Leo Sher, M.D.

A research work, “Characterizing Veteran suicide decedents that were not classified as high-suicide-risk” has been published in Psychological Medicine online ahead of print (1).

According to a 2017 research work (2), approximately 90% of patients treated in the U.S Department of Veterans Affairs (VA) hospitals that go on to die by suicide do not meet the high-risk criteria and therefore do not receive targeted suicide prevention services designed for high-risk veterans. In this study, the authors used national VA data to focus on patients that were not classified as high-risk but died by suicide.

All VA patients who died by suicide in 2017 or 2018 were included in the analysis. The authors determined whether patients were classified as high-risk using the VA’s machine learning risk prediction algorithm. After excluding these patients, the authors used principal component analysis to identify moderate-risk and low-risk patients and investigated demographics, service-usage, diagnoses, and social determinants of health differences across high-, moderate-, and low-risk subgroups.

The authors found that high-risk patients tended to be younger, White, unmarried, homeless, and have more mental health diagnoses compared to moderate- and low-risk patients. Moderate- and low-risk patients tended to be older, married, Black, and Native American or Pacific Islander, and have more physical health diagnoses compared to high-risk patients. Low-risk patients had more missing data than higher-risk patients. The authors suggest that their findings raise concerns about reliance on machine learning risk prediction models.

References

1. Levis M, Dimambro M, Levy J, Dufort V, Fraade A, Winer M, Shiner B. Characterizing Veteran suicide decedents that were not classified as high-suicide-risk. Psychol Med. 2024 Sep 16:1-10. doi: 10.1017/S0033291724001296. Epub ahead of print.
2. Kessler RC, Hwang I, Hoffmire CA, McCarthy JF, Petukhova MV, Rosellini AJ, Sampson NA, Schneider AL, Bradley PA, Katz IR, Thompson C, Bossarte RM. Developing a practical suicide risk prediction model for targeting high-risk patients in the Veterans health Administration. Int J Methods Psychiatr Res. 2017 Sep;26(3):e1575. doi: 10.1002/mpr.1575. Epub 2017 Jul 4.