Screening for Alcohol Use Disorders in the Nursing Home

Abstract

To determine the prevalence of Alcohol Use Disorders (AUDs) among residents of a Veterans Affairs (VA) nursing home (NH) using DSM-III-R criteria for alcohol abuse and dependence, and to examine the demographic variables associated with AUDs among NH residents. A third objective was to assess the sensitivity, compared with DSM-III-R criteria, of three screening tests for AUDs in the NH: the CAGE, the MAST-G, and the two-question instrument developed by Cyr and Wartman. A cross-sectional design, with DSM-III-R criteria determined by the alcohol module of the Diagnostic Interview Schedule (DIS) as the criterion standard. Patients older than age 50 admitted consecutively to a VA NH, n = 117. Sensitivities, specificities, positive predictive values for the CAGE, MAST-G, and Cyr and Wartman Screening questionnaires; receiver operating characteristic (ROC) curves for the CAGE and MAST-G. Forty-nine percent of study participants met DSM-III-R criteria for lifetime alcohol abuse or dependence (18% active, 31% inactive). The sensitivities and specificities of the three screening questionnaires were as follows: CAGE-82% and 90%; MAST-G-93% and 65%; Cyr and Wartman-70% and 92% respectively. The area under the ROC curve was 0.94 for the CAGE and 0.90 for the MAST-G. The prevalence of lifetime alcohol abuse and dependence was high in this VA NH population. Both the CAGE and MAST-G are sensitive to AUDs in this setting. The areas under the ROC curves were not significantly different and indicate both tests discriminated well between NH residents with and without AUDs.