Brief alcohol intervention is moving with apparent ease from demonstration outcome research to clinical practice. With benefits to public health in mind, nurses and doctors are being encouraged to give advice to all patients drinking above recommended limits. This paper critically examines the impact of this strategy on practitioners and their patients. It is argued that, firstly, practitioners might not be entirely satisfied with the evidence about effectiveness. Secondly, they might have difficulty interpreting the evidence about harmful consumption when talking to patients, particularly those without alcohol-related problems. Thirdly, the recommended framework for intervention, advice-giving, is potentially flawed. A number of possible solutions to these difficulties are raised for discussion. These include the widening of brief intervention to include severely dependent drinkers and the discussion of any health behaviours or concern expressed by the patient. These changes might best be achieved by adopting an intervention framework based on the principles of patient-centred medicine.