Objective: Atypical antipsychotics are playing an increasing role in the treatment of bipolar disorder. The objective of this study was to assess the medication treatment patterns and costs associated with different atypical antipsychotics. Methods: PharMetrics Integrated Database for medical and pharmacy claims was used to assess medication patterns and healthcare costs associated with atypical antipsychotics in the treatment of bipolar disorder. Patients who initiated on olanzapine, risperidone, quetiapine or ziprasidone as monotherapy or in combination with other bipolar medications between 01/2003 and 01/2004 were followed for 1 year. Pair-wise group comparisons were made between olanzapine and other atypical antipsychotics using Wilcoxon without adjustment, log linear regression model with adjustment, and propensity score-adjusted bootstrapping methods. Results:Among 1516 patients with bipolar disorder, olanzapine (n = 507, 51 %) was significantly (p < 0.01) more likely to be initiated as the mono-bipolar than risperidone (n = 424, 40%), quetiapine (n = 463, 36%) or ziprasidone (n = 122, 25%). Post-initiation, olanzapine was used as the mono-bipolar medication for significantly (p < 0.01) more days (73.4) than risperidone (52.9), quetiapine (56.2) and ziprasidone (36.6). Annual healthcare costs incurred by patients with bipolar disorder varied from $14 216 for risperidone, $15 208 for olanzapine, $18 087 for quetiapine to $18 729 for ziprasidone treatments. No statistically significant differences in the annual healthcare costs were observed between olanzapine and risperidone treatments. Statistically significant differences between olanzapine and quetiapine were observed in two of the three models compared (p < 0.01, Wilcoxon; p = 0.024, log linear; p = 0.390, propensity score-adjusted bootstrapping) and between olanzapine and ziprasidone in one of the three models (p < 0.01, Wilcoxon; p = 0.068, log linear; p = 0.394, propensity score-adjusted bootstrapping). Limitations: Those arising from the data source and nature of retrospective assessments. Potential bias may also exist due to the presence of confounding factors and unobserved conditions and characteristics. As such, results of this study need to be considered in the context of its limitations and generalizability should be reserved to similar patient populations. Conclusions: Antipsychotic medication use patterns were statistically significantly different among atypical antipsychotics in the usual treatment of bipolar disorder. Olanzapine appears to be more likely used as a mono-bipolar medication compared with risperidone, quetiapine, and ziprasidone. The annual healthcare costs associated with the treatment of bipolar disorder by olanzapine and risperidone were similar, and the costs of these treatments were lower than with quetiapine or ziprasidone.
