OBJECTIVES: This retrospective claims-based study compared atypical antipsychotics to typical antipsychotics and to each other with respect to risk of hospitalization and inpatient costs among Ohio Medicaid patients with schizophrenia. METHODS: Relative risks of hospitalization for mental illness between enrollees with diagnosed schizophrenia treated with any of the atypical antipsychotics (risperidone, olanzapine, quetiapine, ziprasidone) or any of the leading traditional antipsychotics (haloperidol, perphenazine, thioridazine, thiothixene) were assessed. Cox proportional hazard regression controlled for age, gender, diagnosis, prior hospitalization and ER use, substance dependence/abuse, antipsychotic dose, use of other psychotropics,and other health needs. Differences in length of stay and inpatient costs (charges) were also assessed. RESULTS: Patients treated with quetiapine had a 33% lower risk of hospitalization for mental illness compared with patients treated with typical antipsychotics (HR 0.672; P=0.0413), equating to $389 lower expected annual inpatient charges per patient for quetiapine. Other pair-wise comparisons of antipsychotic-related hospitalization were not statistically significant. Risperidone was associated with significantly longer hospital stays than the typicals (1.78, P=0.0301), resulting in $303 higher expected annual inpatient charges per patient. Risperidone also had significantly longer lengths of stay than olanzapine, 1.88 days (P=0.006), resulting in higher inpatient charges ($320 annual per patient). Other differences in length of stay were not significant at P<0.05. CONCLUSIONS: Quetiapine was associated with a lower risk of hospitalization for mental illness compared with typical antipsychotics, resulting in substantially lower inpatient charges. While risperidone may also have a lower risk of hospitalization than the typicals (22% lower, ns P=0.0932), length of stay was significantly longer, contributing to higher inpatient charges. Risperidone also had significantly longer lengths of stay than olanzapine with associated higher inpatient costs, and may also have longer stays than ziprasidone (3.61 days, ns P=0.0619).