0590 Clinical Predictors of Positive Airway Pressure Therapy Adherence: A Comparative Analysis of Physiological and Patient-Reported Outcomes

Abstract Introduction Positive Airway Pressure (PAP) therapy remains the gold standard for Obstructive Sleep Apnea (OSA) treatment, yet sustained short and long-term adherence remains a significant clinical challenge. Although various physiological, symptomatic, and device-related variables have been investigated for the prediction of adherence, the relative impact of each of these variables remains debated. This study assessed the impact of multiple variables—ranging from disease severity to subjective symptom scores—on 90-day PAP adherence to identify independent predictors of therapy adherence within a veteran population. Methods We conducted a retrospective chart review of 110 adult patients initiated on PAP therapy at the James A. Haley Veterans’ Hospital in Tampa, Florida, between July and September 2025. All patients utilized the ResMed AirSense 10 Auto-PAP device. We analyzed the relationship between 90-day objective adherence and nine independent variables: Age, BMI, baseline Apnea-Hypopnea Index (AHI), SpO2 Nadir, Epworth Sleepiness Scale (ESS), Insomnia Severity Index (ISI), NOSE score, Residual AHI, and 95% Leak. Statistical analysis included Multiple Linear Regression (MLR) to assess combined effects, followed by Simple Linear Regression (SLR) for individual variables. Results The cohort had a mean age of 58 (∓14) years and a mean BMI of 31.65 (∓5.06) kg/m2. The average 90-day adherence rate for the group was 46%. The MLR model incorporating all nine variables did not reach statistical significance (p = 0.12). Within this multi-variable model, age approached significance (p = 0.09), while AHI showed a trend toward significance (p = 0.14). However, a subsequent SLR revealed a statistically significant positive correlation between baseline AHI and 90-day adherence (p = 0.013; Adjusted R2 = 0.0465). Other variables, including subjective symptom scores (ESS, ISI, NOSE) and device metrics, did not show statistically significant correlations with adherence. Conclusion Our multi-variable model did not yield significant predictors, but individual targeted analysis identified baseline AHI as a statistically significant predictor of 90-day adherence. Notably, this model's low explanatory power suggests that adherence is primarily driven by non-measured factors. Future evaluation of predictors of PAP adherence in this population may benefit from further exploration of behavioral, psychological, and socio-economic variables to improve predictive utility and identify high-risk patients. Support (if any)