0554 Racial Disparities in Healthcare Resource Utilization among Older Adult Medicare Beneficiaries With Comorbid OSA and Depression

Abstract

Abstract Introduction Depression and obstructive sleep apnea (OSA) are highly comorbid and independently associated with increased economic burden. Evidence suggests important racial disparities in diagnosis and treatment of each condition. This study sought to evaluate the economic burden of occult OSA among Black and White older adults with depression. We hypothesized that the impact of occult OSA on healthcare resource utilization would differ by race. Methods Data were derived from a 5% sample of Medicare administrative claims data. Inclusion criteria were age >64 years, 12 months continuous enrollment prior to OSA diagnosis, depression, and Black or White race. Undiagnosed OSA and pre-existing depression were defined using ICD-10 codes during the 12 months prior OSA diagnosis. HCRU was defined as monthly counts of inpatient, outpatient, and emergency department (ED) visits in each of the 12 months before OSA diagnosis. The relationship between undiagnosed OSA and HCRU was modeled using negative binomial regression. Inverse probability of treatment (IPT) weighting was used to balance covariates between groups. Effect modification by race was evaluated using an interaction term; estimates were stratified by race. Results Of 39,192 beneficiaries with pre-existing depression, 4,353 suffered undiagnosed OSA (n=325 Black adults, 7.5%). Relative to White beneficiaries with undiagnosed OSA, Black beneficiaries with undiagnosed OSA were more likely to be female (76.6% vs. 68.6%, p=.002) and to suffer multimorbidity (>7 comorbid conditions [71.7% vs 52.7%, p<.001]). In IPT weighted models and relative to Black non-sleep disordered controls, Black beneficiaries with undiagnosed OSA had higher inpatient (RtR 1.89; 95% CI 1.33, 2.71), outpatient (RtR 1.38; 95% CI 1.04, 1.83) and ED (RtR 1.83; 95% CI 1.22, 2.75) utilization. Relative to White non-sleep disordered controls, White beneficiaries with undiagnosed OSA demonstrated higher inpatient (rate ratio (RtR) 1.51; 95% CI 1.38, 1.65)), outpatient (RtR 1.18; 95% CI 1.10, 1.26) and ED (RtR 1.45; 95% CI 1.34, 1.57) utilization. The interaction between race and undiagnosed OSA was significant in all models (p<.001). Conclusion Among Medicare beneficiaries with pre-existing depression, comorbid undiagnosed OSA is associated with increased HCRU across multiple points of service. The strength of this association is greater among Black (vs White) beneficiaries. Support (if any) ResMed Foundation