In a controlled evaluation of general practitioner (GP)-based brief intervention, 378 excessive drinkers identified opportunistically by screening in 40 group practices in metropolitan Sydney were assigned to groups receiving: (i) a five-session intervention by the GP (the Alcoholscreen Program); (ii) a single session of 5 minutes' advice by the GP plus a self-help manual (minimal intervention); (iii) an alcohol-related assessment but no intervention; (iv) neither intervention nor assessment. Among all patients allocated to receive it, the Alcoholscreen Program did not result in a significantly greater reduction in consumption at follow-up than control conditions but patients offered Alcoholscreen reported a significantly greater reduction in alcohol-related problems in the period to 6 months follow-up. A greater proportion of patients who returned for the second Alcoholscreen visit were drinking below recommended levels at follow-up than in the remainder of the sample. There was no evidence that minimal intervention or alcohol-related assessment were effective in reducing alcohol consumption or problems. Implications for further research into GP-based brief interventions are discussed.