Resources


Database Restricted Access

MIMIC-IV-Ext-Apixaban-Trial-Criteria-Questions

Elizabeth Woo, Michael Craig Burkhart, Emily Alsentzer, et al.

We created 23 questions resembling eligibility criteria from the apixaban clinical trial and evaluated them on a random sample of 100 patient notes from MIMIC-IV. We release the 2300 total question-answer pairs as a dataset here.

clinical q and a evaluation set clinical trial eligibility

Published: April 30, 2025. Version: 1.0.0

Challenge Credentialed Access

ArchEHR-QA: A Dataset for Addressing Patient's Information Needs related to Clinical Course of Hospitalization

Sarvesh Soni, Dina Demner-Fushman

A dataset for grounded question answering (QA) from electronic health records (EHRs).

question answering electronic health record patient portals clinicians

Published: Jan. 1, 2026. Version: 1.3

Database Credentialed Access

RaDialog Instruct Dataset

Chantal Pellegrini, Ege Özsoy, Benjamin Busam, et al.

Image-based instruct data for Chest X-Ray understanding and analysis.

medical image understaning radiology chatbot radiology report generation radiology assistant large vision-language models

Published: July 12, 2024. Version: 1.1.0

Database Credentialed Access

CXReasonBench: A Benchmark for Evaluating Structured Diagnostic Reasoning in Chest X-rays

Hyungyung Lee, Geon Choi, Jung Oh Lee, et al.

CheXStruct is an automated pipeline that derives structured diagnostic reasoning steps from chest X-rays. CXReasonBench builds on this to evaluate whether models perform clinically grounded, multi-step reasoning beyond final diagnoses.

evaluation chest x-ray benchmark structured chest x-ray qa intermediate reasoning steps structured reasoning grounded reasoning diagnostic reasoning structured diagnostic pipeline

Published: Oct. 23, 2025. Version: 1.0.1

Database Restricted Access

MIMIC-IV-Ext-DiReCT

Bowen Wang, Jiuyang Chang, Yiming Qian

A diagnostic reasoning dataset designed to evaluate the performance of large language models in aligning with human doctors when making diagnoses from clinical notes.

Published: Jan. 21, 2025. Version: 1.0.0

Database Credentialed Access

MIMIC-IV-Ext-Instr: A Dataset of 450K+ EHR-Grounded Instruction-Following Examples

Zhenbang Wu, Anant Dadu, Mike Nalls, et al.

This dataset contains 450K open-ended instruction-following examples generated using GPT-3.5 based on the MIMIC-IV EHR database.

large language models medical question answering instruction tuning

Published: Sept. 9, 2025. Version: 1.0.0