Booster effects and mechanisms of web-based personalised normative feedback alcohol intervention for college students: A pragmatic randomised controlled trial.

ABSTRACT

BACKGROUND: To evaluate the effects of booster and no booster versions of web-based alcohol Personalised Normative Feedback (PNF) and whether descriptive norms mediated and/or participant motivation moderated the effectiveness of the intervention in real world conditions (i.e. no financial incentives). METHODS: Pragmatic randomised controlled trial with 1-, 3-, and 6-month assessments. Brazilian college students reporting alcohol use in the last 12 months (N=931) were recruited from May/2020 to December/2022 and allocated to 1) No booster/single PNF(S-PNF); 2) Booster/multiple PNF(M-PNF); or 3) Assessment-only control. We applied Helmert coding [1 Any intervention (S-PNF or M-PNF) vs. Control; and 2 S-PNF vs. M-PNF]. PRIMARY OUTCOMES: typical number of drinks/week and maximum number of drinks/week; secondary outcomes: drinking frequency and number of consequences. Three-months assessment was the primary interval. Descriptive norms were tested as mediator. Interest, importance, and readiness to change were examined as moderators. RESULTS: Compared to control, any intervention did not influence primary outcomes at 3-months or 6-months, but did at 1-month, when reduced typical drinking (IRR0.77, 95%CI0.66;0.90) and maximum number of drinks (IRR0.69, 95%CI0.58;0.82). There was an intervention effect on the consequences at 3-months. No differences were observed between S-PNF and M-PNF. No mediation effects were found at 3-months. At 6-months, there was an indirect effect on typical drinking through norms at 3-months (b=-0.82, 95%CI-2.03;-0.12) and effects on maximum drinks through norms at 1-month (b=-0.54, 95%CI-1.65;-0.02). No support for moderation was found. CONCLUSIONS: Intervention reduced alcohol drinking at 1 month only and was not effective thereafter. Mechanisms of effect remain unclear.