Objective: To examine gender disparities in the diagnosis of bipolar disorder (BD) within a privately insured population in the United States and investigate potential contributing factors for these gender differences. Methods: This retrospective cohort study utilized 2005–2017 claims data from the MarketScan® Commercial Claims and Encounters database. The study cohort included subjects, aged 10–64 years, who had a minimum of 1-year continuous insurance coverage and no record of a BD diagnosis before cohort entry. We examined the gender difference in BD diagnosis rate, overall and by subgroups. We then used Cox regression models to evaluate the gender effect on time to first BD diagnosis, and the potential moderators of gender effect. Results: The study cohort consisted of 97,193,443 subjects; 0.45% of subjects were diagnosed with BDs after cohort entry with males having a lower diagnosis rate than females (0.36% vs. 0.54%). The Cox regression analysis indicated that males were less likely to be diagnosed with BDs (unadjusted Hazard Ratio, HR [95% CI]: 0.69 [0.68–0.69]) and gender difference remained significant after adjusting for demographics, comorbidity and healthcare utilizations (adjusted HR [95% CI]: 0.77 [0.76–0.77]). Gender disparity was consistently strong among most age groups, but varied in other demographic subgroups. Conclusions: Even though the prevalence of BDs is approximately equal between genders in the general population, our study found a much lower diagnosis rate in men compared to women for a privately insured U.S. population. Future studies aimed at identifying and understanding the barriers to diagnosis of BDs in men are warranted.

Original languageEnglish (US)
Pages (from-to)48-58
Number of pages11
JournalBipolar Disorders
Issue number1
StatePublished - Feb 2022

All Science Journal Classification (ASJC) codes

  • Psychiatry and Mental health
  • Biological Psychiatry


Dive into the research topics of 'Gender disparity in bipolar disorder diagnosis in the United States: A retrospective analysis of the 2005–2017 MarketScan Commercial Claims database'. Together they form a unique fingerprint.

Cite this