Methods
This page outlines the methods used for this report.
Overweight and obesity
Overweight and obesity were based on Body Mass Index (BMI) which was derived using measured height and weight data from the ABS’s National Health Surveys (see Data sources). Height and body composition are continually changing for children and adolescents as they grow. Different BMI cut-off points based on age and sex are used when assessing their BMI at a population level (Cole et al. 2000).
The following BMI ranges were used to categorise overweight and obesity:
- Overweight: 25.0 to 29.9 kg/m2
- Obesity: 30.0 kg/m2 or higher
- Severe obesity: 35 kg/m2 or higher.
Abdominal overweight and obesity
Abdominal overweight and obesity were based on waist circumference measurements from the ABS’s National Health Surveys (see Data sources). The analysis of abdominal overweight and obesity is limited to adults due to a lack of consensus on the definition for children and adolescents.
The following waist circumference measurements were used to categorise abdominal overweight and obesity:
- Abdominal overweight in males is defined as having a waist circumference measurement of 94 cm or more. In females, it is defined as having a waist circumference measurement of 80 cm or more.
- Abdominal obesity in males is defined as having a waist circumference of 102cm or more for men. In females, it is defined as having a waist circumference measurement of 88cm or more.
Crude and age-standardised prevalence estimates are presented as percentages. Crude prevalence, as a percentage, is defined as the number of people with a particular characteristic, divided by the number of people in the population of interest, multiplied by 100.
All prevalence estimates in this report are weighted estimates that use person weights allocated to each survey participant by the ABS.
The jack-knife weight replication method was used to derive the standard error (SE) for each estimate, using replicate weights provided by the ABS.
The statistical significance of any difference in prevalence (percentage) estimates between people across time or population groups (for example, between age groups, socioeconomic quintile, or sex) was assessed using 95% confidence intervals. Confidence intervals were calculated for survey data in this report.
Age-standardised prevalence estimates are presented to remove the influence of age when comparing populations with different age structures. This is necessary because rates of overweight or obesity vary (usually increasing) with age.
The age-standardised proportions in this report have been directly age-standardised to the 2001 Australian standard population.
The birth cohort analysis in this report did not track the same individual over time. Rather, birth cohorts were constructed using cross-sectional survey data representing the Australian population at various time points.
This approached treated, for example, survey participants aged 25–34 in 2011–12 and survey participants aged 35–44 in 2022–24 as representative of the same group of people, that is, those born in 1978–1987, as they aged over these two time points.
The analysis looked at both 10-year birth cohorts and 5-year birth cohorts. This allowed us to look in greater detail of the changes in the prevalence of the overweight and obesity, particularly in the latest 2 time points of 2017–18 and 2022–24.
10-year birth cohorts
The following ABS data sources were used for 10-year birth cohort analysis (see Data sources):
- 2022–24 NHMS
- 2011–12 AHS
- 1995 NNS.
Year of birth was approximated by subtracting age at survey from the survey year. For the 1995 NNS, the analysis assigned survey year as 1995. For the 2011–12 AHS, interviews were conducted over 2011 and 2012, and the 2022–24 NHMS, interviews were conducted from January 2022 to April 2024. Details of which year an individual was interviewed in were not available. For these surveys, the analysis assigned survey year as 2012 for the 2011–12 survey and 2022 for the 2022–24 survey. The use of these years maintained a gap of 10 years between the latest two years of surveys.
Records where then grouped into cohorts based on approximated year of birth, using 10-year spans. The width of the spans was chosen to ensure that there was no overlap of birth cohorts at the time points compared. Table 1 shows the birth cohorts used for 10-year birth cohort analysis.
| Birth cohort | Age in 19951 | Age in 20122 | Age in 20223 |
|---|---|---|---|
| 1918–1927 | 68–77 | .. | .. |
| 1928–1937 | 58–67 | 75–84 | .. |
| 1938–1947 | 48–57 | 65–74 | 75–84 |
| 1948–1957 | 38–47 | 55–64 | 65–74 |
| 1958–1967 | 28–37 | 45–54 | 55–64 |
| 1968–1977 | 18–27 | 35–44 | 45–54 |
| 1978–1987 | 8–17 | 25–34 | 35–44 |
| 1988–19974 | 0–7 | 15–24 | 25–34 |
| 1998–2007 | .. | 5–14 | 15–24 |
| 2008–2017 | .. | .. | 5–14 |
Notes:
1. The survey used is the 1995 NNS.
2. The survey used is the 2011–12 AHS
3. The survey used in the 2022–24 NHMS.
4. For the 1995 NNS, this birth cohort is 1988–1995 and includes children aged 2–7.
5-year birth cohorts
The following ABS data sources were used for 5-year birth cohort analysis (see Data Sources):
- 2022–24 NHMS
- 2017–18 NHS.
Year of birth was approximated by using the same method as the 10-year birth cohorts. For the 2017–18 NHS, interviews were conducted from July 2017 to June 2018. Details of which year an individual was interviewed in were not available. For this reason, the analysis assigned the survey year as 2017. This maintained a gap of 5 years between the 2017–18 NHS and the 2022–24 NHMS, where 2022 was assigned as the year. Table 2 shows the birth cohorts used for 5-year birth cohort analysis.
| Birth cohort | Age in 20171 | Age in 20222 |
|---|---|---|
| 1933–1937 | 80–84 | .. |
| 1938–1942 | 75–79 | 80–84 |
| 1943–1947 | 70–74 | 75–79 |
| 1948–1952 | 65–69 | 70–74 |
| 1953–1957 | 60–64 | 65–69 |
| 1958–1962 | 55–59 | 60–64 |
| 1963–1967 | 50–54 | 55–59 |
| 1968–1972 | 45–49 | 50–54 |
| 1973–1977 | 40–44 | 45–49 |
| 1978–1982 | 35–39 | 40–44 |
| 1983–1987 | 30–34 | 35–39 |
| 1988–1992 | 25–29 | 30–34 |
| 1993–1997 | 20–24 | 25–29 |
| 1998–2002 | 15–19 | 20–24 |
| 2003–2007 | 10–14 | 15–19 |
| 2008–2012 | 5–9 | 10–14 |
| 2013–2017 | .. | 5–9 |
Notes:
1. The survey used is the 2017–18 NHS.
2. The survey used is the 2022–24 NHMS.
For each birth cohort, in each survey year, the following metrics were calculated:
- Obese, defined by BMI
- Severely obese, defined by BMI (for adults only)
- Abdominally obese, defined by waist circumference (for adults only).
The proportions were calculated for total persons for both 5 and 10-year birth cohorts. The proportions were calculated by sex for 10-year birth cohorts only.
The statistical significance of any differences in prevalence estimates between each birth cohort at each age, and also within a birth cohort as people aged, was assessed using z scores.
Comparisons of regions in this report use the ABS 2021 Australian Statistical Geography Standard (ASGS) 5 remoteness areas.
The 5 remoteness areas are Major cities, Inner regional, Outer regional, Remote, Very remote. These areas are defined using the Accessibility/ Remoteness Index for Australia Plus (ARIA+), which is a measure of a location from relative access to services. Noting that the national health surveys exclude Very remote Australia, so these are not included in results in this publication for the general population.
In some instances, data for remoteness areas have been combined because of small sample sizes. For example, Remote has been combined with Outer Regional Australia for results presented by remoteness areas.
For further information see the ASGS Remoteness Structure.
Socioeconomic classifications in this report are based on the ABS Socio-Economic Indexes for Areas 2021 (SEIFA 2021), specifically, the Index of Relative Socio-economic Disadvantage (IRSD). Geographic areas are assigned a score based on social and economic characteristics of that area, such as income, educational attainment, public sector housing, unemployment and jobs in low-skill occupations. The IRSD relates to the average disadvantage of all people living in a geographical area. It cannot be presumed to apply to all individuals living in the area.
For the analyses in this report, the population is divided into 5 socioeconomic areas, with roughly equal populations (each around 20% of the total), based on the level of disadvantage of the statistical local area of their usual residence. The first group includes the 20% of areas with the highest levels of relative disadvantage (referred to as Group 1, most disadvantaged), while the last group includes the 20% of areas with the lowest levels of relative disadvantage (referred to as Group 5, least disadvantaged).
The IRSD values used in this report are based on the 2021 Census of Population and Housing.
This report includes the proportion of adults aged 18 years and over who were classified as overweight or obese, by Primary Health Network (PHN). PHNs are local organisations that connect health services across a specific geographic area, with the boundaries defined by the Australian Government Department of Health, Disability and Ageing.
Proportions have been age standardised to the 2001 Australian population to account for differences in the age structure of the population in different areas.
The quality of estimates from the NHS can vary across PHN areas, as the survey was not specifically designed to produce estimates at this level of geography.
Proportions that have a margin of error that is 10 percentage points or greater have been indicated and these should be used with caution due to the wide confidence interval.
Data for the Northern Territory should be interpreted with caution as Very Remote areas are excluded from the sample.
The relative standard error (RSE) of an estimate is a measure of the error likely to have occurred due to sampling. It is usually expressed as a percentage. The RSEs of the estimates were calculated using the standard errors (SEs):
RSE% = SE(estimate)/estimate x 100
The margin of error (MoE) at the 95% confidence level for each estimate was calculated using 1.96 as the critical value:
MOE = 1.96 x SE(estimate)
The MoE was then used to calculate the 95% confidence interval (CI) around each estimate:
Lower CI = estimate – MOE(estimate)
Upper CI = estimate + MOE(estimate)
The 95% CI is a range of values determined by the variability in data, within which there is a 95% chance that the confidence interval will contain the true value of the population quantity being estimated.
Variation or difference in observed values or rates may be due to a number of causes including, among other things, actual differences in the study’s populations and sampling error. A statistical test of significance indicates how incompatible the observed data are with a specified statistical model. To assess whether differences between estimates are incompatible with a null hypothesis that the survey estimates are normally distributed and that there is no difference between the groups being compared, 95% CIs were used.
A difference between estimates was considered statistically significant if the 95% CIs around the estimates did not overlap. Where there was an overlap between 95% CIs, a test of significance was conducted using the z-score. To do this, the SE of the difference was approximated by:
SE(X-Y) = √(SE(X)^2 + SE(Y)^2)
This standard error is then used to calculate the test-statistic using the following formula:
|X-Y|/ SE(X-Y)
If the absolute value of the test statistic is greater than 1.96 (at the 95% confidence level) then there is evidence of a statistical significance between two estimates. If it is less than 1.96, it cannot be stated with confidence that there is a real difference between the two estimates.
If the 95% CI for the difference between estimates included zero, then the difference was not statistically significant. If it excluded 0, then the difference was considered to be statistically significant.
Cole T, Bellizzi M, Flegal K and Dietz W (2000) 'Establishing a standard definition for child overweight and obesity worldwide: international survey', BMJ, 320:1240–1243.