Meta-Analysis of Psychological and Digital Interventions to Enhance Mental Health and Well-Being in Youth: A Bayesian Umbrella Review.

Résumé

Youth mental health has become a global public health priority, with psychological distress, anxiety, and depressive symptoms increasing sharply over the last decade. Numerous interventions, ranging from mindfulness-based and cognitive behavioral programs to digital applications and peer-support initiatives, have been evaluated through meta-analytic reviews. However, the cumulative evidence remains heterogeneous and dispersed across intervention modalities. The present umbrella meta-analysis synthesized existing meta-analyses on psychological and digital interventions for adolescents and young adults, adopting a Bayesian random-effects framework to quantify the overall effectiveness and heterogeneity of outcomes. Systematic searches were conducted in PubMed, PsycINFO, and Web of Science up to September 2025, using the following syntax: ("meta-analysis" OR "systematic review") AND (adolescent* OR "youth" OR "young people") AND ("mental health" OR "well-being" OR "psychological intervention"). Eligible reviews reported standardized mean differences (Hedges' ) or convertible statistics and targeted mental health or well-being outcomes. Effect sizes were standardized using Hedges' and synthesized under a random-effects framework. They were then pooled using Bayesian random-effects modeling with a Normal (0, 0.5) prior on the grand mean μ and a half-Cauchy (0, 0.5) prior on the heterogeneity variance . Nine eligible meta-analyses ( = 9 aggregated effects, ≈1150 primary studies) met the inclusion criteria. The posterior mean standardized effect was = 0.229 (95% CrI [0.157, 0.301]), indicating a small but credible positive impact of interventions on youth mental health and well-being indicators ( = 0.19 for symptom reduction; = 0.28 for positive well-being). Between-study heterogeneity was non-negligible ( = 0.003; posterior mean = 23%, 95% CrI [0.04%, 74%]), reflecting uncertainty about the true degree of variability across modalities and settings. The posterior probability that > 0 was >0.999, providing strong Bayesian evidence for credible but heterogeneous effects. The findings suggest potentially credible but heterogeneous effects of psychological and digital interventions on youth mental health and well-being outcomes, although the magnitude and consistency of these effects remain constrained by substantial heterogeneity and the breadth of aggregated outcome constructs. Results should be interpreted with appropriate caution.