Regional Variation in Lifetime Probability of Admission to Residential Aged Care in Australia

Abstract

Objectives

This paper aims to apply a novel demographic technique to update – and extend to sub-national regions – estimates of the lifetime probability of admission to residential aged care.

Methods

Making optimal use of Australian data sources on aged care usage, mortality and population, this study adopts a two-population life table approach to produce an updated set of national probability estimates and first-time regional estimates.

Results

The probability of admission generally increases with age: nationally, lifetime probability at age 65 is 50% for women and 37% for men, rising to 55% and 46%, respectively, at age 85. This general pattern varied somewhat across regions.

Discussion

The regional results point to inequities in the uptake of care, thereby informing providers, governments, aged care advocates and anyone interested in equity of access.

Keywords

nursing home aged care equity access policy Australia

Introduction

For many people, moving into residential aged care (RAC, also known as nursing home care) seems a distant and unlikely prospect. Indeed, at any one time, only 4% of Australians aged 65 and over are permanent residents; this rises to 21% of people aged 85 and over (Australian Institute of Health and Welfare, 2019). Yet previous Australian studies have shown that the likelihood of an individual moving into permanent care at some point in their remaining lifetime is as high as 54% for women and 42% for men (Liu, 1998, 2000; Rowland et al., 2002), with similar findings seen internationally (Broad et al., 2015). Hence, understanding and quantifying this demand – particularly in the context of ageing populations – is important for government planners, for providers, and for individuals.

There are four broad approaches to determining the probability of admission: methods based on analysis of place of death recorded on death certificates, methods based on retrospective or prospective cohorts (starting or finishing with death records, respectively), methods based on life table models, and transitional probability models.

Studies using the place of death method first emerged in the United States in the 1970s, initially based on small, localised samples (Kastenbaum & Candy, 1973; Wershow, 1976), then extending to national studies (Ingram & Barry, 1977; Zappolo, 1981) – these studies consistently put the proportion of deaths in nursing home at around 20% of all deaths among older persons. This approach, with an adjustment for hospital deaths amongst recent residents, was recently used in New Zealand (Broad et al., 2015): the data showed that 38% of older deaths were in residential aged care, which converted to an estimated lifetime use of nursing homes (among those aged 65+) of 47%.

A sophisticated study based on deaths was conducted by researchers at the Australian Institute of Health and Welfare (AIHW, 2015), who linked deaths records with aged care usage records for all people aged 65 or older who died in 2011. The study showed that 43% of decedents had used permanent residential aged care at some time in the 8 years before death; adding the current residents would make this rate somewhat higher.

Criticisms of the death certificate method – chiefly regarding the fact that recent discharges from care are not captured (which problem was circumvented in the AIHW study) – led to the alternative approaches. Early proponents of the follow-up method included Palmore (1976) and Vicente et al. (1979), whose lifetime probability estimates (based on small samples but thorough methods) ranged from 26% to 46%.

A challenge in interpreting the results of these early approaches is that the ‘risk’ was ill-defined and not readily comparable across the different study types. A further criticism is that there is little accounting for the changing usage rates as people age, or the increasing attrition from the age groups (McConnel, 1984). The life table method avoids these problems by offering a clear conceptualisation of the risk being computed, and considers the different usage rates and different sizes of a cohort as it ages. Using national data sources, McConnel (1984) estimated the lifetime (at birth) risk of institutionalisation to be 48%, rising to 63% for an individual aged 65 years.

In Australia, Liu (1998; 2000), updated by Rowland et al. (2002), used life table methods to produce estimates of the lifetime probability of first admission to care – and of any use of care – at selected ages, for each sex. They made considerable effort to clarify the concepts being reported, and the later estimates were somewhat below those of McConnel when looking at only permanent care, and similar or above when combining permanent and respite care. For example, the lifetime (at birth) probability of admission to permanent care for females was 42%, and at age 65 was 46%; for permanent + respite care, the probability at birth was 59%, and at age 65 was 64%.

The final approach – typically focused on informing risk assessments for long-term care insurance – draws on multiple data sources to estimate transition probabilities between certain states (such as being well, have a functional limitation, needing institutional care, and death). An early working paper using this method by Dick et al. (1994) produced an estimate of 35% for the risk of being admitted to care beyond age 65, broadly consistent with estimates from the other methods. A more recent study by Hurd et al. (2014) used multiple waves of a longitudinal survey and simulated transitions, resulting in estimates (at age 50) of 65% probability of having at least one stay for females, and 50% for males.

Alas, there are no whole-of-population Australian survey data sources suitable for this approach. Hence, of the four methods described, the life table method is the best option for estimating lifetime probability of care in Australia. Further supporting the choice of this method, there are long-term, detailed administrative data on the use of residential aged care in Australia – enumerated at the person level – from which all the input values can be derived, thereby eliminating any shortcomings identified in earlier (non-Australian) studies. Explicitly, the data yield age-sex-specific rates of first and subsequent admission to care, and rates of exit (to death, hospital, other residential care facility, and return to community). Complete mortality data are also readily available.

The previous studies by Liu and Rowland et al. adopted four simplifications. First, admissions, deaths (in community and in care) and exits were each assumed to occur evenly across an age interval. Second, the number of people in the community was taken to be the overall estimated resident population (ERP), but the ERP includes people in care, which at higher ages can be a large proportion of the ERP. Third, exits to hospital were treated as live discharges, even though a large proportion are known to die in hospital (and neither return to the community nor to RAC). Fourth, admission to care was effectively treated as an ‘absorbing’ state with respect to exiting the cohort of community-dwelling people; that is, people that returned to the community were not re-included in the denominator for calculating subsequent admission rates.

The first two of these simplifications can be resolved by additional calculations using the raw data (see more in Methods). The third issue can be attenuated by including in the count of deaths in care a proportion of the number of exits to hospital. The fourth issue can be addressed by reconceptualising the life table method as a ‘multi-population’ model, as extensively treated by Rogers (1975, 1995), wherein ‘the community’ is one population and residential aged care is the second population. This approach allows for transitions from the community into care and possible return to the community (and potentially subsequent admission, and so on), while also accounting for independent rates of death in the community and in RAC.

Returning to the central concept of the lifetime probability of entering care, of further interest to many people is whether their chance of moving into care is different because of where they live. And to the extent that admission to care (when it is needed) is a function of the availability of places, the regional variation is also of interest to providers, governments, aged care advocates, and anyone interested in equity of access to care.

Hence, the primary aim of this paper is to estimate the lifetime probability of admission to residential aged care for sub-national regions of Australia. A secondary aim is to introduce some enhancements to the traditional life table method of generating such estimates.

Methods

Datasets

The life table method adopted here requires 3 main data inputs: admissions to and exits from RAC; mortality rates for the community-dwelling population and for those in care; and population estimates for those in the community and those in care.

The data were supplied by the AIHW in response to customised requests. Full details of the supplied data are in Appendix 1. Residential care data for calendar years 2017–2019 were sourced from the System for Payment of Aged Residential Care (SPARC): this system collected uniform residential aged care data – at the person level – from 1997 to 2022, so provides complete and accurate records of admissions to and exits from RAC over the study period.

Mortality data for calendar years 2017–2019 (based on year of occurrence) were sourced from the Australian Bureau of Statistics national deaths collection, which collates death registrations data from the 8 Australian states and territories. As death registration is a legal requirement in Australia, the deaths data are considered a complete enumeration of the mortality experience in Australia (ABS, 2022a).

Population data as at June 2018 were sourced from ABS estimated resident population models.

Selected national Census data – on factors known to be associated with higher risk of admission – were sourced from the ABS Census DataPacks.

Regions

The regions used in this analysis are the Statistical Area 4 (SA4) regions from the so-called ‘Main Structure’ of the Australian Statistical Geography Standard (2016 version) as developed by the ABS (2016). There are 107 SA4s in the Standard, but 16 are special codes for ‘offshore/migratory’ areas and ‘no fixed address’ in each state and territory, and 3 are for ‘other territories’, leaving 88 regions in scope for this analysis.

Note that although the SA4 schema aims for relatively uniform regions in terms of population (in metropolitan regions the average is around 365,000 people, whereas in non-metropolitan regions the average is around 194,000), the 88 regions are highly heterogenous in terms of land area and therefore population density. Australia is a vast country – the world’s sixth largest country by land area – with an average population density of around 3 people per square kilometre. However, 67% of the population occupies just 0.7% of the land area comprising the ‘greater capital city’ regions of the eight states and territories (at an average density of 310 per km²). Conversely, just 1.9% of the population resides in the five largest (so-called ‘outback’ regions) that collectively cover 74% of the land area (at an average density of 0.08 per km²). This leads to a more than 75,000-fold variation in the population density across the regions, ranging from 0.07 to 5330 people per km² – this is illustrated in Figure 1.

Figure 1.

Distribution of SA4 regions, with population density. Note: The log scale is used to better show the gradient across regions. A value of –1 corresponds to a density of 0.1 people per km²; a value of 3 corresponds to a density of 1,000 people per km².

The SA4s are a good choice for this study because they are large enough to avoid volatile rates – particularly for constructing the life tables – but small enough to provide some degree of ‘local’ information; they are also specifically designed to align with natural labour market areas across the country, making them useful for planning purposes. At the federal level, Aged Care Planning Regions (ACPRs) are used for aspects of planning and management of aged care services; since 2018 there have been 73 Regions covering the whole country. The ACPRs are built up from the smaller SA2 units in the Australian Statistical Geography Standard, and, for a handful of ACPRs, the correspondence is preserved after aggregating up to the SA4 level. Each of the ‘outback’ SA4s neatly overlaps two or more ACPRs, and in metropolitan areas this is typically inverted (a single ACPR overlaps two or more SA4s), or the correspondence breaks down. The relationship between ACPRs and SA4s is depicted in the map at Appendix Figure A1. Notwithstanding the potential utility of the ACPRs, the SA4 regions are a better choice for this study because the population, mortality and Census data are not readily available for ACPRs.

For admissions and community mortality, the SA4 pertained to the person’s address in the community; for residents and exits, the SA4 pertained to the location of the facility.

Life Table Analysis

As noted above, the approach here follows Rogers’ multi-population life table method (Rogers, 1975, 1995) to allow for transitions from the community into care and possible return to the community (and potentially subsequent admission, and so on), while also accounting for independent rates of death in the community and in RAC.

Because of the very low numbers of people and events related to RAC under the age of 40 (e.g. 71 admissions out of a total of approximately 214,000 (0.03%) over three years), rates for RAC-related events under 40 years of age are set to zero. At the other end of the age scale, although data were supplied for the oldest age groups up to 100+, the small counts of events and population estimates meant that there was significant volatility in the upper ages; hence, the final (open) age interval was set at 90+ (aggregating the 90–94, 95–99 and 100+ age groups across all measures).

Where appropriate, annual rates were calculated as the average over 3 years, using the midpoint of the period in the denominator. To further improve the stability of mortality rates, a Brass logistic model was fitted to the life table for each region, using the relevant state and territory mortality rates as the reference standard. The fitted l_x values from each region’s whole-population life table were then input to the two-population model.

Detailed methodological notes are provided in Appendix 2.

Ethics Approval

This study uses administrative by-product data that – although provided in response to a customised data request – could otherwise be made available in the public domain; hence, no institutional ethics review was required.

Results

The analysis is based on 396,645 aged care ‘events’ (admissions, deaths, discharges to hospital and returns to community) over the 2017–2019 calendar years, for people aged 40 years or older at the time of event; details of each type of event are shown in Appendix Table A1. (The underlying whole-population life tables are based on a total of 479,486 deaths over the same period).

The probability of admission generally increases with age (Table 1): nationally, lifetime probability (at age 0) was 47% for females and 33% for males. At age 65, the (remaining lifetime) probability is only slightly higher at 50% and 37%, respectively, rising to 55% and 46%, respectively, at age 85.

Table 1.

Lifetime Probability of Admission to Permanent Residential Aged Care, by Sex, Selected Ages, Australia (With Regional Summary), 2017–19.

	0	40	65	70	75	80	85	90
	Age
Measure	Female
National	0.47	0.47	0.50	0.51	0.53	0.55	0.55	0.50
Regional
Median	0.50	0.51	0.53	0.55	0.56	0.59	0.60	0.53
Lowest	0.25	0.26	0.27	0.28	0.30	0.33	0.31	0.23
Highest	0.68	0.69	0.73	0.75	0.77	0.80	0.82	0.73
	Male
National	0.33	0.34	0.37	0.39	0.41	0.43	0.46	0.45
Regional
Median	0.35	0.36	0.39	0.41	0.43	0.45	0.47	0.46
Lowest	0.21	0.22	0.25	0.27	0.28	0.29	0.27	0.21
Highest	0.48	0.49	0.53	0.55	0.58	0.62	0.66	0.69

Note. The full national life table calculations are shown in Appendix Table A2, and summary results for each SA4 are shown in Appendix Table A3.

This general pattern applied across all regions, but with substantial variation by region (Figure 2, focused on geographic distribution; Figure 3, focused on numerical distribution). For example, at age 0, the probability of admission for women varied from 25% to 68%, and for males from 21% to 48%. At age 65, the (remaining lifetime) probability ranged from 27% to 73% for females and from 25% to 53% for males. As for the national ( = average) results, the median rate across the regions peaked at age 85, as did the highest rate for females, but the highest regional rate for males was at age 90.

Figure 2.

Lifetime probability of admission at age 65, by sex, SA4 regions, 2017–19.

Figure 3.

Lifetime probability of admission, by sex, by age, SA4 regions, 2017–19. Note: each dot represents a single SA4 region (for each sex).

In very broad terms, the probability of admission was lower for regional and remote areas than for capital city areas. For three of the five large greater capital city areas – greater Melbourne, greater Perth and greater Adelaide – there was a degree of uniformity of results for the regions within that area; for the other two – greater Sydney and greater Brisbane – the degree of variation mirrored the diversity of the non-capital zones for that jurisdiction.

For example, in greater Melbourne the probability of admission for females aged 65 ranged from 48% to 62% (males 34% to 44%), compared with a state-wide low of 41% for females and 28% for males. In greater Sydney the probability for females aged 65 ranged from 41% to 67% (males 32% to 48%), compared with a state-wide low of 41% for females and 29% for males. Results for each region are shown in Appendix Table A2.

For more than three-quarters of regions, the lifetime probability of admission for females was higher than for males across all ages. The main exception, for around one-tenth of regions, was that the probability at age 90 was higher for males than for females. Detailed results by region, for selected ages, are given in Appendix Table A3.

At the national level, the probability of admission was higher at age 85 than 90 for both females and males. However, there were some deviations from this pattern across the regions: for females, seven regions departed from the national pattern such that the probability was higher at age 90 than at 85; for males, one-fifth of regions (18 of 88) varied from the national result.

Given the median age of admission to permanent residential care is around 85 for women and men, it follows that the underlying life expectancy of people in a region is a factor in the likelihood of admission to care. Indeed, at age 65 – when the median life expectancy is 22.6 years for females and 19.9 for males – there is a moderate positive correlation between (remaining) life expectancy and probability of admission (Pearson’s R = 0.38 for females, 0.49 for males – see Figure 4(a)).

Figure 4.

Correlation of lifetime probability of admission with (a) life expectancy (e₆₅) and (b) provision of residential aged care places, age 65, by sex, SA4 regions. Note: each dot represents a single SA4 region (for each sex).

As the maps in Figure 2 show, and noted above, the probability of admission broadly reduces with increasing remoteness from the capital city zones. This largely corresponds with the supply of residential aged care places: there is generally lower (proportional to population) supply in rural and remote areas compared with metropolitan areas. Notwithstanding these parallel patterns, there is only a modest positive correlation between probability of admission (at 65 or older) and supply of residential aged care places in a region (Pearson’s R = 0.13 for females, 0.35 for males – see Figure 4(b)).

Discussion

This study has produced new estimates of the lifetime probability of admission to permanent residential aged care, leveraging the detailed, person-level data available from the Australian aged care system.

The two-population life table approach is strongly suited to this analysis, given there is natural ‘migration’ between the community and care settings, and the mortality rates are dramatically different in each population. The high degree of alignment with previous Australian results (Liu, 1998, 2000; Rowland et al., 2002) further supports the suitability of the approach.

The general drop in probability of admission for the oldest females is consistent with earlier Australian studies (Liu, 1998, 2000; Rowland et al., 2002). Those authors suggested that the decline may be affected by a higher mortality such that someone at risk of entering care dies before admission. This could be a factor, but the life table model shows that, at least for females, there are more deaths among those admitted than not admitted at higher ages. Further, Rahman et al. (2019) – using linked longitudinal data on women only – showed transition from the community to residential aged care was 44% more likely than transition from the community to death. They also showed that baseline age was significantly associated with an increased hazard of transition from the community into care (HR = 1.26, 95% CI = 1.19–1.34), and that this hazard was higher than for transition to death (HR = 1.12, 95% CI = 1.07–1.18).

Hence, the results may reflect a particular point in history, in that this group may have atypical resilience arising from their early experiences of the Great Depression; and it possibly reflects a large cohort of mid-20^th century ‘healthy migrants’ (acknowledging that any initial effects have been seen to diminish over time (Renzaho, 2016)).

Any of these explanations also need to be attenuated in the light of the regional results, in that there were many departures from the national pattern (as reported above). While difficult to quantify, some of this variation may arise from more volatility in the underlying age-specific rates seen in smaller populations. Smoothing the input rates (à la de Beer (2012)) did not appear to suppress the degree of variation.

The relatively low correlation of admission probability with life expectancy, and with supply of places, is a little surprising, and suggests that local factors might be influential in encouraging, or facilitating, uptake of care.

Other well-established predictors of aged care use include level of morbidity (particularly functional ability and dementia), marital status, and availability of a carer (Banaszak-Holl et al., 2004; Cepoiu-Martin et al., 2016; Gaugler et al., 2007; Luppa et al., 2010; Rahman & Byles, 2020; Rahman et al., 2019). The Australian Census provides various self-report measures of these factors (or proxies for them) at the small-area level. Initial analysis of three indicators for people aged 65 or older (proportion reporting dementia and/or stroke, lone households as a proportion of total households, and proportion reporting core activity limitation need for assistance) showed inconsistent correlation with lifetime probability of admission: dementia – Pearson’s R = 0.25 for females, 0.30 for males; disability – R = –0.01 for females, 0.09 for males; living alone – R = 0.09 for females, –0.41 for males. Further analysis is required to explain these varied results, particularly the large negative association with living alone for males.

The regional results presented here are meaningful only to the extent that people don’t move region to take up residential care, as the analysis is grounded in aligning the mortality experience in an area with the take up of aged care in an area. Analysis published by the Royal Commission into Aged Care Quality and Safety (2020) suggests that this is not problematic for metropolitan and more populous regional areas, as 84% and 79%, respectively, of people did not move greater than 25 km to take up residential care. For remote regions, the chances of moving are much higher, with as much as 53% of people from very remote areas moving greater than 100 km to enter care. Although it is possible that these distances are still contained within the SA4 region (the largest SA4 is more than 1.3 million square kilometres), it is possible that there is a mismatch of the mortality and admissions data in some areas. This mismatch could be slightly exacerbated were ACPRs to be used as the unit of analysis, as there are at least two ACPRs that had no residential care places in the reference period, meaning that people would have to have moved to take up care.

Regional results regarding the correlation with the supply of places are also complicated by the extent to which neighbouring regions might supplement supply. Quantifying this effect would be challenging, requiring linked, geocoded data from separate aged care data systems – which are not routinely produced – along with sophisticated Geographical Information System analysis capability. Accordingly, it is beyond the scope of this paper. Nonetheless, at a cursory level, by visual inspection of Figure A1 – and with reference to the drive distance analysis summarised above – the supplementation effect is likely to be minimal in the outback areas, moderate in the regional/remote areas, and high in the metropolitan areas. The choice of analysis unit – ACPR versus SA4 – may not be material for this issue, but would be informed by the proposed complex geospatial analysis.

The reference period for this study (2017–19) immediately precedes the COVID-19 pandemic, so it is only possible to speculate on the impact of the disease on the results for a later period. Many of the inputs to this study – notably mortality (in the community and in care), admission to care, and return to the community – could all be affected by COVID-19 in different ways. For example, to the extent that death in the community is a ‘competing event’ for admission to care, the overall excess mortality due to the pandemic would have reduced admission rates. Further, given prominent media representations of outbreaks in Australian residential care facilities, some people who might otherwise have been admitted could have deferred their admission or sought alternative care arrangements; and some people already in care might have returned (at least temporarily) to the community setting. The comprehensive public health response to the pandemic within the aged care sector resulted in a reduction in excess mortality in care during 2020 compared with 2019 (Department of Health and Aged Care, 2023), slightly boosting the chances of an individual returning to the community.

Another impact on future results is likely to be the rapid expansion in the supply of home care packages, which is a personalised package of government-subsidised care provided to people in their own home. In 2018–19 (that is, towards the end of the reference period for this study), there were a total of around 125,000 packages allocated to individuals, compared with about 214,000 residential care places. In the five years to 2022–23, the home care program grew to nearly 278,000 packages, compared with around 225,000 residential care places. Further study will be required to assess the extent to which the expansion of the home care sector will reduce demand for residential care, and how regional market factors might influence the relative demand for home or residential care.

This study has several strengths. First, it uses a large, person-level, time-series dataset to aggregate records of entry to and exit from aged care, overlaid with regionally matched mortality and population data. Second, use of the Rogers’ multi-regional approach accounts for returns to the community and subsequent entry to care (or death). Third, using the same source data set, it derives empirical ‘separation’ values to tighten up inputs for the life table analysis.

The main limitation of this study is the quality of the underlying population data at a regional level – particularly for men at older ages, as noted by Wilson and Terblanche (2018). These authors applied the Extinct Cohort and Projected Survivor Ratio methods to produce new estimates of the older populations, but such estimates are not available at the regional level. The other factor affecting quality of estimates at the regional level is that direct measures of inter-regional migration are not available, so official population statistics rely on alternative data sources on migration to inform the regional estimates (ABS, 2022b). Further analysis is required to assess the impact of these issues on the overall results.

A second limitation is that many of the ‘small areas’ are not particularly small (as noted above) and for the most part only partially align with the Aged Care Planning Regions, as described above and illustrated in Figure A1. The choice of (regional) unit of analysis invariably involves compromises, and for this study the availability of suitable data was paramount in that choice. The overall results are designed to show the nature and extent of regional variation to highlight possible inequities in access; not using the Aged Care Planning Regions does not materially reduce the utility of these results for that purpose.

Third, the life table approach assumes that the mortality and event rates recorded in the study period apply for the whole lifetime of each cohort, and this is patently untenable in a changing policy environment – particularly where the rapidly emerging policy emphasis is on supporting people to live at home for as long as possible.

Conclusion

The application of Rogers’ two-population approach, along with other methodological enhancements – all applied at the small-area level – has yielded the most rigorous and comprehensive set of lifetime probability estimates for Australia to date. The results are important for individuals, providers and (government) planners as they highlight the true risk of future admission and dispel the ‘4% fallacy’. To wit, on average, a female aged 85 in Australia has a 56% chance of being admitted to residential aged care in her remaining lifetime, and a male aged 85 has a 47% chance. Depending on the region, this chance could be as high as 82% for females (at age 85) and 69% for males (at age 90).

As highlighted earlier, future research could focus on the use of more refined small-area population estimates to underpin the core life tables and the two-population life table analysis.

The analysis dataset may also be a useful input to projection models, wherein population projections and changing patterns of aged care use could be applied to produce a series of life tables to estimate lifetime probability of admission for, say, a 5–10-year horizon, at the regional level.

These results inform aged care planning by quantifying the possible future demand for residential aged care services, and highlighting the variation in probability of admission relatively independent of traditional planning measures. Hence, the regional variation is of interest to providers, governments, aged care advocates and anyone interested in equity of access to care.

Supplemental Material

Supplemental Material - Regional Variation in Lifetime Probability of Admission to Residential Aged Care in Australia

Supplemental Material for Regional Variation in Lifetime Probability of Admission to Residential Aged Care in Australia by Mark Cooper-Stanbury in Journal of Aging and Health

Footnotes

Acknowledgements

I am grateful to the following individuals and organisations for their assistance with this study: Dr. Tom Wilson, University of Melbourne, for assistance with the fundamental demographic approach; Prof Jeromey Temple, University of Melbourne, for general support in my program of study; and The Australian Institute of Health and Welfare, for provision of data in response to a customised request.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Transparency and Openness Promotion Statement

The dataset and associated materials for this study are not available for replication purposes because (a) the data were provided to me in response to a customised data access request and not otherwise published by the data owner, and (b) more analysis is planned on the same dataset. Notwithstanding that the data are not directly accessible by other researchers, I have included detailed methodological notes, and also included the details of the data request which researchers could use to request the same data. The study was not pre-registered, as it is not a clinical trial or systematic review.

ORCID iD

Mark Cooper-Stanbury

Supplemental Material

Supplemental material for this article is available online.

References

ABS.

(2016). Australian statistical Geography standard (ASGS): Volume 1—main structure and greater capital city statistical areas , July 2016. https://www.abs.gov.au/ausstats/abs@.nsf/Lookup/bySubject/1270.0.55.001∼July_2016∼Main_Features∼ABS_structures∼10007

ABS . (2022a). Deaths, Australia methodology. https://www.abs.gov.au/methodologies/deaths-australia-methodology/2021

ABS . (2022b). National, state and territory population methodology. https://www.abs.gov.au/methodologies/national-state-and-territory-population-methodology/sep-2022

AIHW . (2015). Use of aged care services before death (Data Linkage Series No. 19. AIHW.

AIHW . (2019). Aged care data snapshot—2019 (Version 4). https://www.gen-agedcaredata.gov.au/Resources/Access-data/2019/September/Aged-care-data-snapshot—2019

Banaszak-Holl

Fendrick

A. M.

Foster

N. L.

Herzog

A. R.

Kabeto

M. U.

Kent

D. M.

Straus

W. L.

Langa

K. M.

(2004). Predicting nursing home admission: Estimates from a 7-year follow-up of a nationally representative sample of older Americans. Alzheimer Disease and Associated Disorders, 18(2), 83–89. https://doi.org/10.1097/01.wad.0000126619.80941.91

Broad

J. B.

Ashton

Gott

McLeod

Davis

P. B.

Connolly

M. J.

(2015). Likelihood of residential aged care use in later life: A simple approach to estimation with international comparison. Australian and New Zealand Journal of Public Health, 39(4), 374–379. https://doi.org/10.1111/1753-6405.12374

Cepoiu-Martin

Tam-Tham

Patten

Maxwell

C. J.

Hogan

D. B.

(2016). Predictors of long-term care placement in persons with dementia: A systematic review and meta-analysis. International Journal of Geriatric Psychiatry, 31(11), 1151–1171. https://doi.org/10.1002/gps.4449

Department of Health and Aged Care . (2023). COVID-19 outbreaks in Australian residential aged care facilities – 22 December 2023. https://www.health.gov.au/resources/publications/covid-19-outbreaks-in-australian-residential-aged-care-facilities-22-december-2023?language=en

10.

Dick

Garber

A. M.

MacCurdy

(1994). Forecasting nursing home utilization of elderly Americans. In Wise

D. A.

(Ed.), Studies in the economics of ageing (pp. 365–394). University of Chicago Press. https://www-nber-org.eu1.proxy.openathens.net/system/files/working_papers/w4107/w4107.pdf

11.

Gaugler

J. E.

Duval

Anderson

K. A.

Kane

R. L.

(2007). Predicting nursing home admission in the U.S.: A meta-analysis. BMC Geriatrics, 7(1), 13. https://doi.org/10.1186/1471-2318-7-13

12.

Hurd

M. D.

Michaud

P.-C.

Rohwedder

(2014). The lifetime risk of nursing home use. In Wise

D. A.

(Ed.), Discoveries in the economics of aging (pp. 81–109). University of Chicago Press. https://doi.org/10.7208/chicago/9780226146126.003.0003

13.

Ingram

D. K.

Barry

J. R.

(1977). National statistics on deaths in nursing homes interpretations and implications. The Gerontologist, 17(4), 303–308. https://doi.org/10.1093/geront/17.4.303

14.

Kastenbaum

Candy

S. E.

(1973). The 4 per cent fallacy: A methodological and empirical critique of extended care facility population statistics. The International Journal of Aging and Human Development, 4(1), 15–21. https://doi.org/10.2190/28nk-enpj-m7pt-gktk

15.

Liu

(1998). The probability of nursing home use over a lifetime (Welfare Division Working Paper no. 16, Issue. AIHW. https://www.aihw.gov.au/getmedia/84093821-8999-4949-b07f-dc4ef7b9e0b3/wwp16.pdf.aspx?inline=true

16.

Liu

(2000). The probability of nursing home use over a lifetime in Australia. International Journal of Social Welfare, 9(3), 169–180. https://doi.org/10.1111/1468-2397.00126

17.

Luppa

Luck

Weyerer

König

H.-H.

Brähler

Riedel-Heller

S. G.

(2010). Prediction of institutionalization in the elderly. A systematic review. Age and Ageing, 39(1), 31–38. https://doi.org/10.1093/ageing/afp202

18.

McConnel

C. E.

(1984). A note on the lifetime risk of nursing home residency. The Gerontologist, 24(2), 193–198. https://doi.org/10.1093/geront/24.2.193

19.

Palmore

(1976). Total chance of institutionalization among the aged. The Gerontologist, 16(6), 504–507. https://doi.org/10.1093/geront/16.6.504

20.

Rahman

M. M.

Byles

J. E.

(2020). Older women’s patterns of home and community care use and transition to residential aged care: An Australian cohort study. Maturitas, 131, 28–33. https://doi.org/10.1016/j.maturitas.2019.10.004

21.

Rahman

M. M.

Efird

J. T.

Byles

J. E.

(2019). Transitioning of older Australian women into and through the long-term care system: A cohort study using linked data. BMC Geriatrics, 19(1), 286. https://doi.org/10.1186/s12877-019-1291-z

22.

Renzaho

A. M. N.

(2016). Migration and the healthy migrant effect in Australia: Current knowledge, gaps and opportunity for future research. In Renzaho

A. M. N.

(Ed.), Globalisation, migration and health (pp. 363–389). World Scientific. https://doi.org/10.1142/9781783268894_0011

23.

Rogers

(1975). Introduction to multiregional mathematical demography. John Wiley and Sons.

24.

Rogers

(1995). Multiregional demography: Principles, methods and extensions. John Wiley and Sons Ltd.

25.

Rowland

Liu

Braun

(2002). The probability of using an aged care home over a lifetime (1999–00). Australasian Journal on Ageing, 21(3), 117–122. https://doi.org/10.1111/j.1741-6612.2002.tb00431.x

26.

Royal Commission into Aged Care Quality and Safety . (2020). How far do people move to access aged care? [Research Paper](Research Paper 16). https://agedcare.royalcommission.gov.au/sites/default/files/2020-11/research_paper_16_-_how_far_do_people_move_to_access_aged_care.pdf

27.

Vicente

Wiley

J. A.

Carrington

R. A.

(1979). The risk of institutionalization before death. The Gerontologist, 19(4), 361–367. https://doi.org/10.1093/geront/19.4.361

28.

Wershow

H. J.

(1976). The four percent fallacy: Some further evidence and policy implications. The Gerontologist, 16(1 Pt1), 52–55. https://doi.org/10.1093/geront/16.1_Part_1.52

29.

Wilson

Terblanche

(2018). New estimates of Australia's centenarian population. International Journal of Population Data Science, 3(1), 447. https://doi.org/10.23889/ijpds.v3i1.447

30.

Zappolo

(1981). Discharges from nursing homes: 1977 National Nursing Home Survey (Vol 13). (Data from the National Heath Survey Series. https://stacks.cdc.gov/view/cdc/12937/cdc_12937_DS1.pdf

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.45 MB