Get access

Ethnicity and socioeconomic status predict initial continuous positive airway pressure compliance in New Zealand adults with obstructive sleep apnoea


  • Funding: R. Lory is receiving payment for the research project from the Asthma and Respiratory Foundation of New Zealand.

  • Conflict of interest: None.


Background:  Understanding factors that contribute to low continuous positive airway pressure (CPAP) compliance will lead to improvements in the long-term outcome of patients with obstructive sleep apnoea (OSA) syndrome. Both cultural and socioeconomic factors are likely to be important but have not been systematically studied.

Aim:  To examine the effect of ethnicity and socioeconomic status on initial CPAP usage for people with OSA in New Zealand.

Methods:  We retrospectively collected demographic, clinical and CPAP treatment-related data on patients undergoing a 1-month CPAP trial for a 10-month period. We compared objectively measured CPAP usage (by anova) with self-identified ethnicity; levels of socioeconomic deprivation (NZDep06 index), Epworth Sleepiness Scale (ESS) and Apnoea-Hypopnoea Index (AHI).

Results:  A total of 214 patients with a mean age of 52.7 (±11.8) years, mean AHI 57.3 (±35.8) events per hour and mean ESS 13 (±5.58)/24 made up the cohort. CPAP usage which averaged 5.13 ± 2.34 h per night was significantly lower in patients of non-European ethnicity (= 0.019 univariate) and remained significant after socioeconomic status was added to the model (= 0.048). Patients living in the most socioeconomically deprived areas showed significantly lower compliance with CPAP on univariate analysis (= 0.024, NZDep06 scores 1&2, average 5.3 per night compared to score NZDep06 scores 9&10, average 4.3 h per night), but this effect was no longer significant once ethnicity was added to the model (= 0.28).

Conclusion:  CPAP usage in New Zealand is affected by both ethnicity and level of socioeconomic deprivation. We recommend further research to unravel specific cultural and socioeconomic reasons for the variance reported.