Searching for treatment outcome measures for use across trials

Objective: Inconsistencies in outcome measures across studies of treatment efficacy have made comparisons among them difficult. As a result, there is interest in identifying one or more measures that might be recommended as universal indicators of outcome. The present article seeks to identify drinking, psychosocial and/or biological variables that could be candidates for use as universal indicators of change following alcoholism treatment. Method: The primary data set included 128 alcohol-dependent men and women who were randomly assigned to cognitive-behavioral or interactional group treatment for 26 weekly sessions. Results: The greatest changes following treatment were seen in measures of drinking and drinking consequences. Correlational analyses indicated that changes from baseline in drinking consequences were significantly associated with changes in drinking. Psychosocial and biological indicators showed much smaller changes from baseline, and these were only weakly associated with changes in drinking, indicating that they are not sensitive measures of treatment-related change. The overall pattern of these findings was replicated in the Project MATCH data set. Conclusions: It was concluded that drinking frequency and intensity measures, as well as a measure of drinking consequences, may be useful as universal indicators of alcohol treatment outcome, but that the other psychosocial and biological measures studied in these two data sets are not strong candidates for this purpose.