Recognition and Diagnosis of Obstructive Sleep Apnea in Older Americans

Tiffany J. Braley, MD, MS; Galit Levi Dunietz, PhD, MPH; Ronald D. Chervin, MD, MS; Lynda D. Lisabeth, PhD, MPH; Lesli E. Skolarus, MD, MS; James F. Burke, MD, MS


J Am Geriatr Soc. 2018;66(7):1296-1302. 

In This Article


Inclusion and exclusion criteria, and sample sizes of Round 3 and sleep module participants are presented in a flowchart in Supplementary Figure S1. Demographic and clinical characteristics of the 3,195 live NHATS participants are listed in Table 1.

Figure S1.

Flow Chart of participants in the National Health and Aging Trends Survey (NHATS) 2013; In the final analytic sample, the 1,052 respondents reflect the unweighted sample frequencies and represent 7,082,963 older Americans in the general population.

Of the 1,052 sleep module participants, 56% (95% CI=53–59%) were at risk of OSA based on a surrogate STOP–BANG score of 3 or more. Of these at–risk individuals, only 8% (95% CI=5–11%) received OSA evaluation with PSG or HSAT; 94% (95% CI=87–100%) of those evaluated received an ICD–9–coded OSA diagnosis, and 82% (95% CI=65–99%) of those were prescribed PAP equipment (Figure 1). Of the remainder of respondents (44%) with surrogate STOP–BANG scores of 2 or less, 2% received OSA evaluations (95% CI=0.8–4%), 90% of whom received an ICD–9 coded OSA diagnosis (95% CI=88–92%) and 50% of whom were prescribed PAP equipment (95% CI=38–62%). Table 2 summarizes the proportion of sleep module participants who received OSA testing based on each surrogate STOP–BANG score. Proportions of respondents who endorsed sleep–related symptoms or characteristics not included in STOP–BANG and likelihood of evaluation are Summarized in Supplementary Table S2.

Upon recalculation of OSA risk with the age item dropped from the surrogate STOP–BANG score (Supplementary Figure S2), the proportion of at–risk respondents remained high (20%), and of this 20%, 13% underwent a sleep study (87% of those at risk of OSA did not receive a sleep study). Ninety–four percent of the 13% of individuals in this group who underwent sleep studies were diagnosed with OSA.

Figure S2.

Proportion of OSA recognition and treatment among 'At Risk' NHATS participants (PSG=polysomnography, HSAT=home sleep apnea testing) after scoring the age item as "0" to assess the potential influence of the age item toward OSA risk. N=1,052 respondents reflects the unweighted sample frequencies that represent 7,082,963 older Americans in the general population.

In analyses of Round 3 sample (n=3,195), older age, male sex, being married or cohabiting, higher BMI, use of a mobility device, pain, cardiovascular disease, and diabetes mellitus were independently associated with OSA diagnosis, but education, race, and depressive symptoms were not (Table 3). Multivariate models suggested significant associations between OSA diagnosis, male sex, and BMI.

Bivariate analyses compared characteristics—those with potential to influence the likelihood of OSA evaluation with PSG or HSAT—of the 92% of unevaluated respondents and the 8% who were evaluated. Independent drivers and those who needed help getting around in and outside the house were more likely to be evaluated. Beneficiaries with a BMI greater than 30.0 kg/m2 were also more likely to be evaluated. OSA evaluation was not associated with presence of dementia, depression, diabetes mellitus, pain, cancer, or stroke.