Use of Data Linkage Method to Compile Life Tables for Aboriginal and Torres Strait Islander Australians

Abstract

IntroductionAlmost all deaths of Aboriginal and Torres Strait Islander Australians are registered, but some are not identified and recorded as such when registered. Therefore, registration data underestimates the actual number of deaths and death rates of Aboriginal and Torres Strait Islander Australians, and consequently their application can result in overestimates of life expectancy.
 Hence registered deaths needs to be adjusted before deriving life tables and life expectancy estimates.
 Objectives and ApproachObjective: To develop a data-driven method for estimating factors for adjusting registered deaths.
 Approach: The Australian Bureau of Statistics (ABS) linked Census records to all deaths that were registered within a year following the Census. Adjustment factors were estimated by directly comparing Indigenous status of a person reported in Census and death registration data. The ABS then used an established and widely used method to compile life tables, by making up-front adjustments to registered deaths by the factors derived from the Census and deaths linked dataset.
 ResultsAboriginal and Torres Strait Islander deaths identification rate in registration data was estimated to be 89% at the national level in 2016-2017. Over the same period, identification rates varied considerably by state/territory and ranged between 0.69%-1.04%. It clearly showed the need to adjust for under/over identification in Aboriginal and Torres Strait Islander death registrations.
 Conclusion / ImplicationsStatistical data linkage methods can successfully be used to improve Aboriginal and Torres Strait Islander Australians life expectancy estimates. This has implications for accurately measuring and monitoring the Council of Australian Governments’ ‘Closing the Gap’ target in life expectancy estimates between Aboriginal and Torres Strait Islander and non-Indigenous populations.