Health-adjusted life expectancy
For ABDS 2024, the estimation of health-adjusted life expectancy (HALE) uses a three-step process:
- Estimate the number of deaths in 2024. These estimates are used to calculate the mortality rates which underpin estimates of life expectancy.
- Calculate life expectancy in 2024 by constructing a life table based on the above mortality rates. For ABDS 2024, AIHW calculated its own life table for 2024.
- Estimate HALE in the usual manner using Sullivan’s method (see Jagger et al. 2014). Here the estimated life expectancies in 2024 by sex and age are applied to the estimated YLD rates for 2024. Further information can be found in Australian Burden of Disease Study: methods and supplementary material 2018 (AIHW 2021).
Estimation of the number of deaths in 2024
Deaths for 2024 were estimated using a log-linear (Poisson) regression model. Deaths were modelled by sex and single year of age. The model included historical time series of quarterly (all cause) deaths from 2011 to 2022 (excluding 2020), provisional deaths for 2023 and 2 quarters of 2024 (January to June) (ABS 2024b), estimated resident populations for 2011 to 2023 and for the first quarter for 2024 (ABS 2024a). In this time series, the 2022 deaths were adjusted for the undercount which occurs when using the latest year of deaths data (ABS 2022) by using the year of registration of death rather than the year the death occurred for 2022.
The method described above for estimating deaths in 2024 differed to that used in ABDS 2023 where the estimated number of deaths was the sum of provisional deaths in the first half of 2023 and the estimated deaths in the second half of the year based on linear trends of monthly deaths. The log-linear methodology was preferred because the log-linear model is widely used to estimate counts and the projection approach is consistent with the nowcasting approach used for projecting burden rates in ABDS.
AIHW validated the estimated number of deaths in 2024 using a number of different methods. Firstly, the age and sex-specific trends in deaths over the period 2011 to 2024 were assessed. Secondly, projected deaths estimated for 2023 derived from the above log-linear (Poisson) regression model were compared to published ABS provisional deaths for 2023 and the estimates were within 1% difference. Thirdly, the total number of deaths estimated for 2024 was considered reasonable based on ABS provisional deaths for available months in 2024 and predicted deaths for the remaining months in 2024 (if assuming similar monthly trends to those observed in the latter part of 2023).
Calculation of life expectancy in 2024
A full life table (by sex and single year ages from 0 to 100 years or more) was constructed using the mortality rates as estimated above. Other components of the life table that required estimation of inputs for 2024 were:
- The probability of dying before the first birthday. This was estimated by applying the latest available (2023) crude birth rate and sex-ratio at birth from birth registrations collated by the ABS (ABS 2023).
- The average period lived for infants dying before their first birthday. This was estimated for 2024 using the latest available deaths data (2022) from the National Mortality Database (NMD). The average period lived by infants is calculated from the NMD and the methodology is comparable to estimates used elsewhere (for example, Office for National Statistics (ONS) 2019).
- The amount of time lived past 100 years for centenarian deaths. This was estimated using a standard approach of assuming that the probability of death in the age group 100+ is 1.0 and the amount of time lived is estimated as the ratio of the number of survivors to age 100 divided by the death rate for the age group 100+.
The remaining life table calculations are derived from the predicted deaths rates in 2024 and the above estimated inputs.
Calculation of HALE
HALE for 2024 was calculated using the same method (Sullivan’s) as in previous studies: by applying the projected 2024 YLD rates by age and sex to the projected 2024 life expectancy estimates by age and sex. For other ABDS reference years, the 2016–2018 life table (ABS 2019) was used for 2018 HALE estimates, the 2013–2015 life table (ABS 2016) for 2015 HALE estimates, the 2010–2012 life table (ABS 2013) for 2011 HALE estimates and the 2002–2004 life table (ABS 2005) for 2003 HALE estimates.
ABS (Australian Bureau of Statistics) (2005) Life tables, 2002–2004, ABS, Australian Government, accessed 1 July 2022.
ABS (2013) Life tables, 2010–2012, ABS, Australian Government, accessed 1 July 2022.
ABS (2016) Life tables, 2013–2015, ABS, Australian Government, accessed 1 July 2022.
ABS (2019) Life tables, 2016–2018, ABS, Australian Government, accessed 1 July 2022.
ABS (2022) Causes of Death, Australia methodology, ABS, Australian Government, accessed 18 September 2024.
ABS (2023) Births, Australia, ABS, Australian Government, accessed 31 October 2024.
ABS (2024a) National, state and territory population, March 2024, ABS, Australian Government, accessed 31 October 2024.
ABS (2024b) Provisional mortality statistics, Jan–Jun 2024, ABS, Australian Government, accessed 15 October 2024.
AIHW (Australian Institute of Health and Welfare) (2021) Australian Burden of Disease Study: methods and supplementary material 2018, AIHW, Australian Government, accessed 19 September 2024.
Jagger C, Van Oyen H and Robine J (2014) Health expectancy calculation by the Sullivan method: a practical guide (4th edition), Newcastle University, accessed 12 September 2022.
ONS (Office for National Statistics) (2019) Guide to calculating national life tables: Explanation of the methodology used to create the national life tables, ONS, accessed 7 November 2024.