Engagement and Adherence in Digital Mental Health: An Umbrella Review of Definitions and Measures

Digital Mental Health Interventions (DMHIs) show significant potential in improving mental well-being, particularly among marginalized populations with limited access to traditional services. However, persistent challenges around engagement and adherence hinder their overall effectiveness. Despite a growing body of literature, there remains no consensus on how these concepts are defined or measured. This complicates evaluation, implementation, and the development of best practices.

This umbrella review synthesizes evidence from systematic reviews, scoping reviews, and meta-analyses to clarify how engagement and adherence in DMHIs are defined and measured. It also examines additional factors that affect user retention. Following PRISMA 2020 and its extension for umbrella reviews, a systematic search was conducted across six databases (Web of Science, CINAHL, MEDLINE, PsycINFO, Cochrane, and ProQuest), alongside manual searches in BMC Digital Health. After deduplication, 2,196 records underwent title and abstract screening, with 170 full texts assessed for eligibility. Supplementary searching yielded an additional 69 studies for full-text review.

Data extraction captures study characteristics, population details, and definitions and metrics used. Methodological quality is assessed using the JBI Critical Appraisal Checklist. Employing a convergent mixed methods synthesis, the review identifies prevailing themes, inconsistencies, and gaps in the conceptualization and measurement of engagement (e.g., app usage patterns, self-reported metrics) and adherence (e.g., module completion, login frequency). It also examines individual, contextual, and technological factors influencing these behaviours.

Findings aim to inform the design of user-centered DMHIs and offer recommendations for defining, measuring, and reporting engagement and adherence. The review is currently ongoing and is expected to be completed prior to the conference. The protocol is registered with PROSPERO (CRD42025637603) and available as a preprint on JMIR Research Protocols (10.2196/preprints.73438).