OBJECTIVES: The purpose of this study is to profile the drug treatment patterns of bipolar patients and compare the costs associated with patients in alternative treatment groups. METHODS: The study sample was drawn from a de-identified administrative claims database of approximately 1.8 million beneficiaries from 1999 to 2002, and including medical, drug, and disability claims. Patients aged 18 to 65 years who had at least 2 bipolar diagnoses, or 1 bipolar diagnosis and 1 prescription for a mood stabilizer, were included (N = 3499). Costs were measured from an employer's perspective. All costs were adjusted to 2002 dollars using Consumer Price Indices. T-tests were conducted to estimate statistical significance. RESULTS: Forty-five percent of patients in the sample received no specific drug treatment for bipolar disorder in the first 2 months following their diagnose, but 62% of this "no-treatment" group received antidepressants. Fifty percent of the sample took mood stabilizers in the first 2 months. Conventional antipsychotics were rarely used, whereas atypical antipsychotics were usually taken concomitantly with mood stabilizers. After the first observed bipolar episode, patients on mood stabilizer monotherapy incurred increases in medical costs (36%) and drug costs (59%), and a slight decrease in work-loss costs (-4.8%); patients on atypical antipsychotics had a decrease in medical costs (16.8%), an increase in drug costs (66%), and a decrease in work-loss costs ( 13%). On average, therapy with atypical antipsychotics could have saved $4796 annually, reflecting $4640 less in medical services, $158 more in drug costs, and $314 less in indirect work-loss costs compared with mood stabilizer monotherapy. CONCLUSIONS: Bipolar patients were largely untreated or treated inappropriately (with antidepressant monotherapy). Appropriate combination therapy with atypical antipsychotics and a mood stabilizer may reduce both direct health care costs and indirect work-loss costs.