Efficacy of Digital Health Interventions and Mobile Applications in Mental Healthcare: A Systematic Literature Review

Abstract

Background: Mental disorders impose a large and growing burden, and the supply of trained professionals is far below need. Digital health interventions (DHIs), including smartphone applications, web-based programmes, conversational agents, virtual reality and sensor-enabled tools, are widely proposed as scalable ways to close this gap. Evidence on their efficacy is expanding quickly but is uneven, and it is often disconnected from evidence on engagement, quality, privacy and equity.

Objective: To critically synthesise evidence on the efficacy of DHIs and mobile applications in mental healthcare and to analyse that evidence from clinical, user-engagement, ethical-regulatory, health-system, equity and cultural perspectives.

Methods: We applied a protocol following PRISMA 2020 for a review-of-reviews design with supporting primary trials. Eligible sources were meta-analyses, systematic reviews, randomised trials and large observational analyses published from 2009 onward. Methodological quality was appraised with AMSTAR 2 and the Cochrane RoB 2 tool. Certainty was judged in the spirit of GRADE. Results were combined by narrative synthesis with a descriptive comparison of standardised effect sizes.

Results: App-based interventions show small-to-moderate average benefits for depression (g ≈ 0.28–0.38) and anxiety (g ≈ 0.30–0.33). These are comparable to or slightly above the effect of self-guided web-based cognitive behavioural therapy (g ≈ 0.27) and below the effects reported for guided formats. Benefits are larger when interventions are guided, grounded in established psychological techniques and compared against inactive controls. Engagement is weak in real-world settings. Quality, evidence claims and privacy practices in commercial app stores are inconsistent. Evidence for severe mental illness and for young people is encouraging but thin.

Conclusion: DHIs are best understood as scalable, modestly effective components of stepped and blended care, not stand-alone replacements for clinical services. We propose a four-pillar framework (efficacy, engagement, equity and ethics) to guide evaluation, implementation and policy.