Composite outcome measures in alcoholism treatment research: problems and potentialities.

Alcoholism treatment researchers have increasingly incorporated multiple measures representing multiple dimensions in assessing treatment outcomes. However, no satisfactory model exists for examining the dynamic interrelationships of multiple measures in determining clinically meaningful and interpretable outcomes. One approach to analyzing multiple outcomes is to combine them into a composite measure. In their most elementary form, composite measures combine alcohol consumption and consumption-related problems dimensions into treatment outcome classifications. This paper discusses conceptual and methodological issues in studies reporting composite outcomes and provides a model and recommendations for using composite measures in alcohol treatment research.