OBJECTIVES: To estimate clinical outcomes and associated cost of care of treatment with risperidone versus olanzapine in patients with chronic schizophrenia or schizoaffective disorders up to one year following therapy initiation. METHODS: A Markov model was developed to estimate the number of patients who experience side effects (i.e., extrapyramidal symptoms [EPS], prolactin-related disorders, weight gain, and diabetes) of antipsychotic therapies, relapse of psychiatric symptoms as well as discontinuation of antipsychotic therapy following these events at one year; associated costs of care were also calculated. Parameter estimates were based on findings from a randomized, controlled, clinical trial of risperidone and olanzapine and other published and unpublished sources. Analyses were undertaken using second-order Monte Carlo simulation techniques with 10,000 individual trials. RESULTS: The expected number of patients remaining on initial therapy at one year was higher for risperidone (76.3% versus 44.7% for olanzapine); the expected number of months on therapy was lower for olanzapine (8.0 vs 10.5 for risperidone). Therapy discontinuation was primarily driven by patients experiencing increases in body weight exceeding 5 kg since therapy initiation. Expected mean total costs per month on therapy were 8% higher for olanzapine ($2,198 vs $2,033 for risperidone). CONCLUSIONS: Therapy discontinuation at one year was lower for risperidone than for olanzapine. Expected costs of care per month of therapy were also lower.