Statistical and consensus-based strategies for grouping consumers in mental health level-of-care schemes.

Mental health Level-of-Care schemes differ in the emphasis placed on statistically based versus consensus-based strategies for delineating consumer groups. Statistical strategies define groups on the basis of empirical distinctions within the caseload. Consensus strategies group consumers according to an a priori framework of consumer characteristics determined by experts to differentiate the appropriate types and intensities of service. This article compares the performance of two grouping strategies-one statistical and one consensus-on three criteria: programmatic relevance, measurement adequacy, and homogeneity of consumer groups. It explores the advantages of integrating elements of each strategy.