Understanding who improves after psychiatric hospitalization.

This study attempts to identify variables associated with improved functioning after psychiatric hospitalization. The relationship between 25 clinical and demographic variables and improvements and functioning at discharge of 309 consecutive patients in an open ward over 2 years were studied using stepwise logistic regression. Measures of patients functioning at discharge were score above or below 50 on the (Global Assessment of Functioning scale) (GAF) and 3-point scale of functioning 1 week after discharge (needs constant supervision, functions independently at home, functions on the job). We created a logistic regression model of outcome based on the information available at admission and one based also on information at time of discharge (i.e., medication, therapy and length of stay etc.). Discharge models predicted 6-10% better than admission models. Discharge models correctly predicted whether GAF level was above 50 in 72% of the cases and return to preadmission functioning in 71% of the cases. We also created a model of functioning at discharge using a hierarchical logistic regression model. It correctly predicted 74% of the lowest level of functioning, 47% of the middle level and 68% of the highest level. The best predictors of improved functioning after discharge were preadmission levels of functioning, psychotherapy in hospital, absence of organic brain syndrome, attendance at occupational therapy and hospitalization less than 4 months. Diagnosis had almost no discernible effect on outcome of hospitalization.