Citation

Papini S, Hsin H, Kipnis P, Liu VX, Lu Y, Sterling SA, Iturralde E. JAMA Psychiatry 2023; ePub(ePub): ePub.

Copyright

(Copyright © 2023, American Medical Association)

DOI

10.1001/jamapsychiatry.2022.5063

PMID

36790780

Abstract

Innovative prevention strategies are needed to reduce US suicide rates, which have been steadily increasing with a recent disproportionate increase among Black and Hispanic populations.1 Predictive models of suicide risk have been developed using machine learning with electronic health records (EHRs).2 Some models achieve the performance needed to cost-effectively target high-risk individuals.3 However, recent work applying a suicide-death prediction model found excellent performance at the population level (indexed by area under the receiver operating characteristic curve [AUC] = 0.82), but much lower AUC for the American Indian or Alaskan Native subsample (AUC = 0.60).4 We examined whether similar disparities exist in the prediction of suicide attempts, which are far more common and have detrimental effects on individuals and health care systems, even when nonfatal.

Language: en