To the Editor.
—The article by Dr Fleming and colleagues1 demonstrates that advice protocols partially modified drinking behavior and improved health outcomes for at-risk problem drinkers in a primary care setting. However, numerous studies have revealed no effect or only weak positive effects of advice on lifestyle changes.2 The encouraging results in the study on problem drinking by Fleming et al may have been because of an intervention effect. The study did not include a control group that was given nonrelevant medical advice. Members of the no-advice control group had a 20% reduction in alcohol use, perhaps because they were questioned about alcohol use 4 times during a 12-month period. A multinational study conducted by the World Health Organization found similar results.3As Dr Parish indicates in his Editorial,4 those who argue that counseling by physicians affects a range of problem behaviors and that strong evidence supports counseling recognize