metadata

license: cc-by-4.0
language:
  - es
tags:
  - casimedicos
  - explainability
  - medical exams
  - medical question answering
  - extractive question answering
  - squad
  - multilinguality
  - LLMs
  - LLM
pretty_name: casimedicos-squad
configs:
  - config_name: es
    data_files:
      - split: train
        path:
          - data/es/es_train_casimedicos_squad.json
      - split: validation
        path:
          - data/es/es_dev_casimedicos_squad.json
      - split: test
        path:
          - data/es/es_test_casimedicos_squad.json
task_categories:
  - question-answering
size_categories:
  - 1K<n<10K

Antidote CasiMedicos in SQuAD Format for Explanatory Argument Extraction

We present a new multilingual parallel medical dataset of commented medical exams which includes not only explanatory arguments for the correct answer but also arguments to explain why the remaining possible answers are incorrect.

Furthermore, this dataset allows us to setup a novel extractive task which consists of identifying the explanation of the correct answer written by medical doctors. In order to do so we leverage the SQuAD extractive QA paradigm to automatically evaluate performance of language models to identify the explanation of the correct answer in medical exams without relying on costly manual evaluation by medical experts.

The data source consists of Resident Medical Intern or Médico Interno Residente (MIR) exams, originally created by CasiMedicos, a Spanish community of medical professionals who collaboratively, voluntarily, and free of charge, publishes written explanations about the possible answers included in the MIR exams. The aim is to generate a resource that helps future medical doctors to study towards the MIR examinations. The commented MIR exams, including the explanations, are published in the CasiMedicos Project MIR 2.0 website.

We have extracted, clean, structure and annotated the available data so that each document in casimedicos-squad includes the clinical case, the correct answer, the multiple-choice questions and the commented exam written by native Spanish medical doctors. The comments have been annotated with the span in the text that corresponds to the explanation of the correct answer (see example below).

Furthermore, the original Spanish data has been translated to create a parallel multilingual dataset in 4 languages: English, French, Italian and Spanish.

casimedicos-squad splits
train	404
validation	56
test	119

📖 Paper:Explanatory Argument Extraction of Correct Answers in Resident Medical Exams
💻 Github Repo (Data and Code): https://github.com/ixa-ehu/antidote-casimedicos
🌐 Project Website: https://univ-cotedazur.eu/antidote
Funding: CHIST-ERA XAI 2019 call. Antidote (PCI2020-120717-2) funded by MCIN/AEI /10.13039/501100011033 and by European Union NextGenerationEU/PRTR

Example

The example above shows a document in CasiMedicos containing the textual content, including Clinical Case (C), Question (Q), Possible Answers (P), and Explanation (E). Furthermore, for casimedicos-squad we annotated the span in the explanation (E) that corresponds to the correct answer (A).

Data Explanation

The following attributes composed casimedicos-raw:

id: unique doc identifier.
year: year in which the exam was published by the Spanish Ministry of Health.
question_id_specific: id given to the original exam published by the Spanish Ministry of Health.
full_question: Clinical Case (C) and Question (Q) as illustrated in the example document above.
full answer: Full commented explanation (E) as illustrated in the example document above.
type: medical speciality.
options: Possible Answers (P) as illustrated in the example document above.
correct option: solution to the exam question.

Additionally, the following jsonl attribute was added to create casimedicos-exp:

explanations: for each possible answer above, manual annotation states whether:
1. the explanation for each possible answer exists in the full comment (E) and
2. if present, then we provide character and token offsets plus the text corresponding to the explanation for each possible answer.

The process of manually annotating the corpus consisted of specifying where the explanations of the correct and incorrect answers begin and end. In order to obtain grammatically complete correct answer explanations, annotating full sentences or subordinate clauses was preferred over shorter spans.

Citation

If you use this data please cite the following paper:

@misc{goenaga2023explanatory,
      title={Explanatory Argument Extraction of Correct Answers in Resident Medical Exams}, 
      author={Iakes Goenaga and Aitziber Atutxa and Koldo Gojenola and Maite Oronoz and Rodrigo Agerri},
      year={2023},
      eprint={2312.00567},
      archivePrefix={arXiv}
}

Contact: Iakes Goenaga and Rodrigo Agerri HiTZ Center - Ixa, University of the Basque Country UPV/EHU