A Retrospective Cross-Sectional Analysis of the Humanistic and Economic Burden of Bipolar I Disorder

Purpose: This study quantified the burdens of bipolar I disorder (BP-I) by examining patient characteristics, health-related quality of life (HRQoL), health care resource utilization (HCRU), and costs of patients with versus without BP-I. Additionally, these outcomes were assessed across BP-I severity levels.

Methods: A retrospective, cross-sectional analysis of the 2020 National Health and Wellness Survey was conducted. Adults who self-reported a physician diagnosis of BP-I were assigned to the BP-I cohort, with severity-specific subgroups (mild, moderate, severe) created for analysis. A separate cohort of participants without BP-I or MDD was used for comparison. Exclusion criteria included a schizophrenia diagnosis. Bivariate analyses compared demographic and socioeconomic characteristics between cohorts. HRQoL (Short Form-36v2 Health Survey [SF36v2] mental and physical component scores, EuroQol Five-Dimension Visual Analogue Scale [EQ-5D VAS]), HCRU (health care provider visits, emergency department visits, hospitalizations), and annualized costs (direct and indirect) were evaluated for participants with versus without BP-I as well as across BP-I severity subgroups using multivariate analyses adjusted for key baseline differences. Because BP-I is often misdiagnosed as MDD, outcomes were evaluated in a subgroup of participants with MDD who according to the Mood Disorder Questionnaire screened as having probable BP-I (ie, potentially misdiagnosed BP-I) and were compared with the BP-I severity subgroups.

Findings: Cohorts included 818 participants with BP-I (mild = 336, moderate = 285, severe = 197) and 53,021 participants without BP-I. Participants with BP-I reported significantly lower HRQoL scores on the SF-36v2 and EQ-5D VAS (all measures, P < 0.001), and increasing BP-I severity was predictive of declining HRQoL. Participants with BP-I had significantly greater HCRU (all measures, P < 0.05) than participants without BP-I and increasing BP-I severity was associated with greater HCRU versus the mild BP-I cohort (all measures, P < 0.05). Participants with BP-I incurred significantly greater total direct (P < 0.01) and indirect (P < 0.001) costs versus participants without BP-I. Direct costs were incrementally higher across BP-I severity, while indirect costs were high across all groups but did not differ significantly. Participants with potentially misdiagnosed BP-I (n = 302) had similar HRQoL to those with mild-to-moderate BP-I and similar HCRU and direct costs to those with mild BP-I.

Implications: These results demonstrate the substantial clinical and economic burdens associated with BP-I, and these negative impacts generally increase with BP-I severity. The study also suggests that despite not having the diagnosis of BP-I, burdens of potentially misdiagnosed patients are similar to those with mild-to-moderate BP-I. Together, these results reveal substantial and diverse unmet needs among adults with BP-I.