This research proposes to examine the relative effectiveness of drug abuse treatment types for various client subgroups. Treatment types will be jointly defined by modality (outpatient methadone, outpatient drug free, residential), length of treatment and other treatment characteristics such as the range and intensity of specific treatment services. Client subgroups will be defined primarily by drug use (e.g., opioid or nonopioid users) and secondarily by treatment level of criminal involvement, number of drug related problems, psychological health, age, and sex) that might be expected to moderate the relationship between treatment types and treatment outcomes. Treatment outcomes will include measures of drug use, depression, and involvement in illegal activities using one-year followup data from the Treatment Outcome Prospective Study (TOPS). The TOPS data base is a comprehensive data set that represents both a large number of treatment types and client subgroups. Client subgroups will be defined and principal components analysis will be used to understand the structure and dimensionally within each client subgroup. The problem of client selectivity across and within modality will be addressed by discriminant function analysis or multi-logit analysis. The client characteristics that discriminate among treatment types for a given client subgroup will be used as control variables in subsequent multivariate regression models of treatment effectiveness. Multivariate regression models will be developed for each client subgroup. The primary emphasis will be estimating the effects of treatment types on outcomes for each of the client subgroups. The overall goals is to determine if certain treatment types are more effective than others for particular client subgroups.
