Prevalence of Obstructive Sleep Apnea Among Veterans and Nonveterans.

Abstract

Purpose: Understanding disease prevalence can inform treatment and resource needs across populations. This study aimed to identify the prevalence of sleep apnea (OSA) among veterans and nonveterans.

Design: The national Comparative Health Assessment Interview Study, cross-sectional survey using probability-based sampling frames.

Setting: Surveys completed by Internet or phone.

Subjects: 15,166 veterans (40% response rate) and 4,654 nonveterans (57% response rate).

Measures: Self-report of healthcare provider-based diagnosis of OSA.

Analysis: Calculation of prevalence of OSA using statistical weighting to allow for direct comparison between veterans and nonveterans. Secondary analyses evaluated OSA by deployment status among veterans and compared average age of OSA diagnosis and differences in OSA prevalence among veterans and nonveterans stratified by gender, marital status, race/ethnicity, and posttraumatic stress disorder diagnosis.

Results: OSA diagnosis was more than twice as prevalent among veterans (21%, 95% CI 20%-22%) than nonveterans (9%, 95% CI 8%-10%; aOR: 2.56, 95% CI 2.22-2.95, P < .001). Deployment was associated with higher odds of OSA among veterans (aOR: 1.64, 95% CI 1.43-18.7, P < 001.) Veterans were diagnosed with OSA on average 5 years earlier than nonveterans.

Conclusion: Veterans have a high prevalence rate of OSA, highlighting the importance of veterans' access to treatment. OSA is likely underdiagnosed in nonveterans, particularly among racial/ethnic minoritized groups. Future research should investigate disparities in access to diagnostic testing for racial/ethnic minority nonveterans and/or risk factors for OSA among racial/ethnic minority veterans. The increased odds of OSA among those with PTSD highlights in the importance of early referral for OSA testing by providers as well as development of trauma-informed strategies to promote OSA treatment adherence. Limitations include a bias toward underestimation of true disease prevalence due to self-report of diagnosis.