Technical notes
The main source of information for this web report is the AIHW’s Disease expenditure database. It contains estimates of spending by Australian Burden of Disease Study condition, age group, and sex for hospital services (including public hospital admitted patients, public hospital emergency departments, public hospital outpatient services and private hospital admitted patients), primary health care (including general practitioner services, allied health services, benefit paid pharmaceuticals and dental services) and referred medical services (including specialist services, medical imaging and pathology).
The methods used for estimating disease spending is a mixture of ‘top-down’ and ‘bottom-up’ approaches, where total spending across the health system is estimated and then allocated to the relevant conditions based on the available service use data.
Although this approach produces consistency, good coverage and totals that add up to known expenditure, it is not as comprehensive for any specific disease as a detailed ‘bottom-up’ analysis, which would include the actual costs incurred for that disease. A lack of amenable data sources means that a more granular ‘bottom-up’ analysis is not possible.
Estimates in the Disease Expenditure Database have been derived by combining information from the:
- National Hospital Morbidity Database (NHMD)
- National Public Hospitals Establishments Database (NPHED)
- National Non-admitted Patient Emergency Department Care Database (NNAPEDC)
- National Non-admitted Patient Databases (aggregate, NAPAGG, and unit record, NAPUR)
- National Hospital Costs Data Collection (NHCDC)
- Private Hospital Data Bureau (PHDB) collection
- Bettering the Evaluation and Care of Health (BEACH) survey
- Medicare Benefits Schedule (MBS)
- Pharmaceutical Benefits Scheme (PBS)
- Health Expenditure Database.
It is not technically appropriate or feasible to allocate all spending on health goods and services by disease. For example, neither administration expenditure nor capital expenditure can be meaningfully attributed to any particular condition due to their nature.
This study includes payments from all sources of funds, such as the Australian and State and Territory Governments, Private Health Insurance, and out of pocket payments by patients.
Some components of recurrent spending are allocated differently between the AIHW Health expenditure Australia database, and the disease expenditure study. This approach was taken to reflect patterns of healthcare use for particular conditions, which is the focus of this body of work, rather than health funding arrangements. Spending estimates in hospitals are slightly higher than in the Health Expenditure Database. This is discussed further in the accompanying methodology report.
Expenditure information is added to hospital activity data for every admitted patient record in the NHMD, all emergency department presentations in the NNAPEDC, and all service events in the National Non-admitted Patient Databases. Data sets have been constructed for all private hospital admitted patient separations. Aggregated data sets by sex, age group, state/territory and SA3 geographical area, including patient co-payments, have been created for MBS services by provider specialty and subgroup, and pharmaceuticals by Anatomical Therapeutic Classification (ATC). All of the data sets include expenditure estimates for each ABDS condition.
Changes to methodology compared to the 2019–20 study
In this 2020–21 study compared to previous disease expenditure studies, there were changes to the methods used for MBS mapping, PBS mapping, Emergency Department (ED) analysis, Non-admitted patient analysis (NAP) and the identification of COVID-19 cases. The methodology changes are outlined briefly below. For further details on the methods used, refer to Health system spending on disease and injury in Australia: Overview of analysis and methodology 2020-21.
MBS mapping
The mapping of MBS was refined. In addition to allocating MBS items to broader ABDS groups, they were also now mapped to specific ABDS conditions if the item description allowed such allocation.
This improvement allowed for a more precise allocation of expenditure at the condition level, resulting in a more accurate analysis.
PBS mapping
The mapping of the PBS was refined. This was necessary due to certain categories of pharmaceuticals being under-represented in GP prescribing patterns, often because prescriptions are typically written by specialists or because the medication was newly listed on the PBS. To overcome this, a direct mapping approach was used to associate PBS medications with specific conditions, focusing on the drugs' generic names. The primary objective was to identify the most frequent medical uses of these medications. This streamlined approach allowed for a more efficient analysis, guided by specific criteria and assumptions.
Note, there was a variation in the method used for diabetes conditions in 2020–21. The allocation of spending for diabetes was based on the findings of the Fremantle study. This study was preferred over the BEACH study due to its more recent data and specific focus on diabetes.
ED analysis
The analysis for ED cases involved merging total cost buckets (previously used average cost buckets) obtained from the Independent Hospital and Aged Care Pricing Authority (IHACPA) data with each hospital separation. This merging was based on factors such as the hospital and AR-DRG recorded for each separation.
If records did not match these specifications, costs were assigned based on state totals, ensuring comprehensive coverage.
NAP analysis
The NAP analysis included merging total cost buckets from IHACPA data (previously used average cost buckets) with unit record separations. The merging was done based on establishment ID and Tier 2 classifications.
Like ED, records that did not find a match based on these criteria were further merged using either state or Tier 2 totals.
Additionally, clinic types were directly allocated to conditions using demographic patterns observed in the data. Seven clinics were directly allocated, with 4 clinics specifically related to COVID–19.
Identification of COVID-19 cases
Analysis for private and public hospitals:
The 11th version of the International Classification of Diseases (ICD) codes, including U07.1, U07.2, U07.4, U07.5, and U07.7, were used in the diagnostic field of both public and private admitted hospitals to identify COVID-19 cases.
Analysis for ED cases:
The analysis included both confirmed COVID-19 cases using the designated codes and ruled-out cases (U06.0) in the emergency department.
Introduction of Tier 2 clinic classes:
For NAP, four Tier 2 clinic classes (10.21, 20.57, 40.63, 30.09) were established to capture and track the diagnosis, treatment, and COVID-19 vaccination activities in outpatient clinics.
MBS mapping for keyword search:
A mapping file was created for MBS areas to search for specific keywords in the item descriptions. The keywords used in the search were 'COVID-19', 'SARS-COV-2', and 'COVID'. This process enabled the identification of 26 MBS items that could be linked to COVID-19 based on the presence of these keywords in the item descriptions.
The AIHW continually seeks to improve the methods used to produce these estimates. Estimates for disease expenditure are subject to revision. Hence the most recently published results are not directly comparable with previously published data.