Connect
MJA
MJA

Burden of disease and injury in Australia in the new millennium: measuring health loss from diseases, injuries and risk factors

Stephen J Begg, Theo Vos, Bridget Barker, Lucy Stanley and Alan D Lopez
Med J Aust 2008; 188 (1): 36-40. || doi: 10.5694/j.1326-5377.2008.tb01503.x
Published online: 7 January 2008

Information on the magnitude and distribution of health problems in a population is important for health policy decision making. Popular epidemiological measures such as mortality, incidence and prevalence are available for many health problems, but can not be compared across causes as indicators of population-level health. Summary measures of population health, on the other hand, extend the utility of descriptive epidemiology by combining information on mortality and non-fatal health problems into a common measure that can be used to provide a comprehensive picture of the health status of a population.1

We present here a reanalysis of a large body of work that used summary measures to describe the health of Australians in the new millennium.2 The research on which it is based follows a comparable study for the year 1996 reported in the Journal in 2000.3 Both studies use a particular summary measure — the “disability-adjusted life year” (DALY) — to quantify health loss from a comprehensive set of diseases, injuries and health risks of public health importance in Australia. The DALY, in turn, has its origins in an assessment of global health for the World Bank.4,5 One DALY is equivalent to one lost year of healthy life and represents the gap between current health status and an ideal situation of the whole population living into old age in full health. This gap is referred to here as “health loss”, rather than the less accurate but more commonly used term “burden of disease”.

The DALY combines the descriptive epidemiology of each health condition of interest with a multidimensional numerical weighting for the severity of that condition. As the weighting given to each dimension implies a judgement about its relative importance to the total measure, the DALY has obvious normative characteristics that make it not necessarily compatible with other classifications of health (eg, the World Health Organization’s International classification of functioning, disability and health6). For this reason, others have highlighted the importance of limiting interpretation of the DALY to the specific purposes for which it is being used7 — which, in this case, is as a comparative measure of health loss.

Our article provides an assessment of the magnitude and distribution of health problems in Australia in order to identify key opportunities for health gain. Our specific objectives were to calculate:

Methods

Health loss was estimated for a comprehensive set of diseases and injuries of public health importance in Australia, using DALYs as the outcome measure. Diseases and injuries were the smallest reported unit of disaggregation, and are referred to here as “specific causes” or “conditions”. Each is mutually exclusive and belongs to one of 22 “broad cause groups”, most of which correspond to chapter-level headings of the International classification of diseases.8 Each broad cause group, in turn, belongs to one of three broad clusters: (a) communicable, maternal, neonatal and nutritional conditions; (b) non-communicable diseases; and (c) injuries. Further details on methods and assumptions are provided elsewhere.2

DALY estimates

DALYs were calculated by applying severity weights (range, 0–1) to the estimated number of incident cases and average duration for each condition. Weights were derived from two sources,4,11 with extrapolations based on alternative methods in some cases. Adjustments were made to account for the possibility of two or more conditions occurring simultaneously in the same person, either by chance or because the conditions are related. These corrections were achieved by determining numbers of people for every combination of causes of ill health as measured by various surveys and hospital admission data.

Health risk assessment

Past and current exposure to 14 selected risk factors (listed in Box 3) were analysed for their contribution to health loss in 2003. Analyses were based on the theoretical framework developed for the WHO-initiated Comparative Risk Assessment project.12 This approach incorporates a “hypothetical minimum” as the alternative exposure distribution against which health loss is calculated, and uses continuous rather than categorical measures of exposure where appropriate. Results were also calculated for the combined effect of health risks.

Results

Key findings are presented here at two levels of aggregation: “broad cause groups” and “specific conditions”. Both levels are referred to as “causes” and are ranked in terms of “leading” causes compared with others at the same level of aggregation.

Discussion

Our findings emphasise that, despite steady improvements in Australia’s health over the past decade, significant opportunities for further progress remain at the beginning of the 21st century.

The strength of our analysis is that it is based on an internally consistent assessment of the incidence, prevalence, duration and mortality for a mutually exclusive and comprehensive set of diseases and injuries of importance in Australia. Health loss from these causes was quantified for different periods, subpopulations and risks to health using methods that incorporate fatal and non-fatal health outcomes and include adjustments to account for individuals who simultaneously experience multiple conditions. Health loss is likely to be over-estimated without such corrections, as the severity weights used to derive DALYs were originally determined for health states in isolation, without reference to coexisting conditions.13

A potential limitation is that the severity weights used in our analysis were derived from international sources4,11 and applied without evidence of their validity in Australia. However, studies conducted elsewhere suggest that there are only minor variations across populations in the values people ascribe to different health states.4

We have not quantified uncertainty in our analysis, although a qualitative assessment suggests it is unlikely to be excessive. Overall, about half of the total estimated health loss is due to mortality, for which estimates are fairly robust. Of the remainder, half is due to non-fatal outcomes from conditions for which reasonably good data are available (including cardiovascular disease, cancers, diabetes, common mental disorders and injuries), leaving a quarter with varying and probably higher levels of uncertainty. Precision varies between causes, with estimates for hearing loss, neurological conditions, osteoarthritis and cirrhosis being the most inaccurate.

Our results are not directly comparable with previous DALY estimates for Australia,3 owing to the different methods used. First, a number of the epidemiological models in our analysis benefit from more accurate inputs, particularly the cardiovascular disease models, which incorporated linked data from Western Australia. Second, unlike in the previous analysis, the comorbidity adjustments here capture the dependent nature of certain health states (eg, diabetes increases the risk of heart disease). Third, the current risk attribution methods incorporate a number of methodological advances absent from previous health risk analyses.3,14,15 Because of this lack of comparability, we back-calculated estimates for 1993 based on methods that were consistent with estimates for 2003.

Several implications for policy are worth emphasising. All of the health risks examined here are amenable to modification through intervention, and together explain a large proportion of health loss in Australia. In addition, the large health differentials between subpopulations are due, in part, to differential exposure to these risks. Significant health gains are likely to be achieved through realistic changes to future levels of exposure to health risks, given that even small changes in distribution of exposures can lead to substantial reductions in population-level risk.16

The predicted strong growth in DALY rates associated with diabetes is notable in that it is mostly due to increasing body mass. Given that current strategies have failed to mitigate this risk, new approaches are critical. The impact of increasing diabetes incidence will be magnified by reductions in case fatality from cardiovascular disease through successful strategies to reduce smoking and lower cholesterol levels and blood pressure.2,17 Increased survival will result in a greater number of people with diabetes developing other health conditions such as renal failure, retinopathy, neuropathy and peripheral vascular disease. Notwithstanding the apparent intractability of diabetes, further reductions in cardiovascular disease could be achieved, given that most of the health loss from this condition continues to be explained by exposure to known health risks.

The much higher DALY rates in the NT compared with other jurisdictions are largely explained by a higher concentration of Indigenous people in the NT. Health loss in this particular population is considered elsewhere.18,19

Several areas for further research flow from this work. First, health loss and expenditure under a “business as usual” approach to health risk management have been projected into the future,2,20 and such analyses could usefully be extended to include various “what if?” risk-reduction scenarios. Second, simulation methods have been used elsewhere to quantify uncertainty in DALY estimates,21 and would enhance interpretability if applied to these findings. Third, developments in health state valuation methods could, if applied in Australia, increase confidence in the use of the DALY as a valid comparative measure of health loss.

Finally, our analysis is undermined, to some degree, by significant gaps in Australia’s health information infrastructure. In particular, there is limited information on mental disorders, neurological conditions, hearing loss, chronic respiratory diseases and musculoskeletal disorders. Even more importantly, Australia, unlike other countries, has no mechanism for regularly collecting measurement data on biomedical indicators such as body mass, blood pressure, and blood glucose and cholesterol levels. Better and more frequent monitoring in each of these areas would strengthen future comparative assessments of health in Australia, thus enhancing their value for policy and program development.

3 Health loss* attributable to 14 selected risk factors, by selected broad cause group, Australia, 2003

Broad cause group


Cancers

CVD

Mental disorders

Injuries

Diabetes mellitus

All causes


Total health loss (DALYs lost/1000 people)

25.1

23.8

17.6

9.3

7.2

132.4

Attributable health loss — individual (%)

Tobacco use

20.1%

9.7%

na

0.5%

na

7.8%

High blood pressure

na

42.1%

na

na

na

7.6%

High body mass

3.9%

19.5%

na

na

54.7%

7.5%

Physical inactivity

5.6%

23.7%

na

na

23.7%

6.6%

High blood cholesterol levels

na

34.5%

na

na

na

6.2%

Alcohol consumption

3.1%

4.7%

9.7%

18.1%

na

2.3%

Low consumption of fruit and vegetables

2.0%

9.6%

na

na

na

2.1%

Illicit drug use

na

< 0.1%

8.0%

3.6%

na

2.0%

Occupational exposures and hazards

3.1%

0.4%

na

4.7%

na

2.0%

Intimate partner violence

0.5%

0.3%

5.5%

2.5%

na

1.1%

Child sexual abuse

< 0.1%

< 0.1%

5.8%

1.4%

na

0.9%

Urban air pollution

0.8%

2.7%

na

na

na

0.7%

Unsafe sex

1.0%

na

na

na

na

0.6%

Osteoporosis

na

na

na

2.4%

na

0.2%

Attributable health loss — combined (%)§

32.9%

69.3%

26.9%

31.7%

60.1%

32.2%


CVD = cardiovascular disease. DALY = disability-adjusted life year. na = not applicable.

* Expressed as DALYs lost per 1000 people.

“Attributable” health loss is health loss that is explained by past and current exposure to health risks. This is distinct from “avoidable” health loss, which is health loss that might be averted through future changes in exposure to a health risk.

Attributable health loss within each column is expressed as a percentage of total DALY rates for that column.

§ Figures for combined effects are not necessarily column totals because risk factors can share common causal pathways.

  • Stephen J Begg1
  • Theo Vos2
  • Bridget Barker3
  • Lucy Stanley4
  • Alan D Lopez5

  • School of Population Health, University of Queensland, Brisbane, QLD.


Correspondence: s.begg@sph.uq.edu.au

Acknowledgements: 

Our research was funded by the Australian Government Department of Health and Ageing (DoHA). We gratefully acknowledge contributions from the Australian Institute of Health and Welfare (AIHW) and the project steering committee, which included representatives from the DoHA, the Australian Bureau of Statistics, state and territory health departments, and the AIHW.

Competing interests:

None identified.

  • 1. Murray C, Salomon J, Mathers C. A critical examination of summary measures of population health. In: Murray C, Salomon J, Mathers C, Lopez A, editors. Summary measures of population health: concepts, ethics, measurement and applications. Geneva: World Health Organization, 2002.
  • 2. Begg S, Vos T, Barker B, et al. The burden of disease and injury in Australia, 2003. Canberra: Australian Institute of Health and Welfare. (AIHW Cat. No. PHE 82.) http://www.aihw.gov.au/publications/index.cfm/title/10317 (accessed Sep 2007).
  • 3. Mathers CD, Vos ET, Stevenson CE, Begg SJ. The Australian Burden of Disease Study: measuring the loss of health from diseases, injuries and risk factors. Med J Aust 2000; 172: 592-596. <MJA full text>
  • 4. Murray CJL, Lopez AD, editors. The global burden of disease: a comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020. Vol I. Cambridge, Mass: Harvard School of Public Health, on behalf of the World Health Organization and the World Bank, 1996.
  • 5. Murray CJL, Lopez AD. Global health statistics: a compendium of incidence, prevalence and mortality estimates for over 200 conditions. Vol II. Cambridge, Mass: Harvard School of Public Health, on behalf of the World Health Organization and the World Bank, 1996.
  • 6. World Health Organization. International classification of functioning, disability and health. http://www.who.int/classifications/icf/site/icftemplate.cfm (accessed Oct 2007).
  • 7. Australian Institute of Health and Welfare. Disability and its relationship to health conditions and other factors. Canberra: AIHW, 2004. (AIHW Cat. No. DIS 37.) http://www.aihw.gov.au/publications/dis/drhcf/drhcfc01.pdf (accessed Sep 2007).
  • 8. World Health Organization. International statistical classification of diseases and related health problems. 10th revision. 2nd ed. http://www.who.int/bookorders/anglais/detart1.jsp?sesslan=1&codlan=1&codcol=15&codcch=1592 (accessed Oct 2007).
  • 9. Barendregt JJ, Van Oortmarssen GJ, Vos T, Murray CJ. A generic model for the assessment of disease epidemiology: the computational basis of DisMod II. Popul Health Metr 2003; 1: 4.
  • 10. Australian Bureau of Statistics. Population projections, Australia, 2002 to 2101. Canberra: ABS, 2003. (ABS Cat. No. 3222.0.)
  • 11. Stouthard ME, Essink-Bot M, Bonsel GJ, et al. Disability weights for diseases in The Netherlands. Rotterdam: Department of Health, Erasmus University Rotterdam, 1997.
  • 12. Ezzati M, Lopez AD, Rodgers A, Murray CJL, editors. Comparative quantification of health risks: global and regional burden of disease attributable to selected major risk factors. 2 vols. Geneva: World Health Organization, 2004.
  • 13. Mathers CD, Iburg KM, Begg S. Adjusting for dependent comorbidity in the calculation of healthy life expectancy. Popul Health Metr 2006; 4: 4.
  • 14. English DR, Holman CDJ, Milne E, et al. The quantification of drug caused morbidity and mortality in Australia. Canberra: Commonwealth Department of Human Services and Health, 1995.
  • 15. Ridolfo B, Stevenson C. The quantification of drug-caused mortality and morbidity in Australia, 1998. Canberra: Australian Institute of Health and Welfare, 2001. (AIHW Cat. No. PHE 29; Drug Statistics Series No. 7.)
  • 16. Rodgers A, Ezzati M, Vander Hoorn S, et al. Distribution of major health risks: findings from the Global Burden of Disease study. PLoS Med 2004; 1: e27.
  • 17. Taylor R, Dobson A, Mirzaei M. Contribution of changes in risk factors to the decline of coronary heart disease mortality in Australia over three decades. Eur J Cardiovasc Prev Rehabil 2006; 13: 760-768.
  • 18. Zhao Y, Guthridge S, Magnus A, Vos T. Burden of disease and injury in Aboriginal and non-Aboriginal populations in the Northern Territory. Med J Aust 2004; 180: 498-502. <MJA full text>
  • 19. Vos T, Barker B, Stanley L, Lopez A. The burden of disease and injury in Aboriginal and Torres Strait Islander peoples 2003. Brisbane: School of Population Health, University of Queensland, 2007.
  • 20. Begg S, Vos T, Goss J, Mann N. An alternative approach to projecting health expenditure in Australia. Aust Health Rev. In press.
  • 21. Mathers C, Salomon J, Ezzati M, et al. Sensitivity and uncertainty analyses for burden of disease and risk factor estimates. In: Lopez A, Mathers C, Ezzati M, et al, editors. Global burden of disease and risk factors. Washington, DC: World Bank and Oxford University Press, 2006.

Author

remove_circle_outline Delete Author
add_circle_outline Add Author

Comment
Do you have any competing interests to declare? *

I/we agree to assign copyright to the Medical Journal of Australia and agree to the Conditions of publication *
I/we agree to the Terms of use of the Medical Journal of Australia *
Email me when people comment on this article

Online responses are no longer available. Please refer to our instructions for authors page for more information.