Predictors for readmission risk in schizophrenia.

Readmission risk was assessed at the first and subsequent discharges in a total Danish national sample consisting of 8705 first admitted patients who had been discharged alive at least once with a diagnosis of schizophrenia. Predictors for readmission risk were identified using the Cox proportional hazards model. Following the first discharge, 19% of the surviving patients had not been readmitted after 10 years of follow-up. Readmission risk increased with the number of previous admissions. At the first discharge readmission risk decreased with increasing age and was significantly predicted by clinical subtype and gender. At later discharges (5th, 10th, and 15th) the effect of these variables gradually disappeared. At the 15th discharge readmissions were mainly predicted by the duration of the latest admission and discharge periods. Both the increase in readmission risk with the number of previous admissions and the evolving pattern of predictors for readmission risk are interpreted as supporting the existence of a smaller subpopulation among schizophrenic patients with frequent relapses.