OBJECTIVES: To examine time to re-hospitalisation and re-hospitalisation rates of patients who had been discharged from psychiatric hospitals after starting treatment with either risperidone or olanzapine. METHODS: A multi-centre retrospective cohort study was carried out in 9 centres in 3 countries. Re-hospitalisation status was monitored in 393 schizophrenia patients previously discharged on one of the two study drugs. Data were collected using a pre-defined; standardised data collection template. The proportion of re-hospitalised patients were compared between risperidone and olanzapine using the Cochran-Mantel-Haenszel test stratified by centre. Time to re-admission was assessed and compared between the two groups of patients using Kaplan-Meier curves, log-rank test and Cox proportional hazard models. RESULTS: Median follow up was 1282 days in the risperidone group and 1207 days in the olanzapine group, ranging from 93 to 2985 days across both groups. The proportion of re-hospitalised patients was lower for patients discharged on risperidone (59%) than for patients discharged on olanzapine (66%) (p=0.089). Reason for re-hospitalisation was similar in both groups with 65% and 68% of all re-admissions in the risperidone and olanzapine group respectively due to relapse of schizophrenia, with 42% (192 cases) of all re-hospitalisations linked to non-compliance. The time to first re-admission was longer in the risperidone group (KM median estimate: 925 days; 95%CI: 660-1249 days) than in the olanzapine group (526 days; 95%CI: 433-865 days). The difference was statistically significant (p=0.028). The overall risk of a first re-admission was lower after discharge on risperidone than on olanzapine (Hazard ratio: 0.71; 95% CI: 0.55-0.93). CONCLUSIONS: Starting hospitalised acute schizophrenia patients on treatment with risperidone rather than on olanzapine appears to have a positive impact on time to next hospitalisation.