Citation

Devylder JE. JAMA Netw. Open 2022; 5(5): e2212106.

Copyright

(Copyright © 2022, American Medical Association)

DOI

10.1001/jamanetworkopen.2022.12106

PMID

35560056

Abstract

Elsewhere in JAMA Network Open, Wilimitis and colleagues1 set out to resolve a central debate in suicide prevention: do face-to-face risk screenings or electronic health record-based machine learning algorithms provide greater public health benefit? They approach this question by proposing and testing a third option, combining the use of these 2 modes. Their data show clearly that in a large clinical cohort of adults, the combined use of the Columbia Suicide Severity Rating Scale (C-SSRS)2 and a real-time machine learning model was statistically superior to either method alone in predicting subsequent visits for suicidal ideation or attempts, suggesting that each approach offers benefits that are, at least to some extent, additive and independent. However, a purely statistical interpretation may leave out some other relevant considerations that may lead an institution or service to choose one approach or the other (or both, as proposed in this study).

Direct face-to-face screening has some notable disadvantages compared with machine learning approaches, namely the "commonality of patients denying SI [suicidal ideation] despite being at high risk,"1 as stated by the study authors, as well as the overall clinical burden of administering and scoring the screening tools. While the combined method (screening plus machine learning) mitigates the issue of underreporting suicidal ideation, it does not counter the issue of clinical load. This clinical burden can be substantial for a relatively complex tool such as the C-SSRS, although it can notably be reduced by opting for a simpler instrument, such as the Ask Suicide-Screening Questions (ASQ), which was developed for adolescents but recently validated for use with adults as well.3 However, a substantial benefit of the face-to-face approach, which cannot be replicated with machine learning, is that face-to-face screening in itself constitutes a clinical contact and a potential point of entry into a higher level of care. In clinical settings, an individual who has never been asked about suicide may use the screening experience as a time to reach out directly for additional support; this potential benefit has not yet been quantified in the research literature but may yield clinical benefits that otherwise go undetected when examining these screening tools as predictive instruments.

Language: en