Suicide and self-harm prevention on social media: A systematic review of interventions and user-led practices

Social media platforms are commonly used to seek help for suicide and self-harm, yet their role in suicide prevention remains underexamined. This systematic review updated an earlier review and synthesises 11 years of empirical research (2014–2025), examining the use of social media for self-harm and suicide prevention. A systematic search of five databases was conducted in April 2025, and data on the prevention activity or intervention and outcomes were extracted. The Mixed Methods Appraisal Tool was used to assess study quality. From 16,476 records, 75 studies met inclusion criteria and were classified as: describing intervention development/format (17%), intervention evaluation (19%), user experiences (34%) or online behaviours (29%). Significantly more intervention studies were identified in this update compared with the original review and while evidence generally supports their safety, acceptability and feasibility, evidence was largely limited to short-term outcomes. Further, the heterogeneity of approaches and outcome measures limited our ability to determine which approaches had greatest benefit. Community-led practices (e.g., peer support) provided social connection, although benefits coexisted with risks including exposure to graphic and instructional content. Users described social media platforms as important avenues for help-seeking, particularly when offline support was inaccessible. Few studies evaluated intervention effectiveness or measured changes in self-harm or suicidal behaviour, and most relied on small, non-representative samples. • Limited research to date has examined social media-based suicide prevention approaches. • This review synthesises 75 studies on social media-based suicide prevention (2014–2025). • Interventions show promise, but evidence remains limited to short-term outcomes. • Peer support fostered connection and relief alongside risks including harmful content exposure. • Few studies measured behavioural outcomes; rigorous longitudinal research is needed.