Learning to Ask Like a Physician
Existing question answering (QA) datasets derived from electronic health records (EHR) are artificially generated and, as a result, fail to capture realistic physician information needs. We present Discharge Summary Clinical Questions (DiSCQ), a newly curated question dataset composed of 2,000+ ques...
Main Author: | |
---|---|
Other Authors: | |
Format: | Thesis |
Published: |
Massachusetts Institute of Technology
2022
|
Online Access: | https://hdl.handle.net/1721.1/144613 |
_version_ | 1811078279770144768 |
---|---|
author | Lehman, Eric |
author2 | Szolovits, Peter |
author_facet | Szolovits, Peter Lehman, Eric |
author_sort | Lehman, Eric |
collection | MIT |
description | Existing question answering (QA) datasets derived from electronic health records (EHR) are artificially generated and, as a result, 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 will release this dataset (and all code to reproduce baseline model results) to facilitate further research into realistic clinical QA and QG. |
first_indexed | 2024-09-23T10:57:10Z |
format | Thesis |
id | mit-1721.1/144613 |
institution | Massachusetts Institute of Technology |
last_indexed | 2024-09-23T10:57:10Z |
publishDate | 2022 |
publisher | Massachusetts Institute of Technology |
record_format | dspace |
spelling | mit-1721.1/1446132022-08-30T03:45:31Z Learning to Ask Like a Physician Lehman, Eric Szolovits, Peter Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Existing question answering (QA) datasets derived from electronic health records (EHR) are artificially generated and, as a result, 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 will release this dataset (and all code to reproduce baseline model results) to facilitate further research into realistic clinical QA and QG. S.M. 2022-08-29T15:59:38Z 2022-08-29T15:59:38Z 2022-05 2022-06-21T19:25:42.893Z Thesis https://hdl.handle.net/1721.1/144613 0000-0001-9919-2257 In Copyright - Educational Use Permitted Copyright MIT http://rightsstatements.org/page/InC-EDU/1.0/ application/pdf Massachusetts Institute of Technology |
spellingShingle | Lehman, Eric Learning to Ask Like a Physician |
title | Learning to Ask Like a Physician |
title_full | Learning to Ask Like a Physician |
title_fullStr | Learning to Ask Like a Physician |
title_full_unstemmed | Learning to Ask Like a Physician |
title_short | Learning to Ask Like a Physician |
title_sort | learning to ask like a physician |
url | https://hdl.handle.net/1721.1/144613 |
work_keys_str_mv | AT lehmaneric learningtoasklikeaphysician |