Predicting outcome for mental hospital patients: who, what to ask.

Outcomes for 62 hospitalized mental patients were predicted, utilizing various background information, diagnoses, and staff ratings of behavior, as independent variables. Diagnoses and previous admissions correlated weakly with the outcomes, viz., length of stay, subsequent readmissions, and post-hospital adjustment. Staff ratings were more reliable, and of these, the open ward aides' estimates generally were more effective than those of admission ward personnel or professionals. Staff predictions and judgments of patients' behaviors yielded multiple correlations explaining over 60 per cent of the variation in each outcome measure. In addition, the patient's judged competence and posthospital living situation were associated with his or her use of outpatient facilities.