The actual relapse rates were 24.0%, 15.8%, and 7.8% in the 4-week, 26-week, and no-dose-reduction groups, respectively. Except for the favorable relapse rate in the no-dose-reduction group, the other two are in the range of the rates for schizophrenia patients receiving risperidone (12%–25%) found in earlier studies (11, 40, 41). In this study, our hypothesis was supported by the finding that the no-dose-reduction group was associated with the lowest relapse risk, followed by the 26-week and 4-week groups, and there was also a significantly longer estimated time to relapse for the no-dose-reduction group than for the other two groups. This advantage for the no-dose-reduction group in the estimated time to relapse remained even after the covariates were controlled, supporting the notion that the maintenance dose for risperidone should be the dose found to be effective in the acute phase of treatment (4). We cannot compare these results with earlier findings because we know of no similar published study.