• Open Access

Avoidable mortality trends in Aboriginal and non-Aboriginal populations in the Northern Territory, 1985-2004


Correspondence to:
Shu Qin Li, Health Gains Planning, NT Department of Health and Families, PO Box 40596, Casuarina NT 0811. Fax: (08) 8985 8075; e-mail: shu.li@nt.gov.au


Objectives: To analyse rates of avoidable mortality in Aboriginal and non-Aboriginal residents of the Northern Territory (NT) from 1985 to 2004, in order to assess the contribution of health care to life expectancy improvements.

Methods: Australian Bureau of Statistics (ABS) death registration data for NT residents were used to identify ‘avoidable’ deaths, with further separation into three categories of conditions amenable to either medical care or health policy, and a category for ischaemic heart disease (IHD). A Poisson regression model was used to calculate the average annual change in avoidable mortality by sex and Aboriginality in the NT compared with Australia as a whole.

Results: In the 20 years between 1985 and 2004, avoidable mortality rates fell 18.9% in NT Aboriginal people, 61.1% in NT non-Aboriginal people and 59.5% in Australians overall. NT Aboriginal people continued to experience higher avoidable mortality than other Australians and the disparity increased over time. Most of the decline in avoidable mortality for Aboriginal Territorians occurred for conditions amenable to medical care.

Conclusion: Medical care has made a significant contribution to improvements in Aboriginal life expectancy in the NT, however, reductions in avoidable mortality from IHD and conditions amenable to health policy have been variable.

Implications: The results highlight the need for ongoing investment in comprehensive programs incorporating appropriate health policy interventions and management of chronic diseases.

In the 17 years to 2000, life expectancy in Australia increased from 79 to 83 years for women and from 72 to 78 years for men.1 Although considerable attention has justifiably been given to the substantial gap in life expectancy between non-Aboriginal and Aboriginal Australians, a recent article demonstrated that for the Northern Territory (NT) there has been a significant improvement in Aboriginal life expectancy since the 1960s.2 Between 1967 and 2004, life expectancy at birth in the NT increased from 54 to 68 years for Aboriginal women, and from 52 to 60 years for Aboriginal men. For Aboriginal women, this represented a 5.6 year reduction in the gap between their life expectancy and that of all Australian women.

There is also evidence that despite an emerging epidemic of chronic disease in the NT Aboriginal population,3 the increase in Aboriginal mortality rates from ischaemic heart disease (IHD) and diabetes has slowed since 1990 and the mortality rate from chronic obstructive pulmonary disease (COPD) has fallen.4 While these figures give some cause for optimism, from a health service planning perspective it is useful to separate the contribution of the health care system, from the more general contribution of social determinants of health, to these outcomes.

The concept of ‘avoidable mortality’ builds on Rutstein's pioneering work in 1976 on ‘amenable mortality’, and the extensive work on ‘preventable mortality’, which recognised that some conditions were responsive to interventions such as health care, lifestyle change and environmental modification.5–7 Avoidable mortality provides a methodology to assess the extent to which declining mortality rates can be attributed to health care including medical interventions, public health programs and population-based health strategies. This approach labels as ‘avoidable’ deaths from conditions for which death could be avoided by timely and effective health care.8 It needs to be acknowledged that the sensitivity of this method may vary with the selection of causes of death, and that there is currently no ‘gold standard’ of causes that are used when analysing this population health indicator.7 Although debate about the list of conditions that constitute avoidable mortality has limited the capacity for international comparisons of the effectiveness of different approaches to health care,8–10 it remains a useful epidemiological tool to compare trends in avoidable and non-avoidable mortality.

A recent analysis of avoidable mortality in Australia reported that between 1968 and 2001 health care played a significant role in overall mortality reductions, with avoidable mortality rates falling 70% and non-avoidable mortality rates falling 34%.8 However, the benefits of health care have not been distributed equally throughout the population, with people of higher socio-economic status receiving disproportionate benefit, leading to a widening of health inequalities.11 Overseas data suggest a similar disparity in avoidable mortality for certain ethnic groups, with Māori and Pacific Islands people in New Zealand experiencing avoidable mortality rates more than twice those of the total population10 and migrants in the Netherlands also experiencing higher mortality from conditions deemed to be avoidable.12 To date, there has been little information available on avoidable mortality rates for Aboriginal Australians.

Vigorous and collective efforts have been made over the past two decades to tackle excess mortality among the NT Aboriginal population, including the development of the NT Preventable Chronic Disease Strategy, which has prioritised the prevention and management of chronic diseases.13 This study aimed to analyse avoidable mortality rates in Aboriginal and non-Aboriginal Territorians, examine trends over time, and measure the impact of health care on mortality reductions in the NT in order to assess the success of current programs and to inform continuing efforts to reduce Aboriginal mortality.



This study used Australian Bureau of Statistics (ABS) death registration data for both NT and Australia. For the NT Aboriginal population the coverage and identification of deaths are known to be close to 100% from 1988 when Indigenous status was first recorded on NT death registration forms.2,14 For the years 1967 to 1987, extensive data validation has been previously performed manually for each individual record, and where necessary, Indigenous status was derived from names, residence, place of birth and other information from death registrations forms.2,15 The near complete identification of Indigenous deaths in the NT mortality data since the 1960s provided reliable data for long-term trend analysis. NT population data were based on the ABS 2001 Census.16

The ABS codes the underlying cause of death using the International Classification of Diseasesand Related Health Problems, 9th revision (ICD-9-CM)17 for deaths registered from 1981 to 1996, and the 10th revision (ICD-10-AM)18 for deaths registered from 1997 onwards. To ensure comparability with other Australian avoidable mortality analyses,8 an age limit of 74 years was applied for this study.

Following Korda and Butler,8 avoidable deaths were classified into three categories (Appendix A): conditions amenable to medical care (such as asthma, pneumonia and breast cancer), conditions responsive to population-based health policies (smoking, alcohol consumption and motor vehicle accidents) but considered to lack effective treatment once the conditions have developed (such as lung cancer, chronic liver disease and cirrhosis, and motor vehicle accidents), and ischaemic heart disease (IHD), which is considered partly amenable to medical care but primarily responsive to health policy. Deaths from remaining conditions were classified as ‘non-avoidable’ by health care interventions.

Statistical analysis

Annual avoidable and non-avoidable mortality rates by year of death, sex and Indigenous status were calculated from 1985 to 2004, and age-adjusted using the 2001 Australian estimated resident population.

Poisson regression was used to estimate the average annual changes in avoidable mortality by sex and Aboriginality in the NT, compared with the overall Australian population. The Poisson regression model was further used to estimate the annual change in avoidable mortality by the three categories and major conditions of interest, such as diabetes and stroke. Various interaction terms were added to the Poisson regression model to test differences in annual changes between NT Aboriginal and all Australians, and between NT non-Aboriginal and all Australians.

Stata software (version 8; Stata Corp, College Station, Tex, US) was used for all statistical analysis. Ethics approval was granted by the Human Research Ethics Committee of the NT Department of Health and Families and the Menzies School of Health Research.


Overall trends in avoidable mortality

During the 20-year period from 1985 to 2004, 6,411 NT residents died from conditions that were considered to be avoidable (Table 1). These comprised 40.5% of total NT deaths. Although Aboriginal people account for 28% of the NT population, they experienced 54% of total avoidable deaths.

Table 1.  Number of avoidable deaths by Aboriginality, Northern Territory, five year period, 1985 to 2004.
YearNT AboriginalNT non-AboriginalTotal

Time trends in avoidable and non-avoidable mortality for the NT and all Australians are shown in Figure 1. Avoidable mortality rates in the NT population were consistently higher, and fell more slowly, than those in all Australians during the study period. The avoidable mortality rates in NT residents fell from 371 to 195 deaths per 100,000 person years (47.6%) and the rates in all Australians fell from 269 to 109 deaths per 100,000 person years (59.6%). Using the Poisson regression model, the average annual reduction in avoidable mortality was 3.5% (95% CI: 3.11%-3.92%) in the NT and 4.6% (95% CI: 4.51%-4.60%) in all Australians. This difference in average annual reduction was statistically significantly (p<0.05). The average annual reduction for non-avoidable mortality was significantly higher (2.6%, 95% CI: 2.16%-2.98%) for NT residents than for all Australians (1.5%, 95% CI: 1.47%-1.56%).

Figure 1.

Avoidable and non-avoidable mortality, NT and Australia, 1985 to 2004

Trends in total avoidable mortality by Aboriginality and sex

Avoidable mortality rates declined over time in all population groups (Table 2), however the rates in NT Aboriginal people started at nearly three times the comparable Australian rates, and the disparity increased over time. Avoidable mortality rates fell from 638 in 1985 to 517 deaths in 2004 per 100,000 person years (18.9%) in Aboriginal Territorians, from 290 in 1985 to 113 in 2004 deaths per 100,000 person years (61.1%) in non-Aboriginal Territorians, and from 269 in 1985 to 109 in 2004 deaths per 100,000 person years (59.5%) in all Australians.

Table 2.  Age-adjusted avoidable mortality rate by sex and Aboriginality, NT and Australia, five year period, 1985 to 2004.
 NT AboriginalNT non-AboriginalAustraliaRatio
 Rate95% CIRate95% CIRate95% CINT Aboriginal to AustraliaNT non-Aboriginal to Australia
Number of deaths per 100, 000 population

The average annual reduction in avoidable mortality rates was 1.7% (95% CI: 1.10%-2.23%) in the total NT Aboriginal population with Aboriginal males experiencing a fall of 1.8% (95% CI: 1.03%-2.51%) and Aboriginal females a fall of 1.4% (95% CI: 0.56%-2.33%). This compares with an estimated average annual avoidable mortality reduction for the NT non-Aboriginal population of 4.7% (95% CI: 4.10%-5.25%). The overall Australian average annual reduction in avoidable mortality rate was 4.5% (95% CI: 4.47%-4.56%), comprising 4.8% (95% CI: 4.73%-4.84%) in males and 4.2% (95% CI: 4.11%-4.25%) in females. The differences in average annual reduction rate between NT Aboriginal population and other Australians were statistically significant (p<0.05).

Trends in three categories of avoidable mortality

Throughout the study period, mortality rates from conditions amenable to medical care were four to five times higher in Aboriginal Territorians than Australians as a whole. These rates declined substantially from 1985 to 2004 in all population groups (Table 3). The estimated average annual reduction in the avoidable mortality rate for these conditions for Aboriginal and non-Aboriginal Territorians was 3.2% (95% CI: 2.45%-3.99%) and 3.8% (95% CI: 2.76%-4.77%) respectively, compared with 3.6% (95% CI: 3.57%-3.70%) for all Australians. These differences were not statistically significant (p>0.05). The decline in mortality rates for medical care conditions was observed for most common conditions, with the notable exception of some chronic diseases including diabetes, nephritis and nephrosis.

Table 3.  Age-adjusted avoidable mortality rate from conditions amenable to medical care, by sex and Aboriginality, five year period, NT and Australia, 1985 to 2004.
 NT AboriginalNT non-AboriginalAustraliaRatio
 Rate95% CIRate95% CIRate95% CINT Aboriginal to AustraliaNT non-Aboriginal to Australia
Number of deaths per 100, 000 population

Mortality rates from avoidable conditions responsive to health policy were three to four times higher for Aboriginal Territorians than for Australians as a whole. This ratio persisted throughout the study period. Further, while mortality rates from these causes declined consistently over time in non-Aboriginal Territorians and in all Australians, with a narrowing of the gap between non-Aboriginal Territorians and all Australians, there was only marginal change in the NT Aboriginal population (Table 4). For Aboriginal males there was a minor improvement, while for Aboriginal females there was an initial increase in mortality rates, with no overall improvement for Aboriginal women across the entire 20 years.

Table 4.  Age-adjusted avoidable mortality rate from conditions amenable to health policy, by sex and Aboriginality, five year period, NT and Australia, 1985 to 2004.
 NT AboriginalNT non-AboriginalAustraliaRatio
 Rate95% CIRate95% CIRate95% CINT Aboriginal to AustraliaNT non-Aboriginal to Australia
Number of deaths per 100, 000 population

Table 5 demonstrates a steady decline in Australian and NT non-Aboriginal mortality from IHD between 1985 and 2004. In contrast, IHD mortality in the NT Aboriginal population increased between 1985 and 1994, by 23.9% for males and 33.3% for females. This was followed by a marginal decline of 2.2% for Aboriginal males and 5.3% for Aboriginal females between 1995 and 2004. This has led to a substantial increase in the gap in mortality rates from IHD between Aboriginal Territorians and both non-Aboriginal Territorians and all Australians.

Table 5.  Age-adjusted avoidable mortality rate from ischaemic heart disease by sex and Aboriginality, NT and Australia, five year period, 1985 to 2004.
 NT AboriginalNT non-AboriginalAustraliaRatio
 Rate95% CIRate95% CIRate95% CINT Aboriginal to AustraliaNT non-Aboriginal to Australia
Number of deaths per 100, 000 population


The results of this analysis confirm that, while Aboriginal Territorians continue to experience significantly higher avoidable mortality than non-Aboriginal people, the health care system has made a considerable contribution to the fall in NT Aboriginal mortality rates between 1985 and 2004. This is demonstrated by both the substantial decline in deaths from conditions within the medical care category and by the more rapid fall in avoidable than non-avoidable causes of death.8 Nevertheless, the reduction in avoidable mortality rate for Aboriginal Territorians was less than one-third that of non-Aboriginal Territorians confirming that improved outcomes have not been shared equally.

The largest impact on avoidable mortality in NT Aboriginal people was for conditions amenable to medical care. This improvement is consistent with the significant improvement in perinatal survival, and the reduction in mortality from conditions such as congenital malformations, stroke and hypertensive conditions, pneumonia and asthma, and infectious diseases. A broad range of medical care interventions such as an increased number of births occurring in hospitals,19 improved neonatal and paediatric care, and the establishment of prenatal screening for congenital abnormalities, have contributed to the decline in infant mortality in the NT; while the substantial decline in deaths from pneumonia and other infectious diseases are consistent with national trends and likely relate to the use of effective antibiotics and high rates of immunisation uptake.8 The reduced mortality rates from stroke and hypertensive conditions are consistent with comprehensive management including effective drug therapies, improved intensive care, the development of dedicated ‘stroke units’, and surgical procedures such as carotid endarterectomy. Despite its small primary and specialist health care workforce, the NT achieved equivalent reductions in avoidable mortality from medical care indicators to the rest of Australia. The improvements did not occur for all conditions, and there remains substantial potential for continuing improvement, most notably through improvement associated with diabetes. Although not statistically significant, the rate reduction in conditions amenable to medical care was lower in NT Aboriginal people than for other Australians, implying that they are benefiting less from available treatments. This has implications for improvement of access, more culturally appropriate service delivery and more sustainable staffing.

The fall in avoidable mortality nationally, and in the NT non-Aboriginal population, has primarily been driven by a reduction in mortality attributable to IHD and health policy. This has occurred by the combination of health policy interventions including campaigns to reduce smoking and improve diet, along with medical care interventions such as improved control of cardiovascular risk factors and specialised services such as acute coronary care.8 In contrast, both male and female Aboriginal Territorians experienced an overall increase in avoidable mortality from IHD between 1985 and 1994. The reasons for this are likely to be multifactorial and include a higher prevalence of smoking among the NT Aboriginal population as well as barriers to accessing primary and acute medical care.4,20–23 Encouragingly, since 1995, and particularly since 1999, mortality from IHD in Aboriginal Territorians has started to decrease. This corresponds to the period of development and implementation of the NT Preventable Chronic Disease Strategy, which combines health promotion and primary health care interventions to improve outcomes for Aboriginal people.13

This study also suggests that Aboriginal Territorians have benefited less from the impact of health policy interventions than non-Aboriginal Territorians and Australians overall. Although the avoidable mortality rates from conditions responsive to health policy fell significantly among Australians overall, and almost halved for NT non-Aboriginal Territorians, the rates fell only slightly for NT Aboriginal men and, increased slightly for NT Aboriginal women. This increase was particularly marked between 1985 and 1995 and was largely driven by an increased mortality rate for lung cancer.22 The increase in lung cancer mortality for Aboriginal women occurred at the same time as the increase in mortality from IHD for Aboriginal women, both are likely due to the increasing prevalence of smoking. The NT has the highest rates of smoking and alcohol consumption in Australia.19,21,22,24 Although the prevalence of smoking and alcohol consumption is higher for Aboriginal than non-Aboriginal Territorians, the latter group continue to engage in these risk behaviours more frequently than Australians overall.19,21 The fact that, despite this, avoidable mortality from health policy conditions halved for non-Aboriginal Territorians suggests that policies to prevent the health sequelae of risk behaviours were better targeted towards, or perhaps more culturally appropriate for, non-Aboriginal than Aboriginal people. Aboriginal women in particular seem to have not benefited from these programs. This again highlights the need for culturally appropriate public health policies incorporating comprehensive anti-smoking campaigns, targeted alcohol interventions and reduction in road crashes. To ensure the success of public health policies and programs, collective efforts from all other sectors, and at all levels are required.

There are a number of limitations to this study. Firstly, the death registration data from the ABS used for the analysis code only underlying cause of death. This means that other contributing causes of death are not included, leading to the possibility that chronic diseases such as renal failure that are unlikely to be recorded as the underlying cause of death may be under-represented. A second limitation is that approximately 6% of deaths that occurred in 2004 were not available at the time of the analysis due to the late registration, however, as the analysis was mainly based on the Poisson regression model examining trends based on the whole 20-year dataset, the effect on the analysis is negligible. The decrease in non-avoidable mortality in the NT compared with little change for total Australia may relate to some Territorians with serious chronic disease choosing to leave the Territory for medical or social reasons. With increased time from removal, these people are less likely to be identified as NT residents on the death registration forms. This possible selective emigration is more likely in the NT non-Aboriginal population than Aboriginal population, who have minimal levels of interstate migration. This issue warrants further study by linking NT hospital separation records to the National Death Index.

The classification of avoidable conditions is also subject to limitations. Although some conditions can clearly be labelled as ‘medical care indicators’, the outcomes of many conditions are affected by both medical care and health policy. This is currently only recognised for IHD, which is analysed separately. However, potential examples are cerebrovascular disease (which is amenable to medical treatment and smoking policy) and perinatal outcomes (which may partially reflect maternal smoking and alcohol behaviours during pregnancy). It is also possible that a broader range of health policies, such as those concerning healthy eating and physical activity, impact on mortality for conditions such as IHD and diabetes. The inability to separate these effects for all conditions may have led to an underestimation of the importance of health policy in generating reductions in avoidable mortality. The New Zealand literature attempts to overcome this limitation by attributing to each condition the proportion of mortality that can be avoided by primary, secondary and tertiary prevention.10 Although this method is complex and currently controversial, its further development may help clarify the specific aspects of health care that can best contribute to falls in avoidable mortality.

Despite the limitations, this study demonstrates that health care, particularly medical care, has had a significant impact in reducing avoidable mortality for both Aboriginal and non-Aboriginal Territorians over the period from 1985 to 2004. Given the challenges of delivering primary and specialist health care to remote areas of the NT, this is a significant achievement.25 Important exceptions and disparities remain, with IHD, diabetes and conditions responsive to health policy such as lung cancer, chronic liver disease and motor vehicle trauma still contributing significantly to the excess burden of mortality among the NT Aboriginal population. It is well-recognised that many social and environmental factors such as low educational levels and employment status, lying outside the health care boundary also contribute to mortality from IHD and other conditions amenable to medical care and health policies. Therefore, control measures warrant a systematic approach to tackle the root of excess mortality. The decline in non-avoidable mortality in the NT over this time period also suggests that improvements in social conditions have a continued role to play in reducing the disparity between Aboriginal and non-Aboriginal people, and highlights the importance of multi-sectoral approaches to improve Aboriginal health.

  • image(AppendixA:)

[ Avoidable causes of death.8 ]