Resources

This version adds DICOM-format mammography images to TN-Mammo alongside the previously available JPEG images. Dataset identifiers have been replaced with UUIDs, while the labels, annotations, and underlying data remain unchanged.

A dataset of ICU admission notes with single-variable counterfactuals (demographics and vital signs) enabling controlled evaluation of medical LLM sensitivity, robustness, and clinical reasoning to meaningful patient data changes.

This dataset contains annotations of 120 MIMIC-CXR radiology report pairs, each with an original report and an LLM rewrite following a different reporting practice. Radiologists rated each report for realism and each pair for clinical equivalence.

MIMIC-IV-Ext-DischargeBench is a dataset for pre-deployment evaluation of LLM discharge educators through patient simulation. This contains seed files (demographics, extracted medical profiles) and the simulation logs from the paper.

This database allows the study of the feasibility of using ECG-capable smartwatches as diagnostic tools, focusing on their compliance with clinical standards and their ability to measure critical ECG parameters beyond AF detection.

A retrospective dataset of continuous vital sign time-series recorded from dogs and cats during general anesthesia, collected from a veterinary hospital's anesthesia machines and patient monitors.

This project utilizes data derived from the MIMIC-III, MIMIC-IV, eICU-CRD, SICdb and ZhejiangProvinceICUdb databases. The ethics approval of this dataset follows from that of these source datasets. Furthermore, all authors have obtained the necessary

CARMEN-I is a Spanish corpus of 2,000 clinical records from Hospital Clínic, Barcelona. It covers COVID-19 patients and comorbidities, serving as a resource for training clinical NLP models and researchers in NLP applied to clinical documents.

A curated dataset of paired radiology reports derived from MIMIC-CXR, with controlled clinical rewrites and expert annotations. Designed for meta-evaluation of report evaluation metrics, focusing on clinical significance, error types, and severity.

488 MIMIC-IV cases contain initial vignettes, discharge summaries, test names and diagnoses; 71 test cases are clinician-reviewed. Designed primarily for SDBench, they also support diagnostic benchmarks and clinical information extraction.