017 331 342 10.33182/ml.v14i3.347 Corr P. Democracy and dollars: Use of demography in distribution Wilson T. Charles-Edwards E. Bell M. Demography for planning and policy: Australian case studies 2016 Springer Cham CSSE COVID-19 Dashboard in. Center for systems Science and Engineering (CSSE) 2020 Johns Hopkins University Baltimore, Maryland Dobson J. Latham A. Salt J. On the move? Labour migration in times of recession 2009 Policy Network. London Ferguson N. Laydon D. Nedjati-Gilani G. Imai N. Ainslie K. Baguelin M. Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand 2020 Imperial College London London, UK 10.25561/77482 Fordham E. Will coronavirus cause a baby boom? FOXBusiness 2020 https://www.foxbusiness.com/lifestyle/coronavirus-baby-boom-covid-19-1 Foster T.B. Rooted or stuck? The Causes and consequences of American Mobility decline 2016 University of Washington Washington State Foster A. China takes aim at Australian universities with latest boycott news.com.au 2021 https://www.news.com.au/finance/business/other-industries/china-takes-aim-at-australian-universities-with-latest-boycott/news-story/7ac39c1990c7fa44b98ddaf36e6a5aa3 van der Gaag N. van Wissen L. Economic determinants of internal migration rates: A comparison across five european countries Tijdschrift voor Economische en Sociale Geografie 99 2 2008 209 222 10.1111/j.1467-9663.2008.00454.x Goldlust R. Is there a rush of tree changers on the way? The Canberra times, Canberra 2020 https://www.canberratimes.com.au/story/6801494/is-there-a-rush-of-tree-changers-on-the-way/#gsc.tab=0 Goldstein J.R. Lee R.D. Demographic Perspectives on mortality of Covid-19 and other Epidemics in NBER working paper No. 27043 2020 National Bureau of Economic Research Cambridge, MA Graham-McLay C. Sex toy sales triple during New Zealand's coronavirus lockdown The Guardian 2020 https://www.theguardian.com/world/2020/apr/09/sex-toy-sales-triple-during-new-zealands-coronavirus-lockdown Gray E. Evans A. Geographic variation in parity progression in Australia Population, Space and Place 24 2 2018 e2080 10.1002/psp.2080 Hare J. International students won't return until 2022 Financial review 2021 https://www.afr.com/work-and-careers/education/international-students-won-t-return-until-2022-20210330-p57fau Hassan B. “Boomeranging” generation moves home due to COVID-19 Finder 2020 https://www.finder.com.au/press-release-may-2020-boomeranging-generation-moves-home-due-to-covid-19 Henley J. Lockdowns return or are extended as third wave of Covid sweeps Europe The Guardian 2021 https://www.theguardian.com/world/2021/mar/22/lockdowns-return-extended-third-wave-covid-europe Hugo G. Khoo S.-E. McDonald P. Voigt-Graf C. Temporary skilled migration to Australia: The 457 visa sub-class People and Place 11 4 2003 27 40 Kalemba S.V. Bernard A. Charles-Edwards E. Corcoran J. Decline in internal migration levels in Australia: Compositional or behavioural effect? 2020 e2341 10.1002/psp.2341 Khalil S. Coronavirus: Australia to close Victoria-New South Wales border British Broadcasting Company 2020 https://www.bbc.com/news/world-australia-53303317 Kontis V. Bennett J.E. Rashid T. Parks R.M. Pearson-Stuttard J. Guillot M. Magnitude, demographics and dynamics of the effect of the first wave of the COVID-19 pandemic on all-cause mortality in 21 industrialized countries Nature Medicine 26 12 2020 10.1038/s41591-020-1112-0 Levy B.L.M.T. Daniel Perez A. Why did people move during the great recession? The Role of Economics in Migration Decisions. The Russell Sage Foundation Journal of the Social Sciences 3 3 2017 100 125 10.7758/rsf.2017.3.3.05 28547003 Luppi F. Arpino B. Rosina A. The impact of COVID-19 on fertility plans in Italy, Germany, France, Spain and UK SocArXiv 2020 10.31235/osf.io/wr9jb May 22 Martin P. Recession and Migration: A New Era for Labor Migration? Vol. 43 2009 671 691 10.1111/j.1747-7379.2009.00781.x 3 Matysiak A. Sobotka T. Vignoli D. The great recession and fertility in Europe: A sub-national analysis European Journal of Population 2020 10.1007/s10680-020-09556-y Milne G. Xie S. Poklepovich D. O'Halloran D. Yap M. Whyatt D. Effectiveness of second wave COVID-19 response Strategies in Australia 2020 10.1101/2020.11.16.20232843 medRxiv, 2020.2011.2016.20232843 Moaven L. Brown J. COVID-19 “baby boom” Medical Journal of Australia 2021 10.5694/mja2.51010 n/a (n/a) Murray A. A COVID-19 baby boom? More like a baby bust C. G. T. Network Vol. 2020 2020 https://newseu.cgtn.com/news/2020-06-20/A-COVID-19-baby-boom-More-like-a-baby-bust-RsvSpJsKYM/index.html ONS, & Statistics Canada In-depth review of population projections Conference of European Statisticians 2014 United Nations Economic Commission for Europe Geneva, Switzerland Papademetriou D. Terrazas A. Immigrants and the current economic crisis: Research evidence, policy challenges, and implications Migration policy Institute, Washington DC 2009 Rafi B. Lewis L. Indian higher education students in Australia: Their patterns and motivations Australian Journal of Education 57 2 2013 157 173 10.1177/0004944113485837 RBA Statement on Monetary policy, may 2020 Reserve Bank of Australia, Sydney 2020 https://www.rba.gov.au/publications/smp/2020/may/ Rees P. Spatial population analysis using movement data and accounting methods: Theory, models, the ‘MOVE’ program and examples Working paper 404/Computer manual 23 1984 University of Leeds UK Rees P. Boden P. Dennett A. Stillwell J. Jasińska M. Jong A.d. Demographic and migratory flows affecting European regions and cities. Applied research project 2013/1/3. Annex D6 report on scenarios and a database of scenario drivers, Annex D7 regional population dynamics: A report assessing the effects of demographic developments on regional competitiveness and cohesion Programme T.E. Luxemburg 2010 https://www.espon.eu/demifer Remeikis A. Australia plans staggered reopening of international borders in second half of year The Guardian, Australia 2021 https://www.theguardian.com/australia-news/2021/apr/18/australia-plans-staggered-reopening-of-international-borders-in-second-half-of-year Saks R.E. Wozniak A. Labor reallocation over the business cycle: New evidence from internal migration Journal of Labor Economics 29 4 2011 697 739 10.1086/660772 Science Alert There's one really Shameful aspect of the global vaccine rollout, says UN Chief 2021 https://www.sciencealert.com/un-chief-criticizes-wealthy-countries-for-selfishly-stockpiling-vaccines-to-everyone-s-detriment Sobotka T. Skirbekk V. Philipov D. Economic recession and fertility in the developed world Population and Development Review 37 2 2011 267 306 www.jstor.org/stable/23043283 22066128 Stanaway F. Irwig L.M. Teixeira-Pinto A. Bell K.J. COVID-19: How many Australians might have died if we'd had an outbreak like that in England and Wales? Medical Journal of Australia 2020 (Preprint) Stone L. Short-run fertility responses to mortality events: A look to the past Applied Demography, Special Issue on COVID-19 2020 Stone L. Will the coronavirus spike births? Institute for Family studies (IFS) 2020 https://ifstudies.org/blog/will-the-coronavirus-spike-births Tilly C. The impact of the economic crisis on international migration: A review Work, Employment & Society 25 4 2011 675 692 10.1177/0950017011421799 Trias-Llimós S. Bilal U. Impact of the COVID-19 pandemic on life expectancy in Madrid (Spain) Journal of Public Health 2020 10.1093/pubmed/fdaa087 Wilson T. Can international migration forecasting be improved? The case of Australia Migration Letters 14 2 2017 285 299 10.33182/ml.v14i2.333 Wilson T. Bell M. Comparative empirical evaluations of internal migration models in subnational population projections Journal of Population Research 21 2 2004 127 160 10.1007/BF03031895 Wilson T. Temple J. Charles-Edwards E. Will the COVID-19 pandemic affect population ageing in Australia? Journal of Population Research 2021 1 15 World Bank COVID-19 crisis through a migration lens Migration and development brief Vol. 21 2020 (World Bank, Washington D.C.) Worldometers Worldometers Coronanvirus updates Vol. 2021 2021 Dover Delaware, U.S.A https://www.worldometers.info/coronavirus/#countries Zagon C. International travel could resume by October if Australia's vaccine rollout goes to plan 9News 2021 nine.com.au https://www.9news.com.au/national/coronavirus-international-travel-could-resume-by-october-australia-vaccine-rollout-plan/f8735248-bca0-4b7e-8c8b-3a651deb6fc6 Zambrano E. Bueermann G. Sullivan D. Travel restrictions, points of entry, and mobility data: Impact in COVID-19 data models and needed solutions for proximity, location and mobility data Migration policy Practice Vol. X 2020 60 64 2

0

PMC9753126

NO-CC CODE

2022-12-16 23:25:30

no

Appl Geogr. 2021 Sep 6; 134:102506

utf-8

Appl Geogr

10.1016/j.apgeog.2021.102506

oa_other

==== Front Appl Geogr Appl Geogr Applied Geography (Sevenoaks, England) 0143-6228 0143-6228 Elsevier Ltd. S0143-6228(21)00133-8 10.1016/j.apgeog.2021.102517 102517 Article COVID-19 exacerbates unequal food access Kar Armita a Motoyama Yasuyuki b Carrel Andre L. bc Miller Harvey J. ad Le Huyen T.K. a∗ a Department of Geography, The Ohio State University, Columbus, OH 43210, USA b Knowlton School of Architecture, City and Regional Planning Section, The Ohio State University, Columbus, OH 43210, USA c Department of Civil, Environmental and Geodetic Engineering, The Ohio State University, Columbus, OH 43210, USA d Center for Urban and Regional Analysis, The Ohio State University, Columbus, OH 43210, USA ∗ Corresponding author. 154 N Oval Mall, Columbus, OH 43210, USA. 9 7 2021 9 2021 9 7 2021 134 102517102517 9 10 2020 24 6 2021 5 7 2021 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Inequality to food access has always been a serious problem, yet it became even more critical during the COVID-19 pandemic, which exacerbated social inequality and reshaped essential travel. This study provides a holistic view of spatio-temporal changes in food access based on observed travel data for all grocery shopping trips in Columbus, Ohio, during and after the state-wide stay-at-home period. We estimated the decline and recovery patterns of store visits during the pandemic to identify the key socio-economic and built environment determinants of food shopping patterns. The results show a disparity: during the lockdown, store visits to dollar stores declined the least, while visits to big-box stores declined the most and recovered the fastest. Visits to stores in low-income areas experienced smaller changes even during the lockdown period. A higher percentage of low-income customers was associated with lower store visits during the lockdown period. Furthermore, stores with a higher percentage of white customers declined the least and recovered faster during the reopening phase. Our study improves the understanding of the impact of the COVID-19 crisis on food access disparities and business performance. It highlights the role of COVID-19 and similar disruptions on exposing underlying social problems in the US. Keywords Food desert Shopping travel COVID-19 Equity Economic development ==== Body pmc1 Introduction Grocery stores are a key destination in everyday travel. Access to healthy and fresh foods is tied to community health outcomes; ideally, all residents in a city should have access to grocery stores within reasonable travel distance and time. However, the disparity in food access is closely connected to systematic segregation and redlining, both for consumers and retailers (Bower et al., 2014; Reich, 2016; Shannon, 2020; Vargas, 2021). Food stores typically determine their store locations by targeting customers from specific socio-economic groups. For example, the number of dollar stores, mostly SNAP authorized, expanded by 62% in the United States over the last decades, mainly targeting the economically disadvantaged and racially diverse areas (Shannon, 2020). Consequently, the distribution of supermarkets and grocery stores largely varies by the pre-existing socio-economically segregated pattern of US neighborhoods. Low-income and racially diverse neighborhoods are often served only by low-priced discount stores with affordable but less healthy food products and lack supermarkets and grocery chains with healthier foods. This situation has given rise to a large body of literature measuring the impacts of this uneven distribution of grocery shopping opportunities, often framed around the concept of food deserts (Walker et al., 2010). According to this line of research, residents of food deserts have no choice other than to consume unhealthier food or to travel long distances to access supermarkets that carry a larger or healthier variety of foods. Underserved communities also often experience disparities in transportation due to transportation services being inadequate or unaffordable (Farber et al., 2016; Lucas, 2019) as a result of a long history of land use and zoning policies that prioritize automobile travel and single-family housing in suburbs. This further exacerbates the problem for households with limited food and transportation budgets and results in very different travel patterns of underserved populations. Thus, recent research has measured food access as a function of both the distance to food stores from residential areas and individual travel patterns and activity spaces (Li & Kim, 2020; Shannon, 2016; Widener et al., 2013). However, most studies measured food access at a single time point in time; only a handful of studies have considered longitudinal changes in food access and food shopping travel. This has resulted in a limited body of research that captures the impact of temporal variations in transportation service availability, transit schedules, lifestyle, or similar. The COVID-19 pandemic disrupted food shopping patterns in an unprecedented way and exacerbated existing problems with food deserts and transportation access. Not only did it cause profound economic hardships to disadvantaged populations and businesses, it also altered work and general travel patterns through limited out-of-home activities and substantial decreases in transit service supply. However, it is not well understood how the impacts differed across population segments, neighborhoods, and types of grocery stores. In light of the abruptness of the shifts brought about by COVID-19, it is critical to investigate the impacts on food access using data with a fine temporal resolution to capture changes in mobility patterns. The purpose of this study is to assess and investigate changes in travel to supermarkets and grocery stores due to the COVID-19 pandemic. We pursue three objectives: first, to quantify the impact of the pandemic on visits to different types of grocery stores throughout several phases of the pandemic; second, to investigate spatio-temporal changes in origin-destination travel patterns by different demographics; and third, to examine the influence of broader socio-economic and built environment factors on these changes by store type to account for price and customer segments. Our paper contributes to understanding food shopping access and travel patterns, especially during a disruptive time such as the COVID-19 pandemic. Our work is among the few studies in the food access domain that apply observed data on travel and food shopping behavior with fine temporal resolutions. Additionally, by using an aggregate, city-wide large sample of travel data, this study adds to a holistic understanding of food shopping travel and complements past studies that used observed individual travel data. In addition, our paper distinguishes types of stores based on scale (i.e., local vs. large merchandise stores, general vs. specialized stores) and price, accounting for market segmentation, and further shedding light on food access issues faced by low-income and racially diverse populations. Finally, the COVID-19 pandemic has revealed stresses, weaknesses, and social inequities in many of our social and economic systems; analysis of its impacts on food shopping travel for staple foods can help illuminate these in our food systems. 2 Literature review 2.1 Food deserts and social inequality A large number of studies concerning groceries and healthy food access have focused on either retail store locations or food deserts (Jiao et al., 2012; Shaw, 2006; Widener et al., 2011; Zenk et al., 2005), which the United States Department of Agriculture (USDA) defines as “low-income census tracts with a substantial number or share of residents with low levels of access to retail outlets selling healthy and affordable foods” (USDA, 2019). Past studies consistently found that inner cities with higher rates of poverty, unemployment, and vacant units tend to be food deserts (Dutko et al., 2012; Giang et al., 2008; Guy et al., 2004; Semple & Giguere, 2018). This happens largely because of the locations of grocery stores: larger supermarkets with a variety of food options tend to locate in suburban and wealthier areas, while smaller stores dominate in inner cities with low-income population (Chung & Myers, 1999; Moore & Diez Roux, 2006; Zenk et al., 2005). In addition, grocery stores tend to avoid the clusters of fast-food establishments and restaurants that are located in inner cities, further exacerbating the food desert issue (Leslie et al., 2012). As a result, minority neighborhoods in inner cities are often served with independent and small grocery stores, discount stores, and regional supermarkets to fill the gap by large chains (Doussard, 2013; LeDoux & Vojnovic, 2013; Raja et al., 2008). Nonetheless, reduced access to healthy foods for the low-income population is exacerbated by higher prices for food items charged by smaller stores (Johnson et al., 1996). Shannon (2020) maintained that retail redlining, the discriminatory practice that avoids locating grocery stores in disadvantaged neighborhoods, further pushed low-income and racially diverse populations to rely on low-price retailers (e.g., dollar stores) for food shopping. These locational patterns are associated with income level, race, and ethnicity. For instance, affluent black neighborhoods in Atlanta have lower access to food stores than white counterparts at the same income level (Helling & Sawicki, 2003). 2.2 The built environment and travel patterns for food shopping The assumption underlying the aforementioned studies is that people purchase food from nearby areas. Hence, areas of residence, such as census tracts or block groups, without healthy and affordable food options are considered disadvantaged. An alternative approach may consider the travel distance from store locations, such as 500m (or 1/3 of a mile) as the distance beyond which people will not walk (Wrigley et al., 2004). However, people's grocery travel patterns are not necessarily constrained by simple geographic units or boundaries of their residential areas, distance, or estimated travel time. Many people travel for food outside of their immediate neighborhoods, even when nearer grocers were available (Shannon & Christian, 2017; Zenk et al., 2011). These findings are in line with results by Li and Kim (2020), who argued that individual-level activity spaces were more relevant to food accessibility than the residential neighborhood. Therefore, using observed mobility data is critical for the analysis of food access to capture true travel patterns and look beyond the neighborhood level (Chen & Kwan, 2015; Christian, 2012; Shannon, 2016; Widener et al., 2013). Accessibility to multimodal transportation is another key determinant of travel for food shopping. Past literature demonstrated the differences in grocery travel and accessibility to supermarkets considering the availability of multimodal transportation modes (e.g., automobiles, transit, and walking) and its temporal variabilities (Farber et al., 2014; Widener et al., 2015, 2017). Even low-income residents with transportation disadvantages employ alternate travel strategies, including bus, carpool, or including food shopping as a part of trip chaining (Hallett & McDermott, 2011; Shannon, 2016; Ver Ploeg et al., 2015). Furthermore, recipients of food assistance programs, who are by definition low-income, traveled to stores that were approximately twice as far as the nearest major supermarket (LeDoux & Vojnovic, 2013; Ver Ploeg et al., 2015). Lastly, attributes of the built environment and an array of travelers’ characteristics have been found to be significant factors in food access (Ewing & Cervero, 2001). Accessibility to destinations, including shops, can be affected not only by locational attributes (e.g., store location, the road network, availability of other activities, property prices) and store attributes (e.g., store size and number of employees), but also by characteristics of travelers (e.g., personal preferences for travel, mode choices, physical and financial constraints) (Helbich et al., 2017; Miller, 2018). Widener et al. (2013) and Shannon and Christian (2017) suggested that trips for food shopping are commonly chained with work trips, making it necessary to consider food retail locations in proximity to jobs. 2.3 Limitations and emerging challenges in studying the impacts of COVID-19 While existing studies have provided some understanding of food accessibility and inequity, the absence of a temporal dimension and observed travel patterns is a major limitation when attempting to apply their findings to estimate the impacts of the COVID-19 pandemic. Specifically, most past studies focused on static analyses instead of changes in accessibility experienced by communities or people over time due to events such as store closures and neighborhood changes. The few studies that considered temporal changes were primarily regional analyses of store closures or changes over a long period. Guy et al. (2004) found that higher-income areas generally got more stores and choices, while lower-income areas faced store closures in Cardiff, UK, from 1989 to 2001. In a case study of Ypsilanti, Michigan, Semple and Giguere (2018) found that areas designated as ‘food deserts’ shifted over the course of 40 years from predominantly African American neighborhoods to both African American and low-income white neighborhoods. Perhaps the most insightful study was conducted by Shannon et al. (2018), who investigated changes in consumers' shopping behavior relative to the locations of stores accepting food assistance programs during the Great Recession from 2008 to 2012. A few recent studies investigated the impact of COVID-19 at a high level, i.e., by examining the general decline in business activity. However, they fell short of analyzing how the pandemic and changes in grocery travel patterns may be related to neighborhood or other socio-economic factors. Bartik et al. (2020) suggested that the economic impacts of COVID-19 varied by the scale of business, and small businesses were generally more affected. YelpInc (2020) suggested similar findings from analyzing store closures by store types only at the metropolitan level (e.g., type of restaurants and cuisines). The US Census Bureau (2021) has been conducting a weekly data collection effort, namely the Household Pulse Survey, regarding the level of food sufficiency in the previous seven days during this pandemic at the state level and large metropolitan areas. However, these studies do not inform accessibility and travel patterns at the intra-city level. In summary, a range of research has advanced the understanding of food accessibility, retail locations, travel behavior, and social inequity. Yet, it has also shown that it is critical to include observed travel behavior when studying food access and go beyond the simple food desert measure based on preidentified geographic units, such as Census tracts, or assumed travel distances. Moreover, temporal changes in food shopping travel patterns need to be considered using individual-level travel data, coupled with the built environment and socio-economic characteristics of neighborhoods. This will help address the current lack of knowledge about the impact of an economic shock on individual-level access to grocery shopping over time. 3 Methods 3.1 Study area and data Our study area is Franklin County, Ohio, within which the majority of the Columbus, Ohio metropolitan area is located. With a population of 1.3 million (US Census Bureau, 2019b), this area has a community and economic activities that are more diverse than most other areas in Ohio. We used four types of data in this study: (1) store locations and characteristics, (2) store visits, (3) characteristics of incoming travelers as inferred from their origins, and (4) local characteristics of store locations (at the census block group level). The focus of this study was to understand access to supermarkets and grocery stores that people primarily use for buying staple foods. The selection of stores was mainly based on identifying businesses with NAICS code 445110 that represents supermarkets and other grocery stores and excludes convenience stores, and this sub-sector shares 95 percent of NAICS 4451 grocery stores, according to County Business Patterns 2019 (US Census Bureau, 2019a). We further included large merchandise stores that sell groceries, namely, Walmart, Sam's Club, Costco, and Target. We enriched this store dataset using geographic locations of point-of-interest data from SafeGraph. Among the 438 stores in Franklin County, we successfully matched and retained 393 stores (90%) for this analysis. We also performed a manual inspection of the dataset to ensure that our dataset includes all major local grocery stores in Columbus. We categorized stores using the USDA classification as an outline, which includes warehouse stores, supercenters, supermarkets, chain stores, and other types (Cho & Volpe, 2017, p. 32) and considering approaches from past research (Leslie et al., 2012; Shannon, 2020; Shannon et al., 2018). We categorized grocery stores into four types: big-box grocery stores, mid/high-end grocery stores, dollar stores, and local stores based on their retailer brands, NAICS categorization, and store characteristics (employee size and sales volume). Big-box grocery stores are warehouse stores and supercenters that sell grocery products (selected from NAICS 452311), including Walmart, Sam's Club, and Costco. Mid/high-end grocery stores include supermarkets (selected from NAICS 445110) that are regional brands (retailer franchises expanded in multiple US states) with a larger employee size (25 or more) and sales volume ($5million or more), including Kroger, Target, Trader Joe's, Whole Foods Market, Giant Eagle, and Meijer. The third category, dollar stores, is designated as a single category of chain stores, including all stores of respective discount chains, namely Dollar General, Family Dollar, and Dollar Tree, as well as some independent discount stores (selected from NAICS code 452319). Although often considered a combined category of convenience stores, gas stations, and pharmacies in past research (Shannon et al., 2016, 2018), dollar stores sell staple foods such as grains, meat, fruits, vegetables, and dairy products. Thus they may replace supermarkets in socially disadvantaged neighborhoods (Kelloway, 2018; Shannon, 2020). We did not consider non-traditional food stores such as convenience stores as people prefer to visit supermarkets over these stores for buying staple foods. Especially in Columbus, around 87% (65 out of 75) of the convenience stores (NAICS code 445120) are located within gas stations that are less likely to serve the purpose of primary grocery shops. Finally, we categorized the local supermarkets, independent grocery stores, limited assortment supermarkets, superettes, and specialty food stores as local stores. Similar to mid/high-end grocery stores, this category also contains stores from NAICS code 445110 (Supermarkets and Other Grocery Stores). Unlike the mid/high-end grocery stores category, these stores tend to be smaller and operate within the region. Specifically, 85% of them have the employee size of less than 25 and sales volume of less than $5 million, which are below the thresholds used for mid/high-end groceries. A few local supermarkets are also included in this category as they are local to Columbus and have not expanded their business into any other regions. This category comprises small-scale, local grocery stores, local farmer's markets, meat shops, fish markets, and international grocery stores. In summary, our dataset includes 15 big-box stores, 88 mid/high-end grocery stores, 104 dollar stores, and 186 local stores. We provided the detailed store selection criteria in the Appendix (Table A1). We obtained data on characteristics of each store (e.g., employee size, sales volumes) from InfoGroup, 2019 dataset (InfoGroup, 2019). This dataset categorizes both big-box and mid/high-end stores based on the type of retail products (e.g., supermarket, gasoline stations, health care and vision, coffee shops). We only used the supermarket/supercenter section of the InfoGroup dataset to ensure that these stores sell grocery products. Fig. 1 illustrates the spatial distribution of different store types in Franklin County. Big-box stores are generally located in suburban areas, with a major concentration of such stores in western and northeastern Columbus. Although mid/high-end grocery stores seem to have a balanced distribution across Columbus, underserved areas located in the central, eastern, and southeastern Columbus do not have any nearby big-box or mid/high-end grocery stores that residents can easily access (Colombo et al., 2012). However, a large concentration of dollar stores and local stores can be found in these neighborhoods. Apart from these neighborhoods, the majority of remaining dollar stores and local stores are found to be clustered in relatively low-income areas (Colombo et al., 2012).Fig. 1 Spatial distribution of grocery stores in Franklin County, Ohio, categorized by store type. Fig. 1 We obtained weekly visitor counts at each store from SafeGraph (SafeGraph, 2020) for the time period from January 06 to June 01, 2020 and enriched the data with travelers’ origin and destination data from Streetlight (StreetLight Data, Inc., 2020). Streetlight uses mobile-phone-based locations and timestamps to gather trip information and aggregates origin-destination (OD) flow data using user-defined geographic boundaries. The origins are defined as the geographic area from which the device users started moving, and the destinations represent the geographic area where the device remains still for a certain period of time. In this study, we used the parcels of store locations as destinations and census block groups as origins. Although the origins could be home, work, or any locations, there is a high chance that most travelers performed home-based grocery shopping trips during the COVID-19 pandemic due to the restrictions on non-essential travel and more work-from-home opportunities, especially for well-off communities. The OD flow data indicate the volumes and origins of visitors from a particular census block group (O) to the parcels where the stores are located (D). Travelers’ income, race, and travel time are available for each OD pair. We estimated the average travel time to stores as the weighted average of the travel times of OD flows reaching that store using the following equation:WeightedaveragetraveltimetostoreD=∑i=1ntOiD∗VOiD∑i=1nVOiD where, tOiD = average travel time from origin i to the destination of interest VOiD = traffic volume from origin i to the destination of interest. We obtained data at the census block group level, such as population density, from the 2014–2018 American Community Survey 5-year estimates (US Census Bureau, 2020), and built environment characteristics from the Smart Location Database (Ramsey & Bell, 2014). Table 1 provides an overview of the explanatory variables used for this study.Table 1 List of explanatory variables. Table 1Variable name Description Level of measurement Unit Data source Reference Store characteristics Number of employees As estimated by InfoGroup Store Count Infogroup Bartik et al. (2020) and Miller (2018) Sales volume As estimated by InfoGroup Store Thousands of dollars per year Infogroup Bartik et al. (2020) Travelers' characteristics Average trip length Weighted average trip duration of OD flows to the store parcel Parcel Minutes StreetLight Ewing and Cervero (2001) and Miller (2018) Percentage household income less than $50k Mean percentage of travelers with a household income of less than $50k across the OD flows to the store parcel Parcel Percentage StreetLight Hallett and McDermott (2011) and Widener et al. (2015) Percentage of white travelers Mean percentage of white travelers across the OD flows to the store parcel Parcel Percentage StreetLight Moore & Diez Roux (2006) Local characteristics of store locations Population density Total population per square mile Census block group Number per square mile American Community Survey Ewing and Cervero (2001) Area type Categorized by activity density (total number of jobs and dwellings per acre): - Rural: activity density ≤ 0.5 - Suburban: activity density > 0.5 and < 6 - Urban: activity density ≥ 6 Census block group Category Smart Location Database Ewing and Cervero (2001) and Widener et al. (2015) Job density Total number of jobs per acre Census block group Number per acre Smart Location Database Miller (2018) and Widener et al. (2013) Road density Miles of road network per square mile Census block group Miles per square mile Smart Location Database Hallett and McDermott (2011) and Miller (2018) Multimodal road density Miles of road network with multimodal facilities per square mile Census block group Miles per square mile Smart Location Database Farber et al. (2014) and Widener et al. (2015) Intersection density Multimodal intersections with four or more road segments per square mile Census block group Number per square mile Smart Location Database Miller (2018) Low wage workers Percentage of low wage workers within a census block group Census block group Percentage Smart Location Database Miller (2018) We determined three study periods based on the observed travel patterns manifested in the data. Although the state-wide stay-at-home order was effective from March 22, 2020, our data show a decline in store visits starting from March 16. This early decline in food shopping travel may have been driven by the perception of risk, which is shaped by news and media in addition to the announcement of regulations such as the stay-at-home order. In early March 2020, the news focused on new measures such as halting in-person instruction at public universities and K-12 schools as well as upcoming stay-at-home orders implemented by local establishments and the situation in other states and countries. People started panic-buying and stockpiling grocery products, and foot traffic in retail stores declined even before the official lockdown. Similarly, regardless of the official reopening date for all retail businesses (May 12, 2020), our dataset showed that the lockdown effect started to dissipate after April 20, 2020, with visitor numbers to stores starting to increase. Therefore, we chose the date ranges that reflect changes in observed travel patterns rather than the presence of official orders. We defined three phases between January 06 to May 31, 2020: (1) the pre-lockdown phase between January 06, 2020 and March 15, 2020, (2) the lockdown phase between March 16, 2020 and April 19, 2020, and (3) the initial reopening phase between April 20, 2020 and May 31, 2020. It is worth noting that we labeled these three phases based on the store visit patterns exhibited in our data (Section 4.1). 3.2 Exploratory analysis: spatial and temporal changes in visitors and OD flows to food stores We performed a set of exploratory analyses to demonstrate the changes in the study area across time and space. First, we analyzed temporal changes in store visitor numbers for each store type, both in absolute and relative terms. The absolute changes represent the weekly average number of store visitors by store type. The relative changes represent the percentage of changes in the weekly average of store visitors relative to the first week of the study period (January 06 – January 12) by store type. Second, we visualized the spatial changes in the weekly numbers of visitors to stores. For each of the three phases, we calculated the average weekly visitors to each store. Then, we mapped the changes between the pre-lockdown and the lockdown phase, and between the pre-lockdown and the reopening phase. Changes are represented with proportional symbols, calculated relative to the highest and lowest value of the dataset. We also performed an OD flow analysis to identify changes in the distribution of origins of incoming traffic to store locations. For each OD pair, our analysis considered the average daily traffic originating from the origin block groups and destined to the parcels associated with the stores. 3.3 Modeling temporal changes in store visits at store locations Our variable of interest was the difference in average weekly store visits to each store between (1) the pre-lockdown and lockdown phases and (2) between the pre-lockdown and initial reopening phases. Many of the stores in our sample experienced decreases in store visits, but a small number of stores experienced increases. We excluded four stores that had no changes in store visits during this period and averaged zero or one visitors per week, which may be due to measurement errors. For the rest of the stores, we set positive values of changes to 0 and assigned the absolute values for negative changes in traffic. A hurdle model is a suitable statistical method for count data where the zero and non-zero values are generated by two different processes (Cameron & Trivedi, 2013). A binary logit model captures the first process that generates zero counts, when a threshold (hurdle) is not passed. A truncated negative binomial model captures the second process that generates positive counts when the hurdle is passed. In this study, we estimated two hurdle models to quantify the changes in store visits to determine influential factors affecting the changes. The first model quantifies the effects of COVID-19 during the initial lockdown period, and the second model quantifies the eventual recovery of store visits, immediately before and after statewide restrictions were lifted. Each hurdle model includes (1) a binary logit submodel that determines the probability that a store had a decline in store visits, and (2) a truncated negative binomial submodel that estimates the magnitude of traffic change, given that a store had a decrease in store visits. We controlled for various independent variables in the hurdle models, including store characteristics, visitors’ socio-economics, and characteristics of the store location as guided by relevant studies (Table 1). We included the various categories of store types to account for their different levels of traffic, store characteristics, and most importantly, to understand how low-income and racially diverse populations changed their travel behavior during this disruption. 4 Results 4.1 Summary of temporal changes in store traffic Fig. 2 illustrates the absolute and relative changes in average weekly traffic by store type for the study period. The average weekly traffic started to decline in the week beginning on March 16 and continued to decline until the week of April 13 (lockdown phase). Average weekly visitors started to increase from the week of April 20 onward (reopening phase). The average changes in store visits during the lockdown period were dominated by changes in big-box store visits and mid/high-end grocery store visits (Fig. 2a).Fig. 2 Absolute changes (a; top) and relative changes (b; bottom) in average weekly store visits, categorized by store types. Fig. 2 The visitor data for each store type are normalized to the number of visitors in the week of January 6. It shows that relatively speaking, the declining trend of store visitors for all store types has the same pattern during the lockdown phase (Fig. 2b). However, the growth of visitors beginning in the week of April 20 is higher for big-box grocery stores than other stores, especially during the time period of April 20 to May 10 (the end of the week of May 4). Table 2 summarizes the variables used for modeling, categorized by store type. The average employee size and sales volume are the highest for big-box grocery stores and the lowest for dollar stores. Most big-box grocery stores belong to block groups with lower population density (153 people/sq. mi), lower road density (11.9%), and lower percentages of low-wage workers (39%) than the other types of stores.Table 2 Descriptive statistics of the variables used in the change analysis. Table 2 Big-box grocery stores Mid/high-end grocery stores Dollar stores Local stores Mean S.D. Mean S.D. Mean S.D. Mean S.D. Store characteristics Employee size 251 121 120 85 10 11 15 35 Sales volume 39748 19178 26795 19656 1319 706 6696 19156 Local characteristics of store locations Number of stores in urban areas 3 45 43 84 Number of stores in suburban areas 12 42 59 98 Number of stores in rural areas 0 1 2 4 Job density 3.40 3.00 5.30 6.58 3.25 7.56 3.80 6.52 Percent low-wage workers 39% 12% 34% 13% 35% 16% 33% 16% Road density 11.90 3.72 14.43 6.37 14.77 6.26 15.70 6.13 Multimodal road density 2.37 1.48 2.91 1.61 2.61 1.66 2.79 1.90 4-way intersection density 2.87 2.13 5.70 7.73 6.39 9.39 8.31 12.91 Population density 153 121 260 309 276 215 309 217 Average weekly visitors Pre-lockdown 1040 262 344 253 80 48 47 48 Lockdown 759 203 254 203 55 34 32 36 Initial reopening 835 231 264 226 59 36 33 35 Change in average weekly visitors Lockdown vs. pre-lockdown −281 101 −90 86 −25 24 −14 20 Reopening vs. pre-lockdown −205 94 −80 79 −21 21 −14 20 Average trip length (minutes) Pre-lockdown 27 2 24 2 24 2 24 4 Lockdown 23 3 23 3 23 3 24 5 Initial reopening 24 3 23 4 23 4 23 4 Household income less than $50k Pre-lockdown 50.8% 8.0% 47.6% 10.4% 56.4% 10.6% 54.2% 12.7% Lockdown 52.6% 7.3% 48.8% 9.7% 57.6% 10.4% 55.1% 13.8% Initial reopening 52.2% 7.7% 48.5% 10.4% 57.4% 11.7% 55.2% 14.0% Percentage of white travelers Pre-lockdown 70.3% 8.9% 73.7% 10.3% 62.9% 14.0% 65.8% 16.5% Lockdown 68.1% 9.3% 72.2% 11.0% 61.7% 14.0% 65.6% 17.0% Initial reopening 68.8% 9.5% 72.6% 11.5% 61.9% 15.2% 65.7% 16.7% Percentage of travelers with no high school or high school degree Pre-lockdown 43.0% 9.7% 38.7% 10.8% 49.3% 10.9% 45.2% 12.2% Lockdown 43.9% 9.4% 39.5% 10.2% 49.9% 10.4% 46.2% 13.6% Initial reopening 43.0% 9.9% 39.3% 10.5% 50.1% 11.4% 46.6% 13.5% During the pre-lockdown phase, the average weekly visitors to big-box, mid/high-end, and dollar stores were 22.1, 7.3, and 1.7 times higher than the average weekly visitors to local stores, respectively. Although on average, all types of stores had a decline in visitors during the lockdown and reopening phases, the big-box grocery stores received 23.7 times and 25.3 times more average weekly visitors than local stores. There was little to no difference in customers’ travel time to all three store types: on average, people spent 23–27 min to access grocery stores. With regard to shoppers’ demographics before lockdown, the percentages of low-income travelers (i.e., with annual household incomes below $50,000) are lowest for mid/high-end grocery stores (47.6%) and highest for dollar stores (56.4%). The percentage of white customers was highest at mid/high-end grocery stores (73.7%) and lowest at dollar stores (62.9%). The socio-economic composition of visitors by store type did not drastically change across the three study periods. 4.2 Exploratory analysis of spatial and temporal changes in store visits Fig. 3 shows the changes in average weekly store visitors between phases. The change from the pre-lockdown phase to the lockdown phase, as reflected in the proportional symbols, is very low for dollar stores and local stores as compared to big-box grocery stores and mid/high-end grocery stores (Fig. 3a). During the reopening phase (Fig. 3b), the proportional change for mid/high-end, dollar, and local stores is greater as compared to the changes in the big-box grocery stores.Fig. 3 Changes in average weekly store visitors; a) between pre-lockdown and lockdown phase (top), and b) between pre-lockdown and initial reopening phase (bottom). Fig. 3 Fig. 4, Fig. 5 illustrate the absolute changes in OD flows to stores between the three phases. Darker lines represent larger declines. Between the pre-lockdown and the lockdown phases, the decline in visits to big-box and mid/high-end grocery stores from nearby origins was higher than that from distant origins (Fig. 4). Although big-box stores had the highest decline during the lockdown phase, they recovered faster than other types of stores, mostly from short-distance visits (Fig. 4b). For mid/high-end grocery stores, the decline in OD flow was more pronounced in the reopening phase from nearby origins and some distant origins, such as the northeastern parts of Columbus (Fig. 4d).Fig. 4 Changes in OD flows to big-box stores (top panel) and to mid/high-end grocery stores (bottom panel) between pre-lockdown and lockdown phase (a, c) and between pre-lockdown and initial reopening phase (b, d). Fig. 4 Fig. 5 Changes in OD flows to dollar stores (top panel) and to local stores (bottom panel) between pre-lockdown and lockdown phase (a, c) and between pre-lockdown and initial reopening phase (b, d). Fig. 5 There are distinctive patterns in OD flow changes among the store types. Neighborhoods without supermarkets (i.e., food deserts) had smaller changes in OD flows to dollar stores and local stores during both lockdown and initial reopening phases (Fig. 5). For other neighborhoods, most dollar stores experienced a higher decline in incoming flow during the lockdown phase; some of these declines persisted in the reopening phase (Fig. 5a and b). Similarly, the northwestern part of Columbus (west of downtown and the university district), containing mostly high-income neighborhoods, experienced a major decline in flows to the local stores that are in close proximity to mid/high-end grocery stores (Fig. 5c and d). 4.3 Modeling temporal changes in store visits Table 3 presents the estimation results of the hurdle models. The binary logit submodel estimates the probability of a store experiencing a decline in traffic during both the lockdown and initial reopening phases. The table reports the coefficient estimates, p values, and odds ratio (OR). The model results indicate that the probability that a store located in an urban area experienced a decline in traffic was 7.53 times and 7.28 times more than that of a store located in a rural area during the lockdown and initial reopening phases respectively. Job density around the store (at census block group level) significantly affected traffic changes during COVID-19 (ORlockdown = 0.94, ORreopening = 0.95, p < 0.05). Although the effects of travel time and income were statistically insignificant, we found that stores with higher percentages of white travelers were less likely to have a decline in store visits during the reopening phase (ORreopening = 0.02, p < 0.05).Table 3 Hurdle model results: decline in store visits through the lockdown and reopening phases. Table 3 Model 1: Decline between pre-lockdown and lockdown phases Model 2: Decline between pre-lockdown and initial reopening phases Binary logit submodel Coef. p value OR Coef. p value OR Intercept 13.51 0.99 7.4e+05 19.65 0.99 3.4E+08 Store types (ref: big-box grocery stores) Mid/high-end grocery stores −12.55 0.99 0.00 −14.78 0.99 0.00 Dollar stores −12.59 0.99 0.00 −15.56 0.99 0.00 Local stores −14.33 0.99 0.00 −16.61 0.99 0.00 Store characteristics Employee size 0.01 0.19 1.01 0.00 0.72 1.00 Travelers' characteristics Trip duration 0.01 0.82 1.01 −0.01 0.89 0.99 Low income population (%) 3.32 0.28 27.66 0.03 0.99 1.03 White population (%) −1.35 0.54 0.26 −4.15** 0.02 0.02 Local characteristics of store locations Suburban area (ref: rural area) 1.57 0.15 4.79 1.55* 0.09 4.70 Urban area (ref: rural area) 2.02* 0.09 7.53 1.99** 0.04 7.28 Population density 0.00 0.89 1.00 0.00 0.69 1.00 Job density −0.06** 0.04 0.94 −0.05* 0.07 0.95 Multimodal road density 0.08 0.61 1.08 0.17 0.20 1.19 4-way intersection density −0.02 0.34 0.98 −0.02 0.26 0.98 Low-income workers (%) −1.30 0.32 0.27 −1.03 0.36 0.36 Truncated negative binomial submodel Coef. p_value IRR Coef. p_value IRR Intercept 6.33*** 0.00 560.11 5.64*** 0.00 281.11 Store types (ref: big-box grocery stores) Mid/high- end grocery stores −1.02*** 0.00 0.36 −0.77** 0.01 0.47 Dollar stores −2.04*** 0.00 0.13 −1.99*** 0.00 0.14 Local stores −2.43*** 0.00 0.09 −2.29*** 0.00 0.10 Store characteristics Employee size 0.00 0.24 1.00 0.00 0.51 1.00 Travelers' characteristics Trip duration −0.01 0.77 0.99 −0.00 0.91 0.99 Low-income population (%) −2.28*** 0.01 0.10 −0.97 0.25 0.38 White population (%) −1.01* 0.08 0.37 −0.50 0.39 0.61 Local characteristics of store locations Suburban area (ref: rural area) 0.73 0.12 2.08 0.22 0.70 1.24 Urban area (ref: rural area) 0.92* 0.06 2.51 0.42 0.46 1.52 Population density 0.00 0.45 1.00 0.00 0.35 1.00 Job density 0.03 0.10 1.03 0.02 0.19 1.02 Multimodal road density 0.00 0.94 1.00 −0.03 0.55 0.98 4-way intersection density −0.01 0.23 0.99 −0.01 0.11 0.99 Low-income workers (%) 0.49 0.21 1.63 0.49 0.23 1.62 Log(theta) 0.25* 0.09 0.17* 0.09 Goodness-of-fit statistics Log-likelihood −1615 −1585 AIC 3292 3232 Significance codes: p value < 0.01: '∗∗∗', p value < 0.05: '∗∗', p value < 0.1: '∗'. Model 1: ntotal = 393 stores (ndecline in visitors = 354 stores; nincrease in visitors = 39 stores). Model 2: n = 393 stores (decline in store visitors: 341 stores, positive or no changes in store visitors 52 stores). The truncated negative binomial submodels account for the magnitude of traffic decline at the stores that experienced a decline during the periods of interest. Table 3 reports the coefficient estimates, p-values, and the incidence rate ratio (IRR). The results indicate that, compared to big-box grocery stores, the average decline in weekly visitors between the pre-lockdown and lockdown phase was 64% (202 visitors) for mid/high-end grocery stores, 87% (73 visitors) for dollar stores, and 92% (49 visitors) for local stores. Model 2 estimates an average decline of 281 weekly store visitors between the pre-lockdown and initial reopening phase, suggesting an overall recovering pattern in grocery stores. However, compared to big-box stores, the average decline in weekly visitors was 53% (132 visitors) for mid/high-end grocery stores, 84% (39 visitors) for dollar stores, and 90% (28 visitors) for local stores. During the lockdown phase, stores located in urban areas experienced a traffic decline 2.5 times higher than those located in rural areas. Also, both the income and racial profiles present a negative association with the magnitude of decline in store visitors. Store visits were likely to decline by a factor of 0.10 with each percentage increase in low-income population (p < 0.01) and to decline by a factor of 0.37 with each percentage increase in white travelers (p < 0.08). However, neither travelers’ characteristics nor local characteristics of stores show statistically significant effects on the decline in store visitors during the reopening phase. We performed a Moran's I test for spatial autocorrelation using the residuals from the hurdle models. We used bandwidths ranging from 500m–3000m with a 500m interval. For big-box grocery stores, mid/high-end grocery stores, and dollar stores, we found no indication of spatial autocorrelation (p > 0.05). This finding indicates that the residuals of both models are not spatially correlated and the observed traffic decline of each store is not influenced by the attributes of nearby stores. For the local stores, our tests indicate the presence of weak spatial autocorrelation with a bandwidth of 2500m (Moran's I = 0.102, p < 0.05 for the hurdle model between the pre-lockdown and lockdown phase; Moran's I = 0.126, p < 0.05 for the hurdle model between the pre-lockdown and reopening phase). The null spatial autocorrelation at a bandwidth of 2500m for local stores is plausible, as local and ethnic grocery stores tend to have small building footprints and customer capacity, at the same time they tend to cluster with local stores in certain areas of Columbus. In summary, we find that our hurdle model results do not need to further account for spatial effects between stores. 5 Discussion Our study investigated spatio-temporal changes in visitor traffic to various types of food stores in Columbus, Ohio, including supermarkets and big-box retailers, before and after the lockdown due to the COVID-19 pandemic. We categorized changes in store visitor numbers throughout three phases that were defined based on observed store visit patterns: the pre-lockdown phase, the lockdown phase, and the initial reopening phase. We observed that at the aggregate level, all types of food retailers experienced a decline in their weekly visitor numbers during the lockdown phase and did not reach their pre-lockdown levels by the end of May 2020. Our major findings are explained in the following paragraphs. First, we found that the decline and recovery of store visits to grocery stores varied by store type. While the decline during the lockdown was larger for big-box retailers and smaller for dollar stores, the recovery was faster for big-box retailers and slower for mid/high-end grocery stores and other types of stores during the initial reopening phase. This finding is consistent with previous economic studies (Bartik et al., 2020) that the magnitude of the economic impacts of COVID-19 was lower for large businesses than for small businesses. Second, during the reopening phase, big-box and mid/high-end grocery stores experienced a recovery of visitors from nearby locations but not of visitors who traveled long distances. In contrast, the visitor declines for dollar stores and local stores primarily took place in wealthier areas, while store visits in low-income areas had smaller changes, even during the lockdown period. One possibility is that some shoppers stocked up on household commodities from large stores and thus reduced their travel during the lockdown phase. Higher-income households may have been more likely to do so than their lower-income counterparts with access to smaller storage areas and smaller food shopping budgets. Third, we found a significant difference in customers’ socio-economic characteristics by store type and locational attributes. As indicated in the exploratory analysis, big-box and mid/high-end grocery stores received more high-income and white travelers than dollar stores and local stores. Meanwhile, our models indicated that while the percentage of low-income customers and the percentage of white customers were negatively associated with traffic declines during the lockdown, each percent increase in low-income customers had a smaller effect on traffic decline than each percent increase in white customers. Additionally, stores with a higher percentage of white customers were more likely to recover during the reopening phase. The link between store traffic and demographics, that is, dollar stores being dominated by low-income shoppers and big-box and mid/high-end stores being dominated by white and higher-income shoppers, may explain the more substantial decline and faster rebound of big-box and mid/high-end stores. Our findings that suggest a smaller magnitude of decline in stores visited by low-income and white populations also imply that access to supermarkets and its association with different neighborhood socio-economic characteristics is a major driver of changes in food shopping patterns. Similar to Moore & Diez Roux (2006) and Zenk et al. (2005), low-income neighborhoods in our study area also tended to lack access to nearby supermarkets. These limited choices for low-income neighborhoods led to a situation in which residents of such neighborhoods continued to travel to dollar stores and local stores for their grocery shopping needs. Furthermore, similar to the findings of LeDoux and Vojnovic (2013) and Ver Ploeg et al. (2015), our OD flow analysis confirmed that the majority of long-distance travel to supermarkets in the pre-lockdown phase originated from low-income areas. By and large, residents of such areas reduced long-distance grocery travel and still did not visit farther-away supermarkets during the reopening phase. Previous studies of Hallett and McDermott (2011) and Shannon (2016) emphasized the dependency of low-income people on multimodal travel options for food access, and Widener et al. (2013) and Shannon and Christian (2017) identified the association between food shopping and travel to work. The restrictions during COVID-19, which impacted the availability of multimodal transportation options and shut down many workplaces, may have imposed barriers on both long-distance travel and travel to work, and limited access to larger supermarkets for underserved low-income neighborhoods. Fourth, we found that the decline of store visitors was associated with locational attributes, namely urban status and job density. The probability of seeing a decline in visitors was higher in urban areas and areas with higher job density. The magnitude of decline was also higher in urban areas during the lockdown phase, but not during the initial reopening phase. In contrast, suburban areas were more likely to see a slower recovery of grocery store visits during the initial reopening phase, but the magnitude of decline was not statistically different from rural areas. Unlike many studies that relied on distance or travel time as a determinant to identify food deserts, we found travel time to be insignificant in our models. This implies that not only physical distance and thus exposure to nearby food retailers (Widener & Shannon, 2014) plays a role in shaping low-income populations’ food shopping travel patterns, but also other social dimensions, such as price and transportation accessibility (Shannon & Christian, 2017; Widener et al., 2013). The strengths of our study include the use of comprehensive data on observed travel patterns with a large and refined spatio-temporal scale. This allows us to infer shopping travel patterns for the population, thus complementing previous work that relies on a sample of individual travel data measured at one or a few time points (e.g., cross-sectional data or longitudinal data with a small number of waves). Our analysis also emphasizes changes in the travel patterns of multiple population segments, considering customers’ income and race, across different store types and store sizes. This provides a holistic view of disparities in food access from both store and household standpoints. This study has several limitations. First, the OD flow data was limited to parcel sizes, resulting in an underestimation of the area size of big-box and mid/high-end grocery stores, which may occupy more than one parcel, and overestimation of the size of dollar and local stores, which may share the same parcel with other businesses. Second, measurement errors exist, as store visitors may have shopped for non-grocery products. Besides, our study assumes that all stores considered in this study sell a variety of staple foods. However, some stores may carry a more limited collection of staple foods than others (e.g., dollar stores), an aspect that our study could not address. Third, online shopping data were not available in this study. For big-box and mid/high-end grocery stores that offer home delivery options to their customers, the number of store visitors may not reflect the true purchase activity. On a related note, store visits do not equate with the amount of consumption. It is possible that a person's store visits may decrease, but the overall spending may increase. Regarding the sampling and data collection practices, our data may have biases, potentially omitting populations with limited access to cell phones and other forms of communication technology. Additionally, data collection practices were not disclosed to us in detail. Future studies may consider coupling these data sets with other types of data to obtain a more comprehensive view of grocery shopping travel. 6 Conclusions This study investigates the impacts of the COVID-19 pandemic on travel patterns for grocery shopping in Columbus, Ohio. We estimated and compared changes in store visitor numbers across different store types to detect discrepancies in the impact of the pandemic among different customer segments and among four types of stores, namely mid/high-end grocery stores, big-box grocery stores, dollar stores, and local grocery stores. We found that COVID-19 exposes the existing disparities in food access and travel of underserved population, and that smaller stores, such as local stores and dollar stores experienced a slower recovery in store visits during the initial reopening phase of COVID-19 as compared to the larger store types. Our findings indicate that residents of low-income neighborhoods and food deserts became further constrained in their access to high-quality food during the pandemic. This highlights the importance of policies to provide or maintain transportation services that allow residents of such neighborhoods to continue accessing healthy food options, or to bring healthier food options to areas with few store choices. 10.13039/100014337 Furthermore , our findings show the importance of local and small-scale stores in providing access to food for low-income neighborhoods, which suggests that policies and relief funds to support such stores would benefit marginalized populations. Our study contributes to enhancing our understanding of how food shopping patterns are driven by socio-economic and built environment characteristics during a major disruption, thus emphasizes the pre-existing structural inequality in the US. Furthermore, it contributes to understanding the resilience of various store types to such a disruption, especially in light of the locational attributes captured by our study. The study can help practitioners and policy makers develop strategies to support the neighborhoods and local businesses that are disproportionately impacted by COVID-19 to recover after the pandemic. Insights from this study can also support preparations for future disruptions and recessions that disproportionately affect smaller businesses and marginalized populations. Lastly, the study demonstrates an analytical framework that can be applied in other cities and contexts. Author contribution statement AK: Methodology, Data Curation, Formal analysis, Investigation, Writing - Original draft, Writing - Review & Editing, Visualization. YM: Conceptualization, Methodology, Writing - Original draft, Writing - Review and Editing. ALC: Conceptualization, Methodology, Investigation, Writing - Original draft, Writing - Review and Editing. HJM: Conceptualization, Writing - Review & Editing. HTKL: Conceptualization, Methodology, Investigation, Writing - Original draft, Writing - Review & Editing, Supervision. Appendix Table A1 Summary of store characteristics and categorization criteria Table A1 Big-box stores Mid/high-end grocery stores Dollar stores Local stores Employee size Min 65 30 4 1 Max 500 350 30 150 Median 250 135 7 5 Mean 251 120 10 15 S.D. 121 85 11 35 Sales Volume (in thousands) Min 10280 7043 600 54 Max 79073 79989 1423 67403 Median 39537 24651 1050 940 Mean 39748 26795 1319 6696 S.D. 19178 19656 706 19156 Store types and selection criteria Warehouse clubs and supercenters that sell groceries Department stores and regional supermarket chains that sell groceries with an employee size greater than 25 and sales volume greater than 5 million. Discount stores of dollar chains Local supermarkets, independent grocery stores, limited assortment supermarkets, superettes, and specialty food stores NAICS code 452311 (warehouse clubs and supercenters) 452210 (Department stores) and 445110 (Supermarkets and Other Grocery (except Convenience) Stores) NAICS code 452319 (All other general merchandise stores) Rest of the stores from 445110 which do not fulfill the criteria of other 3 categories Example store names Walmart, Sam's Club, and Costco Kroger, Target, Trader Joe's, Whole Foods Market, Giant Eagle, and Meijer Dollar General, Dollar store, Family Dollar, and Dollar Tree Raisin rack natural food market, The Hills market, Saraga international grocery, Istanbul supermarket, Yasmin international market Acknowledgments We thank the Ohio Department of Transportation (ODOT) for providing access to the StreetLight data within the state of Ohio. ==== Refs References Bartik A.W. Bertrand M. Cullen Z. Glaeser E.L. Luca M. Stanton C. The impact of COVID-19 on small business outcomes and expectations Proceedings of the National Academy of Sciences 117 30 2020 17656 17666 10.1073/pnas.2006991117 Bower K.M. Thorpe R.J. Rohde C. Gaskin D.J. The intersection of neighborhood racial segregation, poverty, and urbanicity and its impact on food store availability in the United States Preventive Medicine 58 2014 33 39 10.1016/j.ypmed.2013.10.010 24161713 Cameron A.C. Trivedi P.K. Regression analysis of count data Vol. 53 2013 Cambridge university press Chen X. Kwan M.-P. Contextual uncertainties, human mobility, and perceived food environment: The uncertain geographic context problem in food access research American Journal of Public Health 105 9 2015 1734 1737 26180982 Cho C. Volpe R. Independent grocery stores in the changing landscape of the U.S Food retail industry 2017 United States Department of Agriculture (USDA) Christian W.J. Using geospatial technologies to explore activity-based retail food environments Spatial and Spatio-Temporal Epidemiology 3 4 2012 287 295 23149325 Chung C. Myers S.L. Do the poor pay more for food? An analysis of grocery store availability and food price disparities Journal of Consumer Affairs 33 2 1999 276 296 10.1111/j.1745-6606.1999.tb00071.x Colombo L. Halley D. Lindsjo M. Martin M. Reece J. Rogers C. Neighborhoods & community development in Franklin County: Understanding our past & preparing for our future. January Available at: 2012 Http://Kirwaninstitute. Osu. Edu/Docs/CDC_FinalReport. Pdf (Accessed July 1, 2013 Doussard M. Degraded work: The struggle at the bottom of the labor market 2013 U of Minnesota Press Dutko P. Ver Ploeg M. Farrigan T. Characteristics and influential factors of food deserts 2012 Ewing R. Cervero R. Travel and the built environment: A synthesis Transportation Research Record 1780 1 2001 87 114 10.3141/1780-10 Farber S. Morang M.Z. Widener M.J. Temporal variability in transit-based accessibility to supermarkets Applied Geography 53 2014 149 159 10.1016/j.apgeog.2014.06.012 Farber S. Ritter B. Fu L. Space–time mismatch between transit service and observed travel patterns in the wasatch front, Utah: A social equity perspective Travel Behaviour and Society 4 2016 40 48 10.1016/j.tbs.2016.01.001 Giang T. Karpyn A. Laurison H.B. Hillier A. Perry R.D. Closing the grocery gap in underserved communities: The creation of the Pennsylvania fresh food financing initiative Journal of Public Health Management and Practice 14 3 2008 272 279 10.1097/01.PHH.0000316486.57512.bf 18408552 Guy C. Clarke G. Eyre H. Food retail change and the growth of food deserts: A case study of Cardiff 2004 International Journal of Retail & Distribution Management Hallett L.F. McDermott D. Quantifying the extent and cost of food deserts in Lawrence, Kansas, USA Applied Geography 31 4 2011 1210 1215 10.1016/j.apgeog.2010.09.006 Helbich M. Schadenberg B. Hagenauer J. Poelman M. Food deserts? Healthy food access in amsterdam Applied Geography 83 2017 1 12 Helling A. Sawicki D.S. Race and residential accessibility to shopping and services Housing Policy Debate 14 1–2 2003 69 101 10.1080/10511482.2003.9521469 InfoGroup InfoGroup. Infogroup 2019 https://www.infogroup.com/our-data/ Jiao J. Moudon A.V. Ulmer J. Hurvitz P.M. Drewnowski A. How to identify food deserts: Measuring physical and economic access to supermarkets in king county, Washington American Journal of Public Health 102 10 2012 e32 e39 10.2105/AJPH.2012.300675 22897554 Johnson K. Percy S. Wagner E. University of Wisconsin--Milwaukee, Center for Urban Initiatives and Research, & Food System Assessment Project (Milwaukee, Wis. ) Comparative study of food pricing and availability in Milwaukee 1996 UWM Center for Urban Initiatives and Research Kelloway C. Dollar stores are taking over the grocery business, and it's bad news for public health and local economies 2018 December 17 Civil Eats https://civileats.com/2018/12/17/dollar-stores-are-taking-over-the-grocery-business-and-its-bad-news-for-public-health-and-local-economies/ LeDoux T.F. Vojnovic I. Going outside the neighborhood: The shopping patterns and adaptations of disadvantaged consumers living in the lower eastside neighborhoods of Detroit, Michigan Health & Place 19 2013 1 14 10.1016/j.healthplace.2012.09.010 23142639 Leslie T.F. Frankenfeld C.L. Makara M.A. The spatial food environment of the DC metropolitan area: Clustering, co-location, and categorical differentiation Applied Geography 35 1–2 2012 300 307 10.1016/j.apgeog.2012.07.008 Li J. Kim C. Exploring relationships of grocery shopping patterns and healthy food accessibility in residential neighborhoods and activity space Applied Geography 116 2020 102169 Lucas K. A new evolution for transport-related social exclusion research? Journal of Transport Geography 81 2019 102529 10.1016/j.jtrangeo.2019.102529 Miller E.J. Accessibility: Measurement and application in transportation planning Transport Reviews 38 5 2018 551 555 10.1080/01441647.2018.1492778 Moore L.V. Diez Roux A.V. Associations of neighborhood characteristics with the location and type of food stores American Journal of Public Health 96 2 2006 325 331 16380567 Raja S. Ma C. Yadav P. Beyond food deserts: Measuring and mapping racial disparities in neighborhood food environments Journal of Planning Education and Research 27 4 2008 469 482 10.1177/0739456X08317461 Ramsey K. Bell A. Smart location database 2014 Washington, DC Reich A. Walmart's consumer redlining Contexts 15 4 2016 74 77 10.1177/1536504216685128 SafeGraph SafeGraph | POI data, business listings, & foot-traffic data SafeGraph https://www.safegraph.com/weekly-foot-traffic-patterns 2020 Semple H. Giguere A. The evolution of food deserts in a small midwestern city: The case of Ypsilanti, Michigan: 1970 to 2010 Journal of Planning Education and Research 38 3 2018 359 370 Shannon J. Beyond the supermarket solution: Linking food deserts, neighborhood context, and everyday mobility Annals of the Association of American Geographers 106 1 2016 186 202 10.1080/00045608.2015.1095059 Shannon J. Dollar stores, retailer redlining, and the metropolitan geographies of precarious consumption Annals of the Association of American Geographers 2020 1 19 10.1080/24694452.2020.1775544 Shannon J. Bagwell-Adams G. Shannon S. Lee J.S. Wei Y. The mobility of food retailers: How proximity to SNAP authorized food retailers changed in Atlanta during the Great Recession Social Science & Medicine 209 2018 125 135 10.1016/j.socscimed.2018.05.046 29859969 Shannon J. Christian W.J. What is the relationship between food shopping and daily mobility? A relational approach to analysis of food access Geojournal 82 4 2017 769 785 10.1007/s10708-016-9716-0 Shannon J. Shannon S. Adams G.B. Lee J.S. Growth in SNAP retailers was associated with increased client enrollment in Georgia during the Great recession Health Affairs 35 11 2016 2100 2108 10.1377/hlthaff.2016.0324 27834252 Shaw H.J. Food deserts: Towards the development of a classification Geografiska Annaler - Series B: Human Geography 88 2 2006 231 247 StreetLight Data, Inc. StreetLight data: Transportation Analytics on demand. StreetLight data 2020 https://www.streetlightdata.com/ US Census Bureau County business patterns (CBP). The United States census Bureau 2019 https://www.census.gov/programs-surveys/cbp.html US Census Bureau 2019 U.S. Census Bureau QuickFacts Franklin County, Ohio https://www.census.gov/quickfacts/franklincountyohio US Census Bureau 2014-2018 American community Survey 5-year estimates 2020 https://data.census.gov/cedsci/? US Census Bureau Household Pulse Survey. Census.Gov 2021 https://www.census.gov/householdpulsedata USDA 2019 October 31 USDA ERS - Documentation https://www.ers.usda.gov/data-products/food-access-research-atlas/documentation/ Vargas T.L. Consumer redlining and the reproduction of inequality at dollar general. Qualitative sociology 2021 10.1007/s11133-020-09473-w Ver Ploeg M. Mancino L. Todd J.E. Clay D.M. Scharadin B. Where do Americans usually shop for food and how do they travel to get there? Initial findings from the national household food acquisition and purchase Survey 2015 Walker R.E. Keane C.R. Burke J.G. Disparities and access to healthy food in the United States: A review of food deserts literature Health & Place 16 5 2010 876 884 10.1016/j.healthplace.2010.04.013 20462784 Widener M.J. Farber S. Neutens T. Horner M.W. Using urban commuting data to calculate a spatiotemporal accessibility measure for food environment studies Health & Place 21 2013 1 9 23395918 Widener M.J. Farber S. Neutens T. Horner M. Spatiotemporal accessibility to supermarkets using public transit: An interaction potential approach in Cincinnati, Ohio Journal of Transport Geography 42 2015 72 83 10.1016/j.jtrangeo.2014.11.004 Widener M.J. Metcalf S.S. Bar-Yam Y. Dynamic urban food environments American Journal of Preventive Medicine 41 4 2011 439 441 10.1016/j.amepre.2011.06.034 21961473 Widener M.J. Minaker L. Farber S. Allen J. Vitali B. Coleman P.C. Cook B. How do changes in the daily food and transportation environments affect grocery store accessibility? Applied Geography 83 2017 46 62 10.1016/j.apgeog.2017.03.018 Widener M.J. Shannon J. When are food deserts? Integrating time into research on food accessibility Health & Place 30 2014 1 3 25145664 Wrigley N. Warm D. Margetts B. Lowe M. The Leeds “food deserts” intervention study: What the focus groups reveal 2004 International Journal of Retail & Distribution Management Yelp Inc Yelp: Local economic impact report. Yelp 2020 https://www.yelpeconomicaverage.com/business-closures-update-sep-2020 Zenk S.N. Schulz A.J. Israel B.A. James S.A. Bao S. Wilson M.L. Neighborhood racial composition, neighborhood poverty, and the spatial accessibility of supermarkets in metropolitan Detroit American Journal of Public Health 95 4 2005 660 667 15798127 Zenk S.N. Schulz A.J. Matthews S.A. Odoms-Young A. Wilbur J. Wegrzyn L. Gibbs K. Braunschweig C. Stokes C. Activity space environment and dietary and physical activity behaviors: A pilot study Health & Place 17 5 2011 1150 1161 21696995

0

PMC9753127

NO-CC CODE

2022-12-16 23:25:30

no

Appl Geogr. 2021 Sep 9; 134:102517

utf-8

Appl Geogr

10.1016/j.apgeog.2021.102517

oa_other

==== Front J Neurol Sci J Neurol Sci Journal of the Neurological Sciences 0022-510X 1878-5883 Elsevier B.V. S0022-510X(21)02787-8 10.1016/j.jns.2021.120085 120085 Clinical Short Communication Integrating neurology and pharmacy through telemedicine: A novel care model Li Hanlin a⁎ Naqvi Imama A. b Tom Sarah E. c Almeida Barbara b Baratt Yuliya a Ulane Christina M. b a Department of Pharmacy, NewYork-Presbyterian Hospital, USA b Department of Neurology, Columbia University Irving Medical Center, NewYork-Presbyterian Hospital, USA c Departments of Neurology and Epidemiology, Columbia University Irving Medical Center, USA ⁎ Corresponding author at: Lead Clinical Pharmacy Manager, Ambulatory Care, NewYork-Presbyterian Hospital, 622 West 168th Street, New York, NY 10032, USA. 9 12 2021 15 1 2022 9 12 2021 432 120085120085 13 9 2021 23 11 2021 6 12 2021 © 2021 Elsevier B.V. All rights reserved. 2021 Elsevier B.V. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Teleneurology had been best studied in acute stroke care, but the Coronavirus (COVID)-19 pandemic has highlighted applicability in outpatient practice. Telepharmacy is a convenient method for pharmacists to provide medication management to enhance care. Studies in the outpatient space suggest non-inferiority of teleneurology to increase access to specialized care for patients in rural locations. The role of telemedicine based interdisciplinary collaborations in a metropolitan and under-resourced setting has not been explored. We describe our approach to a teleneurology-telepharmacy collaboration at an urban academic medical center. Since its implementation pre-COVID, the program has expanded and transformed to serve the community further. Keywords Teleneurology Telepharmacy Ambulatory care ==== Body pmc1 Introduction Telemedicine is fast gaining acceptance as a health care delivery model. Access to neurological care in some areas is in part due to a shortage of neurologists. Telemedicine holds promise as a means for a cost-effective and efficient manner in which to increase access to care. Teleneurology was implemented most widely in acute stroke care before the COVID-19 pandemic. In the outpatient setting, teleneurology was utilized mainly to increase specialized neurological care for patients in remote or rural locations. [1,2] The pandemic has incentivized further infrastructure development and policies to expand these efforts. [3] Multiple specialties have reported the noninferiority of telemedicine when evaluating access to care and health outcomes. [4] Additionally, data suggest feasibility, acceptability, and cost savings for teleneurology in the outpatient setting. [5] This shift offers an opportunity to advance healthcare delivery and the potential to promote interdisciplinary collaboration. [6] A team-based care approach incorporating pharmacy can prevent medication errors and enhance clinical engagement to improve patient health outcomes. [7] Telemedicine may be a resource-efficient approach to improve health care through collaborative pharmacist services for neurology patients. [8] Telepharmacy interventions may positively impact disease management for specific disorders regarding utilization, self-management, and adherence. [7,9]. However, literature is scarce on the broader incorporation of telepharmacy services among patients with chronic general neurological conditions. At our urban academic center, we serve the local community of patients who experience several barriers to care, including physical and cognitive disability due to chronic neurological disease, socioeconomic constraints, and limited digital and health literacy. Interdisciplinary care and the need to create digital health equity has great potential for under-resourced populations. [10] We developed and implemented a novel clinical care model to improve access to neurological and interdisciplinary care. The purpose of this report is to describe our experience of a teleneurology-telepharmacy collaboration, with a focus on refining and sustaining telepharmacy services to patients with chronic neurological conditions through the pandemic and beyond. 2 Methods 2.1 Program launch Outpatient telemedicine in our general neurology practice began as a pilot in 2018 with three goals: 1) explore the feasibility of converting in-person follow-up appointments to telemedicine, 2) introduce telepharmacy services to enhance neurological care delivery, and 3) assess patient receptiveness to telehealth. A single-center, descriptive study was completed to evaluate the program and its objectives. This study was reviewed and approved by the Institutional Review Board at NewYork-Presbyterian Hospital/Columbia University Irving Medical Center. For the purposes of our program, we define teleneurology and telepharmacy as two-way video-conferencing between patient and neurologist, and patient and pharmacist, respectively. Existing patients in the Adult General Neurology Ambulatory Care Network Clinic were screened and approached for potential conversion from in-person to video follow-up appointments. The structure of the paired video visits was as follows (Fig. 1A):Fig. 1 Patient-centered teleneurology-telepharmacy workflow. Fig. 1 1st video appointment - telepharmacy: patient meets with the clinical pharmacist who provides patient care and completes documentation. 2nd video appointment – teleneurology: patient meets with the neurologist for clinical follow-up assessment. Prior to visit, neurologist has reviewed telepharmacy documentation and is prepared to discuss any issues related to medication adherence, tolerance and effectiveness. An internal questionnaire developed through pharmacist and physician group feedback assessed patient experience and acceptability of the new care model. Patients who completed video visits received a phone call from the team to participate in an anonymous survey. The questions evaluated the patient's comfort and satisfaction level around the quality of visits (Supplemental material). 2.2 Current state As circumstances changed due to the coronavirus pandemic, our teleneurology and telepharmacy model adapted (Fig. 1B). Building on the pilot program, collaboration has continued and undergone several changes with Plan-Do-Study-Assess learning theory implementation into a patient-tailored program. [12] Beginning in March 2020 through June 2021, all neurology visits were converted from in-person to telemedicine to ensure patient, physician and staff safety during the height of the pandemic. Telepharmacy services were optimized to meet patient and practice needs. Based on information gathered from the pilot, the timing of the telepharmacy visit was adjusted to occur after the teleneurology visit, by referral from the neurologist or pharmacist self-referral based on criteria developed with careful consideration of goals and objectives of our collaborative efforts (Table 2). In the most recent expansion phase, telepharmacy services are now integrated into an interdisciplinary secondary stroke prevention clinic [6]. The team developed a collaborative drug therapy management agreement for the co-management of hypertension in stroke patients under New York State law. [11] The agreement allows the pharmacist to address medication issues during care transitions and triages to follow-up services as appropriate to optimize post-discharge stroke patient care. 3 Results During the initial program launch in May 2018 through March 2019, a total of 103 video visits (teleneurology and telepharmacy combined) with 56 unique patients were completed. A significant reduction in time spent per visit was observed. The average duration for an in-person neurology follow-up appointment at the clinic was approximately 90 min from check-in to check-out; the average combined duration of teleneurology plus telepharmacy appointments was approximately 35 min. Patient perception of the new model was assessed through the survey. Seventeen (30%) patients were reached and agreed to participate. Most agreed that the quality of the video visit was better than or the same as that of an in-person visit, and felt comfortable speaking with the neurologist and pharmacist during the video visit. With the onset of the pandemic and the change in our teleneurology-telepharmacy model, a total of 103 telepharmacy visits were completed between April 2020 through January 2021. Clinical and demographic characteristics of patients participating in telepharmacy visits since the pandemic are shown in Table 1 . Approximately half of the patients were non-English speaking, a majority were women (76%), most were seen for general neurological conditions such as migraine (79%), and the remainder (21%) were seen for various subspecialty conditions.Table 1 Characteristics of neurology patients completing telepharmacy visits (n = 103)*. Table 1Age, mean (SD) 48 (18) Female, n (%) 79 (77%) Male, n (%) 24 (23%) Primary language, n (%)  English 54 (52%)  Spanish 37 (36%)  Other 12 (12%) Type of neurological diagnoses, n (%)  General 81 (79%)  Subspecialty 22 (21%) Number of referral criteria met for telepharmacy, n (%)  One 68 (66%)  Two 27 (26%)  Three or more 8 (8%) Mode of visit, n (%)  Video 100 (97%)  Phone 3 (3%) Number of patients completing one or more telepharmacy visits, n (%)  One 103 (100%)  Two 18 (17%)  Three or more 3 (3%) *Data from April 2020–January 2021. The patient-centered criteria developed for referral to telepharmacy and data regarding the frequency of utilization are shown in Table 2 . Most patients were referred for a single criterion (84%). For those meeting two or more criteria, many required multiple telepharmacy appointments to meet the needs.Table 2 Criteria utilization for telepharmacy referral.⁎ Table 2Criteria for referral to telepharmacy N (% of all referral criteria) Drug titration 69 (39%) Drug initiation 42 (24%) Medication reconciliation 38 (21%) Drug discontinuation 14 (8%) Medication education 13 (7%) Complex medication regimen 1 (0.5%) Medication Adherence 1 (0.5%) ⁎ Patient may be referred for more than one criterion, total number of referral criteria n = 178. 4 Discussion We have described the implementation and feasibility of telemedicine in an underserved patient population at an urban academic center. Our novel collaboration provided more efficient care and was generally well-accepted by patients. Our outpatient practices had to be drastically adapted due to the pandemic. Most significantly, this entailed a complete transition to telemedicine. Given this change and from our experience with the pilot teleneurology and telepharmacy program, we modified our workflow to better suit the needs of our patients. Criteria for referral to telepharmacy were created to reflect the standard treatment plans and changes that occur during a neurology visit. We demonstrated that both neurologist-referral and pharmacist review and self-referral effectively recruited patients for telepharmacy visits. Our clinic serves a diverse population with chronic neurological conditions. Patients with general neurological diagnoses, such as migraines, neuropathy, back pain, and those with sub-specialty conditions including epilepsy, stroke, neuro-immunology participated in telepharmacy. Patients have often prescribed complicated medication changes with the initiation of medications, titration of doses, and or discontinuation of therapy all at the same time. The telepharmacy visit reinforces these changes, answers drug-related questions, and likely helps to ensure adherence and effectiveness of the regimen. Timely identification of medication-related issues and communication with the neurology team expedites problem resolution, and we presume this enhances the quality of care. Patient perception of teleneurology and telepharmacy was evaluated using an internally-developed survey during the program's initial phase. Understanding the needs of our diverse population is crucial to optimize health-related information provided and adjust our processes. [10] Education tools that are language-specific and designed with community participatory design to promote medication adherence may be one such approach. Routine assessment of patient and provider experience with telemedicine may be considered to improve workflows further. There are opportunities to grow the telepharmacy services. In particular, integrated care with pharmacy-guided support to address barriers to medication adherence has been shown to augment blood pressure control in hypertensive patients. [12] Such an collaboration may improve secondary stroke prevention through vascular risk factor control. Our center developed a comprehensive agreement covering multiple associated disease states under the stroke umbrella, allowing pharmacists to augment care in these patients by titration of medications in between specialist appointments. Sub-specialties such as epilepsy and multiple sclerosis may also leverage and benefit from telepharmacy services. [13,14] Our study has a small data sample and measured process metrics. Focus groups to understand the barriers and promotors among stakeholders in the community may inform our collaborative approach to include diverse populations. In the future, we hope to evaluate clinical outcomes of patients with chronic neurological diseases and the role of teleneurology and telepharmacy in these outcomes. To our knowledge, this is the first study to describe a patient-centered teleneurology-telepharmacy interdisciplinary collaboration in an outpatient neurology clinic at an urban academic center with an underserved patient population. The further development of a sustainable model of teleneurology and telepharmacy integration has the potential to improve access to care, quality of care and efficiency of care. Author contributions HL and CMU conceived and implemented the program and collected the data. BA coordinated patients and providers. SET performed data analysis. HL, YB, and IAN drafted a significant portion of the manuscript and carried out the workflow. All authors discussed the results and contributed to the final manuscript. Declaration of competing interest All authors have no relevant conflicts of interest. Appendix A Supplementary data Supplementary material Image 1 Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.jns.2021.120085. ==== Refs References 1 Dorsey E.R. Teleneurology and mobile technologies: the future of neurological care Nat. Rev. Neurol. 14 5 2018 285 297 29623949 2 Schneider R.B. Biglan K.M. The promise of telemedicine for chronic neurological disorders: the example of Parkinson’s disease Lancet Neurol. 16 7 2017 541 551 28566190 3 Smith A.C. Telehealth for global emergencies: implications for coronavirus disease 2019 (COVID-19) J. Telemed. Telecare 26 5 2020 309 313 32196391 4 Halbert K. Bautista C. Telehealth use to promote quality outcomes and reduce costs in Stroke care Crit. Care Nurs. Clin. North Am. 31 2 2019 133 139 31047088 5 Kruse C.S. Telehealth and patient satisfaction: a systematic review and narrative analysis BMJ Open 7 8 2017 e016242 6 Telehealth After Stroke Care: Integrated Multidisciplinary Access to Post-stroke Care (TASC) [cited 2021 05/07/2021]; Available from https://clinicaltrials.gov/ct2/show/NCT04640519 2020 7 Basaraba J.E. Pharmacists as care providers for Stroke patients: a systematic review Canadian J. Neurol. Sci. 45 1 2018 49 55 8 Le T. Toscani M. Colaizzi J. Telepharmacy: a new paradigm for our profession J. Pharm. Pract. 33 2 2020 176 182 30060679 9 McAlister F.A. Case management for blood pressure and lipid level control after minor stroke: PREVENTION randomized controlled trial Can. Med. Assoc. J. 186 8 2014 577 584 24733770 10 Rodriguez J.A. Clark C.R. Bates D.W. Digital health equity as a necessity in the 21st century cures act era Jama 323 23 2020 2381 2382 32463421 11 NYSED.gov Office of the Professions Education Law Article 137, Pharmacy [cited 2021 05/07/2021]; Available from: http://www.op.nysed.gov/prof/pharm/article137.htm 2021 12 McManus R.J. Effect of self-monitoring and medication self-titration on systolic blood pressure in hypertensive patients at high risk of cardiovascular disease: the TASMIN-SR randomized clinical trial Jama 312 8 2014 799 808 25157723 13 May A. Morgan O. Quairoli K. Incorporation and impact of a clinical pharmacist in a hospital-based neurology clinic treating patients with multiple sclerosis Int J MS Care 23 1 2021 16 20 33658901 14 Fogg A. An exploratory study of primary care pharmacist-led epilepsy consultations Int J Pharm Pract 20 5 2012 294 302 22953768

34915405

PMC9753134

NO-CC CODE

2022-12-16 23:25:30

no

J Neurol Sci. 2022 Jan 15; 432:120085

utf-8

J Neurol Sci

10.1016/j.jns.2021.120085

oa_other

==== Front Water Res Water Res Water Research 0043-1354 1879-2448 Elsevier Ltd. S0043-1354(21)00039-7 10.1016/j.watres.2021.116841 116841 Article Stable dechlorination of Trichloroacetic Acid (TCAA) to acetic acid catalyzed by palladium nanoparticles deposited on H2-transfer membranes Cai Yuhang ab1 Long Xiangxing ac1 Luo Yi-Hao a⁎ Zhou Chen a Rittmann Bruce E. a a Biodesign Swette Center for Environmental Biotechnology, Arizona State University, Tempe, AZ 85287-5701, United States b College of Power and Energy Engineering, Harbin Engineering University, Harbin 150001, PR China c Nanosystems Engineering Research Center for Nanotechnology-Enabled Water Treatment, School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe, AZ 85287-3005, United States ⁎ Corresponding author. 1 Yuhang Cai and Xiangxing Long contributed equally to this work. 15 1 2021 15 3 2021 15 1 2021 192 116841116841 18 11 2020 9 1 2021 12 1 2021 © 2021 Elsevier Ltd. All rights reserved. 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Trichloroacetic acid (TCAA) is a common disinfection byproduct (DBP) produced during chlorine disinfection. With the outbreak of the Coronavirus Disease 2019 (COVID-19) pandemic, the use of chlorine disinfection has increased, raising the already substantial risks of DBP exposure. While a number of methods are able to remove TCAA, their application for continuous treatment is limited due to their complexity and expensive or hazardous inputs. We investigated a novel system that employs palladium (Pd0) nanoparticles (PdNPs) for catalytic reductive dechlorination of TCAA. H2 was delivered directly to PdNPs in situ coated on the surface of bubble-free hollow-fiber gas-transfer membranes. The H2-based membrane Pd film reactor (H2−MPfR) achieved a high catalyst-specific TCAA reduction rate, 32 L/g-Pd/min, a value similar to the rate of using homogeneously suspended PdNP, but orders of magnitude higher than with other immobilized PdNP systems. In batch tests, over 99% removal of 1 mM TCAA was achieved in 180 min with strong product selectivity (≥ 93%) to acetic acid. During 50 days of continuous operation, over 99% of 1 mg/L influent TCAA was removed, again with acetic acid as the major product (≥ 94%). We identified the reaction pathways and their kinetics for TCAA reductive dechlorination with PdNPs using direct delivery of H2. Sustained continuous TCAA removal, high selectivity to acetic acid, and minimal loss of PdNPs support that the H2−MPfR is a promising catalytic reactor to remove chlorinated DBPs in practice. Graphical abstract Image, graphical abstract Keywords Disinfection byproducts Trichloroacetic acid (TCAA) Palladium nanoparticle (PdNP) Hollow fiber membrane Catalytic dechlorination ==== Body pmc1 Introduction Chlorination, the most widely applied method for water disinfection, forms numerous disinfection byproducts (DBPs) (Hrudey, 2009). The second group of most common DBPs (after trihalomethanes) are the haloacetic acids (HAAs), including mono-, di-, and tri- chloroacetic acids (MCAA, DCAA and TCAA) (Hong et al., 2013; Krasner et al., 2006). Exposure to HAAs in drinking water is associated with an increased risk of cancer (Boorman et al., 1999; Villanueva et al., 2004), and recent studies suggest that HAA exposure increases antibiotic resistance in bacteria (Li et al., 2016; Lv et al., 2014). The U.S. Environmental Protection Agency (EPA) posted a maximum contaminant level (MCL) of 60 μg/L for the sum of five haloacetic acids (HAA5), including three chloroacetic acids (CAAs) and two bromoacetic acids (Pontius, 1999). With the outbreak of the Coronavirus Disease 2019 (COVID-19) pandemic, the use of chlorine for disinfection has increased and is increasing the risks of DBP exposure, especially in raw water (Chang et al., 2020; Wang et al., 2020). Pulp-mill wastewaters have still higher concentrations of HAAs (Hoekstra et al., 1999). An effective method for HAA removal will minimize undesired outcomes associated HAAs in water. Reductive methods for HAAs removal have included chemical, electrochemical, and sonochemical methods. For instance, Wang et al. (2008) reported reductive dechlorination of TCAA using membrane-immobilized palladium/iron, with acetic acid (AA) (60%), DCAA (14%), MCAA (26%) as the products. However, the TCAA-removal efficiency decreased to below 60% after 8 cycles. Using a three-dimensional graphene−copper-foam electrode with an applied cathode potential of −1.2 V, Mao et al. (2016) achieved 95% removal of 500 mg/L TCAA within 20 min at pH 6.8, with acetic acid (AA) (70%) and MCAA (25%) as the major products at 180 min. Esclapez et al. (2015) reported the dechlorination of TCAA via high-frequency sonoelectrochemical degradation using a Pb cathode, with MCAA (20~40%) and DCAA (10~50%) as the major products. Although these studies provide promising results for TCAA removal, toxic byproducts, including DCAA and MCAA, were produced using these methods, and they did not demonstrate continuous treament. Furthermore, chemical inputs, such as an electrolyte (e.g. Na2SO4), were required for most of electrochemical methods. Among all reductive methods, catalytic reductive dechlorination using palladium nanoparticle (PdNP) is especially promising (De Corte et al., 2011; Hennebel et al., 2009), since H2 is as the only input needed; not needed are the external energy or chemical inputs required in most other methods, e.g., solar radiation, sonic power, hydro-peroxide or iron (Feijen-Jeurissen et al., 1999; Han et al., 2017; Jewell and Davis, 2006; Liu et al., 2001). H2 is nontoxic and readily available, can be purchased in bulk or formed on-site by electrolysis or methane reforming, and has mature handling methods that ensure it safe use (Dincer, 2012; Lv et al., 2012; Wegner et al., 2006). H2 can be used as reductant to convert dissolved Pd2+ ion to zero-valent PdNPs through self-catalyzed reduction. PdNPs are effective because they strongly adsorb H2 and activate it to the reactive H* atoms on the PdNP surface. The reactive H* atom has been proven useful for reducing a wide range of oxidized contaminants: e.g., trichloroethene, nitrate, and chlorophenols (El-Sharnouby et al., 2018; Liu et al., 2001; Zhou et al., 2017), as well as dehalogenation of DBPs: e.g., trihalomethanes and HAAs (Mao et al., 2018; Xu et al., 2020). Efficient H2 delivery is mandatory because of its low solubility in water and combustibility. Delivering H2 with nonporous gas-transfer membranes is a well-studied approach for safe and controllable H2 delivery to biofilms, using the H2-based membrane biofilm reactor (H2−MBfR) (Lv et al., 2020; Zhao et al., 2013; Zhou et al., 2017). In this study, we develop and demonstrate a H2-based palladium-film reactor (H2−MPfR), a modification of the H2−MBfR that utilizes PdNPs instead of biofilm. We in situ synthesized PdNPs that self-assembled on the membrane surface. Using TCAA as a model compound, the membrane achieved safe and controllable H2 delivery to the coated PdNPs for catalytic dechlorination of halogenated DBPs. We elaborate the reaction kinetics and pathways for TCAA catalytic reductive dechlorination. 2 Materials and methods 2.1 System configuration and setup Fig. 1A is a schematic of the one-column bench-scale H2−MPfR, which shared the configuration of previous studies (Zhao et al., 2013; Zhou et al., 2017). It had a total working volume of 35 mL and contained one bundle of 120 identical hollow-fiber membranes in a glass tube (6-cm internal diameter and 24-cm length). The hollow-fiber membrane used in this study was made by nonporous polypropylene and had a 200-μm outer diameter, ~105-μm inner diameter, and ~52-μm wall thickness (Teijin, Ltd., Japan). The 120 fibers provided 0.018 m2 surface area. H2 gas (>99.9%) was supplied to both ends of the fiber bundle at a pressure controlled by a pressure regulator. A peristaltic recycling pump (MasterflexⓇ, USA) was set at 50 mL/min to mix the water in the reactor, and the sampling point was along the pumping tube.Fig. 1 (A). Schematic of the single-column H2−MPfR. (B) Pictures of membranes in the reactor during PdNPs coating using Na2PdCl4 solution with H2 pressure of 10 psig. (C) Pictures of membranes with different PdNPs loadings. Fig 1 All solutions were prepared using analytical-grade reagents (Millipore Sigma, USA) and Millipore ultra-pure water (specific conductivity of 18.2 MΩ). The synthetic DBP water for batch tests and continuous operation was made with 33 mg/L NaCl (imitating low Cl− level in tap and raw water), 1 mM KH2PO4, 1 mM Na2HPO4 (as buffer with initial pH~7.2), and the CAAs (concentration depended on the experiment) dissolved in ultra-pure water. The synthetic DBP water used for pH-controlled batch tests was buffered with a 10-mM KH2PO4 plus NaOH to fix the pH at 3, 5, 7, 9 and 11. The pH changes during all batch and continuous operation were less than 0.3 units. 2.2 In situ deposition of PdNPs on the H2-transfer membranes We deposited PdNPs on the H2-transfer membrane using an in situ reduction method. We first filled the reactor with an Na2PdCl4 solution of 0.1 to 5 mM. We then turned on the H2 supply with a pressure of 10 psig (1.69 atm absolute pressure) and started the recirculation pump. Autocatalytic reduction of Pd2+ to PdNPs occurred directly on the membranes, and complete reduction and deposition required 6 to 12 h. The initial Pd2+ concentration affected the duration of the reduction and deposition process; details of the duration for different initial Pd2+ concentration are listed in Supplementary Material Table S1. Fig. 1B shows how the solution turned from dange to clear, while the membranes became black due to the coated PdNPs. Fig. 1C shows the membranes coated with different surface loadings of PdNPs. Aqueous-phase measurement of Pd2+ gave negligible concentration at the end of the reduction and deposition period. We washed the reactor twice with deionized water to remove any suspended PdNPs and residual Pd2+. 2.3 Characterization of the membrane-supported PdNPs After depositing PdNPs on the membrane surface, we used a scalpel to cut off 2 or 3 fiber pieces, each 0.5 to 1 cm long, for solid-state characterization. We employed scanning transmission electron microscopy (STEM; JEM-ARM200F), tranmission electron microscopy (TEM; Philips CM 12), and energy dispersive X-ray microanalysis (EDX) coupled with STEM to characterize the membrane coatings. The chemically fiexed fibers were sectioned using a Leica Ultracut-R microtome at 70 nm thick for the TEM and STEM analysis. Details of chemical fixation are described by Zhou et al. (2014). 2.4 Batch and continuous tests For each batch tests, we purged the MPfR with pure N2 gas for 15 min before filling the MPfR with synthetic medium containing DBP. All batch kinetic tests were conducted in triplicate using three individual H2−MPfRs with the same conditions. Except for batch tests using different PdNP loading and H2 pressure, all the batch tests were conducted with MPfR coated from a solution of 1 mM Pd2+ and supplied with 3 psig H2. In continuous operation, the same synthetic medium was fed into the MPfR using a peristaltic pump with a constantly controlled flow rate. In order to evaluate any impacts of other constituents in the water, we used tap water (Tempe AZ, USA) to make synthetic medium (instead of ultrapure water) in the final stage. The tap water had Total Dissolve Solids (TDS), Hardness, and Dissolved Organic Carbon (DOC) concentrations of 560 mg/L, 250 as mg CaCO3/L and 1.6 mg/L, respectively. Effluent water samples were collected and measured every day after filtration through 0.22-μm polyvinylidene difluoride syringe filter. When the effluent concentrations of all the susbstrate and the products were stable for three days (a variability less than 5%), we considered the MPfR reached the steady state. 2.5 Analytical methods All aqueous samples for TCAA, DCAA, MCAA, AA, and chloride were first filtered though 0.22-μm polyvinylidene difluoride syringe filter and then quantified using anionic chromatography (IC) (Metrohm 930 Compact IC). The IC had a Metrosep A supp 5 −250/4.0 column and an eluent of 1 mM sodium bicarbonate (NaHCO3) and 3.2 mM sodium carbonate (Na2CO3) with a flow rate of 0.7 mL/min. The detection limits of TCAA, DCAA, MCAA, AA, and chloride were 15, 15, 10, 10, and 3 μg/L, respectively. To verify the actual PdNPs loading on membrane surface, an entire bundle of fiber membranes coated with PdNPs was dissolved through microwave-assisted total digestion (Luo et al., 2020) at the end of catalysts loading batch tests. An inductively coupled plasma mass spectrometry (ICP-MS; PerkinElmer NexIONⓇ 1000) was used to measure the Pd concentration of digestied Pd solutions and effuents taken during continuous operation. 2.6 Kinetic modeling Assuming that the reductions of TCAA, DCAA, and MCAA occurred independently, we set up three reactions (Eqs. (1), (2), and (3), respectively) to simulate the reactions occurring in the H2−MPfR:(1) C2HO2Cl3+(a+2b+3c)H2→aC2H2O2Cl2+bC2H3O2Cl+cC2H4O2+(a+2b+3c)HCl (2) C2H2O2Cl2+(d+2e)H2→dC2H3O2Cl+eC2H4O2+(d+2e)HCl (3) C2H3O2Cl+H2→C2H4O2+HCl In Eq. (2), a, b and c represent the reaction selectivities of TCAA reduction to DCAA, MCAA, and AA, respectively. In Eq. (3), d and e represent the reaction selectivities of DCAA reduction to MCAA and AA, respectively. All reaction rates followed first-order kinetics, Eqs. (4)-(7), with first-order rate constants k1, k2 and k3 obtained from the bath tests with TCAA, DCAA or MCAA as the sole reactant.(4) d[TCAA]dt=−k1[TCAA] (5) d[DCAA]dt=ak1[TCAA]−k2[DCAA] (6) d[MCAA]dt=bk1[TCAA]+dk2[DCAA]−k3[MCAA] (7) d[AA]dt=ck1[TCAA]+ek2[DCAA]+k3[MCAA] Then, the first-order rate constants were used as independent parameters to calculate product selectivity using the results of TCAA-dechlorination batch tests. The model, used to calculate the first-order rate constants and product selectivities, was executed in MATLAB 7.0. The sum of squared relative residual (SSRR) (Cristóvão et al., 2009) was used to assess the goodness of fit between the model results and the experimentally measured values (Supplementary Material, Table S2):(8) SSRR=∑i=1n[Sm,i−Se,iSe,i]2 where n is the number of samples in the experiments, and Sm,i and Se,i are component concentrations in the model and experiment, respectively. 2.7 Specific activity, selectivity, and statistical analyses The catalyst-specific activity (L/g Pd-min) was calculated by dividing the first-order rate constant (1/min) by the Pd loading (g Pd/L: the weight of PdNPs coated on membrane divided by the reactor working volume). Catalyst-specific activity is a commonly metric of the kinetics of PdNPs-catalyzed reactions (Fritsch et al., 2003; Nutt et al., 2006). The product selectivity was calculated by dividing the product concentration (at the end of batch tests or in the effluent) with the reactant concentration (at the beginning of batch tests or in the influent). All error bars in the figures represent the standard deviation of the triplicate results from three individual reactors. 3 Results and discussion 3.1 Characterization by electron microscopy PdNPs coated on the membrane (with initial Na2PdCl4 concentration of 1 mM) were characterized using TEM and STEM coupled with EDX. Fig. 2 A is TEM micrograph of a nonporous polypropylene membrane coated with PdNPs (in the ellipse). We analyzed 94 nanoparticles, from the selected region shown in Fig. 2B, to obtain the particle-size distribution (Fig. 2C) and the EDX spectrum (Fig. 2D). Coated PdNPs displayed an average size of 2.6 nm, with the average particle-size ranging from 1.2 nm to 4.2 nm when the initial Na2PdCl4 concentration used for coating increased from 0.1 mM to 5.0 mM. Our previous work (Luo et al., 2020; Zhou et al., 2017) analyzed the particle size of PdNPs and confirmed that the particle size was uniformly distributed on the same coated membrane. Although several clusters of PdNPs can be observed in Fig. 2B, none of them were aggregated together; the coated PdNPs were well dispersed, suggesting that the coating process did not directly lead to particle aggregation. The EDX spectrum in Fig. 2C indicates a high purity of elemental Pd, since EDX signals corresponding to palladium oxides or other oxidized-palladium species were absent (EDX detection limit = 2%). The particle-size distribution (Fig. 2C) and EDX spectrum (Fig. 2D) confirm that small-size PdNPs well successfully coated on the membranes of the H2−MPfR.Fig. 2 TEM micrograph overviewing the nonporous polypropylene membrane coated with metallic PdNPs highlighted in the elliptical region (A) and STEM micrograph of solid-state PdNPs (B) on membrane surface. A cluster of metallic PdNPs in the squared region (B) was analyzed to obtain the particle size distribution pattern shown in (C) and the EDX spectrum shown in (D). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig 2 3.2 Kinetics of CAAs catalytic dechlorination 3.2.1 Experimental results To investigate the reaction pathways and product selectivity, we conducted batch tests with TCAA, DCAA, or MCAA as the individual reactant. Fig. 3 A shows that 98% of 0.95 mM TCAA was removed in 45 min, while DCAA, MCAA, and AA were produced simultaneously. The concentration of produced DCAA decreased to below the detection limit (15 μg/L) within 150 min. At the end of the batch test, 0.90 mM AA was produced as the main product, along with 0.05 mM MCAA as a minor byproduct. The mass balances of acetic acid and chlorine closed at the end of the batch test (Supplementary Material, Figure S1).Fig. 3 Comparisons of experimental and model-simulated results of batch experiments with 1 mM TCAA and a PdNP loading of 20.6 mg Pd/m2. (A: TCAA, (B) DCAA, and (C) MCAA. The symbols represent the experimental results, and the dashed lines represent the simulated results by first-order modeling. The error bars are the standard deviation of triplicated tests. Fig 3 When 1.06 mM DCAA was added to the reactor as the sole reactant (Fig. 3B), 99% removal of DCAA was achieved within 90 min. At the end of the batch test, 0.24 mM MCAA and 0.75 mM AA were produced as final products. In the batch tests for MCAA (Fig. 3C), only 13% of MCAA was reduced to AA in 180 min. When no H2 was supplied to the membrane, neither reaction nor adsorption of TCAA, DCAA, or MCAA occurred (Supplementary Material, Figure S2). The reductive dechlorinations of TCAA, DCAA, and MCAA were well-represented by first-order reaction kinetics, and their rate constants were 0.08, 0.06, and 0.0015 min−1. These values convert to PdNP-specific catalytic activities of 7.5, 5.6, and 0.1 L/g Pd-min, respectively. The batch-test results with different CAAs reveal that parallel and stepwise reactions happened during TCAA reductive dechlorination. Clear examples of parallel reactions are that TCAA reduction yielded DCAA and AA from the beginning, while DCAA reduction yielded MCAA and AA from the beginning. Stepwise reaction is illustrated by the lag in MCAA production from TCAA due to the need to accumulate DCAA first. 3.2.2 Modeling based on first-order reactions The reaction selectivity values a, b, c, d, and e were estimated using the experimental data from the batch tests of TCAA reductive dechlorination and the first-order rate coefficient reported in the preceding section. The best-fits values were 0.25, 0.01, 0.74, 0.28 and 0.72, respectively, and the model lines in Fig. 3 show that they represented the results well, although DCAA had systematic error for part of the TCAA experiment. The calculated reaction selectivity reveals that only 1% of TCAA was converted to MCAA during parallel reactions of TCAA reduction, while 25% of TCAA was converted to DCAA and subsequently reduced to 28% MCAA and 72% AA; this reinforces the stepwise reaction of TCAA to AA. The relatively poor fit for DCAA in the TCAA experiment probably was a result of an over-estimation of the first-order rate constants of DCAA reductive dechlorination (k2). This may have been caused by competition between TCAA and DCAA for the adsorption site on PdNP surface available to catalytic reduction; thus, the DCAA-reduction rate may have been lower than k2 during the TCAA batch tests. Our model does not account for competition for the adsorption sites on PdNPs. Competition would be a fruitful topic for future experimental and modeling study. 3.3 Pathways of TCAA dechlorination Based on the kinetic results and modeling, we propose the TCAA-transformation pathways in Fig. 4 . Pd-catalyzed reductive dechlorination of TCAA follows the Langmuir-Hinshelwood (LH) mechanism, in which adsorption of the reactants to Pd is the first step (Wu et al., 2018). In the H2−MPfR, H2 diffusing from the hollow fiber membranes adsorbs to the coated PdNPs, without any bubble formation. The adsorbed H2 molecules dissociate to become active H*ads atoms (Eq. (9)), which are very powerful reductive agents for replacing the chlorines of the CAAs (Jewell and Davis, 2006; Lien and Zhang, 2007). Dissolved TCAA also adsorbs onto the Pd0 surface and dissociates to form Pd-C bonds (Eq. (10)). The direct interaction of the Pd-C bonds with an active H*ads atom leads to stepwise Cl− release and reduction of the C all the way to AA (Eqs. (11)–(13)) (Zhou et al., 2010).(9) Pd+H2→(H2)adsPd→2(H*)ads−Pd (10) Pd+C2O2Cl3−→(C2O2Cl3−)ads−Pd (11) (C2O2Cl3−)ads−Pd+2(H*)ads−Pd→(C2HO2Cl2−)ads−Pd+Cl−+H+ (12) (C2HO2Cl2−)ads−Pd+2(H*)ads−Pd→(C2H2O2Cl1−)ads−Pd+Cl−+H+ (13) (C2H2O2Cl1−)ads−Pd+2(H*)ads−Pd→(C2H3O2−)ads−Pd+Cl−+H+ Fig. 4 Proposed pathways of TCAA reductive dechlorination catalyzed by PdNPs. H and TCAA are first adsorbed and dissociated on the PdNP. Dechlorination of TCAA on PdNPs then occurs in parallel to produce DCAA, MCAA, and AA. Any DCAA and MCAA that desorbs can be re-adsorbed further be reduced to AA. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig 4 The direct delivery of H2 through the membrane maintains a high surface site density of active H*ads atoms, and this reduces the possibility of DCAA and MCAA desorption before they are dechlorinated, which maximizes product selectivity to the desired AA. 3.4 Effective conditions for TCAA dechlorination rate and product selectivity 3.4.1 Effect of catalysts loading The solution concentration of Na2PdCl4 during PdNP coating controlled the catalyst loading on the membrane surface. In situ reduction and deposition indicated nearly 100% transfer of Pd2+ to deposited Pd0. Subsequent washing led to a loss of about 10% of the Pd0, determined by measurement of Pd after digestion of the membrane bundles. Therefore, we used 90% of loss of Pd2+ from solution to compute the Pd0 loading on the membrane surfaces; the PdNP surface loadings were 2.1, 10.4, 20.6, 40.8, and 100.2 mg Pd/m2. We used those Pd0 loadings to investigate the impact of catalyst loading on the reaction rate, catalyst-specific activity, and product selectivity in batch experiments. The results are in Fig. 5 .Fig. 5 Effects of palladium coating (A), H2 pressure (B), initial TCAA concentration (C), pH (D) and initial chloride concentration (E) on product selectivity and catalyst-specific dechlorination activity. The error bar presents the standard deviation of the triplicated tests. Fig 5 For kinetics, the first-order reaction rate constant for TCAA reduction (Fig. 5A) increased from 0.03 min−1 to 0.08 min−1 when the surface loading increased from 2 to ≥ 20 mg Pd/m2, but the catalyst-specific activity steadily decreased with higher Pd loading up to 100 mg Pd/m2. However, product selectivity to AA increased with higher surface PdNP loading. Among the different catalyst surface loadings, 20 mg Pd/m2 had the best balance of efficient TCAA removal, selectivity to AA, and the lowest catalyst loading. With lower loading, the coverage of catalysts was insufficient to support fast removal, while higher loading of PdNPs led to aggregation that led to lower PdNP-specific catalytic activity. The PdNPs in the H2−MPfR displayed catalytic-specific activity for dechlorination similar to suspended PdNPs, but 2 to 3 orders of magnitude higher compared with other immobilized PdNPs (El-Sharnouby et al., 2018; Lien and Zhang, 2007; Śrebowata et al., 2016; Zhou et al., 2010). 3.4.2 Effects of H2 pressure and initial TCAA concentration The H2 pressure in the hollow-fiber membrane controls H2 delivery capacity to the catalysts (Tang et al., 2012). Increasing the H2 pressure from 3 to 12 psig (Fig. 5B) did not affect the catalyst-specific activity or the product selectivity. This reveals that H2 pressure as low as 3 psig (or 1.2 atm absolute pressure) was sufficient to deliver enough H2 to support TCAA reductive dechlorination. The lowest H2 gas pressure we used was limited by the range of our pressure regulator. Future work could achieve lower H2 pressure by mixing H2 with N2. Decreasing the initial TCAA concentration from 1 mM to 0.01 mM (Fig. 5C) led to a small decrease in the catalyst-specific activity, from 7.7 to 6.6 L/g Pd-min, but the product selectivity to MCAA decreased from 6.7% to 1.6%. The results in Figs. 5B and 5C suggest that a higher H2-to-TCAA ratio enhanced product selectivity to AA. Also, the efficient direct delivery of H2 to the PdNPs enabled the relatively lower product selectivity to MCAA, compared with other reductive methods (Aslani et al., 2017; Liu et al., 2017; Mao et al., 2018, 2016). 3.4.3 Effects of pH on CAAs dechlorination Batch tests were conducted to investigate the effect of pH on the dechlorination of TCAA (Fig. 5D). With TCAA as the only reactant, increasing the pH from 3 to 11 greatly raised the catalyst-specific activity from 2.4 to 12.9 L/g Pd-min, but product selectivity to MCAA increased from 5.1% to 22.7%. DCAA also remained for pH 3 and 5 because of the low reaction rate. Similar pH effects also occurred when DCAA or MCAA was the only substrate (Supplementary Material, Figure S3 A&B). Based on the proposed dechlorination pathway, we hypothesize that the simultaneous increase of catalyst-specific activity and product selectivity to MCAA was caused by a higher desorption rate of partially dechlorinated products at high pH. A higher pH promotes the ionization of adsorbed CAAs (e.g.(C2H2O2Cl1−)ads−Pd), which increases electrostatic repulsion that accelerates the desorption of products without further dechlorination. In addition, the higher desorption rate opens up more active catalytic sites, which leads to a faster turnover rate. Since the concentration of reactants (CAAs in the liquid and H2 from the membrane) was much higher than the active catalytic site density on the surface of coated PdNPs, the faster turnover rate should have accelerated the catalytic reduction of TCAA at higher pH. 3.4.4 Effects of initial chloride concentration on CAAs dechlorination Increasing the initial chloride concentration from 20 to 300 mg/L (Fig. 5E) did not affect the catalyst-specific activity of TCAA reduction. However, the product selectivity to MCAA increased from 6.7% to 11.3% with higher initial chloride concentration. In contrast to TCAA, higher chloride concentration inhibited the reductive dechlorination of DCAA and MCAA (Figure S3 C&D) and led to decreases of the catalyst-specific activity from 5.6 and 0.2 L/g Pd-min to 4.8 and 0.02 L/g Pd-min, respectively. Chloride ion in aqueous phase may compete with DCAA and MCAA for adsorption onto the surface of coated PdNPs, which would lead to a higher selectivity to MCAA and to inhibition of the DCAA reductive dechlorination. 3.5 Long-term continuous dechlorination Fig. 6A presents the effluent concentrations of TCAA, DCAA, MCAA, and AA during the 50-day continuous test of TCAA reductive dechlorination in the H2−MPfR. Table 1 summarizes the performance results at steady-state for each stage. The effluent concentration of TCAA was always below the detection limit (15 μg/L), with the product selectivity to AA always greater than 90% during continuous operation. From stage I to III, when the influent concentration of TCAA was decreased from 15 mg/L to 1 mg/L, the steady-state average effluent concentration of DCAA and MCAA decreased from 200 μg/L and 309 μg/L to 39 μg/L and <5 μg/L, respectively. With the TCAA concentration of 1 mg/L, a value close to practical raw water contamination (Pressman et al., 2010; Yeh et al., 2014), the sum of effluent concentrations of CAAs was 40 μg/L below the U.S. EPA MCL of 60 μg/L for the sum of five haloacetic acids (Calafat et al., 2003; Krasner et al., 2006). When the hydraulic retention time (HRT) was reduced to 6 and 4 h in stages IV and V, the sums of CAAs were 58 μg/L and 82 μg/L. In stage VI, when the medium was made in tap water instead of ultrapure water, TCAA removal and CAAs in the effluents were similar compared to stage IV, which indicates that other constituents in tap water had no impact on TCAA reductive dechlorination. The results reveal that, for practical applications, the HRT of the bench-scale H2−MPfR can be as low as 6 h to achieve the MCL requirement, and this corresponds to a surface loading of 4.1 mg TCAA/m2/d.Fig. 6 Continuous removal of TCAA in the H2−MPfR with a Pd0 loading of 20.6 mg Pd0/m2 and supplied with 3 psig H2. (A) The effluent concentrations of TCAA, DCAA, MCAA, and AA are marked as blue, green, yellow, and gray circles, respectiveluy. The blue horizontal line represents the influent concentration of TCAA. (B) The effluent concentrations of total palladium (Pd2++ Pd0) and cumulative Pd loss during continuous operation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig 6 Table 1 Average performance of the H2−MPfR for the six steady states in continuous operation. Table 1Continuous operation HydrogenPressure TCAA Effluent Conc. Selectivity HRT Cin Cout Surface Loading Removal DCAA MCAA DCAA MCAA AA Stage hour psig mg/L mg/L mg/m2/d % mg/L mg/L % % % I 8 3 15.0 N.D 43.0 100 0.200 0.309 1.3 2.1 96.6 II 8 3 7.0 N.D 20.3 100 0.106 0.215 1.5 3.1 95.4 III 8 3 1.0 N.D 3.1 100 0.039 0.001 3.9 0.1 96.0 IV 6 3 1.0 N.D 4.1 100 0.052 0.006 5.2 0.6 94.2 V 4 3 1.0 N.D 6.1 100 0.068 0.014 6.8 1.4 91.9 VI 6 3 1.0 N.D 4.1 100 0.041 0.011 4.9 1.7 93.3 N.D. stands for “Not Detectable” or below detection limit; Conc. stands for “Concentration”. Fig. 6B, which presents the effluent concentrations of Pd, reveals that the average total palladium (Pd2++ Pd0) concentration n the effluent was about 2.4 μg/L. This corresponds to a cumulative loss of Pd less than 4% of coated PdNPs on membrane over the 50-day continuous test, and it documents minimal loss of PdNPs during long-term continuous operation of the H2−MPfR. Finally, we compare reactor performance of the H2−MPfR with other catalytic and electrochemical reactors in Table S3, which shows that the H2−MBfR offers competitive reaction kinetics for TCAA dechlorination and better product selectivity towards acetic acid. 4 Conclusion The H2−MPfR, featuring in-situ-coated PdNPs, reductively dechlorinated TCAA with high catalyst-specific rates and high selectivity to AA. Controlling the H2-delivery capacity by the H2 pressure in the fiber membrane lumen enabled efficient and safe H2 supply to the PdNPs at rates required for reductive dechlorination. Over 50-days of continuous operation, an H2−MPfR coated with 20 mg Pd/m2 and supplied with 3 psig H2 achieved greater than 99% removal of 1 mg/L TCAA, with minimal formation of DCAA and MCAA (below U.S. EPA MCL for the five haloacetic acids). Sustained continuous TCAA removal and high selectivity to AA support that the H2−MPfR is promising as a reliable catalytic reactor for removing chlorinated DBPs in practice. Compared to other processes for TCAA removal, the H2-based MPfR offers these benefits: simple design and operation based on the in-situ PdNP coating and membrane-based H2 supply; H2 gas as the only reactant; sustainable continuous operation; and AA as the major product. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix Supplementary materials Image, application 1 Acknowledgements This work was supported by the generous donations from the Swette Family Endowment, the National Science Foundation (EEC-1449500) Nanosystems Engineering Research Center on Nanotechnology-Enabled Water Treatment, the Nanotechnology Collaborative Infrastructure Southwest (NNCI-ECCS-1542160), and ASU's Fulton Chair of Environmental Engineering. We gratefully acknowledge the use of electron microscopic facilities supervised by Mr. David Lowry in the School of Life Science, and by Mr. Karl Weiss and Dr. Manuel Roldan Gutierrez in the LeRoy Eyring Center for Solid State Science, both at Arizona State University. Yuhang Cai also gratefully acknowledges the financial support from 10.13039/501100004543 China Scholarship Council (No. 201906680080). Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.watres.2021.116841. ==== Refs References Aslani H. Nasseri S. Nabizadeh R. Mesdaghinia A. Alimohammadi M. Nazmara S. Haloacetic acids degradation by an efficient Ferrate/UV process: byproduct analysis, kinetic study, and application of response surface methodology for modeling and optimization J. Environ. Manage. 203 2017 218 228 10.1016/j.jenvman.2017.07.072 28783018 Boorman G.A. Dellarco V. Dunnick J.K. Chapin R.E. Hunter S. Hauchman F. Gardner H. Mike C. Sills R.C. Drinking water disinfection byproducts: review and approach to toxicity evaluation Environ. Health Perspect. 107 1999 207 217 10.2307/3434484 10229719 Calafat A.M. Kuklenyik Z. Caudill S.P. Ashley D.L. Urinary levels of trichloroacetic acid, a disinfection by-product in chlorinated drinking water, in a human reference population Environ. Health Perspect. 111 2003 151 154 10.1289/ehp.5644 12573897 Chang A. Schnall A.H. Law R. Bronstein A.C. Marraffa J.M. Spiller H.A. Hays H.L. Funk A.R. Mercurio-Zappala M. Calello D.P. Aleguas A. Borys D.J. Boehmer T. Svendsen E. Cleaning and disinfectant chemical exposures and temporal associations with COVID-19 - National poison data system, United States, January 1, 2020- March 31, 2020. Morb. Mortal Wkly. Rep. 69 2020 496 498 10.15585/MMWR.MM6916E1 Cristóvão R.O. Tavares A.P.M. Loureiro J.M. Boaventura R.A.R. Macedo E.A. Treatment and kinetic modelling of a simulated dye house effluent by enzymatic catalysis Bioresour. Technol. 100 2009 6236 6242 10.1016/j.biortech.2009.07.007 19647998 De Corte S. Hennebel T. Fitts J.P. Sabbe T. Bliznuk V. Verschuere S. van der Lelie D. Verstraete W. Boon N. Biosupported bimetallic Pd–Au nanocatalysts for dechlorination of environmental contaminants Environ. Sci. Technol. 45 2011 8506 8513 21877727 Dincer I. Green methods for hydrogen production Int. J. Hydrogen Energy 37 2012 1954 1971 10.1016/j.ijhydene.2011.03.173 El-Sharnouby O. Boparai H.K. Herrera J. O'Carroll D.M. Aqueous-phase catalytic hydrodechlorination of 1,2-dichloroethane over palladium nanoparticles (nPd) with residual borohydride from nPd synthesis Chem. Eng. J. 342 2018 281 292 10.1016/j.cej.2018.02.006 Esclapez M.D. Tudela I. Díez-García M.I. Sáez V. Bonete P. Towards the complete dechlorination of chloroacetic acids in water by sonoelectrochemical methods: effect of the cathode material on the degradation of trichloroacetic acid and its degradation by-products Appl. Catal. B Environ. 2015 166 167 10.1016/j.apcatb.2014.10.061 66–74 Feijen-Jeurissen M.M.R. Jorna J.J. Nieuwenhuys B.E. Sinquin G. Petit C. Hindermann J.P. Mechanism of catalytic destruction of 1,2-dichloroethane and trichloroethylene over γ-Al2O3 and γ-Al2O3 supported chromium and palladium catalysts Catal. Today 54 1999 65 79 10.1016/S0920-5861(99)00169-8 Fritsch D. Kuhr K. Mackenzie K. Kopinke F.-.D. Hydrodechlorination of chloroorganic compounds in ground water by palladium catalysts: part 1. Development of polymer-based catalysts and membrane reactor tests Catal. Today 82 2003 105 118 Han B. Liu W. Li J. Wang J. Zhao D. Xu R. Lin Z. Catalytic hydrodechlorination of triclosan using a new class of anion-exchange-resin supported palladium catalysts Water Res. 120 2017 199 210 10.1016/j.watres.2017.04.059 28494246 Hennebel T. Verhagen P. Simoen H. Gusseme B.De Vlaeminck S.E. Boon N. Verstraete W. Remediation of trichloroethylene by bio-precipitated and encapsulated palladium nanoparticles in a fixed bed reactor Chemosphere 76 2009 1221 1225 10.1016/j.chemosphere.2009.05.046 19560796 Hoekstra E.J. de Leer E.W.B. Udo A.T. Mass balance of trichloroacetic acid in the soil top layer Chemosphere 38 1999 551 563 10901673 Hong Y. Song H. Karanfil T. Formation of haloacetic acids from dissolved organic matter fractions during chloramination Water Res 47 2013 1147 1155 10.1016/j.watres.2012.11.025 23245540 Hrudey S.E. Chlorination disinfection by-products, public health risk tradeoffs and me Water Res. 43 2009 2057 2092 10.1016/j.watres.2009.02.011 19304309 Jewell L.L. Davis B.H. Review of absorption and adsorption in the hydrogen-palladium system Appl. Catal. A Gen. 310 2006 1 15 10.1016/j.apcata.2006.05.012 Krasner S.W. Weinberg H.S. Richardson S.D. Pastor S.J. Chinn R. Sclimenti M.J. Onstad G.D. Thruston A.D. Occurrence of a new generation of disinfection byproducts Environ. Sci. Technol. 40 2006 7175 7185 10.1021/es060353j 17180964 Li D. Zeng S. He M. Gu A.Z. Water Disinfection Byproducts Induce Antibiotic Resistance-Role of Environmental Pollutants in Resistance Phenomena Environ. Sci. Technol. 50 2016 3193 3201 10.1021/acs.est.5b05113 26928861 Lien H.L. Zhang W.X. Nanoscale Pd/Fe bimetallic particles: catalytic effects of palladium on hydrodechlorination Appl. Catal. B Environ. 77 2007 110 116 10.1016/j.apcatb.2007.07.014 Liu Y. Mao R. Tong Y. Lan H. Zhang G. Liu H. Qu J. Reductive dechlorination of trichloroacetic acid (TCAA) by electrochemical process over Pd-In/Al2O3 catalyst Electrochim. Acta 232 2017 13 21 10.1016/j.electacta.2017.02.071 Liu Y. Yang F. Yue P.L. Chen G. Catalytic dechlorination of chlorophenols in water by palladium/iron Water Res. 35 2001 1887 1890 10.1016/S0043-1354(00)00463-2 11337833 Luo Y.-.H. Zhou C. Bi Y. Long X. Wang B. Tang Y. Krajmalnik-Brown R. Rittmann B.E. Long-term continuous co-reduction of 1, 1, 1-trichloroethane and trichloroethene over palladium nanoparticles spontaneously deposited on H2-transfer membranes Environ. Sci. Technol. 2020 Lv H. Geletii Y.V. Zhao C. Vickers J.W. Zhu G. Luo Z. Song J. Lian T. Musaev D.G. Hill C.L. Polyoxometalate water oxidation catalysts and the production of green fuel Chem. Soc. Rev. 41 2012 7572 7589 10.1039/c2cs35292c 22972187 Lv L. Jiang T. Zhang S. Yu X. Exposure to mutagenic disinfection byproducts leads to increase of antibiotic resistance in pseudomonas aeruginosa Environ. Sci. Technol. 48 2014 8188 8195 10.1021/es501646n 24933271 Lv P.L. Shi L.D. Dong Q.Y. Rittmann B. Zhao H.P. How nitrate affects perchlorate reduction in a methane-based biofilm batch reactor Water Res. 171 2020 115397 10.1016/j.watres.2019.115397 Mao R. Lan H. Yan L. Zhao X. Liu H. Qu J. Enhanced indirect atomic H∗ reduction at a hybrid Pd/graphene cathode for electrochemical dechlorination under low negative potentials Environ. Sci. Nano 5 2018 2282 2292 10.1039/c8en00727f Mao R. Li N. Lan H. Zhao X. Liu H. Qu J. Sun M. Dechlorination of Trichloroacetic Acid Using a Noble Metal-Free Graphene-Cu Foam Electrode via Direct Cathodic Reduction and Atomic H∗ Environ. Sci. Technol. 50 2016 3829 3837 10.1021/acs.est.5b05006 26977556 Nutt M.O. Heck K.N. Alvarez P. Wong M.S. Improved Pd-on-Au bimetallic nanoparticle catalysts for aqueous-phase trichloroethene hydrodechlorination Appl. Catal. B Environ. 69 2006 115 125 Pontius F.W. Complying with future water regulations J. Am. Water Work. Assoc. 91 1999 46 58 10.1002/j.1551-8833.1999.tb08599.x Pressman J.G. Richardson S.D. Speth T.F. Miltner R.J. Narotsky M.G. Hunter E.S. Rice G.E. Teuschler L.K. McDonald A. Parvez S. Krasner S.W. Weinberg H.S. McKague A.B. Parrett C.J. Bodin N. Chinn R. Lee C.F.T. Simmons J.E. Concentration, chlorination, and chemical analysis of drinking water for disinfection byproduct mixtures health effects research: U.S. EPAs four lab study Environ. Sci. Technol. 44 2010 7184 7192 10.1021/es9039314 20496936 Śrebowata A. Tarach K. Girman V. Góra-Marek K. Catalytic removal of trichloroethylene from water over palladium loaded microporous and hierarchical zeolites Appl. Catal. B Environ. 181 2016 550 560 10.1016/j.apcatb.2015.08.025 Tang Y. Zhou C. Van Ginkel S.W. Ontiveros-Valencia A. Shin J. Rittmann B.E. Hydrogen permeability of the hollow fibers used in H2-based membrane biofilm reactors J. Memb. Sci. 407 2012 176 7388 Villanueva C.M. Cantor K.P. Cordier S. Jaakkola J.J.K. King W.D. Lynch C.F. Porru S. Kogevinas M. Disinfection byproducts and bladder cancer: a pooled analysis Epidemiology 15 2004 357 367 10.1097/01.ede.0000121380.02594.fc 15097021 Wang Jiao Shen J. Ye D. Yan X. Zhang Y. Yang W. Li X. Wang Junqi Zhang L. Pan L. Disinfection technology of hospital wastes and wastewater: suggestions for disinfection strategy during coronavirus Disease 2019 (COVID-19) pandemic in China Environ. Pollut. 262 2020 114665 10.1016/j.envpol.2020.114665 Wegner K. Ly H.C. Weiss R.J. Pratsinis S.E. Steinfeld A. In situ formation and hydrolysis of Zn nanoparticles for H2 production by the 2-step ZnO/Zn water-splitting thermochemical cycle Int. J. Hydrogen Energy 31 2006 55 61 10.1016/j.ijhydene.2005.03.006 Wu Z. Pan T. Chai Y. Ge S. Ju Y. Li T. Liu K. Lan L. Yip A.C.K. Zhang M. Synthesis of palladium phosphides for aqueous phase hydrodechlorination: kinetic study and deactivation resistance J. Catal. 366 2018 80 90 10.1016/j.jcat.2018.07.040 Xu L. Bhandari S. Chen J. Glasgow J. Mavrikakis M. Chloroform Hydrodechlorination on Palladium Surfaces: a Comparative DFT Study on Pd (111), Pd (100), and Pd (211) Top. Catal 2020 1 15 Yeh R.Y.L. Farré M.J. Stalter D. Tang J.Y.M. Molendijk J. Escher B.I. Bioanalytical and chemical evaluation of disinfection by-products in swimming pool water Water Res 59 2014 172 184 10.1016/j.watres.2014.04.002 24793115 Zhao H.P. Ontiveros-Valencia A. Tang Y. Kim B.O. Ilhan Z.E. Krajmalnik-Brown R. Rittmann B. Using a two-stage hydrogen-based membrane biofilm reactor (MBfR) to achieve complete perchlorate reduction in the presence of nitrate and sulfate Environ. Sci. Technol. 47 2013 1565 1572 10.1021/es303823n 23298383 Zhou C. Ontiveros-Valencia A. Cornette de Saint Cyr L. Zevin A.S. Carey S.E. Krajmalnik-Brown R. Rittmann B.E. Uranium removal and microbial community in a H2-based membrane biofilm reactor Water Res 64 2014 255 264 10.1016/j.watres.2014.07.013 25073000 Zhou C. Wang Z. Ontiveros-Valencia A. Long M. Lai C. yu, Zhao H. ping, Xia S. Rittmann B.E. Coupling of Pd nanoparticles and denitrifying biofilm promotes H2-based nitrate removal with greater selectivity towards N2 Appl. Catal. B Environ. 206 2017 461 470 10.1016/j.apcatb.2017.01.068 Zhou T. Li Y. Lim T.T. Catalytic hydrodechlorination of chlorophenols by Pd/Fe nanoparticles: comparisons with other bimetallic systems, kinetics and mechanism Sep. Purif. Technol. 76 2010 206 214 10.1016/j.seppur.2010.10.010

33503571

PMC9753135

NO-CC CODE

2022-12-16 23:25:30

no

Water Res. 2021 Mar 15; 192:116841

utf-8

Water Res

10.1016/j.watres.2021.116841

oa_other

==== Front Res Policy Res Policy Research Policy 0048-7333 1873-7625 The Author(s). Published by Elsevier B.V. S0048-7333(21)00199-2 10.1016/j.respol.2021.104403 104403 Short Communication The impact of the pandemic-enforced lockdown on the scholarly productivity of women academics in South Africa Walters Cyrill a⁎ Mehl Graeme G. a Piraino Patrizio b Jansen Jonathan D. a Kriger Samantha c a Stellenbosch University, Faculty of Education, Stellenbosch, 7600, South Africa b University of Notre Dame, Keough School of Global Affairs, Indiana, 46556, USA c Cape Peninsula University of Technology, Faculty of Education, Mowbray, 8000, Cape Town ⁎ Corresponding author. 25 10 2021 1 2022 25 10 2021 51 1 104403104403 1 11 2020 6 10 2021 10 10 2021 © 2021 The Author(s) 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The underrepresentation of women in research is well-documented, in everything from participation and leadership to peer review and publication. Even so, in the first months of the COVID-19 pandemic, early reports indicated a precipitous decline in women's scholarly productivity (both in time devoted to research and in journal publications) compared to pre-pandemic times. None of these studies, mainly from the Global North, could provide detailed explanations for the scale of this decline in research outcomes. Using a mixed methods research design, we offer the first comprehensive study to shed light on the complex reasons for the decline in research during the pandemic-enforced lockdown among 2,029 women academics drawn from 26 public universities in South Africa. Our study finds that a dramatic increase in teaching and administrative workloads, and the traditional family roles assumed by women while “working from home,” were among the key factors behind the reported decline in research activity among female academics in public universities. In short, teaching and administration effectively displaced research and publication—with serious implications for an already elusive gender equality in research. Finally, the paper offers recommendations that leaders and policy makers can draw on to support women academics and families in higher education during and beyond pandemic times. Keywords Female academics Gender in academia Careers Childcare COVID-19 Gender gap ==== Body pmc1 Introduction It is well-documented that there is a gender gap within scientific research and publication (Beaudry and Larivière, 2016; Coe et al., 2019; Huang et al., 2020; Helmer et al., 2017; Holman et al., 2018; Huang et al., 2020; Lerback and Hanson, 2017; Lerchenmueller and Sorenson, 2018; Mason et al., 2013; Mills, 2014) and that the reasons for such inequality are systemic and institutionalized (Coe et al., 2019; Lundine et al., 2019). Studies within the life sciences and STEM fields further demonstrate a gender disparity in these fields (Beaudry and Larivière, 2016; Graddy-Reed et al., 2019; Lerchenmueller and Sorenson, 2018); this gap exists in South Africa as elsewhere (Beaudry et al., 2018; Coe et al., 2019). Emerging evidence suggests that the COVID-19 pandemic has exacerbated this inequality (Amano-Patino et al., 2020; Fazackerley, 2020; Myers et al., 2020; Viglione, 2020), as well as disrupted the research enterprise globally (Adams-Prassl et al., n.d.; Collins et al., 2020; Myers et al., 2020; Nash and Churchill, 2020). However, to date there has been no systematic research that provides a detailed account of, and explanations for, the decline in research activity and outcomes for women academics, particularly outside of the Global North. We believe it is important that the South African experience be represented in the literature, since studies from the United States, Europe (Adams-Prassl et al., n.d.; Amano-Patino et al., 2020; Myers et al., 2020), and Australia (Nash and Churchill, 2020) have dominated published research inside relatively well-resourced institutions. For this study, we conducted a large-scale survey of all female academic staff across South Africa's 26 public universities during the period of the government-enforced lockdown, which began on 26 March 2020 (“Coronavirus: President Ramaphosa announces a 21-day lockdown,” n.d., “COVID-19 South African resource portal,” n.d.), and during which all non-essential businesses, schools, and public universities were closed, and academics were constrained to work from their homes. While the direct health implications of the lockdown have been profound, the disintegration of economic, work, and school structures and the closure of childcare facilities have altered the ways in which academics work. The COVID-19 pandemic and the response to it by female academic staff will affect the higher education sector and scientific establishment for years to come, exacerbating the preexisting dominance of males within scientific and medical fields. This study seeks to elucidate how working female academics are managing the tasks of work while negotiating childcare, homeschooling, cooking, cleaning, and other duties. This study was completed over the period of the various stages of the lockdown, from the initial hard lockdown (“stage 5″) in March through August 2020 (under “stage 2″, since 18 August). During this period, the population of women academics in South Africa ranged between 24,332 and 25,857 people, depending on resignations, secondments, and recruitment of personnel. As of September 1st, a total of 2029 full responses were received from women at different career stages. Thus, an average of 8.3% of the women academics in the national higher education system responded to the survey. The largest numbers of responses per institution were from the University of South Africa (UNISA), with 287 responses; the University of Pretoria (UP), with 185; Stellenbosch University (SU), with 172; and the University of Cape Town (UCT), the continent's top university (“World University Rankings”, 2020), with 111. To protect privacy, respondents are not identifiable beyond their institution, and no response will be attributed to any particular university in this paper. The career stages of respondents were evenly spread, with the largest group of respondents (29.8%) in the 0–5 years range of academic appointment. Although the South African higher education landscape has often been categorized by type of institution (e.g., by language of instruction or historical racial orientation) in order to shed light onto the spectrum of differentiation across universities, this study did not concern itself with the type of institution, but rather examined the sector as a whole; therefore, survey participants were not limited by nationality, race, rank, or terms of appointment. The survey questionnaire consisted of 12 Likert-scale questions, followed by an open-ended section that allowed for detailed, unlimited responses by the participants; this section provided a substantial volume of qualitative data, which were coded by theme and then analyzed. Ethical clearance was granted by the host institution (Stellenbosch University), followed by gatekeeper clearance from the other 25 universities; in one case, the research team was simply given permission by the university to contact its female academic staff directly, whereas in most cases each institution distributed the survey to their female academic staff, providing an introductory cover letter with an electronic link to the survey instrument. This study expands the understanding of how the pandemic is affecting scholarly output, as well as the career trajectory of women in university-based research. Since the South African sample is drawn from a diverse range of historically white and black universities with markedly different research capacities and outputs, it offers unique insights into the academic impacts of COVID-19 in a developing country context with a focus on women academics representing the full range of scholarly disciplines. 2 Results The single most important variable impacting the academic work of female academics appears to be having younger or multiple dependents in the home. Overall, the pandemic appears to have most affected academic work among women with children, with 54% of respondents having children living at home. Further analyses of the data suggests that those who found academic work extremely difficult were those with children under 6 years of age (see Fig. 1 ), as well as those who had children at school. It is evident from the qualitative data that the age and educational stage of children were significant factors in the decline in productivity among female academics. The demands of caring for toddlers, as well as schools’ expectations of homeschooling, took a toll on respondents. Academic mothers were caught up in the demands of competing roles, such as teaching online, nurturing vulnerable students, comforting anxious children, taking care of toddlers, and finding time to do research and writing. Doing academic work was extremely difficult for most.Fig. 1 Share of women who found academic work “extremely difficult” by ages of children in the home Notes: Authors’ calculations from survey data. Total number of respondents = 564 of 2029. Fig 1 The two at-home responsibilities that had the highest impacts on women's academic work during lockdown were childcare (in the case of toddlers) and assisting with schoolwork (in the case of school-age children). While the pandemic seems to have affected women academics in various ways, when respondents were asked which responsibilities (food preparation, housework, etc.) impacted their academic work, it was clear that schoolwork and childcare were the dominant factors. Overall, 42.7% of respondents with children said that schoolwork had a very high impact, and 43.8% said childcare did. While housework and food preparation are significant factors, when the high- and very-high scores were examined closely, childcare was found to account for 74.6% of the responses, with schoolwork at 68.8%, housework at 66.8%, food preparation at 58.9%, and getting supplies at 44.8%. The contrast is starker when one analyses the subset of respondents who had toddlers: 94.5% of respondents with toddlers (children <6) indicated that childcare had a high to very high impact on academic work. Those with toddlers found that other responsibilities also affected their work significantly. On the other hand, respondents with no children felt the impact of other responsibilities to be much lower, as can be seen in Fig. 2 .Fig. 2 Share of women who found that responsibilities had a high/very high impact on academic work by children's age Notes: Authors’ calculations from survey data. A total of 382 women had toddlers (< 6 years old), 798 had school children (between 6 and 18 years old), and 941 had no children. The sum of these respondents does not equal 2029, as some respondents had both toddlers and school children. Fig 2 A key finding of the survey is that the sharp increase in the demands on teaching time during lockdown has effectively displaced the available research time among female academics. Academics perform many different roles, including teaching, research, grant-proposal writing, administrative duties, and other tasks, depending on their rank and discipline. In this survey, the data demonstrate that the distribution of teaching and research was not at all even. Fig. 3 demonstrates that academic time was mostly taken up by teaching online, rather than research. In the qualitative section of the survey, participants lamented the effort required to adjust to the new mode of teaching online. Just over half of the respondents (50.10%) indicated that they spent more than 80% of their total work time teaching online.Fig. 3 Share of women by percentage of time spent on research and teaching related activities Notes: Authors’ calculations from survey data. Total number of valid responses = 1878. Fig 3 The overwhelming majority of women (80.3%) believe that it has been “more” to “much more” difficult for women than for men to do academic work during the lockdown period (see Fig. 4 ). The qualitative analyses suggest that the pandemic has affected researchers differently according to their disciplines. Those in the “bench sciences,” such as chemistry, biological sciences, and biochemistry, were explicit in stating that the closure of laboratories or facilities affected their research productivity. Disciplines that are less lab- and equipment-intensive were also affected; however, these cases were often related to individual circumstances, such as the ability to do fieldwork in particular social science fields.Fig. 4 Share of respondents by perceived relative difficulty of doing academic work for women compared to men Notes: Authors’ calculations from survey data. Total number of respondents = 2029. Fig 4 Fig. 5 shows that a large majority of women academics (72.5%) reported an increase (more/much more) in their administrative workloads during lockdown, with direct implications for available teaching and research time. Only 10.2% of respondents reported that the amount of such work was less (easier/somewhat easier) than before the pandemic lockdown, and 17.3% reported that it remained the same. This result may appear counterintuitive at first, as one might expect that the pause in various activities under lockdown would imply a lighter administrative burden. Our qualitative analysis sheds light on the factors related to this observed increase in administrative tasks. In particular, respondents reported an increase in (i) meetings; (ii) email correspondence; (iii) time devoted to transitioning courses and assessments online; and (iv) time spent cancelling some projects, pivoting others, and reporting requirements on COVID itself. These insights into the escalation of administrative workloads experienced by women are especially important for progress within the scientific enterprise.Fig. 5 Share of respondents by volume of administrative duties during the lockdown Notes: Authors’ calculations from survey data. Total number of respondents = 2029. Fig 5 Most women (75.1%) indicated that doing their academic work (teaching and research) was “somewhat” to “extremely” difficult during the lockdown, while 16.6% reported that it was relatively easier. In further analyses of participants who indicated that work was relatively easier, it became evident that these perceptions were correlated to the following factors: having children and the children's ages; career stages; commuting conditions; and working arrangements prior to lockdown. Moreover, more than half of the women in this study (56.5%) reported having “less time” or “no time” available for academic work, while 31.4% indicated they had more time (some extra/much more) for their academic workload. It is noteworthy that the survey did not find any marked differences between the time available for academic work and the career stage (years of academic incumbency) of respondents, with more experienced academics having only a slight advantage in their available time (see Fig. 6 ).Fig. 6 Share of respondents grouped by career stages and amount of time available to do academic work Notes: Authors’ calculations from survey data. Total number of respondents = 2029. Fig 6 Overall, a total of 40.5% of participants indicated they required much more or significantly more emotional support as working academics to cope with the demands of the job, while 25.8% indicated they required the same amount of support as before (see Fig. 7 ). Several respondents expressed feelings of unending exhaustion, which reduced their ability to focus and to be productive. The feeling of despair and a sense of the unfairness of workload distribution was a key theme emerging from the data. As shown in Fig. 8 , about 48% of participants indicated that they felt “very anxious and concerned” or “somewhat concerned” in continuing with their academic work, given their personal concerns about the pandemic, while (perhaps surprisingly) an exactly equal number felt OK or good. Further analyses of the data make clear that it is the individual circumstances of the female academic that often explain the emotional toll of the enforced lockdown. Themes such as childcare and eldercare added additional and heightened stress levels. These findings are consistent with a recent study of 59 higher education institutions in the UK, which showed evidence of an escalation in poor mental health among university staff (Liz, 2019).Fig. 7 Share of respondents by amount of emotional support needed during the lockdown Notes: Authors’ calculations from survey data. Total number of respondents = 2029. Fig 7 Fig. 8 Share of respondents by feelings of emotional wellness experienced during the lockdown Notes: Authors’ calculations from survey data. Total number of respondents = 2029. Fig 8 The key finding of the study is that the lockdown has had a profound effect on women's academic productivity, with 31.6% reporting having made “no progress” and 21.2% having made only “some progress” towards completing a significant academic product. This will likely affect the prospects of female academics for promotion and advancement. Institutions may need to track these effects and provide support through policies to protect and nurture the sustainability of women's careers in academia. Indeed, many women in the study (48.1%) indicated that the lockdown would impact negatively on their academic career prospects. Leaders in academic institutions need to be aware that female academic staff view the lockdown as yet another barrier to equity, and to consider the effects of the pandemic on career challenges in recruitment and promotion decisions (Fig. 9, Fig. 10 ).Fig. 9 Share of respondents by progress on completing a significant academic product during the lockdown Notes: Authors’ calculations from survey data. Total number of respondents = 2029. Fig 9 Fig. 10 Share of respondents by perceived impact of the lockdown on their career prospects Notes: Authors’ calculations from survey data. Total number of respondents = 2029. Fig 10 As mentioned above, the questionnaire asked whether the respondents felt emotionally well enough to do their academic work, given their concerns about the pandemic. To further analyze the correlates of women's emotional wellness, we estimated a simple logit model, defining the composite variable Stress as being equal to 1 when a respondent reported being either “very anxious and concerned” or “somewhat concerned” emotionally. Table 1 reports the estimated coefficients from a logit model of various factors explaining women's emotional stress. The results show that women who reported experiencing heightened difficulty doing their academic work from home during the lockdown were almost 20% more likely to report being emotionally unwell. In addition, women who experienced an increase in their administrative duties during the pandemic were 7% more likely to report being stressed. Positive and significant coefficients are also found on variables describing some of the principal factors impairing academic work. Women who considered doing housework and helping children with schoolwork as highly distracting were about 8% more likely to report stress. A smaller but also significant association is found between stress and having to procure groceries/supplies for the household. Other significant predictors of women's emotional wellbeing were having more time during lockdown to do academic work (8 percent negative association with stress) and dedicating additional time to research (where a 10% increase in research time is associated with a 1% increase in the likelihood of being stressed).Table 1 What explains women's stress? Table 1VARIABLES Coefficient Academic work difficult 0.196*** (0.0285) Admin duties increase 0.073*** (0.0265) Time during lockdown −0.0819*** (0.0255) Research time 0.00108** (0.000416) Academic stage: 0–5 years 0.0469 (0.0306) Academic stage: 6–10 years 0.0127 (0.0338) Academic stage: 11–15 years −0.0216 (0.0338) No help at home −0.0224 (0.0225) Children −0.0529 (0.0354) Childcare −0.004 (0.0345) Housework 0.0787*** (0.027) Food preparation 0.0261 (0.0269) School work 0.0775** (0.0334) Grocery/supplies 0.0474** (0.0235) Observations 1878 Logit model for binary variable: Stress. Standard errors in parentheses. Reference category for academic stage: 16+ years. *** p<0.01, ** p<0.05, * p<0.1. Table 2 presents the estimates from an ordered logit model of various factors explaining academic productivity. Here, the dependent variable is the response to a question on the respondent's progress towards a significant “academic product” during the lockdown period; the answer could range from no progress at all to full completion, for a total of five categories. Unsurprisingly, women who experienced heightened difficulty in doing their academic work from home during the lockdown were significantly less likely to report being able to make progress on an academic product. The opposite is true for women who reported having had more time during the lockdown to do work; they were significantly more likely to have made progress. Time dedicated to research also has a positive and significant association with making progress on academic work. In addition, table 2 shows that more junior academics were less likely to make progress relative to their more senior colleagues. Finally, women who did not have any help at home also appear to have been penalized in terms of their academic productivity.Table 2 What explains women's academic productivity? Table 2VARIABLES Acad_prod Academic work difficult −0.456*** (0.110) Admin duties increase −0.0438 (0.101) Time during lockdown 0.790*** (0.0976) Research time 0.0241*** (0.00168) Academic stage: 0–5 years −0.296** (0.116) Academic stage: 6–10 years −0.225* (0.123) Academic stage: 11–15 years −0.0399 (0.129) No help at home −0.166* (0.0865) Children −0.120 (0.134) Childcare −0.0817 (0.133) Housework −0.155 (0.105) Food preparation −0.0478 (0.105) School work 0.151 (0.133) Grocery/supplies −0.0992 (0.0913) Observations 1878 Ordered logit model of ordinal variable: academic productivity. Standard errors in parentheses. Reference category for academic stage: 16+ years. *** p<0.01, ** p<0.05, * p<0.1. It should be stressed that the results from Tables 1 and 2 only reveal correlations, not causation. The logit models, however, show several statistically significant associations. These results provide further descriptive evidence of the challenges faced by female academics during the pandemic-enforced lockdown in South Africa. 3 Findings from the qualitative data In the open-ended section of the survey, encoded words and phrases were analyzed using Atlas.ti software. A conventional content analysis was performed, in which codes were extracted from the text data. This form of analysis best suited the study, as it did not bind the researchers to a particular theory or confine them to the counting and comparison of keywords. Instead, the approach was illustrative of a commitment towards understanding the individual and subjective viewpoints of women academics (Flick, 2015). Several powerful themes emerged from the analysis; one particularly strong theme was that of “guilt.” Collins et al. (2020) illustrates emphatically how policy affects the experience of “maternal guilt.” Similarly, many of our respondents struggled to balance employment, motherhood, domestic tasks, and caregiving. The following quotation sums up the overwhelming emotions of the respondents:The lockdown magnified my experiences pre-lockdown as it relates to being a female academic … where most have used this opportunity to reconnect with their children, I have been overwhelmed by feelings of guilt, depression, and anxiety at not being able to juggle everything. A large part of academic guilt described by the respondents related to academic mothers who are caught up in the demands of competing roles, such as teaching online, nurturing vulnerable students, comforting anxious children, taking care of toddlers, and trying to jumpstart research and writing. Several studies have emerged from the community of woman academics reporting patterns of struggle with the increased pressures of balancing parenthood and professional demands (Boncori, 2020; Gourlay, 2020; Guy and Arthur, 2020). The closure of public schools and the loss of formal childcare services during the pandemic is a major reason for the increased pressure on working mothers. Further, the dynamics of using technology and working online are adding to the stresses. For woman academics who are mothers, the threat to emotional wellbeing extends into the home. While the intensity of childcare has been noted in the quantitative data, the open-ended survey indicated that there was an immediate impact on the home as an organized space to accommodate not only the individual woman academic's work, but also on how that contained environment had to be reorganized in relation to others, i.e., children, spouses, partners, parents, and, sometimes, extended family. The participants in this study reported struggling with how best to manage an externally enforced work “flexibility”: clearly, flexible arrangements bring their own challenges, and they are not suitable for everyone. The data indicates that the concepts of home and working from home were fundamentally unsettled by the pandemic-enforced lockdown. The enduring concept of home as a place of refuge from the outside world was replaced with a new and still-unsettled notion of home as a congested, competitive, and constrained place for women's academic work. 4 Discussion Although the respondents in this study are based in South Africa, it is evident from this and prior research that the pandemic has had an effect on the academic enterprise globally. Indeed, circumstances will continue to evolve as the stages of lockdown change, and the full impacts of the pandemic on the scientific enterprise remain to be seen. Concerning the division between time spent teaching versus time spent on research, it is noteworthy that a recent report indicates that, across the African continent, teaching (including supervision of graduate students) takes up an average of 67.9% of academics’ time, while research amounts to 32.1% (Beaudry et al., 2018). There is a recurring complaint within the academy that the hours required for teaching are overwhelming, and that research is expected to be done in one's spare time. This seems to be a major constraint. Respondents gave expression to the harsh reality of advancement in South African universities, as well as to the almost exclusive emphasis that is placed on “research outputs,” even if promotion policy documents pay lip service to the importance of teaching and service in the formal metrics. Even for those aiming for advancement at the senior levels, the prospects are still slim, given the massive increase in workloads during the pandemic lockdown, such as the time-consuming conversion of face-to-face lectures into online learning resources and the demands of caring for small children and managing a household. Anecdotal evidence indicates that there has been increased research productivity during the pandemic across some disciplines, but fewer submissions and publications by female academics (Amano-Patino et al., 2020; Viglione, 2020). It is important to note that the bibliometric data used for these studies cannot capture the career dynamics of teaching, which has had a profound impact on our respondents during the lockdown. Thus, in addressing and improving career prospects of female academics, institutions may need a durable and sustainable approach in alleviating the teaching load. The collapsing of the academic workspace is a new phenomenon. Before the lockdown, some women academics worked from home because it was a comfortable workspace or because it allowed for some form of refuge. Now the situation is different, and this reorganization of space was unanticipated by many. In a recent article shared widely across academia, the social demographer Alessandra Minello (2020) aptly describes the additional social and emotional labor required by women in the academy and how those requirements can block academic advancement:Academic work—in which career advancement is based on the number and quality of a person's scientific publications, and their ability to obtain funding for research projects—is basically incompatible with tending to children … [while] [t]hose with fewer care duties are aiming for the stars. The data in our study correlate with Minello's observations and suggest that there will be consequences in advancement and promotion for South African female academic staff after the enforced lockdown. The findings in this research speak to the precarity of women's academic work as they experience and articulate a sense of instability, and even perilousness, in terms of their academic futures. A major theme that emerges is how women academics’ role as nurturers comes to play a critical part in the intersecting functions of caring for both their students and their families through the period of the pandemic-enforced lockdown. What this study demonstrates is exactly how the emotional, psychological, and educational needs of students draw academic women into extensive nurturing roles, beyond caring for their families, that impact negatively on academic work. It also shows the workings of the symbiotic relationship of giving care (by women academics) and requiring care (by students) in a pandemic. Furthermore, the lockdown has put particular strain on female academics employed on soft funding, as well as those who are in academic appointments conditioned upon the continuation of postgraduate studies. A key factor in maintaining and enhancing the quality of the higher education sector is the quality of the faculty members. We call on institutional leaders, science councils, academic societies, and funding bodies to implement policies to mitigate the career risks that female academics encountered during the enforced lockdown. Importantly, it is not only the introduction of new policies but also the attitudes towards those policies that needs attention. Given the challenge of the unequal effects of the pandemic on female academics, there is a critical need for not only universities but also scientific and medical councils to present a united voice for the support of women academics. Achieving gender equality within the academic enterprise requires a wide set of tools to be utilized and policies to be implemented. It requires institutional commitment, as well as knowledge and competency in effecting organizational change. 5 Potential policies and practices In this section of our paper, we wish to suggest several voluntary policies and practices that could work towards change in achieving gender equality within the academic enterprise. These recommendations are not intended to be an exhaustive list, but rather some of the tools that might provide solutions to the effects of the pandemic on women academics. There is no doubt that institutional leaders and policy makers have a major role to play in shifting the norms, and they must respond to mitigate the impact of the pandemic on woman academics. Acknowledge the problem in university-wide communication. Acknowledgement is an important driver for organizational change and is essential in driving appropriate behaviors in various contexts. It is critical that leadership and policy makers increase awareness of the impossible choices women academics have faced and are facing during the pandemic. Management expectations should be moderated, from the top down, in ways that recognize the exceptional circumstances imposed by the pandemic lockdown. In addition to acknowledging the problem, it is important that communications be clear, consistent, and empathetic throughout and beyond the lockdown crisis. Adjust timelines for the appointment and advancement of women academics, e.g., probation, promotion, and contract renewals. In academia, productivity represents an important determinant in promotion and recruitment. Productivity, particularly in the sciences, means publications. Overall, our study suggests that productivity has declined, and this research makes clear that the pandemic has had a detrimental effect on the productivity of women. Thus, the already well-known “productivity paradox” (Lerchenmueller and Sorenson, 2018) has been exacerbated by the enforced lockdown. Providing for research assistantships to support women academics who are active in research and adjusting application and advancement forms to allow for explanations of lapses in productivity are mechanisms that could address and remove barriers for women academics. A significant number of responses in the open-ended section of the survey expressed feelings of exhaustion and the sense of a reduction in the ability to focus due to childcare and eldercare issues. The lack of provision for childcare appears to be approaching crisis level. Leaders should invest in childcare support (e.g., in the form of salary supplements) for academic women working from home, as well as provide on-campus childcare facilities for women staff. The economist Betsy Stevenson (2020) recently noted of pregnant women and working mothers whose children are too young to manage on their own that “we could have an entire generation of women who are hurt … they may spend a significant amount of time out of the work force, or their careers could just peter out in terms of promotions.” Reflection by leaders of all genders in higher education is required to create a workplace that could prioritize women with children. Also required is a commitment to ongoing institutional research into the problem, so that relevant data can inform senior management deliberations on a regular basis. Legislative approaches to addressing social inequities have been employed in South Africa since the fall of Apartheid. These type of social policy experiences in the Global South might provide insights into the efficacy of, for example, using quotas to bring about change. However, despite legislative prescriptions, in South Africa at this time even the limited gains made in the past decades are at risk of being rolled back, including the incomplete transition of women into truly equal roles in the labor market. Ultimately, what matters more than legislation and policy is the culture surrounding them. Leadership will need the capacity to alleviate anxiety and fear and to project messages of compassion and care. Over the long term, forced structural and cultural changes that could benefit women—a better child care system; more flexible work arrangements; an even deeper appreciation of the sometimes overwhelming demands of childcare and eldercare—could go a long way towards resolving the unequal effects of the enforced pandemic lockdown. 6 Conclusion The pandemic poses a lasting threat to gender equality in academia. One of the earliest studies in the wake of COVID-19 pointed to the pandemic's “substantial impact on research worldwide, which we do not capture” (Myers et al., 2020). This study offers the first comprehensive account of the pandemic's impact in an African country, which not only confirms what we knew from existing research but extends our understanding of the effects of lockdown on women's academic work. In summary, this study gives evidence that the single most important factor that negatively impacts on women's academic work during the pandemic lockdown is the presence of younger children in the home; that the escalating demands of remote teaching and administration effectively displaced time for research, writing and publication; and that most women perceive that doing academic work from home has been “more” to “much more” difficult than for men. What we now know from this study is that the burden of inequality during the pandemic lockdown has immediate consequences for women's emotional health and wellness, and longer-term implications for their academic career prospects, given the sharp decline in their research productivity in the COVID-19 period. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix Supplementary materials XML, APPLICATION 1 Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.respol.2021.104403. ==== Refs References Amano-Patino, N., Faraglia, E., Giannitsarou, C., Hasna, Z., 2020. Who is doing new research in the time of COVID-19? Not the female economists [WWW Document]. URL https://voxeu.org/article/who-doing-new-research-time-covid-19-not-female-economists (accessed 9.19.20). Beaudry C. Larivière V. Which gender gap? Factors affecting researchers’ scientific impact in science and medicine Res. Policy 45 2016 1790 1817 Beaudry, C., Mouton, J., Prozesky, H., 2018. The next generation of scientists in Africa. Boncori I. The never-ending shift: a feminist reflection on living and organizing academic lives during the coronavirus pandemic Gender Work Organ 27 2020 677 682 Coe I.R. Wiley R. Bekker .L..-G. Organisational best practices towards gender equality in science and medicine Lancet 393 2019 587 593 30739694 Collins C. Landivar L.C. Ruppanner L. Scarborough W.J. COVID-19 and the gender gap in work hours Gender, Work Organ 2020 10.1111/gwao.12506 n/a Fazackerley A. Women's research plummets over lockdown-but articles from men increase Guard 2020 May 12 Flick U. Introducing Research methodology: a Beginner's Guide to Doing a Research Project 2nd ed. 2015 Sage Publications London Gourlay L. Quarantined, sequestered, closed: theorising academic bodies under Coronavirus lockdown Postdigitakl Sci. Educ. 2 2020 791 811 Graddy-Reed A. Lanahan L. Eyer J. Gender discrepancies in publication productivity of high-performing life science graduate students Res. Policy 48 2019 103838 Guy B. Arthur B. Academic motherhood during Covid-19. Navigating our dual roles as educators and mothers Gender Work Organ 27 2020 887 899 Helmer M. Schottdorf M. Neef A. Battaglia D. Gender bias in scholarly peer review Elife 6 2017 e21718 28322725 Holman L. Stuart-Fox D. Hauser C.E. The gender gap in science: how long until women are equally represented? PLoS Biol. 16 2018 10.1371/journal.pbio.2004956 Huang J. Gates A.J. Sinatra R. Barabási A.-.L. Historical comparison of gender inequality in scientific careers across countries and disciplines Proc. Natl. Acad. Sci. 117 2020 4609 4616 10.1073/PNAS.1914221117 32071248 Lerback J. Hanson B. Journals invite too few women to referee Nature 541 2017 455 457 28128272 Lerchenmueller M.J. Sorenson O. The gender gap in early career transitions in the life sciences Res. Policy 47 2018 1007 1017 Liz, M., 2019. Pressure vessels: the epidemic of poor mental health among higher education staff. Lundine J. Bourgeault I.L. Clark J. Heidari S. Balabanova D. Gender bias in academia Lancet 393 2019 741 743 Mason M.A. Wolfinger N.H. Goulden M. Do babies matter?: Gender and Family in the Ivory Tower 2013 Rutgers University Press Mills M.J. Gender and the Work-Family experience: an Intersection of Two Domains 2015th ed. 2014 Springer International Publishing AG Cham 10.1007/978-3-319-08891-4 Cham Minello A. The pandemic and the female academic Nature 2020 Myers K.R. Tham W.Y. Yin Y. Cohodes N. Thursby J.G. Thursby M.C. Schiffer P. Walsh J.T. Lakhani K.R. Wang D. Unequal effects of the COVID-19 pandemic on scientists Nat. Hum. Behav. 2020 1 4 31965067 Nash M. Churchill B. Caring During COVID-19: A gendered Analysis of Australian university Responses to Managing Remote Working and Caring Responsibilities 2020 Gender, Work Organ Stevenson, B. 2020. How the child care crisis will distort the economy for a generation. Viglione G. Are women publishing less during the pandemic? Here's what the data say Nature 581 2020 365 366 32433639 World University Rankings 2020 | Times Higher Education (THE) [WWW Document], 2020. URL https://www.timeshighereducation.com/world-university-rankings/2020/world-ranking#!/page/0/length/25/sort_by/rank/sort_order/asc/cols/stats.

0

PMC9753136

NO-CC CODE

2022-12-16 23:25:31

no

Res Policy. 2022 Jan 25; 51(1):104403

utf-8

Res Policy

10.1016/j.respol.2021.104403

oa_other

==== Front J Public Econ J Public Econ Journal of Public Economics 0047-2727 0047-2727 Elsevier B.V. S0047-2727(21)00107-9 10.1016/j.jpubeco.2021.104471 104471 Short Communication The impact of the Federal Pandemic Unemployment Compensation on job search and vacancy creation☆ Marinescu Ioana ab⁎ Skandalis Daphné cd⁎ Zhao Daniel e a University of Pennsylvania, United States b NBER, United States c University of Copenhagen, Denmark d IZA, Germany e Glassdoor, Inc., United States ⁎ Corresponding author at: University of Pennsylvania, United States. 13 7 2021 8 2021 13 7 2021 200 104471104471 7 4 2021 17 6 2021 21 6 2021 © 2021 Elsevier B.V. All rights reserved. 2021 Elsevier B.V. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. During the COVID-19 pandemic, the Federal Pandemic Unemployment Compensation (FPUC) increased US unemployment benefits by $600 a week. Theory predicts that FPUC should decrease job applications, while the effect on vacancy creation is ambiguous. We estimate the effect of FPUC on job applications and vacancy creation week by week, from March to July 2020, using granular data from the online jobs platform Glassdoor. We exploit variation in the proportional increase in benefits across local labor markets. To isolate the effect of FPUC, we flexibly allow for different trends in local labor markets differentially exposed to the COVID-19 crisis. We verify that trends in outcomes prior to the FPUC do not correlate with future increases in benefits, which supports our identification assumption. First, we find that a 10% increase in unemployment benefits caused a 3.6% decline in applications, but did not decrease vacancy creation; hence, FPUC increased labor market tightness (vacancies/applications). Second, we document that tightness was unusually depressed during the FPUC period. Altogether, our results imply that the positive effect of FPUC on tightness was likely welfare improving: FPUC decreased competition among applicants at a time when jobs were unusually scarce. Our results also help explain prior findings that FPUC did not decrease employment. Keywords Unemployment insurance Job vacancies Job applications COVID-19 ==== Body pmc1 Introduction In March 2020, the coronavirus COVID-19 led to a dramatic surge in business closures and job losses. To address this crisis, the Federal Pandemic Unemployment Compensation (FPUC) was voted on March 27 as part of the CARES Act. FPUC provided unemployed workers with an additional $600 a week in unemployment benefits, until the end of July 2020. This represented an unprecedented increase in unemployment insurance generosity: 76% of unemployed workers had a replacement rate above 100%, i.e. collected higher unemployment benefits than their prior wage (Ganong et al., 2020). Economic theory predicts that more generous benefits decrease job search effort and could decrease vacancy creation (Landais et al., 2018b). Therefore, FPUC might have increased unemployment and dampened economic activity, in particular during the summer of 2020, when many lockdown measures were lifted. Yet, the early evidence suggests that FPUC had at most limited effects on employment (Bartik et al., 2020, Altonji et al., 2020, Dube, 2020). Why? We show that the FPUC did indeed have a negative effect on search effort. However, we provide two pieces of evidence that can explain why FPUC did not decrease employment. First, FPUC did not decrease vacancy creation—not even during the re-opening phase in June-July 2020. Second, labor market tightness was unusually low in March-July 2020: since jobs were already receiving an unusually high number of applications, a decrease in applications likely had limited effects on the number of workers hired. Identifying the effect of FPUC on local labor markets is challenging. One can exploit the large disparities in the proportional increase in unemployment benefits: for workers who were eligible to lower regular weekly benefits levels before the CARES Act, the additional $600 per week represents a larger relative increase. However, the passage of the FPUC coincided with exceptionally large and brutal changes in the labor market: just before the implementation of FPUC, the number of job listings collapsed (Forsythe et al., 2020a), while the number of unemployed workers skyrocketed. What is more, low-wage workers were most likely to lose their jobs, and these workers are precisely the ones who experienced the largest increase in benefits with the FPUC. Our empirical analysis aims at addressing these potential confounding factors, exploiting very granular data on labor market outcomes. We collect detailed data on job applications and job postings from Glassdoor.com, for January 2018 to July 2020. We measure search effort and vacancy creation by counting the number of applications and new job vacancy postings each week, in each local labor market —defined as the interaction of state, occupations and industries and wage deciles. We compute potential replacement rates using the calculator in Ganong et al. (2020). To strengthen the external validity of our analysis, we re-weigh observations such that each State×industry×occupation reflects the proportion of the US labor force in that labor market in the Current Population Survey (CPS) for 2018–2020. In the first part of the paper, we estimate the impact of FPUC on search effort, job vacancies and labor market tightness in March-July 2020. We exploit differences in proportional increases in benefits across local labor markets. We include several control variables to flexibly account for changes in the local labor market that were unrelated to the FPUC. We allow for market-specific seasonal variation, using week-of-year by local labor market fixed effects. We include state by week fixed effects to control for differential COVID-19 related policies across states, and other factors that change at the state level. We control for industry by week fixed effects, which accounts for the differential impact of the crisis by industry. Additionally, we allow for different trends in local labor markets that were differentially exposed to unemployment during the COVID-19 crisis prior to the CARES Act. Our identifying assumption is that, in the absence of policy change, the outcome variables would have evolved similarly in local labor markets with different increases in UI, conditional on controls. To support the credibility of our identification assumption, we show that future increases in FPUC do not correlate with trends in job applications, job vacancies or labor market tightness in the weeks prior to the enactment of the FPUC. We also confirm in robustness checks that our main results hold in the construction sector, where job loss before FPUC was not systematically correlated with earnings (and thus the increase in the replacement rate), and hence identification might be less challenging. We find that a 10% increase in the benefit replacement rate due to the FPUC leads to a 3.6% decline in job applications. At the same time, FPUC had no effect on job vacancies. The absence of an effect on vacancy creation indicates that employers’ recruiting was not limited by workers’ lower search effort or potentially higher wage demands, but rather by other factors, consistent with job rationing models (Michaillat, 2012). Taking the effect on applications and vacancies together, a 10% increase in the benefit replacement rate increased labor market tightness by 3.3%. We then analyze the timing of the effect. One could have expected to see a larger impact of FPUC during the reopening phase when most lockdowns were lifted (June-July 2020) and economic activity was picking up. However, FPUC still had no effect on vacancy creation. Further, the effect of FPUC on job applications decreased from May until the end of FPUC in July 2020, suggesting that job seekers might have increased their search effort in anticipation of FPUC expiration. Our results are similar when we exclude periods of lockdown, or when we focus on teleworkable occupations, which are less affected by social distancing measures. In markets where workers are more likely to be on temporary layoffs, the effects of FPUC on search effort are significantly smaller. This is consistent with the idea that workers are less likely to search when they expect to be recalled, and suggests that the prevalence of temporary layoffs during the COVID-19 crisis might have attenuated the effect of unemployment insurance on job search. In the second part of the paper, we describe the context in which FPUC was implemented, as it is crucial to assess the impact of FPUC on unemployment and welfare. We document the evolution of seasonally adjusted applications, vacancies, and labor market tightness, relative to their levels in January-February 2020. The number of applications increased by 4.4% during the FPUC period, despite the negative effect of FPUC on search effort. This probably reflected in part the drastic increase in the number of unemployed workers. At the same time, job vacancies declined by 26%. As a result, labor market tightness decreased by 31%. This implies that, on average, employers got more applicants for their vacancies. Even in the markets most affected by FPUC (4th quartile of the increase in replacement rate due to FPUC), tightness during the FPUC period of March-July 2020 was 19% lower than in January-February 2020. During the re-opening period (June-July 2020), tightness was back to its pre-pandemic level, even though many unemployed workers were receiving FPUC. We finally discuss how our results contribute to understanding the welfare impact of the FPUC. Landais et al. (2018b) formalize that increases in unemployment insurance improve welfare when unemployment insurance increases tightness and tightness is inefficiently low. We demonstrate that this is likely the case for FPUC. First, we show that FPUC did not affect vacancy creation, and hence increased tightness by decreasing applications. Second, we document that labor market tightness was particularly low during the FPUC period of March-July 2020 relative to January-February 2020: this suggests that tightness in the absence of FPUC may have been inefficiently low. Importantly, these results hold for the reopening phase (June-July 2020). Taken together, our results hence indicate that the effect of FPUC on labor market tightness was likely welfare improving, during the whole period of FPUC. Our paper focuses on welfare considerations that are related to labor market tightness. For a comprehensive welfare analysis of FPUC, one should also consider the gains associated with consumption smoothing for UI recipients and the costs associated with their longer unemployment spell. Moreover, beyond the framework of Landais et al. (2018b), the FPUC was potentially also valuable to society by keeping some workers outside of the labor force and limiting the transmission of the virus (Fang et al., 2020), and by boosting consumption and employment through a stimulus effect (Ganong et al., 2021). These additional welfare considerations would likely reinforce our conclusion regarding the desirability of high unemployment insurance in March-July 2020. Our article contributes to the literature on the impact of unemployment benefit levels on job search, and vacancy creation. The increase in the level of benefits with FPUC was exceptionally large (Schmieder and von Wachter, 2016). Yet, our findings are consistent with the estimated effect of unemployment insurance on search effort in prior literature (Krueger and Mueller, 2010, Fradkin and Baker, 2017, Lichter and Schiprowski, 2020). Although most models predict that more generous unemployment insurance should decrease search, one important question in general equilibrium models is how unemployment benefits affect vacancy creation, and hence labor market tightness (Landais et al., 2018b). In standard models, higher unemployment benefits diminish job search and exert upward pressure on wages by raising the outside option of unemployed workers, thereby discouraging vacancy creation (Pissarides, 2000). In “job rationing” models, vacancy creation primarily depends on the marginal product of labor, and is hence unaffected by unemployment insurance (Michaillat, 2012). Our finding that FPUC did not affect vacancy creation is consistent with job rationing. What is the economic intuition? In the context of the COVID-19 crisis, social distancing measures may have decreased labor productivity. Other factors less specific to the COVID-19 crisis might have also contributed to job rationing, since we detect no effect on vacancy creation even after lockdown measures were lifted, and even for teleworkable occupations. Our evidence of job rationing is consistent with what was found in other contexts (Lalive et al., 2015, Marinescu, 2017, Landais et al., 2018a). Second, we contribute to the body of work analyzing the effects the FPUC on employment. Several empirical studies found no association between FPUC and employment (Bartik et al., 2020, Altonji et al., 2020, Finamor and Scott, 2021, Dube, 2020). We provide a potential explanation for the limited effect of FPUC on employment: FPUC did not affect vacancy creation, and its large negative effect on search effort happened at a time when returns to search were low. This complements other explanations for the limited employment effect in the literature. In their models, Boar and Mongey, 2020, Petrosky-Nadeau, 2020, Mitman and Rabinovich, 2021 emphasize that the effect of FPUC on search behavior should be attenuated because unemployed workers anticipated its expiration. Although we find a substantial effect of FPUC on search effort, our finding that it decreased between May and July is consistent with the idea that workers reacted less when they anticipated the expiration. Ganong et al. (2021) show that the FPUC had a positive stimulus effect on employment through increased consumption, which could offset the negative effect on job search effort. Last, our analysis contributes to the study of the labor market during COVID-19 (Bartik et al., 2020, Gupta et al., 2020, Cheng et al., 2020, Fairlie et al., 2020, Montenovo et al., 2020). We add to work investigating trends in the applications and job postings during COVID in Sweden (Hensvik et al., 2020), and job postings in the US (Forsythe et al., 2020a, Campello et al., 2020, Forsythe et al., 2020a, Forsythe et al., 2020b). 2 Data We combine several datasets from Glassdoor, one of the world’s largest jobs and career sites. We use job vacancy listings and job applications from Glassdoor’s online jobs platform in January 2018 to July 2020 for the U.S. We also exploit information on users’ self-reported salaries to measure replacement rates. 2.1 Data on job listings and job applications Glassdoor contains job openings that companies post directly to the site, or that are collected on company career sites or third-party job boards. Glassdoor then ties job listings data to specific industries and canonical occupations that can be mapped into O*NET-SOC codes. We focus on new job listings posted each week rather than the total stock of active job listings. The flow is more responsive to sharp changes in policy and labor market conditions, and captures a disproportionate share of application activity. To measure search effort, we take the count of applications, defined as a user starting an application by clicking on the “Apply Now” button.1 Applications provide a proxy of search effort, under the assumption that the probability of applying on Glassdoor’s website rather than using other search channels is relatively stable over time. We allocate applications into an industry and occupation category based on jobs characteristics, and into a state based on applicants’ address. Applications on Glassdoor don’t come exclusively from unemployment benefits recipients, therefore our estimates at the local labor market level also capture the search effort of job seekers who do not receive UI. This is the right measure for the purpose of investigating overall labor market tightness.2 Note that these market level estimates could be smaller than estimates of the microeconomic impact on the search of UI recipients, as job seekers who do not receive UI should react less to changes in unemployment insurance. Glassdoor’s job listings capture a large portion of online U.S. job listings, which themselves comprise a substantial percentage of all job openings.3 Job listings at very small businesses or managed by unions may be underrepresented (Chamberlain and Zhao, 2019). As a result, compared to the Current Population Survey (CPS) labor force by industry in 2018–2020, some sectors like Construction are underrepresented, while other sectors like Information Technology are overrepresented (Table A.1). We hence use weights in all our analysis, based on the average number of workers in the labor force in the industry, occupation and state from the CPS for January 2018 - July 2020. We confirm that weighted job listings evolve very similarly — the correlation is 0.6 for 2019–2020 — to vacancies in the representative Job Openings and Labor Turnover Survey (JOLTS) (Fig. A.2). In particular, the magnitude of the decrease in job listings in April and May 2020 is the same in the two data sources. 2.2 Data on earnings and unemployment benefits We determine the level of prior earnings for all jobseekers in our sample, using information self-reported by Glassdoor users. Although it is not required to search for jobs on the platform, some Glassdoor users report information on their wage, including base salary and additional compensation like tips and cash bonuses.4 We take the average earnings for each state-occupation, using a very disaggregated occupation definition (O*Net 6-digit codes, 416 categories). Then, we determine the replacement rate that jobseekers eligible for unemployment insurance should receive, with and without the FPUC. We use the calculator created by Ganong et al. (2020), which gives unemployment benefits based on individuals’ state of residence and pre-displacement quarterly earnings, according to UI guidance provided by the Department of Labor: it allows us to compute replacement rates without FPUC, based on jobseekers’ state, imputed prior earnings, and assuming that they worked continuously during the four quarters preceding unemployment.5 Finally, we add $600 to standard weekly benefits to compute the replacement rate with the FPUC. 2.3 Panel at the week and local labor market level We define a local labor market as the interaction of state, 2-digit occupations (76 categories in our sample), 2-digit industries (12 categories in our sample) and wage deciles. We take a narrow definition to be able to capture a lot of the variation in replacement rates and add granular fixed effects to our regressions. We keep local labor markets with at least 1000 applications in 2018–2020, such that we can detect variation in weekly applications count (our results are robust to using alternative thresholds). We arrange our data as a panel at the calendar week and local labor market level: we track the count of applications, of vacancies, labor market tightness, and the average replacement rates, each week in each local labor market. We present descriptive statistics in Table A.2. 3 Institutional background and empirical strategy 3.1 Institutional background The CARES Act was signed into law on March 27 2020. It enacted the Federal Pandemic Unemployment Compensation (FPUC), a $600 supplement to weekly unemployment benefits for all workers eligible for unemployment insurance.6 The $600 FPUC amount was chosen to raise the UI replacement rate to 100% for the average U.S. worker. Because of technical issues in state unemployment insurance systems, it was not possible to tailor FPUC to each worker’s prior earnings. Once the CARES Act was voted, unemployed workers could anticipate their upcoming replacement rate, and potentially adjust their behavior accordingly. But it took a few weeks, until the end of April, for all states to start paying out benefits (Bartik et al., 2020). The last FPUC payment was the week of July 20, 2020; after this, eligible workers could still receive the regular amount of unemployment benefits. 3.2 Empirical strategy Variation in replacement rate increase For identification, we exploit differences in relative increases in replacement rate across local labor markets. It is easy to see that a fixed $600 increase in weekly benefits B generates a relative increase in replacement rate RepFPUCRep (or in benefits BFPUCB) that is negatively correlated with pre-displacement earnings w:RepFPUCRep=BFPUC/wB/w=BFPUCB=1+600B=1+600Rep×w For each local labor market, we compute the logged ratio RepFPUCRep, representing the relative increase in replacement rate. There is substantial variation in this measure, and the measure is negatively correlated with earnings levels (Fig. A.3). Identification strategy: Our empirical strategy consists in analyzing the changes in logged labor market outcomes that are associated with this logged replacement rate ratio around the time when the FPUC was voted: in the absence of confounding factors, larger changes in local labor markets that experience a larger increase in replacement rate measure the effect of the FPUC.7 Note that our strategy is not designed to capture the indirect effects of FPUC going through consumption: such stimulus effects primarily affect jobs producing goods and services that see the largest increase in consumption, not necessarily jobs where workers experience the largest increase in replacement rates. The identification of the effect of FPUC is threatened by any factor that affects the outcomes around the CARES Act and correlates with the increase in replacement rate. The enactment of FPUC coincided with an exceptionally large and abrupt negative labor market shock due to COVID-19 (Fig. A.1). This shock disproportionately hit low-paying jobs, and hence workers who could experience the largest increase in replacement rates with FPUC. Therefore, failing to account for this shock would likely bias estimates of the effect of FPUC. We address this challenge in several ways. First, we include for several controls to absorb the effect of the COVID-19 crisis, or other potential confounders. We include week-of-year×local labor market fixed effects, to flexibly allow for seasonal variation in each local labor market. As the COVID-19 crisis hit different states and industries in different ways, we include time×state fixed effects, and time×industry fixed effects. We also directly quantify the exposure to the COVID-19 crisis in each local labor market: we estimate the risk of unemployment immediately prior to the CARES Act. Using the monthly CPS for January to March 2020 matched with the Outgoing Rotation Group 2019 (ORG) (Flood et al., 2020), we predict the probability of being unemployed for individuals in the labor force, in a logit model including industry, occupation, wage decile, and state fixed effects (see Table A.3). We use the estimated coefficients to predict the risk of unemployment in our sample, i.e the “exposure”. We allow for different trends in markets differentially exposed to the COVID-19 crisis in our estimation model, by including exposure decile×calendar week fixed effects.8 We identify the effect of FPUC on labor market outcomes under the assumption that, in the absence of a change in unemployment insurance, the outcome variables would have evolved similarly in local labor markets with different replacement rate increases, but in the same state, industry, and exposure category.9 If this condition is not satisfied, our estimates may also capture, for example, the differential response of low-wage local labor markets to the COVID-19 shock. Second, we test the plausibility of our identifying assumption by analyzing the correlation between the evolution of outcomes before the CARES Act, and the magnitude of the benefits increase induced by FPUC. If our identification strategy is valid, our pre-CARES Act estimates should not be significant. This is equivalent to the test of the parallel trend assumption in a differences-in-differences setup. Third, we conduct a robustness check where we focus on a segment of the economy where identification problems appear less severe. Specifically, we analyze the correlation of prior earnings with the probability of becoming unemployed at the onset of the COVID-19 crisis, by sector, using the CPS for January-March 2020, matched with ORG 2019. Consistent with descriptive statistics presented in Fig. A.1, there is a strong negative relationship between prior earnings and the risk of unemployment (Table A.4), but this correlation is small and insignificant in the Construction sector (in the Mining sector as well, but there are virtually no applications for this sector on Glassdoor). We hence re-estimate the effect of FPUC for Construction, using the same empirical model without including week by state, week by industry and week by exposure fixed effects.10 The estimates give the causal effect of FPUC, without imposing any functional form to the confounding effect of the COVID-19 crisis on outcomes, under the alternative identifying assumption that there is no substantial confounding factor in this sector. One might want to analyze changes in labor market outcomes around the time of the expiration of the FPUC, rather than at the beginning, to have a more stable economic environment. However, identifying the effect of the FPUC around its expiration is difficult, as there should be no sharp change in outcomes then. Indeed, finite duration unemployment insurance should generate large changes in the search behavior of UI recipients at the start of the UI spell, but the effect should gradually decrease until exhaustion (van den Berg, 1990, Marinescu and Skandalis, 2021, DellaVigna et al., 2020). Moreover, if the impact of FPUC on jobseekers includes liquidity effects rather than only moral hazard effects (Chetty (2008)), it should persist after FPUC expiration, until savings are exhausted. Estimation models: We estimate the following model:(1) ln(Yt,s,o,i,w)=∑τ≠-1βτ·1t=FPUC+τ·lnRepFPUCReps,o,i,w+λweek(t),s,o,i,w+ρt,s+ϕt,i+ψt,Exposure(s,o,i,w)+εt,s,o,i,w lnRepFPUCReps,o,i,w gives the relative increase in the replacement rate associated with FPUC in each local labor market, defined as the intersection of a state s, an occupation o, an industry i and a wage decile w. ln(Yt,s,o,i,w) denotes the logged outcome variable in calendar week t (to avoid missing values, we add 1 to each outcome before taking the log). 1t=FPUC+τ indicates that the calendar week t is τ weeks before the start of FPUC (for τ<0), or after (for τ⩾0). We take as reference period τ=-1, the week before the vote of FPUC (the FPUC was voted on the week of March 23, 2020). The βτ are the coefficients of interest. We include week-of-year×local labor market fixed effect, λweek(t),s,o,i,w ; we let time t fixed effects vary by state (ρt,s), industry (ϕt,i) and Exposure decile (ψt,Exposure(s,o,i,w)). εt,s,o,i,w is the error term. We cluster standard errors at the state level, and observations are weighted by the average number of workers in the labor force in the industry, occupation and state from the CPS for January 2018- July 2020. 4 The impact of FPUC on the labor market 4.1 Main results Pre-FPUC trends: Fig. 1 presents the correlation between labor market outcomes and the increase in UI generosity induced by FPUC, week by week. There is no significant correlation between applications and future benefits increases in weeks prior to the adoption of FPUC. This supports the credibility of our identifying assumption that, in the absence of FPUC, the evolution of outcomes would be similar in markets that experienced different benefits increases, conditional on the controls included. Under this assumption, the coefficients for the period after the adoption of FPUC can be interpreted as the causal effect of FPUC.Fig. 1 The impact of FPUC on applications, vacancies and tightness, week by week. Notes: This Figure reports the estimates from regressions of the logged count of applications, the logged count of vacancies, and the logged labor market tightness on the interaction of each week around the enactment of the FPUC and the potential increase in UI generosity lnRepFPUCRep. The coefficients before the enactment of the FPUC help test our identification assumption. The coefficients after represent the elasticity of the outcome with respect to benefits levels. We use a panel at the calendar week and local labor market level. We include week-of-year×local labor market, time×state, time×industry and time×Exposure deciles fixed effects. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Thin lines deNote 95% confidence intervals, based on robust SE clustered at the state level. The effect of FPUC: The increase in unemployment insurance generosity leads to a decrease in applications immediately after the CARES Act, which remains visible until FPUC expires (Fig. 1, (1)). Quantitatively, Table 1 shows that the corresponding elasticity of applications with respect to the benefit replacement rate is −0.36 (col. (1)): a 10% increase in the replacement rate leads to a 3.6% decline in job applications, consistent with the theoretical prediction that a higher benefit replacement rate decreases search effort. In contrast, FPUC had no clear effect on job vacancies (Fig. 1, (2)). While there is a decrease in the number of vacancies immediately after FPUC was voted and until mid-April, it is only transitory and imprecisely estimated. Since FPUC decreased applications and had no effect on vacancies, it increased labor market tightness (Fig. 1, (3)). Consistently, we see in column (2) of Table 1 that the elasticity of vacancies with respect to the replacement rate is close to zero and statistically insignificant. Column (3) shows that the elasticity of tightness is therefore close to the opposite of the elasticity of applications: a 10% increase in the benefit replacement rate leads to a 3.34% increase in labor market tightness.Table 1 Elasticity of search with respect to benefits levels. Applications Vacancies Tightness (ln) (ln) (ln) (1) (2) (3) lnRepFPUCRep×FPUC period −0.361*** −0.027 0.334*** (0.071) (0.069) (0.106) Exposure decile×Time FE ✓ ✓ ✓ Industry×Time FE ✓ ✓ ✓ State×Time FE ✓ ✓ ✓ Labor market×Week of year FE ✓ ✓ ✓ No. of Obs 1,523,252 1,523,252 1,523,252 Notes: This Table reports the estimated elasticity of job applications, vacancy creation and labor market tightness with respect to unemployment benefits levels. The estimates are obtained from the regression of logged labor market outcomes on the logged increase in benefits interacted with a dummy for the period between the vote of FPUC and its expiration, in a panel at the calendar week and local labor market level. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Robust standard errors clustered at the state level are in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). 4.2 Additional results We first analyze the timing of the effect of FPUC on search effort. In Fig. 1 (panel (1)), we see that the negative effect on search strongly increased in the first 3 weeks, possibly due to delays in initial benefit receipt. The effect then remained roughly constant until June, and decreased until the end of FPUC. The estimated elasticity is −0.44 in May, at its largest; in July, as the extra benefits are about to expire, the elasticity is significantly smaller (in absolute value) at −0.33 (Table 2 , col. 1).11 Our results are consistent with theory predicting that job seekers increase their search effort as benefits expiration approaches (Marinescu and Skandalis, 2021), and therefore that temporary increases in benefits have smaller effects on job search when workers anticipate their expiration (Boar and Mongey, 2020, Petrosky-Nadeau, 2020, Mitman and Rabinovich, 2021).Table 2 Additional results on the impact of FPUC on applications. Applications (ln) (1) (2) (3) (4) lnRepFPUCRep×FPUC period, March −0.144* (0.079) lnRepFPUCRep×FPUC period, April −0.385*** (0.075) lnRepFPUCRep×FPUC period, May −0.436*** (0.065) lnRepFPUCRep×FPUC period, June −0.395*** (0.080) lnRepFPUCRep×FPUC period, July −0.326*** (0.085) lnRepFPUCRep×FPUC period×Lockdown −0.356*** (0.070) lnRepFPUCRep×FPUC period×Open −0.368*** (0.075) lnRepFPUCRep×FPUC period, Telework −0.382*** (0.100) lnRepFPUCRep×FPUC period, Not telework −0.311*** (0.086) FPUC period × Telework 0.131 (0.141) lnRepFPUCRep×FPUC period, High temp layoff −0.173** (0.080) lnRepFPUCRep×FPUC period, Low temp layoff −0.414*** (0.075) FPUC period × High temp layoff −0.361*** (0.107) P-Value Test May  = July .013 P-Value Test Lockdown  = Open .702 P-Value Test Telework  = Not telework .554 P-Value Test High temp layoff  = Low temp layoff .009 Exposure decile×Time FE ✓ ✓ ✓ ✓ Industry×Time FE ✓ ✓ ✓ ✓ State×Time FE ✓ ✓ ✓ ✓ Labor market×Week of year FE ✓ ✓ ✓ ✓ No. of Obs 1,523,252 1,523,252 1,496,988 1,523,252 Notes: This Table reports the estimates from the regression of logged search effort on various independent variables, in a panel at the calendar week and local labor market level—where local labor markets are defined as as the interaction of state, 2-digit occupations and 2-digit industries and wage deciles. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Robust standard errors clustered at the state level are in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). Over time, the economic environment also changed with many states adopting lockdown measures in March 2020 that were later relaxed. As these measures could hinder economic activity, one could expect the impact of unemployment insurance on the labor market to be attenuated during lockdowns. The elasticity of search with respect to unemployment insurance was slightly smaller when states implemented a lockdown, but the difference is small and insignificant (Table 2 col. (2)). This suggests that the difference in the effect of unemployment insurance on search over time is not explained by lockdown measures. One could expect that the reaction of job seekers to the increase in unemployment insurance was smaller than in normal times, because the risk of contamination with COVID-19 imposed constraints on search and on work. Yet, we find little evidence for this: in Table 2 (col. 3), we show that the effect of FPUC on search effort is similar in “teleworkable” and non-“teleworkable” occupations (defined as in Dingel and Neiman (2020)). The situation of unemployed workers during the COVID-19 crisis was also atypical in that a large fraction of them were on temporary layoff (Birinci et al., 2021). We estimate the chance to be on temporary layoff in the CPS for January to July 2020 conditional on being unemployed, in a logit model with two-digit occupation, two-digit industry, state and wage decile fixed effects. We predict the probability of temporary layoff among unemployed workers in each local labor market based on the estimated coefficients. In Table 2 (col. (4)), we show that the effect of FPUC on search effort is significantly smaller in markets with a high (above median) chance of temporary layoff (elasticity of −0.17) than in the others (elasticity of −0.41). This suggests that workers who expect to be recalled search less irrespective of their unemployment insurance. When estimating the FPUC effect by quartile of the probability of temporary layoff (Table A.6), we find the same overall pattern, and detect a negative effect of FPUC on vacancies in labor markets with the lowest prevalence of temporary layoffs. This suggests that the prevalence of temporary layoffs has contributed to attenuating the effect of FPUC on job search and vacancy creation. 4.3 Robustness checks We present the estimates obtained when we add, one by one, the controls included in our preferred specification, in Fig. A.4, Fig. A.5 and Fig. A.6. All control variables contribute to making the pre-trends disappear, in line with the economic intuitions behind our estimation strategy. For instance, when we only control for seasonality, the coefficients in panel (2) of Fig. A.4 exhibit a steep increase in the weeks before the CARES Act, which might reflect the abrupt surge in unemployment among low-wage workers. Although the different controls seem to improve our identification of the effect of FPUC, it is reassuring to note that our qualitative conclusions do not crucially depend on one particular specification. The estimates obtained in all the specifications in Fig. A.4 are suggestive of a negative effect of FPUC on job search. We don’t find negative coefficients for vacancy creation in any of the specifications in Fig. A.5; we sometimes obtain positive coefficients, which could capture a stimulus effect. Finally, all specifications in Fig. A.6 are suggestive of a positive effect of FPUC on tightness. We then analyze more flexibly the relationship between labor market outcomes and the increase in unemployment benefits induced by FPUC in Table A.7. We look separately at the evolution of outcomes in local labor markets at different quartiles of the distribution of relative increase in benefits in columns (1)-(3). As expected, the largest increase in unemployment benefits was associated with the largest decrease in applications. The effect for the third and fourth quartile is essentially the same, suggesting there is no additional decrease in applications for very high increases in the replacement rate. In columns (4)-(6) we estimate the effect of the absolute increase in replacement rates. The results are qualitatively similar. We also show that our results are robust to using alternative samples of local labor markets in Table A.8. Finally, to confirm our main results, we focus on the construction sector, where job loss before FPUC was not systematically correlated with earnings. While the estimates are naturally less precise than in the full sample, we see similar qualitative patterns: FPUC decreased applications, had no effect on vacancies, and therefore ultimately increased labor market tightness (Fig. A.7). These results are reassuring: our conclusions hold in a segment of the economy where there are no substantial factors confounding the identification of the effect of FPUC. 5 Labor market tightness during the FPUC 5.1 The evolution of labor market tightness We present the evolution of seasonally adjusted applications, new vacancy postings, and labor market tightness relative to their baseline levels in Jan-Feb 2020 in Fig. 2 . The number of vacancies strongly declined both before and right after FPUC. In contrast, the number of job applications exhibited limited change. This suggests that the various factors that could influence search in that period (the increase in unemployment, FPUC, the infection risk, uncertainty, etc.) largely compensated each other.12 Therefore, the evolution of vacancies drove the evolution of labor market tightness: between March and May 2020, both dropped way below their baseline level (Fig. A.9 shows that tightness was always above the May 2020 level in the preceding year).Fig. 2 Changes in applications, vacancies, and labor market tightness relative to Jan-Feb 2020 (seasonally adjusted). Notes: The Figure presents the seasonally adjusted changes in the total weekly count of applications, the weekly count of new posted vacancies, and labor market tightness, relative to their baseline levels in Jan-Feb 2020. The Figure is obtained by regressing the logged variables on the calendar week coefficients and week-of-year fixed effects to control for seasonal variation, and then subtracting to each calendar week coefficient the average of estimates for Jan-Feb 2020. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Quantitatively, seasonally adjusted labor market tightness declined by 30.6% during the FPUC period of March-July 2020 relative to January-February 2020 (Table 3 , col. 3, upper panel), reflecting a labor market more favorable to recruiting firms, despite the large increase in unemployment benefits13 . During the whole FPUC period, applications slightly increased by 4.4% (col. 1, upper panel), while vacancies declined by 26.2% (col. 2). During the reopening phase, in June-July 2020, seasonally adjusted applications, vacancies, and labor market tightness were similar to their level in Jan.-Feb. 2020 (col. 1–3, lower panel). Overall, the evidence is not consistent with significant hiring difficulties during the reopening phase.Table 3 Changes in labor market tightness during the period of the FPUC, relative to Jan-Feb 2020. Potential increase in UI Sample: All Q1 Q2 Q3 Q4 Outcome: Applications Vacancies Tightness Tightness Tightness Tightness Tightness (1) (2) (3) (4) (5) (6) (7) FPUC period 0.044** −0.262*** −0.306*** −0.413*** −0.334*** −0.287*** −0.190*** (0.018) (0.022) (0.021) (0.023) (0.024) (0.046) (0.032) Week of year FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ No. of Obs 1,540,598 1,540,598 1,540,598 613,586 412,050 297,346 217,616 FPUC, Mar-May 0.076*** −0.448*** −0.524*** −0.547*** −0.558*** −0.547*** −0.444*** (0.021) (0.023) (0.021) (0.023) (0.021) (0.048) (0.035) FPUC, Jun-Jul 0.004 −0.030 −0.034 −0.246*** −0.055 0.039 0.126*** (0.019) (0.022) (0.025) (0.028) (0.040) (0.054) (0.037) Week of year FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ No. of Obs 1,540,598 1,540,598 1,540,598 613,586 412,050 297,346 217,616 Notes: This Table reports changes in the count of job applications, the count of vacancies, and labor market tightness during the period of the FPUC, relative to their baseline levels in Jan-Feb 2020. In the upper panel, we present estimates for the total period of FPUC, and for two sub-periods in the lower panel. The estimates are obtained by regressing each logged outcome on dummy variables for the FPUC period and for all calendar weeks outside of Jan-Feb 2020, in a panel at the calendar week and local labor market level. We include week-of-year fixed effects to control for seasonal variation. The estimates are obtained in the full sample in col. (1)-(3), and for local labor markets in each quartile of the distribution of potential UI increase in col. (4)-(7). We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Robust standard errors clustered at the state level are in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). 5.2 Heterogeneity One could be concerned that recruitment difficulties increased in the local labor markets most affected by FPUC. We present the evolution of tightness by quartile of the increase in replacement rate due to FPUC, in columns 4–7 of Table 3 and graphically in Fig. A.11. We show that even in the markets most affected by FPUC, seasonally adjusted tightness was 19% lower during the FPUC period than at baseline (Table 3, col. 7). During the re-opening phase, tightness for the fourth quartile was slightly higher (12.6%) than at baseline. For the three lower quartiles (col. 4–6), tightness during reopening remained below or at baseline. Finally, Fig. A.12 and Fig. A.13 show that the evolution of applications, vacancies and labor market tightness are qualitatively similar across sectors. The magnitudes of the fluctuations differ by sector in expected ways: for example, during the lockdown, vacancies in sectors that were directly affected like leisure & hospitality and transport & logistics understandably decreased more than in other sectors. In leisure & hospitality, applications also declined during the FPUC period, suggesting that workers might have decreased their search due to infection risks. In construction, which we analyzed above, the evolution of tightness is similar to that in other sectors. 6 Discussion 6.1 Model of the labor market One important question in general equilibrium models is how unemployment benefits affect vacancy creation (Landais et al., 2018b). Our finding that FPUC did not affect vacancy creation is consistent with job rationing, where vacancy creation primarily depends on the marginal product of labor—rather than wages, or recruiting costs (Michaillat, 2012). In job rationing models, unemployment insurance has more limited effects on employment: because of the limited vacancy response, UI increases labor market tightness, i.e. decreases the competition among job seekers and raises the returns to each unit of search effort. Additionally, labor market tightness was particularly low when FPUC was implemented, such that decreasing search effort likely had limited effects on unemployment. Our findings hence help explain why prior studies have found no effect of FPUC on unemployment (Bartik et al., 2020, Altonji et al., 2020, Finamor and Scott, 2021, Dube, 2020). 6.2 The impact of FPUC on welfare Our results speak to aspects of the impact of FPUC on welfare that have to do with labor market tightness. Landais et al. (2018b) formalize that the welfare impact of unemployment insurance depends on three elements: first, its impact on labor market tightness; second, the impact of tightness on welfare, and third, the trade-off between consumption smoothing and moral hazard. Our paper provides evidence on the first two. We show that FPUC increased tightness. And we document that tightness at that time was very low relative to January-February 2020, hence likely sub-optimal. Note that the level of tightness in January-February 2020 provides a natural benchmark — as the labor market immediately before the crisis was considered healthy — but it may not have been exactly optimal (see Landais et al. (2018a) for an empirical assessment of the optimal tightness level, and Forsythe et al. (2020b) for a discussion of the evolution of tightness up to the COVID-19 crisis). Our results together suggest that FPUC increased welfare by bringing labor market tightness back towards a more efficient level. Intuitively, it strongly decreased the cost of unemployment for job seekers without a commensurate increase in recruitment difficulties for employers. Importantly, this conclusion also holds for the reopening phase (June-July 2020). During this period, one could have expected a clearer negative impact of FPUC on vacancy creation, since economic activity was no longer hindered by lockdown measures. However, we find no evidence of this. One could also have feared labor market tightness to be particularly high during this period, since firms’ recruiting was picking up while many unemployed workers were receiving FPUC. We show that labor market tightness was just back to its pre-pandemic level, despite the positive effect of FPUC on tightness. In the absence of FPUC, the counterfactual level of tightness would likely have remained lower than at baseline. While the positive effects of UI extensions on tightness were likely welfare-improving in March-July 2020, they could be detrimental if tightness were far above its optimal level. Our analysis focuses on labor market tightness. Fully assessing the welfare impact of unemployment insurance in the framework of Landais et al. (2018b) would also require accounting for the classical trade-off between the positive effect on the consumption of UI recipients and the negative effect on their re-employment. The findings by Ganong et al. (2021) suggest important gains from consumption smoothing in the case of FPUC. Moreover, FPUC potentially had other spillover effects, beyond the labor market externalities modeled in Landais et al. (2018b). First, FPUC likely boosted employment through a stimulus effect (Ganong et al., 2021). Second, the sanitary situation might have made it desirable to have workers temporarily outside of the labor force, to the extent that they were less likely to get infected with COVID-19 (Fang et al., 2020). Overall, these additional considerations also point towards the desirability of high unemployment insurance in March-July 2020. 7 Conclusion During the COVID-19 crisis, a 10% increase in unemployment benefits due to the Federal Unemployment Pandemic Assistance (FPUC) led to a 3.6% decline in Glassdoor job applications. Since the FPUC had no effect on vacancies, its effect on applications led to a commensurate 3.3% increase in labor market tightness (vacancies/applications). However, despite the negative effect of FPUC on applications, seasonally adjusted aggregate job applications remained fairly stable during the FPUC period of March-July 2020. With a 4% increase in aggregate applications and a steep 26% decline in job vacancies relative to January-February 2020, seasonally adjusted labor market tightness declined by 31%. Even in the top 25% of labor markets with the highest increase in the level of unemployment benefits, tightness still declined by 19% during the FPUC period. By decreasing applications, FPUC has thus contributed to lifting up a depressed labor market tightness, likely increasing welfare. In a context of excessive competition for jobs among workers, more generous unemployment insurance reduces wasteful applications and has little effect on employment. Our results can help explain why higher unemployment benefits had no effect on employment after the CARES Act (Bartik et al., 2020, Altonji et al., 2020, Dube, 2020). Appendix A Fig. A.1 Descriptive statistics on the state of the labor market when FPUC was voted. Notes: This Figure shows the evolution of the number of UI claims (new and continuing) in the upper panel. In the lower panel, it shows the evolution of employment rates among workers at different parts of the wage distribution (low wage, middle wage or high wage), relative to their levels in January 2020. The data for the lower panel are described in Chetty et al. (2020), and made publicly available on the website https://tracktherecovery.org. Fig. A.2 Comparison of the measure of new vacancies from JOLTS and Glassdoor. Notes: The Figure presents the variation in the count of new job listing at the monthly level, relative to the average level in January-February 2020. We compare the measure of new vacancies from JOLTS, to the measure from Glassdoor data—which we use in the rest of the paper. For the measure from Glassdoor, we present the weighted sum of job listings, using weights reflecting the size of each State×industry×occupation in the labor force in the CPS. Fig. A.3 Variation in the residualized increase in replacement rate in our data. Notes: This Figure illustrates the variation in the residualized increase in replacement rate in our data, which serves to identify the effect of FPUC. The Figure shows that increases in replacement rates are negatively correlated with pre-displacement weekly earnings, and that substantial variation in increase in replacement rate remains even once we take out the portion explained by state, industry, and exposure decile. They are obtained in the cross section of all the local labor market (state by occupation by industry by wage decile) in our study sample. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS, like in the rest of the analysis. As an example, we show in red the observations corresponding to local labor markets for the occupation “salesmen”. Fig. A.4 The impact of FPUC on applications week by week. Notes: This Figure reports the estimates from regressions of logged count of applications on the interaction of each week around the enactment of the FPUC and the potential increase in UI generosity lnRepFPUCRep. We use a panel at the calendar week and local labor market level —where local labor markets are defined as as the interaction of state, 2-digit occupations and 2-digit industries and wage deciles. All models include calendar week and local labor market fixed effects. We also add various controls in different panels. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Thin lines deNote 95% confidence intervals, based on robust SE clustered at the state level. Fig. A.5 The impact of FPUC on new vacancies week by week. Notes: This Figure reports the estimates from regressions of logged count of new vacancies on the interaction of each week around the enactment of the FPUC and the potential increase in UI generosity lnRepFPUCRep. We use a panel at the calendar week and local labor market level —where local labor markets are defined as as the interaction of state, 2-digit occupations and 2-digit industries and wage deciles. All models include calendar week and local labor market fixed effects. We also add various controls in different panels. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Thin lines deNote 95% confidence intervals, based on robust SE clustered at the state level. Fig. A.6 The impact of FPUC on labor market tightness, week by week. Notes: This Figure reports the estimates from regressions of logged labor market tightness (vacancies/applications) on the interaction of each week around the enactment of the FPUC and the potential increase in UI generosity lnRepFPUCRep. We use a panel at the calendar week and local labor market level —where local labor markets are defined as as the interaction of state, 2-digit occupations and 2-digit industries and wage deciles. All models include calendar week and local labor market fixed effects. We also add various controls in different panels. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Thin lines deNote 95% confidence intervals, based on robust SE clustered at the state level. Fig. A.7 The impact of FPUC on applications, vacancies and tightness in the construction sector, week by week. Notes: This Figure reports the estimates from regressions of the logged count of applications, the logged count of vacancies, and the logged labor market tightness (i.e. ratio of vacancies over applications) in the construction sector on the interaction of each week around the enactment of the FPUC and the potential increase in UI generosity lnRepFPUCRep. The coefficients before the enactment of the FPUC help test our identification assumption. The coefficients after represent the elasticity of the outcome with respect to benefits levels. We use a panel at the calendar week and local labor market level—where local labor markets are defined as as the interaction of state, 2-digit occupations and 2-digit industries and wage deciles. We include calendar week and local labor market fixed effects. We control for seasonality using week-of-year×local labor market fixed effects. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Thin lines deNote 95% confidence intervals, based on robust SE clustered at the state level. Fig. A.8 Changes in applications, vacancies, and labor market tightness relative to Jan-Feb 2020, not seasonally adjusted. Notes:The Figure presents the changes in tightness (ratio of the weekly count of new posted vacancies over weekly count of applications), relative to their baseline levels in Jan-Feb 2020. We regress the logged tightness on the time coefficients, and then subtract the average of the estimates obtained for the period Jan-Feb 2020, such that each coefficient represents the relative variation with respect to the average level in Jan-Feb 2020. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Fig. A.9 Changes in applications, vacancies, and labor market tightness relative to Jan-Feb 2020, for a longer time window. Notes: The Figure presents the seasonally adjusted changes in the 3-weeks moving averages of count of applications, the weekly count of new posted vacancies, and labor market tightness, relative to their baseline levels in Jan-Feb 2020. The Figure is obtained by regressing the logged variables on the calendar week coefficients and week-of-year fixed effects to control for seasonal variation, and then subtracting to each calendar week coefficient the average of estimates for Jan-Feb 2020. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Fig. A.10 Weekly changes in applications, vacancies and tightness relative to Jan-Feb, each year. Notes: The Figure presents the weekly changes in the count of applications, of new posted vacancies and of labor market tightness (ratio of the weekly count of new posted vacancies over weekly count of applications), relative to their baseline levels in Jan.-Feb., such that each coefficient represents the relative variation with respect to the average level in Jan.-Feb. We present them separately for the years 2018, 2019 and 2020. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Fig. A.11 Changes in applications, vacancies and labor market tightness relative to Jan-Feb 2020, by UI increase quartile. Notes: The Figure presents, for each quartile of the increase in the replacement rate of UI, the seasonally adjusted changes in the total weekly count of applications, the weekly count of new posted vacancies, and labor market tightness (ratio of vacancies over total applications), relative to their baseline levels in Jan-Feb 2020. The Figure is obtained by regressing the logged variables on the calendar week coefficients and week-of-year fixed effects to control for seasonal variation, and then subtracting to each calendar week coefficient the average of estimates for Jan-Feb 2020. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Fig. A.12 The evolution of job applications & job postings in specific sectors: Notes: The Figure presents, fr each industry, the seasonally adjusted changes in the total weekly count of applications, the weekly count of new posted vacancies, and labor market tightness (ratio of vacancies over total applications), relative to their baseline levels in Jan-Feb 2020. The Figure is obtained by regressing the logged variables on the calendar week coefficients and week-of-year fixed effects to control for seasonal variation, and then subtracting to each calendar week coefficient the average of estimates for Jan-Feb 2020. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Fig. A.13 The evolution of job applications & job postings in specific sectors: Notes: The Figure presents, for each industry, the seasonally adjusted changes in the total weekly count of applications, the weekly count of new posted vacancies, and labor market tightness (ratio of vacancies over total applications), relative to their baseline levels in Jan-Feb 2020. The Figure is obtained by regressing the logged variables on the calendar week coefficients and week-of-year fixed effects to control for seasonal variation, and then subtracting to each calendar week coefficient the average of estimates for Jan-Feb 2020. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Table A.1 Descriptive statistics: Count of applications and job postings on Glassdoor by industry. Data source Glassdoor data CPS Variable Application count Job posting count US labor per week×industry per week×industry force Mean % Mean % % Agriculture 15 0.00 7 0.00 1.59 Business Service 81,112 17.78 84,050 13.84 12.42 Construction 4,420 0.97 3,776 0.62 7.30 Consumer Service 6,447 1.41 8,135 1.34 4.82 Finance 49,031 10.75 39,249 6.46 6.80 Government 7,047 1.55 8,421 1.39 4.67 Health, Education 68,506 15.02 164,691 27.12 22.51 Information 105,109 23.05 65,283 10.75 1.80 Leisure, hospitality 30,366 6.66 72,982 12.02 9.34 Manufacturing 46,673 10.23 33,052 5.44 9.88 Retail 43,129 9.46 69,722 11.48 12.92 Transportation, warehousing 14,217 3.12 57,889 9.53 5.66 Notes: This Table presents the mean weekly count of applications and new job postings and the proportion of applications and new vacancy postings by industry, in our study sample. In the last column, the Table reports the proportion of US labor force by industry based on the CPS in 2018–2020. Table A.2 Descriptive statistics in our study sample. Mean SD p25 p50 p75 (1) (2) (3) (4) (5) Count of applications, each week in each local labor market 82.2 204.3 12.0 26.0 70.0 Count of vacancies, each week in each local labor market 189.6 659.1 16.0 44.0 126.0 Tightness (vacancies/ applications) 4.7 22.6 0.7 1.4 3.4 Replacement rate without FPUC 43.6 8.4 38.2 45.4 49.2 Replacement rate with FPUC 146.1 47.9 107.3 143.5 173.8 Notes: This Table presents descriptive statistics in our study sample. All variables are computed for each calendar week from January 2018 to July 2020, in each local labor market. We use weights reflecting the size of each State×industry×occupation in the labor force in the CPS. Note that the distribution of potential replacement rates in local labor markets in this Table is not directly comparable to the distribution of replacement rate in the population of unemployed workers computed in Ganong et al. (2020): the level of observation is different since one observation represents one local labor market in our analysis, while one observation represents one worker who lost their job during the COVID-19 crisis in Ganong et al. (2020). However, we note that these statistics are relatively close, with the first quartile of replacement rate ranging around 103%, the median around 145%, third quartile around 195% (Ganong et al., 2020). Table A.3 The probability of unemployment, in January-March 2020. Probability of unemployment WageDecile2 −0.547*** (0.117) WageDecile3 −0.406*** (0.123) WageDecile4 −0.841*** (0.133) WageDecile5 −0.934*** (0.150) WageDecile6 −1.148*** (0.160) WageDecile7 −0.975*** (0.158) WageDecile8 −1.211*** (0.166) WageDecile9 −1.119*** (0.166) WageDecile10 −0.993*** (0.167) State FE ✓ Occupation FE ✓ Industry FE ✓ No. of Obs 49,779 Notes: This table reports the estimates from the logit regressions of a dummy for unemployment on local labor market characteristics: wage deciles, occupation fixed effects, (two-digit) industry fixed effects and state fixed effects. The information on workers’ unemployment status, industry, and state come from the monthly CPS data for January-March 2020, and information on their prior earnings come from the Outgoing Rotation Group data for 2019. The two datasets are matched at the individual level. We use individual survey weights from the CPS 2020. Note that as we use the information on earnings from ORG 2019, we could also have computed alternative weights accounting for the composition of respondents in ORG 2019 and correcting for missing earnings values. We use the standard survey weights for replicability, but we checked that this does not affect our results. Robust standard errors in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). Table A.4 The wage gradient in the probability of unemployment, by industry. Probability of unemployment (1) (2) (3) ln(prior earnings) −0.372*** (0.024) Agriculture × ln(prior earnings) −0.525*** −0.541*** (0.198) (0.208) Mining × ln(prior earnings) −0.061 −0.047 (0.393) (0.395) Construction × ln(prior earnings) −0.096 −0.097 (0.088) (0.089) Manufacturing × ln(prior earnings) −0.531*** −0.553*** (0.097) (0.096) Retail × ln(prior earnings) −0.433*** −0.449*** (0.081) (0.081) Transportation × ln(prior earnings) −0.521*** −0.539*** (0.111) (0.110) Information × ln(prior earnings) −0.248* −0.236* (0.130) (0.134) Finance × ln(prior earnings) −0.266*** −0.277*** (0.057) (0.059) Business services × ln(prior earnings) −0.388*** −0.403*** (0.066) (0.067) Education & health × ln(prior earnings) −0.360*** −0.370*** (0.043) (0.045) Leisure & hospitality × ln(prior earnings) −0.262*** −0.264*** (0.087) (0.088) Consumer services × ln(prior earnings) −0.657*** −0.663*** (0.132) (0.137) Government × ln(prior earnings) −0.603*** −0.575*** (0.136) (0.137) Industry FE – ✓ ✓ State FE – – ✓ No. of Obs 54,023 54,017 54,017 Notes: This table reports the estimates from the logit regressions of a dummy for unemployment on local labor market characteristics: prior earnings, two-digit industries and states. The information on workers’ unemployment status, industry, and state come from the monthly CPS data for January-March 2020, and information on their prior earnings come from the Outgoing Rotation Group data for 2019. The two datasets are matched at the individual level. We use individual survey weights from the CPS 2020. Note that as we use the information on earnings from ORG 2019, we could also have computed alternative weights accounting for the composition of respondents in ORG 2019 and correcting for missing earnings values. We use the standard survey weights for replicability, but we checked that this does not affect our results. Robust standard errors in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). Table A.5 Additional results on the impact of FPUC on vacancies and tightness. Vacancies (ln) Tightness (ln) (1) (2) (3) (4) (5) (6) (7) (8) lnRepFPUCRep×FPUC period, March −0.123 0.021 (0.075) (0.084) lnRepFPUCRep×FPUC period, April −0.247** 0.138 (0.115) (0.133) lnRepFPUCRep×FPUC period, May −0.011 0.425*** (0.069) (0.106) lnRepFPUCRep×FPUC period, June 0.026 0.421*** (0.087) (0.133) lnRepFPUCRep×FPUC period, July 0.111 0.437*** (0.092) (0.132) lnRepFPUCRep×FPUC period×Lockdown −0.109 0.246** (0.066) (0.094) lnRepFPUCRep×FPUC period×Open 0.063 0.430*** (0.084) (0.126) lnRepFPUCRep×FPUC period × Telework 0.079 0.461*** (0.098) (0.153) lnRepFPUCRep×FPUC period × Not telework −0.066 0.245** (0.077) (0.114) FPUC period × Telework −0.096 −0.227 (0.083) (0.150) lnRepFPUCRep×FPUC period × High temp layoff 0.044 0.217 (0.114) (0.139) lnRepFPUCRep×FPUC period × Low temp layoff −0.148 0.267** (0.090) (0.118) FPUC period × High temp layoff −0.193 0.168 (0.171) (0.179) P-Value Test May  = July .047 .852 P-Value Test Lockdown  = Open .004 .008 P-Value Test Telework  = Not Telework .062 .116 P-Value Test High temp layoff  = Low temp layoff .178 .741 Exposure decile×Time FE, Industry×Time FE, State×Time FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Labor market×Week of year FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ No. of Obs 1,523,252 1,523,252 1,496,988 1,523,252 1,523,252 1,523,252 1,496,988 1,523,252 Notes: This Table reports the estimates from the regression of logged vacancies and tightness on various independent variables, in a panel at the calendar week and local labor market level—where local labor markets are defined as as the interaction of state, 2-digit occupations and 2-digit industries and wage deciles. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Robust standard errors clustered at the state level are in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). Table A.6 Additional results on the impact of FPUC, by prevalence of temporary layoff. Applications (ln) Vacancies (ln) Tightness (ln) (1) (2) (3) lnRepFPUCRep×FPUC period × Lowest quartile of temp layoff −0.351*** −0.380*** −0.028 (0.091) (0.083) (0.126) lnRepFPUCRep×FPUC period × 2nd quartile of temp layoff −0.299*** −0.020 0.279* (0.094) (0.123) (0.142) lnRepFPUCRep×FPUC period × 3rd quartile of temp layoff −0.211* 0.077 0.288* (0.107) (0.148) (0.155) lnRepFPUCRep×FPUC period × Highest quartile of temp layoff −0.122 0.001 0.123 (0.103) (0.133) (0.176) FPUC period × 2nd quartile of temp layoff −0.160 −0.398** −0.238 (0.110) (0.165) (0.180) FPUC period × 3rd quartile of temp layoff −0.308** −0.507** −0.198 (0.143) (0.204) (0.185) FPUC period × Highest quartile of temp layoff −0.398** −0.339* 0.059 (0.151) (0.188) (0.244) Exposure decile×Time FE ✓ ✓ ✓ Industry×Time FE ✓ ✓ ✓ State×Time FE ✓ ✓ ✓ Labor market×Week of year FE ✓ ✓ ✓ No. of Obs 1,523,252 1,523,252 1,523,252 Notes: This Table reports the estimates from the regression of logged vacancies and tightness on various independent variables, in a panel at the calendar week and local labor market level—where local labor markets are defined as as the interaction of state, 2-digit occupations and 2-digit industries and wage deciles. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Robust standard errors clustered at the state level are in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). Table A.7 Robustness checks on the impact of FPUC on applications, vacancies and tightness. Applications Vacancies Tightness Applications Vacancies Tightness (ln) (ln) (ln) (ln) (ln) (ln) (1) (2) (3) (4) (5) (6) Q2 of lnRepFPUCRep×FPUC period −0.113*** −0.022 0.091 (0.029) (0.046) (0.058) Q3 of lnRepFPUCRep×FPUC period −0.184*** 0.005 0.189*** (0.037) (0.042) (0.066) Q4 of lnRepFPUCRep×FPUC period −0.188*** 0.007 0.195*** (0.046) (0.037) (0.065) (RepFPUC-Rep)×FPUC period −0.218*** −0.013 0.205*** (0.032) (0.036) (0.045) Exposure decile×Time FE ✓ ✓ ✓ ✓ ✓ ✓ Industry×Time FE ✓ ✓ ✓ ✓ ✓ ✓ State×Time FE ✓ ✓ ✓ ✓ ✓ ✓ Labor market×Week of year FE ✓ ✓ ✓ ✓ ✓ ✓ No. of Obs 1,523,252 1,523,252 1,523,252 1,523,252 1,523,252 1,523,252 Notes: This Table reports the estimates from the regression of logged outcomes on various independent variables, in a panel at the calendar week and local labor market level—where local labor markets are defined as as the interaction of state, 2-digit occupations and 2-digit industries and wage deciles. Rep is the replacement rate. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Robust standard errors clustered at the state level are in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). Table A.8 Robustness check: The impact of FPUC on applications, vacancies and tightness, estimated in different local labor market samples. Sample More than 250 applications More than 500 applications More than 1500 applications Outcome Applications Vacancies Tightness Applications Vacancies Tightness Applications Vacancies Tightness (ln) (ln) (ln) (ln) (ln) (ln) (ln) (ln) (ln) (1) (2) (3) (4) (5) (6) (7) (8) (9) lnRepFPUCRep×FPUC period −0.204*** 0.096 0.300*** −0.272*** 0.045 0.318** −0.370*** −0.067 0.304** (0.067) (0.068) (0.096) (0.079) (0.087) (0.122) (0.086) (0.103) (0.138) No. of Obs 3,157,174 3,157,174 3,157,174 2,300,914 2,300,914 2,300,914 1,170,088 1,170,088 1,170,088 Notes: This Table reports the estimates from the regression of logged outcomes on various independent variables, in a panel at the calendar week and local labor market level—where local labor markets are defined as as the interaction of state, 2-digit occupations and 2-digit industries and wage deciles. We estimate the results on different subsamples, based on the total count of applications in the local labor market in the period 2018–2020 (our main results are estimated on local labor markets with more than 1000 applications in total during 2018–2020). Rep is the replacement rate. We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Robust standard errors clustered at the state level are in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). Table A.9 Changes in labor market tightness during the period of the FPUC relative to Jan-Feb 2020, without accounting for seasonal variation. Potential increase in UI Sample: All Q1 Q2 Q3 Q4 Total Outcome: Applications Vacancies Tightness Tightness Tightness Tightness Tightness (1) (2) (3) (4) (5) (6) (7) FPUC period −0.128*** −0.237*** −0.109*** −0.221*** −0.127*** −0.091* 0.004 (0.021) (0.024) (0.024) (0.022) (0.032) (0.046) (0.042) Week of year FE No No No No No No No No. of Obs 1,540,598 1,540,598 1,540,598 613,586 412,050 297,346 217,616 FPUC, Mar-May −0.117*** −0.415*** −0.298*** −0.356*** −0.325*** −0.310*** −0.201*** (0.022) (0.023) (0.023) (0.023) (0.024) (0.043) (0.043) FPUC, Jun-Jul −0.142*** −0.014 0.128*** −0.052** 0.121** 0.183*** 0.260*** (0.022) (0.027) (0.029) (0.026) (0.046) (0.056) (0.045) Week of year FE No No No No No No No No. of Obs 1,540,598 1,540,598 1,540,598 613,586 412,050 297,346 217,616 Notes: This Table presents the same estimates as Table 3, except that we do not control for seasonal variation. The Table reports changes in the count of job applications, the count of new posted vacancies, and labor market tightness during the period of the FPUC, relative to their baseline levels in Jan-Feb 2020. In the upper panel, we present estimates for the total period of FPUC, while we divide the FPUC period in two sub-periods in the lower panel. The estimates are obtained by regressing each logged outcome on a dummy for the FPUC period, and a dummy for all calendar weeks outside of the baseline period of January-February 2020 in a panel at the calendar week and local labor market level. The estimates are obtained in the full sample in col (1)-(3), and separately for local labor markets in each quartile of the distribution of potential UI increase in col (4)-(7). We use weights to reflect the proportion of the labor force in each State×industry×occupation in the CPS. Robust standard errors clustered at the state level are in parenthesis (* p<0.10, ** p<0.05, *** p<0.010). ☆ We would like to thank Hyeri Choi and Jie Guan for excellent research assistance. We would like to thank for their very helpful comments Antoine Bertheau, Sydnee Caldwell, Pascal Noel, as well as the participants of the seminars in the University of Copenhagen, UCD Dublin, Harvard, Essex/Royal Holloway/Bristol, and in the NBER Labor Spring Meeting. 1 The data is a subset of total applications on Glassdoor. To enhance data quality, Glassdoor’s data processing retains likely people (not bots) who apply after organically searching for jobs on Glassdoor (rather than being redirected to the job posting from elsewhere, e.g. from Glassdoor emails or from paid advertisements on other sites). 2 Tightness is often defined as vacancies/unemployment. As suggested by theory (Landais et al., 2018b), we adjust for search intensity and define tightness as vacancies/applications. 3 The count of job openings reported in Glassdoor’s Job Market Report represents 81% of the count job openings reported in the Job Openings and Labor Turnover Survey. 4 Karabarbounis and Pinto (2018) show that the average and variance of the distribution of salaries within geographical region and within industry are fairly representative when compared against the Quarterly Census of Employment and Wages, and the Panel Study of Income Dynamics. 5 This is the case for most individuals eligible for unemployment insurance, given the average number of days worked reported in the 2019 Annual Social and Economic Supplement (ASEC) of the CPS. 6 The CARES Act also instated an additional 13 weeks of benefits for those whose regular unemployment benefits were exhausted (Pandemic Emergency Unemployment Compensation), and extended the access to unemployment insurance to workers not traditionally eligible, but who were unable to work because of the public health emergency (Pandemic Unemployment Assistance). 7 The estimates would be identical if we instead analyzed the changes in logged outcomes associated with changes in logged replacement rates: we only build this time-invariant replacement rate ratio to be able to estimate coefficients for periods before the FPUC was implemented, and hence test the validity of our identification strategy, as explained in what follows. 8 We use exposure deciles for more flexibility, but our results are similar if we instead control linearly for exposure×week. 9 Extensions in potential benefit duration kick in after July 2020 for most workers, and they only vary by state, so they do not bias our estimates. 10 There is substantial variation in earnings across frequent occupations in the construction sector in CPS-ORG: for instance, they range around $1410 for community association managers; $1020 for electricians; $960 for construction laborers. 11 We present the estimates from the same specifications for vacancies and tightness in Table A.5. 12 In the non seasonally adjusted series (Fig. A.8, left panel), applications decline a bit after FPUC, but this pattern occurs every year (Fig. A.10). 13 Results without controlling for seasonal variation are in Fig. A.9. ==== Refs References Altonji J. Contractor Z. Finamor L. Haygood R. Lindenlaub I. Meghir C. O’Dea C. Scott D. Wang L. Washington E. Employment Effects of Unemployment Insurance Generosity During the Pandemic 2020 24 Bartik A. Bertrand M. Lin F. Rothstein J. Unrath M. Measuring the Labor Market at the Onset of the COVID-19 Crisis SSRN Electron. J. 2020 ISSN 1556-5068. URL https://www.ssrn.com/abstract=3633053 Birinci S. Karahan F. Mercan Y. See K. Labor market policies during an epidemic J. Public Econ. 194 2021 104348 ISSN 0047–2727. URL https://www.sciencedirect.com/science/article/pii/S0047272720302127 Boar, C., Mongey, S., 2020. Dynamic Trade-offs and Labor Supply Under the CARES Act. Technical Report w27727. National Bureau of Economic Research. https://www.nber.org/papers/w27727. Campello, M., Kankanhalli, G., Muthukrishnan, P., 2020. Corporate Hiring under COVID-19: Labor Market Concentration, Downskilling, and Income Inequality. SSRN Scholarly Paper ID 3596655. Social Science Research Network, Rochester, NY. URL https://papers.ssrn.com/abstract=3596655. Chamberlain, A., Zhao, D., 2019. Glassdoor Job Market Report Methodology. Technical report. Glassdoor. URL https://www.glassdoor.com/research/glassdoor-job-market-report-methodology/. Cheng, W., Carlin, P., Carroll, J., Gupta, S., Rojas, F.L., Montenovo, L., Nguyen, T.D., Schmutte, I.M., Scrivner, O., Simon, K.I., Wing, C., Weinberg, B., 2020. Back to Business and (Re)employing Workers? Labor Market Activity During State COVID-19 Reopenings. Working Paper 27419. National Bureau of Economic Research. http://www.nber.org/papers/w27419. Chetty R. Moral hazard versus liquidity and optimal unemployment insurance J. Polit. Econ. 116 2008 173 234 Chetty, R., Friedman, J., Hendren, N., Stepner, M., Team TOI, 2020. The Economic Impacts of COVID-19: Evidence from a New Public Database Built Using Private Sector Data. Technical Report w27431. National Bureau of Economic Research, Cambridge, MA. URL http://www.nber.org/papers/w27431.pdf. DellaVigna, S., Heining, J., Schmieder, J.F., Trenkle, S., 2020. Evidence on Job Search Models from a Survey of Unemployed Workers in Germany: 49. Dingel J.I. Neiman B. How many jobs can be done at home? J. Public Econ. 189 2020 104235 ISSN 0047-2727. URL https://www.sciencedirect.com/science/article/pii/S0047272720300992 32834177 Dube A. The Impact of the Federal Pandemic Unemployment Compensation on Employment: Evidence from the Household Pulse Survey 2020 7 Fairlie, R.W., Couch, K., Xu, H., 2020. The Impacts of COVID-19 on Minority Unemployment: First Evidence from April 2020 CPS Microdata. Working Paper 27246. National Bureau of Economic Research. http://www.nber.org/papers/w27246. Fang, L., Nie, J., Xie, Z., 2020. Unemployment Insurance During a Pandemic. SSRN Scholarly Paper ID 3666545. Social Science Research Network, Rochester, NY. URL https://papers.ssrn.com/abstract=3666545. Finamor L. Scott D. Labor market trends and unemployment insurance generosity during the pandemic Econ. Lett. 199 2021 109722 ISSN 0165–1765. http://www.sciencedirect.com/science/article/pii/S0165176520304821 Flood, S., Warren, R.J., King, M., Rodgers, R., 2020. Integrated Public Use Microdata Series, Current Population Survey: Version 8.0 [dataset]. URL https://doi.org/10.18128/D030.V8.0. Forsythe E. Kahn L.B. Lange F. Wiczer D. Labor demand in the time of COVID-19: Evidence from vacancy postings and UI claims J. Public Econ. 189 2020 104238 ISSN 0047–2727. URL https://www.sciencedirect.com/science/article/pii/S004727272030102X 32834178 Forsythe, E., Kahn, L.B., Lange, F., Wiczer, D.G., 2020b. Searching, Recalls, and Tightness: An Interim Report on the COVID Labor Market. Technical Report w28083. National Bureau of Economic Research. https://www.nber.org/papers/w28083. Fradkin A. Baker S. The impact of unemployment insurance on job search: Evidence from google search data Rev. Econ. Stat. 99 2017 756 768 Tex.date-added: 2018–06-07 15:57:53 +0000 tex.date-modified: 2018–06-07 15:59:49 +0000. Ganong P. Greig F. Liebeskind M. Noel P. Sullivan D.M. Vavra J. Spending and Job Search Impacts of Expanded Unemployment Benefits: Evidence from Administrative Micro Data 2021 72 Ganong, P., Noel, P., Vavra, J., 2020. US Unemployment Insurance Replacement Rates During the Pandemic. SSRN Scholarly Paper ID 3601492. Social Science Research Network, Rochester, NY. URL https://papers.ssrn.com/abstract=3601492. Gupta, S., Montenovo, L., Nguyen, T.D., Rojas, F.L., Schmutte, I.M., Simon, K.I., Weinberg, B.A., Wing, C., 2020. Effects of Social Distancing Policy on Labor Market Outcomes. Working Paper 27280. National Bureau of Economic Research. http://www.nber.org/papers/w27280. Hensvik, L., Le Barbanchon, T., Rathelot, R., 2020. Job Search during the COVID-19 Crisis. SSRN Scholarly Paper ID 3598126. Social Science Research Network, Rochester, NY. URL https://papers.ssrn.com/abstract=3598126. Karabarbounis, M., Pinto, S., 2018. What Can We Learn from Online Wage Postings? Evidence from Glassdoor. SSRN Scholarly Paper ID 3322205. Social Science Research Network, Rochester, NY. URL https://papers.ssrn.com/abstract=3322205. Krueger A.B. Mueller A. Job search and unemployment insurance: New evidence from time use data J. Public Econ. 94 2010 298 307 URL http://ideas.repec.org/a/eee/pubeco/v94y2010i3-4p298-307.html Lalive R. Landais C. Zweimueller J. Market Externalities of Large Unemployment Insurance Extension Programs Am. Econ. Rev. 105 12 2015 3564 3596 Landais C. Michaillat P. Saez E. A Macroeconomic Approach to Optimal Unemployment Insurance: Applications Am. Econ. J. Econ. Policy 10 2018 182 216 ISSN 1945-7731, 1945-774X. URL https://pubs.aeaweb.org/doi/10.1257/pol.20160462 Landais C. Michaillat P. Saez E. A Macroeconomic Approach to Optimal Unemployment Insurance: Theory Am. Econ. J. Econ. Policy 10 2018 152 181 ISSN 1945-7731. https://www.aeaweb.org/articles?id=10.1257/pol.20150088. Lichter, A., Schiprowski, A., 2020. Benefit duration, job search behavior and re-employment. CESifo Working Papers Tex.date-added: 2017-12-12 13:36:52 +0000 tex.date-modified: 2020–05-14 14:57:32 +0000. Marinescu I. The general equilibrium impacts of unemployment insurance: Evidence from a large online job board J. Public Econ. 150 2017 14 29 ISSN 0047–2727. http://www.sciencedirect.com/science/article/pii/S0047272717300348 Marinescu I. Skandalis D. Unemployment Insurance and Job Search Behavior Q. J. Econ. 45 2021 Michaillat P. Do Matching Frictions Explain Unemployment? Not in Bad Times Am. Econ. Rev. 102 2012 1721 1750 Mitman K. Rabinovich S. Whether, When and How to Extend Unemployment Benefits: Theory and Application to COVID-19 2021 41 Montenovo, L., Jiang, X., Rojas, F.L., Schmutte, I., Simon, K., Weinberg, B., Wing, C., 2020. Determinants of Disparities in Covid-19 Job Losses. Technical Report w27132. National Bureau of Economic Research, Cambridge, MA. URL http://www.nber.org/papers/w27132.pdf. Petrosky-Nadeau, N., 2020. Reservation Benefits: Assessing Job Acceptance Impacts of Increased UI Payments. Technical Report 2020-28. Federal Reserve Bank of San Francisco. https://www.frbsf.org/economic-research/publications/working-papers/2020/28/. Pissarides C. Equilibrium Unemployment Theory second ed. 2000 MIT Press Cambridge Schmieder J.F. von Wachter T. The Effects of Unemployment Insurance Benefits: New Evidence and Interpretation Ann. Rev. Econ. 8 2016 547 581 10.1146/annurev-economics-080614-115758 van den Berg G. Nonstationarity in job search theory Rev. Econ. Stud. 57 1990 255 277

0

PMC9753141

NO-CC CODE

2022-12-16 23:25:31

no

J Public Econ. 2021 Aug 13; 200:104471

utf-8

J Public Econ

10.1016/j.jpubeco.2021.104471

oa_other

==== Front Journal of Academic Librarianship 0099-1333 0099-1333 Elsevier Inc. S0099-1333(21)00022-7 10.1016/j.acalib.2021.102331 102331 Article Polish university libraries social networking services during the COVID-19 pandemic spring term lockdown Gmiterek Grzegorz The Faculty of Journalism, Information and Book Studies, University of Warsaw, Bednarska Street 2/4, 00-310 Warsaw, Poland 18 2 2021 5 2021 18 2 2021 47 3 102331102331 7 12 2020 2 2 2021 3 2 2021 © 2021 Elsevier Inc. All rights reserved. 2021 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. During the spring 2020 COVID-19 lockdown, University libraries made available digital content of varying types. This article assesses the scope and breadth of content published by 18 Polish university libraries, how libraries compared in their approach to using social media, and the level of engagement and collaboration with patrons during a time when the physical library was closed. Data collection consisted of gathering social networking site information as used by Polish university libraries with the Fanpage Karma tool. The Fanpage Karma tool allowed the researcher to analyse and compare the fan pages of individual university library social networking sites. The tool simplifies the process to calculate key variables such as the number of user reactions to the library content; the number of posts, comments, likes, and dislikes; the average daily number of posts made; the most popular text, images, videos, links; and the kind of hashtags used by librarians. Findings indicate Polish university libraries most frequently used Facebook during the lockdown followed by Instagram and Twitter. To a much smaller extent, libraries also used YouTube. Not all Polish university libraries made equal use of social media during the lockdown despite the availability of content and ease of use. Keywords Social media Web 2.0 Academic libraries Library 2.0 COVID-19 Libraries and lockdown ==== Body pmcIntroduction Advancements and increased accessibility of digital information provide new opportunities for libraries. This is particularly true of interactive tools integrated with the social media environment commonly referred to as Web 2.0. Increasingly, these tools are becoming resources for university libraries. These applications became noticeably beneficial for libraries during the recent global pandemic as organisations sought new ways to engage with patrons. These communications channels provided a means to promote library news as well as to provide remote access to library resources. Social media use also provides an opportunity to collaborate with users, conduct online discussions, and to offer new ways to engage with university libraries. This article presents the results of social media use by 18 Polish university libraries, including public institutions (MNiSW, 2020), during the spring COVID-19 lockdown. The intention is to determine the scope and breadth of content published by university libraries, how these libraries compared in their approach to using social media, and the level of engagement and collaboration with patrons during a time when the physical library was closed. The research was aimed at answering seven specific research questions:RQ1: What social media tools did university libraries use most often during the lockdown? RQ2: What type of social media content did university libraries make available during the lockdown? RQ3: What social media content was the most popular and which hashtags were most often associated with that content during the lockdown? RQ4: What level of engagement did social media users have with university library social media accounts during the lockdown? RQ5: What was the frequency of social media content made available by individual university librarians during the lockdown? RQ6: Which university libraries were most active on social media during the lockdown? RQ7: Are there significant differences in the use of Web 2.0 tools among university libraries during lockdown? If so, what are they? Literature review Library social media use has become essential to adequately provide library services to a connected world. Social media platforms enable wider communication and cooperation between academic librarians and online users (Hicks, 2012). Social media is also increasingly used in library marketing (AlAwadhi & Al-Daihani, 2019; Cheng, Lam, & Chiu, 2020; Jones & Harvey, 2019; Mustafa, Zainuddin, Idris, & Aziz, 2016; Trucks, 2019). The scientific literature regarding the prevalence of social media use among academics is dominated by analysis of the tools used by librarians, their objectives, and the potential for their effective use (Burclaff & Johnson, 2014; Magoi et al., 2017). Research results indicate that Facebook, Twitter, YouTube, Instagram, LinkedIn, Tumblr, Pinterest, Flickr, and Vine are the most commonly used tools by librarians (Magoi et al., 2017). It is also unsurprising that a large proportion of librarians have a positive view of social media use in achieving library goals (Chua & Goh, 2010). Over the last ten years, social media has increasingly become the primary tool to disseminate information and promote library activities (Harrison, Burress, Velasquez, & Schreiner, 2017). The development and popularity of interactive and social tools have led university libraries to gradually implement IT solutions to better promote library services and resources (McCallum, 2015). Although, despite critics that reference the inefficiency of social media use among libraries (Magoi et al., 2017), it is evident that these tools have become a means for institutions to communicate with and meet the information needs of users (Ofili & Emwanta, 2014). Plus, innovative digital tools encourage younger generations to access library services (Williams, 2018), which is why university libraries need to participate in creating their online social reality (Garofalo, 2013). Effectively using social media is also seen as a critical 21st-century skill expected of academic library staff (Chawner & Oliver, 2013). Polish university libraries use social Internet tools for marketing and promotion of library services, communication, and cooperation with users (Gmiterek, 2012; Gmiterek, 2018; Onak, 2013-14; Scheffs, 2012; Sidorczuk & Gogiel-Kuźmicka, 2012; Stach-Siegieńczuk, 2014; Szmajser-Chylarecka, 2013). However, in the first decade of the 20th century, several academic libraries did not initially see the need for social media coverage (Lamberti & Theus, 2016). Only over time were these institutions convinced of the need to use social media tools in their daily operations. Today, this is the standard (Puksza & Witkowska, 2018) and all university libraries use at least Facebook. The desired goal is to build a committed library community and to establish a dialogue with patrons as well as to effectively adapt to better meet their needs (Jaskowska, 2012). Methods The research used qualitative and quantitative methods. Data collection consisted of gathering social networking site information as used by Polish university libraries with the Fanpage Karma tool (www.fanpagekarma.com). This tool was also used to analyse the collected data, which referred both to information made available by libraries and data coming from users, such as likes, comments, and posts (Fanpage Karma, 2018). The data was collected on June 24–27, 2020. Fanpage Karma is a network service used to monitor activities and content on social networking sites, made available on fan pages (e.g. institutions). The tool gathers public information, meaning there is no concern for confidential data or privacy infringement. The Fanpage Karma tool allowed for an analysis and comparison of fan pages of particular university libraries on the social media used by those libraries. The tool also facilitated a relatively accurate determination of the coefficients relating to the number of users' reactions to the shared library content; the number of posts, comments, likes; average number of shared posts per day; the most popular texts, images, videos, links; hashtags used by the librarians. The Fanpage Karma tool also makes it easier to visualize findings and organise the analysed content using tags. This mechanism can be useful when the user would like to organise the content they make available such as through the thematic categories with which they describe each post. The analysis covered the period from 11 March (first closures of library buildings for their readers) to June 2020, when all libraries were already partially open to users. However, all Polish university libraries were not closed and opened at the same time. For example, the University Library in Warsaw (BUW) was closed on 11 March. On 18 May, some services in the building of this institution resumed (e.g. lending and return of books, scanning of books and magazines for readers, interlibrary loans). The Jerzy Giedroyc University Library in Bialystok was closed on 16 March, and its lending library opened on 5 May. This is in contrast to the University Library in Rzeszow, which closed on 12 March and opened its lending library on 25 May 2020. On 1 June at the latest, the Library of the University of Łódź; the Nicolaus Copernicus University Library in Torun; the Main Library – University of Szczecin; the Scientific Information Centre and Academic Library (CINiBA) opened.1 Considering all the social networking sites university libraries use is a priority; however, in the analysed period the libraries studied did not present much content outside of Facebook. While the Nicolaus Copernicus University Library in Torun has been using the Pinterest service for years, during the analysed period there was no activity on the account. The official library profile or fan page on social networking sites was first identified by the information available on the institution's website. This method of identification was not always effective. In more than a few cases, there was no information about such activity on the library websites. The Library of the University of Lodz, which has a profile on YouTube, has no mention of it on its website much like the Library of Rzeszow University and the Library of the University of Zielona Gora. On the other hand, the University Library University of Warmia and Mazury in Olsztyn did not inform on its website about the librarians' Twitter account. The lack of information about all activities in the above-mentioned services forced additional searches for library profiles or fan pages on the most popular and most frequently used social networking sites. For this purpose, a standard Google search was conducted. Compiled library social media use is available in Table 1 .Table 1 Polish university libraries and the social services they use (June 2020). Table 1Name of the library Facebook Other tools University of Warsaw Library Fan page YouTube, Instagram- University Library in Bialystok Fan page YouTube, Instagram Library of the University of Gdansk Fan page None University Library in Poznan Fan page YouTube, Twitter, Instagram Jagiellonian Library Fan page Twitter, Instagram Library of the University of Lodz Fan page Instagram, Twitter, YouTube Main Library of Maria Curie-Sklodowska University in Lublin Fan page Twitter, Instagram Nicolaus Copernicus University Library in Torun Fan page YouTube, Twitter, Pinterest Main Library of the University of Opole Fan page None Main Library – University of Szczecin Fan page Twitter Library of the University of Silesiaa Fan page Instagram, Twitter, YouTube, Snapchat University Library in Rzeszow Fan page YouTube University Library University of Warmia and Mazury in Olsztyn Fan page Instagram, Twitter, YouTube, Instagram Wrocław University Library Fan page Twitter, YouTube Main Library of Cardinal Stefan Wyszynski University in Warsaw Fan page Twitter, YouTube Library of the University of Zielona Gora Fan page YouTube Library of Kazimierz Wielki University in Bydgoszcz Fan page None University Library in Kielce Fan page YouTube a Information about library profiles and fan pages is available through the website of the Scientific Information Centre and the Academic Library (CINiBA), which is a collaboration of two universities in Katowice: the University of Economics and the University of Silesia. The library social media profiles were added to the Fanpage Karma “dashboard.” No posts were added to Pinterest during the analysed period; therefore, it was not taken into account. Snapchat was also not included due to the specificity of its functions. Eight of the eighteen analysed libraries used Instagram. Unfortunately, Fanpage Karma was unable to analyse four of the Instagram library accounts: the University of Warsaw Library; Jagiellonian Library; Library of the University of Lodz; University Library University of Warmia and Mazury in Olsztyn because they are not public business accounts. Thus, it was not possible to use the data provided by these institutions. These library accounts are not business profiles and Instagram policy does not allow for connecting such accounts to external systems like Fanpage Karma. In short, the data provided by the four library institutions have not been marked as fully public, which means that they cannot be analysed and monitored by external tools. As a result, these four profiles were not considered when carrying out further analysis. The remaining four accounts were analysed using the Fanpage Karma tool. The researcher conducted a review of the social networking sites (Facebook, Twitter, Instagram, and YouTube) for the libraries in terms of the number of posts, their content, comments, user reactions, hashtag used by librarians, and the post-interaction index. The four social media networks provided a means of comparing the approach university libraries used to reach out to patrons during the lockdown. Results and data analysis Results suggest the most popular post among Polish library users was made by the Jagiellonian Library on Facebook announcing its partial reopening published on 20 May. The post collected 229 likes, 15 comments, 26 releases, and a total of 358 reactions from users. The second most engaging post was also made on Facebook by the University of Warsaw Library on 18 May. It also concerned the opening of the institution to patrons. The post gathered 215 likes, 3 comments, 13 releases, and 314 reactions. The third most engaging post, also on Facebook, was by the University of Warsaw Library on breaking a record of book lending published on 11 March. The post gathered 195 likes, 21 comments, 10 releases, and 359 reactions. It can be concluded that the content relating to the functioning of the physical institutions and the services they offer was the most frequently interacted with by patrons. In comparison, the most popular post on Instagram (5th overall) published by the Main Library of Maria Curie-Sklodowska University in Lublin on April 25th, entitled “Reading on the balcony, with a cherry blossom,” encouraged users to read; it gathered 103 likes. The second most favourable Instagram post (9th overall) was published by the University Library in Bialystok and also focused on reading, recommending books from a librarian's collection. The post appeared on 30 March and gained 86 likes and 8 comments. The third most popular Instagram post (11th overall) was also posted by the Bialystok library and featured a librarian's recommended book, which gathered 84 likes. The most popular Twitter post only ranked 225 overall. The post was published by the University Library in Poznan on 30 March and referred to remote access to JSTOR, a consortium of libraries and scientific publishers, available until 30 June gathered 17 likes. Whereas the most popular video on YouTube was material prepared by librarians from the University Library in Bialystok. The video was produced as part of Library Week and presented the work of the Collection Sharing Department gathering 8 likes and displayed a total of 207 times. An analyses of the activity of Polish university libraries on each of the social networking sites are detailed individually below. Facebook In the period studied, the University of Warsaw Library had the largest number of Facebook followers (19,004 users). Next was the Library of the University of Lodz (6280 users), University Library in Poznan (5899 users), Jagiellonian Library (5897 users), and Nicolaus Copernicus University Library in Torun (4960 users). The Main Library of the University of Opole (136 users) and the Main Library – University of Szczecin (243 users) saw the lowest number of Facebook followers. In the analysed period, all libraries published a total of 1171 posts. The largest number of posts, as many as 242, was published by the University Library in Bialystok. These posts were on average three per day. Users added 38 comments to these Facebook posts, but all user reactions (including likes, comments, sharing, etc.) totalled 1615 including 1139 likes. The Facebook posting interaction rate for the University Library in Bialystok was 0.66%. This is the engagement rate in Fanpage Karma tool, which shows “an average amount of how often a fan interacts with the posts of a page. It is calculated by dividing the daily amount of likes, comments and shares by the number of fans” (Wusthoff, 2014). The most popular posts based on the number of likes and dislikes were those related to Librarian's Day (8 May), construction of a new library building (22 April), special collections, and a digitisation workshop (15 May). The 2 April Facebook post referring to the International Children's Book Day received three comments. However, a March 16th post which focused on remote access to the library resources was the most frequently made available (7 times). Throughout the Facebook posts, the most frequently used hashtags are #zostańwdomu (eng. #stayhome) (28 times); #tydzieńbibliotek (eng. #librariesweek) (20 times); and #przystańzbiblioteką (eng. #libraryharbour) (18 times); #zasmakujwbibliotece (eng. #tasteitinthelibrary) (18 times); and #niezostawiamczytelnika (eng. #I'mnotleavingthereader) (18 times). The University Library in Bialystok published the most posts on Wednesdays and Fridays (38 times each); although, there was also content available on Saturdays (27 posts) and Sundays (30 posts each). The most frequently published posts were accompanied by media (132 documents) and links (75 documents). Considering the number of posts made available, the second most engaged with library was the University Library in Poznan, which published 115 posts (1.4 per day on average). Users added a total of 30 comments and 1466 reactions including 892 likes. The interaction rate for the University Library in Poznan was 0.19%. The most popular posts included one on 27 March, referred to the capability of users to order an electronic copy of a publication by academics, doctoral students, and students (26 impressions); 11 May about the resumption of the library's activity; and of Easter Christmas greetings of 9 April. Most comments were gathered by the 24th April post with a riddle concerning the photo presented in the post (4 comments). In the posts, hashtags most often used are #tydzieńbibliotek2020 or #librariesweek2020 (8 times), #zasmakujwbibliotece (eng. #tasteitinthelibrary) (6 times), and #zostańwdomu (eng. #stayhome) (4 times). Most library posts were made on Tuesdays and Thursdays (23 each). The most frequently published posts were accompanied by media (57 documents) and video materials (27 documents). The third most active library on Facebook is the Library of the University of Lodz, which in the analysed period, made 109 posts available (1.3 per day on average). Users added a total of 101 comments and a total of 4225 reactions; most of all of the libraries studied, including 2590 likes. The engagement rate was 0.59%. The three most popular posts were published on 15 March which promoted museums presenting their collections online and on 18 April and 22 May with pictures of the library and wishes for its quick reopening for patrons. The post of 21 March, which referred to the remote ordering of electronic copies from magazines and books received the most comments (32 comments). The post promoting museum collections online was the most frequently made available (89 contributions). In the posts, hashtags used are #bułateam or #libraryoftheuniversityoflodzteam (16 times); #repozytoriumuł (eng. #repositoryoftheuniversityoflodz) (8 times); and #zostańwdomu (eng. #stayhome) (5 times). The library published the most posts on Wednesdays and Thursdays (21 documents each). Most of the posts were accompanied by media (104 documents). The University of Warsaw Library saw the most number of Facebook user comments with 105 overall. This institution was also the second-largest in terms of the number of user reactions to the posts it made available with 3697 in total, including 2218 likes. Meanwhile, the Library of the University of Lodz had the most reactions; 4144 reactions, including 2590 likes. The fewest number of Facebook posts was made by the University Library in Rzeszow with 21 posts and 104 user reactions, including 6 comments and 56 likes; the Library of the University of Zielona Gora with 13 posts and 25 user reactions, including 2 comments and 128 likes; and the Main Library of the University of Opole with 13 posts and 25 reactions, including 14 likes. Considering all libraries, on average they made 0.8 posts per day. Most content was published on Wednesdays (227 posts), Thursdays (211 posts), and Tuesdays (209 posts). A few Facebook posts were made on Sundays (45 posts) and Saturdays (75 posts). The most frequently used hashtags by all the university libraries include: #zostańwdomu (eng. #stayhome) (118 times), #ciniba (53 times), #niezostawiamczytelnika (eng. #Iamnotleavingthereader) (47 times), #bibliotekaotwarta (eng. #openlibrary) (44 times), and #tydzieńbibliotek (eng. #libraryweek) (28 times). Regarding the largest number of reactions, the most popular Facebook posts were two from the University of Warsaw Library and the Jagiellonian Library. The first post published on 11 March by the University of Warsaw Library concerned a record number of document loans (351 reactions). The second post on 20 May by the library from Krakow, referred to the opening of the institution (350 reactions). Among the 18 Polish university library Facebook text postings, the most user reactions (59 reactions) were received by the Main Library of Maria Curie-Sklodowska University in Lublin. The post concerned the resumption of lending and returning books. Among graphic posts, the most popular was on 11 March with a record number of document loans, published by the University of Warsaw Library (351 reactions). As far as video materials are concerned, the most popular was the post published on 15 May by the Nicolaus Copernicus University Library in Torun (79 reactions). The post concerned a meeting devoted to the historical book “Powrót Pomorza w granice Rzeczpospolitej w setną rocznicę 1920 – 2020” or in English “Return of Pomerania to the Borders of the Republic of Poland on The Hundredth Anniversary of 1920 – 2020.” Among the most popular posts published on Facebook, the most “liking” (229 likes) was attributed to the 20 May Jagiellonian Library post, which informed patrons of the reopening of the library. The second most popular post was the University of Warsaw Library 18 May post with 215 likes. This post also concerned the reopening of the library following the lockdown. What is more, considering the 10 most popular posts (analysed in terms of the number of likes and dislikes) the third, fourth, fifth, sixth, and eighth-most liked posts were made by the library in Warsaw. These were posts relating to library services (making the library book drop available, waiving fees for scanning materials, a record number of loans, as well as graphics prepared for the library by University of Warsaw students). The seventh most liked post was by the Library of the University of Lodz on 15th March about museum collections made available online. Meanwhile, the ninth most liked Facebook post was by the Scientific Information Centre and Academic Library on 27th April about the launch of bookstores in the library building. Finally, the tenth most liked post was by the Nicolaus Copernicus University Library in Torun on March 11 referring to an increase in borrowed document limits. The most frequently commented on Facebook posts were from the University Library in Torun and the University of Warsaw Library. In the first case, there were two posts on 15 and 13 May with the highest number of comments (30 and 23 comments respectively). They referred to the previously mentioned meeting devoted to the book “Powrót Pomorza w granice Rzeczpospolitej w setną rocznicę 1920 – 2020” and the author's meeting with Barbara Klicka – poet, writer, and cultural animator. The University of Warsaw Library posts on 15 May and 11 March published news relating to the partial reopening of this institution and the record number of document loans made by users. Instagram Eight libraries used the Instagram service. As previously indicated, Fanpage Karma was unable to analyse four other Instagram accounts because they are not business profiles. Therefore, only four Polish University library Instagram accounts, which published a total of 73 posts, were analysed. The accounts belong to the University Library in Bialystok, University Library in Poznan, Scientific Information Centre and Academic Library, and the Main Library of Maria Curie-Sklodowska University in Lublin. The University Library in Bialystok published most often with 34 total posts. The Library in Poznan had 14 posts and CINiBA 13 posts. The Library of the Maria Curie-Sklodowska University in Lublin published 12 posts in the analysed period. Most users commented on the posts added by the University Library in Bialystok (34 comments). In the case of the other libraries, the number of comments was respectively: Library in Lublin (8 comments), CINiBA (4 comments), and Library in Poznan (3 comments). The most reactions among users were aroused by the posts of the University Library in Bialystok with 1891 reactions, including 1857 likes. The second most popular Instagram account was the Main Library of the Maria Curie-Sklodowska University in Lublin with 883 reactions, including 875 likes followed by CINiBA with 367 reactions, including 363 likes. The Library in Lublin had the largest number of observers (1094 followers) proceeded by the University Library in Bialystok (1081 followers). The CINiBA account saw 630 followers and the University Library in Poznan Instagram account had 590 followers. The Library in Bialystok published an average of 0.4 posts per day. The most popular among users were the posts published between 24 and 30 March, in which librarians recommended books worth reading, and a post on 12 March that called for reflection on wasting food. The most engaging posts included those on 30 March with recommended books; 10 April with Christmas wishes; and 23 March with a photo of a librarian's home workstation. University Library in Bialystok hashtags most often used include: #bu_uwb (eng. #UniversityLibraryUniversityinBialystok) (34 times); #bookstagram (19 times), #instabook (18 times), and #buchstagram (16 times). Instagram posts were most often published on Wednesdays and Fridays (7 posts) and Thursdays (6 posts). The University Library in Poznan published an average of 0.2 posts per day. The most popular entries were those on 20 March with information about ebooks available for free online; 26 April about open resources at the University of Poznań; and 13 May with a post about architecture. Hashtags most frequently used on the University Library in Poznan account include #zasmakujwibibibiboteka (eng. #tasteitinthelibrary) (9 times); #tydzieńbibliotek2020 (eng. #librariesweek2020) (6 times); and #tydzieńbibliotek (eng. #librariesweek) (2 times). CINiBA published an average of 0.2 hashtags per day. The most popular were the entries on 1 April with a primaaprilis joke; 22 May on the 23rd anniversary of the Animal Protection Act passed by the Sejm, and 4 May on the opening of the library's lending library and the capability to use book machines. CINiBA's Instagram account hashtags most often used are #ciniba (11 times), #bibliotekaotwarta (eng. #libraryopened) (7 times), #książka (eng. #book) (2 times). Most frequently, posts were published on Mondays (3 times) and Wednesdays (2 times). The Main Library of Maria Curie-Sklodowska University in Lublin published 0.1 posts a day. The most popular entries were those on 25 April with the title: Reading on the balcony, with cherry blossom; 29 May with a post presenting books of the late Jerzy Pilch, and 22 March with a post encouraging users to boast about their home libraries. Hashtags most frequently used by the Main Library of Maria Curie-Sklodowska University in Lublin include #bookstagram (8 times), #bibliotekaumcs (eng. #libraryofUMCS) (8 times), #libraryumcs (7 times), #instabook (7 times), and #terazitam (eng. #nowandthere) (7 times). Most posts were published on Saturdays (4 times) and Wednesdays (3 times). Table 2 presents aggregate data for all the accounts analysed.Table 2 Use of Instagram by libraries (ranking by number of posts made available). Data collected and processed through the Fanpage Karma tool. Table 2Name of the library Number of posts Number of likes Posts–daily average Number of comments All reactions University Library in Bialystok 34 1857 0,4 34 1891 Main Library of the University of Opole 14 190 0,2 3 193 Scientific Information Centre and the Academic Library (CINiBA) 13 363 0,2 4 367 Main Library of Maria Curie-Sklodowska University in Lublin 12 875 0,1 8 883 Twitter Ten libraries' Twitter accounts with a total of 61 posts published were studied. These accounts included the following institutions: Library of the University of Lodz (652 followers), University Library in Poznan (423 followers), University Library in Torun (316 followers), Jagiellonian Library (305 followers), Main Library of Maria Curie-Sklodowska University in Lublin (214 followers), CINiBA (209 followers), Main Library of Cardinal Stefan Wyszynski University in Warsaw (201 followers), Wroclaw University Library (192 followers), University Library University of Warmia and Mazury in Olsztyn (18 followers), and Main Library – University of Szczecin (7 followers). In general, the Twitter microblogging service was not as popular in the analysed period as Facebook or Instagram. Only five of the above-mentioned institutions were active on Twitter. The University Library in Poznan published frequently with 22 posts. In total, these posts gathered 128 reactions, including 82 likes and 46 retweets. The Library published an average of 0.3 posts a day. The most popular post was on 30 March referring to making the collection of digital documents available by the library; the post gathered 17 likes and 8 retweets. The second most popular post was on 26 March promoting the “Remote accessible resources” toolkit with 10 likes and 5 retweets. The third most favourite post was on 7 May and referred to the centenary jubilee of the University with 6 likes and 3 retweets. In Polish university library published tweets there appeared isolated hashtags. These were, for example, #dzienmatki (eng. #mother'sday), #zasmakujwbibliotece (eng. #tasteitinthelibrary), #tydzieńbibliotek (eng. #librariesweek), #ŚwiatowyDzieńWłasnościIntelektualnej (eng. #World DayofInteractiveProperty). The Jagiellonian Library published 14 tweets gathering 14 user reactions, including 11 likes and 3 retweets. The most popular was the post on 18 May about technical problems with adapting the Library to epidemiological restrictions with 8 likes and 2 retweets, the post on 26 May sending wishes for Mother's Day with 4 likes and 2 retweets, and the post on 27 May referring to the winners of the Adam Lysakowski Scientific Award of the Association of Polish Librarians for 2019 with 4 likes and one retweet. Also, in the case of the Jagiellonian Library Twitter account, there were isolated hashtags used, such as #dzienmatki (eng. #mother'sday), #nagroda (eng. #award), and #webinar. Among the analysed library accounts on Twitter, the Library of the University of Lodz remained least active. The Twitter account published a total of 9 tweets, which gathered 31 reactions, including 21 likes and 10 retweets. The most popular post on 12 March referred to the reception of books by users from the library. The post gathered 8 likes and 4 retweets. The second most popular tweet contained a quote from the late Jerzy Pilch gathering 7 likes and 3 retweets. In the tweets hashtags used include #repozytoriumuł (eng. #repositoryoftheUniversityofLodz), #bułateam (eng. #teamofthelibraryoftheUniversityofŁodz), #motoryzacja (eng. #utomotive), and #uł or (eng. #theUniversityofLodz). Table 3 presents aggregate data for all the accounts analysed.Table 3 Use of Twitter microblogging service by libraries (ranking by number of posts). Data collected and processed through the Fanpage Karma tool. Table 3Name of the library Number of posts Number of likes Posts – daily average The number of likes – average for one tweet All reactions University Library in Poznan 22 82 0,3 3,9 128 Jagiellonian Library 14 11 0,2 0,8 14 Library of the University of Lodz 9 21 0,1 2,3 31 Main Library of Cardinal Stefan Wyszynski University in Warsaw 2 0 0,02 0 0 Scientific Information Centre and the Academic Library (CINiBA) 1 1 0,01 1,0 1 YouTube Nine Polish university libraries have accounts on the YouTube social network. In the analysed period these accounts published 35 videos. However, only five institutions were active on YouTube at the time. The University Library in Bialystok published the most videos with as many as 14 receiving 39 likes. The second most popular YouTube account was CINiBA with 10 videos and 6 likes. Rounding out third and fourth the University Library in Torun provided 6 video sequences with 5 likes and the University of Warsaw Library provided 3 documents with 15 likes. Two materials were published by the University Library in Poznan, which received 2 likes (Table 4).Table 4 Use of YouTube service by libraries (ranking by number of posts). Data collected and processed through the Fanpage Karma tool. Table 4Name of the library Number of posts Number of likes Total number of views University Library in Bialystok 14 39 79 CINiBA 10 6 104 University Library in Toruń 6 5 32 University of Warsaw Library 3 15 299 University Library in Poznan 2 2 18 Taking into account all the posts published in the analysed period, the most popular YouTube video (calculated by the number of impressions) was the manual “How to Log in to Databases Subscribed by the University of Silesia” prepared by CINiBA (535 impressions) and published on 25 March. The second most popular video was the instruction of the University of Warsaw Library published on 27 May that related to the new rules of document lending (397 impressions). Following up in third, by the University of Warsaw Library instruction to log in to the service with ebooks “IBUK Libra” with 271 impressions. Table 5 presents statistics on the video materials made available by the Polish university libraries on YouTube.Table 5 10 most popular posts published by Polish university libraries on YouTube (ranking by the number of views of a given video). Data collected and processed through the Fanpage Karma tool. Table 5Name of the library Date of posting The topic of the post Number of views Number of likes CINiBA 25.03 Login instructions (bases) 535 3 University of Warsaw Library 27.05 New rules for rented documents 397 5 University of Warsaw Library 6.05 Login instructions (IBUK Libra) 271 5 University of Warsaw Library 6.05 Instructions for changing your account password 228 5 University Library in Bialystok 11.05 Lending Department (Libraries Week 2020) 206 8 CINiBA 25.03 Database instructions (EBSCOhost) 185 0 University Library in Bialystok 11.05 Librarians in leisure time (Libraries Week 2020) 115 6 University Library in Bialystok 8.05 Acquisitions and Collection Development Section 113 4 University Library in Bialystok 8.05 Call for Library Week (2020) 97 2 CINiBA 25.03 Instructions for using the Academic Research Source eJournals database 92 0 In the case of the Library in Bialystok, all the YouTube posts referred to Library Week and the functioning of individual branches of this institution. The most popular post with 206 impressions and 8 likes was the material about the library making the collections available online. The next most popular posts focused on one librarian's gardening interests garnering 115 impressions and 6 likes and the collection accessibility branch with 113 impressions and 4 likes. The Scientific Information Centre and Academic Library (CINiBA), presented in video materials on YouTube a set of tutorials on how to use the library's resources. The most popular turned out to be the manual for logging into databases subscribed by the University of Silesia published on 25 March with 535 impressions. Next favourable was the instruction to use databases on the EBSCOhost platform with 185 impressions and then the manual for using the Academic Research Source eJournals database with 92 impressions. As part of Library Week, the University Library in Torun published six video documents. The most popular being two that presented Omeka software and tips on how to install and configure the software as well as how to create virtual displays. The materials were viewed 103 times in total. The Library also presented through YouTube video a record of the author's meeting with the poet and writer Barbara Klicka. The material was viewed by users 56 times. Conclusion The increased use of Web 2.0 tools by Polish university libraries is a phenomenon that has evolved over the last fifteen years. These tools have also become essential for communicating with library patrons. This analyses show that Facebook was actively used by all 18 Polish university libraries studied during the period in question. Individual librarians also posted the largest amount of multimedia content, which met with the greatest response from users, considering the response of all the services analysed. Facebook is therefore the most frequently used social network service by Polish university libraries. In several cases, it is currently also one of the basic tools of communication with online users, as well as a platform for publishing multimedia content related to information and library services as well as collected resources. The next most used social media network by university libraries is Instagram. While significantly less active than Facebook, 8 libraries regularly posted content. Importantly, four of these institutions host fan pages that are not business profiles and have not been marked as fully public, which is directly linked to the lack of detailed analysis of their content. In the case of Twitter, university libraries rarely used the opportunity to post on Twitter. Also, only five institutions were active during the analysed period on Twitter. Their activity was sporadic and met with limited response from users. Even more rarely, libraries used the YouTube service infrequently. Often, the content was directly related to a specific event, such as Library Week, or a desire to present library services, collections, and library-friendly artists. YouTube is among one of the few social media networks that doesn't allow communicating as much through text. Rather, it serves to complement the content presented through Facebook and Instagram, emphasizing the role of the former. Findings show that not all Polish university libraries have made equal use of the opportunities provided by social media, whether in terms of the number of tools or the amount of content made available through them. Nevertheless, libraries did make digital content available for patrons during the COVID-19 spring lockdown. Posts included texts, graphic information, video sequences, and links to direct the user's attention to the documents, events, and services offered by the library to engage with patrons while the physical location was closed. Funding This work was supported by 10.13039/501100006445 University of Warsaw [research mini-grants numbers PSP 501-D127-20-0004316]. CRediT authorship contribution statement Grzegorz Gmiterek: Conceptualization, Methodology, Investigation, Writing - Original draft. Declaration of competing interest None. 1 Information on the closure and opening of libraries was provided by data published on the institutional websites and on social media. ==== Refs References AlAwadhi S. Al-Daihani S.M. Marketing academic library information services using social media Library Management 40 3/4 2019 228 239 10.1108/LM-12-2017-0132 Burclaff N. Johnson C. Developing a social media strategy: Tweets, pins, and posts with a purpose College & Research Libraries News 75 7 2014 366 369 10.5860/crln.75.7.9156 Chawner B. Oliver G. A survey of New Zealand academic reference librarians: Current and future skills and competencies Australian Academic & Research Libraries 44 1 2013 29 39 10.1080/00048623.2013.773865 Cheng W.W.H. Lam E.T.H. Chiu D.K.W. Social media as a platform in academic library marketing: A comparative study The Journal of Academic Librarianship 46 5 2020 102188 10.1016/j.acalib.2020.102188 Chua A.Y.K. Goh D.H. A study of Web 2.0 applications in library websites Library and Information Science Research 32 3 2010 203 211 10.1016/j.lisr.2010.01.002 Fanpage Karma Fanpage Karma Retrieved from http://www.fanpagekarma.com/features 2018 Garofalo D.A. Building communities: Social networking for academic libraries 2013 Chandos Publishing Gmiterek G. Biblioteka w środowisku społecznościowego Internetu: biblioteka 2.0 2012 Wydawnictwo SBP Gmiterek G. Biblioteka w Środowisku Społecznościowego Internetu i Interaktywnych Technologii (w 2017 r.) Chałubińska-Jentkiewicz K. Kakareko K. Sobczak J. Biblioteki i archiwa na jednolitym rynku cyfrowym 2018 Silva Rerum 27 44 Harrison A. Burress R. Velasquez S. Schreiner L. Social Media Use in Academic Libraries: A Phenomenological Study The Journal of Academic Librarianship 43 3 2017 248 256 10.1016/j.acalib.2017.02.014 Hicks A. Review of “social networking tools for academic libraries” Collaborative Librarianship 4 4 2012 190 191 Jaskowska B. Bądźmy tam gdzie oni! Elementy social media marketingu w bibliotece Kudrawiec J. Bibliotekarze bez bibliotek czyli bibliotekarstwo uczestniczące: III Ogólnopolska Konferencja Naukowa Białystok, 18–20 maja 2011 2012 Wydawnictwo Uniwersytetu w Białymstoku 295 310 Jones M.J. Harvey M. Library 2.0: The effectiveness of social media as a marketing tool for libraries in educational institutions Journal of Librarianship and Information Science 51 1 2019 3 19 10.1177/0961000616668959 Lamberti M.A. Theus M. Media społecznościowe w polskich bibliotekach, archiwach i muzeach Biblioteka 20 29 2016 183 206 10.14746/b.2016.20.11 Magoi J.S. Aspura M.K.Y.I. Abrizah A. Social media engagement in developing countries: Boon or bane for academic libraries? Information Development 2017 10.1177/2F0266666917748985 McCallum I. Use of social media by the library: Current practices and future opportunities. A white paper from Taylor & Francis The Australian Library Journal 64 2 2015 161 162 10.1080/00049670.2015.1040364 MNiSW Wykaz uczelni publicznych nadzorowanych przez Ministra właściwego ds. szkolnictwa wyższego - publiczne uczelnie akademickie - Portal Gov.pl Retrieved from https://www.gov.pl/web/nauka/wykaz-uczelni-publicznych-nadzorowanych-przez-ministra 2020 Mustafa A. Zainuddin I.N. Idris S.R.A. Aziz M.F.A. Social Media Promotional Tools in Academic Library International Journal of Computer Theory and Engineering 8 3 2016 260 264 10.7763/IJCTE.2016.V8.1055 Ofili D.N. Emwanta M.G. Facebook as an information service delivery tool: Perspectives of library staff at the University of Benin, Nigeria African Journal of Library, Archives and Information Science 24 2 2014 195 202 Onak J. Serwis Facebook jako narzędzie promocji bibliotek Folia Bibliologica 55–56 2013-14 161 169 10.17951/fb.2013.55.161 Puksza K. Witkowska E. Facebook, Instagram, blog… i co dalej? – Biblioteka Uniwersytecka im. Jerzego Giedroycia w Białymstoku w mediach społecznościowych Biuletyn EBIB 178 2018 Retrieved from http://ebibojs.pl/index.php/ebib/article/view/83/84 Scheffs M. Facebook jako forma promocji biblioteki Toruńskie Studia Bibliologiczne 1 8 2012 117 140 10.12775/TSB.2012.007 Sidorczuk A. Gogiel-Kuźmicka A. Web 1.0, Web 2.0, czy może już Web 3.0? — narzędzia i technologie informacyjno-komunikacyjne stosowane na stronach WWW bibliotek technicznych szkół wyższych w Polsce Biuletyn EBIB 2 129 2012 Retrieved from http://www.ebib.pl/images/stories/numery/129/129_sidorczuk.pdf Stach-Siegieńczuk A. Wykorzystanie narzędzi Web 2.0 przez wybrane biblioteki uczelniane najlepszych uczelni odnotowanych w QS World University Rankings 2013 Toruńskie Studia Bibliologiczne 2 13 2014 161 188 10.12775/TSB.2014.023 Szmajser-Chylarecka D. Facebook, jako forma promocji, na przykładzie polskich bibliotek Forum Bibliotek Medycznych 6/2 12 2013 319 327 Trucks E. Making Social Media More Social: A Literature Review of Academic Libraries’ Engagement and Connections Through Social Media Platforms Joe J. Knight E. Social Media for Communication and Instruction in Academic Libraries 2019 Information Science Reference Hershey, PA https://digitalcommons.du.edu/libraries_facpub/25 Williams M.L. The adoption of Web 2.0 technologies in academic libraries: A comparative exploration Journal of Librarianship and Information Science 52 1 2018 137 149 10.1177/2F0961000618788725 Wusthoff K. Social media KPIs explained Retrieved from https://blog.fanpagekarma.com/2014/07/01/social-media-kpis-explained/ 2014

0

PMC9753146

NO-CC CODE

2022-12-16 23:25:31

no

2021 May 18; 47(3):102331

utf-8

nan

10.1016/j.acalib.2021.102331

oa_other

==== Front Journal of Academic Librarianship 0099-1333 0099-1333 Elsevier Inc. S0099-1333(21)00037-9 10.1016/j.acalib.2021.102346 102346 Article Readiness for Online Learning during COVID-19 pandemic: A survey of Pakistani LIS students Rafique Ghulam Murtaza a⁎ Mahmood Khalid b Warraich Nosheen Fatima b Rehman Shafiq Ur c a Department of Information Management, University of Sargodha, Sargodha, Pakistan b Institute of Information Management, University of the Punjab, Lahore, Pakistan c Deanship of Library Affairs, Imam Abdulrahman Bin Faisal University, Dammam, Saudi Arabia ⁎ Corresponding author. 1 3 2021 5 2021 1 3 2021 47 3 102346102346 28 12 2020 15 2 2021 22 2 2021 © 2021 Elsevier Inc. All rights reserved. 2021 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. This study was designed to determine the perceived online learning readiness (OLR) of LIS (Library and Information Sciences) / IM (Information Management) students in Pakistan during COVID-19 pandemic. A quantitative approach based on survey method was used to collect data from 340 LIS students from nine public sector universities of Pakistan through an online questionnaire. The collected data was analyzed using the SPSS and AMOS. The findings revealed that LIS students were not fully personalized and successful in decisions about their online educational activities during COVID-19 pandemic. However, they were motivated to learn through online learning and felt confident in performing basic functions of computers and internet. A significant difference of opinion was observed on ‘computer/internet self-efficacy’ and ‘online communication self-efficacy’ based on respondents' gender. Similarly, students from different levels of degree programs reported significantly different computer, internet, and online communication self-efficacy and learning motivation. Moreover, the age and grades of respondents were noted to be strong predictors of their OLR. These findings would be helpful for library schools, universities, and faculty members in Pakistan to improve the quality of online education and implement clear policies and guidelines. This study provides some theoretical and practical implications based on the findings. Keywords Online learning COVID-19 Library and Information Science (LIS) Online Learning Readiness (OLR) Students Pakistan ==== Body pmcIntroduction The Corona Virus Disease 2019, commonly referred to as COVID-19, which appeared in the first quarter of the year 2020 and quickly spread all over the world has, indubitably, forced the global learner community to transition from the traditional in-class method of learning to a mode of online learning within a very short period of time. The COVID-19 pandemic has brought the world to a standstill, entire cities have been locked down, and people have been restricted to their homes in order to stop or slow the spread of this disease. Despite this challenging situation, most academic institutions in the world have tried to ensure the continuity of the learning process. They have shifted to an online mode of learning, where the students and teachers interact with each other using different kinds of technological tools and techniques. This method of learning is also referred to as e-learning. Even though some institutions had been experimenting with e-learning before the pandemic, it is only now that its full benefit has been realized. It provides the students with a lot of opportunities and benefits such as convenience (Poole, 2000), flexibility (Chizmar & Walbert, 1999), time-saving, teamwork, as well as opportunities to collaborate with others across physical boundaries (M.L. Hung, Chou, Chen, & Own, 2010). It also allows students to have more control over their learning activities and to make decisions about their routine classwork in terms of space, pace, depth, breadth, and time management (Stansfield, McLellan, & Connolly, 2004). In Pakistan, like many other countries of the world, the Higher Education Commission (HEC) and Higher Education Department (HED) have mandated that all public and private sector educational institutes should conduct all their teaching and learning activities online until the curve of the spread of COVID-19 is flattened (Higher Education Department, 2020). Consequently, universities in Pakistan have been actively trying to transform their pedagogical teaching and learning activities into a virtual model. Most of them have developed efficient online learning systems and have transitioned to online learning. Teachers have uploaded their lessons, projects, groups work, and reading material into these online learning management systems, and virtual classes have been organized using various videoconferencing applications such as Zoom, Google Meet, Microsoft Teams, Webex, etc. It cannot be denied that online learning was the best solution to this unprecedented situation. However, it does have some drawbacks, as Chung, Subramaniam, and Dass (2020) asserted, it cannot replicate the direct face-to- face human contact, or the level of social engagement one experiences when in a classroom. These challenges may cause the students to feel that something is lacking, and result in decreased student engagement and interaction with a resultant subpar learning experience. Technology has played a pivotal role during the e-learning process. Various Information & Communication Technology (ICT) tools such as a desktop computer, laptop, smartphone, smart device, an internet connection, and online learning platforms (software/mobile apps) are essential for e-learning. The students have to have access to some of these to successfully participate in e-learning. Furthermore, they can use different kinds of communication tools to communicate with one another and their teachers to easily share information and knowledge. There are various asynchronous (threaded discussion, emailing) and synchronous (live chat, live audio/video call, instant messaging) tools, that are widely available and very user-friendly. Thus, online learning provides a computer-mediating environment for sharing one's perspective with others confidently and comfortably. However, it needs the users to have a certain degree of technical training and aptitude for virtual learning. A certain degree of preparedness is essential if users are to gain maximum benefit from this virtual mode of learning. This has raised the question of whether the students in Pakistan are ready and prepared to easily adopt e-learning and cope with the challenges associated with it. This study was designed to explore the level of online learning readiness (OLR) of students in Pakistan. Additionally, it also investigated whether any demographic factors (gender, age, various levels of LIS degree program, and their grades) made an impact on their OLR. Several studies have been conducted to determine the online learning readiness (OLR) of students (Afrianti & Aditia, 2020; Chung et al., 2020; Firat & Bozkurt, 2020; Forson & Vuopala, 2019; Hergüner, Son, Son, & Dönmez, 2020; Joosten & Cusatis, 2020; Zgheib, AlDaia, Serhan, & Melki, 2019), but none have been conducted in Pakistan or the context of Library and Information Sciences (LIS) students. Therefore, this study was designed to investigate the online learning readiness of LIS students currently enrolled at HEC recognized public sector universities of Pakistan. Research questions The following research questions were framed to address the research objectives:RQ1 – What has been the perceived level of readiness towards online learning (OLR) of LIS students during the COVID-19 pandemic? RQ2 – Does gender predict the perceived OLR of LIS students? RQ3 – Is there any significant impact of age on their OLR? RQ4 – Does the level of various LIS programs such as BS, MA, M. Phil., and Ph. D. affect the LIS students' perceived OLR during the COVID-19 pandemic? RQ5 – Do the grades achieved by LIS students significantly influence their readiness for online learning during COVID-19? Theoretical perspectives and literature review Researchers have been trying to measure the level of online learning readiness (OLR) of students ever since the advent of online learning. The first concept of online learning readiness was proposed by Warner, Christie, and Choy (1998). They defined it as 1) students' preference for classroom instructional method against the face-to-face learning, 2) students' confidence in using different kinds of technology, internet, and especially computer-mediated tools for communication in online learning, and 3) students' engagement in their autonomous learning. McVay, 2000, McVay, 2001, later developed a 13-item scale to measure the readiness of students towards online learning. In the McVay questionnaire, the students' attitudes and behavior were taken as predictors. P.J. Smith, Murphy, and Mahoney (2003) conducted a study using M. McVay's (2001) Readiness for Online Learning questionnaire and reported that the students' self-management of learning and their level of comfort with e-learning were the two main factors that predicted their success. However, these two factors did not comprehensively cover all dimensions of students' readiness towards online learning. To expansively understand the core of online learning readiness, researchers put their efforts into developing more dimensions that would broadly cover all necessary aspects of online learning. Previous studies have found that the technical skills needed to perform computer and computer-based tasks were also the key determinant factors of students' performance in a web-based learning environment (Peng, Tsai, & Wu, 2006). Similarly, the students' perceptions regarding the internet were also related to their attitudes and behavior towards online learning (C.C. Tsai & Lin, 2004). Another important factor affecting the students' OLR was found to be their ability to manage their time. In 2010, M.L. Hung et al. (2010) developed a comprehensive scale to measure the readiness of students regarding online learning. The scale covered all aspects of OLR and had five dimensions: 1) computer/internet self-efficacy, 2) self-directed learning, 3) learner control, 4) motivation for learning and 5) online communication self-efficacy. The conceptual model of OLR has been structured around the following dimensions: Computer & internet self-efficacy Self-efficacy is a person's particular set of beliefs that determine how well one can execute a plan of action in prospective situations (A. Bandura, 1977). As online learning is delivered through online networks, therefore, it is essential to determine the perception of students about ICTs, and to assess their competencies in using these technologies for online learning. The underlying theory of assessing self-efficacy is the ‘social cognitive theory’ which provides the basis for understanding self-efficacy beliefs through cognitive, motivational, affective, and decisional processes (Bandura, 1977, Bandura, 1986, Bandura, 1997). Accordingly, several scales have been developed to measure the computer and internet self-efficacy of individuals. A 10-item instrument by Compeau and Higgins (1995) has identified that computer self-efficacy had a significant impact on computer-use outcomes, computer user's emotional reactions, and actual computer use. Similarly, Eastin and LaRose (2000) pointed out that internet self-efficacy was not merely uploading or downloading files but was also related to the ability of an individual to apply his/her higher-level skills in troubleshooting and problem-solving technical problems while using the internet. M.J. Tsai and Tsai (2003) found that students with high internet self-efficacy performed and learned better than those with lower internet self-efficacy during online learning. Self-directed learning Knowles (1975) defined self-directed learning (SDL) as the process of taking the initiative to understand one's learning needs, establish learning goals, identify human and material resources needed for learning, choose and implement the appropriate learning strategy, and evaluate learning outcomes. Based on Knowles' work, Guglielmino (1977) developed a scale, the Self-Directed Learning Readiness Scale (SDLRS), to help determine students' learning needs and personality traits, as well as promote their autonomy. Garrison (1997) also developed a comprehensive model of SDL and defined SDL as “an approach that helps stimulate students' assumption of personal responsibility and collaborative control over the cognitive (self-monitoring) and contextual (self-management) processes in constructing and confirming meaningful and worthwhile learning outcomes” (p. 21). Online learning has been steadily growing with the rapid development of ICTs. Therefore, it has become imperative that distance learning students learn to be proactive, act as independent learners, and prepare themselves for the e-learning experience. Lin and Hsieh (2001) argued that successful online students made decisions on their own to meet their needs by utilizing their existing knowledge and learning goals. It helps self-directed students take responsibility for their learning and be more enthusiastic about their learning activities. Learner control This dimension is also very important to understand one's readiness towards online learning. In contrast to the traditional mode of learning, where students have direct access to textbooks and other physical forms of information, they have more options, flexibility, and freedom in the e-learning environment. The students can control the content, sequence, and pace of learning (Reeves, 1993). Broadly speaking, learner control is “the degree to which a learner can direct his or her own learning experience and process” (Shyu & Brown, 1992, p. 3). This concept of learner control has evolved with the rapid development of ICTs. According to the Component Display Theory of Merrill (1983) and the Elaboration Theory of Reigeluth and Stein (1983), it is an essential element for effective learning that may boost the students' performance. Further, Merrill (1983) described that students should be given complete control of the sequence of the instructional material so that they could discover how to make instructional decisions and experience the results of those decisions. However, in an e-learning environment, there seems to be no instructional sequence (M.L. Hung et al., 2010). L.-C.C. Wang and Beasley (2002) found that learner control had an impact on the task performance of students in a web-based learning environment. Therefore, those students who were empowered by their own learning decisions exhibited a better performance during the online learning setting, as compared with those who were not. Learning motivation In any educational setting, motivation is a factor that significantly affects a student's attitude and behavior towards learning (Fairchild, Jeanne Horst, Finney, & Barron, 2005). Active learning is a mixture of two invisible variables: cognition and motivation (Pintrich & Schunk, 2002). Motivationally oriented (intrinsic and extrinsic) students tend to perform better academically than those who lack the motivation to learn (Ryan & Deci, 2000). Intrinsic motivation helps a student develop cognitively, physically, and socially. It is associated with a lower dropout rate, higher-quality learning, and better learning strategies (Deci & Ryan, 1985). Extrinsic motivation, on the other hand, is related to the attainment of rewards i.e., obtaining high academic grades, awards and prizes. In Garrison's model (1997), it has been proposed that learning has two forms of motivational aspects: 1) the perceived value of learning and 2) the anticipated success in learning. Motivation is intertwined with doing something willingly without any external pressure and has been taken reciprocally with responsibility by most researchers. However, to sustain their motivation, students must become active learners with a strong desire to learn (Candy, 1991). Ryan and Deci (2000) reported that students felt free to determine their learning paths in an online learning environment as a result of their motivation. Online communication self-efficacy In an online learning environment, students require various computer-mediated tools to perform their educational activities (Palloff & Pratt, 1999). It is empirically evident that shy and hesitant students perform better in an online learning setting than in a traditional learning environment. Therefore, it is imperative that students have opportunities to interact and communicate with other students and their instructors during web-based learning (M. McVay, 2000). The successful students communicate with each other using computer-mediated tools and raise questions in an online discussion to understand their subject or concepts in depth. In case of connectivity issues or burn-out situations, the students should take advantage of the opportunity to work with other students online. Past studies have also concluded that online communication self-efficacy was necessary for students to prevent limitation of online communication as well as isolation in online learning (M.L. Hung et al., 2010). Students' readiness towards online learning during the COVID-19 pandemic Since the start of the COVID-19 pandemic, academic institutes in most countries have transferred their learning and teaching activities from a physical model to an online one. Ever since then, researchers have been trying to determine the factors that could affect the readiness of students towards online learning. Consequently, there has been an abundance of literature published on this topic in recent days. Naji et al. (2020) recently conducted a study on engineering students to determine the factors that affected their readiness towards online learning during the COVID-19 pandemic. They found that four factors had an impact on their level of readiness: 1) initial preparedness and motivation for online learning, 2) self-efficacy beliefs about online learning, 3) self-directed online learning, and 4) support for online learning. Callo and Yazon (2020) reported that familiarity and capability regarding online learning, preparation of the online learning experience, device and connectivity, self-efficacy, and prior experience with technology significantly influenced the preparedness of Polytechnique students for online learning modality in the context of COVID-19. Furthermore, they stated that the readiness of students and teachers towards online learning could be determined through their capability to access and use technology as well as their e-learning self-efficacy. Shawaqfeh et al. (2020) investigated the online distance learning experience of pharmacy students in the Kingdom of Saudi Arabia during the outbreak of COVID-19 and found that the pharmacy students had a receptive attitude towards gaining an education in an online learning environment during the quarantine period of COVID-19. However, they also identified some challenges for the students such as a lack of motivation, boredom during class, information overload, and lack of digital skills, etc. They emphasized the need for computer training for pharmacy students so that they could learn the skills and tools needed to be effective and successful learners during these unprecedented times. Similarly, Kalkan (2020) examined the e-learning readiness of university students in Turkey using the e-learning readiness scale of Yurdugül and Demir (2017). He found that computer, internet, and online communication self-efficacy were the top-ranked factors that significantly affected the e-learning readiness of students, followed by self-learning, learning control and motivation. Allam, Hassan, Sultan, Mohideen, and Kamal (2020) surveyed students of communication and media studies to explore their readiness towards online learning during the outbreak of COVID-19. They revealed that while the study participants had computer/internet literacy, they lacked the motivation to learn online and engage in self-directed learning. Neupane, Sharma, and Joshi (2020) investigated the OLR of medical students during the COVID-19 pandemic and found that the medical students were ready for learning online during the lockdown situation and had sufficient technological facilities and skills to utilize these computer-mediated tools in their learning process. Kalman, Esparza, and Weston (2020) collected data from the students who were enrolled in a chemistry course. They concluded that adaptability, organizational skills, and self-awareness were some of the personal characteristics that enabled the students to succeed and excel as online students. Similarly, Lee (2020) explored the OLR of Malaysian students during the pandemic. They reported that female students and students enrolled in a degree program were more comfortable with online learning than the male students and those studying for a diploma. Furthermore, the students shared that if given the choice, they would prefer onsite classes over online ones. They concluded that, overall, students were ready for online learning during a pandemic. Research design and procedures A quantitative approach based on a survey method was used for the purpose of this research study. LIS students currently enrolled at HEC recognized public sector universities of Pakistan were selected purposively as a unit of analysis. A sample of 385 was drawn from the intended population using the following formula given by Wrenn, Stevens, and Loudon (2002):n=Z2p.qe2 n=1.962.5×.5.052=384.16 where:n = Sample size, Z = Value from normal distribution table for desired confidence level (i.e. corresponding to the chosen alpha level – for 0.05 is 1.96) p = Obtained population proportion (i.e. 50%) and q = l-p e = Error of sampling or desired precision = ±0.05 There are 09 library schools in the various public sector universities of Pakistan. A sample size of 43 per library school was calculated using the equal size sampling technique as shown in Table 1 .Table 1 Population and responses. Table 1SN University name Equal size sample Responses received (%) 1 University of the Punjab, Lahore 43 43 (100) 2 University of Sargodha, Sargodha 43 43 (100) 3 Islamia University, Bahawalpur 43 41 (95.0) 4 University of Okara, Okara 43 7 (16.3) 5 University of Peshawar, Peshawar 43 42 (97.7) 6 Khushal Khan Khatak University, Karak 43 43 (100) 7 University of Karachi, Karachi 43 41 (95.0) 8 University of Baluchistan, Quetta 43 41 (95.0) 9 Allam Iqbal Open University, Islamabad 43 39 (90.7) Total 387 340 (87.9) The scale for data collection was adopted from a study by M.L. Hung et al. (2010) and was slightly modified per the pandemic situation (Annexed). This data collection instrument had 18 items covering five dimensions related to the online learning readiness of students: 1) computer/internet self-efficacy (3 items), 2) self-directed learning (5 items), 3) learner control (3 items), 4) motivation for learning (4 items), and 5) online communication self-efficacy (3 items). Demographic information such as gender, age, level of LIS program i.e., BS, MA, M. Phil., or Ph. D. and Grade Point Average (GPA) of the respondents in the previous semester were added to the questionnaire. The final questionnaire was then designed in Google forms and made available via an online link. This was to ensure maximum reach of the survey, keeping in view the lockdown situation in the country. The link for the online survey was sent to the heads of the concerned departments, and coordinators of students' affairs (CSAs) to be disseminated among concerned students for data collection. Participants were also approached through personal contacts of the researchers as well as the friends-of-friends method. A total of 340 (87.9%) responses were received after an extensive follow-up. The collected data were imported to Statistical Package for Social Sciences (SPSS) version 22 for the necessary statistical analyses. The validity, reliability, and correlation matrix of the measures The data collection instrument's construct validity was assessed using the convergent and discriminant forms of validity. Composite reliability (CR) and Average Variance Extracted (AVE) were calculated in this regard. The value of CR was calculated and the resultant value of all the dimensions was found to be more than 0.70. Whereas the resultant values of AVE ranged from 0.484 to 0.587. The values of CR and AVE were above the threshold value of CR = 0.60 and 0.50 = AVE suggested by Byrne (2016). However, the value of AVE for the two constructs (learner control and learning motivation) was slightly lower than these cut-off values. It was also noted that the values of AVE were less than the values of CR for each dimension (Table 2 ).Table 2 Correlation, validity and reliability of measures. Table 2SN OLR Dimensions CR AVE α value CIS SDL LC MLF OCS 1 Computer/Internet self-efficacy 0.810 0.587 0.839 1 2 Self-directed learning 0.758 0.513 0.832 0.590⁎⁎ 1 3 Learner control 0.501 0.496 0.854 0.483⁎⁎ 0.594⁎⁎ 1 4 Motivation for learning 0.772 0.484 0.834 0.575⁎⁎ 0.607⁎⁎ 0.460⁎⁎ 1 5 Online communication self-efficacy 0.751 0.504 0.824 0.617⁎⁎ 0.587⁎⁎ 0.536⁎⁎ 0.664⁎⁎ 1 Cohen's criterion: r = 0.10 (small effect); r = 0.30 (medium effect), and r = 0.50 (large effect). ⁎⁎ Correlation is significant at the 0.01 level (2-tailed). The Cronbach's alpha value was calculated to check the internal consistency and reliability of the eighteen scale items. The alpha value was found to be 0.90, which indicated good consistency between the various items of the scale. Furthermore, this α value was above the recommended value of ≥0.70 (Hair, Babin, Anderson, & Black, 2018). Dimension-wise alpha value was also calculated, as shown in Table 2. Pearson's Moment correlation was applied to determine the relationship between the five dimensions of the OLR scale. The results revealed that all of the dimensions were positively and significantly correlated with each other at a p-level of 0.01. Further, Cohen's (1988) criterion was used to assess the strength of the association between these dimensions. According to this criterion, computer/internet self-efficacy was strongly correlated with self-directed learning (r = 0.590⁎⁎), motivation for learning (r = 0.575⁎⁎), and online communication self-efficacy (r = 0.617⁎⁎), while its relationship was medium with learner control (r = 0.483⁎⁎), and LC was moderately correlated with learning motivation (r = 0.460⁎⁎) (Table 2). Model testing results As this study used a pre-validated instrument for data collection, therefore Confirmatory Factor Analysis (CFA) was run to validate it on LIS students in the Pakistani context. To confirm the hypothetical model of the study, AMOS (Analysis of Moment Structure) version 21 was used. For this purpose, the values of x 2 /df (chi-square/degree of freedom), GFI (Goodness of Fit Index), AGFI (Adjusted Goodness of Fit Index), CFI (Comparative Fit Index), and RMSEA (Root Mean Square Error of Approximation) were calculated. The results of the above-mentioned model fit indices showed that the values were in acceptable ranges, as recommended by Brown (2015) (Table 3 ). It was concluded that the study model was a good fit in the Pakistani context; however, two items (SDL2 and LC2) had poor factor loading (< 0.50) (Fig. 1 ). Overall, the results of factor loading showed that all of the items were statistically significant and each item in the scale was successfully loaded (≥ 0.50) under the latent dimension (Fornell & Larcker, 1981) (Table 5).Table 3 Model fit indices. Table 3 x2/df GFI AGFI CFI RMSEA Cut-off values ≤ 3 ≥ 0.90 ≥ 0.90 ≥ 0.90 ≤ 0.08 Model Fit Indices 162.464 / 93 = 1.747 0.947 0.923 0.968 0.047 Fig. 1 Measurement model and factor loading. Fig. 1 Results Demographic profile of the respondents The respondents were asked about their gender, age, level of program of study, and GPA in the previous semester to collect their demographic information. The results showed that the majority of the study participants (n = 194, 57.1%) were female, while 146 (42.9%) were male. The majority of participants (n = 309, 90.9%) were also young adults with aged less than 30 years, 24 (7.1%) were between 31 and 40 years of age, while only 7 (2.1%) were older than 40 years. Almost 42% of participants were enrolled in BS-LIS/BS-IM degree programs and the same percentage were enrolled in the MLIS/MA-IM programs. 12.4% (n = 42)) of the participants were working towards an M. Phil. and only 3.8% (n = 13)) were pursuing a Ph.D. in LIS/IM. They were also asked about their grades in the previous semester and it was found that the majority of the respondents (n = 241, 70.9%) had secured more than a 3.00 GPA, 98 (28.8%) had a GPA between 2.01 and 3.00, while only one participant had a GPA less than 2.00 in the previous semester (Table 4 ).Table 4 Demographic composition of respondents. Table 4SN Demographic variables Frequency Percentage (%) 1 Gender Male 146 42.9 Female 194 57.1 2 Age brackets (years) Up to 20 49 14.4 21–30 260 76.5 31–40 24 7.1 41 and above 7 2.1 3 Level of LIS Program BS-LIS / BS-IM 142 41.8 MLIS / MA-IM 143 42.1 M. Phil. 42 12.4 Ph.D. 13 3.8 4 GPA (previous semester) ≤ 2.00 1 0.3 2.01–3.00 98 28.8 3.01–4.00 241 70.9 RQ1 – LIS students' perceived Online Learning Readiness during COVID-19 The respondents were given a set of 18 items to determine their perception of OLR during COVID-19. The responses of the participants for each item with their mean (M) and standard deviation (SD) are presented in Table 5 . The results showed that the top ranked dimension of OLR of LIS students during COVID-19 was a motivation for learning with a mean score of 3.88 (0.717). Further, the results unveiled that LIS students were motivated to learn, open to new ideas, and liked to share their ideas with other class fellows during COVID-19 online learning. The second ranked dimension was computer/internet self-efficacy (M = 3.66, SD = 0.906) followed by self-directed learning (M = 3.62, SD = 0.672), and online communication self-efficacy (M = 3.53, SD = 1.011). However, learner control remained the lowest ranked dimension of OLR for LIS students with a mean score of 3.37 (0.761). The results showed that the majority of the participants agreed on having sufficient computer and internet skills. They felt confident in performing the basic functions using the Microsoft (MS) Office suite i.e., MS Word, MS Excel, and MS PowerPoint (M = 3.70, SD = 1.044). They were proficient in managing different software tools used in online learning (M = 3.44, SD = 1.083) and in retrieving the relevant information (M = 3.83, SD = 1.070) (Table 5).Table 5 Perceived Online Learning Readiness during COVID-19 (N = 340). Table 5ID Statements M SD Factor Loading Computer/Internet self-efficacy (CIS) 3.66 0.906 CIS1 I feel confident in performing the basic functions of Microsoft Office programs (MS Word, MS Excel, and MS PowerPoint). 3.70 1.044 0.764 CIS2 I feel confident in my knowledge and skills of how to manage software for online learning. 3.44 1.083 0.726 CIS3 I feel confident in using the Internet (Google, Yahoo) to find or gather information for online learning. 3.83 1.070 0.805 Self-directed learning (SDL) 3.62 0.672 SDL1 I carry out my own study plan. 3.66 0.929 0.609 SDL2 I seek assistance when facing learning problems. 3.54 0.969 0.405 SDL3 I manage time well. 3.43 1.055 0.626 SDL4 I set up my learning goals. 3.74 0.912 0.733 SDL5 I have higher expectations for my learning performance. 3.76 0.973 0.678 Learner control (LC) 3.37 0.761 LC1 I can direct my own learning progress 3.54 0.903 0.683 LC2 I am not distracted by other online activities when learning online (instant messages, Internet surfing). 3.03 1.186 0.398 LC3 I repeated the online instructional materials on the basis of my needs. 3.54 0.978 0.540 Motivation for learning (MFL) 3.88 0.717 MFL1 I am open to new ideas. 3.77 0.922 0.754 MFL2 I have the motivation to learn. 3.84 0.956 0.748 MFL3 I improve from my mistakes. 4.01 0.910 0.676 MFL4 I like to share my ideas with others. 3.90 0.949 0.520 Online communication self-efficacy (OCS) 3.47 0.886 OCS1 I feel confident in using online tools (email, discussion) to effectively communicate with others. 3.53 1.127 0.771 OCS2 I feel confident in expressing myself (emotions and humor) through text. 3.53 1.011 0.630 OCS3 I feel confident in posting questions in online discussions. 3.37 1.117 0.721 Scale: 1 = Strongly Disagree, 2 = Disagree, 3 = Neutral, 4 = Agree, 5 = Strongly Agree. Bold data signifies Poor factor loading (< 0.50). RQ2 – gender and perceived LIS students' OLR An independent sample t-test was applied to see the difference of opinion between students based on their gender. The results revealed that there was a significant difference of opinion between male and female students on two out of the five dimensions under investigation, i.e., computer/Internet self-efficacy (CIS) (0.007 < 0.05) and online communication self-efficacy (OCS) (0.048 < 0.05) at p-level of 0.05. The opinion of LIS students on both of these dimensions were higher in the male students (for CIS: M = 3.81, SD = 0.941; for OCS; M = 3.58, SD = 3.39) than in the female students (Table 6 ).Table 6 Summary table of results. Table 6Variables Statistics applied Computer/Internet self-efficacy Self-directed learning Learner control Motivation for learning Online communication self-efficacy Test statistics p-value Test statistics p-value Test statistics p-value Test statistics p-value Test statistics p-value Gender Independent sample t-test 0.007⁎⁎ 0.731 0.438 0.679 0.194 0.973 0.939 0.754 0.048⁎ 0.120 Age Simple linear regression 0.35⁎⁎ (R2 = 0.039) 0.000 0.009 (R2 = 0.005) 0.181 0.004 (R2 = 0.001) 0.619 0.012 (R2 = 0.008) 0.108 0.023⁎ (R2 = 0.018) 0.013 Level of LIS Program One-way ANOVA 0.001⁎⁎ 0.01 0.275 0.05 0.262 0.05 0.045⁎ 0.05 0.047⁎ 0.05 Previous semster GPA Simple linear regression 0.039 (R2 = 0.000) 0.723 0.165⁎ (R2 = 0.012) 0.044 −0.010 (R2 = 0.000) 0.918 0.229⁎⁎ (R2 = 0.020) 0.009 0.110 (R2 = 0.003) 0.310 ⁎ Significant at the 0.05 level (2-tailed). ⁎⁎ Significant at the 0.01 level (2-tailed). RQ3 & RQ5 – impact of LIS students' age and grade on their OLR A simple linear regression analysis was run to assess the impact of age and grades (previous semester) of LIS students on their OLR during COVID-19 (Table 6). The results observed a positive significant impact of age on two dimensions of the OLR i.e., computer/internet self-efficacy (β = 0.35**, p = .000) and online communication self-efficacy (β = 0.023*, p = .013 ≤ 0.05). These findings depict that the older students performed better in using computers and the internet as compared to the younger ones. Further, the results showed that the older students were more confident in using online tools (email, discussion) to effectively communicate with others, in expressing their emotions and humor through text, and were more capable of posting questions during online discussions (Table 6). The students' grades also exhibited a significant positive influence on self-directed learning (β = 0.165⁎, p = .044 ≤ 0.05) and learning motivation (β = 0.142⁎⁎, p = .009 ≤ 0.001). These findings highlight the positive impact of a high GPA. The students with higher grades were more likely to be self-learners and more motivated to explore new ideas during online learning in the time of pandemic than those students with lower grades (Table 6). RQ4 – impact of various degree programs on their perceived OLR To measure the impact of various levels of LIS degree programs (BS, MS, M. Phil., pH. D.) on their perceived OLR during COVID-19 pandemic, a one-way ANOVA test was applied. A significant difference of opinion was observed on computer/internet self-efficacy (F = 5.268, Sig. = 0.001** < 0.000), motivation for learning (F = 2.710, Sig. = 0.045* < 0.05), and online communication self-efficacy (F = 2.682, Sig. = 0.047* < 0.05) (Table 6). However, difference of opinion among other dimensions remained insignificant. To further explore the significant difference between the groups (BS, MA, M. Phil., and Ph.D.), a post-hoc Tukey test was run. This analysis revealed that there was a significant difference of opinion between the students enrolled in BS and M. Phil. (Sig. = 0.038* < 0.05) and Ph.D. students (Sig. 0.018* < 0.05) on computer/internet self-efficacy. The opinion of Ph.D. students was dominant over BS and M. Phil. students. Discussion The overall findings of the study revealed that Pakistani LIS students were sufficiently prepared for online learning during the COVID 19 lockdown in the country. They were motivated to learn online, were receptive to new ideas, learned from their mistakes, and were willing to interact and engage with their fellow students while learning online. These findings are similar to those of M.L. Hung et al. (2010), Saadé, He, and Kira (2007), and Hsu, Wang, and Levesque-Bristol (2019) who have reported that motivation played a vital role in online learning. Furthermore, the findings revealed that LIS students possessed a relatively good level of self-efficacy while performing basic functions on the computer using MS Office suite (MS Word, MS Excel, MS PowerPoint, etc.), managing and using different kinds of software necessary for online learning, and using the internet. Alqurashi (2016) and C.L. Tsai, Cho, Marra, and Shen (2020) also concluded that computer self-efficacy was essential for online learning and was significantly correlated with the success of the online learning of students. LIS students further reported that they were proficient users of online tools such as email and chat to communicate with others effectively and were confident in expressing their emotions and humor through text. In an online learning environment, the interaction of students and teachers mostly occurs through computer-mediated tools often called asynchronous tools. The study depicted that LIS students' self-efficacy related to online communication was relatively good. This finding is in line with the findings of Yasin and Ong (2020) who concluded that online communication self-efficacy in a blended learning environment could promote the OLR of students. However, it in contradiction with those of Estira (2020) and Cigdam and Yildirim (2014), who found that online communication self-efficacy of students was comparatively less important. The results further indicated that the students felt that they had less control over their learning environment and time management. These findings are similar to those of M.L. Hung et al. (2010) and Naji et al. (2020) who have reported that learner control was a lower rated dimension of OLR among students as compared to other dimensions. The reason might be that online learning is different from traditional face-to-face learning where there is a high possibility of disruption e.g. students engaging in disruptive activities such as playing online games, internet surfing, chatting or instant messaging with friends, etc. Time management is an important aspect of online learning, and students should devote adequate time to their respective courses, participate in group discussions by posting messages, and submit their work on time (Roper, 2007; P. Wang, Wu, Wang, Chi, & Wang, 2018). The lack of control over their learning environment might result in poor performance of students, therefore, H.T. Hung and Yuen (2010), and L.-C.C. Wang and Beasley (2002) have suggested that students who had appropriate control over their learning setting might exhibit better academic performance in a web-based learning setting as compared with those who did not have any control. As things stand, no one is certain about the reopening of educational institutions. The world is still struggling with the second wave of COVID 19 (Sultan, 2020). In these challenging times, almost all nations have transitioned to online learning. This has resulted in additional responsibility for the students to take ownership of their education, to properly manage their time, and control their learning environment so that the educational systems and processes can continue to run smoothly. Control over their learning environment would allow the students to make decisions about their learning, individualize the selection of media, manage time, and control their educational content. In relationship analyses, the study observed a significant difference of opinion on two dimensions of OLR (computer/internet self-efficacy and online communication self-efficacy), based on the LIS students' gender. On these dimensions, the perception of male students was higher than the female students. These findings are similar to those of Sakal (2017), who found a significant difference in online communication in men. Kay (2009) also depicted that the perception of male students was stronger than female students on the interactive classroom communication systems. This greatly contributed to the learning process of the male students. However, these findings are not compatible with the findings of Chung et al. (2020) and M.L. Hung et al. (2010) who did not find any significant difference in attitude and behavior of male and female students on all five dimensions of the OLR scale. One possible explanation of this inconsistency might be the socio-cultural and socio-economic differences between the participants of these studies. Furthermore, the present study findings have unveiled a significant difference in students' readiness towards their computer, internet, and online communication self-efficacy and learning motivation depending on the level of their program of study. The perception of M. Phil. and Ph.D. students was higher than BS and MLIS/MIM students. This was also an anticipated finding, as it is generally assumed that the self-efficacy of students tends to improve as they progress in their studies. Additionally, these findings might be the result of policies set forth by the HEC of Pakistan making ICTs related courses compulsory for students in higher degree programs. Another reason could be that most LIS students with higher level degrees (M. Phil. and Ph.D.) were working also working as professional librarians and thus had more advanced IT skills. Likewise, research students were more engaged in different kinds of information seeking and research activities and, therefore, had a better computer, internet, and online communication self-efficacy as compared to other students, as asserted by Naveed and Mahmood (2019). Furthermore, age appeared to be a strong predictor of two of the dimensions of online readiness (computer and internet self-efficacy, and online communication self-efficacy). These findings are not surprising. Generally, a student's learning experience, confidence, and competencies also improve with increasing age. Chung et al. (2020) and Lee, Yeung, and Ip (2016) have also reported that matured students tended to exhibit a greater readiness for online learning than the younger students. Moreover, grades of LIS students are also reported as being a significant predictor of two dimensions of OLR (self-directed learning and motivation for learning). It seems that the students who scored higher grades in their previous semester exhibited a higher readiness towards self-directed learning and learning motivation compared to those who had lower grades. These findings show that students possessing higher grades were more confident in executing their study plans, seeking timely assistance, managing their time, setting learning goals, and had higher expectations for their learning performance. This finding highlights that high academic achievers were more innovative, motivated, and more likely to share their ideas with their fellows and teachers. These findings are congruent with the findings of M.L. Hung et al. (2010) who concluded that grades were a strong predictor of self-directed learning and learning motivation. Theoretical and practical implications This study would be a good addition in the area of online education particularly, during a pandemic situation. This is a baseline study in the context of Pakistani LIS students that would open new horizons of exploration for future researchers. This investigation has explored the readiness of students regarding online learning during the COVID-19 pandemic through the OLR scale developed by M.L. Hung et al. (2010). The results have shown that the OLR scale was not fully applicable to the LIS students in an emergency as it had some convergent validity issues on two of its dimensions (learner control and learning motivation). Therefore, there was a need to develop a new scale or modify this one for investigating the LIS students in a pandemic like situation for all future research studies. Furthermore, it is suggested that some additional personal (urban/rural, race, marital status, etc.) and academic factors should be included in the scale while determining the OLR of LIS students. This study may be replicated on students of other disciplines to compare their preparedness for online learning during COVID-19 for holistic findings. This study has some practical implications for heads of LIS schools, university administrators, and policymakers: 1) The study participants have indicated a lack of self-efficacy in posting questions during an online discussion. This had some serious implications for their poor academic performance. Therefore, LIS departments must arrange training and orientation sessions for their students to improve their online communication self-efficacy. This would enable them to fully participate in the online learning experience gain the maximum benefits from it. L.-C.C. Wang and Beasley (2002) have also claimed that such interventions would result in better academic performance among students. 2) Since the LIS students reported a lack of confidence in managing the various software tools used for online learning, therefore, the university IT department must conduct training/orientation sessions to teach about these tools. 3) As the students claimed a lack of control over their learning and had time management issues, so the course instructor should try harder to engage every student in task-based online group discussions. This would encourage student engagement and discourage their involvement in other disruptive activities like chatting, texting, online gaming, etc. during an online class. 4) Lastly, the university administration should play a key role in this regard by establishing a strong system of oversight to monitor student activities during online classes. The BS and MA students' perception about their computer/internet self-efficacy and online communication self-efficacy was weaker than that of M. Phil. and Ph.D. students. This result may be used by policymakers to develop and offer short ICTs related courses for BS and MA students. Such courses would help improve their ICTs proficiency and allow them to cope with the challenges presented by online learning during the prevailing COVID-19 pandemic. Limitations and future research directions This research has certain limitations. Firstly, the study explored the OLR of LIS students enrolled in the nine public sector universities in Pakistan, therefore, its results may not be generalized to the students of other disciplines. Secondly, the sample was drawn via an equal size purposive sampling technique (a form of nonprobability sampling) that could create the issue of generalizability. Thirdly, the study adopted the self-assessment method to collect data about the students' perceived online learning readiness. This raises the issue of bias. Lastly, it is empirically evident that the individuals tend to overestimate their self-efficacy (Botes, 2016; Schlösser, Dunning, Johnson, & Kruger, 2013), therefore, the theory of the Dunning-Kruger effect should be taken into consideration while interpreting this study's results. The present study suggests some topics for future research, for instance, a survey of the viewpoint of LIS teachers who are currently involved in online education should be conducted. A mixed-methods study could be carried out once this pandemic is over, i.e., a post-pandemic study. Moreover, a qualitative study exploring the opinions of students and teachers would be a worthy endeavor. Conclusion During the COVID-19 pandemic, LIS students were not much personalized and successful in their decisions about their online educational life; however, they were motivated to learn in this e-learning environment. Female LIS students' computer/internet and online communication self-efficacy was lower than their male counterparts. Postgraduate students (M. Phil. and Ph.D.) exhibited a higher readiness towards computer/internet, online communication self-efficacy, and learning motivation than undergraduate (BS) and graduate students (MA). Furthermore, the age and grades (GPA) of LIS students appeared to be strong predictors of OLR dimensions during emergencies such as COVID-19. The following is the supplementary data related to this article.Annexure 1 Readiness for online learning during COVID-19 pandemic: A survey of Pakistani LIS students. Annexure 1 CRediT authorship contribution statement All persons who meet authorship criteria are listed as authors, and all authors certify that they have participated sufficiently in the work to take public responsibility for the content, including participation in the concept, design, analysis, writing, or revision of the manuscript. Furthermore, each author certifies that this material or similar material has not been and will not be submitted to or published in any other publication before its appearance in the Journal of Academic Librarianship. ==== Refs References Afrianti N. Aditia R. Online learning readiness in facing the COVID-19 pandemic at MTS Manunggal Sagara Ilmi, Deli Serdang, Indonesia Journal of International Conference Proceedings 3 2 2020 59 66 Allam S.N.S. Hassan M.S. Sultan R. Mohideen A.F.R. Kamal R.M. Online distance learning readiness during covid-19 outbreak among undergraduate students Journal of Academic Research in Business and Social Sciences 10 5 2020 642 657 Alqurashi E. Self-efficacy in online learning environments: A literature review Contemporary Issues in Education Research 9 1 2016 45 52 Bandura A. Self-efficacy: Toward a unifying theory of behavioral change Psychological Review 84 1977 191 215 847061 Bandura A. Social foundations of thought and action: A social cognitive theory 1986 Prentice-Hall Englewood Cliffs, NJ Bandura A. Self-efficacy: The exercise of control 1997 W.H. Freeman New York Botes, J. (2016). The Dunning-Kruger effect and poor performers, available at https://www.hrfuture.net/news-1/the-dunning-kruger-effect-and-poor-performers-21256/. Brown T. Confirmatory factor analysis for applied research 2015 The Guilford Press London Byrne B.M. Structural equation modeling with AMOS: Basic concepts, applications And programming (multivariate applications series) 3rd ed. 2016 Routledge London Callo E.C. Yazon A.D. Exploring the factors influencing the readiness of faculty and students on online teaching and learning as an alternative delivery mode for the new normal Universal Journal of Educational Research 8 8 2020 3509 3518 Candy P.C. Self-direction for lifelong learning: A comprehensive guide to theory and practice 1991 San Francisco Jossey-Bass Chizmar J.F. Walbert M.S. Web-based learning environments guided by principles of good teaching practice Journal of Economic Education 30 3 1999 248 264 Chung E. Subramaniam G. Dass L.C. Online learning readiness among university students in Malaysia amidst COVID-19 Asian Journal of University Education 16 2 2020 46 58 Cigdam H. Yildirim O.G. Effects of students’ characteristics on online learning readiness: A vocational college example Turkish Online Journal of Distance Education 15 3 2014 80 93 Cohen J. Statistical power analysis for the behavioral sciences 2nd ed. 1988 Erlbaum Hillsdale, NJ Compeau D.R. Higgins C.A. Computer self-efficacy: Development of a measure and initial test MIS Quarterly 19 2 1995 189 211 Deci E. Ryan R. Intrinsic motivation and self-determination in human behavior 1985 Plenum Press New York Eastin M.S. LaRose R. Internet self-efficacy and the psychology of the digital divide Journal of Computer-Mediated Communication 6 1 2000 Retrieved from https://academic.oup.com/jcmc/article/6/1/JCMC611/4584219 Estira K.L.A. Online distance learning readiness of business administration students in one state university in the Philippines Journal of Critical Reviews 7 12 2020 826 832 Fairchild A.J. Jeanne Horst S. Finney S.J. Barron K.E. Evaluating existing and new validity evidence for the academic motivation scale Contemporary Educational Psychology 30 3 2005 331 358 Firat M. Bozkurt A. Variables affecting online learning readiness in an open and distance learning university Educational Media International 57 2 2020 112 127 Fornell C. Larcker D.F. Evaluating structural equation models with unobservable variables and measurement error Journal of Marketing Research 18 1 1981 39 50 Forson I.K. Vuopala E. Online learning readiness: Perspective of students enrolled in distance education in Ghana The Online Journal of Distance Education and e-Learning 7 4 2019 277 294 Garrison D.R. Self-directed learning: Toward a comprehensive model Adult Education Quarterly 48 1 1997 18 33 Guglielmino L.M. Development of the self-directed learning readiness scale Unpublished doctoral dissertation 1977 The University of Georgia Athens, GA Hair J.F. Babin B.J. Anderson R.E. Black W.C. Multivariate data analysis 8th ed. 2018 Cengage Boston Hergüner G. Son S.B. Son S.H. Dönmez A. The effect of online learning attitudes of university students on their online learning readiness The Turkish Online Journal of Educational Technology 19 4 2020 102 110 Higher Education Department Government of the Punjab, Notification [SO(Admn)/Misc./2020] 2020 Hsu H.C.K. Wang C.V. Levesque-Bristol C. Reexamining the impact of self-determination theory on learning outcomes in the online learning environment Education and Information Technologies 24 3 2019 2159 2174 Hung H.T. Yuen S.C.Y. Educational use of social networking technology in higher education Teaching in Higher Education 15 6 2010 703 714 Hung M.L. Chou C. Chen C.H. Own Z.Y. Learner readiness for online learning: Scale development and student perceptions Computers & Education 55 3 2010 1080 1090 Joosten T. Cusatis R. Online learning readiness American Journal of Distance Education 34 3 2020 1 14 Kalkan N. Investigation of e-learning readiness levels of university students studying in different departments African Educational Research Journal 8 3 2020 533 539 Kalman R. Esparza M.M. Weston C. Student views of the online learning process during the covid-19 pandemic: A comparison of upper-level and entry-level undergraduate perspectives Journal of Chemical Education 97 9 2020 3353 3357 Kay R.H. Examining gender differences in attitudes toward interactive classroom communications systems (ICCS) Computers & Education 52 4 2009 730 740 Knowles M.S. Self-directed learning: A guide for learners and teachers 1975 Association Press New York Lee S. Sabah student stays overnight in tree to get better Internet connection for online university exams 2020, 09 October Available at: https://www.thestar.com.my/news/nation/2020/06/16/sabah-uni-student-stays-overnight-in-tree-to-get-better-internet-connection-for-online-exams Lee C. Yeung A.S. Ip T. Use of computer technology for English language learning: Do learning styles, gender, and age matter? Computer Assisted Language Learning 29 5 2016 1035 1051 Lin B. Hsieh C.T. Web-based teaching and learner control: A research review Computers & Education 37 4 2001 377 386 McVay M. Developing a web-based distance student orientation to enhance student success in an online bachelor’s degree completion program Unpublished practicum report presented to the Ed.D. Program 2000 Nova Southeastern University Florida McVay M. How to be a successful distance learning student: Learning on the internet 2001 Prentice Hall New York Merrill, M. (1983). Component display theory. In C. Reigeluth (Ed.), Instructional-design theories and models: An overview of their status (pp. 279–334). Hillsdale, NJ: Lawrence Erlbaum Associates. Naji K.K. Du X. Tarlochan F. Ebead U. Hasan M.A. Al-Ali A.K. Engineering students’ readiness to transition to emergency online learning in response to COVID-19: Case of Qatar EURASIA Journal of Mathematics, Science and Technology Education 16 10 2020 1 17 Naveed M.A. Mahmood M. Information literacy self-efficacy of business students in Pakistan Libri 69 4 2019 303 314 Neupane H.C. Sharma K. Joshi A. Readiness for the online classes during covid-19 pandemic among students of Chitwan medical college Journal of Nepal Health Research Council 18 2 2020 316 319 32969401 Palloff R.M. Pratt K. Building learning communities in cyberspace: Effective strategies for the online classroom 1999 San Francisco Jossey-Bass Peng H. Tsai C.C. Wu Y.T. University students’ self-efficacy and their attitudes toward the internet: The role of students’ perceptions of the internet Educational Studies 32 1 2006 73 86 Pintrich P.R. Schunk D.H. Motivation in education: Theory, research, and applications 2nd ed. 2002 Merrill/Prentice Hall Upper Saddle River, NJ Poole D.M. Student participation in a discussion-oriented online course: A case study Journal of Research on Computing in Education 33 2 2000 162 177 Reeves T.C. Pseudoscience in computer-based instruction: The case of lecturer control research Journal of Computer-based Instruction 20 2 1993 39 46 Reigeluth, C. M., & Stein, F. S. (1983). The elaboration theory of instruction. In C. M. Reigeluth (Ed.), Instructional-design theories and models: An overview of their current status, Vol. 1. Hillsdale, NJ: Lawrence Erlbaum Associates. Roper A.R. How students develop online learning skills Educause Quarterly 30 1 2007 62 64 Ryan R.M. Deci E.L. Intrinsic and extrinsic motivations: Classic definitions and new directions Contemporary Educational Psychology 25 1 2000 54 67 10620381 Sakal M. Çevrimiçi Öğrenmede Öğrencilerin Hazirbulunuşluk Düzeylerinin Demografik Özelliklerine Göre Incelenmesi Sosyal ve Beşeri Bilimler Araştırmaları Dergisi 18 39 2017 81 102 Saadé R.G. He X. Kira D. Exploring dimensions to online learning Computers in Human Behavior 23 4 2007 1721 1739 Schlösser T. Dunning D. Johnson K.L. Kruger J. How unaware are the unskilled? Empirical tests of the “signal extraction” counter explanation for the Dunning–Kruger effect in self-evaluation of performance Journal of Economic Psychology 39 2013 85 100 Shawaqfeh M.S. Al Bekairy A.M. Al-Azayzih A. Alkatheri A.A. Qandil A.M. Obaidat A.A. …Muflih S.M. Pharmacy students perceptions of their distance online learning experience during the COVID-19 pandemic: A cross-sectional survey study Journal of Medical Education and Curricular Development 7 2020 1 9 Shyu H.Y. Brown S.W. Learner control versus program control in interactive videodisc instruction: What are the effects in procedural learning? International Journal of Instructional Media 19 2 1992 85 95 Smith P.J. Murphy K.L. Mahoney S.E. Towards identifying factors underlying readiness for online learning: An exploratory study Distance Education 24 1 2003 57 67 Stansfield M. McLellan E. Connolly T. Enhancing student performance in online learning and traditional face-to-face class delivery Journal of Information Technology Education: Research 3 1 2004 173 188 Sultan F. COVID-19 data from last 24h. Cases 908. Positivity percentage >3%. Sixteen deaths. These numbers tell an unmistakable story [Tweet]. Twitter https://twitter.com/fslsltn/status/1321816110233198592?ref_src=twsrc%5Etfw%7Ctwcamp%5Eembeddedtimeline%7Ctwterm%5Eprofile%3Afslsltn%7Ctwcon%5Etimelinechrome&ref_url=http%3A%2F%2Fcovid.gov.pk%2F 2020 October 29 Tsai C.C. Lin C.C. Taiwanese adolescents’ perceptions and attitudes regarding the internet: Exploring gender differences Adolescence 39 156 2004 725 734 15727410 Tsai C.L. Cho M.H. Marra R. Shen D. The Self-Efficacy Questionnaire for Online Learning (SeQoL) Distance Education 2020 1 18 Tsai M.J. Tsai C.C. Information searching strategies in web-based science learning: The role of internet self-efficacy Innovations in Education and Teaching International 40 1 2003 43 50 Wang L.-C.C. Beasley W. Effects of learner control and hypermedia preference on cyber-students’ performance in a web-based learning environment Journal of Educational Multimedia and Hypermedia 11 1 2002 71 91 Wang P. Wu P. Wang J. Chi H.L. Wang X. A critical review of the use of virtual reality in construction engineering education and training International Journal of Environmental Research and Public Health 15 6 2018 1204 29890627 Warner D. Christie G. Choy S. Readiness of VET clients for flexible delivery including on-line learning 1998 Australian National Training Authority Brisbane Wrenn B. Stevens R. Loudon D.L. Marketing research: Test and cases 2002 Routledge London Yasin N.M. Ong M.H. A blended learning model of technology access and technical self-efficacy: Multiple mediator effects on student readiness Journal of Advanced Research in Dynamical and Control Systems 2020 1 13 Yurdugül H. Demir Ö Öğretmen yetiştiren lisans programlarındaki öğretmen adaylarının e-öğrenmeye hazır bulunuşluklarının incelenmesi: Hacettepe üniversitesi örneği Hacettepe Üniversitesi Eğitim Fakültesi Dergisi 32 4 2017 896 915 Zgheib, G., AlDaia, R., Serhan, M., & Melki, A. (2019, November). Factors influencing students' online learning readiness in a middle eastern higher education institution: Implications for online course design. In E-learn: World conference on E-learning in corporate, government, healthcare, and higher education (pp. 1186–1198).

0

PMC9753147

NO-CC CODE

2022-12-16 23:25:31

no

2021 May 1; 47(3):102346

utf-8

nan

10.1016/j.acalib.2021.102346

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(22)00923-0 10.1016/S0140-6736(22)00923-0 Comment Political science and global health policy Gómez Eduardo J a Singh Prerna b Shiffman Jeremy c Barberia Lorena d a College of Health, Department of Community and Population Health, Lehigh University, Bethlehem, PA 18015, USA b Department of Political Science and School of Public Health, Brown University, Providence, RI, USA c Johns Hopkins University, Bloomberg School of Public Health and School of Advanced International Studies, Baltimore, MD, USA d University of São Paulo, Department of Political Science, São Paulo, SP, Brazil 17 5 2022 4-10 June 2022 17 5 2022 399 10341 20802082 © 2022 Elsevier Ltd. All rights reserved. 2022 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcThe drive to make policy more evidence-based has prompted scholars and practitioners to call for removing politics from global health policy making. This stance is neither possible nor is it desirable, because many issues, such as what constitutes a just allocation of health resources, can only be settled legitimately through democratic deliberation. As our new Lancet Series on Political Science and Health1, 2, 3 reveals, politics matters and should become an indispensable part of global health policy discussions. Historically, the concept of power has been central to the study of politics. Political scientists have offered various conceptualisations that are instructive for analysing global health policy making. Harold Lasswell4 provided an early and influential definition, framing politics in terms of the control of resources—”who gets what, when, and how”. Robert Dahl5 offered a compulsory understanding of the concept, arguing that “A has power over B to the extent that he can get B to do something that B would not otherwise do”. Michael Barnett and Raymond Duvall6 called for a more comprehensive understanding of power beyond compulsion, to include constitutive relations—the social processes that define the identity of actors and their relationships, with consequent effects on what these actors can do. Political science provides concepts to help structure analyses of the influence of power and politics on global health policy making. For example, the papers in this Series1, 2, 3 draw on the interconnected concepts of interests, ideas, and institutions.7, 8, 9 Interests refer to the motivations of politicians and civil societal actors as they pursue their agendas and how these affect health policy.10 Institutions pertain to the formal and informal rules and constraints created by individuals, from constitutions to traditions and customs, that shape political life and policy outcomes.11, 12 Ideas refer to beliefs that shape individual behaviour and policy.8, 13, 14 As an example, the second Series paper by Carmen Jacqueline Ho and colleagues2 reveals how achieving universal health coverage (UHC) is a challenging political process. Power can be used to advance ideas and interests, and forge institutions that favour certain groups over others and determine how committed a government is to ensuring health care for all. However, Ho and colleagues also emphasise that implementing UHC is shaped by bureaucratic capacity and the dynamic relationships between policy makers, local officials implementing policy, and non-state actors. This analytical approach helps to explain the challenges of governmental responses to pandemics. The interests of political leaders and senior health officials may diverge due to differences in political, ideological, and scientific beliefs, as seen with the response to COVID-19 in Brazil; when combined with political decentralisation, with state governors and mayors varying in their beliefs and policy response, the differences in institutions and political situations can spell disaster for efforts to control COVID-19.15 Insights from political science are also relevant for understanding the health response in other settings. Ideas of national solidarity have transformed the interests of political actors and made them more likely to prioritise health policy across India.16 Boundary institutions influence ideas of national solidarity, and helped shape the nature of HIV/AIDS policy in Brazil, India, and South Africa.17 And in Mexico, commercial sugar-sweetened beverage industries, which have historically had access to congressional and bureaucratic institutions, hampered the introduction of much needed policies to tackle obesity and non-communicable diseases.18 Other political science concepts offer additional insights into global health policy making. For example, the first Series paper1 draws on policy framing research and provides evidence that the way global health issues are framed—as threats, ethical imperatives, and wise investments—can shape the amount of attention and resources these issues receive from global health organisations and national governments. The concepts of path dependency19 and policy feedbacks processes—ie, how health policies generate supportive coalitions which reinforces existing policies over time—underscore why nations vary in their adoption of health-care legislation.20 This conceptual approach can help to explain why some governments fail to implement new approaches to global health threats, since the political and bureaucratic coalitions that created policies in response to public health challenges in the past—eg, conservative beliefs in the government's limited role in health—strive to maintain these interests at all costs. As seen with the US Government's initial response to COVID-19, the literature on this conceptual approach may provide insight into how path dependent feedback process can obstruct the creation of a comprehensive central government role (overcoming the challenges of decentralisation) in testing, contact tracing, and physical distancing.21 This conceptual approach can also help to explain why some countries were better positioned to respond to COVID-19. Indeed, centralised responses in countries with a history of responding to severe acute respiratory syndrome (SARS), such as those in South Korea and Singapore, are said to have been instrumental to responses in the initial months of the COVID-19 pandemic.22, 23 Political science provides ideas and approaches to research that can enhance our understanding of global health policy and politics. Rather than divorcing politics from policy decision making, political science research emphasises that recognising political power dynamics is crucial in helping to identify why certain public health policies might be more likely to succeed in adoption and implementation. Political scientists also appreciate that political power shapes, and is shaped by, the rise of new policy ideas, institutions, and interests. The papers in this Lancet Series illustrate important contributions from political science, raise new research questions, provide policy-making recommendations, and identify future areas of research. © 2022 franckreporter/Getty Images 2022 EJG led the Lancet Series on Political Science and Health discussed here. This Series was made possible with funding support from the Rockefeller Foundation. EJG received funding from the Rockefeller Foundation for partial (3%) salary costs at his previous institution, King's College London, London, UK. JS received funding from the Rockefeller Foundation for travel expenses to attend a Series meeting. The authors declare no other competing interests. We wish to acknowledge the political scientists and policy practitioners who provided invaluable comments during the meetings organised for this Series: Ken Shadlen, James McGuire, Julia Lynch, Jeff Sturchio, Michael Strauss, Michéle Boccoz, and Rajat Khosla. We wish to thank the Rockefeller Foundation for making the initial meeting at The Lancet office in London, UK, possible through their generous support. ==== Refs References 1 Shiffman J Shawar YR Framing and the formation of global health priorities Lancet 2022 published online May 17. 10.1016/S0140-6736(22)00584-0 2 Ho CJ Khalid H Skead K Wong J The politics of universal health coverage Lancet 2022 published online May 17. 10.1016/S0140-6736(22)00585-2 3 Kickbusch I Liu A Global health diplomacy—reconstructing power and governance Lancet 2022 published online May 17. 10.1016/S0140-6736(22)00583-9 4 Lasswell H Politics: who gets what, when, how? 1950 P Smith Publications New York, NY 5 Dahl RA The concept of power Behavioral Sci 2 1957 202 203 6 Barnett M Duvall R Power in international politics Int Organ 59 2005 39 75 7 Shearer J Abelson J Kouyaté B Why do policies change? Institutions, interests, ideas and networks in three cases of policy reform Health Policy Plan 31 2016 1200 1211 27233927 8 Hall P The role of interests, institutions, and ideas in the comparative political economy of the industrialized nations Lichbach MI Zuckerman AS Comparative politics: rationality, culture, and structure 1997 Cambridge University Press Cambridge 174 207 9 Walt G Health policy: an introduction to process and power 1994 Witwatersrand University Press Johannesburg 10 Wong J Healthy democracies: welfare politics in Taiwan and South Korea 2006 Cornell University Press Ithaca, NY 11 North D Institutions J Econ Perspect 5 1991 97 112 12 Immergut EM Institutions, veto points, and policy results: a comparative analysis of health care J Public Policy 10 1990 391 416 13 Béland D Ideas and social policy: an institutionalist perspective Soc Policy Administr 39 2005 1 18 14 Finnemore M Sikkink K International norm dynamics and political change Int Organ 52 1998 887 917 15 Barberia LG Gómez EJ Political and institutional perils of Brazil's COVID-19 crisis Lancet 396 2020 367 368 32738938 16 Singh P How solidarity works for welfare: subnationalism and social development in India 2015 Cambridge University Press Cambridge 17 Lieberman E Boundaries of contagion 2009 Princeton University Press Princeton, NJ 18 Gómez EJ Coca-Cola's political and policy influence in Mexico: understanding the role of institutions, interests, and divided society Health Policy Plan 34 2021 520 528 19 Mahoney J Path dependency in historical sociology Theory Soc 29 2000 507 548 20 Hacker J The historical logic of national insurance: structure and sequence in the development of British, Canadian, and U.S medical policy Stud Am Political Develop 52 1998 57 130 21 Bali AS He AJ Ramesh M Health policy and COVID-19: path dependency and trajectory Policy Soc 41 2022 83 95 22 Chua AQ Tan MMJ Verma M Health systems resilience in managing the COVID-19 pandemic: lessons from Singapore BMJ Global Health 5 2020 e003317 23 You J Lessons from South Korea's COVID-19 policy response Am Rev Public Administr 50 2020 801 808

35594876

PMC9753152

NO-CC CODE

2022-12-16 23:25:31

no

Lancet. 2022 May 17 4-10 June; 399(10341):2080-2082

utf-8

Lancet

10.1016/S0140-6736(22)00923-0

oa_other

==== Front Technol Forecast Soc Change Technol Forecast Soc Change Technological Forecasting and Social Change 0040-1625 0040-1625 Elsevier Inc. S0040-1625(21)00165-7 10.1016/j.techfore.2021.120733 120733 Article The Covid-19 pandemic and the accommodation sharing sector: Effects and prospects for recovery Gerwe Oksana Brunel University London, Brunel Business School, London, United Kingdom 8 3 2021 6 2021 8 3 2021 167 120733120733 16 2 2021 5 3 2021 © 2021 Elsevier Inc. All rights reserved. 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The Covid-19 pandemic has spread like wildfire across the globe. The hospitality industry, including the accommodation sharing sector, has been one of the hardest hit. Renting an Airbnb property, sharing a room via Couchsurfing and exchanging homes via LoveHomeSwap became almost impossible under the new restrictions. This paper analyses the effects of the Covid-19 crisis on the accommodation sharing sector and conceptually uncovers the underlying reasons for its disruption. We submit that the main strengths of the accommodation sharing sector, which originally drove its rise, became its weaknesses during the pandemic. An asset-light business model, the intermediation of physical transactions via online platforms, a reliance on individually owned and underused properties, and the popularization of access over ownership propelled the initial expansion of this sector. However, these all backfired during the pandemic. The paper outlines potential avenues for the post-pandemic recovery of accommodation sharing and presents future directions for research. Keywords Accommodation sharing Sharing economy Covid-19 pandemic Post-pandemic recovery ==== Body pmc1 Introduction The Covid-19 pandemic has spread like wildfire across the globe, ravaging lives, livelihoods, societies and economies. The unprecedented surge of the virus turned the world upside-down for billions of people in most countries. While the crisis continues, we are still far from being able to fully quantify its devastating human, economic and societal costs. However, we can see that the hospitality industry, including the accommodation sharing sector, was one of the hardest hit (Gössling et al., 2020). Only a few years ago, the sharing economy as a whole, of which accommodation sharing has become an integral part, was predicted to grow to $335 billion in 2025, from $14 billion in 2014 (PwC, 2015). Accommodation sharing companies, including Airbnb, Couchsurfing, LoveHomeSwap, Wimdu and Onefinestay, have become as familiar to travelers as more traditional hospitality providers, such as Marriott International, Hilton Worldwide and Holiday Inn. Fast forward to the beginning of 2020. Global travel came close to a standstill (Farmaki et al., 2020; Škare et al., 2021). According to a report from the UK Office for National Statistics, for example, the hospitality sector in the United Kingdom was hit harder than any other sector in the economy, with accommodation bookings in June 2020 down 92.2% compared to February of the same year.1 Similarly, passenger flow through Heathrow in April showed a year-on-year fall of 97%.2 Overall, aeroplane travel this year is predicted to contract by at least half, compared to the record 4.54 billion passengers carried by the global airline industry in 2019.3 This is hardly surprising, given that in June 2020 around 90% of the world's population was living in countries with largely closed borders. For a global accommodation company like Hilton, with thousands of properties across 118 countries, this resulted in an 81% drop in the second-quarter revenue per available room (RevPAR) .4 In the short-term accommodation sharing sector, the number of bookings for most destinations in May dropped by at least half, compared to mid-February levels (DuBois, 2020). Measures taken to tackle the spread of Covid-19 nearly brought accommodation sharing to a complete halt during the peak of the pandemic. Flight cancellations, border closures, lockdowns and quarantines profoundly disrupted the hospitality industry and many of the accommodation sharing services that have become an essential part of our travel experience in recent years. Renting an Airbnb property at a travel destination, sharing a room via Couchsurfing or exchanging homes via LoveHomeSwap became almost impossible in a world of self-isolation, travel bans, social distancing and strict personal hygiene protocols. The scale of the disruption propelled some to question the very survival of the sharing economy and of accommodation sharing in the post-pandemic world (Rinne, 2020). Why was accommodation sharing hit particularly hard by the spread of Covid-19? What are the prospects for its recovery in the post-Covid-19 world? This conceptual paper provides a systematic overview of the effects of the Covid-19 crisis on the accommodation sharing sector. Building on the existing literature about the sharing economy (Belk, 2014; Frenken and Schor, 2019; Gerwe and Silva, 2020), we uncover the deep underlying reasons for this devastation (Zenker and Kock, 2020). We posit that the unique strengths that have allowed the accommodation sharing sector to reach an unprecedented scale in the last decade, became weaknesses during the Covid-19 crisis. According to our analysis, the Covid-19 outbreak is likely, on the one hand, to expose many of the weak points of sharing businesses, taking some of them to breaking point, affecting suppliers, consumers and platforms in the accommodation sharing sector. On the other hand, the pandemic is likely to push a reset button, bringing long-awaited changes to some of the problems that developed in the sharing economy during the recent boom years, benefiting the sector in the long term. In this paper, we outline the lessons that the accommodation sharing sector has learned from the Covid-19 pandemic, and conceptualise the potential avenues for recovery that we can expect in this sector. This study makes two main contributions to the literature. First, we contribute to the emerging literature on the effects of the Covid-19 pandemic and the prospects for a post-pandemic recovery in the hospitality industry in general and the accommodation sharing sector in particular (Dogru et al., 2020; Dolnicar and Zare, 2020; Gössling et al., 2020; Nicola et al., 2020). Secondly, we contribute to the literature on accommodation sharing (Zervas et al., 2017). Even though pandemics are not new experiences for the hospitality industry (Farmaki et al., 2020), accommodation sharing itself is still relatively young, having emerged only about a decade ago (Botsman and Rogers, 2010), and this is its first major pandemic. By looking more deeply at the unique distinguishing features of the sharing economy (Frenken and Schor, 2019; Guttentag, 2015), we uncover the inherent limitations of this phenomenon that the Covid-19 crisis has exposed, especially in the context of accommodation sharing (Zenker and Kock, 2020). The paper is structured in the following way. After the Introduction, Section Two provides an overview of the sharing phenomenon and the effects of the Covid-19 pandemic on the accommodation sharing sector. Section Three uncovers the reasons why the Covid-19 pandemic turned the strengths of the sharing economy into weaknesses. Section Four explores the potential opportunities and challenges of the post-Covid recovery for accommodation sharing. The last section offers a conclusion and provides suggestions for future research. 2 The effects of the Covid-19 pandemic on the accommodation sharing sector A black swan event of unprecedented proportions, the Covid-19 pandemic, surged across the globe in 2020, wreaking havoc on people's lives, societies and economies, catching populations, businesses and governments off guard. In the early weeks of 2020, China, where the pandemic originated, introduced swift measures to fight the disease: grounding flights, closing borders, isolating epicentres of infection and introducing lockdowns for individuals in affected cities. Soon after, evidence started to emerge about Covid-19 cases on several cruise ships, including the Diamond Princess which ended up being quarantined in Japan. By February 2020, cases of Covid-19 were spreading quickly in Europe, leading to a major lockdown in Northern Italy. What seemed at first to be a disjointed patchwork of local events quickly transformed itself into a worldwide cataclysm. Country after country enacted similar measures to fight the pandemic. Life as it had previously been known came to an abrupt halt. As countries developed plans to fight the pandemic, world economies braced for an imminent blow, which, by some estimates, could yet become the worst peacetime recession in a century (OECD, 2020). Disrupted supply chains, closed businesses, job cuts, furloughs and heavily reduced demand brought devastation to many areas of economic activity (Pantano et al., 2020). The hospitality industry was one of the hardest hit by this crisis. In the United States, for example, in the first week of March 2020, hotel industry revenue per available room fell by 11.6% (Durbin, 2020). Occupancy rates of hotels in China collapsed by 89% in January and by 36% in Germany in March 2020 (Durbin, 2020). All over the world, many traditional hospitality providers had to furlough employees, borrow precautionary funds to weather the storm, seek government aid or temporarily suspend operations altogether (Nicola et al., 2020). The accommodation sharing sector, by now an integral part of the hospitality industry, also suffered a severe blow from the pandemic. The Covid-19 crisis “has gutted” the sector that just a few months earlier was enjoying positive predictions for 2020 (Conger and Griffith, 2020; DuBois, 2020; Gössling et al., 2020). Suppliers, consumers and accommodation sharing platforms have all been negatively affected. Even though the evidence that is already available is often location-specific, it demonstrates the severe downstream effects of the pandemic for global accommodation sharing. On the demand side, weekly bookings on Airbnb between January and March 2020, for example, fell significantly, although this varied across regions, falling by 96% in Beijing, 46% in Seoul and 29% in Milan (DuBois, 2020). On the supply side, in Australia, for instance, the pandemic led to a 70% decrease in income for Airbnb hosts (Chen et al., 2020). As for the accommodation sharing platforms, Airbnb in May 2020 laid off 1900 employees, about 25% of its staff, halved its revenue forecast compared to 2019, slashed costs and raised an additional $1 billion in new emergency funding. Similarly, Couchsurfing saw a 90% drop in demand and, to save its operation, introduced its first-ever membership fee, causing much controversy and discontent amongst its user base. The question inevitably arises: why did as dynamic a sector as accommodation sharing suffer so badly from the Covid-19 crisis? In order to better understand the effects of Covid-19 on accommodation sharing, it is important first to understand the roots and distinctive features of the sharing economy. We argue that the very features that had in the last decade propelled the sharing economy in general, and accommodation sharing in particular, to massive popularity and scale, brought it close to breaking point under the conditions of the pandemic. 2.1 An overview of the sharing economy phenomenon The sharing economy came into existence through the convergence of various technological, economic, social and environmental factors (Botsman and Rogers, 2010). The widespread penetration of the internet and mobile connectivity, together with the development of global positioning system (GPS) technology and mobile apps, allowed companies in different industries to seamlessly connect suppliers and customers in providing and consuming digitally facilitated sharing services (Belk, 2014; Puschman and Alt, 2016). Economic pressures, job losses and the 2008–2009 financial crisis forced individuals to search for alternative sources of income, which the sharing economy companies could offer (Laamanen et al., 2015). Socially, the culture of over-consumption prevalent during the preceding era gave way to more conscientious consumption choices, especially by the millennial generation (Cohen and Munoz, 2016). Finally, environmental pressures, sustainability concerns and the threat of climate change moved public opinion towards more careful use of resources and better stewardship of people's existing assets (Botsman and Rogers, 2010; Hamari et al., 2016). Even though humanity has engaged in sharing since time immemorial (Belk, 2014), ‘the sharing economy’ emerged relatively recently as an umbrella term for a range of business strategies. These strategies generally imply temporary peer-to-peer access to individually owned assets (human or physical) facilitated by a digital platform (Frenken and Schor, 2019; Gerwe and Silva, 2020). In a single decade, some sharing economy companies, including Airbnb, Uber, TaskRabbit, Blablacar and Lyft, have become global household names, changing the way we travel, move around and perform everyday tasks (Taeuscher, 2019). Prior to the Covid-19 pandemic, some of these companies achieved billion-dollar evaluations, went public on the stock market (Uber and Lyft), or were about to do so (Airbnb). According to a report from the World Bank Group, in 2018 the accommodation sharing sector comprised 7% of worldwide accommodation, with about 8 million beds. Based on pre-Covid-19 projections, this sector was predicted to grow by 31% annually between 2013 and 2025, six times higher than the growth rate of traditional accommodation (World Bank, 2018). During the expansion process, different business strategies were adopted by accommodation sharing platforms, using both non-profit and for-profit business models. However, for-profit firms quickly came to dominate the accommodation sharing sector, as with other sectors of the sharing economy (Gerwe and Silva, 2020). As a result, most of the companies that are today associated with accommodation sharing, such as Airbnb, Couchsurfing and LoveHomeSwap, are for-profit organizations. However, one of the main distinctions of the business strategies of these platforms within the dominant for-profit segment concerns the monetary exchange between platform users. For example, accessing a couch or a room via Couchsurfing is free for hosts and guests, but the host has the option of paying a fee for additional services, which generates revenue for the platform. LoveHomeSwap works on a subscription model that allows individuals to put their home into the pool of vacation properties and swap it with other property owners during trips. In this case, the platform generates revenue from user subscriptions but the property owners are not renumerated in the exchange. Airbnb, by contrast, is a fee-based business, typically charging the host 3% of the rental price and the guest about 14%, while the hosts generate income from renting out their properties to the guests. Despite the apparent differences between various sharing economy platforms and their business models, recent research shows that they all share four common features. Firstly, they blend the digital and the physical, as they connect suppliers and consumers virtually through a digital platform, while facilitating transactions in the physical world (Belk, 2014; Botsman, 2013; Frenken and Schor, 2019; Hamari et al., 2016; Sundararajan, 2016). Secondly, they emphasize peer-to-peer transactions, where both the providers and the consumers are ordinary citizens, rather than businesses or professional operators (Frenken and Schor, 2019; Hamari et al., 2016). Thirdly, sharing businesses rely on idle capacity owned by individuals (e.g., physical assets, resources, skills, or time) (Botsman, 2013; Frenken and Schor, 2019; Stephany, 2015). Fourthly, access is the underlying logic behind the transaction, rather than ownership (Belk, 2014; Botsman, 2013; Frenken and Schor, 2019; Hamari et al., 2016; Stephany, 2015). The combination of these four characteristics represents the essence of sharing economy innovation. These features of the sharing economy explain how this way of providing services and goods managed to achieve such scale across mobility, transportation, household help and many other industries in only a decade. This is especially true of accommodation sharing. 2.2 The key features of the sharing economy and their role in the expansion of accommodation sharing 2.2.1 Online platforms that facilitate offline transactions Similar to all sharing economy firms, accommodation sharing platforms are digital multisided platforms (Constantiou, Marton, and Tuunainen, 2017; Henten and Windekilde, 2016) that facilitate interactions between two groups of users – hosts and travellers. They create value by matching the right properties, or listings, with potential guests (Parker et al., 2016). Accommodation sharing platforms do not own the assets that underlie the transactions (Frenken, 2017), i.e. real estate, but make investments (e.g., in advertising and technology) that reduce the barriers to entry for individual property owners willing to participate in sharing activities as providers (Einav et al., 2016). An asset light business model allowed sharing economy based accommodation platforms, such as Airbnb, LoveHomeSwap, Couchsurfing, to achieve incredible scale very rapidly (Frenken, 2017). Since sharing platforms combined digital interaction with transactions in the real, physical world, the question of trust and safety became of paramount importance. Indeed, sharing a home with a complete stranger is a sensitive and a potentially risky undertaking for the host, the guest and even the property. Hence, in addition to superior matching mechanisms and reduced barriers to entry (Davis, 2016), accommodation sharing firms developed efficient mechanism for establishing digital trust on their platforms, such as ratings of hosts and properties, reviews of stays, etc. Thanks to the ingenious trust infrastructure (Calo and Rosenbat, 2017; Schor, 2016), accommodation sharing platforms managed to mitigate the risks of transacting with strangers in the real world, taking the traditional concept of sharing outside the circle of immediate friends and family to a much broader community of users. As a result, in a matter of years we became comfortable with the concept of accommodation sharing, which is evidenced by more than 260 million bookings on Airbnb and 14 million members on Couchsurfing. 2.2.2 Peer-to-peer nature of accommodation sharing transactions Transactions in the accommodation sharing sector are mostly peer-to-peer (Einav et al., 2016). This term refers to transactions that take place between platform participants who are not large companies or organizations. Unlike the traditional hospitality sector, where providers are usually businesses or firms, providers on accommodation sharing platforms are ordinary citizens, individuals, small suppliers or micro players. Small-scale individual participation in accommodation sharing as well as in other sharing economy sectors was part of the original appeal of this type of company. Even the word ‘sharing’ in the term ‘sharing economy’ refers to the individual level of social interaction and human connection, popularized by sharing economy firms, in contrast with traditional business-to-consumer transactions. For example, a personal connection between the host and the guest still remains a major part of the company brand and ethos at Airbnb and Couchsurfing, where the former describes itself as a global marketplace and community “powered by local hosts”5 and the latter highlights the fact that its users “share their lives with the people they encounter, fostering cultural exchange and mutual respect” .6 It is important to note that accommodation sharing has evolved to become increasingly professionalised, in a visible departure from its peer-to-peer origins (Dogru et al., 2020; Frenken and Schor, 2017). As accommodation sharing gained popularity, sharing platforms saw an influx of professional real-estate operators managing large portfolios of properties. For this type of provider, participation in the sharing economy was not a way of generating additional income but a primary business activity. The trend of increasing professionalization allowed accommodation platforms to achieve greater efficiencies in delivering accommodation services to customers. However, this drew harsh criticism and substantial pushback from some activists, the general public, local authorities and residents who felt that sharing economy accommodation businesses caused gentrification, excessive tourism and a shortage of residential rentals in some neighbourhoods or even in entire cities. Searching for growth, Airbnb, for example, went even further and opened its platform not only to individual providers but also to hotels and professional bed-and-breakfast operators, a move that is arguably misaligned with its central idea of staying with the locals and experiencing another destination as a local. Regulatory compliance and the tax obligations of different types of suppliers have also been a consistent issue on accommodation sharing platforms. In response, regulators in different countries have been working on differentiating individual service providers from professional operators on accommodation sharing platforms7 . 2.2.3 Underused capacity: individually owned real estate The third feature that defines the accommodation sharing sector and explains its explosive growth is the reliance on under-utilized capacity, such as an empty room, a vacant vacation property, a second home or apartment. Sharing economy platforms in different sectors have made it easy and cheap for individuals to offer their excess capacity to others (Calo and Rosenblat, 2017), increasing the use and value of privately owned idle assets. Unlike traditional hospitality providers who rely on dedicated real-estate capacity – including hotels, bed-and-breakfasts, resorts and hostels – sharing accommodation platforms rely on a vast pool of individually owned properties that their owners may want to offer to travellers, whether for money (on Airbnb) or gratis (on Couchsurfing). Sharing platforms unlock the value of underused physical assets and bring benefits for the owners, the consumers, the platform and the society at large through better utilization of available resources (Calo and Rosenblat, 2017; Frenken, 2017; Schor and Attwood-Charles, 2017; Sundararajan, 2016). Tapping the pool of underused individually owned properties (Parker et al., 2016) allowed accommodation sharing platforms to grow much faster than traditional hospitality competitors because they did not need to invest in the assets underlying the transaction, i.e. real estate. It should be mentioned, however, that reliance on underused capacity by accommodation sharing platforms has diminished in recent years, as the platforms have become more professionalised. On the one hand, providers who saw the appeal and the economic benefits of participation in accommodation sharing platforms brought not only their empty rooms and vacation homes to the platform but also some dedicated real-estate which had been specifically acquired in order to be rented out on a short-term basis (Slee, 2015). On the other hand, the platforms themselves, such as Airbnb, not only opened their operations to hotels, as mentioned above, but also started to develop branded apartment buildings where rooms and apartments would be exclusively available through their platforms – a clear shift from underused to dedicated capacity.8 Nevertheless, the underused real estate owned by individuals played a pivotal role in the initial expansion of accommodation sharing platforms and continues to form the substantial part of their supply. 2.2.4 Temporary access over ownership The sharing economy is built on the idea of temporary access to an asset rather than its permanent ownership (Botsman and Rogers, 2010). The access rather than ownership ethos of the sharing economy came to represent not only the economic benefits of sharing transactions but also their inherent flexibility, the choice of experiences over accumulation of things and the greater variety available to consumers, all of which have been particularly valued by millennials. The main advantage of this approach is that consumers are able to enjoy a greater variety of services or products while making smaller financial commitments. In the mobility sector, for example, instead of owning a car, in the same week a user can access a bike for shorter rides, rent a car for a longer journey, share a ride with another driver to travel between cities and hire a van to move furniture at the weekend. In the accommodation sharing sector, property sharing has also come to signify the shift from standardization, often equated with staying in a hotel room, to the authenticity of staying with locals and experiencing a travel destination as a local (Guttentag, 2015). To summarise, these four features have substantially contributed to the explosive rise of the accommodation sharing sector, making it not only a viable and economically smart way to find a place to stay but also a popular and trendy one. An asset-light business model, the secure digital intermediation of physical transactions via online platforms, the professional aggregation of underused properties owned by individuals and the popularization of access over ownership all help to explain how Airbnb, Couchsurfing and other accommodation sharing platforms have managed to successfully disrupt the traditional incumbents of the hospitality sector and change the dominant paradigm of the industry (Guttentag, 2015). 3 Covid-19 turns the strengths of the sharing economy into weaknesses Now that we have analysed the unique common features of firms in the sharing economy and discussed the role of each in the expansion of accommodation sharing businesses, we present our analysis of the effects of the Covid-19 pandemic on the sector. We argue that the current crisis has turned every strength of the sharing economy and of accommodation sharing into a weakness. Firstly, the introduction of lockdowns and quarantines severed the link between the digital and the physical in the sharing economy. While the digital interfaces of sharing businesses were still active, unaffected by changes in the real world, the physical aspects of sharing transactions, such as leaving your own home, travelling and entering someone else's property, all became impossible under the new restrictions. For example, Airbnb website functioned continuously throughout the pandemic, but its users were not able to rent (out) properties in areas which had closed down because of the spread of the virus. Thus, even though the virtual connection between hosts and guests was still possible, the delivery and consumption of services that the sharing platform was designed to facilitate came to an abrupt halt. At the platform level, the absence of transactions inevitably led to the implosion of businesses and loss of revenue from fees and premium services. Similarly, individual suppliers of the platforms suddenly found that demand for their properties dried up, causing the cancellation of bookings, lost income and major uncertainties about the future of the platforms, sharing businesses and hospitality industry as a whole. Secondly, the peer-to-peer nature of accommodation sharing changed overnight from an asset to a liability. What looked like a source of authenticity in the pre-Covid-19 world, and a gateway to a local experience and personal connection, became a potential health hazard during the pandemic. The new demands for personal hygiene and safety meant that the peer-to-peer business model suddenly started to look unreliable and even dangerous. Because of the pandemic, cleanliness, sharing a space or simply touching an object that might recently have been touched by someone else seemed to determine whether one caught the virus or passed it on to someone else. As a result, participants in accommodation sharing were no longer comfortable staying in properties that had recently hosted multiple guests or inviting strangers into their homes. The allegedly carefree, relaxed attitude to health-and-safety requirements and the lack of standardization adopted by the sharing economy as part of its easygoing ethos during the boom times now started to look like carelessness (Gerwe, 2020). Covid-19-related health-and-safety requirements inevitably gave a new resonance to the issue of trust that has always been particularly relevant in the context of peer-to-peer transactions (Agag and Eid, 2019; Calo and Rosenblat, 2017; Cheng et al., 2019; Ert et al., 2016; Lee, 2015). In contrast, traditional accommodation providers, such as hotels, with strict and well-developed safety and cleanliness protocols, now started to seem a more reliable and appealing alternative to accommodation sharing. Thirdly, it is not surprising that privately owned assets that had been underused for years, became much better used in the midst of the pandemic. During the lockdown, people all over the world have been very keen to share videos, TikToks and other social media posts showing old guitars, sports equipment, kitchen appliances, easels and gardening tools, which after years of underuse had suddenly been brought out of garages and closets, much to their owners’ delight. This has been even more the case for the lucky owners of gardens, patios and balconies which provided a much-needed escape and a breath of fresh air during the many weeks of lockdown. A spare room overnight turned into a home office used all day, every day. The Covid-19 pandemic and the resulting economic, medical and social uncertainty gave new resonance to the old saying “My house is my fortress”, and forced us to use and appreciate more the resources that we possess. Lastly, the lockdown forced us to re-evaluate the role of asset ownership in our private lives, since it put us in a situation where we could use only what we already had. Indeed, lockdowns and social distancing measures made access to other people's assets impossible, as we could not leave the house and access them. According to anecdotal evidence, people who owned more assets, such as property, a car, IT equipment, garden furniture, a bike, grill or scooter, were able to spend the long weeks of the lockdown more comfortably than those who, before the pandemic, relied on accessing those items through the sharing economy. The measures adopted to fight the pandemic changed the ethos of access-over-ownership back to ownership-over-access. The Covid-19 crisis made us acutely aware that asset ownership allows us not only to use an asset when we need it but also gives the owner greater control over its cleanliness and safety conditions, which are crucial for our health. Moreover, this increases one's economic safety and security, as the assets that one owns may make it more possible to cope with the economic, psychological and social effects of the Covid-19 crisis. While these have been the immediate effects of Covid-19 pandemic on the sharing economy and on the accommodation sharing, we must now ask what kind of recovery can be expected for this sector in the post-pandemic world. 4 The post-covid-19 recovery: opportunities and challenges Based on the unique features of the sharing economy, as well as the general trends that are already emerging today and are forecast for the post-Covid-19 recovery, in the future we can expect significant changes for all the actors in the accommodation sharing sector, including customers, suppliers and platforms. There will also be changes to the broader context of accommodation sharing. In 2004, in a letter to investors, Warren Buffett famously said: “Only when the tide goes out do you discover who's been swimming naked”. In a similar fashion, the Covid-19 pandemic is likely to expose major weaknesses in the accommodation sharing sector and also serve as an impetus for changes that may ensure its more sustainable future. 4.1 Demand: Shifting from global to local The first and most likely effect of the Covid-19 pandemic on the accommodation sharing sector will be a shift in demand from global to local. Early in the crisis, as countries started to close down borders and introduce lockdowns, it became clear that the globalized world would undergo tectonic shifts in the way globalization is perceived and the way it is enacted in economies and societies. Even prior to the Covid-19 outbreak, there was noticeable disillusionment with the idea of globalization. So-called ‘slowbalisation’9 manifested itself in many different ways from Brexit in the United Kingdom to trade wars between the USA and China and the renegotiation of international trade agreements between multiple counterparts. However, despite political moves away from globalization, business supply chains and consumer behaviour were still predominantly global in terms of their overall outlook and practical arrangements. International travel, the consumption of products and services from all over the world, and freedom of movement for work, education and leisure all played a big role in the expansion of accommodation sharing platforms. Measures taken to combat the pandemic disrupted international supply chains and stopped the flow of goods and people between countries and even between different regions of the same country. As we move out of the crisis and the lockdowns are lifted, we are unlikely to see an immediate return of demand for international travel (Dolnicar and Zare, 2020). Even if governments relax all the restrictions that were implemented during the peak of the crisis, overall customer sentiment seems to have shifted towards greater caution. Having collectively experienced the pandemic, even if we are again able to travel to a foreign capital or a remote destination, we may still not want to do so. Staying closer to home is likely to become a more attractive alternative to foreign or remote travel for many, especially if travellers can avoid urban areas with high levels of congestion and move to more rural settings. What this will mean for accommodation sharing will be a shift in demand from global destinations to local, domestic travel. It is therefore not surprising that Airbnb has already launched its ‘Go Near’ campaign, designed to encourage local travel. In addition to health-and-safety considerations, this trend will be exacerbated by the economic difficulties experienced by many individuals around the world caused by loss of jobs and income. Local trips will be more affordable for many customers in the post-Covid-19 world. Early signs of this trend are already noticeable on some accommodation sharing platforms. According to the CEO of Airbnb, Brian Chesky, the demand for domestic accommodation has more than doubled on his platform, to over 80%.10 Almost 60% of customers now book properties up to 300 km from home, compared to only 33% in the pre-Covid-19 period. In addition, since many people continue to work remotely, individual stays are becoming significantly longer. Travellers seem to be combining travel with work, and are enjoying a change of scenery in quiet locations away from urban centres after weeks of isolation without having to rush back to their physical workplace. Furthermore, leisure travel is likely to recover first, while business trips may take a long time to fully return to pre-pandemic levels, as many companies have moved their operations online and will continue to rely at least in part on virtual meetings once the Covid-19 crisis is over. Thus, in the foreseeable future, local, rural and affordable stays for longer periods of time may replace shorter work or leisure trips to large metropolitan centres or weekend getaways to famous cultural destinations. 4.2 Supply: shakeout amongst accommodation providers and the new standards of service Suppliers of properties on the sharing economy platforms are facing two major challenges in the aftermath of the Covid-19 pandemic: the economic impact and the effect on the health-and-safety standards of the accommodation sharing services. Interestingly, the two challenges are likely to have divergent effects on different types of providers across different platforms. Firstly, in terms of the economic impact, all property providers have been hit extremely hard by lockdown and the cancellation of bookings. However, the effect of the pandemic on suppliers has been uneven. Individual providers of shared accommodation on non-commercial platforms, such as Couchsurfing, lost bookings but did not suffer major economic losses, since transactions there do not include monetary compensation for the host. In contrast, property suppliers on commercial platforms, such as Airbnb or Onefinestay, incurred substantial financial losses due to cancellations or lost bookings. The response of platforms to the pandemic increased the losses suffered by the hosts in the accommodation sharing sector. For example, Airbnb, in an attempt to support its customers, instituted free cancellation of bookings from 14 March to 15 July 2020, much to the hosts’ dismay. Furthermore, we can expect the economic effects of the Covid-19 crisis on the two types of providers on commercial sharing platforms – individual peer-suppliers and professional real-estate players – to be very different. With demand in sharp decline, property suppliers that professionally manage dedicated portfolios of short-term tourist rentals, probably acquired with substantial bank loans or requiring high rental payments, may become unable to service their mortgages or pay rent. As a result, in the short to medium term, such properties may have to be sold or repositioned as long-term residential rentals. In large metropolitan centres, such as London or San Francisco, this may be a welcome change for the local pool of residential housing but it will undoubtedly create major disruption for professional players on accommodation sharing platforms. The era of easy money for this type of property suppliers is likely to be over. As a result, the total share of professionals amongst accommodation sharing suppliers may fall (Dolnicar and Zare, 2020). In contrast, individual property owners, who only occasionally rent out an empty room or a second home to guests and whose primary use of those properties is not commercial but personal, may be better able to recover from the Covid-19 crisis once tourism and travel resume. In the medium to long term, the possible shift of professional players to longer lets may benefit peer-suppliers who use accommodation sharing platforms to generate additional income, rather as their main occupation. Reduced competition from professionals may allow individual peer-providers to better weather upcoming recessionary times as their income on sharing platforms will continue to act as a financial cushion in the economic downturn. In a strange twist of events, the Covid-19 pandemic may bring back into the spotlight the original spirit of empowerment of ordinary individuals and human connection that were behind the sharing economy phenomenon a decade ago. Secondly, the pandemic is likely to challenge and improve the health-and-safety standards of the accommodation sharing services. Even before the Covid-19 crisis, there were some voices in favour of tightening health-and-safety standards in this sector. In response, sharing businesses typically pointed to the peer-to-peer nature of accommodation supply, which is much more heterogeneous than the accommodation provided by the traditional hospitality sector. Therefore, they argued, the rigid standards of the latter could not and should not be applied to accommodation sharing. The Covid-19 pandemic is likely to change the hygiene and cleanliness requirements for properties offered on accommodation sharing platforms. This will be an opportunity for the property suppliers to develop a new set of procedures that will ensure the health and safety of both travellers and hosts, and take industry standards to a higher level, benefiting the industry in the long term. Unlike the financial recovery discussed above, professional suppliers may be more likely to implement higher standards of service than their non-professional competitors, as the former may be better equipped to provide better, more consistent and more thorough cleaning services than individual peer-suppliers. Hence, professional suppliers of short-term accommodation that manage to withstand the financial troubles caused by Covid-19 can be expected to receive greater confidence from travellers than their non-professional counterparts. Overall, restoring trust in accommodation sharing is likely to remain a challenge for different platforms and different types of suppliers. More generally, even though previous pandemics, such as SARS in China, showed that the tourism and hospitality industry was able to rebound relatively quickly once they were over (Dombey, 2004), the scale of Covid-19 has called into question the future of the entire global hospitality and travel industry (Farmaki et al., 2020). In order to rebuild trust in the reliability of their services and boost consumer confidence, for example, Airbnb established a new cleaning protocol for its properties in the immediate aftermath of lockdown, including the use of specific cleaning products and a 24-hour break between bookings (Wood, 2020). However, this protocol is optional for the host and its introduction may therefore not fully achieve the desired effect. It is important to recognise that restoring consumer trust and post-pandemic demand in the sector will depend not only on the efforts of suppliers and the platforms but also on the post-pandemic policy response of regulators. As the situation continues to shift and vary in different locations, public policy concerning accommodation sharing activities needs to be targeted and context specific, balancing public health concerns with the economic wellbeing of the population. 4.3 Cost cutting and greater focus on profitability The Covid-19 pandemic is likely to put an end to an unprecedented ten-year boom in the accommodation sharing sector that has propelled many platforms to an immense global scale, ensured spectacular company valuations and generated massive private investment in individual platforms. Even before this crisis, analysts and investors were becoming increasingly concerned about the business fundamentals of sharing companies, as they consistently chose growth over profitability while posting massive operational losses. The disruption caused by Covid-19 may push many sharing economy firms to breaking point, forcing them to thoroughly rethink their operations and strategies, and prioritise profitability and cashflow, instead of growth and a reliance on injections of capital by investors. Some platforms may go under and close down or be acquired by stronger players, leading to consolidation in the accommodation sharing sector. The surviving platforms will have to pay much closer attention to their cost structure, value propositions and core strategies. The bitter pill of Covid-19 is likely to force sharing companies to stop trying to be all things to everyone and focus closely on what they really do best. For example, Airbnb's CEO, Brian Chesky, is already talking about “going back to basics”, closing down non-core hotel and luxury divisions, and prioritising hosts who rent out their own homes rather than professional real-estate operators. If strict cleaning and hygiene protocols are put in place at accommodation sharing properties to a very high standard, if such standards are consistently enforced and well communicated to customers, in the mid-term, accommodation sharing may become more appealing to some travellers than other forms of accommodation, such as hotels, which typically have higher occupancy rates and less control over personal space, even though at the moment the opposite seems to be the case. The economic hardships together with the new health-and-safety requirements and potential shifts in consumer attitudes and behaviour may bring a wave of convergence between traditional hospitality players and accommodation sharing platforms, driving further consolidation in the sector and stimulating further evolution of business models of hospitality service suppliers. 4.4 Prospects of greater sustainability in the sector The tremendous scale achieved by the accommodation sharing sector in the last decade explains its role in the broader hospitality and tourism context. Prior to the Covid-19 crisis, not only consumers, providers and platforms but also entire communities and cities had come to depend on the revenues generated by travellers consuming goods and services at locations that became more accessible and popular because of accommodation sharing (Leung et al., 2019). The rapid recovery of this sector will be important for the resumption and recovery of commercial activities, such as restaurants, bars and shopping, in many local ecosystems. At the same time, some popular tourist attractions were suffering from over-tourism before the pandemic, generated at least in part by the availability of short-term accommodation. Cities including Venice, Barcelona, Amsterdam and Paris found unexpected relief from tourist congestion and overcrowding during the lockdowns. Thus, the Covid-19 outbreak could serve as a catalyst to bring a much needed balance and sustainability to the accommodation sharing playing field (Zenker and Kock, 2020). This is a unique opportunity for regulators and local authorities to implement better practices, rules and norms, balancing the tensions between economic gain and pressures on sustainability from over-tourism at some travel destinations, while encouraging more accommodation sharing elsewhere. Furthermore, regulatory intervention from policymakers may be necessary not only in the introduction of new standards for the accommodation sharing sector, including stricter health-and-safety requirements, but also to ensure compliance with those requirements by both platforms and users. If demand for accommodation sharing in the post-pandemic economy is to return, it will be important to learn the lessons of the Covid-19 crisis and make the wider accommodation sharing ecosystem more sustainable, resilient and safe for all parties involved. 5 Conclusion The Covid-19 pandemic has been described as the biggest crisis of our generation. In a matter of months, our world has been transformed in ways that we could have hardly imagined. The ultimate impact of this catastrophic outbreak will take time to quantify; the human and economic scarring will need years to heal. This conceptual study analyses the effects of the Covid-19 crisis on the accommodation sharing sector and maps out potential avenues for its recovery. By doing so, we contribute to several areas of literature, including research on the effects of the Covid-19 pandemic on the accommodation sharing sector (Dolnicar and Zare, 2020; Nicola et al., 2020); literature on the accommodation sharing economy (Frenken and Schor, 2019; Guttentag, 2015; Mokter, 2020) and broader research into the hospitality industry (Dogru et al., 2020; Zervas et al., 2017). The accommodation sharing sector was one of the many victims of Covid-19. Robust and dynamic just recently, it was brought close to breaking point by the measures taken to fight and stop the spread of the virus. The strengths of the accommodation sharing sector that explain its spectacular rise in the last ten years became its weaknesses during the Covid-19 crisis. Border closures, flight cancellations and lockdowns made sharing a couch on Couchsurfing, exchanging homes via LoveHomeSwap or renting an Airbnb apartment all but impossible. In the aftermath of the pandemic, we can expect drastic changes to affect accommodation sharing customers, suppliers and platforms. As the pandemic first broke out, accommodation sharing platforms showed considerable ingenuity and integrity in finding ways to contribute. For example, Airbnb's community-minded hosts opened up their homes to offer free accommodation for healthcare workers. Participants in other sectors of the sharing economy offered free rides, cars, meal services or recruitment services to help communities through this unprecedented time.11 Such actions are likely to contribute jointly to the post-pandemic legitimacy of the accommodation sharing sector and increase consumer goodwill and trust in its services. As the vaccination gets underway and the world recovers from Covid-19, the accommodation sharing sector will, hopefully, gradually regain its strength and popularity. Airbnb, for example, despite being hit hard by the pandemic, had a very successful initial public offering in the U.S.A. in December 2020, a clear sign of optimism regarding the outlook of accommodation sharing in the medium to long-term. Despite the obvious challenges of the upcoming months and years, the short-term accommodation sharing sector has a unique opportunity to readjust its operations in a more balanced, resilient and sustainable way to benefit all of its stakeholders. Building local communities, forging relationships and helping people save money may once again become the dominant logic of the accommodation sharing sector and the sharing economy at large. As the Covid-19 pandemic is far from over and the recovery of the accommodation sharing sector is still uncertain, further studies are needed on the subject. This paper offers a conceptual analysis of the effects of the Covid-19 pandemic on accommodation sharing. Empirical studies on this subject are urgently needed to gauge the losses incurred in the sector by platforms, suppliers and consumers. It will be worthwhile for future research to compare the losses incurred in the sector by platforms, suppliers and consumers, and to ask which responses by platforms to the pandemic and its aftermath were most and least effective. Given the increasing importance of the health-and-safety standards of accommodation sharing properties, we are likely to expect tighter control from the platform over its participants on both the supplier and the consumer sides. Even before the Covid-19 crisis, the asymmetry of power and control between the platform and its participants was one of the most contentious issues in the sharing economy literature (Calo and Rosenblat, 2017). Early evidence emerging from this crisis suggests that the losses to Airbnb hosts may be up to eight times higher than those to the platform itself (Chen et al., 2020), suggesting the asymmetry of risks and outcomes for the suppliers and the platform itself during an economic downturn. Since the performance of accommodation platforms is closely connected to the availability of the suppliers’ assets, it will be important to gain a greater understanding of the theoretical and practical implications of this imbalance of power and risk in the accommodation sharing sector. In terms of the platform participants, research needs to determine which shifts in attitude and behaviour have already occurred and which can be expected in the future amongst different accommodation platform users. Evidence is already emerging about changes in the perceptions and responses of hosts to Covid-19 (Farmaki et al., 2020). Similar analysis is needed regarding consumer behaviour of the users of accommodation sharing platforms. Answers to these and other questions would provide much needed evidence for policy-makers and regulators as they examine the measures required to help individuals, firms and economies weather the Covid-19 storm and offer targeted solutions for recovery. Generally speaking, we are only now beginning to understand the profound effects of the Covid-19 pandemic on the accommodation sharing sector. Hence, rigorous academic enquiry into its present and future, both empirical and theoretical, is truly essential. Dr Oksana Gerwe is the Associate Professor at the Division of Globalisation, Entrepreneurship and Strategy and the Director of the MBA Programme at Brunel Business School, Brunel University London.  Her research interests range from the sharing economy to peer-to-peer business models to disruptive innovation. Oksana.gerwe@brunel.ac.uk Brunel University London, Brunel Business School 1 https://www.ons.gov.uk/economy/grossdomesticproductgdp/articles/coronavirusandtheimpactonoutputintheukeconomy/june2020. 2 https://www.economist.com/leaders/2020/05/14/has-COVID-19-killed-globalisation. 3 https://www.statista.com/statistics/564717/airline-industry-passenger-traffic-globally/. 4 https://www.marketwatch.com/story/Covid −19-turned-the-hotel-industry-upside-down-but-it-wont-change-what-people-want-2020-08-23. 5 https://news.airbnb.com/about-us/. 6 https://www.couchsurfing.com/about/about-us/. 7 For example, the European Union is trying to distinguish clearly between the two types of participants, working out the different tax, regulatory, legal and other implications for each, based on the frequency of service, profit-seeking motives and the level of turnover generated by the service provider (European Commision, 2016). 8 https://www.businessinsider.com/airbnb-expanding-boutique-hotels-branded-buildings-2019-12?r=USandIR=T. 9 https://www.economist.com/leaders/2019/01/24/the-steam-has-gone-out-of-globalisation. 10 https://www.airbnb.co.uk/resources/hosting-homes/a/may-13-its-time-to-start-looking-ahead-192. 11 https://www.sharingeconomyuk.com/blog/covid-19-the-sharingeconomy-companies-taking-positive-action-to-help-combat-the-coronavirus-outbreak. ==== Refs References Agag G. Eid R. Examining the antecedents and consequences of trust in the context of peer-to-peer accommodation Int J Hosp Manag 81 2019 180 192 Belk R. You are what you can access: sharing and collaborative consumption online J Bus Res 67 8 2014 1595 1600 Botsman R. The sharing economy lacks a shared definition Fast Company 21 2013 2013 Botsman R. Rogers R. What's mine is Yours The rise of Collaborative Consumption 2010 Collins London Calo R. Rosenblat A. The taking economy: uber, information, and power Colum. L. Rev. 117 2017 1623 Chen G. Cheng M. Edwards D. Xu L. COVID-19 pandemic exposes the vulnerability of the sharing economy: a novel accounting framework J. Sustainable Tourism 2020 1 18 Cheng X. Fu S. Sun J. Bilgihan A. Okumus F. An investigation on online reviews in sharing economy driven hospitality platforms: a viewpoint of trust Tourism Manage. 71 2019 366 377 Cohen B. Munoz P. Sharing cities and sustainable consumption and production: towards an integrated framework J Clean Prod 134 2016 87 97 Conger K. Griffith E. The Results are In for the Sharing Economy. They are Ugly 2020 The New York Times https://www.nytimes.com/2020/05/07/technology/the-results-are-in-for-the-sharingeconomy-they-are-ugly.html Constantiu I. Marton A. Tuunainen V.K. Four models of sharing economy platforms MIS Q. Executive 16 4 2017 231 251 Davis, P., 2016. How do Sharing Economy Companies Grow? A Comparison of Internal and External Growth Patterns of Airbnb and Uber. Tennessee Research and Creative Exchange, Knoxville. Dogru T. Mody M. Suess C. Line N. Bonn M. Airbnb 2.0: is it a sharing economy platform or a lodging corporation? Tourism Management 78 2020 104049 Dolnicar S. Zare S. COVID 19 and Airbnb–Disrupting the disruptor Ann Tourism Res. 83 2020 102961 DuBois, D. 2020. Impact of the Coronavirus on Global Short-Term Rental Markets. https://www.airdna.co/blog/coronavirus-impact-on-global-short-term-rental-markets. Durbin D. Airbnb Hotels Seek US Government Aid as Demand Flattens 2020 Associated Press 10.1016/j.ijsu.2020.04.018 Einav L. Farronato C. Levin J. Peer-to-peer markets Annu Rev Econom 8 2016 615 635 Ert E. Fleischer A. Magen N. Trust and reputation in the sharing economy: the role of personal photos in Airbnb Tourism Management 55 2016 62 73 European Commission Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions, A European agenda for the collaborative economy SWD 2016 2016184 final Farmaki A. Miguel C. Drotarova M.H. Aleksić A. Časni A.Č. Efthymiadou F. Impacts of Covid-19 on peer-to-peer accommodation platforms: host perceptions and responses Int J Hosp Manag 2020 102663 Frenken K. Sustainability perspectives on the sharing economy Environ. Innovation Societal Trans. 100 23 2017 1 2 Frenken, K., Schor, J., 2019. Putting the sharing economy into perspective. In A research agenda for sustainable consumption governance. Edward Elgar Publishing. Frenken K. Schor J. Putting the sharing economy into perspective A Research Agenda for Sustainable Consumption Governance 2019 Edward Elgar Publishing Gerwe, O., 2020, April 21. Is the sharing economy over?. Entrepreneurship and Innovation Exchange. https://eiexchange.com/content/is-the-sharingeconomy-over. Gerwe O. Silva R. Clarifying the sharing economy: conceptualization, typology, antecedents, and effects Acad Manage. Perspect. 34 1 2020 65 96 Gössling S. Scott D. Hall C.M. Pandemics, tourism and global change: a rapid assessment of COVID-19 J. Sustainable Tourism 29 1 2020 1 20 Guttentag D. Airbnb: disruptive innovation and the rise of an informal tourism accommodation sector Curr. Issues Tourism 18 12 2015 1192 1217 Hamari J. Sjöklint M. Ukkonen A. The sharing economy: why people participate in collaborative consumption J Assoc Inf Sci Technol 67 9 2016 2047 2059 Henten, A.H., Windekilde, I.M., 2016. Transaction costs and the sharing economy, info, Vol. 18 No. 1, pp. 1-15. https://doi.org/10.1108/info-09-2015-0044. Laamanen M. Wahlen S. Campana M. Mobilising collaborative consumption lifestyles: a comparative frame analysis of time banking Int J Consum Stud 39 5 2015 459 467 Lee C. The sharers’ gently-used clothes Contexts 14 1 2015 17 18 Leung X.Y. Xue L. Wen H. Framing the sharing economy: toward a sustainable ecosystem Tourism Manage. 71 2019 44 53 Nicola M. Alsafi Z. Sohrabi C. Kerwan A. Al-Jabir A. Iosifidis C. Agha M. Agha R. The socio-economic implications of the Coronavirus and COVID-19 pandemic: a review Int. J. Surg. 78 2020 185 193 32305533 OECD Economic Outlook, 2020. Editorial: after the lockdown, a tightrope walk toward recovery. Volume 2020, Issue 1. https://www.oecd-ilibrary.org/economics/oecd-economic-outlook/volume-2020/issue-1_0d1d1e2e-en. Pantano E. Pizzi G. Scarpi D. Dennis C. Competing during a pandemic? Retailers’ ups and downs during the COVID-19 outbreak J Bus Res 2020 Parker G.G. Van Alstyne M.W. Choudary S.P. Platform Revolution: How Networked Markets are Transforming the Economy? and How to Make them Work for You 2016 WW Norton and Company Puschmann T. Alt R. Sharing economy Business Inf. Syst. Eng. 58 1 2016 93 99 PwC, U.K., 2015. The Sharing Economy–Sizing the Revenue Opportunity. https://www.pwc.fr/fr/assets/files/pdf/2015/05/pwc_etude_sharing_economy.pdf. Rinne, A., 2020. Coronavirus: The end of the Sharing Economy, or a New Beginning? https://medium.com/swlh/coronavirus-the-end-of-the-sharingeconomy-or-a-new-beginning-a142acbb7130. Schor J. Debating the sharing economy Journal of Self-Governance and Management Economics 4 3 2016 7 22 Schor J.B. Attwood‐Charles W. The “sharing” economy: labor, inequality, and social connection on for‐profit platforms Sociology Compass 11 8 2017 e12493 p. Škare M. Soriano D.R. Porada-Rochoń M. Impact of COVID-19 on the travel and tourism industry Technol Forecast Soc Change 163 2021 120469 Stephany A. The Business of sharing: Making it in the New Sharing Economy 2015 Springer Sundararajan A. The Sharing economy: The end of Employment and the Rise of Crowd-Based Capitalism 2016 Mit Press Taeuscher K. Uncertainty kills the long tail: demand concentration in peer-to-peer marketplaces Electronic Markets 29 4 2019 649 660 World Bank Group, 2018. Tourism and The Sharing Economy: Policy and Potential of Sustainable Peer-To-Peer Accommodation. http://documents.worldbank.org/curated/en/161471537537641836/pdf/130054-REVISED-Tourism-and-the-Sharingeconomy-PDF.pdf. Zenker S. Kock F. The coronavirus pandemic–A critical discussion of a tourism research agenda Tourism Manage. 81 2020 104164 Zervas G. Proserpio D. Byers J.W. The rise of the sharing economy: estimating the impact of Airbnb on the hotel industry J. Market. Res. 54 5 2017 687 705

0

PMC9753187

NO-CC CODE

2022-12-16 23:25:32

no

Technol Forecast Soc Change. 2021 Jun 8; 167:120733

utf-8

Technol Forecast Soc Change

10.1016/j.techfore.2021.120733

oa_other

==== Front Technol Forecast Soc Change Technol Forecast Soc Change Technological Forecasting and Social Change 0040-1625 0040-1625 Elsevier Inc. S0040-1625(21)00142-6 10.1016/j.techfore.2021.120710 120710 Article Effects of the COVID-19 pandemic on the US stock market and uncertainty: A comparative assessment between the first and second waves Yousfi Mohamed a Ben Zaied Younes b Ben Cheikh Nidhaleddine c Ben Lahouel Béchir d⁎ Bouzgarrou Houssem e a IHEC, Sousse University, Tunisia b EDC Paris Business School, France c ESSCA Paris school of management, France d IPAG Business School Paris, France e ISG, Sousse University, Tunisia ⁎ Corresponding author. 27 2 2021 6 2021 27 2 2021 167 120710120710 31 10 2020 20 2 2021 23 2 2021 © 2021 Elsevier Inc. All rights reserved. 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. This paper makes the first comparative assessment of the impacts of the first and second waves of the ongoing COVID-19 pandemic for the US stock market and its uncertainty. To this end, we investigate the dynamic conditional correlation and the asymmetric impacts of shocks on the correlation between the US and Chinese stock markets before and during the COVID-19 crisis. Furthermore, we analyze and compare the relationship between the COVID-19 pandemic and US returns and uncertainty during the first and second waves of the pandemic. First, we find that the dynamic correlation approach supports the presence of volatility spillovers (contagion effects) between the two stock markets, especially during the rapid spread phase of COVID-19 in the US. Second, the analysis of news impact correlation surfaces shows that the shocks to the US and Chinese markets have asymmetric effects on the correlation between the two markets. Finally, we find a persistent link between US returns, uncertainty, and the COVID-19 pandemic during the first and second waves of the outbreak. Our results prove that the pandemic has shown harmful consequences for financial markets in general and the US economy in particular. Keywords COVID-19 pandemic Comparative assessment First wave Second wave US stock market Economic uncertainty ==== Body pmc1 Introduction The World Health Organization (WHO) declared the COVID-19 outbreak to be a global emergency on January 30, 2020. Nine months later, the COVID-19 pandemic had resulted in roughly 31.12 million confirmed cases and over 950,000 deaths (WHO, 2020). As a result, the governments of the world's largest countries have enforced border shutdowns, travel restrictions, and quarantine, thus sparking fear of an impending financial recession and economic crises. The COVID-19 pandemic has also affected the world economy through the shutdown of financial market indices. The ensuing financial crisis is detectable in the behaviors of different stock market indices. In the United States (US), for example, the S&P 500, DJI Average, and Nasdaq index fell dramatically until the government secured the Coronavirus Aid, Relief, and Economic Security (CARES) Act, at which point the three indices rose by 7.3%,1 7.73%,2 and 7.33%,3 respectively. Moreover, the 10-year US Treasury Bond Yields have dropped to 0.67%.4 Several research papers have recently explored and confirmed the dramatic impact of the COVID-19 crisis on financial markets (see, e.g., Albulescu, 2020; Zaremba et al., 2020; Zhang et al., 2020; Corbet et al., 2020; Akhtaruzzaman et al., 2020; Ashraf, 2020; Sharif et al., 2020; Goodell, 2020). The main purpose of this paper is twofold. First, we model volatility spillovers by examining the conditional correlations between the Chinese and US stock indices before and during the COVID-19 pandemic. Further, we examine the asymmetric impacts of shocks on the correlation between the two markets. To this end, we apply auto regressive moving average dynamic conditional correlation generalized autoregressive conditional heteroskedasticity (ARMA-DCC-GARCH(1,1)) and ARMA-asymmetric DCC-GARCH (ARMA-ADCC-GARCH(1,1)) models, which explicitly consider the leverage effect in financial markets. Second, we examine the persistence in the relationship between COVID-19 metrics (cases and deaths) and the US stock market and its uncertainty using the DCC process and the wavelet approach. This paper contributes to the literature in two ways. First, we shed light on the spillover risk between the Chinese and US stock markets from January 5, 2011 to September 21, 2020, which period includes the COVID-19 pandemic period. Second, we investigate the persistence of the linkage between the global cumulative daily confirmed infection cases and deaths generated by COVID-19 and the US stock market, as well as US uncertainty5 during the first and the second waves of the COVID-19 pandemic (i.e., between January 13, 2020 and September 21, 2020). Our findings suggest that the volatility spillover between the Chinese and US stock markets has been higher during the COVID-19 period compared to the pre-COVID-19 one. We also identify a long and persistent relationship between the US market and the global daily COVID-19 cases and deaths. This relationship includes various levels of US stock market uncertainty. Even after the general quarantine restrictions during the first COVID-19 wave were lifted, the correlation continued during the second wave. Consequently, our results confirm that the COVID-19 health crisis has had harmful consequences for financial markets and the macroeconomic conditions in the US. The pandemic's effect is especially pronounced with respect to increasing economic uncertainty. The main findings of this paper are thus confirming recent empirical results (Kurita and Managi, 2020; Katafuchi et al., 2020; Yoo and Managi, 2020). The rest of the paper is organized as follows. In Section 2, we briefly review the related literature on the socioeconomic impacts of the COVID-19 crisis. Section 3 presents the empirical methodology and the data. We then discuss the results in Section 4. Finally, the robustness check of the main empirical results are presented in Section 5, while Section 6 concludes the paper. 2 Related literature: empirical findings Focusing on recent publications on the financial and socioeconomic impacts of the COVID-19 pandemic, we review the related literature from two perspectives: (1) the transmission volatility during the COVID-19 crisis and (2) the relationship of COVID-19 metrics with stock market performance and economic uncertainty. Several recent studies investigate the financial and socioeconomic impacts of the COVID-19 crisis. For instance, Akhtaruzzaman et al. (2020) use the DCC approach to examine contagion transmission for both financial and nonfinancial firms between China and the G7 economies during the COVID-19 pandemic. They find that both financial and non-financial firms experience a significant increase in the conditional correlations between their stock returns. They also argue that the magnitude of the increase in these correlations is considerably higher for financial firms, indicating their important role in financial contagion transmission between China and the G7 countries. Finally, they show that the optimal hedge ratios increase significantly in most cases, which imply higher hedging costs during the COVID-19 crisis period. Corbet et al. (2020) indicate that, at the beginning of the 2020 COVID-19 pandemic, it came as no surprise that the Chinese markets acted as the epicenter of both physical and financial contagion. Furthermore, Corbet et al. (2021) show the existence of sharp, dynamic, and new correlations related to the term “corona.” Zhang et al. (2020) conclude that the rapid spread of COVID-19 has had dramatic impacts on financial markets worldwide, thus leading to a significant increase in global financial market risk and causing investors to suffer significant losses over a short period of time. Concerning the impact of the pandemic on economic uncertainty, several studies show that the COVID-19 crisis has been characterized by high uncertainty levels affecting all major economies. For instance, Ashraf (2020) uses daily the number of COVID-19 confirmed cases and deaths and panel data analysis techniques to examine the stock market responses to the COVID-19 crisis. He shows that stock markets responded negatively to the growth in the number of confirmed cases, as stock market returns declined as the number of confirmed cases increased. Furthermore, stock markets reacted more proactively to the growth in the number of confirmed cases than to the one in the number of deaths. Based on a wavelet-based approach, Sharif et al. (2020) use daily COVID-19 observations (i.e., number of cases in the US) to investigate the relationship between the recent spread of COVID-19, oil price volatility shocks, the stock market, geopolitical risks, and economic policy uncertainty in the US. They show that the effect of the COVID-19 pandemic on geopolitical risks is substantially higher compared to its impacts on US economic uncertainty. Additionally, Goodell (2020) highlights the possible impacts of COVID-19 on financial markets and institutions and Zaremba et al. (2020) explore the policy responses to the pandemic in 67 countries, demonstrating that non-pharmaceutical interventions significantly increased equity market volatility. Albulescu (2020) uses ordinary least squares (OLS) regression to examine the impact of official announcements of COVID-19 new cases and the fatality ratio on the volatility of the US financial market. He considers both the global and US COVID-19 metrics, and demonstrate that the pandemic enhanced the realized volatility of the S&P 500 index. He also suggests that the prolonged nature of the COVID-19 pandemic is an important source of financial volatility and thus presents a challenge for risk management. This brief literature review focuses on recent developments in the research on the socioeconomic impacts of the COVID-19 pandemic. The following section develops the empirical methodology used to investigate how these impacts manifest in financial markets and in terms of economic uncertainty. 3 Empirical methodology 3.1 Empirical model Several empirical methods have been used to investigate the risk spillovers and estimate correlations between stock market returns. In this paper, we use the multivariate GARCH to model volatility and construct dynamic conditional correlations based on a rolling-window analysis. We chose to use restricted correlation models, such as DCC and ADCC due to their comparative advantages. Namely, they are designed to solve the problems encountered when using the Baba, Engle, Kraft, and Kroner (BEKK) and VECH models due to the presence of a large number of free parameters. The DCC and ADCC models are easier to estimate and are comparatively more robust. They are thus the most appropriate for examining the time-varying correlations between financial products and economic variables (Ciner et al., 2013). Let rt be an n × 1 vector of the sample's asset returns and ARMA(1,1) a process in the mean equation for Rt, where εt is conditional on the set of information and residuals, It−1. The equation can be written as:(1) rt=u+AR1rt−1+MA1εt−1+εt, and the residuals can be modeled as:(2) εt=Ht1/2zt. Ht is the conditional covariance matrix of Rt and zt is an n × 1 identically and independently distributed vector of random errors. In the first step, we estimate the GARCH parameters and then we estimate the dynamic conditional correlations in the second step:(3) Ht=DtRtDt. Ht is the conditional covariance matrix, n × n, and Rt is the conditional correlation matrix.Dt is a diagonal matrix with time-varying standard deviations on the diagonal. Rt and Dt are determined as:(4) Rt=diag(h1.t1/2…..hn.t1/2), (5) Dt=diag(q1.t−1/2….qn.t−1/2)Qtdiag(q1.t−1/2…qn.t−1/2), where h is the expression of the univariate GARCH models, which are used to derive the expression of h on the diagonal matrix (where H is a diagonal matrix). The GARCH(1,1) parameters of Ht can be expressed by:(6) hit=ωi+αiεit−12+βihit−1. Qt is a symmetric positive definite matrix, and can be written as follows:(7) Qt=(1−a−b)Q¯+azt−1zt−1′+bQt−1. Qis the n × n unconditional correlation matrix of the standardized residuals, zit(zit=εit/hit). Parametersa and b are associated with the smoothing process and are used to construct DCCs. The DCC model means return to equilibrium if a+b is less than unity (a+b<1) and positive. The correlation is estimated as follow:(8) ρi.j.t=qi.j.tqi.i.tqj.j.t. We use an ADCC model because the DCC model fails to capture asymmetry effects:(9) hi.t=ωi+αiεi.t−12+βihi.t−1+λiεi.t−12I(εi.t−1). The indicator function, I(εi.t−1), is equal to 1 if εi.t−1<0, and 0 otherwise. The dynamics of Qfor the ADCC model are given as:(10) Qt=(Q¯A′Q¯A−B′Q¯B−G′Q−¯G)+A′zt−1z′t−1A+B′Qt−1B+G′zt−zt′−G, where A, B, and G are n × n parameter matrices and zt−is a vector of zero-threshold standardized errors, which are equal to zt when below 0, and 0 otherwise. Q¯ and Q−¯ are the unconditional matrices of zt and zt−, respectively. To test these models and their relevance for our research question, we construct a daily dataset collected from different sources. The next sub-section presents the data and describes their proprieties. 3.2 Data To perform empirical investigations, we used daily data from the US S&P 500 index, which measures the stock performance of the 500 large companies traded on the exchange, and on Chinese stocks from the CSI 300 index. We chose the CSI 300 index because it is a capitalization-weighted stock market index designed to replicate the performance of the top 300 stocks traded on the Shanghai and Shenzhen stock markets. It has also been frequently and commonly used as a representative index to measure the overall performance of the Chinese stock market (Chen et al., 2013). The CSI 300 prices are denominated in CNY while the S&P 500 ones are denominated in USD. We collected daily data for the S&P 500 index and the CSI 300 index from January 5, 2011 to September 21, 2020, meaning our sample period covers the COVID-19 crisis as well. The choice of the starting date and the analysis period is justified by our research objective and supported by the availability of the data for both indices. Further, we assembled another dataset on the cumulative cases of COVID-19 infections and deaths, US stock market performance (S&P 500), VIX, and EPU. The VIX measures the volatility of the stock market (S&P 500) and represents the stock market's expectations of volatility over the next 30 days. Higher VIX values represent more uncertainty or fear in the market, while lower values indicate less market uncertainty. The EPU is a new measure of uncertainty, developed by Baker et al. (2020), which is based on the frequency of newspaper references to the number and size of federal tax code provisions set to expire in future years; it also captures the disagreement among economic forecasters about policy relevant variables and economic policy uncertainty. The S&P 500 index data were collected from the Federal Reserve Bank of St. Louis and the CSI data from Yahoo Finance. The VIX was also obtained from Yahoo Finance. The US EPU index can be freely downloaded from the EPU website.6 Finally, we obtained data on the global cumulative daily COVID-19 cases and deaths from the WHO. To perform a comparative assessment of the socioeconomic impacts between the first and second COVID-19 waves, we used a sample from January 13, 2020 to September 21, 2020. The different data series show different patterns as follows (see Fig. 1 ). The US and Chinese stock markets experienced a significant drop during the COVID-19 pandemic. By contrast, the two measures of US economic uncertainty (VIX and EPU) show a sharp rise around the start of the COVID-19 pandemic. The higher values represent more uncertainty or fear for both the S&P 500 and the overall economic situation, their rise coinciding with the rising numbers of COVID-19 infection cases and deaths globally.Fig. 1 Time series plots of daily series (financial markets indices, VIX, cases, and deaths). Fig 1 Using these daily series, we calculated the daily returns as follows:100×ln(pt/pt−1), where pt is the daily closing price or settlement. The descriptive statistics for the S&P 500 and CSI 300 series are presented in Table 1 .Table 1 Descriptive statistics of the S&P 500 and CSI 300 daily data. Table 1 S&P 500 CSI 300 No. obs. 2292 2292 Min −12.76522 −9.15444 Max 8.96832 7.42630 Range 21.73354 16.58074 Median 0.06903 0.02373 Mean 0.04119 0.01703 S.E. mean 0.02360 0.03075 Var. 1.27646 2.16792 Std. dev. 1.12980 1.47239 Coef. var. 27.43135 86.48180 JB 31,000 2700 Prob. < 2e-16 < 2e-16 ARCH (12) 910 260 Prob. < 2e-16 < 2e-16 Note: S.E., Var, Coef. var., and Std. dev., stand for standard error, variance, coefficient of variance, and standard deviation. JB is the Jarque-Bera test with the null hypothesis of normality. ARCH is the autoregressive heteroskedasticity test. < 2e-16 indicates small p-values and the rejection of a null hypothesis at the 1% significance level. The means of both stock returns (China and the US) both before and during the COVID-19 pandemic represent positive daily returns. The coefficients of variation indicate that the CSI 300 has a higher variability and the S&P 500 a lower variability. The S&P 500 also has a lower standard deviation. The Jarque-Bera and ARCH-LM test reveal normality and heteroskedasticity issues. The Jarque-Bera test reveals that each series is far from being normally distributed. The ARCH-LM(12) test show strong evidence of ARCH effects, meaning that all series exhibit strong clustering. Table 2 shows the descriptive statistics of the S&P 500, US EPU, VIX, and world cumulative daily COVID-19 cases and deaths for the first and second waves.Table 2 Descriptive statistics of daily data S&P 500, US EPU, VIX and COVID-19 cases/deaths. Table 2 First wave of COVID-19 S&P 500 EPU VIX Cases Deaths Min −12.76522 −77.8809 −26.6228 0.000 0.000 Max 8.96832 144.8161 38.2167 178.161 113.635 Range 21.73354 222.6970 64.8394 178.161 113.635 Median 0.18418 −1.1564 −1.2072 5.585 5.907 Mean −0.01693 0.9279 0.7322 11.546 12.775 S.E. mean 0.31413 3.1539 1.1193 2.217 2.060 Var 9.96637 1004.6517 126.5385 496.431 428.403 Std. dev. 3.15696 31.6962 11.2489 22.281 20.698 Coef. var. −186.49388 34.1587 15.3629 1.930 1.620 JB 54 61 45 5200 830 Prob. 2e-12 5e-14 2e-10 < 2e-16 < 2e-16 ARCH (12) 37 13 23 75 67 Prob. 2e-04 0.4 0.03 4e-11 1e-09 Second wave of COVID-19 Min −6.0753 −63.1028 −12.2421 0.8317 0.45687 Max 3.1015 58.6848 39.1709 7.3858 5.85851 Range 9.1768 121.7876 51.4130 6.5541 5.40164 Median 0.3590 0.5865 −1.1041 1.8011 1.11765 Mean 0.1348 −0.8796 −0.1874 2.3205 1.48414 S.E. mean 0.1338 2.6225 0.7409 0.1555 0.09679 Var. 1.8088 694.6205 55.4387 2.4437 0.94613 Std. dev. 1.3449 26.3557 7.4457 1.5632 0.97269 Coef. var. 9.9763 −29.9622 −39.7228 0.6737 0.65539 JB 92 0.71 280 73 120 Prob. < 2e-16 0.7 < 2e-16 < 2e-16 < 2e-16 ARCH (12) 6 11 1.2 53 61 Prob. 0.9 0.6 1 5e-07 2e-08 Note: S.E., Var., Coef. var., and Std. dev. stand for standard errors, variance, coefficient of variance, and standard deviation, respectively. JB is the Jarque-Bera test with the null hypothesis of normality. ARCH is the autoregressive heteroskedasticity test. < 2e-16 indicate the rejection of the null hypothesis at the 1% significance level. The time series graphs of returns squared (Fig. 2 ) show how volatility has changed for the S&P 500 and CSI 300 over time. Each series displays several periods of volatility clustering in the pre-COVID-19 period and during the COVID-19 pandemic. The effect is most pronounced for the S&P 500 during the pandemic.Fig. 2 Squared daily data plots. Fig 2 4 Discussion of the empirical results Our modeling strategy is to first estimate four different versions of the DCC and ADCC models. Each version includes a constant in the mean equation and a GARCH(1,1) variance equation. Adjustments were made to include an ARMA(1,1) (AR(1)andMA(1)) term in the mean equation and the distribution choice. The model selection criteria indicate that the best fit is the first version (A) of the DCC and ADCC models with the AR (1) and MA (1) terms in the mean equation estimated with a multivariate t distribution (see Table 3 ) for the S&P 500 and CSI 300. Consequently, both models are estimated with the AR(1) and MA(1) terms in the mean equation. To address the non-normality in the distribution of returns, the DCC and ADCC models are estimated with a multivariate t distribution.Table 3 Different specifications of the DCC and ADCC models. Table 3 DCC A B C D ARMA(1,1) yes no yes no Distribution MVT MVT MV NOR MV NOR No. obs. 2292 2292 2292 2292 Akaike 5.6917 5.7011 5.8394 5.8500 Bayes 5.7468 5.7362 5.8870 5.8776 Shibata 5.6915 5.7010 5.8393 5.8500 Hannan-Quinn 5.7118 5.7139 5.8568 5.8601 Likelihood −6501 −6519 −6673 −6693 ADCC A B C D ARMA(1,1) yes no yes no Distribution MVT MVT MV NOR MV NOR No. obs. 2292 2292 2292 2292 Akaike 5.6674 5.6688 5.8158 5.8217 Bayes 5.7300 5.7114 5.8709 5.8567 Shibata 5.6672 5.6687 5.8156 5.8216 Hannan-Quinn 5.6903 5.6843 5.8359 5.8344 Likelihood −6470 −6479 −6643 −6658 Note: This table presents the diagnostic statistics for each type of DCC and ADCC specification. We model the ARMA-DCC-ADCC-GARCH process between the US (S&P 500) and Chinese stock markets (CSI 300). The results are presented in Table 4 . The lag order (1,1) is chosen by minimized information criteria (including the Akaike and the Schwarz information criteria) (Table 3). The results of the mean equation show that the coefficients on all return series are significant at the 10% level.Table 4 Estimation results for the ARMA-DCC-ADCC-GARCH process. Table 4 DCC ADCC Coef. S.E. t-stat Prob Coef. S.E. t-stat Prob uS&P500 0.076282 0.005486 1.390e+01 0.00000 0.074262 0.012511 5.935825 0.000000 ARS&P500 0.959602 0.008604 1.115e+02 0.00000 0.763577 0.114258 6.682928 0.000000 MAS&P500 −1.034701 0.000011 −9.17e+04 0.00000 −0.819141 0.116016 −7.060620 0.000000 ωS&P500 0.027871 0.007215 3.860e+00 0.00012 0.030567 0.006465 4.727788 0.000002 αS&P500 0.185074 0.028642 6.461e+00 0.00000 0.000436 0.017796 0.024515 0.980442 βS&P500 0.808261 0.025064 3.228e+01 0.00000 0.824524 0.024022 34.323514 0.000000 γS&P500 0.294672 0.051119 5.764436 0.000000 λS&P500 4.497313 0.352616 1.275e+01 0.00000 4.709163 0.411914 11.432406 0.000000 uCSI300 0.041053 0.021508 1.907e+00 0.05600 0.038547 0.021795 1.768637 0.076954 ARCSI300 −0.851760 0.086278 −9.83e+00 0.00000 −0.85170 0.086537 −9.834702 0.000000 MACSI300 0.849592 0.087669 9.690e+00 0.00000 0.850468 0.087846 9.681397 0.000000 ωCSI300 0.013565 0.005676 2.390e+00 0.01648 0.015339 0.007077 2.167408 0.030204 αCSI300 0.057572 0.010050 5.728e+00 0.00000 0.051603 0.010820 4.769331 0.000002 βCSI300 0.939849 0.009696 9.693e+01 0.00000 0.937363 0.011737 79.861534 0.000000 γCSI300 0.015100 0.018404 0.820452 0.411958 λCSI300 4.440714 0.365139 1.216e+01 0.00000 4.407926 0.363746 12.118146 0.000000 a 0.002567 0.001322 1.942e+00 0.05207 0.012249 0.014017 0.873901 0.382172 b 0.996613 0.002131 4.675e+02 0.00000 0.933156 0.166451 5.606178 0.000000 c 0.000000 0.016984 0.000009 0.999993 λ 5.123312 0.297466 1.723e+01 0.00000 5.260955 0.329326 15.974936 0.000000 Akaike 5.6917 5.6674 Bayes 5.7468 5.7300 Shibata 5.6915 5.6672 H-Q 5.7118 5.6903 LL −6501 −6470 Notes: DCC and ADCC are estimated using a multivariate normal (MVNORM) distribution. All specifications include a constant and AR(1) and MA(1) terms in the mean equation. The short-term persistence (α) is statistically significant for most variables under the two models. The estimated coefficient on long-term persistence (β) is statistically significant for each series, thus indicating the importance of long-term persistence. The sum of the coefficients for short-term and long-term persistence is less than unit. In each case, the short-term persistence is lower than the long-term one, which indicates that long-term volatility is more intense than short-term volatility. The statistical significance of short- and long-term persistence provides evidence of volatility clustering. We can also see volatility clustering for all variables in Fig. 2. The estimated asymmetric term (γ) is positive and statistically significant for the S&P 500. This means that the negative residuals for the S&P 500 tend to increase variance (conditional volatility) more than positive shocks of the same magnitude, while there is no statistically significant leverage effect for the Chinese market. We then estimated the dynamic conditional correlation coefficients and the results are presented in Table 4. The estimated coefficients, a and b, are positive and statistically significant at the 1% level in each of the two models (except for a in the ADCC model). Their sum is below unity, which indicates that the DCCs return to equilibrium (i.e., are mean-reverting). We can conclude that the DCC models are reasonable and that the volatility of recent returns has a significant influence on the dynamic relationship between the S&P 500 stock market and all variables, as indicated by the considerable value of a. Nonetheless, the values of b are all significant and close to 1 for each series, indicating that the dynamic relationships between the equity market and all other variables are long-term persistent. Our results confirm the long-term relationship between the US and Chinese stock markets. For parameter Shape (λ), which represents the degrees of freedom, the S&P 500 has the highest estimated value. This means that the distribution of CSI 300 stocks has larger tails than the S&P 500 distribution. Shape is equal to the degrees of freedom when the number of degrees of freedom approaches infinity and the form of distribution t approaches that of a normal distribution. The information criteria show that the ADCC is the best fitting model. 4.1 Analysis of dynamic conditional correlations To construct dynamic conditional correlations between the US and Chinese stock markets that are one step ahead, we use rolling window analysis. The estimation window is fixed at 2292 observations and 1000 dynamic conditional correlations one step ahead are also produced. GARCH models are refitted every 20 observations. Considering that the relationship between the US and China changes over time, we explore the time-varying dynamic conditional correlation of the market pair before and during the COVID-19 pandemic. The results are presented in Fig. 3 .Fig. 3 Rolling one-step-ahead dynamic conditional correlations. Fig 3 For comparison, we considered for the S&P 500/CSI 300 pair that the time-varying conditional correlations obtained from the DCC and ADCC models exhibit similar patterns. The results of the time-varying conditional correlations show that the conditional correlation between the two markets fluctuates greatly during the analysis period, meaning that investors adjust their portfolio structures frequently. That the dynamic conditional correlation among the market pair is positive and supports the presence of contagion effects, especially at the beginning of the third quarter of 2020 during the rapid increase in the number of COVID-19 cases in the US. Consequently, during the COVID-19 crisis, the risk spillover between the Chinese and US stock markets was strong. From Fig. 3, the dynamic correlations between the two markets under both models are higher during the COVID-19 period compared to the pre-COVID-19 one. In sum, the correlation between the S&P 500 and CSI 300 stock markets is time-varying and highly volatile, suggesting portfolio managers should change their portfolio structures over time. Moreover, we analyze the news impact correlation surfaces for each pair (see Fig. 4 ). We find that the news impact correlation surfaces of the DCC and ADCC models have similar shapes. For the S&P 500/CSI 300 pair, the results of the news impact correlation surfaces of the DCC or ADCC models show that the shocks to each stock market have asymmetric effects on the correlation between US and Chinese markets. Moreover, the shape of the news impact correlation surfaces produced from the DCC and ADCC models for each pair are convex.Fig. 4 News impact correlation surfaces between S&P 500 and CSI 300. Fig 4 4.2 Persistence of linkage between the S&P 500, EPU, and VIX during the first and second COVID-19 waves To produce a comparative assessment of the contagion effect between the second and first waves of the COVID-19 pandemic, we apply the DCC process to estimate the dynamic conditional correlation coefficients between the US stock market, the US EPU index, the VIX, and global cumulative daily COVID-19 cases and deaths (see Table 5 ). We estimate and compare the persistence of the correlation between the pair during the first and second waves of the COVID-19 pandemic.Table 5 DCC parameters between the US stock market, US uncertainty, and the COVID-19 pandemic. Table 5First wave of the COVID-19 pandemic S&P 500/COVID-19 VIX/COVID-19 US EPU/COVID-19 Coef. Prob. Coef. Prob. Coef. Prob. a 0.144328 0.000004 0.144267 0.000016 0.161976 0.000001 b 0.855672 0.000000 0.855733 0.000000 0.838024 0.000000 Second wave of the COVID-19 pandemic S&P 500/COVID-19 VIX/COVID-19 US EPU/COVID-19 Coef. Prob. Coef. Prob. Coef. Prob. a 0.050281 0.000056 0.049712 0.000000 0.046179 0.000000 b 0.939617 0.000000 0.941952 0.000000 0.925436 0.000000 During the first and second waves of the COVID-19 pandemic, the estimated coefficients, aand b, are positive and statistically significant at the 1% level. Their sum is below 1, which indicates that the DCCs return to equilibrium (i.e., are mean-reverting). Consequently, we can conclude that the DCC models are reasonable and that the volatility of recent returns has a significant influence on the dynamic relationship between the S&P 500 stock market, the VIX, the US EPU index, and the global cumulative daily COVID-19 cases and deaths, which is indicated by the considerable value of coefficient a. Nonetheless, the values of coefficient b are all significant and close to 1 for each pair, indicating that the dynamic linkages between the US stock market, the two measures of US uncertainty, and the COVID-19 pandemic are long-term persistent. Finally, our results confirm the long-term relationship between the US stock market, the VIX, the US EPU. and the global cumulative daily COVID-19 cases and deaths during the first and second waves of the pandemic. Even after the general quarantine restrictions were eased after the first wave, the correlation continued to persist during the second wave. We can conclude that the continued rise in COVID-19 infections and deaths during the first and second waves increased the uncertainty on the US stock market and the economic, with serious financial consequences. Our results confirm the recently published findings of Zhang et al. (2020). 5 Robustness check: wavelet analysis We tested the robustness of the empirical results using the wavelet coherence method, namely focusing on the correlation between the financial and economic variables and the COVID-19 pandemic variables (i.e., total cases and deaths). The wavelet coherence technique identifies specific parts of the domain of time-frequency (see Grinsted et al., 2004), where unexpected and major variations happen in the co-movement patterns of the time series under observation. The equation of the coefficient of the adjusted wavelet coherence, as identified by Torrence and Compo (1998), is expressed as follows:W2(p,q)=|M(M−1Nab(p,q)|2M(M−1|Na(p,q)|2M(M−1|Nb(p,q)|2. M is the smoothing mechanism. The value of the wavelet squared coherence ranges from 0 to 1 (0 ≤ W2(p,q)≤ 1).The range of the squared wavelet coherence coefficient shows the correlation degree. Indeed, when this coefficient is close to 0, it indicates the absence of correlation (no co-movement), and when it is close to 1, it indicates a higher correlation (higher co-movement) that can be considered a scale-specific squared correlation between series. The Monte Carlo method is used to examine the hypothetical allocation of the wavelet coherence. This approach allows us to examine the lead/lag relationship between two series, while avoiding the issue of the squared coherence not being able to distinguish the positive and negative relationships between two series. We applied wavelet coherence to test the correlation and interdependence between daily COVID-19 cases and deaths and US stock (S&P 500) and US uncertainty (VIX and EPU indexes). We estimated and compare the co-movement between each couple of variables during the first and second waves of the-19 pandemic and the wavelet coherence plots are presented in Fig. 5 . The plots represent the estimated wavelet coherence method and the relative phasing of the two series are shown by arrows. The black contour of the estimated plots is the 5% significance level, the red/warm colored region is the one with strong co-movement, whereas the blue/cold colored area represents regions with weak co-movements. The direction of arrows is the direction of the interdependence and causality relationships (Torrence and Webster, 1999; Tiwari, 2013; Yang et al., 2017; Pal and Mitra, 2019; Jiang and Yoon, 2020). When the arrow points to the right, the two variables are positively correlated and when it points to the left, the two variables are negatively correlated. The up-right and down-left (↗↙) arrows mean that the first variable leads the second, whereas the down-right and the up-left (↘↖) arrows indicate that the second variable leads the first. The up (↑) and down (↓) arrows imply that the variable is leading and lagging, respectively.Fig. 5 Wavelet coherence plots, couple-wise estimates during the first and second waves of the COVID-19 pandemic. Fig 5 We start by analyzing the wavelet coherence between each couple of variables during the first wave of COVID-19 outbreak and identify many significant high degrees of co-movement. From Fig. 5, we detect the existence of many small islands that indicate a strong dependence at the beginning, the middle, and the end of the first wave period over short-run frequency bands. The confirmed cases and S&P 500 couple shows strong dependence at the beginning of the first wave period over the 0–4-day frequency bands. The direction of the arrows is up-right (↗) which means that confirmed cases and the US stock market are positively correlated. The plots show another island of higher co-movement between this couple of variables at the end of the wave over the short run. Additionally, in observing the co-movement between cases/deaths and S&P 500, we detect many small islands of a higher degree of dependence. COVID-19 (confirmed cases and deaths) and VIX show several areas of higher coherence. Further, we detect a negative coherence, especially for the first couple at the beginning and the end of the wave over the short run (0–4-day frequency bands). The direction of the arrows is to the left, where the first island shows that confirmed cases lead the VIX (down-left ↙), while the second island shows the existence of an anti-cyclic relationship between infected cases and the VIX. We identify a positive dependence (up-right ↗) between deaths/cases and the VIX over the short run (2-day frequency). Finally, we detect many areas of a higher degree of positive and negative dependence between COVID-19 and the US EPU over the short run (0–2-day frequency bands) and long-run (12 to more than 16-day frequency bands). The direction of the arrows is up-right (↗), indicating that COVID-19 leads the US EPU, while the up-left arrows (↖) mean that there is an anti-cyclic effect between COVID-19 and US EPU. During the second wave of the COVID-19 outbreak, as per Fig. 5, there is a more meaningful and significant degree of coherence between the COVID-19 pandemic and the US stock market and US uncertainty than during the first wave. We identify many large areas with high significant dependence over the short and long run during the second wave. For the COVID-19 and S&P 500 stock, the direction of the arrows is mostly up-right (↗), which means that this couple of variables is positively correlated, and COVID-19 is leading the co-movement. The direction of the arrows for the higher coherence island between COVID-19 and the VIX shows a positive and negative dependence over the short and long run and they point up-right and down-left (↗↙) and down-right and up-left (↘↖), which indicates that COVID-19 leads the VIX and there is an anti-cyclic effect between them. Finally, for the last two couple of variables, the direction of the arrows for strong co-movement areas is down-right and up-left (↘↖), which reveals an anti-cyclic effect between COVID-19 and the US EPU. Our findings thus imply a higher co-movement between the COVID-19 pandemic and the US stock market and uncertainty during the first and second waves of the outbreak, while the degree of dependency is more pronounced during the second wave. The results of the wavelet coherence approach confirm and support the results in Table 5, which reveal the persistence of a linkage between the COVID-19 pandemic and the S&P 500, VIX, and EPU during the first and second waves of the pandemic. Therefore, using two different approaches, we tested and confirmed the robustness of our empirical findings. 6 Conclusions First, our study sheds light on the effects of the COVID-19 pandemic on the risk spillover between the Chinese (CSI 300) and US (S&P 500) stock markets before and during the COVID-19 crisis period. from January 5, 2011 to September 21, 2020. We also studied the asymmetric effects of shocks on the correlation between the two markets. Second, we investigated the linkage between the S&P 500 stock market, as well as two US uncertainty indices (VIX and EPU), and the global cumulative daily COVID-19 cases and deaths during the first and second waves, from January 13, 2020 to September 21, 2020. To achieve our objectives, we applied multivariate GARCH models (DCC-GARCH and ADCC-GARCH), DCC process, and wavelet coherence. We found that the dynamic conditional correlations support the presence of contagion effects, especially during the rapid spread of COVID-19 in the US. The volatility spillover between the Chinese and US markets was higher during the COVID-19 pandemic than before it. The results of the news impact correlations show that the shocks in both markets have asymmetric impacts on the relationship between the US and China stock markets during our sample period, including during the COVID-19 pandemic. We confirm the long-term relationship between the US stock market, the VIX, the US EPU, and the global cumulative daily COVID-19 cases and deaths during the first wave of the pandemic. Even after the general quarantine restrictions were eased following the first COVID-19 wave, the persistence of the correlation continued during the second wave. We confirmed these empirical results using the wavelet coherence methodology and showed that the continued increase in COVID-19 infections and deaths during the first and second waves increased the uncertainty of the US stock market and the overall economy, with serious financial consequences. Authors' contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by [Mohamed Youssfi], [Younes Ben Zaied], [Nidhaleddine Ben Cheikh], [Béchir Ben Lahouel], and [Houssem Bouzgarrou]. The current revised version of the manuscript was written and approved by all authors. Funding Not applicable Availability of data and material Not applicable Code availability Not applicable Declaration of Competing Interest Not applicable. Mohamed YOUSSFI is currently a PhD candidate at the IHEC, University of Sousse Tunisia. His-research interests are econometric modeling, quantitative finance, and economic forecasting. Younes BEN ZAIED is a professor of Finance at EDC Paris Business School. His-research interests include water demand management, environmental economics and policy, demand modeling for water, climate change impacts on agriculture and finance. His-work has appeared in Technological Forecasting and Social Change, Applied Economics, Annals of Operations Research, Journal of International Money and Finance, Finance Research Letters, Bankers, Markets & Investors, Climatic Change, Economics Bulletin, Environmental Modeling & Assessment, Environmental Economics, Journal of Quantitative Economics, etc. Nidhaleddine BEN CHEIKH is a professor of Economics at ESSCA Paris school of management. His-research interests are: the study of the non-linear dynamics of exchange rate shocks, the impact of climate change on water resources and agricultural production, and the impact of oil price fluctuations linked to geopolitical tensions on Middle Eastern countries. In recent years he published several papers on refereed journals including Annals of Operations Research, Journal of International Money and Finance, Finance Research Letters, Bankers, Markets & Investors, Climatic Change, Environmental Modeling & Assessment, Environmental Economics, Economics Bulletin, etc. Béchir BEN LAHOUEL is an associate professor of Management sciences at IPAG Business School Paris. His-research revolves primarily around the environmental management grounded in economic theory, input-output modeling, and social, human and environmental impact analysis. He published his work in refereed academic journals such as Technological Forecasting and Social Change, Journal of Cleaner Production, Finance Research Letters, Bankers, Markets and Investors, Environmental Economics and Policy Studies, Journal of Organizational Change Management, European Business Review, Corporate Governance, Question(s) de management, etc. Houssem BOUZGARROU is an associate professor at the Institut Supérier de Gestion, University of Sousse Tunisia. His-research interests include the analysis of stocks volatility and performance, the economic growth before and after the financial crisis, the dynamic dependence between US indices and commodities prices. He published his work in refereed academic journals such as International Review of Financial Analysis, Research in International Business and Finance, International Journal of Monetary Economics and Finance, Bankers, Markets & Investors, Environmental economics, Journal of Asset Management. Appendix (Table A1 )Table A1 Unit root tests. Table A1ADF test Level S&P 500 CSI 300 EPUUS VIX With constant t-Statistic Prob. −0.5786 0.8728 −1.1273 0.7071 −2.3883 0.1466 −2.3275 0.1646 With constant & trend t-Statistic Prob. −4.9020 0.0003 *** −2.5823 0.2886 −2.2894 0.4370 −2.4781 0.3386 Without constant & trend t-Statistic Prob. 1.4614 0.9648 0.3655 0.7900 −0.9107 0.3205 −0.6460 0.4361 First Difference d(S&P 500) d(CSI 300) d(EPUUS) d(VIX) With constant t-Statistic Prob. −14.8412 0.0000 *** −45.7654 0.0001 *** −22.0476 0.0000 *** −5.8235 0.0000 *** With constant & trend t-Statistic Prob. −14.8441 0.0000 *** −45.7737 0.0000 *** −22.0308 0.0000 *** −5.8589 0.0000 *** Without constant & trend t-Statistic Prob. −14.7400 0.0000 *** −45.7679 0.0001 *** −22.1117 0.0000 *** −5.8335 0.0000 *** PP test Level S&P 500 CSI 300 EPUUS VIX With constant t-Statistic Prob. −0.5381 0.8813 −1.2645 0.6481 −3.3649 0.0136 ** −2.0998 0.2451 With constant & trend t-Statistic Prob. −4.8681 0.0003 *** −2.7379 0.2212 −3.3171 0.0670 * −2.2076 0.4819 Without constant & trend t-Statistic Prob. 1.5614 0.9714 0.3104 0.7754 −1.0202 0.2759 −0.5347 0.4840 First Difference d(S&P 500) d(CSI 300) d(EPUUS) d(VIX) With constant t-Statistic Prob. −57.6733 0.0001 *** −45.7847 0.0001 *** −27.2469 0.0000 *** −18.0532 0.0000 *** With constant & trend t-Statistic Prob. −57.6693 0.0000 *** −45.7923 0.0000 *** −28.7578 0.0000 *** −18.3686 0.0000 *** Without constant & trend t-Statistic Prob. −57.6219 0.0001 *** −45.7896 0.0001 *** −27.2674 0.0000 *** −18.0897 0.0000 *** Notes: *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively; (no) means not significant. This article belong to the special section on Social-Economic Impacts of Epidemic Diseases. 1 S&P 500 index: https://www.marketwatch.com/investing/index/spx 2 Dow Jones industrial average: https://www.marketwatch.com/investing/index/djia 3 NASDAQ composite index: https://www.marketwatch.com/investing/index/comp 4 10 year Treasury rate: https://ycharts.com/indicators/10_year_treasury_rate 5 We use the implied volatility index (VIX) and US Economic Policy Uncertainty Index (EPU). 6 See http://www.policyuncertainty.com ==== Refs References Akhtaruzzaman M. Boubaker S. Sensoy A. Financial contagion during COVID-19 crisis Finance Res. Lett. 38 2020 101604 Albulescu C.T. COVID-19 and the United States financial markets’ volatility Finance Res. Lett. 38 2020 101699 Ashraf B.N. Stock markets’ reaction to COVID-19: cases or fatalities? Res. Int. Bus. Finance 54 2020 101249 Baker S.R. Farrokhnia R.A. Meyer S. Pagel M. Yannelis C. How does household spending respond to an epidemic? Consumption During the 2020 COVID-19 Pandemic 2020 National Bureau of Economic Research Working paper w26949 Chen H. Han Q. Li Y. Wu K. Does index futures trading reduce volatility in the Chinese stock market? A panel data evaluation approach J. Futures Mark. 33 12 2013 1167 1190 Ciner C. Gurdgiev C. Lucey B. Hedges and safe havens: an examination of stocks, bonds, gold, oil and exchange rates Internat. Rev. Finan. Anal. 29 2013 202 211 Corbet S. Larkin C. Lucey B. The contagion effects of the COVID-19 pandemic: evidence from gold and cryptocurrencies Finance Res. Lett. 35 2020 101554 Corbet S. Hou Y. HU Y. Lucey B. Oxley L. Aye Corona! The contagion effects of being named Corona during the COVID-19 pandemic Finance Res. Lett. 38 2021 101591 Goodell J.W. COVID-19 and finance: agendas for future research Finance Res. Lett. 35 2020 101512 Grinsted A. Moore J.C. Jevrejeva S. Application of the cross wavelet transform and wavelet coherence to geophysical time series Nonlinear Process. Geophys. 11 5/6 2004 561 566 Jiang Z. Yoon S.M. Dynamic co-movement between oil and stock markets in oil-importing and oil-exporting countries: two types of wavelet analysis Energy Econ 90 2020 104835 Katafuchi Y. Kurita K. Managi S. COVID-19 with stigma: theory and evidence from mobility data Eco. Disaster. Clim. Change 2020 10.1007/s41885-020-00077-w Kurita K. Managi S. COVID-19 and Stigma: Evolution of Self-restraint behavior, MPRA Paper 2020 University Library of Munich Germany 104042 Pal D. Mitra S.K. Oil price and automobile stock return co-movement: a wavelet coherence analysis Econ. Modell. 76 2019 172 181 Sharif A. Aloui C. Yarovaya L. COVID-19 pandemic, oil prices, stock market, geopolitical risk and policy uncertainty nexus in the US economy: fresh evidence from the wavelet-based approach Int. Rev. Financ. Anal. 70 2020 101496 Tiwari A.K. Oil prices and the macroeconomy reconsideration for Germany: using continuous wavelet Econ. Modell. 30 2013 636 642 Torrence C. Compo G.P. A practical guide to wavelet analysis Bull. Am. Meteorol. Soc. 79 1 1998 61 78 Torrence C. Webster P.J. Interdecadal changes in the ENSO–monsoon system J. Clim. 12 8 1999 2679 2690 World Health Organization, Novel Coronavirus (2019-nCOV) Situation Report, 92, (2020). Retrieved From https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports. Yang L. Cai X.J. Hamori S. Does the crude oil price influence the exchange rates of oil-importing and oil-exporting countries differently? A wavelet coherence analysis Int. Rev. Econ. Financ. 49 2017 536 547 Yoo S. Managi S. Global mortality benefits of COVID-19 action Technol. Forecast. Soc. Change 160 2020 120231 Zaremba A. Kizys R. Aharon D.Y. Demir E. Infected markets: novel coronavirus, government interventions, and stock return volatility around the globe Finance Res. Lett. 35 2020 101597 Zhang D. Hu M. Ji Q. Financial markets under the global pandemic of COVID-19 Finance Res. Lett. 36 2020 101528

0

PMC9753217

NO-CC CODE

2022-12-16 23:25:33

no

Technol Forecast Soc Change. 2021 Jun 27; 167:120710

utf-8

Technol Forecast Soc Change

10.1016/j.techfore.2021.120710

oa_other

==== Front J Crim Justice J Crim Justice Journal of Criminal Justice 0047-2352 0047-2352 Elsevier Ltd. S0047-2352(21)00050-7 10.1016/j.jcrimjus.2021.101830 101830 Article Small area variation in crime effects of COVID-19 policies in England and Wales Langton Samuel Dixon Anthony Farrell Graham ⁎ University of Leeds, United Kingdom ⁎ Corresponding author at: School of Law, University of Leeds, LS2 9JT, United Kingdom. 25 6 2021 July-August 2021 25 6 2021 75 101830101830 20 4 2021 15 6 2021 16 6 2021 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Purpose The aim of this study is to examine small area variation in crime trajectories during the COVID-19 pandemic in England and Wales. While we know how police-recorded crime responded to lockdown policies at the ‘macro’ level, less is known about the extent to which these trends were experienced uniformly at localized spatial scales. Methods Longitudinal k-means clustering is used to unpick local area variation in police notifiable offences across England and Wales. We describe the clusters identified in terms of their spatial patterning, opportunity structures and crime type profile. Results We find that in most small areas, crime remained fairly stable throughout the pandemic. Instead, a small number of meso-level areas contributed a disproportionately large amount to the macro-level trend. These were typically city centers with plentiful pre-pandemic crime opportunities, dominated by theft and shoplifting offences. Conclusion Findings offer support for opportunity theories of crime and for a mobility theory of crime during the pandemic. We explore potential implications for policy, theory and further research. Keywords COVID-19 Clustering K-means Crime decline Crime opportunity theory Pandemic ==== Body pmc1 Introduction Crime rate changes in response to COVID-19 movement restrictions have been widely documented. This includes studies of Australia (Andresen & Hodgkinson, 2020; Payne et al., 2020, Payne et al., 2021), Canada (Hodgkinson & Andresen, 2020), China (Borrion, Kurland, Tilley, & Chen, 2020; Dai, Xia, & Han, 2021), England and Wales (Dixon & Farrell, 2021; Halford, Dixon, Farrell, Malleson, & Tilley, 2020; Langton, Dixon, & Farrell, 2021; Office for National Statistics, 2020), Mexico (de la Mayir, Hoehn-Velasco, & Silverio-Murillo, 2021; Estévez-Soto, 2020), Sweden (Gerell, Kardell, & Kindgren, 2020), and the United States (Abrams, 2020; Ashby, 2020a, Ashby, 2020b; Campedelli, Aziani, & Favarin, 2020; Mohler et al., 2020; Piquero et al., 2020; Stickle & Felson, 2020). The findings are largely consistent with crime opportunity perspectives and the mobility theory of crime during the pandemic (Halford et al., 2020). That is, legally-enforced restrictions on daily activities, mobility and social interactions reduced crime opportunities. As restrictions were relaxed, these opportunities reemerged, and crime began to ‘bounce back’ closer to levels expected without the global pandemic (Langton et al., 2021). While existing studies have provided insight into the impact of lockdown and social distancing on crime, research has almost exclusively been undertaken using macro-level units of analysis, such as cities or countries. Less is known about the local drivers of the lockdown crime drops or the degree to which macro-level trends are masking geographic inequalities in victimization. Pre-pandemic studies examining the long-term crime declines in many countries comprising the international crime drop (Van Dijk, Tseloni, & Farrell, 2012) found significant inequalities (Adepeju, Langton, & Bannister, 2021; Bannister, Bates, & Kearns, 2018; Ignatans and Pease, 2015, Ignatans and Pease, 2016; McVie, Norris, & Pillinger, 2020) including at fine-grained spatial scales (Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017; Andresen & Malleson, 2011; Curman, Andresen, & Brantingham, 2015). These studies would suggest that local areas are unlikely to have experienced lockdown crime trends in unison. Rather, we might expect specific places, typically associated with high ambient populations and plentiful opportunities for crime (Malleson and Andresen, 2015, Malleson and Andresen, 2016), to have driven the wider trend, with most local areas remaining fairly stable. Examining the spatial distribution of the lockdown crime drop (and subsequent resurgence) represents the primary motivation of this paper. We decompose the macro-level trend in police-recorded crime observed in England and Wales between February and August 2020. We deploy non-parametric longitudinal clustering to identify clusters of meso-level units which contributed disproportionately to the nationwide drop and subsequent resurgence in crime during lockdown. The spatial patterning, opportunity structure and crime type profile of these local areas are quantified and summarized for their consistency with expectations from opportunity theories of crime. Relatively little is known about the spatial variation and local drivers of the macro-level trends. To what extent have local areas experienced lockdown trends in unison? This exploration represents a unique test of opportunity perspectives on crime, which would stipulate, for instance, that only a small number of local areas will have driven the lockdown crime drop. Prior to the COVID-19 pandemic, longitudinal studies of crime trends have consistently demonstrated that, in a changing macro-level scenario, most meso or micro-level areas remained remarkably stable. Instead, a disproportionately large volume of the macro-level change is attributable to a small number of units (e.g. Adepeju et al., 2021; Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017; Bannister et al., 2018; Curman et al., 2015; Trickett, Ellingworth, Hope, & Pease, 1995). This finding can be credited to highly localized opportunity structures, which can differ considerably within the same city, and even within the same neighborhood (Andresen & Malleson, 2011; Eck, Gersh, & Taylor, 2000). A shift in these local structures, brought about either through intervention (e.g. hotspot policing) or rapid changes in routine activities (Griffiths & Chavez, 2004) can bring about wider (macro) change in crime rates, even if most units nested within the macro region remain stable. In a lockdown scenario, we might expect these effects to be both exaggerated (in scale) and more instantaneous (in time). For instance, public transport hubs typically act as major crime generators due to the vast congregation of ambient populations in time and space facilitating the convergence of motivated offenders and suitable targets (Newton, 2018). In lockdown, public transport services in some countries including England and Wales were either closed completely, or open only for a limited number of essential purposes. With the necessary convergence of offenders and victims disrupted, often quite literally overnight, we would expect crime in these customarily problematic places to fall considerably, prompting a macro-level decline. By contrast, areas typically devoid of crime opportunity, with say, little or no ambient populations, may remain largely unaffected by lockdown restrictions on mobility, and in turn, contribute little to any macro-level change. In other words, while lockdowns have often been imposed equitably at a city or national level, the effect on crime will likely be moderated by the opportunity structure of local areas. Preliminary evidence from the United States certainly suggests that this may be the case. Using police-recorded crime data in San Francisco and Oakland, Shayegh and Malpede (2020) provided visual descriptive evidence which indicated a degree of geographic variation in pre and post-lockdown crime. In Detroit, a study using a small sample of block units found that, amidst a fall in burglaries following stay at home orders, there was a shift in concentrations away from residential areas towards mixed and non-residential parts of the city (Felson, Jiang, & Xu, 2020). In Chicago, there was evidence of variability in the extent to which lockdown policies impacted upon crime. A small proportion of communities drove the citywide decline, with most areas remaining largely unchanged, and some areas even increasing, bucking the macro-level trend entirely (Campedelli, Aziani, & Favarin, 2020; Campedelli, Favarin, Aziani, & Piquero, 2020). Using regional units of analysis nested within the state of Queensland, Australia, Payne et al. (2021) found a degree of diversity in crime rate trends, suggesting that the lockdown crime drop was not ‘universal’. In England and Wales, early descriptive evidence from the first three months of lockdown suggests that previously high-crime areas may have experienced the steepest relative declines compared to previous years (Dixon, Halford, & Farrell, 2020). Increases in fly-tipping have also been linked to a small number of councils, with trends varying considerably between regions (Dixon & Tlley, 2020). However, there has not been a comprehensive decomposition of local longitudinal variation underpinning lockdown crime drops, or indeed an exploration of the opportunity structures characterizing the local areas which have driven macro-level changes. In this paper, we aim to identify and describe the localized drivers of the lockdown crime drop in England and Wales. We achieve this using 7-months of police-recorded crime data between February and August 2020. First, we summarize the national (macro) trend in terms of crime counts in comparison to previous years. Second, using a non-parametric longitudinal clustering technique, we identify meaningful clusters of meso-level areas which unpick stable (and volatile) local areas underpinning the macro-level trend. Third, we describe the clusters identified in terms of their spatial patterning, opportunity structures, and crime type profile. 2 Data and method To examine localized instability in the lockdown crime drop, we make use of three data sources, namely, open police-recorded crime data, geographic boundaries from Ordnance Survey and the Office for National Statistics (ONS), and data sourced from the Open Street Map API. Each of these are now outlined in turn, followed by an outline of the methods deployed. Code to replicate the data downloads, handling, analyses and visualization reported here are openly available (https://github.com/langtonhugh/covid_spatial). 2.1 Crime data Open police-recorded data on crime and anti-social behavior in England and Wales is published through an online web portal (https://data.police.uk/). Individual records are released on a month-by-month basis for each of the 43 police forces comprising England and Wales. We used a study period spanning February to August 2020 in order to capture the first six months of the nationwide lockdown (March to August) and the one month preceding the change (February). Here, we note that lockdown was initiated on 23 March, making April the first full month of measures. For reference and comparison to historical trends, we obtained data for the same months in 2018 and 2019. Individual records are time-stamped by month – the temporal scale of this study. Due to incomplete data releases from Greater Manchester Police, we excluded data from the Greater Manchester region, collating data from 42 out of 43 forces in England and Wales. Individual open records categorize crime according to thirteen different notifiable offence categories (e.g. burglary, violence and sexual offences, theft from the person, vehicle crime).1 Records also include anti-social behavior (ASB) which usually captures less serious offences such as nuisance behavior and is not a notifiable offence. Individual records were aggregated to create a count measure for ‘notifiable offences (excluding drugs)’ by month at the nationwide (macro) level, and the localized (meso) level, as detailed in the next section. The decision to exclude drug offences follows recognition that drug crime trends, particularly during the COVID-19 lockdown, largely reflect policing proactivity rather than meaningful shifts in criminal behavior (Langton, 2020). Recognizing that aggregating data across crime types can mask variation (Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017) we later decompose our main findings according to the twelve remaining notifiable offences, and report additional analyses broken down by crime type in the Appendix. 2.2 Unit of analysis To provide national context to the main analysis, we use the (macro) geographic region of England and Wales, noting the exclusion of Greater Manchester. For localized (meso) analysis, we aggregate offences to Lower Super Output Area (LSOA). LSOA are a meso-level geographic unit designed for the reporting of official statistics at small geographies (Office of National Statistics, 2021). England and Wales is comprised of 32,844 LSOA designed to be uniform by resident population size. In 2019, the average LSOA housed 1700 people. Data obtained from the open police portal (see previous section) include a pre-assigned field stating the LSOA in which the crime occurred as recorded by the police. Due to the spatial anonymization method used prior to data release, LSOA are the lowest level of aggregation at which we can reasonably assume spatial accuracy across multiple crime types (Tompson, Johnson, Ashby, Perkins, & Edwards, 2015). After removing crimes recorded by Greater Manchester Police, crimes recorded to have occurred within the Greater Manchester region, and four LSOA which contained no crime between 2018 and 2020 (likely due to the spatial anonymization process), our final sample for the meso-level analysis comprised 33,075 LSOA. 2.3 Open Street Map To summarize the opportunity structure of local areas we required a nationwide dataset of theoretically relevant facilities and urban features which could be aggregated at the LSOA level. To this end, we obtained point-level data from the Application Programming Interface (API) for Open Street Map via the osmdata package (Padgham, Lovelace, Salmon, & Rudis, 2017) in R (R Core Team, 2020). Open Street Map is a crowdsourced geospatial database containing a vast array of features which can be used for explaining the temporal and spatial patterning of crime (Langton & Solymosi, 2020). Geographic features are identified by pairs of keys and values which can be used to computationally query the API for geospatial data. Based on existing research examining the opportunity structures of fine-grained spatial scales, we collated the coordinate locations of the following facilities:• Nightlife: pubs, nightclubs, restaurants. • Shops: convenience stores, malls, shoe shops, department stores, clothes shops, electrical shops, supermarkets, chemists, greengrocers. • Public transport: bus stops and railway stations. • Bicycle parking: bicycle parking lots. The point-level data on these features were aggregated to create counts for each facility by LSOA. For simplicity, and due to issues of data sparsity, we sum the counts for each facility according to their overarching description (i.e. nightlife, shops, public transport, bicycle parking). We expect that LSOA containing a high number of facility counts across each domain will have higher pre-pandemic levels of crime, due to the plentiful opportunities for crime, and in turn, steeper declines in crime following lockdown as a result of these opportunities suddenly becoming unavailable. We expect areas with low counts across these domains to have similarly low crime levels pre-lockdown, and thus will remain low and stable following lockdown commencement. 3 Method Analyses to examine the localized variation in the lockdown crime drop are conducted in three principal stages. First, an overview of the nationwide (macro) trend is provided in terms of absolute counts. Second, the macro-level trend is disentangled using non-parametric clustering techniques on the LSOA (meso) units (N = 33,075). Third, the characteristics of each cluster are summarized in terms of their opportunity structures, spatial patterning and crime type profile. Each of these steps is now outlined in turn. 3.1 Macro-level descriptives Macro-level descriptives of count trends notifiable offences (excluding drugs) are reported to provide the context from which we will unmask local (meso) variation. We visualize observed counts between February and August 2020 relative to the same periods in 2018 and 2019. In doing so, we can observe how crime trends changed in the face of lockdown measures in England and Wales (see also Langton et al., 2021). This sets the scene from which we can disentangle the underlying meso-level variation. 3.2 Meso-level clustering To quantify the degree of meso-level uniformity underpinning the macro-level trend, and identify the potential drivers of the lockdown crime drop (and resurgence), we deploy a longitudinal variant of k-means clustering (Genolini, Alacoque, Sentenac, Arnaud, and others, 2015; Genolini & Falissard, 2011). This non-parametric clustering technique has an established role in crime and place research for examining the longitudinal trajectories of local areas in a macro-level crime drop scenario (Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017; Curman et al., 2015). The natural experiment conditions of the COVID-19 lockdown (Stickle and Felson (2020)) make the usage of k-means particularly suitable. Existing research adopting the method has tended to investigate long-term change over years or decades, focusing on the directional homogeneity (e.g. increasing, decreasing or stable) of clusters, rather than short-term volatility (Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017). To this end, it has demonstrated comparable value to model-based techniques such as group-based trajectory modelling (Curman et al., 2015). But, a key strength of k-means is that it is also capable of identifying short-term fluctuation in longitudinal trends (Adepeju et al., 2021). In the lockdown scenario, crime opportunities were withdrawn quite literally overnight, and thus we might expect a similarly rapid and short-term change in crime at meso spatial scales. The ability to unpick these rapid changes represents a key strength of k-means over non-parametric techniques such as anchored k-medoids, which are designed for long-term rather than short-term change (Adepeju et al., 2021) or group-based trajectory modelling, which is limited by polynomial terms (Griffiths & Chavez, 2004). We deploy k-means using the kml package (Genolini, Alacoque, Sentenac, & Arnaud, 2015) in R on notifiable offence (excluding drug) counts for LSOA in England and Wales (N = 33,075) between February and August 2020. To achieve a parsimonious cluster solution while minimizing the risk of missing underlying variation, we proposed potential solutions between two and eight clusters, choosing the final solution based on the Calinski-Criterion (Caliński & Harabasz, 1974). For each potential solution, twenty redraws with different starting conditions were run to ensure that solutions were stable. Potential cluster solution options were also examined using Principal Component Analysis to establish their suitability. We visualize the final cluster solution in a manner which conveys the underlying distribution of observations comprising each cluster at each time point, rather than reporting a summary statistic (e.g. the mean trajectory) in isolation. For each cluster, we also overlay the equivalent trajectories for 2018 and 2019 as a reference point for comparison to a ‘typical’ year. In doing so, we aim to not only identify localized (in)stability in the lockdown crime drop, but also assess the extent to which the trends observed have deviated from previous years. 3.3 Cluster characteristics We expect that the meso-level areas driving the lockdown crime drop will be those with plentiful opportunities for crime. That is, a disproportionately large volume of the decline (and subsequent resurgence) will be attributable to a handful of places which had pre-existing high crime levels as a result of their opportunity structure. Using the measures for opportunity generated from Open Street Map (i.e. nightlife, shops, public transport and bicycle parking), we report descriptive statistics on facility counts for each of the clusters obtained from the k-means analysis. In doing so, we expect to unpick a meaningful pattern which is consistent with the opportunity perspective of crime. To supplement this, we visualize the spatial patterning of the cluster solutions. For brevity and simplicity, we focus on Birmingham, Liverpool, Leeds, Bradford, Sheffield and Cardiff. We have excluded Greater Manchester due to the lack of police data, and given its size, we determined Greater London to warrant an individual case study for future research. Given these exclusions, the six cities we report represent the five most populous cities in England, and the most populous city in Wales. Study regions are defined based on the city names appearing in LSOA name. For the purposes of the visual, one LSOA in Cardiff containing Flat Holm Island, which is off the coast, was removed. Finally, recognizing the unique opportunity structure of specific crimes, we summarize the crime type profile of clusters. For each cluster, we report the percentage breakdown of crimes types. We suspect that clusters will have differing crime type profiles according to the opportunity structures of each grouping. 4 Results 4.1 Nationwide trends To set the context for the localized analysis, Fig. 1 visualizes crime counts between February and August 2020 for notifiable offences excluding drugs before and after lockdown. In April, the first full month of lockdown in England and Wales, we observe a nationwide decline in notifiable offences in comparison to previous years. Upon the relaxation of lockdown rules, crime began to bounce back, and by August, crime had returned to within a range we might have expected without the nationwide lockdown (see also Langton et al., 2021). This trend represents the ‘global’ trend which we will subsequently disentangle using localized analyses.Fig. 1 Notifiable offence (excluding drugs) end of month counts in England and Wales. April was the first full month of lockdown. Fig. 1 4.2 Longitudinal clustering 4.2.1 Cluster trends We deploy non-parametric k-means clustering on LSOA (meso) level geographic units (N = 33,075) comprising England and Wales to decompose the macro-level ‘decline and resurgence’ observed between February and August 2020. Based on an assessment of the Calinski-Criterion statistic (Caliński & Harabasz, 1974) and Principal Component Analysis, we determined the optimal solution to be 6 clusters (see Fig. 2 ). Solid black lines represent the median count for each cluster at any given time point, with the black dotted line showing the mean. Violin plots have been added to convey variation around these points at each time point. These suggest that the mean and median point statistics summarize the underlying data reasonably well, and indicate that clusters are distinct from one another, with little overlap. Additional lines have been added to convey each clusters' mean and median trend in 2018 and 2019 respectively. These trends suggest that the clusters identified using the 2020 study period were distinct and meaningful even in previous years, and provide a relative baseline from which we can compare lockdown trends.Fig. 2 K-means cluster solutions for LSOA notifiable offences (excluding drugs). Distributions refer to 2020 only. Fig. 2 Overall, we note that most LSOA were remarkably stable during the pandemic. Clusters A and B could be described as ‘low crime and stable’, exhibiting fairly low counts throughout the study period and across years. Together, these clusters comprise 89% of LSOA in England and Wales. Even amidst the stark macro-level decline in notifiable offences (see Fig. 1) LSOA in these clusters only experienced marginal average dips in crime. The third largest cluster, cluster C, comprises 10% of LSOA in England and Wales. LSOA in this cluster experienced a more prominent dip in crime, along with cluster D (1.6% of LSOA). Together, we might describe these LSOAs as ‘mid-crime, mid-drop’. In both cases, there is a clear deviation from previous years. Notifiable offences fall between March and April, and then began to converge back to levels observed in previous years. That said, most ‘action’ appears to be occurring amongst a small subset of LSOA. Clusters E and F collectively comprise only 0.33% (N = 110). Yet, their crime counts are much higher, and the decline between March and April is considerable. We might therefore describe these clusters as ‘high crime, major drop’: LSOA with plentiful opportunities for crime in typical times, and in turn, LSOA which are most sensitive to the restriction in opportunities which followed after the imposition of lockdown. To further decompose these clusters, we now disentangle the contribution of each cluster. 4.2.2 Contribution of each cluster To further unpack the contribution of these clusters to the nationwide lockdown crime drop, Fig. 3 plots the monthly change in counts and percentage of total absolute change (i.e. positive or negative) attributable to each cluster. We can use this visual to identify which clusters drove the initial decline and subsequent nationwide resurgence in crime.Fig. 3 Counts and percentage of nationwide change between months attributable to each cluster. Fig. 3 As expected, the vast majority of change across all clusters occurred between March and April. Between these months, notifiable offences experienced a dramatic fall nationwide. That said, the figure demonstrates that this decline did not occur equitably across local areas. Consistent with the cluster solutions trends in Fig. 2, a small number of LSOA (meso) units appear to have contributed disproportionately to the nationwide (macro) trend. For instance, clusters A and B, which comprise 58% and 31% of LSOA in the country, accounted for only 20% and 27% of the total decline between March and April. By contrast, clusters E and F, which together comprised only 0.33% of LSOA (N = 110) contributed to 15% of the nationwide drop in these months. That is, crime in most areas was remarkably stable, given the nationwide volatility, with just a handful of areas driving the nationwide lockdown crime drop. This picture shifts slightly upon the beginning of the resurgence between April and May. Clusters A and B continue to contribute a disproportionately small amount to nationwide change, but the proportion increases, with each contributing 24% and 42% to the increase respectively. The contribution of the smallest clusters E and F is small compared to the initial decline, 2% and 0% (when rounded), respectively. This can be largely attributed to the flattening of their crime count trends between these months. In other words, following their initially steep decline between March and April, crime remained reasonably unchanged between April and May in the ‘high crime, major drop’ areas. Following the period of stabilization, clusters E and F gather pace during the resurgence, as does cluster D. For instance, between June and July, these three clusters collectively account for 27% of the nationwide increase, despite comprising less than 2% of LSOA in the country. July to August is somewhat different. Until that point, each cluster, despite their vastly differing relative contributions to change, were directionally homogeneous. In other words, every cluster declined between February and April, and then increased between April and July. Between July and August, clusters began to diverge. Nationwide, there was a marginal increase in notifiable offence counts (see Fig. 1), but this clearly masks a great deal of localized variability. The ‘low crime and stable’ clusters (A and B) actually declined again between July and August, and the higher crime clusters continued to increase. The change is almost imperceptible when viewing average counts in these LSOA (see Fig. 2) but the sheer size of these units, collectively comprising 89% of England and Wales, had a major impact on nationwide trends. This is suggestive of highly localized change in opportunities as lockdown rules were relaxed: change which is aggregated away at the macro-level. Stable, low crime areas mitigated against further nationwide increases which may, if they had continued, resulted in a higher crime count in August than observed in previous years. These findings can be supplemented with a comparison to the previous year. Fig. 4 compares the clusters identified in 2020 with the equivalent months in 2019. Here, it is clear that the ‘high crime major drop’ clusters (E and F) contributed the most to the lockdown crime drop relative to their size. By way of example, around 27,000 fewer crimes occurred in cluster A–containing 19,162 LSOA–during April 2020 compared to April 2019 (Fig. 4a). This translates to a 24% drop (Fig. 4b), and an average fall per LSOA of 1.39 counts (Fig. 4c). For cluster F, containing just 9 LSOA, approximately 5000 fewer crimes occurred (Fig. 4a), which translates to an 87% drop (Fig. 4b), and an average fall per LSOA of 543 counts (Fig. 4c). In other words: whether comparing month-to-month change in 2020, or comparing equivalent months in 2019, the lockdown crime drop was disproportionately driven by a small number of local areas, with most meso-level units remaining fairly stable, even amidst dramatic macro-level change.Fig. 4 Comparing LSOA clusters between equivalent months in 2020 and 2019 according to (a) absolute difference in counts, (b) percentage difference in counts, (c) average difference in counts. Fig. 4 4.3 Characteristics of clusters 4.3.1 Spatial distribution To provide local context to the cluster solutions identified through the k-means clustering, we visualize the spatial patterning of groupings for six major cities in England and Wales (see Fig. 5 ). We find distinct geographic patterns to the clusters identified. The maps denote a single ‘city center’ based on the LSOA with the highest count of Open Street Map features (i.e. the sum of nightlife, shops, public transport and bicycle parking). Because LSOA are uniform by residential population, geographically small LSOA tend to be areas with higher population density. Here, it is worth noting that the Bradford region contains the city of Bradford, but also the satellite town of Keighley to the north west.Fig. 5 Spatial distribution of clusters by major urban conurbation. LSOA with the highest count of Open Street Map features (in black) denotes city center. Fig. 5 Without exception, the city centers and commercial districts of each city are characterized by ‘high crime, major drop’ clusters E and F, which in turn often neighbor the ‘mid crime, mid drop’ cluster D. By contrast, the ‘low crime and stable’ clusters tend to sit outside of the city centers: cluster B in the suburbs and the sparsely populated cluster A on the periphery of each urban conurbation. As such, changes in crime during lockdown appear to have a distinct geographic pattern: the nationwide drop was not experienced uniformly across space, but rather, particular areas including city centers appear to have driven the macro-level trend, with outer suburbs remaining fairly stable. For a detailed and interactive investigation of the spatial clusters identified for the whole of England and Wales, we refer readers to an openly available online map (anonymous for review). 4.3.2 Opportunity structure We have now identified that, amidst dramatic nationwide change in notifiable offences during lockdown, there has been considerable underlying volatility, with a small number of local areas driving the macro-level trend. From a theoretical perspective, we expect the clusters to have differing opportunity structures. Table 1 reports a series of descriptive statistics based on the opportunity structure of each cluster using the nationwide data obtained from the Open Street Map API. These findings largely support our expectations, namely, that large, stable clusters have few opportunities for crime, while high crime areas which were responsible for a disproportionately large amount of the crime drop contain plentiful features which facilitate crime.Table 1 Descriptive statistics of facilities and features in each cluster. Sourced from Open Street Map Table 1Statistic [A] [B] [C] [D] [E] [F] Nightlife (mean) 0.65 0.88 1.90 7.59 25.05 97.33 Nightlife (median) 0.00 0.00 0.00 4.00 21.00 68.00 Nightlife (SD) 1.29 1.97 3.65 9.67 20.71 51.22 Shops (mean) 1.32 2.23 4.17 11.53 30.99 92.67 Shops (median) 0.00 0.00 1.00 7.00 28.00 88.00 Shops (SD) 2.90 4.28 6.88 13.15 25.12 45.99 Public transport (mean) 5.61 5.99 7.34 13.14 28.42 42.22 Public transport (median) 4.00 5.00 6.00 11.00 24.00 38.00 Public transport (SD) 6.41 5.82 6.24 10.95 19.64 29.30 Bike parking (mean) 0.43 1.06 2.12 6.82 21.85 62.67 Bike parking (median) 0.00 0.00 0.00 3.00 13.00 44.00 Bike parking (SD) 2.84 6.90 5.58 10.06 35.16 44.62 By way of example, there are sparse opportunities for crime in the largest and most stable cluster A. Most LSOA in this cluster have no nightlife facilities, no shops and no bicycle parking. Public transport is available but it is not common: the median LSOA in cluster A only contained four bus stops or railways stations. As we move along to medium and high crime clusters, which had higher pre-existing crime levels, and thus steeper declines during lockdown, these counts markedly increase. For instance, LSOA in cluster E contain a median of 21 nightlife facilities, 28 shops, 24 public transport nodes and 13 bicycle parking spaces. These counts increase further for cluster F, although we note that the cluster contains only nine LSOA. This is consistent with the findings presented earlier which suggest crime declined the most in these clusters: areas previously rich in opportunity, in which crime was pervasive, became the drivers of the lockdown crime drop. 4.3.3 Crime type profile Given the unique longitudinal trends, spatial patterning and opportunity structure of the clusters identified using the aggregate notifiable offences measure, we might expect the clusters identified to have distinct crime type profiles (see Fig. 6 ). In February, before the restrictions on mobility and social interaction, the crime type profiles of each cluster were already distinct. Higher crime clusters (e.g. E, and F) were weighted heavily towards shoplifting and theft. Low crime and stable clusters, by contrast, had higher proportions of criminal damage and arson, burglary and violence and sexual offences.Fig. 6 Crime type profiles of each cluster solution. Fig. 6 Overall, we observe a degree stability in the distribution of crime types both within and between clusters. The largest, stable clusters (A and B) have remarkably similar crime type profiles. Before lockdown, for each of these clusters, the most prevalent crimes were violence and sexual offences, both comprising 41% of total crime. This proportion increased on lockdown commencement, largely at the expense of burglary and vehicle crime, which decreased as a proportion of total crime. In clusters A and B, and indeed across all clusters, the proportion of total crime attributable to public order also increased. These increases may reflect a seasonal effect: crimes such as violence and sexual offences and public order tend to increase between February and August in typical times, while vehicle crime and burglary are usually stable (Langton et al. (2021); see Appendix). This does not, by any means, indicate that these crimes increased during lockdown – counts declined considerably – but rather, these crimes declined less steeply relative to other offence categories. In the case of public order, this may be a result of lockdown-specific activity. Public order includes offences relating to processions and assemblies, and thus may capture gatherings which violated COVID-19 guidelines (Crown Prosecution Service, 2021) and protests such as those relating to Black Lives Matter which were prominent in summer 2020 (Baggs, 2020). Increases in public order offences across these areas would slow and work against a fall in a more typical public order crimes. Notable disparities between clusters arise with regards to theft from the person and other theft in cluster F, which as reported earlier, drove a considerable part of the nationwide decline in crime. Theft in these LSOA was squeezed upon lockdown commencement. In February, theft crime types comprised 53% of total crime, but by April, this figure had dropped to 13%. This occurred despite these crime types typically remaining stable during equivalent times of year (Langton et al. (2021); see Appendix). Somewhat surprisingly, given the spatial distribution and opportunity structures of clusters E and F in city centers, shoplifting did not experience a similar squeeze in these LSOA. 5 Discussion The finding that a small number of areas drove the national crime trends, with most areas otherwise remaining fairly stable, is consistent with other studies. Evidence from Queensland, Australia (Payne et al., 2021) and Chicago, United States (Campedelli, Favarin, et al., 2020) suggests that the aggregate lockdown crime drop masked underlying variation. Likewise, studies from before the pandemic found aggregate crime trends largely attributable to variation in a handful of areas (Andresen, Linning, & Malleson, 2017, Andresen, Curman, & Linning, 2017; Curman et al., 2015; Ignatans and Pease, 2015, Ignatans and Pease, 2016;). Here, we have exploited the quasi-‘natural experiment’ conditions of the pandemic to decompose this variation during both the decline and resurgence stages of a macro-level trend, over a short period of time. Findings suggest that the macro-level effect of lockdown on crime has been moderated by highly localized characteristics. Previously criminogenic areas, characterized by plentiful opportunities for crime, and often located in city centers, were the most susceptible to decline following the introduction of “stay at home” measures. Containing a disproportionately large number of nighttime economy facilities (e.g. pubs, nightclubs), commercial outlets (e.g. malls, convenience stores), public transport nodes (e.g. railway stations, bus stops) and bicycle parking spaces, these areas were where crime was typically concentrated. In March, overnight, as the hospitality sector closed its doors, non-essential shops closed, and transport either closed or became unnecessary, these key features became unavailable or unusable. In turn, these areas contributed a considerable (and disproportionate) amount to the nationwide decline in crime. That being said, it is notable that the local drivers of the lockdown crime drop experienced only a modest resurgence upon the easing of lockdown restrictions. By August, the clusters identified as ‘high crime major drop’ remained well below the crime counts observed in 2018 and 2019, by which point the low and mid-crime clusters had returned to historical levels. This can likely be attributed to the crime type profile of clusters. The drivers of the lockdown crime drop were weighted heavily towards theft (theft from the person, other theft) and shoplifting compared to the low and mid-crime clusters. Instead, low and mid-crime areas were characterized by a larger proportion of violence and sexual offences. Breaking nationwide trends down by crime type, we observe starkly different resurgence trends across these crime types. Whereas violence and sexual offences had returned to expected levels by August, theft and shoplifting had not. This is evidenced by a descriptive comparison to 2018 and 2019 (see Appendix) and more advanced forecasting techniques which account for long-term trends (see Langton et al., 2021). This may reflect the asymmetric manner of lockdown in England and Wales. While it came into force instantaneously and equitably across the country, lockdown was lifted in stages. For instance, a select number of commercial outlets (e.g. gardening centers) were reopened in May, but restrictions on the hospitality sector and public transport remained in place until July. Here we would note some weaknesses in the study. First, the crime measure of ‘notifiable offences (excluding drugs)’ aggregates away some detail in longitudinal trends. The crime type profile findings certainly go some way in addressing this, and we have now set a baseline from which further investigations can be conducted. However, further insight could be gained from replicating this analyses for major crime types which we know to have distinct opportunity structures, such as burglary and theft. Here, we would note that the generalized Gini coefficients by crime type (see Appendix) certainly suggest that there may be differences between crime types, as have preliminary findings using decile clusters (Dixon et al., 2020). Secondly, the Open Street Map API used to quantify the prevalence of crime opportunities in each cluster is imperfect. Studies have found an association between contributions to Open Street Map and contextual characteristics such as poverty and distance from the city center – factors that are also known to be associated with crime. Due to computational issues and API query limits, some features were also unobtainable on a national scale, such as residential building footprints. While we would argue that Open Street Map is unique in its ability to offer open data on opportunity structures nationwide, we would encourage future research to verify these findings using other data sources. 6 Conclusion This paper investigated spatial variation in the COVID-19 crime drop in England and Wales. Using 7-months of police-recorded data on notifiable offences (excluding drugs), we decomposed the nationwide decline in crime according to clusters of local areas, each with distinct longitudinal trajectories. Findings provide substantial evidence to suggest that the nationwide decline in crime during lockdown was driven by a disproportionately small number of areas. We find that these areas, predominately city centers, had pre-existing high levels of crime, likely attributable to their unique opportunity structures, containing numerous nightlife facilities, commercial shops, public transport nodes and bicycle parking spaces. Clusters demonstrate a degree of stability in their crime-type profiles, but the ‘high crime, major drop’ areas appear to be weighted towards the volume crimes of theft and shoplifting. Despite their dramatic declines, these areas experienced a fairly shallow resurgence upon the relaxation of lockdown rules, remaining below historical levels even by the time low and mid-crime clusters had returned to expected levels. The findings lend weight to opportunity theories of crime and to a mobility theory of crime in the pandemic (Halford et al., 2020). Dramatic changes in crime during the pandemic occurred due to movement restrictions which impacted disproportionately upon crime in areas where it was previously concentrated. Competing interests The authors declare that they have no competing interests. Funding The work was funded by the 10.13039/100014013 UKRI open call for research on COVID-19 under grant ES/00445X/1 from the 10.13039/501100000269 Economic and Social Research Council . Availability of data and materials The data are publicly available and retrievable from the open online data portal for England, Wales and Northern Ireland. Code to replicate data downloads, handling and analyses is openly available (https://github.com/langtonhugh/covid_spatial). Appendix A Appendix Fig. 7 Counts by offence type in England and Wales during the lockdown period. Fig. 7 Fig. 8 Monthly generalized Gini coefficient trends by offence type in England and Wales during the lockdown period. Fig. 8 1 Further information on these crime categories can be obtained from https://www.police.uk/pu/contact-the-police/what-and-how-to-report/what-report/. ==== Refs References Abrams D. COVID and crime: An early empirical look, research paper no. 20–49, University of Pennsylvania Carey law School, 14 august 2020 University of Pennsylvania: Institute for Law and Economics Adepeju M. Langton S. Bannister J. Nchored k-medoids: A novel adaptation of k-medoids further refined to measure long-term instability in the exposure to crime Journal of Computational Social Science. 2021 10.1007/s42001-021-00103-1 Andresen M.A. Curman A.S. Linning S.J. The trajectories of crime at places: Understanding the patterns of disaggregated crime types Journal of Quantitative Criminology 33 3 2017 427 449 Andresen M.A. Hodgkinson T. Somehow I always end up alone: COVID-19, social isolation and crime in Queensland, Australia Crime Science 9 25 2020 1 20 Andresen M.A. Linning S.J. Malleson N. Crime at places and spatial concentrations: Exploring the spatial stability of property crime in Vancouver BC, 2003-2013 Journal of Quantitative Criminology 33 2 2017 255 275 Andresen M.A. Malleson N. Testing the stability of crime patterns: Implications for theory and practice Journal of Research in Crime and Delinquency 48 1 2011 58 82 Ashby M. Initial evidence on the relationship between the coronavirus pandemic and crime in the United States Crime Science 9 2020 1 16 Ashby M. Changes for police calls for service during the early months of the 2020 Coronavirus pandemic Policing 14 4 2020 1054 1072 Baggs M. Black lives matter in the UK: “We”re still not being heard’ 2020 BBC News online 25 August 2020, at https://www.bbc.co.uk/news/newsbeat-53812576 Bannister J. Bates E. Kearns A. Local variance in the crime drop: A longitudinal study of neighbourhoods in greater Glasgow, Scotland The British Journal of Criminology 58 1 2018 177 199 Borrion H. Kurland J. Tilley N. Chen P. Measuring the resilience of criminogenic ecosystems to global disruption: A case-study of covid-19 in China PLoS One 15 10 2020 e0240077 Caliński T. Harabasz J. A dendrite method for cluster analysis Communications in Statistics-Theory and Methods 3 1 1974 1 27 Campedelli G.M. Aziani A. Favarin S. Exploring the effects of covid-19 containment policies on crime: An empirical analysis of the short-term aftermath in los Angeles 2020 Campedelli G.M. Favarin S. Aziani A. Piquero A.R. Disentangling community-level changes in crime trends during the covid-19 pandemic in Chicago Crime Science 9 1 2020 1 18 Crown Prosecution Service 6500 coronavirus-related prosecutions in first six months of the pandemic https://www.cps.gov.uk/cps/news/6500-coronavirus-related-prosecutions-first-six-months-pandemic 2021 Curman A.S. Andresen M.A. Brantingham P.J. Crime and place: A longitudinal examination of street segment patterns in Vancouver, bc Journal of Quantitative Criminology 31 1 2015 127 147 Dai M. Xia Y. Han R. The impact of lockdown on police service calls during the COVID-19 pandemic in China Policing 2021 10.1093/police/paab007 Dixon A. Farrell G. A year of COVID-19 and crime in England and Wales Statistical bulletin on crime and COVID-19, issue 14 2021 University of Leeds Dixon A. Halford E. Farrell G. Spatial distributive justice and crime in the COVID-19 pandemic Statistical bulletin on crime and COVID-19, issue 2, 27 July 2020 2020 University of Leeds Dixon A. Tlley N. Fly-tipping during the pandemic Statistical bulletin on crime and COVID-19, issue 13, 26 November 2020 2020 University of Leeds Eck J.E. Gersh J.S. Taylor C. Finding crime hot spots through repeat address mapping Analyzing crime patterns: Frontiers of practice 2000 49 64 Estévez-Soto P.R. Crime and covid-19: Effect of changes in routine activities in Mexico City 2020 Felson M. Jiang S. Xu Y. Routine activity effects of the covid-19 pandemic on burglary in Detroit, march, 2020 Crime Science 9 1 2020 1 7 Genolini C. Alacoque X. Sentenac M. Arnaud C. Kml and kml3d: R packages to cluster longitudinal data Journal of Statistical Software 65 4 2015 1 34 http://www.jstatsoft.org/v65/i04/ Genolini C. Alacoque X. Sentenac M. Arnaud C. others Kml and kml3d: R packages to cluster longitudinal data Journal of Statistical Software 65 4 2015 1 34 Genolini C. Falissard B. KmL: A package to cluster longitudinal data Computer Methods and Programs in Biomedicine 104 3 2011 e112 e121 21708413 Gerell M. Kardell J. Kindgren J. Minor covid-19 association with crime in Sweden Crime Science 9 1 2020 1 9 Griffiths E. Chavez J.M. Communities, street guns and homicide trajectories in Chicago, 1980–1995: Merging methods for examining homicide trends across space and time Criminology 42 4 2004 941 978 Halford E. Dixon A. Farrell G. Malleson N. Tilley N. Crime and coronavirus: Social distancing, lockdown, and the mobility elasticity of crime Crime Science 9 1 2020 1 12 Hodgkinson T. Andresen M.A. Show me a man or a woman alone and I’ll show you a saint: Changes in the frequency of criminal incidents during the covid-19 pandemic Journal of Criminal Justice 69 2020 101706 32834176 Ignatans D. Pease K. Distributive justice and the crime drop Andresen M.A. Farrell G. The criminal act: The role and influence of routine activity theory 2015 Palgrave Macmillan London 77 87 Ignatans D. Pease K. On whom does the burden of crime fall now? Changes over time in counts and concentration International Review of Victimology 22 1 2016 55 63 Langton S. Dixon A. Farrell G. Six months in: Pandemic crime trends in England and wales Crime Science 10 1 2021 1 16 Langton S. Solymosi R. Open street map for crime and place OSF Preprints https://osf.io/a96y7 2020 Langton Samuel Crime and Anti-Social Behaviour in Greater London Statistical bulletin on crime and COVID-19, issue 3, 31 July 2020 2020 University of Leeds Malleson N. Andresen M.A. Spatial-temporal crime hotspots and the ambient population Crime Science 4 10 2015 1 8 Malleson N. Andresen M.A. Exploring the impact of ambient population measures on London crime hotspots Journal of Criminal Justice 46 2016 52 63 de la Mayir B. Hoehn-Velasco J.R.L. Silverio-Murillo A. Druglords don’t stay at home: COVID-19 pandemic and crime patterns in Mexio city Journal of Criminal Justice 72 2021 10.1016/j.jcrimjus.2020.101745 McVie S. Norris P. Pillinger R. Increasing inequality in experience of victimization during the crime drop: Analysing patterns of victimization in Scotland from 1993 to 2014–15 The British Journal of Criminology 60 3 2020 782 802 Mohler G. Bertozzi A.L. Carter J. Short M.B. Sledge D. Tita G.E. …Brantingham P.J. Impact of social distancing during covid-19 pandemic on crime in Los Angeles and Indianapolis Journal of Criminal Justice 101692 2020 Newton A. Macro-level generators of crime, including parks, stadiums, and transit stations 2018 Office for National Statistics Coronavirus and crime in England and Wales: August 2020. Statistical bulletin, 26 August 2020 2020 Office for National Statistics Office of National Statistics Census geography 2021 https://www.ons.gov.uk/methodology/geography/ukgeographies/censusgeography#super-output-area-soa on Economics and the Environment Padgham M. Lovelace R. Salmon M. Rudis B. Osmdata Journal of Open Source Software 2 14 2017 Payne J. Morgan A. Piquero A. COVID-19 and social distancing measures in Queensland, Australia, are associated with short-term decreases in recorded violent crime Journal of Experimental Criminology 2020 10.1007/s11292-020-09441-y Payne J. Morgan A. Piquero A. Exploring regional variability in the short-term impact of COVID-19 on property crime in Queensland, Australia Crime Science 10 7 2021 1 20 Piquero A.R. Riddell J.R. Bishopp S.A. Narvey C. Reid J.A. Piquero N.L. Staying home, staying safe? A short-term analysis of covid-19 on Dallas domestic violence American Journal of Criminal Justice 2020 1 35 R Core Team R: A language and environment for statistical computing 2020 R Foundation for Statistical Computing Vienna, Austria Shayegh S. Malpede M. Staying home saves lives, really! Staying home saves lives, really! 2020 Stickle B. Felson M. Crime rates in a pandemic: The largest criminological experiment in history American Journal of Criminal Justice 45 4 2020 525 536 32837162 Tompson L. Johnson S. Ashby M. Perkins C. Edwards P. UK open source crime data: Accuracy and possibilities for research Cartography and Geographic Information Science 42 2 2015 97 111 Trickett A. Ellingworth D. Hope T. Pease K. Crime victimization in the eighties: Changes in area and regional inequality British Journal of Criminology 35 3 1995 343 359 Van Dijk J.J.M. Tseloni A. Farrell G. The international crime drop: New directions in research 2012 Palgrave Macmillan New York

0

PMC9753224

NO-CC CODE

2022-12-16 23:25:33

no

J Crim Justice. 2021 Jun 25 July-August; 75:101830

utf-8

J Crim Justice

10.1016/j.jcrimjus.2021.101830

oa_other

==== Front J Crim Justice J Crim Justice Journal of Criminal Justice 0047-2352 0047-2352 Elsevier Ltd. S0047-2352(21)00050-7 10.1016/j.jcrimjus.2021.101830 101830 Article Small area variation in crime effects of COVID-19 policies in England and Wales Langton Samuel Dixon Anthony Farrell Graham ⁎ University of Leeds, United Kingdom ⁎ Corresponding author at: School of Law, University of Leeds, LS2 9JT, United Kingdom. 25 6 2021 July-August 2021 25 6 2021 75 101830101830 20 4 2021 15 6 2021 16 6 2021 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Purpose The aim of this study is to examine small area variation in crime trajectories during the COVID-19 pandemic in England and Wales. While we know how police-recorded crime responded to lockdown policies at the ‘macro’ level, less is known about the extent to which these trends were experienced uniformly at localized spatial scales. Methods Longitudinal k-means clustering is used to unpick local area variation in police notifiable offences across England and Wales. We describe the clusters identified in terms of their spatial patterning, opportunity structures and crime type profile. Results We find that in most small areas, crime remained fairly stable throughout the pandemic. Instead, a small number of meso-level areas contributed a disproportionately large amount to the macro-level trend. These were typically city centers with plentiful pre-pandemic crime opportunities, dominated by theft and shoplifting offences. Conclusion Findings offer support for opportunity theories of crime and for a mobility theory of crime during the pandemic. We explore potential implications for policy, theory and further research. Keywords COVID-19 Clustering K-means Crime decline Crime opportunity theory Pandemic ==== Body pmc1 Introduction Crime rate changes in response to COVID-19 movement restrictions have been widely documented. This includes studies of Australia (Andresen & Hodgkinson, 2020; Payne et al., 2020, Payne et al., 2021), Canada (Hodgkinson & Andresen, 2020), China (Borrion, Kurland, Tilley, & Chen, 2020; Dai, Xia, & Han, 2021), England and Wales (Dixon & Farrell, 2021; Halford, Dixon, Farrell, Malleson, & Tilley, 2020; Langton, Dixon, & Farrell, 2021; Office for National Statistics, 2020), Mexico (de la Mayir, Hoehn-Velasco, & Silverio-Murillo, 2021; Estévez-Soto, 2020), Sweden (Gerell, Kardell, & Kindgren, 2020), and the United States (Abrams, 2020; Ashby, 2020a, Ashby, 2020b; Campedelli, Aziani, & Favarin, 2020; Mohler et al., 2020; Piquero et al., 2020; Stickle & Felson, 2020). The findings are largely consistent with crime opportunity perspectives and the mobility theory of crime during the pandemic (Halford et al., 2020). That is, legally-enforced restrictions on daily activities, mobility and social interactions reduced crime opportunities. As restrictions were relaxed, these opportunities reemerged, and crime began to ‘bounce back’ closer to levels expected without the global pandemic (Langton et al., 2021). While existing studies have provided insight into the impact of lockdown and social distancing on crime, research has almost exclusively been undertaken using macro-level units of analysis, such as cities or countries. Less is known about the local drivers of the lockdown crime drops or the degree to which macro-level trends are masking geographic inequalities in victimization. Pre-pandemic studies examining the long-term crime declines in many countries comprising the international crime drop (Van Dijk, Tseloni, & Farrell, 2012) found significant inequalities (Adepeju, Langton, & Bannister, 2021; Bannister, Bates, & Kearns, 2018; Ignatans and Pease, 2015, Ignatans and Pease, 2016; McVie, Norris, & Pillinger, 2020) including at fine-grained spatial scales (Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017; Andresen & Malleson, 2011; Curman, Andresen, & Brantingham, 2015). These studies would suggest that local areas are unlikely to have experienced lockdown crime trends in unison. Rather, we might expect specific places, typically associated with high ambient populations and plentiful opportunities for crime (Malleson and Andresen, 2015, Malleson and Andresen, 2016), to have driven the wider trend, with most local areas remaining fairly stable. Examining the spatial distribution of the lockdown crime drop (and subsequent resurgence) represents the primary motivation of this paper. We decompose the macro-level trend in police-recorded crime observed in England and Wales between February and August 2020. We deploy non-parametric longitudinal clustering to identify clusters of meso-level units which contributed disproportionately to the nationwide drop and subsequent resurgence in crime during lockdown. The spatial patterning, opportunity structure and crime type profile of these local areas are quantified and summarized for their consistency with expectations from opportunity theories of crime. Relatively little is known about the spatial variation and local drivers of the macro-level trends. To what extent have local areas experienced lockdown trends in unison? This exploration represents a unique test of opportunity perspectives on crime, which would stipulate, for instance, that only a small number of local areas will have driven the lockdown crime drop. Prior to the COVID-19 pandemic, longitudinal studies of crime trends have consistently demonstrated that, in a changing macro-level scenario, most meso or micro-level areas remained remarkably stable. Instead, a disproportionately large volume of the macro-level change is attributable to a small number of units (e.g. Adepeju et al., 2021; Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017; Bannister et al., 2018; Curman et al., 2015; Trickett, Ellingworth, Hope, & Pease, 1995). This finding can be credited to highly localized opportunity structures, which can differ considerably within the same city, and even within the same neighborhood (Andresen & Malleson, 2011; Eck, Gersh, & Taylor, 2000). A shift in these local structures, brought about either through intervention (e.g. hotspot policing) or rapid changes in routine activities (Griffiths & Chavez, 2004) can bring about wider (macro) change in crime rates, even if most units nested within the macro region remain stable. In a lockdown scenario, we might expect these effects to be both exaggerated (in scale) and more instantaneous (in time). For instance, public transport hubs typically act as major crime generators due to the vast congregation of ambient populations in time and space facilitating the convergence of motivated offenders and suitable targets (Newton, 2018). In lockdown, public transport services in some countries including England and Wales were either closed completely, or open only for a limited number of essential purposes. With the necessary convergence of offenders and victims disrupted, often quite literally overnight, we would expect crime in these customarily problematic places to fall considerably, prompting a macro-level decline. By contrast, areas typically devoid of crime opportunity, with say, little or no ambient populations, may remain largely unaffected by lockdown restrictions on mobility, and in turn, contribute little to any macro-level change. In other words, while lockdowns have often been imposed equitably at a city or national level, the effect on crime will likely be moderated by the opportunity structure of local areas. Preliminary evidence from the United States certainly suggests that this may be the case. Using police-recorded crime data in San Francisco and Oakland, Shayegh and Malpede (2020) provided visual descriptive evidence which indicated a degree of geographic variation in pre and post-lockdown crime. In Detroit, a study using a small sample of block units found that, amidst a fall in burglaries following stay at home orders, there was a shift in concentrations away from residential areas towards mixed and non-residential parts of the city (Felson, Jiang, & Xu, 2020). In Chicago, there was evidence of variability in the extent to which lockdown policies impacted upon crime. A small proportion of communities drove the citywide decline, with most areas remaining largely unchanged, and some areas even increasing, bucking the macro-level trend entirely (Campedelli, Aziani, & Favarin, 2020; Campedelli, Favarin, Aziani, & Piquero, 2020). Using regional units of analysis nested within the state of Queensland, Australia, Payne et al. (2021) found a degree of diversity in crime rate trends, suggesting that the lockdown crime drop was not ‘universal’. In England and Wales, early descriptive evidence from the first three months of lockdown suggests that previously high-crime areas may have experienced the steepest relative declines compared to previous years (Dixon, Halford, & Farrell, 2020). Increases in fly-tipping have also been linked to a small number of councils, with trends varying considerably between regions (Dixon & Tlley, 2020). However, there has not been a comprehensive decomposition of local longitudinal variation underpinning lockdown crime drops, or indeed an exploration of the opportunity structures characterizing the local areas which have driven macro-level changes. In this paper, we aim to identify and describe the localized drivers of the lockdown crime drop in England and Wales. We achieve this using 7-months of police-recorded crime data between February and August 2020. First, we summarize the national (macro) trend in terms of crime counts in comparison to previous years. Second, using a non-parametric longitudinal clustering technique, we identify meaningful clusters of meso-level areas which unpick stable (and volatile) local areas underpinning the macro-level trend. Third, we describe the clusters identified in terms of their spatial patterning, opportunity structures, and crime type profile. 2 Data and method To examine localized instability in the lockdown crime drop, we make use of three data sources, namely, open police-recorded crime data, geographic boundaries from Ordnance Survey and the Office for National Statistics (ONS), and data sourced from the Open Street Map API. Each of these are now outlined in turn, followed by an outline of the methods deployed. Code to replicate the data downloads, handling, analyses and visualization reported here are openly available (https://github.com/langtonhugh/covid_spatial). 2.1 Crime data Open police-recorded data on crime and anti-social behavior in England and Wales is published through an online web portal (https://data.police.uk/). Individual records are released on a month-by-month basis for each of the 43 police forces comprising England and Wales. We used a study period spanning February to August 2020 in order to capture the first six months of the nationwide lockdown (March to August) and the one month preceding the change (February). Here, we note that lockdown was initiated on 23 March, making April the first full month of measures. For reference and comparison to historical trends, we obtained data for the same months in 2018 and 2019. Individual records are time-stamped by month – the temporal scale of this study. Due to incomplete data releases from Greater Manchester Police, we excluded data from the Greater Manchester region, collating data from 42 out of 43 forces in England and Wales. Individual open records categorize crime according to thirteen different notifiable offence categories (e.g. burglary, violence and sexual offences, theft from the person, vehicle crime).1 Records also include anti-social behavior (ASB) which usually captures less serious offences such as nuisance behavior and is not a notifiable offence. Individual records were aggregated to create a count measure for ‘notifiable offences (excluding drugs)’ by month at the nationwide (macro) level, and the localized (meso) level, as detailed in the next section. The decision to exclude drug offences follows recognition that drug crime trends, particularly during the COVID-19 lockdown, largely reflect policing proactivity rather than meaningful shifts in criminal behavior (Langton, 2020). Recognizing that aggregating data across crime types can mask variation (Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017) we later decompose our main findings according to the twelve remaining notifiable offences, and report additional analyses broken down by crime type in the Appendix. 2.2 Unit of analysis To provide national context to the main analysis, we use the (macro) geographic region of England and Wales, noting the exclusion of Greater Manchester. For localized (meso) analysis, we aggregate offences to Lower Super Output Area (LSOA). LSOA are a meso-level geographic unit designed for the reporting of official statistics at small geographies (Office of National Statistics, 2021). England and Wales is comprised of 32,844 LSOA designed to be uniform by resident population size. In 2019, the average LSOA housed 1700 people. Data obtained from the open police portal (see previous section) include a pre-assigned field stating the LSOA in which the crime occurred as recorded by the police. Due to the spatial anonymization method used prior to data release, LSOA are the lowest level of aggregation at which we can reasonably assume spatial accuracy across multiple crime types (Tompson, Johnson, Ashby, Perkins, & Edwards, 2015). After removing crimes recorded by Greater Manchester Police, crimes recorded to have occurred within the Greater Manchester region, and four LSOA which contained no crime between 2018 and 2020 (likely due to the spatial anonymization process), our final sample for the meso-level analysis comprised 33,075 LSOA. 2.3 Open Street Map To summarize the opportunity structure of local areas we required a nationwide dataset of theoretically relevant facilities and urban features which could be aggregated at the LSOA level. To this end, we obtained point-level data from the Application Programming Interface (API) for Open Street Map via the osmdata package (Padgham, Lovelace, Salmon, & Rudis, 2017) in R (R Core Team, 2020). Open Street Map is a crowdsourced geospatial database containing a vast array of features which can be used for explaining the temporal and spatial patterning of crime (Langton & Solymosi, 2020). Geographic features are identified by pairs of keys and values which can be used to computationally query the API for geospatial data. Based on existing research examining the opportunity structures of fine-grained spatial scales, we collated the coordinate locations of the following facilities:• Nightlife: pubs, nightclubs, restaurants. • Shops: convenience stores, malls, shoe shops, department stores, clothes shops, electrical shops, supermarkets, chemists, greengrocers. • Public transport: bus stops and railway stations. • Bicycle parking: bicycle parking lots. The point-level data on these features were aggregated to create counts for each facility by LSOA. For simplicity, and due to issues of data sparsity, we sum the counts for each facility according to their overarching description (i.e. nightlife, shops, public transport, bicycle parking). We expect that LSOA containing a high number of facility counts across each domain will have higher pre-pandemic levels of crime, due to the plentiful opportunities for crime, and in turn, steeper declines in crime following lockdown as a result of these opportunities suddenly becoming unavailable. We expect areas with low counts across these domains to have similarly low crime levels pre-lockdown, and thus will remain low and stable following lockdown commencement. 3 Method Analyses to examine the localized variation in the lockdown crime drop are conducted in three principal stages. First, an overview of the nationwide (macro) trend is provided in terms of absolute counts. Second, the macro-level trend is disentangled using non-parametric clustering techniques on the LSOA (meso) units (N = 33,075). Third, the characteristics of each cluster are summarized in terms of their opportunity structures, spatial patterning and crime type profile. Each of these steps is now outlined in turn. 3.1 Macro-level descriptives Macro-level descriptives of count trends notifiable offences (excluding drugs) are reported to provide the context from which we will unmask local (meso) variation. We visualize observed counts between February and August 2020 relative to the same periods in 2018 and 2019. In doing so, we can observe how crime trends changed in the face of lockdown measures in England and Wales (see also Langton et al., 2021). This sets the scene from which we can disentangle the underlying meso-level variation. 3.2 Meso-level clustering To quantify the degree of meso-level uniformity underpinning the macro-level trend, and identify the potential drivers of the lockdown crime drop (and resurgence), we deploy a longitudinal variant of k-means clustering (Genolini, Alacoque, Sentenac, Arnaud, and others, 2015; Genolini & Falissard, 2011). This non-parametric clustering technique has an established role in crime and place research for examining the longitudinal trajectories of local areas in a macro-level crime drop scenario (Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017; Curman et al., 2015). The natural experiment conditions of the COVID-19 lockdown (Stickle and Felson (2020)) make the usage of k-means particularly suitable. Existing research adopting the method has tended to investigate long-term change over years or decades, focusing on the directional homogeneity (e.g. increasing, decreasing or stable) of clusters, rather than short-term volatility (Andresen, Curman, & Linning, 2017; Andresen, Linning, & Malleson, 2017). To this end, it has demonstrated comparable value to model-based techniques such as group-based trajectory modelling (Curman et al., 2015). But, a key strength of k-means is that it is also capable of identifying short-term fluctuation in longitudinal trends (Adepeju et al., 2021). In the lockdown scenario, crime opportunities were withdrawn quite literally overnight, and thus we might expect a similarly rapid and short-term change in crime at meso spatial scales. The ability to unpick these rapid changes represents a key strength of k-means over non-parametric techniques such as anchored k-medoids, which are designed for long-term rather than short-term change (Adepeju et al., 2021) or group-based trajectory modelling, which is limited by polynomial terms (Griffiths & Chavez, 2004). We deploy k-means using the kml package (Genolini, Alacoque, Sentenac, & Arnaud, 2015) in R on notifiable offence (excluding drug) counts for LSOA in England and Wales (N = 33,075) between February and August 2020. To achieve a parsimonious cluster solution while minimizing the risk of missing underlying variation, we proposed potential solutions between two and eight clusters, choosing the final solution based on the Calinski-Criterion (Caliński & Harabasz, 1974). For each potential solution, twenty redraws with different starting conditions were run to ensure that solutions were stable. Potential cluster solution options were also examined using Principal Component Analysis to establish their suitability. We visualize the final cluster solution in a manner which conveys the underlying distribution of observations comprising each cluster at each time point, rather than reporting a summary statistic (e.g. the mean trajectory) in isolation. For each cluster, we also overlay the equivalent trajectories for 2018 and 2019 as a reference point for comparison to a ‘typical’ year. In doing so, we aim to not only identify localized (in)stability in the lockdown crime drop, but also assess the extent to which the trends observed have deviated from previous years. 3.3 Cluster characteristics We expect that the meso-level areas driving the lockdown crime drop will be those with plentiful opportunities for crime. That is, a disproportionately large volume of the decline (and subsequent resurgence) will be attributable to a handful of places which had pre-existing high crime levels as a result of their opportunity structure. Using the measures for opportunity generated from Open Street Map (i.e. nightlife, shops, public transport and bicycle parking), we report descriptive statistics on facility counts for each of the clusters obtained from the k-means analysis. In doing so, we expect to unpick a meaningful pattern which is consistent with the opportunity perspective of crime. To supplement this, we visualize the spatial patterning of the cluster solutions. For brevity and simplicity, we focus on Birmingham, Liverpool, Leeds, Bradford, Sheffield and Cardiff. We have excluded Greater Manchester due to the lack of police data, and given its size, we determined Greater London to warrant an individual case study for future research. Given these exclusions, the six cities we report represent the five most populous cities in England, and the most populous city in Wales. Study regions are defined based on the city names appearing in LSOA name. For the purposes of the visual, one LSOA in Cardiff containing Flat Holm Island, which is off the coast, was removed. Finally, recognizing the unique opportunity structure of specific crimes, we summarize the crime type profile of clusters. For each cluster, we report the percentage breakdown of crimes types. We suspect that clusters will have differing crime type profiles according to the opportunity structures of each grouping. 4 Results 4.1 Nationwide trends To set the context for the localized analysis, Fig. 1 visualizes crime counts between February and August 2020 for notifiable offences excluding drugs before and after lockdown. In April, the first full month of lockdown in England and Wales, we observe a nationwide decline in notifiable offences in comparison to previous years. Upon the relaxation of lockdown rules, crime began to bounce back, and by August, crime had returned to within a range we might have expected without the nationwide lockdown (see also Langton et al., 2021). This trend represents the ‘global’ trend which we will subsequently disentangle using localized analyses.Fig. 1 Notifiable offence (excluding drugs) end of month counts in England and Wales. April was the first full month of lockdown. Fig. 1 4.2 Longitudinal clustering 4.2.1 Cluster trends We deploy non-parametric k-means clustering on LSOA (meso) level geographic units (N = 33,075) comprising England and Wales to decompose the macro-level ‘decline and resurgence’ observed between February and August 2020. Based on an assessment of the Calinski-Criterion statistic (Caliński & Harabasz, 1974) and Principal Component Analysis, we determined the optimal solution to be 6 clusters (see Fig. 2 ). Solid black lines represent the median count for each cluster at any given time point, with the black dotted line showing the mean. Violin plots have been added to convey variation around these points at each time point. These suggest that the mean and median point statistics summarize the underlying data reasonably well, and indicate that clusters are distinct from one another, with little overlap. Additional lines have been added to convey each clusters' mean and median trend in 2018 and 2019 respectively. These trends suggest that the clusters identified using the 2020 study period were distinct and meaningful even in previous years, and provide a relative baseline from which we can compare lockdown trends.Fig. 2 K-means cluster solutions for LSOA notifiable offences (excluding drugs). Distributions refer to 2020 only. Fig. 2 Overall, we note that most LSOA were remarkably stable during the pandemic. Clusters A and B could be described as ‘low crime and stable’, exhibiting fairly low counts throughout the study period and across years. Together, these clusters comprise 89% of LSOA in England and Wales. Even amidst the stark macro-level decline in notifiable offences (see Fig. 1) LSOA in these clusters only experienced marginal average dips in crime. The third largest cluster, cluster C, comprises 10% of LSOA in England and Wales. LSOA in this cluster experienced a more prominent dip in crime, along with cluster D (1.6% of LSOA). Together, we might describe these LSOAs as ‘mid-crime, mid-drop’. In both cases, there is a clear deviation from previous years. Notifiable offences fall between March and April, and then began to converge back to levels observed in previous years. That said, most ‘action’ appears to be occurring amongst a small subset of LSOA. Clusters E and F collectively comprise only 0.33% (N = 110). Yet, their crime counts are much higher, and the decline between March and April is considerable. We might therefore describe these clusters as ‘high crime, major drop’: LSOA with plentiful opportunities for crime in typical times, and in turn, LSOA which are most sensitive to the restriction in opportunities which followed after the imposition of lockdown. To further decompose these clusters, we now disentangle the contribution of each cluster. 4.2.2 Contribution of each cluster To further unpack the contribution of these clusters to the nationwide lockdown crime drop, Fig. 3 plots the monthly change in counts and percentage of total absolute change (i.e. positive or negative) attributable to each cluster. We can use this visual to identify which clusters drove the initial decline and subsequent nationwide resurgence in crime.Fig. 3 Counts and percentage of nationwide change between months attributable to each cluster. Fig. 3 As expected, the vast majority of change across all clusters occurred between March and April. Between these months, notifiable offences experienced a dramatic fall nationwide. That said, the figure demonstrates that this decline did not occur equitably across local areas. Consistent with the cluster solutions trends in Fig. 2, a small number of LSOA (meso) units appear to have contributed disproportionately to the nationwide (macro) trend. For instance, clusters A and B, which comprise 58% and 31% of LSOA in the country, accounted for only 20% and 27% of the total decline between March and April. By contrast, clusters E and F, which together comprised only 0.33% of LSOA (N = 110) contributed to 15% of the nationwide drop in these months. That is, crime in most areas was remarkably stable, given the nationwide volatility, with just a handful of areas driving the nationwide lockdown crime drop. This picture shifts slightly upon the beginning of the resurgence between April and May. Clusters A and B continue to contribute a disproportionately small amount to nationwide change, but the proportion increases, with each contributing 24% and 42% to the increase respectively. The contribution of the smallest clusters E and F is small compared to the initial decline, 2% and 0% (when rounded), respectively. This can be largely attributed to the flattening of their crime count trends between these months. In other words, following their initially steep decline between March and April, crime remained reasonably unchanged between April and May in the ‘high crime, major drop’ areas. Following the period of stabilization, clusters E and F gather pace during the resurgence, as does cluster D. For instance, between June and July, these three clusters collectively account for 27% of the nationwide increase, despite comprising less than 2% of LSOA in the country. July to August is somewhat different. Until that point, each cluster, despite their vastly differing relative contributions to change, were directionally homogeneous. In other words, every cluster declined between February and April, and then increased between April and July. Between July and August, clusters began to diverge. Nationwide, there was a marginal increase in notifiable offence counts (see Fig. 1), but this clearly masks a great deal of localized variability. The ‘low crime and stable’ clusters (A and B) actually declined again between July and August, and the higher crime clusters continued to increase. The change is almost imperceptible when viewing average counts in these LSOA (see Fig. 2) but the sheer size of these units, collectively comprising 89% of England and Wales, had a major impact on nationwide trends. This is suggestive of highly localized change in opportunities as lockdown rules were relaxed: change which is aggregated away at the macro-level. Stable, low crime areas mitigated against further nationwide increases which may, if they had continued, resulted in a higher crime count in August than observed in previous years. These findings can be supplemented with a comparison to the previous year. Fig. 4 compares the clusters identified in 2020 with the equivalent months in 2019. Here, it is clear that the ‘high crime major drop’ clusters (E and F) contributed the most to the lockdown crime drop relative to their size. By way of example, around 27,000 fewer crimes occurred in cluster A–containing 19,162 LSOA–during April 2020 compared to April 2019 (Fig. 4a). This translates to a 24% drop (Fig. 4b), and an average fall per LSOA of 1.39 counts (Fig. 4c). For cluster F, containing just 9 LSOA, approximately 5000 fewer crimes occurred (Fig. 4a), which translates to an 87% drop (Fig. 4b), and an average fall per LSOA of 543 counts (Fig. 4c). In other words: whether comparing month-to-month change in 2020, or comparing equivalent months in 2019, the lockdown crime drop was disproportionately driven by a small number of local areas, with most meso-level units remaining fairly stable, even amidst dramatic macro-level change.Fig. 4 Comparing LSOA clusters between equivalent months in 2020 and 2019 according to (a) absolute difference in counts, (b) percentage difference in counts, (c) average difference in counts. Fig. 4 4.3 Characteristics of clusters 4.3.1 Spatial distribution To provide local context to the cluster solutions identified through the k-means clustering, we visualize the spatial patterning of groupings for six major cities in England and Wales (see Fig. 5 ). We find distinct geographic patterns to the clusters identified. The maps denote a single ‘city center’ based on the LSOA with the highest count of Open Street Map features (i.e. the sum of nightlife, shops, public transport and bicycle parking). Because LSOA are uniform by residential population, geographically small LSOA tend to be areas with higher population density. Here, it is worth noting that the Bradford region contains the city of Bradford, but also the satellite town of Keighley to the north west.Fig. 5 Spatial distribution of clusters by major urban conurbation. LSOA with the highest count of Open Street Map features (in black) denotes city center. Fig. 5 Without exception, the city centers and commercial districts of each city are characterized by ‘high crime, major drop’ clusters E and F, which in turn often neighbor the ‘mid crime, mid drop’ cluster D. By contrast, the ‘low crime and stable’ clusters tend to sit outside of the city centers: cluster B in the suburbs and the sparsely populated cluster A on the periphery of each urban conurbation. As such, changes in crime during lockdown appear to have a distinct geographic pattern: the nationwide drop was not experienced uniformly across space, but rather, particular areas including city centers appear to have driven the macro-level trend, with outer suburbs remaining fairly stable. For a detailed and interactive investigation of the spatial clusters identified for the whole of England and Wales, we refer readers to an openly available online map (anonymous for review). 4.3.2 Opportunity structure We have now identified that, amidst dramatic nationwide change in notifiable offences during lockdown, there has been considerable underlying volatility, with a small number of local areas driving the macro-level trend. From a theoretical perspective, we expect the clusters to have differing opportunity structures. Table 1 reports a series of descriptive statistics based on the opportunity structure of each cluster using the nationwide data obtained from the Open Street Map API. These findings largely support our expectations, namely, that large, stable clusters have few opportunities for crime, while high crime areas which were responsible for a disproportionately large amount of the crime drop contain plentiful features which facilitate crime.Table 1 Descriptive statistics of facilities and features in each cluster. Sourced from Open Street Map Table 1Statistic [A] [B] [C] [D] [E] [F] Nightlife (mean) 0.65 0.88 1.90 7.59 25.05 97.33 Nightlife (median) 0.00 0.00 0.00 4.00 21.00 68.00 Nightlife (SD) 1.29 1.97 3.65 9.67 20.71 51.22 Shops (mean) 1.32 2.23 4.17 11.53 30.99 92.67 Shops (median) 0.00 0.00 1.00 7.00 28.00 88.00 Shops (SD) 2.90 4.28 6.88 13.15 25.12 45.99 Public transport (mean) 5.61 5.99 7.34 13.14 28.42 42.22 Public transport (median) 4.00 5.00 6.00 11.00 24.00 38.00 Public transport (SD) 6.41 5.82 6.24 10.95 19.64 29.30 Bike parking (mean) 0.43 1.06 2.12 6.82 21.85 62.67 Bike parking (median) 0.00 0.00 0.00 3.00 13.00 44.00 Bike parking (SD) 2.84 6.90 5.58 10.06 35.16 44.62 By way of example, there are sparse opportunities for crime in the largest and most stable cluster A. Most LSOA in this cluster have no nightlife facilities, no shops and no bicycle parking. Public transport is available but it is not common: the median LSOA in cluster A only contained four bus stops or railways stations. As we move along to medium and high crime clusters, which had higher pre-existing crime levels, and thus steeper declines during lockdown, these counts markedly increase. For instance, LSOA in cluster E contain a median of 21 nightlife facilities, 28 shops, 24 public transport nodes and 13 bicycle parking spaces. These counts increase further for cluster F, although we note that the cluster contains only nine LSOA. This is consistent with the findings presented earlier which suggest crime declined the most in these clusters: areas previously rich in opportunity, in which crime was pervasive, became the drivers of the lockdown crime drop. 4.3.3 Crime type profile Given the unique longitudinal trends, spatial patterning and opportunity structure of the clusters identified using the aggregate notifiable offences measure, we might expect the clusters identified to have distinct crime type profiles (see Fig. 6 ). In February, before the restrictions on mobility and social interaction, the crime type profiles of each cluster were already distinct. Higher crime clusters (e.g. E, and F) were weighted heavily towards shoplifting and theft. Low crime and stable clusters, by contrast, had higher proportions of criminal damage and arson, burglary and violence and sexual offences.Fig. 6 Crime type profiles of each cluster solution. Fig. 6 Overall, we observe a degree stability in the distribution of crime types both within and between clusters. The largest, stable clusters (A and B) have remarkably similar crime type profiles. Before lockdown, for each of these clusters, the most prevalent crimes were violence and sexual offences, both comprising 41% of total crime. This proportion increased on lockdown commencement, largely at the expense of burglary and vehicle crime, which decreased as a proportion of total crime. In clusters A and B, and indeed across all clusters, the proportion of total crime attributable to public order also increased. These increases may reflect a seasonal effect: crimes such as violence and sexual offences and public order tend to increase between February and August in typical times, while vehicle crime and burglary are usually stable (Langton et al. (2021); see Appendix). This does not, by any means, indicate that these crimes increased during lockdown – counts declined considerably – but rather, these crimes declined less steeply relative to other offence categories. In the case of public order, this may be a result of lockdown-specific activity. Public order includes offences relating to processions and assemblies, and thus may capture gatherings which violated COVID-19 guidelines (Crown Prosecution Service, 2021) and protests such as those relating to Black Lives Matter which were prominent in summer 2020 (Baggs, 2020). Increases in public order offences across these areas would slow and work against a fall in a more typical public order crimes. Notable disparities between clusters arise with regards to theft from the person and other theft in cluster F, which as reported earlier, drove a considerable part of the nationwide decline in crime. Theft in these LSOA was squeezed upon lockdown commencement. In February, theft crime types comprised 53% of total crime, but by April, this figure had dropped to 13%. This occurred despite these crime types typically remaining stable during equivalent times of year (Langton et al. (2021); see Appendix). Somewhat surprisingly, given the spatial distribution and opportunity structures of clusters E and F in city centers, shoplifting did not experience a similar squeeze in these LSOA. 5 Discussion The finding that a small number of areas drove the national crime trends, with most areas otherwise remaining fairly stable, is consistent with other studies. Evidence from Queensland, Australia (Payne et al., 2021) and Chicago, United States (Campedelli, Favarin, et al., 2020) suggests that the aggregate lockdown crime drop masked underlying variation. Likewise, studies from before the pandemic found aggregate crime trends largely attributable to variation in a handful of areas (Andresen, Linning, & Malleson, 2017, Andresen, Curman, & Linning, 2017; Curman et al., 2015; Ignatans and Pease, 2015, Ignatans and Pease, 2016;). Here, we have exploited the quasi-‘natural experiment’ conditions of the pandemic to decompose this variation during both the decline and resurgence stages of a macro-level trend, over a short period of time. Findings suggest that the macro-level effect of lockdown on crime has been moderated by highly localized characteristics. Previously criminogenic areas, characterized by plentiful opportunities for crime, and often located in city centers, were the most susceptible to decline following the introduction of “stay at home” measures. Containing a disproportionately large number of nighttime economy facilities (e.g. pubs, nightclubs), commercial outlets (e.g. malls, convenience stores), public transport nodes (e.g. railway stations, bus stops) and bicycle parking spaces, these areas were where crime was typically concentrated. In March, overnight, as the hospitality sector closed its doors, non-essential shops closed, and transport either closed or became unnecessary, these key features became unavailable or unusable. In turn, these areas contributed a considerable (and disproportionate) amount to the nationwide decline in crime. That being said, it is notable that the local drivers of the lockdown crime drop experienced only a modest resurgence upon the easing of lockdown restrictions. By August, the clusters identified as ‘high crime major drop’ remained well below the crime counts observed in 2018 and 2019, by which point the low and mid-crime clusters had returned to historical levels. This can likely be attributed to the crime type profile of clusters. The drivers of the lockdown crime drop were weighted heavily towards theft (theft from the person, other theft) and shoplifting compared to the low and mid-crime clusters. Instead, low and mid-crime areas were characterized by a larger proportion of violence and sexual offences. Breaking nationwide trends down by crime type, we observe starkly different resurgence trends across these crime types. Whereas violence and sexual offences had returned to expected levels by August, theft and shoplifting had not. This is evidenced by a descriptive comparison to 2018 and 2019 (see Appendix) and more advanced forecasting techniques which account for long-term trends (see Langton et al., 2021). This may reflect the asymmetric manner of lockdown in England and Wales. While it came into force instantaneously and equitably across the country, lockdown was lifted in stages. For instance, a select number of commercial outlets (e.g. gardening centers) were reopened in May, but restrictions on the hospitality sector and public transport remained in place until July. Here we would note some weaknesses in the study. First, the crime measure of ‘notifiable offences (excluding drugs)’ aggregates away some detail in longitudinal trends. The crime type profile findings certainly go some way in addressing this, and we have now set a baseline from which further investigations can be conducted. However, further insight could be gained from replicating this analyses for major crime types which we know to have distinct opportunity structures, such as burglary and theft. Here, we would note that the generalized Gini coefficients by crime type (see Appendix) certainly suggest that there may be differences between crime types, as have preliminary findings using decile clusters (Dixon et al., 2020). Secondly, the Open Street Map API used to quantify the prevalence of crime opportunities in each cluster is imperfect. Studies have found an association between contributions to Open Street Map and contextual characteristics such as poverty and distance from the city center – factors that are also known to be associated with crime. Due to computational issues and API query limits, some features were also unobtainable on a national scale, such as residential building footprints. While we would argue that Open Street Map is unique in its ability to offer open data on opportunity structures nationwide, we would encourage future research to verify these findings using other data sources. 6 Conclusion This paper investigated spatial variation in the COVID-19 crime drop in England and Wales. Using 7-months of police-recorded data on notifiable offences (excluding drugs), we decomposed the nationwide decline in crime according to clusters of local areas, each with distinct longitudinal trajectories. Findings provide substantial evidence to suggest that the nationwide decline in crime during lockdown was driven by a disproportionately small number of areas. We find that these areas, predominately city centers, had pre-existing high levels of crime, likely attributable to their unique opportunity structures, containing numerous nightlife facilities, commercial shops, public transport nodes and bicycle parking spaces. Clusters demonstrate a degree of stability in their crime-type profiles, but the ‘high crime, major drop’ areas appear to be weighted towards the volume crimes of theft and shoplifting. Despite their dramatic declines, these areas experienced a fairly shallow resurgence upon the relaxation of lockdown rules, remaining below historical levels even by the time low and mid-crime clusters had returned to expected levels. The findings lend weight to opportunity theories of crime and to a mobility theory of crime in the pandemic (Halford et al., 2020). Dramatic changes in crime during the pandemic occurred due to movement restrictions which impacted disproportionately upon crime in areas where it was previously concentrated. Competing interests The authors declare that they have no competing interests. Funding The work was funded by the 10.13039/100014013 UKRI open call for research on COVID-19 under grant ES/00445X/1 from the 10.13039/501100000269 Economic and Social Research Council . Availability of data and materials The data are publicly available and retrievable from the open online data portal for England, Wales and Northern Ireland. Code to replicate data downloads, handling and analyses is openly available (https://github.com/langtonhugh/covid_spatial). Appendix A Appendix Fig. 7 Counts by offence type in England and Wales during the lockdown period. Fig. 7 Fig. 8 Monthly generalized Gini coefficient trends by offence type in England and Wales during the lockdown period. Fig. 8 1 Further information on these crime categories can be obtained from https://www.police.uk/pu/contact-the-police/what-and-how-to-report/what-report/. ==== Refs References Abrams D. COVID and crime: An early empirical look, research paper no. 20–49, University of Pennsylvania Carey law School, 14 august 2020 University of Pennsylvania: Institute for Law and Economics Adepeju M. Langton S. Bannister J. Nchored k-medoids: A novel adaptation of k-medoids further refined to measure long-term instability in the exposure to crime Journal of Computational Social Science. 2021 10.1007/s42001-021-00103-1 Andresen M.A. Curman A.S. Linning S.J. The trajectories of crime at places: Understanding the patterns of disaggregated crime types Journal of Quantitative Criminology 33 3 2017 427 449 Andresen M.A. Hodgkinson T. Somehow I always end up alone: COVID-19, social isolation and crime in Queensland, Australia Crime Science 9 25 2020 1 20 Andresen M.A. Linning S.J. Malleson N. Crime at places and spatial concentrations: Exploring the spatial stability of property crime in Vancouver BC, 2003-2013 Journal of Quantitative Criminology 33 2 2017 255 275 Andresen M.A. Malleson N. Testing the stability of crime patterns: Implications for theory and practice Journal of Research in Crime and Delinquency 48 1 2011 58 82 Ashby M. Initial evidence on the relationship between the coronavirus pandemic and crime in the United States Crime Science 9 2020 1 16 Ashby M. Changes for police calls for service during the early months of the 2020 Coronavirus pandemic Policing 14 4 2020 1054 1072 Baggs M. Black lives matter in the UK: “We”re still not being heard’ 2020 BBC News online 25 August 2020, at https://www.bbc.co.uk/news/newsbeat-53812576 Bannister J. Bates E. Kearns A. Local variance in the crime drop: A longitudinal study of neighbourhoods in greater Glasgow, Scotland The British Journal of Criminology 58 1 2018 177 199 Borrion H. Kurland J. Tilley N. Chen P. Measuring the resilience of criminogenic ecosystems to global disruption: A case-study of covid-19 in China PLoS One 15 10 2020 e0240077 Caliński T. Harabasz J. A dendrite method for cluster analysis Communications in Statistics-Theory and Methods 3 1 1974 1 27 Campedelli G.M. Aziani A. Favarin S. Exploring the effects of covid-19 containment policies on crime: An empirical analysis of the short-term aftermath in los Angeles 2020 Campedelli G.M. Favarin S. Aziani A. Piquero A.R. Disentangling community-level changes in crime trends during the covid-19 pandemic in Chicago Crime Science 9 1 2020 1 18 Crown Prosecution Service 6500 coronavirus-related prosecutions in first six months of the pandemic https://www.cps.gov.uk/cps/news/6500-coronavirus-related-prosecutions-first-six-months-pandemic 2021 Curman A.S. Andresen M.A. Brantingham P.J. Crime and place: A longitudinal examination of street segment patterns in Vancouver, bc Journal of Quantitative Criminology 31 1 2015 127 147 Dai M. Xia Y. Han R. The impact of lockdown on police service calls during the COVID-19 pandemic in China Policing 2021 10.1093/police/paab007 Dixon A. Farrell G. A year of COVID-19 and crime in England and Wales Statistical bulletin on crime and COVID-19, issue 14 2021 University of Leeds Dixon A. Halford E. Farrell G. Spatial distributive justice and crime in the COVID-19 pandemic Statistical bulletin on crime and COVID-19, issue 2, 27 July 2020 2020 University of Leeds Dixon A. Tlley N. Fly-tipping during the pandemic Statistical bulletin on crime and COVID-19, issue 13, 26 November 2020 2020 University of Leeds Eck J.E. Gersh J.S. Taylor C. Finding crime hot spots through repeat address mapping Analyzing crime patterns: Frontiers of practice 2000 49 64 Estévez-Soto P.R. Crime and covid-19: Effect of changes in routine activities in Mexico City 2020 Felson M. Jiang S. Xu Y. Routine activity effects of the covid-19 pandemic on burglary in Detroit, march, 2020 Crime Science 9 1 2020 1 7 Genolini C. Alacoque X. Sentenac M. Arnaud C. Kml and kml3d: R packages to cluster longitudinal data Journal of Statistical Software 65 4 2015 1 34 http://www.jstatsoft.org/v65/i04/ Genolini C. Alacoque X. Sentenac M. Arnaud C. others Kml and kml3d: R packages to cluster longitudinal data Journal of Statistical Software 65 4 2015 1 34 Genolini C. Falissard B. KmL: A package to cluster longitudinal data Computer Methods and Programs in Biomedicine 104 3 2011 e112 e121 21708413 Gerell M. Kardell J. Kindgren J. Minor covid-19 association with crime in Sweden Crime Science 9 1 2020 1 9 Griffiths E. Chavez J.M. Communities, street guns and homicide trajectories in Chicago, 1980–1995: Merging methods for examining homicide trends across space and time Criminology 42 4 2004 941 978 Halford E. Dixon A. Farrell G. Malleson N. Tilley N. Crime and coronavirus: Social distancing, lockdown, and the mobility elasticity of crime Crime Science 9 1 2020 1 12 Hodgkinson T. Andresen M.A. Show me a man or a woman alone and I’ll show you a saint: Changes in the frequency of criminal incidents during the covid-19 pandemic Journal of Criminal Justice 69 2020 101706 32834176 Ignatans D. Pease K. Distributive justice and the crime drop Andresen M.A. Farrell G. The criminal act: The role and influence of routine activity theory 2015 Palgrave Macmillan London 77 87 Ignatans D. Pease K. On whom does the burden of crime fall now? Changes over time in counts and concentration International Review of Victimology 22 1 2016 55 63 Langton S. Dixon A. Farrell G. Six months in: Pandemic crime trends in England and wales Crime Science 10 1 2021 1 16 Langton S. Solymosi R. Open street map for crime and place OSF Preprints https://osf.io/a96y7 2020 Langton Samuel Crime and Anti-Social Behaviour in Greater London Statistical bulletin on crime and COVID-19, issue 3, 31 July 2020 2020 University of Leeds Malleson N. Andresen M.A. Spatial-temporal crime hotspots and the ambient population Crime Science 4 10 2015 1 8 Malleson N. Andresen M.A. Exploring the impact of ambient population measures on London crime hotspots Journal of Criminal Justice 46 2016 52 63 de la Mayir B. Hoehn-Velasco J.R.L. Silverio-Murillo A. Druglords don’t stay at home: COVID-19 pandemic and crime patterns in Mexio city Journal of Criminal Justice 72 2021 10.1016/j.jcrimjus.2020.101745 McVie S. Norris P. Pillinger R. Increasing inequality in experience of victimization during the crime drop: Analysing patterns of victimization in Scotland from 1993 to 2014–15 The British Journal of Criminology 60 3 2020 782 802 Mohler G. Bertozzi A.L. Carter J. Short M.B. Sledge D. Tita G.E. …Brantingham P.J. Impact of social distancing during covid-19 pandemic on crime in Los Angeles and Indianapolis Journal of Criminal Justice 101692 2020 Newton A. Macro-level generators of crime, including parks, stadiums, and transit stations 2018 Office for National Statistics Coronavirus and crime in England and Wales: August 2020. Statistical bulletin, 26 August 2020 2020 Office for National Statistics Office of National Statistics Census geography 2021 https://www.ons.gov.uk/methodology/geography/ukgeographies/censusgeography#super-output-area-soa on Economics and the Environment Padgham M. Lovelace R. Salmon M. Rudis B. Osmdata Journal of Open Source Software 2 14 2017 Payne J. Morgan A. Piquero A. COVID-19 and social distancing measures in Queensland, Australia, are associated with short-term decreases in recorded violent crime Journal of Experimental Criminology 2020 10.1007/s11292-020-09441-y Payne J. Morgan A. Piquero A. Exploring regional variability in the short-term impact of COVID-19 on property crime in Queensland, Australia Crime Science 10 7 2021 1 20 Piquero A.R. Riddell J.R. Bishopp S.A. Narvey C. Reid J.A. Piquero N.L. Staying home, staying safe? A short-term analysis of covid-19 on Dallas domestic violence American Journal of Criminal Justice 2020 1 35 R Core Team R: A language and environment for statistical computing 2020 R Foundation for Statistical Computing Vienna, Austria Shayegh S. Malpede M. Staying home saves lives, really! Staying home saves lives, really! 2020 Stickle B. Felson M. Crime rates in a pandemic: The largest criminological experiment in history American Journal of Criminal Justice 45 4 2020 525 536 32837162 Tompson L. Johnson S. Ashby M. Perkins C. Edwards P. UK open source crime data: Accuracy and possibilities for research Cartography and Geographic Information Science 42 2 2015 97 111 Trickett A. Ellingworth D. Hope T. Pease K. Crime victimization in the eighties: Changes in area and regional inequality British Journal of Criminology 35 3 1995 343 359 Van Dijk J.J.M. Tseloni A. Farrell G. The international crime drop: New directions in research 2012 Palgrave Macmillan New York

35358420

PMC9753225

NO-CC CODE

2022-12-16 23:25:34

no

Lancet. 2022 Mar 28 30 April-6 May; 399(10336):1692

latin-1

Lancet

10.1016/S0140-6736(22)00368-3

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)00299-3 10.1016/S0140-6736(21)00299-3 Perspectives Easy in, tough out: the dam of health-care racism Choo Esther a a Department of Emergency Medicine, Oregon Health and Science University, Portland, OR 97239, USA 11 2 2021 13-19 February 2021 11 2 2021 397 10274 570570 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcUntil the 1980s, the practice of turning away patients on the basis of their ability to pay, without evaluating or treating them, was legal and frequent in US hospitals. A 1986 study at Cook County Hospital, a public hospital in Chicago, noted that 89% of patients transferred from private hospitals to their facility were Black and Hispanic, and 87% of them were transferred because they did not have adequate insurance. Mortality was disproportionately high among those transferred, likely due to delays in stabilisation and definitive care. This practice was ultimately disrupted—although it continues today—not by institutions recognising its unethical nature and voluntarily halting it, but by the passage of the federal Emergency Medical Treatment and Active Labor Act, which requires hospitals to stabilise and treat patients before transfer, irrespective of insurance status. It took an act of Congress to change this fundamentally racist, inequitable practice. I thought of this history recently because of a conversation with a colleague about hospital equipment that retains components labelled “master” and “slave”. “Why not remove those names?”, I asked. “It's written into the manuals and the systems and the technology”, he told me. “Removing those terms everywhere would take an act of Congress.” Elements within medicine that reinforce racism—eg, racist clinical calculators, patient transfer practices, equipment design, poor representativeness of clinical trials—are put into everyday practice easily, frequently, with surprisingly little challenge, barriers, or argument. The dismantling of racism in medicine, by contrast, is laborious, gradual, and painstaking, often done against great resistance. This easy-in, tough-out system of inequity is like a dam: it is, by nature, designed for accumulation, for retention of racism. When health-care organisations establish new policies, programmes, and decision-making processes with an equity lens, these gradual inflow improvements cannot change what is happening in total when there has not been any corresponding removal of already accumulated policies, programmes, and processes that reinforce racism. © 2021 ErichFend/Shutterstock 2021 This phenomenon also applies to the care individual patients receive. Physician–scientist Tamorah Lewis has described “cumulative deprioritization” in health-care settings: as health professionals make constant decisions about allocating limited resources of supply, time, and attention, “implicit bias and deep subconscious beliefs about who matters and who is more valuable/dispensable creep in”. The many small delays or relative disadvantages in the delivery of care compound and lead to inequitable outcomes by race. And yet this gradual accumulation can go unseen by health-care providers at a macro level. Thus, in patient care, I would argue, the easy-in harms Lewis describes are matched with a tough-out system of corrections. If we recognise the disparities in patients' outcomes on the basis of race and ethnicity, and as we begin to understand the mechanistic, logistical, and structural underpinnings, what actions are needed to prevent or course correct in real time at the bedside? What are the release valves for racism? Medicine recently lost a physician in a scenario that prompts reflection on such questions. Susan Moore was a Black woman and a physician who contracted COVID-19 in November, 2020. During her initial hospitalisation, she posted online, voicing her concerns about facing racial discrimination. She described how pain medication was withheld and how her symptoms were dismissed, leaving her feeling that she had to beg for appropriate treatment. “I put forth and I maintain if I was white”, Moore said in a video, “I wouldn't have to go through that.” Shortly after her discharge, she presented to another hospital, febrile and hypotensive. She was intubated within days and died on Dec 20, 2021. When inequities in care begin to accumulate before our very eyes, what effective countermeasures are needed to dismantle them in the moment? In the aftermath of Moore's tragic death, will medicine begin to identify and use such measures? There is no question that in both health systems and the daily actions of clinicians, we remain far less facile and assured in removing racism than we are in incorporating it. And until that imbalance changes, our statements and aspirations about being anti-racist cannot connect meaningfully to progress. ==== Refs Further reading Schiff RL Ansell DA Schlosser JE Transfers to a public hospital N Engl J Med 314 1986 552 557 3945293 Cerdeña JP Plaisime MV Tsai J From race-based to race-conscious medicine: how anti-racist uprisings call us to act Lancet 396 2020 1125 1128 33038972 @TamorahLewisMD https://twitter.com/TamorahLewisMD/status/1343329088325476352?s=20 Dec 27, 2020 Maybank A Jones CP Blackstock U Crear Perry J Say her name: Dr. Susan Moore The Washington Post Dec 26, 2020

33581813

PMC9753226

NO-CC CODE

2022-12-16 23:25:34

no

Lancet. 2021 Feb 11 13-19 February; 397(10274):570

utf-8

Lancet

10.1016/S0140-6736(21)00299-3

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)00197-5 10.1016/S0140-6736(21)00197-5 Correspondence Breastfeed or be vaccinated—an unreasonable default recommendation Merewood Anne a Bode Lars b Davanzo Riccardo cd Perez-Escamilla Rafael e a Center for Health Equity, Education and Research, Boston University School of Medicine, Boston, MA, USA b Larsson-Rosenquist Foundation Mother-Milk-Infant Center of Research Excellence, University of California San Diego, San Diego, CA, USA c Institute for Maternal and Child Health, Trieste, Italy d Technical Panel on Breastfeeding, Ministry of Health, Rome, Italy e Yale School of Public Health, Yale University, New Haven, CT 06510, USA 4 2 2021 13-19 February 2021 4 2 2021 397 10274 578578 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcBreastfeeding promotes the good health of mothers and infants and is a crucial international public health issue. None of the COVID-19 vaccines currently in phase 3 trials have been trialled in breastfeeding women.1 Pfizer's current recommendation states that breastfeeding women should “ask your doctor or pharmacist for advice before you receive this vaccine”.2 We fear that, like Public Health England's initial recommendation not to vaccinate lactating women,3 many clinicians will recommend against taking the vaccine when breastfeeding, as is the default in the absence of data, as though breastfeeding is a neutral health decision. Those individuals immediately impacted by the advice, of course, are breastfeeding women working as front-line health-care providers and caregivers, who might be required to choose between their own health, their infant's health, and potentially, their job because not being vaccinated might be disadvantageous in the workplace. Although the UK has reversed its stand and now advises offering the vaccine to breastfeeding women,3 concerns remain because the vaccine has not been tested in lactating women, not because of empirical evidence or biological plausibility for harm. However, we want to highlight that human milk is not a vector for severe acute respiratory syndrome coronavirus 2.4 Moreover, the milk contains antibodies that could potentially protect the breastfed baby from COVID-19.5 We need research to determine whether coronavirus vaccines in general, and mRNA vaccines in particular, enter the milk and transfer to the infant. Even if they do, there seems no plausible reason to recommend against vaccination for breastfeeding women. Antibodies generated in response to the vaccine should protect the breastfeeding women and the breastfed infants. Perhaps with this protection in mind, the American College of Obstetricians and Gynecologists stated upfront that ”COVID-19 vaccines be offered to lactating individuals similar to non-lactating individuals when they meet criteria for receipt of the vaccine”. To improve maternal–infant health and maintain public confidence in vaccines in handling this pandemic and preparing for the next, vaccine manufacturers and regulators must work closely with lactation scientists, infectious disease specialists, and public health experts to assess vaccine safety in breastfeeding women at early stages of product development. It is encouraging that many nations, including England, are now adopting a more positive tone around vaccine recommendations for breastfeeding women, but in many cases the finer points of the recommendation will still lie with individual providers or institutions. LB reports grants from the Family Larsson Rosenquist Foundation, and personal fees from Medela, the Nestle Nutrition Institute, the Abbott Nutrition Institute, and Prolacta, outside this Correspondence. All other authors declare no competing interests. ==== Refs References 1 Doshi P Will covid-19 vaccines save lives? Current trials aren't designed to tell us BMJ 371 2020 m4037 2 Pfizer Package leaflet: information for the recipient. COVID-19 mRNA Vaccine BNT162b2 concentrate for solution for injection https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/948518/Information_for_UK_recipients_on_PfizerBioNTech.pdf December, 2020 3 Rimmer A Covid-19: breastfeeding women can have vaccine after guidance turnaround BMJ 372 2021 n64 33419782 4 Chambers C Krogstad P Bertrand K Evaluation for SARS-CoV-2 in breast milk from 18 infected women JAMA 324 2020 1347 1348 32822495 5 Fox A Marino J Amanat F Robust and specific secretory IgA against SARS-CoV-2 detected in human milk iScience 23 2020 101735

33549167

PMC9753227

NO-CC CODE

2022-12-16 23:25:34

no

Lancet. 2021 Feb 4 13-19 February; 397(10274):578

utf-8

Lancet

10.1016/S0140-6736(21)00197-5

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)00024-6 10.1016/S0140-6736(21)00024-6 Comment Research in forced displacement: guidance for a feminist and decolonial approach Singh Neha S a Lokot Michelle a Undie Chi-Chi b Onyango Monica A c Morgan Rosemary d Harmer Anne e Freedman Jane f Heidari Shirin gh a Health in Humanitarian Crises Centre, London School of Hygiene & Tropical Medicine, London WC1H 9SH, UK b Population Council, Nairobi, Kenya c Department of Global Health, Boston University School of Public Health, Boston, MA, USA d Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA e Elrha, London, UK f University of Paris 8, Paris, France g GENDRO, Geneva, Switzerland h Global Health Centre, Graduate Institute of International and Development Studies, Geneva, Switzerland 11 2 2021 13-19 February 2021 11 2 2021 397 10274 560562 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcThe COVID-19 pandemic has deepened inequities and undermined health, human rights, and gender equality for forcibly displaced populations.1, 2 The United Nations Refugee Agency estimates that, at the end of 2019, there were 79·5 million people forcibly displaced as a result of persecution, conflict, violence, human rights violations, or events seriously disturbing public order.3 Evidence about the needs of these populations is crucial to tailor effective and equitable responses, but data collection efforts are faced with considerable new challenges during the COVID-19 pandemic. Many researchers are attempting to overcome such challenges by collecting data remotely, but doing so creates ethical and practical concerns that risk perpetuating gender, racial, and other inequities. For example, the gender divide in mobile phone ownership,4 internet access, and digital literacy creates barriers to data collection from women, further silencing their voices and that of other groups without access to these technologies. Overcrowded living spaces, mobility restrictions, and lack of autonomy over technology use (due to COVID-19, gender norms, or both) exacerbate ethical concerns regarding confidentiality, privacy, and safety during remote data collection. The ongoing pandemic has also exposed persisting power hierarchies between researchers and forcibly displaced populations. These populations experience power asymmetries in their position as the so-called beneficiaries of humanitarian research and action, and are often excluded from participation in decision making about the research intended to benefit their communities. Forcibly displaced women and girls are consistently categorised as vulnerable and needing protection or rescuing, which takes away their agency and power of action, while risking exploitation or abuse by the same humanitarian actors that supposedly aim to protect them. Recent efforts to address these hierarchies through the process of localisation—ie, recognising, respecting, and strengthening leadership by local authorities and the capacity of local civil society in humanitarian action to better address the needs of affected populations and better prepare actors for future humanitarian responses—have been criticised for neglecting the insidious effects of sexism and racism, both intrinsically linked to colonial legacies.5 The perceived urgency to collect data remotely also exposes neocolonial power hierarchies between researchers in affected settings and those in resource-rich settings, where funding is often concentrated. Researchers from advanced economies predominantly define the research questions with little or tokenistic consultation of in-country researchers or communities.6, 7 With COVID-19-related movement restrictions, research can typically only be done by collecting data remotely or by delegating data collection to in-country researchers. In the haste to produce evidence, interactions can become one sided or top down, as those in higher hierarchical positions issue directives to front-line actors. The new nature of these interactions also risks the so-called ethics dumping,8 that is, off-loading risk to in-country researchers by asking them to facilitate data collection under the unique challenges presented by the COVID-19 pandemic. These hurdles demand that researchers confront power hierarchies in knowledge production processes. We propose the application of feminist values to address these concerns. Although there are many feminist strands, feminists are united in seeking to address unequal power hierarchies and striving for social and environmental justice.9, 10 Feminist researchers advocate for intersectional analysis that centres the voices and knowledge of communities, embedding decolonial lenses and ethics of care approaches that value people more than they value data.11, 12 Feminist research explicitly examines gendered and colonial power hierarchies at play in the research process, and is grounded in reciprocal engagement with communities to equalise power dynamics. By advocating a feminist approach, we propose moving beyond the performative dimensions of being gender-sensitive and decolonial, towards understanding what it means to equitably share power within research collaborations in a meaningful way that challenges traditional methods of knowledge production.6 The COVID-19 pandemic presents a crucial opportunity for researchers working with forcibly displaced populations to rethink their traditional approaches. Applying feminist values to data collection during COVID-19 and beyond requires putting at the centre the knowledge of those from whom data are being collected. We provide key recommendations (panel ) and a detailed checklist (appendix) for applying a feminist approach that takes into account ethical, gender, and decolonisation considerations when collecting data in forced displacement settings.Panel Recommendations for a feminist approach during research in forced displacement settings Stage 1: conceptualisation of research and data collection • Establish equitable partnerships to conduct research on topics that are relevant and beneficial to all members of communities Stage 2: funding • Meaningfully involve all researchers in budget preparation and ensure an equitable allocation of resources Stage 3: research design • Consider the political, social, economic, and historical contexts and power hierarchies of the research setting and plan for the meaningful participation of individuals and communities with less power Stage 4: collecting data • Consider how gendered and colonial power hierarchies might be reinforced by capacity building of front-line researchers and engagement with communities • Ensure collection of data on gender to allow for capturing gender and other inequities Stage 5: data analysis and dissemination • Engage front-line researchers and study populations in conducting intersectional gender analysis, as well as in interpretation, writing, and dissemination of findings • Use findings to challenge unjust systems and policies and deliver gender transformative and equitable programmes Please refer to the appendix for more details. At the time of research conceptualisation, applying a feminist approach equates to meaningfully engaging forcibly displaced populations so that research is relevant to their concerns, instead of solely focusing on what researchers believe is important.7 This engagement must include considerations of how gender intersects with other axes of power such as race, ethnicity, or displacement status to shape individual experiences. Taking such steps at the research conceptualisation stage allows ethical approaches to codeveloping recruitment and data collection strategies, treating forcibly displaced populations as more than data providers, and ensuring the participants' privacy and confidentiality.13 Consideration of power hierarchies includes reflection on the dynamics between front-line researchers, who hold power despite being so-called local participants, and communities, leading to concrete steps to reduce power imbalances. Power hierarchies and politics also shape how data are analysed, published, and shared. Choices on which data are deemed relevant, how the analysis is presented, and how authorship is decided are all arenas in which power is exercised to prioritise some voices and silence others. Feminist values emphasise meaningful decision making and relational engagement, from research conceptualisation to publication and beyond. Dismantling well established data collection practices, especially in forced displacement settings, requires a sustained commitment from all parties in the research ecosystem and changes to the architecture that enables these practices. COVID-19 has given us the opportunity to reflect on and challenge long-existing power hierarchies within research—a process that is needed to address lingering colonial and patriarchal power relations and avoid ethical pitfalls. We believe that applying a feminist lens is not merely about demolishing problematic structures, but also about collaboratively building up new ones for a more just world. Supplementary Material Supplementary appendix NSS reports salary support from the RECAP project by UK Research and Innovation as part of the Global Challenges Research Fund, grant number ES/P010873/1. All other authors declare no competing interests. The thinking underpinning this Comment began in a virtual workshop on the Ethical and Gender Considerations in Remote Data Collection and Research in Forced Displacement Settings, hosted by the authors on June 29, 2020, with the support of the Global Health Centre, the Graduate Institute of International and Development Studies, and GENDRO. ==== Refs References 1 Médecins Sans Frontières Stigma and disrupted care: facing COVID-19 in Bangladesh https://www.msf.ie/article/stigma-disrupted-care-facing-covid-19-bangladesh 2020 2 Mballa C Ngebeh J De Vriese M Drew K Parr A Undie C UNHCR and partner practices of community-based protection across sectors in the East and Horn of Africa and the Great Lakes Region 2020 UNHCR and Population Council Nairobi, Kenya https://knowledgecommons.popcouncil.org/departments_sbsr-rh/1310/ 3 United Nations High Commissioner for Refugees Global trends: forced displacement in 2019 https://www.unhcr.org/uk/statistics/unhcrstats/5ee200e37/unhcr-global-trends-2019.html June, 2020 4 United Nations High Commissioner for Refugees The digital lives of refugees: what's next? https://www.unhcr.org/jo/12182-the-digital-lives-of-refugees-whats-next.html 2019 5 Slim H Is racism part of our reluctance to localise humanitarian action? https://odihpn.org/blog/is-racism-part-of-our-reluctance-to-localise-humanitarian-action/ 2020 6 Brun C Lund R Real-time research: decolonising research practices–or just another spectacle of researcher–practitioner collaboration? Dev Pract 20 2010 812 826 7 Lokot M The space between us: feminist values and humanitarian power dynamics in research with refugees Gend Dev 27 2019 467 484 8 Schroeder D Cook J Hirsch F Fenet S Muthuswamy V Ethics dumping: case studies from North-South research collaborations 2018 Springer Nature Cham, Switzerland 9 Brooks A Hesse-Biber SN An invitation to feminist research Hesse-Bilber SN Feminist research practice: a primer 2007 SAGE Publishing Thousand Oaks, USA 10 Davies SE Harman S Manjoo R Tanyag M Wenham C Why it must be a feminist global health agenda Lancet 393 2019 601 603 30739696 11 Kapilashrami A Hankivsky O Intersectionality and why it matters to global health Lancet 391 2018 2589 2591 30070211 12 Pinet M Leon-Himmelstine C How can COVID-19 be the catalyst to decolonise development research? 2020 Oxfam https://oxfamblogs.org/fp2p/how-can-covid-19-be-the-catalyst-to-decolonise-development-research 13 Calia C Reid C Guerra C Ethical challenges in the COVID-19 research context: a toolkit for supporting analysis and resolution Ethics Behav 1 2021 1 16

33581808

PMC9753228

NO-CC CODE

2022-12-16 23:25:34

no

Lancet. 2021 Feb 11 13-19 February; 397(10274):560-562

utf-8

Lancet

10.1016/S0140-6736(21)00024-6

oa_other

==== Front Med Clin (Barc) Med Clin (Barc) Medicina Clinica 0025-7753 1578-8989 Elsevier España, S.L.U. S0025-7753(21)00013-0 10.1016/j.medcli.2020.12.005 Brief Report Cardiac biometric variables and arrhythmic events during COVID-19 pandemic lockdown in patients with an implantable cardiac monitor for syncope work-up Variables biométricas cardiacas y eventos arrítmicos durante el confinamiento por la pandemia de COVID-19 en pacientes portadores de un monitor cardiaco implantable para el diagnóstico de síncopeFrancisco-Pascual Jaume a⁎ Rivas-Gándara Núria a⁎⁎ Santos-Ortega Alba a Pérez-Rodón Jordi a Benito Begoña a Belahnech Yassin a Ferreira-González Ignacio b a Unitat d’Arritmies. Servei de Cardiologia. Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca, Universitat Autònoma de Barcelona, CIBER-CV, Spain b Servei de Cardiologia, Hospital Universitari Vall d’Hebron. Vall d’Hebron Institut de Recerca, Universitat Autònoma de Barcelona, CIBER-ESP, Spain ⁎ Corresponding author. ⁎⁎ Co-corresponding author. 26 2 2021 21 5 2021 26 2 2021 156 10 496499 17 8 2020 2 12 2020 © 2021 Elsevier España, S.L.U. All rights reserved. 2021 Elsevier España, S.L.U. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Aim To assess the changes induced by the COVID-19 lockdown on cardiac biometric variables recorded using an implantable cardiac monitor (ICM) in a patient population monitored for syncope work-up, as well to assess whether there has been an effect on arrhythmic events among the patients. Methods Longitudinal cohort study. We included 245 adult patients monitored with an ICM indicated for the investigation of syncope. The records from days 1 to 12 March 2020 (prior to the institution of lockdown by the state government) with days 16 to 28 March 2020 were compared. Results Daily physical exercise reduced markedly after the imposition of lockdown (132 [55–233] minutes vs. 78 [21–154] minutes). The mean daytime HR prior to lockdown was 77 [69–85] bpm, whereas during lockdown it was 74 [66–81] bpm. During the lockdown period, a drop in the variability in heart rate (114 [94–136] ms vs. 111 [92–133] ms) was observed. Although the incidence of AF was similar over both periods, the daily AF burden was significantly higher post-lockdown (405 [391–425] minutes vs. 423 [423–537] minutes). No differences in the number of other arrhythmias were found. Conclusions The establishment of mandatory lockdown has led to a marked drop in daily physical activity in this population which probably explains changes observed in other cardiac biometric variables. Although, in the short term, we have not documented an increased risk of arrhythmia, we cannot rule out an effect in the medium to long term or in other populations of at-risk patients. Objetivo Evaluar los cambios inducidos por el confinamiento durante la pandemia de COVID-19 en las variables biométricas cardiacas registradas, utilizando un monitor cardíaco implantable (ICM) en una población de pacientes monitorizada para el diagnóstico de síncope, así como evaluar si ha habido un efecto sobre los eventos arrítmicos. Métodos Estudio de cohorte prospectivo. Se incluyeron 245 pacientes adultos monitorizados con un ICM indicado para la investigación del síncope. Se compararon los registros de los días uno al 12 de marzo del 2020 (antes del establecimiento del confinamiento por parte del gobierno estatal) con los días 16 al 28 de marzo del 2020. Resultados El ejercicio físico diario se redujo notablemente después de la imposición del confinamiento (132 [55 a 233] vs. 78 [21 a 154] min). La frecuencia cardiaca diurna media antes del confinamiento fue de 77 (69 a 85) lpm, mientras que durante el mismo fue de 74 (66 a 81) lpm. Durante el período de confinamiento, se observó una disminución de la variabilidad de la frecuencia cardiaca (114 [94 a 136] vs. 111 [92 a 133] ms). Aunque la incidencia de fibrilación auricular (FA) fue similar en ambos períodos, la carga diaria de FA fue significativamente mayor después del bloqueo (405 [391 a 425] vs. 423 [423 a 537] min). No se encontraron diferencias en el número de otras arritmias. Conclusiones El establecimiento de un confinamiento obligatorio ha provocado un marcado descenso de la actividad física diaria en esta población, lo que probablemente explica los cambios observados en otras variables biométricas cardiacas. Si bien, a corto plazo, no se ha documentado un aumento del riesgo de arritmia, no podemos descartar un efecto a medio-largo plazo o en otras poblaciones de pacientes de riesgo. Keywords COVID-19 Lockdown Physical exercise Arrhythmic burden Palabras clave Covid-19 Confinamiento Actividad física Carga arritmica ==== Body pmcBackground On March 14, 2020, the Spanish state government declared a state of emergency and instituted mandatory home lockdown for the population with a view to combatting the exceptional situation of the SARS-CoV-2 pandemic. This state of affairs, which has doubtless had an effect on controlling the epidemic, has also had collateral implications in many aspects of today's society. Several of these changes may have a significant impact on the population's health. For example, we have observed a clear reduction in pollution levels in major cities following the reduction of travel.1 To date, the impact on arrhythmic burden, physical exercise and cardiac biometric variables during lockdown and its possible consequences is unknown. Prolonged electrocardiographic monitoring is mainly indicated for the work-up of syncope,2 palpitations3 and cryptogenic stroke.4, 5 New generations of implantable cardiac monitors (ICMs) are subcutaneous implantable devices allowing not only for electrocardiographic monitoring, but also for other biometric variables as daily physical exercise. The goal of this study was to assess the changes induced by the COVID-19 lockdown on cardiac biometric variables recorded using an ICM in a patient population monitored for syncope work-up, as well to assess whether there has been an effect on arrhythmic events among the patients. The study complies with the Helsinki declaration and was approved by the center's ethics committee. Methods In this longitudinal study carried out in a tertiary Spanish hospital, we included all consecutive adult patients monitored with a latest generation ICM (Reveal Linq™, Medtronic, Inc. Minneapolis. USA) indicated for the investigation of syncope as per applicable clinical practice guidelines2 and who had remotely sent the device records corresponding to the period from 1 to 28 March 2020 to the hospital. To summarize, ICM is indicated in patients in whom the full syncope work-up does not reach a certain diagnosis and they are considered at risk due to the clinical characteristics of the syncope, the presence of structural heart disease or conduction disorders.2, 6 In selected patients with repeated reflex syncope it is indicated to guide treatment. The ICM was programmed with the preconfigured settings for syncope. With these settings, the detection of atrial fibrillation is based both on irregularity of the RR interval and a P wave morphology indicated with a minimum episode duration of 6 min. HR variability was measured by the device calculating the median ventricular interval every 5 min. It then calculates and plots a variability value (in ms) for each day (supplementary figure). We compared the records from days 1 to 12 March 2020 (prior to the institution of lockdown by the state government) with days 16–28 March 2020. The categorical variables are presented as an absolute number (N) and percentages and the continuous quantitative variables are presented as a median and interquartile range [IQR]. For the statistical analysis, we used paired data analysis. 95% confidence interval (CI) for the median of the differences between during and prior the lockdown period is reported. Non-paired data analysis was used where appropriate. All of the statistical tests were performed using Stata, version 15.1.0 (StataCorp LLC College Station, Texas, USA.). Results We included a total of 245 patients. Table 1 summarizes the patients’ baseline characteristics. The median patient age at the time of inclusion was 64.5 years [IQR 45.0–79.0] and 129 (53.1%) were female.Table 1 Baseline characteristics. Table 1Variablea Total patients (n = 245) Clinical characteristics  Age (years) 64.5 [45.0–79.0]  Female (n, %) 129 (53.1%)  Hypertension (n, %) 121 (49.8%)  Dyslipidemia ((n, %) 90 (37.0%)  Diabetes (n, %) 46 (18.9%)  Active smoking (n, %) 32 (13.2%)  CKD (GFR < 60) (n, %) 37(15.2%)  History of stroke (n, %) 31 (12.8%)  Peripheral artery disease (n, %) 22 (9.1%)  Ischemic heart disease (n, %) 48 (19.8%)  History of ST elevation (n, %) 20 (8.2%)  History of HF (n, %) 29 (11.9%)  History of AF (n,%) 48 (19.8%)  LVEF (%) 59 [55–62] Medication  Beta blocker (n, %) 83 (34.2%)  ACEI/ARB (n, %) 114 (46.9%)  Calcium channel blocker (n, %) 29 (11.9%)  Diuretics (n, %) 70 (28.8%) CKD: chronic kidney disease. GFR: glomerular filtration rate. HF: heart failure. LVEF: left ventricular ejection fraction. aThe quantitative variables are expressed as a median [interquartile range]. Daily physical exercise reduced markedly after the imposition of lockdown (132 [55–233] minutes prior the lockdown vs. 78 [21–154] minutes during to lockdown, median of the differences 33 [7–67] minutes; 95% CI from 25 to 41 min) (Panel A, Fig. 1 ). The mean daytime HR prior to lockdown was 77 [69–85] bpm, whereas during lockdown it was 74 [66–81] bpm (median of the differences of 3 [0.2–5] bpm; 95% CI from 2 to 3.3 min). During the lockdown period, a slight but significant drop in the variability in heart rate (114 [94–136] ms vs. 111 [92–133] ms; median of the differences of 3 [−3–9] ms; 95% CI from 1 to 5 ms) was observed. This drop was more pronounced over the first few days of lockdown, and tended to normalize subsequently (Panel C Figure).Fig. 1 (A) Time graph showing the evolution of the average of daily physical activity, measured in minutes. Daily physical exercise reduced markedly after the imposition of lockdown. (B) Time graph showing the evolution of the mean daytime heart rate, measured in beats per minute. The mean daytime HR prior to lockdown was higher than during lockdown. (C) Time graph showing the evolution of the mean variability of the HR, measured in milliseconds. During the lockdown period, a slight drop in the variability in heart rate was observed. This drop was more pronounced over the first few days of lockdown and tended to normalize subsequently. (D) Time graph showing the average daily AF burden in patients who presented with AF crises, measured in minutes. Daily AF burden of the episodes was higher post-lockdown. The dashed red line marks March 14, 2020, the beginning of mandatory confinement. (AF: atrial fibrillation). In a total of 45 (18.4%) patients, we recorded atrial fibrillation (AF) during the study period. Of these, 7 (16%) were permanent and the rest were paroxysmal. Although the incidence of AF was similar over both periods, the duration (daily AF burden) of the episodes was higher post-lockdown (405 [391–425] min vs. 433 [423–537] min, with an increase of 28 min; 95% CI from 18 to 45 min) (Panel D and Table 2 ).Table 2 Number of patients with automatic events and activations due to symptoms detected by the ICM prior and during the lockdown. Table 2 Prior to the lockdown During the lockdown AF/AT 36 (14.7%) 34 (13.9%) Bradycardia 7 (2.9%) 6 (2.4%) Asystole 19 (7.8%) 18 (7.3%) NSVT 10 (4.1%) 10 (4.1%) Symptoms 6 (2.4%) 5 (2.0%) AF: atrial fibrillation; AT: Atrial tachycardia; NSVT: unsustained ventricular tachycardia. A total of 7 patients activated the device due to symptoms. Their distribution was homogeneous over both periods. Similarly, we did not find any significant differences in the number of asymptomatic arrhythmias detected by the device (Table 2). Discussion To our knowledge, this is the first case series exploring changes in cardiac biometric variables physical exercise and arrhythmic burden secondary to mandatory lockdown in the context of the SARS-CoV-2 epidemic and exploring its implications on arrhythmic burden. In our project, based on a patient cohort monitored with an ICM as part of syncope work-up, we observed a marked reduction in daily physical exercise after declaration of the state of emergency. This fact probably explains the change observed in other biometric variables examined, such as mean daytime HR, which also significantly dropped in parallel with physical activity. In this sense, it is of interest to mention the evolution of heart rate variability (HRV). HRV is a measurement related to activity of the autonomic nervous system on cardiac function and it has been shown to be of prognostic value in different pathologies, as well as in the general population.7 It is known that prolonged sedentary lifestyle significantly reduces HRV.8, 9 In our study, we observed a marked drop in HRV just after the start of lockdown, which tended to normalize progressively over the following days, which suggests there is a physiological adaptation to the situation. The clinical and prognostic implications of this behavior are unknown. In our project, we carried out an exploratory study on the effect on arrhythmic events. We did not find any significant differences in this population during the first days of lockdown, other than a greater duration of paroxysms of AF, although it would be interesting to assess the possible implications that these changes may have in the medium to long term. Conclusions The declaration of the state of emergency and the instauration of mandatory lockdown have led to a marked drop in daily physical activity in this population which probably explains changes observed in other cardiac biometric variables. Although, in the short term, we have not documented an increased risk of arrhythmia, we cannot rule out an effect in the medium to long term or in other populations of at-risk patients. The data provided should be taken into account when planning similar strategies in the future should they be necessary. Conflicts of interest The authors have no conflicts to disclose. Appendix A Supplementary data The following are the supplementary data to this article: Acknowledgments The authors would like to thank Mr. Raúl Aguilar for his essential assistance in collecting the data, as well as Mr. Felix Ballesteros, Mr. Gonzalo Sánchez and Ms. Elena Fierro from Medtronic España for their technical support, as well as all of the staff in the arrhythmia unit for their daily work with these patients. Appendix A Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.medcli.2020.12.005. ==== Refs References 1 Tobías A. Carnerero C. Reche C. Massagué J. Via M. Minguillón M.C. Changes in air quality during the lockdown in Barcelona (Spain) one month into the SARS-CoV-2 epidemic Sci Total Environ 726 2020 138540 10.1016/j.scitotenv.2020.138540 32302810 2 Brignole M. Moya A. de Lange F.J. Deharo J.-C. Elliott P.M. Fanciulli A. Practical instructions for the 2018 ESC Guidelines for the diagnosis and management of syncope Eur Heart J 39 2018 1883 1948 10.1093/eurheartj/ehy037 29562304 3 Francisco-Pascual J. Santos-Ortega A. Roca-Luque I. Rivas-Gándara N. Pérez-Rodón J. Milà-Pascual L. Diagnostic yield and economic assessment of a diagnostic protocol with systematic use of an external loop recorder for patients with palpitations Rev Esp Cardiol (Engl Ed) 72 2019 473 478 10.1016/j.rec.2018.04.007 29805092 4 Pagola J. Juega J. Francisco-Pascual J. Moya A. Sanchis M. Bustamante A. Yield of atrial fibrillation detection with Textile Wearable Holter from the acute phase of stroke: pilot study of crypto-AF registry Int J Cardiol 251 2018 45 50 10.1016/j.ijcard.2017.10.063 29107360 5 Kirchhof P. Benussi S. Kotecha D. Ahlsson A. Atar D. Casadei B. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS Rev Esp Cardiol (Engl Ed) 70 2017 50 10.1016/j.rec.2016.11.033 28038729 6 Roca-Luque I. Francisco-Pasqual J. Oristrell G. Rodríguez-García J. Santos-Ortega A. Martin-Sanchez G. Flecainide versus procainamide in electrophysiological study in patients with syncope and wide QRS duration JACC Clin Electrophysiol 5 2019 212 219 10.1016/j.jacep.2018.09.015 30784693 7 Task Force of the European Society of Cardiology the North American Society of Pacing Electrophysiology Heart rate variability: standards of measurement, physiological interpretation and clinical use Circulation 93 1996 1043 1065 8598068 8 Zaffalon Júnior J.R. Viana A.O. de Melo G.E.L. de Angelis K. The impact of sedentarism on heart rate variability (HRV) at rest and in response to mental stress in young women Physiol Rep 6 2018 e13873 10.14814/phy2.13873 30238692 9 Pope Z.C. Gabriel K.P. Whitaker K.M. Chen L.Y. Schreiner P.J. Jacobs D.R. Association between objective activity intensity & heart rate variability Med Sci Sports Exerc 52 2020 1314 1321 10.1249/MSS.0000000000002259 32427750

33642036

PMC9753229

NO-CC CODE

2022-12-16 23:25:34

no

Med Clin (Barc). 2021 May 21; 156(10):496-499

utf-8

Med Clin (Barc)

10.1016/j.medcli.2020.12.005

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)01956-5 10.1016/S0140-6736(21)01956-5 Correspondence Telehealth use in antenatal care? Not without women's voices Galle Anna a Semaan Aline b Asefa Anteneh b Benova Lenka b a Department of Public Health and Primary Care, Ghent University, Ghent 9000, Belgium b Department of Public Health, Institute of Tropical Medicine, Antwerp, Belgium 14 10 2021 16-22 October 2021 14 10 2021 398 10309 14051406 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcKirsten R Palmer and colleagues1 assessed integrated telehealth for antenatal care in Australia during the early COVID-19 pandemic. However, the estimated 50% reduction of in-person consultations does not represent the proportion of telehealth consultations received by women. Women included in the intervention gave birth between March 23 and July 26, 2020, which is equivalent to, at most, 4 months of a telehealth-integrated antenatal care schedule. Although not presented, the average duration of antenatal follow-up was probably 2 weeks (implementation period) and 6 weeks (integrated care period), allowing for a maximum of two telehealth visits with three face-to-face consultations. This limited exposure at the end of pregnancy is unlikely to show significant differences in outcomes and we are concerned that the conclusion of no compromise to pregnancy outcomes is premature. We need rigorous studies assessing the implementation of telehealth in comparison with a face-to-face model throughout the entire pregnancy. Furthermore, the investigators' recommendation to adopt telehealth beyond the pandemic fails to consider dimensions of care quality and equity. Research shows that care quality is compromised by incorporating telehealth into routine maternity care.2, 3 High user satisfaction rates with telehealth should be interpreted within the context of the pandemic's restrictive measures and women's intent to reduce the risk of SARS-CoV-2 infection, because qualitative evidence shows that pregnant women who received telephone consultations felt distressed due to scarce face-to-face contact with health-care providers.4 More research is needed on women's perspectives of respectful and quality care during any antenatal care schedule that uses telehealth. Additionally, relying on telehealth can contribute to exacerbating inequalities in maternal health,5 in which financial barriers, technological illiteracy, and mistrust lead to excluding vulnerable women.2 Although Palmer and colleagues show that partial use of telehealth in antenatal care appeared to be a non-inferior alternative to prevent disruption of care during the period of COVID-19 restrictions in the Australian context, unjustified compromises to high-quality, person-centred, and equitable care should not be acceptable as a way forward. © 2021 BSIP/Universal Images Group/Getty Images 2021 We declare no competing interests. ==== Refs References 1 Palmer KR Tanner M Davies-Tuck M Widespread implementation of a low-cost telehealth service in the delivery of antenatal care during the COVID-19 pandemic: an interrupted time-series analysis Lancet 398 2021 41 52 34217399 2 Galle A Semaan A Huysmans E A double-edged sword—telemedicine for maternal care during COVID-19: findings from a global mixed-methods study of healthcare providers BMJ Global Health 6 2021 e004575 3 Asefa A Semaan A Delvaux T The impact of COVID-19 on the provision of respectful maternity care: findings from a global survey of health workers medRxiv 2021 published online May 9 https://www.medrxiv.org/content/10.1101/2021.05.05.21256667v1 (preprint). 4 Wilson AN Sweet L Vasilevski V Australian women's experiences of receiving maternity care during the COVID-19 pandemic: a cross-sectional national survey Birth 2021 published online June 27. 10.1111/birt.12569 5 Ukoha EP Davis K Yinger M Ensuring equitable implementation of telemedicine in perinatal care Obstet Gynecol 137 2021 487 492 33543895

34656220

PMC9753323

NO-CC CODE

2022-12-16 23:25:38

no

Lancet. 2021 Oct 14 16-22 October; 398(10309):1405-1406

utf-8

Lancet

10.1016/S0140-6736(21)01956-5

oa_other

==== Front Nurs Outlook Nurs Outlook Nursing Outlook 0029-6554 1528-3968 Elsevier Inc. S0029-6554(21)00161-5 10.1016/j.outlook.2021.06.016 Editorial Eliminating inequities to strengthen nursing capacity and expertise Snethen Julia A. PhD, RN, FAAN Associate Editor, Nursing Outlook Professor and PhD Program Director University of Wisconsin-Milwaukee College of Nursing 28 8 2021 July-August 2021 28 8 2021 69 4 495497 © 2021 Elsevier Inc. All rights reserved. 2021 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcDuring 2020 the world faced the challenge of COVID-19, which required the nursing profession to rapidly assume new or additional routines to address a health crisis that was alarming, and initially seemed insurmountable. The pandemic was outside of any experiences in most of our lifetimes, yet nurses demonstrated their agility as they used their knowledge and abilities to swiftly change course or trouble shoot strategies for managing the virus—but it was exhausting. Many inequities present in society and in healthcare were highlighted during the COVID-19 pandemic that will require nurse's knowledge and problem solving capacity to resolve. Eileen Sullivan-Marx (2020), the president of the American Academy of Nursing, shared in an AAN statement entitled: Racism Affects Health and Wellness and It Must be Addressed that “violence, discrimination, and racism have a direct impact on determinants of health, exacerbate health inequities, and can lead to long-term trauma” (Sullivan-Marx, 2020, para 1). Similarly, the CDC (2021) has reported that communities of color have been subject to systemic racism, which negatively affects the health of individuals, both physically and mentally, and leads to health inequities. The American Academy of Nursing as well as professional nursing organizations around the country are stepping up in a myriad of ways to address issues of violence, racism, and health inequities. Currently The Future of Nursing 2020-2030: Charting a Path to Achieve Health Equity (National Academies of Sciences, Engineering and Medicine, 2021) highlights challenges that nursing as a discipline must address. According to the committee, nursing needs to continue to explore a path towards addressing health inequities by working together to eliminate violence, discrimination, and racism. Further, the nursing profession, regardless of whether nurses are working at the bedside, in the classroom, administration or within a nursing community partnership, are encouraged to work aggressively towards that goal. Health equity is a broad and complex issue, which appears to be an insurmountable task for nurses to address all at once. As the discipline is considering The Future of Nursing 2020-2030 report, it is important to contemplate how best to chart the path towards health equity. One overarching goal discussed in the Consensus Study Report Highlights is for the “achievement of health equity in the United States built on strengthened nursing capacity and expertise” (National Academies of Sciences, Engineering and Medicine, 2021, May, para 2). Strengthening nursing capacity and expertise requires nurses to also be “aware of social injustices and the systemic racism that exist in much of nursing” (Moorley, Darbyshire, Serrant, Ali & De Souza, 2020, p. 2450). Social injustices and racism affect nursing and are a nursing workforce issue, which continue to emerge regarding retention of a diverse workforce (Carthon, Travers, Hounshell, Udoeyo, & Chittams, 2021; Clary-Muronda & Hawkins, 2021). Effectively addressing health equity requires nurses to strive to eliminate inequities within the discipline. According to Clary-Muronda and Hawkins (2021) the health of the workplace is instrumental in reducing inequalities amongst nurses, which ultimately leads to greater diversity in the workplace. The authors suggest that a diverse workforce can increase patients’ positive health outcomes. Strategies are needed to reduce inequities in nursing and thereby promote a diverse workforce to strengthen “nursing capacity and expertise” (National Academies of Sciences, Engineering and Medicine, 2021, May, para2). Strategies for reducing inequities in nursing must start with the recruitment of students, and using a holistic process to increase equitable admissions of students in nursing programs around the U.S. Academic environments and opportunities within nursing programs, including financial aid, need to be equitable for the success of all students. Curriculums in nursing should include content threaded throughout that addresses ending social injustices and systemic racism. However, addressing social injustices and systemic racism cannot be limited to educational programs, as novice nurses must be integrated into equitable practice arenas. Establishing an equitable and healthy work environment in practice is a strategy that needs to start with the orientation process. The orientation process is foundational for nurses beginning a new role (Jakubik, Weese, Eliades, & Hugh, 2017; Regan et al., 2017). According to Bastian (2019) “equity is all about recognizing that everyone has had different opportunities and barriers in their lives, and then doing the work to provide people with what they need to be successful.” (para 2). Nurse leaders must review the structure and curriculum of orientation to ensure that it is crafted to assess the goals of each orientee. The assessments can be utilized to equitably assign each nurse to the activities/ opportunities and resources they need to be successful. Ensuring nurses experience equitable opportunities for their professional advancement should not end upon completion of orientation. As a newly employed nurse or continuing colleague, maintaining an emphasis on professional goals (Bastian, 2019) should be ongoing – at least annually- to encourage successful outcomes. Nursing organizations, including the AAN and Sigma, have developed leadership initiatives, which can be instrumental in facilitating each nurse's professional growth and advancement. Professional development opportunities should be funded in Department of Nursing budgets as essential resources. Mentoring is another strategy that has been effective in nursing practice (Jakubik, Weese, Eliades, & Huth, 2017) and can be used for alleviating inequities (Bastian, 2019). Working with a mentor enables nurses to develop greater competence in their clinical judgement and increase autonomy in practice. Rao, Kumar and McHugh (2017) suggest that autonomy is essential to a good work environment, while at the same time increasing satisfaction in practice (Carthon et al., 2021) which can lead to greater retention of diversity in nursing and improved patient outcomes. Additionally, mentors can facilitate the mentees navigation of their career, including goal development, educational opportunities, and engagement in decision-making within the organization. Individual nurses should also consider reducing inequities in the discipline by becoming a sponsor for a novice or junior nurse. Sponsorship is a process that goes beyond mentoring, and focuses by assisting the protégé to expand their skills, make connections and being a professional advocate for them (Bastian, 2021; Hewlett, 2011; Schawbel & Hewlett, 2013, Hewlett, 2019). Sponsorship is an effective way to promote professional growth and equitable outcomes (Hewlett, 2011; Schwebel & Hewlett, 2013). This is especially important for underrepresented nurses as “for many underrepresented identities, there is an unfortunate shortage of people who look like them in the roles that they might aspire to” (Bastian, 2019, para 3). Becoming a sponsor is beneficial for the protégé, while also enabling the sponsor to continue with their professional advancement and job satisfaction (Schwebel & Hewlett, 2013; Hewlett, 2019). Engagement of nurses in an organization needs a process that ensures diversity, equity and inclusion, which can be effective through shared governance. Shared governance processes and structures of committees may vary across organizations, including equitable procedures for engagement, time allotted to participation, and leadership opportunities. Participation by all in shared governance allows nurses voices to be heard, and nurses should strive to hear all voices as they take part in decision-making and responsibility for their practice (O'Grady & Clavelle, 2021; Kutney-Lee et al., 2016). An alternative to consider for equitable engagement of nurses within an organization is to review the shared governance committees available for membership or leadership roles. In academe, faculty might want to consider identifying all colleagues who are eligible and qualified to participate on each shared governance committee. Obtain each faculty members preferences for service and leadership, and instead of holding elections, randomly select committee members. And then have leaders or their mentors personally talk with nurses who are hesitant to step forward and encourage them to participate. Faculty governance can enact a process for establishing mentors to support committee members and leaders as they rotate through and learn their varying roles. Leadership on a committee could advance through a predetermined rotation of governance roles such as member, co-chair, the chair role, with the leader rotating off becoming the mentor for the new chair. Education continues to be foundational to the discipline of nursing. Across venues within nursing, book clubs involve educating nurses on a range of topics including diversity, equity, inclusion, social justice and racism, which affect health inequities in the U.S. The nature of the book clubs can vary, though many have incorporated reading a selected text and then spending time dialoguing with colleagues in an effort to learn from each other, and expand their knowledge of factors contributing to health inequities. Given the increasing access to online platforms for meetings, colleagues within and across institutions can engage in discussion sessions. Discussions and open dialogues that emerge from book clubs can be uncomfortable, challenging, enlightening, and sometimes discouraging. However, ongoing dialogue is an important opportunity, even when it is uncomfortable for individuals as they share their views and insights, while learning from others perspectives. At one nursing program, the faculty and students were very creative and had graduate students facilitating the book club dialogues, and incorporated that process into their DNP project (C. Klingbeil, personal communication, June 10, 2021). As nursing continues to examine The Future of Nursing 2020-2030: Charting a Path to Achieve Health Equity, there are multiple paths that lead towards health equity. One path leading towards health equity entails reducing inequities in nursing in order to promote the growth of a diverse workforce that will strengthen nursing's capacity and expertise. Establishing that path will require nurses to examine all avenues within the discipline for ensuring diversity, equity, and inclusion. Arriving on the path will require each and every one of us to identify the most effective strategies for promoting equity in nursing. Not only are strategies needed, but nurses must make sure they are not only implemented, but evaluated for effectiveness in continuing to strengthen nursing's capacity and expertise. ==== Refs References Bastian R. How mentorship can make workplaces more equitable Forbes 2019, August 26 https://www.forbes.com/sites/rebekahbastian/2019/08/26/how-mentorship-can-makeworkplaces-more-equitable/?sh=3717d78726c2 Carthon J.M.B. Travers J.L. Hounshell D. Udoeyo I. Chittams J. Disparities in Nurse Job Dissatisfaction and Intent to Leave Implications for Retaining a Diverse Workforce Journal of Nursing Administration 51 6 2021 310 317 10.1097/NNA.0000000000001019 33989239 Centers for Disease Control and Prevention (n.d.) Racism and Health Office of Minority Health & Health Equity 2021 https://www.cdc.gov/healthequity/racism-disparities/index.html Clary-Muronda V. Hawkins J.E. Global Impact: Workplaces can change the world Nursing Centered 2021, April 8 https://nursingcentered.sigmanursing.org/topics/global-impact/stories/ Hewlett S.A. The Real Benefit of Finding a Sponsor Harvard Business Review 2011, January 26 2 4 https://hbr.org/2011/01/the-real-benefit-of-finding-a Hewlett S.A. The sponsorship effect: How to be a better leader by investing in others 2019 Harvard Business Review Press Jakubik L.D. Weese M.M. Eliades A.B. Huth J.J. Mentoring in the career continuum of a nurse: Clarifying purpose and timing Pediatric Nursing 43 3 2017 149 152 ISSN: 0097-9805 Kutney-Lee A. Germack H. Hatfield L. Kelly S. Maguire P. Dierkes A. Del Guidice M. Aiken L.H. Nurse engagement in shared governance and patient and nurse outcomes Journal of Nursing Administration 46 11 2016 605 612 10.1097/NNA.0000000000000412 27755212 Moorley C. Darbyshire P. Serrant L. Ali P. De Souza R. Dismantling structural racism: Nursing must not be caught on the wrong side of history Journal of Advanced Nursing 76 10 2020 2450 2453 10.1111/jan.14469 32692444 National Academies of Sciences, Engineering and Medicine The Future of Nursing 2020-2030: Charting a Path to Achieve Health Equity 2021 The National Academies Press Washington, D. C. 10.17226/25982 National Academies of Sciences, Engineering and Medicine Consensus Study Report Highlights The Future of Nursing 2020-2030: Charting a Path to Achieve Health Equity 2021, May https://www.nap.edu/resource/25982/Highlights_Future%20of%20Nursing_4.30.21_final.pdf O'Grady T.P. Clavelle J.T. Transforming shared governance toward professional governance for nursing Journal of Nursing Administration 51 4 2021 206 211 10.1097/NNA.0000000000000999 33734180 Rao A.D. Kumar A. McHugh M. Better Nurse Autonomy Decreases the Odds of 30-Day Mortality and Failure to Rescue Journal of Nursing Scholarship 49 1 2017 73 79 10.1111/jnu.12267 28094907 Regan S. Wong C. Laschinger H.K. Cummings G. Leiter M. MacPhee M. Rheaume A. Ritchie J.A. Wolff A.C. Jeffs L. Young-Ritchie C. Grinspun D. Gurnham M.E. Foster B. Huckstep S. Ruggolo M. Shamian J. Burkoski V. Wood K. Read E. Starting out: Qualitative perspectives of new graduate nurses and nurse leaders on transition to practice Journal of Nursing Management 25 4 2017 246 255 https://onlinelibrary-wiley-com.ezproxy.lib.uwm.edu/doi/pdfdirect/10.1111/jonm.12456 28244181 Schawbel D. Hewlett S.A. Sylvia Ann Hewlett: Find A Sponsor Instead Of A Mentor Forbes 2013, September 10 https://www.forbes.com/sites/danschawbel/2013/09/10/sylvia-annhewlett-find-a-sponsor-instead-of-a-mentor/?sh=12a004c81760 Sullivan-Marx E.M. Racism affects health and wellness and it must be addressed American Academy of Nursing 2020, June 1 https://www.aannet.org/news/press-releases/position-statement-on-racism

34465416

PMC9753324

NO-CC CODE

2022-12-16 23:25:38

no

Nurs Outlook. 2021 Aug 28 July-August; 69(4):495-497

utf-8

Nurs Outlook

10.1016/j.outlook.2021.06.016

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(22)00660-2 10.1016/S0140-6736(22)00660-2 Perspectives Cracks in the foundation: how COVID-19 showed our failures Bassett Mary T 14 4 2022 16-22 April 2022 14 4 2022 399 10334 14611461 Galea Sandro The Contagion Next Time2022Oxford University Press9780197576427 272£18·99, US$24·95© 2022 Elsevier Ltd. All rights reserved. 2022 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcIn 1963, James Baldwin, one of the USA's greatest essayists, published The Fire Next Time. The book's title comes from a slave song—”God gave Noah the rainbow sign/No more water but the fire next time”. Baldwin's words sounded a warning that the USA needed to confront its racial hierarchy by embracing racial equality or doom its future. Sandro Galea's book, The Contagion Next Time, is titled in a homage to Baldwin and it also sounds a warning. Galea's central argument is that vulnerability to COVID-19 lies with a societal failure to recognise that the foundation of health rests on a healthy everyday life and not simply in the provision of health care. He ponders why this key lesson is not at the centre of pandemic discourse, which instead focuses on vaccination and treatment. In often lyrical prose, Galea roams across history, culture, literature, moral values, economics, politics, and personal pandemic experience. Although situated in a global context, the book's focus is the USA. Galea considers especially the enduring impacts of racism on health and the centrality of structural racism to understanding the USA. Globally, there were more than 6 million COVID-19 deaths by early April, 2022. Amid this carnage, few have asked why, if this novel virus was the spark, there was so much dry grass. The Contagion Next Time is suffused with Galea's anguish that failure to grapple with this question will cost more lives. Worse, the world may fare even more poorly when the next contagion arrives. After the omicron (B.1.1.529) surge, which occurred after this book was written, the USA is now approaching a staggering 1 million COVID-19 deaths. An analysis in 2021 by the Lancet Commission on public policy and health in the Trump era, to which both Galea and I contributed, found that about 40% of US COVID-19 deaths “could have been averted had the US death rate mirrored the weighted average of the other G7 nations”. In February, 2022, the per capita cumulative COVID-19 death rate in the USA exceeded that of other wealthy nations and during omicron placed the USA in a league all its own. Just what is going on? Galea, who is Dean of the Boston University School of Public Health, USA, seeks the answers with erudition and passion. He begins with reflections of an array of relevant data. But most of the book reads as a journey towards understanding why the USA has fared so poorly. The answers have little to do with the scientific breakthroughs that brought us COVID-19 vaccines in such record time. Galea points to the US obsession with individualism, making achievement of health a personal project, rather than a collective one, and an attachment to technology that has given priority to medical care over public health. He also posits widespread lack of compassion, which erodes solidarity. Galea does not shrink from naming an aversion to complexity in some quarters that made policy making difficult in the pandemic. COVID-19 demanded high-impact decisions that rested on imperfect, incomplete information. A lack of humility may also have contributed to at times contradictory public pronouncements. And there is a failure to confront racism (embedded in the founding of the USA), marginalisation, and socioeconomic inequality. Galea ends by observing that choosing health will mean “reorienting our social, economic and political priorities” to support our collective wellbeing. But there are questions that go unasked and unanswered. Just why would a belief that medical care creates health persist, faced with decades of public health thinking that has shown the minor share of population health attributable to clinical care? The WHO Commission on Social Determinants of Health began its 2008 final report, “Social justice is a matter of life and death.” It would seem time to ask, who benefits from these unhealthy arrangements? And there are answers. At their root lies an unfettered pursuit of profit that is not good for health. The US departure from the health performance of its peer nations began in around 1980 as market-oriented policies triumphed. Nicholas Freudenberg in his book At What Cost: Modern Capitalism and the Future of Health confronts directly how with present predatory capitalism in the USA there is widespread damage to the public's health. Haven’t we seen this also with COVID-19? Millions lost their jobs while the stock market flourished. Income inequality increased as life expectancy plunged precipitously among racially marginalised groups in the USA. Meanwhile, equitable access to COVID-19 treatments and vaccines is not a reality for many people in low-income and middle-income countries. Debates continue over patent protections. The impact of anti-communism in the USA has limited interrogation of capitalism. But the USA is a capitalist country. Nearly 60 years ago, Baldwin sounded the alarm for the enduring harm of racism to the future of the USA, a lesson which Galea has amplified. Perhaps a more courageous book that contemplates the next contagion would go a step further in sounding the alarm for the health effects of modern capitalism. ==== Refs Further reading Baldwin J The fire next time 1963 The Dial Press New York, NY Mueller B Lutz E U.S. has far higher Covid death rate than other wealthy countries The New York Times Feb 1, 2022 https://www.nytimes.com/interactive/2022/02/01/science/covid-deaths-united-states.html Woolhandler S Himmelstein DU Ahmed S Public policy and health in the Trump era Lancet 397 2021 705 753 33581802

0

PMC9753327

NO-CC CODE

2022-12-16 23:25:38

no

Lancet. 2022 Apr 14 16-22 April; 399(10334):1461

utf-8

Lancet

10.1016/S0140-6736(22)00660-2

oa_other

==== Front Sci Total Environ Sci Total Environ The Science of the Total Environment 0048-9697 1879-1026 The Author(s). Published by Elsevier B.V. S0048-9697(21)00845-7 10.1016/j.scitotenv.2021.145778 145778 Article Long-term analysis of the relationships between indoor and outdoor fine particulate pollution: A case study using research grade sensors Mendoza Daniel L. abc⁎ Benney Tabitha M. d Boll Sarah e a Department of Atmospheric Sciences, University of Utah, 135 S 1460 E, Room 819, Salt Lake City, UT 84112, USA b Department of City & Metropolitan Planning, University of Utah, 375 S 1530 E, Suite 220, Salt Lake City, Utah 84112, USA c University of Utah School of Medicine, Pulmonary Division, 26 N 1900 E, Salt Lake City, UT 84132, USA d Department of Political Science, University of Utah, 260 S Central Campus Drive, Salt Lake City, UT 84112, USA e State of Utah, Division of Facilities Construction and Management, 4315 S 2700 W, Floor 3, Salt Lake City, UT 84129, USA ⁎ Corresponding author. 12 2 2021 1 7 2021 12 2 2021 776 145778145778 26 12 2020 3 2 2021 5 2 2021 © 2021 The Authors 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The growing concern of air quality and its associated health-related impacts has led to increased awareness of pollutant exposure. Most human populations spend the majority of their time indoors and the COVID-19 pandemic has likely exacerbated this behavior. While significant amounts of research have focused on outdoor air quality, to date there have been no studies that examined simultaneous long-term trends on indoor and outdoor air quality on a site using research-grade sensors. We measured fine particulate matter (PM2.5) for a year using sensors located on the rooftop, air handling room, and indoor office space in a building and captured the impacts of three types of regularly occurring elevated pollution events: wintertime atmospheric inversions, wildfires, and fireworks. The events had different magnitudes and durations, and infiltration rates varied for each event leading to dissimilar indoor air pollution levels. The building's air handling unit and different environmental conditions (lower indoor humidity and temperature during the winter) combined to reduce indoor pollution from inversion events however, particulate matter from wildfires and fireworks infiltrated at higher rates. Together, this suggests possible intervention strategies, such as ventilation rates and filter upgrades, that could be used to mitigate contaminant intrusion during elevated pollution events. This year-long study illustrates an array of ways that elevated pollution events interact with the protective effects that buildings have against air pollution for its occupants. Furthermore, we show that outdoor air pollution is an important variable to consider when studying indoor air quality as contaminant infiltration is strongly dependent on the specific pollution source. Graphical abstract Unlabelled Image Keywords Indoor air quality Outdoor air quality PM2.5 Elevated pollution event Research grade sensors Particulate matter Editor: Pavlos Kassomenos ==== Body pmc1 Introduction In the United States, more than 141 million people live in areas with unhealthy levels of air pollution, and pollutant exposure rates are more likely to be determined by race and socioeconomic status than any other factor (Bell and Ebisu, 2012; Clark et al., 2014). While most people are aware of the harms of air pollution, fewer people are aware that their indoor air quality may be worse than their outdoor air quality (Chen and Zhao, 2011). These findings are relevant because humans normally spend 80% or more of their time indoors (Jenkins et al., 1992), and as a result of the COVID-19 pandemic, many people will continue working from home into the future (Bick et al., 2020). As a result, indoor air quality is of growing importance. Prior studies in this area have contributed greatly to our understanding of air quality in an urban setting (Giani et al., 2020) and how the built environment interacts with pollution-specific events (Baek et al., 1997). To contribute further to this body of work, we use a network of research grade fine particulate matter (PM2.5) sensors placed both inside and outside a building to quantify emissions trends around this pollutant. A complete table of abbreviations used is found in Appendix A Table A.1. PM2.5 is of significant interest because studies increasingly suggest that it contributes to a range of illnesses including asthma, chronic obstructive pulmonary disease (COPD), heart disease, pneumonia, depression, low birth weight, and increased mortality (Brauer, 2010; DeVries et al., 2017; Hackmann and Sjöberg, 2016; Liu et al., 2009; McCreanor et al., 2007; Pirozzi et al., 2018a; Pirozzi et al., 2018b). Children are especially susceptible to the health and developmental impacts of air pollution (Mendoza et al., 2020; Mullen et al., 2020) because of their unique biological vulnerabilities, age-related patterns of exposure, and lack of control over their own environmental circumstances as they may spend more time outside than adults (Landrigan et al., 2010). With increasing wildfire events (Abatzoglou and Williams, 2016; Mallia et al., 2015), elevated PM2.5, which was historically a wintertime phenomenon due to atmospheric inversions (Bares et al., 2018; Whiteman et al., 2014), is now also becoming a health concern during the warmer seasons. For these reasons, PM2.5 is an important pollutant to study and understand for a range of stakeholders. To best understand potential exposure trends for PM2.5, we study the short- and long-term relationship between indoor and outdoor air quality in an urban building in Utah's Salt Lake Valley (SLV). We capture three different types of elevated pollution events: winter inversion, seasonal wildfire, and local fireworks. Despite the common belief that being indoors will protect individuals from poor air quality, we find that indoor PM2.5 concentrations vary greatly and are source specific. Indoor conditions may be just as harmful to human health, especially during volatile pollution episodes, which can be misleading because they are often shorter events. Such findings have serious implications for building air quality, urban planning, public health, air quality policy, and other related policies (e.g., school recess policies, warning or awareness programs for the young, elderly, or health vulnerable populations, etc.). This study is novel for at least four reasons. First, this project used research grade PM2.5 sensors for this study. While low-cost sensors can be helpful for improving air quality campaigns because of their differential response time (Bulot et al., 2020) and ability to track sources of pollution (Popoola et al., 2018), questions remain over the accuracy and reliability of the data they produce, especially if humidity and temperature variability are present and with complex emissions sources and concentration profiles (Bulot et al., 2020). It has become clear that low-cost sensors are no longer adequate and increasingly, research grade sensors are necessary at varied scales for improved data reliability and validity (Mead et al., 2013). The use of research grade sensors, therefore, has the potential to provide a far more complete assessment of the high-granularity air quality structure generally observed in the urban environment, and could ultimately be used for quantification of human exposure, air quality monitoring, public health, and legislative purposes. Second, while studies of indoor (Tran et al., 2020) or outdoor air quality are common in the literature, due to data complexity issues, the cost of sensors, and the technical maintenance required to manage sensors, it is uncommon to find research that captures both types of ambient air quality simultaneously without using estimation methods. Instead, most work in this area (Giani et al., 2020; Ljungman et al., 2018; Tsai et al., 2019) uses various forms of estimation to measure air quality and while this approach can be used for overall daily exposure, it lacks detail about elevated pollution events, which are highly random. In fact, one major pollution event discovered in our data was a private event featuring fireworks. Such events, while potentially lethal to some vulnerable groups, would simply be missed by most major studies. Third, this research uses sensor measurements to illustrate multiple regularly occurring pollution events, that are intense enough to rise above potential noise from other sources including anthropogenic emissions from the highly-trafficked roads and large emissions associated with commercial buildings, over a full year using direct measurements. Focused studies on single pollution events (Shen et al., 2020) have been foundational in establishing the importance of these elevated events on human health. Likewise, important advances in the field have come from research that attempts to study multiple pollution events over time, but this work has relied on satellite data (Ljungman et al., 2018) and pollution simulation techniques (Tsai et al., 2019), not real-time pollution measurements. To address this, we study and present detailed measurements of three types of elevated pollution events with the building as the unit of analysis to illustrate the importance of granularity and direct measurement in the study of air quality. Fourth, this study is novel because it measures long-term air quality trends in fine temporal detail to understand what may be missed in estimation type studies. For instance, studies that focus on detailed snapshots (Apte et al., 2017) and short-term air quality impacts in detail (Sunyer et al., 2017) have been vital for illustrating the associations between daily variation in air pollution and a variety of effects. For instance, Sunyer et al. (Sunyer et al., 2017) show the association between daily traffic pollution and lower attention in children, suggesting that short term exposure to air pollution can have consequential impacts. Land use regression techniques have proved to be a fast and accurate means for estimating long term and daily emissions exposure in urban settings (Dons et al., 2014). However, these approaches, like estimation techniques (Lipfert et al., 2006) and classic dispersion calculation approaches (Clench-Aas et al., 1999), lack the ability to model or capture short term major pollution events, which can be deadly because of their quick onset and severity. If only short-term or estimation approaches are used to study this phenomenon, the impact of pollution, which is generally small on a day-to-day basis, will be grossly underestimated. Instead, we must move beyond estimation to provide data that reflects real measurement of spatial and temporal variations. To address these gaps in the literature, this research attempts to combine highly granular, long- and short-term air quality measurements using research grade sensors located at indoor and outdoor sites (e.g., inside the building, in the air handling room of the building, and on the outside) of the same building. Through this indoor/outdoor approach, we were able to analyze, in sharp detail, a variety of pollution events. 2 Methods To study the relationship between indoor and outdoor ambient air, three research-grade, Met One Instruments ES-642 particulate sensors, with an inlet sharp cut cyclone used for selective measurement of PM2.5, (Met One Instruments, 2013) were located at the Unified State Laboratories in Taylorsville, Utah (Fig. 1 ), an urban setting in the SLV. The study site is near three heavily trafficked roads, Bangerter Highway, Interstate 215, and Redwood Road, as well as a community college and retail stores. A regulatory air quality sensor is located approximately 10 km northeast of the study site.Fig. 1 Air quality sensor locations at the Unified State Laboratories (USL), Taylorsville, Utah. USL is marked by the black circle, the shopping mall and community college are marked by red and blue circles, respectively, and the Hawthorne regulatory air quality sensor (approximately 10 km northeast of the USL site) is marked by a dashed black circle. Major north-south highways near the USL site are indicated by dark blue arrows – from left to right: Bangerter Highway, Interstate 215, Redwood Road. Map courtesy of Google Maps. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Fig. 1 The Unified State Laboratory Building, Module 1, is an 82,000 square foot, three-story laboratory building including a Biosafety Level (BSL) 3 laboratory as well as support areas including some office space. The air supply and exhaust systems include two supply air handling units (AHUs), two general exhaust AHUs and five dedicated exhaust fans for the BSL 3 labs, perchloric acid exhaust and radioisotope exhaust. Further information on the building's heating, ventilation, and air conditioning (HVAC) system is found in Appendix B.1 All three sensors for this project have a tolerance of 1 μg m−3 (Met One Instruments, 2013) and read data continuously at 10-s time intervals. The sensors were placed on the rooftop, in the air intake room, and within an internal office space. All three were installed on April 30, 2018 and were online until May 30, 2019. Since the study took a little over a year, no calibration was performed during the experiment time frame. The manufacturer states that factory calibration is generally necessary after 2 years of use and these sensors have been tested and found to robustly match reference sensors (Mendoza et al., 2019). The data was stored locally in a Raspberry Pi 3 computer that was directly connected to and read the data from the ES-642 sensor via a serial cable. The data was then downloaded from the Raspberry Pi 3 to a laptop for analysis using an ethernet cable. Data processing and analysis was done using R Version 3.6.3 software (R Core Team, 2019). We aggregated the data to 1-min resolution, by using the arithmetic mean, and we established a reference baseline by averaging the ten lowest average pollution days, as measured by the rooftop sensor readings. In order to limit the influence of spurious data, we used the “Winsorize” function found in the DescTools R package (Signorell et al., 2016). Winsorizing a set of data involves limiting the range of data by replacing extreme values to a pre-determined maximum and minimum. For this study we set the default minimum value to the 5% quantile and maximum to the 95% quantile so any values outside that range were replaced to either the 5% or 95% quantile. We examined the impact of three types of elevated pollution events: wintertime inversions, wildfire episodes, and public/private fireworks events on sensor readings. Appendix A includes a complete timeline of critical pollution events captured through this project. We also compared the hourly air quality date from the regulatory sensor against the hourly-aggregated rooftop sensor data for the whole study period as well as for the elevated pollution events. 3 Results 3.1 Baseline weekday and weekend cycles Table 1 and Fig. 2 show the results for weekday and weekend PM2.5 values for the ten lowest pollution days based on the rooftop readings. The weekday diurnal cycle (Fig. 2.a) shows the impact of pollution from the two rush hour peaks (8-10 am and 6-8 pm). The weekend diurnal cycle (Fig. 2.b) shows elevated PM2.5 early in the morning (midnight – 8 am) and late in the evening (8 pm – midnight), which could be attributed to social and recreational activities as the study location is near several well-visited sites including a shopping mall and community college. Elevated evening pollutant concentrations are also a result of a lower atmospheric boundary layer due to the colder nighttime temperatures. While the rooftop and air handling room readings tracked relatively closely to each other (Fig. 2.c–d), the indoor sensor consistently reads between a quarter to half as high as the outdoor sensors. The mean daily concentration was approximately 1 μg/m3 for the rooftop and air handling room, and 0.5 μg/m3 for the office for both weekdays and weekends. In this research, we consistently used a linear statistical model to associate all PM2.5 relationships between different sensors as this method provided the best fit for the data. Related statistical values are also provided in Table 1.Table 1 Intercept, coefficient, and R2 values for each baseline day; p-values are listed in parentheses, and for R2 values, these are the p-values of the F-statistic. For statistically significant results: *** = p ≤ 0.001. Table 1Event Indoor vs. rooftop readings Air handling room vs. rooftop readings Intercept Coefficient R2 Intercept Coefficient R2 Weekday 0.18385 (<2e-16)*** 0.29152 (<2e-16)*** 0.1956 (<2.2e-16)*** −0.62078 (<2e-16)*** 1.53373 (<2e-16)*** 0.3788 (<2.2e-16)*** Weekend 0.14290 (<2e-16)*** 0.41853 (<2e-16)*** 0.5295 (<2.2e-16)*** 0.28319 (4.34e-14)*** 0.75317 (<2e-16)*** 0.362 (<2.2e-16)*** Fig. 2 Minute-resolved baseline weekday and weekend cycles using the ten lowest rooftop pollution days: a) Weekday and b) Weekend, and comparison between indoor and outdoor pollutant concentrations for: c) Weekday and d) Weekend. Fig. 2 3.2 Impact of wintertime inversions on air quality Fig. 3.a presents the impact of a typical persistent cold air pool, commonly known as a wintertime inversion, on air quality. The dashed horizontal lines represent air quality index (AQI) level cutoffs (United States Environmental Protection Agency, 2016).2 A gradual buildup had started a few days earlier (December 4th) with partial daily reductions, but on December 7th the pollutant levels stabilized above orange (“unhealthy for sensitive groups”) air quality. December 8th was the first day when the outdoor air quality reached red (“unhealthy”) levels outdoors and yellow (“moderate”) levels indoors. December 9th and 10th followed a similar pattern as December 8th however, PM2.5 had begun to accumulate indoors because the levels stayed consistently above yellow while the outdoor sensors read orange. December 11th experienced a partial clearing out during the middle of the day and on December 12th there was a complete clear out due to a mild snow event. Although the indoor pollutant concentration is generally about one third as high as the outdoor pollutant concentration, there is a possibility that longer events could generate larger indoor pollutant buildups. In addition, the indoor air pollution readings often crossed into the yellow air quality level and were ten times greater than the baseline air day throughout the five-day pollution event. For this inversion event, the air handling room sensor was offline as there was maintenance taking place there at the time.Fig. 3 Minute-resolved impact of elevated pollution events on air quality: a) December 7th–12th inversion, b) August 23rd-24th wildfire, c) 4th of July fireworks, and d) Private fireworks event on August 17th. The red, blue, and black lines correspond to the rooftop, air handling room, and office sensors, respectively. The horizontal lines correspond to AQI levels. The vertical scales are different for each event. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Fig. 3 3.3 Impact of wildfire events on air quality The summer of 2018 was characterized by a large number of wildfire episodes, particularly in the American West. While there were a few local wildfires, the majority of pollutants traveled from California, Oregon, and Washington State due to the prevailing westerly winds. During the period of August 23rd-24th, three wildfires were active in California: 1) Stone (Modoc County), 2) Mill Creek 1 (Humboldt County), and 3) Front (San Luis Obispo and Santa Barbara Counties). Fig. 3.b displays the PM2.5 patterns that occur with the confluence of these wildfire events. During this elevated pollution event, the concentration differences between the indoor and outdoor sensors were much smaller (approximately 10–15% less) than for referenced baseline days (Fig. 2) or inversion events (Fig. 3.a), both of which show a difference of about 50%. For nearly 48 h, indoor air quality reached levels considered problematic for health compromised populations (orange) and nearly reached levels considered unsafe for all populations (red). Wildfire episodes are expected to increase in the upcoming decades (Schoennagel et al., 2017) and these results suggest that such events may need special policies (e.g., higher efficiency filters or lower intake of outside air) to better protect residents if buildings offer little air quality protection from wildfire events. 3.4 Impact of fireworks on air quality 3.4.1 4th of July fireworks Fig. 3.c presents the impact of 4th of July fireworks on air quality by examining July 4th and July 5th, 2018. The vertical axis range on this figure is more than twice as large as the wildfire event (Fig. 3.b) and more than three times larger than for the inversion days (Fig. 3.a). The July 4th firework displays started at 10 pm in Salt Lake County and the air quality effects are noticeable almost immediately. For approximately 3 h, the air quality remained red with a few spikes into purple (“very unhealthy”) levels. Spikes on the red line happening sporadically throughout the 2-day study period were likely fireworks released by individuals living nearby. During the most polluted hours (10 pm – 1 am), the indoor air quality reached orange air quality levels. It was only after 8 am on July 5th that indoor air quality returned to pre-fireworks levels. Since firework pollution is a combination of primary (smoke) and secondary particle material produced by additional chemical reactions, the effect on indoor air quality is due to a combination of both. As this pollution event illustrates, firework events are typically shorter than other types of pollution events, lasting only a few hours at most. However, even in this short period we found significantly elevated, and unhealthy, air quality levels both indoors and outdoors. 3.4.2 Private event fireworks On August 17th, 2018 a private event released fireworks approximately 8 km from the USL site (Fig. 3.d). The fireworks display started at 9 pm and lasted about 30 min. While its impact on air quality – in terms of both magnitude and duration of the accompanying increase in PM2.5 levels – was markedly lower than the July 4th event (Fig. 3.d), the indoor pollution levels still increased to orange for several minutes. Thus, the pattern was consistent with the other firework example, but reflected the smaller scale of this firework event. 3.5 Relationships between indoor and outdoor air quality Table 2 and Fig. 4 show the relationship between indoor and outdoor air quality for the elevated pollution events. As illustrated in Table 2, the relative ratio of the indoor to outdoor readings are all statistically significant to the p ≤ 0.001 level, with each elevated pollution event presenting a unique pattern. For instance, the slope of the coefficient for wildfires is more than double that of inversions. This suggests that indoor pollutant concentration is approximately 77% of rooftop air pollution, as evidenced by the coefficient in Table 2. This may suggest that wildfire related particulate matter is more easily able to avoid being filtered by building HVAC systems or is more chemically stable than inversion particulate matter. Thus, wildfires may warrant special warning systems for health vulnerable populations. The indoor air quality implications of the firework events closely resembles that of inversions in terms of the infiltration rate, however the absolute magnitude of pollution is markedly larger, albeit of shorter duration. These findings suggest that outdoor air pollution levels may be helpful when trying to predict or model indoor air quality because contaminant infiltration appears to be related to outdoor pollution sources or types.Table 2 Intercept, coefficient, and R2 values for each study event; p-values are listed in parentheses, and for R2 values, these are the p-values of the F-statistic. For statistically significant results: *** = p ≤ 0.001. Table 2Event Indoor vs. rooftop readings Air handling room vs. rooftop readings Intercept Coefficient R2 Intercept Coefficient R2 Inversion 0.216571 (9.31e-05)*** 0.294152 (<2e-16)*** 0.843 (<2.2e-16)*** Wildfire −0.522233 (<2e-16)*** 0.771900 (<2e-16)*** 0.9884 (<2.2e-16)*** −0.328231 (7.87e-12)*** 0.993484 (<2e-16)*** 0.9936 (<2.2e-16)*** 4th of July 0.061556 (0.242) 0.279183 (<2e-16)*** 0.864 (<2.2e-16)*** −0.289460 (1.52e-08)*** 1.007242 (<2e-16)*** 0.9887 (<2.2e-16)*** Private fireworks 2.868339 (<2e-16)*** 0.382894 (<2e-16)*** 0.6623 (<2.2e-16)*** −0.392637 (5.13e-11)*** 1.039981 (<2e-16)*** 0.968 (<2.2e-16)*** Fig. 4 Minute-resolved relationship between indoor and outdoor air quality for study events: a) December 7th–12th inversion, b) August 23rd-24th wildfire, c) 4th of July fireworks, and d) Private fireworks event on August 17th. The symbol color corresponds to AQI levels for the y-axis (office or air handling room) data. The axes are different for each panel. Fig. 4 During the inversion event (Fig. 4.a), the office pollutant concentration was approximately 29% of the outdoor concentration. On the other hand, the wildfire event (Fig. 4.b) resulted in the office concentration being approximately 78% of the rooftop reading. The fireworks events (Fig. 4.c–d) display a combination of the patterns from the inversion (Fig. 4.a) and wildfire (Fig. 4.b) events. The office sensor readings were similar to the rooftop readings until about 40 μg/m3 when the office air pollution leveled off while the rooftop pollutant concentration continued to increase. This was likely a result of a combination of building filtration as well as the decay of secondary PM2.5. During the wildfire (Fig. 4.b) and fireworks events (Fig. 4.c–d), the air handling and rooftop readings were nearly identical since little filtration takes place between the two locations. 3.6 Comparison against regulatory sensor data The regulatory sensor data (Hawthorne) and USL rooftop sensor readings are shown in Table 3 , Fig. 5 , and Appendix C Fig. C.1, Fig. C.2. While the full time series showed the closest correlation, the elevated pollution events resulted in relatively similar patterns. It must be noted that the two sensors are located nearly 10 km apart and in different urban typologies, with the USL sensor surrounded by more industrial and commercial buildings, whereas the Hawthorne sensor is in a residential neighborhood. Furthermore, events such as fireworks are strongly localized and factors such as tree cover and land use can impact air quality (Mendoza, 2020) resulting in marked differences between air quality readings taken at different sites within the same urban area. Despite these factors, the close association between the USL and Hawthorne readings (Appendix C) provide confirmation that the USL sensor data are representative of the ambient air pollution.Table 3 Intercept, coefficient, and R2 values for each study period; p-values are listed in parentheses, and for R2 values, these are the p-values of the F-statistic. For statistically significant results: * = p ≤ 0.05 and *** = p ≤ 0.001. Table 3Study period Rooftop vs. regulatory sensor readings Intercept Coefficient R2 Complete −0.680770 (<2e-16)*** 0.949783 (<2e-16)*** 0.6326 (<2.2e-16)*** Inversion 19.6005 (1.41e-09)*** 0.7471 (<1.86e-09)*** 0.2386 (1.858e-09)*** Wildfire −1.78651 (0.194) 1.34973 (<2e-16)*** 0.9269 (<2.2e-16)*** 4th of July −7.3477 (0.0224)* 1.2636 (9.67e-10)*** 0.56 (9.674e-10)*** Private fireworks −0.9470 (0.593) 0.8906 (8.71e-06)*** 0.606 (8.712e-06)*** Fig. 5 Hour-resolved air quality: a) Full study period with the green and black lines correspond to the Hawthorne Division of Air Quality and USL rooftop sensors, respectively, and b) Direct comparison via X-Y plot. The symbol color corresponds to AQI levels for the y-axis (USL rooftop) data. The axes are different for each panel. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Fig. 5 4 Discussion and conclusions 4.1 Discussion Despite the common belief that the built environment is a protective measure in terms of air quality, our research suggests that the relationship between indoor and outdoor air quality varies in relevant and consequential ways. We compared an average baseline day to three types of random, but elevated pollution events and found that the relationship between indoor and outdoor pollution is pollutant-specific and indoor conditions may be just as harmful to human health, especially in the short term. Each case we presented illustrates how the patterns of indoor and outdoor air quality adopt specific patterns based on each unique pollution event and additional research will be needed to further classify the regularity of these patterns. Since the “at-home” work force is expected to grow following the pandemic, indoor air quality is of growing importance and public policy may need to be more adaptive and nuanced to assure human health and well-being is protected during elevated pollution events. The findings from this research have substantial technical and public health implications. 4.1.1 Technical implications Since the SLV has been found to be in non-compliance for air quality standards in both PM2.5 and ozone, the state has been developing a State Implementation Plan (SIP) to address this issue (Utah Department of Environmental Quality, 2018; Utah Division of Air Quality, 2018). Technical knowledge and baseline data on PM2.5 is of great value locally and, in this study, we add to this technical understanding. For instance, we confirmed that during inversion events, the majority of PM2.5 readings are from secondary particulate matter (Baasandorj et al., 2017; Lareau et al., 2013), which is particulate matter that is not directly emitted but instead is generated through chemical reactions facilitated by stable meteorological conditions. When the conditions are drastically changed indoors due to increased temperature and decreased relative humidity, a large amount of PM may decompose into its precursors. While it is possible that the outdoor PM2.5 mass concentrations observed during this time are slightly overestimated due to the high relative humidity present, the magnitude of the differences observed between the outdoor and indoor sensor are significantly beyond the potential impact on concentrations giving us a high degree of confidence in our findings. However, wildfire smoke is primary (directly emitted) PM that does not dissociate due to changing conditions and is able to pass through building filtration, without decomposing, leading to orange air quality levels indoors. This could lead to significant health concerns for sensitive groups include young, elderly, and populations with prior conditions including asthma, cystic fibrosis (CF), and COPD. Wildfire episodes are expected to increase in the upcoming decades (Schoennagel et al., 2017) and our research shows that buildings offer little air quality protection from wildfire events. These technical considerations could be helpful to local and state authorities. Considering the impacts of outdoor pollution events on indoor air quality, this research also has important implications for building health, filtration systems, and ventilation schedules, especially in public spaces where the health and well-being of its occupants must be assured (e.g., government buildings, schools, hospitals, etc.). Through the use of long-term direct measurement in air quality sensing, we have demonstrated that not all pollutants are created equal in terms of their ability to enter a building. Our research, and others like it, will be essential in the effort to design and implement pollution abatement measures and quantifying their effects to best inform the policy process. Having a baseline inventory of the emissions processes found in urban settings can also help to establish risk assessment and management programs. Therefore this (and future) studies can advance (or facilitate) modeling to predict indoor air quality based on outdoor pollutant levels and nature of event (i.e., wildfire, inversions, fireworks). 4.1.2 Public health and policy implications Our study also has important implications for public policy and public health. This research suggests that air pollution policy many need to differentiate and carefully define policy reflecting the trajectory and implications of different elevated pollution events in order to best protect the public in reflection of these unique characteristics. For example, tracking air quality at this level of granularity could greatly enhance our ability to highlight geographical areas whose inhabitants are most exposed to poor air quality or to identify how many people in an area are exposed to concentrations of pollution exceeding air quality guidelines. Research on children is advancing in this area to better understand short-term (Mendoza et al., 2020; Mullen et al., 2020) and lifetime exposure impacts (Berhane et al., 2011), but other groups are also at risk of higher-than-average exposure rates. Our findings could be used to advance such research by describing the exposure of population subgroups such as children, elderly, medically vulnerable groups (e.g., COPD and asthma patients), and essential workers (e.g., military personnel, first responders). 4.1.3 Future research A follow-up study is currently underway at the Utah State Hospital in Provo, Utah. An air quality sensor has been located on the rooftop of the Pediatric building. Two indoor air quality sensors have been placed in rooms belonging to the dormitory and daycare wings as they have different air handlers and will assess the impact of different ventilation and filtration technologies on indoor air quality. This study will help in the understanding of potential benefits of different HVAC systems and identify possible improvements to increase occupant health and safety. Future study on the performance of various filtration options and the differences between 100% outside air systems and recirculated air systems would be beneficial to understanding effective ways to mitigate pollution travelling from outside into workspaces. 4.2 Conclusions This year-long study illustrates an array of ways that elevated pollution events interact with the protective effects that buildings have against air pollution for its occupants. The current results show that while the Unified State Laboratories building provided a relatively protective environment for its occupants during wintertime inversion events, the indoor air quality was comparable to the outdoor air quality during wildfire and fireworks events. These differences are partially attributable to the physical and chemical cycles responsible for the generation and disassociation of secondary particulate matter during wintertime inversion events. It is also likely that some of the wildfire and fireworks smoke particles are too small to be filtered with the current building filtration systems. CRediT authorship contribution statement Daniel Mendoza: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Funding acquisition. Tabitha M. Benney: Conceptualization, Methodology, Formal analysis, Writing – original draft, Writing – review & editing. Sarah Boll: Validation, Resources, Writing – review & editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix A Event timeline and table of abbreviations 2018• April 30: Sensors installed. • July 4–5: Firework event (Fig. 3.c). • August 17: Private firework event (Fig. 3.d). • August 23–24: Wildfire Event (CA, MO, CO – all had major fire events) (Fig. 3.b). • September 15: Wildfires push air quality beyond orange into red. • November 22: Lingering wildfire pollution is cleared out around 7 am until noon as average wind speed went from 2 to 3 kph for the week, up to 25 kph and cleared the smoke out temporarily. • December 7–12: Winter Inversion Event (Fig. 3.a). • December 17: High wind gusts and associated air quality cleanup. 2019• January 8–16: Substantial inversion (not included). • January 30: Mini inversion event. • May 30: Study ends. Table A.1 Table of abbreviations. Table A.1Air Handling Unit AHU Biosafety Level BSL Cystic fibrosis CF Chronic obstructive pulmonary disease COPD Heating, Ventilation, and Air Conditioning HVAC Fine particulate matter PM2.5 State Implementation Plan SIP Salt Lake Valley SLV Unified State Laboratories USL Appendix B Summary of HVAC system in the Utah Unified State Laboratory Building, Module 1 Air is delivered to lab spaces by phoenix air valves and by variable air volume box terminal units for office spaces.3 Both air delivery valves include hydronic reheat coils, indicating that air delivered to a space can be warmed directly prior to delivery to the space. The air flow is a single pass 100% outside air design. The air is pulled by the air handling units (AHUs) through a set of louvers, filters, a heat recovery coil, a glycol preheat coil, passes through the AHU and is then pushed through indirect evaporative cooling coil, a chiller water cooling coil, a direct evaporative cooling coils, a second set of louvers (see Fig. B.1) before being pushed further through the duct work to the various terminal units and delivered to the various spaces through a diffusion grill.Fig. B.1 Air pathway through the Air Handling Units at Unified State Labs. Fig. B.1 This could be relevant to the amount of particulate matter that travels from the outside to workspaces as it averaged 10 points of resistance to the air flow that is delivered to the space, in addition to the filter bank which is designed to remove particulate matter from the air stream. This air path is very typical for the laboratory building. For a typical office building this air path is also very common with the one difference that office buildings try to recover air that has already been conditioned and recirculate it as an energy savings measure. Minimum air flows for various configurations of ventilations systems and building typologies are standardized through ANSI/ASHRAE Standard 62.1–2019.4 It is possible that in a recirculated air delivery system less pollutants would enter the workspace and worthy of further investigation. Appendix C Comparison between USL air quality sensors and a nearby regulatory sensor Fig. C.1 Hour-resolved impact of elevated pollution events on air quality: a) December 7th–12th inversion, b) August 23rd-24th wildfire, c) 4th of July fireworks, and d) Private fireworks event on August 17th. The green and black lines correspond to the Hawthorne Division of Air Quality and USL rooftop sensors, respectively. The horizontal lines correspond to AQI levels. The vertical scales are different for each event. Fig. C.1 Fig. C.2 Figure C.1 Hour-resolved air quality readings relationship between USL rooftop and Hawthorne Division of Air Quality sensors for study events: a) December 7th–12th inversion, b) August 23rd-24th wildfire, c) 4th of July fireworks, and d) Private fireworks event on August 17th. The symbol color corresponds to AQI levels for the y-axis (USL rooftop) data. The axes are different for each panel. Fig. C.2 Acknowledgements Dr. Erik Crosman, Department of Life, Earth and Environmental Sciences, West Texas A&M University; Dr. Alex Jacques, Department of Atmospheric Sciences, University of Utah; Ryan Bares, Department of Atmospheric Sciences and Global Change & Sustainability Center, University of Utah; Samuel R. Baty, Department of Political Science, University of Utah. Funding sources Utah Division of Facilities Construction and Management. 1 Rocky Mountain Power Savings and Incentive Report Unified State Lab Persistent Commissioning Report; Ezra Nielsen, June 28, 2019. 2 The 24-h average AQI breakpoints along with the respective index colors are (in μg/m3): 0.0–12.0, “Good” (green); 12.1–35.4, “Moderate” (yellow); 35.5–55.4, “Unhealthy for sensitive groups” (orange); 55.5–150.4, “Unhealthy” (red); 150.5–250.4, “Very Unhealthy” (purple); 250.5+, “Hazardous” (maroon). 3 Rocky Mountain Power Savings and Incentive Report Unified State Lab Persistent Commissioning Report; Ezra Nielsen, June 28, 2019. 4 ANSI/ASHRAE Standard 62.1-2019, Ventilation for Acceptable Indoor Air quality, https://www.ashrae.org/technical-resources/bookstore/standards-62-1-62-2 ==== Refs References Abatzoglou J.T. Williams A.P. Impact of anthropogenic climate change on wildfire across western US forests Proc. Natl. Acad. Sci. 113 2016 11770 11775 27791053 Apte J.S. Messier K.P. Gani S. Brauer M. Kirchstetter T.W. Lunden M.M. High-resolution air pollution mapping with Google street view cars: exploiting big data Environ. Sci. Technol. 51 2017 6999 7008 28578585 Baasandorj M. Hoch S.W. Bares R. Lin J.C. Brown S.S. Millet D.B. Coupling between chemical and meteorological processes under persistent cold-air pool conditions: evolution of wintertime PM2. 5 pollution events and N2O5 observations in Utah’s salt Lake Valley Environ. Sci. Technol. 51 2017 5941 5950 28468492 Baek S.-O. Kim Y.-S. Perry R. Indoor air quality in homes, offices and restaurants in Korean urban areas—indoor/outdoor relationships Atmos. Environ. 31 1997 529 544 Bares R. Lin J.C. Hoch S.W. Baasandorj M. Mendoza D.L. Fasoli B. The wintertime Covariation of CO2 and criteria pollutants in an Urban Valley of the Western United States J. Geophys. Res.-Atmos. 123 2018 2684 2703 Bell M.L. Ebisu K. Environmental inequality in exposures to airborne particulate matter components in the United States Environ. Health Perspect. 120 2012 1699 1704 22889745 Berhane K. Zhang Y. Linn W.S. Rappaport E.B. Bastain T.M. Salam M.T. The effect of ambient air pollution on exhaled nitric oxide in the Children’s health study Eur. Respir. J. 37 2011 1029 1036 20947676 Bick A. Blandin A. Mertens K. Work from Home after the COVID-19 Outbreak 2020 Brauer M. How much, how long, what, and where: air pollution exposure assessment for epidemiologic studies of respiratory disease Proc. Am. Thorac. Soc. 7 2010 111 115 20427581 Bulot F.M.J. Russell H.S. Rezaei M. Johnson M.S. Ossont S.J.J. Morris A.K.R. Laboratory comparison of low-cost particulate matter sensors to measure transient events of pollution Sensors 20 2020 2219 32326452 Chen C. Zhao B. Review of relationship between indoor and outdoor particles: I/O ratio, infiltration factor and penetration factor Atmos. Environ. 45 2011 275 288 Clark L.P. Millet D.B. Marshall J.D. National patterns in environmental injustice and inequality: outdoor NO2 air pollution in the United States PLoS One 9 2014 e94431 Clench-Aas J. Bartnova A. Bøhler T. Grønskei K.E. Sivertson B. Larssen S. Air pollution exposure monitoring and estimating. Part I. Integrated air quality monitoring system J. Environ. Monit. 1 1999 313 319 11529128 DeVries R. Kriebel D. Sama S. Outdoor air pollution and COPD-related emergency department visits, hospital admissions, and mortality: a meta-analysis COPD: J. Chron. Obstruct. Pulmon. Dis. 14 2017 113 121 Dons E. Van Poppel M. Panis L.I. De Prins S. Berghmans P. Koppen G. Land use regression models as a tool for short, medium and long term exposure to traffic related air pollution Sci. Total Environ. 476 2014 378 386 24486493 Giani P. Castruccio S. Anav A. Howard D. Hu W. Crippa P. Short-term and long-term health impacts of air pollution reductions from COVID-19 lockdowns in China and Europe: a modelling study Lancet Planetary Health 4 2020 e474 e482 32976757 Hackmann D. Sjöberg E. Ambient air pollution and pregnancy outcomes—a study of functional form and policy implications Air Qual. Atmos. Health 2016 1 9 28111595 Jenkins P.L. Phillips T.J. Mulberg E.J. Hui S.P. Activity patterns of Californians: use of and proximity to indoor pollutant sources Atmos. Environ. Part A 26 1992 2141 2148 Landrigan P.J. Rauh V.A. Galvez M.P. Environmental justice and the health of children Mt Sinai J. Med. 77 2010 178 187 20309928 Lareau N.P. Crosman E. Whiteman C.D. Horel J.D. Hoch S.W. Brown W.O. The persistent cold-air pool study Bull. Am. Meteorol. Soc. 94 2013 51 63 Lipfert F. Wyzga R. Baty J. Miller J. Traffic density as a surrogate measure of environmental exposures in studies of air pollution health effects: long-term mortality in a cohort of US veterans Atmos. Environ. 40 2006 154 169 Liu L. Poon R. Chen L. Frescura A.-M. Montuschi P. Ciabattoni G. Acute effects of air pollution on pulmonary function, airway inflammation, and oxidative stress in asthmatic children Environ. Health Perspect. 117 2009 668 674 19440509 Ljungman P.L. Li W. Rice M.B. Wilker E.H. Schwartz J. Gold D.R. Long-and short-term air pollution exposure and measures of arterial stiffness in the Framingham heart study Environ. Int. 121 2018 139 147 30205320 Mallia D.V. Lin J.C. Urbanski S. Ehleringer J. Nehrkorn T. Impacts of upwind wildfire emissions on CO, CO2, and PM2.5 concentrations in Salt Lake City, Utah J. Geophys. Res.-Atmos. 120 2015 147 166 McCreanor J. Cullinan P. Nieuwenhuijsen M.J. Stewart-Evans J. Malliarou E. Jarup L. Respiratory effects of exposure to diesel traffic in persons with asthma N. Engl. J. Med. 357 2007 2348 2358 18057337 Mead M. Popoola O. Stewart G. Landshoff P. Calleja M. Hayes M. The use of electrochemical sensors for monitoring urban air quality in low-cost, high-density networks Atmos. Environ. 70 2013 186 203 Mendoza D.L. The relationship between land cover and Sociodemographic factors Urban Sci. 4 2020 68 Mendoza D.L. Crosman E.T. Mitchell L.E. Jacques A. Fasoli B. Park A.M. The TRAX light-rail train air quality observation project Urban Sci. 3 2019 108 Mendoza D.L. Pirozzi C.S. Crosman E.T. Liou T.G. Zhang Y. Cleeves J.J. Impact of low-level fine particulate matter and ozone exposure on absences in K-12 students and economic consequences Environ. Res. Lett. 15 11 2020 Met One Instruments I. ES-642 Dust Monitor Operation Manual, 2013. Mullen C. Grineski S.E. Collins T.W. Mendoza D.L. Effects of PM2.5 on third grade Students’ proficiency in math and English language arts Int. J. Environ. Res. Public Health 17 2020 6931 32971971 Pirozzi C.S. Jones B.E. VanDerslice J.A. Zhang Y. Paine R. III Dean N.C. Short-term air pollution and incident pneumonia. A case–crossover study Ann. Am. Thorac. Soc. 15 2018 449 459 29283681 Pirozzi C.S. Mendoza D. Xu Y. Zhang Y. Scholand M. Baughman R. Short-term particulate air pollution exposure is associated with increased severity of respiratory and quality of life symptoms in patients with fibrotic Sarcoidosis Int. J. Environ. Res. Public Health 15 2018 1077 29861437 Popoola O.A. Carruthers D. Lad C. Bright V.B. Mead M.I. Stettler M.E. Use of networks of low cost air quality sensors to quantify air quality in urban settings Atmos. Environ. 194 2018 58 70 R Core Team. A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; \. URL https://www.R-project.org 2019. Schoennagel T. Balch J.K. Brenkert-Smith H. Dennison P.E. Harvey B.J. Krawchuk M.A. Adapt to more wildfire in western North American forests as climate changes Proc. Natl. Acad. Sci. 114 2017 4582 4590 28416662 Shen J. Zhao Q. Cheng Z. Huo J. Zhu W. Zhang Y. Evolution of source contributions during heavy fine particulate matter (PM2. 5) pollution episodes in eastern China through online measurements Atmos. Environ. 117569 2020 Signorell A. Aho K. Alfons A. Anderegg N. Aragon T. DescTools: Tools for Descriptive Statistics. R Package Version 0.99. 18 2016 R Foundation for Statistical Computing Vienna, Austria Sunyer J. Suades-González E. García-Esteban R. Rivas I. Pujol J. Alvarez-Pedrerol M. Traffic-related air pollution and attention in primary school children: short-term association Epidemiol. (Cambridge, Mass.) 28 2017 181 Tran V.V. Park D. Lee Y.-C. Indoor air pollution, related human diseases, and recent trends in the control and improvement of indoor air quality Int. J. Environ. Res. Public Health 17 2020 2927 32340311 Tsai D.-H. Riediker M. Berchet A. Paccaud F. Waeber G. Vollenweider P. Effects of short-and long-term exposures to particulate matter on inflammatory marker levels in the general population Environ. Sci. Pollut. Res. 26 2019 19697 19704 United States Environmental Protection Agency. Air Quality Index (AQI) Basics, 2016. Utah Department of Environmental Quality. Utah Division of Air Quality 2017 Annual Report, 2018. Utah Division of Air Quality Utah's Non Attainment Area Locator Shapefiles 2018 Utah Division of Air Quality editor Whiteman C.D. Hoch S.W. Horel J.D. Charland A. Relationship between particulate air pollution and meteorological variables in Utah’s Salt Lake Valley Atmos. Environ. 94 2014 742 753

33647662

PMC9753328

NO-CC CODE

2022-12-16 23:25:38

no

Sci Total Environ. 2021 Jul 1; 776:145778

utf-8

Sci Total Environ

10.1016/j.scitotenv.2021.145778

oa_other

==== Front Nurs Outlook Nurs Outlook Nursing Outlook 0029-6554 1528-3968 Elsevier Inc. S0029-6554(21)00098-1 10.1016/j.outlook.2021.03.024 Letter to the Editor Re-visioning the image of nursing McGinity Ann Scanlon PhD, RN, FAAN ⁎ Houston Methodist Health Care System, Houston, Texas ⁎ Corresponding author: Ann Scanlon McGinity, Houston Methodist Health Care System, 749 Honeysuckle Lane, Quitman, AR 72131 3 5 2021 July-August 2021 3 5 2021 69 4 526527 7 3 2021 22 3 2021 29 3 2021 © 2021 Elsevier Inc. All rights reserved. 2021 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcThe platform for innovation in healthcare and the redesign of nursing's image has been birthed in the COVID 19 pandemic. Though we have witnessed the horrendous morbidity, suffering and mortality of this pandemic globally, the time has also afforded us some of the greatest opportunities to quickly engage and redesign the work we are doing and has allowed us to publicly re- image the contributions of nursing in diverse areas of healthcare. This disruption has allowed us to seek novel ways of examining our nursing profession and to begin the redesign of our roles in transforming healthcare services. The public has long seen and acknowledged that nurses are the number one trusted profession. That platform is the most valuable one from which we now can continue to evolve the image of nursing to that we propose for the future. The current image was formed in the early days of our profession around service of nurses in conflict and war. The often-violent war metaphors by which we have built our nursing profession revolve around these images: fighting battles, fighting for patients, fighting disease, working in the trenches, following doctor's orders. These metaphors produced the image of nursing to be one in which the nurse is perceived as loyal, obedient, protective, subservient, and a follower of orders. Additionally, the nurse is also seen as compassionate, caring, nurturing, protecting, and mothering. Those images today remain the predominant ones that the public think about when they think of nursing: individuals with integrity, warmth, and humanity. Many of the images presented in the media during the COVID crisis were of men and women whose faces were bruised from wearing masks, who were exhausted and tired, who were sacrificing their own good for the good of others. Heroes. These are all important aspects of who and what nurses are and do. Highlighting only parts of one's professional identity however does a disservice to the intellectual capabilities of nurses as well as to our artistry. The call to action now is to reframe and expand the platform from which our image must evolve. This requires global consensus of nurses on the elements that will constitute the emerging image of nursing and what is the preferred future state of our profession. The key to changing image is to first envision the future state. Leading and being key players in the transformation of health care requires that consensus on this future state and on the attributes and competencies needed for future clinicians be identified. National nursing organizations must align and focus on this as their priority. These organizations need to reach consensus collaboratively on this new image in order to effectively carry out their own specific charters and impact the culture of nursing. The cultures of these national and global nursing organizations must align and design strategically an image that will propel our profession forward over the next decade. These groups need to work to communicate and brand the image standard and develop leaders who collectively will be responsible and accountable for creating the culture that presents nurses as scientists, economists, entrepreneurs, epidemiologist and anthropologists. These national and international nursing organizations must establish standards around education, clinical competencies, leadership, research, practice, innovation etc. Global nursing organizations must collectively play a leadership role if this is to be achieved. Nursing leaders are at the table where critical discussions and decision impacting healthcare transformation occur. What oftentimes is not present are the leadership competencies of our leaders to effectively be heard at that table. Leadership development from the moment a student nurse is selected must be visioned and lived in all curricula to which that individual is exposed. The candidate seeking nursing as a profession must bring the personal values of caring, compassion, and integrity: our responsibility as educators and organizational leaders must then hone that individual's professional identity as a clinician, scientist, epidemiologist etc. Personal values are brought to the profession while their professional identity is learned and lived thru cultures that nurses are exposed to throughout their professional life-long experiences. Our nursing organizations need to objectively assess their own inadequacies to date and our failures in creating leaders and cultures that produce talented and exceptional professionals every time. We must move from philosophizing about changing culture to activation as our success is based on how quickly we can pivot in these rapidly changing environments. Silence, fear and intimidation must be acknowledged where it exists and replaced with courageous, capable and accountable individuals who will be respected for their meaningful contributions at whatever table they sit at in their organizations. Culture change and therefore imaging occurs with the leaders of our profession in both academia and practice. Much work is currently being addressed at the individual nurse level however for the greatest change to occur and in the quickest amount of time, it is the leaders who strategically and then operationally set the collective expectations for our future as a profession. This is their primary work if we are to capture the opportunities that COVID-19 has provided to our profession. Focusing the energy of those many bright and talented nurses that support these future images and who are currently actively striving to create innovative and collaborative efforts focused on improving care to their communities can only move us to greater action and impact in the work of healthcare transformation. The immediate call to action for our leaders of national nursing organizations is to convene an Advisory Board of all these current leaders with an equal representation of nursing executives of healthcare organizations with the specific goal of redesigning the image of nursing of the future. This collaborative activity needs to result in a vision and an image of our professional identity that will provide a roadmap for our future that results in attracting, retaining and energizing our current and future workforce as they embark on their work of transforming healthcare. Credit Statement This is to attest that the thoughts put forth in the Letter to the Editor are credited to the author of that letter alone.

33958203

PMC9753364

NO-CC CODE

2022-12-16 23:25:40

no

Nurs Outlook. 2021 May 3 July-August; 69(4):526-527

utf-8

Nurs Outlook

10.1016/j.outlook.2021.03.024

oa_other

==== Front Chemosphere Chemosphere Chemosphere 0045-6535 1879-1298 Elsevier Ltd. S0045-6535(22)00520-3 10.1016/j.chemosphere.2022.134027 134027 Article Human health risk assessment of heavy metal and pathogenic contamination in surface water of the Punnakayal estuary, South India Selvam S. a∗ Jesuraja K. ab Roy Priyadarsi D. c Venkatramanan S. d Khan Ramsha e Shukla Saurabh e Manimaran D. a Muthukumar P. a a Department of Geology, V.O. Chidambaram College, Thoothukudi, 628008. Tamilnadu, India b Regsitration No: 18212232061030, Affiliated to Manonmaniam Sundaranar University, Tirunelveli, 627 012, Tamil Nadu, India c Instituto de Geología, Universidad Nacional Autónoma de México (UNAM), Ciudad Universitaria, Ciudad de México, CP 04510, Mexico d Department of Disaster Management, Alagappa University, Karaikudi, Tamil Nadu, India e Faculty of Civil Engineering, Institute of Technology, Shri Ramswaroop Memorial University, Barabanki, UP, 225003, India ∗ Corresponding author. 14 3 2022 7 2022 14 3 2022 298 134027134027 10 10 2021 25 1 2022 15 2 2022 © 2022 Elsevier Ltd. All rights reserved. 2022 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Variation in levels of toxic heavy metals in river system during the COVID-19 pandemic lockdown might potentially assist in development of a public health risk mitigation system associated with the water consumption. The water quality of Punnakayal estuary in the Thamirabarani River system from the south India, a vital source of water for drinking and domestic purposes, industrial usage, and irrigation was assessed here. A comparitive assessment of physico-chemical variables (pH, EC, TDS, DO, BOD, turbidity and NO3), microbiological parameters (total coliform bacteria, fecal coliform bacteria, fecal streptococci and escherichia coli) and toxic metals (As, Cr, Fe, Cu, Zn, Cd, and Pb) suggested a decrease of 20% in the contaminant ratio during the lockdown period in comparison to the pre-lockdown period. The Health risk assessment models (HQ, HI, and TCR) highlighted carcinogenic and non-carcinogenic hazards for both children and adults through the ingestion and dermal adsorption exposures. The HI values for both As and Cr exceeded the acceptable limit (>1) during the lockdown period, but the potential risk for children and adults remained low in compaisio with the pre-lockdown period. Our results suggested that the Thamirabarani River system remained hostile to human health even during the lockdown period, and it requires regular monitoring through a volunteer water quality committee with private and government participations. Graphical abstract Image 1 Keywords COVID-19 pandemic Water quality Toxic metals Health risk hazard India Handling Editor: Derek Muir ==== Body pmc1 Introduction The outbreak of corona virus or COVID-19 has led to a complete lockdown in 213 countries (World Health Organization, 2020; Bherwani et al., 2020; Gautam and Trivedi, 2020; Sivakumar, 2020). The first case of COVID-19 was reported in China (Travaglio et al., 2021; Kachroo, 2020; Selvam et al., 2020a; Sinha, 2020) and subsequently, it caused an epidemic of acute respiratory syndrome (SARS-CoV-2) all over the globe (Gautam and Hens, 2020). Till now, India has a total number of cases of 1,32,05,926 with a mortality rate of 1.28% i.e., 1,68,436 deaths. The Tamil Nadu state in south India has 33,659 active cases with a mortality rate of 1.40% i.e., 12,863 deaths (Source: https://www.mygov.in/covid-19/). This epidemic has also caused irreversible changes to the socio-economic and environmental conditions as the prolonged closure of various industrial edifices had detrimental effects on the economic prosperity. However, the progressive effects on different ecosystems and environment were applauded. For example, Selvam et al. (2020a, India), Lian et al. (2020, China), Nakada and Urban (2020, Brazil), Ropkins and Tate (2021, UK) and Stratoulias and Nuthammachot (2020, Thailand) documented better air quality. Similarly, the upgrades in surface water qualities was noticed by Selvam et al. (2020b, India, Qiu et al. (2020, China), and Kassem and Jaafar (2020, Lebanon). Dutta et al. (2020) and Patel et al. (2020) documented water quality improvements in the Ganges and Yamuna rivers as a result of the lockdown in India. The water pollution has been evaluated through estimation of heavy metals (Selvam et al., 2020a). It has two vital sources, i.e. (i) natural (ii) anthropogenic. The natural resources include metals released from rock weathering and their eventual discharge to the water bodies (Ravindran and Selvam, 2014). The anthropological sources include the emission of heavy metals by emancipation of industrial effluents, fossil fuels/industrial burning, and discharge of sewage into the surface water bodies (Selvam and Sivasubramanian, 2012; Singaraja et al., 2015; Panneerselvam et al., 2021). Recent studies have reported that human health risk assessment using Heavy Metal Pollution Index (HMPI), and Heavy Metal Toxicity Load (HMTL) approaches, carcinogenic and non-carcinogenic health risk approach and human health risk index analysis. The chronic exposure to low concentrations of metals like Pb, Cd, Cr might cause brain and kidney damage, and other chronic kidney disease (CKD). Heavy metals like Cd, Pb, Cu, Ni, U, As, and Fe, are the main nephrotoxic heavy metals that can cause tubular damage and glomerulopathies (Bineshpour et al., 2020; Karaouzas et al., 2020; Mukherjee et al., 2020; Proshad et al., 2020; Tokatli and Ustaoğlu, 2020). Thamirabarani River system is one of the vital water sources in the state of Tamil Nadu (south India) for drinking, irrigation and middle scale industrial usages (Source: https://en.wikipedia.org/wiki/Thamirabarani_River). Past researches in this region have reported heavy metal contamination in this river and related estuaries through the industrial (tannery) effluent, municipal sewage effluent, and urban development (Selvam et al., 2015; Singaraja et al., 2015; Muthukumaravel et al., 2021). Therefore, the general assumption was that the reduced activities of above sources during the lockdown would have improved the surface water quality of the Thamirabarani channel. However, there has been no study to assess the effects of COVID-19 lockdown. Thus, the purpose of this study was to (i) estimate hydro-chemical variances between the lockdown and pre-lockdown periods, (ii) assess the pollution vulnerability of industrial and anthropological demeanor on surface water quality, (iii) measure possible health hazards to children and adults from domestic, irrigational and livestock usages of surface water during the lockdown and their differences with the pre-lockdown period, (iv) identify the pollution sources and estimate the reduction percentage of toxic heavy metals, and (v) provide the community with suggestions and solutions to protect the water system through eco-friendly environmental activities. 2 Study area The Thamirabarani River basin is located in the coastal district of Thoothukudi, Southern Tamil Nadu, India. This district was detached from the adjacent Tirunelveli district in 1986 on the basis of primary augmentation of industrial efficiencies and relevant coastal economic evolutions (Selvam et al., 2013). Geographically, it extends between N 8.5838889–8.6194444 latitude and E 77.9225–78.1297222 longitude and includes 223.32 sq.km of the total delta basin (Fig. 1 ). The Thamirabarani River originates from the Pothigai hills of the Western Ghats and flows through the districts of Thoothukudi and Tirunelveli before joining the Bay of Bengal. It provides water for irrigation and electricity production (Arisekar et al., 2020). The primary crops include paddy, banana, ground nut, brinjal, ragi, sorghum, coconut, pulses, ginger, tea and rubber in upper reaches of the river course (Arisekar et al., 2018).Fig. 1 Location map of the study area along with sampling points. Fig. 1 The delta region is underlined by the Archaean gneisses followed by granites and charnockites (Jesuraja et al., 2021). Alluvium (Quaternary), Dune Terri sands and Tertiary sediments outcrop in the delta region and the marine and fluvial-marine deposits are present along the coast (Narayanaswamy and Lakshmi, 1967; Magesh et al., 2016; Satheeskumar et al., 2020). The primary land use encompasses agricultural lands, barren lands and salt pans other than emigrations and aquifers (Satheeskumar et al., 2020). Both the groundwater and surface water are vulnerable to anthropogenic activities related to fishing and tourism (Selvam et al., 2021), both of which generated annual revenue of 8.9 crores in 2017–2018 (source: http://www.townpanchayat.in/tiruchendur). All chemical plants are pertinent to salt industry, petrochemicals and plastics industry (Jesuraja et al., 2021). Small scale industries belong to paper, soft drink manufacturing, textile, PVC pipe manufacturing, and soap manufacturing. 3 Sampling and analysis A total of 20 water samples were collected uniformly from the Punnakayal estuary of Thamirabarani River basin, and subsequently located in Fig. 1 using GPS (HANNA 2130). These samples were collected in the pre lockdown period (28–29 January 2020), and during the COVID-19 lockdown (6–7 May 2020). The water samples were collected from 10 cm below the water surface and stored in 2 L polyethylene containers. As per the guidelines of American Public Health Association (APHA), we used ultrapure HNO3 for on-site acidification (pH < 2) to avoid microbial activity and adsorption/precipitation on the bottle walls (APHA, 1995). Different physical parameters such as pH, total dissolved solids (TDS), electrical conductivity (EC), turbidity, and dissolved oxygen (DO) were measured using deluxe water and soil analysis kit (model no: 191). In laboratory, the UV–visible spectrophotometer (Systronic), quantified NO3 and the AAS (Atomic Absorption Spectroscopy Perkin Elmer, Elan Drce) measured the absorptions of metals (Cr, Cu, Fe, Cd, Pb and Zn) and metalloids (As) by using The NIST (National Institute of Standards and Technology) standard (1640a) for the QA/QC (Quality Assurance and Quality Control) resolution.“Cetripur” was used for the multi-element (Merck) calibration. The MPN (Most Probable Number) method (ISO, 2000) and estimation of total coliform, fecal coliform, Escherichia coli, and fecal streptococci determined the microbial quality. The MPN method also estimated the number of coliforms of lactose enzymes that produce gas per 100 ml of water sample. 4 Computations of metal pollution codes 4.1 Heavy metal pollution index (HMPI) HMPI comprehensively assessed the influence of each dissolved heavy metal on overall surface water quality (Mohan et al., 1996; Vetrimurugan et al., 2016; Jorfi et al., 2017; Wagh et al., 2018; Rezaei et al., 2019; Karaouzas et al., 2020). It was computed using the formula:(1) HMPI=∑i=1nWiQi∑i=1nWi Where n and Wi refer to number of heavy metals and unit weight of the ith heavy metal, respectively. In step 1, the sub-index (Qi) of ith heavy metal was computed using equation (2):(2) Qi=∑i=1n|Mcon−Ii|Si−Ii×100 Where Mcon (μg/L) refers the computed value of ith heavy metal. Si is the standard permissible of ith metal for drinking purpose (World Health Organization, 2017) for the heavy metals (μg/L) and Ii refers to the ideal limits of ith heavy metal. In step 2, the unit weight (Wi) of each metal was assessed using the equation (3):(3) Wi=kSi Where k refers to proportionality constant and it is considered as 1 for all the metals (Wanda et al., 2012; Qu et al., 2018). Finally, HMPI was computed using the Eq. (1) and it classified the heavy metal pollution in surface water bodies into three categories such as low contamination (HMPI<15), medium contamination (HMPI: 15–30) and high contamination (HMPI>30, Edet and Offiong, 2002; Wanda et al., 2012; Qu et al., 2018; Zakir et al., 2020). 4.2 Human health risk index analysis Consumption of drinking water contaminated with toxic metals increases the risk of non-carcinogenic and carcinogenic diseases in humans (Bineshpour et al., 2020: Qu et al., 2018). We utilized the methods specified by U.S. Environmental Production Agency (USEPA) to appraise the non-carcinogenic (As, Cr, Fe, Cu, Zn, Cd, and Pb) and carcinogenic risks from the dissolved metals (As, Cr, Cd, and Pb) by following USEPA (2013). 4.2.1 Non-carcinogenic health risk approach USEPA (2004) initiated a health risk technique for measuring the non-cancerous human health risks from the heavy metal elements in groundwater and surface water through ingestion, inhalation and exposure to skin. Primarily, the risk was caused by direct water intake and absorption or skin contact (Saha et al., 2017; Qu et al., 2018; Mukherjee et al., 2020; Saleem et al., 2019; USEPA, 2020). It computes the pollutant dose consumed in human using chronic daily intake (CDI), which reflects the dose of pollutants in kilogram per day captivated through the digestion pathway (CDIingestion)and dermal absorption (CDIdermal) using Eqs. (4), (5)), respectively (USEPA, 2011; Zhang et al., 2017; Jehan et al., 2020; Tokatli and Ustaoğlu, 2020).(4) CDIingestion=Conwater×(IR×EF×ED)(BW×AT) (5) CDIdermal=Conwater×(SA×Kp×ET×EF×ED×CF)(BW×AT) where, Conwater refers to trace metal concentration in surface water (μg/L). Table 1 reveals the non-carcinogenic health impact parameters and their input assumptions used for estimating the exposure to heavy metals through intake and skin absorption.Table 1 Values of parameters used for calculating health risk assessment through oral and dermal exposures. Table 1Parameters Units Children Adult Reference Ingestion Rate (IR) L/day 0.64 2 Xiao et al., 2019 Exposure Frequency (EF) days/year 365 Subba Rao et al., 2019 Exposure Duration (ED) years 6 70 USEPA, 2011 Body Weight (BW) kg 20 70 Tokatli and Ustaoğlu, 2020 Averaging Time (AT) days 2190 25550 Jehan et al., 2020 Skin Area (SA) cm2 6600 18000 Tokatli and Ustaoğlu, 2020 Permeability Coefficient (Kp) cm/h 0.002 for Cr and 0.001 for other metals Qu et al., 2018 Exposure Time (ET) h/day 1 0.58 Naz et al., 2016 Conversion Factor (CF) L/cm3 0.001 Tokatli and Ustaoğlu, 2020 Reference dose (RfD) (μg/kg/day) Ingestion: 0.3 for As, 1.4 for Pb, 0.5 for Cd, 40 for Cu, 300 for Zn, 300 for Fe, 3 for Cr Wu et al., 2009 Dermal absorption: 0.123 for As, 0.42 for Pb, 0.005 for Cd, 12 for Cu, 60 for Zn, 45 for Fe, 0.015 for Cr In second step, we calculated the Hazard quotient (HQ) from CDI (CDIingestion and CDIdermal) and RfD (RfD ingestion and RfD dermal) using Eq. (6) (Saha and Paul, 2019; Imran et al., 2019);(6) HQingestion/HQdermal=CDIingestion/CDIdermalRfDingestion/RfDdermal At the final step, the total potential non-carcinogenic risks were appraised by estimating the hazard index (HI) by using Eq. (7) (Rupakheti et al., 2017; Jehan et al., 2020; Karthikeyan et al., 2021);(7) HI=HQingestion+HQdermal=CDIRfd The toxic metals with HI and HQ of >1 can have adverse effects and with <1 have no adverse effects on human health (USEPA, 1989; Vetrimurugan et al., 2016, 2017; Yang et al., 2017; Mohammadi et al., 2019). 4.2.2 Carcinogenic health risk approach We considered the heavy metals as carcinogenic to humans (As) and likely to be carcinogenic to humans (Cd) in order to assess the carcinogenic and non-carcinogenic risks according to the IARC report (IARC, 2013). Both Pb and Cr were included in the non-carcinogenic risk assessment. Even though, Cu, Fe and Zn are not classified in the IARC report, they were involved in the non-carcinogenic risk assessment (e.g.Chan et al., 1998). The carcinogenic risks (CR) were evaluated using the following equations (Eqs. (8), (9))) (Benhaddya, 2020):(8) CR=CDI×CSF (9) TCR=CRingestion+CRdermal The standard assumption values of cancer slope factor (CSF) to measure the risks are 0.0005, 0.0015, 0.0061 and 0.0000085 ppb/day for Cr, As, Cd and Pb, respectively (Gao et al., 2019; Tokatli and Ustaoğlu, 2020). The acceptable or tolerable carcinogenic risk range is 0.000001–0.0001. If CR or TCR of an element exceeds 0.0001, the effect might be detrimental on human health (Qu et al., 2018; Mohammadi et al., 2019; Gao et al., 2019; Tabassum et al., 2019). 4.3 Heavy metal toxicity load approach HMPI or HI (HMPI>100 or HI > 1) indicate the suitability of water as it reveals the accurate ratio of surplus metal (Saha and Paul, 2019; Proshad et al., 2020). It concedes many ideas for predicting and mitigating the pollution of water bodies. This estimation technique predicts the concentration of excess metal in water and the amount that must be removed to make it harmless for human use. HMTL (Heavy metal toxic load) estimates toxic heavy metal in a water source that seduces human health, and it was computed using the following equation (8) (Saha and Paul, 2019):HMTL=∑i=1nc.HIS where c, n and HIS denote the concentration of heavy metal, number of heavy metals and risk severity score, respectively. HIS scores for As (1676), Cr (1149), Cu (805), Zn (913), Cd (1318), and Pb (1531) were considered from ATSDR (2018). They are based on the frequency of hazardous material occurrence on National Priority List (NPL), prepared by ATSDR. The permissible limit of HMTL was below 5888.527 mg/l (Saha and Paul, 2019; Proshad et al., 2020). The HIS score was multiplied by acceptable limit of specific concentration, which is considered the permissible limit for toxicity load and the permissible toxicity load is given in Table 9. HTML Result identifies that the surplus percentage of removal of heavy metals from surface aquifers beyond the permissible toxicity load is essential for human health. 5 Results and discussion We assessed water quality of the Punnakayal estuary during the lockdown and pre-lockdown periods using the physic-chemical variables, dissolved metals and microbiological parameters. 5.1 Physicochemical parameters Mean pH of the pre-lockdown (7.78), and lockdown (7.63) periods did not show significant difference (Table 2 ). Higher pH of TSW - 11, 12, 20 during the lockdown period was due to warmer water temperatures. Both the mean EC and TDS of the lockdown period were 22.04% and 29.96% lower than the pre-lockdown. TDS of 75% of surface water samples in the pre-lockdown and 50% of samples of lockdown interval exceeded the permissible limit of WHO (2017; 1000 mg/l) (Fig. 2 ). These changes are due to absence of agricultural activities during the first phase of lockdown as well as the dilution effect from rainfall. It also revealed the reduction in water consumption for the industrial purposes. Compared to the Central Pollution Control Board (CPCB) (1979) and BIS (1982) water quality standards, about 45% of both the pre-lockdown (1.26–9.26 mg/l) and lockdown (1.11–6.89 mg/l) samples had DO in class C, revealing unsuitability for drinking. The mean BOD in both periods (5.4 mg/l in pre-lockdown and 4.8 mg/l in lockdown) showed no significance difference and it grouped only few samples in class A, representing the suitability for drinking without conventional treatment but after disinfection (Source: Central Pollution Control Board (CPCB), 1979 and BIS, 1982). Turbidity limits before and during the lockdown are 1.56–9.65 (NTU) and 1.21–6.62 (NTU), respectively. Most samples from both periods exceeded the acceptable limit of BIS (2012). Higher turbidity loads also revealed natural causes such as erosion of more silt and mud and anthropogenic causes such as agriculture, sand mining, construction, and algae from domestic wastes. NO3 − is a prime contaminant in agro-terrains (Adimalla et al., 2018; Zhang et al., 2018; Chandrasekar er al., 2021). It was quantified as 35–72 mg/l during the pre-lockdown and 12–56 mg/l during the lockdown. Comparison with World Health Organization (2017) suggests 50% of pre-lockdown and only 15% of lockdown samples were under high risk. We detected 27% discrepancy between the two periods and plenty of samples of the pre-lockdown period showed high health risk (e.g.Adimalla and Qian, 2019, Fig. 3 ).Table 2 Statistical measures and evaluation of physico-chemical, heavy metals and biological parameters during COVID-19 before and after lockdown period against WHO, USEPA and BIS standards. Table 2Parameters Unit Pre-lockdown Lockdown World Health Organization (2017) USEPA (2009) BIS (2012) % of samples exceed the World Health Organization (2017) Limits Decreased Variation Decreased Variation in % Min Max Mean Min Max Mean Pre-lockdown Lockdown pH – 7.40 8.20 7.78 7.10 8.2 7.63 6.5–8.5 6.5–8.5 6.5–8.5 – – 0.16 1.99 EC μs/cm 550 3568 1448.55 421 2899 1129.30 1500 – 45 25 319.25 22.04 TDS mg/l 456 3215 1718.10 425 1956 1203.40 1000 500 500 75 55 514.7 29.96 DO mg/l 1.26 9.26 5.07 1.11 6.89 4.01 – – – – – 1.06 20.90 BOD mg/l 2.30 9.50 5.50 1.50 9.6 4.80 – – – – – 0.70 12.70 Turbidity NTU 1.56 9.65 4.97 1.21 6.62 3.39 – – 1 – – 1.59 31.90 No3 mg/l 35.0 72.00 52.95 12.0 56 38.35 50 – 45 50 15 14.6 27.57 As mg/l 0.01 0.10 0.06 0.001 0.098 0.05 0.01 – 0.01 100 100 0.014 23.20 Cr mg/l 0.01 0.08 0.05 0.01 0.07 0.04 0.05 0.1 0.01 40 15 0.009 20.88 Fe mg/l 0.12 0.46 0.29 0.10 0.32 0.22 0.3 0.3 0.3 45 5 0.070 24.48 Cu mg/l 0.12 0.30 0.22 0.11 0.30 0.20 2 1.3 1 – – 0.017 7.81 Zn mg/l 0.57 1.00 0.83 0.52 0.999 0.76 – 5 15 – – 0.063 7.60 Cd mg/l 0.00 0.002 0.00 0.00 0.001 0 0.003 0.005 0.01 – – 0.001 41.67 Pb mg/l 0.006 0.019 0.012 0.004 0.009 0.007 0.01 0.015 0.05 65 – 0.005 40.98 Total coliform bacteria MPN/ml/l 6.20 189 94.76 1.20 95 58.91 – – – – – 35.85 37.83 Faecal coliform bacteria MPN/ml/l 8.90 195 97.10 4 85 42.92 – – – – – 54.18 55.80 Escherichia coli MPN/ml/l 0 77 37.75 0 62 30.45 – – – – – 7.3 19.34 Faecal streptococci CFU/ml/l 0 10 4.25 0 10 2.95 – – – – – 1.3 30.59 Fig. 2 Spatial map for TDS distribution during Pre-lockdown and lockdown. Fig. 2 Fig. 3 Interpretation of Nitrate risk during Pre-lockdown and lockdown. Fig. 3 5.2 Metal concentrations Concentrations of chromium (Cr), copper (Cu) and zinc (Zn) showed <20% difference between the lockdown and pre-lockdown periods (Table 2 and Fig. 4 ). However, the concentrations of arsenic (As), iron (Fe), lead (Pb) and cadmium (Cd) decreased >20% during the lockdown in comparison to the pre-lockdown period.Fig. 4 Heavy metal occurrence difference on bar chart during Pre-lockdown and lockdown. Fig. 4 Arsenic: It is one of the most dangerous toxic components, and can lead to immune disorders, reproductive dysfunction and skin cancer (Kabata-Pendias and Szteke, 2015; Kacmaz, 2020; Tokatli and Ustaoğlu, 2020). Concentration of 0.012–0.099 mg/l in pre-lockdown and 0.001–0.098 mg/l during the lockdown suggested that all of them exceeded the permissible limit (World Health Organization, 2017) for drinking (0.01 mg/l). Despite the COVID-19 lockdown, the anthropological activities related to poultry waste, fertilizer plants, brick making plants, pot design making plants and agricultural practices continued in this region (Selvam et al., 2014a, 2017) Chromium: The maximum Cr concentration was 0.08 mg/l in pre-lockdown samples and reduced to 0.07 mg/l in the lockdown samples. About 45% of the pre-lockdown samples exceeded the permissible limit (0.05 mg/l) and only 15% of the lockdown samples exceeded the permissible limit (Fig. 5 ). In both the phases, all samples near the coast exceeded the allowable limit of 0.05 mg/l (World Health Organization, 2017). Reduction in chromium concentration during the lockdown indicated decline in the activities related to the chemical industries (e.g. DCW Industry) (Selvam et al., 2017). Similarly, there was less utilization of petroleum product in heavy vehicle workshops of this region (e.g. Hua et al., 2016; Sakthivel et al., 2016).Fig. 5 Spatial variation map for chromium distribution during Pre-lockdown and lockdown period. Fig. 5 Iron: For local aquifers the main contributors of iron (Fe) are industrial effluent, acid-mine drainage, and sewage. In this research the iron (Fe) concentration is varied from 0.123 to 0.458 mg/l pre lockdown and 0.102–0.321 mg/l during the lockdown (Fig. 6 ). This result shows that 55% of pre-lockdown samples have an acceptable limit of 0.3 mg/l (World Health Organization, 2017), while 5% of samples exceed the allowable limit during the lockdown period. The maximum occurred value of Fe on COVID-19 phases is very close to permissible limit (0.3 mg/l) and shows that this may be due to the shutting of metallurgical industries during the COVID-19 lockdown period, from which wastes are discharged into water bodies and from landfills (Milivojević et al., 2016).Fig. 6 Spatial variation map of iron distribution during Pre-lockdown and lockdown period. Fig. 6 Copper: This trace element is an important nutrient for the human body (Muhammad et al., 2014; Samantara et al., 2017). We detected almost no change between the two COVID-19 stages with mean values of 0.218 mg/l during pre-lockdown and 0.201 mg/l during the lockdown. None of them exceeded the drinking limits (2 mg/l) of the World Health Organization (2017) standard. Zinc: Absence of Zn affects the metabolism and the immune system, resulting in infections in humans, anemia and birth defects in pregnant women and delayed sexual maturity in men (ATSDR, 2005; Samantara et al., 2017; Karunanidhi et al., 2021). Concentration of Zn in pre-lockdown samples (0.568–0.999 mg/l) and lockdown period (0.523–0.999 mg/l) remained similar. All of them remained suitable for drinking (<5 mg/l). It showed the absence of potential sources of Zn such as industries related to rubber, paint, bronze, die-casting metals, brass and other alloys in this region. Cadmium: It ranged from 0 to 0.002 mg/l in pre-lockdown samples and 0–0.001 mg/l in the lockdown samples. Both remained within the permissible limits (World Health Organization, 2017; <0.003 mg/l). The two periods (pre-lockdown and lockdown) showed a significant difference of up to 42%. Cadmium can flow from phosphate fertilizers into the soil and surface water, and it is also sourced from cadmium-based batteries and cadmium coated materials (ATSDR, 2008; Tokatli and Ustaoğlu, 2020). Since there is no homeostatic mechanism to control, the exposure to very low levels of Cd can cause adverse overall effects on humans (Carter and Fernando, 1979). Lead: It varied from 0.006 to 0.019 mg/l with an average of 0.012 mg/l in the pre-lockdown and between 0.004 and 0.009 mg/l with an average of 0.007 mg/l during the lockdown. About 65% of the surface water samples of the pre-lockdown interval exceeded the recommended limit (0.01 mg/l), but all the lockdown samples were within the allowable range for drinking (Fig. 7 ). Industrial discharges from foundries, battery production amenities, contaminated land runoffs and sewage are the main bases of Pb in the Thoothukudi coastal region. (Selvam et al., 2017). Deficiency or minimal disposal of effluents from these industries and manufacturing unit and reduction in petroleum related transportation during the lockdown interval might have led to lower Pb concentration.Fig. 7 Spatial variation map of lead distribution during Pre-lockdown and lockdown period. Fig. 7 5.3 Microbiological parameters The expressive statistics of microbiological concentration are presented in Table 2. Total coliform bacteria population varied between 6.2 and 189 MPN ml/l and 1.2–95 MPN ml/l in the lockdown and pre-lockdown samples, respectively. The maximum population of fecal coliform bacteria was 42.92 MPN ml/l in the lockdown and 97.09 MPN ml/l in the pre-lockdown period. Escherichia coli bacterial population ranged from 0 to 62 MPN ml/l in the lockdown samples and 0 to 77 MPN ml/l in the pre lockdown samples. The reduction in bacteria during the lockdown period was due to closure of fishing companies and tourism contamination (Fig. 8 ). However, all fecal streptococci population was low (Est <10) both in the pre-lockdown and lockdown samples. Escherichia coli Total coliforms, and faecal coliforms are contributed to the water system by humans and other warm-blooded animals. They survive the sewage treatment plants in large numbers and protect their pathogens for a longer period (Selvam et al., 2017). Selvam et al. (2020a) have observed that the closure of industries and other commercial activities in the study area provide favorable condition for the growth of large bacterial population.Fig. 8 Statistical significance of biological parameters (MPN ml/l) in groundwater samples before and during the lockdown period related to COVID-19. Fig. 8 5.4 Pollution Indices Based on the method used by Edet and Offiong (2002), we evaluated the HMPI (heavy metal pollution index) for the pre-lockdown and lockdown phases for As, Cr, Fe, Cu, Zn, Cd, and Pb (Table 3 ). The computed HPI values varied between 15.18 to 81.25, and 10.67 to 46.92 with an average values of 51.81 and 31.61 in the pre-lockdown samples and lockdown samples, respectively. As per the classifications, about 15% and 85% of the pre-lockdown samples were categorized as medium and high pollution, respectively. Similarly, about 20%, 20% and 60% of the lockdown samples were grouped in low pollution, medium pollution and high pollution groups, respectively. Among all the heavy metals, Cd and Pb played important roles in adjusting the HMPI. Results showed no significant depravity between the two COVID-19 phases except for the sampling stations at the central part of the study region (TSW 7 - (15.18–10.67), TSW 8 - (46.75–11.47), TSW 9 - (48.04–12.25), and TSW 10 - (46.36–14.73). The HMPI values of the central part were transferred from the high pollution to low pollution class and there was no change in the rest as they were continued to receive sewage and municipal effluents as well as pollutant from agricultural activities (i.e. Phosphate fertilizers) (Fig. 9 ).Table 3 Heavy metal pollution index (HMPI) and decreased percentage of studied metal in Thamirabarani River for Pre-lockdown and Lockdown phase. Table 3Sampling Point Pre-lockdown phase Lockdown phase Decreased % (Pre lockdown – Lockdown) Pre lockdown Degree of pollution as per HMPI scale Lockdown Degree of pollution as per HMPI scale TSW 1 48.23 High Pollution 35.97 High Pollution 25.41 TSW 2 76.41 High Pollution 39.72 High Pollution 48.02 TSW 3 25.60 Medium Pollution 19.62 Medium Pollution 23.36 TSW 4 34.97 High Pollution 30.03 Medium Pollution 14.14 TSW 5 81.25 High Pollution 39.62 High Pollution 51.23 TSW 6 38.65 High Pollution 30.08 Medium Pollution 22.17 TSW 7 15.18 Medium Pollution 10.67 Low Pollution 29.68 TSW 8 46.75 High Pollution 11.47 Low Pollution 75.46 TSW 9 48.04 High Pollution 12.25 Low Pollution 74.49 TSW 10 46.36 High Pollution 14.73 Low Pollution 68.23 TSW 11 78.95 High Pollution 37.42 High Pollution 52.60 TSW 12 81.25 High Pollution 39.62 High Pollution 51.23 TSW 13 69.68 High Pollution 43.38 High Pollution 37.74 TSW 14 43.56 High Pollution 39.49 High Pollution 9.35 TSW 15 29.79 Medium Pollution 22.12 Medium Pollution 25.77 TSW 16 59.43 High Pollution 34.55 High Pollution 41.86 TSW 17 50.72 High Pollution 44.07 High Pollution 13.10 TSW 18 51.87 High Pollution 46.92 High Pollution 9.55 TSW 19 66.81 High Pollution 42.67 High Pollution 36.14 TSW 20 51.86 High Pollution 43.40 High Pollution 16.32 Min 15.18 Low Pollution (0%) 10.67 Low Pollution (20%) 9.35 Max 81.25 Medium Pollution (15%) 46.92 Medium Pollution (20%) 75.46 Mean 51.81 High Pollution (85%) 31.61 High Pollution (60%) 36.85 Fig. 9 Result of HMPI values display on spatial map for Pre-lockdown and lockdown period. Fig. 9 5.5 Health risk assessment The hazard indices (HI) of non-carcinogenic risk and carcinogenic risk were based on hazard quotients (HQ) of the ingestion and dermal adsorption pathways. They showed the total potential human health risks on children and adults from various heavy metals. 5.5.1 Non-carcinogenic health risk The non-carcinogenic risk for children and adults was evaluated for the toxic As, Cr, Fe, Cu, Zn, Cd and Pb (Table 4, Table 5 ). In the pre-lockdown period for children, the HQ ingestion ranged from 1.280–10.560, 0.107–0.853, 0.013–0.049, 0.098–0.239, 0.061–0.107, 0–0.128 and 0.137–0.434 for As, Cr, Fe, Cu, Zn, Cd and Pb, respectively. Similarly, the dermal pathway HQ ranged from 0.032–0.266, 0.440–3.520, 0.001–0.003, 0.003–0.008, 0.003–0.005, 0.000–1.320 and 0.001–0.004 for As, Cr, Fe, Cu, Zn, Cd and Pb, respectively (Table 4). For children, the HQ ingestion of As exceeded the limit (>1) in all samples and HQ dermal result showed values beyond the hazard quotient limit in 75% samples for Cr and 35% samples for Cd. During lockdown period and for children, the HQ ingestion values ranged from 0.107–10.453, 0.107–0.747, 0.011–0.034, 0.089–0.236, 0.056–0.107, 0.000–0.064 and 0.091–0.206 for As, Cr, Fe, Cu, Zn, Cd and Pb, respectively. The dermal pathway HQ ranged from 0.003–0.263, 0.440–3.080, 0.001–0.002, 0.003–0.008, 0.003–0.005, 0.000–0.660 and 0.001–0.002 for As, Cr, Fe, Cu, Zn, Cd and Pb, respectively. In the lockdown period, the HQ ingestion pathway of As also exceeded the limit (>1) in 100% of the samples and HQ dermal results of Cr remained above the hazard quotient limit in 60% of the surface water samples.Table 4 Non-carcinogenic risk (HQ) (mg/kg/day) among children in the Thamirabarani River water before and during COVID-19 lockdown. Table 4Children Sampling Point Pre-lockdown Lockdown HQingestion HQdermal HQingestion HQdermal As Cr Fe Cu Zn Cd Pb As Cr Fe Cu Zn Cd Pb As Cr Fe Cu Zn Cd Pb As Cr Fe Cu Zn Cd Pb TSW 1 4.800 0.213 0.027 0.151 0.095 0.064 0.320 0.121 0.880 0.002 0.005 0.005 0.660 0.003 2.667 0.213 0.020 0.145 0.085 0.064 0.183 0.067 0.880 0.001 0.005 0.004 0.660 0.002 TSW 2 8.320 0.320 0.049 0.125 0.081 0.128 0.389 0.209 1.320 0.003 0.004 0.004 1.320 0.004 4.160 0.320 0.027 0.115 0.071 0.064 0.206 0.105 1.320 0.002 0.004 0.004 0.660 0.002 TSW 3 5.973 0.853 0.028 0.158 0.070 0.000 0.206 0.150 3.520 0.002 0.005 0.004 0.000 0.002 5.973 0.747 0.022 0.153 0.065 0.000 0.137 0.150 3.080 0.002 0.005 0.003 0.000 0.001 TSW 4 4.800 0.320 0.023 0.127 0.063 0.064 0.137 0.121 1.320 0.002 0.004 0.003 0.660 0.001 3.627 0.320 0.021 0.122 0.060 0.064 0.091 0.091 1.320 0.001 0.004 0.003 0.660 0.001 TSW 5 8.853 0.747 0.042 0.205 0.101 0.128 0.411 0.223 3.080 0.003 0.007 0.005 1.320 0.004 4.480 0.533 0.026 0.200 0.100 0.064 0.183 0.113 2.200 0.002 0.007 0.005 0.660 0.002 TSW 6 5.867 0.213 0.025 0.126 0.061 0.064 0.183 0.148 0.880 0.002 0.004 0.003 0.660 0.002 4.053 0.213 0.015 0.122 0.056 0.064 0.091 0.102 0.880 0.001 0.004 0.003 0.660 0.001 TSW 7 3.733 0.213 0.013 0.098 0.070 0.000 0.137 0.094 0.880 0.001 0.003 0.004 0.000 0.001 3.733 0.213 0.011 0.089 0.065 0.000 0.091 0.094 0.880 0.001 0.003 0.003 0.000 0.001 TSW 8 1.920 0.107 0.017 0.121 0.070 0.064 0.343 0.048 0.440 0.001 0.004 0.004 0.660 0.004 1.280 0.107 0.012 0.106 0.061 0.000 0.137 0.032 0.440 0.001 0.004 0.003 0.000 0.001 TSW 9 1.280 0.107 0.021 0.132 0.073 0.064 0.366 0.032 0.440 0.001 0.005 0.004 0.660 0.004 0.107 0.107 0.012 0.127 0.064 0.000 0.160 0.003 0.440 0.001 0.004 0.003 0.000 0.002 TSW 10 5.973 0.320 0.025 0.138 0.084 0.064 0.274 0.150 1.320 0.002 0.005 0.004 0.660 0.003 3.520 0.320 0.015 0.132 0.070 0.000 0.137 0.089 1.320 0.001 0.005 0.004 0.000 0.001 TSW 11 6.613 0.427 0.034 0.239 0.095 0.128 0.434 0.166 1.760 0.002 0.008 0.005 1.320 0.004 3.520 0.320 0.024 0.230 0.083 0.064 0.183 0.089 1.320 0.002 0.008 0.004 0.660 0.002 TSW 12 8.853 0.747 0.042 0.205 0.101 0.128 0.411 0.223 3.080 0.003 0.007 0.005 1.320 0.004 4.480 0.533 0.026 0.200 0.100 0.064 0.183 0.113 2.200 0.002 0.007 0.005 0.660 0.002 TSW 13 10.133 0.427 0.035 0.170 0.084 0.128 0.274 0.255 1.760 0.002 0.006 0.004 1.320 0.003 9.493 0.213 0.028 0.101 0.070 0.064 0.206 0.239 0.880 0.002 0.003 0.004 0.660 0.002 TSW 14 7.040 0.533 0.033 0.212 0.107 0.064 0.206 0.177 2.200 0.002 0.007 0.005 0.660 0.002 5.973 0.320 0.028 0.189 0.102 0.064 0.183 0.150 1.320 0.002 0.006 0.005 0.660 0.002 TSW 15 6.720 0.853 0.034 0.231 0.099 0.000 0.229 0.169 3.520 0.002 0.008 0.005 0.000 0.002 6.293 0.533 0.031 0.212 0.091 0.000 0.183 0.158 2.200 0.002 0.007 0.005 0.000 0.002 TSW 16 2.347 0.640 0.028 0.205 0.098 0.128 0.206 0.059 2.640 0.002 0.007 0.005 1.320 0.002 1.707 0.640 0.023 0.236 0.088 0.064 0.137 0.043 2.640 0.002 0.008 0.005 0.660 0.001 TSW 17 9.493 0.533 0.031 0.231 0.095 0.064 0.274 0.239 2.200 0.002 0.008 0.005 0.660 0.003 9.493 0.320 0.028 0.196 0.087 0.064 0.206 0.239 1.320 0.002 0.007 0.004 0.660 0.002 TSW 18 10.560 0.853 0.027 0.205 0.105 0.064 0.251 0.266 3.520 0.002 0.007 0.005 0.660 0.003 10.453 0.640 0.027 0.189 0.102 0.064 0.206 0.263 2.640 0.002 0.006 0.005 0.660 0.002 TSW 19 7.040 0.640 0.034 0.186 0.104 0.128 0.251 0.177 2.640 0.002 0.006 0.005 1.320 0.003 5.973 0.533 0.032 0.170 0.102 0.064 0.206 0.150 2.200 0.002 0.006 0.005 0.660 0.002 TSW 20 9.813 0.640 0.043 0.227 0.105 0.064 0.274 0.247 2.640 0.003 0.008 0.005 0.660 0.003 8.960 0.533 0.034 0.186 0.107 0.064 0.183 0.225 2.200 0.002 0.006 0.005 0.660 0.002 Min 1.280 0.107 0.013 0.098 0.061 0.000 0.137 0.032 0.440 0.001 0.003 0.003 0.000 0.001 0.107 0.107 0.011 0.089 0.056 0.000 0.091 0.003 0.440 0.001 0.003 0.003 0.000 0.001 Max 10.560 0.853 0.049 0.239 0.107 0.128 0.434 0.266 3.520 0.003 0.008 0.005 1.320 0.004 10.453 0.747 0.034 0.236 0.107 0.064 0.206 0.263 3.080 0.002 0.008 0.005 0.660 0.002 Mean 6.453 0.485 0.031 0.174 0.088 0.076 0.279 0.162 2.000 0.002 0.006 0.005 0.780 0.003 5.023 0.388 0.023 0.161 0.081 0.044 0.163 0.126 1.600 0.002 0.006 0.004 0.450 0.002 % of samples exceed the limit 100 Nil Nil Nil Nil Nil Nil Nil 75 Nil Nil Nil 35 Nil 100 Nil Nil Nil Nil Nil Nil Nil 60 Nil Nil Nil Nil Nil Table 5 Non-carcinogenic risk (HQ) (mg/kg/day) among adults in the Thamirabarani River water before and during COVID-19 lockdown. Table 5Adults Sampling Point Pre-lockdown Lockdown HQingestion HQdermal HQingestion HQdermal As Cr Fe Cu Zn Cd Pb As Cr Fe Cu Zn Cd Pb As Cr Fe Cu Zn Cd Pb As Cr Fe Cu Zn Cd Pb TSW 1 4.286 0.190 0.024 0.135 0.085 0.057 0.286 0.056 0.411 0.001 0.002 0.002 0.309 0.005 2.381 0.190 0.018 0.129 0.076 0.057 0.163 0.031 0.411 0.001 0.002 0.002 0.309 0.003 TSW 2 7.429 0.286 0.044 0.111 0.072 0.114 0.347 0.098 0.617 0.002 0.002 0.002 0.617 0.006 3.714 0.286 0.024 0.103 0.063 0.057 0.184 0.049 0.617 0.001 0.002 0.002 0.309 0.003 TSW 3 5.333 0.762 0.025 0.141 0.062 0.000 0.184 0.070 1.646 0.001 0.003 0.002 0.000 0.003 5.333 0.667 0.020 0.136 0.058 0.000 0.122 0.070 1.440 0.001 0.002 0.002 0.000 0.002 TSW 4 4.286 0.286 0.020 0.114 0.056 0.057 0.122 0.056 0.617 0.001 0.002 0.002 0.309 0.002 3.238 0.286 0.019 0.109 0.054 0.057 0.082 0.043 0.617 0.001 0.002 0.001 0.309 0.001 TSW 5 7.905 0.667 0.038 0.183 0.090 0.114 0.367 0.104 1.440 0.001 0.003 0.002 0.617 0.007 4.000 0.476 0.023 0.179 0.089 0.057 0.163 0.053 1.029 0.001 0.003 0.002 0.309 0.003 TSW 6 5.238 0.190 0.022 0.112 0.054 0.057 0.163 0.069 0.411 0.001 0.002 0.001 0.309 0.003 3.619 0.190 0.014 0.109 0.050 0.057 0.082 0.048 0.411 0.000 0.002 0.001 0.309 0.001 TSW 7 3.333 0.190 0.012 0.088 0.062 0.000 0.122 0.044 0.411 0.000 0.002 0.002 0.000 0.002 3.333 0.190 0.010 0.079 0.058 0.000 0.082 0.044 0.411 0.000 0.001 0.002 0.000 0.001 TSW 8 1.714 0.095 0.015 0.108 0.062 0.057 0.306 0.023 0.206 0.001 0.002 0.002 0.309 0.006 1.143 0.095 0.011 0.094 0.054 0.000 0.122 0.015 0.206 0.000 0.002 0.001 0.000 0.002 TSW 9 1.143 0.095 0.019 0.118 0.066 0.057 0.327 0.015 0.206 0.001 0.002 0.002 0.309 0.006 0.095 0.095 0.011 0.114 0.057 0.000 0.143 0.001 0.206 0.000 0.002 0.002 0.000 0.003 TSW 10 5.333 0.286 0.022 0.123 0.075 0.057 0.245 0.070 0.617 0.001 0.002 0.002 0.309 0.004 3.143 0.286 0.013 0.118 0.063 0.000 0.122 0.041 0.617 0.000 0.002 0.002 0.000 0.002 TSW 11 5.905 0.381 0.031 0.214 0.085 0.114 0.388 0.078 0.823 0.001 0.004 0.002 0.617 0.007 3.143 0.286 0.021 0.206 0.074 0.057 0.163 0.041 0.617 0.001 0.004 0.002 0.309 0.003 TSW 12 7.905 0.667 0.038 0.183 0.090 0.114 0.367 0.104 1.440 0.001 0.003 0.002 0.617 0.007 4.000 0.476 0.023 0.179 0.089 0.057 0.163 0.053 1.029 0.001 0.003 0.002 0.309 0.003 TSW 13 9.048 0.381 0.031 0.151 0.075 0.114 0.245 0.119 0.823 0.001 0.003 0.002 0.617 0.004 8.476 0.190 0.025 0.090 0.062 0.057 0.184 0.112 0.411 0.001 0.002 0.002 0.309 0.003 TSW 14 6.286 0.476 0.030 0.189 0.095 0.057 0.184 0.083 1.029 0.001 0.003 0.003 0.309 0.003 5.333 0.286 0.025 0.169 0.091 0.057 0.163 0.070 0.617 0.001 0.003 0.002 0.309 0.003 TSW 15 6.000 0.762 0.031 0.206 0.088 0.000 0.204 0.079 1.646 0.001 0.004 0.002 0.000 0.004 5.619 0.476 0.028 0.189 0.082 0.000 0.163 0.074 1.029 0.001 0.003 0.002 0.000 0.003 TSW 16 2.095 0.571 0.025 0.183 0.088 0.114 0.184 0.028 1.234 0.001 0.003 0.002 0.617 0.003 1.524 0.571 0.020 0.211 0.078 0.057 0.122 0.020 1.234 0.001 0.004 0.002 0.309 0.002 TSW 17 8.476 0.476 0.028 0.206 0.085 0.057 0.245 0.112 1.029 0.001 0.004 0.002 0.309 0.004 8.476 0.286 0.025 0.175 0.078 0.057 0.184 0.112 0.617 0.001 0.003 0.002 0.309 0.003 TSW 18 9.429 0.762 0.024 0.183 0.094 0.057 0.224 0.124 1.646 0.001 0.003 0.003 0.309 0.004 9.333 0.571 0.024 0.169 0.091 0.057 0.184 0.123 1.234 0.001 0.003 0.002 0.309 0.003 TSW 19 6.286 0.571 0.031 0.166 0.093 0.114 0.224 0.083 1.234 0.001 0.003 0.003 0.617 0.004 5.333 0.476 0.028 0.151 0.091 0.057 0.184 0.070 1.029 0.001 0.003 0.002 0.309 0.003 TSW 20 8.762 0.571 0.038 0.203 0.094 0.057 0.245 0.115 1.234 0.001 0.004 0.003 0.309 0.004 8.000 0.476 0.031 0.166 0.095 0.057 0.163 0.105 1.029 0.001 0.003 0.003 0.309 0.003 Min 1.143 0.095 0.012 0.088 0.054 0.000 0.122 0.015 0.206 0.000 0.002 0.001 0.000 0.002 0.095 0.095 0.010 0.079 0.050 0.000 0.082 0.001 0.206 0.000 0.001 0.001 0.000 0.001 Max 9.429 0.762 0.044 0.214 0.095 0.114 0.388 0.124 1.646 0.002 0.004 0.003 0.617 0.007 9.333 0.667 0.031 0.211 0.095 0.057 0.184 0.123 1.440 0.001 0.004 0.003 0.309 0.003 Mean 5.762 0.433 0.027 0.155 0.078 0.068 0.250 0.076 0.935 0.001 0.003 0.002 0.365 0.004 4.485 0.346 0.021 0.144 0.073 0.039 0.146 0.059 0.748 0.001 0.003 0.002 0.210 0.003 % of samples exceed the limit 100 Nil Nil Nil Nil Nil Nil Nil 50 Nil Nil Nil Nil Nil 95 Nil Nil Nil Nil Nil Nil Nil 40 Nil Nil Nil Nil Nil In the pre-lockdown period for adults, the HQ ingestion ranged between 1.143–9.429, 0.095–0.762, 0.012–0.044, 0.088–0.214, 0.054–0.095, 0.000–0.114 and 0.122–0.388 for As, Cr, Fe, Cu, Zn, Cd and Pb, respectively. The HQ through dermal pathway varied between 0.015–0.124, 0.206–1.646, 0.000–0.002, 0.002–0.004, 0.001–0.003, 0.000–0.617 and 0.002–0.007 for As, Cr, Fe, Cu, Zn, Cd and Pb, respectively (Table 5). For the adult, the HQ ingestion of As exceeded the limit (>1) in 100% samples and HQ dermal of chromium remained above the hazard quotient limit in 50% of the samples. During the lockdown period, for adults, the HQ values via intake pathway were 0.095–9.333, 0.095–0.667, 0.010–0.031, 0.079–0.211, 0.050–0.095, 0.000–0.057 and 0.082–0.184 for As, Cr, Fe, Cu, Zn, Cd and Pb, respectively. The HQ dermal pathway were 0.001–0.123, 0.206–1.440, 0.000–0.001, 0.001–0.004, 0.001–0.003, 0.000–0.309 and 0.001–0.003 for As, Cr, Fe, Cu, Zn, Cd and Pb, respectively. During the lockdown, As in about 95% samples and Cr in 40% samples exceeded the hazard quotient limits (>1) of ingestion and dermal contact pathways. The pre-lockdown and lockdown HI indices (HI; sum of all HQ values by intake or skin absorption pathway) for children and adults were higher for the water intake pathway compared to the skin contact pathway (Table 6 ). Two-term HI values for ingestion and skin absorption exposure to As and Cr were higher than the acceptable range of non-carcinogenic metals. However, the pollution impact or pollution rate of lockdown period was relatively lower than the pre-lockdown. For example, the average HI intake for children was 6.616 for As and 2.485 for Cr during the pre-lockdown. They decreased to 5.149 (As) and 1.988 (Cr) during the lockdown. The closure of several industries in this region, limited use of petrochemicals in agriculture and reduction of other anthropological contributions such as discharge of domestic wastewater, municipal waste, and chemical waste from industries during the lockdown might have led to less heavy metal contribution and recued health risks.Table 6 The hazard index (HI) of Non-carcinogenic risk (HQ) (mg/kg/day) among children and Adults in the Thamirabarani River water before and during COVID-19 lockdown. Table 6Sampling Point Pre-lockdown Lockdown HI Children HI Adults HI Children HI Adults As Cr Fe Cu Zn Cd Pb As Cr Fe Cu Zn Cd Pb As Cr Fe Cu Zn Cd Pb As Cr Fe Cu Zn Cd Pb TSW 1 4.921 1.093 0.029 0.156 0.100 0.724 0.323 4.342 0.602 0.025 0.137 0.088 0.366 0.291 2.734 1.093 0.022 0.150 0.089 0.724 0.185 2.412 0.602 0.022 0.134 0.078 0.366 0.166 TSW 2 8.529 1.640 0.052 0.129 0.085 1.448 0.393 7.526 0.903 0.045 0.113 0.074 0.731 0.353 4.265 1.640 0.029 0.119 0.075 0.724 0.208 3.763 0.903 0.029 0.107 0.065 0.366 0.187 TSW 3 6.124 4.373 0.029 0.164 0.073 0.000 0.208 5.404 2.408 0.025 0.144 0.064 0.000 0.187 6.124 3.827 0.024 0.158 0.069 0.000 0.139 5.404 2.107 0.024 0.142 0.060 0.000 0.125 TSW 4 4.921 1.640 0.024 0.132 0.066 0.724 0.139 4.342 0.903 0.021 0.116 0.058 0.366 0.125 3.718 1.640 0.023 0.126 0.063 0.724 0.092 3.281 0.903 0.023 0.113 0.055 0.366 0.083 TSW 5 9.076 3.827 0.045 0.212 0.107 1.448 0.416 8.009 2.107 0.039 0.186 0.093 0.731 0.374 4.593 2.733 0.027 0.207 0.105 0.724 0.185 4.053 1.505 0.027 0.185 0.091 0.366 0.166 TSW 6 6.014 1.093 0.027 0.130 0.064 0.724 0.185 5.307 0.602 0.023 0.114 0.056 0.366 0.166 4.155 1.093 0.016 0.126 0.059 0.724 0.092 3.667 0.602 0.016 0.113 0.051 0.366 0.083 TSW 7 3.827 1.093 0.014 0.102 0.073 0.000 0.139 3.377 0.602 0.012 0.089 0.064 0.000 0.125 3.827 1.093 0.012 0.092 0.069 0.000 0.092 3.377 0.602 0.012 0.082 0.060 0.000 0.083 TSW 8 1.968 0.547 0.018 0.125 0.073 0.724 0.346 1.737 0.301 0.016 0.110 0.064 0.366 0.312 1.312 0.547 0.013 0.109 0.064 0.000 0.139 1.158 0.301 0.013 0.098 0.056 0.000 0.125 TSW 9 1.312 0.547 0.023 0.137 0.077 0.724 0.369 1.158 0.301 0.020 0.120 0.067 0.366 0.332 0.109 0.547 0.013 0.132 0.067 0.000 0.162 0.096 0.301 0.013 0.118 0.059 0.000 0.145 TSW 10 6.124 1.640 0.026 0.142 0.088 0.724 0.277 5.404 0.903 0.023 0.125 0.077 0.366 0.249 3.609 1.640 0.016 0.137 0.074 0.000 0.139 3.184 0.903 0.016 0.122 0.064 0.000 0.125 TSW 11 6.780 2.187 0.037 0.247 0.100 1.448 0.439 5.983 1.204 0.032 0.217 0.087 0.731 0.395 3.609 1.640 0.025 0.238 0.087 0.724 0.185 3.184 0.903 0.025 0.214 0.076 0.366 0.166 TSW 12 9.076 3.827 0.045 0.212 0.107 1.448 0.416 8.009 2.107 0.039 0.186 0.093 0.731 0.374 4.593 2.733 0.027 0.207 0.105 0.724 0.185 4.053 1.505 0.027 0.185 0.091 0.366 0.166 TSW 13 10.388 2.187 0.037 0.175 0.088 1.448 0.277 9.167 1.204 0.032 0.154 0.077 0.731 0.249 9.732 1.093 0.030 0.104 0.073 0.724 0.208 8.588 0.602 0.030 0.093 0.064 0.366 0.187 TSW 14 7.217 2.733 0.036 0.219 0.112 0.724 0.208 6.369 1.505 0.031 0.193 0.098 0.366 0.187 6.124 1.640 0.030 0.195 0.107 0.724 0.185 5.404 0.903 0.030 0.175 0.094 0.366 0.166 TSW 15 6.889 4.373 0.037 0.239 0.104 0.000 0.231 6.079 2.408 0.032 0.210 0.091 0.000 0.208 6.452 2.733 0.034 0.219 0.096 0.000 0.185 5.693 1.505 0.034 0.197 0.084 0.000 0.166 TSW 16 2.406 3.280 0.030 0.212 0.104 1.448 0.208 2.123 1.806 0.026 0.186 0.090 0.731 0.187 1.750 3.280 0.025 0.244 0.092 0.724 0.139 1.544 1.806 0.025 0.219 0.080 0.366 0.125 TSW 17 9.732 2.733 0.033 0.239 0.100 0.724 0.277 8.588 1.505 0.029 0.210 0.088 0.366 0.249 9.732 1.640 0.030 0.203 0.091 0.724 0.208 8.588 0.903 0.030 0.182 0.080 0.366 0.187 TSW 18 10.826 4.373 0.029 0.212 0.111 0.724 0.254 9.553 2.408 0.025 0.186 0.096 0.366 0.229 10.716 3.280 0.028 0.195 0.107 0.724 0.208 9.456 1.806 0.028 0.175 0.094 0.366 0.187 TSW 19 7.217 3.280 0.037 0.192 0.110 1.448 0.254 6.369 1.806 0.032 0.169 0.096 0.731 0.229 6.124 2.733 0.034 0.175 0.107 0.724 0.208 5.404 1.505 0.034 0.157 0.094 0.366 0.187 TSW 20 10.060 3.280 0.045 0.235 0.111 0.724 0.277 8.877 1.806 0.039 0.207 0.096 0.366 0.249 9.185 2.733 0.037 0.193 0.112 0.724 0.185 8.105 1.505 0.037 0.173 0.098 0.366 0.166 Min 1.312 0.547 0.014 0.102 0.064 0.000 0.139 1.158 0.301 0.012 0.089 0.056 0.000 0.125 0.109 0.547 0.012 0.092 0.059 0.000 0.092 0.096 0.301 0.012 0.082 0.051 0.000 0.083 Max 10.826 4.373 0.052 0.247 0.112 1.448 0.439 9.553 2.408 0.045 0.217 0.098 0.731 0.395 10.716 3.827 0.037 0.244 0.112 0.724 0.208 9.456 2.107 0.037 0.219 0.098 0.366 0.187 Mean 6.616 2.485 0.033 0.180 0.092 0.856 0.282 5.838 1.368 0.028 0.158 0.080 0.432 0.254 5.149 1.988 0.025 0.167 0.086 0.494 0.165 4.544 1.094 0.025 0.149 0.075 0.249 0.148 % of samples exceed the limit 100 90 Nil Nil Nil Nil Nil 100 60 Nil Nil Nil Nil Nil 95 90 Nil Nil Nil Nil Nil 95 40 Nil Nil Nil Nil Nil 5.5.2 Carcinogenic health risk The carcinogenic risk (CR) was computed for As, Cr, Cd and Pb and Table 7 presents the total carcinogenic risk (TCR; sum of CR from ingestion and dermal contact exposure) for children and adults. In the pre-lockdown samples and for children, the TCR varied between 0.0005819–0.0048010, 0.0001633–0.0013064, 0–0.0003944 and 0.0000017–0.0000052 for As, Cr, Cd and Pb, respectively. Similarly for adults, the TCR values in the pre-lockdown samples varied between 0.0005171–0.0042658, 0.0001462–0.0011693, 0–0.0003526 and 0.00000155–0.0000047 for As, Cr, Cd and Pb, respectively. In carcinogenic elements the following elements are As, Cr and Cd (>95% samples) beyond the permissible carcinogenic limit. Only Pb (100%) remained within the acceptable or tolerable carcinogenic limit (0.000001–0.0001) for children and adults.Table 7 The total carcinogenic risk (TCR) among children and Adults in the Thamirabarani River water before and during COVID-19 lockdown. Table 7Sampling Point Pre-lockdown Lockdown TCR in Children TCR in Adults TCR in Children TCR in Adults As Cr Cd Pb As Cr Cd Pb As Cr Cd Pb As Cr Cd Pb TSW 1 0.0021823 0.0003266 0.0001972 0.0000039 0.001939 0.0002923 0.0001763 0.0000034 0.0012124 0.0032066 0.0001972 0.0000022 0.0010772 0.0002888 0.0001752 0.0000020 TSW 2 0.0037826 0.0004899 0.0003944 0.0000047 0.0033609 0.0004385 0.0003526 0.0000042 0.0018913 0.0048099 0.0001972 0.0000025 0.0016805 0.0004332 0.0001752 0.0000022 TSW 3 0.0027157 0.0013064 0 0.0000025 0.002413 0.0011693 0 0.0000022 0.0027157 0.0112231 0 0.0000017 0.002413 0.0010108 0 0.0000015 TSW 4 0.0021823 0.0004899 0.0001972 0.0000017 0.001939 0.0004385 0.0001763 0.0000015 0.0016488 0.0048099 0.0001972 0.0000011 0.001465 0.0004332 0.0001752 0.0000098 TSW 5 0.0040251 0.0011431 0.0003944 0.0000049 0.0035764 0.0010231 0.0003526 0.0000044 0.0020368 0.0080165 0.0001972 0.0000022 0.0018097 0.000722 0.0001752 0.0000020 TSW 6 0.0026672 0.0003266 0.0001972 0.0000022 0.0023699 0.0002923 0.0001763 0.0000020 0.0018428 0.0032066 0.0001972 0.0000011 0.0016374 0.0002888 0.0001752 0.0000010 TSW 7 0.0016973 0.0003266 0 0.0000017 0.0015081 0.0002923 0 0.0000015 0.0016973 0.0032066 0 0.0000011 0.0015081 0.0002888 0 0.0000010 TSW 8 0.0008729 0.0001633 0.0001972 0.0000041 0.0007756 0.0001462 0.0001763 0.0000037 0.0005819 0.0016033 0 0.0000017 0.0005171 0.0001444 0 0.0000015 TSW 9 0.0005819 0.0001633 0.0001972 0.0000044 0.0005171 0.0001462 0.0001763 0.0000040 4.85E-05 0.0016033 0 0.0000019 0.0000431 0.0001444 0 0.0000017 TSW 10 0.0027157 0.0004899 0.0001972 0.0000033 0.002413 0.0004385 0.0001763 0.0000030 0.0016003 0.0048099 0 0.0000017 0.0014219 0.0004332 0 0.0000015 TSW 11 0.0030067 0.0006532 0.0003944 0.0000052 0.0026715 0.0005846 0.0003526 0.0000047 0.0016003 0.0048099 0.0001972 0.0000022 0.0014219 0.0004332 0.0001752 0.0000020 TSW 12 0.0040251 0.0011431 0.0003944 0.0000050 0.0035764 0.0010231 0.0003526 0.0000044 0.0020368 0.0080165 0.0001972 0.0000022 0.0018097 0.000722 0.0001752 0.0000020 TSW 13 0.004607 0.0006532 0.0003944 0.0000033 0.0040934 0.0005846 0.0003526 0.0000030 0.0043161 0.0032066 0.0001972 0.0000025 0.0038349 0.0002888 0.0001752 0.0000022 TSW 14 0.0032007 0.0008165 0.0001972 0.0000025 0.0028438 0.0007308 0.0001763 0.0000022 0.0027157 0.0048099 0.0001972 0.0000022 0.002413 0.0004332 0.0001752 0.0000020 TSW 15 0.0030552 0.0013064 0 0.0000028 0.0027146 0.0011693 0 0.0000025 0.0028612 0.0080165 0 0.0000022 0.0025422 0.000722 0 0.0000020 TSW 16 0.0010669 0.0009798 0.0003944 0.0000025 0.0009479 0.0008769 0.0003526 0.0000022 0.0007759 0.0096198 0.0001972 0.0000017 0.0006894 0.0008664 0.0001752 0.0000015 TSW 17 0.0043161 0.0008165 0.0001972 3.298E-06 0.0038349 0.0007308 0.0001763 0.0000030 0.0043161 0.0048099 0.0001972 0.0000025 0.0038349 0.0004332 0.0001752 0.0000022 TSW 18 0.004801 0.0013064 0.0001972 0.0000030 0.0042658 0.0011693 0.0001763 0.0000027 0.0047525 0.0096198 0.0001972 0.0000025 0.0042227 0.0008664 0.0001752 0.0000022 TSW 19 0.0032007 0.0009798 0.0003944 0.0000030 0.0028438 0.0008769 0.0003526 0.0000027 0.0027157 0.0080165 0.0001972 0.0000025 0.002413 0.000722 0.0001752 0.0000022 TSW 20 0.0044615 0.0009798 0.0001972 0.0000033 0.0039641 0.0008769 0.0001763 0.0000030 0.0040736 0.0080165 0.0001972 0.0000022 0.0036194 0.000722 0.0001752 0.0000020 Min 0.0005819 0.0001633 0 0.0000017 0.0005171 0.0001462 0 0.0000015 0.0000485 0.0016033 0 0.0000011 0.0000431 0.0001444 0 0.0000010 Max 0.004801 0.0013064 0.0003944 0.0000052 0.0042658 0.0011693 0.0003526 0.0000047 0.0047525 0.0112231 0.0001972 0.0000025 0.0042227 0.0010108 0.0001752 0.0000022 Mean 0.0029339 0.0007423 0.0002331 0.0000034 0.0026069 0.0006644 0.0002084 0.0000030 0.0022837 0.0058302 0.0001345 0.0000020 0.0020291 0.0005251 0.0001195 0.0000017 % of sample exceeding 100 100 100 Nil 100 100 100 Nil 95 100 100 Nil 95 100 100 Nil In the lockdown samples, the TCR for children varied between 0.0000485–0.0047525, 0.0016033–0.0112231, 0–0.0001972 and 0.0000011–0.0000025 for As, Cr, Cd and Pb, respectively. The TCR values for adults varied between 0.0000431–0.0042227, 0.0001444–0.0010108, 0–0.0001752 and 0.0000009–0.0000022 for As, Cr, Cd and Pb, respectively. Again all the toxic metals (As, Cr and Cd), except for Pb exceeded the carcinogenic range (0.000001–0.0001) causing risk to children and adults. In this study did not show any significant improvements in surface water pollution but compare to pre-lockdown period, during lockdown period the pollution level is much reduced. 5.5.3 Heavy metal toxicity load We computed HMTL to evaluate the concentration of pollutants that might cause non-carcinogenic health risk and furnished the data about the percentage of metals that needs removal from the specific samples (Table 8 ). The ranges of As, Cr, Cu, Zn, Cd, and Pb that are the most threatening to human health were selected from the ATSDR material priority list to calculate HTML (ATSDR, 2017). It varied between 773.70 and 1382.30 mg/l with an average of 1101.78 mg/l for the pre-lockdown samples and between 665.63 and 1311.45 mg/l with a mean of 990.04 mg/l for the lockdown samples. DCW industrial waste leaks and increased usages of agro-based petrochemicals in the study area may lead to increased HTML results. It is necessary to remove 90%, 38%, 47% of As, Cr, and Pb in the pre-lockdown samples. Similarly, almost similar amount of As (90%) and slightly less Cr (29%) must be removed from the lockdown samples to make it suitable for human health. However, Pb remained below the permissible toxicity load in all the lockdown samples. In both the pre-lockdown and lockdown periods, the concentrations of Cu, Zn and Cd were suitable for human activities (Table 9 ).Table 8 Heavy metal toxicity load of the river surface water before and during COVID-19 lockdown. Table 8Sampling Point Pre-lockdown Lockdown Toxicity of heavy metals (mg/l) HMTL Toxicity of heavy metals (mg/l) HMTL As Cr Cu Zn Cd Pb As Cr Cu Zn Cd Pb TSW 1 75.42 22.98 152.15 817.14 1.32 21.43 1090.43 41.90 22.98 145.71 725.84 1.32 12.25 949.99 TSW 2 130.73 34.47 125.58 690.23 2.64 26.03 1009.67 65.36 34.47 115.92 608.06 1.32 13.78 838.91 TSW 3 93.86 91.92 159.39 595.28 0.00 13.78 954.22 93.86 80.43 153.76 557.84 0.00 9.19 895.07 TSW 4 75.42 34.47 128.00 537.76 1.32 9.19 786.15 56.98 34.47 122.36 516.76 1.32 6.12 738.01 TSW 5 139.11 80.43 206.08 867.35 2.64 27.56 1323.16 70.39 57.45 201.25 852.74 1.32 12.25 1195.40 TSW 6 92.18 22.98 126.39 518.58 1.32 12.25 773.70 63.69 22.98 122.36 477.50 1.32 6.12 693.97 TSW 7 58.66 22.98 99.02 595.28 0.00 9.19 785.12 58.66 22.98 89.36 558.76 0.00 6.12 735.88 TSW 8 30.17 11.49 121.56 597.10 1.32 22.97 784.60 20.11 11.49 106.26 518.58 0.00 9.19 665.63 TSW 9 20.11 11.49 132.83 629.06 1.32 24.50 819.30 1.68 11.49 128.00 546.89 0.00 10.72 698.77 TSW 10 93.86 34.47 138.46 716.71 1.32 18.37 1003.18 55.31 34.47 132.83 600.75 0.00 9.19 832.54 TSW 11 103.91 45.96 240.70 812.57 2.64 29.09 1234.86 55.31 34.47 231.84 709.40 1.32 12.25 1044.59 TSW 12 139.11 80.43 206.08 867.35 2.64 27.56 1323.16 70.39 57.45 201.25 852.74 1.32 12.25 1195.40 TSW 13 159.22 45.96 170.66 719.44 2.64 18.37 1116.29 149.16 22.98 101.43 595.28 1.32 13.78 883.95 TSW 14 110.62 57.45 213.33 912.09 1.32 13.78 1308.58 93.86 34.47 189.98 872.83 1.32 12.25 1204.70 TSW 15 105.59 91.92 232.65 845.44 0.00 15.31 1290.90 98.88 57.45 213.33 781.53 0.00 12.25 1163.44 TSW 16 36.87 68.94 206.08 842.70 2.64 13.78 1171.01 26.82 68.94 237.48 751.40 1.32 9.19 1095.13 TSW 17 149.16 57.45 232.65 817.14 1.32 18.37 1276.08 149.16 34.47 197.23 744.10 1.32 13.78 1140.05 TSW 18 165.92 91.92 206.08 900.22 1.32 16.84 1382.30 164.25 68.94 189.98 872.83 1.32 13.78 1311.09 TSW 19 110.62 68.94 186.76 892.91 2.64 16.84 1278.71 93.86 57.45 170.66 872.83 1.32 13.78 1209.89 TSW 20 154.19 68.94 228.62 900.22 1.32 18.37 1371.66 140.78 57.45 187.57 912.09 1.32 12.25 1311.45 Table 9 Percentage of removal of heavy metal to reduce pollution load in the Thamirabarani River surface water with respect to pre-lockdown and lockdown period. Table 9Sampling Point Pre-lockdown Lockdown % of heavy metal removal required % of heavy metal removal required As Cr Cu Zn Cd Pb As Cr Cu Zn Cd Pb TSW 1 77 a a a a 29 59 a a a a a TSW 2 87 a a a a 41 74 a a a a a TSW 3 82 38 a a a a 82 29 a a a a TSW 4 77 a a a a a 70 a a a a a TSW 5 88 29 a a a 44 76 1 a a a a TSW 6 82 a a a a a 73 a a a a a TSW 7 71 a a a a a 71 a a a a a TSW 8 44 a a a a 33 15 a a a a a TSW 9 15 a a a a 38 a a a a a a TSW 10 82 a a a a 17 69 a a a a a TSW 11 84 a a a a 47 69 a a a a a TSW 12 88 29 a a a 44 76 1 a a a a TSW 13 89 a a a a 17 89 a a a a a TSW 14 85 1 a a a a 82 a a a a a TSW 15 84 38 a a a a 83 1 a a a a TSW 16 54 17 a a a a 37 17 a a a a TSW 17 89 1 a a a 17 89 a a a a a TSW 18 90 38 a a a 9 90 17 a a a a TSW 19 85 17 a a a 9 82 1 a a a a TSW 20 89 17 a a a 17 88 1 a a a a Permissible toxicity load (mg/l) 16.76 57.45 1610 2739 3.95 15.31 16.76 57.45 1610 2739 3.95 15.31 a Denotes that those samples are within permissible toxicity load. 6 Remediation for human welfare Our results suggested that the surface water of Punnakayal estuary in the Thamirabarani River system poses severe non-carcinogenic and carcinogenic hazards to human health. The pollution sources such as industrial effluents, domestic wastewater, sand mining and agro-ventures can be minimized by proper management of the surrounding petrochemical, and beverages manufacture units as well as improving the wastewater drainage system. The lockdown period bestowed a good opportunity to understand the paramount importance of nature in our daily lives. Furthermore, it also provided an insight to realize that the conservation and sustainability of natural water systems can be inhibited by effectively managing the pollution sources. According to HTML results, As, Cr and Pb required greater attention in the pre-lockdown samples, whereas only As and Cr were peril to human in the lockdown samples. Therefore, the implementation of pertinent strategy technique of water quality management might help to minimize the pollution of water bodies. In environmental studies various researchers were proposed various innovative solutions for arsenic and chromium remediation, especially Marinho et al. (2019) was discussed various aqueous solutions and the analytical methods used for their detection and quantification of arsenic and chromium elements. This study advocates to form a special panel to routinely monitor the surface water quality and mitigate the risk from exposure to potential heavy metals, especially from As and Cr in the Thamirabarani River ecosystem. 7 Conclusions The present study evaluated the influence of COVID-19 pandemic lockdown on surface water quality of the Punnakayal estuary in the Thamirabarani River system of south India by estimating reduced absorptions of As, Cr, Cu, Cd, Fe, Pb and Zn. Toxic heavy element contamination risk assessment codes, health risk assessment methods and some pollution load approach described the water quality prior to lockdown and the lockdown periods. We did not observe any changes in the order of heavy metal enrichments (Zn > Fe > Cu > As > Cr > Pb > Cd) in both phases as the industrial ejects, domestic sewage and agricultural applications continued during the lockdown period. However, the quantity or impurity ratio was reduced compared to the pre-lockdown period. In the pre-lockdown surface water, the concentrations of Cu, Zn and Cd remained within permissible limits of World Health Organization (2017) in all samples and hazardous As, Cr, Fe and Pb exceeded the permissible limits in 100%, 40%, 45% and 65% samples, respectively. During the lockdown period, As remained similar with 100% samples exceeding the permissible limit, but relatively less samples had Cr (15%) and Fe (5%) above the permissible limits. HQ non-carcinogenic risk on children and adults from the ingestion and skin absorption of hazardous Fe, Cu, Zn, Cd and Pb were within the reliable range for both the periods. HI results, however, indicated more delicate to non-cancer risks in children compared to adults from both As and Cr. TCR values also demonstrated “higher risk of cancer” in children and adults from As, Cr, and Cd, even during the lockdown and “no carcinogenic dilemma” from Pb. Credit author statement S. Selvam: Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration, Methodology, Funding acquisition. K.Jesuraja: Writing – original draft, Data curation, Resources. Priyadarsi D. Roy: Writing – review & editing, Investigation, Project administration. S. Venkatramanan: Data curation, Resources, Software, Visualization. Ramsha Khan- Software, Resources, Formal analysis. Saurabh Shukla: Software, Resources, Formal analysis. D. Manimaran: Software, Resources, Formal analysis. P. Muthukumar: Software, Resources, Formal analysis. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements The first author (S. Selvam) acknowledges the financial support of Department of Science and Technology – SERB- ECR, New Delhi (Ref.No. F.ECR/2018/001749). The authors are also grateful to Shri A.P.C.V. Chockalingam (Secretary), Dr. C.Veerabahu (Principal), of the V.O.C College, Tuticorin, for their supports during this study. ==== Refs References Adimalla N. Li P. Venkatayogi S. Hydrogeochemical evaluation of groundwater quality for drinking and irrigation purposes and integrated interpretation with water quality index studies Environ. Process 5 2 2018 363 383 10.1007/s40710-018-0297-4 Adimalla N. Qian H. Groundwater quality evaluation using water quality index (WQI) for drinking purposes and human health risk (HHR) assessment in an agricultural region of Nanganur, south India Ecotoxicol. Environ. Saf. 176 2019 153 161 10.1016/j.ecoenv.2019.03.066 30927636 APHA (American Public Health Association) American Water Works Association (AWWA), Water Pollution Control Federation 16thedn Standard Methods for the Examination of Water and Waste Water 1995 1268 (Washington, DC) Arisekar U. Shakila R.J. Jeyasekaran G. Shalini R. Kumar P. Malani A.H. Rani V. Accumulation of organochlorine and pyrethroid pesticide residues in fish, water, and sediments in the Thamirabarani river system of southern peninsular India Environ. Nanotechnol. Monit. Manag. 11 2018 100194 10.1016/j.enmm.2018.11.003 Arisekar U. Shakila R.J. Shalini R. Jeyasekaran G. Human health risk assessment of heavy metals in aquatic sediments and freshwater fish caught from Thamirabarani River, the Western Ghats of South Tamil Nadu Mar. Pollut. Bull. 159 2020 111496 10.1016/j.marpolbul.2020.111496 32738642 ATSDR (Agency for Toxic Substances and Disease Registry) Toxicological Profile for Zinc 2005 US Department of Health and Human Services Washington ATSDR (Agency for Toxic Substances and Disease Registry) Toxicological Profile for Cadmium 2008 Department of Health and Human Services Atlanta, GA, USA ATSDR (Agency for Toxic Substances and Disease Registry) Agency for Toxic Substances and Disease Registry, Substance Priority List 2017 https://www.atsdr.cdc.gov/spl/ ATSDR Toxic Substance Portal: Toxicological Profiles 2018 https://www.atsdr.cdc.gov/toxprofiles/index.asp Benhaddya M.L. Human health risk assessment of heavy metals from surface water of Chott Merouane, Algeria Int. J. Environ. Anal. Chem. 2020 1 18 10.1080/03067319.2020.1751145 Bineshpour M. Payandeh K. Nazarpour A. Sabzalipour S. Status, source, human health risk assessment of potential toxic elements (PTEs), and Pb isotope characteristics in urban surface soil, case study: arak city, Iran Environ. Geochem. Health 2020 1 20 10.1007/s10653-020-00778-x BIS Indian Standard Tolerance Limits for Inland Surface Water Subject to Pollution, IS 2296:1982 1982 Bureau of Indian Standards New Delhi BIS Bureau of Indian Standards Specification for Drinking Water IS:10500:91. Revised 2012 2012 Bureau of Indian Standards New Delhi Bherwani H. Gupta A. Anjum S. Anshul A. Kumar R. Exploring dependence of COVID-19 on environmental factors and spread prediction in India npj Climate and Atmospheric Science 3 1 2020 1 13 Carter D.E. Fernando Q. Chemical toxicology. Part II. Metal toxicity J. Chem. Educ. 56 8 1979 490 Central Pollution Control Board (CPCB) Water Quality Criteria Forfreshwater Classification 1979 CPCB Chan S. Gerson B. Subramaniam S. The role of copper, molybdenum, selenium, and zinc in nutrition and health Clin. Lab. Med. 18 4 1998 673 685 9891606 Chandrasekar T. Keesari T. Gopalakrishnan G. Karuppannan S. Senapathi V. Sabarinathan C. Viswanathan P.M. Occurrence of heavy metals in groundwater along the lithological interface of K/T boundary, peninsular India: a special focus on source, geochemical mobility and health risk Arch. Environ. Contam. Toxicol. 80 2021 183 207 10.1007/s00244-020-00803-1 33392777 Dutta V. Dubey D. Kumar S. Cleaning the River Ganga: impact of lockdown on water quality and future implications on river rejuvenation strategies Sci. Total Environ. 743 2020 140756 10.1016/j.scitotenv.2020.140756 32758842 Edet A.E. Offiong O.E. Evaluation of water quality pollution indices for heavy metal contamination monitoring. A study case from Akpabuyo-Odukpani area, Lower Cross River Basin (southeastern Nigeria) Geo J. 57 2002 295 304 Gao B. Gao L. Gao J. Xu D. Wang Q. Sun K. Simultaneous evaluations of occurrence and probabilistic human health risk associated with trace elements in typical drinking water sources from major river basins in China Sci. Total Environ. 666 2019 139 146 10.1016/j.scitotenv.2019.02.148 30798224 Gautam S. Hens L. COVID-19: impact by and on the environment, health and economy Environ. Dev. Sustain. 22 2020 4953 4954 10.1007/s10668-020-00818-7 32837275 Gautam S. Trivedi U. Global implications of bio-aerosol in pandemic Environ. Dev. Sustain. 22 2020 3861 3865 10.1007/s10668-020-00704-2 34172977 Hua Z. Yinghui J. Tao Y. Min W. Guangxun S. Mingjun D. Heavy metal concentrations and risk assessment of sediments and surface water of the Gan river, China Pol. J. Environ. Stud. 25 4 2016 10.15244/pjoes/62100 IARC Agents Classified by the IARC Monographs 1e120 2013 http://monographs.iarc.fr/ENG/Classification/index.php/ Imran U. Ullah A. Shaikh K. Mehmood R. Saeed M. Health risk assessment of the exposure of heavy metal contamination in surface water of lower Sindh, Pakistan SN Appl. Sci. 1 6 2019 1 10 10.1007/s42452-019-0594-1 ISO ISO 9308-1. Water Quality—Detection and Enumeration of Escherichia coli and Coliform Bacteria. Part 1 2000 Membrane filtration methodISO Geneva, Switzerland Jehan S. Ullah I. Khan S. Muhammad S. Khattak S.A. Khan T. Evaluation of the Swat River, Northern Pakistan, water quality using multivariate statistical techniques and water quality index (WQI) model Environ. Sci. Pollut. Control Ser. 27 31 2020 38545 38558 10.1007/s11356-020-09688-y Jesuraja K. Selvam S. Roy P.D. Venkatramanan S. Chung S.Y. Muthukumar P. Nath Akhila V. Groundwater pollution index (GPI) and GIS-based appraisal of groundwater quality for drinking and irrigation in coastal aquifers of Tiruchendur, South India Environ. Sci. Pollut. Res. 28 23 2021 29056 29074 Jorfi S. Maleki R. Jaafarzadeh N. Ahmadi M. Pollution load index for heavy metals in Mian-Ab plain soil, Khuzestan, Iran Data Brief 15 2017 584 590 10.1016/j.dib.2017.10.017 29071297 Kabata-Pendias A. Szteke B. Trace Elements in Abiotic and Biotic Environments 2015 CRC Press Boca Raton 468 Kachroo V. Novel coronavirus (COVID-19) in India: current scenario International Journal of Research and Review 7 3 2020 435 447 Karaouzas I. Kapetanaki N. Mentzafou A. Kanellopoulos T.D. Skoulikidis N. Heavy metal contamination status in Greek surface waters; a review with application and evaluation of pollution indices Chemosphere 2020 128192 10.1016/j.chemosphere.2020.128192 33297157 Karthikeyan S. Arumugam S. Muthumanickam J. Kulandaisamy P. Subramanian M. Annadurai R. Sekar S. Causes of heavy metal contamination in groundwater of Tuticorin industrial block, Tamil Nadu, India Environ. Sci. Pollut. Control Ser. 2021 1 16 10.1007/s11356-020-11704-0 Kacmaz H. Assessment of heavy metal contamination in natural waters of Dereli, Giresun: an area containing mineral deposits in northeastern Turkey Environ. Monit. Assess. 192 2 2020 1 12 10.1007/s10661-019-8057-0 Karunanidhi D. Aravinthasamy P. Deepali M. Subramani T. Shankar K. Groundwater pollution and human health risks in an industrialized region of southern India: impacts of the COVID-19 lockdown and the monsoon seasonal cycles Arch. Environ. Contam. Toxicol. 80 2021 259 276 10.1007/s00244-020-00797-w 33398395 Kassem I.I. Jaafar H. The potential impact of water quality on the spread and control of COVID-19 in Syrian refugee camps in Lebanon Water Int. 45 5 2020 423 429 10.1080/02508060.2020.1780042 Lian X. Huang J. Huang R. Liu C. Wang L. Zhang T. Impact of city lockdown on the air quality of COVID-19-hit of Wuhan city Sci. Total Environ. 742 2020 140556 10.1016/j.scitotenv.2020.140556 32634686 Magesh N.S. Chandrasekar N. Elango L. Occurrence and distribution of fluoride in the groundwater of the Tamiraparani River basin, South India: a geostatistical modeling approach Environ. Earth Sci. 75 23 2016 1 16 10.1007/s12665-016-6293-y Marinho B.A. Cristóvão R.O. Boaventura R.A. Vilar V.J. As (III) and Cr (VI) oxyanion removal from water by advanced oxidation/reduction processes—a review Environ. Sci. Pollut. Control Ser. 26 3 2019 2203 2227 10.1007/s11356-018-3595-5 Milivojević J. Krstić D. Šmit B. Djekić V. Assessment of heavy metal contamination and calculation of its pollution index for Uglješnica River, Serbia Bull. Environ. Contam. Toxicol. 97 5 2016 737 742 10.1007/s00128-016-1918-0 27638703 Mohan S.V. Nithila P. Reddy S.J. Estimation of heavy metals in drinking water and development of heavy metal pollution index Journal of Environmental Science & Health Part A 31 2 1996 283 289 10.1080/10934529609376357 Muthukumaravel K. Pradhoshini K.P. Vasanthi N. Raja T. Abdul Jaleel M. Arunachalam K.D. Musthafa M.S. Ayyamperumal R. Karuppannan S. Rajagopal R. Alfarhan A. Chandrasekaran M. Chang S.W. Ravindran B. Assessment of seasonal variation in distribution and abundance of plankton and ichthyofaunal diversity in relation to environmental indices of Karankadu Mangrove, South East Coast of India Mar. Pollut. Bull. 173 2021 113142 34801890 Muhammad I. Ashiru S. Ibrahim I.D. Salawu K. Muhammad D.T. Muhammad N.A. Determination of some heavy metals in wastewater and sediment of artisanal gold local mining site of Abare Area in Nigeria J. Environ. Treat. Tech 1 2014 174 182 Mohammadi A.A. Zarei A. Majidi S. Ghaderpoury A. Hashempour Y. Saghi M.H. Ghaderpoori M. Carcinogenic and non-carcinogenic health risk assessment of heavy metals in drinking water of Khorramabad, Iran MethodsX 6 2019 1642 1651 10.1016/j.mex.2019.07.017 31372352 Mukherjee I. Singh U.K. Singh R.P. Kumari D. Jha P.K. Mehta P. Characterization of heavy metal pollution in an anthropogenically and geologically influenced semi-arid region of east India and assessment of ecological and human health risks Sci. Total Environ. 705 2020 135801 10.1016/j.scitotenv.2019.135801 31838424 Nakada L.Y.K. Urban R.C. COVID-19 pandemic: impacts on the air quality during the partial lockdown in São Paulo state, Brazil Sci. Total Environ. 730 2020 139087 10.1016/j.scitotenv.2020.139087 32380370 Narayanaswamy S. Lakshmi P. CharnockiticrocksofTirunelvelli district, madras Geological Society of India 8 1967 38 50 Naz A. Mishra B.K. Gupta S.K. Human health risk assessment of chromium in drinking water: a case study of Sukinda chromite mine, Odisha, India Exposure and Health 8 2 2016 253 264 10.1007/s12403-016-0199-5 Panneerselvam B. Muniraj K. Pande C. Ravichandran N. Thomas M. Karuppannan S. Geochemical evaluation and human health risk assessment of nitrate-contaminated groundwater in an industrial area of South India Environ. Sci. Pollut. Res. 2021 1 18 Patel P.P. Mondal S. Ghosh K.G. Some respite for India's dirtiest river? Examining the Yamuna's water quality at Delhi during the COVID-19 lockdown period Sci. Total Environ. 744 2020 140851 10.1016/j.scitotenv.2020.140851 32755777 Proshad R. Islam S. Tusher T.R. Zhang D. Khadka S. Gao J. Kundu S. Appraisal of heavy metal toxicity in surface water with human health risk by a novel approach: a study on an urban river in vicinity to industrial areas of Bangladesh Toxin Rev. 1–17 2020 10.1080/15569543.2020.1780615 Qiu Y. Shi W. Wu W. Water Quality Improvement during the COVID-19 Outbreak in China 2020 10.2139/ssrn.3664832 SSRN 3664832 Qu L. Huang H. Xia F. Liu Y. Dahlgren R.A. Zhang M. Mei K. Risk analysis of heavy metal concentration in surface waters across the rural-urban interface of the Wen-Rui Tang River, China Environ. Pollut. 237 2018 639 649 10.1016/j.envpol.2018.02.020 29525630 Ravindran A.A. Selvam S. Coastal disaster damage and neotectonic subsidence study using 2D ERI technique in Dhanushkodi, Rameshwaram Island, Tamilnadu, India Middle East J. Sci. Res. 19 8 2014 1117 1122 Rezaei A. Hassani H. Hassani S. Jabbari N. Mousavi S.B.F. Rezaei S. Evaluation of groundwater quality and heavy metal pollution indices in Bazman basin, southeastern Iran Groundwater for Sustainable Development 9 2019 100245 10.1016/j.gsd.2019.100245 Ropkins K. Tate J.E. Early observations on the impact of the COVID-19 lockdown on air quality trends across the UK Sci. Total Environ. 754 2021 142374 10.1016/j.scitotenv.2020.142374 33254916 Rupakheti D. Tripathee L. Kang S. Sharma C.M. Paudyal R. Sillanpää M. Assessment of water quality and health risks for toxic trace elements in urban Phewa and remote Gosainkunda lakes, Nepal Hum. Ecol. Risk Assess. 23 5 2017 959 973 10.1080/10807039.2017.1292117 Saha N. Rahman M.S. Ahmed M.B. Zhou J.L. Ngo H.H. Guo W. Industrial metal pollution in water and probabilistic assessment of human health risk J. Environ. Manag. 185 2017 70 78 10.1016/j.jenvman.2016.10.023 Saha P. Paul B. Assessment of heavy metal toxicity related with human health risk in the surface water of an industrialized area by a novel technique Hum. Ecol. Risk Assess. 25 4 2019 966 987 10.1080/10807039.2018.1458595 Sakthivel A. Sridhar S.G.D. Balasubramanian M. PradnyaKatkar S. Trace elements concentration in groundwater of South Chennai , Tamil Nadu , India Indian J. Appl. Res. 6 2 2016 301 305 Saleem M. Iqbal J. Shah M.H. Seasonal variations, risk assessment and multivariate analysis of trace metals in the freshwater reservoirs of Pakistan Chemosphere 216 2019 715 724 30391893 Samantara M.K. Padhi R.K. Sowmya M. Kumaran P. Satpathy K.K. Heavy metal contamination, major ion chemistry and appraisal of the groundwater status in coastal aquifer, Kalpakkam, Tamil Nadu, India Groundwater for sustainable development 5 2017 49 58 10.1016/j.gsd.2017.04.001 Satheeskumar V. Subramani T. Lakshumanan C. Roy P.D. Karunanidhi D. Groundwater chemistry and demarcation of seawater intrusion zones in the Thamirabarani delta of south India based on geochemical signatures Environ. Geochem. Health 2020 1 14 10.1007/s10653-020-00536-z Selvam S. Sivasubramanian P. Groundwater potential zone identification using geoelectrical survey: a case study from Medak district, Andhra Pradesh, India Int. J. Geomatics Geosci. 3 1 2012 55 Selvam S. Manimaran G. Sivasubramanian P. Hydrochemical characteristics and GIS-based assessment of groundwater quality in the coastal aquifers of Tuticorin corporation, Tamilnadu, India Appl. Water Sci. 3 1 2013 145 159 10.1007/s13201-012-0068-8 Selvam S. Manimaran G. Sivasubramanian P. Balasubramanian N. Seshunarayana T.J.E.E.S. GIS-based evaluation of water quality index of groundwater resources around Tuticorin coastal city, South India Environ. Earth Sci. 71 6 2014 2847 2867 10.1007/s12665-013-2662-y Selvam S. Venkatramanan S. Singaraja C. A GIS-based assessment of water quality pollution indices for heavy metal contamination in Tuticorin Corporation, Tamilnadu, India Arabian J. Geosci. 8 12 2015 10611 10623 10.1007/s12517-015-1968-3 Selvam S. Ravindran A.A. Venkatramanan S. Singaraja C. Assessment of heavy metal and bacterial pollution in coastal aquifers from SIPCOT industrial zones, Gulf of Mannar, South Coast of Tamil Nadu, India Appl. Water Sci. 7 2 2017 897 913 Selvam S. Jesuraja K. Venkatramanan S. Chung S.Y. Roy P.D. Muthukumar P. Kumar M. Imprints of pandemic lockdown on subsurface water quality in the coastal industrial city of Tuticorin, south India: a revival perspective Sci. Total Environ. 738 2020 139848 10.1016/j.scitotenv.2020.139848 32574914 Selvam S. Muthukumar P. Venkatramanan S. Roy P.D. Bharath K.M. Jesuraja K. SARS-CoV-2 pandemic lockdown: effects on air quality in the industrialized Gujarat state of India Sci. Total Environ. 737 2020 140391 10.1016/j.scitotenv.2020.140391 32783875 Selvam S. Jesuraja K. Venkatramanan S. Roy P.D. Kumari V.J. Hazardous microplastic characteristics and its role as a vector of heavy metal in groundwater and surface water of coastal south India J. Hazard Mater. 402 2021 123786 10.1016/j.jhazmat.2020.123786 33254795 Singaraja C. Chidambaram S. Srinivasamoorthy K. Anandhan P. Selvam S. A study on assessment of credible sources of heavy metal pollution vulnerability in groundwater of Thoothukudi districts, Tamilnadu, India Water Quality, Exposure and Health 7 4 2015 459 467 10.1007/s12403-015-0162-x Sinha A. One COVID-19 Positive Infects 1.7 in India, Lower than in Hot Zones 2020 Indian Express https://indianexpress.com/article/coronavirus/coronavirus-india-infection-rate-china6321154/ Sivakumar B. COVID-19 and water Stoch. Environ. Res. Risk Assess. 35 3 2020 531 534 32837312 Stratoulias D. Nuthammachot N. Air quality development during the COVID-19 pandemic over a medium-sized urban area in Thailand Sci. Total Environ. 746 2020 141320 10.1016/j.scitotenv.2020.141320 32768789 Subba Rao N. Sunitha B. Sun L. Spandana B.D. Chaudhary M. Mechanisms controlling groundwater chemistry and assessment of potential health risk: a case study from South India Geochemistry 2019 125568 10.1016/j.chemer.2019.125568 Tabassum R.A. Shahid M. Niazi N.K. Dumat C. Zhang Y. Imran M. Khalid S. Arsenic removal from aqueous solutions and groundwater using agricultural biowastes-derived biosorbents and biochar: a column-scale investigation Int. J. Phytoremediation 21 6 2019 509 518 10.1080/15226514.2018.1501340 30924354 Tokatli C. Ustaoğlu F. Health risk assessment of toxicants in Meriç river delta wetland, Thrace region, Turkey Environ. Earth Sci. 79 18 2020 1 12 10.1007/s12665-020-09171-4 Travaglio M. Yu Y. Popovic R. Selley L. Leal N.S. Martins L.M. Links between air pollution and COVID-19 in England Environ. Pollut. 268 2021 115859 10.1101/2020.04.16.20067405 33120349 USEPA Risk Assessment Guidance for Superfund: Interim Final. Office of Emergency and Remedial Response 1989 US Environmental Protection Agency USEPA Risk Assessment Guidance for Superfund Human Health Evaluation Manual (Part E, Supplemental Guidance for Dermal Risk Assessment). EPA/540/R/99/005 vol. 1 2004 Office of Superfund Remediation and Technology Innovation Washington, DC, USA USEPA Risk Assessment Guidance for Superfund. Volume I: Human Health Evaluation Manual (Part E, Supplemental Guidance for Dermal Risk Assessment), EPA/540/R/99 2011 USEPA Mid Atlantic Risk Assessment. Regional Screening Level (RSL) Summary Table 2013 (Washington DC) USEPA National Primary Drinking Water Regulations 2009 United States Environmental Protection Agency EPA816-F-09-004 http://www.epa.gov/safewater/ USEPA Exposure Assessment Tools by Media -Water and Sediment 2020 https://www.epa.gov/expobox/exposure-assessment-tools-media-water-andsediment Vetrimurugan E. Brindha K. Elango L. Human exposure risk assessment due to heavy metals in groundwater by pollution index and multivariate statistical methods: a case study from South Africa Water (Switzerland) 9 4 2016 10.3390/w9040234 Vetrimurugan E. Brindha K. Elango L. Ndwandwe O.M. Human exposure risk to heavy metals through groundwater used for drinking in an intensively irrigated river delta Appl. Water Sci. 7 6 2017 3267 3280 10.1007/s13201-016-0472-6 Wagh V.M. Panaskar D.B. Mukate S.V. Gaikwad S.K. Muley A.A. Varade A.M. Health risk assessment of heavy metal contamination in groundwater of KadavaRiver Basin, Nashik, India Modeling Earth Systems and Environment 4 3 2018 969 980 10.1007/s40808-018-0496-z Wanda E.M. Gulula L.C. Phiri G. Determination of characteristics and drinking water quality index in Mzuzu City, Northern Malawi Phys. Chem. Earth, Parts A/B/C 50 2012 92 97 10.1016/j.pce.2012.09.004 World Health Organization (WHO) Guidelines for Drinking Water Quality: Fourth Edition Incorporating the First Addendum 2017 World Health Organization Geneva World Health Organization (WHO) Coronavirus Disease 2019 (COVID-19): Situation Report 2020 72 Wu B. Zhao D.Y. Jia H.Y. Zhang Y. Zhang X.X. Cheng S.P. Preliminary risk assessment of trace metal pollution in surface water from Yangtze River in Nanjing Section, China Bull. Environ. Contam. Toxicol. 82 4 2009 405 409 10.1007/s00128-008-9497-3 19165409 Xiao J. Wang L. Deng L. Jin Z. Characteristics, sources, water quality and health risk assessment of trace elements in river water and well water in the Chinese Loess Plateau Sci. Total Environ. 650 2019 2004 2012 10.1016/j.scitotenv.2018.09.322 30290343 Yang Y. Wei L. Cui L. Zhang M. Wang J. Profiles and risk assessment of heavy metals in Great Rift Lakes, Kenya Clean 45 3 2017 1600825 Zakir H.M. Sharmin S. Akter A. Rahman M.S. Assessment of health risk of heavy metals and water quality indices for irrigation and drinking suitability of waters: a case study of Jamalpur Sadar area, Bangladesh Environmental Advances 2 2020 100005 10.1016/j.envadv.2020.100005 Zhang X. Xu J. Wang H. Liu X. Yao D. Lv J. Health risk assessment of a reservoir-type water source in the northeastern region of China Hum. Ecol. Risk Assess. 23 4 2017 751 766 10.1080/10807039.2016.1278518 Zhang Y. Wu J. Xu B. Human health risk assessment of groundwater nitrogen pollution in Jinghui canal irrigation area of the loess region, northwest China Environ. Earth Sci. 77 7 2018 273 10.1007/s12665-018-7456-9

35301998

PMC9753365

NO-CC CODE

2022-12-16 23:25:40

no

Chemosphere. 2022 Jul 14; 298:134027

utf-8

Chemosphere

10.1016/j.chemosphere.2022.134027

oa_other

==== Front J Hepatol J Hepatol Journal of Hepatology 0168-8278 1600-0641 Elsevier S0168-8278(22)03263-9 10.1016/S0168-8278(22)03263-9 Article JHEP at a glance (January 2023) 15 12 2022 1 2023 15 12 2022 78 1 e1e15 2020 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmc

0

PMC9753405

NO-CC CODE

2022-12-16 23:25:42

no

J Hepatol. 2023 Jan 15; 78(1):e1-e15

utf-8

J Hepatol

10.1016/S0168-8278(22)03263-9

oa_other

==== Front J Hepatol J Hepatol Journal of Hepatology 0168-8278 1600-0641 European Association for the Study of the Liver. Published by Elsevier B.V. S0168-8278(22)03139-7 10.1016/j.jhep.2022.10.003 From the Editor's Desk From the Editor’s Desk… Burra Patrizia ∗ Tacke Frank Ratziu Vlad Zeuzem Stefan Sangro Bruno Angeli Paolo ∗ Corresponding author. 15 12 2022 1 2023 15 12 2022 78 1 14 10 10 2022 10 10 2022 © 2022 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. 2022 European Association for the Study of the Liver Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcSelection of the month Liver disease-related mortality rates during the COVID-19 pandemic Gao, Lv et al. aimed to determine the impact of the COVID-19 pandemic on people with liver disease in the US using a national death dataset. They compared observed vs. predicted mortality for 2020-2021 based on trends from 2010-2019 with joinpoint and prediction modelling analyses. Age standardised mortality rates (ASMRs) for alcohol-related liver disease (ALD) dramatically increased between 2010-2019 and 2020-2021 (5 times higher annual percentage change), leading to a higher observed ASMR than predicted for 2020 and 2021. The ASMR rise for ALD was higher in certain ethnic groups and particularly for the 25–44-year age group. The ASMR for non-alcoholic fatty liver disease (NAFLD) also increased during the pandemic while the rates for hepatitis B and C decreased. The increased trend in mortality was also higher in females than in males. Therefore, during the COVID-19 pandemic, the impact of viral hepatitis has slowed down, while the increasing mortality trend for ALD and NAFLD has accelerated. Experimental and translational hepatology Duodenal stem cells could be a source of hepatic progenitor cells in regenerative medicine The gut and liver form a closely interlinked functional unit (“gut-liver axis”), but also share a common cellular origin, as these organs develop from the definitive endoderm forming the foregut. Cardinale, Carpino et al. now demonstrate that submucosal glands in the human duodenum contain cells with stem/progenitor markers and genetic similarities to biliary epithelial cells. Intriguingly, these human duodenal submucosal gland cells form organoids in vitro, acquire hepatobiliary markers and engraft into the livers of immunocompromised mice. In this experimental setting of cell transplantation, the duodenal gland-derived “hepatocytes” protect the hosts from liver injury and promote liver regeneration. This work sheds new light on cellular plasticity in the gut-liver axis and may provide important avenues for regenerative medicine in liver diseases. IL6/STAT3 axis dictates the PNPLA3-mediated susceptibility to NAFLD A single nucleotide polymorphism in PNPLA3 is strongly associated with NAFLD and advanced disease. Park, Zhao, Zhang et al. tried to understand the mechanistic link between this PNPLA3 mutation and the development of NAFLD. They used a multicellular liver culture system that utilises human pluripotent stem cell (hPSC)-derived hepatocytes, stellate cells and macrophages expressing either wild-type or mutated PNPLA3 in a lipotoxic milieu reflecting the circulating levels of disease risk factors in individuals with NAFLD. Under these conditions the PNPLA3-mutated hPSCs recapitulated many key aspects of NAFLD including lipid accumulation, oxidative stress, inflammatory response, and stellate cell activation. This was associated with elevated levels of IL6/STAT3 activity, which is consistent with transcriptomics from patient liver biopsies. Moreover, dampening IL6/STAT3 activity alleviated the I148M-mediated susceptibility to NAFLD, while boosting it in wild-type liver cultures enhanced NAFLD development. Selective inhibition of trans signalling by IL6 (achievable pharmacologically) produced the same inhibitory results on NAFLD development. The preclinical model developed here is suitable for mechanistic dissection of genetic variants and the manipulation of the IL6 pathway opens new therapeutic prospects akin to those existing in inflammatory bowel disease. Viral hepatitis Viral clade is associated with severity of symptomatic genotype 3 HEV infections in Belgium, 2010-2018 HEV genotype 3 infections display a wide spectrum of clinical presentations. Host – but not viral – factors are reported to be associated with worse clinical outcomes. Peeters, Schenk, De Somer, et al. analysed the role of viral and host factors on disease presentation using clinical, biochemical, virological, chemokine and histological data from infected individuals in Belgium, with disease signs collected retrospectively over an 8-year timeframe. Subtype assignment was possible for 179/218 viraemic cases, confirming genotype 3 as dominant with an almost equal representation of clades abchijklm and efg. An increased hospitalisation rate and higher peak serum levels of alanine aminotransferase, bilirubin and alkaline phosphatase were found in individuals with clade efg infections on univariate and multivariable analyses. In addition, acute clade efg infections were characterised by higher serum CXCL10 levels and more pronounced liver necro-inflammatory activity. Long-term persistence of HCV resistance-associated substitutions after direct-acting antiviral treatment failure Data on the long-term persistence of HCV resistance-associated substitutions (RASs) after treatment with direct-acting antivirals are limited. Dietz et al. evaluated the persistence of NS3, NS5A and NS5B RASs for up to 5 years after the end of treatment (EOT). After protease inhibitor failure, the frequencies of NS3 RASs were 40-90% at EOT and RASs disappeared rapidly in GT1b and GT3 after follow-up month 3, while RASs were stable (≥60%) in GT1a due to Q80 K. The SOF-resistant NS5B RAS S282T was only found in GT3a patients. NS5A RASs were very common in all GTs after NS5A inhibitor failure (88-95%), and even after 24 weeks follow-up their frequency was ≥65%. However, while RASs in GT1b had a stable course, RASs in GT1a and GT3 declined slightly during follow-up. Effect of variants in LGP2 on MDA5-mediated activation of interferon response and suppression of HDV replication RIG-I-like receptors, including RIG-I, MDA5 and LGP2, sense viral RNA to induce the antiviral interferon response. LGP2, unable to activate the IFN response itself, modulates RIG-I and MDA5 signalling. HDV is sensed by MDA5. Gillich, Zhang, et al. show that LGP2 is essential for the MDA5-mediated IFN response induced upon HDV infection. This induction requires both RNA binding and ATPase activities of LGP2. The IFN response only moderately reduced HDV replication in resting cells but profoundly suppressed cell division-mediated HDV spread. An LGP2 variant (Q425R), predominating in Africans who develop less severe chronic hepatitis D, mediated detectably higher basal and faster HDV-induced IFN responses, as well as stronger HDV suppression. Thus, the natural Q425R LGP2 is a gain-of-function variant and may contribute to an attenuated course of hepatitis D. Reverse inflammaging: long-term effects of HCV cure on biological age Not all sequelae of chronic hepatitis C appear to be completely reversible after sustained virologic response (SVR). Oltmanns et al. aimed to investigate whether chronic HCV infection is associated with epigenetic changes and biological age acceleration and whether this is reversible after SVR. The authors show that individuals with HCV had an overall significant epigenetic age acceleration (EAA) of 3.12 years at baseline compared with 2.61 years in the age-and sex-matched reference group. HCV elimination resulted in a significant long-term increase in DNA methylation, EAA decreased to 1.37 years at long-term follow-up. Interestingly, eight individuals who developed hepatocellular carcinoma after SVR had the highest EAA and showed no evidence of reversal after SVR. NAFLD and alcohol-related liver diseases Neutrophil extracellular traps contribute to liver damage and increase defective low-density neutrophils in alcohol-related hepatitis In alcohol-related hepatitis (AH) neutrophil count increases and is associated with poor clinical outcomes; however, the mechanisms of neutrophil-mediated liver damage remain unclear. Cho et al. have shown that neutrophil depletion significantly ameliorates liver damage in a mouse model of AH. There are several neutrophil subpopulations and by studying serum samples from individuals with AH, the authors showed that high density neutrophils (HDNs) were activated and produce neutrophil extracellular traps (NETs), which contribute to liver damage. They also identified low density neutrophils (LDNs) that have opposing transcriptomic profiles and exhibit exhausted/defective phenotypes. Moreover, alcohol can induce NET release by HDNs, which then switch to an LDN phenotype. Thus, these results provide a mechanistic explanation for the paradox of previously observed neutrophil activation and dysfunction in AH. Since LDNs remain in the circulation and the liver for longer, they could be associated with the high susceptibility to infections observed in those with AH. Further, the authors tested two therapeutic approaches in mice: the first involved antibody-mediated neutrophil depletion, which resulted in the elimination of both NET-producing HDNs and defective LDNs. The other involved G-CSF administration, which prevented NET formation and reduced alcohol-related liver damage. These results are important as they provide new insights into neutrophil-mediated inflammation and liver damage in AH. Partial MCT1 invalidation protects against diet-induced NAFLD and the associated brain dysfunction Some studies have associated NAFLD with mild cerebral dysfunction and cognitive decline. Hadjihambi et al. used a high-fat diet-induced steatosis model and demonstrated anxiety and depression-mediated behaviour in the mice. They also showed low-grade brain tissue hypoxia, likely attributed to the low-grade brain inflammation, and decreased cerebral blood volume, accompanied by morphological and metabolic alterations (higher oxygen consumption) in the microglia and astrocytes, suggesting the early stages of an obesogenic diet-induced encephalopathy. The authors then used the same NAFLD model but in Mct1 haploinsufficient mice and showed that, despite fat accumulation in adipose tissue, these mice were protected from NAFLD and associated cerebral alterations. MCT-1 is a carrier of short-chain fatty acids, ketone bodies, and lactate in several tissues, playing an important role in energy homeostasis. It could therefore be a potential new therapeutic target. Cholestatic disease Secretin alleviates biliary and liver injury during late-stage PBC via restoration of secretory processes Secretin/secretin receptor (SCT/SR) signalling is a major regulator of biliary homeostasis as it modulates the “bicarbonate umbrella”, a protective mechanism whereby cholangiocytes release bicarbonate to reduce damage from toxic bile acids. Kennedy et al. used transgenic mice expressing a dominant-negative form of human transforming growth factor-beta receptor II, which can be used as a model of late-stage primary biliary cholangitis (PBC), as well as serum, bile and liver samples from individuals with stage III/IV PBC. They observed a reduction in SCT and SR expression, in particular in primary PBC cholangiocytes, suggesting a blunted biliary SCT/SR axis in late-stage PBC. The authors demonstrated that, in late-stage PBC, impairment of SCT/SR downregulates the “bicarbonate umbrella”, induces loss of mature cholangiocyte differentiation and mucin secretion and leads to enhanced ductular reaction, triggering liver fibrosis and inflammation. They then studied the effects of SCT administration. SCT reduced hepatic bile acid content, cholangitis activity, immune cell infiltration and collagen deposition. It restored the loss of ductulo-canalicular junction connections, and mucin secretion, and restored bile duct loss, supporting bile flow and bile duct system integrity. Through these choleretic effects and the restoration of the bicarbonate umbrella, secretin could be of therapeutic value for late-stage PBC. Cirrhosis and liver failure Evaluation of CirrhoCare® - a digital-health solution for home management of individuals with cirrhosis Kazankov et al. aimed to assess the feasibility and potential clinical benefits of remote management of individuals with acute decompensation (AD) of cirrhosis. Twenty individuals with cirrhosis and AD (mean age 59±10 years, 14 male, 80 % with alcohol-related aetiology, mean MELD-Na score 16.1±4.2) and 20 controls with no statistically different demographic and clinical variables were enrolled. Heart rate, blood pressure, weight, %body water, cognitive function, self-reported well-being and intake of food, fluid and alcohol were recorded daily with the use of commercially available monitoring devices linked to a smartphone (CirrhoCare® app). Independent external adjudicators assessed appropriateness of CirrhoCare®-based decisions. The length of the follow-up was 10.1±2.4 weeks. Fifteen patients showed good engagement (≥4 readings/week), 2 moderate (2-4/week), and 3 poor (<2/week). Five CirrhoCare®-managed patients had 8 readmissions, and none required hospitalization for >14 days. Sixteen other CirrhoCare®-guided patient contacts were made, leading to clinical interventions that prevented further progression. Appropriateness was confirmed by adjudicators. Controls had 13 readmissions in 8 patients, lasting a median of 7 days with 4 admissions of >14 days. They had 6 unplanned paracenteses compared to 1 in the CirrhoCare® group. Thus, the authors concluded that CirrhoCare® is feasible for community-management of individuals with decompensated cirrhosis and alerts clinicians to new decompensating events. Incidence and predictive factors of recurrent thrombosis in non-cirrhotic portal vein thrombosis Long-term anticoagulation is not recommended by CPGs in non-cirrhotic portal vein thrombosis (NC-PVT) without underlying thrombophilia due to a very low risk of recurrent thrombosis (RT). In this multicentre retrospective observational study, Anna Baiges, Procopet, Silva-Junior et al. aimed to describe the incidence of RT in patients with NC-PVT without an indication for long-term anticoagulation. The second aim was to identify RT risk factors and then verify them in a validation cohort. Risk factors for RT in 64 individuals with NC-PVT of idiopathic/local aetiology were evaluated. In a subgroup of 48, the potential value of additional thrombophilic parameters was analysed. Findings were validated in 70 independent individuals with idiopathic/local NC-PVT. Of the 64 individuals in the original cohort, 17 presented splanchnic and/or extra-splanchnic RT (overall-RT) during follow-up. Splanchnic RT was asymptomatic in 53% of the cases. No clinical or biochemical parameters predicted overall-RT. However, in the 48 individuals with additional comprehensive thrombophilic study, factor VIII ≥150% was the only independent factor predicting overall-RT. In the validation cohort, 19 individuals presented overall-RT, and the predictive value of factor VIII was confirmed. Thus, individuals with idiopathic/local NC-PVT appear to be at risk of overall-RT. Splanchnic RT can be asymptomatic and requires screening for its detection. Values of factor VIII ≥150% may help select patients at high risk of overall-RT who could benefit from long-term anticoagulation. Hepatic and biliary cancer Tumour response to predict overall survival in liver cancer trials: still an open question The possibility to use objective tumour response as a predictor of overall survival (OS) is important for the clinical development of new drugs for hepatocellular carcinoma (HCC). REFLECT is a phase III randomised clinical trial that compared lenvatinib with sorafenib. Kudo et al. analysed the potential association between OS and objective response assessed by investigators (INV) and independent radiologists (IR) using ordinary (RECIST 1.1) and modified RECIST (mRECIST) criteria. OS was significantly better in responders than in non-responders using IR-RECIST 1.1 (0.50), IR-mRECIST (HR 0.61), and INV-mRECIST (HR 0.61). Similarly, survival rates at 2, 4, and 6 months after randomization were also better for responders using the 3 criteria. An exploratory multivariate Cox regression analysis showed that objective response by IR-RECIST 1.1 (HR 0.49) and INV-mRECIST (HR 0.55) were independent predictors of OS. Additional studies are nevertheless needed to confirm surrogacy. Targeting YAP to treat cholangiocarcinoma YAP is a transcriptional co-activator and the effector protein of the Hippo signalling pathway, which regulates a number of cell functions linked to cell proliferation and prevention of apoptosis. Notably, YAP is aberrantly activated in many cancers, including cholangiocarcinoma (CCA). On the other hand, LCK is a protein tyrosine kinase that activates YAP and NTRC 0652 is a novel tyrosine kinase inhibitor with relative selectivity for LCK. Conboy et al. studied the effects of LCK inhibition in vitro and in vivo. NTRC 0652-0 resulted in apoptotic cell death in CCA cell lines and in a subset of patient-derived organoids. In particular, CCA with FGFR2 fusions were identified as a clinically relevant genetic subset. In PDX models of FGFR2 fusion-positive CCA, treatment with NTRC 0652-0 showed acceptable toxicity, decreased YAP tyrosine phosphorylation, and significant antitumor activity. These findings warrant clinical testing of this strategy in patients with CCA and FGFR2 fusions. Liver transplantation Low dose interleukin-2 selectively expands circulating Tregs but fails to promote liver allograft tolerance in humans CD4+CD25+Foxp3+ regulatory T cells (Tregs) are essential to maintain immunological tolerance and have been shown to promote liver allograft tolerance both in rodents and humans. Low dose interleukin-2 (LDIL-2) can expand human endogenous circulating Tregs in vivo, but its role in promoting Treg trafficking to the sites of inflammation is unknown. Lim et al. conducted a clinical trial in stable liver recipients 2-6 years post-transplant to determine the capacity of LDIL-2 to suppress allospecific immune responses and allow for the complete discontinuation of maintenance immunosuppression. One month after initiating LDIL-2, patients exhibiting at least a 2-fold increase in circulating Tregs gradually discontinued immunosuppression over a 4-month period while continuing LDIL-2. All patients achieved a marked and sustained increase in circulating Tregs, but this was not associated with the preferential expansion of donor-reactive Tregs nor the accumulation of intrahepatic Tregs. The trial was terminated after the first six participants failed to reach the primary endpoint due to rejection requiring immunosuppression reinstitution. The authors concluded that the expansion of circulating Tregs in response to LDIL-2 is not sufficient to control alloimmunity and to promote liver allograft tolerance. Patrizia Burra∗ at Department of Surgery, Oncology and Gastroenterology, Padua University Hospital, Padua, Italy. Frank Tacke at Department of Hepatology and Gastroenterology, Charité Universitätsmedizin Berlin, Berlin, Germany. Vlad Ratziu at Insitute for Cardiometabolism and Nutrition, Sorbonne Université and Hospital Pitié Salpêtrière, Paris, France. Stefan Zeuzem at Department of Medicine I, Goethe University Hospital, Frankfurt, Germany. Bruno Sangro at Liver Unit, Clinica Universidad Navarra and CIBEREHD, Pamplona, Spain. Paolo Angeli at Unit of Internal Medicine and Hepatology, University of Padua, Padua, Italy.

0

PMC9753409

NO-CC CODE

2022-12-16 23:25:42

no

J Hepatol. 2023 Jan 15; 78(1):1-4

utf-8

J Hepatol

10.1016/j.jhep.2022.10.003

oa_other

==== Front Health Policy Open Health Policy Open Health Policy Open 2590-2296 Published by Elsevier B.V. S2590-2296(22)00023-5 10.1016/j.hpopen.2022.100088 100088 Article The challenges brought by the COVID-19 pandemic to health systems exposed pre-existing gaps Rosenthal Anat a Waitzberg Ruth bc a Department of Health Policy and Management, School of Public Health, Faculty of Health Sciences, Ben-Gurion University of the Negev, Israel b Department of Health Care Management, Faculty of Economics & Management, Technische Universität Berlin, Germany c The Smokler Center for Health Policy Research, Myers-JDC-Brookdale Institute, Jerusalem, Israel 15 12 2022 15 12 2022 100088© 2022 Published by Elsevier B.V. 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcThe COVID-19 pandemic has brought new challenges to bear on health systems worldwide. No less important, it has exposed many of the existing failures of health systems and exacerbated their negative impact on the health of populations. For example, the pressures over health professionals in regions with inadequate staffing have increased even more, resulting often in burnout or poorer quality of care [1], [2]. Furthermore, the effects of the pandemic widened disparities in access to services and health outcomes across populations and widened the gap between the rich and the poor [3], [4], [5]. While these effects were measured in all health systems, their impact was much greater in health systems with limited resources and infrastructures, which had already been facing challenges in their functioning and performance [6]. As a result of challenges faced by many health systems in many low- and middle-income countries (LMICs), as well as the response of rich countries to the pandemic that tended to forgo global commitments in favor of local responses, the COVID-19 pandemic has also widened the gap between rich and poor countries in terms of resources, infrastructures, and workforce. Consequently, LMICs’ health systems have struggled even more to attain their goals such as of improving population’s health, being responsive to populations’ expectations, improving efficiency of service provision, while protecting the population from financial burdens[7]. Access, or lack thereof, to vaccines were a clear example of these global inequities [8], as well as global travels bans initiated by Western countries affecting the economies and freedom of movements of whole continents. The articles in this special issue of Health Policy Open address the COVID-19 pandemic from multiple angels, regions, and countries, exploring the impact of the pandemic on health systems and the resulting policies. In particular, this special issue showcases two types of challenges that health systems faced during the pandemic: unknown challenges that the pandemic exposed; and previously known challenges and gaps that were exacerbated by the pandemic. These challenges are well exposed by Hossaim et.al, [9] who highlight how the insufficient supply medical products such as oxygen and vaccine doses in low- and middle-income countries (LMICs) raise the likelihood of death of COVID-19 patients. Insufficient supply of medical products and workforce to provide care is not new in LMICs but was exacerbated during the pandemic. Haldane et.al[10] suggest that Uruguay’s early success relative to other Latin-American countries in responding to the pandemic was related to rapidly implementing a suite of economic and social measures, instead of strict border closures and restrictions on movement. They conclude that incentives can be more effective than prohibitions in supporting adherence to public health interventions by ensuring that effective social and economic safety net measures are in place to permit compliance with public health measures. Along similar lines, Velez et.al have also analyzed national COVID-19 pandemic response, focusing on preparedness planning documents from a sample of seven (of the eleven) countries in WHO [11]. While plans described the required resources during the COVID-19 pandemic, none presented a clear description of the priority setting (PS) process (e.g., a formal PS framework, and PS criteria). Most of the plans were incomplete and included only a limited number of quality indicators for effective PS, which highlights the need for further research on how countries operationalize PS. The case was different in the Eastern Mediterranean Region (EMR), where Razavi et.al found that national pandemic plans documented value of explicit priority setting in health system decision-making, but it may not be in the top of mind for decision- and policymakers when responding to health emergencies and public health crises[12]. Health system fragmentation is exacerbated during conflict and contributes to COVID-19 inequities experienced across the EMR. Limited prioritization of vulnerable groups like refugees and migrants in planning documents have long-term health implications and exacerbate the burden of COVID within these groups. Inequities and priority setting were also examined by Aiona et.al [13]who assessed the impact of the age-based COVID-19 vaccine prioritization by ethnicity in Denver, Colorado, and found that this prioritization decisions systematically disadvantaged communities of color irrespective of COVID-19 risk. In addition, in the first three phases of the vaccination rollout, 40% of hospitalizations and 16% of deaths occurred among those meeting age and long-term care facility criteria, and could have been averted. Through an online survey and thematic analysis of public documents and chats, Lotta et.al illustrate how the pandemic added to existing vulnerabilities (mental health and burnout) and created new problems and imbalances in the work of community health workers in Brazil[14]. They conclude that the pandemic not only deteriorated community-health workers’ working conditions, but also their relations with other health professionals (nurses and physicians), and of their ability to carry out their essential work in the public health system. Glass et al explore the possibilities enabled by cross-border agreements in the EU model and their applicability in the North American context. In the context of the COVID-19 pandemic, the authors explore the potential benefits and challenges of such agreements for both patients and healthcare systems and open the door for long-term cooperative policy planning[15]. The workforce shocks of the COVID-19 pandemic and the measures put in place to cope with them are used by Timmons and Morris to address the potential of licensing reforms[16]. An example of a “building back better” reflection, the authors review six alternatives considered to address shortages in primary care professionals – a shortage that existed long before the pandemic but was exacerbated by it. While much of the efforts were focused on health systems, health and welfare of populations were affected by other determinants as well. In their study of household food security in Burkina Faso Traoré et al addressed pattens of food insecurity during the pandemic and identified weak household (those headed by women, farm-depended households, and the poor) to be exposed to the risk of food shortages[17]. However, while these weaknesses predated the pandemic the authors also expose the risks faced by urban households that were more vulnerable to food shortages during the pandemic. This example of preexisting weaknesses compounded by new challenges, lead the way to forward policy planning that accounts for old and new shocks. Looking at the actors participating in the health policy arena before and during the pandemic, Meessen and Perazzi point to the importance of existing mechanisms of participation[18]. Examining the role of national hospital associations in health system governance, the authors highlight the importance of plurality of actors in health policy forums, and their contribution to policy making post-pandemic. Lastly, Faruk et al address the seminal role of Social Network Sites (STS) in the process of coping in responding to the COVID-19 pandemic[19]. In their study in Bangladesh, the authors examined STS use patters as information source during the pandemic and found that the absolute majority of their respondents (90%) have relied on STS for up-to-date news and pandemic information. And while social networks have played a significant role in as a source of information, they have also been seen as a source of panic and misconceptions, pointing to the important role played by STS and the needs of policy makers to address them as both a positive tool, and a negative force to be regulated. While the discourse surrounding the pandemic tended to focus on the direct impact of the pandemic, on health systems, the studies in this special issue highlight that what appeared to be “pandemic-related new challenges” were the very same deficiencies that predated the pandemic. No less important, and as the special issue papers illustrate, many of the challenges that were framed as caused by the pandemic, were in fact, challenges exacerbated by the pandemic. It would be misleading to address these shocks as new, surprising and unpredictable. Rather, they are an expression of long-term systemic failures. The papers in this special issue share a common message that while these shocks disrupted health systems, they bear potential lessons on how to mitigate the long-term failures. While the COVID-19 pandemic did not create many of these problems, it has highlighted and exacerbated many of their impacts, particularly inequities. Health inequities – in and between countries – have long been the grounds for the call to address the social determinants of health as well as the need to strengthen health systems functions and performance. And, while many health systems are still recovering from the damage done by the pandemic, both directly due to the high demands and burden on infrastructure and workforce, or indirectly due to limitations on activities during lockdowns, movement restrictions and redirection resources [20], health systems shocks can also serve as catalysts to long-term policies and planning. Mitigating long-term failures does not mean responding to emergencies or single shocks. Rather, planning for long-term and investing shortcomings previously known from routine times by training workforce, building infrastructure and medical products, eliminating inequities, building strong governance, improving access to high-quality care, to name a few. No less important, the pandemic highlighted the crucial role of reliable and timely data as they are the foundation of long-term planning and evidence-informed policies. Such policies will encourage “building back better” to create resilient health systems able to cope with challenges in both routine times and emergencies to avoid future shocks [21]. ==== Refs References 1 Winkelmann J. Webb E. Williams G.A. Hernández-Quevedo C. Maier C.B. Panteli D. European countries’ responses in ensuring sufficient physical infrastructure and workforce capacity during the first COVID-19 wave Health Policy (New York) 126 5 2022 362 372 2 WHO, “Mental Health and COVID-19: Early evidence of the pandemic’s impact: Scientific brief, 2 March 2022,” Geneve, Mar. 2022. Accessed: Jun. 06, 2022. [Online]. Available: https://www.who.int/publications/i/item/WHO-2019-nCoV-Sci_Brief-Mental_health-2022.1 3 M. Marmot, J. Allen, P. Goldblatt, E. Herd, and J. Morrison, “Build Back Fairer: The COVID-19 Marmot Review,” London, 2020. Accessed: Nov. 22, 2022. [Online]. Available: https://www.instituteofhealthequity.org/resources-reports/build-back-fairer-the-covid-19-marmot-review/build-back-fairer-the-covid-19-marmot-review-executive-summary.pdf 4 Nana-Sinkam P. Kraschnewski J. Sacco R. Chavez J. Fouad M. Gal T. Health disparities and equity in the era of COVID-19 J Clin Transl Sci 5 1 Mar. 2021 10.1017/cts.2021.23 5 Capasso A. Kim S. Ali S.H. Jones A.M. DiClemente R.J. Tozan Y. Employment conditions as barriers to the adoption of COVID-19 mitigation measures: how the COVID-19 pandemic may be deepening health disparities among low-income earners and essential workers in the United States BMC Public Health 22 1 Dec. 2022 1 13 10.1186/S12889-022-13259-W/FIGURES/3 34983455 6 WHO, “The impact of COVID-19 on global health goals,” webpage, May 20, 2021. https://www.who.int/news-room/spotlight/the-impact-of-covid-19-on-global-health-goals (accessed Nov. 27, 2022). 7 World Health Organization, “EVERY BODY’S BUSINESS: STRENGTHENING HEALTH SYSTEMS TO IMPROVE HEALTH OUTCOMES WHO’S FRAMEWORK FOR ACTION,” Geneva, 2007. [Online]. Available: https://www.who.int/healthsystems/strategy/everybodys_business.pdf 8 Kluge H. McKee M. COVID-19 vaccines for the European region: an unprecedented challenge The Lancet 397 10286 2021 1689 1691 9 Moyazzem Hossain M.d. Abdulla F. Rahman A. Challenges and difficulties faced in low- and middle-income countries during COVID-19 Health Policy OPEN 3 2022 100082 36405972 10 Haldane V. Morales-Vazquez M. Jamieson M. Veillard J. Marchildon G.P. Allin S. Learning from the first wave of the COVID-19 pandemic: Comparing policy responses in Uruguay with 10 other Latin American and Caribbean countries Health Policy Open 3 Dec. 2022 100081 10.1016/J.HPOPEN.2022.100081 11 Vélez C.-M. Kapiriri L. Nouvet E. Goold S. Aguilera B. Williams I. Examining priority setting in the National COVID-19 pandemic plans: a case study from countries in the WHO- South-East Asia Region (WHO-SEARO) Health Policy Open 3 2022 100086 36447637 12 Razavi S.D. Noorulhuda M. Marcela Velez C. Kapiriri L. Dreyse B.A. Danis M. Priority setting for pandemic preparedness and response: A comparative analysis of COVID-19 pandemic plans in 12 countries in the Eastern Mediterranean Region Health Policy Open 3 2022 100084 36415539 13 Aiona K. Bacon E. Podewils L.J. Haas M.K. The disparate impact of age-based COVID-19 vaccine prioritization by race/ethnicity in Denver, Colorado Health Policy Open 3 Dec. 2022 100074 10.1016/J.HPOPEN.2022.100074 14 Lotta G. Nunes J. Fernandez M. Garcia Correa M. The impact of the COVID-19 pandemic in the frontline health workforce: Perceptions of vulnerability of Brazil’s community health workers Health Policy OPEN 3 2022 100065 35036911 15 Glass L.T. Schlachta C.M. Hawel J.D. Elnahas A.I. Alkhamesi N.A. Cross-border healthcare: A review and applicability to North America during COVID-19 Health Policy Open 3 Dec. 2022 100064 10.1016/J.HPOPEN.2021.100064 16 Timmons E. Norris C. Potential licensing reforms in light of COVID-19 Health Policy Open 3 Dec. 2022 100062 10.1016/J.HPOPEN.2021.100062 17 Traoré O. Combary O.S. Zina Y.d.D. Households’ basic needs satisfaction during the Coronavirus disease 19 (COVID-19) pandemic in Burkina Faso Health Policy OPEN 3 2022 100060 34877532 18 Meessen B. Perazzi S. The role of national hospital associations in health system governance before and during the COVID-19 pandemic: Findings from an exploratory online survey Health Policy Open 3 Dec. 2022 100077 10.1016/J.HPOPEN.2022.100077 19 Faruk M.O. Devnath P. Kar S. Eshaa E.A. Naziat H. Perception and determinants of Social Networking Sites (SNS) on spreading awareness and panic during the COVID-19 pandemic in Bangladesh Health Policy Open 3 Dec. 2022 100075 10.1016/J.HPOPEN.2022.100075 20 A. B. Hogan et al., “Potential impact of the COVID-19 pandemic on HIV, tuberculosis, and malaria in low-income and middle-income countries: a modelling study,” Lancet Glob Health, vol. 8, no. 9, pp. e1132–e1141, Sep. 2020, doi: 10.1016/S2214-109X(20)30288-6. 21 A. Sagan, E. Webb, N. Azzopardi-Muscat, I. de la Mata, M. McKee, and J. Figueras, Health systems resilience during COVID-19 : lessons for building back better. Brussels: European Observatory on Health Systems and Policies, 2021. Accessed: Oct. 18, 2022. [Online]. Available: https://apps.who.int/iris/rest/bitstreams/1390564/retrieve

0

PMC9753444

NO-CC CODE

2022-12-16 23:25:43

no

Health Policy Open. 2022 Dec 15;:100088

utf-8

Health Policy Open

10.1016/j.hpopen.2022.100088

oa_other

==== Front Curr Res Ecol Soc Psychol Curr Res Ecol Soc Psychol Current Research in Ecological and Social Psychology 2666-6227 The Author(s). Published by Elsevier B.V. S2666-6227(22)00049-1 10.1016/j.cresp.2022.100082 100082 Article COVID-19 vaccine intentions in Aotearoa New Zealand: Behaviour, risk perceptions, and collective versus individual motivations Vinnell Lauren J. 1⁎ Becker Julia S. 1 Hudson-Doyle Emma E. 1 Gray Lesley 12 1 Joint Centre for Disaster Research, Massey University, Wellington, New Zealand 2 Department of Primary Health Care and General Practice, University of Otago, Wellington, New Zealand ⁎ Corresponding Author: Dr. Lauren Jennifer Vinnell, Joint Centre for Disaster Research, Massey University, Wellington Campus, 94 Tasman Street, Mount Cook, Wellington 6021, New Zealand 15 12 2022 15 12 2022 100082© 2022 The Author(s) 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The global SARS-CoV-2 (COVID-19) pandemic presents a pressing health challenge for all countries, including Aotearoa New Zealand (NZ). As of early 2022, NZ public health measures have reduced impacts of the pandemic, but ongoing efforts to limit illness and fatalities will be significantly aided by widescale uptake of available vaccines including COVID-19 booster doses. Decades of research have established a broad range of demographic, social, cognitive, and behavioural factors which influence peoples’ uptake of vaccinations, including a large amount of research in the last two years focused on COVID-19 vaccination in particular. In this study, we surveyed people in New Zealand (N = 660) in May and June of 2021, at which point the vaccine had been made available to high-risk groups. We explored individual versus collective motivations, finding that people who were hesitant about COVID-19 vaccination scored lower on independent self-construals (how people define themselves) but higher on community identity, weaker but still positive perceived social norms, lower general risk of COVID-19 to New Zealanders and higher vaccine risk for both themselves and others, and lower response-efficacy both for personal and collective benefits. Overall, the findings suggest some benefit of collective over individual appeals, but that generally messaging to encourage vaccination should focus on conveying social norms, risk from COVID-19 broadly, and vaccine safety and efficacy. Keywords Vaccine intentions self-construals social norms collective efficacy COVID-19 ==== Body pmcOverview of Global COVID-19 Situation On the 5th of January, 2020, the World Health Organization (WHO) first released information on a new “viral pneumonia” observed in Wuhan, People's Republic of China (WHO, 2021). The first vaccines against COVID-19 were administered in December, 2020. At the time of writing in early 2022, confirmed global cases were over 410 million and the death toll was over 5.8 million, despite over 10 billion vaccine doses having been administered. Case rates remain high in many areas, partly reflecting a concentration of vaccines in developed countries, the highly infectious Omicron strain, and fatigue and frustration with public health measures leading to these either being dropped or ignored. Several reports, typically examining excess mortality rates, suggest that both cases and deaths are vastly underreported (Whittaker et al., 2021). COVID-19 in Aotearoa New Zealand Aotearoa New Zealand (NZ) reported its first confirmed case of COVID-19 on the 28th of February 2020 (Radio New Zealand, 2021). The country closed its borders to all but citizens and permanent residents on the 19th of March (who had to go through managed isolation and quarantine from April) and entered a strict “Level 4” lockdown at 11.59pm on the 25th. During this lockdown people were only allowed to leave home if they were an essential worker, going to a supermarket or to seek healthcare, or staying close by to exercise outside. NZ reported its first death from COVID-19 on the 29th of March. There were three considerable outbreaks, in August 2020, February 2021, and an outbreak of the Delta variant which began in August 2021. The arrival of the Omicron variant ultimately led to the loosening of virtually all restrictions. In October 2020 the government announced its first vaccine pre-purchase agreement. The Pfizer BioNTech vaccine was approved for use in NZ on the 10th of February and the first doses administered on the 20th (see Table 1 for an overview). The government implemented a framework targeting first those most at risk of getting infected (i.e., border workers then their households), followed by those most likely to experience worse outcomes if they were infected (e.g., older people, those with pre-existing health issues, and communities which typically experience poorer health outcomes). The Janssen, AstraZeneca, and Novovax vaccines have since been given provisional approval with the NZ government having made advanced purchase agreements, but these vaccines were not widely offered in NZ.Table 1 Timeline of key dates in NZ's vaccine rollout Table 1Milestone Date First international doses December, 2020 Pfizer approved for used in New Zealand 10th Feb, 2021 Vaccine available to border and health workers 20th Feb, 2021 Vaccine available to high-risk groups May, 2021 Data collection for this study 16th May to 6th June, 2021 Vaccine available to general population (ages 60-64)* 28th July, 2021 Vaccine available for those aged 12 to 15 20th August Vaccine available for all over 12 1st September, 2021 Rollout of boosters to border & health workers 29th November, 2021 Boosters to general population 17th Jan, 2022 Vaccine available for those aged 5 to 11 17th Jan, 2022 Note. All dates are for the Pfizer vaccine. * = general population proceeded in stages based on age brackets. At the time of writing, face masks were mandated in many settings including flights and retail and it was a legal requirement for all businesses to display an official QR code for the country's smartphone tracing app. Ongoing government messaging encourages good hygiene and use of the tracing app. Likely given the low number of cases in the country during the initial vaccine rollout, as well as the apparent success of appeals to act to protect each other during the Level 4 lockdown, the early vaccine promotion material focused primarily on allowing people to either return to more “normal” behaviour or to continue such behaviour as well as the communal benefit of vaccinating. However, print news media presented more arguments for vaccines based on preventing or protecting against disease than on protecting the community (Ashwell & Murray, 2020). Since the time of data collection, vaccine mandates have been introduced (and in 2022, largely removed) for a number of workforces including education, and the Covid Protection Framework - which replaced the Alert Levels system, both now no longer in use - had more lenient restrictions for venues which require patrons to provide their vaccine pass (a scannable QR code available for those who have been fully vaccinated to indicate their status). In this study, we aimed to identify key factors related to intentions to get, or intentions to not get, the COVID-19 vaccine. Many studies have explored a range of different factors; we focused on collective versus individual motivations, including specific views of individual and collective risks as well as general, individual difference factors such as independent and interdependent self-construals (how people define themselves). We briefly review literature on vaccine intentions, moving from less to more contextually specific, followed by a brief introduction of key study variables not already introduced in the reviewed vaccine literature. As part of this latter section, we provide an overview of NZ's cultural context to help appropriate interpretations of the study findings. Vaccine Intentions and Behaviour Research has examined intentions and actual vaccination behaviour covering a broad variety of motivators and inhibitors for a range of viruses, including influenza A (H1N1), commonly referred to as swine flu. While much earlier work focused on confidence in vaccines and health systems, more recent studies have looked at a larger range of factors including perceptions of disease risk, psychological barriers, and collective responsibility (Betsch et al., 2018; Gray et al., 2012). Research in Canada looked at H1N1 vaccine uptake among pregnant women (Fabry et al., 2011). Although the majority (95%) recognized that the vaccine was recommended, only three-quarters had received it. Women in the study were more likely to have received the vaccine if they trusted health professional advice, believed that the vaccine had been adequately tested, and believed that it would be effective. Information source also related to vaccine uptake, with higher rates among those who used official websites and lower rates among those who used other sites such as mainstream media. Other factors which have been shown to be relevant for uptake or intentions to receive an H1N1 vaccine include past behaviour, instrumental attitudes, subjective norms, perceived benefits, and worry about catching swine flu (Gray et al., 2012; Myers & Goodwin, 2012; Teitler-Regev et al., 2011; Yang, 2015). In mid-2020, reported willingness to be vaccinated against COVID-19 ranged between 63% and 88% in a multinational survey (Kerr et al., 2021). Several studies have looked at intentions to receive vaccines for COVID-19 specifically. These studies examine demographic factors such as age and gender (Faasse & Newby, 2020; Karlsson et al., 2021; Latkin et al., 2021; Schwarzinger et al., 2021), behavioural factors such as past vaccine uptake and use of other protective behaviours (Faasse & Newby, 2020; Latkin et al., 2020; Schwarzinger et al., 2021), informational factors including information sources (Faasse & Newby, 2020) and trust in scientists or experts (Faasse & Newby, 2020; Freeman et al., 2021; Kerr et al., 2021), cognitive factors such as outcome expectancy (the belief that the behaviour will lead to the intended or suggested outcome; Anthony et al., 2021; Faasse & Newby, 2020; Freeman et al., 2021); and risk factors such as concern about infection, illness, transmitting to others, and vaccine safety (Anthony et al., 2021; Faasse & Newby, 2020; Karlsson et al., 2021; Kerr et al., 2021; Kwok et al., 2021; Motta et al., 2021; Schwarzinger et al., 2021). Some of this research found significant influences of factors relating to collective motivations (Freeman et al., 2021; Karlsson et al., 2021; Kwok et al., 2021); a recent study explicitly explored the role of community identification, finding that those who more strongly identified with their community had higher willingness to get the vaccine, via a stronger perceived sense of duty to their community (Wakefield & Khauser, 2021). Researchers in the area have recommended public communication focused on altruism (Chou & Budenz, 2020) while others have theorised that collective benefits, despite the small contribution that an individual vaccination makes, lead to a moral obligation and a sense of fairness (Giublini et al., 2018); however, one study demonstrated that perceptions of fairness (and harm) were not related to vaccine hesitancy (Amin et al., 2017). Some research has been conducted in NZ exploring factors relating to COVID-19 vaccine intentions. In a survey conducted in July 2020, Thaker (2021) found that general hesitancy around vaccines, including lack of confidence and perception of risks, and trust in scientists were related with COVID-19 vaccine intentions. Three-quarters of participants (74%) said that they intended to get a vaccine once it was available, consistent with international levels (between 63% and 88%; Kerr et al., 2021). The studies presented above present a broad range of factors which relate to intentions or uptake of vaccines, including for COVID-19 specifically. While some studies include collective motivations, few focus on these types of motivations and mostly do not directly compare them to individual motivations. Vaccination offers a good opportunity to test the extent to which people are motivated more by individual or collective reasons; vaccines offer some protection to the individual but are more effective when many people are vaccinated. There are also risks at the individual level (e.g., side effects) and at the collective level (e.g., outbreaks impacting health systems and the economy). To explore individual versus collective motivations, we drew on a number of other factors from behavioural sciences not widely considered in the above vaccine literature. Other Study Factors The following factors were also included in this study as the represent a range of collective factors which have been known to motivate diverse behaviour such as preparing for earthquakes, but with no or limited application to the behaviour of vaccination. These factors therefore are well tested and supported measures but represent relatively novel considerations in the domain of vaccine intentions. There are undoubtedly other factors which fit these criteria; however, in order to avoid making the survey excessively long (and thereby risking a high non-completion rate or lower quality responding) we had to limit the number of factors included. These factors are therefore intended to suggest whether overall there is a collective versus individual trend and ideally suggest a limited number of key factors to consider but is by no means intended to be exhaustive. Inclusion of Community in Self Adapted from the Inclusion of Other in the Self scale (Aron et al., 1992), the Inclusion of Community in Self scale (ICS) was developed as a way to measure community connectedness with a single item (Mashek et al., 2007). Participants are presented with a series of pairs of circles with differing amounts of overlap; one of the circles represents themselves and the other represents their community. They select the pair of circles which best reflects to what extent their community is part of their sense of self. As with previous versions of the scale (e.g., Aron et al., 1992; Tropp & Wright, 2001), the ICS has been established as a valid and reliable measure (Mashek et al., 2007). Given the benefits to the community of individuals being vaccinated, and the concordant focus on this benefit in NZ messaging, it is important to explore how the extent to which people feel connected to their community might affect their vaccine intentions. Previous research in other domains has found that people are more likely to engage in behaviours which will help their community, such as cleaning waterways, if they identify more strongly with their community (Forsyth et al., 2015). However, authors have suggested that the mechanism for this relationship is the motivation to act in line with common behaviours within the community (Solberg et al., 2010). It is therefore also important to consider the influence of social norms. Social norms People's behaviour is influenced by the common behaviour and beliefs of those in their social groups. Norms can be “actual” (i.e., objective strength) or “perceived” (i.e., what an individual thinks the norms are in their social group). Because perceptions of norms are not always accurate (Rimal & Lapinski, 2015), we are interested in this latter type of norms in this study. Within these two broad categories, there are two commonly identified types of norms: descriptive and injunctive. Perceived descriptive norms are how commonly an individual thinks a behaviour is carried out by members of their social group; if someone thinks that many people like them engage in a behaviour, then they are more likely to also engage in that behaviour (Cialdini et al., 1990). Perceived injunctive norms are beliefs about how strongly and widely approved (or disapproved) a behaviour is; similar to descriptive norms, people are more likely to engage in a behaviour which they think is largely approved of by their social group, and to not engage in a behaviour which is not approved (Cialdini, 2007). It is important to consider these two discrete factors separately as they influence behaviour in different ways (e.g., Hamann et al., 2015; Vinnell et al., 2018). Independent and interdependent self-construals Self-construals refer to how people define themselves (Markus & Kitayama, 1991). Those with stronger independent self-construals define themselves as discrete individuals rather than any connections to others being a core part of their sense of self. Those with stronger interdependent self-construals define themselves more by the groups of which they are members. These construals impact people's behaviour, as those who are more independent rely more on their own beliefs and emotions to make decisions than the beliefs and emotions of others (Markus & Kitayama, 1998). They are then more motivated to act in ways which allow them to be direct, achieve their goals, and express uniqueness. In comparison, those who are more interdependent are more likely to behave in ways based on their perceptions of the beliefs, behaviour, and emotions of those in the groups (Markus & Kitayama, 1998), and to act in ways which benefit their groups and help them to fit in. People have both self-construals but to differing extents and differing levels of use in behavioural decision making, depending on factors including cultural and situational context (Kanagawa et al., 2001; Markus & Kitayama, 1991). These concepts, and scales to measure them, have been developed and critiqued over the decades since their initial introduction (e.g., Gudykunst & Lee, 2003), but they have been shown to influence a large range of behaviours from willingness to pay for ecotourism (Hwang & Lee, 2018) to colour preferences (Jeon et al., 2020). Collective efficacy Collective efficacy refers to the ability of a group of people to work together to achieve a shared goal which benefits all group members (Carbone & McMillin, 2019). This particular factor is relevant in the context of vaccines, as although vaccination offers protection to the individual, it also serves to directly protect those around them (such as those who are particularly vulnerable and/or cannot be vaccinated) by limiting the chances of them spreading disease as well as indirectly by reducing possible spread more widely through communities. Collective efficacy influences a broad range of behaviours, including natural hazard preparation (Becker et al., 2015) and pro-environmental behaviour (Pakmehr et al., 2020). Aotearoa New Zealand's Cultural Context NZ is broadly an individualistic country (Brougham & Haar, 2013) meaning that in theory most people should be more motivated by benefits and risk to themselves than to their family and communities. Approximately 17% of NZ's population are Māori, who tend to be relatively collectivistic (Bennett & Liu, 2018; Brougham & Haar, 2013). However, the ethnic diversity within the country means that it is unlikely that all New Zealanders are more individualistic than collectivistic. Official government communication during the initial stages of the pandemic (in particular the first Level 4 lockdown in mid-2020) made several appeals to New Zealanders’ collective identity. These included repeated use of the phrase “team of five million” to describe the population and the reo Māori phrase “he waka eke noa” which translates approximately to “we are all in the same boat” in reference to the idea that the pandemic is a collective, shared experience. These appeals to collective motivations appear to have been successful based on compliance with restrictions during this initial lockdown (although some factors limited ability to follow recommendations such as physical distancing; Gray et al., 2021). Study Aims We did not formulate hypotheses for this study, as the evidence to suggest one more prominent set of factors (i.e., whether individual or collective motivations would be more prominent) was mixed. The current (at time of data collection) campaign from the government suggests a perspective that people will be more motivated by a collective purpose and through a desire to see life either return to “normal” or to continue as normal, rather than from an individual purpose or from a perception of risk from COVID-19. This latter assumption is likely based on the fact that, currently, the largest impacts from the pandemic felt in NZ are restrictions on movement through the border. Method Participants People 16 years or older and currently living in NZ were recruited primarily through social media channels (namely Twitter and Facebook). This did not allow for participant recruitment to target particular demographic quota, but is a useful mechanism to obtain relatively large sample sizes for limited time and financial investment. Seven hundred and twelve people started the survey. One person did not consent, and one indicated that they were below the minimum age (therefore neither progressed to the first question). A further 50 participants who did not answer any questions beyond these screening questions were removed, leaving a final dataset of 660 individuals. Approximately one third of participants identified as men (32.5%) and two-thirds as women (67.4%). Ages ranged from 18 to 83, with a mean of 39 (SD = 13.93). The sample was well-educated, with 68.5% of participants holding a university degree. Similarly, participants were fairly well-off, with only 14.2% reporting that their household costs slightly or considerably exceeded their income. Approximately half (55%) were responsible for dependants. Most participants reported between two and four people in their household (71.3%). A quarter (25.4%) reported that they are considered high risk for negative COVID-19 outcomes according to the Ministry of Health, while 27.6% said that someone in their household is high risk. Only 6.3% of the participants have had COVID-19 while 28.5% of the participants who have family overseas said that at least one of those family members have had COVID-19; 50.8% of these participants indicated that they had overseas family members who had been vaccinated. Data was collected between the 16th of May and 6th of June, 2021, around the time that the vaccine rollout was extended beyond border workers and healthcare to people who are considered at higher risk from COVID-19. At this stage of the survey, the prior vaccine group had been able to be vaccinated approximately four months. Materials Table 2 presents all key survey variables which were answered using 7-point Likert or Likert-type scales.Table 2 Key survey questions Table 2Scale Item "1" label "7" label Adapted from Interdependent self-construal I maintain harmony in the groups of which I am a member Strongly disagree Strongly agree Gudykunst & Lee (2003) I respect the majority's wishes in groups of which I am a member Strongly disagree Strongly agree Gudykunst & Lee (2003) I respect decisions made by my group Strongly disagree Strongly agree Gudykunst & Lee (2003) It is important to consult close friends and get their ideas before making a decision Strongly disagree Strongly agree Gudykunst & Lee (2003) I will sacrifice my self-interest for the benefit of my group Strongly disagree Strongly agree Gudykunst & Lee (2003) I stick with my group even through difficulties Strongly disagree Strongly agree Gudykunst & Lee (2003) Independent self-construal I prefer to be self-reliant rather than depend on others Strongly disagree Strongly agree Gudykunst & Lee (2003) I should decide my future on my own Strongly disagree Strongly agree Gudykunst & Lee (2003) I take responsibility for my own actions Strongly disagree Strongly agree Gudykunst & Lee (2003) It is important for me to act as an independent person Strongly disagree Strongly agree Gudykunst & Lee (2003) My personal identity is important to me Strongly disagree Strongly agree Gudykunst & Lee (2003) I enjoy being unique and different from others Strongly disagree Strongly agree Gudykunst & Lee (2003) COVID-19 collective beliefs We are a team of 5 million Strongly disagree Strongly agree Original When it comes to COVID-19, we are all in the same boat (he waka eke noa) Strongly disagree Strongly agree Original It took everyone in New Zealand to stop COVID-19 from spreading Strongly disagree Strongly agree Original It will take everyone in New Zealand getting vaccinated to prevent COVID-19 from spreading Strongly disagree Strongly agree Original It is my responsibility to get vaccinated to help keep myself safe Strongly disagree Strongly agree Original It is my responsibility to get vaccinated to help those around me safe Strongly disagree Strongly agree Original Anxiety How anxious are you about falling ill with COVID-19? Not at all A lot Karlsson et al. (2021) How anxious are you about passing COVID-19 on to someone else? Not at all A lot Karlsson et al. (2021) Intentions Do you intend to get a COVID-19 vaccines once it is available to you? Definitely not Definitely yes Original COVID-19 risk perception Please indicate how likely you think each of the following situations are Another large-scale COVID-19 outbreak in New Zealand Extremely unlikely Extremely likely Fassse & Newby (2020) You personally contracting COVID-19 (assuming no vaccination) Extremely unlikely Extremely likely Fassse & Newby (2020) Someone you care about in New Zealand contracting COVID-19 (assuming no vaccination) Extremely unlikely Extremely likely Fassse & Newby (2020) If you were to contract COVID-19, how serious do you think your symptoms would be? Not severe at all Very severe Fassse & Newby (2020) Think of someone you care about. If they were to contract COVID-19, how serious do you think their symptoms would be? Not severe at all Very severe Fassse & Newby (2020) In general, how much of a risk do you think COVID-19 poses to New Zealanders? No risk at all A great deal of risk Original COVID-19 Vaccine risk How likely do you think it is that you personally would experience negative side effects from a COVID-19 vaccine? Extremely unlikely Extremely likely Faasse & Newby (2020) How likely do you think it is that someone you care about would experience negative side effects from a COVID-19 vaccine? Extremely unlikely Extremely likely Faasse & Newby (2020) If you were to experience negative side effects of a COVID-19 vaccine, how serious do you think your symptoms would be? Not severe at all Very severe Faasse & Newby (2020) Think of someone you care about. If they were to experience negative side effects from a COVID-19 vaccine, how serious do you think their symptoms would be? Not severe at all Very severe Faasse & Newby (2020) In general, how much of a risk do you think COVID-19 vaccines pose to New Zealanders? No risk at all A great deal of risk Original Response-efficacy Getting a COVID-19 vaccine will keep me safe Strongly disagree Strongly agree Original Getting a COVID-19 vaccine will keep those around me safe Strongly disagree Strongly agree Original Getting a COVID-19 vaccine will help NZ to return to normal faster Strongly disagree Strongly agree Original Getting a COVID-19 vaccine will help NZ open up the borders sooner Strongly disagree Strongly agree Original Collective efficacy People in my community can be trusted Strongly disagree Strongly agree Carbone & McMillin (2019) People in my community get along with each other Strongly disagree Strongly agree Carbone & McMillin (2019) I live in a close-knit community Strongly disagree Strongly agree Carbone & McMillin (2019) People in my community are willing to help each other Strongly disagree Strongly agree Carbone & McMillin (2019) People in my community share the same values Strongly disagree Strongly agree Carbone & McMillin (2019) Perceived descriptive norms Please indicate how likely you think it is that each of these groups will get a COVID-19 vaccine once it is available to them My friends and family Extremely unlikely Extremely likely Vinnell et al. (2021) People in my community Extremely unlikely Extremely likely Vinnell et al. (2021) People like me Extremely unlikely Extremely likely Vinnell et al. (2021) New Zealanders Extremely unlikely Extremely likely Vinnell et al. (2021) Perceived injunctive norms Please indicate how strongly you think each of these groups approve of getting a COVID-19 vaccine once it is available to them My friends and family Strongly disapprove Strongly approve Vinnell et al. (2021) People in my community Strongly disapprove Strongly approve Vinnell et al. (2021) People like me Strongly disapprove Strongly approve Vinnell et al. (2021) New Zealanders Strongly disapprove Strongly approve Vinnell et al. (2021) Vaccine support Do you support the use of COVID-19 vaccines? Strongly oppose Strongly support Original Do you support the use of vaccines generally (i.e., for illnesses)? Strongly oppose Strongly support Original Note. The intentions question also included two further options: “Can't have the vaccine” and “Already had the vaccine”. Inclusion of Community in Self Scale This single-item scale assesses the amount of overlap participants see between themselves and their community, by selecting one of seven pairs of circles with varying degrees of overlap (see Figure 1 ; Mashek et al., 2007).Figure 1 Inclusion of Community in Self Scale Figure 1 Other related behaviours Participants were asked how often (Never, Sometimes, About half the time, Most of the time, Always) they took steps to protect themselves from COVID-19 (relative to the number of times they're in situations where the steps are recommended). These actions were:- “Wearing a face mask or covering on public transport” - “Keeping a distance of 2 metres from people you don't know (when at Alert Level 2 or higher)” - “Logging places you visit with the COVID-19 Tracer App or other method” - “Washing your hands (or sanitising)” - “Coughing and sneezing into your elbow” - “Staying home if you are sick” Participants were also asked:- “How often do you talk about COVID-19 vaccination with people around you” (Never, Less than once a month, A couple of times a month, Once a week, A couple of times a week, Every day) - “Where do you get most of your COVID-19 information?” (Government/health officials, Friends, Family, News media, Social media, Other) - “How often do you get the yearly flu vaccine?” (Never, Some years, Most years, Every year) Demographics Finally, participants were asked for a range of demographic information. Beyond the typical questions (age, gender, ethnicity, education, income, responsibility for dependants, and number of people in the household), we asked several questions about personal and family COVID-19 experience:- if they are considered at higher risk from COVID-19 - if someone in their household is considered at higher risk from COVID-19 - if they have had COVID-19 - if they have family overseas who have had COVID-19 - if they have family overseas who have had a COVID-19 vaccine Results Although intentions were measured with a 7-point Likert-type scale, few participants scored below the mid-point so the variable1 was not treated as continuous. Instead, participants were grouped into four categories to allow for group means comparisons:- Don't intend to get the vaccine (score of 1 to 4 on the intentions scale; i.e., either do not intend to get the vaccine or intend to not get the vaccine), n = 39 - Might get the vaccine (score of 5 or 6 on the intentions scale), n =113 - Will get the vaccine (score of 7 on the intentions scale), n = 319 - Have already received at least one dose, n = 112 Implications of this decision are discussed in the Limitations section. Where possible, Bonferroni corrections were applied to reduce the familywise error rate and a more conservative alpha threshold for significance of .01 was adopted to reduce whole study error rate. These participants are referred to by their groups throughout the results. In all instances, we are referring to statistical groups for the purpose of conducting, describing, and interpreting analyses. We cannot say based on our data that the participants in each group are unique from each other, rather than simply varying in degree of their intentions. Demographic Differences Age Age significantly differed between intentions groups, F(3, 568) = 14.17, p < .001, η2 = .07. Those in the “might” group were significantly younger (M = 32.31, SD = 9.19) than those in the “don't” group (M = 45.03, SD = 11.38; t(56.35) = 6.30, p < .001, d = 1.30), the “will” group (M = 40.98, SD = 14.12; t(297.87) = 7.34, p < .001, d = 0.67), and the “have” group (M = 40.42, SD = 15.72; t(173.44) = 4.67, p < .001, d = 0.63). There were no other differences between groups on age. Number of people in the household The groups also significantly differed on number of people in the household, F(3, 562) = 7.16, p < .001, η2 = .04. Those in the “might” group had significantly more people in their household on average (M = 3.87, SD = 1.53) than those in the “will” group (M = 3.09, SD = 1.45; t(416) = 4.78, p < .001, d = 0.53) and the “have” group (M = 3.24, SD = 1.63; t(216) = 2.95, p < .01, d = 0.40). Dependents The groups also significantly differed on the proportion of participants who have dependents, Χ2(3) = 13.03, p = .005, Cramer's V = .15. Within the “don't”, “might”, and “have” groups, the majority of participants have dependents (55.3%, 68.5%, and 57.9% respectively), compared to a slight minority in the “will” group (48.9%). High risk There was a marginally significant difference in the ratio of people who are and are not at a higher risk from COVID-19 between the intentions groups, Χ2(3) = 10.46, p = .015, Cramer's V = .14. A larger proportion of people in the “don't”, “might”, and “will” groups were not at higher risk, while a larger proportion of the people in the “have” group were at higher risk (28.1% versus 16.3%). This difference likely reflects the prioritization of higher risk individuals for first access to the vaccine at the point at which this data was collected. Family vaccinated There was a significant difference in the ratio of participants who had COVID-19 vaccinated family members between the intentions groups, Χ2(3) = 11.94, p < .01, Cramer's V = .20. The proportion of people who had vaccinated family members overseas was larger in the “will” group (66.0% versus 55.2%) and the “have” group (19.0% versus 13.3%). However, some cells in this analysis were around or smaller than n = 20 so these results should be interpreted cautiously. Gender There was a marginally significant difference in the gender ratio between the intentions groups, Χ2(3) = 9.31, p = .026, Cramer's V = .13. A larger proportion of the participants who identified as women were in the “will” group (59.2%) compared to those who identified as men who were more spread between the “might” (24.6%), “will” (46.2%), and “have” (22.2%) groups. The proportion who were in the “don't” group was similar between men (7.0%) and women (6.6%). People of other genders did not make up enough of the sample to be included in the analysis. Other factors There were no significant differences between intentions groups in having higher-risk household members or family who have had COVID-19. Some cell sizes were too small for overall interpretation for income, education, and participants who had had COVID-19. General Individual/Collective Identity Means significantly differed between intentions categories for all four variables: Inclusion of Community in Self, F(3, 547) = 7.14, p < .001, ηp 2 = .04; Interdependent self-construal, F(3, 547) = 8.21, p < .001, ηp 2 = .04; Independent self-construal, F(3, 547) = 6.06, p < .001, ηp 2 = .03; and Collective efficacy, F(3, 547) = 11.11, p < .001, ηp 2 = .06. Inclusion of Community in Self Those who might receive the vaccine scored marginally higher on the ICS than those do not intend to, p = .038, and those who strongly intend to, p < .001, suggesting a fairly unique trait of the vaccine hesitant identifying more strongly as a member of their community, particularly compared to those who strongly intend to get the vaccine. It is important to note here that “community” was self-defined, so participants might have been primed by the study topic to think of a community to which they belong which has relevance to the vaccine (i.e., a group where vaccine opinions and behaviours are shared between members). It is possible then that those in the “might” group tend to more often be part of communities which share their own hesitancy, potentially explaining why a stronger community identification does not appear to be motivating vaccine uptake. Self-construals Those who said they will not get the vaccine scored significantly lower on interdependent self-construal than those who might, p = .013 (marginal), those who will, p = .001, and those who already have, p = .000, suggesting that those who see themselves as less connected to those around them are more likely to refuse the vaccine. However, there was no significant difference on interdependent self-construal between those who weakly and those who strongly intend to get the vaccine (i.e., Might vs Will) Interestingly, the hesitant “might” group scored marginally significantly lower on independent self-construal than those who said they will receive the vaccine, p = .020, and those who have already received it, p < .001. Overall, however, participants scored higher on independent self-construals than interdependent. Together, these findings suggest that generally in New Zealand individualistic appeals may be more useful than collective appeals, as were commonly used to promote the vaccine, as these appeals are not likely to shift self-construals but could align with them to influence behaviour (e.g., Kim et al., 2022). This is particularly the case for those who said they will not get the vaccine, though this raises the question of whether an intervention targeted at independent self-construals would be capable of motivating these people to get the vaccine. Collective efficacy Vaccine refusers scored significantly lower on collective efficacy than those who are hesitant, p = .001, those who intend to get the vaccine, p = .001, and those who have already had it, p < .001. The only other difference is between the “will” and the “have” groups, p = .003. It could be that receiving the vaccine (i.e., taking part in a collective effort to address a pressing challenge) led to a small increase in perceptions of collective efficacy. Key here is that perceptions of collective efficacy did not significantly differ between the ”might” and “will” groups, suggesting that hesitation around the vaccine may not be due to low perceptions of collective efficacy (and vice versa, that higher collective efficacy motivates vaccine uptake). Social Norms, Beliefs, and COVID-19 Protective Actions Means significantly differed between intentions categories for all four variables: Collective action beliefs, F(3, 569) = 103.92, p < .001, ηp 2 = .35; Descriptive social norms, F(3, 569) = 83.52, p < .001, ηp 2 = .31; Injunctive social norms, F(3, 569) = 125.10, p < .001, ηp 2 = .40; and other protective actions, F(3, 569) = 51.92, p < .001, ηp 2 = .22. Collective action beliefs Those who say they will not receive the vaccine scored much lower on COVID-19 related collective action beliefs than those who might, will, or have received the vaccine, all p < .001. Crucially, the hesitant group also scored lower than those who will or have already had the vaccine, both p < .001, suggesting that such messaging may not be helpful for convincing this group given that these messages have been used widely and repeatedly already. Perceived social norms There is a similar pattern for descriptive and injunctive norms, where those who say they will not receive the vaccine think others are less likely to get it too and are less approving of the vaccine than those who might, will, or already have had the vaccine, all p < .001. Again, of key interest, those in the hesitant group also scored significantly lower than those in the “will” and “have” groups, both p < .001, suggesting that people who are hesitant think fewer others are getting or will get the vaccine, or approve of getting the vaccine. Social norms therefore present a possible target for uptake campaigns. Other protective actions The same pattern was again seen for other protective actions, with those who say they will not get the vaccine less often undertaking other actions to protect against COVID-19 than all other groups, all p < .001. Importantly, those who are hesitant about the vaccine are also less often undertaking other protective actions than those who will or already have had the vaccine, both p < .001. This finding suggests that communication perhaps should not invoke people's previous COVID-related protection actions as a reason for getting vaccinated. Different to perceived social norms which can be shifted in fairly broad, shallow interventions, communication strategies cannot change people's previous behaviour. However, this finding does suggest that underlying principles of efforts to encourage other protective actions may be applicable to encouraging COVID-19 vaccines, given the apparent relationship between other actions and vaccination intentions. COVID-19 and Vaccination Risk Perceptions Mean scores differed between intention categories for all five risk components: General COVID-19 risk, F(3, 544) = 16.72, p < .001, ηp 2 = .08; personal COVID-19 risk, F(3, 544) = 12.01, p < .001, ηp 2 = .06; collective COVID-19 risk, F(3, 544) = 14.97, p < .001, ηp 2 = .08; personal COVID-19 vaccine risk, F(3, 544) = 11.03, p < .001, ηp 2 = .06; and collective COVID-19 vaccine risk, F(3, 544) = 10.64, p < .001, ηp 2 = .06. COVID-19 disease risk Those who said they will not receive the COVID-19 vaccine saw significantly lower general risk from COVID-19 than those in all three other groups, all p < .001. Interestingly, those in the hesitant group saw marginally significantly less risk compared to those who have had already had the vaccine, p = .018, but not those who said they will have the vaccine, p = .055. This suggests that perceptions of risk of COVID-19 may not be the difference between those who somewhat intend to get vaccinated and those who strongly intend to get vaccinated. However, both of these tests are technically non-significant, so the difference suggested here should be tested further with a larger sample size. Those in the “don't” group saw significantly lower personal and collective risk from COVID-19 than those in all other groups, all p < .001. However, there was no significant difference between any of the other groups for either personal or collective risk, suggesting that low levels of perceived risk from COVID-19 is not a key factor in vaccine hesitancy (though might be in vaccine refusal). COVID-19 vaccine risk Interestingly, while vaccine refusers appear to perceive more risk from the vaccine than those who have or will have the vaccine, the difference from the “will” group was only marginally significant for collective risk, p = .022, and non-significant for personal risk. The difference from the “have” group for both personal and collective vaccine risk was non-significant. However, those in the “might” group scored significantly higher for both personal and collective vaccine risk than those in the “will” and “have” groups, both p < .001, suggesting that information about the safety of the vaccine is important to communicate. This finding is key, as it suggests that reassuring those who are hesitant about the safety of the vaccine may help to increase the strength of their intention and translate that intention into behaviour. Vaccine Response Efficacy Mean scores significantly differed between the intentions groups on all three types of response efficacy: personal, F(3, 568) = 150.53, p < .001, ηp 2 = .44; collective, F(3, 568) = 133.90, p < .001, ηp 2 = .41; and general, F(3, 568) = 101.32, p < .001, ηp 2 = .35. Participants who said they will not receive the COVID-19 vaccine saw the vaccine as significantly less effective than those in the “might”, “will”, and “have” groups. Crucially, those in the “might” group saw the vaccine as significantly less effective than those who said they will or have had it, while these latter two groups did not differ. This pattern was found for all types of response efficacy: personal (“Getting a COVID-19 vaccine will keep me safe”), collective (“Getting a COVID-19 vaccine will keep those around me safe”), and general (an average of “Getting a COVID-19 vaccine will help NZ to return to normal faster” and “Getting a COVID-19 vaccine will help NZ open up the borders sooner”). This suggests that communication of the benefits of the vaccine, regardless of what those benefits are, could be effective. Evidence from behaviour change research and practice suggests that it is easier to strengthen a somewhat positive attitude than to change an attitude from negative to positive (McKenzie-Mohr, 2011). Communication efforts targeting more specific communities could therefore identify a limited number of these beliefs which are held positively but weakly to strengthen. Support for Vaccination Mean scores significantly differed between the intentions categories for both support of the COVID-19 vaccine specifically, F(3, 558) = 190.28, p < .001, and other vaccines generally, F(3, 558) = 61.64, p < .001. As expected, those in the “don't” group were significantly less supportive of the COVID-19 vaccine than those in all three other groups, all p < .001, and, importantly, those in the “might” group also scored lower than those in the “will” and “have” groups, both p < .001. The pattern was slightly different for vaccine support generally, with both the “don't” and the “might” group scoring significantly lower than the “will” and “have” groups, all p < .001, but not significantly different from each other. This suggests that general attitudes towards vaccines contributes to hesitancy about the COVID-19 vaccine in particular, but that there may be specific contextual factors which contribute to refusal (as opposed to hesitancy) for the COVID-19 vaccine different to other vaccines. However, it is also important to note that even those in the “don't” group were either neutral or vaguely supportive; support for the COVID-19 vaccine did not differ from the midpoint of the scale, t(38) = 0.99, p = .16, and was significantly above the midpoint for support of other vaccines, t(38) = 3.07, p < .01. Those in the “might” scored above the midpoint for both COVID-19 vaccines, t(109) = 15.63, p < .001, and other vaccines, t(111) = 14.37, p < .001. Discussion Summary of Results Demographics. There were few strong differences in demographic factors between the intentions groups. Those in the “might” group who are hesitant about the vaccine were significantly younger than all other groups (consistent with higher intentions among older people in previous research; Karlsson et al., 2021) and on average had more people in their household (although logically this should increase intention to get vaccinated, as larger households both increases the likelihood of contracting COVID-19 as well as the possibility of infecting a close other). Other demographic differences, such as proportion of people in each group at higher risk of COVID-19 could be due to the staged nature of the rollout prioritizing people who are more risk from serious illness if they were to become infected. Interestingly several previous studies found significant differences between men and women; most found higher intentions among women than men (Fasse & Newby, 2020; Karlsson et al., 2021; Schwarzinger et al., 2021), although other studies found that women were less likely to intend to get a COVID-19 vaccine (Latkin et al., 2021), including in NZ (Thaker, 2021). Research in the area of risk communication suggests that women tend to have higher risk perceptions and to be more likely to take action to prevent or mitigate that risk (see for example the proposed “white male effect”; Finucane et al., 2000), although gender differences are largely inconsistent or non-evident in risk-related fields such as natural hazard research (Becker et al., 2015). Collective factors Those who are vaccine hesitant scored marginally lower on independent self-construals compared to those who said they will receive the vaccine and significantly lower than those who have already had it, suggesting that those who see themselves more as individuals are more likely to get the vaccine. Counter to assumptions, people might be relying on their own, existing attitudes when deciding whether to get vaccinated. Supporting this suggestion, there were no differences between these groups for individual versus collective risk (of either COVID-19 or the vaccine) as well as the finding that those in the “might” group did not score higher on interdependent self-construals or collective efficacy (except compared to the “wont” group). This “might” group, however, scored significantly higher on the Inclusion of Community in Self scale than all other groups. Higher scores compared to the “don't” group is consistent with previous findings that collective factors such as community identification are positively related with uptake intentions (e.g., Freeman et al., 2021; Kwok et al., 2021; Wakefield & Khauser, 2021). The primary purpose of behavioural interventions is typically to create “will” intentions, rather than “might” intentions. However, shifting people who will not engage in positive behaviours such as getting vaccinated to maybe engaging in those behaviours could also be treated as a positive outcome of any intervention efforts, as it would then be more likely that other influences could prompt those people to act. The “might” group also scoring higher than the “will” group suggests a non-linear relationship between intentions and community identification. Some factors relevant to risk communication and prevention/mitigation behaviour tend to be parabolic rather than linear such that higher levels on the factor does not necessarily translate into more behaviour (i.e., both low and high levels have a similar impact on risk-related perception and behaviour; Solberg et al., 2010). It is possible that previous work has found a linear relationship because the analyses were designed to only identify linear relationships. This finding therefore suggests that further work focusing on a limited number of factors should consider non-linear relationships. However, the reason for our findings being inconsistent with previous research can not be confidently concluded based on the current data. Complicating any potential explanation of this difference is that those who will likely get the vaccine did not differ on interdependent self-construal, a measure conceptually similar to the ICS, than those who might get the vaccine. The former asked participants about their relationship with their group(s) rather than their community. It is possible that people consider their role within the groups to which they belong differently than they do their community. Further, people might see their community as less homogenous, particularly on questions such as vaccination, than the groups they considered for the interdependent self-construal scale, resulting in a consistent impact on intentions of self-construal but not of inclusion of community in self. However, again, it is not possible to reach a firm conclusion based on the data here. This question could be explored in future research, to inform whether there is a difference between appealing to people to consider their group or to consider their community. Social and behavioural factors Across all social factors (collective action beliefs, descriptive and injunctive social norms, and frequency of engaging in protective actions against COVID-19), the “don't” group scored lower than all others, and, importantly, the “might” group scored lower than the “will” and “have” groups. This is broadly consistent with previous research showing positive relationships between vaccine intentions or uptake and past behaviour, other protective actions, and social norms (Faase & Newby, 2020; Latkin et al., 2021; Myers & Goodwin, 2012; Schwarzinger et al., 2021; Yang, 2015). While the findings around social norms could inform communication campaigns, the positive relationship between other protective actions and vaccine intentions reflects an ongoing challenge in other risk areas whereby it is easier to encourage people who have done one action to undertake a further action than it is to encourage people who have done nothing to do something (captured in social science theories such as Bandura's triadic reciprocal determinism; Bandura, 1989). Given that the collective action beliefs considered in this survey are largely comprised of messages which have been used widely and repeatedly (such as the appeals to the “team of 5 million”), it is unlikely that efforts to increase the communication of these to the vaccine hesitant will lead to large changes. However, this group perceived positive (i.e., above the midpoint of the scale) descriptive and injunctive norms. Given that it is easier to strengthen an attitude or perception than to change valence (i.e., from negative to positive), it is possible that communication of either or both types of norms to the vaccine hesitant could increase uptake. This suggestion is discussed further in the following section. Risk Those who say they will not get the vaccine saw less risk from COVID-19 across all three factors (personal, collective, and general), consistent with previous research (Karlsson et al., 2021; Kwok et al., 2021; Motta et al., 2021; Schwarzinger et al., 2021). Of particular interest, those in the “might” group did not differ from the “will” or “have” groups for personal or collective risk but did score lower on general risk. This suggests that efforts to communicate the general risk of COVID-19 to New Zealanders (rather than the individual or someone they are close to) might be effective. Karlsson et al. (2021) found that perceptions of illness severity generally, but not personally, were associated with intentions; together with our findings, this suggests that exploring different dimensions of risk, including who is being perceived as at risk, is important to understand people's motivation for prevention and mitigation actions. Consistent with previous findings that perceptions of vaccine safety are positively related with intentions (Fabry et al., 2011; Freeman et al., 2021; Karlsson et al., 2021; Kwok et al., 2021; Thaker, 2021), those in the “might” group saw more risk on both vaccine risk factors compared to the “will” and “have” groups. These findings suggest that messages which communicate the safety of the vaccine, both for the individual and for those they care about, might also be effective. Response efficacy Response efficacy (related to outcome expectancy and instrumental attitudes) has repeatedly been shown to be a key factor motivating risk prevention and mitigation actions in other areas such as natural hazard preparedness (Vinnell et al., 2021). Across all three response efficacy factors (personal, collective, and general), the same pattern as above was found, where the “might” group scored lower than the “will” and “have” groups (and the “don't” group scored lowest of all), largely consistent with previous research (Fabry et al., 2011; Freeman et al., 2021; Yang, 2015; see Faase & Newby, 2020, for an exception). As with social norms, the “might” group scored above the scale midpoint, suggesting that they do hold positive perceptions of response efficacy, but these perceptions are less positive than those who say they will and those who have already had the vaccine. This suggests that efforts to communicate the benefits of the vaccine, regardless of where it is to keep them safe, other people safe, or to help life return to normal, might be effective. If a particular community is the intended audience for messaging, it would be useful to undertake further research to identify if one particular benefit would be more useful for that community. Vaccine uptake Previous research demonstrated between 63% and 88% of people were willing to receive a COVID-19 vaccine (Kerr et al., 2021); a survey in NZ found 74% of participants would get a vaccine once it was available (Thaker, 2021). In our survey, slightly over half (54%) said that they would definitely receive a vaccine. However, when all participants who responded above the midpoint were considered, this rose to 73.1% where above the midpoint represents those more likely than not to get the vaccine. While it is difficult to compare between our findings and those of Thaker (2021) given the latter study used a dichotomous yes/no measure, these findings are similar. However, in our study we also asked if participants had already received at least one dose. When these participants were considered, 92% of participants either had positive intentions to get the vaccine or had already had a COVID-19 vaccine. At time of writing, 95% of the eligible population in NZ had received at least one dose and 93% were fully vaccinated (Ministry of Health, 2022). This suggests that studies of COVID-19 vaccine intentions might have underestimated the percentage of people who would end up being vaccinated. Limitations Related to the above point, one of the limitations of this study is the lack of a perfect relationship between intentions and behaviour. Intentions are considered one of the best predictors of actual behaviour and are substantially easier to measure (Vinnell et al., 2021), but as indicated it cannot be assumed that intended levels will exactly reflect actual behaviour. This limitation, and the finding that vaccine intentions were considerably lower than actual uptake, provides an important caveat for interpreting future work measuring intentions. Also related is the issue of general high levels of vaccine acceptance in NZ. This meant that the sample was skewed significantly towards people who might or will get a COVID-19 vaccine, with few participants being outright refusers. However, the small group of strongly vaccine refusers are unlikely to be motivated to be vaccinated by public education and advertising campaigns. Thus, although any findings relating to this group need to be interpreted with caution, this group is not the core focus of this research. Rather, evidence suggests that the main differences of interest are between those who might and those who will be vaccinated, as those who are hesitant are more likely to be swayed by public campaigns (McKenzie-Mohr, 2011) and should therefore be the focus of research efforts. This main focus on differences between those who weakly intend and those who strong intend to get the vaccine (i.e., might vs. will) partially but not completely mitigates one of the main limitations of this study. The plan was to treat intention to get the vaccine as a continuous variable, using regression to test which of our explanatory variables associated with intention. However, very few participants did not intend to get the vaccine. We believe that grouping participants based on intention scores, essentially transforming the variable from seven potential groups (one for each point on the Likert scale) to three groups allowed for more robust analyses. We acknowledge that this limits the nuance in the data and therefore limits the certainty of our conclusions. Grouping and labelling individuals may imply a conceptual assumption that participants in each of the statistical groups are unique from those in the other groups, rather than varying from others in the degree of their intent. This assumption does reflect real world approaches to behaviour interventions, however, were people who have moderate intentions to undertake a behaviour (i.e., are not likely to convert that intention into action without intervention, but with intervention are likely to act) are identified and communicated to as a homogenous group (e.g., McKenzie-Mohr, 2011). Little literature has considered this question in detail, however, so it poses a possibility for future research which would be useful within and beyond the current context of vaccination intentions. Further, it is possible that people who are neutral about getting the vaccine (i.e., scored 4) would differ from those who intend not to get the vaccine (i.e., scored 1 to 3) but due to the sample size we chose to group these participants together to allow them to be included in the analysis, as the groups would have been too small to include had they been separated. Again, the focus here is on the differences between those who might and those will get the vaccine, as this is an effective way to inform uptake campaigns. There were also some demographic differences between the groups; however, these differences were generally small and/or marginally significant. Given the limits on inference and statistical robustness mentioned above demographic factors were not controlled for in the analyses. Future work with a larger sample size could explore whether factors such as age, objective risk status, and perceived/self-reported risk status influence intentions to get vaccines. For example, although the proportion of people who were high-risk versus not high-risk was only marginally significantly different between groups, it is possible that because when the survey was conducted the vaccine was only available to high-risk groups that these participants thought more concretely about getting vaccinated than those for whom vaccination was not yet a possibility (and therefore more abstract; see for example construal level theory: Trope & Liberman, 2010). Given the national level of this scale, it is unlikely that the findings would generalise to all types of communities in NZ. Any efforts which intend to target messaging at a specific group should undertake further work to identify which factors are most relevant for that group. Finally, it is unclear the extent to which these patterns would be found for other vaccines and for other behaviours where small actions of individuals lead to a collective benefit beyond the sum of the individual benefits. Recommendations Based on our findings, there are several recommendations that can be made for campaigns aimed to increase uptake of COVID-19 vaccination in NZ. While current uptake levels are high, the vaccination programme is ongoing with a recent rollout of booster doses and future booster doses are likely to be required. Thus, further efforts will be needed to encourage people to continue receiving vaccination doses. While our study was able to identify individual and collective factors which might influence intention to get a COVID-19 vaccine and could inform uptake campaigns, these campaigns should still be pilot tested both to ensure that the factors identified here are causally related to intention and that there is potential to increase vaccination behaviour. These efforts could consider crafting messages targeted to younger people, appealing to community identity, presenting both descriptive and injunctive norms, and presenting clear information about the general risk posed by COVID-19 paired with information about the benefits and safety of vaccines. Uncited References McClure et al., 2001, Schultz et al., 2007 Uncited Link Figure 2, Figure 3, Figure 4, Figure 4, Figure 5Figure 2 Mean scores for general collective and individual factors between vaccine intention groups Note. Scores ranged from a possible 1 to 7. Bars are +/- standard error. INTSC = Interdependent Self Construal, INDSC = Independent Self Construal Figure 2 Figure 3 Mean scores for COVID-19 specific behaviours and behavioural perceptions between vaccine intentions groups Note. Scores ranged from a possible 1 to 7, except for Other protective actions which ranged from 1 to 5. Figure 3 Figure 4 Mean scores for COVID-19 and vaccine risk perceptions between vaccine intentions groups Figure 4 Figure 4 Mean scores for general, personal, and collective response efficacy of COVID-19 vaccination Figure 4 Figure 5 Mean scores for support of COVID-19 vaccines versus vaccines for other diseases or illnesses between intention groups Figure 5 Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data Availability Data will be made available on request. Acknowledgements The authors would like to acknowledge the following sources who provided funding for this project: Kia manawaroa – Ngā Ākina o Te Ao Tūroa (Resilience to Nature's Challenge - National Science Challenge) and the Health Research Council NZ for the project ‘COVID-19 Pandemic in Aotearoa NZ: Impact, inequalities & improving our response’. In addition to our funding sources, the authors would like to acknowledge the survey participants who filled in the survey and the editors and reviewers who provided valuable feedback to strengthen this paper. 1 Distribution of scores on the intentions scale: 1 (Definitely not) = 10 participants; 2 = 11; 3 = 4; 4 = 14; 5= 44, 6 = 69; 7 (Definitely yes) = 319, Can't have = 8; Already had = 112 ==== Refs References Amin A.B. Bednarczyk R.A. Ray C.E. Melchiori K.J. Graham J. Huntsinger J.R. Omer S.B. Association of moral values with vaccine hesitancy Nature Human Behaviour 1 2017 873 880 10.1038/s41562-017-0256-5 Anthony KE. Bagley B. Petrun Sayers E.L. Forbes Bright C. To get vaccinated or not? An investigation of the relationship of linguistic assignment of agency and the intention to obtain the COVID-19 vaccine Atlantic Journal of Communication 2021 10.1080/15456870.2021.1981329 Aron A. Aron E.N. Smollan D. Inclusion of Other in the Self Scale and the structure of interpersonal closeness Interpersonal Relations and Group Processes 63 4 1992 596 612 Ashwell D. Murray N. When being positive might be negative: An analysis of Australian and New Zealand newspaper framing of vaccination post Australia's No Jab No Pay legislation Vaccine 38 2020 5627 5633 10.1016/j.vaccine.2020.06.070 32653274 Bandura A. Social Cognitive Theory Vasta R. Annals of child development. Vol. 6. Six theories of child development 1989 JAI Press Greenwich, CT 1 60 Becker, J. S., Paton, D., Johnston, D. M. (2015). Communication of risk: A community resilience perspective. GNS Science Report 2015/66. 30 p. Bennett S.T. Liu J.H. Historical trajectories for reclaiming an indigenous identity in mental health interventions for Aotearoa/New Zealand – Māori values, biculturalism, and multiculturalism International Journal of Intercultural Relations 62 2018 93 102 10.1016/j.ijintrel.2017.05.005 Betsch C. Schmid P. Heinemeier D. Korn L. Holtmann C. Böhm R. Beyond confidence: Development of a measure assessing the 5C psychological antecedents of vaccination PLoS ONE 13 12 2018 e0208601 10.1371/journal.pone.0208601 Brougham D. Haar J.M. Collectivism, cultural identity and employee mental health: A study of New Zealand Māori Social Indicators Research 114 3 2013 1143 1160 10.1007/s11205-012-0194-6 Carbone J.T. McMillin S.E. Reconsidering collective efficacy: The roles of perceptions of community and strong social ties City & Community 18 3 2019 1068 1085 10.1111/cico.12413 Chou W.-Y.S. Budenz A. Considering emotion in COVID-19 vaccine communication: Addressing vaccine hesitancy and fostering vaccine confidence Health Communication 35 14 2020 1718 1722 10.1080/10410236.2020.1838096 33124475 Cialdini R.B. Descriptive social norms as underappreciated sources of social control Psychometrika 72 2 2007 263 268 Cialdini R.B. Reno R.R. Kallgren C.A. A focus theory of normative conduct: Recycling the concept of norms to reduce littering in public places Journal of Personality and Social Psychology 58 6 1990 1015 1026 Faasse K. Newby J. Public perceptions of COVID-19 in Australia: Perceived risk, knowledge, health-protective behaviors, and vaccine intentions Frontiers in Psychology 11 2020 551004 10.3389/fpsyg.2020.551004 Fabry P. Gagneur A. Pasquier J.-C. Determinants of A (H1N1) vaccination: Cross-sectional study in a population of pregnant women in Quebec Vaccine 29 2011 182401829 10.1016/j.vaccine.2010.12.109 Finucane M.L. Slovic P. Mertz C.K. Flynn J. Satterfield T.A. Gender, race and perceived risk: the ‘white male’ effect Health, Risk & Society 2 2 2000 159 172 Forsyth D.R. van Vugt M. Schlein G. Story P.A. Identity and sustainability: Localized sense of community increases environmental engagement Analyses of Social Issues and Public Policy 15 1 2015 233 252 Freeman D. Loe B.S. Chadwick A. Vaccari C. Waite F. Rosebrock L. Jenner L. Petit A. Lewandowsky S. Vanderslott S. Innocenti S. Larkin M. Giubilini A. Yu L.-M. McShane H. Pollard A.J. Lambe S. COVID-19 vaccine hesitancy in the UK: The Oxford coronavirus explanations, attitudes, and narratives survey (OCEAN) II Psychological Medicine 2021 1 5 10.1017/S0033291720005188 Giubilini A. Douglas T. Savulescu J. The moral obligation to be vaccinated: Utilitarianism, contractualism, and collective easy rescue Medicine, Health Care and Philosophy 21 2018 547 560 10.1007/s11019-018-9829-y 29429063 Gray L. MacDonald C. Mackie B. Paton D. Johnston D. Baker M.G. Community responses to communication campaigns for influenza A (H1N1): A focus group study BMC Public Health 12 1 2012 1 12 22214479 Gray L. Rose S.B. Stanley J. Zhang J. Tassell Matamua N. Puloka V. Kvalsvig A. Wiles S. Murton S.A. Johnston D.M. Becker J.S. MacDonald C. Baker M.G Factors influencing individual ability to follow physical distancing recommendations in Aotearoa New Zealand during the COVID-19 pandemic: A population survey Journal of the Royal Society of New Zealand 51 sup1 2021 S107 S126 10.1080/03036758.2021.1879179 Gudykunst W.B. Lee C.M. Assessing the validity of self-construal scales: A response to Levine et al Human Communication Research 29 2 2003 253 274 Hamann K.R.S. Reese G. Seewald D. Loeschinger D.C Affixing the theory of normative conduct (to your mailbox): Injunctive and descriptive norms as predictors of anti-ads sticker use Journal of Environmental Psychology 44 2015 1 9 Hwang K. Lee J. Antecedents and consequences of ecotourism behavior: Independent and interdependent self-construals, ecological belief, willingness to pay for ecotourism services and satisfaction with life Sustainability 10 3 2018 789 10.3390/su10030789 Jeon E. Han Y. Nam M. How you see yourself influences your color preference: Effects of self-construal on evaluations of color combinations Psychology & Marketing 37 7 2020 989 994 10.1002/mar.21348 Kanagawa C. Cross S.E. Markus H.R. Who am I?” The cultural psychology of the conceptual self Personality and Social Psychology Bulletin 27 2001 90 103 Karlsson L.C. Soveri A. Lewandowsky S. Karlsson L. Karlsson H. Nolvi S. Karukivi M. Lindfelt M. Antfolk J. Fearing the disease or the vaccine: The case of COVID-19 Personality and Individual Differences 172 2021 110590 10.1016/j.paid.2020.110590 Kerr J.R. Schneider C.R. Recchia G. Dryhurst S. Sahlin U. Dufouil C. Arwidson P. Freeman A.L.J. van der Linden S. Correlates of intended COVID-19 vaccine acceptance across time and countries: Results from a series of cross-sectional surveys BMJ Open 11 2021 e48025 10.1136/bmjopen-2020-048025 Kim J.Y. Brockner J. Block C.J. Tailoring the intervention to the self: Congruence between self-affirmation and self-construal mitigates the gender gap in quantitative performance Organizational Behavior and Human Decision Processes 169 2022 104118 10.1016/j.obhdp.2022.104118 Kwok K.O. Li K.-K. Wei W.I. Tang A. Wong S.Y.S. Lee S.S. Influenza vaccine uptake, COVID-19 vaccination intention and vaccine hesitancy among nurses: A survey International Journal of Nursing Studies 114 2021 103854 10.1016/j.ijnurstu.2020.103854 Latkin C.A. Dayton L. Yi C. Colon B. Kong X. Mask usage, social distancing, racial, and gender correlates of COVID-19 vaccine intentions among adults in the US PLoS ONE 16 2 2021 e0246970 10.1371/journal.pone.0246970 Markus H.R. Kitayama S. Culture and the self: Implications for selves and theories of selves Psychological Review 98 1991 224 253 Markus H.R. Kitayama S. The cultural psychology of personality Journal of Cross-Cultural Psychology 29 1998 63 87 Mashek D. Cannaday L.W. Tangney J.P. Inclusion of Community in Self Scale: A single-item pictorial measure of community connectedness Journal of Community Psychology 35 2 2007 257 275 McClure J. Allen M.W. Walkey F. Countering fatalism: Causal information in news reports affects judgments about earthquake damage Basic and Applied Social Psychology 23 2001 109 121 10.1207/S15324834BASP2302_3 McKenzie-Mohr D. Fostering sustainable behavior: An introduction to community based social marketing 3rd ed. 2011 New Society Publishers Canada Ministry of Health. (2022). COVID-19: Vaccine data. https://www.health.govt.nz/our-work/diseases-and-conditions/covid-19-novel-coronavirus/covid-19-data-and-statistics/covid-19-vaccine-data Motta M. Sylvester S. Callaghan T. Lunz-Trujill K. Encouraging COVID-19 vaccine uptake through effective health communication Frontiers in Political Science 3 2021 630133 10.3389/fpos.2021.630133 Myers L.B. Goodwin R. Using a theoretical framework to determine adults’ intention to vaccinate against pandemic swine flu in priority groups in the UK Public Health 126 2012 S53 S56 10.1016/j.puhe.2021.05.024 22784583 Pakmehr S. Yazdanpanah M. Baradaran M. How collective efficacy makes a difference in response to water shortage due to climate change in southwest Iran Land Use Policy 99 2020 104798 10.1016/j.landusepol.2020.104798 Radio New Zealand. (2021). Timeline: The year of Covid-19 in New Zealand. https://www.rnz.co.nz/news/national/437359/timeline-the-year-of-covid-19-in-new-zealand Rimal R.N. Lapinski M.K. A re-explication of social norms, ten years later Communication Theory 25 4 2015 393 409 10.1111/comt.12080 Schultz P.W. Nolan J.M. Cialdini R.B. Goldstein N.J. Griskevicius V. The destructive, and reconstructive power of social norms Psychological Science 18 2007 429 434 17576283 Schwarzinger M. Watson V. Arwidson P. Alla F. Luchini S. COVID-19 vaccine hesitancy in a representative working-age population in France: A survey experiment based on vaccine characteristics Lancet Public Health 6 2021 210 221 10.1016/S2468-2667(21)00012-8 Solberg C. Rossetto T. Joffe H. The social psychology of seismic hazard adjustment: re-evaluating the international literature Natural Hazards and Earth System Sciences 10 2010 1663 1677 10.5194/nhess-10-1663-2010 2010 Teitler-Regev S. Shahrabani S. Benzion U. Factors affecting intention among students to be vaccinated against A/H1N1 influenza: A Health Belief Model approach Advances in Preventive Medicine 2011 353207 10.4061/2011/353207 2011 Thaker J. The persistence of vaccine hesitancy: COVID-19 vaccination intention in New Zealand Journal of Health Communication 26 2 2021 104 111 10.1080/10810730.2021.1899346 33719898 Trope Y. Liberman N. Construal-level theory of psychological distance Psychological Review 117 3 2010 440 463 10.1037/2Fa0018963 20438233 Tropp L.R. Wright S.C. Ingroup identification as the inclusion of ingroup in the self Personality and Social Psychology Bulletin 27 5 2001 585 600 Vinnell L.J. Milfont T.L. McClure J. Do social norms affect support for earthquake strengthening legislation? Comparing the effects of descriptive and injunctive norms Environment and Behavior 51 2018 376 400 10.1177/00139165177 52435 Whittaker C. Walker P.G.T. Alhaffar M. Hamlet A. Djaafara B.A. Ghani A. Ferguson N. Dahab M. Checchi F. Watson O.J. Under-reporting of deaths limits our understanding of true burden of covid-19 British Medical Journal 375 2021 1 5 10.1136/bmj.n2239 World Health Organization. (2021). Timeline: WHO's COVID-19 response. https://www.who.int/emergencies/diseases/novel-coronavirus-2019/interactive-timeline Yang Z.J. Predicting young adults’ intentions to get the H1N1 vaccine: An integrated model Journal of Health Communication 20 1 2015 69 79 10.1080/10810730.2014.904023

0

PMC9753451

NO-CC CODE

2022-12-16 23:25:43

no

Curr Res Ecol Soc Psychol. 2022 Dec 15;:100082

utf-8

Curr Res Ecol Soc Psychol

10.1016/j.cresp.2022.100082

oa_other

==== Front Sleep Med Sleep Med Sleep Medicine 1389-9457 1878-5506 Published by Elsevier B.V. S1389-9457(22)01257-6 10.1016/j.sleep.2022.12.008 Article What did the shifting trends in self-reported sleep duration throughout 2020 mean for social disparities in sleep duration? Sheehan Connor a∗ Li Longfeng b Petrov Megan c a School of Social and Family Dynamics, Arizona State University, USA b Pennsylvania State University, USA c Edison College of Nursing and Health Innovation, Arizona State University, USA ∗ Corresponding author. T. Denny Sanford School of Social and Family Dynamics, Arizona State University, P.O. Box 873701, Tempe, 85287-3701, Arizona, USA. 15 12 2022 15 12 2022 7 9 2022 3 11 2022 9 12 2022 © 2022 Published by Elsevier B.V. 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Recent evidence utilizing online samples indicates that sleep patterns were significantly altered during the initial months of the SARS-CoV-2 (COVID-19) pandemic/lockdown. However, it remains less clear how sleep duration changed in population-based samples, in the later months of 2020, and across subpopulations. Here we used a population-based sample to document sleep duration trends for the entire year of 2020, compared these trends to the previous years of 2013, 2014, 2016, and 2018, and systematically analyzed whether self-reported sleep duration patterns in 2020 varied by sex, race/ethnicity, and educational attainment. Data were from the Behavioral Risk Factor Surveillance System (n = 2,203,861) and focused on Americans aged 18 years and older. Respondents self-reported the hours of sleep they got in a 24-hour period. We fit multinomial and linear regression models to predict the category of sleep duration (six or fewer hours, seven to eight hours (base), and nine or more hours) and the raw reports of sleep duration, net of demographic, socioeconomic, and behavioral health covariates. Results revealed significant increases in sleep duration during the months directly after the COVID-19 lockdown (March and April in particular). However, these increases were short lived; reports of sleep duration reverted to historical levels by the Fall of 2020. We also found that the changes in sleep duration trends in 2020 were similar by sex, race/ethnicity, and educational attainment, cumulatively leading to little impact to disparities in sleep duration. In a dramatic, but brief, alteration of population-level sleep duration patterns, disparities in self-reported sleep duration remained intractable. Keywords Sleep Trends Disparities COVID-19 American adults, 2020 ==== Body pmc1 Introduction Humans need sleep to function [1]. Sleep durations that constitute short sleep durations (i.e., six or fewer hours per 24-hour period) are not only associated with decreased immune [2] and psychological functioning [3], but also are associated with increased risk of cardiometabolic conditions [4] and premature death [5]. Sleep durations that constitute long sleep duration (i.e., nine or more hours per 24-hour period) may indicate an underlying health condition [6], but also are associated with increased risk of frailty [7] and death [8]. Salubrious sleep durations at the population-level are not only distributed in a manner that reflects social inequality [9,10] but can also be dramatically, and potentially unevenly, influenced by exogenous events. An increasingly robust body of research indicates that the initial lockdown in response to the SARS-CoV-2 (hereafter COVID-19) pandemic may have altered population-level sleep duration patterns, at least initially (discussed further below). Here we contribute to this research by utilizing a large population-based survey of more than two million American respondents to systematically analyze how self-reported sleep duration changed throughout the entire year of 2020 compared to previous years going back to 2013 and also examine whether shifting patterns during 2020 varied by sex, race/ethnicity, and educational attainment. Understanding secular trends in sleep duration has important implications for intervention and policy, as documenting trends can help to elucidate exposures that impede or improve population-level sleep. Due to the lockdown that coincided with the emergence of the COVID-19 pandemic, researchers have been resourceful and nimble to gain access to high-quality sleep duration data, however this generally resulted in using data based on online sampling or via apps. These sampling designs have known biases relative to population-based research designs such as sampling participants mostly from urban areas in major global cities or those were more affluent and thus able to afford wearable sensors to capture their sleep data. In comparison to one to two years prior to the pandemic, these studies found that generally between March 2020 and May 2020, sleep duration abruptly increased (from 13.7 to 18.6 minutes), particularly on weekdays, bedtime shifted later, and variance of sleep duration within weeks decreased [[11], [12], [13]]. In contrast to the app-based studies, sleep duration data from the nationally representative National Health Interview Study collected before the pandemic were compared to a representative online survey collected during the pandemic. Minimal differences in average sleep duration were found but short and long sleep durations both increased in 2020 compared to 2018 [14]. The shifts in sleep duration patterns during the initial stages of the pandemic may have been unevenly experienced across the population. In the U.S., sleep duration remains stratified by social characteristics [9,10,[15], [16], [17]], yet the pandemic's economic and social impacts on society culminated in countervailing currents that make the influence of the pandemic and accompanying lockdowns on inequality in sleep duration based on gender, race/ethnicity, and educational attainment less clear. For instance, working aged women tended to sleep less than men prior to the pandemic, which likely results from their disproportionate share of household labor (e.g., child care), increased stressors, and gender discrimination broadly and in earnings in particular [15,18]. The pandemic may have exacerbated these trends as childcare burdens increased dramatically after daycares closed and schools transitioned online [19]. Alternatively, more American men were suddenly working from home or without work and could have more time to help with caregiving and household labor. Sleep duration also varies by race/ethnicity. Black adults and Hispanic adults report significantly shorter sleep durations than non-Hispanic White adults [20]. In 2020, the pandemic disproportionately influenced non-Hispanic Black Americans and Hispanic Americans([21], including their insomnia [22]. Black Americans and Hispanic Americans who were more likely to work “frontline”, on-site jobs, and in service sectors that were economically decimated in the tens of millions [23]. The former may have contributed to less sleep opportunity, whereas the latter may have contributed to more sleep opportunity. Similarly, sleep is stratified by educational attainment, with Americans with more education sleeping significantly more [10,24]. Americans with more education were more likely to work from home and thus suddenly had more time to sleep [25]. In contrast, less educated Americans were more likely to work in subsectors of the economy that were devastated by the pandemic [26] and thus were more inclined to lose their job. In sum, previous research has generally indicated that sleep duration increased in the U.S. and other developed countries soon after the global lockdown (i.e., March and April of 2020). We build on this important work in a few important ways. First, rather than comparing trends to the immediate year or years before, we go back to 2013, providing a multi-year foundation that can rule out idiosyncratic baseline comparisons, broader secular trends, random data fluctuations, and potential seasonal (e.g., day light savings) differences in measurement of sleep [[27], [28], [29]]. Second, we use a large population-based survey in the U.S. that used consistent data collection procedures for each of the time periods of analysis (telephone calls) to document month-by-month differences for the entire 2020 rather than the period just after the lockdown. Third, we document if the trends within 2020 varied across sex, race/ethnicity, and educational attainment. Overall, we aim to provide a comprehensive documentation regarding trends in sleep duration within the whole 2020, how they compare to multiple previous years, and whether specific segments of the population were differentially influenced by the shifting secular sleep patterns. 2 Materials and methods 2.1 Data The data for this investigation came from the Behavioral Risk Factor Surveillance System (BRFSS). The BRFSS is a large population-based survey that utilizes telephone-based sampling of cellphones and landlines of non-institutionalized American adults aged 18+ in all 50 states, Washington DC, Guam, and Puerto Rico. Recent estimates from the American Community Survey suggest that 99% of American households have a telephone number, and thus telephone-based sampling has strong external validity [30]. Telephone numbers within a state are randomly dialed, in a manner that each household in that State has a similar probability of being sampled. All responses are self-reports and there are no proxy respondents. Data are then weighted to adjust for non-response, selection, and demographic characteristics. While the BRFSS is designed to be representative of States, estimates of health outcomes correspond quite well with national level estimates from nationally representative data [31]. The BRFSS is conducted throughout the year and has large sample sizes within each month (see Supplemental Table 1), allowing researchers to utilize a population-based survey and track temporal trends with large samples. Notably, the BRFSS operated continually throughout the pandemic, including in the height of the lockdown of March (n = 43,046) and April of 2020 (n = 34,680). The data are publicly available and thus the project is exempt from IRB review. We combined the 2013, 2014, 2016, 2018, and 2020 BRFSS files, utilizing these years as sleep duration was not collected among the entire sample in other years. Our sampling frame consisted of adults aged 18+ (n = 2,279,828). We also removed those who did not report sleep duration (n = 40,053), or those not interviewed in the specified years (35,914, e.g., those in the 2014 file who were interviewed early in 2015). This sampling protocol provided an analytical sample of 2,203,861 respondents or nearly 97% of the total respondents interviewed in the 2013, 2014, 2018, and 2020 files. We found that women, racial/ethnic minorities, and those with low levels of education were less likely to report their sleep duration. Reassuringly, just over 1% of respondents in any given month in 2020 had missing sleep duration, reducing the likelihood that these missing reports of sleep duration affected the substantive conclusions. 2.2 Measures Respondents were asked how much sleep they get in a 24-hour period, and they could respond between 1 and 24 hours. Consistent with previous research and guidelines [20], we coded those who reported six or fewer hours as “short-sleep,” those who reported seven to eight hours as “normal-sleep,” and those who reported nine or more hours as “long-sleep.” To gauge the sensitivity of our results, we additionally analyzed the continuous specification. Previous research has validated the sleep duration measure utilized in the BRFSS by comparing it to actigraphy and other measures of sleep, and found that the duration measure is a valid measure [32]. BRFSS interview year (2013 as reference) and month (January as reference) were categorical, but given the relative comparison to the reference year and month we also calculated absolute levels of short-sleep and the sleep duration (discussed further below). There were no missing data for the year or month of survey. We coded the sex of the respondent as a dichotomous variable with females coded as “0” and males coded as “1.” In terms of race/ethnicity, we used the respondent reports that were coded as: non-Hispanic White (reference; hereafter White), non- Hispanic Black (hereafter Black), Hispanic, and non-Hispanic Other/multiracial (hereafter Other/multiracial). Educational attainment was coded as: “less than high school (ref),” “high school,” “some college,” and “college or more.” Detailed information regarding the coding of all the covariates is provided in the supplemental materials. 2.3 Data analytic plan We began by calculating weighted descriptive statistics of the percentage reporting short sleep and the reported sleep duration (transformed into HH:MM) for each month from the 2013, 2014, 2016, 2018, and 2020 surveys. We documented short sleep, as it is particularly deleterious in terms of health [5], to be consistent with previous trend research [20], and because we found generally little differences in long sleep throughout the study period. We also calculated the weighted descriptive statistics for all covariates and present them in Supplemental Table 1. Given the polytomous nature and implications of sleep duration we fit a multinomial model [33] with normal (seven-eight hours) sleep duration as the base category compared to short (six or fewer hours) and long sleep (nine or more hours). These models included an interaction term between year of interview and month of interview to examine if the reports of sleep duration each month significantly varied across years. These interactions are important as they allow the examination of trends irrespective of seasonality or year to year fluctuations in sleep per se. In addition to the interaction term, this model accounted for demographic, socioeconomic, and behavioral health covariates. We next implemented an identical procedure with an Ordinary Least Squares regression model predicting the raw reports of sleep duration with an interaction term between month and year accounting for the covariates. We also calculated the absolute levels of short sleep duration and reported sleep duration. We present these changes visually, by implementing the margins command to calculate the predicted probability of reporting short sleep and the predicted value of sleep duration (transformed into HH:MM) with all covariates held at their mean value [34]. To examine if the trends varied by gender, race/ethnicity, and educational attainment we fit separate models with three-way interaction terms (see Supplemental Tables 2–4) including all covariates (e.g., sex was controlled for in the race/ethnicity interaction model). We present the results from 2020 from these models in the form of the predicted probability of reporting short sleep duration, from a model that included no other covariates. Missing data on covariates were handled with Stata's multiple imputation suite as 10 imputed datasets were created and then estimates across datasets were combined using Rubin's rule [35]. Unimputed results were similar to the imputed results. 3 Results Table 1 provides the proportion of Americans who report short sleep duration each month as well as the calculated raw values of sleep duration reported by American adults by month and year. In the first month of the lockdown: March 2020, the proportion reporting short sleep decreased below 30.0% for the first time to 29.3% (95% CI [28.9%, 29.7%]), before decreasing to 28.7% (95% CI [28.3%, 29.2%]) in April of 2020. These decreases represent a 2.7% and 3.7% decrease compared to the respective month in 2018. A 3.7% decrease corresponds to nearly approximately 9 million fewer American adults reporting short sleep duration in April of 2020 compared to April of 2018 and represents a relatively similar amount relative to 2013. The raw reports of sleep duration were largely consistent. Beginning in March 2020 (7:07, 95% CI [7:06, 7:08]) reported sleep durations began to increase abruptly, further increasing in April (7:09, 95% CI [7:08, 7:10]) and May (7:08, 95% CI [7:07, 7:09]) and then reverting to traditionally observed levels in the Fall.Table 1 Self-Reported Short Sleep Duration and Mean Self-Reported Sleep Duration (HH:MM), Adults aged 18+, Behavioral Risk Factor Surveillance System, 2013, 2014, 2016, 2018, and 2020. Table 1 Proportion Reporting Short Sleep (≤6 hrs. per 24-hour period) Duration 2013 2014 2016 2018 2020 % 95% CI % 95% CI % 95% CI % 95% CI % 95% CI January 32.0% 31.4% 32.5% 30.8% 30.3% 31.3% 30.9% 30.3% 31.6% 31.7% 31.0% 32.4% 31.3% 30.6% 32.0% February 32.6% 32.1% 33.0% 30.7% 30.2% 31.2% 31.5% 31.0% 32.0% 31.2% 30.7% 31.7% 32.4% 31.9% 32.9% March 33.1% 32.7% 33.6% 31.0% 30.5% 31.4% 31.4% 31.0% 31.9% 32.0% 31.5% 32.4% 29.3% 28.9% 29.7% April 32.8% 32.4% 33.3% 31.0% 30.5% 31.4% 31.5% 31.1% 32.0% 32.4% 31.9% 32.8% 28.7% 28.3% 29.2% May 33.1% 32.7% 33.6% 31.7% 31.2% 32.2% 31.5% 31.0% 31.9% 33.0% 32.5% 33.5% 29.3% 28.8% 29.7% June 33.1% 32.6% 33.6% 32.0% 31.6% 32.5% 31.8% 31.4% 32.3% 32.9% 32.5% 33.4% 30.3% 29.8% 30.8% July 33.1% 32.7% 33.6% 31.7% 31.2% 32.1% 32.3% 31.9% 32.8% 33.0% 32.5% 33.5% 30.4% 29.9% 30.9% August 33.3% 32.8% 33.7% 31.4% 31.0% 31.9% 32.7% 32.2% 33.1% 32.8% 32.4% 33.3% 30.9% 30.4% 31.5% September 33.4% 33.0% 33.9% 32.5% 32.0% 33.0% 32.7% 32.3% 33.2% 33.8% 33.3% 34.3% 31.7% 31.2% 32.3% October 32.8% 32.4% 33.3% 31.7% 31.2% 32.1% 31.9% 31.4% 32.4% 34.0% 33.5% 34.4% 31.7% 31.1% 32.2% November 32.0% 31.5% 32.4% 31.1% 30.7% 31.6% 32.1% 31.7% 32.5% 33.2% 32.7% 33.6% 30.8% 30.3% 31.2% December 31.8% 31.3% 32.2% 32.4% 31.9% 32.9% 32.4% 31.9% 32.8% 33.5% 33.0% 34.0% 31.6% 31.1% 32.1% Mean Reported Sleep Duration (in HH:MM) 2013 2014 2016 2018 2020 Mean 95% CI Mean 95% CI Mean 95% CI Mean 95% CI Mean 95% CI January 7:04 7:03 7:05 7:05 7:04 7:06 7:04 7:03 7:06 7:04 7:03 7:06 7:06 7:04 7:07 February 7:02 7:02 7:03 7:05 7:04 7:06 7:04 7:03 7:05 7:03 7:03 7:04 7:03 7:02 7:04 March 7:01 7:01 7:02 7:04 7:04 7:05 7:03 7:02 7:04 7:03 7:02 7:04 7:07 7:06 7:08 April 7:02 7:02 7:03 7:04 7:03 7:05 7:03 7:02 7:04 7:03 7:02 7:04 7:09 7:08 7:10 May 7:02 7:01 7:03 7:03 7:02 7:04 7:03 7:02 7:04 7:01 7:00 7:02 7:08 7:07 7:09 June 7:03 7:02 7:04 7:03 7:02 7:04 7:03 7:02 7:03 7:02 7:01 7:03 7:05 7:04 7:06 July 7:02 7:01 7:03 7:03 7:02 7:04 7:02 7:01 7:03 7:01 7:00 7:01 7:05 7:04 7:06 August 7:02 7:01 7:03 7:03 7:02 7:03 7:02 7:01 7:03 7:02 7:01 7:03 7:04 7:03 7:05 September 7:01 7:00 7:02 7:01 7:01 7:02 7:02 7:01 7:02 7:00 6:59 7:01 7:03 7:02 7:04 October 7:02 7:02 7:03 7:03 7:02 7:04 7:03 7:02 7:04 6:59 6:59 7:00 7:04 7:03 7:05 November 7:03 7:03 7:04 7:04 7:03 7:05 7:03 7:03 7:04 7:01 7:00 7:02 7:05 7:04 7:06 December 7:05 7:04 7:05 7:02 7:01 7:03 7:02 7:01 7:03 7:01 7:00 7:02 7:04 7:04 7:05 N = 478,811 458,992 466,777 413,998 385,283 Notes: CI = confidence interval. Behavioral Risk Factor Surveillance System. Data are weighted. Table 2 provides the coefficients from multinomial and OLS regression models with an interaction term between month and year of interview predicting reported sleep duration. Before 2020 there were generally few significant interaction terms, about as many as would be expected by chance with an alpha level of 0.05. However, starting in March of 2020 significant month by year interaction terms were observed as there were significantly lower levels of reported short sleep in March, April, May, June, July, August, September, October, and November (given the comparison to January 2013, we present this graphically below). Consistently, the regression model predicting the raw reports of sleep duration indicated few differences in the interaction term before 2020. However, once again starting in March of 2020 respondents began reporting significantly higher levels of sleep duration and this pattern was similar in April, May, June, July, August, and September.Table 2 Coefficients from Multinomial and Regression Models predicting Self-Reported Sleep Duration, American Adults aged 18+, Behavioral Risk Factor Surveillance System 2013, 2014, 2016, 2018, and 2020. Table 2 Multinomial Models Predicting Categorical Sleep Duration Regression Model Predicting Raw Sleep Duration ≤6 hr. vs. 7–8 hr. (base) ≥9 hr. vs. 7–8 hr. (base) RRR 95% CI RRR 95% CI b 95% CI Year 2013 (Reference) 2014 0.950 0.915 0.985 0.969 0.913 1.028 0.019 −0.005 0.043 2016 0.967 0.928 1.009 0.942 0.881 1.007 −0.010 −0.036 0.017 2018 1.016 0.972 1.062 0.965 0.899 1.036 −0.015 −0.043 0.014 2020 1.032 0.989 1.077 1.040 0.972 1.112 −0.005 −0.033 0.022 Month January (Reference) February 1.012 0.978 1.046 0.974 0.922 1.028 −0.024 −0.046 −0.002 March 1.035 1.001 1.070 0.951 0.901 1.004 −0.039 −0.061 −0.018 April 1.038 1.004 1.074 0.996 0.944 1.052 −0.024 −0.046 −0.003 May 1.058 1.023 1.095 0.999 0.945 1.055 −0.036 −0.058 −0.014 June 1.076 1.040 1.114 1.021 0.966 1.079 −0.026 −0.048 −0.004 July 1.072 1.036 1.108 0.983 0.931 1.038 −0.040 −0.062 −0.019 August 1.083 1.047 1.120 1.010 0.957 1.067 −0.034 −0.056 −0.012 September 1.078 1.041 1.115 0.935 0.884 0.988 −0.061 −0.083 −0.039 October 1.054 1.020 1.091 0.956 0.905 1.009 −0.037 −0.059 −0.016 November 0.999 0.966 1.033 0.937 0.888 0.990 −0.017 −0.039 0.005 December 0.991 0.958 1.026 0.966 0.915 1.020 −0.002 −0.024 0.020 Year X Month 2014 X February 0.997 0.949 1.046 1.007 0.932 1.090 0.006 −0.025 0.037 2014 X March 0.979 0.934 1.026 1.034 0.957 1.116 0.027 −0.003 0.058 2014 X April 0.991 0.944 1.040 0.977 0.903 1.056 −0.004 −0.035 0.027 2014 X May 1.007 0.959 1.057 0.966 0.893 1.046 −0.010 −0.041 0.021 2014 X June 1.013 0.965 1.064 0.968 0.894 1.047 −0.019 −0.051 0.012 2014 X July 1.007 0.959 1.056 0.992 0.918 1.073 −0.013 −0.044 0.018 2014 X August 0.988 0.942 1.037 0.916 0.847 0.991 −0.029 −0.060 0.002 2014 X September 1.044 0.994 1.096 1.024 0.945 1.109 −0.021 −0.053 0.010 2014 X October 1.021 0.972 1.071 1.007 0.931 1.089 −0.020 −0.051 0.011 2014 X November 1.048 0.998 1.099 1.041 0.963 1.126 −0.026 −0.057 0.005 2014 X December 1.110 1.056 1.166 1.034 0.955 1.120 −0.059 −0.091 −0.027 2016 X February 0.999 0.948 1.053 1.034 0.950 1.126 0.027 −0.007 0.060 2016 X March 0.981 0.932 1.033 1.032 0.949 1.123 0.033 −0.001 0.066 2016 X April 1.002 0.951 1.056 1.008 0.926 1.097 0.008 −0.026 0.041 2016 X May 0.984 0.934 1.036 1.001 0.921 1.089 0.016 −0.018 0.049 2016 X June 0.973 0.923 1.025 0.966 0.888 1.051 0.001 −0.033 0.034 2016 X July 1.002 0.952 1.054 0.992 0.912 1.078 0.001 −0.032 0.034 2016 X August 1.006 0.956 1.059 0.983 0.905 1.068 −0.004 −0.037 0.029 2016 X September 1.011 0.960 1.065 1.058 0.972 1.152 0.018 −0.015 0.052 2016 X October 1.025 0.972 1.079 1.039 0.954 1.131 0.006 −0.028 0.039 2016 X November 1.088 1.033 1.145 1.068 0.982 1.161 −0.014 −0.047 0.019 2016 X December 1.087 1.032 1.145 0.976 0.898 1.062 −0.050 −0.083 −0.016 2018 X February 0.952 0.902 1.005 0.959 0.878 1.047 0.018 −0.017 0.053 2018 X March 0.963 0.913 1.016 1.022 0.937 1.116 0.034 0.000 0.069 2018 X April 0.975 0.923 1.029 1.006 0.921 1.099 0.024 −0.011 0.059 2018 X May 0.982 0.930 1.038 0.945 0.864 1.033 −0.009 −0.045 0.026 2018 X June 0.979 0.927 1.034 0.976 0.893 1.067 −0.004 −0.039 0.031 2018 X July 0.983 0.931 1.037 0.969 0.887 1.058 −0.011 −0.045 0.024 2018 X August 0.973 0.922 1.027 0.979 0.896 1.069 0.003 −0.032 0.038 2018 X September 1.015 0.961 1.072 0.991 0.906 1.084 −0.007 −0.043 0.028 2018 X October 1.045 0.991 1.103 1.016 0.931 1.109 −0.032 −0.067 0.002 2018 X November 1.058 1.002 1.117 1.026 0.939 1.121 −0.024 −0.059 0.011 2018 X December 1.081 1.024 1.142 1.011 0.926 1.105 −0.046 −0.081 −0.010 2020 X February 1.008 0.955 1.064 0.958 0.879 1.045 −0.012 −0.047 0.022 2020 X March 0.863 0.819 0.910 0.996 0.916 1.083 0.078 0.044 0.111 2020 X April 0.834 0.790 0.879 1.042 0.958 1.134 0.112 0.078 0.146 2020 X May 0.833 0.789 0.879 0.992 0.910 1.080 0.102 0.067 0.136 2020 X June 0.869 0.823 0.917 0.978 0.896 1.067 0.051 0.016 0.086 2020 X July 0.879 0.832 0.927 0.962 0.882 1.049 0.057 0.023 0.092 2020 X August 0.884 0.837 0.933 0.916 0.840 0.999 0.035 0.001 0.070 2020 X September 0.915 0.867 0.967 0.967 0.885 1.057 0.041 0.005 0.076 2020 X October 0.942 0.893 0.994 1.015 0.931 1.106 0.029 −0.006 0.063 2020 X November 0.941 0.892 0.992 0.998 0.917 1.086 0.026 −0.008 0.060 2020 X December 0.984 0.933 1.038 0.970 0.890 1.057 0.007 −0.027 0.042 Constant 0.585 0.564 0.606 0.105 0.099 0.111 6.949 6.925 6.972 Data Source: Behavioral Risk Factor Surveillance System, 2013, 2014, 2016, 2018, and 2020. N = 2,203,861. Notes: RRR = relative risk ratio, CI = confidence interval. Normal (7–8 hr.) sleep duration is the base category in multinomial models. Significant coefficients (p < 0.05) bolded. Models account for age, gender, race/ethnicity, marital status, number of children, educational attainment, income, home ownership, employment status, smoking status, drinking status, BMI, and exercise status. Fig. 1 provides the estimated predicted probability of reporting short sleep duration among Americans by month and year with the covariates held at their mean values. Starting in March of 2020 the percentage of Americans reporting short sleep duration decreased significantly, and stayed significantly lower through September of 2020, where after it reverted to levels consistent with previous years. Similarly, Fig. 2 provides the marginal predicted values of sleep duration (converted to HH:MM) by month and year with the covariates held at their mean values. While most of the years preceding 2020 were relatively similar in terms of reported sleep duration, starting in March and April of 2020 reported sleep duration increased significantly and remained elevated relative to the other years through September of 2020.Fig. 1 Predicted Probability of Reporting Short Sleep Duration (Six or Fewer Hours), Adults aged 18+, Behavioral Risk Factor Surveillance System, 2013–2020 Note. Error bars represent 95% confidence interval. Model accounts for age, gender, race/ethnicity, marital status, number of children, educational attainment, income, home ownership, employment status, smoking status, drinking status, BMI, and exercise status. Fig. 1 Fig. 2 Predicted Mean of Self-Reported Sleep Duration (HH:MM), Adults aged 18+, Behavioral Risk Factor Surveillance System, 2013–2020 Note. Error bars represent 95% confidence interval. Model accounts for age, gender, race/ethnicity, marital status, number of children, educational attainment, income, home ownership, employment status, smoking status, drinking status, BMI, and exercise status. Fig. 2 3.1 Sex, racial/ethnic, and educational differences We next fit three-way interaction terms between year of interview, month of interview, and separate models for: sex, race/ethnicity, and educational attainment. The full results are presented in Supplemental Table 2 (sex), Supplemental Table 3 (race/ethnicity), and Supplemental Table 4 (educational attainment). We found little systematic evidence of sex differences in trends across years in general and within 2020 in particular. These results are also presented in Fig. 3 A, where men's and women's sleep seemingly changed in concert with one another. Additionally, we found little evidence in racial/ethnic differences in trends within 2020. Fig. 3B consistently shows that the documented racial/ethnic disparities remained large and significant in 2020 and did not seem to change throughout 2020. Fig. 3C similarly indicates little changes in sleep disparities within 2020 by level of educational attainment. Taken together these results suggest that the significant shifts in reported sleep duration in 2020 were generally similar across sex, race/ethnicity, and educational attainment, resulting in little changes to sleep duration disparities.Fig. 3 Predicted Probability of Reporting Short Sleep Duration (Six or Fewer Hours), By Gender (Panel A), Race/Ethnicity (Panel B), and Educational Attainment (Panel C), Adults aged 18+, Behavioral Risk Factor Surveillance System, 2020 Note. Error bars represent 95% confidence interval. No covariates included in the model. See Supplemental Tables 2–4 for results with covariates. Fig. 3 4 Discussion In this study, we analyzed the BRFSS years of 2013, 2014, 2016, 2018, and 2020 collectively leading to a sample of more than 2 million Americans, to document secular sleep duration trends within and between years. In doing so, we paid close attention to how reported sleep duration shifted during the emergence of the COVID-19 pandemic and corresponding lockdown, while documenting gender, racial/ethnic, and educational differences in these trends. We contribute to previous research in three important ways. First, our results, at least substantively, replicated past findings using app data and online samples [[11], [12], [13]] with a population-based sample. In March and April of 2020, we found that reports of short sleep duration (six or fewer hours per 24-hour period), decreased by nearly 4% compared to previous years. Based on our weighted sample size, this estimate corresponds to about 9 million fewer Americans reporting short sleep duration in April of 2020 than in April of 2018, and similar amounts compared to the other years where sleep patterns were generally consistent (e.g., about a 4% decrease in short sleep compared to 2013). This means that for the months immediately after lockdown, about nine million fewer Americans than in previous years were sleeping short sleep durations associated with deleterious health outcomes [4] and risk of death [5]. Correspondingly, we observed significant increases in the raw reports of sleep duration through the Fall of 2020. While our results substantively replicated past findings, the estimates did differ slightly as reported sleep duration increased by 3–7 minutes in March and April of 2020 (depending on the year of comparison), which is less than previous studies using smartphone data one of which estimating an 13.7 minutes increase in sleep duration in Western cities [13] and another estimating 11.3–18.6 minute increases in Europe [11]. The discrepancy in the exact amount could be due to the inclusion of rural Americans, Americans who do not use smart phones/track their sleep, or Americans who live in urban areas that did not as stringently enforce or adhere to the lockdown. That is, our sample is likely to include those whose lives may not have been as dramatically influenced by the lockdown as those who live in highly connected global cities. Accordingly, the population-based results we estimated were more conservative than previous studies. The varying influence of the lockdown on sleep patterns based on adherence to the lockdown remains an important area for subsequent research. Even after accounting for demographic, socioeconomic, and behavioral health covariates, we were statistically unable to explain the increase in trends of sleep duration; however, we offer the following three potential explanations. First, many Americans had more time flexibility than ever as millions of Americans worked from home for the first time, had their schooling moved online, lost their jobs, or underwent some combination of these experiences. While losing a job can be incredibly stressful and past work has shown this has undermined sleep in 2020 [19], work too can be stressful and undermine sleep [36]. Indeed, for many Americans, the lockdown represented the first respite from the stressors of work. The millions of Americans who were suddenly not working for the first time in their adult lives had more time for sleep. In addition, the financial and unemployment stressors in 2020 were also at least somewhat buffered by the CARES act which provided enhanced jobless benefits and additional stimuli. Those Americans working at home suddenly had no commutes or lighter work schedules, providing more time to sleep. Second, the pandemic and lockdown more broadly were extremely stressful [19], and thus sleep quality might have decreased and need to be made up for with increased duration. Indeed, the sleep duration question was asked regarding a typical 24-hour period, thus people could be making up for poor quality sleep with naps, and recent research has illustrated that naps increased considerably in the pandemic [37,38]. Third, people had considerably more flexibility in their social schedules early in the pandemic and spent less time socializing but more time sleeping. Likely, the increases in sleep duration, are a mix of these three factors [13,39]. A second contribution is that we found that the increase in sleep duration among Americans was short lived. By Fall of 2020 sleep patterns were largely similar to previous years. This is a novel contribution given that past research has tended to focus on the period immediately after lockdown or until the early summer [11,12,19]. Thus, we contribute to this work by showing that by Fall of 2020, sleep duration patterns were consistent with how they were before the pandemic. As Americans adjusted to living in the pandemic, and society began to reopen, sleep duration patterns reverted. The increasing reports of short sleep duration in the Fall of 2020 are concerning for population health, as over the course of a few months millions of Americans who were sleeping healthy durations in the months after lockdown began sleeping durations associated with higher attendant risk of negative health outcomes [4] as well as death [5]. The third major finding is that we found no significant differences in sleep duration trends within 2020 based on sex, race/ethnicity, and educational attainment. Sleep disparities were just as intractable during and after the pandemic as before. These findings coincide with previous research suggesting that the influence of COVID-19 specific stressors did not uniquely influence sleep quality by gender, race/ethnicity, or educational attainment [19]. The findings are troubling as during one of the most dramatic and abrupt increases in the reported sleep duration on record, sleep disparities across gender, race/ethnicity, and educational attainment remained [22,38]. That is, Americans seemingly consistently changed their sleep duration in the early stages of the pandemic resulting in stable levels of inequality in sleep as before the pandemic. Similarly, as unhealthy sleep patterns remerged, sleep disparities persisted. Overall our results indicate that health disparities that result from sleep inequalities [40] will likely endure going forward. There are important limitations that must be considered. First, we relied on self-reported sleep duration, which has been shown to be biased relative to objective measures [41]. However, research that compared the questions utilized by the BRFSS, including sleep duration, to objective measures of sleep, including duration as measured by actigraphy [32], found that the BRFSS measure was “valid.” We also have no reason to think that the bias would suddenly and dramatically shift in March of 2020, which would be necessary to explain our results, and are similarly reassured that our substantive results are consistent with online, app, and Fitbit data. Additionally, sleep health is also viewed as an increasingly multidimensional construct that should include self-reports in addition to objective measures [42]. Second, while the BRFSS is designed to be representative of States and has held up well in comparison to national level estimates of other health conditions [31], the extent to which it has external validity at the month level is not clear. Saying that, the large sample sizes during a period of lockdowns where other data collection procedures were untenable, makes its usage alluring for analyzing trends between years and especially within 2020. Data collection was influenced by COVID-19 in the BRFSS, however the cell sizes in March and April of 2020 were consistent with previous years. A third limitation is the lack of a 2019 comparison. As the BRFSS sleep questions are generally conducted bi-annually we lacked data to compare to 2019, and thus perhaps a new seasonal trend could have emerged in 2019. Reassuringly, past researchers who did have 2019 data did not document trends similar to those during the lockdown [12], and we found little other examples of idiosyncratic patterns within the other years we investigated. The data, while population-based are cross-sectional and longitudinal data would provide more internal validity, however we are unaware of any prospective population based monthly data of sleep in the U.S. 5 Conclusions Sleep duration remains not only an important barometer of current population-health, it is likely also an important indicator of subsequent population health. Overall, we add to a collectively robust literature documenting substantial increases in sleep duration the early phases of the COVID-19 pandemic. While we can only speculate, the short-lived increases we documented might have been beneficial for the population-health of Americans by promoting immune system functioning [2] and the ability to cope with contextual stressors as well as adversities during the early phases of the pandemic. However, more research is needed to understand the short and long-term consequences of the short-lived increases in sleep duration. Regardless of the potential consequences, these brief increases did little to abate social inequality in sleep. Overall, our findings illustrate how ingrained sleep disparities are even in the face of an exogenous shock to population-level sleep duration patterns. CRediT authorship contribution statement Connor Sheehan: Conceptualization, Methodology, Software, Data curation, Writing – original draft, Writing – review & editing. Longfeng Li: Conceptualization, Methodology, Writing – original draft, Writing – review & editing, Visualization. Megan Petrov: Writing – review & editing, Supervision. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix A Supplementary data The following is the Supplementary data to this article:Multimedia component 1 Multimedia component 1 Acknowledgements This research was supported in part by funding from the 10.13039/100010063 College of Liberal Arts and Sciences and the T. Denny Sanford School of Social and Family Dynamics at Arizona State University. The contents of this manuscript are solely the responsibility of the authors and do not represent the official views of Arizona State University or Pennsylvania State University. Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.sleep.2022.12.008. ==== Refs References 1 Grandner M.A. Fernandez F.-X. The translational neuroscience of sleep: a contextual framework Science 374 2021 568 573 34709899 2 Ali T. Choe J. Awab A. Wagener T.L. Orr W.C. Sleep, immunity and inflammation in gastrointestinal disorders World J. Gastroenterol. WJG 19 2013 9231 24409051 3 Hamilton N.A. Nelson C. Stevens N. Kitzman H. Sleep and psychological well-being Soc Indicat Res 82 2007 147 163 4 Buxton O.M. Marcelli E. Short and long sleep are positively associated with obesity, diabetes, hypertension, and cardiovascular disease among adults in the United States Soc Sci Med 71 2010 1027 1036 20621406 5 Cappuccio F.P. D'Elia L. Strazzullo P. Miller M.A. Sleep duration and all-cause mortality: a systematic review and meta-analysis of prospective studies Sleep 33 2010 585 592 20469800 6 Patel S.R. Malhotra A. Gottlieb D.J. White D.P. Hu F.B. Correlates of long sleep duration Sleep 29 2006 881 889 16895254 7 Baniak L.M. Yang K. Choi J. Chasens E.R. Long sleep duration is associated with increased frailty risk in older community-dwelling adults J Aging Health 32 2020 42 51 30270714 8 Youngstedt S.D. Jean-Louis G. Long sleep a greater mortality risk than short sleep in older adults J Am Geriatr Soc 59 2011 957 21568977 9 Johnson D.A. Jackson C.L. Williams N.J. Alcántara C. Are sleep patterns influenced by race/ethnicity–a marker of relative advantage or disadvantage? Evidence to date Nat Sci Sleep 11 2019 79 31440109 10 Krueger P.M. Friedman E.M. Sleep duration in the United States: a cross-sectional population-based study Am J Epidemiol 169 2009 1052 1063 19299406 11 Lee P.H. Marek J. Nálevka P. Sleep pattern in the US and 16 European countries during the COVID-19 outbreak using crowdsourced smartphone data Eur J Publ Health 31 2021 23 30 12 Rezaei N. Grandner M.A. Changes in sleep duration, timing, and variability during the COVID-19 pandemic: large-scale Fitbit data from 6 major US cities Sleep Health 7 2021 303 313 33771534 13 Robbins R. Affouf M. Weaver M.D. Czeisler M.É. Barger L.K. Quan S.F. Czeisler C.A. others Estimated sleep duration before and during the COVID-19 pandemic in major metropolitan areas on different continents: observational study of smartphone app data J Med Internet Res 23 2021 e20546 14 Hisler G.C. Twenge J.M. Sleep characteristics of US adults before and during the COVID-19 pandemic Soc Sci Med 276 2021 113849 15 Burgard S.A. Ailshire J.A. Gender and time for sleep among US adults Am Socio Rev 78 2013 51 69 16 Garcia C. Sheehan C.M. Flores-Gonzalez N. Ailshire J.A. Sleep patterns among US latinos by nativity and country of origin: results from the national health interview survey Ethn Dis 30 2020 119 128 31969791 17 Petrov M.E. Lichstein K.L. Differences in sleep between black and white adults: an update and future directions Sleep Med 18 2016 74 81 25754383 18 Li L. Sheehan C.M. Thompson M.S. Measurement Invariance and sleep quality differences between men and women in the pittsburgh sleep quality Index J Clin Sleep Med 15 2019 1769 1776 31855162 19 Cobb R.J. Sheehan C.M. Nguyen A.W. Johnson D. COVID-19 hardships and self-reported sleep quality among American adults in March and April 2020: results from a nationally representative panel study 2022 Sleep Health 20 Sheehan C.M. Frochen S.E. Walsemann K.M. Ailshire J.A. Are US adults reporting less sleep?: findings from sleep duration trends in the National Health Interview Survey Sleep 42 2018 zsy221 2004–2017 21 Garcia M.A. Homan P.A. García C. Brown T.H. The color of COVID-19: structural racism and the disproportionate impact of the pandemic on older Black and Latinx adults J Gerontol Ser B 76 2021 e75 e80 22 Cheng P. Casement M.D. Cuellar R. Johnson D.A. Kalmbach D. Cuamatzi Castelan A. Drake C.L. Sleepless in COVID-19: racial disparities during the pandemic as a consequence of structural inequity Sleep 45 2022 zsab242 34788453 23 Gezici A. Ozay O. How race and gender shape covid-19 unemployment probability 2020 Available SSRN 3675022 24 Sheehan C.M. Walsemann K.M. Ailshire J.A. Race/ethnic differences in educational gradients in sleep duration and quality among US adults SSM-Popul Health 12 2020 100685 25 Chen J.T. Testa C. Waterman P. Krieger N. Intersectional inequities in COVID-19 mortality by race/ethnicity and education in the United States, January 1, 2020–January 31, 2021. Harv Cent Popul Dev Stud 21 2021 26 Daly M.C. Buckman S.R. Seitelman L.M. The unequal impact of COVID-19: why education matters FRBSF Econ Lett 17 2020 1 5 27 Allebrandt K.V. Teder-Laving M. Kantermann T. Peters A. Campbell H. Rudan I. Wilson J.F. Metspalu A. Roenneberg T. Chronotype and sleep duration: the influence of season of assessment Chronobiol Int 31 2014 731 740 24679223 28 Basner M. Dinges D.F. Sleep duration in the United States 2003–2016: first signs of success in the fight against sleep deficiency? Sleep 41 2018 zsy012 29 Hisler G.C. Muranovic D. Krizan Z. Changes in sleep difficulties among the US population from 2013 to 2017: results from the National Health Interview Survey Sleep Health 5 2019 615 620 31685440 30 Popeo P. Federal communications commission 2020 31 Nelson D.E. Powell-Griner E. Town M. Kovar M.G. A comparison of national estimates from the national health interview survey and the behavioral risk factor surveillance system Am J Publ Health 93 2003 1335 1341 32 Jungquist C.R. Mund J. Aquilina A.T. Klingman K. Pender J. Ochs-Balcom H. van Wijngaarden E. Dickerson S.S. Validation of the behavioral risk factor surveillance system sleep questions J Clin Sleep Med 12 2016 301 310 26446246 33 Hoffmann J.P. Generalized linear models: an applied approach 2004 Pearson College Division 34 Williams R. Using the margins command to estimate and interpret adjusted predictions and marginal effects STATA J 12 2012 308 331 35 Allison P.D. Missing data 2001 Sage publications 36 Burgard S.A. Ailshire J.A. Putting work to bed: stressful experiences on the job and sleep quality J Health Soc Behav 50 2009 476 492 20099452 37 Dai W. Zhou J. Li G. Zhang B. Ma N. Maintaining normal sleep patterns, lifestyles and emotion during the COVID-19 pandemic: the stabilizing effect of daytime napping J Sleep Res 30 2021 e13259 38 Petrov M.E. Pituch K.A. Kasraeian K. Jiao N. Mattingly J. Hasanaj K. Youngstedt S.D. Buman M.P. Epstein D.R. Impact of the COVID-19 pandemic on change in sleep patterns in an exploratory, cross-sectional online sample of 79 countries Sleep Health 7 2021 451 458 34193394 39 Blume C. Schmidt M.H. Cajochen C. Effects of the COVID-19 lockdown on human sleep and rest-activity rhythms Curr Biol 30 2020 R795 R797 32693067 40 Goosby B.J. Straley E. Cheadle J.E. Discrimination, sleep, and stress reactivity: pathways to African American-White cardiometabolic risk inequities Popul Res Pol Rev 36 2017 699 716 41 Lauderdale D.S. Knutson K.L. Yan L.L. Liu K. Rathouz P.J. Self-reported and measured sleep duration: how similar are they? Epidemiol Camb Mass 19 2008 838 845 42 Buysse D.J. Sleep health: can we define it? Does it matter? Sleep 37 2014 9 17 24470692

0

PMC9753452

NO-CC CODE

2022-12-16 23:25:43

no

Sleep Med. 2022 Dec 15; doi: 10.1016/j.sleep.2022.12.008

utf-8

Sleep Med

10.1016/j.sleep.2022.12.008

oa_other

==== Front J Comput Sci J Comput Sci Journal of Computational Science 1877-7503 1877-7503 Elsevier B.V. S1877-7503(22)00285-X 10.1016/j.jocs.2022.101926 101926 Article D-Cov19Net: A DNN based COVID-19 detection system using lung sound Chatterjee Sukanya Roychowdhury Jishnu Dey Anilesh ⁎ Department of Electronics and Communication Engineering, Narula Institute of Technology, Agarpara, India ⁎ Corresponding author. 15 12 2022 15 12 2022 10192618 5 2021 14 11 2021 27 11 2022 © 2022 Elsevier B.V. All rights reserved. 2022 Elsevier B.V. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The limitations of proper detectors for COVID-19 for the proliferating number of patients provoked us to build an auto-diagnosis system to detect COVID-19 infection using only one parameter. Our designed model is based on Deep Convolution Neural Network and considers lung/respiratory sound as the deterministic input for our approach. 'D-Cov19Net' has been trained with 23,592 recordings, begetting an AUC of 0.972 and sensitivity of 0.983 after 100 epochs. The model can be of immense utility in biomedical technology due to its significant accuracy, simplicity, user convenience, feasibility, and faster detection while maintaining social distancing. Keywords COVID-19 Detection Convolution Neural Network (CNN) Deep Learning Lung/Respiratory sound Auto-diagnosis system ==== Body pmc1 INTRODUCTION COVID-19 pandemic has taken its toll on millions of lives worldwide since its unfortunate outbreak [1], [2], [3]. The worsening situation has led medical, engineering, and related researchers to brainstorm ideas to stop the rapidly increasing death counts with each passing day. As a need of the hour, various effective and ingenious methods to detect the presence of the deadly virus in the human body at the earliest possible stage have also been proposed and devised. One widely used, 'gold standard,' reliable and effective laboratory test for COVID-19 detection is Reverse Transcription Polymerase Chain Reaction (RT-PCR) [4], [5]. But, due to the rapidly increasing demand, the proper kit required to perform RT-PCR worldwide, alongside constraints like timing for specimen collection and the inconvenience caused due to the generation of false-negative results [5], [6], the quest for another effective yet more reliable method went on. Another pitfall of RT-PCR lies in the fact that abnormalities, later recorded as COVID-19 infection, were observed in Chest Radiograph or Chest Tomography scans but were reported to be negative in some cases RT-PCR [7], [8], [9], [10], [11]. Ghoshal and Tucker [12] proposed a neural network with the utility of Bayesian Convolutional Neural Networks (BCNN), with a significantly large dataset. A lightweight Deep Neural Network (DNN) has been proposed by Li et al. [13], bearing accuracy of 88%. Xu et al. [14] devised a 3-D Deep Learning model after data classification, wherein the accuracy level was 87.6%. Shan et al. [15] devised a method for segmenting and quantifying the infected regions. J. D. Arias et al. [16] opted for detection using deep learning with chest X-Ray as a parameter. Bai et al. [17] devised a method to distinguish the COVID-19 infection from pneumonia by using an enormous dataset comprising 132 583 CT slices and a bearing accuracy of 96%. Despite remarkable accuracy and efficiency, these methods were insufficient to ensure clinically acceptable detection [7] for the increasing number of patients. Moreover, the feasibility and user convenience were yet to be improved in some of the cases. Therefore, the thrust for finding a yet more affordable and feasible tool that could help in early detection of COVID-19 infection even in the areas where biomedical testing is limited or error-prone was in process. Keeping the constraints in mind, we have developed our architecture, named 'D-Cov19Net', to detect the infection, using lung sounds as the parameter for detection. The model has been developed with a deep learning approach and is presently ahead of the pathological tests in terms of convenience and accuracy. Moreover, the use of lung sound as the sole parameter for detection makes our approach stand apart from the developed methods of detecting COVID-19 with the aid of Artificial Intelligence. D-Cov19Net considers lung sounds the input time-series signal for the model and converted to a 2D equivalent signal. Then it is processed by the separable convolution layers of the network and has been embellished with the proper usage of ReLu activation [18] and Softmax functions [19] to bring forth an accuracy of 97.22%, which can certainly play a significant role in the detection of the infection. Moreover, the sole purpose of our architecture does not remain with detection solely, but also with the auto-diagnosis report in the form of a spectrogram and its analysis to provide the user with a coherent picture of the severity of the disease in the user under consideration. The proposed model can function efficiently and bear significant accuracy, reducing the time complexity without any human intervention in the midst. Apart from these, it performs on a non-contact basis, thus ensuring the maintenance of the COVID-19 norms of social distancing. The overall model described in this paper may not bear the highest possible accuracy, but it can be surely said that it is much more feasible and cost-effective as it treats lung sound as the sole physiological parameter to the inserted by the user. The rest of the task is performed by an auto-diagnosis system, which can generate a report to ensure the detection of the infection at an early stage, which is a crucial parameter for the patient's recovery. Moreover, the development of the designed tool will yield more academically useful applications that could facilitate biomedical research and engineering domain. This method can serve well in areas especially where pathological tests are difficult to carry for such an enormous number of people. The architecture can be developed further to detect other diseases, especially respiratory diseases, incorporating required changes. All of the points have been summarised in Fig. 1.Fig. 1 The reasons and benefits for building the model. Fig. 1 The materials and methodology used in our designed method have been elaborated in Section II. Section III highlights the results and discussions. Finally, the paper is concluded in Section IV. The references have been cited in section V. 2 MATERIALS AND METHODS 2.1 DATASET COLLECTION Primarily the data has been collected from Kaggle [20] and others by manual collection from different sources [21], [22], [23], [24]. The origin of the Kaggle dataset can be traced from Portugal and Greece. 920 recordings have been taken into account. A total of 5.5 hours of recording has been collected from 126 patients, varying in duration from 10 s to 90 s. It can be further classified into 506 crackles and wheezes, 886 wheezes, and 1864 crackles, thus accounting for 6898 respiratory cycles and ultimately 920 recordings. The dataset consisted of both clean and noisy recordings to bring forth the conditions that occur in reality. The dataset has been collected from children, adults, and elderly persons. The respiratory sounds have been captured using the digital stethoscope and other techniques. The positions from where the recordings have been done include Trachea (Tc), Anterior left (Al), Anterior right (Ar), Posterior left (Pl), Posterior right (Pr), Lateral left (Ll), and Lateral right (Lr). These data are primarily used for the 'normal' and 'others' category in our dataset. The ERS dataset [21] consists of 20 different case recordings of auscultated lung sounds. Among these 20 recordings, there are 11 wheezes, five crackles, two recordings of patients with pneumonia and lung cancer, and two recordings of patients with pleural bleeding and pleural effusion. The Medzcool dataset [22] is a sample dataset that consists of auscultated lung sounds of COVID-19. These consist of wheezes, fine crackles, and coarse crackles, making a total of 4 samples, and each sample varies in duration from 5 s to 10 s. The EMT dataset [23] is also a sample dataset of auscultated lung sound, which includes bronchial breath, coarse crackles, fine crackles, diminished breath sound, expiratory wheeze, pleural rub, rhonchi, stridor, vesicular breathing sound, which makes a total of 9 samples and each sample has a duration of 10 s to 20 s. These data are used for the serious lung diseases in the 'others' category and the confirmed cases of the 'Covid-19' category. The SARS-Cov-2 dataset [24] consists of auscultated lung sounds of 30 different COVID-affected patients. The auscultation process includes a Bluetooth digital stethoscope (Stemoscope) used on six bilateral pulmonary fields to record the required sound. Two are at the basal field, two at the middle field, and two at the upper field, both on anterior and posterior position, respectively. For a detailed recording of auscultated sound, a frequency range of 20 Hz to 1 kHz was used. Every subject was instructed to breathe deeply for the 30 s, including 2 s of constant inspiration and 2 s of constant expiration. During the measurement process, subjects were instructed to sit upright or have a bed elevation of 45 to 90 degrees in case of the forbidden patients. These data are primarily used for the 'Covid-19' category in which all the subjects are confirmed cases of COVID-19 disease. Finally, data augmentation was applied to the existing real data to enhance our dataset. For augmentation, time- stretching techniques at multiple levels along with slight pitch variation were used. In case of time stretching, we slowed down and sped up the audio signal by a factor of t = 0.5 to 1.5 (0.5, 0.75, 1.25, 1.5 are the respective rates which were used), and for pitch variation, we pitched up and down the audio signal by a factor of p = -1 to 1 (-1, -0.5, 0.5, 1 are the respective factors which were used) semitone which was exported as a single data and then added to the original dataset. Thus, the final data in our dataset increased from 983 to 23,592 which is 24 times the original dataset, thus increasing varience in the data set yet retaining the features of a specific sample data. 2.2 D-Cov19Net: THE PROPOSED MODEL The D-Cov19Net has been designed based on tiled Convolution Neural networks [25]. The data obtained is normalized, as one of the effective pre-processing techniques before feeding into the architecture, to beget the required output, with the least possible time complexity [26]. The Convolution function can be generally stated as [27],(1) (f⁎g)(t)≜∫−∞∞f(τ)g(t−τ)dτ… Where, (τ) is a weighting function, and τ is the parameter defining the function (t). But, as our signal was digital in nature so the input values will be discrete, thus changing the formula to,(2) (f⁎g)(t)=Δ∑τ=−∞∞f(τ)g(t−τ)… The 1D time-series signal has been converted to a 2D equivalent signal, resembling the kernel size of the input layer. Thus, if we consider our input to be 'I' and denote the kernel by 'K,' the equation can be mathematically stated as to where the symbols have their usual meanings [27],(3) (f⁎g)(t)=Δ∑nI(m,n)K(i−m,j−n)… The respiratory sounds are to be recorded by the lung auscultation process with a digital stethoscope, and then it is converted into a two-dimensional form by Mel Spectrogram. The architecture uses the zero-padding technique [28] to ensure that feature learning efficiently takes place. For the sake of enhancing the efficiency of our proposed architecture and optimizing the excess computational power, the depth-wise separable design based on 2D Convolutional layers has been implied. D-Cov19Net is designed to be inspired by the Xception architecture [29], with numerous alterations to beget more accurate results and make our proposed method more convenient and effective. In D-Cov19Net, each convolution layer has been optimized using batch normalization [30] and ReLU activation functions, except for the last convolution layer. For avoiding complexions, two 2D convolution layers (3×3) were stacked in the first layer to ensure lightness and proper functionality. The first two layers have been optimized using batch normalization and ReLU activation. Batch normalization helps accelerate the learning rate and makes the network more stable through normalization of the input layer by re-centring and re-scaling the concerned data. Our algorithm introduced three different depth-wise separable convolution sections with the input bearing kernel size of 128x128x1. After pre-processing, the dataset was configured to fit into the model, and batch normalization was included to prevent overfitting or underfitting and to have a stable learning process. ( Fig. 2)Fig. 2 The illustration of the data augmentation where Si is the initial data. Fig. 2 Here, we consider J(V, K) as the loss function that we want to minimize, now, due to the utilization of the backpropagation function [31], we will find a tensor G with its elements. Next, we have computed the derivatives concerning the weights in the kernel. The mathematical function, which has been instrumental in making our architecture, can be stated mathematically as,(4) ∂∂Ki,j,k,lJ(V,K)=∑m,nGi,m,nVj,(m−1)×s+k,(n−1)×s+l… To backpropagate the error further down, we have computed the gradient concerning V. Several separable layers have been introduced in the middle flow section, with eight stacked units to obtain a deeper network for more accurate results. The batch normalization has been used in a specific manner where each entry controlled separable layer is Batch normalized except the exiting layer. These beget more trainable parameters without overfitting the layers. The categorical cross-entropy loss function has been incorporated to calculate the loss of a sample from the three classes by computing the following sum:(5) Loss=−∑i=1outputsizeyi.logyˆi… Where y i denotes the ith scalar value in the model output, y i denotes the corresponding target value, and the output size denotes the number of scalar values in the model output. In the exit flow, two dense or fully connected layers have been introduced, the penultimate layer being ReLU activated and the last dense layer is Softmax activated with three output classes "Covid," "Normal," "Others." After the pre-processing of the input signal is completed, converting the one-dimensional signal into two- dimensional form by spectrogram, which is further reshaped to 128×128, the model predicts the output class. The result so obtained acts as an auto-diagnosis report, thus detecting the COVID-19 infection in an individual. The model has been trained with the pre-processed data incorporating the backpropagation algorithm and Adam optimization [32]. The model has been trained on 100 epochs. The model's architecture has been summarized and presented conveniently using flow charts ( Fig. 3), and the model summary has also been included ( Table 1).Fig. 3 (from left) Brief illustration of the proposed network; (from top right): Illustration of Convolution section A, Illustration of Convolution section B, Illustration of Convolution section C. Fig. 3 Table 1 Model Summary. Table 1Layer (type) Output Shape Parameter Connected to input_3 (InputLayer) [(None, 128, 128, 1)] 0 conv2d_12 (Conv2D) (None, 64, 64, 32) 288 input_3[0][0] batch_normalization_57 (BatchNo (None, 64, 64, 32) 128 conv2d_12[0][0] re_lu_90 (ReLU) (None, 64, 64, 32) 0 batch_normalization_57[0][0] conv2d_13 (Conv2D) (None, 64, 64, 64) 18432 re_lu_90[0][0] batch_normalization_58 (BatchNo (None, 64, 64, 64) 256 conv2d_13[0][0] re_lu_91 (ReLU) (None, 64, 64, 64) 0 batch_normalization_58[0][0] separable_conv2d_88 (SeparableC (None, 64, 64, 128) 8768 re_lu_91[0][0] batch_normalization_59 (BatchNo (None, 64, 64, 128) 512 separable_conv2d_88[0][0] re_lu_92 (ReLU) (None, 64, 64, 128) 0 batch_normalization_59[0][0] conv2d_14 (Conv2D) (None, 32, 32, 128) 8192 re_lu_91[0][0] separable_conv2d_89 (SeparableC (None, 64, 64, 128) 17536 re_lu_92[0][0] batch_normalization_60 (BatchNo (None, 32, 32, 128) 512 conv2d_14[0][0] max_pooling2d_8 (MaxPooling2D) (None, 32, 32, 128) 0 separable_conv2d_89[0][0] add_24 (Add) (None, 32, 32, 128) 0 batch_normalization_60[0][0] max_pooling2d_8[0][0] re_lu_93 (ReLU) (None, 32, 32, 128) 0 add_24[0][0] separable_conv2d_90 (SeparableC (None, 32, 32, 256) 33920 re_lu_93[0][0] batch_normalization_61 (BatchNo (None, 32, 32, 256) 1024 separable_conv2d_90[0][0] re_lu_94 (ReLU) (None, 32, 32, 256) 0 batch_normalization_61[0][0] separable_conv2d_91 (SeparableC (None, 32, 32, 256) 67840 re_lu_94[0][0] batch_normalization_62 (BatchNo (None, 32, 32, 256) 1024 separable_conv2d_91[0][0] re_lu_95 (ReLU) (None, 32, 32, 256) 0 batch_normalization_62[0][0] conv2d_15 (Conv2D) (None, 16, 16, 256) 32768 batch_normalization_60[0][0] separable_conv2d_92 (SeparableC (None, 32, 32, 256) 67840 re_lu_95[0][0] batch_normalization_63 (BatchNo (None, 16, 16, 256) 1024 conv2d_15[0][0] max_pooling2d_9 (MaxPooling2D) (None, 16, 16, 256) 0 separable_conv2d_92[0][0] add_25 (Add) (None, 16, 16, 256) 0 batch_normalization_63[0][0] max_pooling2d_9[0][0] re_lu_96 (ReLU) (None, 16, 16, 256) 0 add_25[0][0] separable_conv2d_93 (SeparableC (None, 16, 16, 728) 188672 re_lu_96[0][0] batch_normalization_64 (BatchNo (None, 16, 16, 728) 2912 separable_conv2d_93[0][0] re_lu_97 (ReLU) (None, 16, 16, 728) 0 batch_normalization_64[0][0] conv2d_16 (Conv2D) (None, 8, 8, 728) 186368 batch_normalization_63[0][0] separable_conv2d_94 (SeparableC (None, 16, 16, 728) 536536 re_lu_97[0][0] batch_normalization_65 (BatchNo (None, 8, 8, 728) 2912 conv2d_16[0][0] max_pooling2d_10 (MaxPooling2D) (None, 8, 8, 728) 0 separable_conv2d_94[0][0] add_26 (Add) (None, 8, 8, 728) 0 batch_normalization_65[0][0] max_pooling2d_10[0][0] re_lu_98 (ReLU) (None, 8, 8, 728) 0 add_26[0][0] REPEATED 8 TIMES separable_conv2d_95 (SeparableC (None, 8, 8, 728) 536536 re_lu_98[0][0] batch_normalization_66 (BatchNo (None, 8, 8, 728) 2912 separable_conv2d_95[0][0] re_lu_99 (ReLU) (None, 8, 8, 728) 0 batch_normalization_66[0][0] separable_conv2d_96 (SeparableC (None, 8, 8, 728) 536536 re_lu_99[0][0] batch_normalization_67 (BatchNo (None, 8, 8, 728) 2912 separable_conv2d_96[0][0] re_lu_100 (ReLU) (None, 8, 8, 728) 0 batch_normalization_67[0][0] separable_conv2d_97 (SeparableC (None, 8, 8, 728) 536536 re_lu_100[0][0] re_lu_101 (ReLU) (None, 8, 8, 728) 0 separable_conv2d_97[0][0] separable_conv2d_98 (SeparableC (None, 8, 8, 728) 536536 re_lu_101[0][0] add_27 (Add) (None, 8, 8, 728) 0 separable_conv2d_98[0][0] add_26[0][0] re_lu_130 (ReLU) (None, 8, 8, 728) 0 add_34[0][0] separable_conv2d_127 (Separable (None, 8, 8, 728) 536536 re_lu_130[0][0] batch_normalization_82 (BatchNo (None, 8, 8, 728) 2912 separable_conv2d_127[0][0] re_lu_131 (ReLU) (None, 8, 8, 728) 0 batch_normalization_82[0][0] conv2d_17 (Conv2D) (None, 4, 4, 1024) 745472 add_34[0][0] separable_conv2d_128 (Separable (None, 8, 8, 1024) 752024 re_lu_131[0][0] batch_normalization_83 (BatchNo (None, 4, 4, 1024) 4096 conv2d_17[0][0] max_pooling2d_11 (MaxPooling2D) (None, 4, 4, 1024) 0 separable_conv2d_128[0][0] add_35 (Add) (None, 4, 4, 1024) 0 batch_normalization_83[0][0] max_pooling2d_11[0][0] separable_conv2d_129 (Separable (None, 4, 4, 1536) 1582080 add_35[0][0] batch_normalization_84 (BatchNo (None, 4, 4, 1536) 6144 separable_conv2d_129[0][0] re_lu_132 (ReLU) (None, 4, 4, 1536) 0 batch_normalization_84[0][0] separable_conv2d_130 (Separable (None, 4, 4, 2048) 3159552 re_lu_132[0][0] re_lu_133 (ReLU) (None, 4, 4, 2048) 0 separable_conv2d_130[0][0] separable_conv2d_131 (Separable (None, 4, 4, 2048) 4212736 re_lu_133[0][0] batch_normalization_85 (BatchNo (None, 4, 4, 2048) 8192 separable_conv2d_131[0][0] re_lu_134 (ReLU) (None, 4, 4, 2048) 0 batch_normalization_85[0][0] global_average_pooling2d_2 (Glo (None, 2048) 0 re_lu_134[0][0] dense_4 (Dense) (None, 2048) 4196352 global_average_pooling2d_2[0][0] dense_5 (Dense) (None, 3) 6147 dense_4[0][0] Total parameters: 33,605,451 Trainable parameters: 33,566,331 Non-trainable parameters: 39,120 3 RESULTS AND DISCUSSIONS Our model has been trained with the pre-processed data by backpropagation based on the Adam optimization algorithm on 100 epochs. 98.3% of training accuracy and 97.2% of validation accuracy were achieved after completion of training ( Fig. 4). The validation accuracy and training accuracy have been depicted graphically in Fig. 5.Fig. 4 Illustration of the proposed method. Fig. 4 Fig. 5 Training loss and accuracy graph. Fig. 5 For instance, we used a real-time lung sound of a subject, 20 years old, which is recorded by lungs auscultation process through a digital stethoscope and fed it into the system for the feature extraction by Mel Spectrogram. After the extraction process is done, the audio is then fed into the model for the prediction. The output was 'Normal' with a prediction accuracy of 99.67%. The output on the spectrogram has been depicted [ Fig. 6(a)]. Similarly, an input of a 1st stage COVID affected a person's lung sound, aged 32 years old was tested, and the output was 'COVID Positive' with a prediction accuracy of 98.79% [Fig. 6(b)]. All the subjects used for testing are confirmed subjects as stated earlier in section II.Fig. 6 (a), 6(b): (from left) spectrogram of the normal lung sound and a COVID affected lung sound. Fig. 6 The main challenge was to give distinct predictions between the 'COVID Positive' and 'Others.' The 'Others' class encompasses the lung sounds of all the other diseases associated with the lungs, and pneumonia was among them. The reason behind using a normal Pneumonia-affected lung sound is that COVID-19 also causes pneumonia with certain deviations. To test this, we gave input on a Pneumonia patient's lung sound. The result came out to be as expected. It gave an output 'Others' with a prediction accuracy of 82.45% [ Fig. 7].Fig. 7 Spectrogram of a Pneumonia affected lung sound. Fig. 7 The sounds used for testing the model were unique and were not present in the training dataset. To predict the overall testing accuracy of our model, we computed a confusion matrix based on four main parameters, which are the True Positive (T.P.), False Positive (F.P.), False Negative (F.N.), and True Negative (F.N.) on three different class each having 60 different test samples. From the computation of the confusion matrix on the test data set, we achieved a sensitivity of 98.33% and a specificity of 96.667%. Again, we computed the confusion matrix based on 'COVID Positive' and 'Others.' Here again, we achieved a sensitivity of 98.33% and a specificity of 96.667%. The results show that our proposed method can detect the true cases of COVID-19 positive with just a 1.67% error rate, and on the other hand, our proposed method can detect the true cases of the COVID-19 negative with just a 3.333% error rate. Thus, our proposed method proves to be highly sensitive and specific even if there are some critical cases. The overall accuracy is calculated through the confusion matrix [33], which is 97.223%, considering the minimum amount of data that we used for the model training. The confusion matrix has been tabulated in Table 2.Table 2 Confusion Matrix for COVID-19 detection. Table 2Class Test 1 (COVID +ve) Test 2 (COVID -ve) COVID-19 TP=59 FN=1 Normal FP=2 TN=58 Others FP=2 TN=58 In addition to that, the analysis based on a histogram has also been performed [34]. The histogram plotting for normal, pneumonia affected, and COVID-19 affected patients have been plotted respectively in Fig. 8(a,b,c). The frequency data of each data point is averaged from all the samples for a specific category, and the resulted data is visualized by histogram. The histogram shows the average dominating frequencies with their respective minimum threshold value. From the figure, we can conclude that, though the histogram plotted for COVID-19 infected case and pneumonia case resemble, certain differences mark the distinction. Thus, all the results obtained so far indicates our method of detection to be efficient and effective.Fig. 8 (a,b,c): (from top left) Normal, pneumonia affected, COVID-19 affected. Fig. 8 Several methods for detecting COVID-19 have been devised in recent times using A.I. A comparison table has been plotted ( Table 3) to compare the accuracy for all the proposed methods using deep learning. It indicates our proposed method to be accurate than the majority of other proposed methods. Moreover, we notice that the detection accuracy of the infection is higher when using speech signals as input. But, we designed this model to be more efficient as it determines the infection in an individual directly by accounting for the acoustic parameters of the lung without increasing complexity.Table 3 Accuracy comparison of different COVID test. Table 3Sl. No. The method used for detection Medical Parameter used for detection Approximate accuracy (in percentage) 1 A.I. (Deep Learning) C.T. Scan [35], [36] 90.8 2 A.I. (Deep Learning) Chest X-Ray [16] 91.5 3 A.I. (Deep Learning) Speech/cough recording [37], [38] 98 4 A.I. (Our proposed method) Lung sound 97.22 To determine the accuracy and account for the validation of the model, we have plotted a ROC curve, and the sensitivity and AUC have been visualized from there [39]. The ROC curve ( Fig. 9) is plotted following the confusion matrix depicted above (Table 3).Fig. 9 ROC Curve. Fig. 9 This result indicates that our model can classify amidst the mentioned classes properly and the True Positive Rate being significantly higher depicts the sign of reliability on the outcome of our model. As a result, our suggested technique can detect and classify COVID-19 disease with excellent specificity and sensitivity. 4 CONCLUSION The rapid spread of COVID-19 worldwide and the increasing number of deaths require urgent actions from all sectors. The key factors behind the rapid spread by continuously changing its phase of the COVID-19 pandemic, such as scarcity of testing and its cost and time-consuming procedure, prompt us to take some urgent action. This study reveals an all-over deployable deep-learning-based preliminary diagnosis tool for COVID-19 using lung sound samples. The innovation and utility of this architecture lie in the fact that it outdoes different existing architectures in terms of efficiency. Another note is that the full implementation has been done successfully with a standard- sized dataset with remarkable accuracy. Moreover, this test for COVID detection is contamination-free, i.e., there will be no false prediction based on the contamination factor. To acquire a proper lung sound sample, the recording should be done in a controlled environment to eliminate the unwanted noise frequencies, resulting in an accurate prediction. And, the total proposed model can be used without any further modification in the areas of medical sciences to test for COVID infection efficiently without violating any preventive norms. The main challenges included distinguishing lung sounds from a pneumonia patient and a COVID-19 infected patient. But, our model proved its efficiency in overcoming the challenge. Moreover, a huge amount of data is not required to train the model. This architecture, with minimal modifications, can be used to function in the same way for some diseases which bear similar symptoms. Even though the results brought forth by the model seem promising, it can always be considered safe to get the analysis verified by some medical personnel, thus ensuring its clinical acceptability. With the involvement of the medical community and required modifications, the scope for building a clinically acceptable model can be accounted for using our proposed work. Our study and proposed model can also act as a building block of various other researches to detect lung diseases with the aid of Artificial Intelligence. With time and further development, this method can turn out to be of immense use in the medical domain for detection of crucial diseases in a glimpse, with minimal effort, that could play a pivotal role in saving the lives of many due to detection within the expected timeframe when medical science will still have something to do and cure the patient, instead of giving up just because, detection of the disease took place after the threshold of time, to save a precious life. Thus, to conclude, all the graphs and illustrations have significantly highlighted our model to be legit with a desired level of accuracy. To account for the validation of our model, the ROC curve has been plotted, and the AUC and sensitivity have been calculated after that. With a significant outcome, i.e., an AUC of 0.972 and a sensitivity of 0.983, it can be concluded that our model is reliable and academically useful. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Anilesh Dey was born in West Bengal, India, in 1977. He received a B.E in Electronics from Nagpur University and M.Tech. (Gold-Medallist) in Instrumentation and Control Engineering from Calcutta University and received PhD. from Jadavpur University. He is working as an Associate Professor and is presently the Head of the Department of Electronics and Communication Engineering at the Narula Institute of Technology, Agarpara, Kolkata. He has been the author and co-author of more than 70 scientific papers in international/national journals and proceedings of the conferences with reviewing committee. He has conducted several research works in the domain of biomedical engineering. Sukanya Chatterjee was born in West Bengal, India, in 2001. She is pursuing her B.Tech graduation in Electronics and Communication Engineering from Narula Institute of Technology, Agarpara, Kolkata Jishnu Roychowdhury was born in West Bengal, India, in 2000. He is pursuing his B.Tech graduation in Electronics and Communication Engineering from Narula Institute of Technology, Agarpara, Kolkata ==== Refs References 1 Ai T. Correlation of chest CT and RT-PCR testing in coronavirus disease 2019 (COVID-19) in China: a report of 1014 cases Radiology 2020 10.1148/radiol.2020200642 2 Arias-Londoño J.D. Gómez-García J.A. Moro-Velázquez L. Godino-Llorente J.I. Artificial Intelligence Applied to Chest X-Ray Images for the Automatic Detection of COVID-19. A Thoughtful Evaluation Approach IEEE Access vol. 8 2020 226811 226827 10.1109/ACCESS.2020.3044858 34786299 3 Bai Harrison X. Wang Robin Xiong Zeng Hsieh Ben Chang Ken Halsey Kasey Tran Thi My. Linh Choi Ji. Whae Wang Dong-Cui Shi Lin-Bo Mei Ji Jiang Xiao-Long Pan Ian Zeng Qiu-Hua Hu Ping-Feng Li Yi-Hui Fu Fei-Xian Huang Raymond Y. Sebro Ronnie Yu Qi-Zhi Atalay Michael K. Liao Wei-Hua Artificial Intelligence Augmentation of Radiologist Performance in Distinguishing COVID-19 from Pneumonia of Other Origin at Chest CT". Radiology 296 3 2020 E156 E165 32339081 4 Bhanja, Samit & Das, Abhishek. (2018). Impact of Data Normalization on Deep Neural Network for Time Series Forecasting. 5 Francesco Bianchi, Giada Bianchi, and Dongho Song, The Long-Term Impact of the COVID-19 Unemployment Shock on Life Expectancy and Mortality Rates, NBER Working Paper No. 28304, December 2020, Revised January 2021, JEL No. C32, E32, I14, J11 6 Boltzmann Ludwig Studien über das Gleichgewicht der lebendigen Kraft zwischen bewegten materiellen Punkten" [Studies on the balance of living force between moving material points] Wiener Berichte 58 1868 517 560 7 [Dataset] EMTprep (2019), Lung Sounds Collection, Available: 〈https://youtu.be/KRtAqeEGq2Q〉. 8 [Dataset] Medzcool (2020), Sounds of Coronavirus (COVID-19) - Lung Sounds, Available: 〈https://www.medzcool.com/〉, https://youtu.be/3Kkp6ZM35As. 9 [Dataset]Pengyu Zhang Bingjian Wang Yan Liu Muge Fan Yong Ji Hao Xu Mengdan Xu Songwen Chen Qing Li Zhi Zhang "Lung Auscultation of Hospitalized Patients with SARS-CoV-2 Pneumonia via a Wireless Stethoscope" In International Journal of Medical Sciences Vol.18 2021 1415 1422 10.7150/ijms.54987 10 [Dataset] Rocha B.M. Filos D. Mendes L. Vogiatzis I. Perantoni E. Kaimakamis E. Natsiavas P. Oliveira A. Jácome C. Marques A. Paiva R.P. Α Respiratory Sound Database for the Development of Automated Classification Precision Medicine Powered by pHealth and Connected Health 2018 Springer Singapore 51 55 11 [Dataset] Anita Simonds, Marc Humbert, Carlos Robalo Cordeiro, Reference Database of Respiratory Sounds," European Respiratory Society (ERS), E-learning resources, Available: 〈https://www.ers-education.org/e-learning/reference-database-of-respiratory-sounds/〉. 12 Fang Y. Sensitivity of chest C.T. for COVID-19: comparison to R.T.–PCR. Radiology 296 2020 200432 13 François Chollet. (2017). Xception: Deep Learning with Depthwise Separable Convolutions. 14 Ghoshal, B. & Tucker, A. Estimating uncertainty and interpretability in deep learning for coronavirus (COVID-19) detection. Preprint at 〈http://arxiv.org/abs/2003.10769〉 (2020). 15 Guan W.J. Ni Z.Y. Hu Y. Liang W.H. China Medical Treatment Expert Group for Covid-19. Clinical Characteristics of Coronavirus Disease 2019 in China. Apr 30 N Engl J Med. 382 18 2020 1708 1720 10.1056/NEJMoa2002032 32109013 16 Harmon S.A. Sanford T.H. Xu S. Artificial intelligence for detecting COVID-19 pneumonia on chest C.T. using multinational datasets Nat Commun 11 2020 4080 32796848 17 Hashemi Mahdi Enlarging smaller images before inputting them into the convolutional neural network: zero- padding vs. interpolation. Journal of Big Data 2019 6 10.1186/s40537-019-0263-7 18 Ian Goodfellow, Yoshua Bengio, & Aaron Courville (2016). Deep Learning. MIT Press. 19 Ioffe, Sergey et al. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. Proceedings of the 32nd International Conference on Machine Learning. PMLR. 20 Kingma, Diederik & Ba, Jimmy. (2014). Adam: A Method for Stochastic Optimization. International Conference on Learning Representations. 21 Laguarta J. Hueto F. Subirana B. COVID-19 Artificial Intelligence Diagnosis using only Cough Recordings. IEEE Open Journal of Engineering in Medicine and Biology 2020 10.1109/OJEMB.2020.3026928 22 LeCun Y. Kavukcuoglu K. Farabet C. Convolutional networks and applications in vision Proceedings of 2010 IEEE International Symposium on Circuits and Systems, Paris, France 2010 253 256 10.1109/ISCAS.2010.5537907 23 Li, X., Li, C. & Zhu, D. COVID-MobileXpert: on-device COVID-19 screening using snapshots of chest X-ray. Preprint at 〈http://arxiv.org/abs/2004.03042〉 (2020). 24 Long C. Diagnosis of the coronavirus disease (COVID-19): rRT–PCR or CT? Eur. J. Radiol. 126 2020 108961 25 Marois G. Muttarak R. Scherbov S. Assessing the potential impact of COVID-19 on life expectancy PLoS ONE 15 9 2020 e0238678 26 Mumbai students developAI-based voice tool to detect COVID-19, 16th April. 27 Nair V. Hinton G.E. "Rectified linear units improve restricted Boltzmann machines," Haifa 2010 807 814 28 Pearson K. Contributions to the Mathematical Theory of Evolution. II. Skew Variation in Homogeneous Material". Philosophical Transactions of the Royal Society A: Mathematical Physical and Engineering Sciences 186 1895 343 414 29 Roberts M. Driggs D. Thorpe M. Common pitfalls and recommendations for using machine learning to detect and prognosticate COVID-19 using chest radiographs and C.T. scans. Nat Mach Intell 3 2021 199 217 10.1038/s42256-021-00307-0 30 Rumelhart David E. Hinton Geoffrey E. Williams Ronald J. Learning representations by back-propagating errors Nature. 323 6088 1986 533 536 31 Salameh J.P. Leeflang M.M. Hooft L. Islam N. McGrath T.A. van der Pol C.B. Frank R.A. Prager R. Hare S.S. Dennie C. Spijker R. Deeks J.J. Dinnes J. Jenniskens K. Korevaar D.A. Cohen J.F. Van den Bruel A. Takwoingi Y. van de Wijgert J. Damen J.A. Wang J. Cochrane COVID-19 Diagnostic Test Accuracy Group, McInnes MD. Thoracic imaging tests for the diagnosis of COVID-19 30th September;9:CD013639 Cochrane Database Syst Rev. 2020 10.1002/14651858.CD013639.pub2 32 Shan F., Gao Y., Wang J., Shi W., Shi N., Han M., et al. (2020) Lung infection quantification of COVID-19 in C.T. images with deep learning. arXiv preprint arXiv:200304655 33 Spackman Kent A. Signal detection theory: Valuable tools for evaluating inductive learning Proceedings of the Sixth International Workshop on Machine Learning. San Mateo 1989 Morgan Kaufmann. CA 160 163 34 Sperrin M. Grant S.W. Peek N. Prediction models for diagnosis and prognosis in COVID-19 BMJ 369 2020 m1464 32291266 35 Stehman Stephen V. Selecting and interpreting measures of thematic classification accuracy." Remote Sensing of Environment. 62 1 1997 77 89 36 Tahamtan Alireza Ardebili Abdollah Real-time RT-PCR in COVID-19 detection: issues affecting the results Expert review of molecular diagnostics vol. 20 5 2020 453 454 10.1080/14737159.2020.1757437 37 Udugama Buddhisha Diagnosing COVID-19: The Disease and Tools for Detection ACS nano vol. 14 4 2020 3822 3835 10.1021/acsnano.0c02624 38 Wang, S. et al. A deep learning algorithm using C.T. images to screen for Corona Virus Disease (COVID-19). medRxiv. Preprint at 〈https://www.medrxiv.org/content/10.1101/2020.02.14.20023028v5〉 39 Xu X., Jiang X., Ma C., Du P., Li X., Lv S., et al. (2020) Deep learning system to screen coronavirus disease 2019 pneumonia. arXiv preprint arXiv:200209334

0

PMC9753453

NO-CC CODE

2022-12-16 23:25:43

no

J Comput Sci. 2022 Dec 15;:101926

utf-8

J Comput Sci

10.1016/j.jocs.2022.101926

oa_other

==== Front J Infect J Infect The Journal of Infection 0163-4453 1532-2742 The British Infection Association. Published by Elsevier Ltd. S0163-4453(22)00705-8 10.1016/j.jinf.2022.12.015 Letter to the Editor Changes of Mycoplasma pneumoniae infection in children before and after the COVID - 19 pandemic, Henan, China Liang Ying Zhang Pin Du Bang Zhang Xianwei ⁎ Hou Guangjun ⁎ Zhang Wancun ⁎ Henan Key Laboratory of Children's Genetics and Metabolic Diseases, Children's Hospital Affiliated to Zhengzhou University, Henan Children's Hospital, Zhengzhou Children's Hospital, Zhengzhou, 450018, China ⁎ Corresponding author: 15 12 2022 15 12 2022 30 11 2022 13 12 2022 © 2022 The British Infection Association. Published by Elsevier Ltd. All rights reserved. 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcDear editor, In this journal, Li et al., and Zhou et al., successively reported the decline of Streptococcus pneumoniae and Haemophilus influenzae infections in children under the impact of the COVID - 19 pandemic1 , 2. Li et al. compared the effect of COVID - 19 on the incidence of Escherichia coli infections in respiratory system and digestive system in children, the results indicated indicate that the COVID - 19 may mainly affect the incidence of respiratory system infection, and has little impact on the incidence of digestive system infection3. Up to now, there was no data on Mycoplasma pneumoniae (M. pneumoniae) infections during the COVID - 19 pandemic. M. pneumoniae is a prokaryotic microorganism without cell wall, insensitive to cell wall antimicrobial agents such as lactam, and transmitted through air droplets, coughing, sneezing and close contact4. M. pneumoniae causes up to 40% of community-acquired pneumonia in children and can develop into serious life-threatening diseases such as refractory mycoplasma pneumoniae pneumonia, necrotizing pneumonia, fulminant pneumoniae and M. pneumoniae encephalitis[5], [6], [7]. In China, macrolide-resistant M. pneumoniae is very common and the prevalence ranges from 83% to 95%, which makes it difficult to treat mycoplasma infection6. Therefore, it is important to dynamically monitor children's M. pneumoniae infection and understand its epidemiological changes so as to formulate preventive strategies. Here we evaluated the changes in M. pneumoniae infection in children before and after the COVID - 19 pandemic, which may help to inform the implementation of clinical prevention strategies. The Henan Children's Hospital was approved as the National Children's Regional Medical Center, Henan Children's Medical Center, and Henan Pediatric Disease Clinical Medical Research Center. In this study, M. pneumoniae infection was monitored in the Henan Children's Hospital from January 1, 2017 to October 31, 2022. From 2017 to 2019, the positive number and positive rate of M. pneumoniae RNA and M. pneumoniae serological tests fluctuated seasonally, while during the two COVID-19 pandemics, the positive number and positive rate of M. pneumoniae RNA and M. pneumoniae serological tests decreased significantly twice (Fig. 1 A and Fig. 2 A). In particular, after the end of the two COVID - 19 pandemics, the positive number and positive rate of M. pneumoniae RNA and M. pneumoniae serological tests continued to decrease for several months, which may inhibit the seasonal upward trend of M. pneumoniae infection. Although the positive number and positive rate of M. pneumoniae RNA and M. pneumoniae serological test in children increased slightly during the recovery period after two COVID - 19 pandemics, it was still lower than that in the same period before COVID - 19 pandemic. Therefore, the epidemic trend of M. pneumoniae infection in children in Henan Province changed before and after the epidemic of COVID - 19.Fig. 1 (A) The positive number of M. pneumoniae RNA and the positive rate of M. pneumoniae RNA from January, 2017, to October, 2022. (B) The number of positive infection of M. pneumoniae RNA form 2017 to 2021. (C) The number of M. pneumoniae RNA positive in different ages from January 2017 to October 2022. (D) The positive rate of M. pneumoniae RNA in different ages from January 2017 to October 2022. Fig 1 Fig. 2 (A) The positive number of M. pneumoniae serological test and the positive rate of M. pneumoniae serological test from January, 2017, to October, 2022. (B) The number of positive infection of M. pneumoniae serological test form 2017 to 2021. (C) The number of M. pneumoniae serological test positive in different ages from January 2017 to October 2022. (D) The positive rate of M. pneumoniae serological test in different ages from January 2017 to October 2022. Fig 2 Furthermore, the total number of M. pneumoniae RNA positive patients over 5 years old accounted for 49% of the total number of M. pneumoniae RNA positive between 0 - 18 y old from 2017 to 2021 (Fig. 1 B), but this proportion was not significant in M. pneumoniae serological tests (Fig. 2 B). In addition, after COVID - 19 pandemic, the positive number and positive rate of M. pneumoniae RNA and M. pneumoniae serological test decreased in < 1y, 1 - 3y, 3 - 5y and 5 - 18y age groups (Fig. 1 C, Fig. 1 D, Fig. 2 C and Fig. 2 D), especially in children over 5 years old, indicating that the COVID-19 pandemic reduced the infection of M. pneumoniae in school-age children. This change may be mainly related to a series of strict measures taken during the COVID-19 pandemic, such as suspension of classes (reduced contact between children), increased awareness of wearing masks and paying attention to hand hygiene. M. pneumoniae infection decreased in children of 0 - 18y during the COVID - 19 pandemic. It also shows that the COVID - 19 pandemic is something we must all face. The epidemic knows no borders, and the virus is the common enemy of mankind. The international community must foster the vision of a community with a shared future for mankind, help each other, jointly address risks and challenges, and jointly safeguard the well-being of people around the world. Therefore, we need to closely observe the epidemic changes of various pathogens affecting the respiratory system before and after the epidemic of COVID - 19. In short, M. pneumoniae infections in children of all ages have declined during the COVID - 19 pandemic. Close monitoring of epidemiological trends helps to prevent M. pneumoniae infection in children, especially in children over 5 years of age. Declaration of competing interest The authors declare no conflict of interests. Acknowledgements This work was funded by the National Natural Science Foundation of China (32201237), China Postdoctoral Science Foundation (2020M672301), Scientific and technological projects of Henan province (222102310270, 222102310109) ==== Refs References 1 Li Y. Guo Y. Duan Y. Changes in Streptococcus pneumoniae infection in children before and after the COVID-19 pandemic in Zhengzhou, China J Infect 85 3 2022 e80 e81 10.1016/j.jinf.2022.05.040 35659542 2 Zhou J. Zhao P. Nie M. Gao K. Yang J. Sun J. Changes of Haemophilus influenzae infection in children before and after the COVID-19 pandemic, Henan, China J Infect 2022 10.1016/j.jinf.2022.10.019 3 Li L. Song C. Li P. Li Y. Changes of Escherichia coli infection in children before and after the COVID-19 pandemic in Zhengzhou, China J Infect 2022 10.1016/j.jinf.2022.11.017 4 Kumar S. pneumoniae Mycoplasma A significant but underrated pathogen in paediatric community-acquired lower respiratory tract infections Indian J Med Res 147 1 2018 23 31 10.4103/ijmr.IJMR_1582_16 29749357 5 Feng S. Chen J.X. Zheng P. Zhang J.Z. Gao Z.J. Mao Y.Y. Status epilepticus associated with Mycoplasma pneumoniae encephalitis in children: good prognosis following early diagnosis and treatment Chin Med J (Engl) 132 12 2019 1494 1496 10.1097/cm9.0000000000000233 31205112 6 Wang M. Wang Y. Yan Y. Zhu C. Huang L. Shao X. Clinical and laboratory profiles of refractory Mycoplasma pneumoniae pneumonia in children Int J Infect Dis 29 2014 18 23 10.1016/j.ijid.2014.07.020 25449230 7 Tong L. Huang S. Zheng C. Zhang Y. Chen Z. Refractory Mycoplasma pneumoniae Pneumonia in Children: Early Recognition and Management J Clin Med 11 10 2022 10.3390/jcm11102824

0

PMC9753454

NO-CC CODE

2022-12-16 23:25:43

no

J Infect. 2022 Dec 15; doi: 10.1016/j.jinf.2022.12.015

utf-8

J Infect

10.1016/j.jinf.2022.12.015

oa_other

==== Front Arch Gerontol Geriatr Arch Gerontol Geriatr Archives of Gerontology and Geriatrics 0167-4943 1872-6976 The Author(s). Published by Elsevier B.V. S0167-4943(22)00294-1 10.1016/j.archger.2022.104907 104907 Article The impact of COVID-19 restrictions on older adults’ loneliness: Evidence from high-frequency panel data in Austria (2020-2022) Stolz Erwin ⁎ Mayerl Hannes Freidl Wolfgang Institute of Social Medicine and Epidemiology, Medical University of Graz, AUSTRIA ⁎ Corresponding author: Erwin Stolz, Institute of Social Medicine and Epidemiology, Medical University of Graz, AUSTRIA 15 12 2022 15 12 2022 1049074 8 2022 13 12 2022 14 12 2022 © 2022 The Author(s) 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Background It is unclear how strong and long lasting the effects of recurring COVID-19 pandemic restrictions are on older adults’ loneliness. Methods 457 retired older Austrians (60+) provided 9,489 repeated observations of loneliness across 30 survey waves of the Austrian Corona Panel Project between March 2020 and March 2022. Ordinal mixed regression models were used to estimate the effect of time-varying country-level strictness of COVID-19 restrictions (stringency index, range=0-100) on older adult's loneliness. Results The proportion of older adults who reported to be often lonely correlated (r=0.45) with the stringency index over time: both peaked during lock-downs (stringency index=82, often lonely=10%-13%) and were lowest during the summer of 2020 (stringency index=36, often lonely=4%-6%). Results from regression models adjusted for the number of new COVID-19 cases and deaths indicate, that when the stringency index increased by one point, the odds for loneliness increased by 2%. Older adults who lived alone were more likely lonely during the pandemic and were more affected by COVID-19 restriction measures compared to those living with others. Conclusions More stringent COVID-19 restrictions were associated with an increase in (situational) loneliness among older adults in Austria, and this effect was stronger among those who lived alone. Efforts should be made to enable older adults, in particular those who live alone, to allow for save in-person contact in case of (future) periods of strict pandemic restriction measures. Keywords loneliness older adults covid-19 pandemic restrictions lock-downs ==== Body pmc1 Introduction After more than two years, the COVID-19 pandemic still represents an acute global health threat, with older adults particularly at risk (Pijls, Jolani, and Atherley, 2021). To break waves of exponential COVID-19 infection rates and to avoid overburdening hospital care, many European governments had initially responded with an array of statutory containment and closure policies (Sharma, Mindermann, and Rogers-Smith, 2021). By restricting in-person social contacts, however, these public health interventions may have led to negative psychosocial side effects among older adults such as loneliness, which can be defined as a perceived discrepancy between the desired and one's existing social relationships (Peplau and Perlman, 1982). Already before the pandemic, older adults had an elevated risk for loneliness due to loss of partners and peers or health-problems (Pinquart and Sorensen, 2001, Savikko, Routasalo, and Tilvis, 2005, Yang and Victor, 2011, Luhmann and Hawkley, 2016, Dahlberg, McKee, and Frank, 2022), and loneliness has long been considered harmful to older adults’ physical and mental health (Courtin and Knapp, 2017, Heikkinen and Kauppinen, 2011, Luo, Hawkley, and Waite, 2012, Buchman, Boyle, and Wilson, 2010, Holt-Lunstad, Smith, and Baker, 2015, Jarach, Tettamanti, and Nobili, 2021, Holwerda, Deeg, and Beekman, 2014, Lee, Pearce, and Ajnakina, 2021). Therefore, loneliness that may have been induced by pandemic restrictions is a legitimate subject of concern. Whether and how strongly COVID-19 restrictions have affected older adults is not yet clear though. While some studies (Wong, Zhang, and Sit, 2020, Tilburg, Steinmetz, and Stolte, 2021, Luchetti, Lee, and Aschwanden, 2020, Macdonald and Hülür, 2021, Krendl and Perry, 2021, Stolz, Mayerl, and Freidl, 2021, Wester, Bovil, and Scheel-Hincke, 2022) suggest an increase in loneliness during the first wave of the pandemic compared to pre-pandemic times, others (Kivi, Hansson, and Bjälkebring, 2021, Röhr, Reininghaus, and Riedel-Heller, 2020, Peng and Roth, 2021, Hansen, Sevenius Nilsen, and Knapstad, 2021) found no changes. We know even less about how loneliness levels changed after the first lock-down. Several studies with multiple measurement points during the early pandemic (Luchetti, Lee, and Aschwanden, 2020, Stolz, Mayerl, and Freidl, 2021, Buecker, Horstmann, and Krasko, 2020, Kotwal, Holt-Lunstad, and Newmark, 2021) suggest that loneliness was higher during and immediately after the first lock-down, but levelled off thereafter. In contrast, a longitudinal study from the US and Canada (Lin, Horta, and Heald, 2022) found stable loneliness during the first five months of the pandemic, and a study from Norway (Hansen, Sevenius Nilsen, and Knapstad, 2021) found that while there were no changes between the time before the pandemic and it's first wave, loneliness increased strongly among older adults during the second wave in late 2020. Importantly, the stringency of pandemic restrictions has not been explicitly measured in most of these studies – although it's variation between countries and over time has been considerable. Furthermore, the effect of pandemic restriction measures on loneliness has not been differentiated from the pandemic threat level itself, i.e. the risk of COVID-19 infection and death, which may cause voluntary social distancing among older adults above and beyond instituted public health restriction measures. Living alone is a known risk factor for loneliness (Savikko, Routasalo, and Tilvis, 2005, Dahlberg, McKee, and Frank, 2022, Hajek and König, 2020, Fierloos, Tan, and Williams, 2021), and older adults who live alone could be particularly affected by COVID-19 related restrictions that limit in-person contacts with individuals from outside the household. Indeed, a few studies reported that during the early pandemic, increases in loneliness were higher among older adults who lived alone (Wong, Zhang, and Sit, 2020, Stolz, Mayerl, and Freidl, 2021, Atzendorf and Gruber, 2021) or were un-partnered (Tilburg, Steinmetz, and Stolte, 2021, Hansen, Sevenius Nilsen, and Knapstad, 2021), but it is again unclear whether these findings extend throughout the pandemic. In sum, we currently know rather little about later periods of the pandemic: It is unclear how strong and long lasting the effects of COVID-19 restrictions on older adults’ loneliness are, and whether those who live alone were particularly affected. Existing research on this topic is often hampered by retrospective survey questions, cross-sectional data, or, in case of longitudinal studies, by long intervals between few repeated measurement time points and by a lacking quantitative measures of the exposure, i.e. the pandemic-related restrictions. Hence, most studies are unable to establish whether a dose-response relationship with loneliness exists, whether this effect has changed over time, and whether older adults who live alone were particularly affected. In the current study, we attempt to improve upon these issues by exploiting high-frequency panel data from Austria, which allows monitoring older adults continuously throughout the entire pandemic in order to assess the real-time impact of the stringency of COVID-19 restrictions on loneliness. 2 Methods 2.1 Data For this study, we used data from the Austrian Corona Panel Project (Kittel, Kritzinger, and Boomgaarden, 2021), a high-frequency online panel survey of the general population in Austria with >1,500 participants per wave conducted by a professional survey agency (Market Agent, Baden, Austria) at the behest of the University of Vienna. 30 waves of online interviews – initially on a weekly basis, then approximately once per month – have been completed between March 27th 2020 and March 25th 2022. To participate, respondents had to be Austrian residents and aged 14 years or older. Respondents were quota sampled from a pre-existing certified (ISO 20252) online panel based on key demographics (age, gender, region, municipality size, educational level) closely mirroring the Austrian resident population. The initial participation rate was 35%, and the retention rates for panelists ranged from 86% in wave 2 to 48% in wave 30. For our current study on older adults’ loneliness, we used data from the 457 retired participants aged 60 years and over who provided a total of 9,489 repeated measurements, resulting in a median number of 23 (IQR=14) interviews per person. 2.2 Variables Loneliness was measured with the same single item in each wave: participants were asked how often they felt lonely during the last week. Possible answer categories included “never”, “on some days”, “multiple times a week”, “almost every day”, and “every day”. Single-item frequency measures of loneliness have been shown to correlate highly with established multiple-item scales (Dahlberg, McKee, and Frank, 2022, Mund, Maes, and Drewke, 2022) and to be reliable (Mund, Maes, and Drewke, 2022). Appropriately, the re-test reliability of the single-item measure of loneliness used in our current study – based on the auto-correlations over the first three weekly repeated measurement occasions ((r12∗r23)/r13) – was very good (=0.87). Due to the limited proportion of answers in the last three categories (3.7%, 2.1%, and 1.8%) and to ease interpretation, we collapsed these to “often lonely” for the subsequent analyses. Individual-level predictor variables included the time of interview since baseline (in weeks), and five time-invariant variables referring to March 2020: living alone (no/yes), age (in years), sex (male/female), high school education (no/yes), and having one or more of the following chronic disease(s): cardiovascular disease, diabetes, hepatitis B, chronic obstructive lung disease, chronic kidney disease or cancer (no/yes). 12 participants (2.6%) had missing values in these variables and were consequently excluded from analysis. To measure the stringency of pandemic-related restrictions, we used the COVID-19 Government Response Stringency Index (SI, (Hale, Angrist, and Goldszmidt, 2021)) as a time-varying, country-level predictor. The stringency index is a sum index based on nine ordinal measures (school closing, workplace closing, canceling of public events, restriction on gathering size, public transport closing, stay at home requirements, restrictions on internal movement, international travel control, and public information campaigns) that quantifies pandemic-related containment and closure policies on a daily basis, ranging from 0 (no restrictions) to 100 (maximum restrictions). Since the loneliness item refers to the last week before each interview, we calculated lagged 7-day smoothed values of the SI. As additional country-level and time-varying predictors, we included the number of new COVID-19 cases (per 1,000) and deaths (per 10) in Austria (lagged 7-day smoothed values; Source: OurWorldInData). 2.3 Statistical Analysis First, we analysed loneliness and the stringency of COVID-19 restrictions descriptively on the aggregate level by plotting the prevalence of categories of loneliness for each survey wave alongside the stringency index across time. We then calculated the Pearson correlation coefficient between the SI and the prevalence of loneliness categories across waves (n=30). Second, we modelled the impact of time-varying country-level SI on repeatedly measured individual-level loneliness using generalized linear mixed regression models. Specifically, we fitted cumulative ordinal logistic regression models (Bürkner and Vuorre, 2019) to handle the three non-equidistant response categories (never lonely, sometimes lonely, and often lonely) as well as the positive skew of the outcome (Liddell and Kruschke, 2018). In the first model, we included time in weeks as a linear predictor to see whether loneliness among older adults has increased during the pandemic. We also included random intercept and slope (week) terms to account for repeated observations nested within respondents. Our core interest, however, was in the overall effect of the time-varying stringency of COVID-19-related restrictions on loneliness. To differentiate the specific effect of restriction measures from the course of the pandemic and its current threat level we adjusted for the number of new COVID-19 cases and deaths. We also wanted to see whether the impact of the SI has changed over the course of the pandemic, which was tested by adding an interaction effect between SI and time (model 2). Finally, we were interested to see whether the effect of SI was moderated by living alone, which we tested by adding an interaction effect between these two variables (model 3). Since the relationship between living alone and loneliness might be confounded by socio-demographics and health (Steptoe, Shankar, and Demakakos, 2013, Cudjoe, Roth, and Szanton, 2020), we also adjusted for these in model 3. For all analyses, we applied demographic weights. All analyses were conducted in R (v4.1.3). Bayesian mixed ordinal regression models were estimated with R-package brms (v2.16.3) (Bürkner, 2017), a front-end for RStan (v2.21.3). Plots were created with R-package ggplot2 (v.3.3.5). 3 Results At baseline (March 2020), the median age of the sample was 69 (IQR=8, range=60-85) years, 57.0% were women, 16.6% had completed high school education, 42.6% had one or more chronic disease(s), and 33.5% lived alone before the onset of the pandemic. The SI varied considerably during the 2-year pandemic period in Austria (grey background in Figure 1 ): restrictions peaked (SI=82) during the first three lock-down periods (March-April 2020, November-December 2020, January 2021) and were lowest during the summer months of 2020 (SI=36) and 2021 (SI=49). New COVID-19 infections remained low in Austria during the first lockdown (<800 cases in March 2020), increased before the second lock-down (∼8,000 cases in November 2020), and peaked toward the end of the observation period (>45,000 cases in March 2022). New COVID-19 deaths remained low in Austria during the initial wave of the pandemic (<20 deaths in April and May 2020) and peaked in late 2020 (∼130 deaths) as well as in late 2021 (∼60 deaths).Figure 1 Prevalence of loneliness (March 2020 - March 2022) n=457, weighted data. Figure shows trajectories of the prevalence of the categories of loneliness (black lines) in % by wave. Answer categories ’multiple times a week’, ’almost every day’, and ’every day’ were summarised as ’often lonely’ for easier presentation and interpretation. Grey background shows the stringency index (SI). Figure 1: Overall, most participants reported to be ‘never’ lonely (73%), or to be lonely only on ‘some days’ (19.4%). On average, 7.6% reported to be often lonely during follow-up. However, the prevalence of feeling often lonely varied considerably over the course of the pandemic (Figure 1). The proportion of older adults who felt often lonely reached its maximum of 10%-12% during the first (March/April 2020) and third (January 2021) lock-down. During the summer months of 2020, when restrictions were lowest, only 4-6% reported to be often lonely. The correlation coefficient between the SI and the prevalence of feeling often lonely across survey waves was r=0.43, but this aggregate-level estimate comes with considerable uncertainty (95%CI=0.09-0.69) as there are just 30 repeated survey waves available. This association, however, varied by living status (Figure 2 ): Older adults who lived alone were not only more likely to report feeling often lonely – maximal proportion was 21-23% during the first lock-down – but their loneliness was also more closely tethered to the SI (r=0.48, 95%-CI=0.14-0.71) compared to those who lived together with others (r=0.14, (95%-CI=-0.23-0.48) who's peak prevalence of loneliness was 11% during the third lockdown.Figure 2 Prevalence of loneliness by living status (March 2020 - February 2022) Weighted data. Figure shows trajectories of the prevalence in % of categories of loneliness (black lines) for those who lived alone (plot A, n = 163) and for those who lived with others (plot B, n = 294). Grey background shows the SI. Figure 2: Results from the regression models (Table 1 ) based on individual-level data show no (linear) overall increase in loneliness across the two-year period, but even a slight decrease. Both new COVID-19 cases (per 1000) and new COVID-19 deaths (per 10) were associated with increased loneliness. Looking at the effect of the SI, we see that a one-point increase (total range in Austria 2020-2022 was 46 points) was associated with a 2%-increase in the odds of being more lonely (model 1). Specifically, the estimated probability to feel lonely ‘on some days’ increased from 5% when the stringency of COVID-19 restrictions was lowest to 13% when the restrictions were strictest (Figure 3 ). For the probability of being often lonely, the maximum difference in the restrictions’ stringency during the observed period translates to an associated estimated maximum change from 0.2% to 0.7%. Overall, the difference between minimum and maximum restriction level translates to a two- to three-fold increase in the probability of feeling lonely. From the results of the second model, it can be seen that the impact of the SI did not change over time. From the third model, we can see that living alone is a strong risk factor for loneliness during the COVID-19 pandemic: older adults who lived alone had a more than 5-times higher chance to be lonely compared to those who lived with others.Table 1 Results from ordinal mixed regression models Table 1: Model 1 Model 2 Model 3 FIXED EFFECTS Treshold 1 3.57 (3.07, 4.07) 3.39 (2.78, 4.01) 3.82 (-0.31, 7.83) Treshold 2 6.71 (6.18, 7.23) 6.53 (5.90, 7.18) 6.96 (2.84, 10.99) Week 0.99 (0.98, 0.99) 0.98 (0.96, 1.00) 0.99 (0.98, 0.99) New COVID-19 cases (per 1000) 1.02 (1.01, 1.03) 1.02 (1.01, 1.04) 1.02 (1.01, 1.03) New COVID-19 deaths (per 10) 1.04 (1.00, 1.07) 1.03 (1.00, 1.07) 1.03 (1.00, 1.06) Stringency index (SI) 1.02 (1.02, 1.03) 1.02 (1.01, 1.03) 1.02 (1.01, 1.03) Stringency index (SI) * week - 1.00 (1.00, 1.00) - Living alone - - 5.60 (2.37, 13.51) Stringency Index (SI) * Living alone - - 1.01 (1.00, 1.02) RANDOM EFFECTS SD Intercept 3.21 (2.89, 3.56) 3.21 (2.89, 3.56) 2.94 (2.64, 3.26) SD Time 0.03 (0.02, 0.03) 0.03 (0.02, 0.03) 0.03 (0.02, 0.03) Corr. Intercept * Time 0.07 (-0.14, 0.27) 0.07 (-0.14, 0.27) 0.10 (-0.10, 0.30) Number of participants = 457, number of repeated observations of loneliness = 9,732. All models adjusted for the number of new COVID-19 cases and new COVID-19 deaths. Model 3 was additionally adjusted for age, sex, education, and chronic diseases. Coefficients are exponentiated log odds, i.e. odds ratios, except for the two thresholds, which act as intercepts in ordinal models. Numbers in parentheses are exponentiated 95% credible intervals. Model fit with Hamiltonian Monte Carlo (HMC) sampling procedure with 3 chains per imputed dataset and 1000 post-warmup iterations per chain. All rˆ-values <1.01. Figure 3 Predicted probability of categories of loneliness by stringency of COVID19 restrictions Results based on ordinal mixed regression models (first row = model 1, second row = model 3) based on weighted data. Note that the y-axis is not identical between the first two columns and the third column. Solid lines based on model 1 are predicted probability estimates based on population-level estimates at mean point during follow-up, and for the mean value of new COVID-19 cases and deaths. Solid lines based on model 3 furthermore refer to women of mean age, without high school-level education, and with chronic diseases. Dashed lines are 95% credible intervals. Figure 3: Also, the effect of the SI seems to be moderated by the living arrangement: older adults who lived alone where somewhat more likely to often feel lonely as pandemic-related restriction measures increased compared to those who lived with others. Specifically, the estimated probability of feeling often lonely more than tripled from the lowest to the highest level of SI among those who lived alone compared to a doubling among those who lived together with partner, children or other household members (Figure 3). 4 Discussion In this paper, we monitored loneliness among retired older adults 60+ in Austria throughout two years of the COVID-19 pandemic based on high-frequency panel data in order to assess the real-time impact of COVID-19 restrictions. In summary, we found that loneliness fluctuated considerably during the pandemic and that these fluctuations are likely in part due to COVID restriction measures: when pandemic restrictions became more stringent, the prevalence of loneliness also increased. Soon after restrictions were loosened, however, the prevalence of loneliness tended to decrease again. The findings of our study are compatible with evidence that loneliness among older adults was higher during the first lock-down in the early pandemic compared to pre-pandemic times (Wong, Zhang, and Sit, 2020– (Wester, Bovil, and Scheel-Hincke, 2022). Our study extends these findings for later periods of the pandemic in Austria, including multiple subsequent lock-downs and differentiating restriction measures from the effects of pandemic threat level, i.e. from COVID-19 infection and death rates. Unlike a study from the US and Canada (Lin, Horta, and Heald, 2022), we detected wave-like patterns of loneliness largely congruent with pandemic restrictions, and unlike a study from Norway (Hansen, Sevenius Nilsen, and Knapstad, 2021), we found no evidence for strong differences in the effect of restrictions on loneliness over time. Instead, we confirmed findings of longitudinal studies from the early pandemic (Luchetti, Lee, and Aschwanden, 2020, Stolz, Mayerl, and Freidl, 2021, Buecker, Horstmann, and Krasko, 2020, Kotwal, Holt-Lunstad, and Newmark, 2021), insofar that increases in loneliness due to lock-downs seem rather short-lived and reverted after restrictions were lifted. We found no evidence in our study that loneliness generally increased or that it chronified over the course of the COVID-19 pandemic. Therefore, restriction-induced loneliness appears mostly situational. In comparison to chronic loneliness, situational loneliness has been found to be less of a risk factor for negative long-term health outcomes (Martín-María, Caballero, and Lara, 2021, Shiovitz-Ezra and Ayalon, 2010). Even so, pandemic-induced loneliness among older adults may already have had negative mental health consequences such as increased depression and anxiety symptoms (Krendl and Perry, 2021, Mayerl, Stolz, and Freidl, 2021), which could be addressed by increased screening for current mental illness – particularly among those with pre-existing mental health problems. The results from our models imply that the pandemic threat level, i.e. new COVID-19 cases and deaths, had an effect on older adults’ loneliness above and beyond the stringency of pandemic-related containment and closure policies introduced by law. This is compatible with evidence from studies (Choi, Farina, and Wu, 2021, Cohn-Schwartz, Vitman-Schorr, and Khalaila, 2021) showing that older adults who voluntarily engaged in physical distancing measures, e.g. avoiding close contact or cancelling social activities, also reported higher levels of loneliness. With regard to the strictness of COVID-19 related restriction measures, we are aware of only two other studies (Wester, Bovil, and Scheel-Hincke, 2022, Atzendorf and Gruber, 2021), both based on the Survey of Health, Ageing, and Retirement in Europe (SHARE), that explicitly used the SI as a fine-grained measure of pandemic restrictions and that assessed it as a determinant of older adults’ loneliness. While one study (Atzendorf and Gruber, 2021) found no association between the number of days with stringent restriction measures (SI>60) across countries and self-reported changes in loneliness, the other (Wester, Bovil, and Scheel-Hincke, 2022) reported an increased risk for loneliness in countries with high SI, and a particularly high increase between pre-pandemic loneliness and loneliness during the summer 2020 for Austria.The latter results (Wester, Bovil, and Scheel-Hincke, 2022) are likely more robust – as loneliness was measured in the same individuals before and during the pandemic rather than to rely on retrospective questions (Atzendorf and Gruber, 2021) – and are compatible with our results. Our study implies an increase in loneliness as restriction measures became more stringent among older adults living alone, who account for 20%-40% of the total older population aged 65+ in Europe and the USA (Cudjoe, Roth, and Szanton, 2020, Tomassini, Glaser, and Wolf, 2004). These results are in line with findings from the early pandemic, that older adults who lived alone had fewer in-person contacts and provided or received less help from others (Fingerman, Ng, and Zhang, 2021), and that pandemic-related increases in loneliness were higher among them (Wong, Zhang, and Sit, 2020, Stolz, Mayerl, and Freidl, 2021, Atzendorf and Gruber, 2021), as well as among those who were un-partnered (Tilburg, Steinmetz, and Stolte, 2021, Hansen, Sevenius Nilsen, and Knapstad, 2021). Therefore, efforts should be made to specifically enable older adults who live alone to have save forms of in-person contact during (future) lock-down periods (Fingerman, Ng, and Zhang, 2021) in order to stay socially connected. In this vein, regular video and phone calls and other electronic contacts have been suggested as a partial remedy (Choi, Farina, and Wu, 2021, Cohn-Schwartz, Vitman-Schorr, and Khalaila, 2021). Next to a number of strengths (high-frequency panel data, fine-grained longitudinal measure of COVID-19-related restrictions over time, adjustment for COVID-19 infection and death rates), there are also two noteworthy limitations to this study. First, we lacked a pre-pandemic baseline rate of loneliness (without any COVID-19 restrictions) for comparison to assess the overall impact of the pandemic. Second, it is unlikely that the quota sample of the current study is truly representative for the population of the general older population in Austria with regard to loneliness – a problem that plagues many studies on this topic (Dahlberg, 2021). Older adults, particularly the oldest old and institutionalized individuals, and those with a low level of education – all of which are more likely to be lonely (Dahlberg, McKee, and Frank, 2022, Steptoe, Shankar, and Demakakos, 2013, Cudjoe, Roth, and Szanton, 2020), particularly during the current pandemic – are difficult to recruit for online interviews (Kelfve, Kivi, and Johansson, 2020). Despite the use of demographic weights, these limitations likely resulted in an underestimation of the prevalence of loneliness, which may, in consequence, also down-bias our effect estimates on loneliness. In other words, we expect the true effect of COVID-19 restriction measures on older adult's loneliness to be even higher than what we were able to document in the current study. 5 Conclusion In this study, we monitored older adults’ (60+) loneliness throughout the pandemic (March 2020-March 2022). Adjusting for the pandemic threat level, we found that increases in pandemic restrictions were associated with an increase in (situational) loneliness among older adults in Austria, particularly among those who lived alone. Declarations Ethics approval and consent to participate The Austrian Corona Panel Project is a social science survey study conducted by a professional survey agency (Market Agent, Baden, Vienna) at the behest of the University of Vienna. An ethical statement was deemed not necessary by the University of Vienna as no patients were examined and no invasive methods used, and there were no risks for survey participants. Survey participants were recruited from an existing and certified (ISO 20252) online panel, which means that the survey agency applies high standards of data security and that the panel is used only for research purposes. Participants were informed about the topic of the survey study, and provided informed consent. Participation in the survey study was voluntary and participants received a small amount of money (1.80€) for each wave they participated in. The current study is a secondary data analysis of the Austrian Corona Panel Project data. Availability of data and materials The dataset supporting the conclusions of this article is available in the Austrian Social Science Data Archive (AUSSDA): https://data.aussda.at/dataset.xhtml?persistentId=doi:10.11587/28KQNS. The R-Markdown code reproducing all analyses, results and this manuscript is also available online via the OSF repository (https://osf.io/aq97s/). Funding The authors received no specific funding for conducting this study. The data collection for the ACPP has been made possible by COVID-19 Rapid Response Grant EI-COV20-006 of the Wiener Wissenschafts- und Technologiefonds (WWTF) and financial support by the rectorate of the University of Vienna. Further funding by the Austrian Social Survey (SSÖ), the Vienna Chamber of Labour (Arbeiterkammer Wien), and the Federation of Austrian Industries (Industriellenvereinigung) is gratefully acknowledged. From October 2020, ACPP continues as a research project funded by the Austrian Science Fund (Grant P33907). Credit author statement Erwin Stolz is the corresponding author, he planned the study, performed all statistical analysis, and wrote the article. Hannes Mayerl contributed to the interpretation of results and critically reviewed the manuscript. Wolfgang Freidl also critically reviewed the manuscript. Conflict of Interest None declared. ==== Refs References Atzendorf J Gruber S. Depression and loneliness of older adults in Europe and Israel after the first wave of Covid-19 Eur J Ageing 2021 10.1007/s10433-021-00640-8 Buchman AS Boyle PA Wilson RS Loneliness and the rate of motor decline in old age: The rush memory and aging project, a community-based cohort study BMC Geriatrics 10 2010 77 20969786 Buecker S Horstmann KT Krasko J Changes in daily loneliness for German residents during the first four weeks of the COVID-19 pandemic Soc Sci Med 265 2020 113541 Bürkner P-C Vuorre M. Ordinal regression models in psychology: A tutorial Advances in Methods and Practices in Psychological Science 2 2019 77 101 Bürkner P-C. Brms: An r package for bayesian multilevel models using stan Journal of Statistical Software 80 2017 1 28 Choi EY Farina MP Wu Q COVID-19 social distancing measures and loneliness among older adults J Geron Psy Sci Soc Sci 2021 gbab009 Cohn-Schwartz E Vitman-Schorr A Khalaila R. Physical distancing is related to fewer electronic and in-person contacts and to increased loneliness during the COVID-19 pandemic among older Europeans Qual Life Res 2021 10.1007/s11136-021-02949-4 Courtin E Knapp M. Social isolation, loneliness and health in old age: A scoping review Health Soc Care Community 25 2017 799 812 26712585 Cudjoe TKM Roth DL Szanton SL The epidemiology of social isolation: National health and aging trends study J Gerontol B Psy Sci Soc Sci 75 2020 107 113 Dahlberg L McKee KJ Frank A A systematic review of longitudinal risk factors for loneliness in older adults Aging & Mental Health 26 2022 225 249 33563024 Dahlberg L. Loneliness during the COVID-19 pandemic Aging & Mental Health 25 2021 1161 1164 33491474 Fierloos IN Tan SS Williams G Socio-demographic characteristics associated with emotional and social loneliness among older adults BMC Geriatrics 21 2021 114 33563228 Fingerman KL Ng YT Zhang S Living alone during COVID-19: Social contact and emotional well-being among older adults J Gerontol B Psy Sci Soc Sci 76 2021 e116 e121 Hajek A König H-H. Which factors contribute to loneliness among older Europeans? Findings from the Survey of Health, Ageing and Retirement in Europe: Determinants of loneliness Arch Gerontol Geriatr 89 2020 104080 Hale T Angrist N Goldszmidt R A global panel database of pandemic policies (oxford COVID-19 government response tracker) Nat Hum Behav 5 2021 529 538 33686204 Hansen T Sevenius Nilsen T Knapstad M Covid-fatigued? A longitudinal study of Norwegian older adults’ psychosocial well-being before and during early and later stages of the COVID-19 pandemic Eur J Ageing 2021 10.1007/s10433-021-00648-0 Heikkinen R-L Kauppinen M. Mental well-being: A 16-year follow-up among older residents in Jyväskylä Arch Gerontol Geriatr 52 2011 33 39 20207429 Holt-Lunstad J Smith TB Baker M Loneliness and social isolation as risk factors for mortality: A meta-analytic review Perspect Psychol Sci 10 2015 227 237 25910392 Holwerda TJ Deeg DJH Beekman ATF Feelings of loneliness, but not social isolation, predict dementia onset: Results from the Amsterdam study of the elderly (AMSTEL) J Neurol Neurosurg Psychiatry 85 2014 135 142 23232034 Jarach CM Tettamanti M Nobili A Social isolation and loneliness as related to progression and reversion of frailty in the survey of health aging retirement in Europe (SHARE) Age Ageing 50 2021 258 262 32915990 Kelfve S Kivi M Johansson B Going web or staying paper? The use of web-surveys among older people BMC Medical Research Methodology 20 2020 252 33032531 Kittel B Kritzinger S Boomgaarden H The Austrian corona panel project: Monitoring individual and societal dynamics amidst the COVID-19 crisis Eur Polit Sci 20 2021 318 344 Kivi M Hansson I Bjälkebring P. Up and about: Older adults’ well-being during the COVID-19 pandemic in a Swedish longitudinal study J Gerontol B Psy Sci Soc Sci 76 2021 e4 e9 Kotwal AA Holt-Lunstad J Newmark RL Social isolation and loneliness among San Francisco bay area older adults during the COVID-19 shelter-in-place orders J Am Geriatr Soc 69 2021 20 29 32965024 Krendl AC Perry BL. The impact of sheltering in place during the COVID-19 pandemic on older adults’ social and mental well-being J Gerontol B Psy Sci Soc Sci 76 2021 e53 e58 Lee SL Pearce E Ajnakina O The association between loneliness and depressive symptoms among adults aged 50 years and older: A 12-year population-based cohort study The Lancet Psychiatry 8 2021 48 57 33181096 Liddell TM Kruschke JK. Analyzing ordinal data with metric models: What could possibly go wrong? Journal of Experimental Social Psychology 79 2018 328 348 Lin T Horta M Heald K Loneliness progression among older adults during the early phase of the COVID-19 pandemic in the United States and Canada J Gerontol B Psy Sci Soc Sci 77 2022 e23 e29 Luchetti M Lee JH Aschwanden D The trajectory of loneliness in response to COVID-19 Am Psychol 75 2020 897 908 32567879 Luhmann M Hawkley LC. Age differences in loneliness from late adolescence to oldest old age Dev Psychol 52 2016 943 959 27148782 Luo Y Hawkley LC Waite LJ Loneliness, health, and mortality in old age: A national longitudinal study Social Science & Medicine 74 2012 907 914 22326307 Macdonald B Hülür G. Well-being and loneliness in Swiss older adults during the COVID-19 pandemic: The role of social relationships The Gerontologist 61 2021 240 250 33258898 Martín-María N Caballero FF Lara E Effects of transient and chronic loneliness on major depression in older adults: A longitudinal study Int J Geriatr Psychiatr 36 2021 76 85 Mayerl H Stolz E Freidl W. Longitudinal effects of COVID-19-related loneliness on symptoms of mental distress among older adults in Austria Public Health 200 2021 56 58 34678551 Mund M Maes M Drewke PM Would the real loneliness please stand up? The validity of loneliness scores and the reliability of single-item scores Assessment 2022 10731911221077227 Peng S Roth AR Social isolation and loneliness before and during the COVID-19 pandemic: A longitudinal study of u.s. Adults older than 50 J Gerontol B Psy Sci Soc Sci 2021 10.1093/geronb/gbab068 Peplau LA Perlman D. Perspectives on loneliness. Loneliness. A Sourcebook of Current Theory, Research and Therapy 1982 John Wiley & Sons 1 8 Pijls BG Jolani S Atherley A Demographic risk factors for COVID-19 infection, severity, ICU admission and death: A meta-analysis of 59 studies BMJ Open 11 2021 e044640 Pinquart M Sorensen S. Influences on loneliness in older adults: A meta-analysis Basic and Applied Social Psychology 23 2001 245 266 Röhr S Reininghaus U Riedel-Heller SG. Mental wellbeing in the German old age population largely unaltered during COVID-19 lockdown: Results of a representative survey BMC Geriatrics 20 2020 489 33225912 Savikko N Routasalo P Tilvis RS Predictors and subjective causes of loneliness in an aged population Arch Gerontol Geriatr 41 2005 223 233 15908025 Sharma M Mindermann S Rogers-Smith C Understanding the effectiveness of government interventions against the resurgence of COVID-19 in Europe Nat Commun 12 2021 5820 34611158 Shiovitz-Ezra S Ayalon L. Situational versus chronic loneliness as risk factors for all-cause mortality International Psychogeriatrics 22 2010 455 462 20003631 Steptoe A Shankar A Demakakos P Social isolation, loneliness, and all-cause mortality in older men and women PNAS 110 2013 5797 5801 23530191 Stolz E Mayerl H Freidl W. The impact of COVID-19 restriction measures on loneliness among older adults in Austria European Journal of Public Health 31 2021 44 49 33338225 Tilburg TG van Steinmetz S Stolte E Loneliness and mental health during the COVID-19 pandemic: A study among Dutch older adults J Gerontol B Psy Sci Soc Sci 76 2021 e249 e255 Tomassini C Glaser K Wolf DA Living arrangements among older people: An overview of trends in Europe and the USA Popul Trends 2004 24 34 15192891 Wester CT Bovil T Scheel-Hincke LL Longirudinal changes in mental health following the COVID-19 lockdown: Results from the Survey of Health, Ageing, and Retirement in Europe Ann Epidem 74 2022 21 30 Wong SYS Zhang D Sit RWS Impact of COVID-19 on loneliness, mental health, and health service utilisation: A prospective cohort study of older adults with multimorbidity in primary care Br J Gen Pract 70 2020 e817 e824 32988955 Yang K Victor C. Age and loneliness in 25 European nations Ageing & Society 31 2011 1368 1388

0

PMC9753455

NO-CC CODE

2022-12-16 23:25:43

no

Arch Gerontol Geriatr. 2022 Dec 15;:104907

utf-8

Arch Gerontol Geriatr

10.1016/j.archger.2022.104907

oa_other

==== Front Vaccine X Vaccine X Vaccine: X 2590-1362 Published by Elsevier Ltd. S2590-1362(22)00108-5 10.1016/j.jvacx.2022.100248 100248 Article COVID-19 vaccination and the skin to deltoid MUSCLE distance in adults with diabetes Doppen Marjan a⁎ Mirjalili Ali b Harwood Matire ac Eathorne Allie a Braithwaite Irene a Bong Jonathan d Kirton Louis ad Semprini Ruth ad Weatherall Mark e Semprini Alex a Kearns Ciléin a Black Melissa a Kung Stacey a Walton Michaela a Beasley Richard ad Hills Thomas af a Medical Research Institute of New Zealand, Medical Research Institute of New Zealand, Private Bag 7902, Wellington 6242, New Zealand b University of Auckland, Building 409, 24 Symonds Street, City Campus, Auckland, New Zealand c Papakura Marae Health Clinic, 29 Hunua Road, Auckland 2110, New Zealand d Capital and Coast District Health Board, 69 Riddiford Street, Newtown, Wellington 6021, New Zealand e University of Otago Wellington, 23A Mein Street, Newtown, Wellington 6242, New Zealand f Auckland District Health Board, Level 7/214 Green Lane West, Greenlane, Auckland 1051, New Zealand ⁎ Corresponding author at: Medical Research Institute of New Zealand, Private Bag 7902, Wellington 6242, New Zealand. 15 12 2022 15 12 2022 10024810 3 2022 2 10 2022 14 12 2022 © 2022 Published by Elsevier Ltd. 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Objectives To estimate the proportion of adult diabetics with a skin to deltoid muscle distance (SDMD) of >25 mm, representing a distance greater than the standard needle length used for intramuscular COVID-19 vaccination, and to assess whether anthropometric measurements predict ultrasound SDMD measurements. Design: Non-interventional cross-sectional study. Setting: Single site, non-clinical setting, Wellington, New Zealand. Participants: One hundred participants (50 females) aged at least 18 years diagnosis with diabetes. All participants completed the study. Main outcome measures The proportions of participants with a SDMD >25mm and a SDMD >20mm (indicating that the needle would not have penetrated at least 5mm into the deltoid, which is considered necessary to ensure deposition of vaccine into muscle); the relationship between anthropometric measurements (body weight, body height, body mass index (BMI), skinfold thickness, arm circumference) and SDMD measured by ultrasound. Results The proportion (95%CI) of participants with a SDMD >25mm was 6/100; 6% (2.2 to 12.6), and the proportion with a SDMD >20mm was 11% (5.6 to 18.8), of which 9/11 had a BMI ≥30kg/m2 and 9/11 were female. The strongest relationships between anthropometric measurements and SDMD were with arm circumference (r=0.76, P<0.001) and BMI (r=0.73, P<0.001). Arm circumference and BMI were the best predictors of SDMD measurements with AUC for ROC curves of 0.99 and 0.94 above the 25 mm cut point, 0.97 and 0.89 above the 20 mm cut point respectively. Conclusions The standard needle length of 25mm is likely to be insufficient to ensure deposition of COVID-19 vaccine within the deltoid muscle in a small but important proportion of obese adults with diabetes. Arm circumference and BMI are simple measurements that could identify those that need a long needle to ensure successful intramuscular vaccine administration. Funding Ruth Maud Ring Spencer Estate; Health Research Council of New Zealand (Independent Research Organisation). Keywords COVID-19 deltoid diabetes needle length vaccination ==== Body pmcIntroduction Diabetes is associated with an increased risk of morbidity and mortality due to COVID-19.[1], [2], [3], [4] Vaccinating people with diabetes against coronavirus is a priority to reduce the burden of disease, risk of hospitalisation and death. The COVID-19 pandemic has seen the development of vaccines that use novel technologies that are administered by intramuscular injection. This route of delivery may be particularly important for mRNA vaccines that use lipid nanoparticle technology, delivered by intramuscular injection, to facilitate the expression of the SARS-CoV-2 spike glycoprotein from muscle cells.[5] It is unclear whether sufficient spike glycoprotein can be produced if the mRNA vaccine is injected into the subcutaneous tissue. This is in contrast to other vaccine technologies, which retain their immunogenicity when delivered via subcutaneous injection.[6] Importantly, approved mRNA COVID-19 vaccines are only licensed for intramuscular injection and the deltoid muscle is the recommended injection site.[7], [8] Diabetes, and in particular Type 2 diabetes, is commonly associated with obesity. There is evidence that obesity may reduce the likelihood of successful injection into the deltoid muscle, due to a larger fat pad thickness at the injection site.[9], [10], [11] It is unclear how vaccine administration with a needle that does not reach the deltoid muscle affects the immunogenicity of COVID-19 mRNA vaccines. For other vaccinations, deposition into subcutaneous fat may slow mobilisation and processing of antigens, which in turn may lead to vaccine failure.[12], [13] Suboptimal vaccine delivery also results in a greater risk of adverse events including local reactions, localised cellulitis, granuloma formation and abscesses.[10], [12], [14] The standard needle length used for the COVID-19 vaccination in New Zealand is 25mm. A longer 38mm needle was originally recommended for ‘larger patients’,[15] and subsequently also for those with ‘a larger arm’.[16] Similarly, in the United Kingdom (UK), ‘The Green Book’ recommends that in larger adults, a longer length (e.g. 38mm) may be required, and an individual assessment as to the length of the needle should be made.[17] It is unclear at what point an adult, or their arm, is large enough that the 38 mm needle would be required. Clear, practical, and evidence-based guidance on how to select the optimal needle length for people receiving an intramuscular COVID-19 vaccine is needed. This is an important issue, particularly for obese individuals and those with diabetes, who are at risk of worse health outcomes should they develop COVID-19.[18], [19] The objective of this study was to estimate the proportion of adults with diabetes with a skin to deltoid muscle distance (SDMD) ultrasound measurement >25mm, at the recommended COVID-19 vaccination site in the non-dominant arm. We also assessed whether anthropometric measurements might predict ultrasound measurements of SDMD, and thereby guide selection of the optimal needle length for vaccine delivery into the deltoid muscle. Methods Study design This was a single site non-interventional cross-sectional study conducted at the Medical Research Institute of New Zealand (MRINZ) in Wellington, New Zealand. The original intent was to undertake the study at the MRINZ-affiliated Papakura Marae Health Centre in Auckland, however this was not possible due to a prolonged government-mandated lockdown of the Auckland region due to COVID-19. All investigations were completed in a single visit of approximately 30 minutes after providing signed informed consent. This study was approved by the Auckland Health Research Ethics Committee (Ref. AH23130) on 29 September 2021. Participants Participants were eligible, regardless of COVID-19 vaccination status, if aged 18 years or older, diagnosed with diabetes of any type and able and willing to provide informed consent prior to participation. Recruitment took place by direct invitation of potential participants on the MRINZ database, through local and national patient organisations, by advertisement on social media, and the MRINZ website. Recruitment continued until 100 participants were enrolled. No stopping criteria applied, provided participants did not withdraw consent before completion of the study. Methods of measurement In addition to date of birth, as part of the eligibility review, participants provided confirmation of diagnosis of diabetes: for which a prescription of diabetes medication or a clinical record was accepted. If requested by the participant, a study investigator contacted their clinical health care provider to obtain confirmation of diagnosis after consent was given. Once enrolled in the study, the following demographic and clinical data were obtained: sex, ethnicity, COVID-19 vaccination status (unvaccinated, partially vaccinated, fully vaccinated), side of non-dominant arm, diabetes treatment regimen and comorbidities. Anthropometric measurements Body height was measured by using a calibrated stadiometer. Body weight and body fat were measured by using the B-587 Body composition Monitor (Tanita, Japan). For accuracy, ‘athletic mode’ was chosen for athletic participants as per the user manual. Derived BMI was calculated. Arm measurements Participants were instructed to expose their non-dominant arm and hang it relaxed by their side. The protocol stated the site for measurements was the deltoid intramuscular vaccination site at the intersection of a line connecting the acromion process and deltoid tuberosity, at the level of the axilla, as recommended in New Zealand guidelines.[20] However, when operationalised, the exact midpoint between the acromion process and deltoid tuberosity was used, as recommended in Australian immunisation guidelines.[21] This latter site was marked with an indelible pen, checked by a second study investigator and corrected if needed. Two consecutive measurements of arm circumference were performed and three measurements of skinfold thickness were taken with a skinfold calliper at the marked recommended injection site. For each participant, three ultrasound images displaying the skin, subcutaneous tissue and fascia and the deltoid muscle were captured and saved using a high-frequency (13-6 MHz) linear transducer (Sonosite X-Porte, Fujifilm, Japan), after using sufficient water-soluble ultrasound transmission gel as an acoustic standoff. The middle of the ultrasound probe was placed at the marked injection site, with minimal pressure, at a 90-degree angle with the skin, in the sagittal plane and a penetration depth of 3.4 cm. Penetration depth setting was increased as required to ensure a sufficient volume of the deltoid muscle was displayed. Ultrasound images were obtained by trained clinical staff (LK, RS, SK). Measurements of the distance (in mm, to the nearest whole mm) between the skin and the fascia of the deltoid muscle were performed by a radiology registrar (JB). Sample size The sample size of 100 was chosen to give a 95% CI for a proportion of plus or minus 10%. Statistical analysis For arm circumference, skinfold thickness and ultrasound measurements, the mean of the repeated measurements was used. Data descriptions used mean and standard deviation (SD); median, 25th and 75th percentiles (interquartile range); and minimum to maximum, for continuous variables; and counts and proportions expressed as percentages for categorical variables. Frequency histograms were used to show the distribution of the continuous variables. LOESS plots showed the relationship between ultrasound measured SDMD and possible predictor variables and the relationships were summarised by linear regression together with R-squared values and correlation coefficients. Estimates of proportions used an exact binomial method. Discrimination for continuous variables for 25mm and 20mm ultrasound SDMD used logistic regression, summarised by the Area under the Curve (AUC) for the Receiver Operating Characteristic (ROC) Curve, and illustrative sensitivity, specificity, and likelihood ratio positive, at various cut-points. These illustrative cut-points were chosen initially for 25mm distance in relation to the 100% sensitivity cut-point and two readings with a greater value (apart from height using lesser values); and by the same criteria for the 20mm distance but with the addition of the same values chosen for the 25mm illustrative cut-points. For skinfold thickness the same illustrative cut-points were used. The estimate of the intra-class correlation coefficient in relation to SDMD for the three investigators was by a mixed linear model estimating the variance components. In a post hoc analysis, simple data descriptions were shown for the participant characteristics and anthropometric variables by sex, and mean values compared by t-tests. SAS (version 9.4, Cary, NC) was used for all statistical analyses. Results There were 50 female and 50 male (n=100) participants with a mean (SD) age of 61 (14.6) years. There was no missing data apart from one participant who was wheel chair dependent in whom body fat percentage could not be recorded. In all participants the diagnosis of diabetes was confirmed by MB or MD as part of eligibility check before being enrolled into the study. Demographic and anthropometric data are shown in Table 1 . Mean (SD) BMI was 30.9 (7.3) kg/m2, mean (SD) body fat was 34.7 (12.4) %, and the mean (SD) SDMD was 14.3 (5.9) mm. Of all participants, 82% were European, 9% were Māori, 8% were Asian, and 1% was of Middle Eastern/Latin American/African ethnicity. 54% of participants were prescribed insulin, 52% metformin, and 39% another oral diabetes medicine. Cardiovascular comorbidity was common, being present in 62%, followed by respiratory disease in 33%. 97% of the study population were fully vaccinated (n=93) or partially vaccinated (n=4) and 3% were unvaccinated against COVID-19.Table 1 Demographic data and anthropometric measurements (n=100) Variable Mean (SD) Median (IQR) Min to Max Age (years), n=100 61 (14.6) 63 (51 to 73) 25 to 85 Height (m), n=100 1.7 (0.1) 1.69 (1.64 to 1.77) 1.47 to 1.93 Weight (kg), n=100 89.4 (22.5) 84.75 (73.65 to 101.1) 48.6 to 171.7 BMI (kg/m2), n=100 30.9 (7.3) 29.9 (25.1 to 35.9) 18.6 to 61.5 Body fat (%), n=99 34.7 (12.4) 36 (25.9 to 43.9) 5.8 to 66.2 Arm circumference (cm), n=100 37.1 (5.8) 35.95 (32.85 to 40.35) 27.2 to 57.6 Skinfold thickness (mm), n=100 30.9 (10.6) 30 (23 to 38) 10 to 55 Skin to deltoid muscle distance (mm), n=100 14.3 (5.9) 13.5 (10 to 16) 5 to 36 Skin to deltoid muscle distance The proportion (95% CI) of the study population with a SDMD >25mm was 6/100: 6% (2.2 to 12.6). The proportion of the study population with a SDMD >20mm, >32mm and >38mm was 11% (5.6 to 18.8), 2% (0.2 to 7.0) and 0% respectively (Figure 1 ).Figure 1 Distribution of the skin to deltoid muscle distance measurements All six participants with a SDMD of >25mm had a BMI ≥ 35kg/m2, and five of the six participants were female (Figure 2 ). Nine of the 11 participants with a SDMD of > 20mm were female, and 9/11 had a BMI ≥ 35kg/m2.Figure 2 The percentage of participants with SDMD >20mm and >25mm by sex and BMI category Relationship between anthropometric measurements and skin to deltoid muscle distance There was a linear relationship between all anthropometric measures and SDMD (Figure 3 ). The strongest relationships to SDMD were with arm circumference, BMI and body fat percentage, with correlation coefficients greater than 0.7. (Table 2 )Figure 3 LOESS plots of the relationship between ultrasound skin to deltoid muscle distance (SDMD) and anthropometric measurements Table 2 Association between skin to deltoid muscle measurements and anthropometric measurements by linear regression Predictor Skin to deltoid muscle distance per unit increase predictor (95% CI) R-squared Correlation coefficient P Arm circumference (cm) 0.77 (0.64 to 0.90) 57.8 0.76 <0.001 Skinfold thickness (mm) 0.28 (0.18 to 0.37) 25.5 0.50 <0.001 BMI (kg/m2) 0.59 (0.48 to 0.70) 53.5 0.73 <0.001 Body fat (%) 0.34 (0.27 to 0.40) 50.7 0.71 <0.001 Height (m) -20.0 (-31.5 to -8.5) 10.9 -0.33 <0.001 Weight (kg) 0.14 (0.09 to 0.18) 28.4 0.53 <0.001 Discrimination of anthropometric measurements for skin to deltoid muscle distance cut point Arm circumference and BMI were the best predictors of SDMD measurements above the 25mm and 20mm cut points with AUC ROC curves of 0.99 and 0.94, and 0.97 and 0.89 respectively (Table 3 , Supplement Figure S1). An arm circumference of ≥45.9cm had a sensitivity of 100% and specificity of 95.7% for a SDMD measurement ≥25mm; the corresponding values for a SDMD measurement ≥20mm were 72.7% and 97.8% respectively (Supplement Table S1). A BMI of ≥36.6 kg/m2 had a sensitivity of 100% and specificity of 85.1% for a SDMD measurement ≥25mm; the corresponding values for a SDMD measurement ≥20mm were 72.7% and 86.5% respectively.Table 3 Area under the receiver operating characteristic curve for skin to deltoid muscle distance Cut-point AUC ROC Predictor 20 mm 25 mm Arm circumference (cm) 0.94 0.99 Skinfold thickness (mm) 0.88 0.81 BMI (kg/m2) 0.89 0.97 Body fat (%) 0.90 0.94 Height (m) 0.66 0.66 Weight (kg) 0.81 0.92 Variance components in relation to investigator There was little evidence of investigator variability in relation to SDMD measurement, with variance components of 0.079 for investigator and 34.38 for residual with an estimated intraclass correlation coefficient of >0.99. Sex differences in anthropometric measurements Women were shorter and had a lower weight than men, but had an increased skinfold thickness, body fat percentage and SDMD (Table 4 ). There was no evidence of a difference in arm circumference, BMI or age between women and men.Table 4 Comparisons of anthropometric measurement variables in relation to sex Variable Female minus Male Estimate (95% CI) P Age (years) -0.02 (-5.86 to 5.82) 0.99 Arm circumference (cm) 0.71 (-1.60 to 3.02) 0.54 Skinfold thickness (mm) 4.68 (0.54 to 8.82) 0.027 BMI (kg/m2) 1.46 (-1.45 to 4.36) 0.32 Body fat (%) 13.6 (9.5 to 17.8) <0.001 Height (m) -0.14 (-0.17 to -0.12) <0.001 Skin to deltoid muscle distance (mm) 5.24 (3.15 to 7.33) <0.001 Weight (kg) -10.34 (-19.06 to -1.62) 0.02 Discussion This study has shown that the standard needle length of 25mm may be insufficient to ensure deposition of COVID-19 vaccine in the deltoid muscle in a small but important proportion of obese adults with diabetes. In total, six percent of participants overall and 12% of participants with a BMI ≥30 kg/m2 had an ultrasound measure of SDMD greater than 25mm at the recommended COVID-19 vaccination site, indicating that in such individuals the vaccine administered with a standard needle may not have reached the deltoid muscle. In 11% of participants overall, and in 20% of participants with a BMI ≥30 kg/m2, the ultrasound-measured SDMD was greater than 20mm, indicating that the use of a 25mm needle may not penetrate at least 5mm into the muscle, as recommended to ensure deposition of the vaccine within the muscle.[10] Measures of BMI and arm circumference at the site of injection were strong predictors of ultrasound measures of SDMD, potentially providing simple, practical and low-cost alternatives to ultrasound for assessment of required needle length at the point of vaccination. These findings extend the current knowledge of the SDMD in adults for which there is insufficient evidence to enable specific recommendations to be made for the appropriate needle length for deltoid intramuscular injection based on demography and anthropometry.[10], [22], [23], [24], [25], [26], [27], [28] The previous observation that female sex is associated with greater SDMD was confirmed in our study.[10], [22], [25], [26] In women, the SDMD was 5.2mm greater than men, together with a higher body fat percentage, despite no evidence of a difference in BMI or arm circumference. This is consistent with the known physiological association of female sex with increased body fat to lean body mass ratio compared with men.[29] These results suggest that different anthropometric cut points for predicting SDMD are required for female and male adults. The relationship between anthropometric measures and SDMD measured by ultrasound was investigated as a basis for determining if cut points for such measures could be used at vaccination sites, recognising that ultrasound would not be a feasible method to use in mass vaccination programmes. We extended the previous reports that BMI was strongly associated with SDMD measured by ultrasound.[10], [22], [25], [26] Arm circumference, BMI and body fat percentage all had strong linear relationships with SDMD, with correlation coefficients greater than 0.7. In this dataset, arm circumference was the best predictor of SDMD measurements. An arm circumference of 45.9 cm or greater had a sensitivity of 100% and a specificity of 96% for a SDMD of >25mm. BMI also had good predictive ability; for example, a BMI of 36.6 kg/m2 or greater had a sensitivity and specificity for a SDMD of >25mm of 100% and 85% respectively. Strengths and weaknesses We studied a population that was predominantly European New Zealanders and so the findings may not be generalisable to other ethnic groups in which differences in body fat distribution may occur.[30] Further study in ethnic groups with high rates of comorbidities such as obesity, and which have greater risk of severe disease with COVID-19 is required. The relatively small dataset including 100 participants limited the precision of our estimates. Compression from the ultrasound probe could distort subcutaneous tissue and result in underestimation of the SDMD.[10], [22] To minimise this risk, minimal pressure was used in obtaining the recordings. In future research the option of utilizing other modalities such as CT or MRI scanning exists, although practical, cost and other issues such as radiation dose with CT scanning may limit their utility. The recommended site of deltoid intramuscular injection varies between national vaccination guidelines.[17], [20], [21], [31] This is potentially an important consideration, given evidence that the precise location used influences the SDMD.[23], [24] The site measured in this study (recommended in Australian 2020 immunisation guidelines), may result in a greater SDMD than a measurement taken at the site recommended in American guidelines,[31] and a smaller SDMD than at the site recommended in New Zealand guidelines.[20] In the UK, ‘The Green Book’ recommends injections within a broad triangular area over the deltoid muscle, within which the SDMD will vary depending on the site chosen.[17] The injection site in local vaccination guidelines should be considered when interpreting the results of this study. In addition to reporting the proportion in whom the SDMD was greater than the standard 25mm needle length, the proportions greater than 20mm, 32mm and 38mm were also reported. The 20mm distance was chosen due to the expert opinion that the vaccine must be delivered at least 5mm within the muscle to ensure adequate intramuscular delivery,[10] whereas the 32mm and 38mm distances were chosen to represent the alternative needle lengths available internationally.[17], [20], [21], [31] Using these criteria, in 11% of participants the 25mm needle length may not have ensured intramuscular delivery, and in 2% and 0% the 32mm and 38mm needles respectively, may not have reached the deltoid muscle at the studied injection site. Public health implications The findings can be viewed from a number of public health perspectives. Firstly, and most importantly, regulatory approval has only been granted for the intramuscular administration of mRNA and viral vector COVID-19 vaccines, indicating that if the vaccine is delivered into the subcutaneous tissue, it is not being administered in accordance with its approved registered use. Secondly, and more generally, it is known that under-penetration of the needle risks reduced immunogenic response to other intramuscular vaccines.[10], [32] The magnitude by which lack of intramuscular delivery of COVID-19 vaccines may reduce the intended immune response to COVID-19 vaccines has not been investigated, although relevant to this study, it has been reported that the antibody response to SARS-CoV-2 BNT162b2 vaccine is higher in low and normal-weight individuals than overweight and obese.[33], [34] However, the length of the needles used was not reported, and there may be confounding by the systemic inflammatory response in obesity, which may also impair immune functioning.[35] Reduced levels of anti-SARS-CoV-2 IgG and neutralising antibodies after COVID-19 mRNA vaccination have been reported in people with Type 2 diabetes compared with healthy people with no effect due to obesity [36]. Recently, Watanabe et al reported that central obesity (as assessed by waist circumference) was associated with lower antibody responses to the same vaccine, but not BMI.[37] As obesity is a major risk factor for increased morbidity and mortality in COVID-19 infection, and is a predictor of required needle length to ensure deposition of the COVID-19 vaccine into muscle, determining the interactions between obesity, COVID-19 vaccine deposition and the COVID-19 vaccine response will be important to resolve. Thirdly, subcutaneous injection may result in an increased risk of adverse effects, such as local tissue reaction, abscesses, and granulomas, compared with intramuscular injection,[10], [12], [14] It is possible that such risk could adversely influence vaccine take up, an important issue with COVID-19 vaccine hesitancy. Fourthly, it is apparent that a 25mm needle should not be a universal needle length for COVID-19 vaccine administration and particularly not for obese adults with diabetes. While recognising that body fat distribution may differ in diabetes,[26], [38] it may be reasonable to infer that this interpretation also applies to obese adults generally, regardless of the diagnosis of diabetes; however, this will need to be the subject of further study. Fifthly, the measurement of arm circumference or BMI may serve as simple and practical predictors of the requirement for needles longer than 25mm, recognising that ultrasound measurement is not feasible for mass vaccination programmes. Finally, the potential for reduced efficacy due to the COVID-19 vaccine not being delivered into the muscle could apply to a substantial proportion of the population. In New Zealand, the prevalence of diabetes is about 5.5% of the adult population,[39] and around one in three adults are obese.[40] The proportion of adolescents and adults in whom a needle longer than 25 mm was used in the New Zealand COVID-19 vaccination programme is currently 2.04%.[37] Our findings suggest a larger proportion of the New Zealand population may benefit from vaccination with a larger needle. In conclusion, the standard needle length of 25mm is likely to be insufficient to ensure deposition of COVID-19 vaccine in the deltoid muscle of a small but important proportion of obese adults with diabetes. The risk of extra-intramuscular deposition of COVID-19 vaccine is greater in females and progressively increases with increasing BMI. Further research is urgently required to inform the relationship between anthropometric characteristics and SDMD in both children and adults in different populations to guide the needle length used in COVID-19 vaccination programmes globally. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability Data will be made available on request. Acknowledgements We thank the volunteers who generously took part in this study. We thank the Ruth Maud Ring Spencer Estate, proudly managed by Perpetual Guardian, for funding the study. Funding: The study was funded by the Ruth Maud Ring Spencer Estate. The MRINZ receives Independent Research Organisation funding from the Health Research Council of New Zealand. Contributions: Concept and design of the study: AM, AS, CK, IB, MH, MW (Mark Weatherall), RB, TH. Acquisition of data: LK, MB, MD, MW (Michaela Walton), RS, SK. Recruitment material and online advertising: CK. Analysis and interpretation of data: AE, CK, JB, MD, MW (Mark Weatherall), RB, TH. Drafting the article: MD, RB, TH. All authors revised the manuscript and have given approval for submission of the final article. ==== Refs References 1 Corona G. Pizzocaro A. Vena W. Rastrelli G. Semeraro F. Isidori A.M. Diabetes is most important cause for mortality in COVID-19 hospitalized patients: Systematic review and meta-analysis Rev Endocr Metab Disord 22 2 2021 275 296 33616801 2 Kumar A. Arora A. Sharma P. Anikhindi S.A. Bansal N. Singla V. Is diabetes mellitus associated with mortality and severity of COVID-19? A meta-analysis Diabetes & Metabolic Syndrome: Clinical Research & Reviews 14 4 2020 535 545 3 Yonas E. Alwi I. Pranata R. Huang I. Lim M.A. Yamin M. Elevated interleukin levels are associated with higher severity and mortality in COVID 19 – A systematic review, meta-analysis, and meta-regression Diabetes & Metabolic Syndrome: Clinical Research & Reviews 14 6 2020 2219 2230 4 Williamson E.J. Walker A.J. Bhaskaran K. Bacon S. Bates C. Morton C.E. Factors associated with COVID-19-related death using OpenSAFELY Nature 584 7821 2020 430 436 32640463 5 Chaudhary N. Weissman D. Whitehead K.A. mRNA vaccines for infectious diseases: principles, delivery and clinical translation Nat Rev Drug Discov 20 2021 817 838 10.1038/s41573-021-00283-5 34433919 6 Cook I.F. Pond D. Hartel G. Comparative reactogenicity and immunogenicity of 23 valent pneumococcal vaccine administered by intramuscular or subcutaneous injection in elderly adults Vaccine 25 2007 4767 4774 10.1016/j.vaccine.2007.04.017 17512098 7 Moderna COVID-19 Vaccine EUA Fact Sheet for Health Care Providers. https://www.fda.gov/media/144637/download (accessed Dec 2021). 8 Medsafe Comirnaty (Pfizer/BioNTech) data sheet | The Immunisation Advisory Centre. https://covid.immune.org.nz/resources/written/medsafe-comirnaty-pfizerbiontech-data-sheet (accessed 3 Dec 2021). 9 Palma S. Strohfus P. Are IM injections IM in obese and overweight females? A study in injection technique Applied Nursing Research 26 2013 e1 e4 10.1016/j.apnr.2013.09.002 24156877 10 Poland G.A. Borrud A. Jacobson R.M. Determination of deltoid fat pad thickness. Implications for needle length in adult immunization JAMA 277 1997 1709 1711 9169899 11 McWilliam P.L. Botwinski C.A. LaCourse J.R. Deltoid intramuscular injections and obesity MedSurg Nursing 23 2014 S4 + 12 Zuckerman J.N. The importance of injecting vaccines into muscle. Different patients need different needle sizes BMJ 321 2000 1237 1238 10.1136/bmj.321.7271.1237 11082069 13 Cook I.F. Subcutaneous vaccine administration - an outmoded practice Human vaccines & immunotherapeutics 17 2021 1329 1341 10.1080/21645515.2020.1814094 32991241 14 Haramati N. Lorans R. Lutwin M. Injection granulomas. Intramuscle or intrafat? Archives of family medicine 3 1994 146 148 10.1001/archfami.3.2.146 7994436 15 Ministry of Health NZ. Instructions for multi-dose vial Pfizer/BioNTech vaccine: preparation and administration. https://covid.immune.org.nz/sites/default/files/2021-04/IMAC_COVID_ED_Pfizer_Instructs-EMAIL4.pdf 16 Instructions for multi-dose vial Pfizer and BioNTech vaccine preparation and administration.pdf. https://covid.immune.org.nz/sites/default/files/2021-07/Instructions%20for%20multi-dose%20vial%20Pfizer%20and%20BioNTech%20vaccine%20preparation%20and%20administration.pdf (accessed 3 Dec 2021). 17 Green-Book-Chapter-4.pdf. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/147915/Green-Book-Chapter-4.pdf (accessed 7 Dec 2021). 18 Popkin B.M. Du S. Green W.D. Beck M.A. Algaith T. Herbst C.H. Individuals with obesity and COVID-19: A global perspective on the epidemiology and biological relationships Obesity reviews : an official journal of the International Association for the Study of Obesity 21 11 2020 e13128 10.1111/obr.13128 32845580 19 Chhabria S. Stanford F.C. A long shot: The importance of needle length in vaccinating patients with obesity against COVID-19 Vaccine 40 2022 9 10 10.1016/j.vaccine.2021.11.008 34794822 20 Zealand N. Ministry of Health accessed 3 Dec 2021 Immunisation handbook 2020 2020 https://www.health.govt.nz/system/files/documents/publications/immunisation-handbook-2020-sep20-v2.pdf 21 Australian Technical Advisory Group on Immunisation (ATAGI)Australian Technical Advisory Group on Immunisation (ATAGI). Administration of vaccines. Australian Immunisation Handbook. 2018.immunisationhandbook.health.gov.au 22 Cook I.F. Williamson M. Pond D. Definition of needle length required for intramuscular deltoid injection in elderly adults: an ultrasonographic study Vaccine 24 2006 937 940 10.1016/j.vaccine.2005.08.098 16191454 23 Nakajima Y. Mukai K. Takaoka K. Hirose T. Morishita K. Yamamoto T. Establishing a new appropriate intramuscular injection site in the deltoid muscle Human vaccines & immunotherapeutics 13 9 2017 2123 2129 28604191 24 Nakajima Y. Fujii T. Mukai K. Ishida A. Kato M. Takahashi M. Anatomically safe sites for intramuscular injections: a cross-sectional study on young adults and cadavers with a focus on the thigh Human Vaccines & Immunotherapeutics 16 1 2020 189 196 31403356 25 Shankar N. Saxena D. Lokkur P.P. Kumar N.M. William N.C. Vijaykumar N. Influence of skin-to-muscle and muscle-to-bone thickness on depth of needle penetration in adults at the deltoid intramuscular injection site Medical journal, Armed Forces India 70 4 2014 338 343 25382907 26 Sim K.H. Hwang M.S. Kim S.Y. Lee H.M. Chang J.Y. Lee M.K. The appropriateness of the length of insulin needles based on determination of skin and subcutaneous fat thickness in the abdomen and upper arm in patients with type 2 diabetes Diabetes & metabolism journal 38 2 2014 120 24851206 27 Jutte L.S. Hawkins J. Miller K.C. Long B.C. Knight K.L. Skinfold Thickness at 8 Common Cryotherapy Sites in Various Athletic Populations Journal of Athletic Training 47 2 2012 170 177 22488282 28 Li A.-H. Yeih D.-F. Chen K.-C. Chu S.-H. Weng C.-S. The thicker the skin fold, the less hematoma development: a novel parameter other than body mass index associated with transradial vascular complication Angiology 58 5 2007 597 602 18024944 29 Lemieux S. Prud'homme D. Bouchard C. Tremblay A. Després J.P. Sex differences in the relation of visceral adipose tissue accumulation to total body fatness The American Journal of Clinical Nutrition 58 4 1993 463 467 8379501 30 Rush E.C. Freitas I. Plank L.D. Body size, body composition and fat distribution: comparative analysis of European, Maori, Pacific Island and Asian Indian adults British Journal of Nutrition 102 2009 632 641 10.1017/S0007114508207221 19203416 31 ACIP Vaccine Administration Guidelines for Immunization | CDC. 2021.https://www.cdc.gov/vaccines/hcp/acip-recs/general-recs/administration.html (accessed 3 Dec 2021). 32 Middleman A.B. Anding R. Tung C. Effect of needle length when immunizing obese adolescents with hepatitis B vaccine Pediatrics 125 2010 e508 e512 10.1542/peds.2009-1592 20142295 33 Pellini R. Venuti A. Pimpinelli F. Abril E. Blandino G. Campo F. Initial observations on age, gender, BMI and hypertension in antibody responses to SARS-CoV-2 BNT162b2 vaccine EClinicalMedicine 36 2021 100928 34109307 34 Magdy Beshbishy A. Hetta H.F. Hussein D.E. Saati A.A. C. Uba C. Rivero-Perez N. Factors Associated with Increased Morbidity and Mortality of Obese and Overweight COVID-19 Patients Biology 9 9 2020 280 32916925 35 Kwok S. Adam S. Ho J.H. Obesity: A critical risk factor in the COVID -19 pandemic Clin Obes 2020 10 10.1111/cob.12403 36 Ali H. Alterki A. Sindhu S. Alahmad B. Hammad M. Al-Sabah S. Robust Antibody Levels in Both Diabetic and Non-Diabetic Individuals After BNT162b2 mRNA COVID-19 Vaccination Front Immunol 12 2021 752233 10.3389/fimmu.2021.752233 37 Watanabe M. Balena A. Tuccinardi D. Tozzi R. Risi R. Masi D. Central obesity, smoking habit, and hypertension are associated with lower antibody titres in response to COVID-19 mRNA vaccine Diabetes Metabolism Res 38 1 2022 10.1002/dmrr.3465 38 Maskarinec G. Raquinio P. Kristal B.S. Franke A.A. Buchthal S.D. Ernst T.M. Body Fat Distribution, Glucose Metabolism, and Diabetes Status Among Older Adults: The Multiethnic Cohort Adiposity Phenotype Study Journal of Epidemiology 32 7 2022 314 322 33642515 39 Indicator: Diabetes (diagnosed, excluding diabetes during pregnancy). https://minhealthnz.shinyapps.io/nz-health-survey-2020-21-annual-data-explorer/_w_827c66c6/#!/explore-indicators (accessed 3 Dec 2021). 40 Indicator: Obese: BMI of 30.0 or greater. https://minhealthnz.shinyapps.io/nz-health-survey-2020-21-annual-data-explorer/_w_827c66c6/#!/explore-topics (accessed 3 Dec 2021).

0

PMC9753456

NO-CC CODE

2022-12-16 23:25:43

no

Vaccine X. 2022 Dec 15;:100248

utf-8

Vaccine X

10.1016/j.jvacx.2022.100248

oa_other

==== Front Vaccine X Vaccine X Vaccine: X 2590-1362 The Author(s). Published by Elsevier Ltd. S2590-1362(22)00109-7 10.1016/j.jvacx.2022.100249 100249 Article Agent-based Model of the Impact of Higher Influenza Vaccine Efficacy on Seasonal Influenza Burden Krauland Mary G ab⁎ Zimmerman Richard K c Williams Katherine V. c Raviotta Jonathan M c Harrison Lee H. d Williams John V. e Roberts Mark S ab a Department of Health Policy and Management, School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA b Public Health Dynamics Laboratory, School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA c Department of Family Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, USA d Center for Genomic Epidemiology, University of Pittsburgh, Pittsburgh, Pennsylvania, USA e Department of Pediatrics, School of Medicine, University of Pittsburgh and University of Pittsburgh Medical Center Children's Hospital of Pittsburgh, Pittsburgh, Pennsylvania, USA ⁎ Corresponding author at: 7132 Public Health, 130 De Soto St, Pittsburgh, PA, USA 15261 15 12 2022 15 12 2022 10024915 3 2022 8 8 2022 14 12 2022 © 2022 The Author(s). Published by Elsevier Ltd. 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Introduction Current influenza vaccines have limited effectiveness. COVID-19 vaccines using mRNA technology have demonstrated very high efficacy, suggesting that mRNA vaccines could be more effective for influenza. Several such influenza vaccines are in development. FRED, an agent-based modeling platform, was used to estimate the impact of more effective influenza vaccines on seasonal influenza burden. Methods Simulations were performed using an agent-based model of influenza that included varying levels of vaccination efficacy (40-95% effective). In some simulations, level of infectiousness and/or length of infectious period in agents with breakthrough infections was also decreased. Impact of increased and decreased levels of vaccine uptake were also modeled. Outcomes included number of symptomatic influenza cases estimated for the US. Results Highly effective vaccines significantly reduced estimated influenza cases in the model. When vaccine efficacy was increased from 40% to a maximum of 95%, estimated influenza cases in the US decreased by 43% to >99%. The base simulation (40% efficacy) resulted in ∼28 million total yearly cases in the US, while the most effective vaccine modeled (95% efficacy) decreased estimated cases to ∼22,000. Discussion Highly effective vaccines could dramatically reduce influenza burden. Model estimates suggest that even modest increases in vaccine efficacy could dramatically reduce seasonal influenza disease burden. Keywords Influenza Agent-based modeling Vaccine ==== Body pmcIntroduction Although vaccines for influenza have been available in the US since the late 1940’s [1], vaccine effectiveness varies and is often modest. The Centers for Disease Control and Prevention (CDC) assesses influenza vaccine effectiveness yearly; in the 2004-5 to 2019-20 seasons, influenza vaccine effectiveness ranged from 10 to 60%, with a mean of ∼40% [2]. Hypothesized reasons for this low effectiveness include poor strain match of vaccine to major circulating strain [3], [4], changes in vaccine during production [5], low immunogenicity [6], and interference from immunity caused by first exposure [7] or by recent vaccination [8], [9], [10]. Development of vaccines using mRNA technology has been proceeding since the 1990s [11]. While initially the technology encountered significant obstacles, many have been resolved over the past decade. mRNA vaccines offer several advantages over conventional vaccines, including rapid vaccine development and manufacturing compared to vaccines grown in cell systems [11], [12]. mRNA vaccines have been shown to elicit strong immune responses in animals [13] and this technology is positioned to become the leader in responding to infectious diseases, particularly newly emerging ones [12]. The technology has been investigated for use against several infectious diseases, including Zika, HIV and rabies [11]. The COVID-19 pandemic accelerated research in mRNA vaccine technology, resulting in development and deployment of mRNA vaccines with efficacy reported as high as 95% at preventing COVID-19 infection [14], [15], [16], [17]. mRNA vaccines for influenza are in development and have shown to be immunogenic in animal models and human trials [18], [19], [20], [21]. The high efficacy of mRNA vaccines against COVID-19 and the success of mRNA vaccines against influenza in animal models suggest that such influenza vaccines may be more effective in humans than the current vaccines. To investigate the impact of higher efficacy vaccines, an agent-based model (ABM) of influenza implemented in the Framework for Reproducing Epidemiological Dynamics (FRED) with varying levels of vaccine efficacy was used to estimate the possible impact of more effective vaccines on seasonal influenza in the US. ABMs have been used extensively to model influenza [22], [23], [24], [25], [26], [27]. This type of model is ideal for investigating vaccine impacts because characteristics such as age, gender and specific susceptibility to disease can be assigned on an individual basis to agents in the simulation population, resulting in highly flexible and granular models. An additional benefit of the FRED platform is that infections result from the interaction of agents over the course of an influenza season. Therefore, the effects of increasing vaccine efficacy are applied not only to the vaccinated agents but to the wider population of agents with whom they interact, through decreased transmission over the entire season. Some evidence indicates that COVID-19 mRNA vaccination caused decreased infectiousness and decreased length of infectious period of breakthrough cases, [28], [29], [30], [31], [32], [33], [34] so this was also investigated for seasonal influenza. A portion of the population has reservations about the safety of mRNA vaccines and so might not utilize an mRNA vaccine but conversely, a more effective vaccine for influenza might also encourage vaccination. To investigate those scenarios, the impact of increased and decreased vaccine uptake was also modeled with varied vaccine efficacy. Methods FRED is an ABM platform that uses census-based synthetic populations in which agents have household demographics, incomes and locations that are statistically realistic at the US census block group level. FRED agents spread infectious conditions through agent interactions in schools, workplaces, neighborhoods and households [35]. FRED models are based upon conditions, which include definitions of states, state transitions and their probabilities and time periods for evaluation of state changes. Conditions can be written to be generic and combined in a simulation. At each timestep, every agent will be in exactly one state in each condition included in the simulation. Conditions can interact directly by allowing a state change in one condition to cause a state change in another condition. FRED has been used to model influenza and other diseases and conditions and has been described in detail [24], [25], [36], [37]. In FRED the reproduction number for an infectious condition is not an input but is generated by the simulation. Each infectious condition has a transmissibility parameter, an infectious period and other characteristics that interact with the specified population to generate a reproduction number in the simulation. The main simulations described here produced an effective reproduction number (Reff) of ∼1.2 in the early phases of the simulation timeline. Comparisons of the base simulations were also performed with higher transmissibility, resulting in Reff of ∼1.4 and ∼1.8. Reproduction number as calculated from FRED results is a proxy for, but not identical to, reproduction number calculated from surveillance data (see Appendix). This study used a modified Susceptible-Exposed-Infectious-Recovered (SEIR) model with the addition of a 1-day pre-symptomatic infectious period (Figure 1 ). Agents in the model can be hospitalized from influenza and hospitalized agents may die. Hospitalization and death occurred at probabilities drawn from published CDC reports [38]. Input parameters and simulation description are included in Table 1 .Figure 1 Influenza model diagram. Influenza was modeled as a modified SEIR, with addition of a pre-symptomatic state, an asymptomatic infectious state, and states representing hospitalization and death. Table 1 Model Inputs Inputs Population 1,218,695 agents derived from Allegheny County census population Influenza Model States E exposed Duration, number of days drawn from a lognormal distribution with μ = 1.9 and σ = 1.23 Ps pre-symptomatic 1 day; infectious at 50% of level of Is Is symptomatic infectious Duration, number of days drawn from a lognormal distribution with μ = 4 and σ = 1.5 in baseline simulation; 75% of infected agents Ia asymptomatic infectious Duration, number of days drawn from a lognormal distribution with μ = 5 and σ = 1.5 in baseline simulation; infectious at 50% of level of Is; 25% of infected agents Hospitalization (probability applied to Symptomatic infectious agents) Rates by age group: age 0-4, 6.97%; age 5-17, 2.74%; age 18-49, 5.61%; age 50-64, 1.06%; age 65+, 9.09% Death (probability applied to Hospitalized agents) Rates by age group: age 0-4, 0.80%; age 5-17, 0.80%; age 18-49, 3.10%; age 50-64, 5.75%; age 65+, 7.73% Prior Immunity applied at simulation start Rates by age group: age 0-4, 20.9%; age 5-17, 15.4%; age 18-49, 11.1%; age 50-64, 13.4%; age 65+, 3.6%, Vaccine Uptake Rates by age group: age 0.5-17, 50.4%; age 18-49, 34.2%; age 50-64 46.8%; age 65+, 68.7% in baseline simulation Vaccine Efficacy Varied 40-95% Simulation Parameters Simulation Period September 15 to May 31 Simulations Per Scenario 100 To capture immunity from prior year infections on simulation start, agents in the model were initialized with immunity to influenza at rates based on CDC reported influenza infections for the 2019-20 season [38] (rates by age group: age 0-4, 20.9%; age 5-17, 15.4%; age 18-49, 11.1%; age 50-64, 13.4%; age 65+, 3.6%, see Appendix). Agents were vaccinated beginning on October 1 of the simulation year. Vaccine uptake was by age group at rates reported by the CDC for 2019-20 [39] (rates by age group: age 0.5-17, probability 0.504; age 18-49 probability 0.342; age 50-64 probability 0.468; age>=65 probability 0.687, see Appendix). Vaccination date for each agent was drawn from a uniform distribution of 1-45 days after October 1 to stagger the timing of immunity. Vaccine immunity is effective 14 days after vaccination in the model. Agents are randomly chosen for application of prior immunity and vaccination (Appendix). Vaccination induced immunity waned at a rate of 7% per month over the simulation [40]. Simulations used a population created from the 2010 Allegheny County Pennsylvania census population. The population consists of ∼1.2 million agents living in urban or suburban areas. The simulation period was one influenza season, extending from September 15 to May 31 (Figure 2 ). Results were scaled to reflect total US population. The influenza season was started in the simulations by inserting 50 cases on November 15. The model included a single strain of influenza, representing a season in which a single type A strain predominated.Figure 2 Simulation timeline. Simulations represented one influenza season, beginning on September 15 and ending on May 31. Prior immunity was applied on simulation start. Vaccination began on October 1. The seasonal outbreak began with seeding of cases on November 15. Vaccination caused a decrease in susceptibility to influenza. Decrease in susceptibility was varied from 40 to 95%. In additional simulations, vaccine efficacy was varied along with a decrease of infectiousness of breakthrough infections by 25 or 50% with or without a 1 or 2 day decrease in length of infectious period for breakthrough cases. An additional set of simulations varied vaccine efficacy and included a 1 or 2 day decrease in length of infectious period for breakthrough infections with no change in degree of infectiousness for those cases. Simulation scenarios are listed in Table 2 .Table 2 Simulation scenarios. Vaccine efficacy Infectious period of breakthrough infections Level of infectiousness of breakthrough infections Base simulation 40-95% Base* Base* Base with decreased infectious period 1 day shorter Base 2 days shorter Base 25% decreased level of infectiousness with varied infectious period base 25% decrease 1 day shorter 25% decrease 2 days shorter 25% decrease 50% decreased level of infectiousness with varied infectious period base 50% decrease 1 day shorter 50% decrease 2 days shorter 50% decrease * Base level of infectiousness produces R ∼1.2 with base infectious period To estimate the impact of increased or decreased vaccine uptake, uptake rates were increased or decreased from CDC reported rates for 2019-20 by 10% and 20% for all age groups. Results were compared with the baseline model at each level of vaccine efficacy. The University of Pittsburgh Institutional Review Board has determined that this study was not human subject research. Results The base model with immunity from prior year infection and vaccination at 40% efficacy with uptake rates similar to those reported by the CDC resulted in a mean over 100 simulations of 28,052,175 symptomatic cases (Std Dev 5,424,113) when scaled to the total US population. From 2010 to 2020, the CDC estimated a range of 9.3 to 41 million symptomatic influenza infections per year in the US (mean 28,230,000, Std Dev 8,680,508) [41]. Modeled increases in vaccine efficacy alone resulted in dramatic decreases in estimated influenza cases (Figure 3 Panel A, Table 3 ). An increase in vaccine efficacy from 40% to 50% decreased estimated cases by 43% from the base scenario to a mean of 15,884,871 (Std Dev 5,859,781). Vaccine efficacy of 80% resulted in a 99% decrease in cases (mean 90,549, Std Dev 109,365) (Appendix Table S3).Figure 3 Impact of Higher Vaccine Effectiveness, Decreased Level of Infectiousness and Shorter Infectious Period on Total Symptomatic Influenza Cases in the US. Base scenarios varied vaccine effectiveness from 40% to 95%. In scenarios representing decreased length of infectious period in breakthrough cases, length of infectious period was decreased by 1 day or 2 days. In simulations representing a lower level of infectiousness in breakthrough infections scenarios, level of infectiousness was decreased by 25% or 50%. In some scenarios, both level of infectiousness and length of infectious period were modified. Table 3 Total US symptomatic influenza cases in base simulation with increased vaccine effectiveness and with lower level of infectiousness and/or shorter period of infectiousness in breakthrough infections (infections occurring in vaccinated agents). Vaccine Effectiveness Simulation Scenario* Base 1-day shorter infectious period 2-day shorter infectious period 40% 28,052,175 (5,424,113) 8,992,725 (4,217,098) 736,736 (715,947) 50% 15,884,871 (5,859,781) 2,908,146 (2,301,825) 283,505 (251,078) 60% 4,328,639 (2,736,138) 671,510 (553,063) 84,031 (73,662) 70% 570,059 (582,682) 145,390 (142,240) 57,641 (48,999) 80% 90,549 (109,365) 46,929 (43,299) 28,142 (17,742) 90% 27,931 (24,225) 24,037 (13,646) 22,653 (10,234) 95% 22,144 (10,970) 21,386 (16,127) 19,233 (8,326) 25% less infectious Base infectious period 1-day shorter infectious period 2-day shorter infectious period 40% 4,482,185 (2,868,150) 723,381 (666,119) 129,384 (122,008) 50% 1,582,181 (1,261,457) 306,846 (266,661) 73,062 (65,945) 60% 346,080 (341,315) 108,201 (101,863) 52,434 (37,994) 70% 105,515 (120,617) 52,044 (45,781) 33,844 (19,784) 80% 42,114 (32,265) 31,077 (18,914) 24,700 (12,505) 90% 25,572 (18,343) 21,871 (9,935) 19,666 (8,053) 95% 20,262 (7,460) 19,572 (8,010) 18,361 (6,873) 50% less infectious Base infectious period 1-day shorter infectious period 2-day shorter infectious period 40% 193,922 (172,039) 95,228 (89,498) 48,621 (33,149) 50% 108,878 (93,709) 49,991 (35,930) 36,885 (22,902) 60% 63,652 (51,335) 43,303 (35,756) 33,379 (17,733) 70% 39,057 (24,108) 31,762 (17,143) 27,614 (13,006) 80% 28,052 (15,346) 24,971 (11,901) 22,436 (9,129) 90% 20,270 (7,826) 20,143 (9,164) 18,524 (6,672) 95% 19,553 (7,677) 18,575 (7,087) 17,579 (5,482) *Mean (standard deviation) over 100 simulations Simulations with higher transmissibility values resulting in Reff in 100 simulations of ∼1.4 and ∼1.8 (Appendix Figure S3, Table S2) resulted in higher infection rates. A given increase in vaccine efficacy reduced cases by a lower percent in higher Reff simulations; however increased vaccine efficacy still resulted in large decreases in infections with reduction to very low levels (>97% reduction) by vaccine efficacy of 80% or above without changes in level of infectiousness or length of infectious period of breakthrough cases. Decreasing the level of infectiousness of breakthrough cases (defined as infections in vaccinated agents) resulted in greater decreases in influenza burden for all levels of vaccine efficacy in the Reff ∼1.2 scenarios (Figure 3 Panel B, Table 3, Appendix Table S3). An increase in vaccine efficacy from 40% to 50% coupled with a 25% decrease in level of infectiousness decreased estimated cases by 94% from the base level with 1,582,181 cases (Std Dev 1,261,457). Vaccine efficacy of 80% with 25% decrease in level of infectiousness resulted in a > 99% decrease in cases (42,114 cases, Std Dev 32,265). A larger decrease in level of infectiousness of breakthrough cases to 50% resulted in even greater reductions in influenza burden. Decreasing the length of the infectious period of breakthrough cases resulted in decreases in influenza burden in addition to that caused by changes in vaccine efficacy (Figure 3 Panels C & D). At 40% vaccine efficacy, a decrease in length of infectious period by 1 day resulted in a 68% decrease in cases over the baseline (8,992,725 cases (Std Dev 4,217,098)). When vaccination efficacy was 80% with a 1 day decrease in length of infectious period, symptomatic cases decreased by >99% to 46,929 (Std Dev 43,299). Decreasing the length of the infectious period by 2 days resulted in larger decreases in symptomatic cases in the simulation (40% vaccine efficacy, 736,736 cases, Std Dev 715,947; 80% vaccine efficacy, 28,142 cases, Std Dev 17,742) (Table 3, Appendix Table S3). Decreasing the level of infectiousness along with decreasing the length of the infectious period of breakthrough cases resulted in the largest decreases in influenza burden (Table 3). With a 50% effective vaccine, reduction in level of infectiousness by 25% combined with decreased period of infectiousness by 1 day gave a 99% reduction in cases from the base simulation, with a mean of 306,846 cases (Std Dev 266,661). At 80% vaccine efficacy with the same reductions in level of infectiousness and decreased infectious period, mean cases dropped to 31,077 (Std Dev 18,914). Reducing the level of infectiousness of breakthrough cases by 50% and/or decreasing the length of the infectious period by 2 days resulted in even greater reductions in influenza burden, with reductions of >99% in all scenarios (Figure 3 Panels C & D, Appendix Table S3). Hospitalizations and deaths in the simulations followed similar patterns to reduction in cases (Appendix Table S3). Increasing or decreasing vaccine uptake resulted in roughly proportional decreases or increases in cases, respectively, for vaccine efficacy values of 40 to 60% (Figure 4 ). A 20% decrease in vaccine uptake had greater impact at 70% vaccine efficacy. At 90% or greater efficacy, differences in uptake had little impact on yearly influenza burden (Appendix Table S4) due to very large decreases in estimated cases.Figure 4 Yearly Symptomatic Influenza Cases in US With Increased Vaccine Effectiveness and Increased or Decreased Vaccine Uptake Base scenarios varied vaccine effectiveness from 40% to 95% and used CDC reported influenza vaccine uptake rates. Vaccine uptake rates were increased or decreased by 10% or 20% across all age groups. Discussion Efficacy results from clinical trials as well as effectiveness results in vaccinated populations for the COVID-19 mRNA vaccines exceeded expectations, with short-term reduction in cases, hospitalizations and deaths of greater than 90% [42], [43]. Development of mRNA influenza vaccines was underway before the COVID-19 pandemic [11] and the success of this type of vaccine for COVID-19 suggests that mRNA vaccines for influenza could be more effective than current vaccines, which have suboptimal vaccine effectiveness [18], [19], [44]. In addition to potentially having a major impact on influenza burden, mRNA vaccines have additional benefits including shorter development and production time and avoidance of the problems associated with egg-based vaccine production (e.g., mutations due to production in eggs). These benefits alone could make the use of mRNA vaccines worthwhile for influenza. In addition to being highly effective, mRNA vaccination for COVID-19 may have decreased the infectivity of breakthrough cases, potentially through decreasing the level of infectiousness and length of infectious period [28], [29], [31], [32]. Some studies have found viral load in vaccine breakthrough cases to be decreased but the results have been mixed and this phenomenon may be SARS-CoV-2 variant dependent [30], [33], [34]. Breakthrough cases may clear virus more quickly and therefore be infectious for a shorter period [33]. Influenza virus shedding may be reduced in amount and duration due to immunity from vaccine or prior infection [45], making a similar phenomenon possible with mRNA influenza vaccines. Our simulation results suggest that increases in vaccine efficacy to levels achieved by COVID-19 mRNA vaccines could markedly reduce influenza burden, with even moderate increases in efficacy potentially having a large impact. Even a modest reduction in infectious period or degree of infectiousness of breakthrough cases would be highly beneficial in vaccines of modest efficacy because many infections would be breakthroughs. Decreases in influenza burden in these simulations reflect not only the impact of a more effective vaccine on the vaccinated agent but also the interruption of transmission; therefore, the impact is much higher than what would be seen only by decreasing the probability of infection for the vaccinated agents. Vaccination benefits not only the vaccinated but also those whom they would infect, propagated onward through the population over the influenza season. In simulations, higher efficacy vaccines had substantial effects without increased vaccine uptake levels above that of recent years. Even a decrease in vaccine uptake from recent levels had little impact on influenza cases in the model when vaccine efficacy was 80% or higher. When higher efficacy was combined with decrease in infectivity and/or decrease in infectious period, the model estimated substantial reductions in disease burden. Although mRNA vaccines have been extremely successful for COVID-19, it is unknown whether they will perform as well for influenza. COVID-19 vaccines may have benefited from the highly antigenic spike protein target and from circulation of relatively few antigenic variants of that target compared to influenza antigen targets. Some vaccines using the spike protein target have not achieved as high a level of efficacy as the mRNA vaccines [46]. Durability of protection with mRNA vaccines is still to be determined. mRNA vaccines also tend to be more reactogenic and may be more expensive. Other technologies, such as self-amplifying RNA [47] may prove to be superior. Ongoing studies will provide the data to evaluate mRNA vaccines for influenza but there is a clear need for more effective influenza vaccines. Strengths and Limitations Modeling has inherent limitations; models are simplifications of reality whose reliability depends on the validity of the model itself and on the accuracy of estimation of parameters upon which the model relies. The usefulness of a model is not in the exact results it produces but in the insights it generates. The simulations described here were designed to estimate the impact of increased vaccine efficacy and of decreases in level of infectiousness and length of infectious period in breakthrough cases due to a more effective vaccine. It may not completely replicate what would occur in a real epidemic, but it provides useful information on the possible impact of those parameters on influenza burden. The influenza model used in the simulations reported here is based upon one that was designed for the 2009 influenza pandemic. It has been validated and used in multiple influenza simulations since that time and so can be considered a robust design [24], [25], [35]. The model has been modified to include a measure of immunity due to prior infection and to allow for vaccination. Parameters for the reported simulations were drawn mainly from official CDC data and therefore represent the best estimates. With high vaccine efficacy, the simulations produced limited hospitalizations and deaths, resulting in large relative standard deviations for these outcomes. This set of simulations does not specifically model mRNA vaccines; it models the impact of increased vaccination efficacy. Therefore, although COVID-19 mRNA vaccines suggested the model, the results are generalizable to any more effective vaccine. Conclusions Even moderate increases in efficacy of influenza vaccines from mRNA or other novel platforms could have a large impact on influenza burden. Further benefits would occur if such vaccines also caused a decrease in infectivity and/or decrease in infectious period. Funding: This work was supported by the Center for Disease Control and Prevention U01-IP001141-01. The University of Pittsburgh Institutional Review Board has determined that this study was not human subject research and is therefore exempt study design. The study sponsor had no role in study design; collection, analysis, and interpretation of data; writing the report; or the decision to submit the report for publication. Financial disclosure: Drs. Zimmerman and Raviotta have research grant funding from Sanofi Pasteur on an unrelated vaccine topic. Dr John Williams serves on the Scientific Advisory Board of Quidel and an Independent Data Monitoring Committee for GlaxoSmithKline, neither involved in the present work. Other authors have declared no financial disclosures. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ==== Refs References. 1 Barberis I. Myles P. Ault S.K. Bragazzi N.L. Martini M. History and evolution of influenza control through vaccination: from the first monovalent vaccine to universal vaccines J Prev Med Hyg. 57 3 2016 E115 E120 27980374 2 Past Seasons Vaccine Effectiveness Estimates. Centers for Disease Control and Prevention. https://www.cdc.gov/flu/vaccines-work/past-seasons-estimates.html. Published 2020. Updated January 29,2020. Accessed August 9, 2021. 3 Belongia E.A. Kieke B.A. Donahue J.G. Effectiveness of inactivated influenza vaccines varied substantially with antigenic match from the 2004–2005 season to the 2006–2007 season J Infect Dis. 199 2 2009 159 167 19086915 4 Tricco A.C. Chit A. Soobiah C. Comparing influenza vaccine efficacy against mismatched and matched strains: a systematic review and meta-analysis BMC Med. 11 2013 153 23800265 5 Zost S.J. Parkhouse K. Gumina M.E. Contemporary H3N2 influenza viruses have a glycosylation site that alters binding of antibodies elicited by egg-adapted vaccine strains Proc Natl Acad Sci U S A. 114 47 2017 12578 12583 29109276 6 Schotsaert M. Garcia-Sastre A. Inactivated influenza virus vaccines: the future of TIV and QIV Curr Opin Virol. 23 2017 102 106 28505524 7 Lewnard J.A. Cobey S. Immune History and Influenza Vaccine Effectiveness Vaccines (Basel) 6 2 2018 8 Ohmit S.E. Petrie J.G. Malosh R.E. Influenza vaccine effectiveness in the community and the household Clin Infect Dis. 56 10 2013 1363 1369 23413420 9 Skowronski D.M. Chambers C. Sabaiduc S. A Perfect Storm: Impact of Genomic Variation and Serial Vaccination on Low Influenza Vaccine Effectiveness During the 2014–2015 Season Clin Infect Dis. 63 1 2016 21 32 27025838 10 Thompson M.G. Naleway A. Fry A.M. Effects of Repeated Annual Inactivated Influenza Vaccination among Healthcare Personnel on Serum Hemagglutinin Inhibition Antibody Response to A/Perth/16/2009 (H3N2)-like virus during 2010–11 Vaccine. 34 7 2016 981 988 26813801 11 Pardi N. Hogan M.J. Porter F.W. Weissman D. mRNA vaccines - a new era in vaccinology Nat Rev Drug Discov. 17 4 2018 261 279 29326426 12 Erasmus J.H. Fuller D.H. Preparing for Pandemics: RNA Vaccines at the Forefront Mol Ther. 28 7 2020 1559 1560 32579882 13 Thess A. Grund S. Mui B.L. Sequence-engineered mRNA Without Chemical Nucleoside Modifications Enables an Effective Protein Therapy in Large Animals Mol Ther. 23 9 2015 1456 1464 26050989 14 Frenck R.W. Jr. Klein N.P. Kitchin N. Safety, Immunogenicity, and Efficacy of the BNT162b2 Covid-19 Vaccine in Adolescents N Engl J Med. 385 3 2021 239 250 34043894 15 Hall V.J. Foulkes S. Saei A. COVID-19 vaccine coverage in health-care workers in England and effectiveness of BNT162b2 mRNA vaccine against infection (SIREN): a prospective, multicentre, cohort study Lancet. 397 10286 2021 1725 1735 33901423 16 Polack F.P. Thomas S.J. Kitchin N. Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine N Engl J Med. 383 27 2020 2603 2615 33301246 17 Teo S.P. Review of COVID-19 mRNA Vaccines: BNT162b2 and mRNA-1273 J Pharm Pract. 8971900211009650 2021 18 Bahl K. Senn J.J. Yuzhakov O. Preclinical and Clinical Demonstration of Immunogenicity by mRNA Vaccines against H10N8 and H7N9 Influenza Viruses Mol Ther. 25 6 2017 1316 1327 28457665 19 Feldman R.A. Fuhr R. Smolenov I. mRNA vaccines against H10N8 and H7N9 influenza viruses of pandemic potential are immunogenic and well tolerated in healthy adults in phase 1 randomized clinical trials Vaccine. 37 25 2019 3326 3334 31079849 20 Freyn A.W. Ramos da Silva J. Rosado V.C. A Multi-Targeting, Nucleoside-Modified mRNA Influenza Virus Vaccine Provides Broad Protection in Mice Mol Ther. 28 7 2020 1569 1584 32359470 21 Lutz J. Lazzaro S. Habbeddine M. Unmodified mRNA in LNPs constitutes a competitive technology for prophylactic vaccines NPJ Vaccines. 2 2017 29 29263884 22 DePasse J.V. Nowalk M.P. Smith K.J. Does cost-effectiveness of influenza vaccine choice vary across the U.S.? An agent-based modeling study Vaccine. 35 32 2017 3974 3981 28606814 23 Guo D. Li K.C. Peters T.R. Snively B.M. Poehling K.A. Zhou X. Multi-scale modeling for the transmission of influenza and the evaluation of interventions toward it Sci Rep. 5 2015 8980 25757402 24 Kumar S. Grefenstette J.J. Galloway D. Albert S.M. Burke D.S. Policies to reduce influenza in the workplace: impact assessments using an agent-based model Am J Public Health. 103 8 2013 1406 1411 23763426 25 Kumar S. Piper K. Galloway D.D. Hadler J.L. Grefenstette J.J. Is population structure sufficient to generate area-level inequalities in influenza rates? An examination using agent-based models BMC Public Health. 15 2015 947 26400564 26 O'Neil C.A. Sattenspiel L. Agent-based modeling of the spread of the 1918–1919 flu in three Canadian fur trading communities Am J Hum Biol. 22 6 2010 757 767 20721982 27 Whitman J. Jayaprakash C. Stochastic modeling of influenza spread dynamics with recurrences PLoS One. 15 4 2020 e0231521 28 Harris R.J. Hall J.A. Zaidi A. Andrews N.J. Dunbar J.K. Dabrera G. Effect of Vaccination on Household Transmission of SARS-CoV-2 in England N Engl J Med. 385 8 2021 759 760 34161702 29 Klompas M. Understanding Breakthrough Infections Following mRNA SARS-CoV-2 Vaccination JAMA. 326 20 2021 2018 2020 34734985 30 Levine-Tiefenbrun M. Yelin I. Katz R. Initial report of decreased SARS-CoV-2 viral load after inoculation with the BNT162b2 vaccine Nat Med. 27 5 2021 790 792 33782619 31 Regev-Yochay G. Amit S. Bergwerk M. Decreased infectivity following BNT162b2 vaccination: A prospective cohort study in Israel Lancet Reg Health Eur. 7 2021 100150 32 Shah A.S.V. Gribben C. Bishop J. Effect of Vaccination on Transmission of SARS-CoV-2 N Engl J Med. 385 18 2021 1718 1720 34496200 33 Singanayagam A. Hakki S. Dunning J. Community transmission and viral load kinetics of the SARS-CoV-2 delta (B.1.617.2) variant in vaccinated and unvaccinated individuals in the UK: a prospective, longitudinal, cohort study Lancet Infect Dis. 2021 34 Petter E, Mor O, Zuckerman N, et al. Initial real world evidence for lower viral load of individuals who have been vaccinated by BNT162b2. medRxiv. 2021. 35 Grefenstette J.J. Brown S.T. Rosenfeld R. FRED (a Framework for Reconstructing Epidemic Dynamics): an open-source software system for modeling infectious diseases and control strategies using census-based populations BMC Public Health. 13 2013 940 24103508 36 Chu K.H. Shensa A. Colditz J.B. Integrating Social Dynamics Into Modeling Cigarette and E-Cigarette Use Health Educ Behav. 47 2 2020 191 201 32090652 37 Liu F. Enanoria W.T. Zipprich J. The role of vaccination coverage, individual behaviors, and the public health response in the control of measles epidemics: an agent-based simulation for California BMC Public Health. 15 2015 447 25928152 38 Estimated Influenza Illnesses, Medical visits, Hospitalizations, and Deaths in the United States — 2019–2020 Influenza Season. Centers for Disease Control and Prevention. https://www.cdc.gov/flu/about/burden/2019-2020.html. Published 2021. Updated June 2, 2021. Accessed August 11, 2021. 39 Flu Vaccination Coverage, United States, 2019–20 Influenza Season. Centers for Disease Control and Prevention. https://www.cdc.gov/flu/fluvaxview/coverage-1920estimates.htm. Published 2020. Updated October 1, 2020. Accessed August 13, 2021. 40 Ferdinands J.M. Fry A.M. Reynolds S. Intraseason waning of influenza vaccine protection: Evidence from the US Influenza Vaccine Effectiveness Network, 2011–12 through 2014–15 Clin Infect Dis. 64 5 2017 544 550 28039340 41 Disease Burden of Flu. Centers for Disease Control and Prevention. https://www.cdc.gov/flu/about/burden/index.html. Published 2022. Updated October 1, 2020. Accessed January 13, 2022. 42 Pilishvili T. Gierke R. Fleming-Dutra K.E. Effectiveness of mRNA Covid-19 Vaccine among U.S. Health Care Personnel N Engl J Med. 385 25 2021 e90 43 Thompson MG, Burgess JL, Naleway AL, et al. Interim Estimates of Vaccine Effectiveness of BNT162b2 and mRNA-1273 COVID-19 Vaccines in Preventing SARS-CoV-2 Infection Among Health Care Personnel, First Responders, and Other Essential and Frontline Workers - Eight U.S. Locations, December 2020-March 2021. MMWR Morb Mortal Wkly Rep. 2021;70(13):495-500. 44 Pardi N. Parkhouse K. Kirkpatrick E. Nucleoside-modified mRNA immunization elicits influenza virus hemagglutinin stalk-specific antibodies Nat Commun. 9 1 2018 3361 30135514 45 Maier H.E. Nachbagauer R. Kuan G. Pre-existing Antineuraminidase Antibodies Are Associated With Shortened Duration of Influenza A(H1N1)pdm Virus Shedding and Illness in Naturally Infected Adults Clin Infect Dis. 70 11 2020 2290 2297 31300819 46 Voysey M. Clemens S.A.C. Madhi S.A. Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK Lancet. 397 10269 2021 99 111 33306989 47 Vogel A.B. Lambert L. Kinnear E. Self-Amplifying RNA Vaccines Give Equivalent Protection against Influenza to mRNA Vaccines but at Much Lower Doses Mol Ther. 26 2 2018 446 455 29275847

0

PMC9753457

NO-CC CODE

2022-12-16 23:25:43

no

Vaccine X. 2023 Apr 15; 13:100249

utf-8

Vaccine X

10.1016/j.jvacx.2022.100249

oa_other

==== Front Med Image Anal Med Image Anal Medical Image Analysis 1361-8415 1361-8423 The Author(s). Published by Elsevier B.V. S1361-8415(22)00350-4 10.1016/j.media.2022.102722 102722 Article Bilateral adaptive graph convolutional network on CT based Covid-19 diagnosis with uncertainty-aware consensus-assisted multiple instance learning Meng Yanda a Bridge Joshua a Addison Cliff b Wang Manhui b Merritt Cristin c Franks Stu c Mackey Maria d Messenger Steve d Sun Renrong e Fitzmaurice Thomas f McCann Caroline g Li Qiang h Zhao Yitian hi⁎ Zheng Yalin aj⁎⁎ a Department of Eye and Vision Science, University of Liverpool, Liverpool, United Kingdom b Advanced Research Computing, University of Liverpool, Liverpool, United Kingdom c Alces Flight Limited, Bicester, United Kingdom d Amazon Web Services, 60 Holborn Viaduct, London, United Kingdom e Department of Radiology, Hubei Provincial Hospital of Integrated Chinese and Western Medicine, Hubei University of Chinese Medicine, Wuhan, China f Adult Cystic Fibrosis Unit, Liverpool Heart and Chest Hospital NHS Foundation Trust, Liverpool, UK g Radiology, Liverpool Heart and Chest Hospital NHS Foundation Trust, UK h The Affiliated People’s Hospital of Ningbo University, Ningbo, China i Cixi Institute of Biomedical Engineering, Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Science, Ningbo, China j Liverpool Centre for Cardiovascular Science, University of Liverpool and Liverpool Heart & Chest Hospital, Liverpool, UK ⁎ Corresponding author. ⁎⁎ Corresponding author at: Department of Eye and Vision Science, University of Liverpool, Liverpool, United Kingdom. 15 12 2022 15 12 2022 10272211 4 2022 17 10 2022 2 12 2022 © 2022 The Author(s) 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Coronavirus disease (COVID-19) has caused a worldwide pandemic, putting millions of people’s health and lives in jeopardy. Detecting infected patients early on chest computed tomography (CT) is critical in combating COVID-19. Harnessing uncertainty-aware consensus-assisted multiple instance learning (UC-MIL), we propose to diagnose COVID-19 using a new bilateral adaptive graph-based (BA-GCN) model that can use both 2D and 3D discriminative information in 3D CT volumes with arbitrary number of slices. Given the importance of lung segmentation for this task, we have created the largest manual annotation dataset so far with 7,768 slices from COVID-19 patients, and have used it to train a 2D segmentation model to segment the lungs from individual slices and mask the lungs as the regions of interest for the subsequent analyses. We then used the UC-MIL model to estimate the uncertainty of each prediction and the consensus between multiple predictions on each CT slice to automatically select a fixed number of CT slices with reliable predictions for the subsequent model reasoning. Finally, we adaptively constructed a BA-GCN with vertices from different granularity levels (2D and 3D) to aggregate multi-level features for the final diagnosis with the benefits of the graph convolution network’s superiority to tackle cross-granularity relationships. Experimental results on three largest COVID-19 CT datasets demonstrated that our model can produce reliable and accurate COVID-19 predictions using CT volumes with any number of slices, which outperforms existing approaches in terms of learning and generalisation ability. To promote reproducible research, we have made the datasets, including the manual annotations and cleaned CT dataset, as well as the implementation code, available at https://doi.org/10.5281/zenodo.6361963. Keywords COVID-19 Multiple instance learning Graph convolutional network Uncertainty and consensus CT images ==== Body pmc1 Introduction Coronavirus disease (COVID-19) is a highly contagious respiratory infection caused by the new coronavirus SARS-CoV2. The most frequent symptoms of infection in the majority of infected people are fever, dry cough, and malaise (Wang et al., 2020c). Some of these patients quickly deteriorate, developing acute respiratory distress syndrome, septic shock, multiple organ failure, and even death, among other complications (Huang et al., 2020, Li et al., 2020a, Chen et al., 2020b). Nearly 600 million people have been infected worldwide so far, and over 6 million lost their lives, COVID-19 still spreads across the world. Timely and accurate COVID-19 diagnosis is critical for estimating the need for intensive care unit admission, oxygen therapy, prompt treatment, and so on. Despite the large number of deep learning models that have been proposed so far for the diagnosis of COVID-19 using computed tomography (CT) and X-ray, none of them is clinically usable due to methodological flaws and/or underlying biases (Roberts et al., 2021). There is an unmet need of accurate and robust diagnosis models for COVID-19. Given existing COVID-19 related datasets, such as computed tomography (CT), X-ray, etc., previous deep learning based diagnosis methods (Goncharov et al., 2021, Bai et al., 2020b, Li et al., 2020b, Wang et al., 2020b, Han et al., 2020, He et al., 2021a) focus on the identification of three classes: novel coronavirus pneumonia (NCP), normal controls (Normal), and common pneumonia (CP) at either 2-dimensional (2D) or 3-dimensional (3D) level depending on the types of data they have used. Specifically, CT plays an important role in diagnosing and quantifying COVID-19 and other pneumonia (He et al., 2021a, Wang et al., 2021, Yang et al., 2021a, Yao et al., 2021, Xie et al., 2020, Zhu et al., 2021, Chao et al., 2021, Yang et al., 2021b, Xu et al., 2021). The appearances on CT of infective pneumonia can give clues to its aetiology, as certain consolidation patterns are associated with specific pneumonia. Fig. 1 demonstrates axial CT slices comparison between various patterns of pneumonia.Fig. 1 Axial CT slices demonstrate various patterns (red arrows emphasised) of pneumonia. A: consolidation in the posterior right upper lobe and superior right lower lobe showing typical air bronchograms and a segmental/lobar distribution in an individual with bacterial pneumonia. B: multifocal patches of airspace change in the posterior right upper lobe in an individual with viral pneumonia. C: bilateral multifocal ground glass changes in the upper lobes with some smaller reticulonodular opacities, in an individual with COVID-19 pneumonia. The CP group consists of different disease types, normally including community-acquired bacterial pneumonia and viral pneumonia. In detail, community-acquired bacterial pneumonia is described as showing focal segmental or lobar opacities, but may also show ground glass attenuation or centrilobular nodules (Vilar et al., 2004). Viral pneumonia is often described as multifocal, patchy or ground glass consolidation with influenza specifically demonstrating bilateral reticulonodular opacities (Koo et al., 2018, Kim et al., 2002). COVID-19 is associated with ground glass opacities (GGO) and areas of consolidation that are often bilateral and peripheral (Shi et al., 2020, Li et al., 2020d, Hani et al., 2020). However, given the overlap in radiological appearances between etiological agents, with a few exceptions no reliable diagnosis of bacterial or viral origin can be made from CT (Reittner et al., 2003), and attempts to differentiate definitively between COVID-19 and other viral pneumonia by imaging have met previously with similarly limited success (Kim et al., 2021, Bai et al., 2020a). Additionally, the underlying correlations among CT slices are essential in NCP diagnosis or infection detection tasks (Greenspan et al., 2020) but have not been considered enough in existing methods. Thus, we specially design and evaluate the proposed model on COVID-19 CT dataset in this work. The proposed framework is also readily applied to other medical applications where 3D data such as CT or MRI are used. Fig. 2 shows an overview of our proposed methods, where we propose a novel diagnosis framework in an attempt to address four critical difficulties that were rarely discussed or unsolved by earlier CT-based COVID-19 approaches. The four critical issues are discussed and elaborated as follows. Firstly, lung segmentation is an essential step prior to performing the COVID-19 classification task, however, it has received little attention in previous methods. Due to a lack of ground truth masks, previous methods (Wu et al., 2021a, Goncharov et al., 2021) segmented the lungs with pre-trained models on non-COVID datasets, while others (Wang et al., 2020b, Wang et al., 2020a) adopted un-/ weakly-supervised schemes. However, due to the noticeable domain gap and complex appearances of CT images specific to COVID-19 (e.g. severe cases with massive GGO), the major issue is poor segmentation performance, which will compromise the subsequent NCP classification task. As a result, these methods need to manually clean a large number of wrong segmentations, which increase the labour cost and inconvenience for use. Here we manually annotated 7,768 slices from public COVID-19 datasets and trained a segmentation model under a fully-supervised learning mechanism. Our segmentation model can achieve more accurate results than pre-trained models or previous un-/ weakly-supervised methods; please refer to Fig. 6 for the qualitative comparison between our model’s and others’ segmentations. We also prove that, without the lung segmentation, the subsequent diagnosis model may only learn a specific format pattern of different classes rather than the actual radiographic diagnosis characteristics (i.e. GGO for NCP). This may be due to specific CT scanner models, protocol standards, data sources of different classes, etc.. The potential dataset issues related to the lung segmentation process are further discussed in Section 7.1. Secondly, selecting a fixed number of slices from each CT volume is often compulsory as the size of the inputs have to be the same for specific models (Fang et al., 2021, He et al., 2021a, Wang et al., 2021, Li et al., 2020c). Manual selection of CT slices is labour-intensive and time-consuming, which is incompatible with the goal of using AI models. Automatic selection following pre-defined slice sampling rules, on the other hand, may result in a hand-crafted optimisation process. Additionally, possibly infected slices being missed may construct a noisy dataset with intrinsic uncertainty. In this work, we propose automatically selecting reliable CT slices according to the model’s probability prediction on 2D slices. A specially designed Uncertainty-aware Consensus-assisted Multiple Instance Learning UC-MIL model is proposed to achieve such a goal. Our UC-MIL can extract 2D level features for each CT slice and automatically select trustworthy slices. Thirdly, several methods (Shamsi et al., 2021, Mallick et al., 2020, Calderon-Ramirez et al., 2021) have attempted to quantify the uncertainty in the COVID-19 classification task but rarely exploited it during the training process. In other words, they only treated uncertainty as a quantification tool after the model had been trained, which overlooked the potential benefits of uncertainty during the model training process. In general, there are two types of uncertainty (Kendall and Gal, 2017): epistemic uncertainty, which corresponds to the uncertainty in the model parameters and can be addressed when sufficient data is available; and aleatoric uncertainty, which corresponds to the inevitable noisy perturbations presented in the data. Publicly available CT datasets (e.g. (Zhang et al., 2020)) contain inevitable inherent uncertainties and constraints (He et al., 2021a), such as multiple domains data sources, duplicated or noisy slices, damaged data, disordered and incomplete slices, etc.. Alleviating the aleatoric uncertainty and exploiting it during the supervision is significant for the COVID-19 classification tasks. In this work, we propose a UC-MIL to extract 2D level features and select reliable CT slices. Specifically, an uncertainty and consensus estimation module is proposed to assist the supervision process of the multiple instance learning (MIL) models. The underlying motivations are threefold: (1) As discussed before, the inherent uncertainty in the CT dataset may perturb the model learning process. (2) In some NCP cases, there might be only few slices with COVID-19 features. Under the assumptions of class-imbalanced slices distribution in a CT volume, classic MIL might result in the classification decision boundary closer to the uncertain (rare) slices (Li et al., 2021a). (3) Owing to the weakly supervised learning nature of MIL, the model is prone to overfitting to noisy and uncertain slices (Khan et al., 2019), as all the slices from a COVID-19 positive CT volume will have the same positive labels. Nonetheless, because many slices may still look normal, this label assignment may mistakenly introduce label noise and uncertainty.Fig. 2 Overview of the proposed diagnosis framework. Our framework first segments and crops automatically the lung regions from the input raw 3D CT volume. Then, we automatically select trustworthy slices and the corresponding 2D features via the proposed UC-MIL. Afterwards, a graph-based reasoning model BA-GCN is proposed to aggregate and fuse the information (vertices) at 2D and 3D levels simultaneously, which contributes to the final diagnosis. Fourthly, previous COVID-19 related deep learning methods only rely on the extracted features from either 2D or 3D level, for example, 2D CNN models on 2D X-ray images (Zhong et al., 2021, Minaee et al., 2020, Soda et al., 2021) or 3D CNN models for CT volumes (He et al., 2021a, Wang et al., 2021, Zhu et al., 2021). Differently, we propose to aggregate and reason features from 2D and 3D levels concurrently during the model learning process. Specifically, we adopt the pre-trained 2D CNN model of the proposed UC-MIL as the initialised 2D level feature extractor. We also adopt a 3D CNN as the backbone network to extract the 3D level features. Please refer to Section 3.3 for further details. With the 2D and 3D features extracted from the input CT slices, we propose a BA-GCN to aggregate the 2D and 3D information. Previous graph-based methods (Meng et al., 2021c, Luo et al., 2020a, Guo et al., 2021) have proven the superiority of graph-based models on tackling cross-granularity relationships. In this work, we regard 2D and 3D features as the bilateral vertices in a graph. A Graph Convolution Network (GCN) based model is proposed to aggregate information and exchange messages between cross-granularity vertices (2D and 3D). Note that the graph structure and edge relationships between vertices are adaptively learnt during the reasoning process, according to the 2D and 3D level features, respectively. In this way, the proposed BA-GCN can adaptively fuse and reason the bilateral relationships between 2D and 3D vertices. Specifically, in this work, the message exchange and information aggregation within 2D/3D vertices can be considered as ‘inner-granularity’, and between 2D/3D vertices can be considered as ‘cross-granularity’. Our experiments prove that such an adaptively learnt graph can better tackle the cross-granularity relationships and achieves superior classification performance than previous GCN reasoning based methods. Please refer to Section 6.2.2 for more details. In summary, this work makes the following contributions: • This work proves that lung segmentation is an essential and necessary pre-processing step for the COVID-19 classification task on the public CT datasets. We establish the largest lung region mask dataset, with precise manual annotations of lung boundaries on the public COVID-19 CT dataset. Because of its significance we will make them publicly available to promote related research in the community. • We propose an Uncertainty-aware Consensus-assisted Multiple Instance Learning (UC-MIL) model for 2D level feature extraction and automatic selection of reliable CT slices simultaneously. This avoids handcrafted data preparation and also allows the framework to work on CT volumes with an arbitrary number of slices. It also alleviates the effects of inherent noise in public datasets on the learning and the potential uncertainty from the weakly-supervised learning mechanism of MIL. • We propose a Bilateral Adaptive Graph Convolution Network (BA-GCN) to aggregate information and exchange messages between bilateral cross-granularity vertices (2D and 3D levels). An adaptively learned graph structure and edge relationships are built during the graph learning process to fuse and reason the relationships between 2D and 3D vertices. This helps our proposed method consider features at both levels when making inference, thus improving the classification performance. • Extensive experiments show that our framework comprising UC-MIL and BA-GCN outperforms existing related approaches in terms of learning ability on the three largest publicly available COVID-19 CT datasets. In respect of varying dataset sources, we evaluate the generalisation ability of the proposed model on one of them as the external dataset, demonstrating its superior robustness and generalisability to the previous methods. 2 Related works In this section, we review previous COVID-19 related works w.r.t. 2D and 3D level, respectively in several aspects, such as classification, infection segmentation, severity assessment, etc.. Additionally, as lung segmentation is an essential pre-process for the diagnosis, we review and compare previous works with such pre-process in a separate section. Apart from that, GCN related works in biomedical image tasks (segmentation, classification, etc.) have also been discussed. 2.1 COVID-19 diagnosis at 2D level It is known that tackling the NCP diagnosis problem with 2D X-ray or 2D ultrasound images can achieve promising results in many tasks, such as severity assessment (Signoroni et al., 2021, Xue et al., 2021), infection localisation (Vieira et al., 2021, Wang et al., 2021b, Roy et al., 2020, Malhotra et al., 2021) and diagnosis (Zhong et al., 2021, Minaee et al., 2020, Soda et al., 2021, Oh et al., 2020, Aviles-Rivero et al., 2022, Guarrasi et al., 2022, Kumar et al., 2022, Shorfuzzaman and Hossain, 2021, Fan et al., 2021, Bridge et al., 2020). However, compared with CT images, X-ray cannot indicate the significant appearance characteristics of NCP, such as GGO), multi-focal patchy consolidation and bilateral patchy shadows (Zhang et al., 2020). On the other hand, CT images are 3D volumes, which contain correlated spatial information among slices, essential for NCP diagnosis and infection localisation tasks. However, some previous methods (Gao et al., 2021, Wu et al., 2021b, Uemura et al., 2021a, Wang et al., 2020d, Liu et al., 2021b, Fan et al., 2020b, Wang et al., 2020e, Zhou et al., 2020, Hou et al., 2021) overlooked the 3D spatial information and developed 2D deep Learning model for the aforementioned tasks only on selected CT slices. This is mainly due to limited 3D data at the early pandemic stage, and various slice numbers of CT scans from different patients. Thus, it is difficult to develop models that can directly take CT volumes with a random number of slices as input. A potential solution adopted by previous methods (Goncharov et al., 2021, Bai et al., 2020b, Li et al., 2020b, Qian et al., 2020) is to extract 2D features independently for each slice, then combining all slices’ feature maps via pooling operations. Although all the slices are considered, features are still extracted independently, and correlations between slices are not utilised. Other than that, hand-crafted selection of a fixed number of slices is commonly used for most CT based COVID-19 methods. We will discuss these methods in the following section (Section 2.2). 2.2 COVID-19 diagnosis at 3D level Information at 3D level is essential for the tasks related with COVID-19. Most deep learning based models used 3D CT volume as the input, such as classification (He et al., 2021a, Wang et al., 2021, Tan and Liu, 2021), segmentation (Yang et al., 2021a, Yao et al., 2021, Xie et al., 2020, Yang et al., 2021b), disease progression (Zhu et al., 2021, Chao et al., 2021, Xu et al., 2021), etc.. However, all of them need a pre-process to select a fixed number of slices as the input of these models. For example, Fang et al. (2021) selected 64 slices per CT volume as the model’s input. Similarly, He et al., 2021a, Tan and Liu, 2021 utilised different slices sampling rules, including random sampling and symmetrical sampling, to select a fixed number of slices. Then a neural architecture search (NAS) technique was proposed to search 3D models for the NCP diagnosis. Along the same line, Wang et al. (2021) used an equal interval sampling rule to select slices. A joint segmentation and classification model was proposed to indicate 3D lesion regions and NCP diagnosis simultaneously. (Li et al., 2020c) proposed to extract the features of COVID-19 positive and negative samples as the pretext task, then a downstream model was developed to tackle the NCP classification. However, the pre-selection step was not discussed, where a fixed size of 256 × 192 × 56 voxels were cropped from CT volume as the input. Ouyang et al. (2020) proposed a size-balanced slice sampling mechanism to train the model in terms of repeating NCP data with small infections and CP data with large infections in each mini-batch. A pre-selection process of different patients w.r.t. different infection regions (small or large) need to be manually done as well. Excessive manual pre-processes made the whole framework labour-extensive and unsuitable for real-world applications. Despite the cutting-edge performance of the models mentioned above, manual selection of a fixed number of slices is an underrated and rarely discussed issue in the task of COVID-19 with CT. For example, manual selection of CT slices is labour-intensive and time-consuming, which violate the intention of developing AI models. Automatic selection under manually designed slice sampling rules may lead to a handcrafted optimisation process and cause missing potential infected slices, which results in noise data and unsubtle predictions. Furthermore, more hyper-parameters, such as the interval value, are introduced into the developed model, which will impair models generalisability. Differently, we propose a UC-MIL framework to work as an automated trustworthy slice selection module, according to the estimated uncertainty and consensus score during the inference. Thus, our framework can automatically select corresponding slices, eliminating the labour-extensive pre-selection process and meeting real-world applications’ needs. In other words, our model can work with a raw CT volume with an arbitrary number of slices instead of pre-selected stacked CT slices. 2.3 Multiple instance learning Multiple Instance Learning (MIL) based methods (Li et al., 2021b, Chikontwe et al., 2021, He et al., 2021b, Han et al., 2020) play a significant role to address the aforementioned challenges. In detail, a whole CT volume of a patient is considered as a bag of slices (instances) that can be COVID-19 positive or negative. Then a patient-level label is given to train the model under the weakly-supervised learning mechanism. Most of the aforementioned MIL based methods are inspired from (Ilse et al., 2018), where an attention mechanism is proposed to learn a scoring system among different instances for the patient-level inference. For example, Li et al. (2021b) proposed an attention-based MIL framework for the task of NCP severity assessment, where the instance-level attention module assigns attention scores to different instances automatically during inference. Along the same line, Chikontwe et al. (2021) and Han et al. (2020) both exploited the instance-level attention mechanisms in the task of NCP diagnosis. In contrast, we propose to research the uncertainty and interpretability learning of the MIL model. A scoring system among different slices is achieved by the uncertain value of each instance’s probability predicted by our UC-MIL model. On the other hand, previous MIL based methods only rely on the extracted features of 2D instance levels. The attention module can only be seen as a weighting system among the embeddings of bags; the underlying correlations between instances are still understudied. Nevertheless, the correlations are essential in NCP diagnosis or infection detection tasks (Greenspan et al., 2020). In our proposed framework, the UC-MIL works for feature extraction and reliable slices selection in the first stage. Moreover, we developed a 3D volume-based BA-GCN model in the second stage to simultaneously exploit the 2D pixel-level features and 3D slice-level correlations for a better diagnosis performance. 2.4 Segmentation before classification To mitigate the influence of non-lung region in CT slices, a standard pipeline will be to segment the lung region as a prerequisite before the NCP diagnosis (Wu et al., 2021a, Wang et al., 2020b, Wang et al., 2020a, Zhao et al., 2021b). For example, Wu et al. (2021a) and Goncharov et al. (2021), segmented the lung regions using a pre-trained U-net on other disease (non-COVID) dataset, such as NSCLC (Kiser et al., 2020) and LUNA16 (Team, 2011), then directly applied it to the COVID-19 CT datasets, (e.g. CC-CCII (Zhang et al., 2020) or MosMed (Morozov et al., 2020)). However, NSCLC and LUNA16 are CT datasets containing epithelial lung cancers, which differ noticeably from CC-CCII (Zhang et al., 2020) and MosMed (Morozov et al., 2020). The domain gaps between these datasets will cause poor segmentation performance of the pre-trained model, which in turn compromises the NCP diagnosis performance. Differently, Wang et al. (2020b) utilised an unsupervised method (Liao et al., 2019) to extract the connected component activation regions, which are regarded as the lung regions. However, the segmentation performance is relatively poor. It is due mainly to the distinct appearance of NCP CT slices from other normal ones, such as GGO, multi-focal patchy consolidation and patchy bilateral shadows. Thus, they had to manually clean a large number of failure cases. On the other hand, Wang et al. (2020a) followed (Wang et al., 2019), used primitive thresholding and connected-component labelling algorithms to obtain a binary lung mask that indicates the coarse lung regions. Then, a sub-image was cropped to contain lung regions covered by the convex hull of the lung masks. They treated the rough mask as the ground truth to train a model to segment the lungs, which led to inevitable noisy training data because of the inaccurate lung regions. In summary, the aforementioned methods either adopted a pre-trained model or un-/ weakly-supervised methods to segment the lung region due to the lack of ground truth. The primary issue is poor segmentation performance, which perturbs the following NCP diagnosis task. Again, some methods need to clean the wrong segmentations manually, which increases the labour cost and reduces repeatability. On the contrary, we trained our segmentation module with the manually annotated lung masks under a fully-supervised learning mechanism; our segmentation model can achieve highly accurate results. We will make this manual annotation dataset publicly available. For more details about the dataset, readers are referred to Section 4.2. 2.5 Uncertainty-assisted COVID-19 diagnosis In recent years, the uncertainty and interpretability of deep learning models have been explored in several different computer vision tasks, such as scene understanding (Meng et al., 2021b, Zhang et al., 2021) and medical image analysis (Yu et al., 2019, Ji et al., 2021, Wang et al., 2020f, Luo et al., 2020b), etc.. Quantifying the uncertainty is crucial for COVID-19 classification task since publicly available CT datasets contain inherent constraints, such as multiple domains of data sources, limited dataset size, etc.. (Shamsi et al., 2021) proposed a transfer learning-based framework with the help of quantified uncertainty to address the COVID-19 diagnosis problem. They estimated the epistemic uncertainty with an ensemble learning scheme (Lakshminarayanan et al., 2017). Differently, Mallick et al. (2020) developed a deep uncertainty-aware classifier using a probabilistic generalisation of the non-parametric KNN approach. The proposed probabilistic neighbourhood component analysis method maps samples to latent probability distributions and then minimises a form of nearest-neighbour loss to develop classifiers. Then they estimated the uncertainty in terms of a threshold of the fraction of correctly classified examples. On the other hand, Calderon-Ramirez et al. (2021) researched the underlying capability of unlabeled data to improve the reliability of uncertainty. They estimated the uncertainty with the Monte Carlo Dropout (Le et al., 2018) methods under the MixMatch (Berthelot et al., 2019) semi-supervised learning scheme. Although these aforementioned methods studied the uncertainty in the diagnosis of COVID-19 cases, the estimated uncertainty is only used as a quantification tool at the inference stage, which overlooked the potential benefits of uncertainty during the model training. Instead, we exploit the value of uncertainty throughout the training process. Specifically, an uncertainty-aware consensus-assisted training mechanism is proposed to help the model produce more reliable predictions. Please read Section 3.2.2 for more details. 2.6 Graph-based diagnosis and reasoning Graph-based reasoning algorithms have been studied in the recent years. Benefits from Graph Neural Network (GNN)’s superior ability of information propagation and message exchange, it achieved promising results in segmentation ((Meng et al., 2020a, Meng et al., 2020b, Huang et al., 2021, Meng et al., 2021c, Meng et al., 2021a, Zhang et al., 2019), classification (Liang et al., 2018, Chen et al., 2019, Rhee et al., 2018, Chen et al., 2020a, Noh et al., 2020, Hao et al., 2021) and reconstruction (Zhao et al., 2021a, Yao et al., 2019, Wickramasinghe et al., 2020, Kong et al., 2021, Chen et al., 2021) tasks in the field of biomedical images analysis. Graph based techniques (Di et al., 2021, Aviles-Rivero et al., 2022, Liu et al., 2021a) have been used to tackle COVID-19 related tasks as well. For example, Aviles-Rivero et al. (2022) proposed a graph diffusion model that reinforces the natural relation among tiny labelled sets and vast unlabeled data in a semi-supervised learning scheme. Specifically, the graph is built on initial embeddings of the network, where each node represents an image, to produce pseudo labels, which is used for the semi-supervised NCP classification task. Moreover, Kumar et al. (2022) combined CNN and GCN to learn the relation-aware representation from the NCP X-ray images. Along the same line, Di et al. (2021) proposed a hypergraph model for the diagnosis of NCP. In detail, various types of features (e.g. regional features and radiographic features) are extracted from CT images for each case (CT volume). Then, the relationship among different cases was formulated by a hypergraph structure. Again, each case represented a vertex (node) in the hypergraph. Similarly, Liu et al. (2021a) proposed a distance-aware pooling procedure along with the GCN to aggregate the slice level feature into the patient level gradually. The CT scan is converted to a densely connected graph, where each slice represents a vertex (node) in the graph. The problem becomes a graph classification task, and each graph represents a different patient (CT volume). The aforementioned methods shared a similar idea: each instance (single slice or whole CT volume) was represented as a vertex in the proposed graph. A subsequent graph reasoning mechanism then propagates the vertex and edge information among instance levels. However, there are some fundamental limitations: (1) the instance level features are reasoned individually. For example, work by Liu et al. (2021a) focused only on slice level feature reasoning by a graph; the same situation happened in the works of Aviles-Rivero et al. (2022) and Di et al. (2021) on patient levels (whole CT volume or X-ray) as well. This setting limits the graph-based model’s capability to tackle cross-granularity or cross-feature information propagation. In other words, the GCN mentioned above only serves to build a long relationship between instances. However, such functionality can also be achieved by pure CNN based methods, according to the recent development of Non-local methods (Wang et al., 2018a) or Transformer-based methods (Dosovitskiy et al., 2021). (2) For GCN based methods (Kumar et al., 2022, Liu et al., 2021a), they adopted Laplacian smoothing-based graph convolution (Kipf and Welling, 2017), which provided specific benefits in the sense of global long-range information reasoning. However, they estimated the initial graph structure from a data-independent Laplacian matrix. Such matrix is defined by a handcrafted or randomly initialised adjacency matrix (Meng et al., 2021a), which leads a model to learn a specific long-range context pattern (Li et al., 2020e). Differently, our graph-based model considered features from both 2D and 3D level to propagate the cross-granularity information. Also, as seen in previous works, the graph structure can be estimated with the similarity matrix from the input data (Li and Gupta, 2018). We estimate the initial adjacency matrix in an input-dependent way. Specifically, a reasoning mechanism is achieved by propagating information and passing messages among inner-granularity and cross-granularity vertices (2D and 3D). Additionally, the structure of our BA-GCN is adaptively built during the graph reasoning according to the information of 2D and 3D levels. Thus, the graph representations can be adaptively learnt in an input-dependent way instead of the pre-defined hand-craft one from the previous methods. Please read Section 3.3.2 for more details. Notably, a recent work (Zhao et al., 2020) built the adjacency matrix based on the instance features in a bag under MIL paradigm, which can also be regarded as input-dependent. However, they handcrafted the adjacency matrix weights, and the major difference between Ours and theirs are threefold: (1) Zhao et al.  (Zhao et al., 2020) built a binary adjacency matrix with edge weight values of 0 or 1 to indicate whether the vertices are connected or not. However, the similarity among vertices is overlooked. Differently, Ours exploited the relationship among vertices’ own correlation and can indicate the similarity of different vertices with normalised edge weights between 0 and 1. (2) Zhao et al.  (Zhao et al., 2020) introduced a hyper-parameter (γ) to determine if two vertices are connected or not, according to their Euclidean distance. Conversely, Ours does not introduce any hyper-parameter and only relies on the vertices’ own correlations. (3) Ours constructs a fully-connected graph with every vertex connecting to one other, while Zhao et al.  (Zhao et al., 2020) did not because of the potential edge weight of 0. Fig. 3 Illustration of the proposed method’s pipeline. In addition to the lung segmentation and region cropping, the two stage diagnosis mechanism w.r.t. UC-MIL and BA-GCN is shown on the top and bottom, respectively. Seg represents the lung region segmentation; UC score denotes the estimated uncertainty and consensus scores. Notably, the non-lung regions were masked out from the raw CT data by using our lung segmentation model before input into the UC-MIL. The 2D/3D level of vertices are initialised by the feature maps at 2D/3D level, which are extracted from UC-MIL and MF-Net backbone, respectively. 3 Methods Fig. 3 shows the proposed method’s pipeline. It contains three sub-tasks: (1) lung region segmentation, (2) reliable CT slices selection and COVID-19 classification on 2D levels (UC-MIL), (3) COVID-19 classification at both 2D and 3D levels (BA-GCN). Given an input CT volume with an arbitrary number of slices, we first segmented the lung regions for each slice, then fed the segmented CT volume into a UC-MIL model to learn and extract the relevant features at 2D slice level under a weakly supervised learning mechanism. After that, we selected D slices from each CT volume according to the predicted probability of UC-MIL model. The D slices and the corresponding 2D features extracted from UC-MIL are regarded as the input for the proposed BA-GCN. The BA-GCN learns the features on the 3D volume level (D slices) and also propagate the information from 2D level features among different vertices in the bilateral graph. Notably, the hyper-parameter D is empirically set as 16 in this work. The details of each task and the developed models are elaborated as follows. 3.1 Lung segmentation Because our intention was primarily the task of COVID-19 classification, here we only utilised classic methods, such as UNet (Ronneberger et al., 2015), UNet++ (Zhou et al., 2019), and other cutting-edge methods, such as PraNet (Fan et al., 2020a), RBA-Net (Meng et al., 2020a), CABNet (Meng et al., 2020b), GRB-GCN (Meng et al., 2021c), and BI-GConv (Meng et al., 2021a). We trained those models with the annotated slices at the 2D level, and applied the trained model on the rest unannotated images then cropped the lung regions. After that, the CT volume containing lungs only is ready for the following COVID-19 diagnosis task. Please note that lung segmentation process is essential and necessary in the task of COVID-19 classification primarily due to the dataset issue. Please refer to Section 7.1 for further details. 3.2 UC-MIL for diagnosis on 2D level To develop a comprehensive COVID-19 classification model, we built a UC-MIL model to learn the diagnosis features at 2D level. In the MIL paradigm (Amores, 2013, Dietterich et al., 1997, Maron and Lozano-Pérez, 1998), unlabeled instances belong to labeled bags of instances. The goal is to predict the label of a new bag or the label of each instance. We will elaborate the mechanisms of the proposed UC-MIL in the following subsections. 3.2.1 Multiple instance learning We denote a patient’s CT volume as a bag and the slices herein as instances, following the standard MIL formulation. We associate the bag label with the corresponding instances. In other words, all instances from the same bag have the same label and are considered discriminatory. Nonetheless, this assignment may inadvertently add label noise in positive bags due to the possibility of a certain number of slices being negative. Thus, exploiting the discriminative training samples is essential under this circumstance. Here, ‘discriminative’ represents that the true hidden label of the instance is the same as the true label of the bag. Let X={X1,X2,…,XN} as the dataset containing N bags. Each bag Xi={xi,1,xi,2,…,xi,Ni} consists of Ni instances, where Xi,j={xi,j,yi}, xi,j is the j-th instance, yi denotes its associated label in the i-th bag. Please note, Ni may differ due to different number of slices in different CT volumes. The label Yi of bag Xi is given by: (1) Yi=0,iff∑iyi=01,otherwise. Generally, a MIL based prediction model contains an appropriate transformation f and a permutation-invariant transformation g (Wang et al., 2018b, Ilse et al., 2018, Li et al., 2021a). Thus, the MIL’s prediction for bag Xi is defined as: (2) P(Xi)=g(f(xi,1),f(xi,2),…,f(xi,Ni)). With respect to the choice of f and g, there are generally two types: 1.) Instance-based approach. f is an instance classifier that assigns a score to each instance, and g is a pooling operator (e.g. max pooling) that fuses the instance scores to obtain a bag score. Specifically, a 2D CNN was trained to predict the class probability of each instance. A few instances with higher responses were selected and performed back-propagation during training. An iteration process was used with a new set of discriminative instances until convergence. 2.) Embedding-based approach. f is an instance-level feature extractor that maps each instance to an embedding; g is an aggregation operator that produces a bag embedding from the instance embeddings and outputs a bag score based on the bag embedding. The embedding-based method generates a bag score from a bag embedding supervised by the bag label. The discriminative and non-discriminative instances’ embeddings contribute differently to the overall bag prediction (Wang et al., 2018b). However, it is typically more challenging to identify the discriminative instances that activate the classifier, compared with instance-based approaches (Liu et al., 2017). 3.2.2 Uncertainty-aware consensus-assisted multiple instance learning All the previous MIL based COVID-19 diagnosis methods (Li et al., 2021b, Chikontwe et al., 2021, Han et al., 2020) are embedding-based methods, which are adapted from (Ilse et al., 2018). In this work, we take another direction and propose an instance-based UC-MIL method. Our experimental results prove that the proposed method outperforms previous instance-based and embedding-based methods on two different evaluation settings (learning ability and generalisation ability). Additionally, we conducted extensive ablation studies to determine the backbone network of the proposed MIL method. More experimental details are referred to Section 6.2.1. Previous instance-based MIL methods (Wu et al., 2015, Pinheiro and Collobert, 2015, Hou et al., 2016, Perdomo et al., 2018, Qiu and Sun, 2019, Campanella et al., 2019, Zhang et al., 2022) achieved promising results on different medical image classification tasks, such as whole slide image classification, optical coherence tomography image classification, etc.. However, two significant challenges remain for these works. Firstly, the distribution of instances in the positive bags may be extremely imbalanced when only a tiny proportion of instances are positive, and models are prone to misclassify those positive instances as negative, especially when a simple aggregation operator, such as max-pooling, is used. This is because, under the assumptions of MIL and imbalanced instances in a bag, max-pooling might result in the classification decision boundary closer to the uncertain (rare) instances (Li et al., 2021a). Secondly, as discussed above, all the instances from the same bag have the same label and are considered discriminatory. Nonetheless, this assignment may inevitably add label noise into positive bags due to the possibility of a certain amount of slices being negative. Due to such a weakly supervised learning mechanism, the model is prone to overfitting to noisy and uncertain instances, resulting in poor generalisability in real-world clinical practice. Additionally, instances with high uncertainty have a disproportionate presence in the classification space, making it difficult to generalise learnt limits to new test examples. (Khan et al., 2019). To solve this problem, we specifically design an uncertainty estimation module and a consensus achievement module into the standard instance-based MIL model training pipeline, where an uncertainty-aware consensus-assisted supervision process is conducted. Firstly, to quantify the reliability of each instance’s prediction, we adopt Shannon Entropy (Shannon and Weaver, 1949) as the metric to measure the randomness of the information (Shannon, 2001), which is referred to as the uncertainty in this work. Formally, given a C-dimensional softmax predicted class score Pxi,j(C) from an input instance xi,j, the uncertainty Ixi,j is defined as: (3) Ixi,j=−∑c=1CPxi,j(C)⊙logPxi,j(C), where ⊙ is Hadamard Product; C is the number of classes. In practice, we perform T times stochastic forward passes on each instance classifier under random dropout and Gaussian noise perturbed input for each input instance. Note that T is empirically set as 8 in this work. Therefore, under such self-ensemble mechanism, we obtain a set of softmax probability vectors: {Pxi,jt}t=1T, then the mean predicted class score P~xi,j(C) is given as: (4) P~xi,j(C)=1T∑t=1TPxi,jt, thus based on Eq. (3) we can obtain the uncertainty I~xi,j for input instance xi,j as : (5) I~xi,j=−∑c=1CP~xi,j(C)⊙logP~xi,j(C). With the quantified uncertainty I~xi,j for instance xi,j, we normalise I~xi,j into [0,1] then perform element-wise broadcasting multiplication between I~xi,j and softmax predicted class score Pxi,j(C). In this way, uncertainty-weighted probability prediction PI~xi,j(C) for each instance xi,j is calculated as: (6) PI~xi,j(C)=I~xi,j⊗Pxi,j(C), where ⊗ denotes the element-wise broadcasting multiplication. In other words, the operator g in our UC-MIL will consider the reliability of each f(xi,Ni) in Eq. (2), and only the trustworthy slides are considered for the model to learn the features. Secondly, under a certain perturbation, network predictions for memorised features that learned from noise change significantly, while those for generalised features do not (Lee and Chung, 2020). In other words, the predictions of a generalisable instance classifier should be robust to input perturbation, and the predicted class score that changes significantly under a certain perturbation hence highly suggests a noisy instance (Li et al., 2021a). Thus, we quantify the consensus regarding the standard deviation over a self-ensembling models’ multiple outputs, with the same input but under various perturbations. Formally, for an instance xi,j, given a set of softmax probability vectors {Pxi,jt}t=1T and the mean predicted class score P~xi,j(C), the standard deviation Pˆxi,j(C) of the predicted class score is defined as: (7) Pˆxi,j(C)=1T∑t=1T(Pxi,jt−P~xi,j(C))2, which is regarded as the metric of consensus in this work. With such quantified consensus achievement, we exclude the uncertain instances so as to guide the model to learn from more reliable instances. More specifically, the reliable instances jr in bag Xi are selected iif Pˆxi,j(C) is smaller than a threshold γ. Formally, for bag Xi, the trustworthy instances set Ω is given by: (8) Ω={xi,j|Pˆxi,j(C)<γ}. Notably, we perform extensive experiments to tune the hyper-parameter γ value, which is empirically set as 0.02 in this work. The number of trustworthy slices in Ω ranges from 16 to 45 for all of the data used in this work. This comes with the advantage that our framework can deal with CT volumes with any arbitrary number of slices. Combining the uncertainty and the consensus scores discussed before, the whole optimisation procedure of the proposed UC-MIL methods in a single bag (Xi) can be found in Algorithm 1. An iteration process was used with a new set of bags (X1,…,XN) to update the parameter of the instance classifier until convergence. In this way, whether a retrieved discriminative instance is trustworthy or noisy can be differentiated by the model during the training. The learnt classifier considers the uncertainty level of the instance predictions to re-adjust boundaries (i.e., providing more room to uncertain samples). This improves the generalisation ability of the proposed model for either imbalanced instances or weakly supervised learning mechanisms (Khan et al., 2019). Furthermore, our experiments prove that with the UC-MIL training, our model outperforms previous instance-level MIL methods by a large margin in the evaluation of generalisation ability. Notably, previous instance-level MIL methods conduct a promising classification results in the seen data (i.e. the evaluation of learning ability), however, drop dramatically on unseen data (i.e. the evaluation of generalisation ability). During the training, we adopted the same method used in (Campanella et al., 2019), which selects the top instances with maximum prediction probability within a bag as the bag’s prediction. Such bag-level aggregation derives directly from the standard multiple instance assumption and is generally referred to as ‘max-pooling’ (Campanella et al., 2019) and is shown in Fig. 3. With the proposed UC-MIL, we obtain temporary patient-level diagnosis results in the first stage. However, the instance level features are learned individually during the whole training process. In other words, only 2D level of information is considered in UC-MIL. Thus we aggregate both 2D and 3D features in the subsequent BA-GCN, which helps to make the diagnosis more reliable and accurate. 3.3 Diagnosis at both 2D and 3D levels In this section, we demonstrate the proposed BA-GCN w.r.t. the COVID-19 diagnosis at both 2D and 3D levels. As discussed in Section 2.2, the correlations between different CT slices are essential for theCOVID-19 diagnosis. For bag Xi, we select the top D instances (slices) according to the ranked order of uncertainty-aware consensus-assisted instance prediction probability (PI~xi,jr(C)) from the corresponding trustworthy set Ω. Then we stack the slices along the depth channel as the 3D input for the proposed BA-GCN. In this way, we can automatically select a fixed number of reliable slices from each CT volume, which avoids the labour-intensive manual selection process or other hand-craft slice sampling strategies that are adopted by the previous methods (Ouyang et al., 2020, Li et al., 2020c, Wang et al., 2021, He et al., 2021a, Fang et al., 2021). Additionally, the extracted slice-level features of UC-MIL are used as the 2D feature maps input for the proposed BA-GCN. Specifically, for each of the D slices classifier in UC-MIL, we extract the feature map before the pooling layer and add an 1×1 convolution layer to reduce the channel size to 128. Then for each CT volume, we stack all the corresponding D feature maps along the depth channel as the ‘2D’ input for the proposed BA-GCN. We represent the ‘2D’ input as X2D in this work. Notably, X2D has the size of D × 128 × 7 × 7. The size format follows (D × C × H × W), where D is number of slices; C is channel size; H and W represents height and width of feature maps, respectively. There are two primary modules in the proposed BA-GCN: (1) Backbone Network, (2) Bilateral Adaptive Graph Reasoning Module. The details for each of them are elaborated as follows. 3.3.1 Backbone network We firstly input 3D CT volumes as the inputs into a backbone network to extract features and learn the correlations between different slices at the 3D level. Different from previous methods (Li et al., 2020c, Wang et al., 2021, Ouyang et al., 2020), where the 3D extensions of ResNet (He et al., 2016) or Inception-Net (Szegedy et al., 2015) are used as the backbone, we adopt Multi-Fiber Network (MF-Net) (Chen et al., 2018) due to its superior ability to extract discriminative features in recognition tasks. MF-Net (Chen et al., 2018) is a sparsely connected 3D CNN backbone that costs a minimal computational overhead, but brings a boosted representation capability of features. The multiple separated lightweight residual units, called fibers, can effectively reduce the number of connections within the network and enhance the model efficiency. The advantage of MF-Net fits in and benefits our model in this specific task. Our ablation study results also prove that the MF-Net based backbone outperforms ResNet or Inception-Net variants in this work. Specifically, the 3D Multi-Fiber Units can enhance the model efficiency while effectively reducing the number of computations. In detail, we extract the feature map before the pooling layer, then add a 1×1×1 convolution layer to reduce the channel size to 128, and save it for the subsequent information aggregation process in the proposed BA-GCN. We refer to this feature maps as X3D in this work. Notably, the X3D has the same size as X2D, with D × 128 × 7 × 7. 3.3.2 Bilateral adaptive graph reasoning module Given the feature maps extracted from UC-MIL as 2D level’s information (X2D) and the feature maps extracted from MF-Net Backbone as 3D level’s information (X3D), we propose a bilateral adaptive graph to aggregate the features from both 2D and 3D levels. In detail, a graph reasoning module is achieved by information exchange and propagation among different granularity levels of vertices. Additionally, our graph structure and the edge relationship are adaptively learnt during the reasoning process according to the 2D and 3D level features’ own information. Thus, a bilateral adaptive graph representation can be learnt in an input-dependent way, rather than predefined hand-craft ones (Di et al., 2021, Aviles-Rivero et al., 2022, Liu et al., 2021a). Classic Graph Convolution We begin with a review of classic graph convolution. Given a graph G = (V, E), the normalised Laplacian matrix is defined as L=I−D−12AD−12, where I is the identity matrix, A is the adjacency matrix, and D is a diagonal matrix representing the degree of each vertex in V, such that Dii=∑jAi,j. Because the graph’s Laplacian is a symmetric and positive semi-definite matrix, L may be diagonalised using the Fourier basis U∈RN×N, resulting in L=UΛUT. Thus, the Fourier space spectral graph convolution of i and j may be described as i∗j=U((UTi) ⊙ (UTj)). The columns of U correspond to the orthogonal eigenvectors U=[u1,…,un], and Λ=diag([λ1,…,λn])∈RN×N is a diagonal matrix with non-negative eigenvalues. Due to the fact that U is not a sparse matrix, this operation is computationally inefficient. Defferrard et al. (2016) hypothesised that the convolution operation on a graph may be characterised by constructing spectral filtering with a kernel gθ in Fourier space through a recursive Chebyshev polynomial. The filter gθ is parameterised as a Chebyshev polynomial expansion of order K, such that gθ(L)=∑kθkTk(Lˆ), where θ∈RK is a vector of Chebyshev coefficients, and Lˆ=2L/λmax−IN is the rescaled Laplacian. Tk∈RN×N is the Chebyshev polynomial of order K. Kipf and Welling (2017) further simplified the graph convolution to gθ=θ(Dˆ−12AˆDˆ−12), where Aˆ=A+I, Dˆii=∑jAˆij, and θ are the only remaining Chebyshev coefficients. The corresponding graph Laplacian adjacency matrix Aˆ is handcrafted, causing the model to learn a specific long range context pattern rather than the input-related one (Li et al., 2020e). Thus, we refer to the classic graph convolution (Kipf and Welling, 2017) as handcrafted input-independent graph convolution.Fig. 4 Overview of the proposed BA-GCN, Bilateral Adaptive Graph Convolution (BA-GConv) and Bilateral Adaptive Adjacency Matrix (A~). Bilateral Adaptive Graph Convolution Given X2D∈RN2d×C and X3D∈RN3d×C, where C is the channel size; N2d=H2d×W2d×D and N3d=H3d×W3d×D are the number of spatial locations of 2D and 3D level of input features, which are referred to as the number of vertices. Note that H, W and D represent the height, width, and depth of the corresponding level of feature map, respectively. Firstly, we construct the bilateral adjacency matrix (A~) in an adaptive way. The vertices of 2D and 3D (X2D, X3D) contribute to the adjacency matrix construction concurrently and adaptively. In detail, we stack them together and represent it as Xall∈R(N3d+N2d)×C, which is regarded as the input vertices of BA-GConv (shown in Equation. (12)). Then, we implement two matrices (Λ~c and Λ~s) to execute channel-wise attention on the dot-product distance and to quantify spatially weighted relations between various input vertices embeddings, respectively. For example, Λ~c(Xall)∈RC×C is the matrix that contains channel-wise attention on the dot-product distance of the input vertex embeddings; Λ~s(Xall)∈R(N3d+N2d)×(N3d+N2d) is the spatial-wise weighting matrix, measuring the spatial relationships among different vertices. (9) Λ~c(Xall)=(MLP(Poolc(Xall)))T⋅(MLP(Poolc(Xall))), where ⋅ denotes matrix product; Poolc(⋅) is the global max pooling for each vertex embedding; MLP(⋅) is a multi-layer perceptron with one hidden layer. On the other hand, (10) Λ~s(Xall)=(Conv(Pools(Xall)))⋅(Conv(Pools(Xall)))T, where Pools(⋅) represents the global max pooling for each position in the vertex embedding along the channel axis; Conv(⋅) is a 1 × 1 convolution layer. The data-dependent adjacency matrix A~ is given by spatial and channel attention-enhanced input vertex embeddings. We initialise the bilateral adjacency matrix A~∈R(N3d+N2d)×(N3d+N2d) as: (11) A~=ψ(Xall,Wψ)⋅Λ~c(Xall)⋅ψ(Xall,Wψ)T+ ζ(Xall,Wζ)⋅ζ(Xall,Wζ)T⊙Λ~s(Xall), where ⋅ represents matrix product; ⊙ denotes Hadamard product; ψ(Xall,Wψ)∈R(N2d+N3d)×C and ζ(Xall,Wζ)∈R(N2d+N3d)×C are both linear embeddings; Wψ and Wζ are learnable parameters. Fig. 4 shows a detailed demonstration of the bilateral adjacency matrix A~. Please note that the different granularity levels of relationships among vertices from 2D and 3D (X2D, X3D) are exploited in this bilateral graph, where the graph is adaptively built up according to the multi-granularity vertices’ own correlations in a data-dependent way. With the constructed A~, the normalised Laplacian matrix is given as L~=I−D~−12A~D~−12, where I is the identity matrix; D~ is a diagonal matrix that represents the degree of each vertex, such that D~ii=∑jA~i,j; notably a softmax is applied on A~ for normalised adjacency weights. We calculated degree matrix D~ with the same way that is used in (Meng et al., 2021a, Li et al., 2020e), to override the computation overhead. Given computed L~, with Xall as the input vertex embeddings, we formulate the single-layer BA-GConv as : (12) Y=σ(L~⋅Xall⋅WG)+Xall, where WG∈RC×C denotes the trainable weights of the BA-GConv; σ is the ReLu activation function. Additionally, we include a residual connection to preserve the features of input vertices. Y∈R(N3d+N2d)×C is the output vertex features. Empirically, Three layers of the proposed BA-GConv with residual connections build up a graph reasoning module (BA-GCN shown in Fig. 4). After the BA-GCN, a convolution layer is added to reduce the channel size to one. Two layers of MLP with ReLu and Softmax as the activation functions respectively are used to aggregate the output vertices features and predict the final patient-level diagnosis probability. 4 Experiments 4.1 Datasets In this work, we perform experiments on three currently largest publicly available COVID-19 CT datasets: CC-CCII (Zhang et al., 2020), MosMed (Morozov et al., 2020) and COVID-CTset (Rahimzadeh et al., 2021). All of the three datasets are used in PNG format in this work. The total number of slices per CT volume ranges from 16 to 375. They are utilised to evaluate the learning ability and generalisation ability of our proposed model, respectively. The details of these three datasets w.r.t. two evaluation settings are shown in Table 1. The CC-CCII (Zhang et al., 2020) dataset contains three classes of NCP, CP and Normal and the other two datasets only contain two classes of NCP and Normal. We evaluate the learning ability of our proposed model on CC-CCII (Zhang et al., 2020) dataset. On the other hand, we evaluate the generalisation ability of our proposed model . Firstly, in order to eliminate the effect of imbalanced data class distribution, we combine the Normal class’s data of CC-CCII (Zhang et al., 2020) with all of MosMed (Morozov et al., 2020) dataset as Train & Val dataset. Then, the COVID-CTset (Rahimzadeh et al., 2021) dataset is treated as the External Test dataset. We have shown this data setting in the middle of Table 1. We elaborate the details of each datasets below. • CC-CCII (Zhang et al., 2020). The original CC-CCII dataset contains a total of 617,775 slices of 6,752 CT volume from 4,154 patients. However, it has several problems, such as corrupted data, duplicated and noisy slices, incomplete slices, non-unified data type, etc.. Please see Fig. 5 for details. Considerable effort has been made to build a clean dataset for training and evaluation. We have manually checked the whole dataset and removed the noisy data (damaged, duplicated and non-unified). Note that we only use complete scans with volume scan per patient to avoid information leakage during training and evaluation. After addressing the above problems, we build a clean CC-CCII dataset, which consists of 172,314 slices of 2,355 scans from 2,355 patients (shown in Table 1). Apart from the issues above, CC-CCII provided pre-segmented CT slices only but without original CT slices for part of the dataset. For example, in the clean CC-CCII, 59,256 slices of 740 volume from 740 patients are pre-segmented, and the rest 113,058 slices of 1,615 scans from 1,615 patients are not. Our experimental results proved that lung segmentation pre-process is necessary for the task of COVID-19 classification, especially for models trained on CC-CCII (Zhang et al., 2020) datasets. The details of the potential dataset issue related to the lung segmentation pre-process are discussed in Section 7.1. Besides, some qualitative visualisation results, such as GradCAMs (Selvaraju et al., 2017), are shown in Fig. 10 to prove the importance of lung segmentation pre-process in this task. To address the non-segmentation problem, we segmented the lungs of the non-segmented slices with our trained model. Compared with the pre-segmented lung slices of CC-CCII (Zhang et al., 2020), our model can segment more accurate lung regions. Qualitative results and comparisons are shown in Fig. 6. As illustrated, our segmentation can generate a more smooth lung boundary and conduct fewer false positive predictions. • MosMed (Morozov et al., 2020). MosMed dataset was collected from March 2020 to April 2020, within the outpatient CT centres in Moscow outpatient clinics, Moscow, Russia. The CT scans were performed on Canon (Toshiba) Aquilion 64 units with standard scanner protocols and 8 mm inter-slice distance. The dataset contains 36,753 slices of 1,110 volume from 1,110 patients. Specifically, 28,188 slices of 856 volume are NCP cases, and the rest 8,565 slices of 254 volume are Normal. Additionally, 50 CT volume were annotated on the region of infection areas such as GGO and consolidation. However, the ground truth of lung region segmentation is not provided. We segmented the lung regions of all the slices with our trained segmentation model and used the cropped slices as the clean dataset for the COVID-19 classification task. Please note that the data were provided in NIfTI format by Morozov et al. (2020), which were converted to PNG format, where a window (window center: -600HU, window width: 1200HU) was applied for re-scaling and normalising the pixel values. • COVID-CTset (Rahimzadeh et al., 2021). COVID-CTset dataset was collected from Negin radiology located at Sari in Iran between March 5th to April 23rd, 2020. This medical center uses a SOMATOM Scope model and syngo CT VC30easyIQ software version for capturing and visualising the lung HRCT radiology images from the patients. The dataset contains 63,849 slices of 377 volumes from 377 patients. Specifically, 15,589 slices of volumes scans are NCP cases, and the rest 48,260 slices of volumes scans are Normal. We randomly select 95 out of 282 Normal volumes to construct a balanced external test dataset. Again the ground truth of lung region segmentation is not provided. Thus, pre-segmentation is performed with our trained segmentation model to build a clean dataset with cropped slices. Table 1 Descriptions of the three COVID-19 CT datasets. Cleaned CC-CCII (Zhang et al., 2020), MosMed (Morozov et al., 2020) and COVID-CTset (Rahimzadeh et al., 2021) are three currently largest publicly available COVID-19 CT datasets. # Patient and # Slices represent the number of patient and slices, respectively. Train & Val represent the subset that contains train and validation datasets. Note that we randomly select 10% of Train & Val as the validation datasets. Datasets Classes # Patients # Slices Train & Val Test Train & Val Test CC-CCII NCP 414 133 24,255 10,330 CP 773 186 59,080 12,509 Normal 675 174 50,874 15,266 Total 1,862 493 134,209 38,105 MosMed + NCP 856 95 28,188 15,589 CC-CCII + Normal 929 95 59,439 12,718 COVID-CTset Total 1,785 190 97,627 28,307 Fig. 5 Examples of problematic slices from the original CC-CCII (Zhang et al., 2020) dataset. Those noisy data will inevitably introduce perturbations into both the lung segmentation task and the COVID-19 diagnosis task. 4.2 Annotation of COVID-19 CT images As discussed in Section 2.4, previous methods, such as (Wu et al., 2021a) and Goncharov et al. (2021), pre-trained the lung segmentation model from non-COVID datasets (i.e. cancer nodule segmentation datasets: NSCLC (Kiser et al., 2020) and LUNA16 (Team, 2011)), then applied it to the COVID-19 CT scans. The domain gap between different datasets would cause significantly performance drop. For example, the GGO regions are typical characterises of NCP cases, which is an unseen feature in the cancer nodule dataset. Thus, their pre-trained models are likely to treat it as background (similar examples are shown in the top left and top right of Fig. 6). To address the challenges and train a robust lung segmentation model, four trained medical students (trainee doctors after training on the annotation tasks) from the University of Liverpool manually annotated 7,768 slices of NCP, CP, and Normal scans from CC-CCII (Zhang et al., 2020) datasets. In detail, the boundaries of the left and right lungs are traced via Labelme (Wada, 2016) annotation tool. Among the annotated 7,768 slices, 6,045 slice of 190 patients are NCP, 1,202 slices of 10 patients are CP, 521 slices of 10 patients are Normal. In this way, our annotated slices contain NCP, CP and Normal examples, which addresses the domain gap between the train and test dataset. 4.3 Evaluation metrics Segmentation Metrics Typical segmentation metrics, such as Dice similarity score (Dice), Mean Absolute Error (MAE) and Balanced Accuracy (B-Acc), are applied. 95% confidence intervals were calculated using 2000 sample bootstrapping for Dice, MAE, and B-Acc. Specifically, B-Acc is the mean value of Sensitivity and Specificity; MAE is used to measure the pixel-wise error between the segmentation and ground truth. MAE is defined as: (13) MAE=1w×h∑xw∑yh|Sp(x,y)−Sgt(x,y)|, where, w and h are the width and height of the ground truth GTs, and (x, y) denotes the coordinate of each pixel in GTs. Classification Metrics Typical classification metrics, such as Sensitivity, Specificity, F1 score (F1), Precision, Receiver Operating Characteristic Curves (ROC Curve), Area Under the ROC Curve(AUROC), are used for the evaluation of classification. In particular, F1 is introduced to eliminate the interference of data imbalance. 95% confidence intervals were calculated using De Long’s method (DeLong et al., 1988) for AUROC and using 2000 sample bootstrapping for Sensitivity, Specificity, F1 and Precision. 4.4 Experimental details In this section, we describe the experimental implementation details for the lung segmentation and COVID-19 classification tasks, respectively. All the training processes are performed on an Amazon Web Services p3.8xlarge node with four Tesla V100 16GiB GPUs and our workstation with four GEFORCE RTX 3090 24GiB GPUs. All the test experiments are conducted on a local workstation with Intel(R) Xeon(R) W-2104 CPU and Geforce RTX 2080Ti GPU. Notably, we have conducted extensive experiments to evaluate the sensitivity of the hyper-parameters, where γ has been set at 0.1, 0.05, 0.02, 0.01, 0.005, and T has been set at 2, 4, 6, 8, 10. In conclusion, we found no significant difference in diagnostic performance with paired t-test (p > 0.05) in the two evaluation settings, which proves that our model is robust to the hyper-parameters. Thus, we set the value of γ and T at 0.02 and 8 empirically, respectively. 4.4.1 Lung segmentation Implementation Details The original slice image is resized into 224 × 224 from 512 × 512by bilinear interpolation for CT slices and by nearest neighbour interpolation for binary annotation masks. To augment the dataset, we randomly rotate and horizontally flip the training dataset with the probability of 0.3. The rotation ranges from −30 to 30 degree. Besides, a random crop of size 112 × 112 are also applied both on the input image and ground truth during the training. Among all of our annotated data, 60% of which are randomly selected as Train dataset, 10% are Val dataset and 30% are Test dataset. The network is trained end-to-end by an Adam optimiser (Kingma and Ba, 2014) for around 400 epochs, with a start learning rate of 0.01 and a cosine decay schedule (Loshchilov and Hutter, 2017). The batch size is set at 126. We adopt standard Dice Loss (Milletari et al., 2016) for training the lung segmentation model. 4.4.2 COVID-19 classification Implementation Details. The input image size is 224 × 224 after lung segmentation. Similarly, to augment the dataset, we randomly rotate, horizontally and vertically flip the training dataset with the probability of 0.3. The rotation ranges from −30 to 30 degree. 10% of the Train & Val dataset are randomly selected as the validation dataset. The network is trained end-to-end for 400 epochs, with a start learning rate of 1e-4 and a cosine decay schedule (Loshchilov and Hutter, 2017). The optimiser is an Adam optimiser (Kingma and Ba, 2014), the batch size is set at 48 and 36, for 2D and 3D COVID-19 diagnosis training, respectively. We adopt standard Cross Entropy Loss for both 2D and 3D COVID-19 classification respectively. Fig. 6 Qualitative comparison of pre-segmented slices and our segmentation results on CC-CCII (Zhang et al., 2020) dataset. The top row is the pre-segmented slices that are provided by CC-CCII and the bottom row shows our segmentation examples on un-segmented cases. Red bounding boxes indicate the pre-segmented slices’ false positive or false negative predictions. In particular, the top left and top right examples illustrate a typical false negative prediction, where the potential GGO regions may be treated as background, as the patient-level label for this case is COVID-19 positive. Such false negative segmentation would perturb the subsequent COVID-19 classification model training because there is no infection areas or diagnosis characteristics left in the segmented CT slices. On the other hand, our segmentation model can produce a complete lung region, even when there is a large number of infection regions (e.g. GGO). Please note that CC-CCII only provides the pre-segmented CT slices without the original ones, thus we cannot intuitively compare the segmentation results with the same examples.. Table 2 Quantitative segmentation results of the lung regions on CT slices. The performance is reported as Dice (%), B-Acc (%) and MAE (%). 95% confidence intervals are presented in brackets. We performed experiments with classic segmentation methods such as U-Net (Ronneberger et al., 2015), U-Net++ (Zhou et al., 2019), and cutting-edge methods such as PraNet (Fan et al., 2020a), RBA-Net (Meng et al., 2020a), CABNet (Meng et al., 2020b), GRB-GCN (Meng et al., 2021c) and BI-Gconv (Meng et al., 2021a). Notably, we sampled 120 vertices for CABNet (Meng et al., 2020b) and RBA-Net (Meng et al., 2020a) to construct a smooth boundary. Methods Metrics Dice (%)↑ B-Acc (%)↑ MAE (%)↓ U-Net 95.7 (93.2, 97.6) 96.9 (95.0, 98.4) 1.49 (1.12, 1.68) textitU-Net++ 94.1 (92.2, 96.0) 95.0 (93.2, 97.5) 1.98 (1.56, 2.23) PraNet 95.2 (94.0, 96.6) 96.0 (95.1, 98.0) 1.55 (1.38, 1.68) RBA-Net 96.2 (95.2, 98.0) 96.8 (95.9, 98.0) 1.45 (1.29, 1.56) CABNet 95.4 (93.8, 96.7) 96.4 (94.7, 98.1) 1.60 (1.42, 1.78) GRB-GCN 96.6 (94.9, 97.9) 96.7 (95.8, 97.9) 1.50 (1.32, 1.68) BI-GConv 96.3 (94.8, 98.0) 96.5 (94.7, 98.2) 1.52 (1.34, 1.69) 5 Results Table 3 Quantitative comparisons between Ours and previous 3D CT based COVID-19 diagnosis methods, such as CCT-Net (Goncharov et al., 2021), C19C-Net (Bai et al., 2020b), COVNet (Li et al., 2020b), DeCoVNet (Wang et al., 2020b), ASCo-MIL (Han et al., 2020)). The performance is reported as F1 (%), Precision (%), Specificity (%), Sensitivity (%), AUROC (%). 95% confidence intervals are presented in brackets. Methods Learning ability Generalisation ability F1 (%)↑ Precision(%)↑ Specificity(%)↑ Sensitivity(%)↑ AUROC(%)↑ F1 (%)↑ Precision(%)↑ Specificity(%)↑ Sensitivity(%)↑ AUROC(%)↑ CCT-Net 76.8 (72.9, 80.6) 83.5 (81.0, 86.1) 84.4 (81.2, 87.4) 78.1 (74.6, 81.5) 96.1 (94.4, 97.1) 71.6 (62.8, 79.2) 86.6 (77.8, 94.2) 90.5 (84.2, 96.0) 61.1 (50.5, 71.3) 85.9 (80.1, 90.7) C19C-Net 66.2 (61.5, 70.8) 71.2 (66.2, 75.9) 80.0 (76.8, 82.9) 70.2 (66.3, 74.0) 86.7 (83.9, 88.4) 70.4 (63.5, 76.2) 56.8 (48.7, 64.3) 29.5 (20.2, 38.8) 92.6 (87.4, 97.8) 80.0 (73.2, 86.0) COVNet 59.6 (54.9, 64.6) 73.7 (64.5, 79.4) 75.5 (71.5, 78.7) 68.0 (63.9, 72.0) 87.5 (84.8, 89.3) 33.6 (22.4, 43.7) 70.0 (52.0, 85.7) 90.5 (84.0, 96.0) 22.1 (14.1, 30.6) 71.5 (63.7, 78.6) DeCoVNet 91.2 (88.5, 93.7) 91.6 (89.1, 94.0) 95.0 (93.4, 96.5) 91.3 (88.6, 93.7) 97.5 (96.7, 98.6) 68.8 (59.6, 76.2) 87.1 (78.2, 94.9) 91.6 (85.7, 96.7) 56.8 (46.5, 67.0) 85.1 (79.2, 90.2) ASCo-MIL 76.5 (72.5, 80.6) 79.6 (76.1, 82.9) 86.2 (83.8,88.4) 77.9 (74.2, 81.5) 91.2 (88.9, 93.0) 60.7 (50.7, 69.7) 88.0 (78.4, 96.1) 93.7 (88.5, 97.9) 46.3 (36.4, 56.6) 82.1 (75.9, 87.8) Ours 94.9 (93.0, 96.8) 95.1 (93.3, 96.9) 97.1 (95.9, 98.2) 94.9 (93.1, 96.8) 98.7 (97.6, 99.4) 88.0 (82.3, 92.7) 96.3 (91.5, 100.0) 96.8 (92.9, 100.0) 81.1 (72.9, 88.7) 91.8 (84.6, 93.3) Fig. 7 ROC Curve comparisons between Ours and previous 3D CT based COVID-19 diagnosis methods, such as CCT-Net (Goncharov et al., 2021), C19C-Net (Bai et al., 2020b), COVNet (Li et al., 2020b), DeCoVNet (Wang et al., 2020b), ASCo-MIL (Han et al., 2020)). Two evaluation settings of learning Ability and Generalisation Ability are presented. 5.1 Lung segmentation Fig. 6 shows the qualitative lung segmentation result of the pre-segmented slices (provided by CC-CCII) and our segmentation results on CC-CCII. Table 2 shows the quantitative results of classic segmentation models, such as U-Net (Ronneberger et al., 2015), U-Net++ (Zhou et al., 2019), and cutting-edge methods such as PraNet (Fan et al., 2020a), RBA-Net (Meng et al., 2020a), CABNet (Meng et al., 2020b), GRB-GCN (Meng et al., 2021c) and BI-Gconv (Meng et al., 2021a). There are no significant differences between these models. Among them, GRB-GCN (Meng et al., 2021c) achieves the best performance of 96.6% Dice, outperforming U-Net (Ronneberger et al., 2015) and U-Net++ (Zhou et al., 2019) by 0.9 and 2.7%. 5.2 COVID-19 diagnosis This section provides the classification results in two evaluation settings with the pre-segmented COVID-19 CT data. Firstly, we train, validate and test our model on CC-CCII dataset (seen data) only, where there are three classes such as Normal, NCP, and CP. In this way, the learning ability of our model can be illustrated on the seen data. Secondly, in order to address the unbalanced classes issue of Mosmed, we combine the Normal class’s data from CC-CCII and all of the data from MosMed, to train and validate our model, while test on COVID-CTset (unseen data). There are two classes in this setting, such as Normal and NCP. In this way, we demonstrate the generalisation ability of our model on the unseen data. Generalisability is essential for the real-world COVID-19 diagnosis task, because of different domains of data w.r.t. scanning machine types, protocol standards, data sources. The details of data settings in these two schemes can be found in Table 1. The quantitative comparison results on respective test datasets of two evaluation settings are shown in Table 3, where previous 3D CT based COVID-19 diagnosis methods such as CCT-Net (Goncharov et al., 2021), C19C-Net (Bai et al., 2020b), COVNet (Li et al., 2020b), DeCoVNet (Wang et al., 2020b), ASCo-MIL (Han et al., 2020) are presented. Notably, their results are reproduced by using their open-source code, and experiments are conducted under the same settings as Ours with our pre-segmented lung CT images. Fig. 8 Qualitative comparisons between Ours, C19C-Net (Bai et al., 2020b), COVNet (Li et al., 2020b), ASCo-MIL (Han et al., 2020) and DeCoVNet (Wang et al., 2020b). Specifically, attention heatmaps visualisation of Grad-CAM on NCP patients are presented in each row. Ours has a more precise and comprehensive activate area that encompasses more diagnosis characteristics, including GGO, multi-focal patchy consolidation and bilateral patchy shadows. 5.2.1 Learning ability Table 3 shows the quantitative comparison results in terms of the learning ability between Ours and previous 3D CT based COVID-19 diagnosis methods on CC-CCII dataset. Ours obtains an average of 94.9 F1, which outperforms the pooled 2D slice features based methods, such as CCT-Net (Goncharov et al., 2021), C19C-Net (Bai et al., 2020b), COVNet (Li et al., 2020b) by 23.6%, 43.4% and 59.2 %, respectively. In addition, Ours outperforms the 3D level CNN based approaches DeCoVNet (Wang et al., 2020b) by 4.1%, outperforms the attention score based MIL method ASCo-MIL (Han et al., 2020) by 24.1%. Fig. 7 demonstrates the ROC Curve comparison between the aforementioned methods. Ours achieves the best AUROC of 98.7%. Notably, the macro-averaged performance (aka unweighted mean of per-class performance) of three classes with one vs rest calculation setting1 on learning ability is presented in Table 3 and Fig. 7. 5.2.2 Generalisation ability To evaluate the generalisation ability of the proposed model, we evaluate and compare Ours with previous 3D CT based COVID-19 diagnosis approaches with external test data (unseen data). The generalisation ability part of Table 3 shows the quantitative results. Ours achieves the best F1 of 88.0%, which outperforms the cutting-edge COVID-19 diagnosis methods DeCoVNet (Wang et al., 2020b) and ASCo-MIL (Han et al., 2020) by 27.9% and 45.0%. Fig. 7 shows the ROC Curve comparison. Ours achieves the best AUROC of 91.8%. 5.2.3 Attention heat maps visualisation Fig. 8 demonstrates the attention heat maps generated by using the gradient-weighted class activation mapping (Grad-CAM) (Selvaraju et al., 2017). Specifically, Grad-CAM results on different slices of different NCP patients are presented in each row of the figure. We compare Ours with previous methods such as C19C-Net (Bai et al., 2020b), COVNet (Li et al., 2020b), ASCo-MIL (Han et al., 2020), CCT-Net (Goncharov et al., 2021), and present them in each column. Ours has a more accurate and comprehensive activate area that covers more diagnosis characteristics, such as GGO, multi-focal patchy consolidation and bilateral patchy shadows, which are highlighted within red bounding box in the figure. Notably, all the compared methods in Fig. 8 adopted at least the same D slices as ours to make the inference and prediction. Specifically, C19C-Net (Bai et al., 2020b) and COVNet (Li et al., 2020b) used the same selected D slices, which is also aligned with their original implementation. ASCo-MIL (Han et al., 2020) and DeCoVNet (Wang et al., 2020b) used all of the slices in a CT scan to make the inference, thus includes the selected D slices. Table 4 Computational efficiency. Model size, FLOPs, and inference time of different 3D CT based COVID-19 diagnosis methods on a 224 × 224 ×D input volume. CCT-Net C19C-Net COVNet DeCoVNet Ours Params (M) 24.8 23.8 23.5 0.35 15.0 FLOPs (G) 67.1 39.0 65.8 28.9 35.0 Inference Time (s) 1.2 1.2 1.1 1.1 1.1 5.2.4 Computational efficiency Table 4 presents the number of parameters (M), floating-point operations (FLOPs) and inference time (s) of the compared models. Notably, ignoring the slices selection process of the first stage, we represent the proposed BA-GCN as Ours to compare with other methods in the Table 4. Ours adopted a light-weight backbone network of MF-Net to extract the 3D level of features, which leads to a relatively smaller model size as 15.0 M parameters. 6 Ablation study We conduct thorough ablation studies, and all the results demonstrate our model’s effectiveness. As an illustration, the ablation results for the lung segmentation and model components are elaborated as follows. Table 5 Ablation study of lung segmentation on CCT-Net (Goncharov et al., 2021), DeCoVNet (Wang et al., 2020b) and Ours. w/o Seg represents without lung segmentation pre-process; w/ Our seg represents adopting our fully supervised lung segmentation method. The performance is reported as F1 (%), AUROC (%). 95% confidence intervals are presented in brackets, respectively. Methods Learning ability Generalisation ability F1 (%)↑ AUROC (%)↑ F1 (%)↑ AUROC (%)↑ CCT-Net, w/o Seg 82.3 (80.0, 84.6) 97.2 (95.2, 98.5) 51.0 (48.7, 54.0) 65.3 (63.2, 68.1) CCT-Net 75.0 (73.0, 78.1) 95.1 (92.5, 97.7) 69.0 (66.9, 72.1) 83.8 (81.1, 85.9) CCT-Net, w/ Our seg 76.8 (72.9, 80.6) 96.1 (94.4, 97.1) 71.6 (62.8, 79.2) 85.9 (80.1, 90.7) DeCoVNet, w/o Seg 93.9 (91.0, 95.5) 99.2 (97.1, 99.8) 57.3 (55.8, 60.2) 76.5 (74.1, 78.8) DeCoVNet 88.7 (86.0, 90.3) 95.4 (93.2, 97.7) 66.7 (64.4, 68.9) 82.0 (80.0, 84.7) DeCoVNet, w/ Our seg 91.2 (88.5, 93.7) 97.5 (96.7, 98.6) 68.8 (59.6, 76.2) 85.1 (79.2, 90.2) Ours, w/o Seg 96.8 (94.7, 98.9) 99.4 (98.1, 99.7) 71.3 (69.2, 73.5) 84.3 (82.0, 86.6) Ours 94.9 (93.0, 96.8) 98.7 (97.6, 99.4) 88.0 (82.3, 92.7) 91.8 (84.6, 93.3) 6.1 Need of lung segmentation pre-process Lung segmentation is an essential pre-processing step in this task. Please note that the original CCT-Net (Goncharov et al., 2021) adopted a pre-trained lung segmentation model on other CT datasets (non-COVID) and the original DeCoVNet (Wang et al., 2020b) used an unsupervised approach to segment the lung regions as the pre-process for subsequent classification task. In this experiment, we used our pre-segmented lung CT images (w/ Our seg) to provide a more accurate cropped lung regions for their methods. Table 5 shows that w/ Our seg can boost their original classification performance of F1 by 2.4%, 2.8% and 3.8%, 3.1% in the Learning Ability and Generalisation Ability experiment settings, respectively. This can demonstrate the importance and the benefits of our fully-supervised lung-segmentation model in the task of 3D CT based COVID-19 classification. Additionally, Table 5 shows that the three methods without lung pre-segmentation (w/o Seg) can produce a better classification performance on the non-segmented CT data under the Learning Ability experiment setting, than the one with segmentation w/ Our Seg. However, the qualitative results (Fig. 10) prove that, such model trained on non-segmented data, can only learn a specific format pattern of different classes rather than the real radiographic diagnosis characteristics (i.e. GGO for NCP), because of specific scanning machine types, protocol standards, data sources of different classes. Also, due to the evaluation setting under Learning Ability of test on seen data, such specific format patterns also exist in the test dataset, which helps the models achieve ‘excellent’ classification results, rather than learning the real diagnosis features. Differently, under the experiment setting of Generalisation Ability, those methods w/o Seg conducts a terrible classification performance because an external test dataset (unseen data) is introduced to evaluate the trained model, where the aforementioned specific format patterns do not exist. This further demonstrates the importance of pre-segmentation, generalisation ability and external test dataset (unseen data) in this task. More visualisation comparisons and discussions related to this challenge are refereed to Section 7.1 and Fig. 10. 6.2 Model components This section presents the results of our ablation study on model components. We evaluate the effectiveness of the proposed UC-MIL, BA-GCN modules, and present the quantitative results in Table 6. Firstly, we remove the BA-GCN and keep the rest of our model, conduct Ours w/o BA-GCN in the Table. Secondly, we replace UC-MIL with random and symmetrical slice sampling rules to select the fixed number of slices for each CT scan in the same manner as (He et al., 2021a). In these two cases, the proposed bilateral graph model becomes an unilateral graph model, because there is no 2D feature information included in the vertices features. Specifically, for both evaluation settings (Learning and Generalisation Abilities), BA-GCN helps our model gain an average 9.4% performance boost w.r.t. F1 and 4.3% performance boost w.r.t AUROC; UC-MIL outperforms the hand-crafted slice sampling rules, e.g. random and symmetrical, by 12.5% and 11.9% F1 on average, respectively. Additionally, we conduct extensive experiments to evaluate the effectiveness of the proposed components inside UC-MIL and BA-GCN modules respectively, such as backbones, Uncertainty-Aware mechanism, Consensus-Assisted mechanism, BA-GConv layers, etc.. The experimental results are elaborated as follows, which prove their effectiveness. Table 6 Ablation study on the effectiveness of the proposed UC-MIL and BA-GCN. The performance is reported as F1 (%), AUROC (%). 95% confidence intervals are presented in brackets, respectively. Methods Learning ability Generalisation ability F1 (%)↑ AUROC (%)↑ F1 (%)↑ AUROC (%)↑ Ours w/o BA-GCN 82.6 (79.0, 86.0) 94.6 (92.9, 96.2) 81.0 (73.9, 87.2) 83.9 (77.1, 89.8) Ours w/o UC-MIL (random) 91.5 (89.1, 93.3) 97.4 (95.5, 98.8) 72.6 (70.9, 74.1) 87.0 (85.2, 88.7) Ours w/o UC-MIL (symmetrical) 91.9 (90.1, 93.3) 97.9 (95.8, 98.8) 73.1 (71.9, 75.2) 86.8 (84.8, 88.0) Ours 94.9 (93.0, 96.8) 98.7 (97.6, 99.4) 88.0 (82.3, 92.7) 91.8 (84.6, 93.3) Table 7 Ablation study on the effectiveness of the UC-MIL’s backbone networks and the proposed Uncertainty-aware Consensus-assisted mechanism. Specifically, we respectively replace the proposed UC-MIL to another two classic MIL methods, such as (Campanella et al., 2019) (w/ Instance-based) and Ilse et al. (2018) (w/ Embedding-based). The performance is reported as F1 (%), AUROC (%). 95% confidence intervals are presented in brackets, respectively. Methods Learning ability Generalisation ability F1 (%)↑ AUROC (%)↑ F1 (%)↑ AUROC (%)↑ Backbone w/ ResNet34 93.2 (90.9, 95.5) 98.0 (96.1, 99.1) 86.8 (84.7, 88.1) 90.5 (88.7, 92.0) w/ ResWide50 93.3 (91.7, 95.0) 97.7 (95.4, 98.9) 86.0 (84.2, 88.1) 90.2 (88.4, 92.3) w/ EfficientNetB3 90.2 (88.1, 92.3) 96.0 (93.9, 98.0) 84.6 (82.7, 86.1) 88.1 (86.5, 89.7) w/ Res2Net50 91.7 (89.9, 93.2) 96.8 (94.7, 98.0) 85.0 (83.1, 87.2) 88.7 (86.6, 89.9) Ours 94.9 (93.0, 96.8) 98.7 (97.6, 99.4) 88.0 (82.3, 92.7) 91.8 (84.6, 93.3) Component w/o Uncertainty 92.2 (90.1, 94.1) 97.0 (95.8, 98.1) 86.2 (94.9, 88.3) 90.2 (88.1, 92.1) w/o Consensus 92.2 (90.4, 94.6) 97.7 (95.1, 98.6) 85.8 (83.3, 87.0) 89.4 (87.2, 90.5) w/ Instance-based 88.7 (86.0, 90.1) 95.5 (93.3, 96.8) 81.5 (80.0, 83.1) 85.7 (83.3, 87.1) w/ Embedding-based 89.9 (87.4, 91.2) 95.9 (93.3, 97.6) 83.0 (81.0, 85.2) 87.0 (85.1, 88.9) Ours 94.9 (93.0, 96.8) 98.7 (97.6, 99.4) 88.0 (82.3, 92.7) 91.8 (84.6, 93.3) 6.2.1 UC-MIL Backbone Network We conduct experiments to evaluate the effectiveness of different backbone models in the proposed UC-MIL. We adopt several classic 2D classification backbones, such as ResNet (He et al., 2016) variants (e.g. 18, 34, 50, 101), and cutting-edge classification backbones such as ResWide (Zagoruyko and Komodakis, 2016) variants (e.g. 50, 101), ResNeXt (Xie et al., 2017) variants (e.g. 50, 101), EfficientNet (Tan and Le, 2019) series (e.g. B0, B3, B5, B7) and Res2Net (Gao et al., 2019) variants (e.g. 50, 101). For each model’s variants, we present the best performance in Table 7 for an intuitive comparison. Ours achieves the best performance of 94.9% and 88.0% F1 with ResNeXt50 and ResNet18 as the backbone in Learning Ability and Generalisation Ability settings, respectively. Uncertainty & Consensus Mechanism We evaluate the effectiveness of the proposed Uncertainty-aware mechanism and Consensus-assisted mechanism respectively. In detail, we remove each of them correspondingly and remain the rest of the model unchanged, which are represented as w/o Uncertainty and w/o Consensus in Table 7. As a result, the reliable slices selection process will rely on the ranked order of consensus-assisted instance probability (Pxi,jr(C), xi,jr∈Ω) and the ranked order of uncertainty-aware instance probability (PI~xi,j(C)), respectively. Specifically, Uncertainty-aware and Consensus-assisted modules boost the performance of F1 by 2.9% and 2.9% on Learning Ability and 2.0% and 2.6% on Generalisation Ability, respectively. Multiple Instance Learning To further verify the usefulness of the proposed UC-MIL, we respectively replace it with another two classic MIL methods, Campanella et al. (2019) (w/ Instance-based) and Ilse et al. (2018) (w/ Embedding-based), shown in Table 7. Notably, w/ Instance-based can be seen as our UC-MIl but without Uncertainty and Consensus mechanisms. As for w/ Embedding-based (Ilse et al., 2018), as we discussed in Section 2.1, all previous 3D CT based COVID-19 diagnosis methods (Li et al., 2021b, Chikontwe et al., 2021, Han et al., 2020) adopted its attention scoring system. Specifically, we adopted the same backbone framework as Ours, but a trainable attention score-based pooling mechanism from (Ilse et al., 2018). In detail, two fully-connected layers with Softmax as the activation functions are applied to learn a weighted average of instances (low-dimensional embeddings). We trained those two models (Campanella et al., 2019, Ilse et al., 2018) with all of the training CT slices/instances under the same experiment settings as ours. In Table 7, Ours outperforms w/ Instance-based and w/ Embedding-based by an average of 7.5% and 5.8% F1 on both evaluation settings. Notably, for w/ Instance-based (Campanella et al., 2019), we selected the top D instances (CT slices) according to the ranking of the predicted probability of instances, which is straightforward to implement and has been adopted by previous MIL methods (Campanella et al., 2019, Su et al., 2022, Kraus et al., 2016). On the other hand, for w/ Embedding-based (Ilse et al., 2018), we used the ranked attention weights to select the corresponding top D instances, which is similar to the previous methods (Ilse et al., 2018, Li et al., 2021a, Shao et al., 2021). Those selected top D instances were then used as the 3D input for our proposed BA-GCN. As we have noted, previous instance-level MIL methods yield promising classification results in the seen data (i.e. the evaluation of Learning Ability), but poor on unseen data (i.e. the evaluation of Generalisation Ability). On the other hand, Ours can achieve more consistent results on the unseen data, with the benefit of the proposed uncertainty-aware and consensus-assisted mechanisms. Table 8 Ablation study on the effectiveness of the BA-GCN’s backbone networks and the proposed Bilateral Adaptive Graph Convolution. Specifically, we respectively replace the proposed BA-GConv layer to another three cutting-edge graph reasoning based classification layers, such as SGR (Liang et al., 2018), DualGCN (Zhang et al., 2019) and GloRe (Chen et al., 2019). The performance is reported as F1 (%), AUROC (%). 95% confidence intervals are presented in brackets, respectively. Methods Learning ability Generalisation ability F1 (%)↑ AUROC (%)↑ F1 (%)↑ AUROC (%)↑ Backbone w/ 3D-ResNet50 93.2 (91.4, 95.1) 98.0 (96.3, 98.9) 87.1 (86.2, 88.7) 91.0 (89.7, 92.8) w/ 3D-ResNeXt50 93.2 (91.7, 95.5) 97.7 (95.8, 98.8) 87.3 (85.8, 89.9) 91.2 (89.7, 93.0) w/ 3D-EfficientNetB0 90.7 (87.1, 91.3) 95.5 (92.8, 97.0) 85.2 (82.5, 84.0) 89.9 (87.0, 91.2) Ours (w/ MF-Net) 94.9 (93.0, 96.8) 98.7 (97.6, 99.4) 88.0 (82.3, 92.7) 91.8 (84.6, 93.3) Component w/ SGR 90.3 (88.2, 92.8) 95.5 (93.0, 97.7) 85.7 (83.6, 87.6) 86.0 (83.9, 88.0) w/ DualGCN 91.1 (89.4, 92.9) 96.0 (94.1, 97.8) 85.9 (83.8, 87.1) 86.3 (84.9, 87.7) w/ GloRe 90.8 (88.5, 92.0) 95.9 (93.1, 97.0) 86.1 (84.2, 88.2) 86.6 (84.1, 88.0) Ours 94.9 (93.0, 96.8) 98.7 (97.6, 99.4) 88.0 (82.3, 92.7) 91.8 (84.6, 93.3) 6.2.2 Bilateral adaptive graph convolution network Backbone Network. We conduct experiments to evaluate the effectiveness of different backbone models in the proposed BA-GCN. We adopt several classic classification backbones, such as 3D-ResNet (He et al., 2016) variants (e.g. 18, 34, 50), and cutting-edge classification backbones such as MF-Net (Chen et al., 2018), 3D-EfficientNet (Tan and Le, 2019) variants (e.g. B0, B3, B5) and 3D-ResNeXt (Xie et al., 2017) variants (e.g. 50, 101). For each model’s variants, we present the best performance in Table 8 for an intuitive comparison. Ours achieves the best performance of 94.9% and 88.0% F1 with MF-Net as the backbone in Learning Ability and Generalisation Ability settings, respectively. Graph Convolution To further verify the usefulness of the proposed BA-GCN, we respectively replace it to another three cutting-edge graph-based reasoning methods, such as SGR (Liang et al., 2018), DualGCN (Zhang et al., 2019) and GloRe (Chen et al., 2019), shown in Table 8. In detail, we retain the same input vertices (Xall) and replace the proposed BA-GConv layer to their corresponding graph convolution layers, where SGR makes use of the knowledge graph mechanism, DualGCN investigates the coordinate space and feature space graph convolution, and GloRe makes use of the projection and re-projection mechanisms to reason about the relationships of different regions. In this way, the compared GCNs will consider both 2D and 3D levels of information from the input vertices. Table 8 shows that Ours achieves more accurate and reliable results, and outperforms SGR, DualGCN and GloRe by an average of 3.9%, 2.9% and 3.4% F1 on both the evaluation settings. Fig. 9 CT slices are randomly selected from different patients. The top and bottom rows represent Normal and NCP classes, respectively. Red bounding box highlights the differences between the scanner beds in the two classes. 7 Discussion 7.1 Hidden challenges of the COVID-19 dataset CC-CCII (Zhang et al., 2020) is now the largest public available 3D CT dataset for the COVID-19 diagnosis, with patients’ CT scans of NCP, CP and Normal classes. Many previous methods (He et al., 2021a, Wu et al., 2021a, Tan and Liu, 2021, Hou et al., 2021) reported evaluation results on it, however, rarely discussed the importance of pre-segmentation process. In Table 5, we present a better quantification results of training the proposed model without pre-segmentation process (Ours, w/o Seg), than the one with the pre-segmentation (Ours). Similar circumstances are also observed with previous methods in Table 5, such as CCT-Net, w/o Seg and DeCoVNet, w/o Seg. However, the trained models without pre-segmentation may only learn a specific format pattern of different classes, rather than the true radiographic diagnosis characteristics (i.e. GGO for NCP), because of specific scanning machine types, protocol standards, data sources for different classes in the dataset. For example, in Fig. 9 we show ten randomly selected CT slices of different patients from Normal and NCP classes. The model can easily learn the difference between the specific scanner bed part of different classes (highlighted with red bounding box). To further prove the necessity of pre-segmentation in this task, we visualise the trained model’s attention heat maps, which are generated by using the Grad-CAM. In Fig. 10, it shows that the models without pre-segmentation (Ours, w/o Seg) look at other regions (e.g. scanner bed) rather than the diagnosis characteristics part (e.g. GGO) of the lungs in the NCP CT images. Fig. 10 Qualitative comparison of Grad-CAM on the same input with and without pre-segmentation step. Models without pre-segmentation (Ours, w/o Seg) attend to other regions (e.g. scanner bed) rather than the discriminatory parts (e.g. GGO) of the lung regions in the NCP CT images. 7.2 Limitations of the proposed model Our proposed model achieves accurate classification results on three largest public available CT dataset w.r.t. Learning Ability and Generalisation Ability evaluation settings. However, one limitation of our model is two-stage, which requires a relatively longer inference time or training time compared to other one-stage methods. This is because we proposed UC-MIL for 2D feature extraction and trustworthy slices selection on the first stage, then, we propose BA-GCN to extract 3D features, and aggregate the 2D and 3D information for a more comprehensive level of feature reasoning on the second stage. Such a design increases the diagnostic accuracy but also consumes more time to infer and train. Compared to the other methods in Table 3, Ours takes 30.0 more hours on average for training the first stage of UC-MIL due to MIL’s specific training mechanism. This is similar to ASCo-MIL (Han et al., 2020), which is also a MIL-based method. However, we believe the training model process is often one-off, while inference speed plays a more important role in evaluating the algorithm and applying to the real applications. Specifically, Ours requires approx 0.26 s more inference time per 3D CT volume for both evaluation settings on average. In addition, if ignoring the slice selection process in the first stage for both Ours and other compared methods, we have demonstrated in the Table 4 that all the methods have a similar inference time. However, w.r.t. the diagnosis of COVID-19, the diagnostic accuracy would matter more than the inference speed. This highlights the need of a trade-off between accuracy and running time when applying AI models to real world applications. On the other hand, our proposed UC-MIL works as the automatically reliable CT slices selection step in the first stage, rather than the handcrafted slice sampling rules or manual slices selection of previous methods. In other words, previous methods also belong to the two-stage pipeline, where they need to select CT slices in a handcrafted way in the first stage. However, our method can automatically work with raw CT images without any manually designed pre-processing steps. 7.3 Future work Future studies building on this work should may wish to focus on the first stage of reliable slices selection, as the second stage of graph-based 2D/3D feature reasoning processes will rely mainly on the selected slices and the 2D features from the first stage as the input. Consequently, a collection of noisy input slices will inevitably introduce noise into the second stage and in turn perturbing the training process. The ablation study experiments of UC-MIL and BA-GCN in Table 7 and Table 8 of the original manuscript further support this view, that is, unreliable slices of the first stage lead to lower performance in the diagnosis of second stage, especially in the generalisation ability evaluation. A potential concern of using automatically selected top D CT slices of UC-MIL as the 3D input for the 3D CNN backbone in BA-GCN could be that the non-adjacent top D CT slices may lack abundant spatial correlations along the channel axis, which may lead to insufficient usage of the potential of 3D CNN. To address this concern, we have experimentally demonstrated that such 3D input can be used to boost the COVID-19 diagnosis performance via the extracted 3D features in both Learning Ability and Generalisation Ability settings in TABLE. 8, compared to the one without 3D features (Ours w/o BA-GCN in Table 6). Also, the same circumstance occurred and has been observed by many previous CT-based COVID-19 diagnosis studies (He et al., 2021a, Tan and Liu, 2021, Wang et al., 2021, Fang et al., 2021, Ouyang et al., 2020, Li et al., 2020c), where they sampled a fixed number of slices from adjacent CT slices, to form a 3D input volume with non-adjacent CT slices. Moreover, they have all proved that such 3D volume can be used for 3D CNN to extract COVID-19 diagnosis-related features and also achieve satisfying results. An extensive analysis of the relations between 3D CNN and non-adjacent CT slices’ effectiveness will be of interest in future studies. 8 Conclusion We have proposed a novel and comprehensive framework for diagnosing COVID-19 using CT scans of an arbitrary number of slices. It takes advantage of both 2D and 3D features of CT images by utilising the proposed UC-MIL and BA-GCN modules. Our experiments have demonstrated that our framework can locate the diagnosis characteristics in both seen and unseen evaluation settings by the graph-based information aggregation of trustworthy 2D and 3D features. Our approach is anticipated to be widely applicable to real-world applications. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix A Supplementary data The following is the Supplementary material related to this article. MMC S1 . Data availability Data will be made available on request. Acknowledgments This study was funded by EPSRC Impact Acceleration Account (IAA) funding and Amazon Web Services . 1 https://scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html#sphx-glr-auto-examples-model-selection-plot-roc-py. Appendix A Supplementary material related to this article can be found online at https://doi.org/10.1016/j.media.2022.102722. ==== Refs References Amores J. Multiple instance classification: review, taxonomy and comparative study Artificial Intelligence 201 2013 81 105 Aviles-Rivero A.I. Sellars P. Schönlieb C.-B. Papadakis N. GraphXCOVID: explainable deep graph diffusion pseudo-labelling for identifying COVID-19 on chest X-rays Pattern Recognit. 122 2022 108274 Bai H.X. Hsieh B. Xiong Z. Halsey K. Choi J.W. Tran T.M.L. Pan I. Shi L.-B. Wang D.-C. Mei J. Performance of radiologists in differentiating COVID-19 from viral pneumonia on chest CT Radiology 2020 Bai H.X. Wang R. Xiong Z. Hsieh B. Chang K. Halsey K. Tran T.M.L. Choi J.W. Wang D.-C. Shi L.-B. Artificial intelligence augmentation of radiologist performance in distinguishing COVID-19 from pneumonia of other origin at chest CT Radiology 296 3 2020 E156 E165 32339081 Berthelot D. Carlini N. Goodfellow I. Papernot N. Oliver A. Raffel C.A. MixMatch: a holistic approach to semi-supervised learning Adv. Neural Inf. Process. Syst. 32 2019 Bridge J. Meng Y. Zhao Y. Du Y. Zhao M. Sun R. Zheng Y. Introducing the GEV activation function for highly unbalanced data to develop COVID-19 diagnostic models IEEE J. Biomed. Health Inf. 24 10 2020 2776 2786 Calderon-Ramirez S. Yang S. Moemeni A. Colreavy-Donnelly S. Elizondo D.A. Oala L. Rodríguez-Capitán J. Jiménez-Navarro M. López-Rubio E. Molina-Cabello M.A. Improving uncertainty estimation with semi-supervised deep learning for COVID-19 detection using chest X-ray images IEEE Access 2021 Campanella G. Hanna M.G. Geneslaw L. Miraflor A. Silva V.W.K. Busam K.J. Brogi E. Reuter V.E. Klimstra D.S. Fuchs T.J. Clinical-grade computational pathology using weakly supervised deep learning on whole slide images Nature Medicine 25 8 2019 1301 1309 Chao H. Fang X. Zhang J. Homayounieh F. Arru C.D. Digumarthy S.R. Babaei R. Mobin H.K. Mohseni I. Saba L. Integrative analysis for COVID-19 patient outcome prediction Med. Image Anal. 67 2021 101844 Chen, Y., Kalantidis, Y., Li, J., Yan, S., Feng, J., 2018. Multi-fiber networks for video recognition. In: Proceedings of the European Conference on Computer Vision. ECCV, pp. 352–367. Chen B. Li J. Lu G. Yu H. Zhang D. Label co-occurrence learning with graph convolutional networks for multi-label chest X-ray image classification IEEE J. Biomed. Health Inf. 24 8 2020 2292 2302 Chen X. Ravikumar N. Xia Y. Attar R. Diaz-Pinto A. Piechnik S.K. Neubauer S. Petersen S.E. Frangi A.F. Shape registration with learned deformations for 3D shape reconstruction from sparse and incomplete point clouds Med. Image Anal. 2021 102228 Chen, Y., Rohrbach, M., Yan, Z., Shuicheng, Y., Feng, J., Kalantidis, Y., 2019. Graph-based global reasoning networks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 433–442. Chen N. Zhou M. Dong X. Qu J. Gong F. Han Y. Qiu Y. Wang J. Liu Y. Wei Y. Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study Lancet 395 10223 2020 507 513 32007143 Chikontwe P. Luna M. Kang M. Hong K.S. Ahn J.H. Park S.H. Dual attention multiple instance learning with unsupervised complementary loss for COVID-19 screening Med. Image Anal. 2021 102105 Defferrard, M., Bresson, X., Vandergheynst, P., 2016. Convolutional neural networks on graphs with fast localized spectral filtering. In: Advances in Neural Information Processing Systems. pp. 3844–3852. DeLong E.R. DeLong D.M. Clarke-Pearson D.L. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach Biometrics 1988 837 845 3203132 Di D. Shi F. Yan F. Xia L. Mo Z. Ding Z. Shan F. Song B. Li S. Wei Y. Hypergraph learning for identification of COVID-19 with CT imaging Med. Image Anal. 68 2021 101910 Dietterich T.G. Lathrop R.H. Lozano-Pérez T. Solving the multiple instance problem with axis-parallel rectangles Artificial Intelligence 89 1–2 1997 31 71 Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., Houlsby, N., 2021. An image is worth 16x16 words: transformers for image recognition at scale. In: International Conference on Learning Representations. Fan D.-P. Ji G.-P. Zhou T. Chen G. Fu H. Shen J. Shao L. Pranet: carallel reverse attention network for polyp segmentation International Conference on Medical Image Computing and Computer-Assisted Intervention 2020 Springer 263 273 Fan Y. Liu J. Yao R. Yuan X. COVID-19 detection from X-ray images using multi-kernel-size spatial-channel attention network Pattern Recognit. 2021 108055 Fan D.-P. Zhou T. Ji G.-P. Zhou Y. Chen G. Fu H. Shen J. Shao L. Inf-net: automatic COVID-19 lung infection segmentation from CT images IEEE Trans. Med. Imaging 39 8 2020 2626 2637 32730213 Fang C. Bai S. Chen Q. Zhou Y. Xia L. Qin L. Gong S. Xie X. Zhou C. Tu D. Deep learning for predicting COVID-19 malignant progression Med. Image Anal. 72 2021 102096 Gao S. Cheng M.-M. Zhao K. Zhang X.-Y. Yang M.-H. Torr P.H. Res2net: a new multi-scale backbone architecture IEEE Trans. Pattern Anal. Mach. Intell. 2019 Gao K. Su J. Jiang Z. Zeng L.-L. Feng Z. Shen H. Rong P. Xu X. Qin J. Yang Y. Dual-branch combination network (DCN): towards accurate diagnosis and lesion segmentation of COVID-19 using CT images Med. Image Anal. 67 2021 101836 Goncharov M. Pisov M. Shevtsov A. Shirokikh B. Kurmukov A. Blokhin I. Chernina V. Solovev A. Gombolevskiy V. Morozov S. CT-based COVID-19 triage: deep multitask learning improves joint identification and severity quantification Med. Image Anal. 71 2021 102054 Greenspan H. Estépar R.S.J. Niessen W.J. Siegel E. Nielsen M. Position paper on COVID-19 imaging and AI: from the clinical needs and technological challenges to initial AI solutions at the lab and national level towards a new era for AI in healthcare Med. Image Anal. 66 2020 101800 Guarrasi V. D’Amico N.C. Sicilia R. Cordelli E. Soda P. Pareto optimization of deep networks for COVID-19 diagnosis from chest X-rays Pattern Recognit. 121 2022 108242 Guo, L., Tang, L., Chen, T., Zhu, L., Nguyen, Q.V.H., Yin, H., 2021. DA-GCN: a domain-aware attentive graph convolution network for shared-account cross-domain sequential recommendation. In: The 30th International Joint Conference on Artificial Intelligence (IJCAI). Han Z. Wei B. Hong Y. Li T. Cong J. Zhu X. Wei H. Zhang W. Accurate screening of COVID-19 using attention-based deep 3D multiple instance learning IEEE Trans. Med. Imaging 39 8 2020 2584 2594 32730211 Hani C. Trieu N.H. Saab I. Dangeard S. Bennani S. Chassagnon G. Revel M.-P. COVID-19 pneumonia: a review of typical CT findings and differential diagnosis Diagn. Interv. Imaging 101 5 2020 263 268 32291197 Hao J. Liu J. Pereira E. Liu R. Zhang J. Zhang Y. Yan K. Gong Y. Zheng J. Zhang J. Uncertainty-guided graph attention network for parapneumonic effusion diagnosis Med. Image Anal. 2021 102217 He, X., Wang, S., Chu, X., Shi, S., Tang, J., Liu, X., Yan, C., Zhang, J., Ding, G., 2021a. Automated model design and benchmarking of deep learning models for COVID-19 detection with chest CT scans. In: Proceedings of the AAAI Conference on Artificial Intelligence. 35, (6), pp. 4821–4829. He, K., Zhang, X., Ren, S., Sun, J., 2016. Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 770–778. He K. Zhao W. Xie X. Ji W. Liu M. Tang Z. Shi Y. Shi F. Gao Y. Liu J. Synergistic learning of lung lobe segmentation and hierarchical multi-instance classification for automated severity assessment of COVID-19 in CT images Pattern Recognit. 113 2021 107828 Hou, L., Samaras, D., Kurc, T.M., Gao, Y., Davis, J.E., Saltz, J.H., 2016. Patch-based convolutional neural network for whole slide tissue image classification. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 2424–2433. Hou J. Xu J. Jiang L. Du S. Feng R. Zhang Y. Shan F. Xue X. Periphery-aware COVID-19 diagnosis with contrastive representation enhancement Pattern Recognit. 118 2021 108005 Huang, H., Lin, L., Zhang, Y., Xu, Y., Zheng, J., Mao, X., Qian, X., Peng, Z., Zhou, J., Chen, Y.-W., et al., 2021. Graph-BAS3Net: boundary-aware semi-supervised segmentation network with Bblateral graph convolution. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 7386–7395. Huang C. Wang Y. Li X. Ren L. Zhao J. Hu Y. Zhang L. Fan G. Xu J. Gu X. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China Lancet 395 10223 2020 497 506 31986264 Ilse M. Tomczak J. Welling M. Attention-based deep multiple instance learning International Conference on Machine Learning 2018 PMLR 2127 2136 Ji, W., Yu, S., Wu, J., Ma, K., Bian, C., Bi, Q., Li, J., Liu, H., Cheng, L., Zheng, Y., 2021. Learning calibrated medical image segmentation via multi-rater agreement modeling. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 12341–12351. Kendall, A., Gal, Y., 2017. What uncertainties do we need in bayesian deep learning for computer vision?. In: Advances in Neural Information Processing Systems. pp. 5574–5584. Khan, S., Hayat, M., Zamir, S.W., Shen, J., Shao, L., 2019. Striking the right balance with uncertainty. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 103–112. Kim E.A. Lee K.S. Primack S.L. Yoon H.K. Byun H.S. Kim T.S. Suh G.Y. Kwon O.J. Han J. Viral pneumonias in adults: radiologic and pathologic findings Radiographics 22 suppl_1 2002 S137 S149 12376607 Kim S.-H. Wi Y.M. Lim S. Han K.-T. Bae I.-G. Differences in clinical characteristics and chest images between coronavirus disease 2019 and influenza-associated pneumonia Diagnostics 11 2 2021 261 33567507 Kingma, D.P., Ba, J., 2014. Adam: a method for stochastic optimization. In: International Conference on Learning Representations. Kipf, T.N., Welling, M., 2017. Semi-supervised classification with graph convolutional networks. In: International Conference on Learning Representations. Kiser K. Ahmed S. Stieb S. Mohamed A. Elhalawani H. Park P. Doyle N. Wang B. Barman A. Fuller C. Data from the thoracic volume and pleural effusion segmentations in diseased lungs for benchmarking chest CT processing pipelines The Cancer Imaging Archive 10 2020 Kong F. Wilson N. Shadden S.C. A deep-learning approach for direct whole-heart mesh reconstruction Med. Image Anal. 2021 Koo H.J. Lim S. Choe J. Choi S.-H. Sung H. Do K.-H. Radiographic and CT features of viral pneumonia Radiographics 38 3 2018 719 739 29757717 Kraus O.Z. Ba J.L. Frey B.J. Classifying and segmenting microscopy images with deep multiple instance learning Bioinformatics 32 12 2016 i52 i59 27307644 Kumar A. Tripathi A.R. Satapathy S.C. Zhang Y.-D. SARS-net: COVID-19 detection from chest X-rays by combining graph convolutional network and convolutional neural network Pattern Recognit. 122 2022 108255 Lakshminarayanan B. Pritzel A. Blundell C. Simple and scalable predictive uncertainty estimation using deep ensembles Adv. Neural Inf. Process. Syst. 30 2017 Le M.T. Diehl F. Brunner T. Knol A. Uncertainty estimation for deep neural object detectors in safety-critical applications 2018 21st International Conference on Intelligent Transportation Systems ITSC 2018 IEEE 3873 3878 Lee, J., Chung, S.-Y., 2020. Robust training with ensemble consensus. In: International Conference on Learning Representations. Li Q. Guan X. Wu P. Wang X. Zhou L. Tong Y. Ren R. Leung K.S. Lau E.H. Wong J.Y. Early transmission dynamics in Wuhan, China, of novel coronavirus–infected pneumonia N. Engl. J. Med. 2020 Li, Y., Gupta, A., 2018. Beyond grids: learning graph representations for visual recognition. In: Advances in Neural Information Processing Systems. pp. 9225–9235. Li, B., Li, Y., Eliceiri, K.W., 2021a. Dual-stream multiple instance learning network for whole slide image classification with self-supervised contrastive learning. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 14318–14328. Li L. Qin L. Xu Z. Yin Y. Wang X. Kong B. Bai J. Lu Y. Fang Z. Song Q. Using artificial intelligence to detect COVID-19 and community-acquired pneumonia based on pulmonary CT: evaluation of the diagnostic accuracy Radiology 296 2 2020 E65 E71 32191588 Li Y. Wei D. Chen J. Cao S. Zhou H. Zhu Y. Wu J. Lan L. Sun W. Qian T. Efficient and effective training of COVID-19 classification networks with self-supervised dual-track learning to rank IEEE J. Biomed. Health Inf. 24 10 2020 2787 2797 Li K. Wu J. Wu F. Guo D. Chen L. Fang Z. Li C. The clinical and chest CT features associated with severe and critical COVID-19 pneumonia Investigative Radiology 2020 Li, X., Yang, Y., Zhao, Q., Shen, T., Lin, Z., Liu, H., 2020e. Spatial pyramid based graph reasoning for semantic segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 8950–8959. Li Z. Zhao W. Shi F. Qi L. Xie X. Wei Y. Ding Z. Gao Y. Wu S. Liu J. A novel multiple instance learning framework for COVID-19 severity assessment via data augmentation and self-supervised learning Med. Image Anal. 69 2021 101978 Liang, X., Hu, Z., Zhang, H., Lin, L., Xing, E.P., 2018. Symbolic graph reasoning meets convolutions. In: Proceedings of the 32nd International Conference on Neural Information Processing Systems. pp. 1858–1868. Liao F. Liang M. Li Z. Hu X. Song S. Evaluate the malignancy of pulmonary nodules using the 3-d deep leaky noisy-or network IEEE Trans. Neural Netw. Learn. Syst. 30 11 2019 3484 3495 30794190 Liu C. Cui J. Gan D. Yin G. Beyond COVID-19 diagnosis: prognosis with hierarchical graph representation learning International Conference on Medical Image Computing and Computer-Assisted Intervention 2021 Springer 283 292 Liu J. Dong B. Wang S. Cui H. Fan D.-P. Ma J. Chen G. COVID-19 lung infection segmentation with a novel two-stage cross-domain transfer learning framework Med. Image Anal. 74 2021 102205 Liu Y. Gadepalli K. Norouzi M. Dahl G.E. Kohlberger T. Boyko A. Venugopalan S. Timofeev A. Nelson P.Q. Corrado G.S. Detecting cancer metastases on gigapixel pathology images 2017 arXiv preprint arXiv:1703.02442 Loshchilov I. Hutter F. SGDR: stochastic gradient descent with warm restarts Int. Conf. Learn. Represent 2017 Luo A. Li X. Yang F. Jiao Z. Cheng H. Lyu S. Cascade graph neural networks for rgb-d salient object detection European Conference on Computer Vision 2020 Springer 346 364 Luo L. Yu L. Chen H. Liu Q. Wang X. Xu J. Heng P.-A. Deep mining external imperfect data for chest X-ray disease screening IEEE Trans. Med. Imaging 39 11 2020 3583 3594 32746106 Malhotra A. Mittal S. Majumdar P. Chhabra S. Thakral K. Vatsa M. Singh R. Chaudhury S. Pudrod A. Agrawal A. Multi-task driven explainable diagnosis of COVID-19 using chest X-ray images Pattern Recognit. 2021 108243 Mallick, A., Dwivedi, C., Kailkhura, B., Joshi, G., Han, T.Y.-J., 2020. Can your AI differentiate cats from COVID-19? sample efficient uncertainty estimation for deep learning safety. In: ICML 2020 Workshop on Uncertainty & Robustness in Deep Learning. Maron O. Lozano-Pérez T. A framework for multiple-instance learning Adv. Neural Inf. Process. Syst. 1998 570 576 Meng Y. Meng W. Gao D. Zhao Y. Yang X. Huang X. Zheng Y. Regression of instance boundary by aggregated CNN and GCN European Conference on Computer Vision 2020 Springer 190 207 Meng Y. Wei M. Gao D. Zhao Y. Yang X. Huang X. Zheng Y. CNN-GCN aggregation enabled boundary regression for biomedical image segmentation International Conference on Medical Image Computing and Computer-Assisted Intervention 2020 Springer 352 362 Meng, Y., Zhang, H., Gao, D., Zhao, Y., Yang, X., Qian, X., Huang, X., Zheng, Y., 2021a. BI-GConv: boundary-aware input-dependent graph convolution for biomedical image segmentation. In: The British Machine Vision Conference. BMVC. Meng, Y., Zhang, H., Zhao, Y., Yang, X., Qian, X., Huang, X., Zheng, Y., 2021b. Spatial uncertainty-aware semi-spervised crowd counting. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 15549–15559. Meng, Y., Zhang, H., Zhao, Y., Yang, X., Qiao, Y., Ian, J.C.M., Huang, X., Zheng, Y., 2021c. Graph-based region and boundary aggregation for biomedical image segmentation. In: IEEE Transactions on Medical Imaging. Milletari F. Navab N. Ahmadi S.-A. V-net: fully convolutional neural networks for volumetric medical image segmentation 2016 Fourth International Conference on 3D Vision (3DV) 2016 IEEE 565 571 Minaee S. Kafieh R. Sonka M. Yazdani S. Soufi G.J. Deep-COVID: predicting COVID-19 from chest X-ray images using deep transfer learning Med. Image Anal. 65 2020 101794 Morozov S.P. Andreychenko A.E. Blokhin I.A. Gelezhe P.B. Gonchar A.P. Nikolaev A.E. Pavlov N.A. Chernina V.Y. Gombolevskiy V.A. Mosmeddata: data set of 1110 chest CT scans performed during the COVID-19 epidemic Digital Diagnostics 1 1 2020 49 59 Noh K.J. Park S.J. Lee S. Combining fundus images and fluorescein angiography for artery/vein classification using the hierarchical vessel graph network International Conference on Medical Image Computing and Computer-Assisted Intervention 2020 Springer 595 605 Oh Y. Park S. Ye J.C. Deep learning COVID-19 features on cxr using limited training data sets IEEE Trans. Med. Imaging 39 8 2020 2688 2700 32396075 Ouyang X. Huo J. Xia L. Shan F. Liu J. Mo Z. Yan F. Ding Z. Yang Q. Song B. Dual-sampling attention network for diagnosis of COVID-19 from community acquired pneumonia IEEE Trans. Med. Imaging 39 8 2020 2595 2605 32730212 Perdomo O. Otálora S. González F.A. Meriaudeau F. Müller H. Oct-net: a convolutional network for automatic classification of normal and diabetic macular edema using sd-oct volumes 2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018) 2018 IEEE 1423 1426 Pinheiro, P.O., Collobert, R., 2015. From image-level to pixel-level labeling with convolutional networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 1713–1721. Qian X. Fu H. Shi W. Chen T. Fu Y. Shan F. Xue X. M3 Lung-sys: A deep learning system for multi-class lung pneumonia screening from CT imaging IEEE J. Biomed. Health Inf. 24 12 2020 3539 3550 Qiu J. Sun Y. Self-supervised iterative refinement learning for macular OCT volumetric data classification Comput. Biol. Med. 111 2019 103327 Rahimzadeh M. Attar A. Sakhaei S.M. A fully automated deep learning-based network for detecting COVID-19 from a new and large lung CT scan dataset Biomed. Signal Process. Control 1746-8094 2021 102588 10.1016/j.bspc.2021.102588 URL: https://www.sciencedirect.com/science/article/pii/S1746809421001853 Reittner P. Ward S. Heyneman L. Johkoh T. Müller N.L. Pneumonia: high-resolution CT findings in 114 patients European Radiology 13 3 2003 515 521 12594553 Rhee, S., Seo, S., Kim, S., 2018. Hybrid approach of relation network and localized graph convolutional filtering for breast cancer subtype classification. In: Proceedings of the 27th International Joint Conference on Artificial Intelligence. pp. 3527–3534. Roberts M. Driggs D. Thorpe M. Gilbey J. Yeung M. Ursprung S. Aviles-Rivero A.I. Etmann C. McCague C. Beer L. Common pitfalls and recommendations for using machine learning to detect and prognosticate for COVID-19 using chest radiographs and CT scans Nat. Mach. Intell. 3 3 2021 199 217 Ronneberger O. Fischer P. Brox T. U-net: Convolutional networks for biomedical image segmentation International Conference on Medical Image Computing and Computer-Assisted Intervention 2015 Springer 234 241 Roy S. Menapace W. Oei S. Luijten B. Fini E. Saltori C. Huijben I. Chennakeshava N. Mento F. Sentelli A. Deep learning for classification and localization of COVID-19 markers in point-of-care lung ultrasound IEEE Trans. Med. Imaging 39 8 2020 2676 2687 32406829 Selvaraju, R.R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., Batra, D., 2017. Grad-CAM: visual explanations from deep networks via gradient-based localization. In: Proceedings of the IEEE International Conference on Computer Vision. pp. 618–626. Shamsi A. Asgharnezhad H. Jokandan S.S. Khosravi A. Kebria P.M. Nahavandi D. Nahavandi S. Srinivasan D. An uncertainty-aware transfer learning-based framework for COVID-19 diagnosis IEEE Trans. Neural Netw. Learn. Syst. 32 4 2021 1408 1417 33571095 Shannon C.E. A mathematical theory of communication ACM Sigmobile Mobile Comput. Commun. Rev 5 1 2001 3 55 Shannon C.E. Weaver W. The Mathematical Theory of Communication 1949 Illinois Press Urbana, Illinois 117 Shao Z. Bian H. Chen Y. Wang Y. Zhang J. Ji X. Transmil: Transformer based correlated multiple instance learning for whole slide image classification Adv. Neural Inf. Process. Syst. 34 2021 2136 2147 Shi H. Han X. Jiang N. Cao Y. Alwalid O. Gu J. Fan Y. Zheng C. Radiological findings from 81 patients with COVID-19 pneumonia in wuhan, China: a descriptive study The Lancet Infect. Dis 20 4 2020 425 434 32105637 Shorfuzzaman M. Hossain M.S. MetaCOVID: a siamese neural network framework with contrastive loss for n-shot diagnosis of COVID-19 patients Pattern Recognit. 113 2021 107700 Signoroni A. Savardi M. Benini S. Adami N. Leonardi R. Gibellini P. Vaccher F. Ravanelli M. Borghesi A. Maroldi R. BS-net: learning COVID-19 pneumonia severity on a large chest X-ray dataset Med. Image Anal. 71 2021 102046 Soda P. D’Amico N.C. Tessadori J. Valbusa G. Guarrasi V. Bortolotto C. Akbar M.U. Sicilia R. Cordelli E. Fazzini D. AIforCOVID: predicting the clinical outcomes in patients with COVID-19 applying AI to chest- X-rays. an italian multicentre study Med. Image Anal. 74 2021 102216 Su Z. Tavolara T.E. Carreno-Galeano G. Lee S.J. Gurcan M.N. Niazi M. Attention2majority: Weak multiple instance learning for regenerative kidney grading on whole slide images Med. Image Anal. 79 2022 102462 Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., Rabinovich, A., 2015. Going deeper with convolutions. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 1–9. Tan M. Le Q. Efficientnet: rethinking model scaling for convolutional neural networks International Conference on Machine Learning 2019 PMLR 6105 6114 Tan, W., Liu, J., 2021. A 3D cnn network with bert for automatic COVID-19 diagnosis from CT-scan images. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 439–445. Team N.L.S.T.R. Reduced lung-cancer mortality with low-dose computed tomographic screening N. Engl. J. Med. 365 5 2011 395 409 21714641 Uemura T. Näppi J.J. Watari C. Hironaka T. Kamiya T. Yoshida H. Weakly unsupervised conditional generative adversarial network for image-based prognostic prediction for COVID-19 patients based on chest CT Med. Image Anal. 73 2021 102159 Vieira P. Sousa O. Magalhes D. Rablo R. Silva R. Detecting pulmonary diseases using deep features in X-ray images Pattern Recognit. 2021 108081 Vilar J. Domingo M.L. Soto C. Cogollos J. Radiology of bacterial pneumonia Eur. J. Radiol. 51 2 2004 102 113 15246516 Wada K. labelme: image polygonal annotation with python 2016 https://github.com/wkentaro/labelme Wang J. Bao Y. Wen Y. Lu H. Luo H. Xiang Y. Li X. Liu C. Qian D. Prior-attention residual learning for more discriminative COVID-19 screening in CT images IEEE Trans. Med. Imaging 39 8 2020 2572 2583 32730210 Wang X. Deng X. Fu Q. Zhou Q. Feng J. Ma H. Liu W. Zheng C. A weakly-supervised framework for COVID-19 classification and lesion localization from chest CT IEEE Trans. Med. Imaging 39 8 2020 2615 2625 33156775 Wang, X., Girshick, R., Gupta, A., He, K., 2018a. Non-local neural networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 7794–7803. Wang C. Horby P.W. Hayden F.G. Gao G.F. A novel coronavirus outbreak of global health concern Lancet 395 10223 2020 470 473 31986257 Wang X. Jiang L. Li L. Xu M. Deng X. Dai L. Xu X. Li T. Guo Y. Wang Z. Joint learning of 3D lesion segmentation and classification for explainable COVID-19 diagnosis IEEE Trans. Med. Imaging 2021 Wang Z. Liu Q. Dou Q. Contrastive cross-site learning with redesigned net for COVID-19 CT classification IEEE J. Biomed. Health Inf. 24 10 2020 2806 2813 Wang G. Liu X. Li C. Xu Z. Ruan J. Zhu H. Meng T. Li K. Huang N. Zhang S. A noise-robust framework for automatic segmentation of COVID-19 pneumonia lesions from CT images IEEE Trans. Med. Imaging 39 8 2020 2653 2663 32730215 Wang X. Tang F. Chen H. Luo L. Tang Z. Ran A.-R. Cheung C.Y. Heng P.-A. UD-MIL: uncertainty-driven deep multiple instance learning for OCT image classification IEEE J. Biomed. Health Inf. 24 12 2020 3431 3442 Wang J. Wang J. Wen Y. Lu H. Niu T. Pan J. Qian D. Pulmonary nodule detection in volumetric chest CT scans using CNNs-based nodule-size-adaptive detection and classification IEEE Access 7 2019 46033 46044 Wang Z. Xiao Y. Li Y. Zhang J. Lu F. Hou M. Liu X. Automatically discriminating and localizing COVID-19 from community-acquired pneumonia on chest X-rays Pattern Recognit. 110 2021 107613 Wang X. Yan Y. Tang P. Bai X. Liu W. Revisiting multiple instance neural networks Pattern Recognit. 74 2018 15 24 Wickramasinghe U. Remelli E. Knott G. Fua P. Voxel2mesh: 3D mesh model generation from volumetric data International Conference on Medical Image Computing and Computer-Assisted Intervention 2020 Springer 299 308 Wu X. Chen C. Zhong M. Wang J. Shi J. COVID-AL: the diagnosis of COVID-19 with deep active learning Med. Image Anal. 68 2021 101913 Wu Y.-H. Gao S.-H. Mei J. Xu J. Fan D.-P. Zhang R.-G. Cheng M.-M. Jcs: an explainable COVID-19 diagnosis system by joint classification and segmentation IEEE Trans. Image Process. 30 2021 3113 3126 33600316 Wu, J., Yu, Y., Huang, C., Yu, K., 2015. Deep multiple instance learning for image classification and auto-annotation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 3460–3469. Xie, S., Girshick, R., Dollár, P., Tu, Z., He, K., 2017. Aggregated residual transformations for deep neural networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 1492–1500. Xie W. Jacobs C. Charbonnier J.-P. Van Ginneken B. Relational modeling for robust and efficient pulmonary lobe segmentation in CT scans IEEE Trans. Med. Imaging 39 8 2020 2664 2675 32730216 Xu G.-X. Liu C. Liu J. Ding Z. Shi F. Guo M. Zhao W. Li X. Wei Y. Gao Y. Cross-site severity assessment of COVID-19 from CT images via domain adaptation IEEE Trans. Med. Imaging 2021 Xue W. Cao C. Liu J. Duan Y. Cao H. Wang J. Tao X. Chen Z. Wu M. Zhang J. Modality alignment contrastive learning for severity assessment of COVID-19 from lung ultrasound and clinical information Med. Image Anal. 69 2021 101975 Yang D. Xu Z. Li W. Myronenko A. Roth H.R. Harmon S. Xu S. Turkbey B. Turkbey E. Wang X. Federated semi-supervised learning for COVID region segmentation in chest CT using multi-national data from China, Italy, Japan Med. Image Anal. 70 2021 101992 Yang Z. Zhao L. Wu S. Chen C.Y.-C. Lung lesion localization of COVID-19 from chest CT image: a novel weakly supervised learning method IEEE J. Biomed. Health Inf. 25 6 2021 1864 1872 Yao J. Cai J. Yang D. Xu D. Huang J. Integrating 3D geometry of organ for improving medical image segmentation International Conference on Medical Image Computing and Computer-Assisted Intervention 2019 Springer 318 326 Yao Q. Xiao L. Liu P. Zhou S.K. Label-free segmentation of COVID-19 lesions in lung CT IEEE Trans. Med. Imaging 2021 Yu L. Wang S. Li X. Fu C.-W. Heng P.-A. Uncertainty-aware self-ensembling model for semi-supervised 3D left atrium segmentation International Conference on Medical Image Computing and Computer-Assisted Intervention 2019 Springer 605 613 Zagoruyko, S., Komodakis, N., 2016. Wide residual networks. In: The British Machine Vision Conference. BMVC. Zhang J. Fan D.-P. Dai Y. Anwar S. Saleh F. Aliakbarian S. Barnes N. Uncertainty inspired RGB-D saliency detection IEEE Trans. Pattern Anal. Mach. Intell. 2021 Zhang, L., Li, X., Arnab, A., Yang, K., Tong, Y., Torr, P.H., 2019. Dual graph convolutional network for semantic segmentation. In: The British Machine Vision Conference. BMVC. Zhang K. Liu X. Shen J. Li Z. Sang Y. Wu X. Zha Y. Liang W. Wang C. Wang K. Clinically applicable AI system for accurate diagnosis, quantitative measurements, and prognosis of COVID-19 pneumonia using computed tomography Cell 181 6 2020 1423 1433 32416069 Zhang, H., Meng, Y., Zhao, Y., Qiao, Y., Yang, X., Coupland, S.E., Zheng, Y., 2022. DTFD-MIL: double-tier feature distillation multiple instance learning for histopathology whole slide image classification. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Zhao, T., Cao, K., Yao, J., Nogues, I., Lu, L., Huang, L., Xiao, J., Yin, Z., Zhang, L., 2021a. 3D graph anatomy geometry-integrated network for pancreatic mass segmentation, diagnosis, and quantitative patient management. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 13743–13752. Zhao C. Xu Y. He Z. Tang J. Zhang Y. Han J. Shi Y. Zhou W. Lung segmentation and automatic detection of COVID-19 using radiomic features from chest CT images Pattern Recognit. 2021 108071 Zhao, Y., Yang, F., Fang, Y., Liu, H., Zhou, N., Zhang, J., Sun, J., Yang, S., Menze, B., Fan, X., et al., 2020. Predicting lymph node metastasis using histopathological images based on multiple instance learning with deep graph convolution. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 4837–4846. Zhong A. Li X. Wu D. Ren H. Kim K. Kim Y. Buch V. Neumark N. Bizzo B. Tak W.Y. Deep metric learning-based image retrieval system for chest radiograph and its clinical applications in COVID-19 Med. Image Anal. 70 2021 101993 Zhou L. Li Z. Zhou J. Li H. Chen Y. Huang Y. Xie D. Zhao L. Fan M. Hashmi S. A rapid, accurate and machine-agnostic segmentation and quantification method for CT-based COVID-19 diagnosis IEEE Trans. Med. Imaging 39 8 2020 2638 2652 32730214 Zhou Z. Siddiquee M.M.R. Tajbakhsh N. Liang J. UNet++: redesigning skip connections to exploit multiscale features in image segmentation IEEE Trans. Med. Imaging 2019 Zhu X. Song B. Shi F. Chen Y. Hu R. Gan J. Zhang W. Li M. Wang L. Gao Y. Joint prediction and time estimation of COVID-19 developing severe symptoms using chest CT scan Med. Image Anal. 67 2021 101824

0

PMC9753459

NO-CC CODE

2022-12-16 23:25:44

no

Med Image Anal. 2022 Dec 15;:102722

utf-8

Med Image Anal

10.1016/j.media.2022.102722

oa_other

==== Front Journal of Building Engineering 2352-7102 2352-7102 Elsevier Ltd. S2352-7102(22)01734-X 10.1016/j.jobe.2022.105728 105728 Article Investigation on the cross-infection control performance of interactive cascade ventilation in multi-scenario of winter Li Han Lan Yuer Ma Xiuqin Kong Xiangfei ∗ Fan Man School of Energy and Environmental Engineering, Hebei University of Technology, China ∗ Corresponding author. 15 12 2022 15 12 2022 10572828 8 2022 5 12 2022 10 12 2022 © 2022 Elsevier Ltd. All rights reserved. 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. With the wide spread of COVID-19, numerous cases demonstrate that proper ventilation method can reduce the cross-infection risk obviously. Interactive cascade ventilation (ICV) as a recently proposed ventilation method, the advantage of indoor environment construction has been proven. However, few studies are conducted to investigate the virus prevention and control characteristics of ICV, which is particularly important under epidemic normalizing. Hence, this study explored and compared the cross-infection control performance of three ventilation strategies, namely mixing ventilation (MV), stratum ventilation (SV), and interactive cascade ventilation (ICV), with a validated CFD model. A typical office was selected as the background scene, where an infected person coughs, sneezes with standing or sitting at different positions. Exposure doses, health infection risk, and disease burden (DB) were employed as the evaluation indicators under different ventilation methods of multi-scenario. The research results indicated that the average aerosol exposure dose among the human respiratory region under ICV was 0.29 g/day, which was reduced by 67 % and 50 % compared with MV and SV. In addition, only in ICV can the health infection risk meets the EPA standard. The average disease health burden for exposed persons under ICV was 0.93 × 10−6 DALYs pppy, which was 37 % and 70 % lower than SV and MV. The findings obtained from this study confirm that ICV performs excellently in reducing the cross-infection risk, providing the theoretical basis for future epidemic prevention and control. Keywords COVID-19 Ventilation strategies Interactive cascade ventilation Cross-infection risk Virus prevention and control performance ==== Body pmcNomenclature COVID-19 Coronavirus disease MV Mixing ventilation SV Stratum ventilation ICV Interactive cascade ventilation DB Disease health burden WHO World Health Organization DV Displacement ventilation HVAC Heating Ventilation Air Conditioning U⇀ velocity vector Γφ effective diffusion coefficient S⇀φ source term (kg/m·s) U⇀P particle velocity (m/s) Ri net rate of production ν⇀S gravitational settling velocity of the particle (m/s) Dp Brownian diffusion coefficient CD drag coefficient ρP density of the particle (kg/m3) Re Reynolds number Spp per person of minimum required ventilation (m3/s) SPA ventilation required per unit of floor area (m/s) Qi energy generated by interior loads (w) Qs energy of supply air (Cooling load) (w) EC exposure aerosol concentration (kg/m3) BR breathing rate (m3/day) AG aerosol ingestion rate (%) d daily dose(kg/day) Py(d) infection per person UFAD Under-floor air distribution RANS Reynolds Average Navier-Stokes QMRA quantitative microbial risk assessment P(y) d Annual probability of infection U.S. EPA U.S. Environmental Protection Agency CFD computational fluid dynamics RNG Re-Normalization Group φ solving variables (i.e., velocity, temperature and concentration) t flow time (s) ρ Density (kg/m3) Yi local mass fraction of particles (%) SYi rate of the source (m2/s3) μeff Effective viscosity (kg/m·s) εp particle turbulent diffusion coefficient dp diameter of particles (m) ρa density of the ambient air (kg/m3) (Sa)min minimum required ventilation (m3/s) O number of cooling load with different actual occupancy at different cooling loads A floor area (m2) Qe energy generated by exterior loads (W) OA% share of the total room load of outdoor air pppy per-person-per-year T daily exposure duration (hour) Pi(d) risk of infection per daily exposure (DALYs pppy) k model parameter n frequency of exposure 1 Introduction The widespread outbreak of coronavirus disease (COVID-19) shows an unprecedented impact on global society and human health. To quickly stop the spread, countries adopt various measures such as social distancing and restrictions on access to entertainment venues.etc. With outdoor activities being greatly limited, the time people spend indoors is obviously increased [1]. According to the research results [2], the novel coronavirus can be transmitted by direct contact and airborne transmission. In an enclosed space, the virus particles are first released by coughing and sneezing of an infected people. Then these virus particles polymerize into aerosols and droplets and begin to spread in the air [3]. Hence, the indoor ventilation patterns strongly impact the airborne transmission of pathogens. Proper ventilation method can effectively prevent virus transmission and reduce the cross-infection risk. While the aerosol transmission can be even intensified under an unreasonable ventilation method, leading to a large number of virus-positive samples circulating in the air [4]. Numerous cases during the pandemic are still caused by the poor indoor environment due to the unreasonable ventilation methods. As reported by the WHO in 2022 [5], large outbreaks of the epidemic usually occur in a crowded indoor settings, such as restaurants, choir practices, fitness classes, nightclubs, offices and places of worship. Talking, shouting, breathing heavily or singing loudly can facilitate the spread of the virus. Based on the WHO Statista [6], at least 17 million people in the European Region have experienced long COVID-19 in the first two years of the pandemic, and millions may have to live with it for years to come. Fortunately, physical measures such as wearing masks [7], increasing social distancing [8] and adding protective barriers [9] can effectively reduce the transmission risk of aerosol particles. However, these physical measures cannot be last for all the day in an enclosed space with people taking different behaviors. Therefore, it is urgent to seek for a proper ventilation method to block the spread of the virus and reduce the cross-infection risk in an effective way. Many studies have researched the aerosol diffusion characteristic under different ventilation methods. Ren et al. [10] studied three airflow organization strategies of mixing ventilation in novel coronavirus pneumonia prefabricated wards. It proposed that the pollutant particle size was a key factor on the removal efficiency of different ventilation methods. The configuration with inlets and outlets installed on the opposite sidewalls showed the highest removal efficiency for small particle pollutants. He et al. [11] compared the exhaled droplet transmission for different diameters of 0.8 μm, 5 μm and 16 μm between occupants under different ventilation strategies in a typical office room. The research results indicated that displacement ventilation performed best concern protecting the exposed manikin from the pollutants exhaled by the polluting manikin without personalized ventilation. Lu et al. [12] contrastively investigated the pollutant distribution and exposure risk in a ward served by stratum ventilation, mixing ventilation, under-floor air distribution, and displacement ventilation, respectively. The research results demonstrated that stratum ventilation can obviously reduce the exposure risk of the healthcare workers. Research conducted by Kong et al. [13] also showed that stratum ventilation can minimize airborne viral contamination. Lin et al. [14] compared the particle diffusion of three representative scenarios in a classroom under displacement ventilation and stratum ventilation through numerical simulations, and pointed out that stratum ventilation showed best on the anti-airborne infection performance. It follows that COVID-19 poses new challenges to the current air conditioning systems. In addition to the current thermal comfort and energy-saving, the ability of virus prevention and control has also become the important indicator of HVAC system under the normal epidemic situation. Recently, Li et al. [15] proposed a novel ventilation method named interactive cascade ventilation (ICV) and indicated ICV showed good performance on indoor comfort and energy-saving. ICV as a novel ventilation method innovatively introduces different grade energy to obtain air jets with various temperatures. The air supply inlets of ICV are installed at different heights of the same wall. The upper supply air inlets provide lower temperature airflow while the lower supply air inlets provide higher temperature airflow. The thermal buoyancy force generated by the temperature difference between jets of the ICV can prevent the upper layer of fresh air from floating up fast. Hence, more clean air stay in the respiratory layer under ICV due to its special air flow characteristics, which is supposed to reduce the risk of cross-infection in the room. However, current studies on the HVAC system are mainly focused on the performance and thermal comfort. Few studies are conducted to investigate the virus prevention and control characteristics of ICV, which is particularly important under epidemic normalizing. Therefore, this paper conducted an in-depth comparative study on the cross-infection control performance in a typical office served by mixing ventilation (MV), stratum ventilation (SV) and ICV, respectively. In this study, a typical office served by MV, SV and ICV was selected as the background environment. Reynolds Average Navier-Stokes (RANS) method was adopted to conduct the numerical calculation for investigating the cross-infection control performance of different ventilation methods by considering an infected person coughed, sneezed with standing or sitting at different positions. Quantitative microbial risk assessment (QMRA) was introduced to explore the health risk and disease burden (DB) associated with exposure to aerosol environments [16]. The findings obtained from this study can provide guide and new ideas for the further HVAC system design by considering the cross-infection control ability for a healthy indoor environment. 2 Methodology 2.1 Ventilation method In this study, three ventilation methods, namely mixing ventilation (MV), stratum ventilation (SV) and interactive cascade ventilation (ICV) were selected as the research objects. Usually, MV is used to provide a uniform indoor environment by diluting to reduce air pollutant concentrations [17]. SV is proposed to realize a non-uniform environment, which delivers fresh air directly to the breathing region to make the air pollutant concentration much lower. ICV as a novel ventilation method is also employed to build a non-uniform indoor environment. Two air jets with different temperatures of ICV are sent from the inlets at various heights at the same sidewall. And different grades of energy are introduced into the ICV to obtain the two air jets, as shown in Fig. 1 . The thermal buoyancy force caused by the temperature gradient between jets can slow the warm air coming up. The upper warm air with a lower temperature (24 oC) can produce a downward settling force on the lower warm air with a higher temperature (26 oC). Conversely, the lower jet also exerts a lifting force on the upper jet. Therefore, ICV can make more fresh air delivered from the upper jet stay in the breathing region, reducing the cross-infection risk. As shown in Fig. 1 (b), the two jets locate in the superimposed region will change their original jet trajectories due to the interaction between jets. During the outbreak, ICV can adopt the mode of upper fresh air coupled with lower return air to realize a clean breathing layer with less energy consumption. The studies conducted by Li et al. [15] and Kong et al. [18] have demonstrated that ICV can overcome the defect of warm air floating on to a certain extent, which also enhance the effective fresh air volume meanwhile. However, the virus prevention and control ability of ICV is still a gap.Fig. 1 The airflow distribution of interactive cascade ventilation. Fig. 1 2.2 Model description The physical models of three ventilation modes are presented in Fig. 2 . The size of the office room is set as 9.8 (length)×12.4 (width)×2.6 m (height). The volume of the office room is about 316 m3, referring to the research model conducted by Ren et al. (2021) [17]. The office is equipped with 43 workstations, including one staff member and one table. The staff member is simulated with a size of 0.4 (length)×0.3 (width)×1.1 m (height). And the ahead (including neck) size is 0.2 (length)×0.2 (width)×0.2 m (height). The mouth size is 3 cm×2 cm. The heat dissipation of the staff members is also considered in the models. A constant temperature of 37 °C is used as the thermal boundary condition to simulate the human body temperature. Five row desks with the size of 1.2 (length)×0.7 (width)×0.8 m (height) are located in the office room. Air supply parameters are determined in accordance with GB50736-2012 [19] and indoor heat requirements. For MV, six air supply inlets (M1-M6) and one return outlet (E1) are located in the ceiling (c.f. Fig. 2b). For SV, four air supply inlets (S1-S4) with a diameter of 0.3 m are installed at 1.2m of the side wall (c.f. Fig. 2c). For ICV, the sidewall air inlets are circular diffuser with a diameter of 0.2 m. The upper air inlets (ICV 1, 3, 5, 7) are located at the height of 1.2 m and the lower air inlets (ICV 2, 4, 6, 8) are located at 0.7 m. The exhaust louvres (E2-E5) of ICV are 1.2 m above the floor on the opposite wall (c.f. Fig. 2d). More detailed information has been presented in Table 1 [17,20].Fig. 2 (a) Layout of open office (b) Mixing ventilation (c) Stratum ventilation (d) Interactive cascade ventilation. Fig. 2 Table 1 Numerical Boundary condition settings. Table 1Ventilation methods Supply air temperate (°C) Supply air velocity (m/s) Turbulence Intensity (%) ACH (h−1) Supply volume (m3/s) size (m) MV 25 0.3 5 5.12 0.45 0.5×0.5 SV 25 2 5.12 0.45 φ = 0. 3 ICV Upper inlet:24 1.5 5 5.12 0.45 φ = 0.2 Lower inlet:26 1.5 Mouth 37 Cough:13 4 – – 0.02×0.03 Sneeze:50 Outlet Outflow MV 0.2×0.2 ICV/SV 0.2×0.3 An infected person is set to produce virus-containing droplets or aerosols by coughing with a speed of 13 m/s [21] or sneezing of 50 m/s [22], along with the behaviors of sitting or standing meanwhile. The airflow temperature during coughing and sneezing is 37 oC [23]. The initial droplet aerosols are assumed as spherical with the density of 998 kg/m3[24][24]. The size of droplet nuclei produced by human coughing and sneezing is 1 μm, whose evaporation can be neglected according to the previous studies [25]. The turbulence intensity is calculated based on equation (1):(1) I=0.16ReDH0.125 where ReDH is the Reynolds number derived from the hydraulic diameter as the characteristic length. The material emission component ReDH is calculated as:(2) ReDH=ρuDHμ where ρ is the density, considering air density as 1.29 kg/m3 in this study; u is the flow velocity, m/s; DH is the hydraulic diameter, m; μ is the kinematic viscosity, kg/m·s. The kinematic viscosity of air and droplet nuclei is 1.7894×10−5 kg/m·s and 0.0011 kg/m·s, respectively. Hence, the turbulence intensity at the supply air terminal and infector mouth is 4.86 % and 3.92 %. For the three ventilation methods, the external wall surface temperature is −15 °C. The air supply volume is designed as 0.45 m3/s. And the ventilation flow rate is calculated as 5.12 h−1 considering the room volume as 315.95 m3. Considering the heat transfer through the building envelops, internal load, and cold air penetration load, the total heating load is 1.94 kW. The heat input through the ventilation system is 2.08 KW. The detailed information of the related parameters has been list in Table 2 . It can be seen that it is heat balanced for all three systems.Table 2 Detailed information of load calculation. Table 2Surface Materials Heat transfer coefficient (W/m2∙K) Size Tn(oC) Tw (oC) West/East wall Concrete hollow blocks 0.83 12.4 × 2.6 (m) 22 −15 North/South wall 0.85 9.8 × 2.6 (m) floor 0.3 9.8 × 12.4 (m) ceiling 0.45 9.8 × 12.4 (m) staff 80W heat source 43 (per) This study conducts the stable numerical analysis of coughing or sneezing for further research. Actually, they are transient. This method may overestimate the airborne transmission/infection risk to some extent. On average, a healthy person sneezes four times per day and coughs two times per day [26]. However, if a patient is infected with the respiratory infectious disease, such as COVID-19 or SARS, the coughing and sneezing will occur more frequently. A reasonable ventilation method can minimize the infection risk to surrounding people in cases where the exposure is high. The worst environment caused by infected people should be considered to maximize the protection of the uninfected. Hence, this study replaces transient simulation with steady-state simulation for further research. In addition, the relevant research also points out that the analytical velocity of saliva droplets produced by coughing is overestimated [27]. 2.3 Governing equations The analysis of particle diffusion and deposition is performed by computational fluid dynamics (CFD) simulations with software tool ANSYS Airpak 3.0.16 [28]. In this study, incompressible flow and steady-state condition are considered and solved by Reynolds-averaged Navier-Stokes (RANS) turbulence model, which is better for calculating the indoor airflow environment [29]. Standard wall functions are introduced to deal with near-wall condition. A discrete coordinate radiation model is used to simulate the radiative heat exchange between human body surfaces, luminaires and other interior surfaces. The convergence criterion is set as 10−4 for the momentum residuals and 10−6 for the mass residuals, turbulent kinetic energy residuals, turbulent dissipation residuals, energy residuals and radiation intensity residuals. The conservation equation of mass, momentum, energy, K and ε is shown as follows:(3) ∂(ρφ)∂t+div(ρU⇀φ)=div(Γφgradφ)+S⇀φ where φ is the solving variables (i.e. velocity, temperature and concentration); t is the flow time; ρ is the density; U⇀ is the velocity vector; Γφ represents the effective diffusion coefficient; S⇀φ is the source term [30]. More detailed information about the governing equation is shown in Table 3 .Table 3 Diffusion terms and source terms in the governing equation. Table 3Item Variable Γφ S⇀φ Continuity 1 0 0 Velocity μi μeff=μ+μt −∂p∂xi+sui Temperature T μPr+μtσT sT Kinetic energy k αKμeff GK+GB−ρε Dissipation rate ε αεμeff C1εεk(GK+C3εGB)−C2ερε2k−Rε Concentration C μSc+μtσc sc μi is the component in the i direction of velocity; Pr and σT are laminar and turbulent Prandtl numbers; σT is taken as 0.85; αK = αε = 1.393 is the inverse turbulent Prandtl number; P is the static pressure; Sui is the external and gravitational body forces in the i direction; ST is the volumetric heat source; GK and GB represent the turbulent kinetic energy caused by the mean velocity gradient and buoyancy, respectively; Rε is the source term of the reformulation; C1ε = 1.42, C2ε = 1.68, C3ε = tanh|μiμk|; and σc = 0.7. Euler and Lagrangian methods are two basic theoretical methods to solve particle dispersion. Both of them can well predict particle dispersion under steady-state condition [31]. The drift-flux model based on Eulerian-Eulerian method has been validated for simulating particles in indoor flow fields, which takes into account Brownian diffusion, turbulent diffusion, and gravitational settling. The controlling equation for particle concentration is similar to the Navier-Stokes equation, combining the particle gravitational settling effect into convective term [32]. For solving the dispersion of particles, the species transport model is used:(4) ∂(ρYi)∂t+Δ(ρ(U⇀P+ν⇀S)Yi)=Δ(μeffσcΔYi)+SYi where Yi is the local mass fraction of particles; (U⇀P) is the particle velocity; ν⇀S is the gravitational settling velocity of the particle; μeff is the effective viscosity, the total of molecular and turbulent viscosity; σc is the Prantl number, usually set as 1; SYi is the rate of the source [33]. The above equations are discretized directly into algebraic equations by the finite volume method with second order accuracy. Buoyancy effects are considered by the Boussinesq model. The methods for analyzing the performance of different ventilation methods used in this study have been confirmed and verified in other similar studies[33][[33], [34], [35]][35]. The key assumptions in the simulations are as follows:1. Without considering the effect of particles on the turbulent flow, the interaction between the carrier air and particles can be treated as a unidirectional coupling. 2. The diameter of exhaled aerosols ranges from 1μm to 1000μm, with most droplet nuclei falling in the 1–10μm size range [36]. Droplets smaller than 10 μm can reach the alveolar region more efficiently, thus causing a greater infection risk to susceptible individuals [37]. Therefore, droplet nuclei of 1μm are chosen to be simulated. The evaporation process can be neglected due to its transient occurrence time [38]. 3. Aerosol condensation does not change particle size due to low particle loading. 4. The rebound and resuspend of droplets are ignored. 5. For particles 1.0μm in size, the Brownian diffusion coefficient Dp is much weaker than kinetic viscosity and the turbulent diffusion coefficient. Therefore, ρ(Dp+εp) is replaced by μeff in Eq (4). Eq (5) can be obtained as follows:(5) |vs|=[43∙gdpCD∙(ρP−ρa)ρa]1/2 where CD is the drag coefficient, dp is the diameter of particles, ρP and ρa are the density of the particle and ambient air. The directions of settling velocity and gravity are the same. The drag coefficient CD is derived from the Stokes equation (Re < 1) or the modified equation (1<Re < 1000):CD=24ReRe＜1 (6) CD=24Re(1+0.15Re2/3)1＜Re＜1000 where Re is the Reynolds number according to the relative velocity of particle and air. 2.4 Grid-independent tests To exclude the influence of the grid number on the calculation results, three groups of the coarse, medium, and fine grids are set up for the grid independence analysis. The coordinate points for test are defined as (x, y)=(4.55 m, 1.55 m) with the alternative heights of 0.2, 0.4, 0.6, 0.8, 1.0, 1.2, 1.4, 1.6, 1.8, 2.0, 2.2 and 2.4 m. The validation data for the measurement points are obtained from experiments. The mesh grid number for the three simulation cases is 2,934,123, 4,212,373, and 8,913,257, respectively. Fig. 3 presents the comparison results of different grid numbers. The results demonstrate that the simulation results employing 4,212,373 and 8,913,257 meshes are in good agreement. The deviation is less than 6 %. Considering the long computation time due to high grid number, this study selects the grid number of 4,212,373 for further calculations.Fig. 3 Grid independence analysis for different grid numbers. Fig. 3 2.5 Evaluation index Quantitative microbial risk assessment (QMRA) [39] is conducted in accordance with the National Academy of Sciences Risk Assessment framework, including hazard identification, exposure assessment, dose response assessment, and risk characterization [40]. This study investigated the airborne transmission characteristic of virus-containing aerosols exhaled by coughing or sneezing of the infector under different ventilation methods. Risk characterization is based on the dose response model. The health risk levels recommended by the U.S. Environmental Protection Agency (2005) for infections and the criteria for disability-adjusted life years (DALYs) recommended by the World Health Organization (2008) are introduced to assess health risk in this study. The U.S. EPA benchmark is ≦ 10−4 infection cases per-person-per-year (pppy), and the WHO benchmark is ≦ 10−6 DALYS pppy [41]. 2.5.1 Hazard identification The following aspects should be considered when assessing the cross-infection risk in exposed aerosols: (a) different stages of the COVID-19 outbreak; (b) percentage of people infected; and (c) different ventilation methods. 2.5.2 Exposure assessment The daily dose (kg/day) of SARS-CoV-2 aerosols inhaled by the staffs in an office is given by [42].(7) d=EC×BR×T×AG where EC is the exposure aerosol concentration, kg/m3; BR is breathing rate, m3/day. Usually, the breathing rate of one adult man is 18.65 m3/day [43]. T is the daily exposure duration, h. This study considers 8 h for the occupational exposure. AG is the aerosol ingestion rate, %. 2.5.3 Dose response assessment The risk or probability of getting infected through intake of pathogens is estimated with dose–response model. The exponential dose response model for human coronavirus suggested by Watanabe et al. (2010) is as follows [44]:(8) Pi(d)=1−e−killd where Pi(d) is the infection risk per daily exposure of staff to aerosols/droplets of SARS-CoV-2, DALYs pppy; d is the daily dose, kg/day; and k ill is the model parameter, which is considered as the disease response endpoint value of Equation (8), 5.39 × 10−2 is employed in this study [44]. The annual infection risk of SARS-CoV-2 per person Py(d) is estimated by Equation (9)[45].(9) Py(d)=1−[1−Pi(d)]n where n is the exposure frequency. In this research, it is defined as 365. The disease burden (DB) can be obtained by Equation (10)[43].(10) DB=Py×HB where HB values for 0.00427 according to the previous research results [46]. 3 Study cases To investigate the cross-infection control performance of different ventilation methods, various scenarios with an infected person coughing or sneezing with the behaviors of standing or sitting (c.f. Fig. 4 b) at different locations of the room are considered in this study. Three typical locations, namely closing to the air supply inlets, closing to the air return outlets, and the center of the room, are considered (c.f. Fig. 4a). Based on this, a total of 36 working conditions are analyzed subsequently. Detailed information of the study cases are provided in Table 4 .Fig. 4 Scenario settings (a) XY plan; (b) YZ pan. Fig. 4 Table 4 Detailed information of the study cases. Table 4case Air distribution scenario Source location case Air distribution scenario Source location Case1 MV Sit cough Source 1 Case13 MV Sit cough Source 2 Case2 Sit sneeze Case14 Sit sneeze Case3 Stand cough Case15 Stand cough Case4 Stand sneeze Case16 Stand sneeze Case5 SV Sit cough Case17 SV Sit cough Case6 Sit sneeze Case18 Sit sneeze Case7 Stand cough Case19 Stand cough Case8 Stand sneeze Case20 Stand sneeze Case9 ICV Sit cough Case21 ICV Sit cough Case10 Sit sneeze Case22 Sit sneeze Case11 Stand cough Case23 Stand cough Case12 Stand sneeze Case24 Stand sneeze Case25 MV Sit cough Source 3 Case29 SV Sit cough Source 3 Case26 Sit sneeze Case30 Sit sneeze Case27 Stand cough Case31 Stand cough Case28 Stand sneeze Case32 Stand sneeze Case33 ICV Sit cough Case35 ICV Stand cough Case34 Sit sneeze Case36 Stand sneeze 4 Results and discussion 4.1 Model validation Hang et al. [47] and Duguid et al. [48] point out that the particles with diameters between 0.5 and 10 μm can remain in the air for longer time and their transport pattern is similar to gases. The droplet nuclei containing pathogens exhaled by infectors are particulate matter. Bivolarov et al. [49] confirmed that tracer gas can be reliably used to simulate small particles in airborne measurements by comparing the human exposure at the concentrations of tracer gas and monodisperse particle (0.07, 0.7 and 3.5 μm), respectively. Hence, tracer gas CO2 is introduced to simulate the virus-bearing droplet nuclei exhaled by an infector in this study. And the species models are validated by the experimental data. It is noted that the research method can be accepted even though the particle diffusion and deposition models are not been verified. The experiment is carried out in an environmental chamber with the size of 4.9 × 4.8× 2.17 m. The size of air inlets and outlets of MV is 0.5× 0.5 m. The air supply inlet diameter of SV and ICV is 0.3 m, and the air outlet is 0.5× 0.5 m. The air supply parameters used in the simulations are the same as experiments. Specific parameter settings of supply air are presented in Table 5 . Cardboard boxes with dimensions of 0.54×0.29×1.2 m are used to simulate manikins. The Cardboard boxes are fitted with electric heat tape to realize a surface temperate of 37 oC to consider the heat dissipation of the people, and ensuring to heat the CO2 gas pipe meanwhile. Hence, the temperature of CO2 is similar to that of the air exhaled by the infector. The experimental method has been proven to be an effective alternative for investigating the interaction between occupants and indoor environment [50]. Two 42 W LED lamps are installed on the ceiling. Detailed information of the building envelops has been listed in Table 2.Table 5 Detailed air supply parameters. Table 5Ventilation method ACH (h−1) Supply air volume (m3/h) Supply air temperature (oC) Supply air velocity (m/s) MV 5.2 540 25 0.3 SV 5.2 540 25 1.2 ICV 5.2 540 24/26 1.2 To avoid oversupply or undersupply of conditioned air, the method of Anand [51] et al. is introduced to determine the minimum required supply air flow rate based on occupancy and load information. The fresh air flow rate for the target area can be calculated as follows:(11) (Sa)min=Spp×O+SIL where (S a)min is the minimum required ventilation rate, (m3/s); S pp is the minimum required ventilation (0.0025 m3/s) per person; O is the number of occupants. The material emission component S IL is calculated as:(12) SIL=SPA×A where S PA is the required ventilation per floor area, m3/s per m2; A is floor area, m2. The simplified heat balance model is shown as follow:(13) Qi+Qe=Qs where Q i is the energy generated by interior loads, W; Q e is the energy generated by exterior loads, W; Q s is the energy of supply air, W.(14) (Sa)=(Sa)minOA% where OA % is the exterior load percentage of the total load. By calculating, S a is 0.16 m3/s, which is met the minimum requirements of 0.15 m3/s. In the experiments, temperature, velocity, and CO2 concentration of different heights are recorded at measurement points (c.f. Fig. 5 b). The accuracy of T-type thermocouple is ±0.1 °C in the range of −200∼260 °C, and hot wire anemometer measurement (TESTO440 MODE TYPE 06280152 SERIAL No. 61183753) is ±0.07 m/s in the range of 0.02–3.0 m/s. The CO2 concentration is measured by Telaie 7001 with an accuracy of ±50 ppm in the range of 0–10000 ppm. CO2 tanks and flowmeters are employed to release gas through a hole at 1.1 m of the rectangular thermal manikin to simulate the respiration. The locations of the measurement points are presented in Fig. 5 (a). The experimental results at test point are compared with the corresponding simulation data, which are shown in Fig. 6 .Fig. 5 The location of the measurement point (a) Schematic of the vertical view; (b) Real image. Fig. 5 Fig. 6 Comparisons of simulated and experimental at the measuring points (a) Velocity; (b) temperature; (c) CO2 concentration. Fig. 6 It can be seen that most of the numerical values are in good agreement with the experimental data. The maximum relative error of velocity, temperature and CO2 concentration is 9.2 %, 4.3 % and 5.4 %, respectively. It is mainly caused by the instrument error, measurement error and slight fluctuation of the outdoor conditions. The comparison results indicate that the accuracy of the model is reliable and can be used for further research. 4.2 Exposure dose of different pollution sources Fig. 7, Fig. 8 present the velocity distributions with infected people sitting to cough at pollution source 1. Under MV, uniform environment can be obtained. However, most of the clean warm air is concentrated at the ceiling due to the thermal buoyancy. Under SV, the air supply inlets can directly deliver fresh air to the breathing zone. Nevertheless, the air velocity and temperature are also decayed with distance increasing. It can be concluded that the people far away from the inlets cannot get enough fresh air. Aiming at the problems exposed in the stratum ventilation, ICV introduces two jets with different temperature. Depending on the interaction between the jets, ICV can realize the direction change of thermal buoyancy, thus to alleviate the warm air rising. It can be seen that a more comfortable indoor environment with “sandwich” characteristic remaining more clean warm air in the breathing zone is provided under ICV.Fig. 7 Velocity distributions (Y = 6.1m) with infector sitting to cough at pollution sources 1 under different ventilation methods. Fig. 7 Fig. 8 Velocity distributions (Z = 1.1m) with infector sitting to cough at pollution sources 1 under different ventilation methods (a) MV (b) SV (c) ICV. Fig. 8 The effect of the pollution source location on the ventilation performance is extremely obvious. As shown in Fig. 9 , significant variations exist among the exposure bioaerosol concentrations with one infector sitting to cough at source 1, source 2, and source 3 served by different ventilation methods. It can be seen that the dispersion of the pollutants changes with the pollutant source locations.Fig. 9 Exposure aerosol concentration distributions at source 1, source 2 and source 3 under different ventilation methods. Fig. 9 Under MV (c.f. Fig. 8(a–c)), the pollutant concentration range becomes more widely when the infector is located at source 2. With the concentration of 0.012 kg/m3 as the center of the pollution source 2, the diffusion radius can reach at least 9m. It is due to that the contamination source is located at the room corner, which is far from the return air outlets. And the supply air carries the pollutants throughout the room. Source 1 also presents the wide diffusion range. The pollutant transmission radius is around 7 m with a concentration limit of 0.012 kg/m3. Source 3 is located near the return air outlet. Hence, the exhaled aerosol with virus can be quickly sucked and cleared out. Fig. 9a demonstrates that the diffusion radius is only 3 m with the concentration boundary of 0.012 kg/m3. Fig. 9(d–f) presents the exposure bioaerosol concentration distribution under SV. It demonstrates that the pollutant source location shows a minor effect on the concentration distribution under SV. The pollutant diffusion radius is all about 3 m with the exhaled aerosol concentration of 0.012 kg/m3 as the boundary. However, when the infector is closed to the supply air inlets, it is easier to cause the aerosol with virus to be delivered to the human respiratory region, leading to a relatively high exposure of the staff across from and in the same row as the infector. Under ICV, Fig. 9(g–i) indicate that the variation of the source location has no significant effect on the aerosol dispersion compared to MV and SV. And the pollutant diffusion radius can be reduced to 2–3 m under ICV. The average contaminant exposure dose at the nose of a healthy staff, who sits across from the infector, is contrastively presented in Fig. 10 . The results reveal that the contaminant exposure dose is the highest under MV while lowest under ICV. In addition, the infector located at source 2 can lead to the highest exposure dose under MV. Due to the similar airflow organization of SV and ICV, the exposure dose to virus-containing aerosols caused by sneezing was both higher when the infector locates at source 3 under the two ventilation methods.Fig. 10 Comparison of exposure doses under different ventilation methods. Fig. 10 4.3 Annual infection probability Considering the infector coughing/sneezing with sitting/standing at various locations, the annual infection probability Py(d) for bioaerosol health risk under different ventilation methods is shown in Fig. 11 . The U.S. EPA benchmark is ≦ 10−4 infection cases per-person-per-year (pppy). According to the previous research [52], the calculation results and criteria can be expanded by a multiple of 103 to make the expression clearer, which has no impact on the judging of the indicators. Hence, 0.1 presented in Fig. 11 is the healthy infection risk benchmark after index unitization. It can be seen that the health infection risk under ICV is the lowest among the three ventilation methods. Ventilation methods show noteworthy effect on the indoor health risk. Also, the location and behavior of infected individuals can further influence ventilation performance and lead to the differentiation of human health risks.Fig. 11 Heat map of annual infection probability under various ventilation scenarios. (a) Locations of pollutant source; (b) Source 1; (c) Source 2; (d) Source 3. Fig. 11 Under ICV, the aerosols exhaled by the infector coughing at sources 1 and 2 can be effectively diluted by the clean air, which makes the infection risk lower and meet the baseline. Due to the source 3 is closed to the supply air inlets of SV and ICV, the infection risk is higher than the U.S. EPA standard in all scenarios. Compared with SV, there are two heights of the ICV air supply inlets to form a particular airflow distribution with “sandwich” characteristic. Although the aerosol containing virus is quickly diffused to the breathing region under ICV, clever airflow pattern can effectively control the spread of the virus, realizing a much lower health infection risk for exposed individuals compared with SV and MV. As for MV and SV, the health infection risk for the exposed person is always above the benchmark. It can be concluded that the indoor health infection risk hardly meet U.S. EPA standards served by MV and SV. Even so, the overall health infection risk under SV is still less than MV. It shows the superiority of non-uniform environment construction technology in epidemic prevention and control. The research results of the annual infection probability demonstrate that the indoor health infection risk varies greatly under different ventilation methods. Hence, reasonable ventilation methods should be introduced as the appropriate control strategy for reducing the cross-infection risk to an acceptable level, which is especially important when the epidemic is normalized. 4.4 Disease burden Fig. 12 presents the assessment results of disease burden (DB) in various scenarios. It can be seen that the estimations of Py(d) and DB are nearly identical. WHO benchmark of DB is ≦ 10−6 DALYS pppy. Similar to the annual infection probability, the calculation results and criteria of DB are expanded by a multiple of 105 to make the expression clearer, which has no impact on the judging of the indicators. 0.1 employed in Fig. 12 is the DB benchmark after index unitization. The disease health burden can be acceptable if the value is less than 0.1. Hence, the red numbers shown in Fig. 12 represent that the DB exceeds the allowed value under the corresponding scenario (ventilation methods, infector behavior, and infector location).Fig. 12 Heat map of disease burden under various ventilation scenarios (a) Locations of pollutant source; (b) Source 1; (c) Source 2; (d) Source 3. Fig. 12 Under ICV, the lowest disease health burden can be realized in all the discussed scenarios. It can be concluded that ICV is more effectivity on reducing aerosol concentrations in the breathing region of exposed individuals. When an infected person is sitting or standing, the aerosols exhaled through coughing can be effectively diluted by the clean air. However, the exhaled airflow is much larger when the infector sneezes. Hence, the position of the infector obviously affects the burden of disease borne of the exposed person when the infector sneezes. With the infector locating in the center of the room (source 1) and near the supply air inlets (source 3), the burden of the exposed person increases significantly. Due to the source 2 is located near the exhaust vents, the aerosol concentration in the breathing zone can be quickly reduced owe to the peculiar flow patter of ICV. Therefore, the health burden of disease for exposed individuals can all be within allowed values for both standing and seated aerosol exhalation when the infector is at source 2. As for SV, the calculation results of the disease burden of exposed individuals indicate that ventilation performance of SV is slightly less effective than ICV. Compared with MV, SV can better deal with aerosol pollutants in the respiratory region when the infector coughs. In addition, the position of the infector has a significant effect on DB under SV. When an infected person sneezes with a large velocity of exhaled air, SV is less effective in removing aerosols than ICV, while stronger than MV. It is also noted that the health burden of disease on the exposed person is unbearable with the infector near the air inlets (source 3) under SV. It can be known that the infector should be avoided near the air inlets as far as possible, so as not to cause a wide range of infection under SV. However, no such limit exists under ICV. Unfortunately, it is difficult to ensure the disease health burden on exposed individuals within tolerable limits in almost all scenarios under MV. The value is still over the benchmark under a conservative estimate. At source 3, the high velocity aerosols generated by the infector sneezing are directly sent into the exhaust vent, thus to realize an acceptable burden of the exposed people. In general, ICV shows excellent advantages in reducing aerosols concentration in the respiratory region of exposed people. The indoor exposed population burden mostly meets the WHO baseline requirements under ICV from the perspective of human health. However, this research only studied the potential impact of the health risks in some specific scenarios. Deeper research on the real-life and more scenarios of public health protection are still need to be investigated in the future. 5 Conclusions This study introduced a typical office as the background scene, where an infected person coughs, sneezes with standing or sitting at different positions. The cross-infection control performance of three ventilation strategies, namely mixing ventilation (MV), stratum ventilation (SV), and interactive cascade ventilation (ICV), was explored and compared with a validated CFD model. The related conclusions can provide a basis for designing and selecting the reasonable ventilation systems under the normal situation of the epidemic. The main conclusions of this study are as follows:(1) Compared with SV and ICV, the cross-infection control performance of MV is the weakest. The location and behavior of the infector show a significant impact on the exposed person at other locations under MV. The diffusion range of pollutants under MV is even increased three-fold over ICV. When the pollution source was near the air supply inlets, the pollutant diffusion range becomes more expansive, increasing the exposure doses by 52 % under discussed ventilation methods. The exposure does under ICV is 0.29 g/day, which can be reduced by 67 % and 50 % compared with MV and SV, respectively. (2) The health risk of the exposed person under ICV is 0.21×10−3 DALYs pppy, which is consistently lower than SV by 32 % and MV by 69 %. The health risk is greatest under MV while minimum under ICV for all exposed people in all discussed scenarios. (3) The average disease health burden for exposed persons under ICV is 0.93 × 10−6 DALYs pppy, which is consistently 37 % and 70 % lower than SV and MV, respectively. Compared to coughing, aerosols produced by sneezing of the infector under MV hardly meet the disease burden criterion. However, the percentages of the scenarios that meet the criterion under SV and ICV are 33 %and 75 %, respectively. It can be concluded that ICV is a feasible ventilation strategy for the crowded indoor environment, which can effectively reduce the cross-infection risk and improve virus prevention and control. However, it is noted that the concentrated low velocity air supply inlets are employed in the MV model in this study, which may weaken the evaluation effect of virus control of MV. Hence, it is also worth studying on the effect of different air inlet arrangement and supply air velocity on virus prevention and control under different ventilation methods. Moreover, this study is focused on the virus prevention and control performance of the ventilations in winter scenario, which is the high outbreak season. Nevertheless, the load distribution, airflow distribution, and thermal buoyancy direction are different and more complex in summer scenario. The related conclusions based on the heating conditions cannot be applied directly under cooling condition. Therefore, future work will be comprehensively conducted under the summer conditions to improve the main conclusions obtained at this stage. Author statement Han Li: Conceptualization, Methodology, Formal analysis, Investigation, Data Curation, Writing - Original Draft. Yuer Lan: Conceptualization, Methodology, Formal analysis, Investigation, Writing - Review & Editing. Xiuqin Ma: Investigation, Resources. Xiangfei Kong: Conceptualization, Methodology, Funding acquisition, Writing-Review & Editing. Man Fan: Investigation, Resources, Review & Editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix A Supplementary data The following is the Supplementary data to this article:Multimedia component 1 Multimedia component 1 Data availability Data will be made available on request. Acknowledgment This work is supported by 10.13039/501100001809 National Natural Science Foundation of China (Project No. 52008147), Hebei Province Funding Project for Returned Scholars, China (Project No. C20190507) and Fundamental Research Funds of 10.13039/501100010850 Hebei University of Technology (Project No. JBKYTD2003). Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.jobe.2022.105728. ==== Refs References 1 Zheng Wandong Hu Jingfan Wang Zhaoying Li Jinbo Fu Zheng Li Han Jurasz Jakub Chou S.K. Yan Jinyue COVID-19 impact on operation and energy consumption of heating, ventilation and air-conditioning (HVAC) systems Adv. Appl. Energy 3 2021 100040 10.1016/j.adapen.2021.100040 2 Pease Leonard F. Wang Na Salsbury Timothy I. Underhill Ronald M. Flaherty Julia E. Vlachokostas Alex Kulkarni Gourihar James Daniel P. Investigation of potential aerosol transmission and infectivity of SARS-CoV-2 through central ventilation systems Build. Environ. 197 15 June 2021 107633 10.1016/j.buildenv.2021.107633 33531734 3 Quinones Jhon J. Doosttalab Ali Sokolowski Steven Voyles Richard M. Castano Victor Zhang Lucy T. Castillo Luciano Prediction of respiratory droplets evolution for safer academic facilities planning amid COVID-19 and future pandemics: a numerical approach J. Build. Eng. 54 15 August 2022 104593 10.1016/j.jobe.2022.104593 4 Ren Chen Cao Shi-Ji Haghighat Fariborz A practical approach for preventing dispersion of infection disease in naturally ventilated room, J. Build. Eng. 48 1 May 2022 103921 10.1016/j.jobe.2021.1039219 5 Statista WHO advice for the public: Coronavirus disease (COVID-19) https://www.who.int/emergencies/diseases/novel-coronavirus-2019/advice-for-public 2022 6 Statista WHO/Europe urges countries to take post COVID-19 condition seriously by urgently investing in research, recovery, and rehabilitation https://www.who.int/europe/news/item/13-09-2022-at-least-17-million-people-in-the-who-european-region-experienced-long-covid-in-the-first-two-years-of-the-pandemic–millions-may-have-to-live-with-it-for-years-to-come 2022 7 Liao Mingrui Liu Huayang Wang Xi Hu Xuzhi Huang Yuhao Liu Xuqing Brenan Keith Mecha Jared Nirmalan Mahesan Lu Jian Ren A technical review of face mask wearing in preventing respiratory COVID-19 transmission Curr. Opin. Colloid Interface Sci. 52 April 2021 101417 10.1016/j.cocis.2021.101417 33642918 8 Choi Younhee Song Doosam Yoon Sungmin Koo Junemo Comparison of factorial and Latin hypercube sampling designs for meta-models of building heating and cooling loads Energies 512 2 2021 14 10.3390/en14020512 9 Ai Zhengtao Ming Mak Cheuk Naiping Gao Niu Jianlei Mitigating Tracer gas is a suitable surrogate of exhaled droplet nuclei for studying airborne transmission in the built environment Build. Simulat. 13 2020 489 496 10.1007/s12273-020-0614-5 10 Ren Juan Wang Yue Liu Qibo Liu Yu Numerical study of three ventilation strategies in a prefabricated COVID-19 inpatient ward Build. Environ. 188 2021 107467 10.1016/j.buildenv.2020.107467 33223598 11 He Qibin Niu Jianlei Gao Naiping Zhu Tong Wu Jiazheng CFD study of exhaled droplet transmission between occupants under different ventilation strategies in a typical office room Build. Environ. 46 2011 397 408 10.1016/j.buildenv.2010.08.003 32288015 12 Lu Yalin Oladokun Majeed Lin Zhang Reducing the exposure risk in hospital wards by applying stratum ventilation system Build. Environ. 183 2020 107204 10.1016/j.buildenv.2020.107204 13 Kong Xiangfei Guo Chenli Lin Zhang Duan Shasha He Junjie Ren Yue Ren Jianlin Experimental study on the control effect of different ventilation systems on fine particles in a simulated hospital ward Sustain. Cities Soc. 73 2021 103102 10.1016/j.scs.2021.103102 34189016 14 Lin Zhang Wang Jinliang Yao Ting Chow T.T. Investigation into anti-airborne infection performance of stratum ventilation Build. Environ. 54 August 2012 29 38 10.1016/j.buildenv.2012.01.017 15 Li Han Li Jinchao Fan Man Wang Zhaoying Li Wei Kong Xiangfei Study on the performance of interactive cascade ventilation oriented to the non-uniform indoor environment requirement Energy Build. 253 2021 111539 10.1016/j.enbuild.2021.111539 16 Mbanga Joshua Luther King Abia Akebe Gyamfi Amoako Daniel Essack Sabiha Y. Quantitative microbial risk assessment for waterborne pathogens in a wastewater treatment plant and its receiving surface water body BMC Microbiol. 20 1 2020 1 12 10.1186/s12866-020-02036-7 31896348 17 Ren Chen Xi Chang Wang Junqi Feng Zhuangbo Nasiri Fuzhan Cao Shi-Jie Haghighat Fariborz Mitigating COVID-19 infection disease transmission in indoor environment using physical barriers Sustain. Cities Soc. 74 November 2021 103175 10.1016/j.scs.2021.103175 34306996 18 Kong Xiangfei Wang Zhaoying Fan Man Li Han Analysis on the performance of interactive cascade ventilation for space heating based on non-uniform indoor environment demand Build. Environ. 219 2022 109244 10.1016/j.buildenv.2022.109244 19 GB50736 Design code for heating ventilation and air conditioning of civil buildings 2012 Ministry of Housing and Urban-Rural Development China 20 Zhang Sheng Lin Zhang Ai Zhengtao Wang Fenghao Cheng Yong Huan Chao Effects of operation parameters on performances of stratum ventilation for heating mode Build. Environ. 148 2019 55 66 10.1016/j.buildenv.2018.11.001 21 Ai Zhengtao Ming Mak Cheuk Naiping Gao Niu Jianlei Mitigating Tracer gas is a suitable surrogate of exhaled droplet nuclei for studying airborne transmission in the built environment Build. Simulat. 13 2020 489 496 10.1007/s12273-020-0614-5 22 Bu Yunchen Ooka Ryozo Kikumoto Hideki Oh Wonseok Recent research on expiratory particles in respiratory viral infection and control strategies: a review Sustain. Cities Soc. 73 2021 103106 10.1016/j.scs.2021.103106 34306994 23 Bin Zhao., Zhang Zhao, Xianting Li, Dongtao Huang, Comparison of Diffusion Characteristics of Aerosol Particles in Different Ventilated Rooms by Numerical Method, ASHRAE Transactions 110, 88-95. https://www.researchgate.net/publication/283863540 24 Leng J.W. Wang Q. Liu K. Sustainable design of courtyard environment: from the perspectives of airborne diseases control and human health Sustain. Cities Soc. 62 2020 102405 10.1016/j.scs.2020.102405 32834938 25 Owen M.K. Ensor D.S. Airborne particle sizes and source found in indoor air Atmos. Environ. 26A 12 1992 2149 2162 10.1016/0960-1686(92)90403-8 26 Busco Giacomo Yang Se Ro Seo Joseph Hassan Yassin A. Sneezing and asymptomatic virus transmission [J] Phys. Fluids 32 2020 10.1063/5.0019090 073309 27 Zhu Shengwei Kato Shinsuke Yang Jeong-Hoon Study on transport characteristics of saliva droplets produced by coughing in a calm indoor environment Build. Environ. 41 2006 1691 1702 10.1016/j.buildenv.2005.06.024 28 Alonso Maria Justo Andresen Trond Frydenlund Frode Widell Kristina N. Improvements of air flow distribution in a freezing tunnel using Airpak Procedia Food Sci. 1 2011 1231 1238 10.1016/j.profoo.2011.09.183 29 Chen Qingyan Comparison of different k-ε models for indoor air flow computations, Numerical Heat Transfer Part B: Fund. 28 3 1995 353 369 10.1080/10407799508928838 30 Lu Yalin Oladokun Majeed Lin Zhang Reducing the exposure risk in hospital wards by applying stratum ventilation system Build. Environ. 183 2020 107204 10.1016/j.buildenv.2020.107204 31 Zhang Zhao Qingyan Chen, Comparison of the Eulerian and Lagrangian methods for predicting particle transport in enclosed spaces Atmos. Environ. 41 25 2007 5236 5248 10.1016/j.atmosenv.2006.05.086 32 Gao Naiping Niu Jianlei Morawska Lidia Distribution of respiratory droplets in enclosed environments under different air distribution methods Build. Simulat. 1 2008 326 335 10.1007/s12273-008-8328-0 33 He Qibin Niu Jianlei Gao Naiping Zhu Tong Wu Jiazheng CFD study of exhaled droplet transmission between occupants under different ventilation strategies in a typical office room Build. Environ. 46 2011 397 408 10.1016/j.buildenv.2010.08.003 32288015 34 Lin Zhang Wang Jinliang Yao Ting Chow T.T. Zhu Tong Wu Jiazheng Investigation into anti-airborne infection performance of stratum ventilation Build. Environ. 54 2012 29 38 10.1016/j.buildenv.2012.01.017 35 Gao Naiping Niu Jianlei Morawska Lidia Distribution of respiratory droplets in enclosed environments under different air distribution methods Build. Simulat. 1 2008 326 335 10.1007/s12273-008-8328-0 36 Nicas M. Nazaroff W.W. Hubbard A. Toward understanding the risk of secondary airborne infection: emission of respirable pathogens J. Occup. Environ. Hyg. 2 2005 143 154 10.1080/15459620590918466 15764538 37 Qian H. Li Y. Nielsen P.V. Hyldgaard C.E. Wong T.W. Chwang A.T.Y. Dispersion of exhaled droplet nuclei in a two-bed hospital ward with three different ventilation systems Indoor Air 16 2006 111 128 10.1111/j.1600-0668.2006.00449.x 16507039 38 Leng Jiawei Wang Qi Liu Ke Sustainable design of courtyard environment: from the perspectives of airborne diseases control and human health Sustain. Cities Soc. 62 2020 102405 10.1016/j.scs.2020.102405 32834938 39 Shi Kuang-Wei Wang Cheng-Wen Jiang Sunny C. Quantitative microbial risk assessment of Greywater on-site reuse Sci. Total Environ. 635 2018 1507 1519 10.1016/j.scitotenv.2018.04.197 29710672 40 National Research Council Risk assessment in the federal government: managing the process https://refhub.elsevier.com/S0048-9697(15)00352-6/rf0035 41 Lim Keah-Ying Hamilton Andrew J. Jiang Sunny C. Assessment of public health risk associated with viral contamination in harvested urban stormwater for domestic application Sci. Total Environ. 523 2015 95 108 10.1016/j.scitotenv.2015.03.077 25863500 42 Ali Wajid Yang Ya-fei Gong Ling Yan Cheng Cui Bei-bei Emission characteristics and quantitative health risk assessment of bioaerosols in an indoor toilet after flushing under various ventilation scenarios Build. Environ. 207 2022 108463 10.1016/j.buildenv.2021.108463 43 Jahne M.A. Rogers S.W. Holsen T.M. Grimberg S.J. Quantitative microbial risk assessment of bioaerosols from a manure application site Aerobiologia 31 1 2015 73 87 10.1007/s10453-014-9348-0 44 Watanabe T. Bartrand T.A. Weir M.H. Omura T. Haas Development of a dose-response model for SARS coronavirus Risk Anal. 30 7 July 2010 10.1111/j.1539-6924.2010.01427.x 45 Jahne M.A. Rogers S.W. Holsen T.M. Grimberg S.J. Quantitative microbial risk assessment of bioaerosols from a manure application site Aerobiologia 31 1 2015 73 87 10.1007/s10453-014-9348-0 46 Fan Chiao-Yun Fan Jean Ching-Yuan Yang Ming-Chin Lin Ting-Yu Chen Hsiu-Hsi Liu Jin-Tan Yang Kuen-Cheh Estimating global burden of COVID-19 with disability-adjusted life years and value of Journal of the Formosan Medical Association 120 S106eS117 10.1016/j.jfma.2021.05.019 2021 47 Hang Jian Li Yuguo Ching W.H. Wei Jianjian Jin Ruiqiu Liu Li Xie Xiaojian Potential airborne transmission between two isolation cubicles through a shared anteroom Build. Environ. 89 2015 264 278 10.1016/j.buildenv.2015.03.004 32288029 48 Duguid J.P. The size and the duration of air-carriage of respiratory droplets and droplet-nuclei[J] J. Hyg. 44 6 1946 471 479 10.1017/S0022172400019288 20475760 49 Bivolarova M. Ondracek J. Melikov A. Zdimal V. A comparison between tracer gas and aerosol particles distribution indoors: the impact of ventilation rate, interaction of airflows, and presence of objects, Indoor Air 27 2017 1201 1212 10.1111/ina.12388 28378912 50 Huan Chao Wang Fenghao Lin Zhang Wu Xiaozhou Ma Zhenjun Wang Zhihua Zhang Linhua An experimental investigation into stratum ventilation for the cooling of an office with asymmetrically distributed heat gains Build. Environ. 110 2016 76 88 10.1016/j.buildenv.2016.09.031 51 Anand Prashant Cheong David Sekhar Chandra Computation of zone-level ventilation requirement based on actual occupancy, plug and lighting load information Indoor Built Environ. 29 2020 558 574 10.1177/1420326X19875802 52 Lu Yalin Niu Dun Zhang Sheng Chang Han Lin Zhang Ventilation indices for evaluation of airborne infection risk control performance of air distribution plug and lighting load information Build. Environ. 222 2022 109440 10.1016/j.buildenv.2022.109440 35937047

0

PMC9753475

NO-CC CODE

2022-12-16 23:25:44

no

2022 Dec 15;:105728

utf-8

oa_other

==== Front Physiol Behav Physiol Behav Physiology & Behavior 0031-9384 1873-507X Published by Elsevier Inc. S0031-9384(21)00250-X 10.1016/j.physbeh.2021.113561 113561 Editorial Introduction to ingestive behavior research across the generations (society for the study of ingestive behavior collection 2020) Tracy Andrea L. a⁎ Temple Jennifer L. b a Department of Psychology, Grinnell College, Grinnell, IA 50112, USA b Department of Exercise and Nutrition Sciences, University at Buffalo, Buffalo, NY 14214, USA ⁎ Corresponding author. 19 8 2021 1 11 2021 19 8 2021 241 113561113561 © 2021 Published by Elsevier Inc. 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcThe Covid-19 pandemic led to many unprecedented changes and challenges in the year 2020, including the cancellation of many scientific meetings. Although the Society for the Study of Ingestive Behavior was unable to gather in-person and share our findings as originally planned, we wanted to continue the tradition of the annual Article Collection in Physiology & Behavior highlighting outstanding research in the field. This special issue would typically contain papers from contributors to the Annual Meeting, but in the absence of the Annual Meeting, we saw an opportunity to reflect on the history of research into the psychology and physiology of eating and drinking behaviors and to look forward to where this work is headed. SSIB prides itself on its supportive environment for young scientists. One hallmark of this is that strong mentorship and collaboration between junior and senior investigators are highly valued by the Society and its members. Thus, we were excited to invite long-time members of SSIB to work with their trainees and up-and-coming colleagues to write manuscripts highlighting novel advances they are making in their respective research areas, to review their own past contributions, and to speculate on the future of research in the field of ingestive behavior from a cross-generational perspective. In this Article Collection, we are pleased to present the work of many members who have made outstanding contributions to the Society, both through their research and service, including ten current or former SSIB presidents (Harry Kissileff, Barbara Rolls, Tim Moran, Marion Hetherington, Allen Levine, Thomas Lutz, Linda Rinaman, Suzanne Higgs, Derek Daniels, and Rick Samson). Ingestive behavior scientists have long framed discussions of food intake in terms of factors that promote eating and those that inhibit eating. Epstein & Carr tackle both sides of this equation, assessing the relationships between individual differences in food reinforcement, habituation, and their effects on starting and ending meals [1]. Another factor known to increase food intake is the presence of others and, in this issue, Ruddock, Brunstrom, & Higgs review the factors involved in this “social facilitation of eating” and discuss a potential evolutionary mechanism that may help us understand this phenomenon [2]. Chawner & Hetherington tackle the perennial question of how to get kids to eat their vegetables, evaluating the experimental literature and assessing how these techniques might be further adapted in real world applications [3]. On the inhibitory side of the food intake equation, Cunningham & Rolls discuss their development of the Reasons Individuals Stop Eating Questionnaire (RISE-Q) and how this has lead to the Satiation Framework, a dynamic set of behavioral and psychological processes contributing to meal termination [4]. Honegger, Lutz, & Boyle present the use of a novel glucose clamp paradigm in order to look into the biological aspects of food reduction and find that the ability of amylin to reduce food intake is lessened under conditions of hypoglycemia [5]. Food palatability plays a substantial role in intake as well, interacting with biological systems as shown by Head and colleagues, who demonstrate the combined effects of oxytocin and naloxone to reduce consumption of foods high in fat and sugar [6]. Armitage, Iatridi, & Yeomans look at individual differences in food preference, describing three phenotypes of “sweet liking” in humans and how this influences awareness of interoceptive appetitive states and the effects on intake [7]. Psychological contexts, such as stress, have long been shown to alter food choice and intake and, in this issue, Hamm et al. present findings from a novel application of a virtual portion selection paradigm to assess “stress eating" and demonstrate that this visual selection of virtual foods is a valid and useful strategy under these conditions [8]. You will also find an excellent set of clinical perspectives on how we understand and treat both eating disorders and obesity. Smith & Moran [9] describe the role of gastrointestinal peptides in eating disorders with an eye toward pharmacotherapy, while Melles, Spix, & Jansen [10] propose the use of inhibitory learning therapies to target avoidance behaviors in anorexia nervosa. Eichen et al. present evidence for the use Compensatory Cognitive Training in conjuction with behavioral weight loss strategies in advocating for the use of their NEXT (Novel Executive Function Training for Obesity) paradigm to treat obese individuals [11]. Finally, Chen, Singh, & Lowe tackle the “food restriction wars” in attempt to bring together multiple models of food restriction and dieting and understand their current and future role in eating disorder and obesity interventions [12]. Ingestive behavior is not limited to food intake. Mietlicki-Baase, Santollo, & Daniels review the role dopamine in appetitive and consummatory motivation for fluid ingestion, and evaluating the role of neuropeptides traditionally viewed as relevant for food intake and sex differences in generating a framework for future work in this area [13]. Control of ingestive behavior is not limited to the role of the brain, as demonstrated by the work presented here from Raybould & Zumpano [14]. Their work demonstrates the clear and critical role for gut microbiota in the regulation of food intake and body weight and the critical need for furthering this research on the gut-brain axis for a fuller understanding of the mechanisms controlling ingestive processes. Physiological systems involved in ingestive behavior frequently play roles in other behavioral and biological processes that interact with food intake. As investigated in novel experiments here, Edwards, Dolezel, & Rinaman show how food deprivation states influence passive avoidance memory and demonstrate the effect of sex and the role of A2 noradrenergic neurons in this behavior [15]. Yosten and collegues’ work reviews the role of the CART (cocaine- and amphetamine-regulated transcript) peptide in food and fluid ingestion, pain sensation, and other psychological processes, emphasizing the importance of the identification of the GPR160 receptor as a key for future research on the function of this peptide and potential pharmacotherapeutic approaches [16]. Though the range of topics covered in this issue represents only a small fraction of the research carried out by the members of the Society for the Study of Ingestive Behavior over the 34 years since its incorporation, it exemplifies the rigorous, creative, and important science being conducted in this field. We can definitively conclude from this Ingestive Behavior Research Across the Generations that the past, present, and future of SSIB is bright. ==== Refs References 1 Epstein Leonard Carr Katelyn Food reinforcement and habituation to food are processes related to initiation and cessation of eating Physiology & Behavior 239 2021 113512 10.1016/j.physbeh.2021.113512 In this issue 2 Ruddock Helen K Brunstrom Jeffrey M Higgs Suzanne The social facilitation of eating: why does the mere presence of others cause an increase in energy intake? Physiology & Behavior 240 2021 113539 10.1016/j.physbeh.2021.113539 In this issue 3 Chawner L R Hetherington M M Utilising an integrated approach to developing liking for and consumption of vegetables in children. Physiology & Behavior 236 2021 113493 10.1016/j.physbeh.2021.113493 In this issue 4 Cunningham Paige M Rolls Barbara J The Satiation Framework: Exploring processes that contribute to satiation Physiology & Behavior 236 2021 113419 10.1016/j.physbeh.2021.113419 In this issue 5 Honegger Miriam Lutz Thomas A Boyle Christina N Hypoglycemia attenuates acute amylin-induced reduction of food intake in male rats Physiology & Behavior 237 2021 113435 10.1016/j.physbeh.2021.113435 In this issue 6 Head Mitchell A Levine Allen S Christian David G Klockars Anica Olszewski Pawel K Effect of combination of peripheral oxytocin and naltrexone at subthreshold doses on food intake, body weight and feeding-related brain gene expression in male rats Physiology & Behavior 2021 113464 10.1016/j.physbeh.2021.113464 In this issue 7 Armitage Rhiannon M Iatridi Vasiliki Yeomans Martin R Understanding sweet-liking phenotypes and their implications for obesity: Narrative review and future directions Physiology & Behavior 235 2021 113398 10.1016/j.physbeh.2021.113398 In this issue 8 Hamm Jeon D Klatzkin Rebecca R Herzog Musya Tamura Shoran Brunstrom Jeffrey M Kissileff Harry R Recalled and momentary virtual portions created of snacks predict actual intake under laboratory stress condition Physiology & Behavior 238 2021 113479 10.1016/j.physbeh.2021.113479 In this issue 9 Smith Kimberly R Moran Timothy H Gastrointestinal peptides in eating-related disorders Physiology & Behavior 238 2021 113456 10.1016/j.physbeh.2021.113456 In this issue 10 Melles Hanna Spix Michelle Jansen Anita Avoidance in Anorexia Nervosa: Towards a research agenda Physiology & Behavior 238 2021 113478 10.1016/j.physbeh.2021.113478 In this issue 11 Eichen Dawn M Pasquale Ellen K Twamley Elizabeth W Boutelle Kerri N Targeting executive function for weight loss in adults with overweight or obesity Physiology & Behavior 240 2021 113540 10.1016/j.physbeh.2021.113540 In this issue 12 Chen Joanna Y Singh Simar Lowe Michael R The food restriction wars: Proposed resolution of a primary battle Physiology & Behavior 240 2021 113530 10.1016/j.physbeh.2021.113530 In this issue 13 Mietlicki-Baase Elizabeth G Santollo Jessica Daniels Derek Fluid intake, what’s dopamine got to do with it?Author links open overlay panel Physiology & Behavior 236 2021 113418 10.1016/j.physbeh.2021.113418 In this issue 14 Raybould Helen E Zumpano Danielle L Microbial metabolites and the vagal afferent pathway in the control of food intake Physiology & Behavior 240 2021 113555 10.1016/j.physbeh.2021.113555 In this issue 15 Edwards Caitlyn M Dolezal Tyla Rinaman Linda Sex and metabolic state interact to influence expression of passive avoidance memory in rats: Potential contribution of A2 noradrenergic neurons Physiology & Behavior 239 2021 113511 10.1016/j.physbeh.2021.113511 In this issue 16 Yosten Gina L C Haddock Christopher J Harada Caron M Almeida-Pereira Gislaine Kolar Grant R Stein Lauren M Hayes Matthew R Salvemini Daniela Samson Willis K Past, present and future of cocaine- and amphetamine-regulated transcript peptide Physiology & Behavior 235 2021 113380 10.1016/j.physbeh.2021.113380 In this issue

34419457

PMC9753481

NO-CC CODE

2022-12-16 23:25:44

no

Physiol Behav. 2021 Nov 1; 241:113561

utf-8

Physiol Behav

10.1016/j.physbeh.2021.113561

oa_other

==== Front J Infect Chemother J Infect Chemother Journal of Infection and Chemotherapy 1341-321X 1437-7780 Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. S1341-321X(22)00327-0 10.1016/j.jiac.2022.12.004 Original Article Validation of a specialized evaluation system for COVID-19 in Japan: A retrospective, multicenter cohort study Furuhata Hiroki a∗ Araki Kenji b a Graduate School of Medicine and Veterinary Medicine, University of Miyazaki, 5200 Kibara Kiyotake-cho, Miyazaki, 8891692, Japan b Department of Patient Advocacy Center, University of Miyazaki Hospital, 5200 Kibara Kiyotake-cho, Miyazaki, 8891692, Japan ∗ Corresponding author. Department of Hospital Institutional Research, University of Miyazaki Hospital, Kibara 5200, Kiyotake-cho, Miyazaki City, Miyazaki Prefecture, Japan. 15 12 2022 15 12 2022 8 9 2022 22 11 2022 7 12 2022 © 2022 Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. All rights reserved. 2022 Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Introduction Evaluation of a severity grade (SG) is important to classify patients for efficient use of limited medical resources. This study validates two existing evaluation systems for the prevention of the coronavirus disease 2019 (COVID-19) in Japan: a criterion of SG and a list of 14 specialized underlying diseases (SUDs). Methods A retrospective cohort was created using electronic medical records from 18 research institutes. The cohort includes 6,050 COVID-19 patients with two types of diagnosis information as follows: SG at hospitalization among mild, moderate I, moderate II, and severe and aggravation after hospitalization. Results A crude mortality rate and an aggravation rate increased by the worsening of SG in the COVID-19 cohort. The transition of the aggravation rate was notable for COVID-19 patients with SUD. A conditional probability of the mortality given the aggravation in the COVID-19 cohort was 87.4% compared to mild or moderate patients (approximately 21%–45%) who have the possibility of the aggravation. An odds ratio of the mortality and aggravation information about the SUD list was higher than other variables. Conclusions We demonstrated the possibility of improving the criteria of SG by including the SUD list for more effective operation of the criteria of SG. Furthermore, we demonstrated the importance of the prevention of the aggravation based on the conditional probability, and the possibility of predicting the aggravation using the risk factors. Keywords COVID-19 Infection Severity Underlying disease ==== Body pmc1 Introduction 1.1 Background Evaluation of a severity grade (SG) is important to classify patients with an infectious disease not limited to the coronavirus disease 2019 (COVID-19) for isolating them. In fact, one study summarized various factors to evaluate the severity of COVID-19 in early stage of its pandemic [1]. In addition, various studies have analyzed details of these factors such as cytokines [2,3], hypertension [4], coexistence of cancer [5], and lymphopenia [6]. Moreover, studies have demonstrated that a typical negative lifestyle habit could constitute the risk factor such as obesity [[7], [8], [9]] and smoking [10]. Considering that one study discussed a prediction model as a helpful tool for evaluating the severity and prognosis of a patient [11], these factors could become explorative variables of this model. However, applying this model to an actual diagnosis by each new inpatient is difficult because these variables are always not recorded in electronic medical records (EMRs) for the execution of this model. Similarly, some studies have applied the existing risk scoring system of pneumonia on COVID-19 patients, such as the A-DROP score [[12], [13], [14]], regarding various meta-analysis studies that have evaluated the negative effects of typical risk factors on COVID-19 patients with pneumonia [[15], [16], [17], [18], [19]]. On the other hand, various guidelines for COVID-19 have been published for situations such as an overview [20,21], a treatment plan by each severity [22,23] and that of an individual situation (e.g., perinatal medicine [24], inpatients without an intensive care [25], and drug recommendation [26]). Similarly, the Ministry of Health, Labour and Welfare provides the guideline of COVID-19 within a criterion of SG among mild, moderate I, moderate II, and severe [27]. In particular, we believe that the criterion of severe patients is important for the efficient operation of limited medical resources because most severe patients have to wear a mechanical ventilator (MV) or be admitted to an intensive care unit (ICU). Therefore, it is necessary to validate the criterion of SG to be able to classify patients irrespective of their nationality. In addition to the criterion of SG, the Ministry of Health, Labor and Welfare lists 14 specialized underlying diseases (SUDs) [28] to classify those who should prioritized for vaccination. We believe that SUD is useful to classify severe patients more efficiently because previous studies have demonstrated that certain diseases could become the risk factor in COVID-19 patients [[4], [5], [6]]. However, the criterion of SG and the SUD list are now independently operated in Japan. Therefore, we expect to improve the criterion of SG by including the list. 1.2 Study objective We aim to validate two existing evaluation systems—the criteria of SG and the SUD list—that were designed in Japan toward an efficient and appropriate provision of medical treatment under conditions of limited medical resources. 2 Patients and methods 2.1 Study design and participants A retrospective cohort of patients treated from April 1, 2019, to September 30, 2021, was created using EMRs collected from 26 cooperative research institutes. These records contain the Diagnostic Procedure Combination (DPC) data submitted by acute hospitals for the social insurance system in Japan. The DPC data included fundamental patient information, such as sex, age, and disease, and a record of hospital charges by treatment. Although 27 institutes originally participated in our study, one institute was excluded due to the lack of the DPC data. Since we aim to evaluate the situation of each institute, this study does not use a sample size estimation that a typical clinical trial is required to include. Fig. 1 depicts the analysis dataset creation process for our data analysis using the cohort from the 26 institutes. There are four exclusion criteria: (1) any research institute completely lacks monthly data of hospital charges; (2) any individual record not showing COVID-19; (3) any missing value in a body mass index (BMI) that the SUD list requires to record; or (4) any past hospitalization in patients whose hospitalization time is multiple. The 10th revision of the International Statistical Classification of Diseases and Related Health Problems (ICD-10) is used to extract patients with COVID-19. Those who recorded U07.1 or U07.2 in the main disease, diseases in which the most medical resources were used, or diseases that were a direct trigger of hospitalization are considered as COVID-19 patients. The analysis dataset was created following this exclusion and was divided following four SGs evaluated at hospitalization (see Table 1 ).Fig. 1 Data processing flowchart. Fig. 1 Table 1 Definition of severity criteria. Table 1Severity Symptom Participants (n = 6,050) Number % Mild Nothing in below 4,662 77.1 Moderate I All 346 5.7 With dyspnea 4 0.1 With pneumonia 343 5.7 Moderate II Necessity of oxygen inhalation 788 13.0 Severe All 254 4.2 Admission to ICU 0 0.0 Necessity of MV 254 4.2 Abbreviations: ICU, Intensive care unit; MV, Mechanical ventilator. 2.2 Definition of the criterion of SG and the SUD list using data items Table 1 highlights a definition of SG of COVID-19 at hospitalization using existing data items. The description “all” in the table refers to those who fall under one or more detail items (e.g., with dyspnea and with pneumonia in moderate I). Initially, patients who wore MV on the first day of hospitalization or who were admitted to the ICU, retrospectively, were considered severe patients. Next, those who were administered with oxygen on the first day of hospitalization were considered as moderate II patients. Finally, patients whose comorbidity is dyspnea (ICD-10 is R06.0) or pneumonia (ICD-10 is J10.0 or J11.0, or the first three headings of ICD-10 is J12 to J18) were considered as moderate I patients, and the remaining patients were considered as mild patients. Although saturation of percutaneous oxygen (SpO2) is used to classify mild or moderate patients, this study did not use it because most patients did not record its value upon hospitalization in our existing database within the DPC data or otherwise, such as through a laboratory inspection record. Table 2 highlights a definition of the 14 SUDs based on the code of ICD-10 of the comorbidity at hospitalization except SUD8, SUD9, SUL10, SUD12, and SUD14 due to the lack of corresponding codes for these four diseases. Furthermore, we classified patients into four subgroups using their age and BMI (I: age <65 & BMI <30; II: age <65 & BMI ≥ 30; III: age ≥ 65 & BMI <30; and IV: age ≥ 65 & BMI ≥ 30).Table 2 Definition of fourteen SUDs. Table 2SUD Corresponding code (ICD-10) Participants (n = 6,050) Number % SUD1 Chronic respiratory disease J31.x, J32.x, J35.x, J37.x, J40.x-J47.x, J68.4, J70.1, J70.3, J95.3, J96.1 282 4.7 SUD2 Chronic heart disease within high blood pressure I10.x-I15.x, I25.x, I31.x, I48.2 826 13.7 SUD3 Chronic nephric disease N18.x 214 3.5 SUD4 Chronic liver disease K71.3, K71.4, K71.5, K72.1, K73.x 6 0.1 SUD5 Diabetes E10.x-E14.x 1,030 17.0 SUD6 Blood disease except iron deficiency anemia D51.x-D77.x 125 2.1 SUD7 Immunodeficiency within cancer Cxx.x, D80.x-D89.x 328 5.4 SUD8 Treatment with something to make immune function be worse Unavailable SUD9 Neural or neuromuscular disease caused by immunodeficiency Unavailable SUD10 Physical decline caused by neural or neuromuscular disease Unavailable SUD11 Chromosomal aberration Q90.x-Q99.x 2 0.0 SUD12 Patients with both physical and intellectual severe disability Unavailable SUD13 Sleep apnea syndrome G47.3 20 0.3 SUD14 Severe mental disorder Unavailable Abbreviations: ICD-10, The 10th revision of the International Statistical Classification of Diseases and Related Health Problems; SUD, Specialized underlying disease. Notes: The digit “x” in the corresponding code such as “J31.x” and “Cxx.x” means that the numeric code in ICD-10 is not used to decide whether patients contract SUD or not. 2.3 Statistical methods Our primary analysis compared a crude mortality rate (CMR) and an aggravation rate (AGR) by the cohort to evaluate whether the criteria of SG can classify patients. Here, the aggravation is defined as those who were not severe at hospitalization but characterized as being severe patients in Table 1 after hospitalization. Therefore, AGR is defined for only mild or moderate patients at hospitalization as a ratio of aggravated patients and all patients. First, we compared CMR by SG at hospitalization. We expected this comparison to demonstrate the increase of CMR by the worsening of SG. Second, we compared AGR and CMR in aggravated patients. We expected this comparison to demonstrate the influence of the aggravation on the increase of CMR. In similar, we dealt with the comparison by whether patients contract any SUD or not. Here, a P value is calculated using Fisher's exact test for the comparison above. Finally, we calculated a conditional probability (CP) of the mortality given SG at hospitalization or the aggravation using the Bayes' theorem as follows:P(Si|M)=P(M)P(M|Si)P(M)P(M|S1)+P(M)P(M|S2)+P(M)P(M|S3) P(A|M)=P(M)P(M|A)P(M)P(M|A)+P(M)P(M¯|A) P(M|X)=P(M∩X)P(X)={N(M∩X)N(T)}/{N(X)N(T)} whereas Si means three SGs (1: mild, 2: moderate I, and 3: moderate II), M means the mortality, A means the aggravation, X mean Si or A, and N(∙) means the number of those who fall under the value in brackets, where especially N(T) means the number of all patients. Each N(∙) can be seen in Table 4 . We expected this calculation to demonstrate which is more important in mild or moderate patients, the mortality by SG at hospitalization or the aggravation following hospitalization. Therefore, CP was calculated without severe patients.Table 3 CMR and AGR by SG and SUD. Table 3SUD SG All patients zAggravated patients Participants (n, %) Death (n) CMR (%) P value Participants (n, %) AGR (%) Death (n) CMR (%) P value AGR CMR All All 6,050 100.0 416 6.9 <0.001 350 100.0 5.8 107 30.6 <0.001 0.122 Mild 4,662 77.1 234 5.0 239 68.3 5.1 69 28.9 Moderate I 346 5.7 27 7.8 29 8.3 8.4 6 20.7 Moderate II 788 13.0 85 10.8 82 23.4 10.4 32 39.0 Severe 254 4.2 70 27.6 No All 3,942 100.0 220 5.6 <0.001 181 100.0 4.6 55 30.4 <0.001 0.733 Mild 3,125 79.3 113 3.6 <0.001 126 69.6 4.0 31 24.6 <0.001 0.011 Moderate I 186 4.7 14 7.5 12 6.6 6.5 3 25.0 Moderate II 491 12.5 54 11.0 43 23.8 8.8 21 48.8 Severe 140 3.6 39 27.9 Yes All 2,108 100.0 196 9.3 <0.001 169 100.0 8.0 52 30.8 0.003 0.416 Mild 1,537 72.9 121 7.9 113 66.9 7.4 38 33.6 Moderate I 160 7.6 13 8.1 17 10.1 10.6 3 17.6 Moderate II 297 14.1 31 10.4 39 23.1 13.1 11 28.2 Severe 114 5.4 31 27.2 Abbreviations: AGR, Aggravation rate; CMR, Crude mortality rate; SG, Severity grade; SUD, Specialized underlying disease. Notes: The P value in the rows that the column SG shows “Mild” means the results of comparison between the corresponding rate by SG at hospitalization. The word “All” means a comparison between the corresponding rates by whether patients contract any SUD. Table 4 CP of the mortality given SG or the aggravation. Table 4Given variable Category Participants (n, %) CP (%) Survival Mortality Total SG Mild 4,428 76.4 234 4.0 4,662 80.4 21.3 Moderate I 319 5.5 27 0.5 346 6.0 33.1 Moderate II 703 12.1 85 1.5 788 13.6 45.7 Aggravation No 5,207 89.8 239 4.1 5,446 94.0 12.6 Yes 243 4.2 107 1.8 350 6.0 87.4 Total 5,450 94.0 346 6.0 5,796 100.0 Abbreviations: CP, Conditional probability; SG, Severity grade. Our secondary analysis estimated an odds ratio (OR) of the mortality and the aggravation to explore a risk factor using a logistic regression in those who were not severe at hospitalization. Since this analysis would explore a potential risk of the mortality or the aggravation, severe patients were excluded for this analysis. This analysis includes both univariate and multivariate analyses. If all categories record OR with statistical significance (i.e., P value is less than 0.05), the corresponding explorative variable is included in the multiple analysis. However, sex, age, and BMI are excluded despite being statistically significant in the univariate analysis if the explorative variables classified in the SUD or A-DROP score created from these three variables (e.g., subgroup in SUD vs age and BMI) also show the statistical significance, retrospectively. The reason is that there is a high possibility of missed OR caused by a strong correlation between these three variables and variables in SUD or A-DROP score. In the secondary analysis, there are three classifications of these data items. The first is a data item about SUD: (1) the four subgroups defined by age and BMI; (2) number of contracting SUDs. The second is two items about the A-DROP score that can be created \using data items in our database as follows: (1) “A: Age” extracts male patients whose age is ≥ 70 or female patients whose age is ≥ 75; (2) “O: Orientation” extracts patients who record any value in the Japan Coma Scale at hospitalization. The third is a data item about patient basic characteristics: (1) sex; (2) age; (3) BMI; (4) a smoking index; (5) an activities daily living (ADL) at hospitalization. All statistical analyzes were performed using the R programming version 4.1.2, with the P value of <0.05 indicating statistical significance. 3 Results This study included 6,050 participants in the data analysis. Table 3 shows CMR and AGR by each SG. The number of aggravated patients is an internal number of all patients. Table 3 indicates that CMR and AGR increased by the worsening of SG despite the consideration of SUD. On the other hand, there was no diffrence of CMR in patients with aggravation by each SG. Furthermore, AGR in patients with SUD was higher than in those without SUD. Table 4 is a cross-tabulation to calculate CP of the mortality given SG or the aggravation in those who are not severe at hospitalization. CP is known as an inverse probability, one of the most popular frameworks of the Bayes’ theorem. We calculated CP, considering a probability of the cause (SG or the aggravation) given the effect (the mortality). Table 4 indicates that CP of the mortality given the aggravation is remarkably higher than that of mild or moderate patients. Table 5 shows OR of the mortality and the aggravation by each risk factor in those who were not severe at hospitalization. If OR was >2 with the statistical significance, the corresponding cell was enhanced. Supplementally, Appendix AAppendix A shows CMR and AGR by each risk factor in Table 5. The appendix B shows summary statistics of these risk factors in a numeric value. Table 5 indicates the risk factors of both the mortality and aggravation using OR when the reference category that we assumed was the lowest risk category. In particular, subgroups (II and IV) and ADL (abnormal) would be important to detect higher risk patients because their OR was >2 in both objective (mortality and aggravation) and regression styles (univariate and multiple).Table 5 OR estimation by each risk factor. Table 5(a) Mortality Risk factor Univariate Multiple Classification Variable Category Estimation [95% CI] P Value Estimation [95% CI] P Value SUD Subgroup I Ref. Ref. II 2.24 [1.35, 3.71] 0.001 2.08 [1.25, 3.46] 0.004 III 6.10 [4.52, 8.22] <0.001 3.30 [2.21, 4.93] <0.001 IV 8.56 [4.48, 16.33] <0.001 4.68 [2.36, 9.27] <0.001 Number of 0 Ref. Ref. SUDs 1 1.67 [1.31, 2.14] <0.001 1.05 [0.81, 1.36] 0.704 >1 2.14 [1.57, 2.93] <0.001 1.21 [0.87, 1.69] 0.245 A-DROP A: Age No Ref. Ref. Score Yes 3.88 [3.10, 4.86] <0.001 1.2 [0.87, 1.65] 0.274 O: Orientation No Ref. Ref. Yes 2.75 [2.05, 3.70] <0.001 1.28 [0.94, 1.75] 0.111 Basic Sex Male Ref. characteristic Female 0.85 [0.68, 1.07] 0.164 Age <20 0.03 [0.02, 0.03] <0.001 (years) 20 to 64 Ref. 65 to 74 3.55 [2.58, 4.88] <0.001 >74 5.71 [4.34, 7.52] <0.001 BMI <18.5 0.06 [0.05, 0.07] <0.001 18.5 to <25 Ref. 25 to <30 1.11 [0.86, 1.45] 0.419 30 to <35 0.86 [0.54, 1.35] 0.502 35 to <40 1.00 [0.43, 2.32] 0.996 ≥40 2.27 [1.01, 5.07] 0.045 Smoking index 0 Ref. Ref. >0 1.43 [1.15, 1.78] 0.001 1.42 [1.13, 1.78] 0.002 ADL Normal Ref. Ref. Abnormal 5.57 [4.23, 7.33] <0.001 3.51 [2.62, 4.70] <0.001 (b) Aggravation Risk factor Univariate Multiple Classification Variable Category Estimation [95% CI] P Value Estimation [95% CI] P Value SUD Subgroup I Ref. Ref. II 2.49 [1.74, 3.56] <0.001 2.22 [1.54, 3.20] <0.001 III 2.04 [1.60, 2.61] <0.001 1.17 [0.90, 1.53] 0.241 IV 4.62 [2.53, 8.44] <0.001 2.67 [1.42, 4.99] 0.002 Number of 0 Ref. Ref. SUDs 1 1.74 [1.36, 2.21] <0.001 1.31 [1.02, 1.69] 0.036 >1 2.14 [1.57, 2.93] <0.001 1.52 [1.10, 2.12] 0.012 A-DROP A: Age No Ref. Score Yes 1.19 [0.95, 1.50] 0.128 O: Orientation No Ref. Ref. Yes 1.73 [1.24, 2.41] 0.001 0.98 [0.69, 1.38] 0.898 Basic Sex Male Ref. Ref. characteristic Female 0.53 [0.42, 0.68] <0.001 0.62 [0.47, 0.80] <0.001 Age <20 0.05 [0.04, 0.06] <0.001 (years) 20 to 64 Ref. 65 to 74 2.46 [1.91, 3.17] <0.001 >74 1.05 [0.79, 1.39] 0.746 BMI <18.5 0.05 [0.04, 0.06] <0.001 18.5 to <25 Ref. 25 to <30 2.04 [1.59, 2.63] <0.001 30 to <35 2.18 [1.51, 3.13] <0.001 35 to <40 3.05 [1.66, 5.58] <0.001 ≥40 2.96 [1.32, 6.64] 0.008 Smoking index 0 Ref. Ref. >0 1.72 [1.38, 2.15] <0.001 1.42 [1.11, 1.80] 0.004 ADL Normal Ref. Ref. Abnormal 5.01 [3.84, 6.53] <0.001 4.74 [3.58, 6.27] <0.001 Abbreviations: ADL, Activities daily living; BMI, Body mass index; CI, Confidence interval; OR, Odds ratio; SUD, Specialized underlying disease. 4 Discussion 4.1 Study contribution Initially, there was the effective operation of the criterion of SG because CMR worsened SG upon hospitalization. In particular, CMR in severe patients with COVID-19 was remarkably higher (27.6%) than among all patients (6.9%). This was appropriate because this value was identical to those who were admitted to the ICU (25.7%) [29]. Furthermore, CMR in COVID-19 patients (6.9%) was remarkably lower than that of the avian influenza virus infection (approximately 60%) [30], the Middle East respiratory syndrome (upper 20%, retrospectively) [31], and severe acute respiratory syndrome (approximately 11%, retrospectively) [32]. In addition, we demonstrated that the aggravation after hospitalization was more important than the severe diagnosis at hospitalization with regarding Table 4. The aggravation and the severe diagnosis may be a same condition as wearing MV or admitting to ICU. However, its meaning is totally different because this condition cannot be expected for the aggravation despite being expectable for severe diagnosis at hospitalization. In other words, the appropriate time for the MV or ICU is different for aggravated patients (after hospitalization) and patients with the severe diagnosis (upon hospitalization). This difference indicates inefficient use of limited medical resources within MV and ICU, because the use of MV or ICU for aggravated patients is an unexpected situation in comparison to the initial treatment plan at hospitalization that neither MV nor ICU would be implemented on these patients. During the COVID-19 pandemic, their efficient use is a critical issue to improve the quality of treatment due to a huge demand of MV with inequity problem of its distribution [33]. Moreover, various diseases discussed a new type of MV with both a lower cost of introduction and user friendliness [[34], [35], [36], [37], [38]]. Therefore, we believe that the prevention of the aggravation is required to solve this issue because the prevention can decrease unexpected use of MV and ICU. For example, the inclusion of the SUD list in existing SG criteria (cf. Table 3) could detect those who have a higher possibility of the aggravation at hospitalization because we demonstrated that both CMR and AGR were higher in patients with SUD compared to those without SUD (see Table 2). Moreover, Category IV in the four subgroups in the SUD list showed a higher OR of aggravation with statistical significance (see Table 5). Since this category means patients whose age is ≥ 65 and BMI is ≥ 30 form the highest risk group, this result would be clinically appropriate and contribute to improving the existing evaluation criteria of SG. Finally, we discuss a risk prediction model of the aggravation in addition to mortality as a solution of the above issue using various risk factors in Table 5. Although there are some cases of the model for the mortality by COVID-19 using methods of machine learning [[39], [40], [41]], few studies have developed that of the aggravation. In fact, one study recently explored a possibility of a biomarker of the aggravation as a candidate for an explorative variable in the model [42]. However, we demonstrated that certain risk factors could be available to predict the aggravation as same as the mortality such as the four groups in SUD and ADL. Since these factors are recorded though creating the DPC database as a routine process of claiming medical fees, we believe that these factors could be easily implemented on the prediction model compared to the previous studies. Besides the four groups in SUD and ADL, we also explored the variables that could efficiently detect higher-risk patients using the data items in our database because of opaque risk factors in COVID-19 as an emerging infectious disease. For instance, we combined some risk factors into one explorative variable such, as A: age in the A-DROP score classification, using sex and age. Although explorative variables other than the two variables below did not show a remarkable value (OR was >2) in the mortality and the aggravation, they showed the value in only one of the mortality and the aggravation. Therefore, we believe that these other variables would have a possibility of improving the prediction model. 4.2 Limitations A major limitation was not being able to use a value from a laboratory inspection despite including SpO2 in the criterion of SG. In comparison to data items in the DPC database, this value was not recorded every day. This occurred as a consequence of the insufficient introduction of EMRs in Japan. In fact, studies have discussed that small- or moderate-scale hospitals cannot introduce EMRs like large scale hospitals [43,44]. However, this limitation did not have a critical negative influence on our data analyses because SpO2 was not necessary to extract severe patients as shown in Table 1. On the contrary, there were some minor limitations in this study. The first was that our secondary analysis did not consider the adjustment of confounding factors. However, this could be solved in future work because there are various data analyzing methods for the prediction model within this adjustment by a rapid progress of an information technology. The second minor limitation was that the influence of future variants of COVID-19 that was detected after was not discussed because our database only recorded information until September 30, 2021, and there were no data items that could identify an individual strain such as the Alpha strain. The third was that it was not possible to observe the influence of vaccination history because the database did not include it, though this history was strongly expected to decrease mortality, especially among the elderly. 4.3 Conclusion In conclusion, we demonstrated that the criteria of SG could effectively identify patients with a higher risk of mortality as those who were severe at hospitalization. Also, we defined patients with aggravation as those who were not severe at hospitalization but wore MV or were admitted to ICU after hospitalization. Moreover, we found that the occurrence of aggravation after hospitalization was a serious risk factor of mortality rather than a severe diagnosis at hospitalization. Finally, we explored the risk factors of mortality and aggravation. This exploration indicated the possibility of our future work building a high-performance prediction model. Funding This study was supported by a scholarship donation from the Kansai Economic Federation to the 10.13039/501100005683 Kyoto University . Kyoto University provided a part of this donation with the Japan Medical Network Association as a research cooperation institution. Author statements HF developed the study design, collected data, tackled the data analysis, and wrote the draft of the manuscript. KA was the chief investigator and supervised overall process of the study. All authors contributed to the writing of the final manuscript. Ethics approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the Kyoto University Graduate School and Faculty of Medicine, Ethics Committee (approval number, R2963) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent Informed consent was obtained by an opt-out method. The research executive office noted details of this study on their website. The office can consider that the informed consent is obtained unless participants request the office not to use their information. Declaration of competing interest HF has no competing interests. KA received a research fund as the director of the Japan Medical Network Association from the Kyoto University. Appendices Table A CMR and AGR by each risk factor Table ARisk factor All patients (n = 6,050) Without severe patients (n = 5,796) Classification Variable Category Participants (n, %) Mortality (n) CMR (%) Participants (n, %) Aggravation (n) AGR (%) SUD Subgroup I 2,878 47.6 70 2.4 2,792 48.2 108 3.9 II 547 9.0 32 5.9 520 9.0 47 9.0 III 2,526 41.8 299 11.8 2,394 41.3 181 7.6 IV 99 1.6 15 15.2 90 1.6 14 15.6 Number of 0 3,942 65.2 220 5.6 3,802 65.6 181 4.8 SUDs 1 1,492 24.7 131 8.8 1,415 24.4 113 8.0 >1 616 10.2 65 10.6 579 10.0 56 9.7 A-DROP A: Age No 4,123 68.1 166 4.0 3,961 68.3 227 5.7 Score Yes 1,927 31.9 250 13.0 1,835 31.7 123 6.7 O: Orientation No 5,488 90.7 314 5.7 5,342 92.2 307 5.7 Yes 562 9.3 102 18.1 454 7.8 43 9.5 Basic Sex Male 3,624 59.9 270 7.5 3,445 59.4 254 7.4 characteristic Female 2,426 40.1 146 6.0 2,351 40.6 96 4.1 Age <20 245 4.0 0 0.0 245 4.2 0 0.0 (years) 20 to 64 3,180 52.6 102 3.2 3,067 52.9 155 5.1 65 to 74 1,091 18.0 100 9.2 1,017 17.5 118 11.6 >74 1,534 25.4 214 14.0 1,467 25.3 77 5.2 BMI <18.5 733 12.1 52 7.1 707 12.2 29 4.1 18.5 to <25 3,185 52.6 203 6.4 3,069 53.0 137 4.5 25 to <30 1,486 24.6 114 7.7 1,410 24.3 123 8.7 30 to <35 473 7.8 31 6.6 447 7.7 41 9.2 35 to <40 112 1.9 8 7.1 105 1.8 13 12.4 ≥40 61 1.0 8 13.1 58 1.0 7 12.1 Smoking index 0 3,083 51.0 175 5.7 2,980 51.4 137 4.6 >0 2,967 49.0 241 8.1 2,816 48.6 213 7.6 ADL Normal 3,149 52.0 68 2.2 3,133 54.1 72 2.3 Abnormal 2,901 48.0 348 12.0 2,663 45.9 278 10.4 Abbreviations: ADL, Activities daily living; AGR, Aggravation rate; BMI, Body mass index; CMR, Crude mortality rate; SUD, Specialized underlying disease. Table B Summary statistics of each risk factor in a numerical value Table BVariable Mean SD Median Min Max Number of SUDs 0.5 0.7 0 0 5 Age (years) 58.8 20.8 60 0 104 BMI 23.73 5.44 23.4 0.0 60.5 Smoking index 1,982.4 3,750.8 0 0 9999 Abbreviations: BMI, Body mass index; SD, Standard deviation; SUD, Specialized underlying disease. Acknowledgements We would like to thank the JCHO Hokkaido Hospital, the Kitaimi Red Cross Hospital, the Tesshokai Kameda Medical Center, the Tosenkai Keiju Medcal Center, the University of Fukui Hospital, the Shizuoka General Hospital, the Japanese Red Cross Aichi Medical Center Nagoya Daiichi Hospital, the Nagahama Red Cross Hospital, the Kyoto University Hospital, the Keishinkai Kyoto Kitukawa Hospital, the Japan Baptist Medical Foundation Japan Baptist Hospital, the Hirakata Kohsai Hospital, the Osaka Red Cross Hospital, the Hospital of Hyogo College of Medicine, the Kobe City Medical Center General Hospital, the Hyogo Prefectural Amagasaki General Medical Center, the Wakayama Medical Center, the Shouwakai Brain Attack Center Ota Memorial Hospital, the Heiseishisenkai Kokura Memorial Hospital, the Yame General Hospital, the Saga Prefectural Medical Center Koseikan, the Kumamoto Rosai Hospital, the University of Miyazaki Hospital, the Miyazaki Prefectural Miyazaki Hospital, the Zenjinkai Miyazaki Zenjikai Hospital, the Miyazaki Prefectural Nichinan Hospital, the Zenjinkai Miyazaki Zenjikai Hospital, and the Life Data Initiative, Inst. for data collection. In addition, we would like to thank Enago (www.enago.jp) for the English language review. ==== Refs References 1 Marin B.G. Aghagoli G. Lavine K. Yang L. Siff E.J. Chiang S.S. Predictors of COVID-19 severity: a literature review Rev Med Virol 31 2020 e2146 10.1002/rmv.2146 2 Del Valle D.M.D. Kim-Schulze S. Huang H.H. Beckmann N.D. Nirenberg S. Wang B. An inflammatory cytokine signature predicts COVID-19 severity and survival Nat Med 26 2020 1636 1643 10.1038/s41591-020-1051-9 32839624 3 Han H. Ma Q. Li C. Liu R. Zhao L. Wang W. Profiling serum cytokines in COVID-19 patients reveals IL-6 and IL-10 are disease severity predictors Emerg Microb Infect 9 2020 1123 1130 10.1080/22221751.2020.1770129 4 Shibata S. Arima H. Asayama K. Hoshide S. Ichihara A. Ishimitsu T. Hypertension and related diseases in the era of COVID-19: a report from the Japanese society of hypertension task force on COVID-19 Hypertens Res 43 2020 1028 1046 10.1038/s41440-020-0515-0 32737423 5 Robilotti E.V. Babady N.E. Mead P.A. Rolling T. Perez-Johnston R. Bernardes M. Determinants of COVID-19 disease severity in patients with cancer Nat Med 26 2020 1218 1223 10.1038/s41591-020-0979-0 32581323 6 Tan L. Wang Q. Zhang D. Ding J. Huang Q. Tang Y.Q. Lymphopenia predicts disease severity of COVID-19: a descriptive and predictive study Signal Transduct Targeted Ther 5 2020 33 10.1038/s41392-020-0148-4 7 Cai Q. Chen F. Wang T. Luo F. Liu X. Wu Q. Obesity and COVID-19 severity in a designated hospital in Shenzhen, China Diabetes Care 43 2020 1392 1398 10.2337/dc20-0576 32409502 8 Deng M. Qi Y. Deng L. Wang H. Xu Y. Li Z. Obesity as a potential predictor of disease severity in young COVID-19 patients: a Retrospective Study Obesity (Silver Spring) 28 2020 1815 1825 10.1002/oby.22943 32602202 9 Gao F. Zheng K.I. Wang X.B. Sun Q.F. Pan K.H. Wang T.Y. Obesity is a risk factor for greater COVID-19 severity Diabetes Care 43 2020 e72 e74 10.2337/dc20-0682 32409499 10 Karanasos A. Aznaouridis K. Latsios G. Synetos A. Plitaria S. Tousoulis D. Impact of smoking status on disease severity and mortality of hospitalized patients with COVID-19 infection: a systematic review and meta-analysis Nicotine Tob Res 22 2020 1657 1659 10.1093/ntr/ntaa107 32564072 11 Wynants L. Calster B.V. Collins G.S. Riley R.D. Heinze G. Schuit E. Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisal BMJ 369 2020 m1328 10.1136/bmj.m1328 32265220 12 Kuroda S. Matsumoto S. Sano T. Kitai T. Yonetsu T. Kohsaka S. External validation of the 4C Mortality Score for patients with COVID-19 and pre-existing cardiovascular diseases/risk factors BMJ Open 11 2021 e052708 10.1136/bmjopen-2021-052708 13 Fan G. Tu C. Zhou F. Liu Z. Wang Y. Song B. Comparison of severity scores for COVID-19 patients with pneumonia: a retrospective study Eur Respir J 56 2020 2002113 10.1183/13993003.02113-2020 14 Miyashita N. Nakamori Y. Ogata M. Fukuda N. Yamura A. Ishiura Y. A warning related to predicting the severity of COVID-19 pneumonia using the A-DROP scoring system J Infect Chemother 28 2022 359 360 10.1016/j.jiac.2021.12.010 34930625 15 Huang I. Lim M.A. Pranata R. Diabetes mellitus is associated with increased mortality and severity of disease in COVID-19 pneumonia-A systematic review, meta-analysis, and meta-regression Diabetes Metabol Syndr: Clin Res Rev 14 2020 395 403 10.1016/j.dsx.2020.04.018 16 Pranata R. Lim M.A. Huang I. Raharjo S.B. Lukito A.A. Hypertension is associated with increased mortality and severity of disease in COVID-19 pneumonia: a systematic review, meta-analysis and meta-regression J Renin Angiotensin Aldosterone Syst April-June 2020 1 11 10.1177/1470320320926899 17 Kasraeian m Zare m Vafaei H. Asadi N. Faraji A. Bazrafshan K. COVID-19 pneumonia and pregnancy; a systematic review and meta-analysis J Matern Fetal Neonatal Med 35 2022 1652 1659 10.1080/14767058.2020.1763952 32429786 18 Zheng Y. Wang L. Ben S. Meta-analysis of chest CT features of patients with COVID-19 pneumonia J Med Virol 93 2021 241 249 10.1002/jmv.26218 32579236 19 Hariyanto T.I. Putri C. Arisa J. Situmeang R.F.V. Kurniawan A. Dementia and outcomes from coronavirus disease 2019 (COVID-19) pneumonia: a systematic review and meta-analysis Arch Gerontol Geriatr 93 2021 104299 10.1016/j.archger.2020.104299 20 World Health Organization Therapeutics and COVID-19 https://apps.who.int/iris/rest/bitstreams/1449398/retrieve 2022 21 Nicola M. O'Neill N. Sohrabi C. Khan M. Agha M. Agha R. Evidence based management guideline for the COVID-19 pandemic - review article Int J Surg 77 2020 206 216 10.1016/j.ijsu.2020.04.001 32289472 22 Gandhi R.T. Lynch J.B. del Rio C. Mild or moderate covid-19 N Engl J Med 383 2020 1757 1766 10.1056/NEJMcp2009249 32329974 23 Berlin D.A. Gulick R.M. Martinez F.J. Severe covid-19 N Engl J Med 383 2020 2451 2460 10.1056/NEJMcp2009575 32412710 24 Flannery D.D. Puopolo K.M. Perinatal COVID-19: guideline development, implementation, and challenges Curr Opin Pediatr 33 2021 188 194 10.1097/MOP.0000000000000997 33534424 25 Nielsen Jeschke K.N. Bonnesen B. Hansen E.F. Jensen J.-U.S. Lapperre T.S. Weinreich U.M. Guideline for the management of COVID-19 patients during hospital admission in a non-intensive care setting Eur Clin Respir J 7 2020 1761677 10.1080/20018525.2020.1761677 26 Agarwal A. Rochwerg B. Lamontagne F. Siemieniuk R.A.C. Agoritsas T. Askie L. A living WHO guideline on drugs for covid-19 BMJ 370 2020 m3379 10.1136/bmj.m3379 32887691 27 Ministry of Health, Labor and Welfare Treatment guideline of COVID-19 (in Japanese) https://www.mhlw.go.jp/content/000936655.pdf 2022 version 8.0 28 Ministry of Health, Labor and Welfare A request to provide information about vaccination for those who contract an underlying disease (in Japanese) https://www.mhlw.go.jp/content/000783864.pdf 2022 29 Quah P. Li A. Phua J. Mortality rates of patients with COVID-19 in the intensive care unit: a systematic review of the emerging literature Crit Care 24 2020 285 10.1186/s13054-020-03006-1 32498689 30 Poovorawan Y. Pyungporn S. Prachayangprecha S. Makkoch J. Global alert to avian influenza virus infection: from H5N1 to H7N9 Pathog Glob Health 107 2013 217 223 10.1179/2047773213Y.0000000103 23916331 31 Park J.E. Jung S. Kim A. Park J.E. MERS transmission and risk factors: a systematic review BMC Publ Health 18 2018 574 10.1186/s12889-018-5484-8 32 Chan K.S. Zheng J.P. Mok Y.W. Li Y.M. Liu Y.N. Chu C.M. SARS: prognosis, outcome and sequelae Respirology 8 Supplement 2003 10.1046/j.1440-1843.2003.00522.x S36–S40 33 Schmidt H. Roberts D.E. Eneanya N.D. Rationing, racism and justice: advancing the debate around ‘colourblind’ COVID-19 ventilator allocation J Med Ethics 48 2022 126 130 10.1136/medethics-2020-106856 33408091 34 Fang Z. Li A.I. Wang H. Zhang R. Mai X. Pan T. AmbuBox: a fast-deployable low-cost ventilator for COVID-19 emergent care SLAS Technol 25 2020 573 584 10.1177/2472630320953801 32882150 35 Vasan A. Weekes R. Connacher W. Sieker J. Stambaugh M. Suresh P. MADVent: a low-cost ventilator for patients with COVID-19 Med Devices Sens 3 2020 e10106 10.1002/mds3.10106 36 Raymond S.J. Baker S. Liu Y. Bustamante M.J. Ley B. Horzewski M.J. A low-cost, highly functional, emergency use ventilator for the COVID-19 crisis PLoS One 17 2022 e0266173 10.1371/journal.pone.0266173 37 Kwon A.H. Slocum A.H. Jr. Varelmann D. Nabzdyk C.G.S. MIT E-Vent Team Rapidly scalable mechanical ventilator for the COVID-19 pandemic Intensive Care Med 46 2020 1642 1644 10.1007/s00134-020-06113-3 32588066 38 El Majid B. El Hammoumi A. Motahhir S. Lebbadi A. El Ghzizal A. Preliminary design of an innovative, simple, and easy-to-build portable ventilator for COVID-19 patients Euro Mediterr J Environ Integr 5 2020 23 10.1007/s41207-020-00163-1 39 Yan L. Zhang H.-T. Goncalves J. Xiao Y. Wang M. Guo Y. An interpretable mortality prediction model for COVID-19 patients Nat Mach Intell 2 2020 283 288 10.1038/s42256-020-0180-7 40 Bertsimas D. Lukin G. Mingardi L. Nohadani O. Orfanoudaki A. Stellato B. COVID-19 mortality risk assessment: an international multi-center study PLoS One 15 2020 e0243262 10.1371/journal.pone.0243262 41 Kuno T. Sahashi Y. Kawahito S. Takahashi M. Iwagami M. Egorova N.N. Prediction of in‐hospital mortality with machine learning for COVID‐19 patients treated with steroid and remdesivir J Med Virol 94 2022 958 964 10.1002/jmv.27393 34647622 42 Iwamura C. Hirahara K. Kiuchi M. Ikehara S. Azuma K. Shimada T. Elevated Myl9 reflects the Myl9-containing microthrombi in SARS-CoV-2–induced lung exudative vasculitis and predicts COVID-19 severity Proc Natl Acad Sci U S A 119 2022 e2203437119 10.1073/pnas.2203437119 43 Yasunaga H. Imamura T. Yamaki S. Endo H. Computerizing medical records in Japan Int J Med Inf 77 2008 708 713 10.1016/j.ijmedinf.2008.03.005 44 Takeshita K. Takao H. Imoto S. Murayama Y. Improvement of the Japanese healthcare data system for the effective management of patients with COVID-19: a national survey Int J Med Inf 162 2022 104752 10.1016/j.ijmedinf.2022.104752

0

PMC9753483

NO-CC CODE

2022-12-16 23:25:44

no

J Infect Chemother. 2022 Dec 15; doi: 10.1016/j.jiac.2022.12.004

utf-8

J Infect Chemother

10.1016/j.jiac.2022.12.004

oa_other

==== Front IJID Reg IJID Reg IJID Regions 2772-7076 The Author(s). Published by Elsevier Ltd on behalf of International Society for Infectious Diseases. S2772-7076(22)00152-7 10.1016/j.ijregi.2022.12.003 Article Characteristics and clinical outcomes of COVID-19 in children: A hospital-based surveillance study in Latin America's hardest-hit city Jarovsky Daniel MD 12⁎ Fongaro Giuliana de Freitas MD 1 Zampol Renata Mazzotti MD 1 de Oliveira Thales Araújo MD 1 Farias Camila Giuliana Almeida MD 1 da Silva Daniella Gregória Bomfim Prado MD 1 Cavalcante Denis Tadeu Gomes MD 1 Nery Sabrina Bortolin MD 1 de Moraes José Cassio 2 Junior Francisco Ivanildo de Oliveira MD 1 Almeida Flávia Jacqueline MDPhD 12 Sáfadi Marco Aurélio Palazzi MDPhD 12 1 Hospital Infantil Sabará, São Paulo, Brazil 2 Santa Casa de São Paulo School of Medical Sciences ⁎ Correspondence: Daniel Jarovsky, Alameda Jau 585/121, Jardim Paulista, São Paulo, SP, 01420001, Brazil, Mobile: +5511-992779288 15 12 2022 15 12 2022 © 2022 The Author(s). Published by Elsevier Ltd on behalf of International Society for Infectious Diseases. 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Background : In 2020, Brazil became the coronavirus disease (COVID-19) pandemic's epicenter in Latin America, resulting in an unparalleled health catastrophe. Nevertheless, comprehensive clinical reports in Brazilian children are unavailable. Methods : This retrospective, hospital-based, active-surveillance study was performed to identify pediatric patients with COVID-19 presenting at a private academic medical center in a large urban area between March 2020 and March 2021. Clinical and demographic information was analyzed among those requiring hospitalization, those with severe illness, and those with clinical syndromes. Results: 964 cases of symptomatic conditions were evaluated : 17.7% required hospitalization, 27.5% of which were classified as severe/critical. Acute bronchiolitis and pneumonia were the most frequent causes of hospitalization among severe cases. Twenty-seven hospitalized children fulfilled the diagnostic criteria for multisystem inflammatory syndrome (median age: 29 months; 85.2% non-severe). Significant coexisting condition was present in 29% of hospitalized children. Risk of hospitalization was higher in children with ≥1 comorbidity, age <2 years, or obesity. Increased risk of severe disease was described among those with leucopenia, leukocytosis, or any significant comorbidity. No deaths occurred. Conclusion : Though most children with COVID-19 experienced mild disease and no deaths occurred, a significant proportion required hospitalization and developed severe illness. Obesity, young age, underlying comorbidity, leucopenia, and leukocytosis were risk factors for hospitalization or severe disease. Keywords SARS-CoV-2 COVID-19 MIS-C PIMS-TS hospital-based surveillance Pediatric Brazil ==== Body pmcINTRODUCTION The coronavirus disease (COVID-19) pandemic is the most severe global public health threat since the 1918 influenza A H1N1 pandemic. With a population of 210 million individuals, Brazil, an upper-middle-income country, was the first Latin American country with confirmed COVID-19 cases (late February 2020) [1]. As a large urban area of more than 12 million inhabitants, São Paulo has been the epicenter of Brazil's coronavirus pandemic. In the first half of 2020, Brazil became the pandemic's epicenter in Latin America, responsible for the second largest death toll in the world [2,3]. Furthermore, starting in late 2020, the country suffered from a second wave associated with the emergence of a new variant of concern, Gamma VOC, (also known as lineage P.1) [4]. Subsequently, Brazil has plunged into an unparalleled health catastrophe: despite Brazil containing less than 3% of the world's population, nearly one in every nine deaths worldwide occurred there [5]. Since the start of the pandemic, the country has confirmed more than 34.5 million cases, including 3.2 million hospitalizations and 686.000 deaths associated with COVID-19 [6]. In addition, by June 2022, children and adolescents under 19 years of age accounted for more than 41.000 hospitalizations and over 3000 deaths [7]. These numbers indicate a case fatality rate (CFR) as high as 7.2% among children and adolescents hospitalized with COVID-19 (data from SIVEP-Influenza Epidemiological Surveillance Information System, Brazilian Ministry of Health) and multisystem inflammatory syndrome in children (MIS-C) [3], which is up to four times that observed in the United States [8]. Extensive worldwide data suggest lower rates of severe or critical COVID-19, significantly lower risk of death from the disease, and higher rates of asymptomatic infection in children than in adults [9–13]. The underlying reasons for these considerable differences remain unclear, but age-related differences in expression of angiotensin-converting enzyme 2 receptors and differences in innate and adaptive immunity might play a role [14]. Despite the lower risk for severe outcomes in children, it is essential to acknowledge the consequences of the disease burden on children, including hospitalizations, deaths, long COVID, and MIS-C, particularly in low and middle-income countries (LMIC). Contrasting with the abundant data among adult and pediatric COVID-19 cases worldwide, comprehensive reports about Brazilian children are unavailable. Further, understanding the risk factors associated with hospitalization and severity of COVID-19 could guide clinicians to better diagnostic and management strategies. We present the clinical, laboratory, and radiographic characteristics of symptomatic pediatric patients with confirmed SARS-CoV-2 infection presenting to a large private pediatric referral hospital found in the heart of the Brazilian pandemic. MATERIALS AND METHODS Setting: This hospital-based descriptive surveillance analysis defines the epidemiology of pediatric COVID-19 in Hospital Infantil Sabará (Sabará Children's Hospital; HIS), a 145-bed private hospital providing tertiary care in metropolitan São Paulo, Brazil (Figure 1 ). With around 100 000 urgent consultations and over 6 000 admissions every year, HIS is the second largest private pediatric medical center in the country [15].Figure 1 Map of São Paulo Metropolitan area (population in 2018: 21.571.281) [51] with the approximate geographic location of residence of outpatient (blue) and hospitalized patients (red). Hospital Infantil Sabará's location is plotted in yellow. Figure 1 Study design: From March 28, 2020, through March 31, 2021, we included all patients under 18 years of age with SARS-CoV-2 detected in nasopharyngeal swab specimens using a reverse-transcription polymerase chain reaction (RT-PCR) and with symptoms consistent with COVID-19 (Figure 2 ). As per hospital protocol, a nasal swab was performed only on symptomatic children if clinical and epidemiological information raised suspicion of SARS-CoV-2 infection. RT-PCR testing began on February 15, 2020, with the first positive test on March 28 (Figure 3 ). Asymptomatic children were not tested, even if they presented as close contact to a confirmed COVID-19 source. To determine an association between recent SARS-CoV-2 infection and post-infection manifestations (herein MIS-C), an immunological testing was performed using an anti-SARS-CoV-2 ELISA assay for the detection of IgA/IgM, IgG and total antibodies [16], independently of the molecular evaluation.Figure 2 Flow chart of study population selection, including patient recruitment and exclusion criteria. Abbreviations: RT-PCR: reverse transcription-polymerase chain reaction; HIS: Hospital Infantil Sabará; MIS-C: Multisystem inflammatory syndrome in children. Figure 2 Figure 3 Number of positive RT-PCR tests among outpatients (blue), hospitalized patients (red), and multisystem inflammatory syndrome in children (MIS-C) cases (green) according to epidemiological week [52] among 964 children with COVID-19 at Hospital Infantil Sabará (HIS) (left axis). Textboxes present the relevant events occurring in SCH during the COVID-19 pandemic (black), main nonpharmaceutical interventions (yellow), school-related interventions (red), and vaccination against COVID-19 by age in Sao Paulo State (green). The gray columns in the background represent COVID-19 cases in all ages in the São Paulo Metropolitan area (right axis) [53] . By overlapping the pediatric cases from HIS with overall cases (i.e., pediatric and non-pediatric cases) in the same geographic area, we identified a disproportionate number of infections in adults during the first months of the pandemic, which flattened as non-pharmacological measures and restrictions took place. An apparent increase in infections occurred in mid-late October 2020, coinciding with the gradual reopening of elementary and high schools. This pattern persisted during the remaining weeks of the study, which included the initial period of Gamma variant circulation. Figure 3 SARS-CoV-2 infection evaluation: laboratory testing was requested upon individual medical assessments under hospital protocols. All laboratory evaluation was conducted at the Diagnósticos da America S.A. RT-PCR testing was performed according to Charité-Berlin protocol [17]. In addition, Xpert® Xpress SARS-CoV-2 (Cepheid, Sunnyvale, California, USA) was used in limited samples. Evaluation of other respiratory pathogens: the evaluation of additional respiratory viruses (through rapid RSV and influenza A/B immunochromatographic tests) and group A Streptococcus in the nasopharynx (through rapid immunoenzymatic test) is not standardized by the hospital and was requested exclusively upon individual medical assessment. Additionally, such tests are not covered by private medical care. Even more expensive multiplex RT-PCR is more frequently requested following hospitalization (please refer to Supplemental file for technical information regarding this RT-PCR assay). Radiologic evaluation: chest radiographs and CT scans were requested upon individual medical decisions. When performed, the exams were analyzed by unblinded radiologists that interpreted independently. Data sources and analysis: Clinical and demographic information was systematically extracted from electronic medical records for each case and included age, sex, underlying medical conditions that posed risk for severe outcomes, exposure to COVID-19 cases (confirmed, suspected, or unknown - when no symptomatic disease was identified among close contacts irrespectively of laboratory confirmation), symptoms experienced, radiologic evaluation, laboratory evaluation, hospitalization, need for intensive care, supportive treatments, pharmacologic therapy, and outcomes. Race and ethnicity were self-reported based on fixed categories. For children requiring serial laboratory evaluations the most altered value was considered. Disease severity among inpatients and outpatient was classified as follows [18]: mild (individuals who had any signs or symptoms of COVID-19 except shortness of breath, dyspnea, or abnormal chest imaging), moderate (individuals who showed evidence of lower respiratory disease during clinical assessment or imaging and had SpO2 ≥ 94% on room air), severe (individuals who had SpO2 < 94% on room air or lung infiltrates >50% at chest CT), and critical (individuals who had respiratory failure, septic shock, and/or multiple organ dysfunction). MIS-C was defined using the World Health Organization case definition for multisystem inflammatory disorders [19]. Statistical data analysis: Due to the large number of predictor variables to be analyzed for each outcome, a pre-selection was performed through bi-varied analysis. Chi-square test or Fisher's exact test were used to evaluate the association between categorical variables and the binary outcomes, as appropriate, while the T-Student or Mann-Whitney test were used to compare the categorical variables and binary outcomes. Significant predictors with p-value ≤ 0.2 in a bivariate analysis were exported to the multivariable logistic regression model, where three separate analyses were conducted for each outcome: (i) factors associated with hospitalization, (ii) factors associated with clinical severity among hospitalized children (i.e., non-severe versus severe cases), and (iii) factors associated with predefined clinical syndromes among hospitalized children (i.e., “Respiratory syndrome”, “MIS-C”, and “Other clinical syndromes”). Multivariate logistic regression analysis was carried out using stepwise forward technique to find the combined effects of the variables. Odds ratios (ORs) and 95% confidence intervals (CIs) were determined; the significance level was set at 0.05 (two-tailed test) for all tests. Cases with unknown/blank information were excluded from the analysis. Categorical variables were reported as absolute numbers and percentages, while continuous variables were expressed as median and interquartile ranges (IQRs), if abnormally distributed. Statistical analyses were performed using IBM SPSS Statistics version 13.0. Ethics: Study approval was obtained from the Research Ethics Committee of the José Luiz Egydio Setúbal Foundation review board (approval number 42080620.5.0000.5567). Given the study's purely descriptive and retrospective nature, written informed consent was waived. Data were collected anonymously, analyzed, and reported only in aggregate form. RESULTS Study population: From the first detected COVID-19 case in HIS, a total of 964 symptomatic SARS-CoV-2-related conditions from 962 unique patients were evaluated. 17.7% (n=171) were hospitalized for at least 1 day (Figure 2) and 95.3% (n=163) of those stayed for ≥48h. Geographical case distribution is depicted in Figure 1. Overall, the median age was 44.7 months (IQ25-75=16.4-104.7), and 55.4% (n=534) were male. White ethnicity was the most prevalent (74.5%; n=718), followed by black/African descent (11.9%; n=115), and Asian (3.6%; n=35). Case distribution was inversely related with age: children under 2 years comprised more than one third of all infections (34.2%; n=330), followed by children aged 2–5 years (24.8%; n=239), 5–10 years (21.6%; n=208), and over 10 years (19.4%; n=187). Known exposure to SARS-CoV-2 was documented in 80.7% (n=778) of the cases and a family contact was the source in 75.7% (n=730). Marked differences in most of these epidemiological variables were seen when comparing inpatients and outpatients (Table 1 ).Table 1 Clinical, demographic, and laboratory characteristics of pediatric patients with COVID-19 evaluated at Hospital Infantil Sabará between March 2020 and March 2021. Reference values and ranges for parameters are provided by the local laboratory. Table 1 All patients Clinical syndrome among hospitalized patients Characteristics Hospitalized Outpatient p value MIS-C Respiratory Other syndromes p value Number of patients (%) 171 17.7% 793 82.3% 27 15.8% 86 50.3% 58 33.9% Age, median (IQR), months 21.6 (7.2-53.8) 51.5 (19.3-108.3) < 0.001 29 (19-69) 16 (4-47) 21.6 (3-54) < 0.001 Age distribution — no. (%) < 0.001 0.444 <2 yr 91 53.2% 239 30.1% 10 37.0% 50 58.1% 31 53.4% 2-5 yr 40 23.4% 199 25.1% 8 29.6% 18 20.9% 14 24.1% 5-10 yr 16 9.4% 192 24.2% 5 18.5% 7 8.1% 4 6.9% >10 yr 24 14.0% 163 20.6% 4 14.8% 11 12.8% 9 15.5% Sex — no. (%) 0.217 0.318 Female 69 40.4% 361 45.5% 10 37.0% 38 44.2% 21 36.2% Male 102 59.6% 432 54.5% 17 63.0% 48 55.8% 37 63.8% Race/ethnicity, No. (%) 0.693 0.421 Black 1 0.6% 5 0.6% 0 0.0% 1 1.2% 0 0.0% Afrodescendant 13 7.6% 96 12.1% 1 3.7% 8 9.3% 4 6.9% White 137 80.1% 581 73.3% 25 92.6% 66 76.7% 46 79.3% Asian 6 3.5% 29 3.7% 0 0.0% 6 7.0% 0 0.0% NA 14 8.2% 82 10.3% 1 3.7% 5 5.8% 8 13.8% Significant underlying medical conditions — no. (%) 50 29% 153 19% 0.001 < 0.001 Asthma or recurrent wheezing 25 14.6% 82 10.3% 0 0.0% 22 25.6% 3 5.2% < 0.001 Prematurity 10 5.8% 7 0.9% 0 0.0% 6 7.0% 4 6.9% 0.392 Chronic neurologic disease 11 6.4% 28 3.5% 0 0.0% 7 8.1% 4 6.9% 0.099 Genetic/chromossomal disease 6 3.5% 7 0.9% 1 3.7% 4 4.7% 1 1.7% 0.423 Onocologic disease 3 1.8% 0 0.0% 0 0.0% 1 1.2% 2 3.4% 0.49 Congenital cardiopathy 5 2.9% 12 1.5% 1 3.7% 3 3.5% 1 1.7% 0.524 Chronic pulmonary disease (not asthma) 5 2.9% 4 0.5% 0 0.0% 5 5.8% 0 0.0% 0.14 Obesity 3 1.8% 1 0.1% 0 0.0% 3 3.5% 0 0.0% 0.382 Severity of Illness and outcomes Severe and critical 47 27.5% - - - 4 14.8% 40 46.5% 3 5.2% < 0.001 Required intensive care 57 33.3% - - - 14 51.9% 39 45.3% 4 6.9% < 0.001 Length of stay, PICU, median (IQR), days 4.5 (2.3-8) - - - 6 (3.3-8.0) 4 (2-8) 4 (3.5-6) - Length of stay, hospital, median (IQR), days 4 (3-6) - - - 6 (4.5-8.5) 4 (3-7) 3 (2-4) - Symptoms presented during disease course — no. (%) Median duration of symptoms/signs at ER evaluation (IQR), days 3.0 (2-5) 3.0 (2-4) 0.004 7.0 (4-9) 3.0 (2-5) 2.0 (1-4) 0.004 Fever 131 76.6% 525 66.2% 0.008 27 100.0% 61 70.9% 43 74.1% < 0.001 Cough 72 42.1% 416 52.5% 0.014 3 11.1% 62 72.1% 7 12.1% < 0.001 Rhinorrhea/Nasal congestion 65 38.0% 469 59.1% < 0.001 5 18.5% 54 62.8% 6 10.3% 0.127 Sneezing 13 7.6% 113 14.2% 0.019 0 0.0% 11 12.8% 2 3.4% < 0.001 Tachypnea on admission 41 24.0% 21 2.6% < 0.001 1 3.7% 38 44.2% 2 3.4% 0.083 Tachycardia on admission 18 10.5% 7 0.9% < 0.001 5 18.5% 11 12.8% 2 3.4% < 0.001 Shortness of breath 48 28.1% 26 3.3% < 0.001 0 0.0% 47 54.7% 1 1.7% < 0.001 Hypoxia (O2 <92% as measured by pulse oximetry) 31 18.1% 4 0.5% < 0.001 1 3.7% 28 32.6% 2 3.4% 0.329 Cyanosis 9 5.3% 0 0.0% < 0.001 0 0.0% 6 7.0% 3 5.2% 0.023 Reduced feeding or difficulty feeding 68 39.8% 169 21.3% 0.01 16 59.3% 28 32.6% 24 41.4% 0.023 Abdominal pain 28 16.4% 72 9.1% 0.197 13 48.1% 2 2.3% 13 22.4% < 0.001 Diarrhea 43 25.1% 164 20.7% < 0.001 9 33.3% 10 11.6% 24 41.4% 0.020 Nausea or vomiting 53 31.0% 119 15.0% < 0.001 13 48.1% 15 17.4% 25 43.1% 0.002 Dehydration 25 14.6% 14 1.8% < 0.001 7 25.9% 5 5.8% 13 22.4% 0.004 Fatigue/mialgia 21 12.3% 114 14.4% 0.351 10 37.0% 7 8.1% 4 6.9% < 0.001 Drowsiness/Irritability 44 25.7% 51 6.4% < 0.001 10 37.0% 16 18.6% 18 31.0% 0.018 Headache 16 9.4% 181 22.8% < 0.001 4 14.8% 6 7.0% 6 10.3% 0.322 Sore throat 16 9.4% 144 18.2% 0.002 6 22.2% 6 7.0% 4 6.9% 0.018 Anosmia 4 2.3% 39 4.9% 0.072 0 0.0% 4 4.7% 0 0.0% 0.401 Ageusia 4 2.3% 39 4.9% 0.072 0 0.0% 4 4.7% 0 0.0% 0.401 Neurologic symptoms 9 5.3% 6 0.8% < 0.001 1 3.7% 0 0.0% 8 13.8% < 0.001 Meningeal signs 1 0.6% 0 0.0% 0.178 0 0.0% 0 0.0% 1 1.7% 0.122 Rash 24 14.0% 33 4.2% < 0.001 15 55.6% 4 4.7% 5 8.6% < 0.001 Cervical adenopathy 5 2.9% 1 0.1% 0.001 5 18.5% 0 0.0% 0 0.0% < 0.001 Oral abnormalities 17 9.9% 48 6.1% 0.069 12 44.4% 2 2.3% 3 5.2% < 0.001 Non-supurative conjunctivitis 10 5.8% 5 0.6% < 0.001 10 37.0% 0 0.0% 0 0.0% < 0.001 Extremities abnormalities 11 6.4% 0 0.0% < 0.001 10 37.0% 0 0.0% 1 1.7% < 0.001 General laboratory evaluation Age-matched anemia 50/158 31.6% 6/85 7.1% < 0.001 19/25 76.0% 17/81 21.0% 14/52 26.9% < 0.001 WBC per µL, median 8330 (5925-12950) 7815 (5925-10175) 0.172 13000 (8040-19200) 7600 (5700-12100) 7920 (5140-12200) - Leucopenia (<4000 cells/μL) 14/156 9.0% 3/88 3.4% 0.101 2/25 8.0% 8/78 10.3% 2/26 7.7% 0.898 Left shift 17/158 10.8% 2/85 2.4% 0.017 6/25 24.0% 6/81 7.4% 5/52 9.6% 0.063 Absolute neutrophil count (cells/μL), median 3525 (1890-6440) 2998 (1815-5965) 6198 (3450-8532) 3300 (1890-5800) 3745 (1880-6485) - Absolute lymphocyte count (cells/μL), median 2595 (1630-4220) 3170 (2045-4525) 0.246 2553 (1311-3387) 2500 (1490-4300) 2930 (1800-4275) - Neutrophil to lymphocyte ratio, median (nv not established) 1.4 (0.6-3.2) 1.2 (0.4-2.1) - 1.8 (1.1-4.6) 1.5 (0.6-3.2) 1.3 (0.5-2.7) - Age-matched lymphopenia (cells/µL) 57/153 37.3% 15/90 16.7% 0.001 7/21 33.3% 33/80 41.3% 17/52 32.7% 0.227 Atypical lymphocytes 92/158 58.2% 43/84 51.2% 0.178 16/25 64.0% 43/81 53.1% 33/52 63.5% 0.597 Platelets/µL, median 284000 (210000-383000) ###### (217000-3432500) 0.184 ###### (82000-434000) ###### (249000-371000) ###### (211000-350000) Increased platelets count (>450.000/µL) 20/157 12.7% 5/84 6.0% 0.050 6/25 24.0% 10/81 12.3% 4/51 7.8% 0.405 Reduced platelets count (<150.000/µL) 136/157 86.6% 79/84 94.0% 0.937 19/25 76.0% 70/81 86.4% 47/51 92.2% < 0.001 AST (U/L), median 39 (29-65) 44 (27-43) 0.424 48 (32-81) 35 (31-64) 28 (24-41) 0.153 ALT (U/L), median 24 (17-44) 22 (17-32) 0.999 30 (21-57) 24 (17-44) 20 (17-24) 0.323 Laboratory evidence of cardiac involvement Increased CPK (nv 30 a 135 U/L) 1/19 5.3% - - - 1/12 8.3% 0/5 0.0% 0/2 0.0% 0.767 Increased troponin I (nv ≤53 ng/L) 8/22 36.4% - - - 6/15 40.0% 2/6 33.3% 0/1 0.0% 0.119 Increased proBNP (nv < 100 pg/mL) 8/21 38.1% - - - 7/15 46.7% 1/4 25.0% 0/2 0.0% 0.517 Laboratory evidence of systemic inflamation Median C-reactive protein level, mg/dL 1.5 (0.3-6.8) 0.6 (0.1-2.4) - 12.6 (5.9-20.8) 0.8 (0.2-2.6) 1.2 (0.2-4.7) - Elevated C-reactive protein (>3.0 mg/dL) 56/157 35.7% 11/63 17.5% 0.002 21/26 80.8% 17/80 21.3% 18/51 35.3% < 0.001 Elevated C-reactive protein (>10.0 mg/dL) 26/157 16.6% 3/63 4.8% 0.028 15/26 57.7% 4/80 5.0% 7/51 13.7% < 0.001 ESR (mm/h), median 53.0 (25-81) 11.0 (9-13) 0.004 72 (63-105) 23 (15-36) 30 (23-42) - Elevated ESR (≥40 mm/h) 27 58.1% 0 0.0% 0.041 21/21 100.0% 3/15 20.0% 3/7 42.9% < 0.001 Leucocytosis (≥15.000 mm³) 26/158 16.5% 5/85 5.9% 0.01 9/25 36.0% 12/81 14.8% 5/52 9.6% 0.003 Elevated DHL (>237 UI/L) 35/41 85.4% - - - 18/20 90.0% 12/13 92.3% 5/8 62.5% 0.002 Elevated D-dimer (> 0,50 µg/mL) 122/122 100.0% - - - 5/5 100.0% 65/65 100.0% 52/52 100.0% 0.006 Altered fibrinogen (nv 200-400 mg/dL) 24/38 63.2% - - - 19/21 90.5% 3/12 25.0% 2/5 40.0% 0.001 Altered coagulopathy 7/35 20.0% - - - 5/15 33.3% 1/14 7.1% 1/6 16.7% 0.081 Albumin level ≤3.5 g/dL 108/108 100.0% - - - 5/5 100.0% 57/57 100.0% 46/46 100.0% 0.215 Increased IL-6 (nv <7 pg/mL) 15/20 75.0% - - - 14/14 100.0% 1/5 20.0% 0/1 0.0% 0.001 Exposure to SARS-CoV-2 — no. (%) < 0.001 0.689 Family cluster 109 63.7% 621 78.3% 17 63.0% 58 67.4% 34 58.6% Contact with other suspected case 6 3.5% 42 5.3% 1 3.7% 4 4.7% 1 1.7% Unidentified source of infection 56 32.7% 130 16.4% 9 33.3% 24 27.9% 23 39.7% Laboratory evidence of COVID-19, contact with SARS-CoV-2 and detection of other respiratory pathogens Positive test for any other respiratory pathogen 22/82 26.8% 2/81 2.5% < 0.001 2/12 16.7% 20/51 39.2% 0/19 0.0% - Positive multiplex RT-PCR respiratory panel 15/35 42.9% 0/0 #DIV/0! - 2/4 50.0% 13/27 48.1% 0/4 0.0% - RSV (rapid test) 10/22 35.7% 1/10 10.0% - 0/1 0.0% 10/24 41.7% 0/3 0.0% - RSV (any test) 15/22 68.2% 1/2 50.0% - 0/2 0.0% 15/20 75.0% 0/0 #DIV/0! - Influenza A/B (rapid test) 0/23 0.0% 0/17 0.0% - 0/2 0.0% 0/17 0.0% 0/4 0.0% - Group A Streptococcus (rapid test) 0/28 0.0% 1/66 1.5% - 0/7 0.0% 0/9 0.0% 0/12 0.0% - Radiologic evaluation — no. (%) No radiologic evaluation 40 23.4% 575 72.5% - 6 22.2% 0 0.0% 34 58.6% - Chest CT performed 18 10.5% 7 0.9% - 2 7.4% 14 16.3% 2 3.4% - Chest CT abnormalities 14/18 77.8% 1/7 14.3% - 0/2 0.0% 12/14 85.7% 2/2 100.0% - Ground-glass opacity 11 61.1% 1 14.3% - 0 0.0% 9 64.3% 2 100.0% - Consolidations 9 50.0% 0 0.0% - 0 0.0% 8 57.1% 1 50.0% - Other 5 27.8% 0 0.0% - 0 0.0% 4 28.6% 1 50.0% - Chest CT abnormalities with normal X-ray 4/14 0/1 - 0/0 3/12 1/2 - X-ray performed 130 76.0% 213 26.9% - 21 77.8% 85 98.8% 24 41.4% - X-ray performed, no chest CT 113 86.9% 211 99.1% - 19 90.5% 72 84.7% 22 91.7% - Any X-ray abnormalities 78/130 60.0% 75/213 35.2% < 0.001 11/21 52.4% 54/85 63.5% 13/24 54.2% 0.271 Perihilar peribroncovascular thickening 65/130 50.0% 65/213 30.5% - 9/21 42.9% 44/85 51.8% 12/24 50.0% - Pulmonary opacities 24/130 18.5% 11/213 5.2% - 3/21 14.3% 19/85 22.4% 2/24 8.3% - Atelectasis 8/130 6.2% 4/213 1.9% - 1/21 4.8% 5/85 5.9% 2/24 8.3% - Maximum respiratory and vasoactive support None 124 72.5% - - - 23 85.2% 46 53.5% 55 94.8% < 0.001 Supplemental oxygen 25 14.6% - - - 1 3.7% 23 26.7% 1 1.7% 0.002 High-flow nasal cannula (HFNC) 13 7.6% - - - 0 0.0% 12 14.0% 1 1.7% 0.016 CPAP or BiPAP 4 2.3% - - - 0 0.0% 4 4.7% 0 0.0% 0.275 Intubation/tracheostomy ventilation 5 2.9% - - - 3 11.1% 1 1.2% 1 1.7% 0.016 Vasoactive support 4 2.3% - - - 3 11.1% 0 0.0% 1 1.7% 0.002 Advanced therapy (iNO, ECMO, prone ventilation) 2 1.2% - - - 1 3.7% 0 0.0% 1 1.7% 0.122 Pharmacologic therapy None 74 43.3% 723 91.2% < 0.001 4 14.8% 28 32.6% 42 72.4% < 0.001 Macrolide 29 17.0% 19 2.4% < 0.001 2 7.4% 26 30.2% 1 1.7% < 0.001 Tocilizumab 3 1.8% 0 0.0% - 3 11.1% 0 0.0% 0 0.0% 0.002 Intravenous immune globulin 19 11.1% 0 0.0% - 16 59.3% 1 1.2% 2 3.4% < 0.001 Oseltamivir 19 11.1% 1 0.1% < 0.001 2 7.4% 17 19.8% 0 0.0% 0.083 Systemic glucocorticoids 54 31.6% 24 3.0% < 0.001 14 51.9% 36 41.9% 4 6.9% < 0.001 Antibacterial therapy (other than macrolide) 65 38.0% 30 3.8% < 0.001 17 63.0% 34 39.5% 14 24.1% 0.017 Abbreviations: IQR: interquartile range; MIS-C: Multisystem inflammatory syndrome in children; PICU: pediatric intensive care unit; ER: emergency room; WBC: white blood cells; AST: aspartate transaminase; ALT: alanine aminotransferase; CPK: creatine phosphokinase; D-dimer: fibrin degradation product D; ESR: erythrocyte sedimentation rate; RT-PCR: reverse transcription polymerase chain reaction; RSV: respiratory syncytial virus; ECMO: extracorporeal membrane oxygenation; INO: inhalational nitric oxide. Clinical characteristics and outcomes: A coexisting condition was present in 21.1% (n=203) of all children. Asthma/recurrent wheezing, chronic neurologic disease, and prematurity were disproportionately represented among both outpatient and hospitalized children (Table 1). Twenty-five of 57 patients (43.8%) admitted to PICU reported at least one significant underlying disease. The additional relevant clinical conditions are described in Table 2. Among outpatients, upper respiratory tract infection (URTI) (n=364/793) and flu-like syndrome (n=216/793) were the leading presentations, comprising 73.1% of all discharged children altogether. Acute gastroenteritis (12%; n=95/793) and fever without source (5.5%; n=44/793) were less frequent presentation syndromes.Table 2 Significant underlying chronic conditions according to disease severity among pediatric patients with COVID-19 evaluated at Hospital Infantil Sabará between March 2020 and March 2021. Table 2Underlying chronic conditions Hospitalized Discharged Severe/critical (n=47) Non-severe (n=124) Outpatient (n=793) Asthma/recurrent wheezing 14 (29.8%) 11 (8.9%) 164 (20.7%) Chronic pulmonary disease 4 (8.5%) 1 (0.8%) 8 (1.0%) Genetic syndrome 5 (10.6%) 1 (0.8%) 14 (1.8%) Congental cardiopathy 3 (6.4%) 2 (1.6%) 24 (3.0%) Chronic neurologic disease 6 (12.8%) 5 (4.0%) 56 (7.1%) Oncologic disease 0 - 3 (2.4%) 0 - Obesity 2 (4.3%) 1 (0.8%) 2 (0.3%) Diabetes 0 - 0 - 4 (0.5%) Chronic kidney disease 2 (4.3%) 0 - 6 (0.8%) Chronic rheumatologic disease 0 - 0 - 2 (0.3%) Prematurity 4 (8.5%) 6 (4.8%) 14 (1.8%) Children requiring hospitalization were divided into two groups, according to the clinical course: non-severe hospitalized cases (72.5%; n=124/171) and severe/critical cases (27.5%; n=47/171) requiring PICU care (Table 1). The clinical syndromes in this subgroup were more evenly distributed among nine conditions (Figure 4 ). A significant number of severe and critical patients experienced respiratory disease: acute bronchiolitis (38.3%; n=18/47) and pneumonia (23.4%; n=11/47) were the most frequent causes of hospitalization, followed by flu-like syndrome (12.7%; n=6/47), asthma exacerbation (10.6%; n=5/47), and MIS-C (8.5%; n=4/47). No deaths related to COVID-19 or MIS-C occurred in HIS prior to this manuscript's acceptance date.Figure 4 Distribution of clinical syndromes according to disease severity among 964 children with COVID-19 at Hospital Infantil Sabará. Outpatients were considered mild. Abbreviations: URT: upper respiratory tract; MIS-C: Multisystem inflammatory syndrome in children. Figure 4 COVID-19 in very young infants: Thirty-four symptomatic infections occurred in children under 2 months of age, and 26 were hospitalized (15.2% of all inpatients) - four were late preterm babies and all had significant comorbidities. The most frequent clinical syndromes causing hospitalization were acute bronchiolitis (n=6), URT infection (n=5), and fever without source (n=8). Severe cases comprised 23.5% (n=8), and respiratory syncytial virus (RSV) was co-detected in two of these babies. URT infection (n=3) and exanthematous disease (n=2) were the main presentation forms among outpatients. Viral co-detections: Testing for other respiratory pathogens was performed in 17.0% of all children (n=164/964), including 10.3% of outpatients (n=82/793) and 48% of inpatients (n=82/171). Crude positivity rate was 16.5% (n=27/164) but differed significantly among those hospitalized (26.8%, n=22/82 vs. 6.1%, 5/82; p=0.001). RSV was detected exclusively in inpatients, with 75% (n=15/20) of those hospitalized with respiratory symptoms. RSV was also the leading co-pathogen in acute bronchiolitis, with 15 detections out of 28 tests performed irrespective of disease severity, and in 10 of 16 severe cases tested for additional pathogens. Among the 29 children with severe respiratory disease, three viruses were identified: 11 RSV (one co-detection, rhinovirus) and five rhinoviruses (two co-detection: RSV and adenovirus). Seven positive tests in patients hospitalized with non-severe infection detected RSV (n=4) and three rhinoviruses (one co-detection, adenovirus). Clinical markers for hospitalization: Factors that increased the risk of hospitalization are described in Table 1 and included chronic pulmonary disease, but not asthma or recurrent wheezing (p=0.002), genetic or chromosomal disease (p=0.003), oncologic disease (p=0.006), and obesity (p=0.019). Symptomatic infections occurring after the identification of the gamma variant did not result in a higher risk of hospitalization (p=0.434). The adjusted odds of hospitalization were 2.0 (95% CI 1.23–3.47) and 22.9 (95% CI 1.67–313.8) among children with at least one significant comorbidity and obesity, respectively, than those with no underlying medical conditions. Additional conditions associated with increased risk of hospitalization included age under 2 years (OR: 6.6; 95% CI 4.0–11.0), nausea or vomiting (OR: 4.5; 95% CI 2.2–9.2), abdominal pain (OR: 4.07; 95% CI 2.1–10.1), and rash (OR: 3.5; 95% CI 1.7–7.3). Risk markers for severe clinical illness: The clinical factors associated with increased risk of severe disease included asthma/recurrent wheezing (p<0.001), presence of any significant comorbidity (p<0.001), prematurity (p=0.013), any gastrointestinal symptoms (p=0.009), and age under 2 years (p=0.033). The only laboratory markers differentiating severe cases from non-severe ones were higher erythrocyte sedimentation rate (median 62 mm/h vs. 21 mm/h, respectively; p=0.007) and increased white blood cells (median 10500/mm3 vs. 7640/mm3, respectively; p=0.036). The adjusted odds of severe disease were 3.0 (95% CI 1.09–8.6) among cases with at least one significant comorbidity versus those with none. Additional laboratory markers associated with increased severity included leucopenia (<4000 cells/μL) (OR: 7.3; 95% CI 1.4–37.5) and leukocytosis (OR: 6.0; 95% CI 1.7–21.2). Radiologic findings: A radiologic evaluation was performed in 36.2% of all children (n=349) (Table 1). Any pulmonary abnormality was present in 35.2% of discharged patients, in contrast with 60% among the inpatients (p<0.001). The most frequent findings in both subgroups were perihilar peribroncovascular thickening, pulmonary opacities, and atelectasis. A chest CT scan was performed in 25 children (2.6%): 60% (n=15) had a significant underlying condition, and 18 were hospitalized. Ground-glass opacity with at most moderate involvement (25 to 50%) of the lung parenchyma was the most frequent finding among those with an abnormal report (n=12; 48,0%), followed by consolidations (n=9; 36%). However, these abnormalities differed significantly between inpatients (n=11; 61.1% and n=9; 50% respectively) and outpatients (n=1; 14,3% and n=0, respectively). Radiologic signs of complicated pneumonia (i.e., pleural effusion or empyema, necrotizing pneumonia, or intrapulmonary abscesses) were not described in any patient. Administered therapy: Most non-hospitalized patients were discharged from the ER with a prescription for symptomatic treatment only (Table 1 and Figure 5 ). Forty-nine outpatients (6.2%) had an oral β-lactam or macrolide antibiotic prescribed. Antibiotic justifications were acute otitis media (n=10), presumed urinary tract infection (n=3), “abnormalities” seen on chest X-ray (n=2), and treatment of group A streptococcal pharyngotonsillitis (n=1). Only 7.7% of all children presenting with flu-like syndrome or pneumonia (n=20/258) and 32.5% of severe respiratory illness (n=13/40) had oseltamivir prescribed. Overall, 8.1% of patients (n=78/964) had been prescribed a systemic glucocorticoid: 3% of outpatients (n=24/793) and 31.6% of hospitalized (n=54/171), of whom 29 had severe clinical presentation. No patient received specific SARS-CoV-2 antiviral therapy (such as remdesivir) or chloroquine/hydroxychloroquine.Figure 5 Distribution of clinical syndromes and pharmacology therapies according to disease severity among 964 children with COVID-19 at Hospital Infantil Sabará. Outpatients were considered mild. Figure 5 MIS-C: 27 hospitalized children fulfilled the diagnostic criteria for MIS-C. The median duration of fever was 7 days (range: 3–15), and the most frequent clinical findings ae described in Table 1. RT-PCR confirmed COVID-19 and serologic evidence of SARS-CoV-2 infection was seen in 29.6% (n=8) and 59.3% (n=16) of patients tested, respectively. Only three cases (11.1%) showed positive epidemiologic link and both tests negative. Three of the four severe/critical cases were characterized by intense clinical courses including mechanical ventilation in a prone position, vasoactive support, dialysis (n=1), and therapy with tocilizumab, a monoclonal antibody against the interleukin-6 receptor. Six children (22.2%) were treated with high-dose IVIG only, four (14.8%) used systemic glucocorticoids only, ten (37%) used both medications in combination, and seven children (26%) did not receive any treatment. All MIS-C patients had echocardiographic evaluations. Abnormal findings were seen in two-thirds of cases (n=18), while two or more abnormalities occurred in 22.2% (n=6). As expected, the presence of any abnormal finding was significantly more common in children with MIS-C than in those with acute COVID-19 (n=29; p=0.001): pericardial effusion (PE) was the most frequent finding (37.0%; n=10 vs. 6.9%; n=2), followed by coronary artery dilation (22.2%; n=6 vs. 3.4%; n=1), mitral valve regurgitation (14.8%; n=4 vs. 3.4%; n=1), and left ventricular dilation (7.4%; n=2 vs. n=0). Coronary artery aneurysms were not detected. DISCUSSION Globally the pediatric population represents around 15% of all COVID-19 cases [20], and less than 1% of documented critical and fatal cases [11,21]. However, LMIC, including several Latin American countries, experience more significant pediatric disease fatality than high-income nations [22–24]. Brazil, for example, is one of the world's leading countries in pediatric COVID-19 deaths, accounting for almost one of every four deaths worldwide before Omicron [25]. Despite these concerning numbers, during more than 2 years of the COVID-19 pandemic, limited comprehensive pediatric data have been published on SARS-CoV-2 infection in Brazilian children. To the best of our knowledge, this is the largest study to evaluate the clinical, laboratory, and radiographic characteristics of pediatric COVID-19 in a Brazilian population. As the disease epicenter in Brazil and the largest pediatric hospital in the country's southeast region, the high prevalence of SARS-CoV-2 infection in Metropolitan Sao Paulo offered a unique opportunity to describe the varied pediatric forms of the disease. Two remarkable findings in our study stand out. First, before Delta and Omicron predominance, pediatric COVID-19 resulting in hospitalization ranged from 0.1% to 1.5% in the USA and some European Union countries [20,26]; however, our rates exceeded such countries’ by 10–15 times. Even higher rates were reported by a multicenter pediatric research network, comprising five Latin American countries, where an overwhelming 47% were admitted to hospital [23]. Our data support the current evidence of a more severe disease in Latin/Hispanic children, one of the largest and most populated areas in the world. Similarly, PICU admission occurred in fewer than 13% of hospitalized pediatric patients in Europe and North America [26–29], as opposed to our rate of 27.5%. However, we believe the clinical severity is a more correct measurement of intensive care need, as almost one-fifth of the PICU admissions at our hospital were required due to IVIG administration. This study strengthens the concept that the risk of hospitalization and severe COVID-19 outcome is substantially elevated for children with underlying risk factors [27,30]. Asthma, neurodevelopmental disorders, congenital cardiac anomalies, prematurity (particularly in young children), leukopenia, and lymphocytopenia are among the strongest risk factors for hospitalization and severe disease [31]. This is the first study from Latin America to report pediatric obesity as a risk factor for COVID19 hospitalization. Persistent immune dysregulation (including dysfunctional innate T cells), chronic inflammation, and endothelial dysfunction are among the postulated underlying mechanisms [32–34]. Our research identified thrombocytosis as an additional factor infrequently described in children. While thrombocytopenia is a well-recognized negative prognostic factor in COVID-19 and MIS-C, platelets ≥450  ×  109/L has been reported in less than 5% [35,36] and is usually associated with reduced overall in-hospital mortality and mechanical ventilation needs [36]. Additionally, atypical lymphocytes in the peripheral blood (regarded as immunologically activated T cells) were associated with pneumonia and need for supplemental oxygen [37]. However, despite documented in almost 60% of the children requiring hospitalization, it did not distinguish the hospitalized from non-hospitalized cohorts or between clinical syndromes in our study. Contrasting with the high hospitalization and PICU rates, the absence of fatal outcomes was the second and probably most intriguing finding. No deaths directly or indirectly related to SARS-CoV-2 infection were documented at our hospital as of November 2022, even during the emergence of Omicron BA.1/BA.2 and BA.5 lineages. While this information could be seen as reassurance of the low severity clinical presentation among children, our numbers diverge from the expected case fatality rates from national and international data [7,25]. Additionally, common cardiovascular complications such as shock and cardiac collapse, considered cardinal acute cardiovascular manifestations in classic episodes of MIS-C, [38] were rarely found in our patients. Such numbers contrast with the Brazilian Ministry of Health data, where the overall CFR of 7% is among the highest in the world [3] and once again could be related to the discrepancies in quality and timely medical care in most Brazilian cities. It is unclear why MIS-C severity differs between countries and populations, but viral characteristics and low specificity of case definitions may be involved. Our data suggest that the diagnostic criteria in use might not be specific enough to predict severe or lethal cases [39,40]. Despite early promotion by the Brazilian government of ineffective drugs against COVID-19 (including hydroxychloroquine, azithromycin, and ivermectin), none of the HIS patients received these medications for COVID-19. However, despite the exceptionally low frequency of antibacterial prescriptions for outpatients, we have identified a disproportionate use of macrolide derivatives, such as azithromycin, among those who received antibiotics. Not shown to improve survival in COVID-19 patients [41,42], recurrent macrolide use affects the gut resistome (the body's largest reservoir of antimicrobial resistance genes) and has the potential to propagate antibiotic resistance in young children [43]. The findings of this study are subject to several limitations. HIS is a private hospital unassociated with Brazil's free and universal Unified Health System, limiting access to those with supplementary private health insurance. In Brazil, where more than 70% the population does not have private medical insurance [44], poverty, inequality, and social determinants of health create favorable conditions for transmission of infectious diseases, including COVID-19 [45,46]. Therefore, as a population with disproportionate access to high-quality medical care, we may have experienced an unintended selection bias [47]. Second, we had a small number of black participants in this analysis, which is consistent with the racial profile of the population attended at HIS. We identified that ethnicity was not a predictor for hospitalization or deaths among our children, which contrasts with previous publications demonstrating race, particularly Black, Indigenous, and Hispanic, is an independent risk factor for negative outcomes in pediatric COVID-19 and MIS-C [31,46,48–50]. Third, data collection through in-depth abstraction of routine clinical documentation is subject to incomplete reporting. Since most clinical data were recovered from electronic records and attending clinicians did not use standardized forms, some inputs were excluded from our dataset. This drawback was minimized by carefully reviewing all available medical records. Fourth, as we have evaluated COVID-19 in a broad but well-defined region in São Paulo, our findings may underestimate the extent of the disease in other regions. The marked climatic diversity and latitudinal differences in the pattern of viral circulation may differ substantially in a country such as Brazil. Moreover, testing strategies may differ regionally, resulting in geographic differences in disease hospitalization rates and outcomes. Fifth, we were not able to evaluate vaccine effectiveness against severe disease or hospitalization as immunization of children from 5 to 11 years became available in the country only in late January 2022. Finally, our findings describe the epidemiology and characteristics of the disease during the circulation of two primary SARS-CoV-2 variants. Continuous hospital-based surveillance of new variants should bring new information on COVID-19 in children, including the circulation of highly transmissible strains (such as Omicron) and effect of pediatric vaccination. With pediatric absolute and proportionate COVID-19 cases on the rise since late December 2021, pediatric disease is cause for increased concern given children's vulnerabilities. Multicenter studies to detect regional disparities in the disease course and replicate our findings in different settings are urgently necessary, especially following the Omicron wave of infections. DISCLAIMER The authors' findings and conclusions are those of the authors and do not necessarily represent the views of the institutions cited in the article. CONFLICT OF INTEREST The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. AUTHOR CONTRIBUTIONS DJ, FJA, and MAPS contributed to the conception and design of the study. All authors contributed to collecting the data. DJ organized the database. DJ wrote the first draft and the definitive version of the manuscript. MAPS and FJA wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version. FUNDING SOURCES None. ACKNOWLEDGMENTS The authors thank Erika Tiemi Fukunaga for providing statistical advice. REFERENCES [1] Rodriguez-Morales AJ, Gallego V, Escalera-Antezana JP, Méndez CA, Zambrano LI, Franco-Paredes C, et al. COVID-19 in Latin America: The implications of the first confirmed case in Brazil. Travel Med Infect Dis 2020;35:101613. https://doi.org/10.1016/j.tmaid.2020.101613. [2] COVID-19 situation reports n.d. https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports (accessed April 27, 2020). [3] Ministério da Saúde, Secretaria de Vigilância em Saúde. Boletim Epidemiológico Especial COVID19 - Semana Epidemiológica 24 (07 a 13/06/2021) 10-sep-21. [4] Tracking SARS-CoV-2 variants n.d. https://www.who.int/emergencies/emergency-health-kits/trauma-emergency-surgery-kit-who-tesk-2019/tracking-SARS-CoV-2-variants (accessed August 16, 2021). [5] Ritchie H, Mathieu E, Rodés-Guirao L, Appel C, Giattino C, Ortiz-Ospina E, et al. Coronavirus Pandemic (COVID-19) Country Profile: Brazil. Our World Data 2020. [6] Ministério da Saúde, Secretaria de Vigilância em Saúde. Boletim Epidemiológico Especial COVID19 - Semana Epidemiológica 23 (05 a 11/06/2022) 2022. [7] Ministério da Saúde. Boletins Epidemiológicos COVID-19 2022. https://www.gov.br/saude/pt-br/centrais-de-conteudo/publicacoes/boletins/boletins-epidemiologicos/covid-19/boletins-epidemiologicos-covid-19 (accessed April 18, 2022). [8] Martins-Filho PR, Barberia LG. The unjustified and politicized battle against vaccination of children and adolescents in Brazil. Lancet Reg Health Am 2022;8:100206. https://doi.org/10.1016/j.lana.2022.100206. [9] Tagarro A, Epalza C, Santos M, Sanz-Santaeufemia FJ, Otheo E, Moraleda C, et al. Screening and Severity of Coronavirus Disease 2019 (COVID-19) in Children in Madrid, Spain. JAMA Pediatr 2020. https://doi.org/10.1001/jamapediatrics.2020.1346. [10] Dong Y, Mo X, Hu Y, Qi X, Jiang F, Jiang Z, et al. Epidemiology of COVID-19 Among Children in China. Pediatrics 2020;145. https://doi.org/10.1542/peds.2020-0702. [11] Lu X, Zhang L, Du H, Zhang J, Li YY, Qu J, et al. SARS-CoV-2 Infection in Children. N Engl J Med 2020;382:1663–5. https://doi.org/10.1056/NEJMc2005073. [12] Zimmermann P, Curtis N. Coronavirus Infections in Children Including COVID-19: An Overview of the Epidemiology, Clinical Features, Diagnosis, Treatment and Prevention Options in Children. Pediatr Infect Dis J 2020;39:355–68. https://doi.org/10.1097/INF.0000000000002660. [13] Wu Z, McGoogan JM. Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72 314 Cases From the Chinese Center for Disease Control and Prevention. JAMA 2020. https://doi.org/10.1001/jama.2020.2648. [14] Patel AB, Verma A. Nasal ACE2 Levels and COVID-19 in Children. JAMA 2020;323:2386–7. https://doi.org/10.1001/jama.2020.8946. [15] Fundação José Luiz Egydio Setúbal. Relatório de Atividades 2018 - Hospital Infantil Sabará 2019. https://fundacaojles.org.br/relatorio-de-atividades-2018/sabara-hospital-infantil/ (accessed June 19, 2020). [16] Beavis KG, Matushek SM, Abeleda APF, Bethel C, Hunt C, Gillen S, et al. Evaluation of the EUROIMMUN Anti-SARS-CoV-2 ELISA Assay for detection of IgA and IgG antibodies. J Clin Virol Off Publ Pan Am Soc Clin Virol 2020;129:104468. https://doi.org/10.1016/j.jcv.2020.104468. [17] Corman VM, Landt O, Kaiser M, Molenkamp R, Meijer A, Chu DK, et al. Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR. Eurosurveillance 2020;25. https://doi.org/10.2807/1560-7917.ES.2020.25.3.2000045. [18] National Institutes of Health. Clinical Spectrum of SARS-CoV-2 Infection 2020. [19] WHO. Multisystem inflammatory syndrome in children and adolescents with COVID-19 2020. [20] American Academy of Pediatrics, Children's Hospital Association. Children and COVID-19: State-Level Data Report. Child COVID-19 State-Level Data Rep 2022. http://www.aap.org/en/pages/2019-novel-coronavirus-covid-19-infections/children-and-covid-19-state-level-data-report/ (accessed March 3, 2022). [21] Zachariah P, Johnson CL, Halabi KC, Ahn D, Sen AI, Fischer A, et al. Epidemiology, Clinical Features, and Disease Severity in Patients With Coronavirus Disease 2019 (COVID-19) in a Children's Hospital in New York City, New York. JAMA Pediatr 2020:e202430. https://doi.org/10.1001/jamapediatrics.2020.2430. [22] Kitano T, Kitano M, Krueger C, Jamal H, Al Rawahi H, Lee-Krueger R, et al. The differential impact of pediatric COVID-19 between high-income countries and low- and middle-income countries: A systematic review of fatality and ICU admission in children worldwide. PloS One 2021;16:e0246326. https://doi.org/10.1371/journal.pone.0246326. [23] Antúnez-Montes OY, Escamilla MI, Figueroa-Uribe AF, Arteaga-Menchaca E, Lavariega-Saráchaga M, Salcedo-Lozada P, et al. COVID-19 and Multisystem Inflammatory Syndrome in Latin American Children: A Multinational Study. Pediatr Infect Dis J 2021;40:e1–6. https://doi.org/10.1097/INF.0000000000002949. [24] Madani S, Shahin S, Yoosefi M, Ahmadi N, Ghasemi E, Koolaji S, et al. Red flags of poor prognosis in pediatric cases of COVID-19: the first 6610 hospitalized children in Iran. BMC Pediatr 2021;21:563. https://doi.org/10.1186/s12887-021-03030-2. [25] UNICEF. Child mortality and COVID-19 2022. https://data.unicef.org/topic/child-survival/covid-19/ (accessed April 20, 2022). [26] Bundle N, Dave N, Pharris A, Spiteri G, Deogan C, Suk JE, et al. COVID-19 trends and severity among symptomatic children aged 0-17 years in 10 European Union countries, 3 August 2020 to 3 October 2021. Euro Surveill 2021;26. https://doi.org/10.2807/1560-7917.ES.2021.26.50.2101098. [27] Ward JL, Harwood R, Smith C, Kenny S, Clark M, Davis PJ, et al. Risk factors for PICU admission and death among children and young people hospitalized with COVID-19 and PIMS-TS in England during the first pandemic year. Nat Med 2022;28:193–200. https://doi.org/10.1038/s41591-021-01627-9. [28] Shekerdemian LS, Mahmood NR, Wolfe KK, Riggs BJ, Ross CE, McKiernan CA, et al. Characteristics and Outcomes of Children With Coronavirus Disease 2019 (COVID-19) Infection Admitted to US and Canadian Pediatric Intensive Care Units. JAMA Pediatr 2020;174:868–73. https://doi.org/10.1001/jamapediatrics.2020.1948. [29] Uka A, Buettcher M, Bernhard-Stirnemann S, Fougère Y, Moussaoui D, Kottanattu L, et al. Factors associated with hospital and intensive care admission in paediatric SARS-CoV-2 infection: a prospective nationwide observational cohort study. Eur J Pediatr 2022;181:1245–55. https://doi.org/10.1007/s00431-021-04276-9. [30] Shi Q, Wang Z, Liu J, Wang X, Zhou Q, Li Q, et al. Risk factors for poor prognosis in children and adolescents with COVID-19: A systematic review and meta-analysis. EClinicalMedicine 2021;41:101155. https://doi.org/10.1016/j.eclinm.2021.101155. [31] Woodruff RC, Campbell AP, Taylor CA, Chai SJ, Kawasaki B, Meek J, et al. Risk Factors for Severe COVID-19 in Children. Pediatrics 2021:e2021053418. https://doi.org/10.1542/peds.2021-053418. [32] Korakas E, Ikonomidis I, Kousathana F, Balampanis K, Kountouri A, Raptis A, et al. Obesity and COVID-19: immune and metabolic derangement as a possible link to adverse clinical outcomes. Am J Physiol Endocrinol Metab 2020;319:E105–9. https://doi.org/10.1152/ajpendo.00198.2020. [33] Frydrych LM, Bian G, O'Lone DE, Ward PA, Delano MJ. Obesity and type 2 diabetes mellitus drive immune dysfunction, infection development, and sepsis mortality. J Leukoc Biol 2018;104:525–34. https://doi.org/10.1002/JLB.5VMR0118-021RR. [34] Li Y, Woods K, Parry-Strong A, Anderson RJ, Capistrano C, Gestin A, et al. Distinct Dysfunctional States of Circulating Innate-Like T Cells in Metabolic Disease. Front Immunol 2020;11:448. https://doi.org/10.3389/fimmu.2020.00448. [35] Wool GD, Miller JL. The Impact of COVID-19 Disease on Platelets and Coagulation. Pathobiol J Immunopathol Mol Cell Biol 2021;88:15–27. https://doi.org/10.1159/000512007. [36] Lucijanic M, Krecak I, Soric E, Sedinic M, Sabljic A, Derek L, et al. Thrombocytosis in COVID-19 patients without myeloproliferative neoplasms is associated with better prognosis but higher rate of venous thromboembolism. Blood Cancer J 2021;11:189. https://doi.org/10.1038/s41408-021-00585-2. [37] Sugihara J, Shibata S, Doi M, Shimmura T, Inoue S, Matsumoto O, et al. Atypical lymphocytes in the peripheral blood of COVID-19 patients: A prognostic factor for the clinical course of COVID-19. PloS One 2021;16:e0259910. https://doi.org/10.1371/journal.pone.0259910. [38] Valverde I, Singh Y, Sanchez-de-Toledo J, Theocharis P, Chikermane A, Di Filippo S, et al. Acute Cardiovascular Manifestations in 286 Children With Multisystem Inflammatory Syndrome Associated With COVID-19 Infection in Europe. Circulation 2021;143:21–32. https://doi.org/10.1161/CIRCULATIONAHA.120.050065. [39] Godfred-Cato S. COVID-19–Associated Multisystem Inflammatory Syndrome in Children — United States, March–July 2020. MMWR Morb Mortal Wkly Rep 2020;69. https://doi.org/10.15585/mmwr.mm6932e2. [40] Rhys-Evans S. Call for a universal PIMS-TS/MIS-C case definition. Arch Dis Child 2022;107:e10. https://doi.org/10.1136/archdischild-2021-322829. [41] RECOVERY Collaborative Group. Azithromycin in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet 2021;397:605–12. https://doi.org/10.1016/S0140-6736(21)00149-5. [42] PRINCIPLE Trial Collaborative Group. Azithromycin for community treatment of suspected COVID-19 in people at increased risk of an adverse clinical course in the UK (PRINCIPLE): a randomised, controlled, open-label, adaptive platform trial. Lancet 2021;397:1063–74. https://doi.org/10.1016/S0140-6736(21)00461-X. [43] Doan T, Worden L, Hinterwirth A, Arzika AM, Maliki R, Abdou A, et al. Macrolide and Nonmacrolide Resistance with Mass Azithromycin Distribution. N Engl J Med 2020;383:1941–50. https://doi.org/10.1056/NEJMoa2002606. [44] Instituto Brasileiro de Geografia e Estatística. Pesquisa Nacional de Saúde 2019 2019. https://www.ibge.gov.br/estatisticas/sociais/saude/9160-pesquisa-nacional-de-saude.html?=&t=microdados (accessed October 22, 2021). [45] Quinn SC, Kumar S. Health inequalities and infectious disease epidemics: a challenge for global health security. Biosecurity Bioterrorism Biodefense Strategy Pract Sci 2014;12:263–73. https://doi.org/10.1089/bsp.2014.0032. [46] Oliveira EA, Colosimo EA, Simões E Silva AC, Mak RH, Martelli DB, Silva LR, et al. Clinical characteristics and risk factors for death among hospitalised children and adolescents with COVID-19 in Brazil: an analysis of a nationwide database. Lancet Child Adolesc Health 2021;5:559–68. https://doi.org/10.1016/S2352-4642(21)00134-6. [47] Moore JT. Disparities in Incidence of COVID-19 Among Underrepresented Racial/Ethnic Groups in Counties Identified as Hotspots During June 5–18, 2020 — 22 States, February–June 2020. MMWR Morb Mortal Wkly Rep 2020;69. https://doi.org/10.15585/mmwr.mm6933e1. [48] Moreira A, Chorath K, Rajasekaran K, Burmeister F, Ahmed M, Moreira A. Demographic predictors of hospitalization and mortality in US children with COVID-19. Eur J Pediatr 2021;180:1659–63. https://doi.org/10.1007/s00431-021-03955-x. [49] Lee EH, Kepler KL, Geevarughese A, Paneth-Pollak R, Dorsinville MS, Ngai S, et al. Race/Ethnicity Among Children With COVID-19-Associated Multisystem Inflammatory Syndrome. JAMA Netw Open 2020;3:e2030280. https://doi.org/10.1001/jamanetworkopen.2020.30280. [50] Alsaied T, Tremoulet AH, Burns JC, Saidi A, Dionne A, Lang SM, et al. Review of Cardiac Involvement in Multisystem Inflammatory Syndrome in Children. Circulation 2021;143:78–88. https://doi.org/10.1161/CIRCULATIONAHA.120.049836. [51] Empresa Paulista de Planejamento Metropolitano S/A (EMPLASA), Governo do Estado de São Paulo. Região Metropolitana de São Paulo 2019. https://www.emplasa.sp.gov.br/RMSP (accessed June 19, 2020). [52] Sistema de Informação de Agravos de Notificação (SINAN), Ministério da Saúde do Brasil. SINANWEB - Calendários Epidemiológicos 2021. http://portalsinan.http://portalsinan.saude.gov.br/calendario-epidemiologico/171-calendario-epidemiologico-2021 (accessed June 19, 2022). [53] e-SUS VE/DVE/COVISA/SMS-SP. TabNet Win32 2.7 - COVID19 e-SUS-VE Síndrome Gripal n.d. http://tabnet.saude.prefeitura.sp.gov.br/cgi/tabcgi.exe?secretarias/saude/TABNET/RCOVID19/covid19.def (accessed June 19, 2022). Appendix Supplementary materials Image, application 1 Image, application 2 Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.ijregi.2022.12.003.

0

PMC9753484

NO-CC CODE

2022-12-16 23:25:44

no

IJID Reg. 2022 Dec 15; doi: 10.1016/j.ijregi.2022.12.003

utf-8

IJID Reg

10.1016/j.ijregi.2022.12.003

oa_other

==== Front Infect Prev Pract Infect Prev Pract Infection Prevention in Practice 2590-0889 Published by Elsevier Ltd on behalf of The Healthcare Infection Society. S2590-0889(22)00066-X 10.1016/j.infpip.2022.100265 100265 Article A Simulation-Based PPE Orientation Training Curriculum for Novice Physicians Greaves Spencer W. a Alter Scott M. a Ahmed Rami A. b Hughes Kate E. c Doos Devin b Clayton Lisa M. a Solano Joshua J. a Echeverri Sindiana d Shih Richard D. a Hughes Patrick G. a∗ a Department of Emergency Medicine, Florida Atlantic University Charles E. Schmidt College of Medicine b Department of Emergency Medicine, Division of Simulation, Indiana University School of Medicine c Department of Emergency Medicine, University of Arizona d Clinical Skills Simulation Center, Florida Atlantic University Charles E. Schmidt College of Medicine ∗ Corresponding author. Florida Atlantic University at Bethesda Health, Department of Emergency Medicine, GME Suite, Lower Level, 2815 South Seacrest Blvd, Boynton Beach, FL 33435, Tel: (561) 733-5933, Fax: (866) 617-8268, 15 12 2022 15 12 2022 10026516 8 2022 24 10 2022 9 12 2022 © 2022 Published by Elsevier Ltd on behalf of The Healthcare Infection Society. 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Background Personal protective equipment (PPE) is effective in preventing coronavirus disease (COVID-19) infection. Resident knowledge of proper use and effective training methods is unknown. We hypothesise that contamination decreases and knowledge increases after a formalised PPE educational session. Methods Participants included first year interns during their residency orientation in June 2020. Before training, participants took a knowledge test, donned PPE, performed a simulated resuscitation, and doffed. A standardised simulation-based PPE training of the donning and doffing protocol was conducted, and the process repeated. Topical non-toxic highlighter tracing fluid was applied to manikins prior to each simulation. After doffing, areas of contamination, defined as discrete fluorescent areas on participants’ body, was evaluated by ultraviolet light. Donning and doffing were video recorded and asynchronously rated by two emergency medicine (EM) physicians using a modified Centers for Disease Control and Prevention (CDC) protocol. The primary outcome was PPE training effectiveness defined by contamination and adherence to CDC sequence. Results Forty-eight residents participated: 24 internal medicine, 12 general surgery, 6 EM, 3 neurology, and 3 psychiatry. Before training, 81% of residents were contaminated after doffing; 17% were contaminated after training (P<0.001). The most common contamination area was the wrist (50% pre-training vs. 10% post-training, P<0.001). Donning sequence adherence improved (52% vs. 98%, P<0.001), as did doffing (46% vs. 85%, P<0.001). Participant knowledge improved (62% to 87%, P <0.001). Participant confidence (P<0.001) and preparedness (P<0.001) regarding using PPE increased with training. Conclusion A simulation-based training improved resident knowledge and performance using PPE. Key words COVID-19 simulation PPE training fluorescent tracer quality improvement ==== Body pmcIntroduction Due to the bedside nature of the profession, physicians and other healthcare workers are at an increased risk of occupational exposure to communicable diseases [1]. This has become especially important during the SARS-CoV-2 pandemic where hundreds of healthcare workers have died [2]. Unfortunately, the growth of the world population, urbanisation, and international trade has increased the risk of global spread of infectious disease likely making it easier for pandemics to evolve [3]. As such, an important aspect in pandemic responses will be the effective use of personal protective equipment (PPE) by healthcare workers. Despite this, there is evidence that suggests that many physicians do not have formal training in proper use of PPE [4]. There is limited published research on formal effective curricula to educate medical students prior to exposure to hazards in the clinical environment of residency training [5]. The core focus of medical school curricula is to prepare the graduate for medical practice. However, an assessment has shown the majority of medical students have inadequate, incomplete, or non-existent PPE training [6]. This lack of PPE training for recent healthcare graduates is evident across many different medical education settings globally [7]. We describe a simulation-based method for PPE training that can be easily implemented in quality improvement programs. Our study evaluates PPE training effectiveness defined by contamination and adherence to Centers for Disease Control and Prevention (CDC) sequence. Methods Study Design, Participants and Setting This study was a prospective observational study of quality improvement measures implemented in residency orientation training in June 2020. First year resident physicians who needed American Heart Association (AHA) life support certifications (advanced cardiac life support, basic life support, etc.) were enrolled. There were three distinct phases to this process: pre-training, training, and post-training (Figure 1 ). All phases of the study were conducted at a single university simulation center. The study was determined not to be human subjects research by the university’s institutional review board.Figure 1 Phases of training. Figure 1 In the first phase, participants were asked to fill out a survey to establish baseline knowledge of PPE usage and previous training in medical school. Next, participants were asked to don PPE in the manner they felt was most effective while being video recorded. Participants then performed cardiopulmonary resuscitation (CPR) and positive pressure ventilation using a bag valve mask (BVM) on manikins (unmodified CAE iStan, CAE Healthcare INC, Sarasota, FL). Topical non-toxic invisible highlighter fluid was used as a tracer [8], which was applied with equal amounts distributed on the face, neck, and chest. After performing two rounds of compressions and one round of ventilations, participants were asked to doff PPE while being video recorded. Participants were then examined under long wave ultraviolet (UV) light for evidence of the tracer, indicating areas of contamination (Figure 2 ).Figure 2 Ultraviolet light assessment of contamination. Figure 2 The second phase consisted of a standardised 30-minute formative training program utilising modified CDC guidelines on donning and doffing procedures of PPE. Participants were asked to reflect on their performance during the simulated code and give their opinion why they may have had contamination. The CDC training program included a step-by-step simulation-based review of proper PPE the donning and doffing sequence with deliberate practice and feedback (Supplement 1). The final phase was similar to the first. Participants were asked to don PPE, perform compressions and ventilations, and then doff their PPE prior to examination for contamination under long-wave UV light. Participants then completed a post-training survey including a knowledge test and a curricular evaluation. Measurements and Outcomes The primary study outcome was to evaluate PPE training effectiveness defined by contamination and adherence to CDC sequence. The secondary outcomes were to assess participant knowledge on PPE usage and isolation precaution types and to evaluate participant perceptions of the training and use of fluorescent tracer to visualise errors. There were three sets of data gathered during this study: participant pre- and post-training survey data, contamination data, and video data of donning and doffing procedures for participants before and after training. Survey data included questions on previous PPE education, precaution types, previous clinical experience, pre- and post-training knowledge on PPE usage, as well as confidence with PPE utilisation (Supplement 2). Contamination was defined as discrete areas of fluorescence separated by greater than 3 cm. Two independent assessors were used to judge contamination. Discrepancies were discussed and adjudicated and a third independent observers was involved, if necessary. The number of contaminated areas was recorded for each specific body area, including face, neck, chest, arms, forearms, wrists, hands, and legs. Video data was scored by two of five randomly assigned independent raters, all emergency medicine attending physicians, using a summarised scoring sheet of CDC PPE recommendations (Supplement 3). Each rater assessed performance of specific and defined aspects of PPE donning or doffing. Donning was scored from one to five points and doffing was scored from one to seven points. The evaluators completed a video training session and inter-rater reliability was assessed using intraclass correlation coefficient (ICC). Final ICC was 0.96, indicating excellent inter-rater reliability. Data management of all measurements were preformed using spreadsheet software (Microsoft Excel, Microsoft Corporation, Redmond, WA). Analysis General summary statistics were performed on participant survey data. Descriptive statistics on overall and specific body part contamination were performed and compared for participants before and after training. Donning and doffing scores were compared with respect to training and comparative categorical univariate statistics. Descriptive and univariate statistics were performed using Stata Release 16 (StataCorp, College Station, TX) and SPSS Statistics for Windows, version 27 (IBM Corp., Armonk, NY). Results A total of 48 residents completed the training course: 24 internal medicine, 12 general surgery, 6 emergency medicine, 3 neurology, and 3 psychiatry. Survey Survey data revealed that 71% of participants had not undergone any PPE training. Only 23% had hands-on general PPE training in addition to either video-based content training, written, or didactic based training. Only 23% of participants had any COVID-19 specific PPE training and only 6% had any clinical experience with COVID-19 patients. Many participants stopped clinical rotations in March 2020 (46%) when they were medical students. Furthermore, only 15% of participants had a rotation in the past three months prior to residency. Of note, 15% had not had clinical exposure in the past six months prior to residency.48% of residents had never used PPE in that time period and 79% of those participants donned and doffed PPE less than 10 times. Overall participant knowledge scores increased after training (62% vs. 87%; difference 25%; 95% CI: 19%-31%; P<0.001). Knowledge of doffing sequence significantly increased, but no significant change was noted in donning sequence. Participant knowledge of transmission of tuberculosis, chickenpox, SARS, Norwalk virus and West Nile virus improved after the education, but no significant change in COVID-19 (Table I ).Table 1 Knowledge Questions Results Summary, mean (95% CI). Table 1Knowledge question Pre-training correct Post-training correct Difference P-value Donning sequence 81% 92% 10% (-3%-24%) 0.133 Doffing sequence 52% 94% 42% (27%-56%) <0.001 Tuberculosis transmission 75% 98% 23% (11%-35%) 0.001 Chickenpox transmission 27% 54% 27% (12%-43%) 0.001 SARS transmission 69% 94% 25% (10%-40%) 0.002 Norwalk virus transmission 50% 90% 40% (23%-56%) <0.001 West Nile virus transmission 75% 94% 19% (6%-32%) 0.005 COVID-19 transmission 67% 81% 15% (-3%-33%) 0.109 Overall score 62% 87% 25% (19%-31%) <0.001 Participant confidence in donning and doffing PPE safely and preparedness to appropriately use PPE when caring for COVID-19 patients significantly increased after training (Table II ). The post-training survey data showed that 44% felt that practicing donning and doffing was one of the most beneficial parts of the training while 46% of people felt the most beneficial aspect was the use of a fluorescent tracer to visualise errors. 98% of participants felt the use of fluorescent tracer was helpful and 92% felt it improved their ability to visualise errors during the donning and doffing process. Only 4% felt that the training would have no effect on their future PPE use. Overall, 98% of participants agreed or strongly agreed that all incoming residents should receive PPE training prior to starting work in the healthcare setting.Table 2 Opinion questions. Table 2 Top 2 boxesa, n (%) Difference (95% CI) Median (IQR) P-value Pre Post Pre Post Confidence in PPE donning/doffing 6 (13%) 48 (100%) 88% (78%-97%) 3 (2-3) 5 (4-5) <0.001 Concerned of COVID-19 exposure 17 (35%) 13 (27%) -8% (-24%-7%) 3 (2-4) 3 (2-4) 0.311 Prepared to use PPE 8 (17%) 47 (98%) 81% (70%-93%) 2 (2-3) 5 (4-5) <0.001 Residents should receive PPE education 47 (98%) 47 (98%) 0% (-6%-6%) 5 (5-5) 5 (5-5) 0.279 a Responses of 4 or 5 on 5-point Likert scale. Contamination Prior to training, 81% (39/48) of participants had contamination compared to 17% (8/48) after training (P<0.001). Half of all participants had at least one wrist contaminated, which was the most common area of contamination. The hand and forearm were the next two most common areas of contamination. With training, all areas of contamination reduced, with no contamination of the face after PPE training (Table III ). Prior to training, statistically significant correlations between contaminated body areas were noted of the arm and face, wrist and face, and hand and wrist (Table IV ).Table 3 Subjects with Contamination Results Summary, n (%). Table 3Body surface Pre-training contamination Post-training contamination Difference (95% CI) P-value Face 10 (21%) 0 (0%) -21% (-33%--9%) 0.001 Neck 10 (21%) 0 (0%) -21% (-33%--9%) 0.001 Chest 4 (8%) 0 (0%) -8% (-16%-0%) 0.044 Arm 3 (6%) 2 (4%) -2% (-12%-7%) 0.659 Forearm 11 (23%) 5 (10%) -13% (-27%-2%) 0.083 Wrist 24 (50%) 5 (10%) -40% (-55%--24%) <0.001 Hand 14 (29%) 0 (0%) -29% (-43%--16%) <0.001 Leg 0 (0%) 0 (0%) 0% Any area 39 (81%) 8 (17%) -65% (-80%--49%) <0.001 Table 4 Correlations between contaminated body areas pre training. Table 4 Face Neck Chest Arm Forearm Wrist Hand Leg Face - -0.14 0.03 0.29∗ -0.04 0.31† 0.12 - Neck -0.14 - 0.03 0.08 -0.16 0.00 0.01 - Chest 0.03 0.03 - 0.23 -0.16 0.00 -0.03 - Arm 0.29∗ 0.08 0.23 - 0.06 0.08 0.02 - Forearm -0.04 -0.16 -0.16 0.06 - 0.05 -0.24 - Wrist 0.31† 0.00 0.00 0.08 0.05 - 0.37‡ - Hand 0.12 0.01 -0.03 0.02 -0.24 0.37‡ - - Leg - - - - - - - - ∗ Significant at P=0.044. † Significant at P=0.033. ‡ Significant at P=0.010. Donning and doffing Average total score for donning and doffing of participants significantly increased with training. This was true for both average donning scores (52% vs. 98%; difference 47%; 95% CI: 37%-57%; P<0.001) and doffing scores (46% vs. 85%; difference 39%; 95% CI: 33%-46%; P<0.001). Participants had significantly increased success rate for each step of donning and doffing (Table 5, Table 6 ).Table 5 Successful donning steps before and after training. Table 5 Pre-training success Post-training success Difference (95% CI) P-value 1: Sanitize hands 76% 100% 24% (11%-37%) <0.001 2: Don isolation gown 36% 97% 61% (48%-74%) <0.001 3: Don N95 mask 40% 99% 59% (45%-72%) <0.001 4: Don face shield or goggles 49% 100% 51% (37%-65%) <0.001 5: Put on gloves 56% 100% 44% (29%-58%) <0.001 Total 52% 98% 47% (37%-57%) <0.001 Table 6 Successful doffing steps before and after training. Table 6 Pre-training success Post-training success Difference (95% CI) P-value 1: Remove gloves 26% 100% 74% (62%-87%) <0.001 2: Remove gown 24% 96% 71% (58%-85%) <0.001 3: Leave contaminated area 6% 49% 43% (27%-58%) <0.001 4: Sanitize hands 52% 70% 18% (0%-36%) 0.048 5: Remove face shield or goggles 74% 96% 21% (10%-32%) <0.001 6: Remove/waste N95 83% 99% 16% (7%-25%) 0.001 7: Sanitize hands 54% 86% 32% (15%-49%) 0.001 Total 46% 85% 39% (33%-46%) <0.001 Discussion Our study highlights the need for PPE training for newly graduated medical students entering residency. Prior PPE use knowledge was poor, and PPE education is an essential aspect of orientation prior to clinical work. It is also important for clinicians to know the transmission modes of virulent diseases for proper body substance isolation in avoiding contamination. During donning and doffing of PPE, frequent contamination occurred and a pattern was identified in this study. Wrist contamination was the most common area and correlated with face contamination. Most importantly, PPE training had clear effect in reducing contamination after doffing. Our study also found that the most common error was poor adherence to PPE donning and doffing sequence. The simulation-based training was positively perceived by participants. Post-survey data showed it improved confidence and preparedness of PPE usage for clinical care of COVID-19 patients. Previous studies have utilised ultraviolet fluorescent tracer during simulated patient history and physical exams, however this study is unique in its use of CPR as the defining simulation event [9]. Despite this difference, our results indicate that during PPE usage, the wrist is the major source of contamination which has also been shown in previous research [[10], [11], [12], [13]]. Our study further shows that contamination of the wrists and arms appears to correlate with contamination of the face, a result not reported previously. This is important as face contamination likely increases the risk of mucous membrane exposure and infection since healthcare workers do not routinely wash their faces after doffing their PPE [13]. There are a variety of unique errors occur during PPE use due to its numerous and complex sequence of steps to properly perform [13, 14]. Most errors occurred because of failure to adhere to the recommended step-by-step protocol (e.g. not sanitizing hands before placing isolation gown) or improper technique. Many of these errors occurred because participants did not secure the gowns behind their bodies correctly during donning, allowing for the gown to become dislodged during CPR. The investigators noticed that this appeared to result in significantly more touching between the gown and the provider’s hands as attempts were made to readjust the gown. Previous research suggests this lack of proper sequence is not uncommon and that errors in doffing are high risk for contamination since these occur after attending to the patient [15]. The extremely low number of participants that underwent hands-on PPE training prior to starting internship is surprising and reflects the need for training. This is highlighted by participant feedback regarding the usefulness of the training program and is also shown in previous simulation-based training in medical education [16]. Post-curricular evaluations overwhelmingly suggest that tracer utilisation was a highly beneficial component of training methodology as it provided objective feedback regarding contamination. While previous research suggests fluorescent tracer may not delineate all clinically relevant areas of contamination [12], it appears to be effective in estimating how effectively someone utilises PPE and their risk for potential contamination [13, 17, 18]. Further, the most beneficial aspect of this training regimen utilising tracer may lie in its ability to give real-time visual feedback as someone is performing simulated training exercises. Previous research also supports this type of training as both direct feedback and repetition of educational goals is effective in trainee understanding [19]. While it is not clear if improved PPE use will translate into future reductions in workplace-caused infections, an immediate secondary benefit is the reduced anxiety regarding the risk of work-related occupational exposure. Previous research has also shown this type of benefit [20]. Additionally, there is a general increased psychological risk with prolonged use of PPE [21]. With a better knowledge on the donning and doffing technique and lessened anxiety regarding iatrogenic exposure, clinicians may be more willing to remove and replace PPE for breaks. This is especially important as resident interns are particularly prone to excessive stress, anxiety and burnout [[22], [23], [24], [25]]. This study has several important limitations. While our results suggest that PPE training is both highly effective and essential to proper usage, it is unclear if this leads to decreased actual transmission of communicable diseases. However, during a pandemic, it is unlikely that a randomized trial of PPE training would be undertaken for ethical reasons. Further, this study only included physicians. It did not study non-physicians. However, there is no clear reason that these individuals would not also benefit from this type of training. Another limitation is that use of the fluorescent tracer is based on contact and not aerosolization. There are tracers that simulate aerosol exposure and may be more applicable for viral transmission. Lastly, the training and testing occurred on the same day and it is unclear what interval of re-training may be needed to solidify long-term habits regarding PPE usage. Conclusions A simulation-based curriculum utilizing fluorescent tracer demonstrated improved confidence, improved knowledge, reduced contamination, and improved performance of donning and doffing PPE in simulated scenarios for first year residents. Funding No funding to declare. Ethics approval University IRB reviewed study and determined it not to be human subjects research. Declaration of Interests Research support: This research received no external financial or non-financial support. Relationships: There are no additional relationships to disclose. Patents and Intellectual Property: There are no patents to disclose. Other Activities: There are no additional activities to disclose. Authors’ CRediT criteria for inclusion Spencer W. Greaves: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data Curation, Writing – Original Draft, Writing – Review and Editing, Visualization Scott M. Alter: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Data Curation, Writing – Original Draft, Writing – Review and Editing Rami A. Ahmed: Conceptualization, Methodology, Validation, Investigation, Writing – Original Draft, Writing – Review and Editing Kate E. Hughes: Conceptualization, Methodology, Validation, Investigation, Writing – Original Draft, Writing – Review and Editing Devin Doos: Conceptualization, Methodology, Validation, Investigation, Writing – Original Draft, Writing – Review and Editing Lisa M. Clayton: Conceptualization, Methodology, Validation, Investigation, Writing – Original Draft, Writing – Review and Editing Joshua J. Solano: Conceptualization, Methodology, Validation, Investigation, Writing – Original Draft, Writing – Review and Editing Sindiana Echeverri: Conceptualization, Methodology, Investigation, Resources, Writing – Original Draft, Writing – Review and Editing Richard D. Shih: Methodology, Investigation, Writing – Original Draft, Writing – Review and Editing Patrick G. Hughes: Conceptualization, Methodology, Validation, Investigation, Resources, Writing – Original Draft, Writing – Review and Editing, Visualization, Supervision. Appendix A Supplementary data The following are the Supplementary data to this article: Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.infpip.2022.100265. ==== Refs References 1 Haagsma J.A. Tariq L. Heederik D.J. Havelaar A.H. Infectious disease risks associated with occupational exposure: a systematic review of the literature Occup Environ Med 69 2012 140 146 10.1136/oemed-2011-100068 22006935 2 Gouda D. Singh P.M. Gouda P. Goudra B. An overview of health care worker reported deaths during the COVID-19 pandemic J. Am. Board Fam. Med. 34 2021 S244 S246 10.3122/jabfm.2021.S1.200248 33622846 3 Allen T. Murray K.A. Zambrana-Torrelio C. Morse S.S. Rondinini C. Di Marco M. Breit N. Olival K.J. Daszak P. Global hotspots and correlates of emerging zoonotic diseases Nat. commun. 8 2017 1 10 10.1038/s41467-017-00923-8 28232747 4 John A. Tomas M.E. Cadnum J.L. Mana T.S. Jencson A. Shaikh A. Zabarsky T.F. Wilson B.M. Donskey C.J. Are health care personnel trained in correct use of personal protective equipment? Am. J. Infect. Control. 44 2016 840 842 10.1016/j.ajic.2016.03.031 27181222 5 Villamagna A.H. Bonura E.M. Infectious diseases simulation for medical students: experiential instruction on personal protective equipment MedEdPORTAL 16 2020 1103 10.15766/mep_2374-8265.11031 6 John A. Tomas M.E. Hari A. Wilson B.M. Donskey C.J. Do medical students receive training in correct use of personal protective equipment? Med. Educ. Online. 22 2017 1264125 10.1080/10872981.2017.1264125 7 Amin T.T. Al Noaim K.I. Saad M.A.B. Al Malhm T.A. Al Mulhim A.A. Al Awas M.A. Standard precautions and infection control, medical students’ knowledge and behavior at a Saudi university: the need for change Glob. J. Health Sci. 5 2013 114 10.5539/gjhs.v5n4p114 23777728 8 Link M.S. Berkow L.C. Kudenchuk P.J. Halperin H.R. Hess E.P. Moitra V.K. Neumar R.W. O'Neil B.J. Paxton J.H. Silvers S.M. White R.D. Yannopoulos D. Donnino M.W. Part 7: adult advanced cardiovascular life support: 2015 American Heart Association guidelines update for cardiopulmonary resuscitation and emergency cardiovascular care Circulation 132 2015 S444 S464 10.1161/CIR.0000000000000261 26472995 9 Drew J.L. Turner J. Mugele J. Hasty G. Duncan T. Zaiser R. Cooper D. Beating the spread: developing a simulation analog for contagious body fluids Simul. Healthc. 11 2016 100 105 10.1097/SIH.0000000000000157 27043095 10 Tomas M.E. Cadnum J.L. Mana T.S. Jencson A.L. Donskey C.J. Seamless suits: reducing personnel contamination through improved personal protective equipment design Infect. Control Hosp. Epidemiol. 37 2016 742 744 10.1017/ice.2016.79 27068020 11 Verbeek J.H. Rajamaki B. Ijaz S. Sauni R. Toomey E. Blackwood B. Tikka C. Ruotsalainen J.H. Selcen Kilinc Balcie F. Personal protective equipment for preventing highly infectious diseases due to exposure to contaminated body fluids in healthcare staff Cochrane Database Syst. Rev. 4 2020 CD011621 10.1002/14651858.CD011621.pub4 32293717 12 Bell T. Smoot J. Patterson J. Smalligan R. Jordan R. Ebola virus disease: the use of fluorescents as markers of contamination for personal protective equipment IDCases 2 2014 27 30 10.1016/j.idcr.2014.12.003 26793445 13 Tomas M.E. Kundrapu S. Thota P. Sunkesula V.C. Cadnum J.L. Mana T.S.C. Jencson A. O'Donnell M. Zabarsky T.F. Hecker M.T. Ray A.J. Wilson B.M. Donskey C.J. Contamination of health care personnel during removal of personal protective equipment JAMA Intern. Med. 175 2015 1904 1910 10.1001/jamainternmed.2015.4535 26457544 14 Kwon J.H. Burnham C.A.D. Reske K.A. Liang S.Y. Hink T. Wallace M.A. Shupe A. Seiler S. Cass C. Fraser V.J. Dubberke E.R. Assessment of healthcare worker protocol deviations and self-contamination during personal protective equipment donning and doffing Infect. Control Hosp. Epidemiol. 38 2017 1077 1083 10.1017/ice.2017.121 28606192 15 Beam E.L. Gibbs S.G. Boulter K.C. Beckerdite M.E. Smith P.W. A method for evaluating health care workers’ personal protective equipment technique Am. J. Infect. Control. 39 2011 415 420 10.1016/j.ajic.2010.07.009 21255874 16 Abas T. Juma F.Z. Benefits of simulation training in medical education Adv. Med. Educ. Pract. 7 2016 399 400 10.2147/amep.S110386 27486352 17 Zamora J.E. Murdoch J. Simchison B. Day A.G. Contamination: a comparison of 2 personal protective systems CMAJ 175 2006 249 254 10.1503/cmaj.060094 16880444 18 Casanova L. Alfano-Sobsey E. Rutala W.A. Weber D.J. Sobsey M. Virus transfer from personal protective equipment to healthcare employees’ skin and clothing Emerg. Infect. Dis. 14 2008 1291 1293 10.3201/eid1408.080085 18680659 19 McGaghie W.C. Issenberg S.B. Petrusa E.R. Scalese R.J. A critical review of simulation‐based medical education research: 2003–2009 Med. Educ. 44 2010 50 63 10.1111/j.1365-2923.2009.03547.x 20078756 20 Norton E.J. Georgiou I. Fung A. Nazari A. Bandyopadhyay S. Saunders K.E.A. Personal protective equipment and infection prevention and control: a national survey of UK medical students and interim foundation doctors during the COVID-19 pandemic J. Public Health 43 2020 67 75 10.1093/pubmed/fdaa187 21 Centers for Disease Control and Prevention, The Physiological Burden of Prolonged PPE Use on Healthcare Workers during Long Shifts. https://blogs.cdc.gov/niosh-science-blog/2020/06/10/ppe-burden/, 2020 (accessed 18 June 2022). 22 Butterfield P.S. The stress of residency: a review of the literature Arch. Intern. Med. 148 1988 1428 1435 3288162 23 Navines R. Olive V. Ariz J. Lopez J. Tortajada M. Varela P. Valdes M. Martin-Santos R. Stress and burnout during the first year of residence training in a university teaching hospital: preliminary date Dual Diagn. Open Acc. 1 2016 17 https://doi:10.21767/2472-5048.100017 24 Lue B.H. Chen H.J. Wang C.W. Cheng Y. Chen M.C. Stress, personal characteristics and burnout among first postgraduate year residents: a nationwide study in Taiwan Med. Teach 32 2010 400 407 10.3109/01421590903437188 20423259 25 Thomas N.K. Resident burnout JAMA 292 2004 2880 2889 10.1001/jama.292.23.2880 15598920

0

PMC9753485

NO-CC CODE

2022-12-16 23:25:44

no

Infect Prev Pract. 2022 Dec 15;:100265

utf-8

Infect Prev Pract

10.1016/j.infpip.2022.100265

oa_other

==== Front Int J Hum Comput Stud Int J Hum Comput Stud International Journal of Human-Computer Studies 1071-5819 1071-5819 Elsevier Ltd. S1071-5819(22)00200-2 10.1016/j.ijhcs.2022.102982 102982 Article Virtual nature experiences and mindfulness practices while working from home during COVID-19: Effects on stress, focus, and creativity Ch Nabil Al Nahin a⁎ Ansah Alberta A. a Katrahmani Atefeh b Burmeister Julia c Kun Andrew L. a Mills Caitlin d Shaer Orit c Lee John D. b a University of New Hampshire, Electrical and Computer Engineering, 105 Main Street, Durham, 03824, NH, USA b University of Wisconsin-Madison, Industrial and Systems Engineering, 702 West Johnson Street, Madison, 53715, WI, USA c Wellesley College, Computer Science, 106 Central St, Wellesley, 02481, MA, USA d University of New Hampshire, College of Liberal Arts, 105 Main Street, Durham, 03824, NH, USA ⁎ Corresponding author. 15 12 2022 15 12 2022 10298214 3 2022 11 12 2022 13 12 2022 © 2022 Elsevier Ltd. All rights reserved. 2022 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. In this study, we focus on the impact of daily virtual nature experiences combined with mindfulness practices on remote workers’ creativity, stress, and focus over an extended period (9 weeks) during the COVID-19 pandemic. Our results show a positive effect of virtual reality (VR) nature experience on increasing focus and reducing stress. When VR nature and mindfulness practices were combined, we also found an increase in convergent thinking task performance. Our findings demonstrate that 10-minute daily exposure to VR nature and mindfulness practices could compensate for some of the adverse effects of working remotely by improving some aspects of workers’ well-being and creativity. Keywords Working from home COVID-19 Stress Creativity VR Mindfulness ==== Body pmc1 Introduction The COVID-19 pandemic has drastically changed the way we live and work (Rudnicka et al., 2020). Many people were suddenly forced to work from home, a change that neither workers nor organizations have planned or prepared for. Although previous studies show that working from home may have some benefits like increasing productivity (Hunter, 2019, Bloom et al., 2015) and worker satisfaction (Bloom et al., 2015), the situation for workers during the pandemic was often challenging. The change was abrupt and many people were not prepared to work from home. Without a proper work setup at home, many workers had to improvise with limited resources (Newbold et al., 2021). Sometimes that meant working from the couch or kitchen table or sharing a small workspace with others in the home. A major challenge for some was juggling work with personal life while surrounded by ambient distractions. In particular, parents with younger children faced additional obstacles since the closure of schools meant they had to take care of the children during working hours (Gorlick, 2020). Restrictions on traveling and social gatherings made it harder to get a break from this difficult situation. Studies examining work-from-home arrangements during the pandemic indicate some negative consequences of such arrangements. Some workers find themselves spending more hours working than before the pandemic (DeFilippis et al., 2020, Teodorovicz et al., 2021). Even if the workers do not spend more total time working, they sometimes extend the workday by taking breaks during the day and working later at night (Teodorovicz et al., 2021). To make things worse, at the end of the long working day many workers struggle to unplug from work (Routley, 2020). This might indicate an erosion of boundaries between work and personal life, a situation that increases work-related physical and emotional exhaustion among workers (Palumbo, 2020). Workers also report challenges to concentrate while working from home because of the surrounding distractions. Household chores, noisy neighbors, and interruptions by children or pets make it difficult for remote workers to focus on work and might lead to lower productivity (Mark et al., 2017). Lack of social interactions and of new and exciting experiences might make the life of remote workers more monotonous, which in turn may impact the creativity of workers (Peppercorn, 2020). Despite facing many challenges, a recent survey found that the vast majority of workers want to have the option to continue working from home for the rest of their careers (Routley, 2020). They stated that a flexible schedule, the ability to work from anywhere, and no need to commute are the main reasons behind this choice (Routley, 2020). Many organizations are adopting a hybrid workplace model where employees can perform at least part of the work remotely. This helps the companies to attract top talent and reduce operational costs at the same time (Barbuto et al., 2020). Because more than one-third of the jobs in the United States can be performed entirely from home (Dingel and Neiman, 2020) and both employers and employees could benefit from work-from-home arrangements, it is clear that remote work arrangements will continue (Barrero et al., 2021). So it is crucial to explore ways to help remote workers deal with challenges presented by the COVID-19 pandemic in particular, and by working-from-home arrangements more generally. For this reason, we conducted a longitudinal field study focusing on the impact of daily virtual nature experiences combined with mindfulness practices on remote workers’ creativity, perceived stress, and focus over an extended period of time (9 weeks) during the Covid-19 pandemic. Over the years researchers have investigated the effect of nature exposure (real and virtual) and mindfulness practices on stress, focus, and creativity (Atchley et al., 2012, Palanica et al., 2019). However, most studies were cross-sectional and were conducted in controlled laboratory environments. In contrast, this article expands on the existing literature by reporting findings from a field study focusing on the long-term effects of daily at-home virtual nature experience combined with mindfulness practice on remote workers. Even though in such a study we lose some experimental control compared to cross-sectional laboratory experiments, our experimental design increases ecological validity by observing remote workers in real-life environments without risking their well-being. The general hypothesis guiding this study is that an at-home intervention, which combines exposure to nature through a VR headset with mindfulness practice (using a mobile app) enhances the creativity and focus of remote workers and reduces stress. In the following, we present results from the study we conducted, which show a positive effect of VR nature experiences on increasing focus and reducing stress. When VR and mindfulness practice were combined, we also found an increase in convergent thinking task performance. Our findings demonstrate that 10-20 min of daily exposure to VR nature and mindfulness practice could compensate for some of the negative effects of working remotely by reducing stress, enhancing focus, and convergent thinking. 2 Related work Well-being and productivity are two major considerations for effective remote work arrangements during the COVID-19 pandemic. Prior research suggests that stress, focus, and creativity are linked with the well-being and productivity of workers. Stress is an indicator of well-being (Lazarus and Folkman, 1984) and also a major concern for workers and organizations worldwide (World Health Organization et al., 2011). Difficulties in focusing on work can negatively affect productivity (Mark et al., 2017). The creativity of the workers is vital for solving personal and work-related problems. Working from home, especially during the COVID-19 pandemic, has been shown to increase workers’ stress levels (Galanti et al., 2021, Gómez et al., 2020). Frequent switching between work and personal life might make it harder for remote workers to focus on a task. Moreover, the uniformity of everyday life of remote workers can negatively impact their creativity (Peppercorn, 2020, Amabile, 1998). 2.1 Stress, focus, and creativity According to the World Health Organization (WHO), health is not just the absence of disease, but rather “a state of complete physical, mental and social well-being” (World Health Organization et al., 1995). This definition explicitly links well-being with health and underlines its importance. Over the years, researchers introduced various definitions and descriptions for well-being. Dodge et al. defined well-being as a balance between someone’s resource pool and the challenges they face (Dodge et al., 2012). When someone experiences an environment where their resources are not enough to meet the challenges they face, it endangers their well-being and the relationship between the person and the environment can be defined as stress (Lazarus and Folkman, 1984). Experiencing stress can cause both physical and psychological health problems (Byrne et al., 2007, Fink, 2016, Stansfeld et al., 1997). Stress is associated with sleep difficulties (Kim and Dimsdale, 2007, Waters et al., 1993, Lundh and Broman, 2000, Van Reeth et al., 2000), increased blood pressure (Gasperin et al., 2009, Kulkarni et al., 1998), and depression in some people (Van Praag, 2004, Hammen, 2005, Caspi et al., 2003). Previous studies also found evidence of deleterious effects of stress on the heart which may lead to cardiovascular diseases (Dimsdale, 2008, Roux, 2003, Steptoe and Brydon, 2009, Schnall et al., 1994). Stress can also be related to work. A poll conducted by European Agency for Safety and Health at Work (EU-OSHA) found that as much as 50% of workers in Europe find their workplace stressful (European Agency for Safety and Health at Work, 2011). Work-related stress is considered a major challenge for organizations around the world as it can negatively affect workers’ health and productivity (World Health Organization et al., 2011). Research suggests that high job-strain is associated with elevated blood pressure (Landsbergis et al., 1994), cardiovascular diseases (Johnson and Hall, 1988), musculoskeletal problems (Houtman et al., 1994), and psychiatric disorders (Stansfeld et al., 1997). Stress-related issues are among the leading causes of missing working days and retiring early from work (European Agency for Safety and Health at Work, 2011, World Health Organization et al., 2011). While well-being is an important consideration for remote workers, so is productivity. The ability to maintain focus while working from home impacts the productivity of the workers. Focus has been shown to depend on criteria such as activity involvement, time of day, and type of work (Mark et al., 2014). Research shows that workers are often interrupted during work and that switching focus between tasks may be difficult (Czerwinski et al., 2004, González and Mark, 2004). Increasing focus in the workplace increases productivity (Mark et al., 2017). Creativity is the capacity to produce original and useful ideas or work (Sternberg and Lubart, 1999). According to Runco & Jaeger, for an idea to be considered creative, it needs to be original, and it also needs to effectively solve the problem (Runco and Jaeger, 2012). Workers rely heavily on their creativity to deal with work-related challenges. Creativity is often assessed in terms of divergent and convergent thinking (Runco and Acar, 2012, Cropley, 2006). Divergent thinking is the capacity for exploring multiple potential answers or solutions to a given problem. People can come up with different valid answers which are novel and unusual for the same question (Cropley, 2006). On the other hand, convergent thinking is the capability to narrow in on a single answer to a given situation. Convergent thinking helps people find the correct or best solution for a clearly defined problem (Cropley, 2006). The emphasis here is on accuracy and how quickly one can find the solution. It is most effective in real-life situations where a correct solution already exists and needs to be worked out by applying logical reasoning (e.g. multiple choice tests). Measuring creativity is a challenging task (Baer and McKool, 2009, Batey et al., 2010, Kaufman et al., 2008). Over the years, researchers have adopted different approaches to measuring creativity (Said-Metwaly et al., 2017). This resulted in the development of various tools and assessment techniques (Torrance, 1966, Amabile, 1982, Guilford, 1968, Guilford, 1950, Mednick, 1968, Guilford, 1978, Urban, 2005). However, a review of different creativity tests shows that each test has some advantages and disadvantages and using multiple tests may provide a more accurate assessment of creativity (Cropley, 2000, Said-Metwaly et al., 2017). Studies show that experiencing unusual and unexpected events or situations can enhance creativity (Ritter et al., 2012, Martindale, 1972, Goertzel et al., 1978), while lack of interactions with coworkers and limited exposure to new situations can negatively affect the creativity of the workers (Peppercorn, 2020, Amabile, 1998). Thus, working from home, which can limit experiencing new situations as well as interactions with others, can undermine creativity. 2.2 Positive effects of nature One way to counteract the negative effects of working from home might be by spending time in nature. Spending time in nature can improve health, wellness, and creativity (Palanica et al., 2019). People who spend 120 min per week in nature reported a higher level of health and well-being (White et al., 2019). Nature helps reestablish our mental equilibrium, which in turn helps us recover from stress (Ulrich et al., 1991, Grahn and Stigsdotter, 2003). This process was named “environmental self-regulation” by Korpela (1989). Nature experience also helps people think less about the negative aspects of life (Bratman et al., 2015). Even watching three-dimensional videos of tree-covered streets (Jiang et al., 2016), driving along such roads (Parsons et al., 1998), or sitting in a room with tree views (Hartig et al., 2003) can help recover from stress faster. Spending time in nature has been shown to decrease symptoms of attention-deficit/hyperactivity disorder in children (Taylor et al., 2001), and lowers aggressive behaviors (Guite et al., 2006). People who spend more time in green spaces are less depressed and are less likely to have high blood pressure (Shanahan et al., 2016, Hartig et al., 2003). Exposure to nature can also increase concentration (Taylor et al., 2001) and enhance the ability to maintain focus (Berto, 2005). Studies show nature experience improves working memory (Berman et al., 2008) and helps people perform better on attention tests (Berman et al., 2008, Berto, 2005, Li and Sullivan, 2016). Attention Restoration Theory (ART) (Kaplan, 1995) emphasizes natural environments central to restoring attentional resources depleted by work. Even exposure to natural lights can improve the creativity and productivity of office workers (Peters, 2015). The benefits of experiencing nature can be gained through both short- and long-term exposure to nature. As short as a 40-second micro-break in nature can help people perform better in cognitive tests (Lee et al., 2015). On the other hand, 4 day long nature exposure can improve creativity and problem-solving task performance by as much as 50% (Atchley et al., 2012). Many of the therapeutic effects of being in nature do not require people to be in direct contact with nature. Observing nature through a window (Kaplan, 2001, Li and Sullivan, 2016) or looking at an artificial window (Radikovic et al., 2005), viewing pictures or videos of nature (Wooller et al., 2016, Jiang et al., 2016), experiencing nature through virtual reality (VR) (Yu et al., 2018, Li et al., 2021, Anderson et al., 2017, Liszio et al., 2018), and even listening to nature-based soundscapes (Newbold et al., 2017) can provide similar benefits. Researchers found that watching nature videos can increase connectedness to nature, positive emotions, and attentional ability (Mayer et al., 2009), and exposure to VR nature video can increase physiological arousal and improves positive mood levels (Browning et al., 2020). Even though in both studies the effects were more dramatic for actual nature experience, the benefits of virtual nature experience were also significant. Immersive experiences, such as virtual reality environments that include sounds and images, promote stress recovery more than only images of nature (Annerstedt et al., 2013) and also enhance creativity (Fleury et al., 2021). These studies demonstrate how exposure to nature can improve physical and mental health, productivity, and creativity. However, these studies did not focus on remote workers, and the effects of nature exposure were not evaluated over an extended period of time. Based on these studies we hypothesize that using VR headsets to experience nature scenes may negate some of the undesirable effects of working from home, especially for workers who have limited opportunities to go out in nature. 2.3 Positive effects of mindfulness In addition to nature experiences, one effective method to enhance well-being and creativity is mindfulness practice. Mindfulness can improve the well-being and productivity of remote workers by helping them establish a strict boundary between work and personal life (Toniolo-Barrios and Pitt, 2021). Over the years researchers defined mindfulness in various ways. Most of the definitions describe mindfulness as a state of mind in which a person is aware or paying attention to the present, both internally and to the external environment (Brown et al., 2007, Herndon, 2008, Kabat-Zinn, 2005, Dane, 2011). Bishop et al. defined mindfulness as “non-elaborate, nonjudgmental, present-centered awareness in which each thought, feeling, or sensation that arises are acknowledged and accepted as it is” (Bishop et al., 2004). Despite the inherent challenges of promoting and measuring mindfulness (Davidson and Kaszniak, 2015), there is substantial evidence of its positive effect on health, productivity, and creativity (Brown et al., 2007, Klainin-Yobas et al., 2016, Baer, 2003, Dane, 2011, Lebuda et al., 2016). Mindfulness practice can enhance psychological well-being (Klainin-Yobas et al., 2016), help processing one’s negative emotions (Shepherd and Cardon, 2009) and improve mood (Broderick, 2005). A clinical intervention study conducted by Brown and Ryan shows that increases in mindfulness can result in lower stress levels over time (Brown and Ryan, 2003). Since mindfulness can lead to a better attentional focus on the task, it is also associated with improved task performance (Kersemaekers et al., 2018, Good et al., 2016) in different types of jobs. Mindfulness has been linked with job performance among restaurant workers (Dane and Brummel, 2014), health care professionals (Beach et al., 2013), and managers working in leadership roles (Reb et al., 2014). A wide body of research also links mindfulness to creative thinking (Greenberg et al., 2012, Ostafin and Kassman, 2012). Mindfulness can improve the ability to solve insight problems (Ostafin and Kassman, 2012, Ren et al., 2011) and to find novel solutions to given problems (Greenberg et al., 2012). This indicates that mindfulness can affect creativity in terms of both convergent and divergent thinking. We are also seeing uses of different technologies to support mindfulness practice (Terzimehić et al., 2019) and the popularity of applications developed for this purpose (Daudén Roquet and Sas, 2018, Salehzadeh Niksirat et al., 2017) suggests that more and more people are realizing its benefits. These studies demonstrate that mindfulness has many benefits in terms of enhancing well-being, productivity, and creativity. Given the challenges remote workers are facing during the COVID-19 pandemic, it is important to examine if mindfulness can help them in similar ways. Our contributions focus on examining mindfulness practice as complementary to virtual nature experiences. 3 Method We conducted a longitudinal field study to explore how VR nature experiences combined with mindfulness practices affect the stress, focus, and creativity of remote workers during the COVID-19 pandemic. The dependent variables in this study are focus, stress, and creativity as measured by daily questionnaire data collected using text messages. The independent variables are engagement in nature experiences through VR (yes/no), and engagement in mindfulness practice (yes/no). In our nine-week study, the participants were asked to participate in each of the following three phases: • Weeks 1—3: No Intervention: Text check-in • Weeks 4—6: VR + No Mindfulness: 10 min of VR practice daily and text check-in • Weeks 7—9: VR + Mindfulness: 10 min VR + 10 min mindfulness practice daily and text check-in The first three weeks are for collecting baseline data, weeks four to six are for evaluating the effects of the VR nature experience, and the last three weeks are for evaluating the effects of the VR nature experience combined with mindfulness practice. We decided not to include a phase for only mindfulness practice in the study for two reasons: (1) the benefits of mindfulness are well documented. The expected results would not contribute new knowledge; (2) in order to measure the impact of mindfulness only (without VR), some participants would have had to give up either using VR for nature exposure or mindfulness practice in the later phases of the study. Since we conducted the study during the challenging time of COVID-19, we did not want the participants to give up a potentially helpful intervention. Instead, we decided to investigate whether combining mindfulness practice with VR nature exposure yields additional benefits compared to just experiencing nature through VR. The study was conducted between October 2020 and January 2021. Data from the study was used in another paper where researchers investigated the stability of an individual’s creativity over time (Katrahmani et al., 2022), but did not report on the effect of different interventions on stress, focus, and creativity. The study was approved by the University of New Hampshire Institutional Review Board (IRB). 3.1 Task The participants were asked to check in using text messages and answer questions (see Section 3.4) every weekday during the 9-week study. They also completed a survey before starting each 3-week phase. Fig. 1 shows the steps of the study. During the first three weeks of the study, participants were only required to check in via text messages. For the next three weeks, participants experienced nature through a VR headset. They were instructed to use an application (Guided Relaxation VR on Oculus Go Cubicle Ninjas, 2021) to experience nature scenes for 10 min every weekday and check in via texts after that. Each nature scene presented participants with the sights and sounds of a dynamically changing place in nature, such as a waterfall or a beach with waves (Fig. 2). During the weekdays of the final three weeks, participants experienced both natures through VR for about 10 min and practiced mindfulness for about 10 min. They then checked in via text messages. For practicing mindfulness, participants selected a mindfulness session using the HealthyMinds application (Healthy Minds Innovations, Inc, 2021). We provided the following options for mindfulness sessions in the HealthyMinds application for the participants to choose from: 1. Calm in the midst of chaos: This practice aims to help people to calm their minds by connecting to their inner resilience. 2. Clarity in uncertain times: This practice aims to help people to re-frame things with insight and appreciation. 3. A true break: This practice aims to help people to take a break by focusing on doing nothing at all. We did not track which session participants were choosing. The mindfulness practice sessions we selected in that application were similar in nature, where a speaker guided the participant through meditative practices focusing on breathing, awareness, etc. These mindfulness sessions were guided through audio only, no visuals were used. The sessions we selected were also similar in duration (about 10 min). Since the participants were practicing mindfulness daily for three weeks, giving them a list of options allowed them to try various mindfulness sessions instead of practicing the same session every day. The aim was to ensure that each mindfulness practice session for all the participants was similar in nature and duration. The HealthyMinds application can be installed on both Android and iOS systems. This application has been shown to enhance well-being in a clinical trial (Goldberg et al., 2020). After completing the 9-week study, participants were invited to participate in a semi-structured exit interview (via Zoom) regarding their experience in the study including the use of VR headset and mindfulness practice throughout the study. Fig. 1 Steps and timeline of the study. Fig. 2 Nature scenes participants used for virtual nature experience. 3.2 Procedure We conducted a screening survey that was distributed through popular social media platforms. The survey included questions about people’s living environment, work arrangements, experience with virtual reality, and general anxiety. The goal of the screening survey was to recruit participants who lived in urban cities, did not own a VR headset, and had a relatively low anxiety rating. After recruiting the participants, we distributed a video describing the study and asked them to set a time to meet with us via Zoom. Participants were required to watch the video before the meeting. At the meeting, we discussed the requirements for the study and answered questions. We sent an Oculus Go virtual reality headset to each participant. They received this headset towards the end of the first three weeks of the study and used it for the final six weeks. Before the start of each new 3-week phase, we sent instruction videos to participants explaining what they were expected to do during that period of the study. The videos included instructions about how to set up the VR headset, install applications, and select specific nature scenes, duration, and no background music for experiencing nature through VR. We also met (via Zoom) with the participants individually before the start of each 3-week phase to discuss their role in the study, troubleshoot, and answer questions. Participants could also reach us via email if they experienced any difficulties or had concerns. On each weekday during the study, participants received an automated text message in the morning reminding them to complete the daily tasks. The daily text check-ins were then initiated by the participants at their convenience, anytime during the day by sending a text message to a given number. We used Twilio (Twilio Inc, 2021), a cloud communications platform, to automate the participants’ text check-in process. After sending the initial text message, participants received questions via text messages regarding stress, focus, and creativity (see Section 3.4). After sending each question, the system waited for the participant to answer via text message before sending the next question. In addition to the daily reminder text, participants received another reminder text if they missed check-ins on two consecutive days. 3.3 Participants For this study, we recruited 20 remote workers from urban and suburban areas in five U.S. states: Massachusetts, Washington, New York, Illinois, and Mississippi. We selected people who do not own a VR headset, live in urban or suburban areas, and performed all or some of their work from home. The recruited people were less likely to experience nature in their day-to-day life, therefore more likely to be benefited from VR nature experience and mindfulness practice. Another selection criterion was a general anxiety score below the clinical threshold (not higher than 50) on the State-Trait Anxiety Inventory (STAI) measure (range 20–80) (Spielberger, 2010). Although anxiety is more likely a dimensional construct without strict cut-off points, this cut-off has nevertheless been used in other studies (Kim et al., 2010). Our screening survey did not include a question about gender, as we did not use gender as a selection criterion. Fifty-eight people applied (filled out the screening survey) to take part in the study. The first 20 people meeting the selection criteria were recruited to participate in the study (19 women and one man). Historically women participants are underrepresented in VR research (Peck et al., 2020, Stanney et al., 2020). Thus, our participant population presented an opportunity to contribute to VR literature by studying VR with more women participants. It is also important to consider the major impact of working from home during the pandemic specifically on women (Ibarra et al., 2020, Brower, 2021, Molla, 2021) and that anxiety is substantially higher in women than in men (Bekker and van Mens-Verhulst, 2007). This indicates that our study participants are directly included in the target audience for the interventions we investigate. Participants were between the ages of 25–68 (M = 43.45 years, SD = 12.85). Fourteen participants (out of 20 participants) also participated in a semi-structured exit interview at the end of the study. The participants kept the VR headset as compensation for the study. They also received a $15 Amazon gift card after completing each 3-week phase of the study. Those who participated in the exit interview received an additional $15 Amazon gift card. 3.4 Measures We tracked participants’ creativity and self-assessed stress and focus over nine weeks. The participants reported their focus by responding to the question: “On a scale of 1 to 7 (1 being lowest and 7 being highest), how focused are you on your current task?”. Similarly, to measure their stress we asked “On a scale of 1 to 7 (1 being very relaxed and 7 being very stressed), how stressed do you feel today?”. We used adapted versions of two different creativity measures, the Remote Associates Test (RAT) (Mednick, 1968) and the Alternate Uses Task (AUT) (Guilford, 1978), which measure participants’ convergent and divergent thinking ability, respectively. We chose these measures because they are well-validated and widely used in the creative problem-solving literature (Carroll, 1968, Colzato et al., 2012, Mednick, 1968). The RAT measures creativity in terms of the ability to make associations where the participants are given three words and they need to come up with a word that connects all three given words. An example of a RAT question we asked is “Look at the three given words and find a fourth word that is related to all three: fly/clip/wall”. In contrast, the AUT measures the ability to generate diverse ideas where the participants have to find as many uses for a given object as possible within a certain period of time. An example of an AUT question we asked is “Name all possible uses of the item mentioned below within 2 min. Please reply with only 1 use (1 or 2 words) per text message. Broom”. Both measures were adapted to fit a short, mobile-friendly version for the current study. 3.5 Data collection After the participants completed their daily activities, they were asked to check in via text messages using their smartphones. During the check-in process, they were asked four questions (see Section 3.4). The first two questions were intended to assess their stress and focus levels, and the last two questions were intended to assess their creativity in terms of convergent and divergent thinking. The timestamps and received text messages were stored in a cloud-based spreadsheet on Airtable (Airtable, 2021). In the recorded online exit interviews we asked the participants open-ended questions about their general experience with the study, virtual nature experience, and benefits to their well-being. To achieve an understanding of the virtual nature experience we asked some questions on the VR experience such as: Tell us about the VR experience. Did you enjoy it? Do you feel that it was beneficial? How? Did you experience any discomfort? How was the VR equipment? We also asked some questions about the combined VR and mindfulness experience- Tell us about the combined VR and mindfulness experience. Did you enjoy it? Do you feel that it was beneficial? How? Which one did you like more? (VR/mindfulness) and why? 3.6 Data processing During their daily text check-ins, participants reported their level of focus and stress with a number between 1 and 7 with 1 being the lowest level of focus and stress and 7 being the highest. For the RAT, their responses were scored 1 (correct) and 0 (incorrect). The response times were also calculated as the time difference between the timestamps of RAT text sent and received. For the AUT, participants’ responses were scored on fluency and originality. The fluency score represents how many ideas a participant generated for a task. To measure fluency, the total number of uses mentioned by the participant for a given object was calculated. The originality score represents how unique the generated ideas are. To find the originality score, at first, the semantic similarity between the given object and each use named by the participant was calculated using spaCy (Honnibal and Montani, 2017), a library for natural language processing. Semantic similarity scores range from −1 to 1, with higher scores indicating higher similarity and therefore lower originality. The semantic distance score was calculated by subtracting the semantic similarity score from 1 which produced originality scores ranging from 0 to 2 with higher scores indicating higher originality. Maximum semantic distance scores were calculated from all the uses mentioned by a participant for a given object. 3.7 Data analysis 3.7.1 Daily check-ins For analyzing the daily check-ins data, we used linear mixed models (using the open-source R package lme4 Bates et al., 2015) to assess the relationship between intervention types and self-reported level of focus and stress, RAT response time, and AUT-related measures (total number of uses, maximum semantic distance). Due to the binomial distribution of RAT response correctness, we used a logistic mixed model to assess the relationship between RAT response correctness and intervention types. For all the models, interventions were included as a fixed effect, and participants were included as a random effect. Day number was treated as a competing exposure and was included as a fixed effect to control for its effect on dependent variables. An interaction term between intervention and day number was included; however, we note that we only include day number and interaction term in our models to control for the effects of time in the study. That is, including both the day number and the interaction term will help us understand if the intervention types have a main effect that explains unique variance above and beyond time in the study, but we do not have sufficient power to investigate any potential interactions in the current work (see Section 6 for more on this). Interventions and participants were treated as a factor and day were treated as a numeric. All the models are presented in Table 1. For each model, we performed the analysis of variance using Chi-square tests. We also compared each pair of intervention types by comparing estimated marginal means using the Tukey method. A comparison of intervention types is shown in Table 2. 3.7.2 Exit interviews Two researchers first reviewed the autogenerated transcripts of the recorded interviews. Corrections were made where necessary. The researchers individually read through the transcripts and each generated a code-book with themes from the recorded responses. The themes from the separate code books were discussed as a team and were consolidated into 17 codes. The researchers independently coded all data, and then discussed the transcripts for participants with the least inter-coder agreement. Cohen’s kappa was calculated with 97% inter-coder reliability. For quantifying the exit interview responses, we used a binary coding system (1/0) to indicate the presence or absence of a particular effect. For example, in the category positive VR experience, a rating of ‘1’ by a coder would indicate a participant had a positive VR experience while a rating of ‘0’ would indicate that the participant did not have a positive VR experience. Table 1 Parameters of all the mixed-effects models (the estimates of the models are conditional effects corresponding to day=0). Creativity Convergent thinking (RAT) Divergent thinking (AUT) Focus (Scale 1–7) Stress (Scale 1–7) Correctness probability (logit) Response time (Sec) Total number of uses (Fluency) Max. semantic distance (Originality) Predictors Estimates 95% CI Estimates 95% CI Estimates 95% CI Estimates 95% CI Estimates 95% CI Estimates 95% CI (Intercept) No intervention 4.39 [3.82, 4.96] 4.14 [3.68, 4.59] 0.45 [−0.38, 1.28] 47.25 [28.30, 66.21] 9.20 [7.87, 10.53] 0.82 [0.79, 0.85] VR only 0.13 [−0.29, 0.55] −0.97 [−1.42, −0.52] −0.54 [−1.43, 0.35] −5.6 [−26.74, 15.55] −0.49 [−1.4, 0.38] −0.02 [−0.06, 0.02] VR+Mindfulness 0.29 [−0.08, 0.65] −0.86 [−1.25, −0.47] 0.41 [−0.41, 1.22] −11.73 [−30.25, 6.78] −2.16 [−2.93, −1.39] 0.04 [0.007, 0.07] Day −0.01 [−0.03, 0.01] −0.04 [−0.07, −0.02] −0.05 [−0.09, 0] 0.3 [−0.78, 1.38] −0.05 [−0.1, −0.007] 0.0002 [−0.002, 0.002] VR only:Day 0.01 [−0.02, 0.05] 0.06 [0.02, 0.09] 0.06 [−0.01, 0.14] −0.03 [−1.77, 1.72] −0.03 [−0.1, 0.05] 0.001 [−0.002, 0.004] VR+Mindfulness:Day 0.006 [−0.03, 0.04] 0.03 [−0.006, 0.06] 0.08 [0.003, 0.15] −0.98 [−2.59, 0.64] 0.10 [0.04, 0.17] −0.003 [−0.006, −0.0002] Random effects σ σ σ σ σ σ Participants 1.18 0.84 1.44 32.38 2.79 0.04 Residual 1.1 1.18 55.54 2.31 0.1 Table 2 Weighted mean and standard deviation of measures for each intervention and pairwise comparison of estimated means for intervention types. No Intervention VR only VR+Mindfulness No Intervention - VR only No Intervention - VR+Mindfulness VR only - VR+Mindfulness Mean (SD) Mean (SD) Mean (SD) Effect size (Cohen’s d) p Effect size (Cohen’s d) p Effect size (Cohen’s d) p Focus (Scale 1–7) 4.26 (1.05) 4.54 (1.31) 4.64 (1.41) 0.23 0.025 0.31 0.001 0.07 0.70 Stress (Scale 1–7) 3.72 (0.97) 3.31 (0.92) 3.1 (1.11) 0.43 0.001 0.60 <0.001 0.21 0.21 Convergent thinking (RAT) Correctness probability 0.5 0.51 0.69 0.89 <0.001 <0.001 Response time (sec) 49.98 (42.92) 45.02 (33.99) 30.06 (28.7) 0.13 0.48 0.55 <0.001 0.48 0.008 Divergent thinking (AUT) Total number of uses (fluency) 8.66 (2.87) 7.79 (2.82) 7.65 (3.25) 0.31 <0.001 0.33 <0.001 0.05 0.31 Max semantic distance (originality) 0.82 (0.05) 0.81 (0.05) 0.83 (0.05) 0.13 0.79 0.25 0.62 0.39 0.29 4 Results We recorded a total of 812 check-ins out of the 900 expected check-ins during the 9 weeks of the study (20 participants × 9 weeks × 5 days a week). After cleaning the data and removing the non-relevant data points (non-relevant answers, answers during weekends, incomplete check-ins, etc.), 749 data points were extracted for analysis purposes. 4.1 Stress We found a significant effect of intervention type on the stress level of the participants (χ2=32.08, p<0.001). The control variables were also significant: we found a significant main effect of the day (χ2=5.95,p=0.015) as well as the interaction between day and intervention type (χ2=9.5,p=0.0087). Post-hoc tests for the main effect of interest showed that VR nature experience and VR+mindfulness practice reduced stress compared to no intervention. Further, Cohen’s effect size value (d=0.43&d=0.60) suggested low to moderate practical significance. However, the addition of mindfulness practice did not reduce stress significantly compared to only experiencing nature in VR (Table 2), thus suggesting that both intervention types involving VR had some benefit to stress. 4.2 Focus We found that intervention type was a significant factor for predicting the focus level of the participants (χ2=14.13, p<0.001). This indicates that the focus level of the participants was influenced by the interventions. We did not find a significant main effect of day and also no interaction between day and intervention. Post-hoc tests showed that VR nature experience and VR+mindfulness practice improved focus levels compared to no intervention. Further, Cohen’s effect size value (d=0.23&d=0.31) suggested small practical significance. However, the addition of mindfulness practice did not improve focus levels significantly compared to just using VR for experiencing nature (Table 2). 4.3 Creativity assessment We explored the effects of different interventions on participants’ creativity in terms of convergent and divergent thinking. 4.3.1 Convergent thinking To assess participants’ convergent thinking, we analyzed the response time and correctness probability of their reply to the remote associates test (RAT). We found that the intervention type was a significant factor for predicting the probability of getting the correct answer to a RAT question (χ2=28.74, p<0.001) as well as for the response time (χ2=19.83, p<0.001). This indicates that convergent thinking was influenced by the interventions. We did not find a significant main effect of day or interaction between day and intervention type. Post-hoc pairwise comparisons revealed that using VR for nature experience did not significantly improve convergent thinking ability (both in terms of increasing the probability of coming up with the correct answer and reducing the response time) but VR+mindfulness practice improved convergent thinking ability compared to no intervention and only experiencing nature through VR (Table 2). Cohen’s effect size value (d=0.55&d=0.48) suggested small to medium practical significance for change in response time. 4.3.2 Divergent thinking We assessed participants’ divergent thinking ability in terms of fluency (number of uses) and originality (semantic distance) of their replies to the alternate uses task. We found that intervention type was a significant factor for predicting the number of uses (χ2=34.44, p<0.001) but not the maximum semantic distance. This indicates that the interventions influenced the divergent thinking ability of the participants in terms of fluency but not originality. In terms of control variables, we did not find a significant main effect of day, but we did observe a significant interaction between day and intervention type for the total number of uses (χ2=14.18, p<0.001) and maximum semantic distance (χ2=8.02,p=0.02). Post-hoc analysis of the main effect of intervention type indicated that the number of uses decreased significantly for both experiencing nature through VR and VR+mindfulness practice compared to no intervention, but no significant difference was observed between the intervention types for maximum semantic distance (Table 2). Cohen’s effect size value (d=0.31&d=0.33) suggested small practical significance for change in the number of uses (fluency). In sum, divergent thinking ability did not change significantly in terms of originality (semantic distance) between intervention types, but it actually worsened in terms of fluency (number of uses) for VR nature experience and VR+mindfulness practice. 4.4 Participant experience and perceived benefits In this section, we discussed our findings from the qualitative analysis of the post-study (exit) interviews with 14 participants. 4.4.1 VR experience The VR experience of the participants was coded based on three categories: the effect of the VR equipment on the experience, the experience with the VR nature scene, and the effect of the study structure on the VR experience. When asked about the VR experience, 71%(10/14) participants described having a positive experience with the VR headset describing the experience using terms such as relaxing, calming, and ability to travel. For example, participant 12 mentioned that “for me being able to be in the scene and really be able to be present for ten minutes was great, but I loved, I loved VR”. While 29%(4/14) had negative experiences due to factors such as technical difficulties in setting up the headset, the repetitive nature scene, and not being able to choose their own nature scenes. 64%(9/14) participants had a positive experience with the VR nature scenes; participant 12 shared: “I loved all of the scenes, especially the water and just the peacefulness and serenity”, while 29%(4/14) had a negative experience. Participant nine describes it as “the clouds they’re not moving and they were they were just like little specks on white things” and participant 10 also described the scenes as “some sort of streams dystopian science fiction world in which you’re in nature, but it’s never going to change”. For 29%(4/14) of participants, the study structure contributed to a negative experience. For example, participant two expressed this sentiment by saying: “I would like to be able to choose. I think if I had been able to choose from one or two, from two or three different scenes every day? I think the variety would have kept me more interested”. 4.4.2 Perceived VR benefits During the exit interview, participants were asked if they experienced any benefits from the VR task. 64%(9/14) of participants reported feeling ‘calm’, ‘relaxed’, or ‘refreshed’ as a benefit of the VR nature experience and 50%(7/14) described the VR nature benefits as ‘an escape’. For example, participant 14 stated, “I felt that calming and soothing, VR is kind of a nice way to escape out of the walls of your home”. Some participants also shared that the VR experience allowed them to feel focused as participant 13 describes it, “I would feel just more relaxed and focused”. 4.4.3 VR plus mindfulness experience The VR plus mindfulness experience was coded into two main categories: Whether participants liked the combination of VR and Mindfulness and whether they favored one over the other. 50%(7/14) of participants liked the combination of VR and Mindfulness. Participants with a positive combined experience shared that the VR experience prepared them for the mindfulness practice. Participant 3 shared “having the nature first, let me really like relax into it and be in a better place to do the guided meditation”. Factors contributing to a negative experience included the added time for the combined tasks and some participants preferring one practice over the other. Participant 13 shared, “I’d say the biggest challenge of that was just the added time. Because like I said, I mean 20 min is not that much, but I, like, it’s funny how hard it is to find 20 min of time that I can just like completely be for myself uninterrupted”. Participant one shared: ”I could have done without the VR part of it. And the length of time was, it was hard to fit that into my schedule”. Despite the added time, some participants experienced a benefit of the combined experience as described by participant 14: “honestly, before I would do it, I’d be like, Oh, I don’t really have time to do this or I’m not in the mood to do this but then once you get into it, it just kinda, kinda resets like hitting the reset button, I guess”. 21%(3/14) participants preferred mindfulness to VR and 21%(3/14) preferred VR to mindfulness. 4.4.4 Perceived benefits of VR plus mindfulness The perceived benefits of VR and mindfulness were coded according to participants feeling ‘calm’ or ‘relaxed’ and feeling ‘focused’ or ‘centered’. 86%(12/14) of participants reported feeling more calm and relaxed; “they both relaxed me”, participant nine described, as well as participant eight who described feeling “grounded” and “I guess a little bit refreshed”. 50%(7/14) of participants reported feeling more focused or centered after the combined experience. “It made me feel more focused, ready for the day at hand and the challenges that it was going to bring. And again, I mentally could check off that I did something to try to center myself a little bit before the start of an unpredictable day, which most of December was anyway”, participant seven shared. 5 Discussion This remote nine-week study examined how experiencing nature through a virtual reality device, and then also engaging in mindfulness practices, affected the focus, stress, and creativity of remote workers. The results showed that the VR nature experience increased participants’ focus. However, adding mindfulness practices to the VR nature experience did not provide additional benefits. In addition to improving focus, the study also revealed that the VR nature experience decreased participants’ stress levels. Similarly, adding mindfulness did not affect the stress level significantly highlighting the potential benefits of VR nature experiences on their own. In general, we can conclude that a 10-minutes daily experience of nature through a VR headset may increase focus and reduce stress for remote workers over time. These findings were further supported by participants’ responses during their exit interviews. Most of the participants stated that the VR nature experience helped them to relax and some of them thought of it as an escape from the situation they are in. In terms of creativity, we assessed both convergent and divergent thinking using adapted versions of the Remote Associates Test (RAT) and Alternative Uses Task (AUT), respectively. Results suggested that VR experiences did not improve convergent thinking. However, adding a daily mindfulness practice to the VR experience improved convergent thinking both in terms of the rate of getting correct answers and decreased response time. In contrast, neither the VR experience nor the mindfulness practice improved divergent thinking, as assessed by fluency and originality. The fluency scores were lower during the VR nature experience and combined VR mindfulness phase. This suggests that the same interventions may have different effects on different aspects of creativity. Further, this may be an artifact of our creativity measures since we adapted two measures that are more traditionally used in single-session experiments rather than over multiple months. 5.1 Implications for designing interventions Our results are encouraging for a number of reasons. The brief daily VR and mindfulness interventions were well-received by participants, who remained engaged in the study for nine weeks, and told us in exit interviews that the interventions provided them with positive experiences. This is indeed good news because we expect that many workers will continue to work from home, partly because many firms see a number of advantages in working-from-home arrangements (Neeley, 2021, Barrero et al., 2021). The workers of these firms are likely to enjoy not having to spend time commuting (Kun et al., 2020), but they could still be at risk of not experiencing nature or interacting with co-workers often enough. Simple interventions like the ones we tested can be effective in supporting the general well-being of these remote workers, as well as their work productivity. Designers can also use our results as an indication that VR nature experiences can reduce stress and improve focus for remote workers. Of course, we only tested one implementation of VR nature experiences, and all of our participants engaged in these experiences in a uniform way, spending a few minutes on them each workday. Additional research can shed light on how to design VR experiences for different workers, where relevant variables to test include types of nature experiences (e.g. walking through nature vs. observing static or moving images), VR quality, and time spent in VR. In addition, some of our participants expressed a desire to control the details of interventions, like the time and duration of the intervention, selecting virtual nature scenes, and mindfulness practice options. It would be important to assess how personal user preferences, and the ability to control different aspects of the interventions, affect the outcomes of the interventions. An additional important implication for design is that any particular intervention might have different effects on creativity depending on whether it is related to convergent or divergent thinking. Tools and techniques developed for enhancing the creativity of workers might be more useful if focused on enhancing either convergent or divergent thinking based on the types of jobs or tasks on hand. We want to note that our decision to trade off statistical power (relatively small sample size) for increased ecological validity and longer-term observation allowed us to collect data under real-world conditions, over a relatively long period of time (9 weeks). Our study became part of our participants’ everyday routines for multiple weeks. This gives us confidence that designs that are based on our implications could be well received by users, as well as applied in work-from-home settings using commercially available VR headsets. As we discussed in the related work section, there is already substantial evidence for the therapeutic effects of both real and virtual nature exposure and the results from our study extend these results by demonstrating that daily brief (10 min) interventions over an extended period of time are sufficient for positive impact on stress, focus, and some aspects of creativity during a time of crisis like the COVID-19 pandemic. 6 Limitations Conducting longitudinal VR studies remotely is challenging (Ratcliffe et al., 2021a, Ratcliffe et al., 2021b). Even though on average, participants were highly compliant (daily task completion rate dropped below 70% for only one day of the study), there were some issues that may have impacted the study. One issue was that maintaining experimental control was a challenge. In the instructions for the participants we tried to specify all the factors that may influence the study (nature scenes, background music, duration, etc.), but we could not verify that all the participants were following these instructions. Another challenge was that our study was conducted in the US in a time frame that included the US presidential election, Christmas, Thanksgiving, New Year’s Eve, and varying COVID-19 restrictions. All these events may have affected the participants differently, which may have introduced noise in the data. We used self-reported measures instead of standard questionnaires or physiological measures for assessing perceived stress and focus. We chose this approach because it made it easier for participants to report their perceived focus and stress level using simple text messages on a daily basis for nine weeks. This experimental design decision indicates a trade-off between using validated measures and lowering barriers for participants to report data on a daily basis. Our sample size was relatively small (n = 20). The fact that we see significant results despite having relatively low power suggests that the observed effects are robust for practical purposes. Some of the null effects may be significant if replications are completed with a larger sample, but this does not detract from the overall findings presented. We also observed some interactions with Day in the study that we did not follow up on within the current paper due to this limitation in sample size. Future work, however, should be designed with power in mind so that we can ensure that our null results are not a type 2 error and other factors like the age of participants and effects of time can be studied more intentionally. Moreover, the stability of our creativity measures within individuals over a longer period of time needs further investigation as well, particularly since these were adapted for quick, mobile-friendly usage (Baer, 1994, Magnusson and Backteman, 1978, Katrahmani et al., 2022). For example, a recent study found that the two measures of creativity used here are quite variable over time, indicating that these measures are not simply tracking a stable individual difference measure (Katrahmani et al., 2022). This variability suggests that we can expect to see some fluctuations in creativity scores (at least for these tasks) over time and that they may be influenced by various situational and cognitive-affective factors. However, we also note that it may be worthwhile using other creativity tasks that are specifically designed for longitudinal data collection, or ones that are more directly related to workplace creativity. Another limitation of the study was that the order of interventions was not counterbalanced. In order to do that, some participants would have had to give up using VR headsets or mindfulness practice, or both in the later phases of the study. Since the pandemic already put a lot of mental pressure on people, we did not want to ask the participants to give up practices that might help them cope with the added stress. For this reason, we could not have both VR-only and Mindfulness only phases in our study. Since many researchers have investigated the effect of mindfulness practices on creativity and well-being (Brown et al., 2007, Keng et al., 2011), we decided not to include a phase in which the participants only focused on mindfulness practices without using the VR. For our analysis, we did not consider the novelty effect of using VR technology. Even though none of the participants owned a VR headset, we expect the novelty effect to be small because of repeated exposure to VR technology throughout the study (Huang et al., 2021). An additional limitation of our study is that almost all of our participants were women (19/20), which may limit the generalization of our results. Additional work is necessary to assess how different people in different contexts would be able to benefit from VR nature experiences and mindfulness practices. However, considering the major impact of working from home during the pandemic specifically on women (Ibarra et al., 2020, Brower, 2021, Molla, 2021) and that anxiety is substantially higher in women than in men (Bekker and van Mens-Verhulst, 2007), we believe it is important to consider such interventions with population that is likely to benefit from them. It is also important to note that considering women are underrepresented in VR research (Peck et al., 2020, Stanney et al., 2020), we believe that the contribution of the study is important even if it cannot be generalized to all workers. Despite these limitations, this study contributes to exploring how exposure to VR nature scenes and mindfulness practice can improve the focus, stress, and creativity of remote workers. 7 Conclusion In conclusion, our study shows that experiencing nature through VR for as little as 10 min a day can help people focus on a task and also reduce stress, but it may not have any effect on convergent creative thinking. Adding mindfulness practice to the nature VR experience, may not affect focus or stress, but it can improve convergent creative thinking. Both VR nature experience and mindfulness practice may diminish some aspects of divergent thinking (fluency) and may not have any effects on other aspects (originality). Taken together, findings from this study can help researchers explore different ways in which VR interventions could be used to improve the well-being of remote workers. CRediT authorship contribution statement Nabil Al Nahin Ch: Conceptualization, Data curation, Writing, Data analysis. Alberta A. Ansah: Conceptualization, Data curation, Data analysis. Atefeh Katrahmani: Conceptualization, Writing. Julia Burmeister: Conceptualization. Andrew L. Kun: Conceptualization, Supervision. Caitlin Mills: Conceptualization, Supervision. Orit Shaer: Conceptualization, Supervision. John D. Lee: Conceptualization, Supervision. Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Andrew L. Kun, Orit Shaer, John D. Lee reports financial support was provided by National Science Foundation. Data availability Data will be made available on request. Acknowledgments This work was in part supported by NSF grants CMMI- (University of New Hampshire: 1840085; Wellesley College: 1840031; University of Wisconsin-Madison : 1839484). ==== Refs References Airtable Airtable 2021 https://airtable.com. (Online Accessed 16 January 2022) Amabile T.M. Social psychology of creativity: A consensual assessment technique J. Personal. Soc. Psychol. 43 1982 997 Amabile T.M. How to kill creativity 1998 (Online Accessed 3 August 2021) Anderson A.P. Mayer M.D. Fellows A.M. Cowan D.R. Hegel M.T. Buckey J.C. Relaxation with immersive natural scenes presented using virtual reality Aerosp. Med. Hum. Perform. 88 2017 520 526 28539139 Annerstedt M. Jönsson P. Wallergård M. Johansson G. Karlson B. Grahn P. Hansen Å.M. Währborg P. Inducing physiological stress recovery with sounds of nature in a virtual reality forest — Results from a pilot study Physiol. Behav. 118 2013 240 250 10.1016/j.physbeh.2013.05.023 23688947 Atchley R.A. Strayer D.L. Atchley P. Creativity in the wild: Improving creative reasoning through immersion in natural settings PLoS One 7 2012 e51474 Baer J. Performance assessments of creativity: Do they have long-term stability? Roeper Rev. 17 1994 7 11 Baer R.A. Mindfulness training as a clinical intervention: A conceptual and empirical review Clin. Psychol. Sci. Pract. 10 2003 125 Baer J. McKool S.S. Assessing creativity using the consensual assessment technique Handbook of Research on Assessment Technologies, Methods, and Applications in Higher Education 2009 IGI Global 65 77 Barbuto A. Gilliland A. Peebles R. Rossi N. Shrout T. Telecommuting: Smarter workplaces 2020 Barrero J.M. Bloom N. Davis S.J. Why Working from Home Will Stick: Technical Report 2021 National Bureau of Economic Research Bates D. Mächler M. Bolker B. Walker S. Fitting linear mixed-effects models using lme4 J. Stat. Softw. 67 2015 1 48 10.18637/jss.v067.i01 Batey M. Chamorro-Premuzic T. Furnham A. Individual differences in ideational behavior: Can the big five and psychometric intelligence predict creativity scores? Creat. Res. J. 22 2010 90 97 Beach M.C. Roter D. Korthuis P.T. Epstein R.M. Sharp V. Ratanawongsa N. Cohn J. Eggly S. Sankar A. Moore R.D. A multicenter study of physician mindfulness and health care quality Ann. Fam. Med. 11 2013 421 428 24019273 Bekker M.H.J. van Mens-Verhulst J. Anxiety disorders: Sex differences in prevalence, degree, and background, but gender-neutral treatment Gend. Med. 4 2007 S178 S193 18156102 Berman M.G. Jonides J. Kaplan S. The cognitive benefits of interacting with nature Psychol. Sci. 19 2008 1207 1212 19121124 Berto R. Exposure to restorative environments helps restore attentional capacity J. Environ. Psychol. 25 2005 249 259 Bishop S.R. Lau M. Shapiro S. Carlson L. Anderson N.D. Carmody J. Segal Z.V. Abbey S. Speca M. Velting D. Mindfulness: A proposed operational definition Clin. Psychol. Sci. Pract. 11 2004 230 241 Bloom N. Liang J. Roberts J. Ying Z.J. Does working from home work? Evidence from a Chinese experiment Q. J. Econ. 130 2015 165 218 Bratman G.N. Hamilton J.P. Hahn K.S. Daily G.C. Gross J.J. Nature experience reduces rumination and subgenual prefrontal cortex activation Proc. Natl. Acad. Sci. 112 2015 8567 8572 26124129 Broderick P.C. Mindfulness and coping with dysphoric mood: Contrasts with rumination and distraction Cogn. Ther. Res. 29 2005 501 510 Brower T. Women and the pandemic: Serious damage to work, health and home demands response Forbes 2021 (Online Accessed 9 September 2021) Brown K.W. Ryan R.M. The benefits of being present: Mindfulness and its role in psychological well-being J. Personal. Soc. Psychol. 84 2003 822 Brown K.W. Ryan R.M. Creswell J.D. Mindfulness: Theoretical foundations and evidence for its salutary effects Psychol. Inq. 18 2007 211 237 Browning M.H. Mimnaugh K.J. van Riper C.J. Laurent H.K. LaValle S.M. Can simulated nature support mental health? Comparing short, single-doses of 360-degree nature videos in virtual reality with the outdoors Front. Psychol. 10 2020 2667 32010003 Byrne D.G. Davenport S.C. Mazanov J. Profiles of adolescent stress: The development of the adolescent stress questionnaire (ASQ) J. Adolesc. 30 2007 393 416 16750846 Carroll J.B. Reviews: Guilford, JP the nature of human intelligence. New York: McGraw-Hill, 1967. 538 + xii pp. $14.75 Am. Educ. Res. J. 5 1968 249 256 Caspi A. Sugden K. Moffitt T.E. Taylor A. Craig I.W. Harrington H. McClay J. Mill J. Martin J. Braithwaite A. Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene Science 301 2003 386 389 12869766 Colzato L.S. Szapora A. Hommel B. Meditate to create: The impact of focused-attention and open-monitoring training on convergent and divergent thinking Front. Psychol. 3 2012 116 22529832 Cropley A.J. Defining and measuring creativity: Are creativity tests worth using? Roeper Rev. 23 2000 72 79 Cropley A. In praise of convergent thinking Creat. Res. J. 18 2006 391 404 Cubicle Ninjas A. Guided relaxation VR 2021 https://www.oculus.com/experiences/go/2052864981411233/?ranking_trace=0_2052864981411233_SEARCH_5db56e2a-98ed-49be-b600-0cb025a42c43. (Online Accessed 26 May 2021) Czerwinski, M., Horvitz, E., Wilhite, S., 2004. A diary study of task switching and interruptions. In: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. pp. 175–182. Dane E. Paying attention to mindfulness and its effects on task performance in the workplace J. Manag. 37 2011 997 1018 Dane E. Brummel B.J. Examining workplace mindfulness and its relations to job performance and turnover intention Hum. Relat. 67 2014 105 128 Daudén Roquet, C., Sas, C., 2018. Evaluating mindfulness meditation apps. In: Extended Abstracts of the 2018 CHI Conference on Human Factors in Computing Systems. pp. 1–6. Davidson R.J. Kaszniak A.W. Conceptual and methodological issues in research on mindfulness and meditation Am. Psychol. 70 2015 581 26436310 DeFilippis E. Impink S.M. Singell M. Polzer J.T. Sadun R. Collaborating During Coronavirus: The Impact of COVID-19 on the Nature of Work: Technical Report 2020 National Bureau of Economic Research Dimsdale J.E. Psychological stress and cardiovascular disease J. Am. Coll. Cardiol. 51 2008 1237 1246 18371552 Dingel J.I. Neiman B. How many jobs can be done at home? J. Public Econ. 189 2020 104235 Dodge R. Daly A.P. Huyton J. Sanders L.D. The challenge of defining wellbeing Int. J. Wellbeing 2 2012 European Agency for Safety and Health at Work R. Psychosocial Risks and Stress at Work 2011 European Agency for Safety and Health at Work Retrieved from: https://osha.Europa.Eu/En/Themes/Psychosocial-Risks-and-Stress Fink G. Stress, definitions, mechanisms, and effects outlined: Lessons from anxiety Stress: Concepts, Cognition, Emotion, and Behavior 2016 Elsevier 3 11 Fleury S. Blanchard P. Richir S. A study of the effects of a natural virtual environment on creativity during a product design activity Think. Ski. Creat. 40 2021 100828 Galanti T. Guidetti G. Mazzei E. Zappalà S. Toscano F. Work from home during the COVID-19 outbreak: The impact on employees’ remote work productivity, engagement, and stress J. Occup. Environ. Med. 63 2021 e426 Gasperin D. Netuveli G. Dias-da Costa J.S. Pattussi M.P. Effect of psychological stress on blood pressure increase: A meta-analysis of cohort studies Cad. Saude Publica 25 2009 715 726 19347197 Goertzel M.G. Goertzel V. Goertzel T.G. 300 Eminent Personalities: A Psychosocial Analysis of the Famous 1978 Jossey-Bass San Francisco Goldberg S.B. Imhoff-Smith T. Bolt D.M. Wilson-Mendenhall C.D. Dahl C.J. Davidson R.J. Rosenkranz M.A. Testing the efficacy of a multicomponent, self-guided, smartphone-based meditation app: Three-armed randomized controlled trial JMIR Ment. Health 7 2020 e23825 Gómez S.M. Mendoza O.E.O. Ramírez J. Olivas-Luján M.R. Stress and myths related to the COVID-19 pandemic’s effects on remote work Manag. Res.: J. Iberoam. Acad. Manag 2020 González, V.M., Mark, G., 2004. “Constant, constant, multi-tasking craziness” managing multiple working spheres. In: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. pp. 113–120. Good D.J. Lyddy C.J. Glomb T.M. Bono J.E. Brown K.W. Duffy M.K. Baer R.A. Brewer J.A. Lazar S.W. Contemplating mindfulness at work: An integrative review J. Manag. 42 2016 114 142 Gorlick A. The productivity pitfalls of working from home in the age of COVID-19 Stanf. News 2020 Grahn P. Stigsdotter U.A. Landscape planning and stress Urban For. Urban Green. 2 2003 1 18 Greenberg J. Reiner K. Meiran N. “Mind the trap”: Mindfulness practice reduces cognitive rigidity PLoS One 7 2012 e36206 Guilford J.P. Creativity. American psychology 1950 Guilford J.P. Creativity, Intelligence, and Their Educational Implications 1968 EDITS/Knapp San Diego, CA Guilford J.P. Alternate Uses 1978 Sheridan supply Company Guite H.F. Clark C. Ackrill G. The impact of the physical and urban environment on mental well-being Public Health 120 2006 1117 1126 17097120 Hammen C. Stress and depression Annu. Rev. Clin. Psychol. 1 2005 293 319 17716090 Hartig T. Evans G.W. Jamner L.D. Davis D.S. Gärling T. Tracking restoration in natural and urban field settings J. Environ. Psychol. 23 2003 109 123 Healthy Minds Innovations, Inc T. Healthy minds program 2021 https://apps.apple.com/us/app/healthy-minds-program/id1326310617#?platform=iphone. (Online Accessed 26 May 2021) Herndon F. Testing mindfulness with perceptual and cognitive factors: External vs. internal encoding, and the cognitive failures questionnaire Personal. Individ. Differ. 44 2008 32 41 Honnibal, M., Montani, I., 2017. spaCy 2: Natural language understanding with Bloom embeddings, convolutional neural networks and incremental parsing. 7, 411–420 (in press). Houtman I.L.D. Bongers P.M. Smulders P.G.W. Kompier M.A.J. Psychosocial stressors at work and musculoskeletal problems Scand. J. Work Environ. Health 1994 139 145 8079136 Huang W. Roscoe R.D. Johnson-Glenberg M.C. Craig S.D. Motivation, engagement, and performance across multiple virtual reality sessions and levels of immersion J. Comput. Assist. Learn. 37 2021 745 758 Hunter P. Remote working in research: An increasing usage of flexible work arrangements can improve productivity and creativity EMBO Rep. 20 2019 e47435 Ibarra H. Gillard J. Chamorro-Premuzic T. Why WFH isn’t necessarily good for women Harv. Bus. Rev. 7 2020 Jiang B. Li D. Larsen L. Sullivan W.C. A dose-response curve describing the relationship between urban tree cover density and self-reported stress recovery Environ. Behav. 48 2016 607 629 Johnson J.V. Hall E.M. Job strain, work place social support, and cardiovascular disease: A cross-sectional study of a random sample of the Swedish working population Am J Public Health 78 1988 1336 1342 3421392 Kabat-Zinn J. Coming to Our Senses: Healing Ourselves and the World Through Mindfulness 2005 Hachette UK Kaplan S. The restorative benefits of nature: Toward an integrative framework J. Environ. Psychol. 15 1995 169 182 Kaplan R. The nature of the view from home: Psychological benefits Environ. Behav. 33 2001 507 542 Katrahmani A. Ansah A.A. Ch N. Burmeister J. Kun A.L. Mills C. Shaer O. Lee J.D. Proceedings of the Human Factors and Ergonomics Society Annual Meeting Vol. 66 2022 SAGE Publications Sage CA Los Angeles, CA 2290 2294 Kaufman J.C. Baer J. Cole J.C. Sexton J.D. A comparison of expert and nonexpert raters using the consensual assessment technique Creat. Res. J. 20 2008 171 178 Keng S.-L. Smoski M.J. Robins C.J. Effects of mindfulness on psychological health: A review of empirical studies Clin. Psychol. Rev. 31 2011 1041 1056 21802619 Kersemaekers W. Rupprecht S. Wittmann M. Tamdjidi C. Falke P. Donders R. Speckens A. Kohls N. A workplace mindfulness intervention may be associated with improved psychological well-being and productivity. A preliminary field study in a company setting Front. Psychol. 9 2018 195 29541039 Kim W.-S. Byeon G.-J. Song B.-J. Lee H.J. Availability of preoperative anxiety scale as a predictive factor for hemodynamic changes during induction of anesthesia Korean J. Anesthesiol. 58 2010 328 20508787 Kim E.-J. Dimsdale J.E. The effect of psychosocial stress on sleep: A review of polysomnographic evidence Behav. Sleep Med. 5 2007 256 278 17937582 Klainin-Yobas P. Ramirez D. Fernandez Z. Sarmiento J. Thanoi W. Ignacio J. Lau Y. Examining the predicting effect of mindfulness on psychological well-being among undergraduate students: A structural equation modelling approach Personal. Individ. Differ. 91 2016 63 68 Korpela K.M. Place-identity as a product of environmental self-regulation J. Environ. Psychol. 9 1989 241 256 Kulkarni S. O’Farrell I. Erasi M. Kochar M.S. Stress and hypertension WMJ 97 1998 34 38 Kun A. Sadun R. Shaer O. Teodorovicz T. Where did the commute time go Harv. Bus. Rev. 2020 Available Online: https://hbr.org/2020/12/where-did-the-commute-time-go. (Accessed 3 January 2021) Landsbergis P.A. Schnall P.L. Warren K. Pickering T.G. Schwartz J.E. Association between ambulatory blood pressure and alternative formulations of job strain Scand. J. Work Environ. Health 1994 349 363 7863299 Lazarus R.S. Folkman S. Stress, Appraisal, and Coping 1984 Springer publishing company Lebuda I. Zabelina D.L. Karwowski M. Mind full of ideas: A meta-analysis of the mindfulness–creativity link Personal. Individ. Differ. 93 2016 22 26 Lee K.E. Williams K.J.H. Sargent L.D. Williams N.S.G. Johnson K.A. 40-Second green roof views sustain attention: The role of micro-breaks in attention restoration J. Environ. Psychol. 42 2015 182 189 Li, J., Ma, Y., Li, P., Butz, A., 2021. A Journey Through Nature: Exploring Virtual Restorative Environments as a Means to Relax in Confined Spaces. In: Creativity and Cognition. pp. 1–9. Li D. Sullivan W.C. Impact of views to school landscapes on recovery from stress and mental fatigue Landsc. Urban Plan. 148 2016 149 158 Liszio S. Graf L. Masuch M. The relaxing effect of virtual nature: Immersive technology provides relief in acute stress situations Annu. Rev. Cyberther. Telemed. 16 2018 87 93 Lundh L.-G. Broman J.-E. Insomnia as an interaction between sleep-interfering and sleep-interpreting processes J. Psychosom. Res. 49 2000 299 310 11164054 Magnusson D. Backteman G. Longitudinal stability of person characteristics: Intelligence and creativity Appl. Psychol. Meas. 2 1978 481 490 Mark, G., Iqbal, S., Czerwinski, M., 2017. How blocking distractions affects workplace focus and productivity. In: Proceedings of the 2017 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium on Wearable Computers. pp. 928–934. Mark, G., Iqbal, S.T., Czerwinski, M., Johns, P., 2014. Bored mondays and focused afternoons: The rhythm of attention and online activity in the workplace. In: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. pp. 3025–3034. Martindale C. Father’s absence, psychopathology, and poetic eminence Psycho Rep. 31 1972 843 847 Mayer F.S. Frantz C.M. Bruehlman-Senecal E. Dolliver K. Why is nature beneficial? The role of connectedness to nature Environ. Behav. 41 2009 607 643 Mednick S.A. The remote associates test J. Creative Behav 1968 Molla R. For women, remote work is a blessing and a curse Vox 2021 (Online Accessed 9 August 2021) Neeley T. Remote Work Revolution: Succeeding from Anywhere 2021 HarperCollins Newbold, J.W., Luton, J., Cox, A.L., Gould, S.J.J., 2017. Using nature-based soundscapes to support task performance and mood. In: Proceedings of the 2017 CHI Conference Extended Abstracts on Human Factors in Computing Systems. pp. 2802–2809. Newbold J.W. Rudnicka A. Cook D. Cecchinato M. Gould S. Cox A.L. The new normals of work: A framework for understanding responses to disruptions created by new futures of work Hum.-Comput. Interact. 2021 1 24 10.1080/07370024.2021.1982391 Ostafin B.D. Kassman K.T. Stepping out of history: Mindfulness improves insight problem solving Conscious. Cogn. 21 2012 1031 1036 22483682 Palanica A. Lyons A. Cooper M. Lee A. Fossat Y. A comparison of nature and urban environments on creative thinking across different levels of reality J. Environ. Psychol. 63 2019 44 51 Palumbo R. Let me go to the office! An investigation into the side effects of working from home on work-life balance Int. J. Public Sect. Manag 2020 Parsons R. Tassinary L.G. Ulrich R.S. Hebl M.R. Grossman-Alexander M. The view from the road: Implications for stress recovery and immunization J. Environ. Psychol. 18 1998 113 140 Peck T.C. Sockol L.E. Hancock S.M. Mind the gap: The underrepresentation of female participants and authors in virtual reality research IEEE Trans. Vis. Comput. Graphics 26 2020 1945 1954 Peppercorn S. How to stay creative when life feels monotonous Harv. Bus. Rev. 2020 (Online Accessed 26 May 2021]) Peters A. Google is trying to improve its workplaces with offices inspired by nature Fast Company 2015 Radikovic, A.S., Leggett, J.J., Keyser, J., Ulrich, R.S., 2005. Artificial window view of nature. In: CHI’05 Extended Abstracts on Human Factors in Computing Systems. pp. 1993–1996. Ratcliffe, J., Soave, F., Bryan-Kinns, N., Tokarchuk, L., Farkhatdinov, I., 2021a. Extended Reality (XR) remote research: A survey of drawbacks and opportunities. In: Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems. pp. 1–13. Ratcliffe, J., Soave, F., Hoover, M., Ortega, F.R., Bryan-Kinns, N., Tokarchuk, L., Farkhatdinov, I., 2021b. Remote XR Studies: Exploring Three Key Challenges of Remote XR Experimentation. In: Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems. pp. 1–4. Reb J. Narayanan J. Chaturvedi S. Leading mindfully: Two studies on the influence of supervisor trait mindfulness on employee well-being and performance Mindfulness 5 2014 36 45 Ren J. Huang Z. Luo J. Wei G. Ying X. Ding Z. Wu Y. Luo F. Meditation promotes insightful problem-solving by keeping people in a mindful and alert conscious state Sci. China Life Sci. 54 2011 961 965 22038009 Ritter S.M. Damian R.I. Simonton D.K. van Baaren R.B. Strick M. Derks J. Dijksterhuis A. Diversifying experiences enhance cognitive flexibility J. Exp. Soc. Psychol. 48 2012 961 964 Routley N. 6 charts that show what employers and employees really think about remote working World Economic Forum 2020 https://www.weforum.org/agenda/2020/06/coronavirus-covid19-remote-workingoffice-employees-employers Roux A.V.D. Residential environments and cardiovascular risk J. Urban Health 80 2003 569 589 14709706 Rudnicka A. Newbold J.W. Cook D. Cecchinato M.E. Gould S. Cox A.L. Eworklife: Developing effective strategies for remote working during the COVID-19 pandemic Eworklife: Developing Effective Strategies for Remote Working During the COVID-19 Pandemic 2020 The New Future of Work Online Symposium Runco M.A. Acar S. Divergent thinking as an indicator of creative potential Creat. Res. J. 24 2012 66 75 Runco M.A. Jaeger G.J. The standard definition of creativity Creat. Res. J. 24 2012 92 96 Said-Metwaly S. Van den Noortgate W. Kyndt E. Approaches to measuring creativity: A systematic literature review Creat. Theor. Res. Appl. 4 2017 238 275 Salehzadeh Niksirat, K., Silpasuwanchai, C., Mohamed Hussien Ahmed, M., Cheng, P., Ren, X., 2017. A framework for interactive mindfulness meditation using attention-regulation process. In: Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems. pp. 2672–2684. Schnall P.L. Landsbergis P.A. Baker D. Job strain and cardiovascular disease Annu. Rev. Public Health 15 1994 381 411 8054091 Shanahan D.F. Bush R. Gaston K.J. Lin B.B. Dean J. Barber E. Fuller R.A. Health benefits from nature experiences depend on dose Sci. Rep. 6 2016 1 10 28442746 Shepherd D.A. Cardon M.S. Negative emotional reactions to project failure and the self-compassion to learn from the experience J. Manage. Stud. 46 2009 923 949 Spielberger C.D. State-trait anxiety inventory The Corsini Encyclopedia of Psychology 2010 Wiley Online Library 1 Stanney K. Fidopiastis C. Foster L. Virtual reality is sexist: But it does not have to be Front. Robot. AI 7 2020 4 33501173 Stansfeld S.A. Fuhrer R. Head J. Ferrie J. Shipley M. Work and psychiatric disorder in the Whitehall II study J. Psychosom. Res. 43 1997 73 81 9263933 Steptoe A. Brydon L. Emotional triggering of cardiac events Neurosci. Biobehav. Rev. 33 2009 63 70 18534677 Sternberg R.J. Lubart T.I. The concept of creativity: Prospects and paradigms Handb. Creat. 1 1999 3 15 Taylor A.F. Kuo F.E. Sullivan W.C. Coping with ADD: The surprising connection to green play settings Environ. Behav. 33 2001 54 77 Teodorovicz T. Sadun R. Kun A.L. Shaer O. How does working from home during COVID-19 affect what managers do? Evidence from time-use studies Hum.-Comput. Interact. 2021 1 26 10.1080/07370024.2021.1987908 Terzimehić, N., Häuslschmid, R., Hussmann, H., Schraefel, M., 2019. A Review & Analysis of Mindfulness Research in HCI: Framing Current Lines of Research and Future Opportunities. In: Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems. pp. 1–13. Toniolo-Barrios M. Pitt L. Mindfulness and the challenges of working from home in times of crisis Bus. Horiz. 64 2021 189 197 33041346 Torrance E.P. The Torrance Tests of Creative Thinking-Norms-Technical Manual Research Edition-Verbal Tests, Forms A and B-Figural Tests, Forms A and B 1966 Twilio Inc E.P. Twilio 2021 https://www.twilio.com/. (Online Accessed 16 January 2022) Ulrich R.S. Simons R.F. Losito B.D. Fiorito E. Miles M.A. Zelson M. Stress recovery during exposure to natural and urban environments J. Environ. Psychol. 11 1991 201 230 Urban K.K. Assessing creativity: The test for creative thinking-drawing production (TCT-DP) Int. Educ. J. 6 2005 272 280 Van Praag H.M. Can stress cause depression? Prog. Neuropsychopharmacol. Biol. Psych. 28 2004 891 907 Van Reeth O. Weibel L. Spiegel K. Leproult R. Dugovic C. Maccari S. Physiology of sleep (review)–interactions between stress and sleep: From basic research to clinical situations Sleep Med. Rev. 4 2000 201 219 Waters W.F. Adams S.G. Jr. Binks P. Varnado P. Attention, stress and negative emotion in persistent sleep-onset and sleep-maintenance insomnia Sleep 16 1993 128 136 8446832 White M.P. Alcock I. Grellier J. Wheeler B.W. Hartig T. Warber S.L. Bone A. Depledge M.H. Fleming L.E. Spending at least 120 minutes a week in nature is associated with good health and wellbeing Sci. Rep. 9 2019 1 11 30626917 Wooller J.-J. Barton J. Gladwell V.F. Micklewright D. Occlusion of sight, sound and smell during Green exercise influences mood, perceived exertion and heart rate Int. J. Environ. Health Res. 26 2016 267 280 26600402 World Health Organization J.-J. Constitution of the world health organization 1995 World Health Organization J.-J. Occupational Health. Psychosocial Risk Factors and Hazards 2011 World Health Organization Retrieved from: http://www.who.int/occupationalhealth/topics/riskspsychosocial/en Yu C.-P. Lee H.-Y. Luo X.-Y. The effect of virtual reality forest and urban environments on physiological and psychological responses Urban For. Urban Green. 35 2018 106 114

0

PMC9753486

NO-CC CODE

2022-12-16 23:25:44

no

Int J Hum Comput Stud. 2022 Dec 15;:102982

utf-8

Int J Hum Comput Stud

10.1016/j.ijhcs.2022.102982

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)00026-X 10.1016/S0140-6736(21)00026-X Comment Offline: The cosmopolitan state Horton Richard 7 1 2021 9-15 January 2021 7 1 2021 397 10269 8181 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcBritain feels broken. Broken by COVID-19. Broken by Brexit. Broken by the lies that have been dripped into British political debates these past 5 years. Those scientists and doctors who voted to remain in the European Union—perhaps based on a belief in the public value of research and clinical networks that have been painstakingly built over half a century between the UK and its European neighbours—were accused of Project Fear. But now that the UK has finally left the EU, with the thinnest of thin deals, one can look back and say that the arguments to leave were themselves based on nothing but fear—fear of European expansion, fear of the costs of EU membership, fear of immigration, fear of the Euro, and fear of the European Court of Justice. The British Prime Minister says that we should be celebrating an “amazing moment” in the history of our country. But living under an intensifying lockdown in response to a coronavirus epidemic that is manifestly out of control, many might not be able to comprehend that “amazing moment”. Instead, what they might feel is a profound melancholy for a country that has wounded itself, insulted its friends, and harmed a future generation that will bear the burden of their mistakes. © 2021 Thierry Monasse/Pool/Bloomberg/Getty Images 2021 In the late 18th century, Europe was also in turmoil. A Declaration of Independence was signed on July 4, 1776, announcing the separation of 13 American colonies from British rule. The measures that Americans took to free themselves from Britain passed like an electric shock through European nations. The French monarchy was presiding over unsustainable debt and unfair taxes. If the American people could overthrow a corrupt government, why not the French? Meanwhile, the Enlightenment was creating conditions for a raft of inventions and discoveries that challenged longstanding assumptions about the organisation of society. Amid this political and intellectual turbulence, in 1784 the German philosopher Immanuel Kant wrote a short essay with far-reaching implications—Idea for a Universal History with a Cosmopolitan Intent. He sought to show that despite the complicated and unpredictable nature of individuals, there might be a “guiding thread” that directs human progress. Kant set out nine theses to prove his case. First, our natural capacities develop completely and in conformity with their end—eg, an organ that is not intended to be used is a contradiction. Second, the full possibilities of human reason can never be fulfilled in one individual, but only for our entire species across “a perhaps incalculable sequence of generations”, each passing its knowledge to the next. Third, we have the gifts of reason and free will to produce everything from ourselves. Fourth, the means that nature uses to bring about the development of human capacities is antagonism—that is, our “unsocial sociability”, our need to live harmoniously among others but our equal need to fulfil our own individual desires and aspirations. Fifth, the greatest challenge we face in achieving the highest attainable development of our capacities is the creation of “a universal civil society”. Sixth, each of us needs an authority to set limits on our freedom. Seventh, “The problem of establishing a perfect civil constitution depends on the problem of law-governed external relations among nations.” Kant envisages a “federation of peoples”, a “cosmopolitan state in which the security of nations is publicly acknowledged.” Eighth, “One can regard the history of the human species...as the realisation of a hidden plan of nature to bring about [a] perfect national constitution, as the sole state in which all of humanity's natural capacities can be developed.” Nature's “supreme objective”, according to Kant, is “a universal cosmopolitan state.” And finally, although some might consider this “universal history of the world” impossibly romantic, the history of human progress does offer evidence that it might be true. © 2021 Leon Neal/Getty Images 2021 Kant doesn't believe that history proceeds mechanically according to some predestined plan. But he does suggest that if the natural endpoint of human society is cosmopolitanism, it might prove to be “some small motivation” to national leaders as they reflect on their ambitions. At a moment of pandemic and political turmoil, Kant's notion of a “guiding thread” might indeed be comforting. Despite the current worldwide division and disorder, nations and peoples will reunite, polarisations will diminish, and a spirit of cooperation will be renewed. © 2021 Wikimedia (public domain) 2021

33422248

PMC9753495

NO-CC CODE

2022-12-16 23:25:44

no

Lancet. 2021 Jan 7 9-15 January; 397(10269):81

utf-8

Lancet

10.1016/S0140-6736(21)00026-X

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(20)32757-4 10.1016/S0140-6736(20)32757-4 Perspectives Thinking across disciplinary boundaries in a time of crisis Kneebone Roger ab Schlegel Claudia c a Surgical Education, Division of Surgery, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London, London, UK b Centre for Engagement and Simulation Science, Academic Surgery, Chelsea and Westminster Hospital, London SW10 9NH, UK c Simulation Center, Berner Bildungszentrum Pflege, Bern, Switzerland 7 1 2021 9-15 January 2021 7 1 2021 397 10269 8990 © 2020 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcThe COVID-19 pandemic forced clinicians and researchers to look beyond traditional professional boundaries, working at a speed that would have previously been unthinkable. A widespread shortage of essential equipment in the early stages of the pandemic prompted health-care staff to search outside their clinical domains for new solutions. Reports from the UK and beyond showed clinicians, engineers, and racing car teams collaborating to design and produce respiratory support apparatus such as ventilators and continuous positive airway pressure machines for treating seriously ill patients. At the same time, a shortage of personal protective equipment prompted tailors and textile workers to come up with new ways of producing gowns and face masks. Such collaborations depend on effective communication between clinicians, scientists, engineers, and those from the creative and craft industries. The novelty and challenge of such collaborations highlight the pigeon-holed nature of much medical education. As clinician educators we believe that collaboration beyond the world of medicine is essential, not only in times of crisis but also as part of normal training. Yet a widely held assumption that everything a learner needs to know about a field can be gained from those already working within it holds powerful sway. We challenge that assumption, proposing that clinicians can learn from experts outside medicine whose ways of doing resonate with medical practice but whose expertise often goes unrecognised. While the value of collaborating with engineers and non-clinical scientists is easy to see, there is great scope too for collaboration with those in the visual and performing arts. For example, one consequence of the COVID-19 pandemic has been a radical shift from face-to-face to remote consultation. Clinicians and patients have developed unfamiliar ways of interacting. We suggest there is much for health professionals to learn from creative performers, such as musicians, actors, and close-up magicians. These performers can be adept at capturing and shaping each audience member's attention via remote technology, sustaining engagement throughout the performance, and ensuring that all participants perceive the experience as worthwhile. Yet experts in such disciplines have different “languages”, and clinical learners are often ill-equipped to communicate effectively across disciplinary boundaries. The sociologist Harry Collins distinguishes between contributory expertise (being able to do something yourself) and interactional expertise (fluency in the technical language of a specialism without being expert in it yourself). Medical training focuses on the former, sometimes at the expense of the latter. Of course there is a huge corpus of domain-specific knowledge and skills that medical students and other clinicians in training need to master. In the early phases this entails a necessary funnelling into scientific and clinical study, elbowing out perspectives from beyond medicine and science. In the UK educational system in particular, there is a danger that the languages of the arts, social sciences, and engineering will be overlooked in that process. Becoming a skilled, safe, knowledgeable, and compassionate clinician involves an ontological transformation, an internal journey that often takes years. Developing one's professional identity—what it is to be a medical doctor as well as to master what a medical doctor must know and be able to do—is a complex process whose maturation may not fit with assessments of knowledge and component competencies. Each clinician's professional identity requires the integration of many kinds of knowing, not least the ability to recognise the impact of one's own physical and mental state when caring for patients and how this may be affected by extraordinary circumstances such as war, natural disasters, or pandemics. Awareness is growing of the value of study that transcends disciplinary boundaries—of finding ways to broaden out the funnel of medical learning. Many programmes bring medicine, science, and engineering together, encouraging students to develop solutions to practical challenges. Medical humanities programmes invite students to connect with the worlds of literature, poetry, and the visual and performing arts, among other fields, acknowledging that valuable insights into clinical practice come from this humanist perspective. Yet such programmes are typically dominated by words and images, with much less emphasis on physicality and doing. Trainees are seldom invited to move outside the world of medicine to gain practical insights from other kinds of experts as they enter postgraduate medical education. For example, craftspeople in wood, stone, glass, textiles, and metal have an understanding of the material world that reaches back centuries. Performers—whether in safety-critical industries such as aviation or offshore drilling, or in music, dance, theatre, or sport—offer insights into how people work together under pressure. Based on their own apprenticeships and subsequent experiences, these experts, like those in medicine and nursing, understand the value of repetitive work which can seem boring at the time but lays the groundwork for eventual mastery. Through practice and performance they develop an understanding of their work that defies verbal description. Experts in such fields are skilled at recognising how their skills can degrade when they are tired, irritated, fearful, or unwell. We argue for creating perforations in the funnel of learning so that exposure to such insights can become part of mainstream medical education. © 2021 Pedro Vilela/Getty Images 2021 Many such experts gain their mastery through prolonged exposure to materials, tools, and techniques. They develop a capacity for attentiveness on the basis of close observation and sensory awareness. These qualities are crucial in medicine too, although often overshadowed by a focus on factual knowledge or outsourced to imaging and other technologies. Conceptual knowledge and embodied knowledge develop at different rates. Theoretical understanding and factual grasp can take place rapidly, one stage leading quickly to the next. The knowledge of the hand takes much longer. Time and again, the craftspeople in our studies described how their understanding of what they were attempting to do outstripped their ability to do it. Whether you're learning to make a jacket, shape a vase, repair an engine, engrave a piece of glass, or perform close-up magic, there's no substitute for sustained and repetitive work in the workplace—and this applies to the operating theatre, the ward, and the consulting room too. Our exploratory work over several years has brought together clinicians, scientists, craftspeople, and leaders in the visual and performing arts to identify common themes. These have included small-scale working, working with precious materials, and time-critical performance. Subsequent work in small groups has investigated promising themes in more detail. For example, an exploration of risk, error, and recovery brought together an upper gastrointestinal cancer surgeon, a polar explorer, a combat pilot, and a classical guitarist. Despite their different domains of expertise, all described recognising when “something feels wrong'', reaching a state of provisional safety before spending time in analysing the cause. The combat pilot described how he once sensed before touchdown that something “wasn't right”. Rather than trying to work out why, his immediate response was to put on power, gain height and “go round again” while he reviewed the situation and considered his options from a place of temporary safety. On this occasion, he realised that despite going through his pre-landing checklist he had not actually pulled the lever that would lower the aircraft's undercarriage. Continuing to land would have been fatal. The other experts all described strategies for reaching a safe space in their fields—whether repositioning abdominal organs and “starting again” when confronted by unexpected problems in complex cancer surgery; building a temporary ice shelter when becoming lost and exhausted in the Arctic; or “reverting to tempo” with other players in a musical performance to allow a breathing space before resuming solo display. For each expert this temporary refuge allowed them to rethink, review, and restart. This was especially relevant in the case of low-probability, high-severity events that most professionals are unlikely to have experienced before. Awareness that one is entering a risky state and knowing how to achieve a temporary state of safety is a characteristic of expert medical performance. Similar experiences characterise many branches of medicine and nursing. Though these broader aspects of clinical practice are widely understood by individual clinicians, they are seldom systematically addressed within institutions or curricula. Recognising the expertise that lies outside medicine is a means of perforating the funnel within which medical doctors tend to learn. Unless we pay attention to the craftsmanship and performance of medicine as well as its science, we overlook essential aspects of clinical practice. We owe it to our patients, our students, and ourselves to ensure this does not happen. Yet we should not need a worldwide crisis such as COVID-19 to trigger such collaboration. Invaluable expertise is all around us, hiding in plain sight. We have a responsibility to our students to ensure they develop the awareness and skills to engage with experts outside medicine, recognising complementary skills and ways of thinking. The challenge is to bring such expertise into our educational programmes and working lives so patients and clinicians can profit from it together. Although COVID-19 has disclosed fault lines in our educational system, the pandemic has also provided inspiring examples of how we might continue to do things differently once this crisis is over. Roger Kneebone is the author of Expert: Understanding the Path to Mastery (Viking Penguin, 2020). ==== Refs Further reading Pollina E Piovaccari G Ferrari and Fiat look helping Italy make ventilators in coronavirus crisis May 8, 2020 Reuters http://www.reuters.com/article/us-health-coronavirus-ventilators-italy/exclusive-ferrari-and-fiat-look-at-helping-italy-make-ventilators-in-coronavirus-crisis-idUSKBN2162YT Nanrah G The Nottingham textile company making face masks to protect key workers outside the NHS. NottinghamshireLive http://www.nottinghampost.com/news/nottingham-news/nottingham-textile-company-making-face-4044589 April 15, 2020 Kneebone R The vanishing art of doing BMJ 364 2019 k5326 Collins H Evans R Rethinking expertise 2007 University of Chicago Press Chicago, IL Kneebone R Schlegel C Spivey A Science in hand: how craft informs lab work Nature 564 2018 188 189 30531883 Pallasmaa J The thinking hand: existential and embodied wisdom in architecture 2009 Wiley Chichester

33422253

PMC9753501

NO-CC CODE

2022-12-16 23:25:45

no

Lancet. 2021 Jan 7 9-15 January; 397(10269):89-90

utf-8

Lancet

10.1016/S0140-6736(20)32757-4

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(20)32757-4 10.1016/S0140-6736(20)32757-4 Perspectives Thinking across disciplinary boundaries in a time of crisis Kneebone Roger ab Schlegel Claudia c a Surgical Education, Division of Surgery, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London, London, UK b Centre for Engagement and Simulation Science, Academic Surgery, Chelsea and Westminster Hospital, London SW10 9NH, UK c Simulation Center, Berner Bildungszentrum Pflege, Bern, Switzerland 7 1 2021 9-15 January 2021 7 1 2021 397 10269 8990 © 2020 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcThe COVID-19 pandemic forced clinicians and researchers to look beyond traditional professional boundaries, working at a speed that would have previously been unthinkable. A widespread shortage of essential equipment in the early stages of the pandemic prompted health-care staff to search outside their clinical domains for new solutions. Reports from the UK and beyond showed clinicians, engineers, and racing car teams collaborating to design and produce respiratory support apparatus such as ventilators and continuous positive airway pressure machines for treating seriously ill patients. At the same time, a shortage of personal protective equipment prompted tailors and textile workers to come up with new ways of producing gowns and face masks. Such collaborations depend on effective communication between clinicians, scientists, engineers, and those from the creative and craft industries. The novelty and challenge of such collaborations highlight the pigeon-holed nature of much medical education. As clinician educators we believe that collaboration beyond the world of medicine is essential, not only in times of crisis but also as part of normal training. Yet a widely held assumption that everything a learner needs to know about a field can be gained from those already working within it holds powerful sway. We challenge that assumption, proposing that clinicians can learn from experts outside medicine whose ways of doing resonate with medical practice but whose expertise often goes unrecognised. While the value of collaborating with engineers and non-clinical scientists is easy to see, there is great scope too for collaboration with those in the visual and performing arts. For example, one consequence of the COVID-19 pandemic has been a radical shift from face-to-face to remote consultation. Clinicians and patients have developed unfamiliar ways of interacting. We suggest there is much for health professionals to learn from creative performers, such as musicians, actors, and close-up magicians. These performers can be adept at capturing and shaping each audience member's attention via remote technology, sustaining engagement throughout the performance, and ensuring that all participants perceive the experience as worthwhile. Yet experts in such disciplines have different “languages”, and clinical learners are often ill-equipped to communicate effectively across disciplinary boundaries. The sociologist Harry Collins distinguishes between contributory expertise (being able to do something yourself) and interactional expertise (fluency in the technical language of a specialism without being expert in it yourself). Medical training focuses on the former, sometimes at the expense of the latter. Of course there is a huge corpus of domain-specific knowledge and skills that medical students and other clinicians in training need to master. In the early phases this entails a necessary funnelling into scientific and clinical study, elbowing out perspectives from beyond medicine and science. In the UK educational system in particular, there is a danger that the languages of the arts, social sciences, and engineering will be overlooked in that process. Becoming a skilled, safe, knowledgeable, and compassionate clinician involves an ontological transformation, an internal journey that often takes years. Developing one's professional identity—what it is to be a medical doctor as well as to master what a medical doctor must know and be able to do—is a complex process whose maturation may not fit with assessments of knowledge and component competencies. Each clinician's professional identity requires the integration of many kinds of knowing, not least the ability to recognise the impact of one's own physical and mental state when caring for patients and how this may be affected by extraordinary circumstances such as war, natural disasters, or pandemics. Awareness is growing of the value of study that transcends disciplinary boundaries—of finding ways to broaden out the funnel of medical learning. Many programmes bring medicine, science, and engineering together, encouraging students to develop solutions to practical challenges. Medical humanities programmes invite students to connect with the worlds of literature, poetry, and the visual and performing arts, among other fields, acknowledging that valuable insights into clinical practice come from this humanist perspective. Yet such programmes are typically dominated by words and images, with much less emphasis on physicality and doing. Trainees are seldom invited to move outside the world of medicine to gain practical insights from other kinds of experts as they enter postgraduate medical education. For example, craftspeople in wood, stone, glass, textiles, and metal have an understanding of the material world that reaches back centuries. Performers—whether in safety-critical industries such as aviation or offshore drilling, or in music, dance, theatre, or sport—offer insights into how people work together under pressure. Based on their own apprenticeships and subsequent experiences, these experts, like those in medicine and nursing, understand the value of repetitive work which can seem boring at the time but lays the groundwork for eventual mastery. Through practice and performance they develop an understanding of their work that defies verbal description. Experts in such fields are skilled at recognising how their skills can degrade when they are tired, irritated, fearful, or unwell. We argue for creating perforations in the funnel of learning so that exposure to such insights can become part of mainstream medical education. © 2021 Pedro Vilela/Getty Images 2021 Many such experts gain their mastery through prolonged exposure to materials, tools, and techniques. They develop a capacity for attentiveness on the basis of close observation and sensory awareness. These qualities are crucial in medicine too, although often overshadowed by a focus on factual knowledge or outsourced to imaging and other technologies. Conceptual knowledge and embodied knowledge develop at different rates. Theoretical understanding and factual grasp can take place rapidly, one stage leading quickly to the next. The knowledge of the hand takes much longer. Time and again, the craftspeople in our studies described how their understanding of what they were attempting to do outstripped their ability to do it. Whether you're learning to make a jacket, shape a vase, repair an engine, engrave a piece of glass, or perform close-up magic, there's no substitute for sustained and repetitive work in the workplace—and this applies to the operating theatre, the ward, and the consulting room too. Our exploratory work over several years has brought together clinicians, scientists, craftspeople, and leaders in the visual and performing arts to identify common themes. These have included small-scale working, working with precious materials, and time-critical performance. Subsequent work in small groups has investigated promising themes in more detail. For example, an exploration of risk, error, and recovery brought together an upper gastrointestinal cancer surgeon, a polar explorer, a combat pilot, and a classical guitarist. Despite their different domains of expertise, all described recognising when “something feels wrong'', reaching a state of provisional safety before spending time in analysing the cause. The combat pilot described how he once sensed before touchdown that something “wasn't right”. Rather than trying to work out why, his immediate response was to put on power, gain height and “go round again” while he reviewed the situation and considered his options from a place of temporary safety. On this occasion, he realised that despite going through his pre-landing checklist he had not actually pulled the lever that would lower the aircraft's undercarriage. Continuing to land would have been fatal. The other experts all described strategies for reaching a safe space in their fields—whether repositioning abdominal organs and “starting again” when confronted by unexpected problems in complex cancer surgery; building a temporary ice shelter when becoming lost and exhausted in the Arctic; or “reverting to tempo” with other players in a musical performance to allow a breathing space before resuming solo display. For each expert this temporary refuge allowed them to rethink, review, and restart. This was especially relevant in the case of low-probability, high-severity events that most professionals are unlikely to have experienced before. Awareness that one is entering a risky state and knowing how to achieve a temporary state of safety is a characteristic of expert medical performance. Similar experiences characterise many branches of medicine and nursing. Though these broader aspects of clinical practice are widely understood by individual clinicians, they are seldom systematically addressed within institutions or curricula. Recognising the expertise that lies outside medicine is a means of perforating the funnel within which medical doctors tend to learn. Unless we pay attention to the craftsmanship and performance of medicine as well as its science, we overlook essential aspects of clinical practice. We owe it to our patients, our students, and ourselves to ensure this does not happen. Yet we should not need a worldwide crisis such as COVID-19 to trigger such collaboration. Invaluable expertise is all around us, hiding in plain sight. We have a responsibility to our students to ensure they develop the awareness and skills to engage with experts outside medicine, recognising complementary skills and ways of thinking. The challenge is to bring such expertise into our educational programmes and working lives so patients and clinicians can profit from it together. Although COVID-19 has disclosed fault lines in our educational system, the pandemic has also provided inspiring examples of how we might continue to do things differently once this crisis is over. Roger Kneebone is the author of Expert: Understanding the Path to Mastery (Viking Penguin, 2020). ==== Refs Further reading Pollina E Piovaccari G Ferrari and Fiat look helping Italy make ventilators in coronavirus crisis May 8, 2020 Reuters http://www.reuters.com/article/us-health-coronavirus-ventilators-italy/exclusive-ferrari-and-fiat-look-at-helping-italy-make-ventilators-in-coronavirus-crisis-idUSKBN2162YT Nanrah G The Nottingham textile company making face masks to protect key workers outside the NHS. NottinghamshireLive http://www.nottinghampost.com/news/nottingham-news/nottingham-textile-company-making-face-4044589 April 15, 2020 Kneebone R The vanishing art of doing BMJ 364 2019 k5326 Collins H Evans R Rethinking expertise 2007 University of Chicago Press Chicago, IL Kneebone R Schlegel C Spivey A Science in hand: how craft informs lab work Nature 564 2018 188 189 30531883 Pallasmaa J The thinking hand: existential and embodied wisdom in architecture 2009 Wiley Chichester

33338438

PMC9753502

NO-CC CODE

2022-12-16 23:25:45

no

Lancet. 2021 Dec 15 9-15 January; 397(10269):79-80

latin-1

Lancet

10.1016/S0140-6736(20)32671-4

oa_other

==== Front Energy Build Energy Build Energy and Buildings 0378-7788 1872-6178 Elsevier B.V. S0378-7788(22)00894-5 10.1016/j.enbuild.2022.112723 112723 Article Impact of the COVID-19 on Electricity Consumption of Open University Campus Buildings-The Case of Twente University in the Netherlands Xu Sheng a Cheng Bin a⁎ Huang Zefeng b Liu Tao c Li Yuan d Jiang Lin a Guo Wei e Jie Xiong a a School of Civil Engineering and Architecture, Southwest University of Science and Technology, Mianyang 621010, China b School of Urban Design, Wuhan University, Wuhan 430072, China c School of Earth Sciences, Tsinghua University, Beijing 100084, China d School of Architecture and Civil engineering,Xiamen University, 361005,China e Department of Architecture, Deyang Installation Technician College, Deyang, 618099 China ⁎ Corresponding author. 15 12 2022 15 12 2022 11272326 10 2022 27 11 2022 11 12 2022 © 2022 Elsevier B.V. All rights reserved. 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Since the COVID -19 outbreak, the restrictive policies enacted by countries in response to the epidemic have led to changes in the movement of people in public places, which has had a direct impact on the use and energy consumption of various public buildings. This study was based on electricity consumption data for 25 on-campus public buildings at 1-hour intervals between January 2020 and June 2022 at Tewnte University in the Netherlands, and after the data were climate-corrected by multiple regression analysis, the changes in EU and EUI for various types of buildings were compared for different restriction periods using ANOVA, LSD and t-tests. And additionally, further analyzed the changes and reasons for the electricity consumption of various public buildings on campus and customers' electricity consumption behavior in a period of time after the lifting of the epidemic restriction policy. The results of ANOVA analysis show that the restriction policy has a significant effect on teaching, sports, and cultural buildings, and the electricity intensity of the three types of buildings is reduced by 0.28, 0.09, and 0.07 kwh/m2/day respectively under the strict restriction policy; The t-test results show that during the restriction period, all building types, except for living and academic buildings, show a significant decreasing trend, with the teaching buildings having the greatest energy saving potential, with an average daily EU reduction of 1088kwh/day and an EUI reduction of 0.075kwh/ m2/day. The above findings provide a case study of a complete cycle of energy consumption changes in university buildings under similar epidemic restriction policies before and after the epidemic restriction, and inform the electricity allocation policies of university and government energy management authorities. Keywords Open university campus Energy consumption COVID-19 restriction intensity Changes in electricity usage behavior ==== Body pmc1 Introduction On January 5, 2020, the World Health Organization (WHO) announced that it had received news of a case of pneumonia of unknown origin from Wuhan, China [1],and officially released the COVIN-19 Joint Operations Investigation Report on February 28, 2020 [2].Each country then developed a series of restrictions based on the reporting guidelines and its own situation to help contain the spread of the epidemic [3], [4], [5], [6].These restrictions directly affect the normal functioning of various industries, including but not limited to healthcare[7],education [8],tourism [9], [10],industry [11], [12],economy [13],energy [14],environment [15], [16], etc. At the same time, a series of changes in the behavior of people's daily activities, such as work and study, have occurred due to restrictive policies. The daily behavior of occupants, in turn, has a significant impact on building energy consumption[17], [18]. Therefore, the restrictive policies during the epidemic will certainly cause changes in building energy consumption. Energy consumption in buildings has a strategic importance in world energy consumption[19], and for this reason, a series of studies have been carried out in the academic community on the changes in energy consumption in different types of buildings during the epidemic. Among them, Hyuna Kang et al. evaluated the energy consumption of different types of buildings in Korea based on big data and found that the overall electricity and natural gas were 4.46% and 10.35% lower compared to the year before the epidemic, and the rate of change in building energy consumption depended on the correlation between building function and neo crown pneumonia[20]. Ahmed Abdeen et al. used statistical methods to analyze electricity data from more than 500 households in Canada and found a 12.8% increase in daily household electricity demand compared to the pre-epidemic period[21]. Usep Surahman et al. studied the use of natural gas and appliances in 311 residential houses in Indonesia through interviews and field surveys, which showed that the average annual energy consumption during the epidemic was elevated by 3.0 GJ compared to the previous period[22]. Energy consumption changed not only in residential buildings, but also in public buildings during the epidemic with significant fluctuations, Matheus Soares Geraldi et al. analyzed and investigated the electricity consumption of municipal buildings during the Floriańopolis epidemic restriction policy and found that the average electricity consumption of health centers, administrative buildings, elementary school and kindergartens decreased by 11.1%, 38.6%, 50.3% and 50.4%, respectively, compared to the pre-epidemic period[23]. ZF Huang and ZH G's statistical analysis of half-hourly electricity consumption and intensity of electricity use in municipal public buildings in Scotland, UK, yielded the same findings as Floriańopolis et al. The intensity of electricity use was significantly reduced in all public buildings except for office buildings, with the lowest reduction in building energy consumption in secondary schools[24]. Based on the current study of building energy consumption during the aforementioned epidemic, it was shown that the most significant reduction in energy consumption was seen in primary and secondary school buildings[25]. Perhaps because the education industry is generally a congregation of students and teachers, the impact during the New Coronavirus was more pronounced. However, university campuses are generally larger and have a greater variety of building types and therefore higher energy demand compared to primary and secondary schools. The restrictions during the epidemic have also had a knock-on effect on university campuses, such as a drop in the mobility of international students[26], Some teachers are unable to teach normally due to city restrictions, and the teaching mode is changed to online teaching mode with the help of modern communication tools such as computers and cell phones[27], [28], [29], [30], A safe distance needs to be maintained even when attending classes on campus during the restriction period[31], and the academic and psychological stress that COVID-19 causes to students, etc[32], [33], [34]. The changes brought about by these shocks also have a direct impact on the energy consumption and carbon emissions of university campus buildings[35], [36], [37]. The energy consumption of university campus buildings occupies an important position in the energy consumption of urban buildings, so it is necessary to conduct a detailed study of the changes in energy consumption of university campus buildings during the epidemic. At present, a few scholars have begun to pay attention to the changes in energy consumption of university campuses during the epidemic. For example, Paula Brumer Franceschini evaluated the existing research methods of campus building performance, and looked forward to the future research topics based on the impact of the new crown epidemic on campus personnel living behavior models[38]. Sharifah Nurain Syed Nasir et al compared the building electricity consumption of a research complex at the National University of Malaysia in 2019-2020 and found that the building's energy consumption decreased by 11% during the covid-19 period[39]. In contrast, Reza Mokhtari et al. used a simulation optimization algorithm to evaluate the effects of the presence of people, air exchange rate, class time and work time on HVAC system energy consumption and the number of people infected with New Crown Pneumonia in a building at Tehran University during the New Crown Pneumonia epidemic, concluding that increasing the ventilation rate of the building during the New Crown Pneumonia epidemic was effective in reducing the number of infections, but the corresponding building energy consumption would also increase[40]. In addition, K. Gaspar et al. investigated the energy consumption of 83 teaching buildings of the Polytechnic University of Catalonia-Barcelona during the covid-19 blockade, The results show that the weather-corrected energy consumption of the entire university campus buildings decreased by 19.3% during the year during the epidemic, and the occupancy rate of academic buildings did not significantly affect the changes in building energy consumption[36]. XC Gui, ZH Gou and others compared the energy consumption of Griffith University in a normal academic year and an epidemic academic year, and concluded that during the new coronary pneumonia period, by shutting down the air conditioners in academic, administrative, teaching, retail and other buildings, the average energy consumption per Weekly energy can save at least 860kwh of energy consumption[41]. However, most of the existing studies on building energy consumption on university campuses during the epidemic period focus on the comparison between the period before the epidemic and the period during the epidemic restriction, and lack a trend analysis on the change of building energy consumption after the lifting of the epidemic restriction policy. In addition, due to the epidemic situation and policies of different countries and the influence of climatic conditions, the energy consumption changes of different university campus buildings during the epidemic also vary greatly. In short, there is still a paucity of research cases exploring the potential for building energy efficiency using changes in building energy consumption data on university campuses during the New Crown Pneumonia outbreak as a case study. As an important part of urban educational resources, university campuses have various public buildings inside the campus for a large number of students. The teaching activities with aggregation type are more sensitive to the restrictive policy of epidemic, and with the intervention of online education, the usage of some public buildings (teaching buildings or cultural centers, etc.) inside university campuses is significantly reduced. If we can understand the impact of the epidemic restriction policy on the energy consumption of different types of public buildings on campus by studying the changes in energy consumption of university campus buildings during the epidemic, we can develop subsequent response plans according to the degree of impact on each type of building, which will make a great contribution to reducing carbon emissions in cities. This paper uses the latest real-time data on the hourly electricity consumption of 25 different types of public buildings on campus in the database of the University of Twente in the Netherlands from 2020 to June 2022 to explore the impact of the new crown epidemic restriction policy on the energy consumption of different types of public buildings on university campuses. It provides a more detailed and complete campus case for building energy consumption research under the covid-19 epidemic, and provides a reference for the government and university energy management departments to formulate policies under similar conditions. The main content of the article is structured as follows: Section 2 introduces the research methods and cases, Section 3 compares and analyzes the corrected data in different periods and stages, and Section 4 discusses the results. Section 5 presents the main conclusions drawn from this study. 2 Methodology 2.1 Definition of Space-Time Boundaries for Study Cases The research subject, Twente University, is located in Enschede, a municipality in the eastern part of the Netherlands, in the interior of the country, with a typical temperate maritime climate. The climate features a relatively mild performance with adequate rainfall. The highest average temperature month of the year is July, with an average temperature of 17.2°C, and the lowest average temperature month is January, with an average temperature of 1.7°C. The maximum average temperature difference between winter and summer is about 15.5°C, so there is a degree of demand for heating and cooling in winter and summer. On February 27, 2020, the Dutch government announced the first coronavirus person[42], Since then, COVID-19 has been spreading widely in the Netherlands, and according to Dutch government statistics, as of August 18, 2022, the number of people infected with COVID-19 in the Netherlands has reached 8.37 million, with 22,565 deaths[43]. In the interest of citizen health and safety, the government began developing measures to curb the spread of the virus on March 12, 2020, and officially issued strict restrictions on March 23, 2020. Twente University immediately announced a series of response measures in accordance with the restrictions issued by the Dutch government. Due to the uncertainty of the epidemic, these measures have also changed with the new restrictions announced by the government, but they are basically based on government rules[44]. Important responses on campus as of now are divided into the following time periods:From March 27 to June 1, 2020, all teaching and examinations will be conducted online, students will study at home, and all university teaching services will be closed until the government announces the relaxation policy. Some services and facilities on the university campus and public transportation will be restored one after another under the specified number of restrictions from June 1, 2020 to August 31, 2020 before the start of the academic year, but the university's teaching is still required to be conducted online (strict restriction period 2020.3.27∼ 2020.8.31). With COVID-19 effectively controlled under the restrictive policy August 31, 2020, Twente University decided to open the school normally under the government's relaxed restrictive policy, but on-campus personnel must wear masks and maintain a safety distance of more than 1.5m, adopting a hybrid online and offline teaching model. During this period, the university closed some public buildings due to the resurgence of the epidemic but the rest of the academic activities were carried out normally. (Relaxed restriction period 2020.9.1 ∼ 2020.12.14). With the rebound of the New Crown epidemic, the university campus began another period of strict control on December 15, 2020, with a change in the educational model to online closure of indoor gymnasiums, libraries, and other similar public facilities, and a curfew policy (strict restriction period 2020.12.15 ∼ 2021.4.25). With the introduction of the vaccine, the situation of COVID-19 improved, the Dutch government lifted the curfew and other restrictive policies on April 26, 2021, and opened various public places one after another, and universities opened some face-to-face education again, but with time and number regulations. The University officially opens on September 1, 2021, during which there are partial restrictions on the hours of operation of some public buildings, and on-campus personnel conduct teaching and learning activities under protective regulations (relaxed restriction period - 2021.4.26 ∼ 2022.2.16). On February 17, 2022, the government officially lifted all restrictions and resumed normal activities as the New Crown vaccine became widespread and the epidemic stabilized. For a more intuitive understanding of the research cycle, this study divided the cycle of response measures during the epidemic into restricted and unrestricted periods based on the campus response strategies and time points described above. Among them, the restriction period mainly includes strict restriction and easy restriction period, and the non-restriction period includes pre-epidemic and post-epidemic. (Table 1 )Table 1 search process time point Year Period Time 2020∼2021 Unrestricted Period(Before the COVID-19)Strictly RestrictedEasting RestrictedStrictly Restricted 2020/01/01∼2020/03/262020/03/27∼2020/08/312020/09/01∼2020/12/142020/12/15∼2021/04/25 2021∼2022/06 Easting RestrictedUnrestricted Period(After the COVID-19) 2021/04/26∼2022/02/162022/02/17∼2022/06/30 Since school vacation dates and public holidays can also have an impact on the use of public buildings on campus, this issue needs to be considered in the analysis. In Fig.1 we have drawn a detailed timeline based on the campus calendar with different phases of restriction levels, schools and public holidays. As can be seen from the chart, the number of vacation days during 2020∼2022 is basically the same, but the specific vacation time varies. This study analyzed the electricity consumption of 25 different types of public buildings in Twente University based on the above timeline framework.Fig. 1 Twente University Calendar Arrangement and Different Restriction Periods 2.2 Data collection Twente University's public building energy consumption data is published in real time on its official website, which includes the hourly electricity consumption of many types of buildings and campus facilities within the University, covering the real-time electricity consumption of 45 public places on campus[45]. These include public lighting, parking lots, cultural centers, teaching buildings, laboratories, facility management rooms, and commercial buildings on campus, among others. CSV format can be exported for analysis at any time as needed. The building electricity consumption (EU) is recorded hourly by the smart meter, which can effectively ensure the accuracy of the data. Due to the lack of electricity consumption data for some time periods in the public data and some sites are not suitable for this research scope, after a series of screening, 25 buildings were finally selected for this analysis. In addition, based on the campus building information and campus map[46],this paper used QGIS software combined with the high-resolution sentinel-12 images provided by ESA (European space Agency), extracted the building base vector area data using random forest deep learning method, and imported into Google Earth (google earth) for correction, finally estimated the building area and determined the building type. Table 2 shows the basic classification of buildings and floor area statistics (Estimated floor area = base area * number of building floors). Although the estimated floor area cannot reach full accuracy, it is sufficient for calculating the building electricity intensity (EUI) based on the same comparison conditions in this paper.Table 2 Final selection of buildings and classification Building type Numbers of case Estimated floor area (m2) Average Value Maximum Value Minimum Value Service building 9 2718 6810 1600 Teaching building 6 14501 38830 3980 Research building 4 2777 7350 600 Sports building 3 3817 6180 1600 Office building 2 5495 9580 1410 Culture building 1 12536 12536 12536 2.3 Climate adjustment data Climate characteristics are an important factor affecting building energy consumption that cannot be ignored[47]. The outdoor temperature in different years is not the same for building energy consumption, and it is not reasonable to use raw data to directly evaluate and compare energy consumption in different years under this objective factor. It is therefore necessary to make climate corrections to the raw data in order to compare building energy consumption over time under the same base conditions [48]. Currently, climate correction has been widely used to evaluate building energy assessment studies during the New Crown Pneumonia outbreak [20], [36], The principle is to perform a multiple regression of building energy consumption against the climate proxies HDD (heating degree days) and CDD (cooling degree days) of the study area. There are two main steps: firstly, the HDD and CDD are calculated based on the local climate data, and secondly, the correction coefficients are derived from the multiple regression analysis. The first climate data used for the calculation were obtained from the website of the local weather station in the Netherlands [49]. Then a reference temperature is set for the calculation of HDD and CDD. The reference temperature varies from country to country due to the different regions and climatic conditions around the world, and in this study the reference temperature is set to 18.3°C according to the international standard recommended by ASHRAE, the American Society of Heating Refrigeration and Air Conditioning Engineers. The simplest way to calculate the average daily temperature is to compare it with the reference temperature, and when the average temperature is lower than the reference temperature, it is classified as HDD, and vice versa, it is classified as CDD. The calculation formula is as follows:(1) HDD=∑i=1n(18.3-Tmeani) (2) CDD=∑i=1n(Tmeani-18.3) In the above equation, where Tmeani=（Tmaxi+Tmini）/2,That is, the average of the daily maximum temperature and the daily minimum temperature. 18.3 is the set reference temperature. After the HDD and CDD are calculated, formula (3-5) is used to calculate the building energy consumption after climate adjustment. Eq. (3) shows the regression analysis between the independent variable (monthly HDD and CDD) and the dependent variable (amount of energy consumed in the building), Eq. (4) then represents the use of the regression coefficients (B1j and B2j) derived from the above multiple regression analysis to calculate the correction coefficients（CORRjk）. Eq. (5) then represents the climate-adjusted energy consumption data obtained by subtracting the correction factor (CORRjk) from the raw monthly consumption data (Yjk) of the different buildings mentioned above.(3) Yjk=b0j+b1j·HDDij+b2j·CDDij (4) CORRjk=b^1j·HDDjk-NHDDj+b^2j·CDDjk-NCDDj (5) Yjka=Yjk-CORRjk Where Yjk represents the electric energy consumption of class j building in month k,b0j represents normal time and day rather than seasonal energy consumption；HDDjk and CDDjk then represent the HDD and CDD of building type j in month k, respectively. b1j represents the regression coefficient of HDDjk；b2j represents the regression coefficient of CDDjk; CORRjk represents the correction coefficient of building category j in month k. NHDDj represents the average HDDj in month k of recent years; NCDDj represents the average CDDj in month k of recent years, and the last 5 years of climate data are used for the calculation in this paper. Yjka then represents the climate-corrected standardized building electrical energy consumption data. 2.4 Analytical method In this study, ANOVA, LSD method and independent sample t-test were used to compare and analyze various types of buildings under different restriction policies during the study period. IBM SPSS Statistics 22 software was used in this analysis process. The variance model is used to determine whether the control variables have a significant effect on the observed variables and has been used to analyze the effect of different surface materials on the surrounding thermal environment[50]. In this paper is used to determine whether there is a significant effect of COVID-19 limit intensity on building electricity consumption and intensity of electricity use. LSD is further used to determine the value of the difference between groups of the same type over different restriction periods, while the independent samples t-test is used to determine whether there is a significant difference in two overall means using the overall sample, and has long been established in similar studies[51]. It was mainly used in this study to supplement the analysis of the changes and differences in EUI occurring in various types of buildings before and after the restriction period and epidemic in different years. Since there is a certain lag effect between the time of restriction policy release and the response of electricity consumption behavior, in order to ensure the real situation of electricity consumption fluctuation under different restriction intensity in the ANOVA process, we screened the electricity consumption data of one stable natural month in each restriction period for ANOVA analysis From T0 before the implementation of the restriction policy until T5 after the lifting of the restriction, for a total of 6 natural months.• T0(2020.2.25∼3.25) • T1(2020.5.15∼6.15) • T2(2020.10.15∼11.15) • T3(2021.2.15∼3.15) • T4(2021.6.15∼7.15) • T5(2022.3.15∼4.15) And before the analysis, it is necessary to conduct normal distribution and homogeneity test of variance on the data to determine whether the variance of power consumption of various buildings under different restriction periods is equal. The analysis results only have statistical significance when they conform to the normal distribution and the Sig value in the variance test is ≥ 0.05. The comparative analysis of LSD further compares the difference values of the same type of building in different restriction periods based on the results of ANOVA analysis. In the final step of the independent samples t-test, the main output parameters in the model include the test statistical observation, the p-value to test the likelihood of difference, and the mean difference value. A significant difference between the two clusters is considered to exist when the p-value is less than 0.05; and vice versa. The flow of the comparative analysis is depicted in detail in Fig.2 .Fig. 2 Analysis and comparison framework Since there is a sample of single buildings in this building classification group, such as cultural buildings. Therefore, it is necessary to verify whether the existing sample size (Table 3 ) can meet the requirements of the t-test. In this study, G power 3.1 was used to carry out a software validation of the existing sample size for the effect size, first selecting in the execution command t test- difference between two independent means, The mode was set to and post hoc analysis, and since this study belongs to the sample size under objective conditions, the effect size d was set to 0.5 is acceptable under the current conditions. The α err prob is generally preset to 0.05, and the output efficacy value Power (1 - β err prob) can be derived after inputting the current sample size. Usually, an output efficacy value greater than 0.8 represents acceptable, and the minimum sample output efficacy value in this study was 0.89, which indicates that the current sample size is valid and sufficient for this study.Table 4. Table 3 Sample size of different types of public buildings on campus Building Type Unrestricted Period(Before the COVID-19) StrictlyRestricted(2020/03/27∼2020/08/31) EastingRestricted(2020/09/01∼2020/12/14) StrictlyRestricted(2020/12/15∼2021/04/25) EastingRestricted(2021/04/26∼2022/02/16) Unrestricted Period(After the COVID-19) Service 765 1413 945 1188 2673 1206 Teaching 510 942 630 792 1782 804 Research 340 628 450 528 1188 536 Sports 255 471 315 396 891 402 Office 170 314 210 264 594 268 Culture 85 157 105 132 297 134 Table 4 The average daily electricity consumption of different types of campus buildings after climate correction Building Type Unrestricted Period/EU(Before the COVID-19) StrictlyRestricted/EU(2020/03/27∼2020/08/31) EastingRestricted/EU(2020/09/01∼2020/12/14) StrictlyRestricted/EU(2020/12/15∼2021/04/25) EastingRestricted/EU(2021/04/26∼2022/02/16) Unrestricted Period/EU(After the COVID-19) Mean/EU Service 6632.48 12156.4 8785 9486.8 9531.2 10767.39 9559.87 Teaching 24091.88 21407.88 25560 23544.53 25185.17 26259.53 24341.49 Research 13691.03 13750.58 14317.32 13737.5 13963.4 13839.67 13883.25 Sports 1433.51 1080.87 1643.04 1226.65 1573.6 1979.58 1489.54 Office 1612.27 1403.09 1429.59 1379.57 1501.1 1620.35 1490.99 Culture 1635.28 1147.35 1500.12 1277.28 1523.65 1864.15 1491.31 3 Analysis 3.1 Basic overview of data After climate-correcting all the data, this paper initially compares the electricity consumption of different types of campus buildings at various stages of the study cycle. Table 4 shows the daily average EU of the six types of campus buildings in this study at six stages. As can be seen from the table, teaching buildings, as the main part of the university campus complex, have the highest EU, with a daily average EU of 21,407.88kwh even during the strict restriction period, followed by research buildings and living buildings, with a total daily average EU of 13,883.25kwh and 9,559.87kwh, respectively. The average daily EU of the three types of buildings, sports, office and culture, is very close, at 1489.54 kwh, 1490.99 kwh and 1491.31 kwh, respectively. As can be seen from the overview of electricity consumption data, teaching and research activities are the main activities on university campuses, and most of the students' or teachers' activities and time are concentrated in the educational buildings, so the EU of teaching buildings is much higher than that of other buildings. In addition, the EU of research buildings is also relatively high, second only to teaching buildings, because research buildings contain many high-powered instruments and equipment, and some equipment may need to run non-stop due to experimental reasons, consuming a lot of electricity in the process of continuous operation. The living buildings are involved in the daily needs of students and faculty such as eating and shopping, and therefore also account for an important part of the energy consumption of campus buildings. Office buildings provide office functions but are not the main functions of the university, while sports and cultural buildings are more often used by students after school, so these three types of buildings account for a relatively low share of the overall electricity consumption of the university campus. And after further comparing the average daily electricity consumption during different restriction periods, it can be found that all other types of the university were affected to some extent during the restriction period, except for research buildings. During the strict restriction period, the EU of teaching, sports, office and cultural buildings decreased significantly, while the EU of living buildings increased, and the academic buildings were not affected much. Due to the large sample size of the daily average EUI, this paper calculates the weekly average of the sample data in order to better show the changes of EU and EUI of buildings in different stages. Fig.3 shows the changes of EUI in different restricted stages of various types of buildings. The red and blue areas represent periods of restriction of different intensities, respectively, and the cyan and gray areas represent non-restricted periods before and after the outbreak. As can be seen from the figure, the EUI of academic buildings is originally higher than that of other types of buildings, and the EUI performance is very smooth in different restriction periods. Life building EUI is second only to academic building, but it is more influenced by the restriction measures, and the EUI change goes to show a more substantial change in different restriction periods, the stronger the restriction period, especially the first strict restriction phase, the overall EUI is inversely proportional to the intensity of the epidemic restriction, while the teaching building EUI is lower than the life building, but it is also influenced by the restriction policy and is proportional to the intensity of the restriction. The EUI performance of sports, office and cultural buildings is low, and the change in EUI picks up somewhat during the unrestricted and loosely restricted periods and improves significantly after the epidemic of exposure to restrictions.Fig. 3 Weekly EUI changes at different stages 3.2 ANOVA and LSD results for EUI of buildings in different periods Because of the large amount of electricity used in various types of buildings, electricity intensity is derived from the constant of electricity consumption divided by floor area, and the results and significance are the same in ANOVA analysis, so this section starts ANOVA analysis with EUI as the base data. Before the analysis of variance, the data has passed the K-S normal distribution test in SPSS. In addition, in order to test the effectiveness of the ANOVA analysis model, the homogeneity test of variance must be carried out to determine which types of buildings have statistical significance in different periods. The test parameters of the dependent variable include, Service, Teaching, Research, Sports, Office, Culture EUI of six types of buildings, and the impact factors are set to a total of six periods from T0 to T5. The results show that the EUI in different periods conforms to the homogeneity of variance test, and only three types of buildings are significant, namely, Teaching (F=10.106, Sig.=0.055), Sports (F=10.106, Sig.=0.055), and Culture (F=58.274, Sig.=0.062). Sig. values of service, research and office of the other three types of buildings are less than 0.05, which does not conform to the homogeneity test of variance, so the statistical results are meaningless. The results of the ANOVA analysis kind show that the sig. values of EUI for all types of buildings in different periods are less than 0.05, indicating that the epidemic restrictions have a limiting effect on the electricity consumption of all types of buildings. Among them, the values of teaching, sports, culture and P of the three types of buildings that conform to the homogeneity of variance are equal to 0.000, which indicates that the power consumption of these three types of buildings in different periods has changed significantly. Further LSD analysis is needed to further explore the most significant differences between these three types of buildings during the restriction period and before the restriction. Table 5 lists the LSD comparison results of three types of buildings in different periods before the epidemic and after the restrictive measures are taken.Table 5 Multiple comparison results of the impact of restriction policies on EUI of Teaching, Sports and Culture buildings in different periods EUI Each period(I) Each period(J) Mean difference(I-J) Standard error Sig. 95% confidence interval Lower limit Upper limit Teaching T0 T1 0.28417* 0.06073 0.000 0.1643 0.4040 T2 0.02600 0.06073 0.669 -0.0938 0.1458 T3 -0.01057 0.06223 0.865 -0.1334 0.1122 T4 -0.01449 0.06120 0.813 -0.1353 0.1063 T5 -0.11600 0.06073 0.058 -0.2358 0.0038 Sports T0 T1 0.09066* 0.02035 0.000 0.0505 -0.1308 T2 0.03451 0.02035 0.092 -0.0056 0.0747 T3 0.12058* 0.02085 0.000 0.0794 0.1617 T4 -0.20202* 0.02051 0.000 -0.2425 -0.1615 T5 -0.08462* 0.02035 0.000 -0.1248 -0.0445 Culture T0 T1 0.06943* 0.00671 0.000 0.0562 0.0827 T2 0.02754* 0.00671 0.000 0.0143 0.0408 T3 0.03322* 0.00688 0.000 0.0197 0.0468 T4 -0.3244* 0.00676 0.000 -0.0458 -0.0191 T5 -0.00820 0.00671 0.223 -0.0214 0.0050 * Significance level<0.05. From the table, it can be seen that the electricity intensity in the restriction period of teaching buildings in T1 phase decreases by 0.28 kwh/ m2/day compared to the EUI in T0 period, and also decreases by 0.09 kwh/ m2/day and 0.07 kwh/ m2/day for sports and cultural buildings, respectively, indicating that the restriction measures have significant effects on the electricity intensity of all three types of buildings. The intensity of sports buildings changed most significantly with the restrictions, decreasing by 0.12kwh/ m2/day in T3 phase, increasing by 0.20kwh/ m2/day in T4 phase during the second period of relaxed restrictions, and improving by 0.08kwh/ m2/day after the restrictions were lifted after the epidemic compared to the period before the restrictions were taken. Cultural buildings received the most intuitive impact under the epidemic restriction intensity, with electricity intensity decreasing by 0.07kwh/ m2/day and 0.03kwh/ m2/day during the two mandatory restriction measures T1 and T3, respectively, until the restriction measures were lifted after the epidemic to return to normal. 3.3 Comparison of strict restrictions and easting restrictions In order to further compare the impact of different restraint intensities on various types of public buildings on campus, this paper uses T-test to further compare the daily average EU and EUI of various types of buildings in different restraint intensity periods during the research period. The results in Table 6 show that the p-values of EU and EUI for different restriction intensities are <0.001 for all building types except living buildings, which are statistically significant and show significant differences. The largest EU daily mean difference is for teaching buildings (2599kwh/day), followed by research (669.6kwh/day), sports buildings (665.4kwh/day), and cultural buildings (416.3), and the smallest EU mean difference is for office buildings (202.8kwh/day), and the mean difference for living buildings is also relatively high, but does not have statistically significant. The above results show that the intensity of the restriction policy is inversely proportional to the EU of other types of public buildings on campus except for living, that is, the higher the intensity of the COVID-19 restrictive policy, the lower the electricity consumption of other types of public buildings on campus except for living. And in the comparison of EUI, it can be found that the largest difference in mean value is for research buildings (0.241kwh/ m2/day), followed by teaching (0.179kwh/ m2/day) and sports buildings (0.174kwh/ m2/day), and the smallest is for office buildings (0.033kwh/ m2/day) and cultural buildings (0.033kwh/ m2/ day). This indicates that the research, teaching, and sports categories have the highest energy saving potential under different intensity of epidemic restriction policies.Table 6 T-test results for EU and EUI in strictly restricted and loosely restricted periods BuildingType EU EUI t-test P-value Mean of differences (kwh/day) t-test P-value Mean of differences (kwh/day) Service 1.869 0.62 755.3 1.869 0.62 0.278 Teaching 8.131 ＜0.001 2599 8.131 ＜0.001 0.179 Research 6.188 ＜0.001 669.6 6.188 ＜0.001 0.241 Sports 20.3 ＜0.001 665.4 20.3 ＜0.001 0.174 Office 4.905 ＜0.001 202.8 4.905 ＜0.001 0.037 Culture 12.59 ＜0.001 416.3 12.59 ＜0.001 0.033 3.4 Comparison before and after the COVID-19 unrestricted period Finally, to understand whether the EU and EUI of each type of public building on campus changed before and after the restriction policy of the New Hall epidemic, this paper further compared the EU and EUI data before and after the unrestricted period using t-tests. The results of the data in Table 7 show that, except for the research and office buildings, the rest of the building types changed significantly before and after the epidemic, with p-values <0.05 and statistically significant. The most significant changes in EU are in living buildings (4135kwh/day) and teaching buildings (2168kwh/day), followed by sports buildings (546kwh/day) and cultural buildings (228.9kwh/day). This indicates that the restriction policy during the New Crown epidemic suppressed people's out-of-home activities and electricity consumption habits to a certain extent. Due to the long-term suppression, people's electricity consumption habits became more frequent than before for a period of time after the lifting of the restriction, resulting in an increase in electricity consumption in the category of living services and buildings. And this is reflected in the mean difference of EUI. The highest mean difference of EUI before and after the epidemic is for living buildings (1.522kwh/ m2/day), followed by teaching (0.149kwh/ m2/day) and sports buildings (0.143kwh/ m2/day). This suggests that the potential for electricity use in living and educational buildings can be considered first in future electricity management decisions.Table 7 T-test results for EU and EUI before the COVID-19 restriction policy was issued and after the restriction was lifted BuildingType EU EUI t-test P-value Mean of differences (kwh/day) t-test P-value Mean of differences (kwh/day) Service 7.965 <0.001 4135 7.965 <0.001 1.522 Teaching 3.558 0.0005 2168 3.558 0.0005 0.149 Research 1.002 0.317 148 1.002 0.317 0.053 Sports 9.118 <0.001 546.1 9.118 <0.001 0.143 Office 0.1001 0.9204 8.088 0.1001 0.9204 0.001 Library 3.582 0.0004 228.9 3.582 0.0004 0.018 4 Discussion 4.1 Impact of Government Restrictions on Electricity Consumption in Public Buildings on Campus To demonstrate that the introduction of restrictions at Twente University's campus is not an isolated reference, and to discuss in more detail the impact of restrictions on electricity consumption in campus public buildings, the campus shutdown index from the Oxford Coronavirus Tracking Project is cited to indicate the intensity of campus restrictions over time [52],The index reflects the different levels of the government’s policy on school closures in the region, mainly including no measures (0), recommended closures (1), conditional closures (2), and full closures (3). The higher the value, the more restrictive the greater the strength. The government policy for schools provides a more comprehensive understanding of the impact of the intensity of restrictions from different dimensions on the electricity consumption of public buildings on university campuses. Fig.4 shows the intensity of restrictions on schools in the Dutch region during the New Crown epidemic, and Fig.5 shows the change in weekly average EU for Twente universities. By comparing Fig.4 and Fig.5, it can be found that from mid-February 2022, with the Dutch government's restrictive policy of total school closures, there was a significant downward trend in EU for all types of buildings on campus, except for research buildings, until June when it began to gradually rebound. This strict restriction occurred three times, in February 2020, December 2020, and December 2021, during which all types of buildings on campus showed a downward trend, but as the number of times increased, the second and third strict restrictions gradually shortened the recovery time of electricity consumption compared to the first. It indicates that the combined online + offline education model developed by the campus has led to the beginning of a gradual reduction in the impact of the epidemic on various public buildings on campus.Fig. 4 Netherlands Regional school policy restricts intensity amid COVID-19 Fig. 5 Weekly average EU for public buildings on the University of Twente campus 4.2 Changes in campus electricity consumption before and after covid-19 COVID-19 has left people living under restrictive policies for the past two years. The prolonged restriction policy has potentially affected people's life and electricity consumption behavior while affecting energy consumption in various buildings. In the case of this study, the EU changes of residential buildings before and after the epidemic were the most obvious. In the T-test results of 3.4, the electricity consumption of residential buildings increased by 4135kwh/day compared with that before the epidemic, which indicates that the use of residential buildings The duration and utilization rate have increased, and this change has a certain relationship with the utilization rate of public buildings for living. Although detailed data on the usage rates of public buildings containing campus amenity categories are not available in detail, this paper uses the data on the usage rates of various amenity services in the Dutch region from Google's Community Mobility Report on New Coronary Pneumonia as an indirect evidence reference for the usage rates of amenity buildings in this study [53]. Fig.6 shows the relative change in the number of visitors in various types of restaurants, shopping centers, cafes, and other public places in the Netherlands region from January 2020.to the present with respect to the date.Fig. 6 Relative change in visitor numbers to retail and entertainment venues It can be seen that from February 26, 2022, when the restriction policy was fully lifted in the Netherlands, the number of visitors to living public places began to rise rapidly and by June 30, 2022, had fully surpassed the pre-epidemic period. On the one hand, this is due to the fact that although personal protection habits were developed, many students' awareness of protection began to decrease after the lifting of restrictions[54], and on the other hand, the curfew policy implemented in the Netherlands during the epidemic had a direct impact on the hours of use of such living buildings, which led to a surge in the number of visitors as the hours of operation increased after the lifting of the ban, and as people broke out after being suppressed by the long-term restriction policy. This reason led to an increase in building occupancy limits and an increase in electricity consumption in such living buildings, which explains the spike in EU in campus living buildings before and after the outbreak in Fig. 5. 4.3 Implication and Limitations The restrictions imposed by the government during COVID-19 have led to changes in the movement of people in public places and the frequency of use of various buildings, and the frequency of use of buildings and the number of visitors has a direct impact on the energy consumption of buildings [41]. And this paper analyzes the EU and EUI changes of different types of campus public buildings at different stages during the epidemic in real time by interpreting the restriction policies of the government and schools, especially analyzing the link of joining the analysis before and after the epidemic, and studying the impact of electricity consumption of various types of campus public buildings before and after the epidemic in a more comprehensive way, which has important reference significance for the government in formulating campus restriction measures and energy saving policies under similar climatic conditions. Although it is partly indicated that the popularity of vaccines has eased the epidemic situation in schools and other institutions[55], the virus is still mutating, and people may be restricted by other activities in the future. Therefore, the opening hours or energy restrictions of buildings should be appropriately adjusted according to the changing patterns of energy consumption of different building types during the epidemic to ensure that more energy-efficient policy measures are developed in the face of similar situations. In addition, with the restriction of the new crown epidemic, online education became one of the main tools of the campus response policy, but according to the available research results, although the energy consumption performance of the teaching buildings decreased somewhat during the epidemic, the overall decrease was small corresponding to their large floor area. But on the other hand, it also shows that online education is still limited to reduce the electricity consumption of education type buildings, and the energy saving potential of education buildings can still be improved. If the energy consumption of student course types can be combined to develop online + offline teaching models to accommodate the relaxed restrictions, it may be possible to improve the energy utilization of campus buildings[37], [40]. The results of the current study show that campus buildings have great potential for energy savings [36], [41], The data cut-off date of this study is June 30, 2022, and the study period includes all the time periods before the outbreak, during the restriction and after the lifting of the restriction, and the actual electricity consumption data of various public buildings on campus are analyzed with the real-time restriction policy and intensity, which provides an exhaustive and complete case study of the energy consumption transformation of campus-type buildings during the outbreak with important practical significance. In addition, by comparing the EU and EUI changes of various campus buildings under different restrictions at the government and school levels, the reasons for the changes in people's electricity consumption behavior in different periods are revealed, which will provide evidence and support for energy managers to deal with electricity domination during the epidemic period. Of course, there are certain limitations in this study, such as the floor area data measured and calculated using QGIS and Google satellite maps may not be accurate enough, and also the study only includes the public type buildings inside the campus, if the analysis can be done in conjunction with the dormitory type buildings, such as Zhou et al.'s assessment of the change in electricity consumption in campus dormitories during the COVID-19[56], the change in building energy consumption of a complete campus living system can be better assessed. In addition, this study only analyzed the impact of the epidemic restriction policy on the electricity consumption of campus buildings between January 1, 2020 and June 30, 2022, which has certain timeliness and limitations. Moreover, since the climatic conditions in different regions also have an impact on the use of different types of buildings, it is necessary to investigate and analyze campus buildings in other climatic regions in the future to improve the study. 5 Conclusion Based on the building energy consumption database of Twente University in the Netherlands, this study compared and analyzed the EU and EUI of different types of buildings at the university at various stages throughout the epidemic period. the impact of restricting policy intensity on various types of buildings on campus was revealed through analysis of variance, LSD, and t-test, and the energy saving potential and energy consumption change patterns of different types of buildings were explored. the main conclusions are as follows.1. In the results of ANOVA analysis shows that the epidemic restriction intensity has a significant effect on three types of buildings, namely teaching, culture and sports, especially under the first restriction policy intensity, the electricity intensity of the three types of buildings decreased by 0.28, 0.09 and 0.07kwh/ m2/day respectively. LSD results show that cultural and sports buildings fluctuate with the intensity of restrictive policies indicating that these two types of buildings are more sensitive to restrictive policies. 2. In the supplemental t-test analysis results, except for living buildings, EU and EUI of teaching, academic, sports, office, and cultural buildings are inversely proportional to the epidemic restriction intensity, i.e., the higher the restriction intensity, the lower the EU and EUI. 3. After the epidemic restriction policy was lifted compared to before the restriction policy was implemented, people's electricity consumption behavior habits changed under the influence of the long-term restriction policy, and the EU and EUI of living, teaching, sports and cultural buildings on campus improved significantly. Especially, the average daily EU of living class buildings improved by 4135kwh/day and EUI improved by 1.522kwh/ m2/day. Although the government has now lifted all restrictions, wearing masks and maintaining social distance has become a habit for many people. After experiencing COVID-19, the surge in electricity consumption of various buildings on campus may only last for a period of time, and people will return to normal after they release their suppressed emotions, or people have completely changed their electricity consumption habits due to the epidemic restriction policy. This state of affairs will continue, but confirming this speculation will take longer and follow-up research to complete. Therefore, the main objective is to adjust the electricity distribution policy according to the existing findings in order to optimize the city's electrical energy load and to improve the efficiency of building energy consumption. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ==== Refs References 1 WHO, Report of the WHO-China Joint Mission on Coronavirus Disease 2019(COVID-19), in, 2020. 2 WHO, COVID-19 - China, in, 2020-1-5 3 Kabir M. Afzal M.S. Khan A. Ahmed H. COVID-19 economic cost; impact on forcibly displaced people Travel Medicine and Infectious Disease 101661 2020 4 Norouzi N. Zarazua de Rubens G. Choupanpiesheh S. Enevoldsen P. When pandemics impact economies and climate change: Exploring the impacts of COVID-19 on oil and electricity demand in China Energy Res Soc Sci 68 2020 101654 5 Tian H. Liu Y. Li Y. Wu C.H. Chen B. Kraemer M.U.G. Li B. Cai J. Xu B. Yang Q. Wang B. Yang P. Cui Y. Song Y. Zheng P. Wang Q. Bjornstad O.N. Yang R. Grenfell B.T. Pybus O.G. Dye C. An investigation of transmission control measures during the first 50 days of the COVID-19 epidemic in China Science 368 6491 2020 638 642 32234804 6 Capano G. Howlett M. Jarvis D.S.L. Ramesh M. Goyal N. Mobilizing Policy (In)Capacity to Fight COVID-19: Understanding Variations in State Responses Policy Soc 39 3 2020 285 308 35039722 7 Sun G.Q. Wang S.F. Li M.T. Li L. Zhang J. Zhang W. Jin Z. Feng G.L. Transmission dynamics of COVID-19 in Wuhan, China: effects of lockdown and medical resources Nonlinear Dyn 101 3 2020 1981 1993 32836805 8 Kabilan M.K. Annamalai N. Online teaching during COVID-19 pandemic: A phenomenological study of university educators’ experiences and challenges Studies in Educational Evaluation 74 2022 9 Shakibaei S. de Jong G.C. Alpkokin P. Rashidi T.H. Impact of the COVID-19 pandemic on travel behavior in Istanbul: A panel data analysis Sustain Cities Soc 65 2021 102619 10 Ren M. Park S. Xu Y. Huang X. Zou L. Wong M.S. Koh S.Y. Impact of the COVID-19 pandemic on travel behavior: A case study of domestic inbound travelers in Jeju Korea, Tour Manag 92 2022 104533 11 Wang Q. Liu C.-Q. Zhao Y.-F. Kitsos A. Cannella M. Wang S.-K. Han L. Impacts of the COVID-19 pandemic on the dairy industry: Lessons from China and the United States and policy implications Journal of Integrative Agriculture 19 12 2020 2903 2915 12 Fisher Ke J.-Y. Otto J. Han C. Customer-Country diversification and inventory efficiency: Comparative evidence from the manufacturing sector during the pre-pandemic and the COVID-19 pandemic periods Journal of Business Research 148 2022 292 303 13 Tang X. Li Z. Hu X. Xu Z. Peng L. Self-correcting error-based prediction model for the COVID-19 pandemic and analysis of economic impacts Sustainable Cities and Society 74 2021 14 Norouzi N. Post-COVID-19 and globalization of oil and natural gas trade: Challenges, opportunities, lessons, regulations, and strategies Int J Energy Res 45 10 2021 14338 14356 34219899 15 Garrido-Cumbrera M. Foley R. Brace O. Correa-Fernandez J. Lopez-Lara E. Guzman V. Gonzalez Marin A. Hewlett D. Perceptions of Change in the Natural Environment produced by the First Wave of the COVID-19 Pandemic across Three European countries Results from the GreenCOVID study, Urban For Urban Green 64 2021 127260 16 Ravindra K. Singh T. Vardhan S. Shrivastava A. Singh S. Kumar P. Mor S. COVID-19 pandemic: What can we learn for better air quality and human health? J Infect Public Health 15 2 2022 187 198 34979337 17 D’Oca S. Hong T. Langevin J. The human dimensions of energy use in buildings: A review Renewable and Sustainable Energy Reviews 81 2018 731 742 18 Maljaee S.S. Khadem Sameni M. Investigating factors affecting university students' use of subway before and after COVID-19 outbreak: A case study in Tehran J Transp Geogr 105 2022 103461 19 Hong T. Yang L. Hill D. Feng W. Data and analytics to inform energy retrofit of high performance buildings Applied Energy 126 2014 90 106 20 Kang H. An J. Kim H. Ji C. Hong T. Lee S. Changes in energy consumption according to building use type under COVID-19 pandemic in South Korea Renew Sustain Energy Rev 148 2021 111294 21 Abdeen A. Kharvari F. O'Brien W. Gunay B. The impact of the COVID-19 on households' hourly electricity consumption in Canada Energy Build 250 2021 111280 22 Surahman U. Hartono D. Setyowati E. Jurizat A. Investigation on household energy consumption of urban residential buildings in major cities of Indonesia during COVID-19 pandemic Energy and Buildings 261 2022 23 Geraldi M.S. Bavaresco M.V. Triana M.A. Melo A.P. Lamberts R. Addressing the impact of COVID-19 lockdown on energy use in municipal buildings: A case study in Florianopolis Brazil, Sustain Cities Soc 69 2021 102823 24 Huang Z. Gou Z. Electricity consumption variation of public buildings in response to COVID-19 restriction and easing policies: A case study in Scotland, U.K Energy Build 267 2022 112149 25 Rodriguez D. Urbieta I.R. Velasco A. Campano-Laborda M.A. Jimenez E. Assessment of indoor air quality and risk of COVID-19 infection in Spanish secondary school and university classrooms Build Environ 226 2022 109717 26 Mok K.H. Xiong W. Ke G. Cheung J.O.W. Impact of COVID-19 pandemic on international higher education and student mobility: Student perspectives from mainland China and Hong Kong Int J Educ Res 105 2021 101718 27 Binali T. Tsai C.-C. Chang H.-Y. University students’ profiles of online learning and their relation to online metacognitive regulation and internet-specific epistemic justification Computers & Education 175 2021 28 Zhu W. Liu Q. Hong X. Implementation and Challenges of Online Education during the COVID-19 Outbreak: A National Survey of Children and Parents in China Early Child Res Q 61 2022 209 219 35919864 29 Muñoz-Carril P.-C. Hernández-Sellés N. Fuentes-Abeledo E.-J. González-Sanmamed M. Factors influencing students’ perceived impact of learning and satisfaction in Computer Supported Collaborative Learning Computers & Education 174 2021 30 Li J. Che W. Challenges and coping strategies of online learning for college students in the context of COVID-19: A survey of Chinese universities Sustain Cities Soc 83 2022 103958 31 Bartolucci A. Templeton A. Bernardini G. How distant? An experimental analysis of students' COVID-19 exposure and physical distancing in university buildings Int J Disaster Risk Reduct 70 2022 102752 32 Wu Y. Yin G. Zhang Y. Experience and Perceptions of Chinese University Students Regarding the COVID-19 Pandemic: A Qualitative Analysis Front Public Health 10 2022 872847 33 Martin A.J. Ginns P. Collie R.J. University students in COVID-19 lockdown: The role of adaptability and fluid reasoning in supporting their academic motivation and engagement Learn Instr 83 2023 101712 34 Heumann E. Helmer S.M. Busse H. Negash S. Horn J. Pischke C.R. Niephaus Y. Stock C. Anxiety and depressive symptoms of German university students 20 months after the COVID-19 outbreak - A cross-sectional study J Affect Disord 320 2023 568 575 36220498 35 Sun L. Kaufman M.F. Sirk E.A. Durga S. Mahowald N.M. You F. COVID-19 impact on an academic Institution's greenhouse gas inventory: The case of Cornell University Journal of Cleaner Production 363 2022 36 Gaspar K. Gangolells M. Casals M. Pujadas P. Forcada N. Macarulla M. Tejedor B. Assessing the impact of the COVID-19 lockdown on the energy consumption of university buildings Energy Build 257 2022 111783 37 Yin Z. Jiang X. Lin S. Liu J. The impact of online education on carbon emissions in the context of the COVID-19 pandemic - Taking Chinese universities as examples Appl Energy 314 2022 118875 38 Franceschini P.B. Neves L.O. A critical review on occupant behaviour modelling for building performance simulation of naturally ventilated school buildings and potential changes due to the COVID-19 pandemic Energy and Buildings 258 2022 39 Nasir S.N.S. Ludin N.A. Radzi A. Junedi M.M. Ramli N. Marsan A. Mohd Z.F.A. Roslan M.R. Taip Z.A. Lockdown impact on energy consumption in university building Environ Dev Sustain 2022 1 20 40 Mokhtari R. Jahangir M.H. The effect of occupant distribution on energy consumption and COVID-19 infection in buildings: A case study of university building Build Environ 190 2021 107561 41 Gui X. Gou Z. Zhang F. Yu R. The impact of COVID-19 on higher education building energy use and implications for future education building energy studies Energy Build 251 2021 111346 42 G.o.t. Netherlands, Man diagnosed with coronavirus (COVID-19) in the Netherlands, in, 2020.2.27. 43 Rijksoverheid, Total number of COVID-19 patients reported as deceased, (2022.8.18). 44 U.o. Twente, Latest Coronavirus Updates （COVID-19）, (2022.2.17). 45 U.o. Twente, University of Twente energy data, in, 2022. 46 U.o. Twente, CAMPUS MAP & CAMPUS BUILDINGS, (2022). 47 Auffhammer M. Mansur E.T. Measuring climatic impacts on energy consumption: A review of the empirical literature Energy Economics 46 2014 522 530 48 Elkhafif M. An iterative approach for weather-correcting energy consumption data Energy Economics 18 3 1996 221 230 49 K.N.M. Insitituut, Daggegevens van het weer in Nederland, in, 2022. 50 Huang Z. Gou Z. Cheng B. An investigation of outdoor thermal environments with different ground surfaces in the hot summer-cold winter climate region Journal of Building Engineering 27 2020 51 Kontokosta C.E. Spiegel-Feld D. Papadopoulos S. The impact of mandatory energy audits on building energy use Nature Energy 5 4 2020 1 8 52 O.W.i. Data, COVID-19: School and Workplace Closures, in 2022 53 O.W.i. Data, COVID-19: Google Mobility Trends, in 2022 54 Zhang N. Liu X. Jin T. Zhao P. Miao D. Lei H. Su B. Xue P. Xie J. Li Y. Weakening personal protective behavior by Chinese university students after COVID-19 vaccination Build Environ 206 2021 108367 55 Ghaffarzadegan N. Effect of mandating vaccination on COVID-19 cases in colleges and universities Int J Infect Dis 123 2022 41 45 35985570 56 Zhou T. Luo X. Liu X. Liu G. Li N. Sun Y. Xing M. Liu J. Analysis of the influence of the stay-at-home order on the electricity consumption in Chinese university dormitory buildings during the COVID-19 pandemic Energy Build 277 2022 112582

0

PMC9753509

NO-CC CODE

2022-12-16 23:25:45

no

Energy Build. 2023 Jan 15; 279:112723

utf-8

Energy Build

10.1016/j.enbuild.2022.112723

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)01558-0 10.1016/S0140-6736(21)01558-0 Perspectives When art mirrors planetary health Lucas Tamara 17 7 2021 17-23 July 2021 17 7 2021 398 10296 202202 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcThe future of the arts in many countries has been plunged into uncertainty during the COVID-19 pandemic. Many museums and galleries have been closed for long periods and their funding hangs by a thread. Like other arts organisations, the planned exhibition schedule at Compton Verney in Warwickshire, UK, was thrown into disarray in 2020. It is therefore remarkable that Compton Verney has recently reopened with a poignant and moving exhibition that could not be more pertinent to COVID-19 and the plight of the planet, despite being conceived and commissioned in 2019. Rebecca Louise Law's Seasons is an installation of dried seasonal flowers that takes the visitor on a journey of a calendar year, capturing nature, climate change, and the planet as experienced by the artist. Law used her own collection of preserved flowers, and plants growing in the extensive grounds of Compton Verney for the installation. They are strung together with fine copper wire, a feat of immense skill and patience. The show begins with Spring: an uplifting room with tens of thousands of dried flowers suspended from the ceiling as a gentle field of daffodils and crocuses. The analogy with planetary health is instant and striking—the raw health of the natural planet is recorded in this exhibition through a lens of art. The delicacy and beauty of the installation intimate life and death, and hit the visitor with a visceral power. Moving through Spring towards Summer is tinged with tragedy and pathos; the dead poppy heads, perfectly preserved pink roses and hydrangeas, and a scented bed of lavender are physically and emotionally overwhelming. The exhibition as a whole also feels like a metaphor for the past year: fragility and upheaval yet the seasons ebb and flow, and life goes on. © 2021 Rebecca Louise Law, Seasons © Compton Verney, 2021 2021 The exhibition's narrative continues through Law's diaries, intricate drawings, and notes that document the development and progress of Seasons. She relates the sudden doubt and chaos of the COVID-19 pandemic that was also accompanied by an appreciation of life and nature. In Law's experience, this ranged from the vast landscapes of Wales to the minutiae of mice eating stored Autumn apples in a shed, and the joy of an early morning walk on the beach or a conversation with a child. Occasional references to COVID-19— the knitting of hearts and the sense of sadness and horror at global events—remind the visitor that there was nothing routine about the past year, although the seasons continue. A short film taken from Law's Instagram stories compresses the year into a visual journey—life comes around again, and renewal is part of nature's seasons, seamlessly accommodating social media in contrast to the raw materials of the rest of the show. The final room of the exhibition concludes with a devastating essay by the artist. Law makes a plea for people to pause their uncontrolled consumption and the human activity that are destroying the planet, and to recognise and respect the beauty of nature and its ability to heal through living in more harmony with and not exploiting the natural world. She puts into words the feelings visitors have walking around her installation: the awe of nature and the diminutive size of human life alongside all other creatures and plants. That this exhibition is so relevant to a world that has suffered and is starting to heal in some places is not only serendipity, but also reflects the brilliance of the curators and leadership at Compton Verney, who appreciate the changing planet and dangerous trajectory that humans are following and are using their platform to great effect. When visual arts and culture can so successfully convey planetary health and the forces at work, they plant a seed of hope that human civilisation is not beyond repair. Seasons could not be more perfect as a first exhibition to see when returning to galleries after an enforced absence. Both shattering and sublime, it will stay with me for a long time to come. Rebecca Louise Law: Seasons Compton Verney, Warwickshire, UK, until Aug 30, 2021 https://www.comptonverney.org.uk/thing-to-do/rebecca-louise-law-seasons/ © 2021 Rebecca Louise Law, Seasons © Compton Verney, 2021 2021

0

PMC9753558

NO-CC CODE

2022-12-16 23:25:47

no

Lancet. 2021 Jul 17 17-23 July; 398(10296):202

utf-8

Lancet

10.1016/S0140-6736(21)01558-0

oa_other

==== Front Rev Neurol (Paris) Rev Neurol (Paris) Revue Neurologique 0035-3787 0035-3787 Elsevier Masson SAS. S0035-3787(21)00794-3 10.1016/j.neurol.2021.12.008 Original Article Long-term outcomes after NeuroCOVID: A 6-month follow-up study on 60 patients Chaumont H. abc* Meppiel E. d Roze E. ce Tressières B. f de Broucker T. d Lannuzel A. abcf on behalf of the contributors to the French NeuroCOVID registry 1 a Service de neurologie, Centre hospitalier universitaire de la Guadeloupe, Pointe-à-Pitre/Abymes, Guadeloupe b Faculté de Médecine de l’université des Antilles, Pointe-à-Pitre, Guadeloupe c Sorbonne Université, Institut national de la Santé et de la Recherche médicale, U 1127, CNRS, Unité Mixte de Recherche (UMR) 7225, Institut du Cerveau, Paris, France d Service de Neurologie, Centre hospitalier de Saint-Denis, Hôpital Delafontaine, Saint-Denis, France e Département de Neurologie, AP–HP, Hôpital de la Pitié-Salpêtrière, Paris, France f Centre d’investigation Clinique Antilles Guyane, Inserm CIC 1424, Pointe-à-Pitre, Guadeloupe ⁎ Corresponding author. Department of Neurology, University Hospital of Guadeloupe, 97139, Pointe-à-Pitre/Abymes, Guadeloupe. 1 Contributors to the French NeuroCOVID registry are listed to Supplementary Table S1. 6 1 2022 January-February 2022 6 1 2022 178 1 137143 27 9 2021 20 11 2021 27 12 2021 © 2021 Elsevier Masson SAS. All rights reserved. 2021 Elsevier Masson SAS Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Background and purpose Long-term outcomes after neurological manifestations due to COVID-19 are poorly known. The aim of our study was to evaluate the functional outcome and identify the risk factors of neurologic sequelae after COVID-19 associated with neurological manifestations (NeuroCOVID). Methods We conducted a multi-center observational study six months after the acute neurological symptoms in patients from the French NeuroCOVID hospital-based registry. Results We obtained data on 60 patients. NeuroCOVID had a negative impact on the quality of life (QoL) of 49% of patients. Age was a predictor of residual QoL impairment (OR: 1.06, 95% CI: 1.01–1.13, p = 0.026). At six months, a significant residual disability was found in 51.7% of patients, and impaired cognition in 68.9% of cases. The main persistent neuropsychiatric manifestations were a persistent smell/taste disorder in 45% of patients, memory complaints in 34% of patients, anxiety or depression in 32% of patients. Conclusions NeuroCOVID likely carries a high risk of long-term neuropsychiatric disability. Long-term care and special attention should be given to COVID-19 patients, especially if they had neurological manifestations during acute infection. Keywords COVID-19 Neurological manifestation NeuroCOVID Post-COVID Syndrome Long-term outcomes ==== Body pmc1 Introduction SARS-CoV-2 infection is characterized by many persistent symptoms or sequelae known beyond two months after the acute phase of mild to severe COVID-19 [1], [2]. Six months after acute infection, 76% of hospital-discharged patients reported persistent symptoms [3]. The most common symptoms are fatigue, muscle weakness, sleep difficulties [3], post-exertional malaise [4], and cognitive dysfunction [4], [5]. Patients have an increased risk of mood and anxiety disorders within three months after COVID-19 infection [6]. The 6-month prevalence of self-reported olfactory dysfunction was evaluated at 60% in mild symptomatic COVID-19 patients [7]. During the acute phase of COVID-19, up to 36% of patients develop neurological manifestations [8]. Typically they consist of mild or nonspecific neurological symptoms such as confusion, dizziness, headache, and myalgia [8]. Less common severe diseases such as encephalopathy, encephalitis, stroke, Guillain-Barre syndrome, cranial nerves palsy, acute cerebellar ataxia, myoclonus, and mixed neurological disorders may also occur [9], [10], [11], [12], [13], [14], [15]. The long-term outcome of patients with neurological manifestations related to SARS-CoV-2 (NeuroCOVID) is poorly known. To evaluate functional outcomes and identify risk factors of sequelae after COVID-19 associated with neurological manifestations, we conducted a multicenter observational study six months after the acute neurological episode in patients from the French NeuroCOVID multicenter registry [16]. 2 Methods 2.1 Patients and study design We included patients from the multicenter registry of 222 adult patients admitted for neurological manifestations associated with COVID-19. In this hospital-based study established during the first French epidemic wave in March and April 2020 [16], patients were considered to have NeuroCOVID in the presence of de novo neurological manifestations occurring five days before to 35 days after the first symptoms of COVID-19. The diagnosis of COVID-19 required either a positive SARS-CoV-2 real-time reverse transcriptase PCR assay result on a nasopharyngeal sample or typical clinical history and chest computed tomographic scan. The illness was classified as mild, moderate, severe, and critical according to the United States National Institutes of Health criteria [17]. Mild Illness for any infectious signs and symptoms of COVID-19 but no shortness of breath, dyspnea, or abnormal chest imaging; Moderate Illness for respiratory clinical or radiological involvement with saturation of oxygen (SpO2) ≥ 94% on room air; Severe Illness for SpO2 < 94% on room air, ratio of arterial partial pressure of oxygen to fraction of inspired oxygen (PaO2/FiO2) < 300 mm Hg, respiratory frequency > 30 breaths/min, or lung infiltrates > 50%; Critical Illness for respiratory failure, septic shock, and/or multiple organ dysfunction. Neurologic manifestations were classified as either involving the central nervous system (CNS) or peripheral nervous system (PNS): encephalopathy (38.3%), stroke (28.4%), encephalitis and meningitis (10.9%), other CNS (7.7%), Guillain-Barre syndrome (6.8%), other PNS (9%) and mixed CNS and PNS (5.9%). Anosmia (3.2%) and ageusia (1.8%) were certainly underestimated as they may have been considered minor manifestations and, as such, not consistently recognized as neurological manifestations. All centers involved in this registry were contacted to participate in the follow-up study. Patients who died during the acute phase of COVID-19 and those followed in centers that declined their participation were not included. Six months after the onset of neurological signs, clinical assessment was scheduled during a telephone interview or by routine medical advice. The study was approved by the French Ile de France VIII ethic committee (2020-A01882-37). 2.2 Data collection and neurological investigations We collected the following data: age, sex, clinical characteristics of COVID-19 defined in the multicenter registry (illness severity, neurological comorbidities, type, and chronology of neurologic manifestations). The severity of disability was graded using the modified Rankin Scale (mRS) [18]. Long-term residual disability was defined as mRS ≥ 2 (level 2 corresponding to a slight disability, able to look after own affairs without assistance, but unable to carry out all previous activities). The 3-level version of the EuroQol five-dimension scale (EQ-5D-3L) was used to measure the quality of life [19]. EQ-5D-3L consists of five questions and a health-related visual analogue scale (VAS). In these questions, the respondents are asked to rate their current health state on five dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) into one of three levels (no problems (1), some problems (2) or a lot of problems (3)). The answers are transformed into a utility value, describing the respondent's health, where utility = 0 equals death and utility = 1 equals perfect health. For example, the answer pattern 11111 would result in an ideal health state (utility = 1), whereas 33333 would result in the worst possible health state. Our study considers a significant alteration of the quality of life if more than one question is at least two points (example: 12211, 21112). The EQ-VAS records the patient's self-rated health on a vertical visual analogue scale where the endpoints are labeled “Best imaginable health state” and “Worst imaginable health state.” The VAS is a quantitative measure of health outcome that reflects the patient's judgment. Evaluation of cognition and mood included Montreal Cognitive Assessment Blind [20] (MoCA-Blind), 35-item version of Cognitive Difficulties Scale (CDS) [21], and Hospital Anxiety and Depression Scale (HADS) [22]. The MoCA-Blind is a version of the MoCA test without the visual elements (first four items). This score evaluates memory, attention, language, abstraction, delayed recall, and orientation. Scoring ranges from 0 to 22, with higher scores indicating better performance. A persistent cognitive dysfunction is defined by a MoCA-Blind score < 19. The 35-item version of CDS is used to assess memory complaints et was considered abnormal if ≥ 15. HADS is a 14-item inventory: seven to evaluate depression and seven to assess anxiety. Each item scores 0–3. After summing up scores, a total score of 0–7 was taken as “no depression/anxiety”, 8–10 was taken as “moderate depression/anxiety”, and 11–21 was taken as “severe depression/anxiety”. Taste and Smell Survey (TSS) [23] was used to detect persistent olfactory or gustatory dysfunction. Persistent olfactory or gustatory dysfunction was present if the olfactory score or gustatory score were ≥ 1. 2.3 Statistical analysis Quantitative variables were summarized as median with interquartile range (IQR) and compared across groups using the Wilcoxon test. Categorical data were expressed as a percentage and compared between groups using the Chi-square test or Fisher exact test, depending on the sample size. Logistic regression models were used to identify factors associated with long-term residual disability, significant alteration of the quality of life, memory complaints, persistent cognitive dysfunction, moderate or severe depression/anxiety, and persistent olfactory or gustatory dysfunction. Factors investigated were the demographic characteristics, neurologic comorbidities, types of neurological manifestations, occurence of neurologic manifestations relative to COVID-19 symptoms, neurological symptoms, and severity of COVID-19. All statistical analyses were performed using R 4.0.2, and significance was considered at the level 5%. 2.4 Data availability Some data will be made available from the corresponding author, upon reasonable request. The data are not publicly available because they contain information that could compromise the privacy of our patients. 3 Results Sixty patients among the 194 who survived a confirmed diagnosis of COVID-19 were included in this study performed within a median of 6 months (IQR 6–7) after the acute phase of NeuroCOVID (Fig. 1 ). Fifty patients (83.3%) had central nervous system (CNS) involvement, five (8.3%) had peripheral nervous system (PNS) disorder. For five of them (8.3%), the mechanism was undetermined (Table 1 ). Twenty-six patients (43.3%) had a severe or critical NeuroCOVID and ten (16.7%) had a history of neurological comorbidities: stroke (n  = 3), neurodegenerative disorder (n  = 3), epilepsy (n  = 2), sequels of Guillain-Barre syndrome (n  = 1), axonal sensitive neuropathy (n  = 1). No difference was observed in patients’ demographic and neurological characteristics included in the follow-up study than those from the initial registry (Supplementary Table S2). The follow-up was done by phone interview (n  = 49) or during a medical visit (n  = 11). Results are shown in Table 2 .Fig. 1 Enrolment and outcome in the French NeuroCOVID multicenter registry. Table 1 Demographic and neurological characteristics. Table 1At acute phase n (%) N 60 Age, median (IQR) 66 (55-73) Male 36 (60.0%) Neurological comorbidities  No 50 (83.3%)  Yesa 10 (16.7%) Severity of COVID-19  Mild/Moderate 34 (56.7%)  Severe/Critical 26 (43.3%) Neurological manifestations in the acute phase  Stroke 18 (30.0%)  Encephalopathy 18 (30.0%)  Encephalitis-Meningitis 8 (13.3%)  Other CNSb 6 (10.0%)  PNSc 5 (8.3%)  Undetermined mechanism 5 (8.3%)  Anosmia 4 (6.7%)  Ageusia 3 (5.0%) Type of stroke  Ischemic stroke 14 (23.3%)  Transient ischemic attack 3 (5.0%)  Intracerebral hemorrhage 1 (1.7%)  Cerebral venous thrombosis 0 (0%) Abbreviations: IQR: Inter Quartile Range; CNS: central nervous system; PNS: peripheral nervous system. a Stroke (n = 3), neurodegenerative disorder (n = 3), epilepsy (n = 2), others (n = 2). b Transient alteration of consciousness (n = 3), isolated seizure (n = 2), and generalized myoclonus with cerebellar ataxia (n = 1). c Peripheral complications of intensive care unit (n = 2), Guillain-Barre syndrome (n = 3). Table 2 Characteristics of NeuroCOVID patients at the follow-up evaluation. Table 2 n/N % 95% CI Disability (mRSa)  Score - median (IQR) 2 (1–3)  0–1 29/60 48.3 35.2–61.6  2–5 31/60 51.7 38.4–64.8 Quality of live (EQ-5D-3Lb)  Not altered 26/51 51.0 36.6–65.2  Altered 25/51 49.0 34.8–63.4 Memory complain (CDS)  Score - median (IQR) 11 (5–18)  Normal (<15) 33/50 66.0 51.2–78.8  Abnormal (≥15) 17/50 34.0 21.2–48.8 Cognitive assessment (MoCA-Blind)  Score - median (IQR) 17 (14–19)  Normal (≥19) 14/45 31.1 18.2–46.6  Abnormal (<19) 31/45 68.9 53.4–81.8 Anxiety - Depression (HADSc)  Anxiety score - median (IQR) 5 (1–9)  Depression score - median (IQR) 2 (1–6)  Not altered 34/50 68.0 53.8–80.5  Altered 16/50 32.0 19.5–46.7 Taste and Smell Survey (TSSd)  Taste score - median (IQR) 0 (0–4)  Smell score - median (IQR) 0 (0–3)  Not altered 28/51 54.9 40.3–68.9  Altered 23/51 45.1 31.1–59.7 Abbreviations: 95% CI: 95% Confidence Interval; IQR: Inter Quartile Range; mRS: modified Rankin Scale; EQ-5D-3L: 3-level version of the EuroQol five-dimension scale; CDS: 35-item version of Cognitive Difficulties Scale; MoCA-Blind: Montreal Cognitive Assessment Blind; HADS: Hospital Anxiety and Depression Scale; TSS: Taste and Smell Survey. a 0: no symptoms at all; 1: no significant disability despite symptoms; able to carry out all usual duties and activities; 2: slight disability; unable to carry out all previous activities, but able to look after own affairs without assistance; 3: moderate disability; requiring some help, but able to walk without assistance; 4: moderately severe disability; unable to walk without assistance and unable to attend to own bodily needs without assistance; 5: severe disability; bedridden, incontinent and requiring constant nursing care and attention. b Significant alteration when more than one question is at least two points (example: 12211, 21112…). c Altered when at least one score, depression or anxiety was > 7. d Altered when at least one score, taste or smell was ≥ 1. Of the 60 patients, 31 (51.7%) had a residual disability (mRS ≥ 2). Slight disability (mRs = 2) was found in ten patients (16.7%) who were able to look after their affairs without assistance but unable to carry out all previous activities. Disability was moderate in 11 patients (18.3%) (mRS = 3). These patients required some help but able to walk without assistance. Ten (16.6%) had a severe residual disability, two of them (3%, mRS = 4) were unable to walk without assistance and unable to attend to their own bodily needs without assistance, and eight (13.3%, mRS = 5) were bedridden and required constant nursing care and attention. Severely disabled patients could not perform tests or questionnaires concerning the quality of life, memory complaints, cognitive and sensory deficit. NeuroCOVID had a negative impact on the quality of life in 49% of cases (25/51 patients). The most altered sub-scores were mobility, with 35% of patients reporting having problems walking, and pain with 43% having moderate pain and 15% severe pain. Thirty-four percent (17/50 patients) had residual memory complaints. The MoCA-Blind score was abnormal (< 19/30) in 68.9% of cases (31/45 patients). Fourteen patients (28%, 14/50 patients) had anxiety symptoms, and eight patients (16%, 8/50 patients) reported symptoms of depression. Overall, 32% (16/50 patients) presented at least one symptom of anxiety or depression. Smell or taste persistent disorder was found in twenty-three patients (45%, 23/51 patients) six months after NeuroCOVID. In multivariate analysis, age was found as a risk factor of impaired quality of life (OR 1.06 per year, 95% CI 1.01, 1.13, P  = 0.026). Neither age, sex, neurological comorbidities, the severity of COVID-19, type of neurological manifestation was associated with residual disability, cognitive impairment, depression or anxiety, or the presence of a sensory deficit. 4 Discussion Our study describes the long-term course of NeuroCOVID. Half of our patients had residual disability and impaired quality of life six months after the acute phase. Cognitive difficulties, anxious and depressive symptoms, and smell/taste disorders were the most frequent neuropsychiatric sequelae. Interestingly, we found that age predicts the long-term alteration of QoL after NeuroCOVID. This emphasizes the need for special attention and long-term care over months in patients with NeuroCOVID, particularly in the oldest patients. Our study has limitations. First, the retrospective nature of the study puts at risk the possibility of recall bias. Second, many patients from the registry refuse to answer the phone interview, introducing a bias of selection. The study group was comparable to the registered population (Supplementary Table S2), which partially limits this bias. The relatively small number of patients could also have led to a lack of statistical power. Additionally, in the absence of a control group of COVID patients with no neurological manifestation, we cannot definitively distinguish the impact of neurological involvement from the consequences of COVID in other organs and functions. Lastly, the lack of comprehensive information on the premorbid functional status of our patients could represent another difficulty for results interpretation. Half of our patients (49%) reported an alteration of their QoL six months after discharge. It is consistent with the findings of a previous 6-month follow-up studies in older COVID-19 patients with or without neurological involvement (median 74 years versus 66 years in our study) discharged from hospital [24]. Age is reported as a predictor of residual impaired quality of life in a previous 3-month follow-up study of COVID-19 patients [25]. In some COVID-19 series, patients are primarily classified according to the severity of the disease without detailed information about specific involvement of various organs, including nervous system [24], [25], [26]. The long-term consequences of neurological impairment on the QoL could be drowned in those of severe overall impairment. In our younger population, a significant recovery of the worsening of the neurological symptoms within six months after the acute phase, a lack of control group, or a lack of statistical power could have prevented us to detect a potential impact of neurological damage on the alteration in the quality of life. In our study, preexisting neurological comorbidities, reported in 17% of patients, were not statistically related to a poor neuropsychiatric outcome at six months. In retrospective COVID-19 hospital-based cohort studies, preexisting neurological diseases were found as an independent risk factor for poor outcomes during the acute phase [27], [28] and at discharge [29], [30], [31]. Furthermore, worsening of preexisting neurological conditions was frequently observed immediately after a COVID-19 [30], [31], [32]. A significant recovery of the worsening of the neurological comorbidity six months after the acute phase or a lack of statistical power could explain this discordant result in our study. Anxiety and depression were present six months after discharge in 10–31% in cohort studies of COVID-19 patients [3], [5], [33], and in 25–46% in a recent prospective study among hospitalized COVID-19 patients with neurological complications [34]. These symptoms were detected in the same proportion in our study (32%), and no risk factors were identified. The severity of COVID-19 disease was previously identified as a risk factor for developing anxiety disorders [3]. It was not found in our study (43% with severe or critical disease), possibly due to the lack of statistical power. Persistent cognitive impairment ranged from 36% and 65% of COVID-19 patients three/four months after discharge [5], [33], [35], and in 50% of patients with neurological complications of COVID-19 during the index hospitalization [34]. The higher proportion in our study (68.9%) could be explained by the use of various cognitive tests. Critically ill patients are known to have a risk of developing cognitive impairment 12 months after discharge due to a longer duration of delirium (mainly sedative-associated-delirium), sepsis, hypoxemia, and doses of sedatives [36]. This condition explains a part of persistence of cognitive impairment in 43% of our patients but is probably not the only cause of this condition since around 60% of patients had persistent cognitive disorders after seven months, even in mild COVID-19 [4]. Prolonged systemic inflammation might predispose to persistent depression and associated neurocognitive dysfunction [37]. Whether prolonged cognitive impairment might be related to pathological changes associated with critical illness-related encephalopathy or specific to SARS-CoV-2 infection remains unclarified. Retrograde neuroinvasion from the olfactory sensory neurons to their downstream brain structures is likely to be involved in selective neurological damage related to SARS-CoV-2 infection [38]. Other possible gateways to the CNS comprise the blood–brain barrier epithelium and brain infiltration by peripheral immune cells [38]. Loss of smell or taste is a frequent symptom of acute COVID-19 [39], [40]. Various non-exclusive pathogenic mechanisms can account for these manifestations, namely obstruction of the olfactory cleft [41], perturbation of supporting cells of the olfactory epithelium [42], alteration of olfactory sensory neurons neurogenesis [43], and secondary neurological damage related to edema in the olfactory bulb [44], [45]. A recent study indicates that olfactory neuroepithelium is an important site of SARS-CoV-2 infection and that SARS-CoV-2 persistence and related inflammation likely account for the prolonged loss of smell in patients [38]. When detected by a specific questionnaire, olfactory dysfunction was found in 51.3% of mild symptomatic COVID-19 patients one month after the acute phase [46] and 60% after six months [7], which is closed to our result (45%) after six months. The balance of patients with smell or taste sequelae is lower (between 10 and 20%) when patients are simply asked about the presence of smell or taste symptoms [47], [48]. NeuroCOVID likely carries a risk of long-term neuropsychiatric disability. Further NeuroCOVID cohort studies using control groups are needed to assess specific long-term outcomes of neurological manifestations associated with COVID-19 infection and clarify their pathogenesis. Disclosure of interest H. Chaumont: received travel grants from PEPS development, Roche and Pfizer. E. Roze: served on scientific advisory boards for Orkyn, Aguettant, Merz-Pharma; received honoraria for speeches from Orkyn, Aguettant, Merz-Pharma, Medday-Pharma, Everpharma, International Parkinson and Movement disorders Society; received research support from Merz-Pharma, Orkyn, Aguettant, Elivie, Ipsen, Everpharma, Fondation Desmarest, AMADYS, Fonds de Dotation Brou de Laurière, Agence Nationale de la Recherche; received travel grant from Vitalair, PEPS development, Aguettant, Merz-Pharma, Ipsen, Merck, Orkyn, Elivie, Adelia Medical, Dystonia Medical Research Foundation, International Parkinson and Movement disorders Society, European Academy of Neurology, International Association of Parkinsonism and Related Disorders. A. Lannuzel: received research support from France Parkinson, PSP France, Agence Nationale de la Recherche, Fonds européen de développement regional, French Ministry of Health, University Hospital of Guadeloupe; received honoraria from Association des Neurologues du Québec and travel grants from Vitalair, PEPS development, Merz-Pharma, International Parkinson and Movement disorders Society. E. Meppiel, B. Tressières, and T. de Broucker declare that they have no competing interest. Appendix A Supplementary data Table S1 Contributors to the NeuroCOVID follow-up study. Table S2 Demographic and neurological characteristics of NeuroCOVID patients. Comparison of participating and non-participating patients in the follow-up evaluation. Appendix A Supplementary data associated with this article can be found, in the online version, at http://doi.dx.org/10.1016/j.neurol.2021.12.008. ==== Refs References 1 Carfì A. Bernabei R. Landi F. for the Gemelli Against COVID-19 Post-Acute Care Study Group Persistent symptoms in patients after acute COVID-19 JAMA 324 2020 603 10.1001/jama.2020.12603 32644129 2 Halpin S.J. McIvor C. Whyatt G. Adams A. Harvey O. McLean L. Postdischarge symptoms and rehabilitation needs in survivors of COVID-19 infection: a cross-sectional evaluation J Med Virol 93 2021 1013 1022 10.1002/jmv.26368 32729939 3 Huang C. Huang L. Wang Y. Li X. Ren L. Gu X. 6-month consequences of COVID-19 in patients discharged from hospital: a cohort study The Lancet 0 2021 10.1016/S0140-6736(20)32656-8 4 Davis H.E. Assaf G.S. McCorkell L. Wei H. Low R.J. Re’em Y. Characterizing Long COVID in an International Cohort: 7 months of symptoms and their impact MedRxiv 2020 10.1101/2020.12.24.20248802 5 Morin L. Savale L. Pham T. Colle R. Figueiredo S. The Writing Committee for the COMEBAC Study Group Four-month clinical status of a cohort of patients after hospitalization for COVID-19 JAMA 325 2021 1525 10.1001/jama.2021.3331 33729425 6 Taquet M. Luciano S. Geddes J.R. Harrison P.J. Bidirectional associations between COVID-19 and psychiatric disorder: retrospective cohort studies of 62 354 COVID-19 cases in the USA The Lancet Psychiatry 0 2020 10.1016/S2215-0366(20)30462-4 7 Boscolo-Rizzo P. Menegaldo A. Fabbris C. Spinato G. Borsetto D. Vaira L.A. Six-month psychophysical evaluation of olfactory dysfunction in patients with COVID-19 Chem Senses 2021 10.1093/chemse/bjab006 bjab006 8 Mao L. Jin H. Wang M. Hu Y. Chen S. He Q. Neurologic Manifestations of hospitalized patients with coronavirus disease 2019 in Wuhan, China JAMA Neurol 77 2020 683 10.1001/jamaneurol.2020.1127 32275288 9 Moriguchi T. Harii N. Goto J. Harada D. Sugawara H. Takamino J. A first case of meningitis/encephalitis associated with SARS-Coronavirus-2 Int J Infect Dis 94 2020 55 58 10.1016/j.ijid.2020.03.062 32251791 10 Poyiadji N. Shahin G. Noujaim D. Stone M. Patel S. Griffith B. COVID-19–associated acute hemorrhagic necrotizing encephalopathy: imaging features Radiology 296 2020 E119 E120 10.1148/radiol.2020201187 32228363 11 Oxley T.J. Mocco J. Majidi S. Kellner C.P. Shoirah H. Singh I.P. Large-vessel stroke as a presenting feature of covid-19 in the young N Engl J Med 382 2020 e60 10.1056/NEJMc2009787 32343504 12 Dinkin M. Gao V. Kahan J. Bobker S. Simonetto M. Wechsler P. COVID-19 presenting with ophthalmoparesis from cranial nerve palsy Neurology 95 2020 221 223 10.1212/WNL.0000000000009700 32358218 13 Toscano G. Palmerini F. Ravaglia S. Ruiz L. Invernizzi P. Cuzzoni M.G. Guillain–Barré Syndrome associated with SARS-CoV-2 N Engl J Med 382 2020 2574 2576 10.1056/NEJMc2009191 32302082 14 Foucard C. San-Galli A. Tarrano C. Chaumont H. Lannuzel A. Roze E. Acute cerebellar ataxia and myoclonus with or without opsoclonus: a parainfectious syndrome associated with COVID-19 Eur J Neurol 2021 10.1111/ene.14726 ene.14726 15 Chaumont H. San-Galli A. Martino F. Couratier C. Joguet G. Carles M. Mixed central and peripheral nervous system disorders in severe SARS-CoV-2 infection J Neurol 267 2020 3121 3127 10.1007/s00415-020-09986-y 32533322 16 Meppiel E. Peiffer-Smadja N. Maury A. Bekri I. Delorme C. Desestret V. Neurologic manifestations associated with COVID-19: a multicentre registry Clin Microbiol Infect 2020 10.1016/j.cmi.2020.11.005 S1198743X20306984 17 COVID-19 Treatment Guidelines Panel. Coronavirus Disease 2019 (COVID-19) Treatment Guidelines. National Institutes of Health. Available at https://www.covid19treatmentguidelines.nih.gov/. Accessed [insert date]. n.d. 18 Bamford J.M. Sandercock P.A. Warlow C.P. Slattery J. Interobserver agreement for the assessment of handicap in stroke patients Stroke 20 1989 10.1161/01.STR.20.6.828 828-828 19 Rabin R. Charro F. de EQ-SD: a measure of health status from the EuroQol Group Ann Med 33 2001 337 343 10.3109/07853890109002087 11491192 20 Nasreddine Z.S. Phillips N.A. Bédirian V. Charbonneau S. Whitehead V. Collin I. The Montreal Cognitive Assessment, MoCA: a Brief Screening Tool for Mild Cognitive Impairment: MOCA: a brief screening tool for MCI J Am Geriatr Soc 53 2005 695 699 10.1111/j.1532-5415.2005.53221.x 15817019 21 McNair D. Kahn R. Self assessment of cognitive deficits. In: Crook J, Ferris S, Bartus R, editors. New Canaan, CT: Mark Powley Associates Inc 1983 Assessment of Geriatric Psy-chopharmacology 137 143 22 Zigmond A.S. Snaith R.P. The Hospital Anxiety and Depression Scale 67 1983 Acta Psychiatrica Scandinavica 361 370 23 Heald A.E. Pieper C.F. Schiffman S.S. Taste and smell complaints in HIV-infected patients AIDS 12 1998 1667 1674 10.1097/00002030-199813000-00015 9764787 24 Walle-Hansen M.M. Ranhoff A.H. Mellingsæter M. Wang-Hansen M.S. Myrstad M. Health-related quality of life, functional decline, and long-term mortality in older patients following hospitalisation due to COVID-19 BMC Geriatr 21 2021 199 10.1186/s12877-021-02140-x 33752614 25 Qu G. Zhen Q. Wang W. Fan S. Wu Q. Zhang C. Health-related quality of life of COVID-19 patients after discharge: a multicenter follow-up study J Clin Nurs 2021 10.1111/jocn.15733 jocn.15733 26 Gautam N. Madathil S. Tahani N. Bolton S. Parekh D. Stockley J. Medium-term outcome of severe to critically ill patients with SARS-CoV-2 infection Clin Infect Dis 2021 10.1093/cid/ciab341 ciab341 27 Benussi A. Pilotto A. Premi E. Libri I. Giunta M. Agosti C. Clinical characteristics and outcomes of inpatients with neurologic disease and COVID-19 in Brescia, Lombardy, Italy Neurology 95 2020 10.1212/WNL.0000000000009848 e910-20 28 Livingston G. Rostamipour H. Gallagher P. Kalafatis C. Shastri A. Huzzey L. Prevalence, management, and outcomes of SARS-CoV-2 infections in older people and those with dementia in mental health wards in London, UK: a retrospective observational study The Lancet Psychiatry 7 2020 1054 1063 10.1016/S2215-0366(20)30434-X 33031760 29 Qin C. Zhou L. Hu Z. Yang S. Zhang S. Chen M. Clinical characteristics and outcomes of COVID-19 patients with a history of stroke in Wuhan, China Stroke 51 2020 2219 2223 10.1161/strokeaha.120.030365 32466735 30 Bianchetti A. Rozzini R. Guerini F. Boffelli S. Ranieri P. Minelli G. Clinical presentation of COVID19 in dementia patients J Nutr Health Aging 24 2020 560 562 10.1007/s12603-020-1389-1 32510106 31 Ji W. Huh K. Kang M. Hong J. Bae G.H. Lee R. Effect of underlying comorbidities on the infection and severity of COVID-19 in Korea: a nationwide case-control study J Korean Med Sci 35 2020 e237 10.3346/jkms.2020.35.e237 32597048 32 Antonini A. Leta V. Teo J. Chaudhuri K.R. Outcome of Parkinson's disease patients affected by COVID-19 Mov Disord 35 2020 905 908 10.1002/mds.28104 32347572 33 van den Borst B. Peters J.B. Brink M. Schoon Y. Bleeker-Rovers C.P. Schers H. Comprehensive health assessment three months after recovery from acute COVID-19 Clin Infect Dis 2020 10.1093/cid/ciaa1750 ciaa1750 34 Frontera J.A. Yang D. Lewis A. Patel P. Medicherla C. Arena V. A prospective study of long-term outcomes among hospitalized COVID-19 patients with and without neurological complications J Neurol Sci 426 2021 117486 10.1016/j.jns.2021.117486 34000678 35 Miskowiak K. Johnsen S. Sattler S. Nielsen S. Kunalan K. Rungby J. Cognitive impairments four months after COVID-19 hospital discharge: Pattern, severity and association with illness variables Eur Neuropsychopharmacol 46 2021 39 48 10.1016/j.euroneuro.2021.03.019 33823427 36 Girard T.D. Thompson J.L. Pandharipande P.P. Brummel N.E. Jackson J.C. Patel M.B. Clinical phenotypes of delirium during critical illness and severity of subsequent long-term cognitive impairment: a prospective cohort study Lancet Respir Med 6 2018 213 222 10.1016/S2213-2600(18)30062-6 29508705 37 Mazza M.G. Palladini M. De Lorenzo R. Magnaghi C. Poletti S. Furlan R. Persistent psychopathology and neurocognitive impairment in COVID-19 survivors: effect of inflammatory biomarkers at three-month follow-up Brain Behav Immun 94 2021 138 147 10.1016/j.bbi.2021.02.021 33639239 38 de Melo G.D. Lazarini F. Levallois S. Hautefort C. Michel V. Larrous F. COVID-19-related anosmia is associated with viral persistence and inflammation in human olfactory epithelium and brain infection in hamsters Sci Transl Med 2021 10.1126/scitranslmed.abf8396 eabf8396 39 Xydakis M.S. Dehgani-Mobaraki P. Holbrook E.H. Geisthoff U.W. Bauer C. Hautefort C. Smell and taste dysfunction in patients with COVID-19 Lancet Infect Dis 20 2020 1015 1016 10.1016/S1473-3099(20)30293-0 32304629 40 Qiu C. Cui C. Hautefort C. Haehner A. Zhao J. Yao Q. Olfactory and gustatory dysfunction as an early Identifier of COVID-19 in adults and children: an International Multicenter Study Otolaryngol Head Neck Surg 163 2020 714 721 10.1177/0194599820934376 32539586 41 Eliezer M. Hamel A.-L. Houdart E. Herman P. Housset J. Jourdaine C. Loss of smell in patients with COVID-19: MRI data reveal a transient edema of the olfactory clefts Neurology 95 2020 10.1212/WNL.0000000000010806 e3145-52 42 Brann D.H. Tsukahara T. Weinreb C. Lipovsek M. Van den Berge K. Gong B. Non-neuronal expression of SARS-CoV-2 entry genes in the olfactory system suggests mechanisms underlying COVID-19-associated anosmia Sci Adv 6 2020 10.1126/sciadv.abc5801 eabc5801 43 Saussez S. Lechien J.R. Hopkins C. Anosmia: an evolution of our understanding of its importance in COVID-19 and what questions remain to be answered Eur Arch Otorhinolaryngol 2020 10.1007/s00405-020-06285-0 44 Laurendon T. Radulesco T. Mugnier J. Gérault M. Chagnaud C. El Ahmadi A.-A. Bilateral transient olfactory bulb edema during COVID-19–related anosmia Neurology 95 2020 224 225 10.1212/WNL.0000000000009850 32444492 45 Aragão M.F.V.V. Leal M.C. Cartaxo Filho O.Q. Fonseca T.M. Valença M.M. Anosmia in COVID-19 Associated with Injury to the olfactory bulbs evident on MRI AJNR Am J Neuroradiol 2020 10.3174/ajnr.A6675 ajnr;ajnr.A6675v1. 46 Boscolo-Rizzo P. Borsetto D. Fabbris C. Spinato G. Frezza D. Menegaldo A. Evolution of Altered sense of smell or taste in patients with mildly symptomatic COVID-19 JAMA Otolaryngol Head Neck Surg 146 2020 729 10.1001/jamaoto.2020.1379 32614442 47 Paderno A. Mattavelli D. Rampinelli V. Grammatica A. Raffetti E. Tomasoni M. Olfactory and gustatory outcomes in COVID-19: a prospective evaluation in nonhospitalized subjects Otolaryngol Head Neck Surg 163 2020 1144 1149 10.1177/0194599820939538 32600175 48 Petersen M.S. Kristiansen M.F. Hanusson K.D. Danielsen M.E. á Steig B. Gaini S. Long COVID in the Faroe Islands - a longitudinal study among non-hospitalized patients Clin Infect Dis 2020 10.1093/cid/ciaa1792

35000793

PMC9753577

NO-CC CODE

2022-12-16 23:25:48

no

Rev Neurol (Paris). 2022 Jan 6 January-February; 178(1):137-143

utf-8

Rev Neurol (Paris)

10.1016/j.neurol.2021.12.008

oa_other

==== Front Blood Blood Blood 0006-4971 1528-0020 The American Society of Hematology S0006-4971(22)08129-0 10.1182/S0006-4971(22)08129-0 Article Table of Contents 15 12 2022 15 12 2022 15 12 2022 140 24 iiiv 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmc

0

PMC9753581

NO-CC CODE

2022-12-16 23:25:48

no

Blood. 2022 Dec 15; 140(24):iii-v

utf-8

Blood

10.1182/S0006-4971(22)08129-0

oa_other

==== Front Blood Blood Blood 0006-4971 1528-0020 The American Society of Hematology S0006-4971(22)08129-0 10.1182/S0006-4971(22)08129-0 Article Table of Contents 15 12 2022 15 12 2022 15 12 2022 140 24 iiiv 2022 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmc

34756712

PMC9753593

NO-CC CODE

2022-12-16 23:26:08

no

Transplant Proc. 2021 Nov 30; 53(9):2645

latin-1

Transplant Proc

10.1016/j.transproceed.2021.10.001

oa_other

==== Front Gynecol Oncol Gynecol Oncol Gynecologic Oncology 0090-8258 1095-6859 Academic Press S0090-8258(21)01456-6 10.1016/j.ygyno.2021.10.012 Poster Abstract #2 The COVID-19 pandemic did not adversely affect clinical trial enrollment in gynecologic oncology trials at a single academic institution Bailey Courtney a Ghamande Sharad b Tran Lynn b Rungruang Bunja c Wheatley Donna b Higgins Robert d a Medical College of Georgia / Augusta Uni b Augusta University c Medical College of, Georgia at Augusta Un d Medical College of GA at Augusta University 27 12 2021 1 2022 27 12 2021 164 1 1818 2021 Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcObjectives The objective of this study was to compare enrollment of patients with gynecologic cancers at our institution in 2019, before the COVID-19 pandemic, with enrollment in 2020, during the pandemic. Methods Clinical trial enrollment data was obtained through the clinical trials office. Patients enrolled in gynecologic oncology therapeutic trials (excluding maintenance trials) in 2019 and 2020 were compared using Wilcoxon Rank Sum testing. The number of patients enrolled in each clinical trial phase (Phase 1–3) and each disease site were also compared between 2019 and 2020. Standard descriptive statistics were used to compare demographic data of the clinical trial enrollees. Results Total patient enrollment for 2019 was 56 patients, and 45 patients enrolled in 2020. There was no statistically significant difference between 2019 and 2020 in the number of patients enrolled in clinical therapeutic trials at our institution by quarter (p-value 0.486). There was no statistically significant difference between the two years in the number of patients enrolled by disease site (p = 0.476) or in the phase of clinical trial in which patients enrolled (p = 0.126). The mean age of patients enrolled was similar (58.5 vs 60.7 years, p = 0.432). The mean distance traveled to our site was also similar between the two years (66.5 vs 76.0 miles, p = 0.687). Conclusions Unlike many other centers throughout the United States, clinical trial enrollment at our institution remained similar during the COVID-19 pandemic compared to the prior year. We attribute the continued enrollment of patients in clinical trials to several factors. These factors include the dedication of our research team to work on-site, the ability of our non-profit patient-accommodation facility to remain open, and the commitment of our gynecologic cancer support group to continue to hold events virtually.Unlabelled Image

34696892

PMC9753595

NO-CC CODE

2022-12-16 23:26:08

no

Gynecol Oncol. 2022 Jan 27; 164(1):18

utf-8

Gynecol Oncol

10.1016/j.ygyno.2021.10.012

oa_other

==== Front Prev Med Prev Med Preventive Medicine 0091-7435 1096-0260 Elsevier Inc. S0091-7435(20)30393-5 10.1016/j.ypmed.2020.106362 106362 Short Communication A longitudinal study of psychological distress in the United States before and during the COVID-19 pandemic Breslau Joshua a⁎ Finucane Melissa L. a Locker Alicia R. a Baird Matthew D. a Roth Elizabeth A. b Collins Rebecca L. b a RAND Corporation, Pittsburgh, PA 15213, USA b RAND Corporation, Santa Monica, CA 90401, USA ⁎ Corresponding author at: RAND Corporation, 4570 Fifth Avenue, Pittsburgh, PA 15213, USA. 31 12 2020 2 2021 31 12 2020 143 106362106362 20 7 2020 7 12 2020 8 12 2020 © 2020 Elsevier Inc. All rights reserved. 2020 Elsevier Inc. Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. The COVID-19 pandemic has caused financial stress and disrupted daily life more quickly than any prior economic downturn and on a scale beyond any prior natural disaster. This study aimed to assess the impact of the pandemic on psychological distress and identify vulnerable groups using longitudinal data to account for pre-pandemic mental health status. Clinically significant psychological distress was assessed with the Kessler-6 in a national probability sample of adults in the United States at two time points, February 2019 (T1) and May 2020 (T2). To identify increases in distress, psychological distress during the worst month of the past year at T1 was compared with psychological distress over the past 30-days at T2. Survey adjusted logistic regression was used to estimate associations of demographic characteristics at T1 (gender, age, race, and income) and census region at T2 with within-person increases in psychological distress. The past-month prevalence of serious psychological distress at T2 was as high as the past-year prevalence at T1 (10.9% vs. 10.2%). Psychological distress was strongly associated across assessments (X2(4) = 174.6, p < .0001). Increase in psychological distress above T1 was associated with gender, age, household income, and census region. Equal numbers of people experienced serious psychological distress in 30-days during the pandemic as did over an entire year prior to the pandemic. Mental health services and research efforts should be targeted to those with a history of mental health conditions and groups identified as at high risk for increases in distress above pre-pandemic levels. Keywords COVID-19 pandemic Mental health Psychological distress Longitudinal study ==== Body pmc1 Introduction There is reason to be concerned about a rapid and possibly sustained negative impact of the COVID-19 pandemic on mental health (Raker et al., 2020). The financial and social stressors, including unemployment and isolation, have affected more people over a shorter period than any prior economic downturn or natural disaster. Evidence suggests that there are negative mental health effects of even short periods of unemployment(Cygan-Rehm et al., 2017), which many people experienced early in the pandemic. In addition, studies have documented immediate adverse effects of natural disasters(Goldmann and Galea, 2014; North, 2014), including pandemics(Brooks et al., 2020) on mental health. The literature also suggests that there are strong predictors of vulnerability to the mental health effects of disasters, including pre-disaster mental health status(Goldmann and Galea, 2014; North, 2014). There is little data yet available on psychological distress during the COVID-19 pandemic and, as is often the case in disaster research(Parker et al., 2019), no studies with longitudinal assessments that allow within-person comparisons with pre-pandemic mental health status. Elevated distress during the pandemic was reported in a web survey conducted in Italy(Mazza et al., 2020). Two studies of U.S. national samples, one of psychological distress and one of depressive symptoms, found threefold higher prevalence of poor mental health status compared to prior periods, but those studies used repeated cross-sectional designs in which data was collected using different samples and methodologies at each time point(Ettman et al., 2020; McGinty et al., 2020). Longitudinal data are needed to prospectively assess meaningful worsening of mental health status during the pandemic, relative to individuals' own pre-pandemic mental health status. Information on the extent and predictors of psychological distress during the early phase of the pandemic will be important for targeting interventions and resources to vulnerable groups and tracking the pandemic's long-term health effects. In this study, we use longitudinal data from a nationally representative sample of the United States adult population to compare psychological distress experienced during the pandemic with the highest level of distress respondents had experienced during a 12-month period prior to the pandemic. This comparison allows us to identify not only individuals with high distress, but those whose distress was meaningfully elevated above the highest level of distress they experienced over an entire year pre-pandemic. Examining within-person change allows us to better understand the mental health impact of the pandemic. 2 Methods 2.1 Sample Data come from the Rand American Life Panel (ALP), a probability-based representative sample of U.S. adults age 20 and over. The baseline (T1) survey, conducted in February 2019, N = 2555, had a participation rate of 65%. Of these, n = 1870, or 73%, were interviewed in May of 2020 (T2). The T2 assessment was conducted about 8 weeks following the declaration of a national emergency related to COVID-19 in the United States on March 13, 2020 (Carman and Nataraj, 2020). All study procedures were approved by the Rand IRB. 2.2 Psychological distress Psychological distress was assessed using the Kessler-6 (K6), a commonly used instrument designed to identify clinically significant psychiatric conditions(Kessler et al., 2003). Internal consistency (α) was 0.91 at T1 and 0.89 at T2. K6 scores were classified using established cut-offs (Furukawa et al., 2008): no/low distress (LD:0–7), mild/moderate distress (MD: 8–12), and serious distress (SD:13–24). K6 score during the worst month of the past year assessed at T1 was compared with psychological distress in the past-30 days assessed at T2. Respondents were considered to have had an increase in distress if they moved from LD at T1to MD or SD at T2 or from MD at T1 to SD at T2. 2.3 Statistical analysis Analyses used sampling weights generated to account for non-response and match the 2019-wave demographics to the 2019 Current Population Survey, as described in ALP technical documentation(Pollard and Baird, 2017). The distribution of psychological distress at T1 and T2 and the probabilities of within-person transitions between levels of psychological distress between T1 and T2 were calculated. Unadjusted comparisons between T1 and T2 were conducted using survey-adjusted chi-square tests. A survey-adjusted logistic regression model was used to estimate associations between an increase in psychological distress at T2 relative to T1 and demographic characteristics at T1 (gender, age, race, and income) and census region at T2. 3 Results Fig. 1 shows the distribution of psychological distress at the two study time points and individual change between categories over time. The prevalence of psychological distress in the past 30-days at T2 did not differ from the prevalence of psychological distress during the worst month of the past year at T1. For instance, the prevalence of SD was 10.9% (95% CI 7.6% -14.0%) during the worst month pre-pandemic and 10.1% (95% CI 6.9% - 13.3%) during the pandemic. The prevalence of SD at T2 did not differ from that at T1for any demographic group examined (p > .3 for all comparisons, data not shown).Fig. 1 Individual Change in Psychological Distress during the COVID-19 Pandemic, American Life Panel, N = 1870 Percents outside the bars represent population prevalence at each time point. Percents within the bars represent the proportions at each level of distress at T1 who transitioned to each level of distress at T2; these proportions are also reflected in the width of the bars. Turquoise bards indicate increases in distress, yellow bars indicate decreases in distress and grey bars indicate no change. Psychological distress assessed with the Kessler-6, with categorization as defined in the text. The pre-COVID-19 assessment was conducted in February 2019 and referred to the worst month of the past year. The COVID-19 assessment was conducted in May 2020 and referred to the past 30 days. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 1 Analysis of within-person change found that psychological distress during the pandemic was strongly related to psychological distress during the worst month of the year at T1 (X2 (4) = 174.6, p < .0001). People with SD at T1 were more likely to have SD at T2 compared with those with LD or MD at T1 (47.9% vs. 17.8% and 3.2% respectively). Although risk of SD is relatively low among those with LD at T1 (3.2%), they comprise a substantial minority of those with SD at T2 (23.4%). An increase in psychological distress at T2 relative to T1 was found in 12.8% (95% CI 9.9%–15.7%) of the sample. Increase in distress was more common among women compared with men, those under 60 compared with those over 60 and Hispanics compared with other racial/ethnic groups (Table 1 ). The associations of increase in distress with gender and age were sustained in the adjusted model, but that with race/ethnicity was not. In addition, adjusted odds of an increase in distress were about twice as high in those with household incomes of $35-$60 K relative to those with incomes over $100 K (OR = 2.2, 95% CI 1.0–4.6) and less than half as high among those in the South relative to those in New England (OR = 0.4, 95% CI 0.2–0.8).Table 1 Increase in psychological distress during the pandemic over the highest level of distress pre-pandemic. Table 1Individual characteristics Increase above T1 X2 p-value Adjusted odd ratio % se OR 95% CI Gender Female 17.7 2.3 4.45 0.035 1.87 (1.1,3.1) Male 10.6 2.2 Reference Age Ages 20–39 20.8 4.6 10.99 0.004 2.4 (1.4,4.2) Ages 40–59 14.4 2.0 1.7 (1.0,2.8) Ages 60 and up 8.7 1.3 Reference Race-ethnicity Hispanic 25.2 6.5 12.49 0.006 1.9 (0.9,4.0) NH-black 11.5 3.2 0.8 (0.4,1.6) NH-other 14.5 4.9 1.0 (0.4,2.4) NH-white 11.8 1.4 Reference Income ≤ $35 K 15.4 2.6 3.09 0.378 1.4 (0.7,3.0) $35 K-$60 K 18.2 3.9 2.2 (1.0,4.6) $60 K-$99 K 14.9 4.3 1.4 (0.7,3.0) ≥ $100 K 10.2 2.4 Reference Census region South 10.0 1.7 7.14 0.068 0.4 (0.2,0.8) Midwest 11.3 2.7 0.6 (0.3,1.2) West 18.4 4.3 0.8 (0.4,1.6) New England 18.7 3.9 Reference OR = odds ratio; se = standard error; CI=Confidence Interval; NH=Non-Hispanic. Increase in psychological distress defined as movement from no/low distress to mild/moderate or severe distress or from mild/moderate to severe distress. 4 Discussion This is the first longitudinal study of psychological distress during the COVID-19 pandemic in a nationally representative sample of U.S. adults. The prevalence of SD, indicative of a clinical need, during the pandemic exceeded levels that would be expected in the absence of the pandemic; the 30-day prevalence of SD in May 2020 did not differ from the past-year prevalence of SD assessed with the same instrument in February 2019. In other words, equal numbers of people experienced SD in 30-days during the pandemic as experienced SD over an entire year prior to the pandemic. For comparison, the 30-day prevalence of SD is typically found to be about half the 12-month prevalence, when both are assessed at the same time(Hedden et al., 2012). Elevated prevalence occurred across all demographic groups examined. Elevated psychological distress has been observed in prior disasters(Goldmann and Galea, 2014), but has never before been seen for a persistent and complex stressor affecting the entire U.S. population. SD at T1 is a strong predictor of SD at T2, consistent with prior research on psychiatric sequelae of natural disasters(North and Pfefferbaum, 2013). In this study, risk for SD during the pandemic among those with SD during a year before the pandemic was almost 3 times higher than among those reporting MD and 15 times higher than among those reporting LD during the pre-pandemic year. People with prior mental health problems are clearly a high-risk group. However, there is substantial variability in within-person change over time; about half (52.1%) of those with SD at T1 had LD or MD at T2 despite the pandemic. It is also important to note that about half of those with SD during the pandemic had not experienced SD during the year prior to the pandemic. Particular attention should focus on groups at elevated risk for increases in distress relative to prior levels. Higher risk of an increase in psychological distress among respondents younger than 60 suggest that distress may be driven more by economic stressors than fears specific to the disease, since older individuals are widely reported to be at higher risk of morbidity and mortality related to the virus. This finding is reinforced by the vulnerability observed in respondents in the low-middle range of household incomes, close to or below the U.S. median and above $35,000. People in this group may have been at most risk for loss of income or stressful employment conditions(McQuarrie, 2020). Risk for an increase in distress was lower in the Souththan in the NortheastThe geographic difference may have been driven by concerns about the virus itself, the economic consequences of social distancing which were implemented more slowly in the southern states(Kates et al., 2020), or even differing politically inflected interpretations of the threat, given that political affiliation is predictive of social distancing policies(Adolph et al., 2020). Further follow-up studies covering the time period during which the pandemic became more widespread in the southern United States will shed further light on this finding. We found that Hispanics were more likely to report an increase in psychological distress than other racial/ethnic groups. This finding adds to evidence from the one prior study of psychological distress during the COVID-19 pandemic that Hispanic population of the US has been disproportionately affected(McGinty et al., 2020). The finding of high risk among Hispanics was not sustained in our adjusted model, although the magnitude of the difference between Hispanics and Non-Hispanic Whites remained large. Our finding of higher risk among women is consistent with prior studies of psychiatric disorders following disasters(North, 2016). Limitations of this study should be considered. The T1 assessment was about 1 year prior to the pandemic and based on a recall period of 12-month prior to T1. While there is no systematic bias in use of this time period, there may have been variations in psychological distress between assessments that are not captured here. Second, participation was limited to individuals who could complete the survey in English. Studies examining psychological distress in non-English speakers during this time are needed. Third, specific exposures and their relationships with change in distress could not be examined in this study. 5 Conclusion The mental health impact of the COVID-19 pandemic is unprecedented with respect to its nation-wide scale. However, during the period of time covered in this survey, the first two months following the declaration of a national emergency in the United States, the epidemiological patterns of mental health effects are more similar than different to those observed in prior studies of the mental health consequences of disasters. Most notably, during this period serious psychological distress was highly elevated in the general population, with particularly high risk among people with prior psychiatric conditions(North, 2016). Clinical services should be targeted to this population. Services can also be targeted to population groups at high risk for elevated psychological distress during the pandemic, including people vulnerable to the economic consequences of social distancing. Prior research suggests that many who experience psychological distress immediately following a disaster return to pre-disaster levels over time(Pietrzak et al., 2012), and a similar pattern has been observed in trajectories of distress in the U.S. since March of 2020(Daly and Robinson, 2020). However the pandemic's influence on economic stressors, disruption of usual activities and subsequent effects on population health may continue for an extended period and affect different regions of the country at different points in time. Tracing patterns of persistence of serious psychological distress will provide important information to guide the national public health response to the COVID-19 pandemic. Providing services each month of the crisis to as many people as typically experience SD in a 12-month period, with normal pathways to care disrupted by social distancing, is an enormous challenge. Policy makers and practitioners may need to plan for strategic resource distribution to address the most serious clinical needs, reduce stress on providers, and direct resources to address persistent economic distress. The following is the supplementary data related to this article.Supplemental Table A Sample characteristics. Supplemental Table A Declaration of Competing Interest None. Acknowledgments This research was supported by grants from the 10.13039/100006545 National Institute on Minority Health and Health Disparities (R01 MD010274, PI: Breslau) and 10.13039/100000025 National Institute of Mental Health (R01 MH104381, PI: Collins) ==== Refs References Adolph C. Amano K. Bang-Jensen B. Fullman N. Wilkerson J. Pandemic Politics: Timing State-Level Social Distancing Responses to COVID-19 2020 medRxiv:2020.03.30.20046326 Brooks S.K. Webster R.K. Smith L.E. Woodland L. Wessely S. Greenberg N. Rubin G.J. The psychological impact of quarantine and how to reduce it: rapid review of the evidence Lancet 395 2020 912 920 32112714 Carman K.G. Nataraj S. 2020 American Life Panel Survey on Impacts of COVID-19: Technical Documentation 2020 RAND Santa Monica, CA Cygan-Rehm K. Kuehnle D. Oberfichtner M. Bounding the causal effect of unemployment on mental health: nonparametric evidence from four countries Health Econ. 26 2017 1844 1861 28497638 Daly M. Robinson E. Psychological distress and adaptation to the COVID-19 crisis in the United States J. Psychiatr. Res. 2020 10.1016/j.jpsychires.2020.10.035 (in press) Ettman C.K. Abdalla S.M. Cohen G.H. Sampson L. Vivier P.M. Galea S. Prevalence of depression symptoms in US adults before and during the COVID-19 pandemic JAMA Netw. Open 3 2020 e2019686-e86 Furukawa T.A. Kawakami N. Saitoh M. Ono Y. Nakane Y. Nakamura Y. Tachimori H. Iwata N. Uda H. The performance of the Japanese version of the K6 and K10 in the world mental health survey Japan Int. J. Methods Psychiatr. Res. 17 2008 152 158 18763695 Goldmann E. Galea S. Mental health consequences of disasters Annu. Rev. Public Health 35 2014 169 183 24159920 Hedden S. Gfroerer J.C. Barker P. Smith S. Pemberton M.R. Saavedra L.M. Forman-Hoffman V.L. Ringeisen H. Novak S.P. Comparison of NSDUH Mental Health Data and Methods with Other Data Sources 2012 Center for Behavioral Health and Statistics and Quality Data Review March Kates J. Michaud J. Tobert J. Stay-At-Home Orders to Fight COVID-19 in the United States: The Risks of a Scattershot Approach 2020 Kaiser Family Foundation Kessler R.C. Barker P.R. Colpe L.J. Epstein J.F. Gfroerer J.C. Hiripi E. Howes M.J. Normand S.L.T. Manderscheid R.W. Screening for serious mental illness in the general population Arch. Gen. Psychiatry 60 2003 184 189 12578436 Mazza C. Ricci E. Biondi S. Colasanti M. Ferracuti S. Napoli C. Roma P. A nationwide survey of psychological distress among Italian people during the COVID-19 pandemic: immediate psychological responses and associated factors Int. J. Environ. Res. Public Health 17 2020 McGinty E.E. Presskreischer R. Han H. Barry C.L. Psychological distress and loneliness reported by US adults in 2018 and April 2020 JAMA. 324 1 2020 93 94 10.1001/jama.2020.9740 32492088 McQuarrie K. The Average Salary of Essential Workers in 2020: How Much Are Essential Workers Earning Compared to the Average Income in their State? Business.org 2020 North C.S. Current research and recent breakthroughs on the mental health effects of disasters Curr. Psych. Rep. 16 2014 481 North C.S. Disaster mental health epidemiology: methodological review and interpretation of research findings Psychiatry 79 2016 130 146 27724836 North C.S. Pfefferbaum B. Mental health response to community disasters: a systematic review JAMA 310 2013 507 518 23925621 Parker A.M. Edelman A.F. Carman K.G. Finucane M.L. On the need for prospective disaster survey panels Disaster Med. Public Health Prep. 2019 1 3 Pietrzak R.H. Tracy M. Galea S. Kilpatrick D.G. Ruggiero K.J. Hamblen J.L. Southwick S.M. Norris F.H. Resilience in the face of disaster: prevalence and longitudinal course of mental disorders following hurricane Ike PLoS One 7 2012 e38964 Pollard M. Baird M.D. The RAND American Life Panel: Technical Description 2017 RAND Corporation Santa Monica, CA Raker E.J. Zacher M. Lowe S.R. Lessons from hurricane Katrina for predicting the indirect health consequences of the COVID-19 pandemic Proc. Natl. Acad. Sci. 117 2020 12595 32424085

33388325

PMC9753596

NO-CC CODE

2022-12-16 23:26:08

no

Prev Med. 2021 Feb 31; 143:106362

utf-8

Prev Med

10.1016/j.ypmed.2020.106362

oa_other

==== Front Microbiol Resour Announc Microbiol Resour Announc mra Microbiology Resource Announcements 2576-098X American Society for Microbiology 1752 N St., N.W., Washington, DC 36346245 00790-22 10.1128/mra.00790-22 mra.00790-22 Genome Sequences plant-microbiologyPlant MicrobiologyDraft Genome Sequence of a Washington Isolate of “Candidatus Phytoplasma pruni” https://orcid.org/0000-0003-4814-4372 Wright Alice A. a alice.wright@usda.gov Harper Scott J. b a Sugarcane Research Unit, USDA Agricultural Research Service, Houma, Louisiana, USA b Department of Plant Pathology, Washington State University, Prosser, Washington, USA Editor Newton Irene L. G. Indiana University, Bloomington The authors declare no conflict of interest. 8 11 2022 12 2022 8 11 2022 11 12 e00790-229 8 2022 27 10 2022 https://doi.org/10.1128/AuthorWarrantyLicense.v1 This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply. ABSTRACT Illumina sequencing of a Prunus avium tree with X-disease symptoms was performed to obtain a draft genome of “Candidatus Phytoplasma pruni.” The genome consists of 14 contigs covering 588,767 bp. This is the first metagenome to be sequenced from the current X-disease epidemic in stone fruit in the Pacific Northwest. Washington Tree Fruit Research Commission (WTFRC) https://doi.org/10.13039/100012487 GR00002025 Wright Alice Ann Harper Scott cover-dateDecember 2022 ==== Body pmcANNOUNCEMENT The causative agent of X-disease in Prunus avium (cherry) and Prunus persica (peach/nectarine), which has reached epidemic levels in stone fruit in the Pacific Northwest, is “Candidatus Phytoplasma pruni” (1, 2). In peaches and nectarines, infection by “Ca. Phytoplasma pruni” causes premature yellowing of leaves and spots of necrosis, producing a shot-hole-like appearance. The pathogen also causes limb dieback and, within approximately 5 years of infection, tree death (1). In cherries, the fruit is affected; at harvest, the fruit is small and misshapen, with poor color and taste and often a reduction in sugar content (1, 2). Like other phytoplasmas, “Ca. Phytoplasma pruni” is a cell wall-less bacterium that currently cannot be cultured independently of its host (3). Developing improved genomic resources for this pathogen is crucial to understanding and managing this disease. To that end, an isolate from the current epidemic was sequenced. Cuttings from a sweet cherry tree expressing classic X-disease symptoms were collected from a commercial orchard in Zillah, Washington. DNA was extracted from woody stem tissue because this is the tissue in which phytoplasma titers are highest (4). Briefly, 100 mg of tissue was frozen at −80°C for at least 1 h prior to homogenization to a fine powder by bead-beating in a TissueLyser (Qiagen, Germantown, MD) at 20 beats/s for 30 s. DNA was then extracted with a DNeasy Plant minikit (Qiagen) according to the manufacturer’s instructions. Phytoplasma presence was confirmed by quantitative PCR (qPCR) as described (4), and the DNA concentration was determined using a Qubit fluorometer (Thermo Fisher Scientific, Hampton, NH). DNA was submitted to Genewiz (South Plainfield, NJ) for library preparation using a NEBNext Ultra library preparation kit (New England Biolabs, Ipswich, MA) and 250-bp paired-end sequencing with a single HiSeq flow cell on a HiSeq 2500 system (Illumina Inc., San Diego, CA, USA). Raw reads were processed with CLC Genomics Workbench v21.0.5 (Qiagen, Hilden, Germany) using default settings unless stated otherwise. Sequencing yielded 309,073,036 raw reads, which were reduced to 309,072,714 reads after adaptor removal and quality assessment; this was performed using the Trim Reads pipeline with the default quality limit of 0.05. De novo assembly was performed on the trimmed reads using the CLC Genomics Workbench de novo assembly tool. Minimum contig length was set to 200 bp, word size to 26, and bubble size to 246. The assembly consisted of 343,722 contigs, with an N50 value of 2,935 bp. BLASTn analysis with a word size of 50 was used to identify 371 phytoplasma-like sequences, which were then imported into Geneious v2022.0.2 (Biomatters, Inc., Auckland, New Zealand) to find open reading frames (ORFs). ORF identification was performed with a minimum size set to 300 bp and the genetic code set to standard. Genes were confirmed and where possible assigned a function based on homology to known phytoplasma genes using BLASTp. Of the 371 contigs, 14 were large enough to contain ORFs; the remainder were small fragments that might have been the result of repetitive regions that could not be resolved by Illumina sequencing. These 14 contigs were composed of a total of 588,767 bp, with a GC content of 27.1%. The average sequencing depth was 18,538×, with 45,212,112 reads mapped to the contigs. Through the Geneious ORF prediction tool and manual curation, 469 genes were annotated. These genes were assigned functions based on homology to genes in other phytoplasma genomes. The estimated size of the “Ca. Phytoplasma pruni” genome is 677 kb (5). This draft genome represents approximately 87% of the full genome. It is similar in size to the draft genome published previously (6); however, there are fewer contigs in the genome of this metagenome. The genome sequenced previously (6) was from a strain that was originally from an infected peach tree in Connecticut (7) and was maintained in periwinkle, making it temporally and geographically separate from the Washington sample. The genome presented here was sequenced directly from an infected cherry in Zillah, Washington, and is representative of the current epidemic in the Pacific Northwest. Additional work, including comparative genomic analysis among geographically diverse isolates, may explain why the current epidemic is so devastating. Data availability. Raw reads have been submitted to the SRA under BioProject accession number PRJNA857992. The assembled and annotated genome has been deposited in GenBank with the accession number JANIEM000000000. ACKNOWLEDGMENTS We appreciate the orchard managers allowing us to collect tissue for this study. This study was funded by the Washington Tree Fruit Research Commission. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA. ==== Refs REFERENCES 1 Uyemoto JK, Kirkpatrick BC. 2011. X-disease phytoplasma, p 243–245. In Hadidi A, Barba M, Candresse T, Jelkman W (ed), Virus and virus-like diseases of pome and stone fruits. APS Press, St. Paul, MN. 2 Wright AA, Shires MK, Beaver C, Bishop G, DuPont ST, Naranjo R, Harper S. 2021. Effect of “Candidatus Phytoplasma pruni” infection on sweet cherry fruit. Phtyopathology 111 :2195–2202. doi:10.1094/PHYTO-03-21-0106-R. 3 Hogenhout SA, Oshima K, Ammar E-D, Kakizawa S, Kingdom HN, Namba S. 2008. Phytoplasmas: bacteria that manipulate plants and insects. Mol Plant Pathol 9 :403–423. doi:10.1111/j.1364-3703.2008.00472.x.18705857 4 Wright AA, Shires MK, Molnar C, Bishop G, Johnson AM, Frias C, Harper S. 2022. Titer and distribution of “Candidatus Phytoplasma pruni” in Prunus avium. Phytopathology 112 :1406–1412. doi:10.1094/PHYTO-11-21-0468-R.35021858 5 Firrao G, Smart CD, Kirkpatrick BC. 1996. Physical map of the western X-disease phytoplasma chromosome. J Bacteriol 178 :3985–3988. doi:10.1128/jb.178.13.3985-3988.1996.8682811 6 Lee I-M, Shao J, Bottner-Parker KD, Gundersen-Rindal DE, Zhao Y, Davis RE. 2015. Draft genome sequence of “Candidatus Phytoplasma pruni” strain CX, a plant-pathogenic bacterium. Genome Announc 3 :e01117-15. doi:10.1128/genomeA.01117-15.26472824 7 Davis RE, Zhao Y, Dally EL, Lee I-M, Jomantiene R, Douglas SM. 2013. “Candidatus Phytoplasma pruni,” a novel taxon associated with X-disease of stone fruits, Prunus spp.: multilocus characterization based on 16S rRNA, secY, and ribosomal protein genes. Int J Syst Evol Microbiol 63 :766–776. doi:10.1099/ijs.0.041202-0.22798643

36346245

PMC9753602

NO-CC CODE

2022-12-16 23:26:09

no

Microbiol Resour Announc.; 11(12):e00790-22

utf-8

Microbiol Resour Announc

10.1128/mra.00790-22

oa_other

==== Front Microbiol Resour Announc Microbiol Resour Announc mra Microbiology Resource Announcements 2576-098X American Society for Microbiology 1752 N St., N.W., Washington, DC 36346245 00790-22 10.1128/mra.00790-22 mra.00790-22 Genome Sequences plant-microbiologyPlant MicrobiologyDraft Genome Sequence of a Washington Isolate of “Candidatus Phytoplasma pruni” https://orcid.org/0000-0003-4814-4372 Wright Alice A. a alice.wright@usda.gov Harper Scott J. b a Sugarcane Research Unit, USDA Agricultural Research Service, Houma, Louisiana, USA b Department of Plant Pathology, Washington State University, Prosser, Washington, USA Editor Newton Irene L. G. Indiana University, Bloomington The authors declare no conflict of interest. 8 11 2022 12 2022 8 11 2022 11 12 e00790-229 8 2022 27 10 2022 https://doi.org/10.1128/AuthorWarrantyLicense.v1 This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply. ABSTRACT Illumina sequencing of a Prunus avium tree with X-disease symptoms was performed to obtain a draft genome of “Candidatus Phytoplasma pruni.” The genome consists of 14 contigs covering 588,767 bp. This is the first metagenome to be sequenced from the current X-disease epidemic in stone fruit in the Pacific Northwest. Washington Tree Fruit Research Commission (WTFRC) https://doi.org/10.13039/100012487 GR00002025 Wright Alice Ann Harper Scott cover-dateDecember 2022 ==== Body pmcANNOUNCEMENT The causative agent of X-disease in Prunus avium (cherry) and Prunus persica (peach/nectarine), which has reached epidemic levels in stone fruit in the Pacific Northwest, is “Candidatus Phytoplasma pruni” (1, 2). In peaches and nectarines, infection by “Ca. Phytoplasma pruni” causes premature yellowing of leaves and spots of necrosis, producing a shot-hole-like appearance. The pathogen also causes limb dieback and, within approximately 5 years of infection, tree death (1). In cherries, the fruit is affected; at harvest, the fruit is small and misshapen, with poor color and taste and often a reduction in sugar content (1, 2). Like other phytoplasmas, “Ca. Phytoplasma pruni” is a cell wall-less bacterium that currently cannot be cultured independently of its host (3). Developing improved genomic resources for this pathogen is crucial to understanding and managing this disease. To that end, an isolate from the current epidemic was sequenced. Cuttings from a sweet cherry tree expressing classic X-disease symptoms were collected from a commercial orchard in Zillah, Washington. DNA was extracted from woody stem tissue because this is the tissue in which phytoplasma titers are highest (4). Briefly, 100 mg of tissue was frozen at −80°C for at least 1 h prior to homogenization to a fine powder by bead-beating in a TissueLyser (Qiagen, Germantown, MD) at 20 beats/s for 30 s. DNA was then extracted with a DNeasy Plant minikit (Qiagen) according to the manufacturer’s instructions. Phytoplasma presence was confirmed by quantitative PCR (qPCR) as described (4), and the DNA concentration was determined using a Qubit fluorometer (Thermo Fisher Scientific, Hampton, NH). DNA was submitted to Genewiz (South Plainfield, NJ) for library preparation using a NEBNext Ultra library preparation kit (New England Biolabs, Ipswich, MA) and 250-bp paired-end sequencing with a single HiSeq flow cell on a HiSeq 2500 system (Illumina Inc., San Diego, CA, USA). Raw reads were processed with CLC Genomics Workbench v21.0.5 (Qiagen, Hilden, Germany) using default settings unless stated otherwise. Sequencing yielded 309,073,036 raw reads, which were reduced to 309,072,714 reads after adaptor removal and quality assessment; this was performed using the Trim Reads pipeline with the default quality limit of 0.05. De novo assembly was performed on the trimmed reads using the CLC Genomics Workbench de novo assembly tool. Minimum contig length was set to 200 bp, word size to 26, and bubble size to 246. The assembly consisted of 343,722 contigs, with an N50 value of 2,935 bp. BLASTn analysis with a word size of 50 was used to identify 371 phytoplasma-like sequences, which were then imported into Geneious v2022.0.2 (Biomatters, Inc., Auckland, New Zealand) to find open reading frames (ORFs). ORF identification was performed with a minimum size set to 300 bp and the genetic code set to standard. Genes were confirmed and where possible assigned a function based on homology to known phytoplasma genes using BLASTp. Of the 371 contigs, 14 were large enough to contain ORFs; the remainder were small fragments that might have been the result of repetitive regions that could not be resolved by Illumina sequencing. These 14 contigs were composed of a total of 588,767 bp, with a GC content of 27.1%. The average sequencing depth was 18,538×, with 45,212,112 reads mapped to the contigs. Through the Geneious ORF prediction tool and manual curation, 469 genes were annotated. These genes were assigned functions based on homology to genes in other phytoplasma genomes. The estimated size of the “Ca. Phytoplasma pruni” genome is 677 kb (5). This draft genome represents approximately 87% of the full genome. It is similar in size to the draft genome published previously (6); however, there are fewer contigs in the genome of this metagenome. The genome sequenced previously (6) was from a strain that was originally from an infected peach tree in Connecticut (7) and was maintained in periwinkle, making it temporally and geographically separate from the Washington sample. The genome presented here was sequenced directly from an infected cherry in Zillah, Washington, and is representative of the current epidemic in the Pacific Northwest. Additional work, including comparative genomic analysis among geographically diverse isolates, may explain why the current epidemic is so devastating. Data availability. Raw reads have been submitted to the SRA under BioProject accession number PRJNA857992. The assembled and annotated genome has been deposited in GenBank with the accession number JANIEM000000000. ACKNOWLEDGMENTS We appreciate the orchard managers allowing us to collect tissue for this study. This study was funded by the Washington Tree Fruit Research Commission. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA. ==== Refs REFERENCES 1 Uyemoto JK, Kirkpatrick BC. 2011. X-disease phytoplasma, p 243–245. In Hadidi A, Barba M, Candresse T, Jelkman W (ed), Virus and virus-like diseases of pome and stone fruits. APS Press, St. Paul, MN. 2 Wright AA, Shires MK, Beaver C, Bishop G, DuPont ST, Naranjo R, Harper S. 2021. Effect of “Candidatus Phytoplasma pruni” infection on sweet cherry fruit. Phtyopathology 111 :2195–2202. doi:10.1094/PHYTO-03-21-0106-R. 3 Hogenhout SA, Oshima K, Ammar E-D, Kakizawa S, Kingdom HN, Namba S. 2008. Phytoplasmas: bacteria that manipulate plants and insects. Mol Plant Pathol 9 :403–423. doi:10.1111/j.1364-3703.2008.00472.x.18705857 4 Wright AA, Shires MK, Molnar C, Bishop G, Johnson AM, Frias C, Harper S. 2022. Titer and distribution of “Candidatus Phytoplasma pruni” in Prunus avium. Phytopathology 112 :1406–1412. doi:10.1094/PHYTO-11-21-0468-R.35021858 5 Firrao G, Smart CD, Kirkpatrick BC. 1996. Physical map of the western X-disease phytoplasma chromosome. J Bacteriol 178 :3985–3988. doi:10.1128/jb.178.13.3985-3988.1996.8682811 6 Lee I-M, Shao J, Bottner-Parker KD, Gundersen-Rindal DE, Zhao Y, Davis RE. 2015. Draft genome sequence of “Candidatus Phytoplasma pruni” strain CX, a plant-pathogenic bacterium. Genome Announc 3 :e01117-15. doi:10.1128/genomeA.01117-15.26472824 7 Davis RE, Zhao Y, Dally EL, Lee I-M, Jomantiene R, Douglas SM. 2013. “Candidatus Phytoplasma pruni,” a novel taxon associated with X-disease of stone fruits, Prunus spp.: multilocus characterization based on 16S rRNA, secY, and ribosomal protein genes. Int J Syst Evol Microbiol 63 :766–776. doi:10.1099/ijs.0.041202-0.22798643

34973716

PMC9753609

NO-CC CODE

2022-12-16 23:26:09

no

Lancet. 2022 Dec 31 1-7 January; 399(10319):23-24

latin-1

Lancet

10.1016/S0140-6736(21)02727-6

oa_other

==== Front Microbiol Resour Announc Microbiol Resour Announc mra Microbiology Resource Announcements 2576-098X American Society for Microbiology 1752 N St., N.W., Washington, DC 36409113 00881-22 10.1128/mra.00881-22 mra.00881-22 Other Genetic Resources genomics-and-proteomicsGenomics and ProteomicsComplete Structural Predictions of the Proteome of African Swine Fever Virus Strain Georgia 2007 Spinard Edward a Azzinaro Paul a Rai Ayushi b Espinoza Nallely a Ramirez-Medina Elizabeth a Valladares Alyssa a Borca Manuel V. a Manuel.Borca@usda.gov https://orcid.org/0000-0002-7894-0233 Gladue Douglas P. a Douglas.Gladue@usda.gov a Plum Island Animal Disease Center, Agricultural Research Service, USDA, Greenport, New York, USA b Oak Ridge Institute for Science and Education (ORISE), Oak Ridge, Tennessee, USA Editor Stedman Kenneth M. Portland State University The authors declare no conflict of interest. 21 11 2022 12 2022 21 11 2022 11 12 e00881-2230 8 2022 6 10 2022 https://doi.org/10.1128/AuthorWarrantyLicense.v1 This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply. ABSTRACT Here, we announce the predicted structures of the 193 proteins encoded by African swine fever virus (ASFV) strain Georgia 2007 (ASFV-G). Previously, only the structures of 16 ASFV proteins were elucidated. USDA | Agricultural Research Service (ARS) https://doi.org/10.13039/100007917 1940-32000-063-00D Borca Manuel V. Gladue Douglas Paul cover-dateDecember 2022 ==== Body pmcANNOUNCEMENT African swine fever (ASF) is a highly infectious and fatal disease of feral and domesticated swine that is the only member of the family Asfarviridae and genus Asfivirus. ASF has spread through several Eurasian countries and, more recently, has reappeared for the first time in over 40 years in the Western Hemisphere, with a positive identification in the Dominican Republic (1). Currently, only the structures of 16 of the nearly 200 predicted ASF virus (ASFV) proteins have been solved via X-ray diffraction, solution nuclear magnetic resonance (NMR), or electron microscopy and are hosted as 52 Protein Data Bank (PDB) files by the Research Collaboratory for Structural Bioinformatics (2–19). In the 14th Critical Assessment of Protein Structure, AlphaFold was determined to demonstrate protein structure accuracy comparable to experimentally resolved structures (20). Accordingly, the structures of all 193 proteins predicted to be encoded by the progenitor strain from the most recent outbreak, ASFV Georgia 2007 (ASFV-G) (GenBank accession number FR682468), were determined using AlphaFold (21). Utilizing the U.S. Department of Agriculture’s (USDA) Agricultural Research Service (ARS) Scientific Computing Initiative (SCINet) Ceres high-performance computing (HPC) cluster or the SCINet/Mississippi State University (MSU) collaborative Atlas HPC cluster, structural predictions were performed for all proteins (except QP509L) by running AlphaFold v2.2.0 using the default databases and the following parameters: model_preset = monomer, db_preset = full_dbs, use_gpu_relax = True, max_template_date = 2020-05-14. Due to hardware limitations, QP509L was instead predicted using the AlphaFold v2.1.0 colab notebook without run relaxation or homologous structures and using a reduced Big Fantastic Database (BFD) (20). The unrelaxed model was then minimized and subjected to molecular dynamics for 1 ns using GROMACS (22–26). The per-residue estimate of confidence (pLDDT) generated using AlphaFold is included in each structure file. The predicted protein structures are essential for the in silico prediction of B-cell and T-cell epitopes and the development of antivirals. Data availability. All PDB structure files can be found on the download page at the website for the Center of Excellence for African Swine Fever Genomics (https://asfvgenomics.com/) or in ModelArchive at https://modelarchive.org/doi/10.5452/ma-asfv-asfvg. Individual ASFV protein structural predictions can also be found in Table 1. TABLE 1 Predicted structures of the 193 proteins encoded by ASFV strain Georgia 2007 Protein ASFV Genomics link ModelArchive link(s) 285L https://asfvgenomics.com/ASFV/285L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-001 A104R https://asfvgenomics.com/ASFV/A104R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-002 A118R https://asfvgenomics.com/ASFV/A118R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-003 A137R https://asfvgenomics.com/ASFV/A137R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-004 A151R https://asfvgenomics.com/ASFV/A151R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-005 A179L https://asfvgenomics.com/ASFV/A179L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-006 A224L https://asfvgenomics.com/ASFV/A224L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-007 A238L https://asfvgenomics.com/ASFV/A238L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-008 A240L https://asfvgenomics.com/ASFV/A240L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-009 A859L https://asfvgenomics.com/ASFV/A859L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-010 ASFV G ACD 00090 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000090.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-011 ASFV G ACD 00120 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000120.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-012 ASFV G ACD 00160 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000160.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-013 ASFV G ACD 00190 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000190.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-014 ASFV G ACD 00210 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000210.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-015 ASFV G ACD 00240 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000240.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-016 ASFV G ACD 00270 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000270.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-017 ASFV G ACD 00300 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000300.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-018 ASFV G ACD 00320 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000320.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-019 ASFV G ACD 00330 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000330.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-020 ASFV G ACD 00350 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000350.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-021 ASFV G ACD 00360 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000360.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-022 ASFV G ACD 00520 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000520.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-023 ASFV G ACD 00600 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2000600.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-024 ASFV G ACD 01020 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2001020.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-025 ASFV G ACD 01870 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2001870.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-026 ASFV G ACD 01940 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2001940.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-027 ASFV G ACD 01960 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2001960.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-028 ASFV G ACD 01980 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2001980.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-029 ASFV G ACD 01990 https://asfvgenomics.com/ASFV/ASFV%20G%20ACD%2001990.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-030 B117L https://asfvgenomics.com/ASFV/B117L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-031 B119L https://asfvgenomics.com/ASFV/B119L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-032 B125R https://asfvgenomics.com/ASFV/B125R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-033 B169L https://asfvgenomics.com/ASFV/B169L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-034 B175L https://asfvgenomics.com/ASFV/B175L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-035 B263R https://asfvgenomics.com/ASFV/B263R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-036 B318L https://asfvgenomics.com/ASFV/B318L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-037 B354L https://asfvgenomics.com/ASFV/B354L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-038 B385R https://asfvgenomics.com/ASFV/B385R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-039 B407L https://asfvgenomics.com/ASFV/B407L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-040 B438L https://asfvgenomics.com/ASFV/B438L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-041 B475L https://asfvgenomics.com/ASFV/B475L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-042 B602L https://asfvgenomics.com/ASFV/B602L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-043 B646L https://asfvgenomics.com/ASFV/B646L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-044 B66L https://asfvgenomics.com/ASFV/B66L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-045 B962L https://asfvgenomics.com/ASFV/B962L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-046 C122R https://asfvgenomics.com/ASFV/C122R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-047 C129R https://asfvgenomics.com/ASFV/C129R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-048 C147L https://asfvgenomics.com/ASFV/C147L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-049 C257L https://asfvgenomics.com/ASFV/C257L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-050 C315R https://asfvgenomics.com/ASFV/C315R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-051 C475L https://asfvgenomics.com/ASFV/C475L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-052 C62L https://asfvgenomics.com/ASFV/C62L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-053 C717R https://asfvgenomics.com/ASFV/C717R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-054 C84L https://asfvgenomics.com/ASFV/C84L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-055 C962R https://asfvgenomics.com/ASFV/C962R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-056 CP123L https://asfvgenomics.com/ASFV/CP123L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-057 CP204L https://asfvgenomics.com/ASFV/CP204L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-058 CP2475L https://asfvgenomics.com/ASFV/CP2475L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-059, https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-060, https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-061, https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-062, https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-063 CP312R https://asfvgenomics.com/ASFV/CP312R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-064 CP530R https://asfvgenomics.com/ASFV/CP530R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-065 CP80R https://asfvgenomics.com/ASFV/CP80R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-066 D1133L https://asfvgenomics.com/ASFV/D1133L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-067 D117L https://asfvgenomics.com/ASFV/D117L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-068 D129L https://asfvgenomics.com/ASFV/D129L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-069 D205R https://asfvgenomics.com/ASFV/D205R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-070 D250R https://asfvgenomics.com/ASFV/D250R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-071 D339L https://asfvgenomics.com/ASFV/D339L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-072 D345L https://asfvgenomics.com/ASFV/D345L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-073 DP238L https://asfvgenomics.com/ASFV/DP238L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-074 DP60R https://asfvgenomics.com/ASFV/DP60R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-075 DP71L https://asfvgenomics.com/ASFV/DP71L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-076 DP79L https://asfvgenomics.com/ASFV/DP79L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-077 DP96R https://asfvgenomics.com/ASFV/DP96R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-078 E111R https://asfvgenomics.com/ASFV/E111R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-079 E120R https://asfvgenomics.com/ASFV/E120R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-080 E146L https://asfvgenomics.com/ASFV/E146L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-081 E165R https://asfvgenomics.com/ASFV/E165R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-082 E183L https://asfvgenomics.com/ASFV/E183L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-083 E184L https://asfvgenomics.com/ASFV/E184L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-084 E199L https://asfvgenomics.com/ASFV/E199L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-085 E248R https://asfvgenomics.com/ASFV/E248R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-086 E301R https://asfvgenomics.com/ASFV/E301R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-087 E423R https://asfvgenomics.com/ASFV/E423R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-088 E66L https://asfvgenomics.com/ASFV/E66L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-089 EP1242L https://asfvgenomics.com/ASFV/EP1242L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-090 EP152R https://asfvgenomics.com/ASFV/EP152R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-091 EP153R https://asfvgenomics.com/ASFV/EP153R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-092 EP296R https://asfvgenomics.com/ASFV/EP296R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-093 EP364R https://asfvgenomics.com/ASFV/EP364R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-094 EP402R https://asfvgenomics.com/ASFV/EP402R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-095 EP424R https://asfvgenomics.com/ASFV/EP424R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-096 EP84R https://asfvgenomics.com/ASFV/EP84R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-097 F1055L https://asfvgenomics.com/ASFV/F1055L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-098 F165R https://asfvgenomics.com/ASFV/F165R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-099 F317L https://asfvgenomics.com/ASFV/F317L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-100 F334L https://asfvgenomics.com/ASFV/F334L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-101 F778R https://asfvgenomics.com/ASFV/F778R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-102 G1211R https://asfvgenomics.com/ASFV/G1211R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-103 G1340L https://asfvgenomics.com/ASFV/G1340L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-104 H108R https://asfvgenomics.com/ASFV/H108R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-105 H124R https://asfvgenomics.com/ASFV/H124R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-106 H171R https://asfvgenomics.com/ASFV/H171R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-107 H233R https://asfvgenomics.com/ASFV/H233R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-108 H240R https://asfvgenomics.com/ASFV/H240R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-109 H339R https://asfvgenomics.com/ASFV/H339R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-110 H359L https://asfvgenomics.com/ASFV/H359L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-111 hypothetical_01 https://asfvgenomics.com/ASFV/hypothetical_01.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-112 hypothetical_02 https://asfvgenomics.com/ASFV/hypothetical_02.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-113 hypothetical_03 https://asfvgenomics.com/ASFV/hypothetical_03.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-114 I10L https://asfvgenomics.com/ASFV/I10L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-115 I177L https://asfvgenomics.com/ASFV/I177L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-116 I196L https://asfvgenomics.com/ASFV/I196L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-117 I215L https://asfvgenomics.com/ASFV/I215L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-118 I226R https://asfvgenomics.com/ASFV/I226R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-119 I243L https://asfvgenomics.com/ASFV/I243L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-120 I267L https://asfvgenomics.com/ASFV/I267L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-121 I329L https://asfvgenomics.com/ASFV/I329L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-122 I73R https://asfvgenomics.com/ASFV/I73R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-123 I7L https://asfvgenomics.com/ASFV/I7L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-124 I8L https://asfvgenomics.com/ASFV/I8L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-125 I9R https://asfvgenomics.com/ASFV/I9R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-126 K145R https://asfvgenomics.com/ASFV/K145R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-127 K196R https://asfvgenomics.com/ASFV/K196R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-128 K205R https://asfvgenomics.com/ASFV/K205R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-129 K421R https://asfvgenomics.com/ASFV/K421R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-130 K78R https://asfvgenomics.com/ASFV/K78R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-131 KP177R https://asfvgenomics.com/ASFV/KP177R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-132 L11L https://asfvgenomics.com/ASFV/L11L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-133 L60L https://asfvgenomics.com/ASFV/L60L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-134 L83L https://asfvgenomics.com/ASFV/L83L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-135 M1249L https://asfvgenomics.com/ASFV/M1249L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-136 M448R https://asfvgenomics.com/ASFV/M448R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-137 MGF 100-1L https://asfvgenomics.com/ASFV/MGF%20100-1L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-138 MGF 100-1R https://asfvgenomics.com/ASFV/MGF%20100-1R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-139 MGF 100-3L https://asfvgenomics.com/ASFV/MGF%20100-3L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-140 MGF 110-10L–MGF110-14L fusion https://asfvgenomics.com/ASFV/MGF%20110-10-L%20-%20MGF110-14L%20fusion.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-141 MGF 110-12L https://asfvgenomics.com/ASFV/MGF%20110-12L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-142 MGF 110-13La https://asfvgenomics.com/ASFV/MGF%20110-13La.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-143 MGF 110-13Lb https://asfvgenomics.com/ASFV/MGF%20110-13Lb.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-144 MGF 110-1L https://asfvgenomics.com/ASFV/MGF%20110-1L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-145 MGF 110-2L https://asfvgenomics.com/ASFV/MGF%20110-2L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-146 MGF 110-3L https://asfvgenomics.com/ASFV/MGF%20110-3L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-147 MGF 110-4L https://asfvgenomics.com/ASFV/MGF%20110-4L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-148 MGF 110-5L-6L https://asfvgenomics.com/ASFV/MGF%20110-5L-6L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-149 MGF 110-7L https://asfvgenomics.com/ASFV/MGF%20110-7L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-150 MGF 110-8L https://asfvgenomics.com/ASFV/MGF%20110-8L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-151 MGF 110-9L https://asfvgenomics.com/ASFV/MGF%20110-9L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-152 MGF 300-1L https://asfvgenomics.com/ASFV/MGF%20300-1L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-153 MGF 300-2R https://asfvgenomics.com/ASFV/MGF%20300-2R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-154 MGF 300-4L https://asfvgenomics.com/ASFV/MGF%20300-4L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-155 MGF 360-10L https://asfvgenomics.com/ASFV/MGF%20360-10L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-156 MGF 360-11L https://asfvgenomics.com/ASFV/MGF%20360-11L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-157 MGF 360-12L https://asfvgenomics.com/ASFV/MGF%20360-12L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-158 MGF 360-13L https://asfvgenomics.com/ASFV/MGF%20360-13L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-159 MGF 360-14L https://asfvgenomics.com/ASFV/MGF%20360-14L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-160 MGF 360-15R https://asfvgenomics.com/ASFV/MGF%20360-15R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-161 MGF 360-16R https://asfvgenomics.com/ASFV/MGF%20360-16R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-162 MGF 360-18R https://asfvgenomics.com/ASFV/MGF%20360-18R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-163 MGF 360-19Ra https://asfvgenomics.com/ASFV/MGF%20360-19Ra.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-164 MGF 360-19Rb https://asfvgenomics.com/ASFV/MGF%20360-19Rb.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-165 MGF 360-1La https://asfvgenomics.com/ASFV/MGF%20360-1La.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-166 MGF 360-1Lb https://asfvgenomics.com/ASFV/MGF%20360-1Lb.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-167 MGF 360-21R https://asfvgenomics.com/ASFV/MGF%20360-21R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-168 MGF 360-2L https://asfvgenomics.com/ASFV/MGF%20360-2L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-169 MGF 360-3L https://asfvgenomics.com/ASFV/MGF%20360-3L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-170 MGF 360-4L https://asfvgenomics.com/ASFV/MGF%20360-4L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-171 MGF 360-6L https://asfvgenomics.com/ASFV/MGF%20360-6L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-172 MGF 360-8L https://asfvgenomics.com/ASFV/MGF%20360-8L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-173 MGF 360-9L https://asfvgenomics.com/ASFV/MGF%20360-9L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-174 MGF 505-10R https://asfvgenomics.com/ASFV/MGF%20505-10R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-175 MGF 505-11L https://asfvgenomics.com/ASFV/MGF%20505-11L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-176 MGF 505-1R https://asfvgenomics.com/ASFV/MGF%20505-1R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-177 MGF 505-2R https://asfvgenomics.com/ASFV/MGF%20505-2R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-178 MGF 505-3R https://asfvgenomics.com/ASFV/MGF%20505-3R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-179 MGF 505-4R https://asfvgenomics.com/ASFV/MGF%20505-4R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-180 MGF 505-5R https://asfvgenomics.com/ASFV/MGF%20505-5R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-181 MGF 505-6R https://asfvgenomics.com/ASFV/MGF%20505-6R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-182 MGF 505-7R https://asfvgenomics.com/ASFV/MGF%20505-7R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-183 MGF 505-9R https://asfvgenomics.com/ASFV/MGF%20505-9R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-184 NP1450L https://asfvgenomics.com/ASFV/NP1450L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-185 NP419L https://asfvgenomics.com/ASFV/NP419L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-186 NP868R https://asfvgenomics.com/ASFV/NP868R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-187 O174L https://asfvgenomics.com/ASFV/O174L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-188 O61R https://asfvgenomics.com/ASFV/O61R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-189 P1192R https://asfvgenomics.com/ASFV/P1192R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-190 Q706L https://asfvgenomics.com/ASFV/Q706L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-191 QP383R https://asfvgenomics.com/ASFV/QP383R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-192 QP509L https://asfvgenomics.com/ASFV/QP509L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-193 R298L https://asfvgenomics.com/ASFV/R298L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-194 S183L https://asfvgenomics.com/ASFV/S183L.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-195 S273R https://asfvgenomics.com/ASFV/S273R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-196 X69R https://asfvgenomics.com/ASFV/X69R.html https://modelarchive.org/doi/10.5452/ma-asfv-asfvg-197 ACKNOWLEDGMENTS This work was supported by research funded by the U.S. Department of Agriculture, Agricultural Research Service (ARS)-CRIS project 1940-32000-063-00D. This research used resources provided by the SCINet project and the AI Center of Excellence of the USDA Agricultural Research Service (ARS project number 0500-00093-001-00-D). We thank the SCINet Virtual Research Support Core (VRSC) in Ames, IA, for hosting the Ceres HPC cluster and the collaboration between MSU and the U.S. Department of Agriculture’s Agricultural Research Service for hosting the Atlas HPC cluster. ==== Refs REFERENCES 1 Gonzales W, Moreno C, Duran U, Henao N, Bencosme M, Lora P, Reyes R, Nunez R, De Gracia A, Perez AM. 2021. African swine fever in the Dominican Republic. Transbound Emerg Dis 68 :3018–3019. doi:10.1111/tbed.14341.34609795 2 Banjara S, Caria S, Dixon LK, Hinds MG, Kvansakul M. 2017. Structural insight into African swine fever virus A179L-mediated inhibition of apoptosis. J Virol 91 :e02228-16. doi:10.1128/JVI.02228-16.28053104 3 Banjara S, Shimmon GL, Dixon LK, Netherton CL, Hinds MG, Kvansakul M. 2019. Crystal structure of African swine fever virus A179L with the autophagy regulator Beclin. Viruses 11 :789. doi:10.3390/v11090789.31461953 4 Chen Y, Chen X, Huang Q, Shao Z, Gao Y, Li Y, Yang C, Liu H, Li J, Wang Q, Ma J, Zhang Y-Z, Gu Y, Gan J. 2020. A unique DNA-binding mode of African swine fever virus AP endonuclease. Cell Discov 6 :13. doi:10.1038/s41421-020-0146-2.32194979 5 Chen Y, Liu H, Yang C, Gao Y, Yu X, Chen X, Cui R, Zheng L, Li S, Li X, Ma J, Huang Z, Li J, Gan J. 2019. Structure of the error-prone DNA ligase of African swine fever virus identifies critical active site residues. Nat Commun 10 :387. doi:10.1038/s41467-019-08296-w.30674878 6 Du X, Gao Z-Q, Geng Z, Dong Y-H, Zhang H. 2021. Structure and biochemical characteristic of the methyltransferase (MTase) domain of RNA capping enzyme from African swine fever virus. J Virol 95 :e02029-20. doi:10.1128/JVI.02029-20. 7 Guo F, Shi Y, Yang M, Guo Y, Shen Z, Li M, Chen Y, Liang R, Yang Y, Chen H, Peng G. 2021. The structural basis of African swine fever virus core shell protein p15 binding to DNA. FASEB J 35 :e21350. doi:10.1096/fj.202002145R.33629764 8 Hakim M, Fass D. 2009. Dimer interface migration in a viral sulfhydryl oxidase. J Mol Biol 391 :758–768. doi:10.1016/j.jmb.2009.06.070.19576902 9 Huang J-W, Niu D, Liu K, Wang Q, Ma L, Chen C-C, Zhang L, Liu W, Zhou S, Min J, Wu S, Yang Y, Guo R-T. 2020. Structure basis of non-structural protein pA151R from African swine fever virus. Biochem Biophys Res Commun 532 :108–113. doi:10.1016/j.bbrc.2020.08.011.32828542 10 Li C, Chai Y, Song H, Weng C, Qi J, Sun Y, Gao GF. 2019. Crystal structure of African swine fever virus dUTPase reveals a potential drug target. mBio 10 :e02483-19. doi:10.1128/mBio.02483-19.31662460 11 Li G, Fu D, Zhang G, Zhao D, Li M, Geng X, Sun D, Wang Y, Chen C, Jiao P, Cao L, Guo Y, Rao Z. 2020. Crystal structure of the African swine fever virus structural protein p35 reveals its role for core shell assembly. Protein Cell 11 :600–605. doi:10.1007/s13238-020-00730-w.32519301 12 Li G, Liu X, Yang M, Zhang G, Wang Z, Guo K, Gao Y, Jiao P, Sun J, Chen C, Wang H, Deng W, Xiao H, Li S, Wu H, Wang Y, Cao L, Jia Z, Shang L, Yang C, Guo Y, Rao Z. 2020. Crystal structure of African swine fever virus pS273R protease and implications for inhibitor design. J Virol 94 :e02125-19. doi:10.1128/JVI.02125-19.32075933 13 Li G, Wang C, Yang M, Cao L, Fu D, Liu X, Sun D, Chen C, Wang Y, Jia Z, Yang C, Guo Y, Rao Z. 2020. Structural insight into African swine fever virus dUTPase reveals a novel folding pattern in the dUTPase family. J Virol 94 :e01698-19. doi:10.1128/JVI.01698-19.31748385 14 Liang R, Wang G, Zhang D, Ye G, Li M, Shi Y, Shi J, Chen H, Peng G. 2021. Structural comparisons of host and African swine fever virus dUTPases reveal new clues for inhibitor development. J Biol Chem 296 :100015. doi:10.1074/jbc.RA120.014005.33139328 15 Liu Q, Ma B, Qian N, Zhang F, Tan X, Lei J, Xiang Y. 2019. Structure of the African swine fever virus major capsid protein p72. Cell Res 29 :953–955. doi:10.1038/s41422-019-0232-x.31530894 16 Liu R, Sun Y, Chai Y, Li S, Li S, Wang L, Su J, Yu S, Yan J, Gao F, Zhang G, Qiu H-J, Gao GF, Qi J, Wang H. 2020. The structural basis of African swine fever virus pA104R binding to DNA and its inhibition by stilbene derivatives. Proc Natl Acad Sci USA 117 :11000–11009. doi:10.1073/pnas.1922523117.32358196 17 Maciejewski MW, Shin R, Pan B, Marintchev A, Denninger A, Mullen MA, Chen K, Gryk MR, Mullen GP. 2001. Solution structure of a viral DNA repair polymerase. Nat Struct Biol 8 :936–941. doi:10.1038/nsb1101-936.11685238 18 Showalter AK, Byeon IJ, Su MI, Tsai MD. 2001. Solution structure of a viral DNA polymerase X and evidence for a mutagenic function. Nat Struct Biol 8 :942–946. doi:10.1038/nsb1101-942.11685239 19 Wu W-J, Su M-I, Wu J-L, Kumar S, Lim L-H, Wang C-WE, Nelissen FHT, Chen M-CC, Doreleijers JF, Wijmenga SS, Tsai M-D. 2014. How a low-fidelity DNA polymerase chooses non-Watson-Crick from Watson-Crick incorporation. J Am Chem Soc 136 :4927–4937. doi:10.1021/ja4102375.24617852 20 Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Žídek A, Potapenko A, Bridgland A, Meyer C, Kohl SAA, Ballard AJ, Cowie A, Romera-Paredes B, Nikolov S, Jain R, Adler J, Back T, Petersen S, Reiman D, Clancy E, Zielinski M, Steinegger M, Pacholska M, Berghammer T, Bodenstein S, Silver D, Vinyals O, Senior AW, Kavukcuoglu K, Kohli P, Hassabis D. 2021. Highly accurate protein structure prediction with AlphaFold. Nature 596 :583–589. doi:10.1038/s41586-021-03819-2.34265844 21 Chapman DA, Darby AC, Da Silva M, Upton C, Radford AD, Dixon LK. 2011. Genomic analysis of highly virulent Georgia 2007/1 isolate of African swine fever virus. Emerg Infect Dis 17 :599–605. doi:10.3201/eid1704.101283.21470447 22 Berendsen HJC, van der Spoel D, van Drunen R. 1995. GROMACS—a message-passing parallel molecular-dynamics implementation. Comput Phys Commun 91 :43–56. doi:10.1016/0010-4655(95)00042-E. 23 Lindahl E, Hess B, van der Spoel D. 2001. GROMACS 3.0: a package for molecular simulation and trajectory analysis. J Mol Model 7 :306–317. doi:10.1007/s008940100045. 24 Van der Spoel D, Lindahl E, Hess B, Groenhof G, Mark AE, Berendsen HJC. 2005. GROMACS: fast, flexible, and free. J Comput Chem 26 :1701–1718. doi:10.1002/jcc.20291.16211538 25 Hess B, Kutzner C, van der Spoel D, Lindahl E. 2008. GROMACS 4: algorithms for highly efficient, load-balanced, and scalable molecular simulation. J Chem Theory Comput 4 :435–447. doi:10.1021/ct700301q.26620784 26 Pronk S, Páll S, Schulz R, Larsson P, Bjelkmar P, Apostolov R, Shirts MR, Smith JC, Kasson PM, van der Spoel D, Hess B, Lindahl E. 2013. GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 29 :845–854. doi:10.1093/bioinformatics/btt055.23407358

36409113

PMC9753626

NO-CC CODE

2022-12-16 23:26:10

no

Microbiol Resour Announc.; 11(12):e00881-22

utf-8

Microbiol Resour Announc

10.1128/mra.00881-22

oa_other

==== Front Microbiol Resour Announc Microbiol Resour Announc mra Microbiology Resource Announcements 2576-098X American Society for Microbiology 1752 N St., N.W., Washington, DC 36377884 00921-22 10.1128/mra.00921-22 mra.00921-22 Genome Sequences applied-and-industrial-microbiologyApplied and Industrial MicrobiologyThe Genome Sequence of an Algal Strain of Nannochloropsis QH25 https://orcid.org/0000-0003-4413-9058 Sanchez Monica R. a monsanc@sandia.gov Biondi Thomas C. b Kunde Yuliya A. b Eng Wyatt b Nalley Jakob O. c Ganuza Eneko c Hovde Blake T. b Corcoran Alina A. d https://orcid.org/0000-0002-8564-1274 Starkenburg Shawn R. b a Sandia National Laboratory, New Mexico, Albuquerque, New Mexico, USA b Los Alamos National Laboratory, Los Alamos, New Mexico, USA c Qualitas Health, Imperial, Texas, USA d New Mexico Consortium, Los Alamos, New Mexico, USA Editor Stajich Jason E. University of California, Riverside The authors declare no conflict of interest. 15 11 2022 12 2022 15 11 2022 11 12 e00921-221 9 2022 2 11 2022 https://doi.org/10.1128/AuthorWarrantyLicense.v1 This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply. ABSTRACT Species of Nannochloropsis are single-celled Stramenopiles commonly used in microalgae-based technologies for the manufacturing of bioproducts. Nannochloropsis oceanica QH25 was isolated from an algal cultivation pond located in Imperial, Texas (USA). We used PacBio continuous long read (CLR) sequencing to produce a highly contiguous 29.34 Mb genome. Connecticut Department of Energy and Environmental Protection (DEEP) https://doi.org/10.13039/100004820 DE #EE0008122 Corcoran Alina A. Starkenburg Shawn R. cover-dateDecember 2022 ==== Body pmcANNOUNCEMENT Microalgae are commonly used to produce bioproducts, including food, feed, fuel, pharmaceuticals, and nutraceuticals. Yet, strains that are adapted to outdoor cultivation are rarely sequenced, limiting our understanding of microalgal cultivar physiology, biology, and ecology. Sequencing and analysis of the genomes of robust microalgal cultivars are needed to support strain development and engineering for use in (microbial) agronomy. The QH25 strain was grown for over 5 years in an outdoor 1.3-acre open pond before it was isolated. In July 2020, the strain was isolated from a 1,527,000-L pond at a production facility (Qualitas Health) in Imperial, Texas, USA (31°26'96.1″N 102°68'32.7″W; elevation, 2,392′ ft). At the time of isolation, the pond was operated at 0.5 g ash free dry weight (AFDW)/L, a depth of 25 to 30 cm, with a flow rate of 30 cm/s. Media was at 32°C, 20 ppt salinity. No known pathogens were present during sample collection. This strain was isolated using single cell flow cytometry, cultivated in standard media, including f/2, and has been cryopreserved with dimethyl sulfoxide (DMSO) (1). Library preparation and sequencing. A high molecular weight DNA extraction protocol was applied similarly to previous work (2). Briefly, cells were centrifuged and washed and pipetted into agarose plug molds. Cells were incubated at 4°C, followed by an incubation in the protoplasting solution overnight at 37°C. An incubation in the lysis buffer with proteinase K followed. Lysed plugs were washed 3× and digested overnight with beta-Agarose. gDNA was purified according to the high salt:phenol:chloroform:isoamyl alcohol protocol and concentrated with AMPure PB beads. The large-insert SMRTbell library was prepared according to the Pacific Biosciences protocol. Sequencing primer v.4 was annealed and Sequel DNA polymerase 3.0 bound to the SMRT bells. Diluted SMRTbell/DNA polymerase complex was sequenced on two SMRT Cells 1M using the sequencing plate 3 and 10-hour movies. Genome assembly. The total number of reads was 1,446,753, resulting in 334× coverage of the 30-Mbp genome. The read N50 was 12.44 Kbp with the max read length of 87.877 Kbp. We aligned the reads to a closely related genome, Nannochloropsis oceanica strain LAMB2011 (GenBank accession number GCA_004519485.1) with bowtie2 (default parameters) (3), keeping the reads that only aligned to the reference using samtools (parameter “-F 4”) (4). Canu v.2.1 was used to read-correct decontaminated reads (5). The resulting corrected reads were fed into Flye v.2.8.2-b1689 using parameter “–pacbio-corr” to perform the assembly and polishing. The 29.34-Mbp assembly consisted of 84 scaffolds and 85 contigs with a scaffold N50 of 1.041 Mbp with a max scaffold length of 1.674 Mbp (6) and a GC content of 54.1%. Data availability. This genome project has been deposited at GenBank under the accession sample number SAMN24587378 containing the assembly JANPWC000000000 reference number (GCA_024741955.1). PacBio data have been deposited in the NCBI Sequence Read Archive (SRA) under the BioProject accession PRJNA793856. Strain Nannochloropsis QH25 is privately deposited at the University of Texas Austin’s UTEX Culture Collection under the same name (QH25) and may become available upon request subject to signature of a material transfer agreement. ACKNOWLEDGMENTS This material is based upon work supported by the U.S. Department of Energy’s Office of Energy Efficiency and Renewable Energy (EERE) under the Bioenergy Technologies Office (BETO) Award Number DE- EE0008902. ==== Refs REFERENCES 1 Lee PA, Martinez KJL, Letcher PM, Corcoran AA, Ryan RA. 2018. A novel predatory bacterium infecting the eukaryotic alga Nannochloropsis. Algal Res 32 :314–320. doi:10.1016/j.algal.2018.04.003. 2 Hoffman CS, Winston F. 1987. A ten-minute DNA preparation from yeast efficiently releases autonomous plasmids for transformation of Escherichia coli. Gene 57 :267–272. doi:10.1016/0378-1119(87)90131-4.3319781 3 Langmead B, Salzberg SL. 2012. Fast gapped-read alignment with Bowtie 2. Nat Methods 9 :357–359. doi:10.1038/nmeth.1923.22388286 4 Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R, 1000 Genome Project Data Processing Subgroup. 2009. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25 :2078–2079. doi:10.1093/bioinformatics/btp352.19505943 5 Koren S, Walenz BP, Berlin K, Miller JR, Bergman NH, Phillippy AM. 2017. May. Canu: scalable and accurate long-read assembly via adaptive k-mer weighting and repeat separation. Genome Res 27 :722–736. doi:10.1101/gr.215087.116.28298431 6 Kolmogorov M, Yuan J, Lin Y, Pevzner PA. 2019. Assembly of long, error-prone reads using repeat graphs. Nat Biotechnol 37 :540–546. doi:10.1038/s41587-019-0072-8.30936562

36377884

PMC9753661

NO-CC CODE

2022-12-16 23:26:13

no

Microbiol Resour Announc.; 11(12):e00921-22

utf-8

Microbiol Resour Announc

10.1128/mra.00921-22

oa_other

==== Front Microbiol Resour Announc Microbiol Resour Announc mra Microbiology Resource Announcements 2576-098X American Society for Microbiology 1752 N St., N.W., Washington, DC 36445150 00723-22 10.1128/mra.00723-22 mra.00723-22 Genome Sequences bacteriologyBacteriologyGenome Sequences of Hemolytic and Nonhemolytic Listeria innocua Strains from Human, Food, and Environmental Sources McIntosh Tori a * Kucerova Zuzana a zik0@cdc.gov Katz Lee S. a Lilley Cullen M. a § Rowe Lori A. b ◊ https://orcid.org/0000-0001-9088-0360 Unoarumhi Yvette b Batra Dhwani b Burnett Elton c Smikle Monica d https://orcid.org/0000-0003-4540-1628 Lee Christine a clee13@cdc.gov a Enteric Diseases Laboratory Branch, Centers for Disease Control and Prevention, Atlanta, Georgia, USA b Biotechnology Core Facility Branch, Centers for Disease Control and Prevention, Atlanta, Georgia, USA c Institute of Parasitology, McGill University, Sainte-Anne-de-Bellevue, Canada d Department of Medical Microbiology, University of the West Indies, Mona, Jamaica Editor Klepac-Ceraj Vanja Wellesley College * Present address: Tori McIntosh, Waterborne Diseases Prevention Branch, Centers for Disease Control and Prevention, Atlanta, Georgia, USA. § Present address: Cullen M. Lilley, Loyola University Chicago Stritch School of Medicine, Chicago, Illinois, USA. ◊ Present address: Lori A. Rowe, Viral Characterization, Isolation, Production, and Sequencing Core, Department of Microbiology, Tulane National Primate Research Center, Covington, Louisiana, USA. The authors declare no conflict of interest. 29 11 2022 12 2022 29 11 2022 11 12 e00723-2221 7 2022 11 11 2022 https://doi.org/10.1128/AuthorWarrantyLicense.v1 This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply. ABSTRACT This report describes genome sequences for nine Listeria innocua strains that varied in hemolytic phenotypes on sheep blood agar. All strains were sequenced using Pacific Biosciences (PacBio) single-molecule real-time (SMRT) chemistry; overall, the average read length of these sequences was 2,869,880 bp, with an average GC content of 37%. HHS | Centers for Disease Control and Prevention (CDC) https://doi.org/10.13039/100000030 McIntosh Tori Kucerova Zuzana Katz Lee S. Lilley Cullen M. Rowe Lori A. Unoarumhi Yvette Batra Dhwani Burnett Elton Smikle Monica Lee Christine cover-dateDecember 2022 ==== Body pmcANNOUNCEMENT The Listeria genus represents a diverse group of Gram-positive organisms that are commonly found in environmental reservoirs. Public health focuses on Listeria monocytogenes, which can cause fatal infections in vulnerable human populations (1, 2). Unlike L. monocytogenes, Listeria innocua typically is nonpathogenic but in rare cases has been linked to human infections (3, 4). Genotypic characterization has identified L. innocua as a close relative of L. monocytogenes (2, 4–6). Although L. innocua is typically nonhemolytic, some isolates within the species exhibit β-hemolysis on blood agar (2, 5, 7, 8). These atypical strains also share virulence factors with L. monocytogenes (2, 5, 7). A genomic assessment was performed with nine L. innocua strains that had been isolated from various sources and were archived at the National Listeria Reference Laboratory at the Centers for Disease Control and Prevention (CDC). Here, we present Pacific Biosciences (PacBio) sequences of nine L. innocua isolates that were received between 1987 and 2015. Genomic DNA was extracted from single-colony isolates using the DNeasy blood and tissue kit (Qiagen, Germany) after overnight growth on Trypticase soy agar II with 5% sheep blood (BD, Germany) at 36°C. Samples were assessed for quality (5 to 10 μg DNA in 100 μL buffer) for submission for PacBio sequencing. The Biotechnology Core Facility Branch at the CDC performed sequencing and raw read analysis. Genomic DNA was sheared to 10 kb or 20 kb using needle shearing. Plasmids were identified as contigs of <0.5 MB, consistent with standard threshold cutoff values. Contigs from strains H0996, 2010L-1951, and 2010L-2059 were linear; all other contigs were circularized. Default parameters were used for all software unless otherwise specified. Libraries were generated with the SMRTbell template preparation kit v1.0 (PacBio, Menlo Park, CA). The 2012L-5520 and 2015L-6726 sample libraries were size selected to 8 kb with the BluePippin system (Sage Science, Beverly, MA). All libraries were bound to polymerase using the DNA/polymerase binding kit P6 v2 (PacBio), loaded on one single-molecule real-time (SMRT) Cell (PacBio), and sequenced with C4 v2 chemistry (PacBio) for 270- or 360-min movies, depending on the length, with an RS II instrument (PacBio). Sequence assembly was performed on the HGAP v3 platform (9), with an average yield of 2,869,880 bp. A phylogenetic tree was inferred on the BioLinux command line using Parsnp v1.2 (10); it included 30 published sequences from the NCBI database annotated as Listeria innocua and was visualized in MEGA7 (6) (Fig. 1). The level of base alignment with the reference genome (H0996) was 73.8%, as assessed with Gingr v1.3 (10). FIG 1 Phylogenetic tree based on the full genome sequences for 39 L. innocua isolates. PacBio sequences for the 9 isolates featured in this study are shown in bold, and their sources are listed in parentheses. Strains exhibiting the atypical hemolysis phenotype are indicated by brackets. The phylogeny was constructed by Parsnp using H0996 as the reference genome. Bootstrap branch support values are depicted beside the root and branches. All isolates that exhibited β-hemolysis were in two distinct clades (Fig. 1). Isolate 2012L-5520 (FSL J1-023) is of unknown origin, F6215 from a food source, and F6217, 2010L-2059, and 2010L-1951 from human infections (Table 1). The description of these atypical strains confirms that L. innocua can contribute to human illness. These genome sequences serve as useful resources for additional characterization of the evolutionary mechanisms of Listeria. TABLE 1 Summary of genomic characteristics from PacBio sequencing for the L. innocua isolates Isolate Host (source) Depth of coverage (×) No. of reads Avg read length (bp) No. of contigs (no. of plasmids) Contig N50 (bp) GC content (%) Total genome size (bp) SRA accession no. BioSample accession no. Nucleotide accession no. F4078 (ATCC 33090) Animal (cow brain) 79.51 15,820 16,676 2a 2,794,052 37.6 2,782,538 SRS12596339 SAMN13941777 NZ_JABXLC000000000 F6215 Food (meat) 178.65 35,563 18,169 2a 2,901,888 37.5 2,939,591 SRS12596348 SAMN13941779 NZ_JABXLE000000000 F6217 Human (peritoneum) 353.39 79,416 17,132 1a 2,884,312 37.4 2,855,969 SRS952935 SAMN03761682 NZ_CP095724.1 2015L-6714 Food (meat) 223.30 49,392 16,497 2a 2,930,612 37.38 2,984,334 SRR21942761 SAMN27548027 NZ_CP095723.1 H0996 (ATCC 51742) Food (cabbage) 394.95 84,980 19,213 3 (1) 2,801,715 37.5 2,795,487 SRR21942762 SAMN10869156 NZ_CP095730.1 2010L-1951 Human (abscess [armpit]) 431.68 72,490 21,319 3 (1) 2,818,897 37.4 2,810,256 SRS952976 SAMN03761736 NZ_CP095727.1 2010L-2059 (PNUSAL000003) Human (knee fluid) 322.32 81,774 18,681 15 (2) 2,340,577 37.4 2,816,380 SRR21982767 SAMN10869157 NZ_CP095726.1 2012L-5520 (FSL J1-023) Unknown 337.92 77,923 16,424 1a 2,826,835 37.5 2,826,835 SRR2157026 SAMN13941775 NZ_JABXLA000000000 2015L-6726 Food (meat) 249.48 66,858 12,899 2a 2,857,050 37.4 2,926,353 SRS1160884 SAMN04263682 NZ_CP095728.1 a Circularized contig. Data availability. The GenBank accession numbers for all strains are provided in Table 1. PacBio sequences are available under BioProject accession number PRJNA212117. All isolates are available upon request. ACKNOWLEDGMENT We thank Jasmine Huffman of the Enteric Diseases Laboratory Branch for her assistance with uploading sequences to NCBI. ==== Refs REFERENCES 1 Stea EC, Purdue LM, Jamieson RC, Yost CK, Truelstrup Hansen L. 2015. Comparison of the prevalences and diversities of Listeria species and Listeria monocytogenes in an urban and a rural agricultural watershed. Appl Environ Microbiol 81 :3812–3822. doi:10.1128/AEM.00416-15.25819965 2 Johnson J, Jinneman K, Stelma G, Smith BG, Lye D, Messer J, Ulaszek J, Evsen L, Gendel S, Bennett RW, Swaminathan B, Pruckler J, Steigerwalt A, Kathariou S, Yildirim S, Volokhov D, Rasooly A, Chizhikov V, Wiedmann M, Fortes E, Duvall RE, Hitchins AD. 2004. Natural atypical Listeria innocua strains with Listeria monocytogenes pathogenicity island 1 genes. Appl Environ Microbiol 70 :4256–4266. doi:10.1128/AEM.70.7.4256-4266.2004.15240309 3 Favaro M, Sarmati L, Sancesario G, Fontana C. 2014. First case of Listeria innocua meningitis in a patient on steroids and eternecept. JMM Case Rep doi:10.1099/jmmcr.0.003103. 4 Perrin M, Bemer M, Delamare C. 2003. Fatal case of Listeria innocua bacteremia. J Clin Microbiol 41 :5308–5309. doi:10.1128/JCM.41.11.5308-5309.2003.14605191 5 Moura A, Disson O, Lavina M, Thouvenot P, Huang L, Leclercq A, Fredriksson-Ahomaa M, Eshwar AK, Stephan R, Lecuit M. 2019. Atypical hemolytic Listeria innocua isolates are virulent, albeit less than Listeria monocytogenes. Infect Immun 87 :e00758-18. doi:10.1128/IAI.00758-18.30670551 6 Kumar S, Stecher G, Tamura K. 2016. MEGA7: Molecular Evolutionary Genetics Analysis version 7.0 for bigger datasets. Mol Biol Evol 33 :1870–1874. doi:10.1093/molbev/msw054.27004904 7 Volokhov DV, Duperrier S, Neverov AA, George J, Buchrieser C, Hitchins AD. 2007. The presence of the internalin gene in natural atypically hemolytic Listeria innocua strains suggests descent from L. monocytogenes. Appl Environ Microbiol 73 :1928–1939. doi:10.1128/AEM.01796-06.17220266 8 Clayton EM, Daly KM, Guinane CM, Hill C, Cotter PD, Ross PR. 2014. Atypical Listeria innocua strains possess an intact LIPI-3. BMC Microbiol 14 :58. doi:10.1186/1471-2180-14-58.24606727 9 Chin C-S, Alexander DH, Marks P, Klammer AA, Drake J, Heiner C, Clum A, Copeland A, Huddleston J, Eichler EE, Turner SW, Korlach J. 2013. Nonhybrid, finished microbial genome assemblies from long-read SMRT sequencing data. Nat Methods 10 :563–569. doi:10.1038/nmeth.2474.23644548 10 Treangen TJ, Ondov BD, Koren S, Phillippy AM. 2014. The Harvest suite for rapid core-genome alignment and visualization of thousands of intraspecific microbial genomes. Genome Biol 15 :524. doi:10.1186/s13059-014-0524-x.25410596

36445150

PMC9753683

NO-CC CODE

2022-12-16 23:26:14

no

Microbiol Resour Announc.; 11(12):e00723-22

utf-8

Microbiol Resour Announc

10.1128/mra.00723-22

oa_other

==== Front Microbiol Resour Announc Microbiol Resour Announc mra Microbiology Resource Announcements 2576-098X American Society for Microbiology 1752 N St., N.W., Washington, DC 36326501 00692-22 10.1128/mra.00692-22 mra.00692-22 Genome Sequences virologyVirologyComplete Genome Sequence of O’nyong Nyong Virus Isolated from a Febrile Patient in 2017 in Uganda Ledermann Jeremy P. a Kayiwa John T. b Perinet Lara C. a Apangu Titus c Acayo Sarah c Lutwama Julius J. b Powers Ann M. a https://orcid.org/0000-0002-2315-4577 Mossel Eric C. a emossel@cdc.gov a Arboviral Diseases Branch, Division of Vector-Borne Diseases, Centers for Disease Control and Prevention, Fort Collins, Colorado, USA b Department of Arbovirology, Emerging and Re-emerging Diseases, Uganda Virus Research Institute, Entebbe, Uganda c Arua Field Station, Uganda Virus Research Institute, Arua, Uganda Editor Matthijnssens Jelle KU Leuven The authors declare no conflict of interest. 3 11 2022 12 2022 3 11 2022 11 12 e00692-223 8 2022 18 10 2022 https://doi.org/10.1128/AuthorWarrantyLicense.v1 This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply. ABSTRACT Despite causing numerous large outbreaks in the 20th century, few isolates of o’nyong nyong virus (ONNV) have been fully sequenced. Here, we report the complete genome sequence of an isolate of ONNV obtained from a febrile patient in northwest Uganda in 2017, designated ONNV UVRI0804. Global Health Security Agenda Mossel Eric C. Ledermann Jeremy P. Kayiwa John Perinet Lara C. Apangu Titus Acayo Sarah Lutwama Julius Julian Powers Ann M. cover-dateDecember 2022 ==== Body pmcANNOUNCEMENT O’nyong nyong virus (ONNV; family Togaviridae, genus Alphavirus) is a human pathogen responsible for large, sporadic epidemics in sub-Saharan Africa (1). The virus is transmitted by Anopheles sp. mosquitoes, especially Anopheles gambiae and Anopheles funestus, and causes an acute, debilitating arthralgic syndrome (2). In 2014, the US Centers for Disease Control and Prevention (CDC) and the Uganda Virus Research Institute (UVRI) initiated an outpatient research study to ascertain vector-borne causes of acute febrile illness among residents of Arua District, in the West Nile region in northwestern Uganda. Among other planned testing of paired acute and convalescent serum samples, an aliquot of collected serum samples was sent to the UVRI Arbovirus Laboratory in Entebbe for arbovirus diagnostics. In June 2017, a sample was collected from an adult female presenting to an enrolling health clinic with fever, chills, headache, and joint pain. IgM enzyme-linked immunosorbent assay (ELISAs) to geographically relevant arboviruses were negative; however, the acute specimen produced cytopathic effect following incubation on Vero cells. An aliquot of RNA extracted from cell culture supernatant (QIAamp Viral RNA mini kit; Qiagen, Germantown, MD) was sent to the Centers for Disease Control and Prevention, Arboviral Diseases Branch, Ft. Collins, CO, for identification and genome sequence determination. Group-specific reverse transcription PCR (RT-PCR) for alphaviruses, orthobunyaviruses, and flaviviruses and Sanger sequencing of the resulting amplicon indicated the presence of ONNV genomic RNA (3–6). Whole-genome sequencing was performed on the MiSeq (Illumina, San Diego, CA) next-generation sequencing (NGS) platform. Briefly, the sample RNA was treated with DNase I (Promega, Madison, WI) and used in the TruSeq total RNA V2 library prep kit (Illumina) to generate a cDNA template. To assess the size profile and quantity, the library was analyzed on a bioanalyzer DNA 1000 chip (Agilent Technologies, Santa Clara, CA). The template was then diluted to the appropriate molar concentration and added into the MiSeq reagent kit v2 (300 cycles) sequencing protocol (Illumina). The CLC Genomics Workbench 12.0 (Qiagen) software was used to run a de novo assembly, using default CLC parameters, of the generated 14.1 million trimmed raw data reads into contigs. BLASTn analysis (7) of the 11,807-nucleotide contig to the NCBI nucleotide collection database (https://blast.ncbi.nlm.nih.gov/) determined it was most similar (98.3% identical) to ONNV isolate SG650 (accession number AF079456), a 1996 human-derived isolate. A comparison of the contig sequence to other ONNV sequences in GenBank listed as a complete genome suggested that it contains the 5′ and 3′ termini. The nonstructural polyprotein open reading frame (ORF) contains an opal stop codon at nucleotides 5,765 to 5,767, as is often seen in ONNV and some other alphaviruses (8). Utilization of this codon results in a 6-amino acid truncation of the nsP3 protein, while read-through results in the production of full-length nsP3 and the nsP4 protein (9). Despite the large number of cases in previous ONNV outbreaks, only five full-length genome sequences of ONNV and Igbo Ora virus, now known to be a strain of ONNV (1), have been deposited into GenBank (accession numbers M20303, KX771232, AF079456, MF409176, and AF079457). Here, we add the complete sequence of a 2017 human-derived isolate, ONNV UVRI0804. Data availability. The annotated whole-genome sequence is available in GenBank, accession number ON595759. Raw NGS reads are available from Sequence Read Archive database, accession number SRX17609960. ACKNOWLEDGMENTS We thank Kiersten Kugeler, Ph.D. and Amy Schwartz, U.S. CDC, coordinators of the Arua Acute Febrile Illness (AFI) Project, and the UVRI Arua Field Station staff. The Arua AFI Project was conducted in accordance with CDC institutional review board (IRB) protocol 6480. The laboratory work was funded in part by the Global Health Security Agenda. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the U.S. Centers for Diseases Control and Prevention or the U.S. Government. ==== Refs REFERENCES 1 Lanciotti RS, Ludwig ML, Rwaguma EB, Lutwama JJ, Kram TM, Karabatsos N, Cropp BC, Miller BR. 1998. Emergence of epidemic O'nyong-nyong fever in Uganda after a 35-year absence: genetic characterization of the virus. Virology 252 :258–268. doi:10.1006/viro.1998.9437.9875334 2 Williams MC, Woodall JP, Corbet PS, Gillett JD. 1965. O'nyong nyong fever: an epidemic virus diseases in East Africa. 8. Virus isolations from Anopheles mosquitoes. Trans R Soc Trop Med Hyg 59 :300–306. doi:10.1016/0035-9203(65)90012-X.14298035 3 Kuno G, Chang GJ, Tsuchiya KR, Karabatsos N, Cropp CB. 1998. Phylogeny of the genus Flavivirus. J Virol 72 :73–83. doi:10.1128/JVI.72.1.73-83.1998.9420202 4 Lambert AJ, Lanciotti RS. 2008. Molecular characterization of medically important viruses of the genus Orthobunyavirus. J Gen Virol 89 :2580–2585. doi:10.1099/vir.0.2008/002253-0.18796727 5 Bryant JE, Crabtree MB, Nam VS, Yen NT, Duc HM, Miller BR. 2005. Isolation of arboviruses from mosquitoes collected in northern Vietnam. Am J Trop Med Hyg 73 :470–473. doi:10.4269/ajtmh.2005.73.470.16103625 6 Powers AM, Brault AC, Shirako Y, Strauss EG, Kang W, Strauss JH, Weaver SC. 2001. Evolutionary relationships and systematics of the alphaviruses. J Virol 75 :10118–10131. doi:10.1128/JVI.75.21.10118-10131.2001.11581380 7 Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local alignment search tool. J Mol Biol 215 :403–410. doi:10.1016/S0022-2836(05)80360-2.2231712 8 Strauss EG, Levinson R, Rice CM, Dalrymple J, Strauss JH. 1988. Nonstructural proteins nsP3 and nsP4 of Ross River and O'Nyong-nyong viruses: sequence and comparison with those of other alphaviruses. Virology 164 :265–274. doi:10.1016/0042-6822(88)90644-7.2834873 9 Strauss JH, Strauss EG. 1994. The alphaviruses: gene expression, replication, and evolution. Microbiol Rev 58 :491–562. doi:10.1128/mr.58.3.491-562.1994.7968923

36326501

PMC9753729

NO-CC CODE

2022-12-16 23:26:15

no

Microbiol Resour Announc.; 11(12):e00692-22

utf-8

Microbiol Resour Announc

10.1128/mra.00692-22

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)00772-8 10.1016/S0140-6736(21)00772-8 World Report New EU health programme comes into force Samarasekera Udani 1 4 2021 3-9 April 2021 1 4 2021 397 10281 12521253 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcEU4Health, a €5·1 billion project spurred by the COVID-19 pandemic, aims to strengthen the response to health crises and challenges across Europe. Udani Samarasekera reports. EU4Health, the EU's ambitious health programme for 2021–27, has become a reality, coming into force on March 26. Dedicated funding of €5·1 billion has been agreed for the project, which aims to protect citizens from cross-border health threats, make medicines more accessible and affordable, and build resilient health systems in the EU. The programme is the EU's response to the COVID-19 pandemic, the biggest challenge it has faced since World War 2. As of March 18, the 27 countries of the EU have reported more than 23 million cases of COVID-19 and over 560 000 deaths, according to the European Centre for Disease Prevention and Control (ECDC). EU health systems, including those of well resourced member states, have struggled to deal with the crisis. “Only a common, joint approach will help us tackle such cross-border health threats”, Stefan de Keersmaecker, spokesperson for the European Commission, told The Lancet. The proposal for the EU4Health programme states that it will support new actions to “fill in gaps that this pandemic has revealed in terms of the development and manufacturing of medicinal products, the adequate supply of equipment in hospitals and sufficient medical human resources, the uptake of digital tools and services that enable continuity of care, and the need to maintain access to essential goods and services in times of crisis”. © 2021 Philippe Turpin/Getty Images 2021 EU4Health funds will go towards creating a strategic reserve of medical supplies and of health-care staff and experts that can be mobilised to prevent or respond to health crises throughout the EU. “Building reserves should allow us to better tackle future pandemics”, says de Keersmaecker. The money will also support the training of health-care professionals for deployment across the EU and increased surveillance of health threats. Beyond preparedness and response, another key area of focus will be strengthening the health systems of the different member states, including investment in health promotion and disease prevention and improving access to health care. “We need resilient health systems that are well equipped to tackle new health challenges”, de Keersmaecker notes. Health experts have welcomed EU4Health. Walter Ricciardi, professor of hygiene and public health at Catholic University of the Sacred Heart (Rome, Italy) and president of the World Federation of Public Health Associations, said: “This is a very important programme. It's the first time ever that the European Commission has tried to work in a concrete manner with member states, providing collaboration and support for health challenges. It is something we as a public health organisation have been advocating for years…Health nowadays is global, and decisions have to be taken at international level, but at European level you have to be at least coordinating your policies if not harmonising them.” François Alla, professor of public health at the University of Bordeaux, France, agrees: “It is essential to build a collective strategy to respond to global crises and challenges. Health must be a priority in this European strategy. This was not the case until now. I therefore welcome this initiative.” The biggest advantage of the policy, Ricciardi says, is its potential to reduce health inequalities. “Europe is the continent where the highest health inequalities happen. Reducing inequalities and granting equitable [health-care] access to citizens is the most important result that can come from the European approach.” EU4Health is the fourth and the largest of the EU health programmes in monetary terms since their inception in 2003. Its funds are more than ten times that of its predecessor. “I think the fact that it's a lot better resourced than the last health programme is very encouraging”, says Sascha Marschang, acting secretary general of the European Public Health Alliance, which represents more than 80 public health non-governmental organisation (NGO), patient group, and health professional members. Commenting on the proposal, he said: “In terms of the language and declared comprehensiveness of the communication, I think that that's pretty good. The thing of course is what does it mean in practice. For example, health inequalities—it's meant to be a general objective of the EU4Health programme, and it's also mentioned as a specific objective. But then when you look at the list of actions, when you are looking for things related to health inequalities it's a little bit less clear; it doesn't really translate into very tangible actions.” de Keersmaecker notes that EU4Health will also help fund the policy priorities under the Commission's vision for a European Health Union, announced by Ursula von der Leyen, president of the European Commission, in her State of the Union address in September, 2020. Cancer is one such priority. €1·25 billion of EU4Health funding will go to Europe's €4 billion Beating Cancer Plan, which aims to tackle the entire disease pathway, with ten flagship programmes spanning from prevention to the quality of life of patients and survivors. Marschang says the plan is “very comprehensive”, but “what we want to avoid is falling into the trap of a siloed approach. That we only look at cancer, that we forgot about all the other non-communicable diseases. That we only look at pandemic preparedness, but we forget about the syndemic, the conditions that are creating the pandemic in the first place.” As part of its European Health Union concept, the Commission is also establishing a new agency, the Health Emergency Preparedness and Response Authority (HERA). “What we do not have is what our American friends have, the so-called BARDA, this authority that allows the Americans to intervene quickly when a pandemic hits, by building reserves of medicines, reserves of medical devices, by immediately contacting the pharmaceutical industry, by testing production capacity, by immediately placing orders. This is why we suggested HERA”, de Keersmaecker explains. Ricciardi welcomes this new agency and the fact that “new power is going to be attributed to ECDC and EMA [European Medicines Agency], which are important bodies” in the implementation of the EU's health priorities. ECDC's mandate will be revised so that it can support the Commission and member states in several areas, including integrated systems to allow real-time epidemiological surveillance and preparedness and response planning, reporting, and auditing. EMA's mandate will be reinforced so that it can facilitate a coordinated EU-level response to health crises by monitoring and mitigating the risk of shortages of critical medicines and medical devices, providing scientific advice on potential medicines for diseases causing crises, coordinating studies to monitor the effectiveness and safety of vaccines, and coordinating clinical trials. Digital transformation of health systems is another funding focus for EU4Health. Marschang notes that the pandemic has provided an opportunity to scale up digital health because many patients and health professionals in Europe needed to use electronic health and mobile health tools for the first time. However, he thinks that health literacy initiatives need to be clearly defined in EU4Health to avoid widening inequities. “Many of the groups who may not enjoy the best quality of health care tend to be the people that are less health literate. They then have issues using the technology and having access to the technology and most of all making sense of the information they find online.” Training for health-care professionals “so they can also work more effectively with patient communities around this issue of the digital transformation of health care” will be key and needs more clarity in EU4Health, Marschang says. Member states, health organisations, and NGOs will be able to apply for EU4Health funding this year following the adoption of the 2021 work programme, which will prioritise crisis preparedness, disease prevention, health systems, digitalisation, and cancer. However, Anniek de Ruijter, associate professor of health law and policy at the University of Amsterdam in the Netherlands, is worried about the bureaucracy of the funding distribution process. “This is an opportunity to really display the EU-added benefit for alleviating some of the inequalities in access to high-quality care across EU member states. Now there will be this whole bureaucratic circus of calls and applications for calls. Would that effort not be better spent if at least a big chunk of the funding is disseminated in direct contact with national public health authorities?”, she asks. de Ruijter is also concerned about the governance of the programme. Although the EU4Health proposal mentions monitoring, reporting, evaluation, and auditing, she questions “how much capacity does the EU actually have here to do the monitoring and auditing in the member states itself?...These things are not at all in a clear form” in the proposal. “I'm just worried that a lot of money will go out, but there's not enough capacity to do the governance part of it.” Additionally, Marschang raised concerns about long-term follow-up. “The problem is, if I think about some of the calls that were released under the previous health programme, the projects themselves might have been good, and the output excellent as well, but then are these outcomes being integrated into national health systems or do they then become forgotten about over time?”, he comments. “When you have European projects, there's a lack of follow-up or lack of money for the national health system to integrate these best practices, these ideas, practically into their health system because their health system might be a slightly different model to the country that was leading on the project.” Ricciardi thinks that delivery and implementation are the crucial focuses now. “At the moment, with the money that has been allocated, it is more than enough to start working and provide concrete results for citizens”, he says. However, he acknowledges, “This is a very complex programme. The challenge now is governance and implementation.”

33812481

PMC9753741

NO-CC CODE

2022-12-16 23:26:16

no

Lancet. 2021 Apr 1 3-9 April; 397(10281):1252-1253

utf-8

Lancet

10.1016/S0140-6736(21)00772-8

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)00770-4 10.1016/S0140-6736(21)00770-4 World Report New Libyan Government faces health challenges Devi Sharmila 1 4 2021 3-9 April 2021 1 4 2021 397 10281 12501250 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcInterim Prime Minister Abdul Hamid Dbeibah must deal with COVID-19, migration, and the legacy of 10 years of conflict. Sharmila Devi reports. Amid hopes for an end to conflict, Libya's first unity government was sworn in on March 15, 2021. However, it faces many challenges, including 1·3 million people in need of humanitarian assistance, a lack of many essential medicines, and many closed health-care centres because of security threats and lack of funding. The new government, led by Abdul Hamid Dbeibah, interim prime minister, emerged from UN talks to replace two rival administrations and will prepare for elections in December, 2021. “11 years of armed conflict and political instability have severely affected Libyan nationals, particularly those who have been displaced, as well as refugees, asylum seekers, and migrants, and have pushed Libya's health-care system to the point of collapse”, Adolphe Fotso, medical coordinator for Médecins Sans Frontières (MSF), told The Lancet. Conflict has engulfed the country since a NATO-backed uprising overthrew Muammar Gaddafi in 2011. Local militias still controlled some areas, while about 20 000 foreign forces and mercenaries were still in the country. A ceasefire has been in place since October, 2020, but has often been flouted. “While the ceasefire agreement continues to hold, there are reports of ongoing fortifications and the setting up of defensive positions… Elements of Islamic State of Iraq and the Levant and other international terrorist groups continue to operate in the country”, Ján Kubiš, UN special envoy for Libya, told the Security Council on March 24. “While the ceasefire has resulted in a dramatic reduction of civilian casualties, UNSMIL [UN Support Mission in Libya] continues to document killings, enforced disappearances, sexual violence, including rape, arbitrary arrests and detentions, attacks against activists and human rights defenders, and hate crimes”, he said. More than half of Libya's health-care facilities that were functioning in 2019 were forced to close in 2020, particularly in rural areas, said WHO in its annual Libya report released on March 15, 2021. Those that remained open suffered frequent electricity cuts. Repeated shortages of critical vaccines disrupted immunisation and put children at risk of diseases such as measles and polio. More than a quarter of a million children missed their doses of essential vaccines. Additionally, more than two-thirds of primary health-care centres faced shortages—including having no antibiotics, analgesics, insulin, or blood pressure medication—and most health-care staff had to wait months to receive their salaries, said the report. COVID-19 was an additional burden and as of March 1, 2021, Libya had reported over 1800 deaths, the second-highest number in north Africa after Tunisia, said the UN. The political and security environment hampered efforts to contain the pandemic's spread, said WHO. In addition to COVID-19 and conflict, Libya faces a third major challenge with the presence of around 570 000 migrants and refugees, according to UN estimates. About 4000 migrants and refugees were intercepted at sea and returned to Libya in the first 2 months of 2021, compared with around 2200 during 2020 for the same reporting period, said the UN Office for the Coordination of Humanitarian Affairs. 3858 migrants and refugees were detained in official detention centres “under extremely poor conditions without due process and restrictions on humanitarian access”, Kubiš told the UN Security Council. Detainees in these centres are not guaranteed access to medical care, which is provided by a handful of humanitarian organisations that have a limited presence in the country. Bianca Benvenuti, MSF advocacy manager for Libya, described a detention centre in a personal testimony shared with The Lancet. “In the main hangar, where people are kept, the only light that enters is from high-up windows, close to the ceiling”, she said. “I entered the main cell with [the] MSF medical team to distribute Plumpy'Nut—a high-calorie peanut paste we usually give to malnourished children. We distribute it to adults in the detention centres because we know they don't receive enough food.” Human rights campaigners have strongly criticised European policies that enable people rescued at sea to be forcibly returned to Libya, where it is not safe to disembark migrants and refugees, said a Council of Europe report, entitled A distress call for human rights: the widening gap in migrant protection in the Mediterranean, released on March 19, 2021. “Member states' support to the Libyan Coast Guard, which contributes to the return of refugees and migrants to Libya, where they undergo serious human rights violations, has been reconsidered in some cases… although this resulted mainly from legal action brought by [non-governmental organisations] and other stakeholders rather than proactive steps by member states”, said the report.

33812479

PMC9753766

NO-CC CODE

2022-12-16 23:26:18

no

Lancet. 2021 Apr 1 3-9 April; 397(10281):1250

utf-8

Lancet

10.1016/S0140-6736(21)00770-4

oa_other

==== Front Lancet Lancet Lancet (London, England) 0140-6736 1474-547X Elsevier Ltd. S0140-6736(21)00629-2 10.1016/S0140-6736(21)00629-2 Comment Migration and health in Latin America during the COVID-19 pandemic and beyond Bojorquez Ietza a Cabieses Báltica b Arósquipa Carlos c Arroyo Juan d Novella Andrés Cubillos e Knipper Michael f Orcutt Miriam g Sedas Ana Cristina h Rojas Karol i a Department of Population Studies, El Colegio de la Frontera Norte, Tijuana, CP 22560, Mexico b Instituto de Ciencias e Innovación en Medicina, Universidad del Desarrollo, Las Condes, Santiago, Chile c Pan American Health Organization, La Molina, Lima, Perú d Pontificia Universidad Católica del Perú, Lima, Perú e Instituto de Salud Pública, Pontificia Universidad Javeriana, Bogotá, Colombia f Institute for the History of Medicine, University Justus Liebig Giessen, Giessen, Germany g Institute for Global Health, University College London, London, UK h Department of Global Health and Social Medicine, Harvard Medical School, Boston, MA, USA i Universidad de Costa Rica, San José, Mercedes, Costa Rica 1 4 2021 3-9 April 2021 1 4 2021 397 10281 12431245 © 2021 Elsevier Ltd. All rights reserved. 2021 Elsevier Ltd Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. ==== Body pmcCOVID-19 has created a syndemic scenario that is deepening pre-existing structural inequalities for migrants in Latin American countries (LACs).1, 2 LACs have been severely affected by COVID-19, and migrants are among the populations most impacted by the heightened humanitarian crisis across the region. Socioeconomic inequalities between migrants and local people have widened, and there are fears that progress towards the Sustainable Development Goals (SDGs) will be reversed. During the COVID-19 pandemic, food insecurity, unemployment, and reduced socioeconomic agency have led to considerable insecurity and anxiety for migrants living in LACs.3, 4 The region faces one of the largest mass migrations worldwide; more than 5·5 million refugees and migrants have left Venezuela, 4·6 million of whom now live in the Latin American region.5, 6 In central America, a combination of criminal and political violence, poverty, and the increasing impact of droughts on farmers forced thousands to move northwards, where many people have become trapped due to restrictive migration policies.7 For decades, migration fuelled the economic and cultural development in the region. Today, migration has increasingly become an issue of risk and precarity. Together with the social and structural disparities that make Latin America a region with one of the highest levels of inequalities in the world,8, 9 the worsening environment for migrants during the pandemic means their basic needs are unmet10 and their social, economic, and cultural capabilities are not realised. In LACs, migrants face barriers in accessing regular health services due to inadequate information, the absence of culturally appropriate care, or insufficient legal provisions. Additionally, there have been challenges in the prevention of COVID-19 among migrant populations as a result of poor public health communication, reduced access to public health prevention measures, and living or working in conditions where it is difficult to isolate.11 The deteriorating determinants of migrants' health and wellbeing, revealed and increased by COVID-19, strain health systems in LACs, and governmental, regional, and international organisations' efforts to advance inclusive health policies.12 COVID-19 has reduced human mobility within and across borders. Border closures and changes to migration policy across Latin America interrupted migrants' movement, leaving thousands stranded across the region. Others were forced to return to the same danger, social exclusion, inadequate health care, and poverty they had fled. Migrants who were trapped in transit, or who were living with undocumented or irregular status, became even more vulnerable to family separation, trafficking, or exploitation.13 In the pandemic migrants have had reduced access to asylum processes and resettlement schemes have been suspended,14 with many asylum seekers and refugees subsequently facing a protection crisis. This diverse region has proven its capacity to cooperate and exchange knowledge and practice in public health,15, 16 and has a strong tradition of social medicine and community health, which is evident in the multiple community-based initiatives that have emerged to combat COVID-19.17, 18 However, the detrimental impacts of COVID-19 on migrants and failure of inclusive health policies and systems have revealed the vulnerability of migrants to the consequences of politics designed without considering migration and human mobility in government agendas. These failures underline the importance of compliance with international standards, such as the Global Compact for Safe, Orderly and Regular Migration19 and the WHO global action plan to promote the health of refugees and migrants.20 Explicit commitment from governments to include migrants in all phases of the evolving public health response to COVID-19 is needed. This response must include ensuring access to vaccination and social protection programmes without discrimination and addressing the broader socioeconomic inequalities that affect migrants' health.21 The forthcoming Ibero-American Summit on April 21, 2021, presents an opportunity for heads of states and governments across the region to make explicit commitment to ensure all migrants, refugees, and asylum seekers, irrespective of age, gender, or migration status, have universal and equitable access to health care and pandemic prevention and response. Sustainable and inclusive approaches to migrant health should go beyond humanitarian response and meet the needs of local and mobile populations alike, in line with the SDG agenda. Yet the regions, countries, and municipalities most affected by COVID-19 also need solidarity and global support in responding to the current global public health crisis and future global health challenges. © 2021 Mads Nissen/Politiken/Panos pictures 2021 This online publication has been corrected. The corrected version first appeared at thelancet.com on April 15, 2021 IB, BC, CA, JA, AC-N, and KR are Lancet Migration Latin American Regional Hub board members; MO is Executive Director of Lancet Migration; BC and MK are co-Leads of Lancet Migration Regional Hub, Latin America; ACS is Lancet Migration Regional Research Fellow, Latin America. We declare no other competing interests. ==== Refs References 1 Horton R Offline: COVID-19 is not a pandemic Lancet 396 2020 874 32979964 2 The Lancet COVID-19 in Latin America: a humanitarian crisis Lancet 396 2020 1463 3 Cabieses B Obach A Urrutia C Encuesta sobre COVID-19 a Poblaciones Migrantes Internacionales en Chile. Informe de Resultados Completo. Universidad del Desarrollo, Servicio Jesuita a Migrantes, Comisión de Migración y Salud, Colegio Médico de Chile, Núcleo Milenio de Resistencia Antimicrobiana MICROB-R, Red de Investigación Interdisciplinaria en enfermedades infecciosas https://www.migracionenchile.cl/wp-content/uploads/2020/07/Reporte-FINAL-ENCUESTA-MIGRANTES-Y-COVID19-18julio2020.pdf July, 2020 4 Espinel Z Chaskel R Berg RC Venezuelan migrants in Colombia: COVID-19 and mental health Lancet Psychiatry 7 2020 653 655 32711697 5 R4V Inter-Agency Coordination Platform for Refugees and Migrants from Venezuela RMRP 2021 regional refugee and migrant response plan for refugees and migrants from Venezuela (January–December 2021) https://r4v.info/es/documents/details/82927 Dec 10, 2020 6 R4V Inter-Agency Coordination Platform for Refugees and Migrants from Venezuela RMRP 2020 dashboard https://r4v.info/es/situations/platform 7 Kauffer E The Mexico–Guatemala border during covid-19: from open border to new assemblage? Borders Globaliz Rev 2 2020 66 8 Dávila-Cervantes CA Agudelo-Botero M Health inequalities in Latin America: persistent gaps in life expectancy Lancet Planet Health 3 2019 e492 e493 31836431 9 Economic Commission for Latin America and the Caribbean Social panorama of Latin America, 2020. LC/PUB.2021/2-P 2021 United Nations Santiago 10 WHO ApartTogether survey: preliminary overview of refugees and migrants self-reported impact of COVID-19 2020 World Health Organization Geneva https://www.who.int/publications/i/item/9789240017924 11 Lancet Migration Situational and policy briefs: COVID-19 and migration https://www.migrationandhealth.org/migration-covid19-briefs 2020 12 Riggirozzi P Grugel J Cintra N Situational brief: perspective on migrant's right to health in Latin America during COVID-19. Protecting migrants or reversing migration? COVID-19 and the risks of a protracted crisis in Latin America https://1bec58c3-8dcb-46b0-bb2a-fd4addf0b29a.filesusr.com/ugd/188e74_543cbb0400824084abcea99479dfa124.pdf?index=true 2020 13 International Organization for MigrationUN Migration COVID-19 analytical snapshot #26: stranded migrants https://www.iom.int/sites/default/files/documents/covid-19_analytical_snapshot_26_-_stranded_migrants_0.pdf 2020 14 International Organization for Migration IOM, UNHCR announce temporary suspension of resettlement travel for refugees https://www.iom.int/news/iom-unhcr-announce-temporary-suspension-resettlement-travel-refugees 2020 15 Pan American Health Organization Annual report of the Director of the Pan American Sanitary Bureau 2020. Saving lives and improving health and well-being 2020 Pan American Health Organization Washington, DC https://iris.paho.org/handle/10665.2/52852 16 Vasquez EE Perez-Brumer A Parker RG Social inequities and contemporary struggles for collective health in Latin America 2020 Routledge London and New York 17 Melo Maciel FB Pereira Campos dos Santos HL Araújo da Silva Carneiro R Community health workers: reflections on the health work process in Covid-19 pandemic times Cien Saude Colet 25 suppl 2 2020 4185 4195 33027355 18 Terán-Puente C Aráuz V Terán G Manual operativo del sistema comunitario de vigilancia epidemiológica Covid-19 2021 Sociedad Ecuatoriana de Salud Pública, GAD Sucre, Universidad Estatal de Milagro UNEMI Quito, Ecuador 19 WHO Proposed Health Component in the Global Compact for Safe, Orderly and Regular Migration 2020 World Health Organization Geneva https://www.who.int/migrants/about/gcm-health-component/en/ 20 WHO Promoting the health of refugees and migrants: draft global action plan, 2019–2023. Report by the Director-General 2019 World Health Organization Geneva https://www.who.int/publications/i/item/promoting-the-health-of-refugees-and-migrants-draft-global-action-plan-2019-2023 May 23, 2019 21 Orcutt M Patel P Burns R Global call to action for inclusion of migrants and refugees in the COVID-19 response Lancet 395 2020 1482 1483 32334651

33812478

PMC9753767

NO-CC CODE

2022-12-16 23:26:18

no

Lancet. 2021 Apr 1 3-9 April; 397(10281):1243-1245

utf-8

Lancet

10.1016/S0140-6736(21)00629-2

oa_other