**7. Hospital Treatment**

As already described, a national collection of hospital in-patient data was the first task of the AIHW on its establishment in 1987. A national minimum dataset was introduced in 1990. Annual data have been published since 1993–1994, and the reporting has become progressively more timely. The collection covers patients in public and private hospitals [25].

Statistics on emergency department presentations are also produced, including principal diagnosis and triage category, as well as demographic characteristics of patients.

Non-inpatient data remain limited to administrative characteristics, with no information ye<sup>t</sup> available on reasons for encounter, diagnoses, or interventions.

Hospital in-patient data have formed the information base for many important health policy developments at national and state levels, including in relation to casemix funding for hospitals, potentially preventable hospitalisation, and quality and safety developments in hospitals. Additionally, equity issues around hospital access and variations in intervention patterns can be explored.

### *7.1. Casemix Funding*

Casemix funding was developed at a national level from the 1980s and was first introduced in Victoria in 1993. Casemix is now referred to in Australia as activity-based funding. Australia adopted the casemix models originally developed in the United States [26]. Casemix is a measure of hospital output for each patient, based on their diagnoses and interventions provided.

The Australian casemix system for acute in-patients (Australian Refined Diagnosis Related Groups) is based on hospital in-patient statistics and a hospital costing survey. The classifications used are an Australian modification of ICD-10 for diagnoses (ICD-10-AM) and the Australian Classification of Health Interventions (ACHI). Supplementary systems exist for sub-acute patients such as rehabilitation and palliative care.

The Australian activity-based funding system, including its supporting classifications, is not in the public domain. Several countries have licences from the Australian Government to use the system in their countries.

#### *7.2. Potentially Preventable Hospitalisation*

Potentially preventable hospitalisation is an indicator of the effectiveness of primary care. There is a range of vaccine-preventable, acute, and chronic conditions for which policymakers believe hospitalisation could be prevented by earlier community-based care, particularly significant diagnoses of interest include complications of diabetes and chronic obstructive pulmonary syndrome (COPD) [1] (Section 5). This is an example where statistics from one sector of the health system can be used as a performance indicator for another sector.

#### *7.3. Quality and Safety*

Selected diagnoses relating to 'hospital-acquired complications' are used as indicators of quality and safety issues in Australian hospitals [27]. There are 16 complications, including pressure injuries, health-care-associated infections, and respiratory complications. Hospital funding arrangements now include an adjustment for hospital-acquired complications, taking into account the non-preventable occurrences of these conditions.

Regrettably, there is duplication of hospital in-patient data collections. The Independent Hospitals Pricing Authority collects data from the States for its activity-based funding (casemix) functions. The Australian Department of Health collects clinical, demographic, and financial information for privately insured in-patients from private health insurers, and, through the Australian Private Hospital Data Bureau, also collects data from private hospitals covering patient demographics, clinical information, and hospital charges. The AIHW's collection covers most of the data items collected in these collections, apart from information on charges to private patients.

#### **8. Primary Health Care**

In contrast to information on hospital in-patients, primary health care statistics have had a chequered history and are a significant weak point in Australian health statistics. Primary health care practitioners, including general practitioners (GPs), nurses, allied health professionals, pharmacists, dentists, and Aboriginal and Torres Strait Islander health workers, provide services in a range of community settings and are critical first points of contact with the health system.

In 1998, a national sample-based data collection of general practice encounters (BEACH) was put in place by the University of Sydney and the AIHW, with wide-ranging support from general practitioners' professional representative groups and in partnership with a number of pharmaceutical companies. This unique partnership offered a publicly accessible dataset and provided information to participating pharmaceutical companies on their products' uses. The paper by Gordon et al. [28] in this issue outlines the development and uses of this data collection.

The Australian Government terminated funding for the BEACH collection in 2016 at short notice and without an alternative data collection in place. Currently, general practice statistical data are limited to extracts from GPs' electronic records, which come from multiple systems without any common data architecture or standards [29]. The samples drawn are not always structured to enable the production of statistics about all elements of the population, especially those with significant health disadvantages, such as Indigenous Australians. The AIHW is leading the development of a National Primary Health Care Data Asset [30] of which statistics about general practice are one component.

### **9. Public Health**

In 1999, a National Public Health Information Plan was published, which focused on the need for improving national health surveys. As a result, information on a range of risk factors has been expanded and systematised over the past 20 years—smoking, alcohol use, exercise, and diet are examples.

The *Australia's Health* series reports on a range of population health indicators, as well as information on national screening programs. Between 2010 and 2013, Australia had a specialist agency focusing on preventive health. It produced a national report, *State of Preventive Health 2013* [31], which has not been repeated. This report brought together the national data on major risk factors, as well as international comparisons.

The National Notifiable Diseases Surveillance System brings together reports on notifiable diseases across Australia. Fortnightly reports are produced. Notifications in respect of Indigenous people made up more than half of the 330,000 notifications in 2016, the latest report available [32].

COVID-19 has seen separate and timely reporting of data on infection cases, hospitalisations, and deaths. It remains to be seen whether this results in the quality and timeliness of notifiable diseases information being improved and strengthened.

### **10. Mental Health**

The paper in this issue by Rosenberg et al. [33] describes the development of mental health statistics since the 1990s. This development occurred outside the processes established under the National Health Information Agreement, even though the Australian Government and State health agencies were cooperating through their mental health experts. A separate governance arrangemen<sup>t</sup> was established through the Mental Health Information Strategy Sub-Committee.

The AIHW publishes an annual review of mental health services and associated resources in Australia [34,35]. Data are drawn from across AIHW data collections and other sources, including the 2007 ABS National Survey of Mental Health and Wellbeing.

#### **11. Medicines: Use and Outcomes**

Comprehensive data on medicines provided to Australians in the community are available from the national Pharmaceutical Benefits Scheme (PBS), established in 1948. Medicines below the cost threshold for the PBS and those provided to public hospital in-patients are not included in the PBS.

The use of these data for pharmaco-epidemiological purposes is described by Pearson et al. [36] in this issue. Increasingly, medicines data are linked to other national datasets (see Section 15, Data Linkage, below). The authors note that studies in Australia are relatively few and do not utilise all of the datasets available; they explore possible paths to facilitate a leap forward in medicine outcome studies.

#### **12. Data on Health and Health Disadvantage for Particular Population Groups**

There have been several references already in this paper to health statistics concerning Indigenous Australians. There are other population groups whose health status and access to health services also need to be monitored, as they experience significant disadvantages in relation to health.

People with disability are one such group. Statistics on people with disability and disability support services have been greatly improved over the past 20 years, although the introduction of the National Disability Insurance Scheme, itself a major social reform, has led to a break in the series of nationally consistent data on disability services, which was collected through the Disability Services National Minimum Dataset from 1991 to 2019 [37]. However, information on the health of people with disability and their access to health services has generally come from health and disability surveys, rather than from health services statistics. The paper by Fortune et al. [38] in this issue discusses this in more detail.

Medicare, Australia's universal health insurance system, gives all Australians the capacity to access high-quality medical and hospital services. The reality is that there is a clear excess burden of disease for lower socio-economic groups, notably in coronary heart disease, lung cancer, chronic kidney disease, and COPD [39]. In addition to survey evidence, data linkage is facilitating the examination of socio-economic variables in relation to health. For example, the ABS now links mortality data and census records, which has allowed examination of mortality due to various health conditions according to household equivalised income, highest educational attainment, and housing tenure [40].

Australia has about a quarter of its population born overseas, and almost half have at least one parent born overseas [41] (p. 271). Many health data collections include country of birth and language spoken at home. However, the AIHW has acknowledged that information on culturally and linguistically diverse (CALD) populations is among Australia's data gaps [1] (p. 6). *Australia's Health 2020* omitted data on CALD populations altogether. The 2018 edition briefly discussed the generally lower age-standardised mortality rates and rates of potentially preventable hospitalisations for people born outside Australia, compared with the Australian-born population. These data gaps have been thoroughly addressed in a recent report by the Federation of Ethnic Communities Councils of Australia (FECCA) [42].

In Australia, prisons and corrective services are the responsibility of the States. Without Australian Government involvement, it took many years for a national effort to report on the health of prisoners, which is the responsibility of State health departments or State correctional services agencies. Since 2009, the AIHW has conducted the National Prisoner Health Data Collection every 3 years. Data reported highlight significant mental health problems, high rates of smoking and drug use, and a high prevalence of disability among prisoners [43].

### **13. Health Registries**

Each Australian State has operated a cancer registry for many years. The AIHW maintains the National Cancer Statistics Clearing House (NCSCH), which was established in 1986 as the national repository of cancer incidence and mortality statistics. The repository is used to produce national cancer statistics. Each jurisdiction uses the national minimum dataset for its reporting. In addition, the jurisdictions collaborate with the AIHW to produce registries for breast, cervical, and bowel cancer screening.

There is now a wide range of clinical registries in Australia. These include clinical quality, disease, immunisation, and product registries. A Framework for clinical quality registries has been developed by the Australian Commission on Safety and Quality in Health Care [44].

### **14. Oral Health**

Dental statistics have been well developed in Australia through a specialist centre at the University of Adelaide, which has worked in collaboration with AIHW. Foundation work in South Australia was built to give a rich picture of child and adult dental health, as well as the dental health of Aboriginal and Torres Strait Islander peoples. The paper in this issue by Amarasena et al. [45] describes a recent national oral health survey and the changes in dental health over the past 30 years.

### **15. Data Linkage**

Data linkage has been mentioned earlier in this paper. Data linkage involves the development of enriched datasets by linking two or more datasets. Data linkage in Australia commenced in Western Australia in the 1990s and now occurs in all State jurisdictions, and is supported by the Public Health Research Network (see Smith et al. in this issue) [46]. Ethical approval is essential for data linkage because linked data can readily produce identifiable data even if the original datasets are de-identified.

The AIHW, as described by Jensen [47] has developed the National Integrated Health Services Information Analysis Asset (NIHSIAA) linking a range of its datasets and other Australian Government datasets, thus bringing together data covering hospitals, Medicare, Pharmaceutical Benefits Scheme, Repatriation Pharmaceutical Benefits Scheme, residential aged care, and the National Death Index.
