Toward a Multi-Level Approach to CPAP Adherence

The diagnosis of obstructive sleep apnea (OSA) in patients referred to a sleep medicine specialist is often not a challenging one. Cardinal symptoms—snoring, witnessed apnea, fragmented sleep and daytime fatigue—are readily elicited, and signs of morbid obesity and upper airway crowding easily identified on exam. Patients meeting appropriate criteria may then be started on continuous positive airway pressure (CPAP) treatment.¹

Once the CPAP appliance is sitting on the bedside table, however, a greater challenge begins: namely, how to ensure adequate adherence to therapy. As U.S. Surgeon General C. Everett Koop once opined in reference to medication adherence, “drugs do not work in patients who do not take them,” so it follows that CPAP therapy will not work unless patients persist in using CPAP nightly, for as many hours as possible, over the course of years. Without optimal CPAP use, patients will fail to achieve the full symptomatic^2,3 and cardiovascular⁴ benefits of therapy.

Clearly what is needed is a comprehensive, longitudinal disease management program for sleep apnea in which the diagnostic work-up and initiation of CPAP therapy are only the first steps in a long-term program. Key components include ensuring adequate mask fit, pressure adjustment, review of CPAP adherence and barriers to adherence, effectiveness reports, weight loss, assessment of occupational and driving safety, and consideration of alternative therapies.⁵ Ideally, for such a disease management program to be most effective, it would be helpful to identify those patients who are unable to adhere to CPAP and why. Bakker and colleagues, in an article in this issue of SLEEP, seek to advance our understanding of this important but challenging latter issue.⁶

In a comprehensive, prospective evaluation of 126 consecutive patients presenting to a regional New Zealand (NZ) sleep clinic, Bakker et al. found that the socioeconomically disadvantaged Māori population demonstrated poorer adherence to CPAP therapy [median 4.7 interquartile range (IQR) 2.2 h/night vs. a median 5.3 IQR 2.6 for the non-Māori population, P = 0.05].⁶ Socioeconomically disadvantaged subjects—those with annual income of less than $20,000 (NZ dollars), those holding a community service card to subsidize health care costs, and those demonstrating the highest level of socioeconomic deprivation on the NZ Deprivation Index—were also less likely to adhere to CPAP for an average of 4 h or more per night. In adjusted analysis, once measures of educational achievement and socioeconomic deprivation were included, the apparent ethnic-based difference in CPAP adherence melted away, leaving only failure to achieve a tertiary education and, to a lesser extent, high levels of socioeconomic deprivation, as risk factors for lower adherence to CPAP. Factors not associated with CPAP adherence included disease severity (apnea hypopnea index), daytime symptoms (baseline and change in the Epworth Sleepiness Scale), an NZ area-based index of socioeconomic deprivation, a standardized assessment of health literacy, and a baseline assessment of the Self-Efficacy Measure for Sleep Apnea (the SEMSA, consisting of risk perception, outcome expectancies and belief in one's ability to use CPAP).⁷

The strengths of the Bakker study lie in its prospective design and its inclusion of a comprehensive range of presumed clinical, socioeconomic, ethnic and behavioral risk factors for CPAP adherence.⁶ Weaknesses include the single center design, which limits generalizability, and small sample size, which allows for a possible Type II error (i.e., failure to detect determinants of CPAP adherence). A third weakness lies in the data presentation, as only the subset of subjects with the highest quantile of socioeconomic deprivation on the NZ socioeconomic index had a significantly decreased odds of adequate CPAP adherence [adjusted OR 0.10 (95% CI 0.02-0.86)].⁶ The socioeconomic index failed to predict adherence over the full range of quantiles, a point somewhat obscured by the reporting of p-values for the individual quantiles rather than for the index as a whole.

The finding by Bakker et al.⁶ that socioeconomic deprivation and educational attainment confound the association of race/ethnicity and CPAP adherence^8,9 is consonant with a recent report linking low neighborhood socioeconomic status and poor CPAP adherence in a U.S. veteran patient population.¹⁰ Given the co-localization of minority race/ethnicity and socioeconomic deprivation in many segments of Western society,¹¹ the report by Bakker and colleagues⁶ highlights the potential pitfall of focusing on race/ethnicity per se as a risk factor for CPAP adherence. Instead, for a complex behavior such as CPAP adherence, reporting of race/ethnicity may serve as a marker for residual confounding of co-existent (but often unrecorded) socioeconomic hardship, cultural or behavioral practices or beliefs, or other environmental challenges, not an etiologic factor in and of itself.¹⁰

Beyond the challenge of disentangling the potential influences of race/ethnicity and socioeconomic deprivation on health behaviors, it is clear that there are many patients from moderate and high socioeconomic strata who also fail to adhere well to CPAP therapy. The report by Bakker et al.⁶ demonstrates the difficulty in predicting CPAP adherence among these subjects as well. Future work, therefore, should focus not only on assessment of baseline variables, but also on barriers to adherence that may only be identified soon after CPAP therapy is initiated.

The powerful ability to track daily CPAP usage opens the door to the development and benchmarking of a series of iterative interventions. Adherence initiatives, founded on potentially modifiable behavioral factors,¹² could be more effective if applied within a framework that acknowledges the influence of household members, the workplace, the local community and society at large.¹³ Such a multi-level or socio-ecological approach has been exemplified in the tobacco cessation, healthy diet, and mammography screening campaigns.¹⁴ In the report by Bakker et al., one component of the health belief model, self-efficacy, measured pre-CPAP treatment, was not associated with adherence.⁶ However, recent reports suggest that the self-efficacy model may be most informative only after exposure to CPAP,¹⁵ thereby highlighting the temporal aspect of behavioral influences.

To conclude, CPAP adherence program initiatives should embrace a multi-level approach that recognizes that individuals executing healthy behaviors, like effective CPAP use, are affected not just by mask fit or disease symptomatology, but by a gamut of interactions between behavioral and environmental factors. Such healthy behavior initiatives, if successful, would benefit both the individual and the community at large.

DISCLOSURE STATEMENT

The authors have indicated no financial conflicts of interest.

CITATION

Platt AB, Patel NP. Toward a multi-level approach to CPAP adherence. SLEEP 2011;34(11):1459-1460.