Benefits and limitations of main national data sources
On this page:
The benefits and limitations of main national data sources for monitoring dementia are summarised in Table 3.1. Data from these sources have data gaps or limitations that provide caveats to interpreting the results. For more details on each data source below, see Appendix.
Table 3.1: The benefits and limitations of main national data sources for monitoring dementia(a)
Survey of Disability, Ageing and Carers (SDAC)
Large survey measuring conditions and causes of disability, and disability-related need for assistance
It records dementia along with other health conditions
Comparable methods over time, allowing for time series analysis
Likely underestimates people with dementia
Unable to assess subgroups of interest (for example, those with younger onset dementia)
No coverage in very remote areas
General practitioners and specialists(b)
No national dementia-specific data currently available
Dementia diagnoses captured in various practice management systems
Suitability of dementia data from practice management systems still being ascertained
Pharmaceutical Benefits Scheme (PBS)
Information on PBS-listed prescription medications, including those for people with Alzheimer’s disease who were prescribed dementia-specific medication
Not all people with dementia are prescribed dementia-specific medication
The PBS currently subsidises dementia-specific medications only for people diagnosed with Alzheimer’s disease
Information about admitted patient activity in Australian hospitals and reason for admission
Inconsistent coding of dementia
Under-diagnosis and under-disclosure of dementia
Emergency department presentations
Information about patient activity in Australian hospital emergency departments and their reason for admission
National coverage of public hospitals with emergency departments
Missing data from private hospitals
Under-diagnosis and under-disclosure of dementia
Aged care assessments
Information on people assessed by Aged Care Assessment Teams to receive a range of aged care services
More likely to identify mild and moderate dementia
Changes over time in how data are held and reported
Detailed health condition and dementia data collection discontinued in transition to AN-ACC(c)
Includes only people who accessed formal aged-care services
Residential aged care
Information relating to the administration of residential aged care subsidies
National coverage of people in permanent residential aged care
May underestimate people with dementia
Incomplete coverage in very remote areas
Income support and allowances
Claims and payments data for recipients of certain government income support and allowances with a medical diagnosis of dementia (and their carers)
Dementia may not be recorded if claim for payment is based on another medical condition
Information on deaths in Australia and their underlying cause of death or associated cause of death
Unlikely that mild-to-moderate dementia will be recorded
(a) There are other data sets, not listed above, that can be used to monitor dementia when linked with the listed databases. However, care must be taken as the limitations listed against each data source are likely to apply to the linked data sets as well.
(b) The Medicare Benefits Schedule (MBS), which captures information on general practitioners (GPs) and specialist services, does not capture dementia diagnosis information. The Bettering the Evaluation and Care of Health (BEACH) program, which captured information on conditions managed by GPs in Australia, ceased in 2016.
(c) AN-ACC is the Australian National Aged Care Classification.
Limitations of the main national data sources for monitoring dementia
Inconsistency in the coding of dementia data items across data sets impacts the quality and comparability of data items such as dementia diagnosis, type and severity. The Dementia National Best Practice Data Set (NBPDS) provides concise and unambiguous definitions for data items related to dementia diagnosis, treatment and management. It aims to standardise the collection of dementia information in Australia across a range of data sets, including those used in clinical practice, clinical registries, epidemiological research and surveys.
Incomplete or consistent coding of electronic medical records may be improved with the use of data mining (using artificial intelligence). AI may also be able to guide assessments to improve diagnostic accuracy.
Inadequate sample size in national surveys, such as the SDAC, limits the analysis that can be undertaken of less common chronic conditions, including analysis by small geographical areas, Very remote areas, and by population groups of interest. Dementia is captured in the SDAC, but not other surveys that exclude people living in aged care facilities. Dementia-specific questions may not be included in surveys due to higher priority questions and time limitations for completing surveys. People with cognitive impairment may be under-represented if they opt out of participating in surveys. Survey age limits may also affect the coverage of people with dementia.
Administrative data sets could be improved with appropriate measures of cultural, ethnic and linguistic diversity, informed by the Standards for Statistics on Cultural and Language Diversity (ABS 2022), to better understand culturally and linguistically diverse (CALD) populations and inform health policy and service planning (FECCA 2020). Most studies of dementia in First Nations people and CALD populations come from site-specific epidemiological studies or national surveys such as the SDAC. Site-specific epidemiological studies are irregular, which stop or limit regular comparisons over time (AIHW 2020). Inadequate coverage in national surveys may limit analysis of dementia in priority population groups. Studies and methods of engagement need to be culturally safe and appropriate for priority populations (including First Nations people and CALD populations) in order to obtain representative data to inform dementia service and policy needs.
While improvements to the content and scope of available national data are needed, there are also broader system and policy-level barriers that need to be overcome, such as a lack of technological infrastructure and informatics mapping, and fragmented service navigation (Burkinshaw et al. 2022). Systemic barriers can contribute to a delay of accurate and timely diagnosis, especially in marginalised groups, which can impact reporting in both administrative and other data collections.
Standalone administrative data are particularly powerful for informing dementia monitoring when linked for the same individual or time or across multiple sources. Data linkage can improve dementia identification when dementia records from different sources are captured, as well as allow for a more wholistic, person-centred view of dementia.
Administrative data can also be linked with population-based studies undertaken for specific research purposes and which provide rich, high-quality data that may be difficult to obtain from administrative data (such as self-reported patient experiences or biometric data).
There are different types of linked data assets, and particularly useful for dementia monitoring are enduring linked data assets since they are more efficient than one-off linkages for specific projects and provide a consistent source for different uses. As of June 2023, key enduring linked data assets for dementia monitoring are the:
- National Integrated Health Services Information Analysis Asset (NIHSI-AA), which brings together national data from 2010–11 onwards on various health care data (hospital admissions, emergency department visits, prescriptions under the PBS and Repatriation Pharmaceuticals Benefit Scheme (RPBS), and MBS services), residential aged care data, and mortality data. These data provide better dementia identification than standalone data sets and a wide range of opportunities to better understand the pathways used among people with dementia. This includes the relationship between health and aged care use, reasons for transitioning into residential aged care, and the influence of patterns of care on outcomes.
- Multi-Agency Data Integration Project (MADIP), which links administrative data from a range of Commonwealth agencies covering diverse areas from tax and social security payments to MBS and PBS data. It also has linked national survey data sets, such as the SDAC. MADIP is missing hospitals and key aged care data that capture dementia information but provides insightful information on other understudied groups with dementia, such as people with younger onset dementia.
It is important to note that enduring linked data assets generally take more time to set up and use than standalone data assets and have more complex data governance and access rules.
Dementia programs data
There is a vast body of data captured by non-governmental organisations on the programs they undertake, which are being increasingly used to understand the impact of community services available to people with dementia and their carers, as well as to the broader Australian community:
- Dementia Australia is the national peak body for people impacted by dementia in Australia and provides a diverse number of activities and support programs, such as education, information, counselling and early-intervention support. Dementia Australia collect rich data on clients accessing their services, but unit record data are not currently available for external analysis.
- Dementia Support Australia (DSA) provides behavioural and clinical support to people living with dementia, their carers and/or health professionals. Services are across metropolitan, regional and remote areas, including tailored First Nations support. DSA collects multifaceted data on their clients. On request, aggregated data sets are available for external analysis.
- Dementia Training Australia (DTA) is a consortium funded by the Australian Government to provide nationwide education and training on the care of people living with dementia. DTA collects data on the numbers of trainings and completions for continuing professional development, Tailored Training Programs and vocational education and training, by workforce type, setting (aged care, community care, primary care, acute care, mixed care), and whether online or in-person training.
ABS (Australian Bureau of Statistics) (2022) Standards for statistics on cultural and language diversity, ABS, Australian Government, accessed 24 March 2023.
AIHW (Australian Institute of Health and Welfare) (2020) Australia’s health 2020: data insights, AIHW, Australian Government, accessed 28 July 2023.
Burkinshaw K, Tsourtos G and Cations M (2022) ‘System and policy-level barriers and facilitators for timely and accurate diagnosis of young onset dementia’, International Journal of Geriatric Psychiatry, e5859, onlinelibrary.wiley.
FECCA (Federation of Ethnic Communities’ Councils of Australia) (2020) If we don’t count it… it doesn’t count! Towards consistent national data collection and reporting on cultural, ethnic and linguistic diversity, accessed 22 March 2023.