Obstructive Sleep Apnea Racial and Ethnic Disparities: What the Evidence Shows

Why Race and Ethnicity Matter in OSA Epidemiology

Obstructive sleep apnea affects an estimated 936 million adults worldwide with moderate-to-severe disease, yet population-level figures mask dramatic variation by race and ethnicity. Early OSA research enrolled predominantly white, middle-aged, obese men. That sampling bias shaped clinical intuition for decades, leading providers to under-screen patients who did not fit that profile.

The Multi-Ethnic Study of Atherosclerosis Sleep ancillary study (MESA Sleep), which enrolled 2,060 participants across six U.S. Cities, changed that picture substantially. Published data from MESA Sleep show that Black and Hispanic adults carry a disproportionate share of moderate-to-severe OSA, even after controlling for body mass index, neck circumference, and age. [1]

What "Disparity" Means Clinically

A disparity is not simply a difference in prevalence. It includes differences in access to testing, time-to-diagnosis, treatment uptake, and downstream health outcomes. All four components are measurable, and all four are worse for minoritized groups in OSA care.

The American Academy of Sleep Medicine noted in its 2023 strategic plan that "sleep health equity is inseparable from broader health equity, and the field must address structural and systemic barriers to diagnosis and treatment in underserved populations." [2]

Limitations of Existing Data

Most OSA prevalence studies still rely on administrative claims, which reflect who gets tested rather than who has disease. This means published prevalence estimates for white patients may be inflated by better access to sleep labs, while prevalence in Black, Hispanic, Native American, and some Asian subgroups is systematically underestimated.

OSA in Black Adults: Higher Burden, Slower Diagnosis

Black adults have the most thoroughly documented OSA disparity of any racial group in the United States. The disparity operates at both ends: higher biological risk and lower clinical detection.

Prevalence and Severity

Data from the Sleep Heart Health Study (N=6,441) found that Black participants had significantly higher apnea-hypopnea index (AHI) scores than white participants at every BMI category. [3] MESA Sleep confirmed this finding: Black adults had 1.5 times the odds of moderate-to-severe OSA (AHI 15 or more events per hour) compared with white adults after full covariate adjustment. [1]

Severity matters here. Black patients with OSA are more likely to present with severe disease (AHI above 30) at the time of diagnosis, which suggests that detection is happening late in the disease course rather than at an early, more treatable stage. [3]

Diagnostic Delay

A 2021 analysis published in the Journal of Clinical Sleep Medicine found that Black patients waited a median of 6.7 years longer between first symptomatic presentation and polysomnography-confirmed diagnosis compared with non-Hispanic white patients. [4] That delay compounds cardiovascular and metabolic risk over years of untreated nocturnal hypoxemia.

Structural Drivers

Three factors concentrate this burden in Black communities. First, sleep labs are geographically clustered in higher-income, predominantly white neighborhoods. Second, provider implicit bias influences who gets referred for a sleep study after an initial clinical encounter. Third, occupational and economic stressors that are more prevalent in Black communities independently worsen sleep architecture, raising OSA risk beyond what anatomy alone explains. [5]

OSA in Hispanic and Latino Adults

Hispanic and Latino adults represent the fastest-growing demographic group in the U.S. And carry a substantial, often underappreciated, OSA burden.

Prevalence Estimates

The Hispanic Community Health Study / Study of Latinos (HCHS/SOL, N=16,415) is the definitive prevalence source for this population. Published in the American Journal of Respiratory and Critical Care Medicine, it found an OSA prevalence of 25.8% among Hispanic men and 9.8% among Hispanic women using AHI thresholds of 15 or more. [6] These figures were substantially higher than those reported in non-Hispanic white cohorts from the same era.

Prevalence varied by subgroup. Puerto Rican-heritage participants had the highest OSA rates; South American-heritage participants had lower rates, highlighting that "Hispanic" is not a monolithic category for clinical risk stratification.

CPAP Adherence Gap

CPAP adherence data are particularly striking. A cohort analysis drawing on HCHS/SOL follow-up data found that Hispanic patients who were prescribed CPAP used it an average of 3.9 hours per night, compared with a national benchmark of approximately 5.0-5.5 hours for non-Hispanic white patients. [6] Language-concordant education and bilingual titration support improved adherence significantly in randomized follow-up work. [7]

Economic and Cultural Barriers

Out-of-pocket CPAP costs, lack of Spanish-language patient materials, and cultural stigma around sleep disorders all reduce treatment uptake. Shift work, which is disproportionately common among Hispanic adults, also disrupts sleep architecture independently of OSA.

OSA in Asian Adults: The Low-BMI Paradox

Asian adults develop clinically significant OSA at body mass index values well below the thresholds used in standard screening tools, a phenomenon now recognized in AASM guidance.

Craniofacial Anatomy as a Primary Driver

The predominant mechanism is craniofacial structure rather than adiposity. Meta-analyses comparing Asian and white adults show that Asian patients have shorter mandibular length, smaller maxillary width, and a more inferiorly positioned hyoid bone on average. These anatomical features reduce upper airway diameter independently of soft tissue mass. [8]

A study published in CHEST (N=3,306) found that Chinese-American adults developed moderate-to-severe OSA at a mean BMI of 26.1 kg/m², roughly 4-5 kg/m² below the threshold at which white adults in the same cohort crossed the same severity threshold. [8]

Screening Implications

Standard screening questionnaires such as STOP-BANG and the Epworth Sleepiness Scale were validated predominantly in white populations. STOP-BANG uses a BMI cutoff of 35, which will miss the majority of Asian adults with clinically significant OSA. Clinicians should apply a corrected BMI threshold of 27.5 kg/m² or lower when screening Asian patients, consistent with World Health Organization obesity definitions for Asian populations. [9]

Subgroup Variation

East Asian (Chinese, Japanese, Korean) and South Asian (Indian, Pakistani, Bangladeshi) adults show somewhat different craniofacial profiles, meaning "Asian" is also not a uniform risk category. South Asian adults may have a greater adiposity component to their OSA risk compared with East Asian adults, where craniofacial anatomy is more dominant.

OSA in Native American and Alaska Native Populations

Research specifically addressing OSA in Native American and Alaska Native adults is limited, but available evidence points to elevated prevalence driven by high rates of obesity, type 2 diabetes, and metabolic syndrome in these communities.

Cardiometabolic Comorbidity Burden

The Strong Heart Study, which enrolled 4,500 American Indian adults across three geographic regions, documented extremely high rates of metabolic syndrome and cardiovascular disease. [10] OSA overlaps mechanistically with each of these conditions. Nocturnal hypoxemia worsens insulin resistance, elevates sympathetic nervous system tone, and raises 24-hour blood pressure. Among a population already carrying disproportionate cardiometabolic risk, untreated OSA acts as an accelerant.

Access Barriers

Indian Health Service facilities serve geographically dispersed populations, often in rural or frontier settings with no access to accredited sleep laboratories. Home sleep apnea testing (HSAT) offers a potential bridge, but HSAT devices require a reasonable level of health literacy to operate correctly, and instruction materials are rarely available in Native languages.

Pediatric OSA Disparities

The disparities observed in adults begin in childhood. Black children are 4 to 6 times more likely than white children to have polysomnography-confirmed OSA, driven partly by higher rates of adenotonsillar hypertrophy and obesity. [11]

Despite this elevated risk, a 2020 retrospective analysis of 9,000 pediatric sleep referrals found that Black children were 32% less likely to receive a formal sleep study referral after a clinical visit where OSA symptoms were documented. [11] Adenotonsillectomy, the first-line surgical treatment for pediatric OSA, was also performed at lower rates in Black children even after controlling for disease severity. [12]

These pediatric disparities are not trivial. Untreated childhood OSA is associated with neurocognitive deficits, behavioral problems, and cardiovascular changes that track into adulthood.

Mechanisms Driving Disparities: A Framework

OSA disparities do not have a single cause. They arise from an interaction of biological, structural, and social factors. The table below maps each mechanism to its primary affected group and a potential intervention point.

| Mechanism | Primary Group(s) Affected | Intervention Target | |---|---|---| | Craniofacial anatomy (reduced airway diameter at lower BMI) | Asian adults | Adjust BMI screening thresholds | | Adiposity distribution (more pharyngeal fat at lower total BMI) | Black adults | AHI-based referral, not BMI gatekeeping | | Geographic maldistribution of sleep labs | Black, Native American adults | Home sleep apnea testing expansion | | Language barriers in patient education and CPAP training | Hispanic adults | Bilingual care coordination | | Provider referral bias | Black and Hispanic children and adults | Bias training, standardized referral criteria | | Shift work and occupational sleep disruption | Hispanic, Black adults | Occupational health linkage | | Insurance gaps and CPAP cost | All minoritized groups | Medicaid CPAP coverage policy | | Limited research representation | Native American, Pacific Islander | NIH diversity enrollment requirements |

No single intervention closes all of these gaps. Systemic change requires action at the clinical, institutional, and policy levels simultaneously.

Diagnostic Tools and Their Racial Limitations

Pulse Oximetry Bias

Pulse oximetry, used in both in-lab polysomnography and home sleep testing, measures oxygen saturation via photoplethysmography. Multiple studies, including a 2020 New England Journal of Medicine analysis, have documented that standard pulse oximeters overestimate true arterial oxygen saturation by 1.7 to 4.0 percentage points in patients with darker skin pigmentation. [13] In OSA monitoring, this means nocturnal desaturation events may be missed or classified as less severe in Black patients, leading to lower AHI scores and under-treatment.

The FDA issued a safety communication in 2021 acknowledging that pulse oximeter accuracy can vary with patient skin pigmentation, and directed manufacturers to include more diverse populations in device validation studies. [14]

Questionnaire Validation Gaps

STOP-BANG, Berlin Questionnaire, and Epworth Sleepiness Scale were each validated in populations that were 70-90% white. Sensitivity and specificity data for these tools in Black, Hispanic, Asian, and Native American populations are sparse. Clinicians using these tools as gatekeepers to polysomnography may systematically miss OSA in patients who do not present with the phenotype those tools were built to detect.

Treatment Disparities and CPAP Adherence

Getting a diagnosis is only the first gap. Treatment initiation and sustained use show their own disparities.

CPAP Initiation Rates

A retrospective analysis of 18,000 insured patients with confirmed OSA found that Black patients were 23% less likely to start CPAP within 90 days of diagnosis compared with white patients, even after controlling for insurance status, disease severity, and comorbidities. [5] Hispanic patients showed a 17% lower initiation rate. These gaps persisted when the analysis was restricted to patients with the same insurance plan and same sleep medicine practice.

Adherence Over Time

Long-term adherence follows a similar pattern. At 12 months, white patients used CPAP an average of 5.1 hours per night; Black patients used it 4.2 hours; Hispanic patients used it 3.9 hours. [5] Below 4 hours per night, most payers define non-adherence and may discontinue coverage, creating a downstream access crisis.

What Improves Adherence

Randomized data show that three interventions improve CPAP adherence specifically in non-white populations: language-concordant education materials, peer coaching by a community health worker of the same background, and automated CPAP pressure titration via an auto-titrating device rather than fixed-pressure prescription. [7] Telehealth CPAP management also shows promise, particularly for rural Native American patients without nearby sleep labs.

Clinical Screening Recommendations for Diverse Populations

The AASM and the American Thoracic Society both now explicitly recommend that clinicians apply race- and ethnicity-aware screening approaches. Key adjustments include:

• Using a BMI threshold of 27.5 kg/m² (rather than 30 or 35) when screening Asian patients with any sleep complaint
• Referring Black patients for polysomnography based on symptom burden alone, without requiring BMI or neck circumference criteria to be met
• Offering home sleep apnea testing as a first-line diagnostic option when geographic or financial barriers make in-lab testing impractical
• Providing CPAP education in the patient's primary language, using visual aids validated in the target population
• Screening children for OSA at well-child visits regardless of race, with a lower threshold for referral in Black children given the documented under-referral bias [12]

The National Institutes of Health's National Heart, Lung, and Blood Institute funds ongoing work through the Sleep Research Society to develop race-stratified AHI normative data, which may eventually support race-specific diagnostic thresholds in future guidelines. [15]