Detection by the CAGE of alcoholism or heavy drinking in primary care outpatients and the general population.

There is a need to improve the diagnosis of alcoholism in clinical settings because alcoholism, particularly in its early stages, is often unrecognized in general medical practice and in hospitals. In this study the CAGE questionnaire was used to detect alcoholism or heavy drinking in three populations, namely, alcoholics in treatment (ALC), primary-care outpatients (PC), and the general population (GP). Nearly all the ALC tested positive on the CAGE (97.8%), both for current (past year) and for lifetime alcohol-related problems. Among the PC subjects, 44.8% tested positive for lifetime alcohol problems, but the prevalence decreased to 17.2% when only past-year problems were considered. Likewise, 38.3% of the GP sample tested positive for lifetime, but half of these did not meet the 1-year recency criterion. Compared to DSM-III-R criteria during the same time intervals, the sensitivity/specificity of the lifetime CAGE was 91.2%/84.0% and 76.9%/85.1% in the PC and GP, respectively. The corresponding sensitivity/specificity of the past-year CAGE was 94.4%/97.0% and 74.6%/91.6%, respectively. Thus, the CAGE is an appropriate screening test for alcohol problems in these two populations, but other confirmatory tests or interviews are necessary to eliminate false positives. There were neither gender nor racial differences in the ALC sample responses to individual CAGE questions. However, there were gender differences in the PC and GP samples, with more males responding yes to each of the questions. The gender differences probably reflected the higher prevalence of heavy drinking and alcoholism among males.