Health spending due to risk factors

A risk factor is any determinant that causes (or increases the likelihood of) one or more diseases or injuries. The ABDS estimates the contribution of a range of risk factors to the burden of disease experienced in the Australian population. This includes risk factors that relate to personal behaviours (such as smoking), environment (such as air pollution) and biomedical risks (such as high blood pressure).  Quantification of the impact of risk factors assists evidence-based decisions about where to direct efforts to prevent disease and injury and to improve population health. This analysis uses burden of disease data and disease expenditure data to estimate the financial cost of the diseases that are due to risk factors, to expand current analysis beyond disease burden.

Estimating risk factor attributable fractions

To estimate the burden of a condition that is due to a risk factor, there needs to be sufficient evidence of a causal link between the risk factor and the linked disease. The following 20 risk factors have been identified as having sufficient evidence of a contribution to the development of any disease measured in the ABDS:

  • air pollution
  • alcohol use
  • bullying victimisation
  • child abuse and neglect
  • dietary risks
  • high blood plasma glucose (including diabetes)
  • high blood pressure
  • high cholesterol
  • high sun exposure
  • illicit drug use
  • impaired kidney function
  • intimate partner violence
  • iron deficiency
  • low birth weight and short gestation
  • low bone mineral density
  • occupational exposures and hazards
  • overweight (including obesity)
  • physical inactivity
  • tobacco use
  • unsafe sex.

The burden of each condition due to a risk factor is estimated using the relative risks determined through epidemiological investigation, population age and size, and the proportion of the population in each exposure group. This risk distribution is compared to a ‘counterfactual’, what would have occurred if exposure to the risk factor had been avoided or had been reduced to its lowest level. This is a theoretical minimum exposure which may not be achievable, feasible or economically viable, for example no overweight people or obesity in the Australian population. Combining this information gives a measure of the population attributable fraction (PAF) of a disease that is due to a risk factor. The proportion the risk factor contributes to a condition was adjusted to account for overlaps between risk factors (such as high blood plasma glucose (including diabetes) and overweight) because each risk factor was calculated independently in the ABDS. This adjustment was made using the joint effect calculation and ensures that the burden allocated is not greater than actual burden. Further details are described in Australian Burden of Disease Study: Methods and Supplementary material 2018.

Estimating spending due to risk factors

Risk factor PAFs represent the proportion of disease burden that can be attributed to a risk factor, while disease expenditure estimates represent the spending on their treatment and management. The proportion of disease burden that is due to a risk factor for a particular disease was used to allocate expenditure due to the risk factor. One benefit of using national datasets around disease expenditure and disease burden is that expenditure attributable to an individual risk factor can be directly compared to that of another, as expenditure on each disease attributed is capped along with total health system expenditure.

To estimate the health spending on a risk factor is relatively simple: total expenditure for diseases by age and sex are multiplied by the corresponding risk factor PAFs, to calculate the percentage of the disease spending that is due to a risk factor. For example, if 90% of the burden of disease due to type 2 diabetes is estimated to be attributable to high blood plasma glucose (HBPG), then 90% of expenditure on type 2 diabetes has been allocated to HBPG.

This assumes that the treatment patterns and outcomes for a specific condition are the same for those with a risk factor as without a risk factor. For example, the spending on treating lung cancer in a smoker is the same as for a non-smoker. For various reasons, this may not necessarily be the case: smoking may increase healing time for wounds (increasing length of stay in hospital), or mean that certain treatment options are not appropriate or available. A literature review was undertaken for each risk factor and disease pair to determine if there is any systemic difference in spending due to the presence of a risk factor for an individual. Through this process 114 studies were identified, and a large number indicated slight differences in disease treatment outcomes or spending may be present. However, many of these studies did not examine conditions for which a risk factor is causally linked in the ABDS, but examined clinical risk factors that may affect treatment.

However, due to differences in study designs and confounding factors, populations, classification of risk factors and types of spending included, or conflicting results between studies, none of the results have been adjusted for such differences in this analysis. It is important to interpret the results with care for these risk factor and disease combinations.

This risk factor spending analysis focuses on quantifying the spending on conditions that are caused by a risk factor, rather than the spending on managing a risk factor. Some risk factors are also coded as conditions within the ICD-10-AM, which therefore classifies them as conditions in the disease expenditure database, such as hypertension (captured in 'other cardiovascular conditions'). This is of particular relevance for pharmaceuticals that reduce blood pressure and cholesterol.