Obstructive sleep apnoea in Aotearoa/New Zealand : an objective and questionnaire-based approach to population prevalence and clinical screening : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Public Health at Massey University, Wellington Campus, New Zealand
The goals of this thesis were to objectively assess the prevalence of obstructive sleep
apnoea syndrome (OSAS) among Māori and non-Māori adults in a community-based
sample, and to develop a questionnaire-based multivariate predictive tool for OSAS, to
help improve referral of patients to specialist sleep services, and prioritise waiting lists.
This research was situated within the wider scope of ethnic inequalities in health
between Māori and non-Māori, and was conducted within a Kaupapa Māori Research
Between August 1999 and June 2001 letters and information were progressively sent
out to 1200 (600 Māori, 600 non-Māori) Wellington residents aged 30-60 years selected
randomly from the electoral rolls. Participants were asked to wear a small sleep
monitoring device (MESAM4) for one night in their own homes and to fill out a sleep
questionnaire. Contemporaneously, sleep and questionnaire data were collected from
510 consecutive patients aged 30-60 years, who were referred to the regional sleep
clinic for suspected OSAS.
In the community sample, OSA was found to be more prevalent among Māori. Among
men, 21.98% of Māori had OSA (RDI≥ 5) compared with 11.37% of non-Māori.
Among women, 6.28% of Māori and 3.02% of non-Māori respectively had OSA (RDI≥
5). The higher risk among Māori appeared to be due to well-recognised risk factors
such as higher body mass index (BMI) and larger neck circumference, rather than
ethnicity per se.
Using the combined data from the community and clinical samples, two clinical
prediction models were developed using logistic regression modelling. One model
(Model 1a) included age, sex, observed apnoeas, self-reported habitual snoring,
subjective excessive daytime sleepiness, and BMI. The second model (Model 2a)
included neck circumference instead of BMI. Model 1a correctly classified 82.50% of
participants (sensitivity 72%, specificity 87%). Model 2a correctly classified 81.10% of
participants (sensitivity 80%, specificity 82%).
This research indicates that OSA is a common problem among New Zealand adults and
that ethnic disparities exist. The results provide important guidance for planning to
meet population needs, by identifying differential needs of specific groups. The
prediction models provided reliable estimates of a priori probability of OSA, and
therefore may be useful tools for screening patients for OSAS.