Datasets:
Tasks:
Text Classification
Formats:
parquet
Languages:
English
Size:
10K - 100K
ArXiv:
Tags:
medical
License:
metadata
dataset_info:
- config_name: conversational_medical
features:
- name: id
dtype: int64
- name: prompt
list:
- name: role
dtype: string
- name: content
dtype: string
- name: completion
list:
- name: role
dtype: string
- name: content
dtype: string
- name: Label
dtype: string
splits:
- name: train
num_bytes: 12239188
num_examples: 4904
- name: dev
num_bytes: 1277389
num_examples: 525
- name: test
num_bytes: 3980928
num_examples: 1578
download_size: 6339380
dataset_size: 17497505
- config_name: conversational_patient
features:
- name: id
dtype: int64
- name: prompt
list:
- name: role
dtype: string
- name: content
dtype: string
- name: completion
list:
- name: role
dtype: string
- name: content
dtype: string
- name: Label
dtype: string
splits:
- name: train
num_bytes: 15221480
num_examples: 4904
- name: dev
num_bytes: 1597949
num_examples: 525
- name: test
num_bytes: 4920175
num_examples: 1578
download_size: 8521486
dataset_size: 21739604
- config_name: processed-medical
features:
- name: id
dtype: int64
- name: topic_id
dtype: int64
- name: Description_Medical-Language
dtype: string
- name: CTR_Context
dtype: string
- name: CTR_Title
dtype: string
- name: CTR_id
dtype: string
- name: Label
dtype: string
splits:
- name: train
num_bytes: 12842220
num_examples: 4904
- name: dev
num_bytes: 1334940
num_examples: 525
- name: test
num_bytes: 4203085
num_examples: 1578
download_size: 6690961
dataset_size: 18380245
- config_name: processed-patient
features:
- name: id
dtype: int64
- name: topic_id
dtype: int64
- name: Description_Patient-Language
dtype: string
- name: CTR_Context
dtype: string
- name: CTR_Title
dtype: string
- name: CTR_id
dtype: string
- name: Label
dtype: string
splits:
- name: train
num_bytes: 12824417
num_examples: 4904
- name: dev
num_bytes: 1334300
num_examples: 525
- name: test
num_bytes: 4179550
num_examples: 1578
download_size: 6696616
dataset_size: 18338267
- config_name: source
features:
- name: id
dtype: int64
- name: topic_id
dtype: int64
- name: statement_medical
dtype: string
- name: statement_pol
dtype: string
- name: premise
dtype: string
- name: NCT_title
dtype: string
- name: NCT_id
dtype: string
- name: label
dtype: string
splits:
- name: train
num_bytes: 15799992
num_examples: 4904
- name: dev
num_bytes: 1652875
num_examples: 525
- name: test
num_bytes: 5134442
num_examples: 1578
download_size: 6841482
dataset_size: 22587309
configs:
- config_name: conversational_medical
data_files:
- split: train
path: conversational_medical/train-*
- split: dev
path: conversational_medical/dev-*
- split: test
path: conversational_medical/test-*
- config_name: conversational_patient
data_files:
- split: train
path: conversational_patient/train-*
- split: dev
path: conversational_patient/dev-*
- split: test
path: conversational_patient/test-*
- config_name: processed-medical
data_files:
- split: train
path: processed-medical/train-*
- split: dev
path: processed-medical/dev-*
- split: test
path: processed-medical/test-*
- config_name: processed-patient
data_files:
- split: train
path: processed-patient/train-*
- split: dev
path: processed-patient/dev-*
- split: test
path: processed-patient/test-*
- config_name: source
data_files:
- split: train
path: source/train-*
- split: dev
path: source/dev-*
- split: test
path: source/test-*
license: cc-by-sa-4.0
task_categories:
- text-classification
language:
- en
tags:
- medical
pretty_name: NLI4PR
size_categories:
- 10K<n<100K
Natural Language Inference for Patient Recruitment (NLI4PR)
Dataset Description
| Links | |
|---|---|
| Homepage: | Github.io |
| Repository: | Github |
| Paper: | arXiv |
| Contact (Original Authors): | Mathilde Aguiar (mathilde.aguiar@lisn.fr) |
| Contact (Curator): | Artur Guimarães (artur.guimas@gmail.com) |
Dataset Summary
MedQA is a large-scale multiple-choice question-answering dataset designed to mimic the style of professional medical board exams, particularly the USMLE (United States Medical Licensing Examination). Introduced by Jin et al. in 2020 under the title “What Disease Does This Patient Have? A Large‑scale Open‑Domain Question Answering Dataset from Medical Exams”, the dataset supports open-domain QA via retrieval from medical textbooks
Data Instances
Source Format
{
"id": "5088",
"topic_id": "39",
"statement_medical": "A 55-year-old white woman comes for a routine checkup. She has no significant medical history and does not use tobacco, alcohol, or illicit drugs. The patient's only medication is an over-the-counter multivitamin. Family history is notable for a hip fracture in her mother. Blood pressure is 130\/80 mm Hg and pulse is 112\/min. She has occasional back pain and lives a sedentary lifestyle with the BMI of 24 Kg\/m2. Plain X-ray of the spine shows mild compression fracture at the level of T10. X-ray absorptiometry studies demonstrate abnormally low bone density in the lumbar vertebrae and T-score values below -2.5, which confirms the diagnosis of osteoporosis.",
"statement_pol": "I'm a 55-year-old white woman and I recently visited my family doctor. I don't smoke anything or drink. I don't have any remarkable medical history. I only use over-the-counter multivitamins to keep myself fresh and energized. My mom had a hip fracture. The doctor took my blood pressure and it was 130\/80 and my pulse was 112\/min. I have annoying back pain from time to time and to be honest I don't exercise much or move much. My BMI is 24. I did a spine X-ray a while ago and my doctor showed me that I have a fracture on one of my vertebrae. I also have a low bone density in my lumbar vertebrae and T-score values below -2.5. The doctor diagnosed me with osteoporosis.",
"premise": "Inclusion Criteria:\n\n - Postmenopausal women and men referred for bone density examination.\n\n Exclusion Criteria:\n\n - Patients unable to sign consent for participation.\nNo condition on gender to be admitted to the trial.\nAccepts Healthy Volunteers\nSubject must be at least 20 Years old.\nSubject must be at most 90 Years",
"NCT_title": "Bindex Ultrasonometer for Osteoporosis Diagnostics",
"NCT_id": "NCT01935232",
"label": "Contradiction"
}
Data Fields
Source Format
TO:DO
Data Splits
TO:DO
Additional Information
Dataset Curators
Original Paper
- Mathilde Aguiar (mathilde.aguiar@lisn.fr) - Université Paris-Saclay, CNRS, Laboratoire Interdisciplinaire des Sciences du Numérique
- Pierre Zweigenbaum - Université Paris-Saclay, CNRS, Laboratoire Interdisciplinaire des Sciences du Numérique
- Nona Naderi - Université Paris-Saclay, CNRS, Laboratoire Interdisciplinaire des Sciences du Numérique
Huggingface Curator
- Artur Guimarães (artur.guimas@gmail.com) - INESC-ID / University of Lisbon - Instituto Superior Técnico
Licensing Information
Citation Information
@misc{aguiar2025ieligiblenaturallanguage,
title={Am I eligible? Natural Language Inference for Clinical Trial Patient Recruitment: the Patient's Point of View},
author={Mathilde Aguiar and Pierre Zweigenbaum and Nona Naderi},
year={2025},
eprint={2503.15718},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.15718},
}
10.18653/v1/2025.cl4health-1.21
Contributions
Thanks to araag2 for adding this dataset.