Umbrella review of psychosocial and ward-based interventions to reduce self-harm and suicide risks in in-patient mental health settings.

Abstract

Background: Understanding what psychosocial interventions can reduce self-harm and suicide within in-patient mental health settings can be challenging, due to clinical demands and the large volume of published reviews.

Aims: To summarise evidence from systematic reviews on psychosocial and ward-level interventions (excluding environmental modifications) for self-harm and suicide that may enhance patient safety in in-patient mental health settings.

Method: We systematically searched Medline, Embase, CINAHL, PsycINFO and CDSR (2013-2023) for systematic reviews on self-harm and suicide prevention interventions that included in-patient data. Review quality was assessed using AMSTAR-2, primary study overlap via an evidence matrix, and evidence strength evaluated (GRADE algorithm). Findings were narratively synthesised, with input from experts-by-experience throughout (PROSPERO ID: CRD42023442639).

Results: Thirteen systematic reviews (seven meta-analyses, six narrative), comprising over 160 000 participants, were identified. Based on quantitative reviews, cognitive-behavioural therapy reduces repeat self-harm by follow-up, and dialectical behaviour therapy decreases the frequency of self-harm. Narrative review evidence suggested that post-discharge follow-up, as well as system and ward-based interventions (e.g. staff training) may reduce suicide and/or self-harm. However, review quality varied, patient involvement was lacking and methodological quality of trials informing reviews was predominately low. Overlap was slight (covered area 12.4%).

Conclusions: The effectiveness of interventions to prevent self-harm and suicide in in-patient settings remains uncertain due to variable quality reviews, evidence gaps, poor methodological quality of primary studies and a lack of pragmatic trials and co-production. There is an urgent need for better, co-designed research within in-patient mental health settings.