Article Text

Download PDFPDF

Why is housing tenure associated with a lower risk of admission to a nursing or residential home? Wealth, health and the incentive to keep ‘my home’
  1. Mark McCann1,
  2. Emily Grundy2,
  3. Dermot O'Reilly3
  1. 1Institute of Child Care Research, Queen's University Belfast, Belfast, UK
  2. 2Department of Population Studies, London School of Hygiene and Tropical Medicine, London, UK
  3. 3Centre for Public Health, Queen's University Belfast, Belfast, UK
  1. Correspondence to Dr Mark McCann, Institute of Child Care Research, Queen's University Belfast, 6 College Park, Belfast BT7 1LP, UK; mark.mccann{at}


Background Previous research has shown that home ownership is associated with a reduced risk of admission to institutional care. The extent to which this reflects associations between wealth and health, between wealth and ability to buy in care or increased motivation to avoid admission related to policies on charging is unclear. Taking account of the value of the home, as well as housing tenure, may provide some clarification as to the relative importance of these factors.

Aims To analyse the probability of admission to residential and nursing home care according to housing tenure and house value.

Methods Cox regression was used to examine the association between home ownership, house value and risk of care home admissions over 6 years of follow-up among a cohort of 51 619 people aged 65 years or older drawn from the Northern Ireland Longitudinal Study, a representative sample of ≈28% of the population of Northern Ireland.

Results 4% of the cohort (2138) was admitted during follow-up. Homeowners were less likely than those who rented to be admitted to care homes (HR 0.77, 95% CI 0.70 to 0.85, after adjusting for age, sex, health, living arrangement and urban/rural differences). There was a strong association between house value/tenure and health with those in the highest valued houses having the lowest odds of less than good health or limiting long-term illness. However, there was no difference in probability of admission according to house value; HRs of 0.78 (95% CI 0.67 to 0.90) and 0.81 (95% CI 0.70 to 0.95), respectively, for the lowest and highest value houses compared with renters.

Conclusions The requirement for people in the UK with capital resources to contribute to their care is a significant disincentive to institutional admission. This may place an additional burden on carers.

  • Care home admission
  • house value
  • longitudinal study
View Full Text

Statistics from


The ageing of the population will present challenges for health and social care services for the foreseeable future. For older people with the most complex and demanding care needs residential and nursing home care remains a necessary component of long-term care provision. The funding and provision of residential long-term care is an important and controversial topic of policy debate. Further understanding of socio-demographic factors associated with admission to a care home is important both to inform this debate and to enable better forecasting of future demand which is essential for planning purposes.

Previous research has clearly established that social factors such as living arrangements and levels of informal care support are important factors moderating the risk of admission associated with age, sex and health status.1 2 Socioeconomic status is also important,3 and a range of papers have shown that risk of admission is lower for those who own rather than rent their accommodation.1 3–5 There are several possible explanations for the association. First, people who own their homes may be in better health than those who do not. Previous studies have shown that in older age groups, wealth, of which tenure is an indicator, is more strongly associated with health disparities than income.6 7 Second, home ownership may indicate greater access to resources that facilitate continued independent living, such as ability to pay for private domestic, personal and nursing care and housing adaptations or the ability to relocate to more suitable housing, for example, a bungalow rather than a house with stairs. A third possibility relates to the requirement in the UK for older people with assets, including housing, over c£24k (c30k Euros) to contribute towards the cost of their care. This threshold is considerably lower than the value of virtually all residential property in Northern Ireland (or elsewhere in the UK), which means that admission to residential or nursing home care may involve selling the family home. The reluctance of older people and their families to do this may stimulate an additional amount or intensity of informal caring.5

The aim of this paper is to explore the association between housing tenure and admission to a care home by considering not just tenure itself but also the value of owner occupied accommodation. We posit that if the association between housing tenure and risk of care home admission reflects differences in access to financial resources, then there should be a graded relationship between housing value and admission. If the relationship is due to difference in health status, there should be a graded relationship that disappears or weakens after adjustment for baseline health. However, if house ownership per se is the determinant factor, then no gradient with house value would be expected.


The cohort

The Northern Ireland Longitudinal Study (NILS) is a data linkage study formed by linkage of health card registration system data to 2001 Census returns and other sources of administrative data for a representative c28% sample of the Northern Ireland population.8 The research reported here is based on analyses of a subset of this database comprising a cohort of 51 619 people who were aged 65 years or older and not living in a care home at the time of the 2001 Census. Personal characteristics drawn from census returns included age, sex, marital status, household composition and two indicators of self-reported morbidity. Information on marital status and household composition were used to derive a combined indicator of living arrangements which distinguished those living alone (with separate categories for different marital status groups), those living just with a spouse or partner, those living with children (but no partner), those living with spouse/partner and children and those living with other relatives or friends. The health measures were drawn from two questions, one on the presence of long-term illness limiting usual activities (LLTI) and another on general health (GH) in the preceding year. The LLTI question had a yes/no response; the GH question offered three responses—good, fairly good or not good. The use of the NILS for research was approved by the Office for Research Ethics Committees Northern Ireland.

Housing tenure and house value

The census question on housing tenure identified the 28% of the cohort who were renting (79% of whom were in socially rented accommodation and 21% privately rented); for the remaining 72%, the capital value of their house was appended from a separate data set. This had been derived as part of an exercise in 2005 to determine the level of local tax payable by each household. In approximately 7% of cases, values could not be linked; these are shown here as a separate category. The capital value of residence and housing tenure were combined to produce a variable showing house value for people who own their property (subdivided into five approximately equal sized groups, see table 1), with separate categories for owners with unvalued homes, and people renting their property. As Cox regression models showed no difference in admission rates between social and private renters, these groups were combined in the main analysis.

Table 1

Distribution, proportions and ORs for reporting fairly/not good health and presence of a limiting long-term illness by house value/tenure

Identification of care homes

The census identifies those resident in various types of communal household, but as some misclassification has been reported9 and in order to ensure comparability of sources used to exclude those resident in care and nursing homes in 2001 and those entering these types of homes during the follow-up period, we drew on linked information from the regional care home inspectorate, which has a statutory responsibility for registering care homes. Information on 399 care homes for older people operating during the study period was linked to the 2001 census by the NILS data custodians. These data were used to exclude 4164 (7.4%) people living in care at the time of the census and to identify those who were admitted over the course of the follow-up period. This register excludes non-institutional care facilities such as sheltered accommodation or special housing schemes.

Statistical analysis

The relationship between cohort and household characteristics and risk of admission during the 6 years of follow-up was investigated using Cox proportional hazards models. Respondents who died during follow-up were right censored at date of death. Age was fitted with its mean centred square in models to account for the sharply increasing gradient in admission risk with increasing age. The linear association between house value and admission risk was assessed excluding renters. Departure from a linear trend was assessed by comparing the linear association to a model also containing a squared term for house value. Proportional hazards assumptions were graphically checked for each variable. Two sets of sensitivity analyses were undertaken: the first used three rather than 6 years of follow-up to assess if time since the assessment of living arrangement status affected the results and the second used only admission to nursing home as the endpoint to see if the results also held for those with greater levels of need. There was little change across these models hence they are not presented here.


Descriptive results

Of the 51 619 individuals in the cohort aged 65 years and older and living in the community at baseline, 2138 (4%) were admitted to care homes in the subsequent 6 years, 1563 women (5%) and 575 men (3%). Table 1 shows the strong relationship between house tenure/value and health status. After controlling for age (in single years) and sex, even people in the lowest band of house value reported lower rates of not good GH (OR 0.72, 95% CI 0.67 to 0.77) and LLTI (OR 0.74, 95% CI 0.70 to 0.79) than renters. There was a strong trend towards better health with increasing house value, so that those in the highest value houses had a third the odds of reporting not good health (OR 0.28, 95% CI 0.26 to 0.29) or an LLTI (OR 0.33, 95% CI 0.31 to 0.35).

Table 2 shows the relationship between housing tenure/value and risk of admission to nursing or residential home care over the follow-up period. Model 1 is adjusted for age and sex, model 2 additionally controls for marital status and living arrangement, model 3 for health status and model 4 for urban/rural residence. Coefficients for the control variables are not shown (available on request) but showed the expected patterns. There was a sharply increasing gradient in admission risk with age, and women were more likely to be admitted than men (HR 1.17, 95% CI 1.06 to 1.30). Those in fairly good (HR 1.41, 95% CI 1.24 to 1.62) and not good health (HR 2.09, 95% CI 1.80 to 2.43) had higher admission rates than those in good health, and people with an LLTI had a higher risk than those without (HR 1.36, 95% CI 1.21 to 1.54). Those living alone had higher admission risks than those living with a partner only (the reference category), whether never married (HR 2.00, 95% CI 1.70 to 2.34), widowed (HR 1.48, 95% CI 1.30 to 1.67), divorced or separated (HR 1.51, 95% CI 1.09 to 2.08) or currently married (HR 1.74, 95% CI 1.20 to 2.52). People living with children had a comparable risk to people living with a partner only (HR 0.98, 95% CI 0.81 to 1.17), while people living with children and a partner had a lower risk (HR 0.73, 95% CI 0.57 to 0.92). Those living with other family/non-family had a slightly higher admission risk (HR 1.25, 95% CI 1.01 to 1.54). People living in rural areas had a lower admission risk than those in urban areas (HR 0.76, 95% CI 0.67 to 0.85; models not shown).

Table 2

RR of care home admission by house value and tenure for 51 619 people, adjusted for demographic, health and household characteristics

After 6 years of follow-up, just over 3% of the owner-occupier group had entered residential or nursing home care compared with 7.3% and 7.4% of social and private renters, respectively. No strong gradient in admission rates with increasing house value was apparent: 3.1% of people from the highest house value group entered care homes compared with 3.4% of those in the lowest valued houses. Only 2.5% of the 7% of homeowners without a linked house value were admitted. Results from the Cox proportional hazard models showed that people living in owner-occupied accommodation had lower risks of admission than renters; even those living in the least expensive houses were less likely to be admitted than those renting (HR 0.70, 95% CI 0.61 to 0.81). There was, however, no evidence of a trend towards lower admission risk for those in more expensive houses with overlapping CIs for all categories of house value. Evidence of a linear trend for reducing admission risk with increasing house price was also assessed using models including only homeowners whose house value was known. In fully adjusted models, there was no evidence of a linear association between admission risk and household value (HR per pound increase in house value 0.98 95%, CI 0.94 to 1.02). Comparing the model using continuous house price to the model using the categories above, there was no evidence of a departure from linear trend (χ2 (3 d.f) =0.77, p=0.86). The apparent difference in admission risk occurs only between homeowners and renters. Adjustment for health status at baseline attenuated the difference in risk of admission between renters and owners but had no effect on the already flat gradients across house value.


To our knowledge this is the first study to specifically consider housing value in investigating associations between housing tenure and risk of admission to a care home. At the start of the paper, we posited that there were three principal explanations for the association between housing tenure and admission risk. The first was that homeowners are wealthier and therefore healthier. Although the study has clearly demonstrated a strong and graded association between house value and two indicators of morbidity, there was little evidence of a house value gradient in admission risk even before accounting for variations in health status. A large difference in admission risk persisted between owners and renters after controlling for health status. These results suggest that the better health of owner occupiers is unlikely to be the main underlying explanation of housing tenure differentials in rates of admission to residential and nursing home care.

The second possible explanation was that people who own their home have more financial resources to buy in extra care or home adaptations enabling them to stay at home for longer. However, if capital assets are generally related to financial resources, then residence in a more expensive house should on average be associated with a reduced risk of admission. Again the absence of a gradient of admissions across house value argues against this as a satisfactory explanation.

The overall conclusion of the analysis was that, other things being equal, people in modest houses are as likely as their peers in more expensive houses to be admitted to a care home. The main difference in admission rates is between those who rent and those who own their homes. Owning a house somehow alters the likelihood of being admitted but after that the value of that house makes little difference. Other studies of various indicators of well-being and resilience in later life have reported apparent advantages for homeowners which, it has been suggested, may reflect either prior characteristics associated with housing choices or the sense of security home ownership may convey. However, we suggest that an important factor may be the deterrent effect of means testing which in many cases requires older homeowners to sell their house if admitted to a residential or nursing home. Many families may choose to significantly increase the amount of informal caregiving (and perhaps paid care) rather than see this asset sold.5 10 The value of the property is unlikely to act as a further incentive, given that the threshold at which payment is set is well below the price of virtually all properties. The importance of charging policies on use of institutional care has also been suggested in studies of international differences which have found differences between countries in the relationship between indicators of socioeconomic status and admissions hypothesised to reflect different national charging and copayment requirements for institutional care.

This is a large-scale longitudinal study that is representative of the Northern Ireland population and includes validated indicators of care home residence. All of the cohort attributes were ascertained at baseline and reverse causation is not a possibility. There are, however, some limitations that are should be mentioned. The health measures in the census are self-reported and are probably only moderately related to limitations in activities of daily living which are more closely related to risk of admission.5 11 12 However, they have been closely related to future well-being and mortality risk,13 and in any event, any residual confounding is likely to result in an underestimation of the differences between renters and owners rather than across house values. A further limitation is that there was no indication of available income for cohort members, and it is recognised that householders may be ‘asset rich but income poor’, though house ownership is generally associated with higher income in the older population. Older people in the UK obtain income most often from state pension, occupational pension, disability benefit, investment income and (self) employment income. Less common sources of income include arts or land royalties, rent or interest payments.14 We also lack information on receipt of care at home, whether from family members or paid providers, which would allow us to explicitly examine associations between housing tenure, wealth and receipt of care at home.

UK long-term care policies have been the subject of considerable debate for some time and are currently being reviewed once again by the Commission on the Funding of Care and Support established in 2010. The conclusion that housing tenure per se may be reducing the risk of admission, presumably at least in part by incentivising family members to provide more care, raises some important questions. Given that levels of owner occupation are increasing, all else being equal, it might imply reducing demand for residential and nursing home care. However, the apparent reluctance of families to use housing capital to contribute to care costs suggests difficulties for proposed funding plans that rely on such pooling of individual and state resources and for various market solutions that also rely on drawing down capital from housing. Apart from the very important issue of funding, there is also considerable debate about what level of support it is reasonable to expect family members and other unpaid carers to provide. The Wanless review of social care for older people suggested that formal services should substitute for ‘filial’ care from children to avoid conflict with adult children's labour market roles.15 The last government's Green Paper on social care also identified the provision of more than 50 h/week of care as unsustainable and potentially harmful to the caregiver.16 More research is needed to assess whether and how current means testing procedures impact on the provision of family care. Finally, we also need to discover whether there are particular features of living in rented accommodation that make remaining at home harder for older people with disabilities.

What is already known on this subject

Homeowners have a lower risk of care home admission than renters. The extent to which this is an indication of better health, better ability to purchase care or greater effort to maintain home ownership is unclear.

What this study adds

This study looked more closely at these three mechanisms by investigating home ownership and house value simultaneously. Home ownership per se rather than the value of one's home is associated with reduced admission risk. This suggests that UK homeowners' requirement to pay for their care is a major disincentive to admission, shifting the burden of care towards informal carers.


We acknowledge the help provided by the staff of the NILS and the NILS Research Support Unit. Many thanks to the RQIA with their help in preparing care home data.


View Abstract


  • The authors alone are responsible for the interpretation of the data.

  • Funding The Northern Ireland Longitudinal Study (NILS) is funded by the Health and Social Care Research and Development Division of the Public Health Agency (HSC R&D Division) and NISRA. The NILS-RSU is funded by the ESRC and the Northern Ireland government.

  • Competing interests None.

  • Ethics approval This study was approved by ORECNI.

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

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.