Existing question answering (QA) datasets derived from electronic health records (EHR) are artificially generated and consequently fail to capture realistic physician information needs. We present Discharge Summary Clinical Questions (DiSCQ), a newly curated question dataset composed of 2,000+ questions paired with the snippets of text (triggers) that prompted each question. The questions are generated by medical experts from 100+ MIMIC-III discharge summaries. We analyze this dataset to characterize the types of information sought by medical experts. We also train baseline models for trigger detection and question generation (QG), paired with unsupervised answer retrieval over EHRs. Our baseline model is able to generate high quality questions in over 62% of cases when prompted with human selected triggers. We release this dataset (and all code to reproduce baseline model results) to facilitate further research into realistic clinical QA and QG: https://github.com/elehman16/discq.
@article{arxiv.2206.02696,
title = {Learning to Ask Like a Physician},
author = {Eric Lehman and Vladislav Lialin and Katelyn Y. Legaspi and Anne Janelle R. Sy and Patricia Therese S. Pile and Nicole Rose I. Alberto and Richard Raymund R. Ragasa and Corinna Victoria M. Puyat and Isabelle Rose I. Alberto and Pia Gabrielle I. Alfonso and Marianne Taliño and Dana Moukheiber and Byron C. Wallace and Anna Rumshisky and Jenifer J. Liang and Preethi Raghavan and Leo Anthony Celi and Peter Szolovits},
journal= {arXiv preprint arXiv:2206.02696},
year = {2022}
}