Accuracy of clinicians' expectancies for psychiatric rehospitalization

Rehospitalizations of 269 patients in a state community mental health center were compared to therapists' expectations at discharge for each patient's readmission within the next 2 years. Analysis of the residual variance (predicted minus observed rehospitalization) indicated that clinicians' prognostic judgments were biased in regard to patients' ethnicity: black patients were considered more likely to be rehospitalized than whites, although the opposite occurred. No evidence of gender-related bias was found. Clinicians' expectancies were influenced unduly by their perceptions of patients' severity of illness and cooperativeness. They apparently ignored the prognostic values of favorable factors such as lack of prior hospitalizations, youthfulness, lack of severe impairment, and living in a middle-class or residentially stable neighborhood. Therapists expected rehospitalization for two-thirds of their patients, but less than half actually returned. Predictors of rehospitalization were prior hospitalizations, age, and instability of the postdischarge neighborhood.