Occult, undiagnosed obstructive sleep apnea is associated with increased health care resource utilization and costs among older adults with comorbid depression: a retrospective cohort study among Medicare beneficiaries

Abstract

Obstructive sleep apnea and depression are highly comorbid among older adults, and each is associated with increased economic costs and health care resource utilization. The purpose of this study was to determine the economic burden of comorbid occult obstructive sleep apnea among a random sample of older adult Medicare beneficiaries in the United States. Among 41,500 participants with preexisting depression and meeting inclusion criteria, 4,573 (11%) had occult OSA. In fully adjusted models, beneficiaries with occult OSA were heavier users of inpatient (rate ratio: 1.53; 95% CI: 1.39, 1.67), outpatient (rate ratio: 1.18; 95% CI: 1.10, 1.27), emergency department (rate ratio: 1.48; 95% CI: 1.35, 1.63), and prescription (rate ratio: 1.09; 95% CI: 1.05, 1.14) services. Mean total costs were also significantly higher among beneficiaries with occult OSA ($44,390; 95% CI: $32,076, $56,703).

Citation:

Wickwire EM, Albrecht JS. Occult, undiagnosed obstructive sleep apnea is associated with increased health care resource utilization and costs among older adults with comorbid depression: a retrospective cohort study among Medicare beneficiaries. J Clin Sleep Med. 2024;20(5):817–819.

INTRODUCTION

By any standard, geriatric depression represents a major public health and economic burden in the United States and worldwide. Identifying factors that can reduce the burden of geriatric depression is an important research priority, with many gaps in knowledge. One modifiable factor that could potentially reduce the burden of geriatric depression is obstructive sleep apnea (OSA). High-quality longitudinal and meta-analytic data confirm the elevated prevalence of OSA in patients with depression.^1,2 Similarly, our prior work among Medicare beneficiaries has demonstrated that occult, undiagnosed OSA is associated with increased risk for incident depression during the 12 months prior to OSA diagnosis.³

Both depression and OSA are associated with a broad range of adverse health consequences among older adults, as well as dramatically increased health care resource utilization (HCRU) and costs.^4–6 However, to our knowledge, no prior study has examined the burden of comorbid OSA in older adults with pre-existing depression. To this end, the purpose of this study was to determine the economic burden of occult, undiagnosed OSA among a national sample of older adult Medicare beneficiaries with pre-existing depression. We hypothesized that, relative to older adults with depression alone, individuals with depression and comorbid occult, undiagnosed OSA would demonstrate greater HCRU and costs across multiple points of service.

METHODS

Participants

Data were obtained from a 5% random sample of Medicare administrative claims from 2006 to 2013. Inclusion criteria included age ≥ 65 years and ≥ 12 months of continuous Medicare coverage for Parts A, B, and D, excluding Part C (ie, Medicare Advantage) prior to OSA diagnosis (or matched index data for controls without a sleep disorder), as well as 24 months of continuous coverage following the index date. All individuals were diagnosed with depression (defined using the Medicare Chronic Conditions Warehouse algorithm) prior to the index date.⁷ Exclusion criteria included any evidence of OSA-related testing, diagnosis, or treatment prior to the index date. This study was approved by the Institutional Review Board at the University of Maryland, Baltimore (HP-00072414).

Obstructive sleep apnea

OSA was defined by receipt of ≥ 1 International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnostic codes on an inpatient or outpatient claim: 780.51, 327.23, 780.57, or 780.53. Undiagnosed OSA was defined as the 12-month period preceding the index date.

Controls without a sleep disorder

Comparison patients with pre-existing depression but without any history of sleep-related diagnosis, treatment, or diagnostic procedure were randomly selected and assigned an index date, such that the distribution of index dates was equal between cohorts.

HCRU and costs

HCRU was operationalized as counts of claims in each of the 12 months leading up to the index date by point of service (inpatient, outpatient, emergency department, and prescription claims). Mean annual costs were calculated in aggregate (total costs). Costs were adjusted to 2022 US dollars (using the inflation calculator available from the US Bureau of Labor Statistics) for reporting.

Data analysis

We compared demographic and clinical characteristics between those with occult OSA and no OSA and tested differences in distributions for statistical significance. Differences in baseline characteristics were balanced between groups using inverse probability of treatment weights (IPTW). We modeled monthly HCRU by point of service using generalized estimating equations with a negative binomial distribution and a log link, accounting for repeated observations and weighted with IPTW. Rate ratios and 95% confidence intervals (CIs) are reported. These rate ratios compare the rates of monthly HCRU between beneficiaries with undiagnosed OSA and controls without a sleep disorder at each point of service, with higher rate ratios indicating increased HCRU. Next, we modeled mean inflated total annual costs using generalized linear models with a gamma distribution and a log link, weighted with IPTW, and estimated the marginal effects and 95% CIs using the delta method. We report mean costs and their 95% CIs.

RESULTS

We identified 41,500 participants meeting the inclusion criteria, of whom 4,573 (11%) had occult OSA (Table 1). Relative to controls, beneficiaries with occult OSA were younger (74 vs 78 years; P < .001) and more likely to be men (31% vs 21%; P < .001), demonstrated a heavier comorbidity burden, and were more likely to have > 7 chronic conditions (55% vs 47%; P < .001). Average total costs during the year leading up to OSA diagnosis were $148,655 vs $85,969 in controls without a sleep disorder over the same period. In IPTW models, beneficiaries with occult OSA were heavier users of inpatient (rate ratio: 1.53; 95% CI: 1.39, 1.67), outpatient (rate ratio: 1.18; 95% CI: 1.10, 1.27), emergency department (rate ratio: 1.48; 95% CI: 1.35, 1.63), and prescription (rate ratio: 1.09; 95% CI: 1.05, 1.14) services. Mean total costs were also significantly higher among beneficiaries with occult OSA (adjusted cost difference = $44,390; 95% CI: $32,076, $56,703).

Table 1.

Association between OSA and health care resource utilization and costs among Medicare beneficiaries ≥ 65 years with depression over 12 months before OSA diagnosis.

	Rate Ratio (95% CI)
	Unadjusted	Stabilized IPTW
Inpatient
No OSA	Reference	Reference
Undiagnosed OSA	1.68 (1.58, 1.78)	1.53 (1.39, 1.67)
Outpatient
No OSA	Reference	Reference
Undiagnosed OSA	1.30 (1.26, 1.35)	1.18 (1.10, 1.27)
Emergency department
No OSA	Reference	Reference
Undiagnosed OSA	1.61 (1.53, 1.69)	1.48 (1.35, 1.63)
Prescriptions
No OSA	Reference	Reference
Undiagnosed OSA	1.16 (1.13, 1.18)	1.09 (1.05, 1.14)
	Marginal Costs (95% CI)
Total mean inflated costs
No OSA	Reference	Reference
Undiagnosed OSA	$62,686 ($48,956, $76,416)	$44,390 ($32,076, $56,703)

n = 41,500. CI = confidence interval, IPTW = inverse probability of treatment weighted, OSA = obstructive sleep apnea.

DISCUSSION

In this national study of Medicare beneficiaries with pre-existing depression, occult, undiagnosed OSA was associated with dramatically increased HCRU and costs across multiple points of service, even after using IPTW to balance exposure groups on important confounders. These are the first data to examine the economic burden of comorbid OSA among older adults with depression and add to the growing body of evidence demonstrating significant economic burden associated with comorbid OSA among older adults.

Depression is among the most common chronic conditions, especially among older adults, with many older adults experiencing a first depressive episode in the second half of life. Although the US Preventive Task Force has found insufficient evidence for routine screening for OSA in asymptomatic individuals,⁸ the clinical practice guidelines of the American Psychiatric Association highlight the importance of comorbid OSA among patients with depression, particularly those who present with daytime sleepiness, fatigue, or treatment-resistant symptoms.⁹ Future research should examine the potential benefit of OSA screening, triage, and treatment for improving mental health and economic outcomes in older adults.⁶ Such insight could provide evidence-based guidance to clinical trialists, health systems leaders, and policy makers seeking to manage limited health resources in the future.

Strengths of our study include a large sample highly representative of the Medicare fee-for-service population,¹⁰ adequate statistical power to examine relationships of interest, and ability to control for a large number of potential confounders and comorbid disease conditions. At the same time, limitations include inability to measure important clinical characteristics, such as OSA severity or depressive symptoms, known underdiagnosis and undertreatment of both OSA and depression in both the occult OSA cohort as well as among controls (which would bias results toward the null), and potential for residual confounding despite rigorous efforts to control for differences between groups including the use of IPTW.

In conclusion, in this nationally representative study of older Medicare beneficiaries with depression, occult OSA was associated with increased HCRU and costs. OSA treatment should be explored as one potential pathway to reduce the clinical and economic burden of depression among older adults.

ABBREVIATIONS

CI

confidence interval

HCRU

health care resource utilization

IPTW

inverse probability of treatment weights

OSA

obstructive sleep apnea

DISCLOSURE STATEMENT

All authors have seen and approved this manuscript. This research was supported by an investigator-initiated grant awarded from the ResMed Foundation to The University of Maryland, Baltimore (Principal Investigator: E.M.W.). E.M.W.’s and J.S.A.’s institution has received research funding from the AASM Foundation, Department of Defense, Merck, ResMed, and the ResMed Foundation. E.M.W. has served as a scientific consultant to DayZz, Ensodata, Eisai, Idorsia, Merck, Nox Health, Primasun, Purdue, and ResMed and is an equity shareholder in WellTap. No other conflicts of interests are declared.