Abstract Introduction Medication adherence derived from Electronic Health Record (EHR) for chronic disease conditions may predict positive airway pressure (CPAP) adherence for obstructive sleep apnea (OSA) treatment. We compared CPAP adherence and one-year medication adherence for hypertension and hyperlipidemia medications prior to CPAP initiation. Methods We identified two adult OSA cohorts with new CPAP initiations in Kaiser Permanente Southern California between 2016 and 2025 with active prescriptions for hypertension (HTN Cohort) or hyperlipidemia (Chol Cohort) at time of CPAP initiation. Medications were classified using their Generic Product Identifier (GPI). EHR-derived baseline characteristics, medication adherence (proportion of days covered using pill days dispensed up to 365 days prior to CPAP initiation), and 90-day CPAP adherence were compared between CMS- adherents (30-day period with ≥70% of days with ≥4 hours use during first 90-days) and non-adherents. Student’s t-tests evaluated group differences. Associations between medication adherence and CPAP adherence were evaluated using Pearson correlation coefficients. Results In the HTN cohort (n=73,132), CMS adherent patients (48.1%) and non-adherent patients (51.9%) were similar in age (60.5±12.3 vs 59.7±13.2 years) and BMI (35.2±7.8 vs 34.8±8.1 kg/m2). Compared with non-adherents, CMS adherents were more often male (59.6% vs 56.4%) and White (41.9% vs 33.8%), and less frequently Hispanic (26.0% vs 33.1%) or Black (7.5% vs 10.7%). CMS adherents had higher AHI4% (33.5±24.6 vs 29.8±23.4) and substantially greater 90-day CPAP use (363.0±109.0 vs 122.0±120.1 min/night). Medication adherence was statistically higher among CMS adherents (78.8±31.8% vs 77.5±31.8%; p 0.0001), although the effect size was minimal, and the correlation between medication adherence and CPAP adherence was negligible (R=0.02; p 0.0001). In the Chol Cohort (n=18,104), CMS adherent and non-adherent patients showed similar demographic and clinical characteristics as the HTN cohort. Medication adherence was again slightly higher in CMS adherents (70.8%±35.6 vs 68.5±36.9%), though the effect size and its correlation with CPAP adherence were negligible (R = 0.03; p 0.0001). Conclusion Although medication adherence was statistically higher in CMS-adherent patients, the effect size was clinically negligible. Medication adherence is unlikely to contribute meaningfully towards development of CPAP adherence prediction models. Support (if any) NIH NHLBI R01 HL161253-01A1
Hwang et al. (2026) studied this question.