A Brief Intervention With Instant Messaging or Regular Text Messaging Support in Reducing Alcohol Use: A Randomized Clinical Trial

IMPORTANCE: Alcohol use is prevalent among university students. Mobile instant messaging apps could enhance the effectiveness of an alcohol brief intervention (ABI), but the evidence is scarce. OBJECTIVE: To evaluate the effectiveness of an ABI plus 3 months of mobile chat-based instant messaging support for alcohol reduction in university students at risk of alcohol use disorder. DESIGN, SETTING, AND PARTICIPANTS: In this randomized clinical trial, 772 students at risk of alcohol use disorder (Alcohol Use Disorders Identification Test [AUDIT] score ≥8) were recruited from 8 universities in Hong Kong between October 15, 2020, and May 12, 2022. Participants were randomly assigned 1:1 to either the intervention or control group. INTERVENTIONS: Both groups received the same ABI at baseline, which consisted of face-to-face or video conferencing with research nurses who delivered personalized feedback based on the participant’s AUDIT risk level, along with a 12-page booklet describing the benefits of alcohol reduction and the harmful effects of alcohol on health and social well-being. The intervention group then received 3 months of chat-based instant messaging support on alcohol reduction guided by behavioral change techniques. The control group received 3 months of short message service (SMS) messaging on general health topics. MAIN OUTCOMES AND MEASURES: All outcomes were self-reported. The primary outcome was alcohol consumption in grams per week at 6 months of follow-up. By definition, 1 alcohol unit contains 10 g of pure alcohol. Secondary outcomes at the 6-month follow-up included changes in AUDIT score, weekly alcohol consumption, intention to drink in the next 30 days, drinking frequency and any binge or heavy drinking in the past 30 days, and self-efficacy of quitting drinking. The primary analysis followed the intention-to-treat principle, and linear regression (reported as unstandardized coefficient B) and logistic regression (reported as odds ratios) were used to compare the primary and secondary outcomes between the intervention and control groups. RESULTS: The study included 772 students (mean [SD] age, 21.1 [3.5] years; 395 females [51.2%]) who were randomly assigned to either the intervention (n = 386) or control (n = 386) group. In the intention-to-treat analysis, the intervention group had lower alcohol consumption in grams per week (B, −11.42 g [95% CI, −19.22 to −3.62 g]; P = .004), a lower AUDIT score (B, −1.19 [95% CI, −1.63 to −0.34]; P