User Engagement Strategies for Digital Medication Adherence Interventions: A Systematic Review and Meta-Analysis

Background: Medication adherence is critical for chronic disease management, yet poor adherence remains an unsolved challenge. Digital health interventions are promising to support medication adherence but are limited by poor user engagement with the tools. Understanding which strategies effectively support user engagement could improve the efficacy and sustainability of these interventions.Purpose: This systematic review and meta-analysis characterized strategies used to support user engagement in digital adherence interventions with the well-established Behavior Change Technique Taxonomy and evaluated whether these strategies were associated with medication adherence and user engagement outcomes. Methods: Electronic databases (Cumulative Index to Nursing and Allied Health Literature (CINAHL), Embase via Ovid, MEDLINE, PsycInfo, and Web of Science Core Collection) were searched (4/18/2025) to identify randomized controlled trials of app-based digital adherence interventions for people with chronic medical conditions taking oral medications, published in English between 2020 and 2025. User engagement strategies were characterized with the Behavior Change Technique Taxonomy. Meta-regression models evaluated associations between specific techniques and study outcomes. Risk of bias was characterized with the Risk of Bias-1 measure. Results: The 60 studies included for review predominantly included adults with cardiovascular disease, hypertension, and type 2 diabetes (median N=117, IQR=72.0-204.5). All but one included at least one user engagement strategy appropriate to characterize with the Behavior Change Technique Taxonomy. The most common techniques were Prompts/cues (90%), Instruction on how to perform the behavior (70%), Adding objects to the environment (35%), Feedback on behavior (28%), Self-monitoring of behavior (25%), Credible source (22%), Goal setting (behavior) (20%), and Self-monitoring outcome(s) of behavior (20%). Prompts/cues were frequently paired with another behavior change technique (54%). User engagement was rarely measured as an outcome, and thus, it was not possible to evaluate this outcome in the meta-analysis. In the 44 studies included in the meta-analysis of medication adherence outcomes, there were no statistically significant associations with user engagement techniques. Conclusions: Digital adherence interventions commonly include strategies to support user engagement, particularly prompts/cues, but it remains unknown whether these strategies translate to actual user engagement. No user engagement strategy was associated with medication adherence outcomes. This may reflect variability across study samples and medication adherence measurement approaches, threats to internal validity, or reliance on ineffective strategies. Further research should continue to evaluate which engagement strategies support sustained digital adherence tool use and improved adherence outcomes.