Abstract
In the Drug Abuse Reporting Program (DARP) research on evaluation of treatment effectiveness, a major effort has been devoted to patient classification using Admission Report data relating to the pretreatment background and status of patients. In two previous evaluation studies based on DARP Cohort 1 samples [1,2], the concept of homogeneous grouping of patients proved to be informative and useful in predicting differential outcomes in treatment. Due to the addition of new agencies and revisions in admission policies in those agencies continuing in the program during Cohort 2, however, there was a question regarding whether or not the same strategy should be employed as for Cohort 1. More specifically, it was a question of whether the patient types developed for the earlier cohort were still fully applicable.