Lock 2006.

Methods	Cluster randomised controlled trial: brief intervention versus standard advice, average cluster size = 2. ITT: yes.
Participants	Setting: UK, general practice. Participants: Aged ≥ 16 years with AUDIT score ≥ 8 for men or ≥ 7 for women; excluded if current major physical or psychiatric illness, or severely alcohol dependent, or severely brain damaged or mentally impaired. Number randomised = 127; 50% male; mean age = 44.1 years; 72% employed, 5% unemployed, 15% retired, 3% students; 3% primary school, 6% some secondary school, 47% completed secondary school, 21% technical or trade certificate, 23% university or tertiary education. At baseline: mean weekly units consumed = 24.6; mean AUDIT score = 10.5.
Interventions	Intervention group (N = 67) received 5 to 10 minute intervention using the 'drink‐less' protocol. This involved structured advice including: standard drink units, recommended low‐risk consumption levels, benefits of cutting down drinking, tips on helping patients reduce consumption, advice on how to set goals, determine action and review progress. Also received a self‐help booklet/diary to take away. Control group (N = 60) received standard treatment comprising nurses advice on cutting down drinking and UK Government Health Education Authority leaflet entitled 'Think About Drink', which contained daily benchmark guides for adult men and women and basic advice on alcohol.
Outcomes	AUDIT score, mean drinks per drinking day, mean units per week, Drinking Problems Index, SF‐12 physical health, health related quality of life measured by the SF‐12. Assessed at 6 and 12 months.
Funding source	Funding for this project was provided by an NHS Executive (Northern & Yorkshire) Research and Development Regionally Commissioned Project Grant.
Declaration of interests	Not reported.
Notes	Loss to follow‐up at 12 months: Intervention group: 31/67 (46.2%). Control group: 18/60 (30.0%). Number of patients assessed = 78 in 40 GP practices.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer generated allocation (p. 428).
Allocation concealment (selection bias)	Low risk	Randomisation was carried out by a member of the team not involved in recruitment or training, all patients in each practice received the same intervention (p. 428).
Blinding of treatment providers	Low risk	Cluster randomised so that treatment providers interacted with only one arm.
Blinding of participants	High risk	Patients knew that alcohol advice was being evaluated (p. 429).
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome measurement was done by researchers who were blind to patient's study group (p. 429).
Incomplete outcome data (attrition bias) All outcomes	High risk	Loss to follow‐up > 30%.
Selective reporting (reporting bias)	Low risk	Outcomes specified in methods are reported.