A three-sample test of a masked alcohol screening questionnaire.

The study tested an expanded version of a masked alcohol screening instrument developed by Wallace and Haines [Use of a questionnaire in general practice to increase recognition of patients with excessive alcohol consumption. British Medical Journal 290, 1949-1953 (1985)]. The alcohol subscales include the CAGE, consumption, trauma, medical advice, past problems, and present problems with alcohol. There is preliminary evidence of the construct validity and reliability of the Health Screening Survey (HSS). As expected the HSS was sensitive in known populations of alcoholics (0.96; 0.95) correctly identifying at least 95% of the alcoholic patients recruited from two treatment centers. Specificity was adequate with identified nonalcoholics (0.80; 0.70). It was also sensitive in a community primary care sample, correctly classifying 78% of the subjects meeting DSM-III criteria for alcohol abuse and/or dependence. There were gender differences in the validity of the subscales in the community sample with alcoholic males more likely to report problems on the questionnaire than alcoholic females. Based on results, the HSS was revised to improve validity.