Midanik L T, Zahnd E G, Klein D
University of California at Berkeley, School of Social Welfare, 94720-7400, USA.
Alcohol Clin Exp Res. 1998 Feb;22(1):121-5.
The purpose of this study is to evaluate a revised CAGE and a new drug version of the CAGE as screeners for risk of heavier or problem alcohol and/or drug use among a population of low-income, pregnant women and adolescents (n = 1147) recruited from 19 agencies in two California counties. Two versions of the CAGE were used in this analysis: (1) the 4-item Alcohol CAGE, using the year before knowing about pregnancy as the timeframe (as opposed to lifetime prevalence), and (2) a newly developed 4-item Drug CAGE using the same timeframe. The two instruments were assessed on sensitivity, specificity, and area under the receiver operating characteristic curve by using self-reported periconceptional heavier or problem alcohol and drug use as the criteria for the 12 months before knowing about the pregnancy. The results indicate that, for the Alcohol CAGE, the cut-point of 1 yielded the highest sensitivity, while maximizing sensitivity and specificity (receiver operating characteristic analysis) for this sample. For heavier drug use, the Drug CAGE had a high sensitivity rate with a cut-point of 3. The Drug CAGE was not a useful screener for periconceptional lighter drug and marijuana use. These results demonstrate the utility of a revised version of the Alcohol CAGE that incorporates a more specific timeframe (the year before pregnancy) to screen pregnant, low-income women for at-risk heavier or problem alcohol use. The Drug CAGE, which uses the same timeframe, seems to be an effective tool for identifying pregnant, low-income women at risk for heavier drug use only. More psychometric works needs to be done to refine the Drug CAGE for detecting at-risk use of lighter drugs and marijuana.
