JMIR Publications

ABSTRACT

Background:

Digital mental health interventions (DMHI) have gained prominence as accessible and cost-effective solutions in workplace settings. However, our previous meta-analysis revealed a concerning trend: despite advancements in technology, the effectiveness of these interventions has not improved over time. Heterogeneity among interventions and sample may cause this stagnation.

Objective:

This review, a first of its kind uses a Bayesian meta-regression to examine intervention characteristics that impact on effectiveness.

Methods:

Methods:

A systematic review of RCTs of employee-based DMHIs. Eligible studies were assessed based on specific criteria, participant and intervention characteristics, and outcome measures. Data extraction and coding were performed, followed by a Bayesian meta-analysis approach. Allowing for a more nuanced evaluation of the effectiveness of intervention features and designs, accounting for uncertainty and prior knowledge in the field.

Results:

81 RCTs were identified involving 98 interventions and ~25,000 participants. Both sample and intervention characteristics contributed to heterogeneity across studies. Stress management (stress ER=9.8) and mindfulness (depression ER=3.0; anxiety ER=3.6) interventions demonstrated more efficacy than CBT-based approaches. Utilising person support showed efficacy across stress, depression and anxiety (ER = 3.9; 5.8; 10.6). Several intervention features, including videos, feedback scores, and reminder texts, were associated with positive mental health outcomes.

Conclusions:

This review provides valuable insights into the optimal design and development of workplace DMHIs. The evidence-based practices offer guidance for developers to address the heterogeneity within interventions. Importantly, findings have the potential to serve as a robust evidence base for app designers, enabling them to create more effective, personalised and engaging DMHIs. Clinical Trial: PROSPERO (CRD42022323301).