Online self-directed mental health interventions: A meta-analysis and systematic review

Background: Access to mental health care is limited by financial, time, and resource restraints. Massive open online interventions (MOOIs), a form of online self-directed mental health interventions, work to address these barriers by providing evidence-based interventions to many individuals at little to no cost. The present meta-analysis and systematic review aimed to assess the efficacy of MOOIs to inform the development of future MOOIs for mental health. Methods: Following the Preferred Reporting for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, a systematic review of research investigating fully online and self-directed mental health interventions was conducted within the PsycINFO, PsycARTICLES, and Academic Search Complete databases. Random effect models were used to calculate pooled effect sizes using mean differences for mental health outcomes to determine the efficacy of MOOIs. Person-level (age, gender, continental location) and program-level (control type, inclusion criteria, therapeutic orientation, treatment components, treatment format, time to follow-up, treatment duration) moderators were examined with mixed-effect or meta-regression models. Results: The meta-analysis included 31 articles with 10397 participants. MOOIs (n = 5235) were more effective at reducing mental health symptoms versus controls (n = 5162). Small intervention effect sizes were observed for anxiety, depression, sleep, and stress at post-treatment. At follow-up, all mental health outcomes were null. Subgroup analyses indicated studies had a larger effect size at post-treatment if they had an inactive control (vs. active), were conducted in Europe (vs. other continental locations), and delivered the intervention through web-based formats (vs. mobile applications). No other moderators were significant. Conclusions: This meta-analysis demonstrates the positive potential impact of MOOIs in reducing mental health symptomology across various populations. However, due to the limited findings, it cannot be concluded which intervention components led to the most favourable outcomes and for whom may benefit most from MOOIs. Further research should investigate which aspects of MOOIs lead to meaningful results to help design and implement more effective self-directed online interventions.