Relationship between social determinants of health and hospitalizations and costs among patients with bipolar disorder 1

BACKGROUND: Bipolar disorder type 1 (BD -1) is a serious episodic mental illness whose severity can be impacted by social determinants of health (SDOH). To date the relationship of social and economic factors with health care utilization has not been formally analyzed using real -world data. OBJECTIVE: To describe patient characteristics and assess the influence of SDOH on hospitalizations and costs in patients with BD -1 insured with commercial and managed Medicaid health plans. METHODS: This retrospective observational study used data from the Medical Outcomes Research for Effectiveness and Economics (MORE2) Registry to identify patients aged 18 years and older with evidence of BD-1 between July 1, 2016, and December 31, 2018. SDOH were linked to patients at the "near neighborhood" level (based on ZIP9 area). Multivariable models assessed the relationship between patient characteristics and hospitalizations (incidence rate ratios [95% CI]) and costs (cost ratios [95% CI]). RESULTS: Of 243,286 patients with BD-1, 62,148 were covered by commercial insurance and 181,138 by Medicaid. Mean ages [+/- SD] were similar (commercial 39.8 [+/- 14.8]; Medicaid 40.1 [+/- 13.6]), with more female patients in both cohorts (commercial 59.8%; Medicaid 65.4%). All-cause hospitalization rates were 21.6% for commercial and 35.1% for Medicaid patients; emergency department visits were 39.7% and 64.3%, respectively. All-cause costs were $15,379 [+/-$27,929] for commercial and $21,474 [+/-$37,600] for Medicaid. Older age was a significant predictor of fewer hospitalizations compared with those aged younger than 30 years, particularly ages 40-49 for both commercial (0.60 [0.57-0.64]) and Medicaid (0.82 [0.80-0.85]). Increasing age was associated with significantly higher costs, especially age 65 and older (commercial 1.37 [1.31-1.44]); (Medicaid 1.43 [1.38-1.49]). Initial treatment with antipsychotics plus antianxiety medications was a significant predictor of higher hospitalizations (commercial 2.12 [1.98-2.27]; Medicaid 1.62 [1.57-1.68]) and higher costs (commercial 1.86 [1.80-1.92]); Medicaid 1.80 [1.76-1.84]). Household income was inversely associated with hospitalizations for Medicaid (<$30,000 [1.16 (1.12-1.19)]; $30,000-$39,999 [1.11 (1.07-1.15)]; $40,000-$49,999 [1.08 (1.05-1.12)]; $50,000-$74,999 [1.06 (1.02-1.09)]). Not speaking English well or at all was associated with 90% higher hospitalizations for commercial patients (1.93 [1.36-2.76]) but 40% fewer hospitalizations for Medicaid patients (0.59 [0.53-0.67]). Low English language proficiency was associated with significantly higher costs for commercial patients (2.22 [1.86-2.64]) but lower costs for Medicaid patients (0.57 [0.53-0.61]). CONCLUSIONS: Medicaid patients with BD -1 had high SDOH burden, hospitalizations, and costs. The association of lower English proficiency with fewer hospitalizations and lower costs in Medicaid patients suggests a potential disparity in access to care. These findings highlight the importance of addressing social risk factors to advance health equity in treatment of mental illness.