Australia’s health data landscape

Introduction 

Data and information are critical for developing evidence-based decision-making – for example, informing timely government responses during crises – and for increasing productivity. Broadly, health data are used to:

• provide clinical care
• generate population health statistics
• help governments and other entities manage the resources and performance of the health system
• develop policy.

Health data have been used for many years to inform decisions about health, both at an individual and a population level. Key examples of the use of population health data in Australia include:

• Knowing the rates of preventive activities (such as cancer screening or immunisation) among different population groups has enabled health promotion campaigns to be targeted at these groups.
• Calculation of population-level cancer survival rates has provided valuable information for both individuals and clinicians. For example, survival statistics may be useful for someone with cancer to understand their prognosis.  
• Rates of biomarkers by population groups has informed general practitioner (GP) guidelines for recommended ages for testing blood pressure, cholesterol and other biomarkers – particularly for Aboriginal and Torres Strait Islander (First Nations) people.
• Monitoring tobacco smoking, risky drinking and food intake has both informed and helped to guage the impact of policies such as plain packaging for tobacco products and health star ratings on food packaging.
• Monitoring infectious diseases facilitates response and recovery – for example, as happened for the COVID-19 pandemic.
• Monitoring health service use (including the use of hospitals and mental health services) during the COVID-19 pandemic enabled service needs and gaps to be understood and planned for.
• Estimates from the Australian Burden of Disease Study for a range of health conditions and risk factors were used to inform and measure progress against the National Preventive Health Strategy 2021–2030.
• Assessment of the impact of the bowel cancer screening program identifying cancers at earlier stages shows that the program reduces the risk of premature death. 

During the last decade, many changes have occurred in both the health and broader data landscape. These have been driven by a range of factors, including increased need as well as availability, accountability and technology. Key factors include:

• the move from paper to electronic recording 
• the advent of ‘big data’ – that is, the extremely large and complex data sets that cannot be effectively processed or analysed using traditional techniques 
• the ability and technology to support data linkage 
• the recognition of the importance of the need to govern, manage and use data well. 

At the same time, the greater availability of data and rapid advancements in technology have led to different challenges, such as ensuring privacy and managing cyber security risks.

Data on the health status of people

A range of data sources are used to understand the health status of Australians – including information about health conditions, medications and treatments, functioning ability and general wellbeing. These data are essential not only to inform policies, programs and service planning but also to assess trends over time and to indicate disparities in health status. 

Data sources used to monitor the health status of people include surveys, administrative data sets, disease-specific registries and disease-surveillance systems (Box HD.1). 

Data on the health status of population groups

Some population groups and communities in Australia face more challenges in accessing health services and information – and experience poorer health – than the broader Australian population. 

The National Preventive Health Strategy 2021–2030 refers to these groups of people as ‘priority populations’; they include, but are not limited to:

• First Nations people
• culturally and linguistically diverse (CALD) people 
• lesbian, gay, bisexual, trans/transgender, intersex, queer and/or other sexuality (LGBTIQ+) and gender diverse people 
• people with mental health conditions 
• people in low socioeconomic groups 
• people with disability 
• people living in rural, regional and remote areas.

Improving data on the health status of population groups

A large amount of data is already available about many of these priority population groups. (See for example, First Nations people, Rural and remote health). However, some areas still have limited data. 

A range of data development work is underway to improve the collection of demographic information in national data sets to identify priority population groups – for example, to include the ABS Standard for Sex, Gender, Variations of Sex Characteristics and Sexual Orientation Variables in data sets (Box HD.2).

Other development work includes data linkage to better identify and report on the health of CALD populations (Box HD.3). 

Data from health systems

Sources of Australia’s health information can be summarised into 3 distinctive categories:

• Primary health: This covers a person’s first contact with the health system, including GPs, pharmacy, dentistry and allied health care for which PBS and Medicare data are key resources. The AIHW’s Medicare Benefits Scheme funded services: monthly data dashboard:
  • reports regularly on the proportion of fees charged by service providers that attracted a subsidy through the MBS, as opposed to being funded by patient contributions (including out-of-pocket payments, private health insurance and other sources)

  • shows the volume of MBS services claimed per person in the population for each broad type of service.&

    This dashboard is updated every month with a 1-month lag (for example, February data are published at the end of March).

• Hospitals: The well-established national hospitals databases track interactions at every public hospital and most private hospitals in Australia. They record granular information, such as time of admission, admission type, demographic information and locality. These databases can show trends reaching back decades.

  One example of where these data are made available is the AIHW’s MyHospitals reporting. MyHospitals explores various topics, such as emergency department care, admitted patient care, elective surgery and the health workforce. It provides vital data on hospital access, activity, safety and quality, and spending.

• My Health Record: This centralised electronic health record system improves the accessibility and coordination of health information to empower the individual. Individuals have the option to create a record to consolidate their key health information, such as allergies, medications, pathology results and immunisation history. Patients and authorised health-care providers can access this information to support informed decision-making and continuity of care. 

Primarily, information collected from health systems informs health-care service delivery; for example, to bill for transactions and to monitor workforce and facility capacity. While useful, these data sources do not necessarily provide all the information needed to monitor the health of the population. They do not always include detailed information about why the health service was used, what occurred at the health service, and what was done – nor do they include all desired demographic information. However, for particular situations, these data sources can be very informative and have very detailed information not available elsewhere – often at the population level and not just a sample. For example, almost all people who experience a non-fatal heart attack will be treated in hospital and their hospital record will contain rich data about that admission.

Data gaps and development

Considerable progress has been made over recent years in developing new sources of, and insights into, information to fill data gaps.

Maternity care

One notable example is in maternity care with the development of the maternity models of care data collection. This collection provides information on:

• the different types of maternity care available
• how many people are using them
• how the outcomes differ by the different care types.

Developing a maternity models of care data collection

Maternity models of care describe how maternity care is provided to women during pregnancy, birth and the postnatal period. The model of care may vary by maternity service and location, as well as between public and private providers.

The AIHW collects models of care information from each maternity service in Australia using the Maternity Care Classification System (MaCCS). This system was developed jointly by the National Perinatal Epidemiology and Statistics Unit at the University of New South Wales and the AIHW, as part of the National Maternity Data Development Project. 

The MaCCS allows the AIHW to report on how many models of care there are, and the characteristics of these models, including, for example, the women for whom they are designed, the carers involved in providing the models of care, and the continuity of care within them. Each model of care can be classified into one of 11 categories based on its characteristics. 

In 2023, maternity services reported around 1,000 models of care across Australia. Most of these (81%) fall into 4 categories: 

1. public hospital maternity care (41% of all models)
2. shared care (15% of models)
3. midwifery group practice caseload care (14% of models)
4. private obstetrician specialist care (11% of models). 

Around 29% of the models have continuity of carer (the same caregiver) across the whole maternity period, 35% have some continuity of carer (for example, in the antenatal period only) and 36% have no continuity of carer (Figure HD.1). 

For more information see Maternity models of care in Australia and METEOR. 

Figure HD.1: The continuity of carer within models of care varies across Australia

Collecting and reporting data on maternity models of care will:

• tell us about the maternity care available to women across Australia, and how this changes over time
• support the inclusion of model of care data elements in the National Perinatal Data Collection. Linking models of care information with perinatal data allows us to explore which models of care women giving birth use, whether they experience continuity of care, and if this varies by maternal characteristics such as age and geographic location (see Maternity models of care infocus)
• allow outcomes for mothers and babies by their model of care to be explored, including for different populations – such as First Nations women, women from rural and remote areas, and women from CALD backgrounds. This will support the monitoring and evaluation of Women-centred care: strategic directions for Australian maternity services. 

Primary care

Another key data gap is in the area of primary care – there is no ongoing national data collection on the reasons for visits to primary care, or for treatments provided. 

The AIHW is committed to a work program that advances the provision of primary health care data by developing processes for their governance, standardisation, collection, analysis and reporting in Australia. This work will ultimately form a National Primary Health Care Data Collection, initially focused on GP activity data, to provide a better understanding of health conditions managed, and outcomes for individuals. 

The AIHW is working with:

• Primary Health Networks to understand the quality and uses of GP data, using dementia as a use case
• the Commonwealth Scientific and Industrial Research Organisation (CSIRO) on the creation and use of national Fast Healthcare Interoperability Resources standards in health-care information exchange
• a multijurisdictional project team to support the National Primary and Acute Care Data Linkage Project (Design Phase). 

The National Primary and Acute Care Data Linkage Project (Design Phase) is co-led by NSW Health, Commonwealth Department of Health and Aged Care and AIHW, in partnership with all state and territory health departments. The project is engaging key stakeholders, such as those from the Primary Health Network, general practice and Aboriginal community controlled health sectors, during the consultation process to inform a blueprint for a hub-and-spoke data linkage system. It is envisaged that de-identified data from general practices would be linked with other health data by leveraging existing infrastructure and successes across jurisdictions, such as the Lumos project in NSW, to provide better insights into patient journeys across the health system.

Data linkage 

Data linkage is the other area where considerable progress has been made in filling data gaps.

Australia has been at the forefront of the development of data linkage systems and the use of linked data. Western Australia led the way in Australia through the establishment in 1995 of the WA Data Linkage System (WADLS) – a system of linkages within and between health and non-health data collections in that state. 

The AIHW has been linking mortality data to the incidence of cancer to estimate cancer mortality rates since 1990. It also produces a large volume of linked data sets to support important medical research and to monitor the health of the population each year. 

The Population Health Research Network (PHRN), established in 2009, is a national collaboration that brings together existing data across Australia and makes the resultant linked data available for research. The network comprises project participants and data linkage units.

The AIHW serves as the PHRN’s National Linkage Unit; the state and territory data linkage units are: 

• the Centre for Health Record Linkage in New South Wales 
• the Centre for Victorian Data Linkage 
• Data Linkage Queensland 
• the Western Australian Data Linkage Branch 
• the SA–NT Datalink 
• the Tasmanian Data Linkage Unit.

Data linkage can allow data and their context to be viewed more comprehensively than is possible by looking at individual data sets in isolation. While the complexity and scale of data linkages has expanded substantially in recent years, the increased capacity to facilitate data sharing within existing legislations has enabled enduring linked data sets to be developed that make data linkage more efficient and accelerate the analysis of the linked data. 

The following sections describe some examples of enduring linked assets.

National Health Data Hub

The National Health Data Hub (NHDH), formerly known as the National Integrated Health Services Information Analysis Asset (NIHSI), was created to provide better insights into a person’s journey through the health system (AIHW 2024b). It includes data on a person’s:

• hospital visits in most jurisdictions (admissions, and outpatient and emergency department services)
• processed claims related to services that qualify for a benefit via the MBS
• processed claims related to prescription medicines that qualify for a benefit via the PBS
• use of aged care services
• immunisation history and death information (via the Immunisation Register and National Death Index, respectively).

A key benefit of the NHDH is the ability to look at person-level information – such as how many times a person is admitted to hospital as opposed to just counting the total number of hospitalisations – as well as understanding a person’s pathway through the health system. 

The following sections on treatment pathways for people hospitalised with acute coronary syndrome and transitions from hospital to residential aged care for people living with dementia give examples of analysis using the NHDH.

Treatment pathways for people hospitalised for acute coronary syndrome

A project conducted using the NHDH provided a snapshot of almost 35,800 people (aged 25 to 84) who survived an acute coronary syndrome (ACS) hospitalisation (AIHW 2024c). The project used data about a person’s interventional procedures and medication use to describe their ‘treatment pathway’ after an ACS hospitalisation.

A key finding was that only 31% of people filled a prescription for all 4 classes of recommended medications within 40 days of surviving an ACS hospitalisation. People who did not initiate the guideline recommended medication (all 4 classes of recommended medications within 40 days) were more likely to be women, be aged 75–84 or have identified prior coronary heart disease. 

Figure HD.2 shows treatment pathways and outcomes for people with a diagnosis of ST-segment elevation myocardial infarction (STEMI), a type of heart attack almost always caused by a complete blockage to a major coronary artery.

Figure HD.2: Treatment pathways and outcomes, among people with STEMI

The findings from this project have implications for policy development and clinical practice. They also provide direction for additional research which is needed to identify why these groups are less likely to follow guideline recommendations.

For more information, see Treatment pathways for people hospitalised for acute coronary syndrome and Medication use for secondary prevention after coronary heart disease hospitalisations: Patient pathways using linked data.

Transitions from hospital to residential aged care for people living with dementia

The NHDH was used to examine movements between residential aged care and hospital for Australians living with dementia who were aged 65 or older and hospitalised in 2017 (Figure HD.3). 

Results included that 1 in 4 people with dementia who were living in the community moved into aged care after a hospital stay. By comparison, 1 in 50 people without dementia who were living in the community moved into residential aged care in the 7 days after being discharged from hospital (Figure HD.4).  

Figure HD.3: Transitions to residential aged care or mortality up to 12 months after first hospitalisation for people living with dementia

Figure HD.4: Transitions to residential aged care or mortality up to 12-months after first hospitalisation for people without dementia

For more information, see Transitions to residential aged care after hospital for people living with dementia.

Person Level Integrated Data Asset

The ABS’s PLIDA combines information on health, education, government payments, income and taxation, employment, and population demographics (including from the ABS Census of Population and Housing) over time. It provides whole-of-life insights on various population groups in Australia, such as:

• the interactions between their characteristics
• their use of services like health care and education
• their outcomes like improved health and employment. 

A key benefit of the PLIDA is the ability to gain insights into the social determinants of health, and population demographics such as education, income and CALD populations. 

Box HD.4 describes a key project using PLIDA.

National Disability Data Asset

The National Disability Data Asset (NDDA) – designed to be an enduring national asset – comprises a collection of linked, de-identified data from across multiple national, state and territory government service systems to inform insights on people with disability and their pathways through services. When operational, the NDDA will be used to:

• provide a more complete picture of the programs and services used by people with disability
• help governments improve these programs and services
• share information about how opportunities and outcomes could be improved
• improve reporting on outcomes for people with disability for Australia’s Disability Strategy 2021–2031.

Australian National Data Integration Infrastructure

The Australian National Data Integration Infrastructure (ANDII) is being collaboratively developed by Australian, state and territory governments to deliver shared national infrastructure for data sharing and integration. It is proposed as a national source of high‑quality and timely linked data for Australian policy makers, analysts and researchers to provide insights for national and local benefit. The ANDII is the underlying infrastructure established to deliver the NDDA. Subject to future agreements and funding, the ANDII could be used to facilitate the creation of other specific data assets on other important policy issues.

The ANDII also includes:

• data governance and streamlined data sharing arrangements will enable the creation of the data asset, as well as workflow management (including data access, and use and release processes), an ANDII ICT Solution to support secure hosting and transfer of data as permitted across the national/state/territory ANDII Network, as well as data linkage, analytical asset build and analytical activity. 

The NDDA is being established using the Data Availability and Transparency Act 2022 (Cwlth), together with a range of other existing legislation. The new national infrastructure will build on and complement existing data integration practices at the national and state/territory level.

Future of data

The health data landscape is changing rapidly. Data are being used more and more in policy making, and how routinely collected data can be better used in this space is increasingly being considered.

Responding to change

Changes are occurring in many areas. We need to ensure that data are collected not only to measure this change but also to understand the impact of this change on other aspects of health.

• One example of this changing landscape is the importance of measuring the uptake of vaping and the relationship between that and rates of smoking. While smoking rates are declining, rates of vaping are increasing, particularly among young people and people from least disadvantaged areas.
• Another example is the increased use of virtual care – for example, delivering health care via telephone or video conferencing. Currently, there are no standards for measuring virtual care. Different jurisdictions have started collecting information in different ways; this is affecting the ability to establish how many admitted patients there are as opposed to how many are receiving their care virtually. Future work is planned to develop standards to ensure consistent collection and reporting of data on virtual care service delivery.

Collecting data currently not harnessed

Additional data are being collected in many settings that are currently not harnessed. Examples include:

• information about children collected through the Maternal and Child Health Program
• information about smoking and vaping status collected as part of the public dental program
• extensive data collected through electronic medical Records (EMRs).

Technological advances

Technological advances are creating opportunities to automate and accelerate many aspects of work. The development of generative artificial intelligence and the advancements in machine learning and natural language processing (NLP) will provide new opportunities to automate the way data are captured, coded, transmitted and reported.

For example, the use of machine learning and NLP will also likely change the way health data are coded in Australia, with a number of studies already undertaken to demonstrate this possibility (Liu et al. 2022). Within the AIHW, machine learning and associated AI tools are being used; they will continue to be used to support the automation of combining data sets and analysis of linked data sets such as predictive modelling, forecasts and projections for health planning and service delivery.

Australia has been active in developing and advancing new standards, terminologies and classifications that will help to deal with the challenges posed by these technical advances – such as SNOMED-CT AU, ICD-11, and health data exchange standards such as fast health interoperability resources (FHIR).

The use of SNOMED-CT AU is rapidly becoming the basis of structured data capture in EMRs in General Practice and hospitals, including its use within the FHIR standards. 

In addition to this, the AIHW is leading the work to consider how Australia can leverage ICD-11 in a range of health care settings. The 11^th revision of the ICD is fully digital for the first time which means it has the potential to integrate with other terminologies and tools to automate classification of health conditions across the healthcare continuum, including primary care (WHO 2022). This would significantly improve our understanding of the health of the population in a consistent way, particularly as models of care change.

Improving timely access to data

Timely data continue to be imperative, particularly for the reporting and management of communicable diseases. A key development in this area is the establishment of the interim Australian Centre for Disease Control (the Australian CDC). Part of its statement of intent is to improve timely access and sharing nationally of consistent data, information and advice. Such action is essential to the work of the Australian CDC to enable rapid risk assessment and response, and to support informed public health decision-making.

Looking to the future

An important focus for the future will be on ensuring that data are collected, stored and analysed in a safe and secure way and that these data are both accessible and ultimately used to inform, manage and monitor health outcomes. Emphasis will be on filling data gaps, particularly across all population groups and geographic areas – both through data linkage and other mechanisms – and on presenting information in a way that is useful and easy to interpret.

Further reading

Related topic summaries

Glossary   Data tables More articles from Australia’s health 2024 data insights