Disease burden estimates

On this page:

2024 estimates

Earlier reference years

Estimation of COVID-19 for 2024

Estimation of lower respiratory infections (including influenza and pneumonia) for 2024

YLL and YLD data quality

2024 estimates

Estimates for years of life lost (YLL) for 2024 were largely based on trend analyses of annual deaths occurring between 2011 to 2022, projected to 2024. A 12-year time period was considered sufficient for producing reliable trends in mortality. Trend analyses were supplemented with ABS provisional deaths for available months (January to June) in 2024 which was used to validate and adjust projected deaths where appropriate. Provisional deaths to September 2024 were also used in the estimation of COVID-19 and lower respiratory infections (LRI) (see sections on COVID and LRI for more detail). 

Estimates for years lived with disability (YLD) for 2024 used trend analysis based on YLD estimates for previous ABDS reference years (2003, 2011, 2015, 2018), 2019 (calculated as part of ABDS 2022), 2020 (calculated as part of ABDS 2023), and 2022 prevalence data where available (for example, hospitalisation data and data from National Survey of Mental Health and Wellbeing (NSMHW)).

For some diseases, the reference period used to inform the trend was restricted (see Table S4 for information about the models used and years included in YLL and YLD estimates for each disease and injury in ABDS 2024). For example, for diseases that were largely impacted by the lockdowns and restrictions put in place during the early part of the COVID-19 pandemic (for example, road traffic injuries), estimates for 2020 were not included in the projection models for YLD estimates. The year 2020 was also excluded from trend analyses for YLL as this did not resemble a typical mortality year (ABS 2020). Other years were excluded where data were considered inappropriate for use in trend analysis, such as due to coding changes, or where data in early years (that is, 2003 for some causes) were not considered robust. See Box 5.1 for examples of coding changes for selected diseases and how this affected trend analyses. 

Box 5.1: Examples of changes to disease coding and guidelines over time  

Substance use disorders and accidental poisoning 

From 2013, the ABS implemented new software for coding causes of death, applied International Classification of Diseases 10th revision (ICD-10) updates and reviewed coding practices. These processes impacted the cause of death output from 2013 onwards for some diseases and conditions. Specifically, for substance use disorders, where a death was due to an accidental drug overdose for a person with a known addiction to the drug, the addiction was reported as the underlying cause of death (that is, ICD-10 codes F10–F19 Mental and behavioural disorders due to psychoactive substance use). Since the coding changes, the drug overdose is captured as the underlying cause of death (X40–X49 Accidental poisoning) and the addiction is retained as an associated cause of death. The result was an increase in deaths due to Accidental poisoning, and a decrease in the number of deaths due to Mental and behavioural disorders due to psychoactive substance use. These changes caused a break in the fatal burden series for these conditions. Comparisons of fatal burden between study years prior to 2013 (that is, 2003 or 2011) and those after (that is, 2015, 2018 or 2024) should take this into consideration. It does not affect comparisons between 2003 and 2011, or between 2015 and 2018 or 2024.

To account for the change in 2013, the 2024 projected estimates of fatal burden (YLL) of substance use disorders were based on trend analysis starting from 2013. 

Dementia and stroke

There was a change in the trend for deaths due to dementia in 2006. The number of deaths due to dementia has increased when comparing data before 2006 with data from 2006 onwards. This increase can be attributed to: 

  • changes in ICD-10 instructions for coding deaths data, which have resulted in assigning some deaths to vascular dementia (F01) that may previously have been coded to cerebrovascular diseases (stroke) (I60–I69) 
  • the increase in reporting dementia as the underlying cause of death accompanied by the decrease in reporting of dementia as an associated cause (Buckley et al. 2019) 
  • legal changes allowing veterans and members of the defence forces to relate death from vascular dementia to relevant service
  • a promotional campaign targeted at health professionals, which is thought to have increased the reported number of dementia deaths among this group (ABS 2014). 

These changes will have an impact on comparisons made between 2003 and study years after 2006. 

To account for these changes, the 2024 burden estimates (both YLL and YLD) for dementia and stroke were based on trend analysis starting from 2011. 

Gestational diabetes    

The International Association of the Diabetes and Pregnancy Study Groups (IADPSG) developed a new consensus guideline for the testing and diagnosis of gestational diabetes in 2010. In 2014, the endorsement of the IADPSG guidelines by the Australasian Diabetes in Pregnancy Society (ADIPS) resulted in a significant change to the practice of testing and diagnosing gestational diabetes in Australia (AIHW 2019). Reflecting international trends, Australian studies found increases in the number of women diagnosed with gestational diabetes following the introduction of the IADPSG guidelines between about 2010 and 2014, of 20% (Laafira et al. 2015), 35% (Moses et al. 2011) and 74% (Cade et al. 2019). A steep increase of the incidence of gestational diabetes was recorded from 2012–13 (AIHW 2019). These changes will have an impact on comparisons made between study years before 2014 (that is, 2003 or 2011) and those after (that is, 2015, 2018 or 2024). It does not affect comparisons between 2003 and 2011, or between 2015 and 2018 or 2024.

To address this change, the 2024 non-fatal burden (YLD) estimates of gestational diabetes were based on trend analysis starting from 2015, which was the closest starting point amongst available ABDS estimates.

For each disease, one of the following projection methods was used to estimate YLL or YLD burden in 2024:

  • log-linear regression (also called Poisson regression or Poisson linear regression) which assumes that rates change at a constant per cent annually (for example, increases by 1% every year);
  • ordinary least-squares regression (also called simple linear regression) which assumes a constant fixed amount of change (for example, 10 YLD) every year (NCI 2022);
  • sex- and age-specific crude rates assumed to be the same between the year with the latest available data (for example, 2022) and 2024, and applied to 2024 population data.  

The COVID-19 pandemic presents an important consideration for the selection of appropriate models given its impacts on the input data sources available, the health system or the disease/injury itself. For example, disease estimates that would otherwise rely on health surveys or screening data sources were likely to be impacted due to restrictions and lockdowns in reference years following the onset of the pandemic. Therefore, selected regression models take into account factors beyond indicators of best fit and incorporate an assessment of appropriateness in consideration of the pandemic data environment. Projections of burden using model inputs only up to and including the year 2019 (that is, prior to the pandemic), are available upon request.

COVID-19 was added to the ABDS 2022 as a new disease, and has since been included in the ABDS. COVID-19 and lower respiratory infections (including influenza and pneumonia) estimates were derived using the most recent available data for 2024. While burden of disease estimates report on lower respiratory infections (including influenza and pneumonia), attributable burden estimates report on lower respiratory infections and influenza separately, consistent with the approach used for ABDS 2018. Further details on these diseases, including caveats and assumptions, are presented below.

Earlier reference years

For YLL, estimates reported for earlier ABDS reference years (2003, 2011, 2015, 2018) have been revised to incorporate changes in mortality coding under the ABS revisions process. Once a year, the ABS revises mortality information for coroner-certified deaths to improve the accuracy of the coding of these deaths. These revisions do not increase the overall number of deaths in any year but may change the distribution of the causes of death. Further information on the ABS mortality revisions process is available on the ABS website. 

For YLD, estimates for earlier ABDS reference years were largely sourced from published estimates from ABDS 2018 (AIHW 2021a). Estimates for 2011, 2015 and 2018 were revised for the mental health conditions sourced from the NSMHW (depressive disorders, anxiety disorders, bipolar affective disorder and alcohol use disorder) to account for changes in prevalence between the 2007 and 2020–22 surveys. YLD estimates for autism spectrum disorders for 2018 were also updated following a revision to WA Intellectual Disability Exploring Answers (IDEA) data which has been linked to the National Disability Insurance Scheme (NDIS), resulting in higher ascertainment of individuals with autism spectrum disorders. YLD estimates for 2018 were revised to allow comparisons with 2024 estimates, however, it should be noted that estimates for 2018 and 2024 are not strictly comparable to estimates for 2015 and earlier years due to this addition of a new ascertainment source in the IDEA.

Estimation of COVID-19 for 2024  

Fatal burden

Methods for calculating fatal burden (expressed as YLL) of COVID-19 used the number of deaths directly due to COVID-19, the ages at which these deaths occurred, and the Global Burden of Disease Study (GBD) 2010 standard reference life table.

Definition and coding of COVID-19 deaths

In the International Classification of Diseases 10th revision (ICD-10), COVID-19 deaths are coded to:

  • ICD-10 code U07.1 – COVID-19 virus identified is used when COVID-19 is confirmed by laboratory testing.
  • ICD-10 code U07.2 – COVID-19 virus not identified is used for suspected or clinical diagnoses of COVID-19 where testing is not completed or inconclusive. 
  • ICD-10 code U10.9 – Multisystem inflammatory syndrome associated with COVID-19. This code is used to identify people who have died from a multi-inflammatory response syndrome associated with COVID-19. 
  • ICD-10 code U09.9 – Post COVID-19 condition. This code is used to link long term conditions including chronic lung conditions that are the result of the virus. These deaths are included as associated cause of death.

In ABDS 2024, deaths coded to U07.1, U07.2 and U10.9 as the underlying cause of death (death directly due to COVID-19) were included in estimating fatal burden. 

Data sources

COVID-19 deaths for 2024 were sourced from the ABS death registration data, which is the official Australian deaths data collected via the state/territory Registrars of Births, Deaths and Marriages. It includes death registration data and medical cause of death information completed by a certifying medical practitioner and is considered a high-quality data source. In early-mid 2020, the ABS started releasing provisional deaths data to monitor the impact of the COVID-19 pandemic. Further information about the completeness and timeliness of the ABS provisional deaths data is available on the ABS website.

Estimating fatal burden in 2024

At the time of analysis, COVID-19 deaths up to September 2024 were available from the ABS provisional deaths report, released on 29 October 2024 (ABS 2024). These data were incomplete for August and September 2024 due to late registrations, and as such data for these months were inflated to account for incompleteness. To estimate the number of COVID deaths for the remaining months of 2024, the monthly changes of COVID-19 mortality rates were modelled from COVID-19 deaths in the previous year (2023) using the percentage change in rates from month-to-month to December.  

The estimated COVID-19 deaths for 2024 were then disaggregated by single year of age and sex, using the age and sex distributions from a customised data request of 2024 provisional deaths (January to June) provided by the ABS. The standard reference life table was then applied to the estimates to derive the YLL at each age.

Non-fatal burden

The input data needed to calculate COVID-19 non-fatal burden estimates should ideally reflect the full coverage of cases, with any under-ascertainment adjusted for with appropriate data, if available. This was the method used for 2022 estimates, however over time, procedures for testing COVID-19 have changed (that is, the move from strict requirements for PCR-based testing to rapid antigen testing and self-reporting), making it difficult to determine the number of COVID-19 cases each year and the level of under-ascertainment. 

Estimating non-fatal burden in 2024

The potential of high under-ascertainment of COVID-19 cases in 2024 and limited data availability at the time of analysis on hospitalised and Intensive Care Unit (ICU) cases nationally, COVID-19 YLL-to-YLD ratios for 2023 (from ABDS 2023) were applied to the COVID-19 YLL estimated for 2024 (which used Provisional deaths) to estimate COVID-19 YLD in 2024 for each age-sex group. This approach was deemed reasonable considering mortality data for COVID-19 is of reasonable quality and the burden due to COVID-19 is predominantly fatal (83% in 2023). The underlying assumption using this method is that the COVID-19 YLL-to-YLD ratio for 2023 will be similar for 2024.  

This method was applied to all age-sex groups where the YLL in 2024 was predicted to be greater than 0. For age groups where there was no predicted YLL (Female 0–34 and 40–44; Male 0–34), a ratio of YLD in 2023 to YLD in 2024 was calculated for each age group and the median value applied to each 5-year age sex group. The median was used to limit the impact of outliers.

Details on the methods and conceptual model for COVID-19 used in ABDS 2023 (which underpins the non-fatal burden estimates for ABDS 2024) can be found in the ABDS 2023 Technical notes. 

Estimation of lower respiratory infections (including influenza and pneumonia) for 2024

Fatal burden

Deaths due to lower respiratory infections (LRIs), including influenza and pneumonia, were sourced from ABS provisional death registration data for 2024 and validated using the Australian Influenza Surveillance Reports. The Australian Influenza Surveillance Reports are compiled from a number of data sources, including laboratory-confirmed notifications to the National Notifiable Diseases Surveillance System (NNDSS); sentinel hospital admissions with confirmed influenza; sentinel influenza-like illness (ILI) reporting from general practitioners; ILI-related community level surveys; and sentinel laboratory testing results. See Department of Health and Aged Care (2024) for more information.    

Deaths for 2024 were derived separately for influenza and other LRIs (including pneumonia). For influenza, a similar method to estimating deaths from COVID-19 was used. Doctor and coroner-certified influenza deaths up to September 2024 were available from the ABS provisional deaths report, released on 29 October 2024 (ABS 2024a). Deaths for August and September 2024 were inflated to account for incompleteness due to late registrations. To estimate deaths from influenza in October to December, the monthly changes of influenza mortality rates were modelled from deaths in the previous year (2023) using the percentage change in rates from month-to-month to December. Monthly deaths from the Australian Influenza Surveillance Reports were compared with the ABS provisional deaths to validate the trend predicted.

For other LRI (excluding influenza), only doctor-certified deaths up to June 2024 were available to the AIHW at the time of analyses (customised data request supplied by the ABS). Pneumonia deaths make up the large majority of these deaths and thus were first estimated using the same methods as described above for influenza. An inflation factor based on the previous 5 years was applied to the doctor-certified pneumonia deaths for 2024 to account for the missing coroner-certified deaths. To estimate the remaining number of LRI deaths in 2024, the ratio of pneumonia to total LRI deaths between 2018 and 2022 from the National Mortality Database (NMD) (0.89), was applied to the estimated number of pneumonia deaths in 2024 to derive the number of LRI deaths for 2024. 

The single year age and sex distribution for influenza and other LRI from the provisional deaths data provided by the ABS for available months in 2024 was applied to the total number of deaths estimated for influenza and other LRIs in 2024. The standard reference life table was then applied to the estimates to derive the YLL at each age. YLL estimates for influenza and other LRI (including pneumonia) have been combined for reporting purposes.

Non-fatal burden

There was no national data available on the incidence of LRIs in 2024 at the time of analysis. As such, the same approach for estimating COVID-19 YLD in 2024 was used for LRI. This was to apply YLL-to-YLD ratios for 2023 (from ABDS 2023) to the YLL estimated for 2024 (based on ABS provisional deaths) to estimate YLD in 2024 for each age-sex group. This method was applied to all age-sex groups where the YLL in 2024 was predicted to be greater than 0. For the age-sex groups where there was no predicted YLL (Female 0–34 and 40–44; Male 0–4 and 10–34), a ratio of YLD in 2023 to YLD in 2024 was calculated and the median value applied to each 5 year age-sex group. The median was used to limit the impact of outliers. The underlying assumption using this method is that the LRI YLL-to-YLD ratio for 2023 will be similar for 2024.

Details on the methods and conceptual model for LRIs used in ABDS 2023 (which underpins the non-fatal burden estimates in ABDS 2024) can be found in detail in the ABDS 2023 Technical notes.

YLL and YLD data quality

To provide information on the quality of estimates, a quality index was developed for the ABDS to rate estimates according to the relevance and quality of source data, and methods used to transform data into a form required for analysis. Generally, the higher the rating, the more relevant and accurate the estimate.

To report on the reliability of projected burden of disease measures, the inclusion of confidence intervals associated with regression estimates was explored. However, these were not presented as these relate to the uncertainty of the regression models and do not reflect the underlying uncertainty associated with data inputs that inform prevalence estimates.

The burden estimates for ABDS 2024 were largely based on trend analyses using the most recent data for major data sources. The quality of input estimates in the ABDS 2024 for earlier reference years (2003, 2011, 2015 and 2018) are the same as the quality presented in the ABDS 2018. The ABDS 2018 estimates were produced using the best available data within the scope and time frame of the study. 

Fatal burden (YLL) estimates were considered to have the highest quality rating for both data and methods used, as they used administrative data from the National Mortality Database, or the ABS provisional deaths supplied. The projections for 2024 were largely based on previous mortality trends. The non-fatal burden (YLD) estimates varied depending on the disease or injury, and the data sources used.

Information about the quality of the YLD 2018 estimates and the data and methods used can be found in Appendix B in the Australian Burden of Disease Study: impact and causes of illness and death in Australia 2018 report (AIHW 2021a) and in the Australian Burden of Disease Study: methods and supplementary material 2018 report (AIHW 2021b). 

An assessment of the quality of YLD estimates for COVID-19 were not available in the ABDS 2018. Lower respiratory infections (including influenza and pneumonia) were adjusted using recent data due to available evidence that these diseases were impacted by COVID-19. To help users understand the potential sources of uncertainty associated with the estimates, the 2-dimensional index developed for the burden estimates was used for these 2 diseases. This index was derived based on:

  • the relevance of the underlying epidemiological data
  • the methods used to transform that data into a form required by this analysis.

The index is scored on a scale from A (highest) to E (lowest). The quality of COVID-19 and lower respiratory infections (including influenza and pneumonia) are discussed below. 

COVID-19

Non-fatal burden estimates for COVID-19 should be used with caution. Due to lack of robust data that captured all or most cases of COVID-19, non-fatal burden due to COVID-19 in 2024 was estimated using the ratio of non-fatal to fatal burden from 2023. This assumes that this ratio is the same between 2023 and 2024. Fatal burden estimates are considered more robust as they were calculated using available mortality data for the first half of 2024, with some assumptions made about the second half. Therefore, modelling non-fatal burden estimates using fatal burden was considered reasonable.     

Data score = D

Method score = D

Lower respiratory infections (including influenza and pneumonia)

Non-fatal burden estimates for LRIs should be used with caution. Due to lack of data that captured all or most cases of LRIs, non-fatal burden was estimated using the ratio of non-fatal to fatal burden from 2023. This assumes that the ratio between 2023 and 2024 are the same. Fatal burden estimates are considered more robust as they were calculated using available mortality data for the first half of 2024, with some assumptions made about the second half. Therefore, modelling non-fatal burden estimates using fatal burden was considered reasonable.   

Data score = D

Method score = D

Previous page
Next page