Although all deaths in Australia are likely to be registered, not all Indigenous deaths are identified as Indigenous when they are being registered. To deal with this problem of under-identification of Indigenous deaths, official life tables for the Indigenous population were based on indirect methods. By looking at the changes in the size of birth cohorts between censuses, these methods estimated cohort-specific mortality during an inter-censal period based on various assumptions about whether or not the population in question was closed to migration. These indirect methods produced inconsistent results. Since 2009, the ABS and AIHW have been producing Indigenous life tables using direct methods.The AIHW method is based on linking death records on the National Death Index (NDI) to corresponding records on the National Hospital Separations Database, the Residential Aged Care Database and the National Perinatal Data Collection. The Indigenous status on the linked records are compared, and an appropriate algorithm developed and used to enhance the Indigenous status on the death records. The enhanced death data are then used to prepare age-specific mortality rates, which then serve as inputs for the preparation of the Indigenous life tables.While Indigenous under-identification is a major problem when producing Indigenous life tables, it is not the only problem. Other problems encountered include:
• the quality of Indigenous identification on the data sets linked
• the quality of some of the input variables on the data sets linked (e.g. name, date of birth etc.)
• the quality of data linkage
• finding additional data sets to link to the NDI in order to enhance the quality of linkage
• lags in death registration
• the choice of a denominator population for the estimation of the age-specific death rates
• numerator-denominator bias
Navigating these issues and finding appropriate solutions to them is a journey in learning.
Tetteh Dugbaza obtained a Ph D in Demography from the Australian National University. Since graduation, he has worked at the Australian Bureau of Statistics, the United Nations Population Fund, and the Australian Institute of Health and Welfare in a wide variety of areas in Demography and Social Statistics.
At the Australian Institute of Health and Welfare, Tetteh has been leading the Enhanced Mortality Database and the Linked Perinatal, Births, Deaths Database projects. The main objective of the Enhanced Mortality Database project is to use data linkage to enhance Indigenous identification on deaths data in order to produce more robust estimates of Indigenous mortality and life expectancy. This work has led to two publications and numerous presentations.For the Linked Perinatal, Births, Deaths Database project, Tetteh and his team are using linked, perinatal, registered births and registered deaths data to model the drivers and differentials in adverse pregnancy and birth outcomes between Indigenous and non-Indigenous women.