Article Text

Cohort profile
Cohort profile: Dementia in the Registry of Senior Australians
  1. Monica Cations1,2,
  2. Catherine E Lang1,
  3. Stephanie A Ward3,4,5,
  4. Maria Crotty2,6,
  5. Craig Whitehead2,6,
  6. John Maddison2,7,
  7. Maria Inacio1,8
  1. 1 Registry of Senior Australians, South Australian Health and Medical Research Institute, Adelaide, South Australia, Australia
  2. 2 College of Medicine and Public Health, Flinders University, Adelaide, South Australia, Australia
  3. 3 Centre for Healthy Brain Ageing, University of New South Wales, Sydney, New South Wales, Australia
  4. 4 Department of Geriatric Medicine, The Prince of Wales Hospital, Sydney, New South Wales, Australia
  5. 5 School of Public Health and Preventative Medicine, Monash University, Clayton, Victoria, Australia
  6. 6 Southern Adelaide Local Health Network, SA Health, Adelaide, South Australia, Australia
  7. 7 Northern Adelaide Local Health Network, SA Health, Adelaide, South Australia, Australia
  8. 8 Division of Health Sciences, University of South Australia, Adelaide, South Australia, Australia
  1. Correspondence to Dr Monica Cations; monica.cations{at}


Purpose Clinical quality registries (CQRs) are being established in many countries to monitor, benchmark, and report on the quality of dementia care over time. Case ascertainment can be challenging given that diagnosis occurs in a variety of settings. The Registry of Senior Australians (ROSA) includes a large cohort of people with dementia from all Australian states and territories identified using routinely collected aged care assessment data. In ROSA, assessment data are linked to information about aged and health service use, medicine dispensing, hospitalisations and the National Death Index. The ROSA dementia cohort was established to capture people for the Australian dementia CQR currently in development who may not be identified elsewhere.

Participants There were 373 695 people with dementia identified in aged care assessments from 2008 to 2016. Cross-sectional analysis from the time of cohort entry (e.g. when first identified with dementia on an aged care assessment) indicates that individuals were 84.1 years old on average, and 63.1% were female. More than 44% were first identified at entry to permanent residential aged care. The cohort recorded more severe cognitive impairment at entry than other international dementia registries.

Findings to date The cohort has so far been used to demonstrate a declining prevalence of dementia in individuals entering the aged care sector, examine trends in psychotropic medicine prescribing, and to examine the impact of dementia on aged care service use and outcomes.

Future plans The ROSA dementia cohort will be updated periodically and is a powerful resource both on its own and as a contributor to the Australian dementia CQR. Integration of the ROSA dementia cohort with the dementia CQR will ensure that people with dementia using aged care services can benefit from the ongoing monitoring and benchmarking of care that a registry can provide.

  • dementia
  • epidemiology
  • geriatric medicine
  • quality in health care
  • statistics & research methods

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

  • The Registry of Senior Australians (ROSA) dementia cohort includes large cohort of people with dementia from all Australian states and territories, with a wide breadth of linked data to allow for monitoring of care.

  • The ROSA dementia cohort does not include people who do not have a diagnosis of dementia or those who have not used government-subsidised aged care services.

  • Entry to the ROSA dementia cohort occurs at identification on aged care eligibility assessments, which can be sometime after dementia symptom onset or formal diagnosis. This limits the potential for monitoring early clinical care.


Registries are powerful tools for research and monitoring of clinical care because they facilitate population-level surveillance over time.1 2 As the global prevalence of dementia rises,3 dementia-related registries are being established internationally to complement clinical research and improve the quality of care for people with this condition.4 Methods of capturing cases of dementia vary between registries but usually include reporting from specialist clinics and hospitals. Importantly for a dementia registry, diagnosis occurs in a variety of settings and therefore capturing the whole population can be challenging.4

Dementia is a common chronic health condition in Australia, affecting an estimated one in 10 people aged over 65 years.5 More than 400 000 people are estimated to be living with dementia in Australia, 25 000 of whom are aged under 65 years.5 The Australian Dementia Network (ADNeT) Clinical Quality Registry (CQR) is a new national dementia CQR established to monitor, benchmark and report on the quality of care for people with mild cognitive impairment (MCI) and dementia over time.6 The ADNeT Registry enrols participants at the point of diagnosis in memory or private specialist clinics and will track longitudinal outcomes via patient and carer reported outcome measures as well as linkage with administrative datasets.

However, many people with dementia or MCI may be diagnosed in other settings.7 Given the high prevalence of dementia among those accessing government-subsidised aged care services,8 existing aged care assessment data has great potential to contribute to the capture of individuals into the ADNeT CQR. Approximately 47% of Australian residential care users and 21% of home care users have a recorded diagnosis of dementia.8 It is estimated that aged care assessments conducted from 2009 to 2015 captured approximately 36% of the estimated total population of people with dementia in Australia (prevalent cases) at the end of 2015.9 Therefore, understanding and studying the cohort with dementia captured within the aged care sector can significantly contribute to our understanding of the individuals that may not be captured earlier for a national CQR. Information about health service use, medicines, hospitalisations, mortality and other information can then be monitored for these individuals over time.

In our current evaluation we have examined (1) the demographic and clinical features of people with dementia using Australian aged care services and the extent to which these are representative of the broader population of people with dementia in Australia and (2) the comparability of data captured in aged care datasets to selected established international dementia registries. This will allow for better understanding of the characteristics and limitations of this cohort for monitoring the quality of care and outcomes for people living with dementia in Australia.

Cohort description

Design and data sources

A national cohort of all non-indigenous Australians aged 65 and over who have accessed government-subsidised aged care services from 1997 to 2017 (and updated regularly) is captured in the Registry of Senior Australians (ROSA). In ROSA, national aged care assessment data are linked with information about aged care service use, health service use, medication dispensing, hospitalisations and death records10 of individuals that entered the aged care sector. Specifically, assessments within the aged care sector are conducted to determine eligibility for government-subsidised services (by Aged Care Assessment Teams (ACAT), established in 2003) or to identify funding requirements in residential aged care (Aged Care Funding Instrument (ACFI), established in 2008). In both assessments assessors are clinically trained medical, nursing or allied health professionals who identify the level of care need based on functional and cognitive limitations.11 12 Data from assessments, as well as aged care service use, are provided to ROSA from the Australian Institute of Health and Welfare National Aged Care Data Clearinghouse.

These data are subsequently linked with information about government-subsidised health service use from the Medicare Benefits Schedule, medicine dispensing from the Pharmaceutical Benefits Scheme (PBS), state-based hospital records, and mortality data from the National Death Index. The ROSA established this cross-sector data linkage for research purposes and aims to assess the effectiveness, appropriateness, and quality of aged care services provided to older individuals over time. In its entirety, the historical ROSA cohort includes over 2.8 million individuals, including 1.2 million who have had aged care eligibility assessments for substantial aged care services like permanent residential care, home care packages, residential respite care and transition care.

Here, we present results of a cross-sectional evaluation of the people with dementia identified in aged care assessment data (hereafter referred to as the ‘ROSA dementia cohort’) between 1 July 2008 and 30 June 2016. The entry point to the ROSA dementia cohort is the first aged care assessment where a recording of dementia was made, though the person may have entered the broader ROSA with an earlier assessment (on which a dementia diagnosis was not recorded). Entry is distinct from the date of diagnosis, which will have occurred earlier and is not known for our cohort. Where a person is identified from medication prescribing records, data from the closest aged care assessment is included here for cohort profiling.

Dementia ascertainment

Dementia is determined from aged care eligibility assessments (conducted by ACAT), assessments for funding in permanent residential care (conducted using the ACFI), and from pharmaceutical data as captured in PBS. In assessments, assessors record up to 10 (ACAT) or up to 3 (ACFI) major diseases or disorders that have an impact on the person’s need for assistance with activities of daily living and social participation, together with documented evidence of a diagnosis from a medical practitioner. Assessors can record one or more types of dementia or have the option to classify the dementia as ‘unspecified’ based on the medical record. In addition, medicines prescribed for the treatment of Alzheimer’s disease (AD) are not dispensed for any other reason. Any person with who has been dispensed donepezil (Anatomical Therapeutic Chemical Classification System code, ATC N06DA02), galantamine (ATC code N06DA04), rivastigmine (ATC code N06DA03) or memantine (ATC code N06D×01) can be classified as having dementia.

Minimum data set

The data available for the ROSA dementia cohort are presented in table 1, with comparable data from other established dementia registries. Registries included for comparison were chosen based on their broad coverage and the availability of data for comparison here. They include one CQR (the Swedish Dementia Registry (SveDem)) and two epidemiological dementia registries (French National Alzheimer Database; Registry of Dementia of Girona, Spain (ReDeGi)).

Table 1

Comparison of the ROSA dementia cohort and international dementia registries, minimum dataset available

ROSA includes comprehensive demographic data and information about aged and healthcare service use, including service entry and exit dates. All prescription-based medicine dispensing is recorded by the PBS, facilitating monitoring of medicine dosage, duration and polypharmacy. Information about family history of dementia, diagnostic procedures or other clinical details (aside from comorbidities) is not available. ROSA also does not include privately funded health service use.

Assessments conducted for financial purposes at entry into residential aged care are repeated when care needs change. These include a standardised neuropsychological assessment. While most dementia registries include the Mini-Mental Status Examination (MMSE)13, copyright restrictions have precluded its widespread clinical use in Australia. Instead, the Psychogeriatric Assessment Scale-Cognitive Impairment Scale (PAS-CIS)14 is conducted where cognitive impairment is suspected or known. The PAS-CIS correlates strongly with the MMSE.15 A Cornell Scale for Depression16 is conducted where symptoms of depression and dysthymia are present. Functional dependence is rated across domains (including nutrition, mobility, personal hygiene, toileting, continence, home maintenance and transport) though a validated measure like those used in other registries is not included.

Cohort characteristics

There were 363 695 people in the ROSA dementia cohort over the capture period (table 2). Detailed demographic data on the cohort appears in online supplemental table 1. The cohort is representative of the geographical spread of the Australian population with the majority living in New South Wales, Victoria and Queensland (76.2%, compared with 77.4% of the general population). Similarly, 33.4% of the ROSA dementia cohort live outside a major city (compared with 28.2% of the general population), 32.9% were born outside Australia (compared with 33.3% of the general population) and 11.5% primarily speak a language other than English (compared with 22.2% of general population households where a non-English language is spoken).17 While comparable on sex, the ROSA dementia cohort is older at entry than other registries and includes more people living in permanent residential care.

Supplemental material

Table 2

Comparison of the ROSA dementia cohort and international dementia registries, demographic characteristics

AD is the most common type of dementia in ROSA, similarly to other registries examined (table 3). On cognitive assessment, which are only completed by those living in permanent residential care (82% of the cohort had a PAS-CIS score available), the ROSA cohorts were more cognitively impaired than the cohorts of other registries.

Table 3

Comparison of the ROSA dementia cohort and international dementia registries on clinical characteristics

People in the ROSA dementia cohort have a median four (IQR=2) other comorbid health conditions. In SveDem, people with dementia recorded a median Charlson comorbid index score of 2 (IQR=2).18 In the ROSA dementia cohort, the most common comorbid conditions were arthritis (35.0%), hypertension (34.9%) and heart diseases (31.7%%). Cerebrovascular disease (15.5%) and hypercholesterolaemia (9.1%) were common. Prevalence of hypertension and hypercholesterolaemia were lower than the Spanish ReDeGi registry (50.6% and 25.1%, respectively). Prevalence of cerebrovascular disease in our cohort was similar to ReDeGi.19 In our cohort, more than 99% of individuals reported at least one activity limitation, most often transport (94.7%), healthcare tasks (92.7%) and social and community participation (92.5%). Individuals in the ROSA dementia cohort lived for an average of 2 years after they were first identified with dementia in the aged care assessment data and were on average 88 years old (SD=6.6 years) at the time of their deaths. In contrast, a recent analysis of the SveDem cohort identified that only 28% of the cohort had died within the median 2-year follow-up period.20 In ROSA, dementia was recorded as the primary cause of death for 27% of the cohort, most commonly unspecified dementia (15.2%). Other common causes of death were heart diseases (21.8%) and cerebrovascular disease (21.1%).

Patient and public involvement

Consumer representatives are part of the governance structure of ROSA and provided oversight for ROSA’s development and now oversee ongoing operations.

Findings to date

To date, the ROSA dementia cohort has been used to demonstrate a declining prevalence of dementia in individuals entering the aged care sector,8 to show the value of residential respite for delaying institutionalisation for people with dementia,21 and to highlight poorer outcomes after hip fracture among those with premorbid dementia than without dementia (publication currently under review). The broader ROSA dataset has also recently been used to develop an Outcome Monitoring System for aged care, with 12 indicators of care quality that can be monitored over time and across geographical areas.22 Most of these indicators are relevant to dementia care, including psychotropic medicine use, hospitalisations, and falls, and work is underway to apply them to the monitoring of care in the ROSA dementia cohort.

Strengths and limitations

The large sample and national coverage provided by ROSA are key strengths of the ROSA dementia cohort. ROSA includes the largest existing population-based sample of people with dementia in Australia and is representative of the population in many ways, including sex, regionality, and cultural and linguistic diversity. An average of 37 661 new cases of dementia are identified in ROSA each year and many of these may not be identified via other sources.9 The ROSA dementia cohort is, therefore, a powerful resource both on its own and as a contributor to the ADNeT CQR.6 A wide breadth of data is available in ROSA and this is expanding as linkage to new state-based data sources continues, including hospitalisations and ambulance use. These data can facilitate monitoring of clinical care and determinants of important outcomes including institutionalisation and mortality over time.

Despite these benefits, there are important limitations to the ROSA dementia cohort. First, we cannot capture people with dementia who do not have a diagnosis, nor those who do not access aged care services. Approximately half of people with dementia in Australia are estimated to receive a diagnosis,7 23 24 and delays in diagnosis are common.7 25 People who do receive a diagnosis tend to have more severe impairment, have insight into their impairment, speak English, live in metropolitan areas and in areas with greater access to health services, have higher levels of education, and be married.26 Also, one in three women and one in two men will not use an aged care service in their lifetime27; ROSA is not able to capture these individuals and these factors introduce a sampling bias to our cohort.

Second, two important groups are not represented in the current ROSA cohort. People who accessed aged care services before 65 years of age but died before turning 65 years old are not included. We estimate that ROSA currently captures 40%–84% of those aged under 65 using aged care (depending on the year). The number of missing cases of dementia attributable to this data gap is likely to be small given that most people with symptom onset prior to 65 years are aged 60–64 years at onset and will age over 65 years with their condition.28 Nonetheless, types of dementia that are most common in younger groups (eg, alcohol-related dementias, frontotemporal dementias, dementia in Huntington’s disease, dementia in Down syndrome) are likely to be underrepresented. ROSA also does not currently include Aboriginal and Torres Strait Islander people, though consultation is under way to enable inclusion of this cohort in future analyses.

Third, the data available in ROSA are not collected for research purposes and therefore may have limited internal validity. While the ‘breadth’ of data is a key strength of this cohort, its relative lack of ‘depth’ is a limitation and the suitability of service use is difficult to assess. The accuracy of clinical and demographic data also relies on assessors who are not necessarily trained in research data collection or in dementia care. Aged care assessors are limited to recording a maximum of 10 health conditions (ACAT) or three mental/cognitive conditions (ACFI) per assessment; whether dementia is considered an important enough comorbidity to be listed is at the assessors’ discretion. In the absence of cognitive assessment, the accuracy of the dementia diagnosis recorded in ROSA is dependent on the skills and resources available to the clinician who made the diagnosis. Additionally, ROSA is not a dementia-specific registry and includes fewer clinical details than available in other cohorts.

Finally, aged care eligibility assessments can occur sometime after dementia symptom onset or formal diagnosis, limiting the potential for monitoring early clinical care. More than 36% of individuals entered the ROSA dementia cohort at entry to or while living in residential aged care, which is likely to be late in their disease path. As such, people living in permanent residential care are overrepresented in ROSA compared with national estimates5 and to other registries that recruit at the time of diagnosis. They are likely to have more functional limitations and comorbid health conditions and to die sooner than other registries, though little data from other registries is available for comparison. Capture of those entering residential care for a dementia CQR is nonetheless important given that many will not be identified elsewhere.


ROSA is the product of a consortium of 13 academic, clinical, industry, consumer representative and public health organisations.10 The consortium oversees ROSA management and use, ensuring that ROSA projects have clinical and public health relevance. Results described here demonstrate that aged care assessment data can be a valuable resource for maximising capture for the ADNeT CQR. Linkage between ROSA and ADNeT will ensure both that monitoring of care can occur early in the disease course, and that people with dementia using aged care services can benefit from the ongoing monitoring and benchmarking of their clinical care.

Other collaboration is welcomed, and researchers interested in collaboration are invited to contact the research team to access all data available in ROSA. In addition to the data described here, similar data are available from aged care users without dementia and state-based hospitalisation data for some Australian states. Data linkage is ongoing and is being updated over time, and we particularly encourage collaboration with those with other datasets that could be linked. Data access is subject to ethical and governance approval.


We would like to acknowledge the Healthy Ageing Research Consortium Investigator Team, the South Australian Health and Medical Research Institute Research Team, and our consumer advisors for ensuring the success of ROSA and support with this study. We also acknowledge the South Australian Government who provide us with support (2017-2021) through the Department for Innovation and Skills, and the Australian Institute of Health and Welfare and SA Health for the provision of the raw data used in the ROSA. The authors would also like to thank the ADNeT Registry Steering Committee for providing governance oversight, strategic direction and advice to the ADNeT Registry project.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Twitter @MonicaCations

  • Contributors MI, MCa and SAW conceptualised the project, obtained funding and assisted with reviewing and editing the manuscript. MCr drafted, reviewed and edited the manuscript. MCa and CEL conducted data analysis. CW and JM provided oversight to the project and assisted with manuscript review and editing. All authors read and approved the final manuscript.

  • Funding This study is funded by the South Australian Government Premier’s Research and Industry Fund (2017-2021; no grant number) and the National Health and Medical Research Council Boosting Dementia Research Grants (GNT1152623). MCa is supported by an Early Career Fellowship from The Hospital Research Foundation. MI is supported by a Mid-Career Fellowship from The Hospital Research Foundation. All authors had final responsibility for the decision to submit for publication.

  • Disclaimer The funders had no role in study design, methods, data collection and analysis, decision to publish or preparation of this manuscript.

  • Competing interests MCa has been employed in the last 5 years to assist with data collection for drug trials funded by Janssen and Merck. All other authors declare no conflicts of interest.

  • Patient consent for publication Not required.

  • Ethics approval Ethical approval was provided by the University of South Australia (reference ID200489), Australian Institute of Health and Welfare (reference EO2018/1/418), South Australian Department for Health and Wellbeing (reference HREC/18/SAH/90) and New South Wales Population and Health Services (reference 2019/ETH12028) Human Research Ethics Committees.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data are available on reasonable request. Researchers interested in collaboration are invited to contact the research team to access the data available in ROSA. In addition to the data described here, similar data are available from aged care users without dementia, and state-based hospitalisation data for some Australian states. Data linkage is ongoing and is being updated over time.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.