Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Aug 15;17(16):5929.
doi: 10.3390/ijerph17165929.

Artificial Intelligence and Suicide Prevention: A Systematic Review of Machine Learning Investigations

Rebecca A Bernert  1 , Amanda M Hilberg  1 , Ruth Melia  1   2 , Jane Paik Kim  1 , Nigam H Shah  3   4 , Freddy Abnousi  5   6
Affiliations

Artificial Intelligence and Suicide Prevention: A Systematic Review of Machine Learning Investigations

Rebecca A Bernert et al. Int J Environ Res Public Health. .

Abstract

Suicide is a leading cause of death that defies prediction and challenges prevention efforts worldwide. Artificial intelligence (AI) and machine learning (ML) have emerged as a means of investigating large datasets to enhance risk detection. A systematic review of ML investigations evaluating suicidal behaviors was conducted using PubMed/MEDLINE, PsychInfo, Web-of-Science, and EMBASE, employing search strings and MeSH terms relevant to suicide and AI. Databases were supplemented by hand-search techniques and Google Scholar. Inclusion criteria: (1) journal article, available in English, (2) original investigation, (3) employment of AI/ML, (4) evaluation of a suicide risk outcome. N = 594 records were identified based on abstract search, and 25 hand-searched reports. N = 461 reports remained after duplicates were removed, n = 316 were excluded after abstract screening. Of n = 149 full-text articles assessed for eligibility, n = 87 were included for quantitative synthesis, grouped according to suicide behavior outcome. Reports varied widely in methodology and outcomes. Results suggest high levels of risk classification accuracy (>90%) and Area Under the Curve (AUC) in the prediction of suicidal behaviors. We report key findings and central limitations in the use of AI/ML frameworks to guide additional research, which hold the potential to impact suicide on broad scale.

Keywords: artificial intelligence; intervention; machine learning; prediction; risk; suicide.

PubMed Disclaimer

Conflict of interest statement

No conflicts are reported for disclosure of potential conflicts of interest for the present report. Dr. Bernert has received financial support for consulting services (Facebook, Inc.; and The California Mental Health Services Oversight and Accountability Commission); no financial support was received for the present manuscript.

Figures

Figure 1
Figure 1
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.
Figure 2
Figure 2
Boxplot of accuracy by suicide outcome. Boxplot of classification accuracy by suicide outcome groupings. Notes: Outcomes assessed suicide death (M = 0.69 (SD = 0.04)); 95% CI (0.58, 0.80), suicide attempt (M = 0.82 (SD = 0.12)); 95% CI (0.71, 0.92), suicide ideation (M = 92 (SD = 0.04)); 95% CI (0.84, 0.99), other-undifferentiated (M = 0.85 (SD = 0.19)); 95% CI (−0.86, 2.57), other-social media (M = 0.84 (SD = 0.04)); 95% CI (0.74, 0.94). CI = confidence interval by outcome.
Figure 3
Figure 3
Boxplot of AUC by suicide outcome. Boxplot of classification area under the curve (AUC) by suicide outcome groupings. Notes: Outcomes assessed suicide death (M = 0.79 (SD = 0.13)); 95% CI (−0.41, 2.0), suicide attempt (M = 0.81 (SD = 0.09)); 95% CI (0.76, 0.87), suicide ideation (M = 0.78 (SD = 0.15)); 95% CI (0.52, 1.03), multiple outcomes (M = 0.87 (SD = 0.08)); 95% CI (0.67-1.06). CI = confidence interval by outcome.
See this image and copyright information in PMC

References

    1. National Center for Injury Prevention and Control, CDC Web-Based Injury Statistics Query and Reporting System (WISQARS), Fatal Injury and Violence Data. [(accessed on 30 November 2018)]; Available online: http://www.cdc.gov/injury/wisqars/index.html.
    1. Rockett I.R., Regier M.D., Kapusta N.D., Coben J.H., Miller T.R., Hanzlick R.L., Todd K.H., Sattin R.W., Kennedy L.W., Kleinig J., et al. Leading causes of unintentional and intentional injury mortality: United States, 2000–2009. Am. J. Public Health. 2012;102:e84–e92. doi: 10.2105/AJPH.2012.300960. - DOI - PMC - PubMed
    1. Shepard D.S., Gurewich D., Lwin A.K., Reed G.A., Jr., Silverman M.M. Suicide and suicidal attempts in the United States: Costs and policy implications. Suicide Life Threat. Behav. 2016;46:352–362. doi: 10.1111/sltb.12225. - DOI - PMC - PubMed
    1. Goldsmith S., Pellmar T., Kleinman A., Bunney W. Reducing suicide: A national imperative (Committee on Pathophysiology and Prevention of Adolescent and Adult Suicide, Board on Neuroscience and Behavioral Health, Institute of Medicine of the National Academies) Proc. Natl. Acad. Sci. USA. 2002 doi: 10.17226/10398. - DOI - PubMed
    1. World Health Organization . Preventing Suicide: A Global Imperative. World Health Organization; Geneva, Switzerland: 2014.

Publication types

LinkOut - more resources