AIM: To determine the rates of psychiatric hospitalisation and to evaluate the predictors of readmissions in a cohort of 887 schizophrenic patients at their "first" psychiatric admission. METHODS: Data were collected using the hospital discharge database of the Lazio region, Italy. The cohort included patients admitted to a psychiatric ward in 1999, with a main diagnosis of schizophrenia and no prior psychiatric hospitalisation during the 5 years preceding the index admission. Psychiatric readmissions were considered up to year 2004. Kaplan-Meier survival curves and logistic procedures were performed to estimate the cumulative readmission incidence and ORs of readmissions for potential confounders. RESULTS: During the 4-year follow-up, 44.3% of the patients were readmitted at least once. An active treatment among community psychiatric facilities at the time of index admission predicted a higher readmission risk. However, for those who experienced their first admission in a private ward readmission risk was consistently higher, whether or not having had such a treatment. CONCLUSIONS: Risk of readmission in a psychiatric ward appears to be related to the level of integration with mental health community facilities, while length of stay is strongly affected by service system variables. Possible uses of medical information systems in this field are described.
AIM: To determine the rates of psychiatric hospitalisation and to evaluate the predictors of readmissions in a cohort of 887 schizophrenicpatients at their "first" psychiatric admission. METHODS: Data were collected using the hospital discharge database of the Lazio region, Italy. The cohort included patients admitted to a psychiatric ward in 1999, with a main diagnosis of schizophrenia and no prior psychiatric hospitalisation during the 5 years preceding the index admission. Psychiatric readmissions were considered up to year 2004. Kaplan-Meier survival curves and logistic procedures were performed to estimate the cumulative readmission incidence and ORs of readmissions for potential confounders. RESULTS: During the 4-year follow-up, 44.3% of the patients were readmitted at least once. An active treatment among community psychiatric facilities at the time of index admission predicted a higher readmission risk. However, for those who experienced their first admission in a private ward readmission risk was consistently higher, whether or not having had such a treatment. CONCLUSIONS: Risk of readmission in a psychiatric ward appears to be related to the level of integration with mental health community facilities, while length of stay is strongly affected by service system variables. Possible uses of medical information systems in this field are described.