Citation

Perlis RH. JAMA Netw. Open 2023; 6(4): e235781.

Copyright

(Copyright © 2023, American Medical Association)

DOI

10.1001/jamanetworkopen.2023.5781

PMID

37022688

Abstract

With the explosion of studies relying on large-scale electronic health records or health claims data, it is easy to forget that diagnostic codes exist for a very specific purpose entirely unrelated to research: to enable health care systems to collect payment. What such codes lack in precision, however, they may make up in quantity: health records-based research enables generation of far larger cohorts in silico than would ever be feasible with prospective data collection, particularly for otherwise difficult-to-capture phenotypes.

One area where health records may prove particularly important is in research understanding the risk factors for and consequences of firearm-related violence, because for 2 decades the US Congress limited funding for such investigation.1 But even when investigators could not prospectively collect data on firearm injuries, these patients kept coming,2 presenting to emergency departments at rates unfathomable to most of the world.

Language: en