Telephone surveys provide reliable information on risk behaviours and health status of Aboriginal and Torres Strait Islander people
Correspondence to: Ms Margo Barr, Centre for Epidemiology and Evidence, NSW Ministry of Health, 73 Miller St, North Sydney, NSW 2055; e-mail: firstname.lastname@example.org
Aboriginal and Torres Strait Islander Australians have life expectancies below that of the general Australian population and are significantly more likely to experience ill health and disability – and reduced quality of life because of ill health – than other Australians.1 Therefore, quality data on the health of Aboriginal and Torres Strait Islander Australians are needed to evaluate programs and interventions, to assess the effectiveness of policies and to inform policy and program development.2
The Australian Bureau of Statistics (ABS) provides risk behaviour and health status estimates for Aboriginal and Torres Strait Islander Australians every six years through the face-to-face National Aboriginal and Torres Strait Islander Health Survey (NATSIHS).3
State-based continuous landline telephone health surveys such as the New South Wales Population Health Survey (NSWPHS),4 designed to provide risk behaviours and health status estimates for the general population, have also been used to report on Aboriginal and Torres Strait Islander Australians.5,6 But how accurate are these estimates using a telephone survey that was not specifically designed for this purpose? To investigate this question, results for Aboriginal and Torres Strait Islander persons from the 2002–2005 NSWPHS were compared with NSW results from the 2004–05 NATSIHS.
Prevalence estimates and standard errors for alcohol non-drinkers; fruit and vegetable consumption; current smoking; influenza and pneumococcal vaccination; self-rated health; current asthma, diabetes or high blood glucose; and overweight or obesity; using the same definitions and for the same age groups (ideally adults 16 years and over), were obtained from the NSW Ministry of Health for the NSWPHS and from the ABS for the NSW-NATSIHS.
Each of the prevalence estimates were then compared using a statistical test as follows:7
- 1The difference d between the two independent surveys prevalence estimates was calculated: d=(ENSWPHS–ENSW-NATSIHS).
- 2The standard error for the difference was calculated: SE(d)=√[SE(ENSWPHS)2+SE(ENSW-NATSIHS)2]
- 3The ratio z were calculated: ratio z=d/SE(d)
- 4The ratio z was then compared to the standard normal distribution. The 95% confidence interval for the difference was also calculated: d−1.96*SE(d) to d+1.96*SE(d)
Results and Discussion
Only alcohol non-drinkers (27.1% v 19.4%; p=0.01) and current smoking (44.6% v 52.9%; p=0.02) differed significantly between NSWPHS and NSW-NATSIHS (Table 1).
Table 1. Risk behaviour and health status comparisons between the 2002–05 NSW Population Health Survey and 2004–05 National Aboriginal and Torres Strait Islander Health Survey.
(18 years and over)
|NSW-NATSIHS||19.4||2.4||12.3|| || || || |
|Recommended fruit consumption||NSWPHS||37.8||2.3||6.1||3.30||3.75||0.88||0.19|
|NSW-NATSIHS||41.1||3.0||7.2|| || || || |
|Recommended vegetables consumption||NSWPHS||9.5||1.3||14.0||1.00||1.84||0.54||0.29|
|NSW-NATSIHS||8.5||1.3||15.1|| || || || |
(18 years and over)
|NSW-NATSIHS||52.9||3.1||5.8|| || || || |
(50 years and over)
|NSW-NATSIHS||48.7||10.3||21.1|| || || || |
(50 years and over)
|NSW-NATSIHS||21.9||6.0||27.3|| || || || |
|Positive self-rated health||NSWPHS||77.3||1.9||2.4||1.20||3.09||0.39||0.35|
|NSW-NATSIHS||76.1||2.4||3.2|| || || || |
|NSW-NATSIHS||17.7||1.7||9.8|| || || || |
|NSW-NATSIHS||8.3||1.1||13.7|| || || || |
|Overweight or obese||NSWPHS||48.8||2.5||5.1||1.50||3.61||0.42||0.34|
|NSW-NATSIHS||50.3||2.6||5.2|| || || || |
The demographic characteristics of the NSWPHS and NSW-NATSIHS samples were also compared to the 2001 Aboriginal and Torres Strait Islander population of NSW.8 The NSWPHS sample contained a higher proportion of older adults, females, and regional and remote adults, and the NSW-NATSIHS sample contained a higher proportion of females, and regional and remote adults.
Finding that the majority of the risk behaviours and all the health status prevalence estimates were not significantly different was encouraging. The higher rate of alcohol non-drinking and lower rate of current smoking in the NSWPHS was likely due to the under-representation of young Aboriginal and Torres Strait Islander men in the NSWPHS telephone survey. Although 98% of Aboriginal and Torres Strait Islander persons had used telephones in the month prior to the National Aboriginal and Torres Strait Islander Social Survey (NATSISS) interview, landline phone ownership still remains low (71% in non-remote and 40% remote) compared to the general public; however, mobile phone ownership is becoming more common (more than 90% in non-remote and more than 80% remote).9 Expanding the NSWPHS sampling frame, as planned in 2012, to include mobile phones could potentially increase access to young Aboriginal and Torres Strait Islander men and improve the accuracy of the estimates for alcohol and smoking.