A just‐in‐time adaptive personalized mobile app intervention produces limited short‐term reductions in heavy episodic drinking frequency in college students: A randomized control trial during the COVID‐19 pandemic

Abstract Background High‐risk drinking among college students is common, and alcohol expectancies and experienced consequences are proximal predictors of use. This study tested short‐ and long‐term efficacy of a personalized just‐in‐time adaptive intervention with daily messages about alcohol use, alcohol expectancies, and consequences delivered via mobile app. Methods Participants were 408 students enrolled at a 2‐ or 4‐year college (75% 4‐year; 64% female) who completed twice‐daily assessments for 21 days and follow‐up surveys at 1‐, 6‐, and 12‐months. Data collection spanned January 2020 through April 2022. Participants were randomized to either an intervention condition receiving daily intervention messages and other related alcohol and expectancy‐focused content via the app or an assessment‐only control condition. Both conditions were administered daily surveys through the app. Results Poisson multilevel models were conducted to examine intervention effects on alcohol outcomes (i.e., drinks per week, heavy episodic drinking frequency, peak estimated blood alcohol concentration, and alcohol‐related consequences, as well as positive and negative alcohol expectancies) at each follow‐up assessment. On average, participants in both conditions reported decreased alcohol use outcomes, consequences, and expectancies at 1‐, 6‐, and 12‐month follow‐up compared to baseline. A single statistically significant difference between participants in the intervention and control conditions was observed at 1‐month follow‐up. Specifically, participants in the intervention condition reported a 17% greater decrease in heavy episodic drinking frequency than participants in the control condition, on average. Conclusions Findings highlight the complexities of developing, implementing, and testing adaptive interventions, particularly within the context of the COVID‐19 pandemic.