This randomized trial evaluated an intervention for reducing at-risk drinking practices in a sample of 307 patients. Eligible drinking patterns included chronic drinking (> or = 2 drinks per day in the past month), binge drinking (> or = 5 drinks per occasion at least twice in the past month), and drinking and driving (driving after > 2 drinks in the past month). Members of the intervention group received a message from their physician during their regularly scheduled visit, a self-help manual, written personalized feedback, and up to 3 telephone counseling calls. Dropout was significantly higher in the intervention than control group.