Reproduce “EHR2Trace: Auditable EHR Data Infrastructure for Patient World Models and Clinical Agents” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xinye Yang, Yuli Wang, Cheng Ting Lin, Harrison Bai
EHR2Trace converts EHR histories into OMOP tables and MEDS event streams, preserving source links, event and availability times, and distinct medication actions. Validation compares saved outputs with source and audit records. Across three datasets, it produced 846.4 million canonical events. In a retrospective MIMIC-IV mortality-proxy experiment, availability-filtered training scored 0.829 AUROC versus 0.642 for backdated training tested on filtered histories.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.