Artificial intelligence, machine learning approach and suicide prevention: a qualitative narrative review

Abstract

The initial step in suicide prevention involves identifying individuals who may be at risk of attempting suicide at an early stage. Utilising artificial intelligence (AI) and machine learning (ML) techniques offers innovative avenues for the early detection of such individuals. Nevertheless, there is a lack of clear information regarding the application of AI and ML in suicide prevention. Our objective is to examine the latest research findings on the utilization of AI/ML in forecasting suicidal tendencies. Authors reviewed four databases (PubMed/MEDLINE, Scopus, Web of Science and SCImago) for studies using AI/ML for suicide prevention published in English from 1 January 2000 to 31 December 2021. Search strings and MeSH were employed for searching terms relevant to suicide prevention and AI/ML. Results of the studies were analysed qualitatively, and information was presented as tables and figures. After removing duplicate articles, out of 434 studies, 21 articles, involving a total of 274,876 participants, met the inclusion criteria and were considered for this review. The results suggested that AI/ML-based suicide prediction models might improve healthcare systems by identifying individuals at high risk of suicide by preventing suicidal attempts. However, further researches are needed to perform AI/ML-based evidence-based assessment tools and determine their validity and reliability for suicide prediction models in different contexts.