Aboriginal[1] child mortality in Australia: Recent levels and covariates
Table of Contents
The subject of infant and child survival continues to be of great interest to policy makers, health planners, social and bio-medical researchers as well as communities and interest groups around the world. Overall, Australia is one of the healthiest and safest countries for young children to live with an infant mortality rate of less than six per thousand live births and an under-five mortality rate of some seven per thousand children (ABS 2004b). However, the same cannot be said with confidence for Australia’s Indigenous population, where morbidity and mortality among young children remain excessive. Available evidence suggests that, at current rates, as many as one in fifty Aboriginal children will die before reaching age one, while an additional eight per thousand will have no chance of reaching their fifth birthday (Kinfu & Taylor 2005).
While the broad profiles of such excess mortality have been known for some time, far less is known about the precise levels and determinants of high mortality in this population. For instance little, if anything, is known about how socioeconomic factors influence child mortality or how child mortality risks vary according to household, spatial and residential patterns across the country (Gray 1988), information that is vital for public policy making. This dearth of information in part reflects the paucity of relevant health and mortality data on Indigenous Australians. Identification of Indigenous origin in death certificates in Australia did not begin in most jurisdictions until the mid-1980s, and was not introduced until much later in Queensland (1996), while hospital separation records in the country continue to suffer from inconsistency in the way they collect information on Indigenous status. The situation is no different with regard to the census, which provides the denominator for most health and mortality related studies: despite an early introduction, in 1971, of a direct question on Indigenous status, no two successive censuses in the country have yet produced consistent counts of the Indigenous population. The problem is further compounded by the fact that none of these data sources, with the exception of the census, provide detailed information on the social, economic and other background characteristics of the study population, which limit their utility for the study of differentials and determinants of mortality and ill-health in this population.
The 2002 NATSISS obtained, for the first time since the 1986 Census, data on the number and survival status of children born to women in and past reproductive age groups. While the 1994 NATSIS, which preceded the current survey, had obtained information on the number of children ever born, no data were collected regarding the survival status of children. The existence of both the number and survival status of children ever born in the recent survey, therefore, provides a rare opportunity for analysing child mortality among Indigenous Australians. Specifically, this set of data enables the estimation of independent child mortality measures and provides an opportunity to isolate the risk factors associated with child mortality in the population. Such analyses will not only help to identify underprivileged groups who experience higher mortality levels in the population, they will also improve our understanding of the determinants of mortality and their inter-relationships. This will form a basis on which proper policy measures for reducing mortality may be developed, selected and improved.
The balance of the chapter is set out as follows. The next section describes the source data, assesses its quality, and identifies issues of concern for the intended analysis. This is then followed by an examination of recent levels and trends in Aboriginal child mortality and the factors associated with it. The last part summarises key findings of the study and recommends strategies for future demographic data collection on Aboriginal Australians.
The starting point for any demographic analyses is a formal evaluation of the quality of the source data. One way of evaluating the quality of the child mortality data collected in the 2002 NATSISS is by computing the average number of children ever born and the proportion of children who have died by the age group of women. Unless there is significant and abrupt change in the scale and patterns of mortality and fertility schedules, both the average number of children ever born and the proportion of children who have died are expected to increase with the age of women. Fig. 6.1 and 6.2 provide the average number of children ever born and the proportion of children who have died by age group of women and jurisdiction obtained from the 2002 NATSISS, respectively.
As can be seen from Fig. 6.1, while for most States/Territories and age groups the average number of children increases with age, for some jurisdictions the reported number of children for some older women often falls short of that reported for younger women. For instance, for the age group 35–39 years in Queensland, 40–44 years in South Australia and 45–49 years in the Northern Territory, Tasmania and the ACT, the average number of children per woman reported in these age groups is lower than that reported for the immediate younger age groups in the respective jurisdictions. The data on the proportion of children who have died (reported in Fig. 6.2) reveals even more anomalies. With the exception of the data for Victoria and for Australia as a whole, the proportion of children who have died, which is expected to increase linearly with age, depicts a more erratic pattern in all jurisdictions. While part of this unexpected picture could be attributed to sampling, the fact that the problem is also manifested in such jurisdictions as Queensland and New South Wales—where sample size is less of an issue—may point a more structural non-sampling-related quality problem in the data. At any rate, what these suggest is that some care needs to be exercised in using the data for mortality analysis beyond the national level. For this reason, the analyses in the subsequent section will focus on the national level.