OBJECTIVE: This study compared treatment patterns and costs for bipolar disorder (BP) patients (recognized and unrecognized) to those of major depression disorder (MDD) patients without a BP claim (non-BP) during the observational period. METHODS: An employer claims database (1998-2001) was used to identify 9009 patients (aged 18-65) diagnosed with MDD and initially treated with antidepressants (AD). A subset of patients was identified as bipolar based on a BP diagnosis claim and/or a mood stabilizer claim. Of these BP patients, unrecognized BP (UBP) patients received their initial BP diagnosis and/or mood stabilizer (MS) prescription after AD initiation, while recognized BP (RBP) patients had these records on/before AD initiation. Medical costs are total payments to providers; indirect costs are payments to employees for lost work time. RESULTS: BP patients accounted for 7.0% of the research sample (3.7% UBP, 3.3% RBP). UBP patients incurred significantly more monthly medical costs in the 12 months following initiation of AD treatment than RBP patients, for both non-BP treatment costs ($1081 versus $683) and total medical costs ($1179 versus $802). MDD total medical costs ($585) were significantly lower than both RBP and UBP costs. In the 6 months before the Index Date, UBP and RBP patient monthly costs for medical care other than BP treatment were similar ($542 versus $504, respectively). MDD total medical costs ($436) were significantly lower than RBP total costs ($631), but not UBP total costs ($542). Monthly indirect costs are significantly greater for UBP and RBP employees compared to MDD employees in the 12 months following the Index Date ($570, $514, and $335, respectively). CONCLUSIONS: Accurate and timely recognition of BP disease is associated with lower overall medical costs and lower indirect work loss costs. More effort is needed to quickly diagnose and effectively treat patients with bipolar disorder.