Social-Media-Based Mental Health Interventions: Meta-Analysis of Randomized Controlled Trials

The study addresses the global mental health crisis marked by high prevalence of anxiety and depression and limited access to services due to stigma, workforce shortages, and cost. Social-media-based interventions offer scalable, low-cost, and accessible support through familiar platforms that can deliver peer support and psychoeducation. The authors note a gap: although many meta-analyses cover online or digital interventions, few rigorously examine social-media-based mental health interventions for the general population. The meta-analysis aims to synthesize randomized controlled trials (RCTs) meeting strict methodological criteria to assess whether social-media-based interventions reduce negative mental health outcomes. Research Questions: RQ1 examines overall impacts of social-media-based RCTs on alleviating negative mental health outcomes (depression, anxiety, stress, negative affect, psychological distress) compared with care as usual (CAU) or waitlist; hypothesis H1 expects effectiveness. RQ2 examines moderation by recruitment type (clinical vs nonclinical), age, control group type (waitlist vs CAU), intervention delivery (self-guided vs human-guided), sex composition, program duration, and program focus (social vs task-oriented); hypothesis H2 expects larger effects with clinical populations, younger age, more passive controls, human guidance, social orientation, more women, and longer duration.

Prior reviews and meta-analyses have extensively covered online, digital, eHealth, computer-based, and internet interventions for mental health, often reporting effectiveness for depression and anxiety. However, only two meta-analyses focused specifically on social-media-based interventions and those were among patients with cancer, alongside a scoping review for children and a review of social networking site interventions for young people. There remains no meta-analysis of rigorously designed social-media-based mental health interventions for general populations, motivating this work. Potential moderators were selected based on findings from broader online/digital intervention literature, including recruitment type, age, control group type, intervention delivery personnel, duration, program orientation (social vs task), and sex.

Registration: Preregistered on Open Science Framework, with three deviations from protocol: broadened population to include all ages with age-group moderators; narrowed interventions to social-media-based only to reduce heterogeneity; focused outcomes on negative mental health (depression, anxiety, stress, negative affect, psychological distress) to avoid heterogeneity with positive outcomes. Search Strategy: Comprehensive search across 7 databases (ERIC, PsycINFO, Scopus, PsycArticles, Communication and Mass Media Complete, PubMed, ProQuest), targeted hand searching via Paperfetcher, and forward/backward citation tracking using CitationChaser, completed by April 2025, yielding 11,658 records. Screening conducted in Covidence. Eligibility Criteria (12): RCTs; ≥30 participants per condition at baseline; interventions delivered largely via social media platforms (e.g., Facebook, Instagram, WhatsApp, WeChat) excluding abstinence interventions; baseline differences <0.25 SD on mental health measures; differential attrition <15%; delivered by nonresearchers; quantitative measures of negative mental health outcomes with data sufficient to compute Hedges g; full-text English and available online; published ≥2005; primary studies; exclude one-item measures; exclude single-session interventions. Screening and Coding: Double-blinded independent screening by at least two authors; dual coding using Google Spreadsheets; conflicts resolved via group discussion; materials and code shared on GitHub. Analytical Plan: Computation of standardized mean differences (Hedges g) using R metafor package (escalc), inverse-variance weighting with Hedges’ adjustments, random-effects meta-regression due to heterogeneity. Moderators (grand-mean centered): recruitment type (clinical vs nonclinical), age group (adolescents <20; early adulthood 20-<40; middle adulthood 40-<60; late adulthood ≥60), control group type (waitlist vs CAU/other), intervention delivery (self-guided vs human-assisted), program duration (weeks), sex composition (≥70% female vs <70%), program orientation (social vs task). Publication bias assessed using selection modeling (weightr) rather than funnel plots/Egger/fail-safe N. Risk of bias assessed with JBI Critical Appraisal Checklist for RCTs, coding criteria 2, 4, 5, 6, 8; double-blinded quality assessment with third reviewer resolution.

Included studies: 17 studies, 22 intervention programs, total sample size 5624; 73 effect sizes (depression n=31, anxiety n=27, stress n=12, negative affect n=2, psychological distress n=1). Overall effectiveness: random-effects mean ES=0.32 (95% CI 0.24–0.45, P<.001). Heterogeneity: I²=88.10% (partial I² between studies 28.87%, within-cluster 59.23%), τ²=0.13; prediction interval −0.38 to 1.08 (reported also as −0.42 to 1.06). Outcome-specific effects: depression ES=0.31 (P<.001, n=31), anxiety ES=0.33 (P=.04, n=27), stress ES=0.69 (P=.02, n=12). Moderator results (meta-regression): • Program focus: social-oriented more effective than task-oriented (β=−0.76 comparing task to social, P=.03); marginal means social 1.20 (P<.01) vs task 0.44 (P=.01). • Control group: waitlist less effective than CAU/active (β=−0.49, P=.02); marginal means CAU 1.37 (P<.01) vs waitlist 0.88 (P<.01). • Delivery: self-guided less effective than human-guided (β=−0.72, P=.02); marginal means human-guided 1.35 (P<.01) vs self-guided 0.63 (P=.01). • Sex: ≥70% female associated with larger effects (β=1.40, P=.01); marginal means ≥70% female 1.81 (P<.01) vs <70% female 0.41 (P=.02). • Age: no significant moderation; comparisons trending but nonsignificant (e.g., adolescents vs middle adulthood P=.06); marginal mean largest for late adulthood (1.92, P<.01) with very limited data. • Clinical vs nonclinical: not significant (β=0.34, P=.17). Publication bias: selection modeling indicated upward adjustment of mean effects (e.g., g≈0.98 for P<.01 cut-off; g≈0.02 for P=.28 cut-off), suggesting significant results more likely reported. Risk of bias: generally low; mean appraisal 9.29/13; varying concealment and blinding across studies. PRISMA screening: 11,658 identified; 4,019 duplicates removed; 7,639 screened; 761 full-text assessed; 17 included; common exclusions: wrong design (n=345), not social media–based (n=112), not negative outcome (n=77), <30 per condition (n=73), abstinence (n=35), differential attrition >15% (n=30), delivered by researchers (n=18), single session (n=18), insufficient data (n=17), baseline difference >0.25 SD (n=16).

Findings address RQ1 by demonstrating that rigorously designed social-media-based RCTs significantly reduce negative mental health outcomes (depression, anxiety, stress), supporting the hypothesis that such interventions are effective. For RQ2, several moderators influenced effectiveness: interventions were more effective when human-guided, socially oriented, with CAU controls, and when samples were majority female, indicating that social connection and human support may be key mechanisms of benefit in social-media contexts. Unexpectedly, programs with waitlist controls were less effective, potentially because waitlisted participants sought alternative treatments. Age and clinical recruitment did not significantly moderate effects, suggesting broad applicability across age groups and settings, though data in some subgroups were limited. The discussion highlights cultural patterns (e.g., female-only recruitment in Muslim-majority countries) and posits that social-media anonymity and support may particularly appeal to women, aligning with theories such as tend-and-befriend. Compared with broader online/digital interventions, social-media-based effects were smaller, possibly reflecting the relative novelty of designs and implementation fidelity, but still meaningful given scalability and accessibility.

This meta-analysis synthesizes rigorous evidence showing that social-media-based mental health interventions can effectively reduce depression, anxiety, stress, negative affect, and psychological distress. Given their scalability, cost-effectiveness, and accessibility, integrating social-media-delivered interventions into routine mental health services could expand reach, especially for individuals facing access barriers. Future research should develop more rigorous RCTs, explore interactions among program orientation, sex, personality, and technology comfort, design three-arm trials to compare CAU vs waitlist more directly, and improve reporting on participant characteristics such as race to assess equity. Policymakers and platforms might collaborate to embed evidence-based interventions within social media to promote equitable mental health support.

Primary limitations include the small number of included studies (n=17) limiting statistical power and degrees of freedom in meta-regression; sparse data in certain outcome and moderator subgroups (e.g., psychological distress, negative affect, late adulthood) warrant cautious interpretation; and selection modeling indicates potential publication bias with upwardly adjusted mean effects. Although risk of bias was generally low due to stringent inclusion criteria, blinding and allocation concealment were variably reported.