Bipolar Disorder Treatment Patterns in the United States: A Retrospective Claims Database Study

Background: The objective of this retrospective study was to describe treatment patterns among patients with bipolar disorder.

Methods: Adults with newly diagnosed bipolar disorder from 2016-2018 were identified using MarketScan Commercial, Medicare Supplemental, and Medicaid healthcare claims databases. Patients were enrolled for 12 months pre- and  ≥6 months post-initial diagnosis. Lines of therapy (LOTs) were continuous treatment periods based on filled prescriptions, ending in one of the following: treatment discontinuation, new medication addition, different type of bipolar diagnosis, or end of enrollment/study period. Data were analyzed descriptively.

Results: There were 40,345 patients identified from the Commercial database. The most common initial episode types were bipolar II (38.1%), bipolar I depression (29.8%), bipolar I mania (12.8%), and bipolar I with mixed features (12.0%). Among all episode types, approximately 90% of patients received treatment (LOT1) and approximately 80% of these patients received  ≥1 additional LOT. There were 2,067 different regimens in LOT1 among 36,587 patients. Across all episode types, the most common medication classes in LOT1 were mood stabilizers (43.8%), antidepressants (42.3%), atypical antipsychotics (31.7%), and benzodiazepines (20.7%); with subsequent LOTs, antidepressant (51.4-53.8%) and benzodiazepine (26.9-27.4%) usage increased. Of patients with bipolar I depression as their first episode, 14.5% used monotherapy antidepressants in LOT1. Results were largely similar for Medicare-insured (n=2,208) and Medicaid-insured (n=49,324) patients.

Conclusions: Antidepressants and benzodiazepines were frequently prescribed despite current guidelines recommending against use or as frontline therapy. These results suggest many clinicians treating bipolar disorder are not using evidence-based prescribing practices, which can adversely affect health outcomes.