Study characteristics of 16 randomised controlled trials of Internet-based interventions for adult alcohol misuse.

Notes:1–4: Risk of bias: 1): adequate sequence generation; 2): report of conditions by an independent party; 3): blinding of outcome assessor or using self-report outcomes only; 4): intention to treat analyses; *: in regression analysis coded as NO; ∧ control group reported twice; &: at risk group drinkers only; aintervention group size; control group size; cfull-care intervention; ?stepped-care intervention; ebrief personalised feedback intervention; fbrief personalised advice intervention; gAlcohol Savvy intervention; hDrinker’s Check-Up intervention; itext message feedback intervention; jtext message assessment intervention; k1 ‘drink’ contains 10 grams of ethanol; lonly a subsample included; mnot all participants randomly allocated; AB alcohol brochure control; AB/H alcohol/health brochure control; AO assessment-only control; AUDIT: Alcohol Use Disorders Identification Test; AUDI-C brief AUDIT, composite measure that consists of respondents’ scores on frequency of drinking, drinks per drinking day and frequency of five or more drinks on one occasion; BSC behavioural self-control training; CBT cognitive-behavioural therapy; CO completers-only analysis; CW comparator website control; DO Study dropout; FAST Fast Alcohol Screening Test; GGT gamma-glutamyl transpeptidase test; HB Health booklet control; HBM Health Belief model; HE health education; ITT intention-to-treat analysis; MI motivational interviewing; PFGS personalised feedback with goal setting; PNF personalised normative feedback; SCM stages of change model; TTM transtheoretical model of change.