Improving Pre-Exposure Prophylaxis Adherence in People at Risk for HIV: Secondary Analysis of a Digital Health Intervention to Enhance User Engagement

Background: Although highly effective HIV pre-exposure prophylaxis (PrEP) is available, its usage and adherence among young men who have sex with men and young transgender women remain low, reducing its overall effectiveness. The study included a 3-arm randomized clinical trial of Prepared, Protected, emPowered (P3), a comprehensive PrEP adherence digital health intervention, compared to an enhanced version, P3+, which incorporates in-app adherence coaching. Objective: This study aims to analyze data from study participants in the P3/P3+ intervention arms to understand how different levels of user engagement with the app's features were associated with adherence to PrEP as well as the costs of each intervention and their relative cost-effectiveness. Methods: Descriptive statistics for study variables at baseline were calculated. To examine the differences in intervention engagement and acceptability by arm, independent samples 2-tailed t tests for continuous variables and a chi-square analysis for categorical variables were conducted. To examine the effect of arm and engagement categories on PrEP adherence at 3 months, three logistic regression analyses were conducted: (1) the effect of arm on PrEP adherence, (2) the effect of predefined engagement categories (high vs moderate and low) on PrEP adherence, and (3) the interaction effect of arm and predefined engagement categories on PrEP adherence, along with the main effects of arm and predefined engagement categories. The study team calculated the average cost per participant and the incremental cost-effectiveness for PrEP adherence and engagement measures. Results: A total of 163 participants were randomized to the P3 intervention (82 to the P3 arm and 81 to the P3+ arm). Participants in the P3+ arm earned higher incentives (US $90.6 vs $75.4; P=.04), had more app log-ins (96.6 vs 76.1; P=.01), used the app on more days (63.3 d vs 53.2 d; P=.04), and spent more time in the intervention (378.8 min vs 186.66 min; P