Psychosocial interventions for non-suicidal self-injury and self-harm outcomes in adolescents: A systematic review and network meta-analysis

Non-suicidal self-injury (NSSI) and self-harm are common during adolescence and are associated with later suicidal behavior, psychiatric morbidity, and functional impairment. Previous evidence syntheses have often combined psychosocial and pharmacological treatments or pooled conceptually distinct self-harm outcomes. We therefore compared psychosocial interventions specifically and synthesized dichotomous NSSI/self-harm occurrence, continuous NSSI/self-harm severity, and broader self-harm events as separate outcome domains. We searched PubMed, Embase, Web of Science, the Cochrane Library/CENTRAL, Scopus, China National Knowledge Infrastructure, Wanfang, and VIP for randomized controlled trials of psychosocial interventions in adolescents or young people reporting NSSI/self-harm outcomes. Continuous outcomes were summarized as standardized mean differences (SMDs) and dichotomous outcomes as odds ratios (ORs). Random-effects frequentist network meta-analysis was conducted with treatment as usual/usual care (TAU/usual care) as the reference node. Risk of bias was assessed separately for each outcome and analysis time point. Disconnected or insufficiently populated networks were not forced into network meta-analysis. The review was registered in PROSPERO (CRD420261442260). The quantitative evidence base comprised 30 randomized controlled trials. The dichotomous NSSI/self-harm network included 10 studies and 10 treatment nodes (I² = 42.1%). Compared with TAU/usual care, culturally adapted CBT/problem-solving intervention (OR = 0.19, 95% CI 0.04 to 0.99), DBT/DBT-A (OR = 0.32, 95% CI 0.11 to 0.93), and group therapy/psychosocial intervention (OR = 0.14, 95% CI 0.02 to 0.96) were associated with lower NSSI/self-harm risk. For continuous NSSI/self-harm outcomes, DBT-based intervention (SMD = -1.42, 95% CI -2.33 to -0.52), CBT-based/third-wave CBT (SMD = -1.21, 95% CI -2.23 to -0.19), and general psychosocial/nursing intervention (SMD = -1.64, 95% CI -2.81 to -0.48) favored lower severity, although heterogeneity was high (I² = 95.1%). The broader dichotomous self-harm network included only two studies. Excluding trials at high risk of bias preserved the direction of the principal binary and continuous estimates, while heterogeneity in the continuous network remained substantial. DBT/DBT-A, CBT/problem-solving interventions, and some group or comprehensive psychosocial interventions may reduce adolescent NSSI/self-harm risk or severity. However, the evidence network was sparse and largely anchored to TAU/usual care, several nodes were supported by single studies, and continuous outcomes were highly heterogeneous. These findings should therefore be interpreted as comparative signals rather than a definitive treatment hierarchy.