Validity of the CAGE questionnaire for screening alcohol-dependent inpatients on hospital wards

Abstract

To verify the validity of the CAGE questionnaire in screening inpatients with alcohol dependence. In a transversal study, 747 medical inpatients hospitalized on general medical wards in the Federal University of Santa Catarina University Hospital were evaluated. Sociodemographic and clinical data were collected and the following instruments were used: the CAGE questionnaire and the Mini International Neuropsychiatry Interview (MINI), the latter being a semi-structured interview used as the gold standard for diagnosing alcohol dependence (according to DSM-IV criteria). Validity indices (sensitivity and specificity) were assessed for the different possible CAGE cut-off points. The ROC curve was used to determine the best cut-off point. The sample was composed of 747 patients. Most were men (66%), white (85%) and married (61%). Mean age was 50 +/- 17 years, and mean level of education was 6 +/- 4 years. According to the MINI, 48 patients (6.6%) were diagnosed as having alcohol dependence. The CAGE questionnaire presented its highest sensitivity (93.8%) when the cut-off point of 0/1 (one or more "positive" responses indicating a positive test) was used. The specificity for this cut-off point was 85.5%. Using the 0/1 cut-off point, the CAGE questionnaire presented good sensitivity (93.8%) and specificity (85.5%) for use in general hospital ward patients. Since it is an easily applied, rapidly executed and inexpensive instrument, it could be useful in screening such patients for alcohol dependence.