Abstract
Summary Within a comprehensive community care system, we examined which patients became hospitalized during long-term treatment. We studied 60 consecutively admitted, and mainly chronic psychotic patients over 5 years. Full and partial hospitalizations were assessed by means of a hospitalization index. This index was significantly lower in the 2nd to 5th years after admission than in the 1st year. Sociodemographic and clinical characteristics failed to predict hospitalizations during the 1st year. Patients' age, gender and last occupational status were related to hospitalizations in the following 4 years. A better prediction of hospitalizations was allowed by treatment data obtained during the 1st year. It is suggested that longterm prognosis should be based mainly upon experience with a patient in a given care system, rather than on a patient's history and characteristics as known before treatment.