Prior research in individual therapy has provided evidence that therapists are poor predictors of client outcome and often misjudge clients' perceptions of the therapeutic relationship. The focus of the current research was to conduct a similar predictive study in a group setting. Group therapists were recruited from a university counseling center and a state psychiatric hospital; 64 group members and 10 group leaders participated in the study. We tested therapist accuracy in predicting client outcome and perceived quality of therapeutic relationship. Results suggested that therapists underestimate the number of clients who deteriorate during therapy and were unable to accurately predict client perceptions of the group relationship, replicating findings from larger samples in the individual literature. Results suggest that using outcome and process measures as feedback tools may be also useful for clinicians leading groups.

The purpose of this article is to describe a procedure to assist in selecting outcome measures for inpatients treated at a state psychiatric hospital. The procedure combines evidence-based criteria from the literature, instruments shown to be sensitive to change in clinical trials, and the perspectives of a multidisciplinary team of researchers, administrators, providers, and patient advocates. Recent efficacy and effectiveness studies were used to identify recurrently used outcome instruments. A computerized search of more than 30 bibliographic databases, such as PsycINFO, MEDLINE, Social SciSearch, and ERIC, was conducted for articles published between 1990 and 2002. Comparisons of the most frequently used instruments were made on seven criteria proposed as best-practice indicators, including sensitivity to change and robust psychometrics. The sample produced 110 measures. Rater-completed instruments were represented more often than patient-completed ones. However, considerable variability across both methods was found on the criteria. The limited resources associated with publicly funded inpatient facilities led to a recommendation to select at least one rater-completed and one patient-completed instrument.

The Brief Psychiatric Rating Scale (BPRS) is the most commonly used outcome measure for the severely and persistently mentally ill (SPMI) population, possessing good interrater reliability, concurrent validity, and a strong factor structure. However, psychometric study of the extended version of the BPRS (the BPRS-E) is limited when compared with earlier versions (BPRS and BPRS-A). This study examined the item, factor, and diagnosis-specific sensitivity to change of the BPRS-E, the most recent version of this popular scale. Assessments were conducted at 90-day intervals with 201 adult psychiatric inpatients at the Utah State Hospital, yielding 786 symptom ratings. Of note was that ratings were conducted by independent assessors who were unaware of patients' treatment status. All but 2 of the 24 BPRS-E items, all 4 factors, and the total score were found to be sensitive to change when comparing patients' admission and discharge scores. Patient diagnosis was not associated with item, factor, or total score sensitivity to change. These findings extend the psychometric support for the BPRS-E and have implications for assessing outcome with the SPMI population.

The SR can be used to estimate work productivity using tools like the Productivity Calculator. Specific relationships among measures and suggestions for future research are discussed.

No abstract