Reducing indigenous mortality in Australia: lessons from other countries

Introduction

We report trends in mortality from these causes of death in indigenous people in Australia, New Zealand (NZ), and the United States (US).  

Methods

Australian data
Over the past 10 years, death registration of Aboriginal and Torres Strait Islander people has reached an acceptable level of completeness (90% or more) in two States only -- Western Australia (WA) and the Northern Territory (NT); we have confined our analysis of indigenous death rates to figures from these two States.

The accuracy of Aboriginal and Torres Strait Islander censuses has been investigated.^8,9 Experimental, adjusted counts have been calculated by the Australian Bureau of Statistics (ABS) for 1986-1991¹⁰ and for 1994¹¹ at the State level. Population counts of indigenous people in the 1996 Census were greater in every State than expected from the 1991 Census and adjusted counts, indicating that more people identified themselves as Aboriginal or Torres Strait Islander than in previous censuses.

The 1985-1996 indigenous population counts we used for WA and the NT were based on ABS experimental counts, adjusted for the 1996 Census. In WA, adjusted 1996 population counts were greater than predictions based on experimental counts for earlier years. Consequently, ABS population counts for WA for 1985-1995 were adjusted upwards so that they were compatible with 1996 figures. In the NT, adjusted 1996 counts were consistent with predictions, and ABS experimental counts and predictions were used.

Annual total numbers of deaths for all Australia and for WA and the NT, and corresponding populations, were obtained from the ABS for 1979-1996.

ICD-9 coding to four digits was used for coding cause of death.¹²

US and NZ data
Numbers of deaths by age, sex and year of registration for Native Americans (including Alaskan natives) living on or near reservations were provided by the United States Indian Health Service for 1972-1992. Population counts were based on US census data, with interpolation between census years.

Annual numbers of deaths for Maori for 1974-1994 were obtained from the NZ Health Information Service, and corresponding population estimates were based on census data from Statistics NZ.

Excess mortality
For a given cause, excess mortality was defined as the difference between the observed number of deaths for 1992-1995 for indigenous people in WA and the NT and the expected number of deaths for that cause calculated from death rates for the WA and NT (indigenous and non-indigenous) populations for the same period. Excess deaths for a given cause were expressed as a percentage by dividing the number of excess deaths for that cause by the number of excess deaths from all causes.

Standardised mortality rates
The World Standard Population 1960¹³ was used for directly standardising death rates. We used Poisson regression to fit trend lines to time series of annual mortality rates, and calculated percentage change over time in annual standardised death rates from smoothed, trend-curve estimates rather than observed values.  

Results

All-cause mortality rates
Figure 1 shows annual all-cause mortality rates and fitted trend lines for Australian indigenous people, Maori, Native Americans and all Australians. The mortality rate from all causes for indigenous people in WA and the NT fell by 9% between 1985 and 1996. Over the same period, the all-cause rate for all Australians fell by 26%. The indigenous rate was 2.5 times the all-Australian rate in 1985 and three times the rate in 1996. The "hump" in the mortality rates for Australian indigenous people could represent a real increase to the 1990s followed by a decrease, but is more likely to be an artifact due perhaps to changing rates of registration of indigenous deaths.

Death rates for NZ and US indigenous people fell relatively rapidly in the 1970s and more slowly thereafter. Between 1974 and 1984 mortality rates in Maori fell 32%, and a further 19% between 1984 and 1994. Mortality rates in Native Americans fell 17% between 1973 and 1982 from levels some 30% lower than those for Maori, and a further 11% between 1982 and 1992. Current mortality rates in Australian indigenous people are comparable with rates in Maori in the early 1970s.

By 1990-1994, the average Australian indigenous all-cause mortality rate was 1.9 times the Maori rate, 2.4 times the US indigenous rate and 3.15 times the all-Australian rate.

Main causes of excess mortality in Aboriginals and Torres Strait Islanders
The ICD-9 coding system groups all causes of death into 17 broad categories. Four groups of conditions accounted for almost 70% of the total excess deaths in the Aboriginal and Torres Strait Islander population of WA and the NT during the five-year period 1992-1996. Circulatory conditions accounted for over a quarter (26%) of all excess deaths, with ischaemic heart disease, cerebrovascular disease and hypertension accounting for most of the circulatory disease excess; the injury and poisoning group, principally transport accidents, homicide and suicide, accounted for 15%, and respiratory conditions, including chronic obstructive airway disease and pneumonia, accounted for 16%; endocrine conditions, largely diabetes, caused a further 10% of excess deaths.

International comparisons of mortality trends for the major causes of excess mortality
Trends in annual standardised mortality rates for the four main causes of excess deaths in indigenous Australians are presented in Figure 2. (Data shown are for Australian indigenous people [1985-1996], Maori [1974-1994], Native Americans [1974-1992], and all Australians [1979-1996].)

Circulatory disease: Current mortality rates for circulatory diseases in Aboriginals and Torres Strait Islanders were 50% higher than corresponding rates for Maori (the next-highest), and 2.5 times the all-Australian rate (Table, below). In the early 1970s, these mortality rates in Maori were above current Australian indigenous rates, but fell rapidly over the ensuing 20 years. Circulatory disease mortality rates in Native Americans have been lower than those in the other groups for the entire period, while the all-Australian mortality rates have held an intermediate course. Mortality rates from ischaemic heart disease in Maori fell in 1974 from above contemporary Australian indigenous rates to a third below by 1990-1994. By 1990-1994, mortality rates from cerebrovascular disease in Maori, after being equivalent in 1974, were down to less than half the corresponding Australian indigenous rates.

Respiratory diseases: Mortality trends in respiratory system diseases were generally similar to those of the circulatory system, with Maori rates being initially very high (at the level of current Aboriginal and Torres Strait Islander rates) and then falling rapidly to levels not greatly exceeding current all-Australian levels. The overall reductions in rates of diseases of the respiratory system in Maori have been due largely to reductions in asthma and pneumonia deaths, which accounted for most respiratory system deaths in 1990-1994. In Australian indigenous people, other chronic obstructive pulmonary diseases and pneumonia accounted for a similar proportion of respiratory disease deaths in the same period.

Injury and poisoning: The pattern here has been somewhat similar, although the relative positions of Native Americans and Maori are reversed. Death rates in Native Americans and Alaskan Natives were one and a half times the Australian indigenous injury and poisoning death rates in the early 1970s, but US rates have now fallen to below the current Australian level. For Aboriginals and Torres Strait Islanders in WA and the NT, there is little evidence of an overall decline (Figure 2). The decline in mortality rates from injury and poisoning in Native Americans has been driven particularly by changes in transport accidents, but also changes in homicide and suicide rates. For Australian indigenous people in WA and the NT, there appear to have been some relatively small recent falls in homicides and transport accident deaths, but there is some evidence that suicide rates are rising.

Endocrine diseases: There has been no substantial improvement in mortality from endocrine diseases (largely diabetes). Death rates from diabetes in Maori fell in the 1970s, but have been rising slowly since the mid 1980s. Diabetes mortality rates in Native Americans have been rising slowly for 20 years. The rate for Australian indigenous people has also risen since the mid 1980s, but much more rapidly. In 1990-1994, the average mortality rate in Australian indigenous people for endocrine diseases was 2.4 times the Maori rate, 3.2 times the Native American rate and eight times the Australian rate (Figure 2).  

Discussion

Several factors may influence the validity of comparisons among these three indigenous groups. A fundamental cause of non-random variation in annual mortality rates for indigenous groups is differences in factors influencing identification of a death as an indigenous death, and those determining whether people identify themselves as indigenous at census time. Special efforts to increase completeness in the 1996 Census resulted in a sharp rise in Australian indigenous population counts. However, similar improvements in ethnic identification on death certificates had not occurred by 1996. Indigenous population sizes over the 12 years studied were adjusted upwards to reflect the 1996 Census increase, which means that the Australian indigenous rates are likely to be lower than their true values.

For all three indigenous groups, deaths were grouped by year of registration rather than year of death, so that they could be compared on a common basis. The extent of year-to-year variation in the proportion of deaths in a given year that were also registered in that year is unknown for Maori and Native Americans. Such variation would add a component of extra-Poisson variability to annual rate figures, but is unlikely to influence long term trends. The degree of late registration of Australian indigenous deaths is known to be relatively large compared with non-indigenous deaths and to vary from year to year. For this reason, apparent trends in Australian indigenous rates may be influenced to some degree by error in endpoint rate values.

Different cause-of-death coding practices may have existed in the three comparison countries over time, leading to different apparent mortality rates for certain diseases. However, comparability of New Zealand and Australian non-indigenous rates suggests that there has been little difference between coding practices in these two countries over time. Furthermore, all-cause rates, not subject to coding variations, display the same patterns and differentials between ethnic groups observed for the separate causes of death.

Although the three indigenous groups each have a different heritage and cultures, they share common experiences in their history. They are minority cultures in affluent nations dispossessed of their country and marginalised. However, Maori and Native Americans have made rapid gains in health and life expectancy over the past two decades. This progress is characterised by an initial rapid fall in death rates, followed by a more gradual decline as levels of the non-indigenous population are approached. Australian indigenous mortality shows little or no evidence of this pattern for any of the major causes of excess deaths. For some causes there is evidence of a recent decline, but no significant trend has yet been established.

The health problems of Australia's indigenous people, and the circumstances responsible, are not unique. Why, then, has the health of Australia's indigenous people failed to match the improvement seen in other countries? Kunitz¹⁴ identified several aspects of Australian postcolonial indigenous experience which he felt contributed to the lack of improvement in health. He argues that Aboriginal affairs have traditionally been first a colonial then a State responsibility; and that there has been constant tension between State interests, particularly relating to land and management of the indigenous population. Furthermore, an official policy of assimilation maintained over many decades discouraged the creation of specialised indigenous health programs. In more recent times, competition for federal health funding by State and non-State bodies has meant that efforts to improve indigenous health have lacked coordination. Another factor, he suggests, is a lack of a sense of a single identity in many present-day indigenous communities, which were created artificially by gathering together people from many different tribal groups.

The Treaty of Waitangi has been central to the relationship between Maori and other New Zealanders, and in the United States treaties established some status for Native Americans in their relationships with the "invading" Europeans, although these treaties were often abused. It is therefore difficult to entirely discount the suggestion that the absence of a treaty is a factor in the relative lack of progress in improving Australian indigenous health. Treaties, no matter how loosely worded, have appeared to play a significant and useful role in the development of health services, and in social and economic issues, for the indigenous people of New Zealand, the United States and Canada.

Syme¹⁵ has hypothesised that the sense of control that people have over their lives and the sense of hope that this creates are important determinants of health status. He links both of these factors to disadvantage. It is arguable that colonial paternalism, an official policy of assimilation, and a lack of formal recognition through treaties have together acted to create and reinforce a sense of powerlessness in Australian indigenous people which is relatively less in other indigenous groups. It may well be that these intangible factors are adversely affecting indigenous health in Australia.

The experience of other indigenous people around the world generates considerable confidence that effective action in Australia will produce substantial changes in indigenous health. Progress is needed in five key areas: infrastructure (land, housing, water supply, education, income, etc), self-determination of health services, access to a network of community-controlled primary healthcare services delivering effective health services for priority issues, an adequate level of resources, and a skilled workforce.¹⁶ Future success requires much greater progress within each of these five areas than has occurred in the past. A greater sense of control may only come from a wider acceptance and recognition of a valued role for Australian indigenous people in Australian society.  

References

1. Kunitz SJ, Streatfield R, Santow G. Health of populations in northern Queensland Aboriginal communities: change and continuity. Hum Biol 1994; 66: 917-943.
2. Veroni M, Gracey M, Rouse I. Patterns of mortality in Western Australian Aboriginals, 1983-1989. Int J Epidemiol 1994; 23: 73-81.
3. Hogg RS. Indigenous mortality: placing Australian Aboriginal mortality within a broader context. Soc Sci Med 1992; 35: 335-346.
4. Ring I. The Elkington Oration. Inequalities in health, the challenge for the nineties. Brisbane: Queensland Health, 1993.
5. Runciman C, Ring I. The health of indigenous people in Queensland, some background information. Brisbane: Queensland Health, 1994.
6. Australian Bureau of Statistics/ Australian Institute of Health and Welfare. The health and welfare of Australia's Aboriginal and Torres Strait Islander peoples. Canberra: ABS/AIHW, 1997 (Catalogue No. 4704.0.)
7. Australian Institute of Health and Welfare. Australia's Health 1996. Canberra: AGPS, 1996.
8. Australian Bureau of Statistics. Estimates of the Aboriginal population: review of data sources. Canberra: ABS, 1993. (Demography working paper: 93/2.)
9. Gray A, Tetfaghiorghis H. Aboriginal population prospects. J Aust Popul Assoc 1993; 10: 81-99.
10. Australian Bureau of Statistics. Experimental estimates of Aboriginal and Torres Strait Islander population (June 1986 to June 1991). Canberra: ABS, 1994. (Catalogue No. 3230.)
11. Australian Bureau of Statistics. National Aboriginal and Torres Strait Islander Survey 1994. Detailed findings. Canberra: ABS, 1995. (Catalogue No. 4190.0.)
12. The Australian version of the International Classification of diseases, 9th revision, clinical modification (ICD-9-CM). 2nd edition. Sydney: National Coding Centre (NCC), 1996.
13. Waterhouse J, Muir C, Correa P, Powell J, editors. Cancer incidence in five continents. Vol III. Lyon: International Agency for Research on Cancer, 1976. (IARC Scientific Publications No. 15)
14. Kunitz SJ. Disease and social diversity: the European impact on the health of non-Europeans. New York: Oxford University Press, 1994.
15. Syme SL. Individual vs. community interventions in public health practice: Some thoughts about a new approach. Health Promotion Matters (Vic Health) July 1997; (2): 2-12.
16. Ring I. An open letter to the President of the Public Health Association. Aust J Public Health 1995: 19: 228-230.