OBJECTIVES: To examine treatment patterns and analyze differences in health services utilization and costs for subjects receiving risperidone, olanzapine or quetiapine. METHODS: This study used administrative claims data to identify continuously enrolled subjects prescribed atypical antipsychotics between January 1, 2000 and December 31, 2000. Subjects were assigned to a diagnostic category based on the appearance of two or more ICD-9 codes for schizophrenia, affective disorder, dementia, anxiety or childhood disorders during the study period. Duration of therapy, compliance, daily dose, daily average consumption and concomitant medication use were analyzed descriptively. Subjects were propensity score matched on baseline characteristics for the purpose of comparing health services utilization and cost by site of service. RESULTS: Of 6471 study subjects, average annual days of therapy were nearly equal between drug groups (184-186 days). However, average annual days of therapy varied widely by diagnostic condition (181 days for anxiety, 270 days for schizophrenia). Concomitant use of psychotropic medication was common for all 3 drugs (81%). Subjects receiving risperidone had an average daily dose below the recommended target dose for schizophrenia as did olanzapine subjects with bipolar disorder. Daily average consumption increased slightly for risperidone and quetiapine subjects over the follow-up period. After matching, there was no significant difference in health services utilization between drug groups. However, risperidone subjects had significantly lower pharmacy costs (<.0001) and total costs (p=0.0181) compared to olanzapine subjects. Subjects with affective disorders had total costs that were significantly higher compared to subjects with schizophrenia or childhood disorders. CONCLUSIONS: Studies comparing cost and utilization among atypical antipsychotics should consider the host of factors that may influence receipt or regiment of care such as diagnostic condition, duration of therapy and dosing.