Burden of disease analysis aims to quantify health loss for all health outcomes, both fatal and non-fatal, and attribute it to a disease or injury category. This is achieved by separately estimating the fatal (YLL) and non-fatal (YLD) burden, according to a defined list of diseases, and summing them. This burden can then be attributed to risk factors selected for inclusion in that part of the analysis.

Detailed methods for estimating each component of burden of disease analysis for the Australian Burden of Disease Study 2018 are described in this report.

In this report

This report describes, as far as practicable, the methods and assumptions used by the Australian Burden of Disease Study (ABDS) 2018 to quantify the fatal and non-fatal effects and causes of diseases and injuries in the Australian population in 2018, 2015, 2011 and 2003

Burden of disease quantifies the gap between a population’s actual health and an ideal level of health in the given year—that is, every individual living in full health for his or her ideal or potential life span—and includes both fatal and non-fatal components. Risk factor analysis allows death and health loss to be attributed to specific underlying (or linked) risk factors
Burden of disease measures include Years of Life lost and Years lived with Disability which are summed to give Disability adjusted life years (DALY). DALYs can be attributed to risk factors and be used in the calculation of Health adjusted life-expectancy (HALE).
This chapter provides detailed information on the methods used to estimate mortality and morbidity for each cause in the 17 disease groups and describes in detail the methods unique to each risk factor included in the ABDS 2018.
In an ideal world, burden of disease estimates would be based on a fully enumerated set of data of all health loss and risk exposure experienced by every person in the population of interest. But in reality, burden of disease estimates are based on models of disease and risk factor epidemiology applied to existing sources of data of varying completeness and quality. In some instances, these 2 components are perfectly matched, but in many cases, there can be differences between the data required by the model and the data available to be analysed, leading to various levels of uncertainty around the estimate.