The relative performance of diverse measures of alcohol abuse and dependence in a community sample

There is broad diversity among the ways in which alcohol abuse and/or dependence are defined and operationalized by researchers. In a community survey of 1,257 adult males and females, eight different measures of alcohol abuse or dependence were used. For analysis, each measure was dichotomized at its usual cutoff point. A factor analysis of distance scores between pairs of measures gave only one interpretable factor representing drinking patterns and social problems. Sensitivity and specificity were calculated for each measure with respect to abuse or dependence, as measured by the DIS-III-R. High maximum levels of drinking and social problems were moderately sensitive and specific to DIS-III-R. These relationships varied, though, for subgroups of age and gender. DSM-III-R alcohol abuse and dependence are widely respected variables but relatively expensive to measure. Alternative approaches for various survey purposes are discussed. Measures to be regarded as alternatives to the DIS-III-R are found to perform better in that role in samples limited by age and gender.