INTRODUCTION: The general practitioner (GP) plays a very important role in early detection of alcoholism. Clinical evaluation is often the only method used, although it has been suggested that any systematic patient history should include the CAGE test. We compare the effectiveness of these two approaches and attempt to determine the applicability of the CAGE test in a general practitioner's usual practice. METHOD: 12 GPs took part in this study during 6 months. They looked for possible alcohol abuse in each new patient by a standard patient history and clinical examination. Patients were randomized into 2 groups, one of which was given the CAGE test and the other not. For each patient in the CAGE group the applicability of the test was quantified by the GP. RESULTS: 416 patients were included; 214 were randomized into the "CAGE group" and 202 into the control group. On a clinical basis, 15 patients in the control group and 16 in the "CAGE group" (14 men, 2 women) were suspected of alcohol abuse. The CAGE test was positive in 15 patients (7%); among these, 6 were not suspect on a clinical basis. In patients aged 18-34, the detection rate of alcohol-related problems more than doubled when the CAGE test was used. The age of the patients influenced performance of the CAGE test and clinical evaluation. Only 2% of women had a positive CAGE test. Administration of the CAGE test was considered easy in 112 patients and average to difficult in 50, while the test was inapplicable with 52 patients. The latter proportion was higher than that observed in institutions (hospitals, outpatient departments) of the same region. Applicability was influenced neither by the sex nor the age of the patients, but varied greatly according to the physician (from 38% to 100%). CONCLUSION: The CAGE test increases the number of patients detected with alcohol problems by 37% and seems to be especially useful when administered to young people. The number of women with alcohol problems is probably underestimated by both clinical evaluation and the CAGE test. Furthermore, physicians in private practice are more reluctant to use the CAGE test systematically than those in a public institution.