Independent review of the ABS' Aboriginal and Torres Strait Islander life expectancy estimates

The ABS produces estimates of life expectancy for Aboriginal and Torres Strait Islander Australians every five years following the Census of Population and Housing to support government policy and reporting about the wellbeing of Aboriginal and Torres Strait Islander people.

The estimation of life expectancy for Australia’s Indigenous population is a complex and challenging exercise, largely due to the definition and identification of the Aboriginal and Torres Strait Islander population and how the population changes over time. The concept collected in the Census asks for people to self-identify whether they are of Aboriginal and/Torres Strait Islander origin. The way that people identify (or are identified) as such has changed over time, which can be explained by demographic and non-demographic factors (see Understanding change in counts of Aboriginal and Torres Strait Islander Australians: Census). The ABS uses Census information to estimate the Aboriginal and Torres Strait Islander population and related statistics such as life expectancy and these estimates are therefore affected by changes in identification over time.

In response to advice from the ABS Aboriginal and Torres Strait Islander Demographic Statistics Expert Advisory Group (ATSIDSEAG), the ABS commissioned an independent review to assess the fitness for purpose of the ABS Aboriginal and Torres Strait Islander life expectancy estimates. The review was commissioned to maintain the quality and relevance of ABS Aboriginal and Torres Strait Islander life expectancy estimates. The review panel comprised Dr Alison Taylor (Chair), Adjunct Professor Tony Barnes, and Professor Yin Paradies. The review was conducted during 2020 and 2021 and focused on the methodology used to produce the 2015-17 ABS Aboriginal and Torres Strait Islander life expectancy estimates. Its findings have been used to guide the production of Aboriginal and Torres Strait Islander life expectancy, 2020–2022, released 29 November 2023.

The ABS summary of the key findings and recommendations from the review are given below.

The ABS releases Aboriginal and Torres Strait Islander life expectancy estimates every five years following the Census of Population and Housing. This release includes life expectancy at birth for Aboriginal and Torres Strait Islander males and females. Following the 2016 Census, the ABS produced life expectancy estimates for Australia, for selected states and territories (New South Wales, Queensland, Western Australia and the Northern Territory), by remoteness area and by the Index of Relative Socio-economic Disadvantage.

The method used to produce ABS Aboriginal and Torres Strait Islander life expectancy estimates following the 2016, 2011 and 2006 Censuses is based on data from the Census of Population and Housing, the Post Enumeration Survey (a short survey conducted shortly after the Census to assess the coverage of the Census and adjust for this in official population estimates) and death registrations. The ABS uses data from the Census and Post Enumeration Survey and demographic modelling to produce population estimates for the Aboriginal and Torres Strait Islander population by age, sex and location, and death registrations provide data on deaths of Aboriginal and Torres Strait Islander people by age, sex and location.

The ABS method uses data linkages between the Census and death registrations, and data linkages between the Census and Post Enumeration Survey, to produce adjusted death registrations that account for potential underreporting and misreporting of Aboriginal and Torres Strait Islander status in registration data. The population estimates and adjusted death registrations are then used to directly calculate life tables and life expectancy via standard demographic methods. A key feature of this method is the conceptual consistency between the numerator (adjusted death registrations) and the denominator (population estimates) used to calculate life expectancy. Using this method, the ABS draws on the strength of the Post Enumeration Survey, which is collected via interview and produces high quality data with limited non-response, to complement death registrations data and the high coverage and data quality of the Census. More information about the methodology can be found at Aboriginal and Torres Strait Islander life expectancy methodology, 2020–2022.

The panel concluded that the ABS national headline life expectancy estimates are satisfactory point-in-time measures of Aboriginal and Torres Strait Islander life expectancy. Further, the panel noted that the national headline estimates likely allow adequate measurement of life expectancy differences between Aboriginal and Torres Strait Islander and non-Indigenous Australians.

The panel noted that, because of a combination of data quality, definition (including identification), classification, measurement, population size and estimation differences, Aboriginal and Torres Strait Islander life expectancy estimates (and changes to these estimates) can never achieve accuracy comparable to total Australia population life expectancy estimates. However, the panel stated that there are no currently used or proposed methods that clearly produce superior national estimates in all respects than the method currently used by the ABS.

A primary quality concern identified by the panel was the width of estimated confidence intervals for point-in-time life expectancy estimates, and the absence of confidence intervals for the gap between life expectancy estimates for Aboriginal and Torres Strait Islander and non-Indigenous Australians. Due to statistical uncertainty in the estimates (among other concerns discussed below), the panel noted concerns about the ability of the national headline estimates to measure statistically significant changes in life expectancy over the preceding five years and noted greater uncertainty for subnational estimates.

The panel found that the introduction of an adjustment to take age-specific identification rates into account likely improved national estimates, while it also highlighted biases in some existing (not age-adjusted) sub-national estimates. As such, the panel stated that the exploration of methods for developing similar age-specific adjustments at sub-national levels should be pursued as a priority. The panel stated that age-specific adjustments for remoteness areas and the Index of Relative Socio-Economic Disadvantage should also be investigated.

The panel noted that the way that people identify (or are identified) determines the Aboriginal and Torres Strait Islander population, and as a result, life expectancy changes over time risk being attributed entirely to improvements in population health, when such change may be in part, due to changes in identification. The panel advised that this risk can be best managed by always presenting change estimates in ways that openly recognise this possibility, and actively seek its minimisation.

The panel recommended changes to the presentation and communication of ABS life expectancy estimates to support interpretation and use of the statistics by stakeholders. The panel recommended consideration of engagement practices to ensure the appropriate range of technical expertise and professional backgrounds was involved, and the mechanisms for engagement was appropriate to the nature of the collaboration needed. The panel highlighted that particular attention should be given to improving engagement with Aboriginal and Torres Strait Islander people.

The panel stated that the importance of accurately estimating Aboriginal and Torres Strait Islander life expectancy, and subsequently measuring the gap with non-Indigenous Australians’ life expectancy was widely recognised. The panel acknowledged that assessing any changes in the life expectancy gap, as an identified national priority, requires the highest possible quality data, methods, and information to inform practical, relevant, and effective policies, planning, monitoring, advocacy, resource allocation and service provision.

The panel made 31 detailed recommendations relating to methodology, possible additional estimates, planning, communication, and engagement.

The ABS thanks the panel for their valuable time and efforts, application of their significant expertise, and dedication to a thorough and considered approach in their conduct of the review. This led to the delivery of a constructive and comprehensive report on ABS Aboriginal and Torres Strait Islander life expectancy estimates.

The ABS has been working with relevant stakeholders and engaging with members of the review panel to implement and explore the recommendations from the review. These investigations have led to new insights and deeper understanding of the methodology used, its applications and their implications.

The latest release of Aboriginal and Torres Strait Islander life expectancy was released on 29 November 2023, and incorporates changes based on the recommendations from the review and subsequent investigations. 

The recommendations and the ABS’ response are summarised below. For more detail about the review, or to request a copy of the review report, please email demography@abs.gov.au.

1. Develop and consider the feasibility of producing sub-national life expectancy estimates, in line with the current method, with age-specific adjustments to Aboriginal and Torres Strait Islander deaths.
2. Develop estimates of the error of the gap between life expectancy for Aboriginal and Torres Strait Islander and non-Indigenous Australians.
3. Produce guidance on the interpretation of changes in life expectancy to accompany all relevant life expectancy estimates.

  4 – 5. Analyse and compare estimates of the error of life expectancy estimates produced with the current method to a method that produces adjustments to Aboriginal and Torres Strait Islander deaths through linkage between the Census and deaths data only (and not linkage with the Post Enumeration Survey).

6. Consider possible improvements to life tables at older ages through addressing potential misreporting of age.

ABS response

The ABS undertook investigations into the feasibility of producing sub-national life expectancy estimates with age-specific adjustments to Aboriginal and Torres Strait Islander deaths (Recommendation 1). Examination of relevant methods found that, under certain conditions, age-specific adjustments produced life expectancy estimates with high errors. Given the outcome of these investigations, the review panel updated their advice and no longer recommended that the ABS pursue the implementation of age-specific adjustments for sub-national life expectancy estimates. The ABS then undertook further investigations into age-specific adjustments and found that a partial age adjustment was possible and able to be reliably applied nationally and sub-nationally. Partial age adjustment has been implemented in the 2020–2022 Aboriginal and Torres Strait Islander life expectancy estimates.

The ABS has developed estimates of the error of the gap between life expectancy for Aboriginal and Torres Strait Islander and non-Indigenous Australians, which will be included in the 2020–2022 Aboriginal and Torres Strait Islander life expectancy release. This release will include other changes to address the recommendations of the review, such as emphasising caution on the interpretation of changes in life expectancy over time.

  7 – 8. Consider the feasibility of producing annual life expectancy estimates, as well as estimates for South Australia and Victoria.

9. Consider the feasibility of producing population projections that include the impact of non-demographic changes, such as changes in the propensity for people to identify as Aboriginal and/or Torres Strait Islander.

ABS response

The ABS explored the feasibility of producing estimates for South Australia and Victoria for the 2020–2022 Aboriginal and Torres Strait Islander life expectancy release. Data for these jurisdictions did not support the production of reliable estimates. The ABS has published the quality criteria used to assess the feasibility of producing sub-national life expectancy estimates, and the outcomes for South Australia and Victoria are detailed within.

Given the method the ABS currently uses to produce Aboriginal and Torres Strait Islander life expectancy estimates is anchored on the five-yearly Census of Population and Housing, the production of annual national estimates will not be further explored at this time.

Working with relevant stakeholders, the ABS will explore the suitability of including an additional projection series that considers the impacts of identification change in the 2011–2036 Aboriginal and Torres Strait Islander population projections release.

10. Seek to work with the Australian Institute of Health and Welfare (AIHW), and other relevant organisations, to clarify the relationship between different life expectancy estimates and reporting.

11 – 12. Planning to determine responses to potential future input data changes (that is, the Census, the Post Enumeration Survey, and deaths data).

ABS response

The ABS and AIHW have been working together to understand differences in methodologies and clarify the relationship between ABS and AIHW Aboriginal and Torres Strait Islander life expectancy estimates. As part of this work, guidance for users on the interpretation of Aboriginal and Torres Strait Islander life expectancy estimates will be prepared and released during 2023.

An important part of the ABS work program is planning for future developments and changes that may impact the production of statistics. The ABS will continue to plan to address potential future changes that may impact the delivery of Aboriginal and Torres Strait Islander life expectancy estimates.

13 – 23. Various changes to the structure and content of the ABS website relating to life expectancy estimates for Aboriginal and Torres Strait Islander Australians.

ABS response

The ABS has re-designed the website content aimed at improving the communication and understanding of Aboriginal and Torres Strait Islander life expectancy estimates. The 2020–2022 Aboriginal and Torres Strait Islander life expectancy release includes an accompanying article which describes in plain English how the estimates are created and guidance for how to interpret them.

24 – 28. Recommendations to enhance engagement with expert stakeholders, data users, organisations with reporting requirements, and Aboriginal and Torres Strait Islander organisations. This includes recommendations to enhance membership and use of the ABS Aboriginal and Torres Strait Islander Demographic Statistics Expert Advisory Group (ATSIDSEAG).

29 – 31. Recommendations to enhance engagement with Aboriginal and Torres Strait Islander people to encourage improved participation, utilisation and understanding of all Aboriginal and Torres Strait Islander statistics, increase engagement with the ABS ATSIDSEAG, and increase opportunities for employment in relevant areas of the ABS.

ABS response

The ABS has reviewed and will continue to regularly review the membership and role of ATSIDSEAG, aiming to increase the number of Aboriginal and Torres Strait Islander members over time. The ABS will also continue to engage closely with the AIHW and other stakeholders relevant to the production of Aboriginal and Torres Strait Islander life expectancy estimates.  

The ABS is committed to working with Aboriginal and Torres Strait Islander people through our statistical collections, return of information to community and a diverse workforce. Alongside other relevant areas of the ABS, the ABS Population Statistics Branch is actively pursuing opportunities to increase engagement with Aboriginal and Torres Strait Islander people and increase the number of Aboriginal and Torres Strait Islander staff.