Abstract
The generalizability of previously isolated prototypical profiles of the Brief Psychiatric Rating Scale (BPRS) was examined in a sample of homeless individuals with both severe mental illness and substance-use problems who were part of a 24-month study that evaluated the effectiveness of various treatment interventions. These prototypical profiles (depressed, actively psychotic, and withdrawn) did generalize to the new sample, with a 59.4% coverage rate. In addition, some of the participants' BPRS profiles (10%) were characterized by negative correlations with the withdrawn profile (termed agitated) and others (17%) by minimal within-profile variability (labelled flat). Overall, with these additions, the coverage of the prototypical profiles was 86.4%. These prototypical profiles were then used to evaluate changes in profile elevation and shape over the course of the study. Generally, changes in both profile elevation and shape were moderated by the particular prototypical profile that the participants resembled. The use of these prototypical profiles in evaluating change permits a more precise analysis of what kind of individuals manifest particular effects. The clinical meaning of the BPRS profile changes observed was also discussed.