A population-based study of obstructive sleep apnea: using probabilistic case ascertainment through health administrative databases

Background — Although polysomnography (PSG) is the diagnostic gold standard, PSG data on obstructive sleep apnea (OSA) presence and severity are rarely available at the population level. Health administrative databases offer broad coverage for OSA case ascertainment but require methods that address inherent misclassification. We utilized a previously validated probabilistic case ascertainment model of OSA to estimate population-level prevalence and correlates of OSA.

Methods — We conducted a population-based retrospective cohort study in Ontario, Canada. First, using a clinical cohort with PSG-confirmed OSA severity (measurement cohort; n = 18 581), we quantified misclassification in probabilistic OSA case definitions. We then applied the externally validated probability models to all adults 18 years and older in Ontario who underwent in-laboratory overnight PSG between 2010 and 2018 (provincial PSG cohort; n = 834 361). OSA prevalence and correlates were estimated using multiple probability thresholds and a threshold-free bootstrap imputation approach (BIA).

Results — Compared to the PSG-based gold standard, the BIA surrogate definition had the lowest average relative bias in OSA prevalence estimation (BIA: 5% versus >38% probability threshold: 11.4%). Using BIA, we estimate a 44.3% prevalence of moderate-severe OSA in the provincial PSG cohort, which in scenario-based projections yields a 3.4%–17.1% general population prevalence dependent on an assumed OSA underdiagnosis rate. OSA prevalence was higher in men, older ages, and previous comorbidity and was stable annually.

Conclusion — Probabilistic case ascertainment using health administrative data enables robust population-level estimation of OSA epidemiology despite the absence of clinical diagnostic data and provides a framework for studying OSA prevalence, correlates, and outcomes over time.