Aging in Place: Literature Review for Bridgepoint and Toronto Central LHIN

Institute for Life Course and Aging

U N I V E R S I T Y O F T O R O N T O

Aging in Place

Bridgepoint/LHIN Literature Review

Prepared by:

The Institute for Life Course and Aging

University of Toronto

Director: Dr. Lynn McDonald

Research Coordinator: Julia Janes

November 23, 2007

Institute for Life Course and Aging University of Toronto

2

TABLE OF CONTENTS

Table of Contents.................................................................................................................................... 2

Executive Summary................................................................................................................................ 5

Background........................................................................................................................................... 5

Findings ................................................................................................................................................ 6 Predisposing Characteristics............................................................................................................ 6 Enabling Characteristics .................................................................................................................. 6 Need Characteristics ........................................................................................................................ 7 Use Characteristics .......................................................................................................................... 7

Limitations............................................................................................................................................ 8 Methodological limitations in the literature: ................................................................................... 8 Content missing in the literature: ..................................................................................................... 8

1. Introduction........................................................................................................................................ 9

1.1 Methodology........................................................................................................................... 10

1.2 Terminology ........................................................................................................................... 11

1.3 Organization of the Literature Review ................................................................................... 12

1.4 Background............................................................................................................................. 13 1.4.1 Canada’s Aging Population ........................................................................................... 13 1.4.2 Seniors’ Living Accommodation..................................................................................... 14

2. Predisposing Characteristics .......................................................................................................... 15

2.1 Demographic Factors: Age, Gender and “Race”.................................................................... 15

2.2 Social Support......................................................................................................................... 16 2.2.1 Marital Status and Living Alone..................................................................................... 17 2.2.2 Adult Children and Other Kin Support........................................................................... 19 2.2.3 Caregiving ...................................................................................................................... 20 2.2.4 Dimensions of Social Support......................................................................................... 23

2.3 Beliefs and Expectations......................................................................................................... 25

2.4 Health Behaviours: Smoking, Alcohol Consumption and Physical Activity ......................... 27

3. Enabling Characteristics................................................................................................................. 28


3

3.1 Familial/Household Resources ............................................................................................... 28 3.1.1 Income............................................................................................................................ 29 3.1.2 Wealth ............................................................................................................................. 31 3.1.3 Education........................................................................................................................ 32 3.1.4 Homeownership .............................................................................................................. 32 3.1.5 Housing Conditions: Quality, Accessibility Accommodations, Assistive Technology and

Supportive/assisted Residential Models ......................................................................... 33 3.1.6 Type and extent of healthcare insurance ........................................................................ 40

3.2 Community, Market and Policy Resources ............................................................................ 41 3.2.1 Housing Market Conditions............................................................................................ 41 3.2.2 Neighbourhood Characteristics and Place Attachment ................................................. 41 3.2.3 Balance of Care: Availability of and Expenditure on Home and Community-Based

Services and Institutional Long-term Care..................................................................... 43

4. Need Characteristics........................................................................................................................ 49

5. Health Care Utilization ................................................................................................................... 54

5.1 Prior Nursing Home Admission and Hospitalization ............................................................. 55

5.2 Other Measures of Use: Paid Helpers, Doctors and Medications .......................................... 56

6. Conclusion ........................................................................................................................................ 57

6.1 Limitations of the Literature................................................................................................... 57

6.2 Summary: Predisposing Characteristics ................................................................................. 61

6.3 Summary: Enabling Characteristics ....................................................................................... 62

6.4 Summary: Need Characteristics ............................................................................................. 63

7. References …………………………………………………………………………………………65

8. Appendices........................................................................................................................................ 77

8.1 Appendix A: TABLES ........................................................................................................... 77 Table 1 Definitions of Accommodation and Care Settings ....................................................... 77 Table 2 Living Arrangements, Ontario, 1996, by Gender (65+, 65-74, 75-84, 85+) ................ 78 Table 3 Living arrangements of seniors aged 65 and over by sex and age group, 2001.......... 78 Table 4 Number of Equations (N) Reporting Positive Significant (+), Negative Significant (–),

and Non-significant (NS) Associations among Predictors and Adverse Outcomes.... 79

8.2 Appendix B: FIGURES .......................................................................................................... 82 Figure 1 Andersen’s Behavioral Model ..................................................................................... 82 Figure 2 Population Projections for Canada, Provinces, and Territories, 2005-2056 ............ 83


4

Figure 3 Age of Toronto’s Senior Population ........................................................................... 84 Figure 4 Percentage of Canadians in Good Health, by Age Group, Household Population,

Aged 65 and Over........................................................................................................... 85


5

EXECUTIVE SUMMARY

Aging in Place: a Review of the Literature

Background

The purpose of the literature review was to document the research that examines the predictors

of the institutionalization of the elderly. The review is grounded in the framework of “aging in place,”

which acknowledges that older adults wish to live in their own communities for as long as possible and

that home and community services will support this aim while being cost-effective.

The review is organized according to Anderson’s Behavioural Model of health services, which

groups predictors into four categories: predisposing, enabling, need and use characteristics.

Predisposing characteristics include demographics, levels and characteristics of social support, health

behaviours, beliefs and expectations. Enabling characteristics include familial and community

resources like neighbourhood context, supply of long-term care facilities and home and community-

based services, the balance of health care expenditures and health policy. Need characteristics include

indicators of self-rated and practitioner-evaluated health status. Use characteristics are essentially

indicators of health care utilization.


6

Findings

Predisposing Characteristics

� Consistent predictors of institutionalization: age, Caucasian, living alone, having

fewer adult children and having low levels of community engagement;

� Modest predictors of institutionalization: lower levels of informal supports

independent of caregiver status, experience of severe loneliness;

� Interventions: targeted to those at the highest risk, as well as their caregivers, taking

advantage of existing care networks;

� Future research: aim to isolate the many dimensions of informal support and evaluate

their effect on different groups and in different contexts.

Enabling Characteristics

� Consistent predictors of institutionalization: not owning a home, living in an area

with a greater supply of nursing home beds;

� Modest predictors of institutionalization: lower household wealth and income in

terms of perceived adequacy rather than dollar value, living in areas with few affordable

small unit rentals, less access to assistive technologies and home modifications; living

in a socially “deprived” and poorly serviced neighbourhood, having “weak ties” to the

neighbourhood;

� Interventions: develop more age-appropriate affordable supportive housing, expand

utility subsidies and emergency funds; enhance neighbourhood health and social service

infrastructure;

� Future research: investigate issues such as availability and access to supports; attempt

to isolate features of “liveable” housing and communities that foster healthy aging and

independent living.


7

Need Characteristics

� Consistent predictors of institutionalization: cognitive impairment, limited activities

of daily living (ADLs); impaired instrumental activities of daily living (IADLs);

depression, dementia and Alzheimer’s disease;

� Modest predictors of institutionalization: self-reported health status, self-reported

level of functioning; stroke and heart disease, digestive systems diseases, urinary and

bowel incontinence;

� Interventions: develop programs that allow aging in place for the cognitively impaired

and the frail elderly, similar to PACE programs;

� Future research: longitudinal, panel research with multiple measures of baseline

factors at a number of time intervals including time of placement.

Use Characteristics

� Consistent predictors of institutionalization: prior nursing home admission and

hospitalization

� Modest predictors of institutionalization: presence of formal help (paid or

professional care/services by doctors, home health aides, PSWs);

� Interventions educate older adults and their caregivers about community-care options;

ensure active involvement in discharge planning and the development of comprehensive

care packages;

� Future research: clarify the explanatory effect of prior use as to whether it is due to

greater acceptance by an older adult because of a previous placement or a bias of

service providers/discharge planners; disentangle the different types of formal help and

the mechanisms that impact residential outcomes.


8

Limitations

Methodological limitations in the literature:

� The prediction of change, based on single cross-sectional baseline measures;

� Predictors of a change may be different from predictors of the type of change;

� Predictors of home and community-based care may be different than those that predict

moves to institutional settings;

� Easily measured predictors (e.g. scale–based health indicators) are more commonly

investigated than those that are more complex (e.g. community, market and policy

resources);

� Few random control trials, random sampling and qualitative studies.

Content missing in the literature:

� Attention to entries and exits from nursing homes to understand what leads to

institutionalization and what facilitates returns to the community;

� Investigation of the interface between hospital discharge/rehabilitation strategies and

nursing home use;

� Examination of the factors behind over and under care, particularly the role of health

policy/expenditures;

� Consideration of the impact of caregiver well-being on the recipients’ placement

outcomes;

� Evaluation of the clinical and quality of life outcomes for various residential options to

understand why a particular setting is the best possible fit for an older adult;

� Evidenced-based evaluation of the many promising community care and supportive

housing programs.


9

1. INTRODUCTION

The ability to predict what factors lead to the institutionalization of older adults is relevant to

the current policy of aging in place in Ontario. The policy for aging in place is significant because the

majority of older adults prefer to age in their own communities while maintaining their health,

autonomy and dignity. At the same time, the policy is advantageous to society as a whole because the

support of older adults with home and community services is cost-effective when compared to

expensive institutional long-term care. In contrast, the dangers of such a policy are the potential for

over or under care and the implications these inappropriate levels of care would have for older adults

and for the functioning of the health care system.

The extent of over care and under care is a clear signal that programs and policy frameworks

would benefit from more extensive implementation of evidence-based practice. The few extant studies

that explore the prevalence and pathways to inappropriate care indicate that a proportion of older

adults reside in more restrictive settings than is necessary or they do not receive the care they need.

Berthelot and colleagues, in their analysis of Canadian data from the National Population Health

Survey, found that 10 percent of adults 65 years of age and older with no disability resided in long-

term care facilities (Berthelot et al., 2000). Although this proportion is modest, it still warrants

questioning whether these individuals could be better accommodated in less restrictive settings.

In an American study that used data from a sample of 3,170 older adults residing in long-term

care facilities, it was estimated that between 15% to 70% could be appropriately cared for in less

restrictive settings (Spector, Reschovsky & Cohen, 1996). Even at the most conservative criterion

level, the proportion of older adults receiving too much care in this American sample is alarming.

Coyte and colleagues in their revision of the forecast produced for the Health Services Restructuring


10

Commission in Ontario noted that "many people who currently are being admitted to Long-Term Care

facilities would be able to receive care in their own homes or in a supportive housing setting" (Coyte et

al., 2002 p.9). It may well be that some older adults find themselves in institutional settings for reasons

other than their health, functional or cognitive problems.

Under care is an equally disturbing phenomenon. A recent study by the American Association

for Retired Persons found that almost one-third of the 865 older adults surveyed reported having unmet

needs for personal assistance (AARP, 2006). Davey and colleagues found that of those Americans in

the highest risk group for nursing home placement, a high proportion (one in ten) were without any

support whatsoever (Davey et al., 2005).

This literature review was prepared for the Toronto Central Local Health Integration Network

Seniors Council. The Council is comprised of representatives from seniors’ agencies, physicians,

aboriginal health leaders, and consumers. It is the body responsible for implementing the “Plan for

Seniors”, a component of the “Integrated Health Services Plan” (Toronto Central Local Health

Integration Network, 2007). A priority is to enable seniors to live independently in the community for

as long as possible (Toronto Central Local Health Integration Network (LHIN, 2007). In considering

this priority, the Council posed the following questions, “What factors prevent seniors from remaining

in their homes and communities?” and “What are the factors that predict the move of older adults to

more restrictive settings?” This review of the literature is was undertaken in response to these

questions.

1.1 Methodology

The search of the literature entailed a scan of 377 journals, books, and government documents

of which 114 were reviewed. All were published between 1993 and 2007. Excluded studies included


11

those that were inappropriate for reasons such as the methodology used, they were primarily about

screening tools, they were dated, or they were not consistent with the intent of the review. The majority

of the publications were from the fields of social work, sociology, psychology, nursing, medicine, and

health economics. Databases reviewed included Google Scholar, Scholar’s Portal, and Web of

Knowledge. Where possible, the search emphasized relevant and current Canadian studies. The review

focussed on predictors of the move of older adults (65 years of age and older) to more restrictive

residential settings. The main search terms were old*, elder*, senior*, Engl*, Canad*, United States,

Europe*, Australi*, health*, predict*, factor*, aging in place, social support, informal support,

caregiving, housing, supportive or assisted, home care, paid help*, neighbourhood, home

modifications, assistive technol*, community-based healthcare, health policy, socio-economic status ,

nursing home and institutionaliz*. Searches yielded documents that dealt with health and socio-

economics status, social support, housing, community and policy environments, health service

utilization, activities of daily living, instrumental activities of daily living, chronic poor health, health

concerns, and health crises. The majority of the studies were quantitative population-based, random

control trials (RCT), and longitudinal studies.

1.2 Terminology

For the purposes of this literature review, the term “senior” refers to those aged 65 and older.

The terms “senior”, “older person”, “older adult,” and “later life” will be used interchangeably,

although in reality, older adults are a very diverse group who age at different rates and in different

ways.

Commonly, institutionalization refers to permanent admission to a more restrictive residential

setting in a healthcare setting or other facility (Hope, Keene, Gedling, Fairburn, & Jacoby, 1998),


12

where there are professional and personal care services available. A facility may be a chronic care

hospital, long-term care institution, such as a nursing home, home for the aged, psycho-geriatric

setting, retirement home (if adjunct services are purchased to maintain independence), or other seniors’

institutions (please see Table 1). For the purposes of this review, the definition of institutionalization

does not include admission for respite care, the use of home care, assisted living/supportive housing, or

enrolment in day programs, although doing so may provide care levels comparable to that offered in

institutional settings

1.3 Organization of the Literature Review

The fundamental components of this literature review include a survey of pertinent contextual

and demographic information as they predict institutionalization. This literature review is organized

according to a widely used model created by Andersen and colleagues (Hancock, Arthur, Jagger, &

Matthews, 2002; Miller & Weissert, 2000) which is appropriate to the population and research

questions guiding this review. According to the model, predictors of institutionalization can be

organized in the following schema: predisposing, enabling, and need characteristics of the individual.

Predisposing characteristics include demographics, levels and characteristics of social support and

beliefs and expectations. Enabling characteristics include familial and community resources (e.g.

neighbourhood context, supply of long-term care facilities and home and community-based services,

the balance of health care expenditures and health policy). Need characteristics include indicators of

self-rated and practitioner-evaluated health such as health status, functional status or disability. More

recent versions of Andersen’s model include “use” characteristics (utilization of health care services)

in the schema (Andersen, 1995; Gelberg, Andersen, & Leake, 2000). Figure 1 describes Anderson’s


13

behavioural model, while Table 4 provides an example of the model using the summary findings from

Miller and Weissert’s recent review of the literature (Miller and Weissert, 2000).

1.4 Background

1.4.1 Canada’s Aging Population

According to Statistics Canada, in 2006, Canadians aged 65 and over comprised 13.7% of

Canada’s population or 4.3 million persons (Statistics Canada, 2007a). In Canada, between 2006 and

2056 (Figure 2), the number of seniors, is expected to increase significantly. In fact, by 2031, seniors

will account for approximately one-quarter of Canada’s population (Statistics Canada, 2005).

Moreover, Toronto’s City Planning Division projects that the overall number of seniors will grow by

38% between 2006 and 2031 (Toronto, 2007b). A similar trend is evident in the Toronto Central

LHIN’s senior population whose growth trajectory until 2016 is depicted in Figure 3.

As is the case with the overall senior population, the number of people in the oldest age groups

is expected to increase in the approaching decades. Canada’s population aged 80 and over will almost

double, from 1.2 million in 2006 to 2 million in 2026 (Statistics Canada, 2007b), and will constitute

the second largest increase (+25%) in population of any age group (Statistics Canada, 2007b) between,

2001 and 2006. By comparison, over the same time period, the same age group in Toronto has grown

by 30%.(Toronto, 2007b).

Today’s cohort of seniors is aging well; they are in better physical and mental health than were

their predecessors. Moreover, their financial situation is improved, a key factor related to good health

(Statistics Canada, 2003b). As seen in Figure 4, approximately 80%, of older adults living at home rate

their health as good or better, although this perception changes as seniors age (Shields & Martel,


14

2006). Regardless, academics and policymakers continue to be concerned about the impact on the

economy of the rapidly increasing 80 plus population (Dang, Antolin, & Oxley, 2001).

1.4.2 Seniors’ Living Accommodation

The increasingly large numbers of older people will likely make more demands on home and

health care services (Robson, 2001), including institutional care, notwithstanding the improvements in

health care and the concomitant increases in Canadians’ life expectancy. Although 90 percent of older

adults live in private households (National Advisory Council on Aging, 1999), and prefer not to live in

institutions (Division of Aging and Seniors, 2006), this percentage diminishes with age: with less than

one-half of seniors over the age of 85 living alone (see Table 3).

As seniors reach older ages, they are much more likely to become institutionalized as seen in

Tables 2 and 3. Approximately 7 percent of all older adults live in long-term care facilities in Canada.

Twice as many aged 75 and older live in long-term care facilities in Canada (National Advisory

Council on Aging , 2005 as cited by (Division of Aging and Seniors, 2006). The Toronto figures are

consistent with National estimates: 7 percent of senior Torontonians are institutionalized (Toronto,

2007a). In Canada (excluding Quebec), there were 5,024 more seniors in institutions1 in 2004-2005

than in 2003-2004 indicating that the number of seniors who live in institutions in Canada (excluding

Quebec), rose by 3.3 percent.

1 Homes for senior citizens, retirement homes, or lodges, where there is no care provided, are not included.


15

2. PREDISPOSING CHARACTERISTICS

Predisposing characteristics are more distal influences on the use of health services and the

likelihood of residing in more restrictive (therefore service rich) settings. This review considers

demographic factors, various dimensions and levels of social support and caregiving, as well as beliefs,

expectations and health behaviours associated with residential outcomes.

2.1 Demographic Factors: Age, Gender and “Race”

A systematic review published in 2000 by Miller and Weissert which focussed on the

predictors of institutionalization among American seniors (Miller & Weissert, 2000), concluded that

gender (female) was not a significant predictor of institutionalization. However, a large (N= 2805), 14-

year Australian, longitudinal study (McCallum, Simons, Simons, & Friedlander, 2005) of non-

institutionalized older adults did find that being female, independent of the effect of advanced age,

increased the risk of institutionalization.

Inconsistent results for the effect of gender may be due to that fact that gender exerts an

influence because women have a greater life expectancy than men. For example, in Canada women, on

average, live longer than do men (82.5 years compared with 77.7 years, in 2004) and they represent

two-thirds of those over age 80 (Statistics Canada, 2005). Consequently, in Canada, gender differences

exist, with more females hospitalized (National Advisory Council on Aging, 1999) and

institutionalized than are males (Statistics Canada, 2003a).

Likewise, a National Institute of Health study (N=236) found that predictors of

institutionalization included gender and age, and specifically age of onset of dementia (Stern et al.,

1997). As well, another study of approximately 9000 Canadian seniors participating in the Canadian

Study of Health and Aging, found that age and gender predicted institutionalization, with age being the


16

best of any predictors (Gutman, Stark, Donald, & Beattie, 2001). Consistent with Gutman and

colleagues findings about age, Miller and Weissert in their systematic review found, in 42 out of 55 of

the studies that age was positively associated with institutionalization, and in 29 out of 33 studies,

being “non-white” decreased the chances of institutionalization (Miller & Weissert, 2000). American

researchers studied 4,646 participants in a day program for seniors age 55 and older whom either

subsequently had been or not been institutionalized (Friedman, Steinwachs, & Rathouz, 2005). They

examined the predictors of nursing home placement occurring within 3 years, and found that among

community-dwelling seniors, advanced age and identifying as “white” were associated with greater

likelihood of institutional living.

2.2 Social Support

Social functioning and its effect on institutionalization are measured in a variety of ways in the

research literature. Informal support is often assumed if a spouse or adult child is present. Being

married (including previous marital status) and the presence of surviving adult children (including the

number of children and their gender) frequently are used as predictors of institutionalization. It is more

common to measure social support in terms of instrumental support (banking, shopping) compared to

psycho-emotional support. Similarly, it is more common to measure support in terms of direct

assistance provided by family/friends versus indirect support through the purchase of services. Other

less common measures of social support are frequency and community engagement indicators such as

casual (intermittent) versus dedicated support (continuing), and social activities such as volunteerism

and attending faith-based groups.

Most studies recognize the complexity of using social functioning as a predictor of

institutionalization (Freedman, 1996; Kersting, 2001; Wilcox, 1995) and acknowledge that the concept


17

is multi-faceted and not consistently operationalized in the literature. Cox (2005) raises another issue:

that quantitative measures overlook subjective assessments of the quality of support provided. Further

complicating the interpretation of the effects of social functioning is that the presence of support or

caregiving is positively associated with the same variables that predict residing in a more restrictive

setting. As a result, some investigations have found that the presence of social support is associated

with greater risk of nursing home admission. For example, Lo Sasso and Johnson (2002) observed that

using a simple model of nursing home entry resulted in the finding that those who received informal

care were significantly more likely to use nursing home services than those who did not receive

informal care. However, when Lo Sasso and Johnson (2002) used a more complex equation, the author

found that help from adult children with ADL was associated with a 57% reduction in nursing home

admissions.

2.2.1 Marital Status and Living Alone

The majority of studies reviewed measure “within household support” such as existence of a

spouse and number of adult children, as well as some analyses of the gender and family status of adult

children. The two most consistent predictors are marital status (being married or not) and living alone

(the presence or absence of other adults in the household). However, some authors caution that these

two measures are frequently co-linear and poorly constructed. For example, Kersting (2001) notes that

“being married” and “not living alone” may tap the same underlying construct of dedicated support in

the home. If this is the case, it may be redundant to include them both in the analyses. Furthermore,

there have been few analyses that have investigated whether the presence of a spouse necessarily

implies instrumental support or whether none or other types of support are provided.

Nevertheless, as predisposing factors, being married and having adult children typically are

associated with a lower risk of nursing home placement, while living alone is a consistently strong


18

predictor of increased risk of institutionalization. One study found that living alone was a powerful

predictor of discharge from a non-acute geriatric hospital to a nursing home (Aditya, Sharma, Allen, &

Vassallo, 2003). Other authors have found a similar significant effect for living alone (Akamigbo &

Wolinsky, 2006; Gaugler , Duval, Anderson & Kane, 2007; Kersting, 2001; Mutchler & Burr, 2003).

Not only is living alone a catalyst for nursing home entry but Grando and colleagues found that living

alone was also a key factor in explaining why older adults remain in a nursing home even when care

needs could be met in a less supportive setting (Grando et al., 2002). This is significant finding given

that, according to 2007 Toronto census data, there were 89,790 seniors living alone in Toronto in 2006,

an increase of 5.4% since 2001(Toronto, 2007a).

Furthermore, the consistently high risk of institutionalization associated with living alone may

make targeting home and community-based support to this group worthwhile. For example, Beland

and colleagues (2006) found that a Quebec model, the System of Integrated Community-based Care

(SIPA), significantly delayed time to nursing home admission, but only for those older adults who

were living alone or those who had few chronic conditions.

Yet some studies yielded non-significant findings for marital status and living alone. For

example, McCallum et al.,(2005) longitudinal study of Australian seniors found both variables to be

non-significant predictors of institutionalization but cautioned that these findings may be due to the

excessively long follow-up period of 14 years during which many of those married at baseline would

have lost a spouse (McCallum et al., 2005). Interestingly, this study investigated the quality of support

not just the frequency or existence of social contact. These more nuanced measures of social support

may have greater explanatory power than the blunter measure of marital status.


19

2.2.2 Adult Children and Other Kin Support

Generally, research shows that non-spousal family involvement with the older adults prevents

institutionalization. One study found that proximity to family was the top priority regarding future

moves for the seniors in their study (Weeks, Branton, & Nilsson, 2005). The researchers also found

that older adults with little family support were significantly more likely to access homemaking

supports and consider group living as a future housing preference suggesting that family support plays

a crucial role in determining future place of residence.

Akamigbo and Wolinsky (2006) found that having two or more children reduces the chances of

nursing home placement. A large scale, cross-sectional study of older women (N = 75,962) found that

women living in institutions were more likely to have had fewer children than were those who were

still living in the community (Burr, Mutchler, & Warren, 2005). Likewise, Burr and colleagues found

that having fewer children was associated with higher odds of institutional living for both widowed,

separated, or divorced women and men (Burr et al., 2005). Gaugler et al., (2007) meta-analyses of 77

longitudinal studies, drawing on 12 data sources with a sample of 175,056, estimated that having more

children was consistently associated with lowered odds of institutionalization (Gaugler et al., 2007).

Kersting (2001) found that residing with a child and the number of siblings were both

associated with lower odds of subsequent nursing home admission. This longitudinal study employed

comprehensive measures of social support, including community engagement, and contact with family

and friends. However, even after expanding his measures of social support, the author did not find a

significant effect for contact with family and friends. This signals that the indicators may not be

measuring the presence of available support or the quality of the support. Kersting adopted Cox’s

(2005) recommendation of including measures of quality and measured the quality of social activity,


20

not just the frequency of engagement. As was expected, Kersting found that the measure of quality was

not only significant, but was the only measure of social support, which reached significance.

Freedman (1996) found that having a daughter reduced the possibility of nursing home

admission by 27 percent, and having a sibling by 21 percent. Freedman goes on to suggest that future

studies should attempt to isolate the mechanisms by which family structure may affect the risk of

institutionalization. Research of this nature would provide evidence that would inform interventions.

Pathways for future exploration include kin providing personal care directly, kin assisting with

obtaining formal home and community-based supports and/or making financial transfers to secure

these supports. Further investigations could include emotional support and how it affects health and

well-being. As well, measurements such as Lubben’s social network scale, which taps the potential for

and quality of support available from kin and non-kin, may help produce more precise analyses of

support (Lubben et al., 2006).

2.2.3 Caregiving

The literature on informal support focuses on the impact of dedicated caregiving to the provider

and the recipient. As noted earlier, the effects of caregiving on institutionalization need to be carefully

examined, as the same factors that are associated with the presence of a caregiver are associated with

risk of nursing home admissions. Therefore, the findings for informal support by a caregiver are often

mixed. For example, Gaugler et al., (2007) meta-analysis found that, overall, the presence of a

caregiver was a strong predictor of nursing home admission (Gaugler et al., 2007). On the contrary,

Miller and Weissert found that in 67 percent of the studies reviewed, there was reduced risk for those

with greater levels of caregiver support (Miller & Weissart, 2000). This latter measure of the relative


21

degree of available caregiving support may be a more effective way of assessing the impact of

caregiving than the blunter measure of whether or not caregiving is available.

Eloniemi-Sulkava and colleagues investigated whether a community-based nurse management

intervention targeted at both the care recipient and caregiver could delay nursing home admission.

Using a randomized control trial, the authors found that the intervention had significant positive impact

in the early months of the two-year study (Eloniemi-Sulkava et al., 2001). However, the effect

continued only for those with the most severe levels of dementia who averaged 647 days prior to

nursing home admission versus 396 days for the control group.

Beland et al., (2006) evaluated the effect of a Quebec-based community intervention, SIPA, for

“frail” older adults on hospital and nursing home admissions (Beland et al., 2006). The researchers

found reductions in hospitalization and in nursing home admission but only for those with few chronic

conditions and for those who lived alone. They also found that the intervention resulted in greater

caregiver satisfaction, as well as no increase in caregiver burden. This study demonstrated that

caregiver satisfaction is critical to sustaining older adults in less restrictive living arrangements, and

that longer community tenure is not necessarily achieved at the cost of increasing caregiver burden.

A time-series longitudinal study predicted institutionalization of Spanish dementia patients

based on caregivers' health and quality of life (Argimon, Limon, Vila, & Cabezas, 2005). The 181

caregivers, averaging age 63, completed a survey that measured social support, as well as patients'

problematic behaviour. The placement rate in nursing homes was 10.5 percent as assessed at six and

twelve months. Patients, whose caregivers continued to enjoy good physical health and were

functionally capable, were less likely to be placed in a nursing home. Similarly, a 5-year longitudinal

study of 123 cognitively impaired seniors and their caregivers highlighted the degree to which


22

caregivers’ burden influences decisions surrounding institutionalization (Strain, Blandford, Mitchell, &

Hawranik, 2003).

Using data from the Canadian Study of Health and Aging, Hébert and colleagues’ longitudinal

study of caregivers of dementia patients (N=326) found that those who were institutionalized were

more likely to have severe Alzheimer's disease and to have a caregiver who was not a spouse or child,

but was over age 60 and who felt an extreme burden because of caregiving (Hébert, Dubois, Wolfson,

Chambers, & Cohen, 2001). The authors recommended assessing caregivers for signs of stress, and

creating interventions to reduce their stress.

Some authors examined whether caregiving is a substitution or a complement to formal paid

supports. Davey et al., (2005) compared the relative uptake of formal and informal care in American

and Swedish samples, and found evidence that in the United States, the relationship between informal

and formal support is more one of substitution, whereas in Sweden, formal care is primary and

informal care is complementary (Davey et al., 2005). This finding highlights how context, whether

socio-cultural or political, can have a significant influence on the quantity and quality of caregiving. In

contrast, Cox suggests that the interaction between informal caregiving and formal paid supports is

determined by caregiver characteristics (Cox, 2005). She explains that many caregivers do not access

formal supports until the burden reaches a critical threshold at which time these supports tend to

supplement deteriorating informal care. However, if caregivers use formal supports preventively rather

than remedially, then the relationship between the two would be more complementary. Such findings

suggest the importance of supports and early outreach to caregivers to extend community tenure for

older adults.

Finally, no discussion of caregiving can ignore the gendered nature of caregiving and the

effects that changing demographics and dynamics of employment will have on informal support. The


23

forecast for institutional and home care for older Ontarians by Coyte et al., (2002), stresses that the

increased participation of women in the workforce has implications for future family caregiving

(Coyte, Laporte, Baranek, & Croson, 2002). About 12 percent of women compared to 6 percent of men

retire to caregive in Canada while about 15 percent of women are double caregivers for aging parents

and children at the same time (McDonald, 2006). Although little attention is paid to retiring to

caregive, income from wages and opportunities for investment are lost, eligibility requirements for

pensions are not met, the accumulation of benefits and interest are frequently forgone, and savings are

spent. These penalties may slow down the move of women to leave the labour force to care. Worse still

is the struggle that many women have to rejoin the labour force when the caregiving is over and all

professional support has evaporated (McDonald, Sussman, & Donahue, 2007).

2.2.4 Dimensions of Social Support

Some of the studies examined expanded measurements of social support to explore specific

characteristics of social support and differences across contexts. Lo Sasso and Johnson’s study found

differential effects for the nature of social support: help with ADL was significant but when help was

broadened to include IADL, the findings were less than significant (Lo Sasso & Johnson, 2002).

Consequently, much of the instrumental support provided by adult children (outside of personal care)

may not be associated with lower risk of nursing home admission.

Most studies take-for-granted the instrumental dimension of social support in mediating

nursing home risk through direct or indirect support for activities of daily living (ADL and IADL).

However, Russell and colleagues focused on the emotional component of support by examining the

dimension of loneliness and the mechanisms by which it influences living arrangements (Russell,

Cutrona & Wallace, 1997). Loneliness is understood to be separate from social isolation in that it refers

to a perceived deficit in the quantity and/or quality of social interaction. The authors found a positive


24

association, over the duration of their four-year longitudinal study, between the level of loneliness and

moves to settings that are more restrictive. However, after controlling for other dimensions of social

support and contact, as well as a host of socio-demographic and health variables, only those

participants scoring in the range of severe loneliness were six times more likely to be admitted to an

institution. The fact that this association was independent of health and mental health factors, and other

measures of social support led the authors to hypothesize that older adults experiencing severe

loneliness may be entering nursing homes with the expectation of accessing a “built in” social network.

Interestingly, even though social support was measured as the availability of family and friends in an

emergency (versus a crude measure of frequency of contact) it did not have a significant effect on

nursing home admissions. In fact, the only significant social support predictor of institutionalization

was attendance or affiliation with a religious group. Although this study and a few others included in

this review have looked at the effect of social support provided by friends and the community at large,

there is a notable gap in the examination of effects for non-kin social support.

As with a number of other studies (Hays, Pieper, & Purser, 2003; Kersting, 2001) reviewed,

community engagement or social activity may be a better measure of social capital than variables

associated with contact with or support provided by friends and family. Hays and colleagues also

attempted to isolate instrumental from emotional support by using both instrumental aid and

availability of a confidante (a proxy for the emotional dimension) to assess impact on

institutionalization, but found that both factors were not significant predictors of either

institutionalization or household expansion (Hays et al., 2003).

One study investigated how different types of care networks affect the health of older adults

with dementia and/or severe mental illness (Wilcox, Jones, & Alldrick, 1995). In this study, higher

need and, as a result, greater risk of institutionalization were significantly associated with low levels of


25

social support. The researchers also found that the most common network type: the “locally integrated

network,” which includes both kin and non-kin support, was associated with longer community tenure

and more active community engagement. Further, the authors suggest that this type of network can

substantially delay the need for residential care and is amenable to support from service providers. For

example, integrating a paid “good neighbours” network into care planning and education and supports

for caregivers is effective. The researchers also note that consideration of the impact on caregivers of

longer community tenure must be evaluated and considered in all community-based long-term care

interventions.

Another area that warrants attention is that of the effect of environmental context on social

support. Beland and colleagues investigated the relative effects of regional differences or

neighbourhood characteristics on various social measures and found that different types of social

networks have varying impacts, depending on the environmental context (Beland et al., 2004). The

researchers compared social functioning between a relatively “deprived” neighbourhood in Montreal,

Quebec and a more affluent sample from Moncton, New Brunswick. Family networks had a significant

positive influence on self-rated health in the “deprived” neighbourhood, but not in the more affluent

city.

2.3 Beliefs and Expectations

Health beliefs as operationalized by self-rated health are powerful predictors. The prior

ongoing debate about whether self-reporting of health status is generally reliable (Bound, 1991), has

been resolved, for the most part. Although, self-reported health ratings are subjective, non objective

measures, it is generally accepted that they are reliable measures of health (Arber & Ginn, 1991). As

several authors note: “Subjective measures are important to consider, because self-reports of poor


26

health have been found to be associated with early mortality (Mossey & Shapiro, 1982; Welin et al.,

1985) and with institutionalization, after controlling for other health and age variables” (Shapiro &

Tate, 1988, p.110).

Other types of beliefs might affect residential outcomes, such as attitudes toward

intergenerational living. The data clearly show that visible minorities, First Nations peoples and

immigrants live in larger households than their counterparts (Thomas and Wister, 1984; Boyd, 1991;

Kamo and Zhou, 1994; Gee, 1999; Kritz, Gurak and Chen, 2000). For example, Burr and Mutchler

found that attitudes regarding co-residence were significant predictors of subsequent living

arrangements and that differences exist across ethno-racial groups (Burr & Mutchler, 1999). Johnson

and Wolinsky investigated the effects of kin and non-kin support across race and marital status as a

predictor of community and home based care (Johnson & Wolinsky, 1996). Non-kin support was

significantly associated with the use of home care for both white and black females and moderately

significant for white males2. However, kin support was significantly associated with reductions in

home care for black females only. This study demonstrates that there is evidence of racial differences

in use of home care, especially between white and black females but whether this applies to

institutionalization is unknown. Consideration of ethno-cultural norms for co-residence and social

support could be a valuable cue to appropriate planning of future moves. This is no small matter since

43 percent of visible minorities settle in Toronto compared to 18 percent in Vancouver and 12 percent

in Montreal (Chappell, McDonald and Stones, 2007).

2 Black males were not included in the analyses, as the numbers were too few for meaningful comparison.


27

Another predictor examined by Akamigo and Wolinsky (2006), is expectations of nursing

home placement as a predictor of subsequent admission. The researchers, using longitudinal data from

Assets and Health Dynamics among the Oldest of Old (AHEAD), modeled expectations for nursing

home placement over 5 years, as well as actual placement for 6,242 Caucasian and African American

adults over 70 years of age. Those who estimated the probability of their future placement in a nursing

home to be within the range of 11-50 percent were significantly more likely to be institutionalised by

the five-year follow-up interview. Surprisingly, the effects were not significant for those who indicated

expectations in the higher range of 51-100 percent. Although the researchers offer no explanation for

this finding, it may be that many of those who indicated expectations of greater than 50 percent odds of

placement died during the five-year interval of the study. Another important aspect of this study was

that the associations between other key predictors, for example, ADL limitations, cognitive function,

and self-rated health, and expectations of subsequent placement were significant. The authors conclude

that not only are expectations for nursing home placement rational, to the extent that they are based on

a reasonable self-assessment, they also are reliable predictors of subsequent placement. These findings

confirm the value of client-centred care planning that acknowledges the role of beliefs and self-

assessment in predicting residential outcomes.

2.4 Health Behaviours: Smoking, Alcohol Consumption and Physical Activity

Surprisingly little research has been conducted on the effect of various health behaviours (e.g.

smoking, alcohol consumption and level of physical activity) on residential outcomes and the research

that does exist, is contradictory. For example, a large-scale longitudinal study of older Australians

found that a significant predictor of lower odds of institutionalization (45% lower risk) was the

consumption of alcohol (any level versus none), while cigarette smoking was non-significant


28

(McCallum et al., 2005). However, other researchers have found the reverse to be true. For example,

smoking had a significant effect on physical decline but alcohol consumption yielded no significant

effect in either direction in a longitudinal analysis of predictors of declines in physical functioning

(Sachs-Ericsson, Schatschneider, & Blazer, 2006). Health behaviours were not modeled in the

Gaugler et al. recent meta-analysis of the literature on nursing home predictors (Gaugler et al., 2007).

Consequently, it would seem that the effects of health behaviours may not be independently associated

with the risk of nursing home placement and are mediated through other health variables such as self-

reported health, disease profile, and functional status.

3. ENABLING CHARACTERISTICS

Enabling characteristics are those contextual factors (household and community “assets and

liabilities”) that influence residential outcomes. For example, this review considers such factors as

household income and wealth, home ownership, housing conditions, home modifications and

technological supports, as well as community, market and policy resources.

3.1 Familial/Household Resources

Income, wealth, and education are variables used to capture the effect of socio-economic status

(SES) on residential outcomes. Income is the most commonly measured variable because it is readily

available but has proven to have inconsistent effects. Wealth, a more difficult measure, is less

frequently used and often is measured as non-housing assets; while education is a relatively commonly

measured variable investigated in the literature. The Miller and Wiessert review (2000) found that,

overall, the research findings did not support that income, wealth or education had significant effects

on the risk of institutionalization. However, other household resources such as owning a home and


29

having structural, technological or human supports in the home decrease the likelihood of movers to

more restrictive settings.

3.1.1 Income

Many individual studies have found lower income to be a significant predictor of institutional

versus community living. Opoku and colleagues (2006), using data from the 1999 National Long-term

Care Study, found that having an annual income of less than $20,000 significantly increased the

likelihood of residing in institutional care. Other studies have found similar effects for income using

American census data (Burr & Mutchler, 2007; Mutchler & Burr, 2003). Russell et al., (1997) also

found lower income levels to be a significant predictor of nursing home admission (Russell, Cutrona &

Wallace, 1997). However, Hays et al., (2003) found that greater household income was a predictor of

household expansion but not of institutionalization (Hays; Pieper & Purser, 2003). Likewise, a cross-

sectional and longitudinal study of 1,425 elderly patients from Leicestershire, England, examined the

risk of institutionalization in relationship to financial status and found no significant effect for income

(Hancock et al., 2002).

Measuring the effects of a decline in income rather than the absolute level of income,

Finlayson (2002), used longitudinal data from the Aging in Manitoba study and found that a decline in

income was one of two variables (the other change in service use) which significantly predicted both

home and nursing home care. The author commented on the consistency of poverty as a predictor even

in the context of universal healthcare and provincial subsidies for long-term care.

However, Kersting (2001), in a study focusing on the impact of poverty on nursing home

admissions, found that problems of measurement led to results that were difficult to interpret. The

researcher contends that the contradictory results about the relationship between income and

institutionalization suggest that the association may not be linear. Low-income persons may have


30

higher institutional rates due to poorer health, and high-income persons may have higher rates due to

the greater accessibility that their wealth may provide, leaving the middle-income persons with lower

rates than either group. Further, Kersting (2001) suggests that a narrow measurement of current income

or wealth cannot capture the cumulative effects of poverty and that a more comprehensive inventory is

required, at least in the United States. Measures of income, education, owning a home and having

primary or complementary health insurance may better convey the relationship between poverty and

institutionalization.

Inconsistencies in the findings for income have been attributed to a number of factors including

inadequate weighting for number of persons within the household, sensitivity around disclosure

leading to under reporting and the exclusion of other financial resources that contribute to overall

wealth. Cox (2005) notes that it is not just disclosure and measurement issues that constrain the effect

of income but that as a dollar measure, it varies across context, living arrangement and need for

supports.

In the same way that perceived quality of health is often a more powerful measure than an

inventory of chronic conditions, Sachs-Ericsson and colleagues (2006) suggest an alternative measure

of SES that relies on a subjective assessment of adequacy of income rather than any single or bundled

quantifiable measure. The authors found that perceived problems with meeting basic needs

significantly influenced the rate of decline in physical functioning over time after controlling for other

measures of SES and for health-related variables. In a previous study (cited in Sachs-Ericsson et al.,

2006) using the same longitudinal data set, perception of meeting basic needs predicted mortality.

Consequently, this variable is likely to be associated with residential outcomes and future research

would need to investigate this relationship further. The researchers suggest that interventions targeted

at alleviating resource deficits may have a positive effect on health and housing outcomes. The


31

interventions may take the form of specific support for basic needs such as adequate affordable

housing, utility subsidies, and emergency funds.

Yet another measurement of income found in the literature is the gap between individual

income levels and neighbourhood SES. Deeg and Thomése (2005) suggest that it is not the absolute

value of income that is the determinant of health and well-being but rather the level of income relative

to the community of reference (neighbourhood SES). Using a large sample, the researchers created

three variables for comparison: “discrepant low” for individuals with low incomes in a neighbourhood

with an overall high SES, “discrepant high” for individuals with a high income in a low SES

neighbourhood and a reference group of matched high income in a high SES neighbourhood. Almost

all of the health indicators were worse (physical functioning and cognitive capacity, as well as greater

loneliness) for the “discrepant low” participants with the exception of depressive symptoms and self-

rated health. This finding paints a more complex picture of the effects of “deprived neighbourhoods”

and their influence on poor health since neighbourhood variables were included with no significant

effect and indications of poor health were found in neighbourhoods with high SES.

3.1.2 Wealth

Akyan (2002), using longitudinal data from the AHEAD study, found a significant effect for

non-housing wealth providing a protective effect on moves to more restrictive settings but only for

women. Lo Sasso and Johnson (2002) found no effect for wealth on institutionalization for either

women or men using the same longitudinal data and the same measurement of wealth but with a

different sampling strategy that excluded older adults without surviving children. These mixed results,

along with the non-significant effects reported in Miller and Weissert’s (2000) review of the literature,

suggest that wealth may have differential effects for older adults with or without surviving children.

Older adults with children may be reluctant to use accumulated wealth to secure home and community-


32

based services (HCBS) because of a desire to maintain a legacy. Another factor confounding

evaluations of the impact of wealth on residential outcomes is the phenomenon of having to “spend

out” assets to qualify for nursing home subsidies.

3.1.3 Education

Several studies have confirmed the findings of a recent review (Miller & Weissert, 2000) and

meta-analysis (Gaugler et al., 2007) of the literature (Lo Sasso & Johnson, 2002; McCallum et al.,

2005) that there is limited evidence of the predictive value of educational attainment on residential

outcomes.

However, Burr and Mutchler’s (2007) longitudinal study did find that having fewer years of

education was a significant predictor of nursing home living but only for unmarried women. In one of

the few studies that investigated both entries, Laditka (1998) found that women with more education

had greater odds of nursing home entries but that they also had greater odds of discharge. Berthelot and

colleagues, using data from the 1996/97 National Population Health Survey, found a very significant

relationship between having no formal education and residence in a long-term care facility for both

men and women (Berthelot, Martel, Legare, Trottier, & Houle, 2000). This relationship between “no

education” and institutionalization was particularly strong for those individuals residing in an

institutional setting but reporting no disability. Further research is necessary to clarify the mechanisms

involved.

3.1.4 Homeownership

Miller and Weissert (2000) concluded that among the studies they reviewed that the majority

(53 percent or 8 out of 15 studies) found that “not owning a home” was associated with greater risk of

institutionalization. A similar positive association for “not owning a home” and institutionalization


33

emerged from Gaugler et al., (2007) meta-analysis of the literature. Greene and colleagues (1993) also

established that homeowners had a much lower probability of institutionalization. The researchers

suggest that homeownership may affect length of community tenure in two ways: as a proxy for

accumulated wealth and as an indication of social attachment to place.

In contrast, the large-scale Australian longitudinal sample found no effect for homeownership

(McCallum et al., 2005). However, this study had a long follow-up interval of 14 years where baseline

measures of homeownership did not reflect residential mobility over time. Kersting (2001), who also

did not find an effect for homeownership, acknowledged the difficulty of using a cross-sectional

measurement of this predictor. Further, research to clarify the magnitude of the effect is necessary, as

well as to determine whether it is a proxy for place attachment or for financial assets (or both).

3.1.5 Housing Conditions: Quality, Accessibility Accommodations, Assistive Technology and Supportive/assisted Residential Models

(Heumann, 2004) contends that factors associated with built environments are a neglected

feature of modeling of risk and of geriatric assessments, yet the impact on well-being and

independence is widely acknowledged. The researcher notes that there is an overwhelming emphasis

on the role of human service systems rather than on built features in supporting the independence of

older adults. Heumann (2004) notes that while service providers will hire a homemaker to take laundry

to a laundromat at $18 dollars per hour (costing a program approximately $2000 per year), there is

little consideration of installing a washer and dryer for less than $1000 that could be used for several

years and reused by new clients. Not only are options such as these cost effective, but they are less

disruptive to the privacy and autonomy of older adults. Findings from a survey of service providers

suggest that, as a group, they are substantially under resourced vis à vis needs associated with the built


34

environment. The majority of service providers reported limited or no access to resources for repair

and adaptation of older adults’ housing (Heumann, 2004).

Wentzel and colleagues, using longitudinal data from the Canadian Study of Health and Aging

(CSHA-1), found that the quality of an older adult’s physical environment was a significant predictor

of time until institutionalization (Wentzel, Rose, & Rockwood, 2001). The data measured the quality

of the built environment according to the degree of disrepair, cleanliness and tidiness to determine

whether the living environment was “ideally” or “less than ideally” maintained. Regrettably, the study

did not investigate accessibility modifications whether structural or technological.

Structural accommodations such as ramps, grab bars, steps or doors into bathing units are

critical to independence, sense of personal control and to personal care yet few studies have examined

their effect on extending community tenure. Assistive technologies such as smart appliances, social

alarms and alerts and motion detectors can significantly enhance mobility, as well as safety and

autonomy for older adults. Holland and Peace (2001) define assistive technology as any item,

equipment, or system that is used to increase or maintain the functional capabilities of an individual.

The researchers note that ethical questions regarding privacy impose limitations with respect to

production of so-called “smart home” technologies that could be construed as intrusive. It would

appear that the same considerations limit the research investigating their effect on health, well-being,

and independence.

Tabbarah and colleagues examined the relationship between various health and socio-

demographic factors and access to home modifications and found that African Americans and

Hispanics had significantly fewer home modifications (Tabbarah, Silverstein, & Seeman, 2000). They

also note that having a higher income, dwelling in an apartment, and living alone were all associated

with a greater likelihood of modifications. Self-rated health and vision, having diabetes, having a


35

stroke, having a fall, breaking a hip, replacing a joint, and having greater ADL disability were also

associated with having home modifications.

Many of the predictors identified in this study (Tabbarah et al., 2000) such as ADL limitations,

living alone and stroke are strongly associated with risk of moves to more restrictive settings. Although

the overall levels of modifications are low, structural accommodations to prolong independence may

have a significant positive impact on extending independent living. Further, this study indicates that

having a “non-white” identity and lower income are associated with less access to modifications of the

built environment, which suggests that ethno-cultural barriers are exacerbated by the prohibitive costs

associated with modifications. Interventions to address these barriers in the form of ethno-culturally

sensitive outreach may extend access. In addition, given the longitudinal data used in this study

(AHEAD), future research should explicitly examine the associations between households with

modifications and residential outcomes.

One of the few studies to systematically investigate the effect of assistive technologies on the

health, well-being and living arrangements of older adults was a two-year random control trial

conducted by Tomita (2007). The intervention group received an enhanced package of smart home

technologies including a computer, Activehome software, lighting system and remote chime for

security/medication, as well as ongoing support from a geriatric nurse specialist trained in information

technologies.

The findings from this study indicated that a significantly greater proportion of the intervention

group was living independently at year two. Other statistically significant effects for the intervention

group were better scores for cognitive function and IADL, as well as positive self-reports associating

the use of smart technologies with a greater sense of agency and control over health and well-being. A


36

limitation to these results was that equivalence tests between groups at baseline showed a slightly

higher mean age for the control group, as well as some differences in disease profiles across groups.

Another aspect of housing conditions investigated in the literature is that of housing type; for

example, private versus publicly funded housing and age segregated versus general population

housing. Freedman (1996) found a significantly higher risk of nursing home admission for older adults

residing in public low income and private age segregated housing as compared to other community–

based living arrangements, even after controlling for age and income. Although there was no

discussion of the implications of this finding, it would appear that the multiple conditions of

disadvantage associated with residing in public housing may enhance the risk of institutionalization

and that the greater monitoring associated with age-segregated housing may facilitate earlier detection

of risk and subsequent placement in settings with higher levels of support.

Although some researchers have examined endorsement by seniors of less restrictive residential

options such as assisted living, supportive or second suite housing, they have not examined whether

endorsement translates into actual moves or how these different housing options affect community

tenure. For example, Weeks and colleagues (2005) found that overwhelmingly, people preferred to

“age in place.” Only half of the study’s sample would consider sheltered, seniors, or congregate

housing. The level of endorsement for these options was considerably greater for those older adults

who indicated no or low family support; again suggesting the pivotal role informal support plays in

decision making regarding residential moves.

Formal supports associated with these alternative residential models, whether they are linked or

delinked (provided by satellite agencies) to the housing, can be critical determinants to sustaining older

adults in the least restrictive settings possible. Although a number of options exist on the continuum

from least restrictive (independent housing) to most restrictive (institutional living) residential settings,


37

the research indicates that awareness is quite low (Diez Roux, 2004; Gibler, Lumpkin, & Moschis,

1997). Gibler and colleagues established that there were a number of barriers to awareness about

residential options including the lack of accessible and up- to-date information and the inconsistent

labelling of the types of living arrangements.

A large sample survey (N= 4800) found that the majority of older adults reported limited

awareness of options between independent living and long-term care facilities; many, in fact, viewed

nursing homes as the only retirement housing option (Diez Roux, 2004). The researcher notes that the

lack of standardization of terminology and poor communication of the differences and benefits

associated with various residential options severely constrains decision making for older adults and

their families. If nursing home care is seen as the only option for extended support, many older adults

will move prematurely to inappropriately restrictive settings.

In a qualitative study, Grando and colleague’s (2002) highlighted that the lack of knowledge of

residential options was a critical factor in why older adults with light care needs reside in nursing

homes. Lack of appropriate information to evaluate risk and quality of care in different settings was

also raised in an analysis of older adults residing in nursing homes who met the criteria for care in less

restrictive settings (Spector, Reschovsky, & Cohen, 1996). The researchers recommended that more

comprehensive and accessible information on residential options is critical to securing appropriate

levels of support in the least restrictive settings possible. Cox (2005) comments that fragmented health

policy planning and lack of integrated information resources leave older adults and their families

struggling to negotiate layers of services delivered by different systems.

Not only is information on residential options limited but housing providers often lack any

coordinated access to supports available through government health ministries and have to resort to

independently seeking resources to support tenants aging in place (Mollica, 2003). Although care


38

planning is influenced by an older adult’s living arrangements, the importance of housing is often

peripheral to discussions of healthy aging.

There appears to be very few studies available that use longitudinal, random control trials or

even quasi-experimental methodologies to examine the effect of supportive/assistive housing on

extending community tenure for older adults. In the grey literature, a recent large-scale American

demonstration project: HOPE for Elderly Independence Demonstration Program (HOPE IV) was

evaluated (Westat, 1996). The intervention was designed to help low-income, “frail,” elderly persons

maintain the highest possible quality of life in the least restrictive environment. The evaluation

employed a quasi-experimental longitudinal design to investigate the impact of a comprehensive care

package for community-dwelling “frail” older adults eligible for government subsidies. Positive

outcomes, across multiple domains of functioning were evident. For example, the intervention group

scored significantly better on major mental health dimensions (anxiety, depression,

behavioural/emotional control, and psychological well-being), social functioning (quantity and quality

of social activities), vitality (energy level and fatigue), and other measures of social well-being.

However, there were no significant differences in nursing home placements between intervention and

control groups. The authors explained that this is common to a number of intervention evaluations,

which have shown significant improvements in social functioning and overall health and well-being

but not in longer-term outcomes such as institutionalization.

A recent Canadian study (Lum, Ruff & Williams, 2005) investigated health and housing

outcomes for older adults living in supportive and social housing. The findings from the report

challenged conventional assumptions about thresholds for institutional care and the peripheral status

assigned to community supports outside of the traditional health sector. Almost all the older adults in

the study met the criteria for placement in a long-term care facility yet with minimal supports like


39

housekeeping, grocery shopping and for some, supports for personal care, they were able to continue to

live in the community. Further, for those living in housing with onsite support, the use of costly

emergency services was reduced, leading the authors to conclude that community supports for aging in

place are not an “add on” to an already overburdened healthcare system but rather a cost effective

alternative to acute and institutional care. The authors estimated that regional average annual costs to

the Ministry of Health and Long-term Care for supportive housing services in Toronto were $6,984

(Lum et al., 2005), while estimates of annual costs to the government for long-term care services were

in the range of $26,000 per year (Ontario Association for Non-profit Homes and Services for Seniors,

2007).

Another study examining the balance of care and residential outcomes estimated the cost of

HCBS as compared to LTC facility placement for a thirteen-week period and found that about half of

those on the LTC waiting list in Waterloo could be effectively and cost effectively cared for with a

HCBS package (Williams, Paul, Devitt-Wilson & Kuluski, 2007). The authors contend that a full three

quarters of those on the LTC wait list could be accommodated if more support options were available

(e.g. affordable supportive housing). Coyte et al., (2002), in the recommendations arising from their

forecasts for long-term care in Ontario, suggest that the province should provide support and incentives

for modifications to enable long-term care providers to create seniors’ congregate housing, such as

assisted living/supportive housing.

Other available research demonstrates that service-enriched housing promotes resident

satisfaction, successfully provides service to frail populations, and supports aging in place (Pynoos,

Liebig, Alley & Nishita, 2004). Pynoos and colleagues suggest that low-intensity programs involving

only service coordination can support aging in place, while higher intensity programs for the more

severely impaired may extend the option of community living for older adults with higher needs.


40

However, the authors found that high-needs older adults in the community were slightly less physically

and cognitively impaired than residents of nursing facilities indicating that institutional care will still

be required for some older adults.

3.1.6 Type and extent of healthcare insurance

The major challenge with evaluating the effects of holding various types of health insurance is

the difficulties in comparisons across jurisdictions. Australia, United States, Western Europe, and

Canada have different healthcare systems and consequently different uptake of private healthcare

insurance. McCallum and colleagues using Australian and Johnson and Wolinsky using American

longitudinal data both found that holding private health insurance did not have a significant effect on

institutionalization (Johnson & Wolinsky, 1996; McCallum et al., 2005). However, Opuko and

colleagues found that having no health insurance whatsoever, increased the odds of residing in a more

restrictive setting 22.4 times more than the risk for those with some sort of health insurance - a risk

greater than any other variable measured - including age, functional or cognitive capacity ( Opoku et

al., 2006).

Miller and Weissert’s (2000) review of the literature found that only one of three studies

examining the effect of private insurance found a significant positive association with

institutionalization. The authors found a similar inconsistent effect for Medicaid subscription with four

of eleven studies finding a positive association with risk for nursing home placement and the

remaining seven finding non-significant effects. Unfortunately, no Canadian study was found that

investigated the effect of having supplementary (private) health insurance on residential outcomes.


41

3.2 Community, Market and Policy Resources

The literature on the predictive value of community characteristics on residential outcomes is

scant. There is an emerging literature on such factors as housing market conditions, neighbourhood

SES (as opposed to measures of individual or household SES), health policy, and balance of care

expenditures, as well as availability of HCBS and institutional long-term care.

3.2.1 Housing Market Conditions

Mutchler and Burr (2003) examined various factors associated with the housing market to

determine the effect of this macro level influence on institutionalization. The authors found that lower

vacancy rates and higher median rents were significantly associated with higher rates of

institutionalization for both unmarried women and men suggesting that scarcity of affordable housing

may force many older adults into lower cost institutional living arrangements. A lower percentage of

small residential units (e.g. bachelor apartments, second suites or garden flats) was significantly

associated with institutionalization for unmarried women. The researchers suggest that housing policy

which encourages the development (or retention) of affordable, smaller unit rentals for older adults

may support longer periods of independent living.

3.2.2 Neighbourhood Characteristics and Place Attachment

Another measure of macro predictors of residential outcomes is that of neighbourhood

characteristics. Motiwala et al. (2007) analysis of predictors of place of death in a population cohort of

Ontario seniors found that seniors living in areas with higher social deprivation were more likely to die

in long-term care facilities than either in the hospital or at home (Motiwala, Croxford, Guerriere, &

Coyte, 2007).


42

Deeg and Thomése (2005) who investigated the gap between neighbourhood SES and income

as a predictor of declines in health provided an alternate view of neighbourhood influences, and, as

mentioned above, this study found that older adults with low incomes in neighbourhoods with high

SES had significantly poorer health. Consequently, it was not so much the neighbourhood

characteristics that affected health and well-being but the gap between individual levels of income and

the overall SES of the neighbourhood. Bartlett and Peele (in Andrews and Phillips, 2005) conclude that

while there are indications that neighbourhood or community-level factors may exert an effect on older

adults ability to age in place, there is very limited research in this area. Studies are required to

investigate issues such as availability and access to supports and that attempt to isolate features of

liveable communities that foster healthy aging and independent living.

Another factor that may influence long-term residential outcomes is residential stability, which

is a proxy for place attachment to either housing and/ or neighbourhood. Using the longitudinal

AHEAD data, Aykan (2002) found that greater residential stability (residing in the same housing for

10 or more years) significantly predicted lower risk of institutionalization for men but not women. The

researcher suggests that residential stability may be a proxy for non-kin informal support but does not

comment on the gendered effect of the findings for this variable. Future research might examine place

attachment and its role in residential outcomes, as well as it meaning across gender and ethno cultural

identity.

In addition to such factors as neighbourhood SES and place attachment, neighbourhood

environments can be accessible and inclusive or be challenging and exclusionary. There is a

considerable body of literature investigating what constitutes an “age-friendly” or “liveable”

community. The World Health Organization has initiated the “Global Age-Friendly Cities Project” in

multiple sites (including 3 Canadian locations) to document the major physical and social barriers to


43

active aging which will be compiled and form the basis for a blueprint for an “Age-friendly City” to be

used by cities around the world (World Health Organization, 2007). Although there is a body of

literature on the indicators of age-friendly environments, there appears to be no research yet that

investigates how these indicators may affect the residential outcomes of older adults.

3.2.3 Balance of Care: Availability of and Expenditure on Home and Community-Based Services and Institutional Long-term Care

The literature on the effects of policy on residential outcomes for older adults tends to focus on

the availability of nursing home and hospital beds, of home and community-based services (HCBS),

and the relative balance of healthcare expenditures on HCBS versus institutional long-term care.

Aykan (2002) investigated the impact of differences in Medicare policy across five states on

the risk of nursing home placement and found no significant effect for any of the health policy

variables. The author offered several reasons for these findings that highlight the difficulties with

assessing policy effects on long-term care outcomes (Aykan, 2002). For example, that payment

protocols cause hospitals to discharge post-acute patients to nursing homes for convalescence therefore

flattening the effect that policies such as HCBS waivers3 have on nursing home admissions. The author

also raises the issue of direction of causality suggesting that odds of nursing home placement may

drive policy rather than be a result of specific policy interventions. Ladkita (1998) found mixed results

for the influence of health policy while evaluating the effect of better health on lifetime nursing home

use. The author found that women who reside in states with more generous nursing home

reimbursement rates (government subsidy to long-term care facilities) were at significantly higher risk

3 States waive certain requirements of the Social Security Act to create home and community–based supports to delay

institutionalization.


44

of institutionalization. However, Miller and Weissert’s (2000) review of the literature found a higher

risk of institutionalization for those residing in states with lower rates of reimbursement but a

significantly greater risk of institutionalization for those older adults residing in states with higher

Medicaid enrolment.

Several authors (c.f. Akyan, 2002; Burr et al., 2005; Hoerger, Picone & Sloane, 1996) have

examined the impact of the number of available nursing home beds as a determinant of odds of

institutional living. In the United States, many states have capped the development of new nursing

home beds and/or limited subsidies for their development. Akyan (2002) used the longitudinal

AHEAD data to determine the influence of these state policies on residential outcomes. The author

found no significant effect for the risk of institutionalization to those residing in states where the

number of beds per capita or the rates of reimbursement were higher (Akyan, 2002). However, the

author acknowledged that state policies might indeed be effective in influencing supply and demand

but that this does not necessarily translate into a measurable effect on the risk of nursing home

placement. Similarly, another investigation of state commitment to home and community-based care

found no significant effect for level of investment in HCBS (2002) and rates of occupancy of nursing

home beds (Burr et al., 2005).

However, other studies did find significantly greater odds of nursing home living for older

adults living in states with greater per capita rates of nursing home beds and with higher levels of

reimbursement (Hoerger et al., 1996). These mixed findings are mirrored in the results of Miller and

Weissert’s review (2000) where four out of seven studies found that the risk of institutionalization rose

with the number of nursing home beds per capita while the remaining studies found no effects.

Not only is the supply of nursing home care a potential predictor of residential outcomes but so

is the availability of HCBS, which exerts a positive influence on the likelihood of residing in less


45

restrictive settings. One study found that older adult’s perception of their ability to age in place was

significantly affected by their proximity to health and social services (Sherman & Combs, 1997).

Another study, using longitudinal data from the National Survey of Families and Households, found

that older adults living in counties with a greater density of community-based geriatric services were

significantly more likely to live independently than in a nursing home (Burr & Mutchler, 2007).

Models of integrated care targeted to enabling older adults with complex needs to remain in the

community as long as possible include the Program of All-Inclusive Care for the Elderly (PACE) and

in Canada, the Comprehensive Home Option of Integrated Care for the Elderly (CHOICE). These

programs provide care through a single portal (one organization which may contract out certain

services) typically a Day Centre with a multidisciplinary team who assess and support clients. A recent

analysis of the records for 4,646 participants aged 55 years or older who were enrolled in PACE

programs during the period from June 1, 1990, to June 30, 1998 found that the cumulative risk of

nursing home admission for PACE respondents was less than 15 percent (Friedman et al., 2005). This

level of risk was evaluated as low considering that a hundred percent of the enrolees were certifiable for

nursing home care. Despite this promising finding, the study did not include a control group and the

risk of admission was only compared to national statistics.

Other models of care emphasize coordination across organizations such as the

Program of Research to Integrate the Services for the Maintenance of Autonomy (PRISMA).

Preliminary results from a pilot of PRISMA (Hébert, Durand, Dubuc & PRISMA Group) indicated

that the control group had a greater risk of institutionalization but that it only just approached significance

(p=0.06).

Greene and colleagues linked the type of support with different need profiles to identify what

combinations of community-based long-term care and need were associated with delays to time of


46

nursing home admission (Greene et al., 1993). The researchers isolated particular types of

interventions that maximized community tenure. For example, they found statistically significant

reductions in nursing home admissions for the following combinations of care and need: nursing

services provided to those using a wheelchair, home-health assistance provided to those who have

cognitive impairment and personal care and housekeeping supports provided to those who experience

functional impairment. The authors contend that these findings indicate that different outcomes are

associated with specific care packages targeted to particular groups. The implications emerging from

these findings are that expansion of HCBS without targeting individual client profiles is neither

effective nor cost effective. More research that focuses on determining the relationship between client

characteristics, HCBS, and institutionalization is critical to achieving the best possible fit and risk

reduction.

Another measure of the effect of health policy on moves to more restrictive settings is the

balance of spending on home and community-based services (HCBS) versus institutional care. Burr

and colleagues examined the effect of the relative spending on HCBS versus long-term care (LTC) in

institutional settings (Burr et al., 2005). The researchers found that in states with a higher proportion of

spending on HCBS, older women were less likely to reside in institutional settings.

Given that HCBS do not fall under the Canadian Health Act, provinces and territories have

developed quite different home care and community-based healthcare spending patterns and programs.

A Canadian study tried to determine whether home care services were an effective substitution for

nursing home care in British Columbia (Chappell, Havens, Hollander, Miller, & McWilliam, 2004).

The researchers compared a matched group of community dwelling older adults to those residing in an

institutional setting and concluded that home and community care was as effective as and more cost

effective than levels of care and costs associated with institutional long-term care.


47

Homecare expenditures represent about four percent of Canadian health expenditures, but are

reaching a relatively high proportion of seniors relative to other jurisdictions (MacAdam, 2000).

MacAdam (2000) reviewed home care programs and spending across several jurisdictions and

concluded that, overall, Germany, Japan, the US, and the U.K. are moving toward a similar goal of

increasing access and options for community-based care and reducing reliance on institutionally based

long-term care. However, the relative shift in spending toward non-institutional healthcare is slow.

Spending on community-based health services is more difficult to track, because it is not specifically

measured in most health reports or tracked in healthcare databases. That said, Guyatt, Yalnizyan, and

Devereaux comment that a government trend toward greater support of community-based care is

implicit in the reduced proportion of health dollars committed to hospital care (Guyatt, Yalnizyan, &

Devereaux, 2002). Better data collection on community-based healthcare usage, outcomes, and

spending is critical to understanding its role in maintaining older adults in the least restrictive settings

possible.

Concerns over relative spending allocations and their implications for decision making have

been raised in studies evaluating health-planning models. For example, Coyte and colleagues argue

that while the balance of funding for long-term care remains weighted toward institutional settings,

there is considerable evidence that consumer preferences do not mirror the allocation of long-term care

(LTC) dollars (Coyte et al., 2002). The researchers found that when preferences for home care were

considered the estimates for the number of nursing home beds required between 1996 and 2018 were

significantly lower: down from 103.4 percent o as low as 46.2 percent and estimates for the demand

for home care were significantly greater: rising from 57.4% to as much as 103.6%. Further, the authors

stress that over expansion of LTC bed capacity influences practices and behaviours, which may result

in greater acceptance of nursing home placement rather than less restrictive options. The implications


48

are less serious in Ontario where the authors reported lower bed ratios per 1000 population for those 75

years and older and a more balanced mix of institutional and in-home care than other jurisdictions in

Canada.

One of the more dynamic policy decisions impacting aging in place is the 1999 Olmstead

decision in the United States. Essentially, the decision mandated State governments and service

providers to accommodate individuals with support needs in the least restrictive setting possible. A few

early evaluations of the impact of the Olmstead decision have emerged (O'Hara & Day, 2001;

Zubritsky, Mullahy, Allen, & Alfano, 2006). One study evaluated the impact of the Olmstead decision

five years after its inception and found that 73 percentof adults surveyed reported that the Olmstead

planning had not had any measurable impact on reducing barriers to community-based service/support

acquisition (Zubritsky et al., 2007). However, when other stakeholders such as service providers and

bureaucrats were included in the analysis, there were moderate improvements in the areas of

community-based services, communication, and coordination across jurisdictions, and access to

affordable housing.

The lack of measurable effects of the Olmstead decision on residential outcomes may be due to

a number of factors. The decision does not specifically mandate greater state funding for HCBS nor

does it set out timelines or accountability frameworks to evaluate the impact of Olmstead Plan. For

example, O’Hara and Day examined state planning and found that Olmstead plans generally do not

address the issue of developing supportive housing strategies (O’Hara & Day, 2001). Further, the

researchers reported that states required assistance in bringing key housing decision makers into

discussions (O’Hara & Day, 2001). Notwithstanding the limited evidence on the impact of the

Olmstead decision, a similar protection is championed by the Ontario Human Rights Commission as

an important mechanism for measuring and monitoring the rate of unnecessary or “undue”


49

institutionalization of persons in order to safeguard the principle of integration over segregation

(Ontario Human Rights Commission, 2007).

4. NEED CHARACTERISTICS

Need factors in Anderson’s behaviour model are more proximate factors that affect the use of

health services and the subsequent outcomes. These factors include indicators of self-perceived and

practitioner–evaluated health. Although, these indicators dominate the research literature, they have

not been studied extensively in the evaluation of the risk for institutionalization. The factors that have

garnered most attention include dementia, cognitive impairment, depression, incontinence, functional

impairment and self-rated health and functional status. Most studies use Cox proportional hazard

models or bivariate logistic regressions to estimate the risk of institutional placement. Several of the

studies were longitudinal and used random sampling or random assignment.

In a fourteen year longitudinal study in Australia of 2805 community dwelling seniors from the

Dubbo study of the elderly, there were 244 nursing home placements (8 percent), comprising of 95

men and 149 women (John McCallum et al., 2005). Forty-four percent of the placements were due to a

primary diagnosis of dementia while for another twenty percent, dementia was a secondary diagnosis

(McCallum et al., 2005). Stroke (sixteen percent) and coronary heart disease followed (fourteen

percent) the other significant predictors of placement. The conclusion was that dementia and disability-

related care burdens rather than other health conditions are the dominant causes of nursing home

placement, in Australia. Other risk factors such as incontinence, impaired respiratory function and

depression were also identified as predictors of placement but these factors were seen to be potentially

amenable to intervention. While the longitudinal aspects of the study are its strength, there is the

problem that baseline measures could have different impacts on nursing home placements at different


50

intervals such as at 5, 10 or 15 years. Similarly, the risk factors included at baseline represented the

knowledge in 1988 and lastly, they could have changed over time.

In a 28-month longitudinal study of 881 of elderly residents in public housing in Baltimore,

(Black et al., 1999) suggest that the two strongest predictors of placement were functional status

(impaired instrumental activities of daily living) followed by cognitive impairments (Black, Rabins, &

German, 1999). In this housing context, the findings made sense because elderly residents of public

housing live alone and have higher rates of cognitive disorders. These individual are more likely to

become unable to live independently (shopping, finances, telephone assistance) before activities of

daily living or gait become impaired. The study however, was not random and relied on the self-report

of people with impaired cognition.

Research by Harris and Cooper (2006) used three datasets from the Health Outcomes Survey,

which randomly sampled 137,000 Medicare and Choice enrolees, The Nursing Home Minimum Data

Set and the Medicare Enrolment Database. The participants were Medicare beneficiaries aged 65 years

and older still living in the community and the measurements included self-reported functional status,

chronic health conditions, demographics and several mood-related questions. After controlling for age,

race, sex, marital status, home ownership, functional status, and co-morbid conditions, people who

identified themselves as feeling sad or depressed most of the time over the previous year were at a

significantly higher risk for nursing home placement (Harris & Cooper, 2006). The researchers also

found that there may be a link between depression and social support, but suggested further study.

In a similar study, using Manitoba data (N=1751) from the Canadian Study of Health and

Aging, depressive symptoms also predicted admission to a long-term care facility (St. John &

Montgomery, 2006). The researchers found that 19.4% of those with depressive symptoms were

institutionalized versus 10.9% of those with no symptoms. However, when functional impairment was


51

considered, this association was no longer apparent, leading the researchers to conclude that other

factors might be operating. The functional impairment is either a confounding variable or a mediating

factor for depressive symptoms (St John & Montgomery, 2006, p. 1795). The researchers also

observed that this finding should serve as a cautionary note for the interpretation of the Harris and

Cooper study that used a crude measure of depression and did not account for confounding variables

(Harris and Cooper, 2006).

The search for an empirically-based tool that predicts the prognosis for either death or nursing

home placement for patients with Alzheimer’s and related dementias was the focus of a National

Institute of Health study (Stern et al., 1997). This prospective cohort study of 236 patients who were

followed up semi-annually for up to seven years had a validation cohort of 105 patients. An algorithm

of predictors for outcomes was developed that included gender, age, Mini-Mental State Examination

(MMSE) scores, age of onset, existence of psychosis, and presence of an accelerated disease course.

The findings indicated that the prediction algorithms were promising for prognosis. The relative risk of

reaching the nursing home endpoint was significantly increased for those who had extrapyramidal

signs of psychotic symptoms measured on the Columbia University Scale for Psychopathology in

Alzheimer’s Disease (CUSPAD) at their initial visit and in patients with younger ages at disease onset.

In a related secondary analysis of 341 randomly assigned Alzheimer’s patients who were enrolled in a

clinical drug trial and followed at three-month intervals for up to two years, researchers found that

nursing home placement closely reflected dementia progression (Knopman et al., 1999). The change

scores on four dementia severity measures from baseline to last measurement (Clinical Dementia

Rating, changes in two of three basic activities of daily living, changes in dependence levels and

changes in the Blessed Dementia Rating Scale) were used to quantify change in dementia severity. All

four severity measures were strongly associated with nursing home placement.


52

In a longitudinal examination of behaviour in dementia of a 100 community dwelling

participants living at home with a carer, a nested case-control study was utilized to investigate the

predictors of institutionalization (Hope et al., 1998). The characteristics which best predicted

institutionalization one year later were excessive night-time activity; immobility or difficulty in

walking; incontinence; being away from a carer for more than 16 hours a week; and being cared for by

a female. Aggressive behaviour was not associated with an increased chance of entry into an institution

one year later, although it was more prevalent four months prior to entering an institution.

Using phase one and two of the longitudinal Canadian Study of Health and Aging, Gutman et

al., (2001) examined the relationship between two self-report health measures and mortality and

institutionalization. The respondents were asked “How would you say your health is these days?”

(HEALTH) and” How much do your health troubles stand in the way of your doing the things you

want to do?” (TROUBLE). Because the two measures address different concepts – general health and

level of functioning of self-reported health, it was found that HEALTH and TROUBLE cannot act as

proxies for each other. Both variables predict institutionalization, but independently of each other.

However, a composite variable of the two factors improved prediction of both institutionalization and

mortality.

Incontinence has been identified as a significant factor in nursing home admissions. A number

of earlier studies have shown that a large proportion of nursing home residents are incontinent of urine

at the time of their admission to nursing home and urinary incontinence is pivotal in family decisions

to move the older adult to an institution (Ouslander, 1990; Ouslander, Kane and Abass, 1982).

Selecting urinary incontinence as a primary criterion for inclusion in their study, Coward et al. (1995),

examined the institutionalization rates of 719, rural versus urban older adults using the Longitudinal

Study on Aging, 1984-1990, (Coward, Horn and Peek, 1995). They indicated that residents of less


53

urbanized counties in the United States were more likely to have a nursing home admission than were

older adults in any other residential context. The hypothesis that rural older persons were less likely to

have access to community based interventions for urinary incontinence which led to their admissions,

was strongly supported, even when controlling for socio-demographic factors, health status and extent

of social network. While other factors might explain the rural respondent’s institutionalization, like the

lack of a geriatrician, or a scarcity of alternative housing, the data raise questions about nursing homes

being over or under utilized, depending geographical location.

Researchers Friedman et al., (2005) in their secondary analysis of 4,646 adults, 55 years of age

and over within the Program of All-inclusive Care for the Elderly (PACE) examined the overall

predictors of nursing home admission. Among the community-dwelling seniors, incontinence of the

bowel, IADL status and age predicted institutionalization. Unlike other studies, poor cognitive status,

number of chronic conditions, activity of daily living deficits, urinary incontinence, several

behavioural disturbances, and duration of program operation did not predict admission. The reasons

may be that there is a selection bias into the PACE program, the program is very effective or the high

level of frailty of the enrolees and their risk might already be so high, distinctions cannot be made. It

also may be that as people become more disabled, they are more affected by their informal support

network. Namely, there is the possibility that in a frail population at high risk, IADLs have a greater

impact on nursing home admission than ADLs (Greene and Onrich, 1990).

Mayo et al. (2005) in their search for improved predictors of institutionalization included

functional status indicators in their indices (Mayo et al., 2005). Data was collected in 1997-1998 from

approximately 100 physicians, and 6,465 of their patients, they examined their limitations in doing

housework, stair climbing and so on. These functional status indicators independently predicted

institutionalization, while controlling for socio-demographics factors and co-morbidity.


54

A Spanish prospective study of 353 patients on a geriatric ward attempts tried to predict, among

the elderly, death while in hospital, prolonged hospital stay, and transfer to an institution upon

discharge for elderly patients (Alarcon, Barcena, Gonzalez-Montalvo, Penalosa, & Salgado, 1999).

Factors that were predictive of institutionalization at discharge were scores on the Red Cross Hospital

Functional Disability Scale, a measure of functional ability.

Strain et al.’s ( 2003) previously mentioned 5-year longitudinal study using 123 cognitively

impaired seniors and their caregivers from the Manitoba Study of Health and Aging resulted in the

identification of risk profiles for institutionalization – high, medium and low. In the screening stage,

cognitive status was assessed using the Modified Mini-Mental State Examination. Those with a score

of 78 or lower were asked to participate in the study. The group at highest risk of institutionalization

were the most cognitively impaired, required the most assistance with IADLs/ADLs and were the most

likely to exhibit behavioural problems. Both the high and medium group tended to wake up at night for

no apparent reasons and slept during the day. Their findings indicated that 75% of the most cognitively

impaired were most likely to be institutionalized 5 years later. While the study had a small sample size,

did not collect information at the time of institutionalization or measured the services used prior to

institutionalization, the predictive value of the risk profiles was established. Strain et al. (2003) note

that the findings suggest the need to target assessments and services differently to the various risk

groups.

5. HEALTH CARE UTILIZATION

Another cluster of predictors, embedded in the Anderson model, that have been shown to affect

residential outcomes is measures of health care utilization. The most common measures utilized in the

research are measures of prior nursing home use, rates of hospitalization , use of paid helpers (home


55

care support), and doctors and medications. By far the most robust findings are for prior nursing home

admissions.

5.1 Prior Nursing Home Admission and Hospitalization

Akyan (2002) found a very strong effect for having been in a nursing home in the year prior to

the baseline interviews for the AHEAD study. The odds of entering into a nursing home almost

quadrupled for women and tripled for men who were previously institutionalized. A similar outcome

for prior nursing home admission was found in a study where a significantly greater rate of

institutional placement was associated with previous institutionalization (Russell et al., 1997). Miller

and Weissert’s (2000) review of the literature found that the majority of studies demonstrated a

significantly greater risk of institutional living for those with a history of nursing home admission and

hospitalization. Likewise, a meta-analysis found a robust effect for prior nursing home placement: a

3.47 times greater likelihood of subsequent admission and a 1.19 times odds ratio of nursing home

placement following hospitalization (Gaugler et al., 2007).

Although prior nursing home and hospital use were strong predictors of nursing home

placement, Akamigbo and Wolinsky (2006) found that neither of the variables were associated with

reported expectations of nursing home placement at baseline. The researchers suggest that

understandings of nursing home placement may be ambiguous, especially as to whether the placement

is short term or permanent. Future research could examine placement time intervals and exits to more

clearly understand what predicts duration of institutionalization.


56

5.2 Other Measures of Use: Paid Helpers, Doctors and Medications

Miller and Weissert’s (2000) review found ambiguous results for the effects of paid helpers on

institutionalization. In thirty-two studies, seventeen found significant positive relationships, four

significant negative effects and eleven studies were non-significant. Gaugler and colleagues’ meta-

analysis found a moderate increase in the likelihood of nursing home placement associated with the

use of formal services (Gaugler et al., 2007). Interestingly, in this meta-analysis, the pooled odds of

institutionalization for formal supports were equal to those for informal support. Finlayson’s analysis

of the Aging in Manitoba longitudinal data found that changes in service use predicted moves to more

restrictive settings and discriminated between expansions of home care and moves to nursing home

care (Finlayson, 2002).

Other studies investigating the role of formal healthcare supports have looked at number of

doctor’s visits and the number of prescription medications as a proxy for disease and chronic

conditions and as a measure of contact with healthcare providers. The number of doctors’ visits was a

significant predictor of nursing home admission in Russell and colleagues study of the role of

loneliness in predicting community versus institutional tenure (Russell et al., 1997). The researchers

also found, that as the number of prescriptions increased, so did the odds of nursing home placement.

In Gaugler and colleagues’ meta-analysis of the literature, the number of medications had a small but

significant positive association with institutional living (Gaugler et al., 2007). Although only two

studies examining medication use were included in Miller and Weissert’s (2000) review, both found a

significant positive association with the probability of nursing home placement. The challenge to

understanding the effects of service use on residential outcomes is that it is difficult to disentangle the

potential protective effect of reducing health declines from the negative risks associated with greater

monitoring of health status.


57

In this vein a, a study of Adult Protective Service (APS) use and nursing home placement in the

United States found that an APS referral for self-neglect was the highest risk factor for placement in a

nursing home, net demographic, medical, functional and social factors associated with nursing home

placement (Lachs et al., 2002). The sample was 2,812 community-dwelling older adults who were 65

years of age or older in the1982 cohort in the New Haven Established Populations for Epidemiologic

Studies in the Elderly. A subset of this cohort (N=202) was referred to protective services over a nine

year follow-up period from cohort inception. Nursing home placement over that time period was also

determined. The rates for nursing home placement was 69.2 percent for self-neglecting respondents,

52.3 percent for mistreated respondents and 31.8 percent for respondents who had no contact with

APS. Proportional hazard models were estimated controlling for the usual factors such as dementia,

urinary incontinence, functional disability and poor social networks. Surprisingly, the most potent

predictor of institutionalization was APS referral for self-neglect followed by APS referral for

mistreatment (Lachs et al., 2002). This finding replicates an earlier and equally rigorous study by

Blenkner (1971) over thirty years ago. While Canada has a different legal framework for neglect and

elder abuse, the outcomes are likely to be much the same since service providers still have to find

alternative housing or community team care in an environment with shrinking resources. The findings

underscore the need for alternative housing and for community team-based care such as that offered by

PACE.

6. CONCLUSION

6.1 Limitations of the Literature

Although this review utilizes Andersen’s model of health service utilization, other models

highlight additional dimensions such as knowledge of services, and still others acknowledge the role of


58

accessibility and acceptability. This review, where possible, has included research in the area of

knowledge of resources and of endorsements/expectations around different residential outcomes but is

nonetheless, limited by the boundaries of the model.

There are many methodological problems with the extant research. Hays et al., (2003)

demonstrated that the single decision of moving into a nursing home is really a two-stage decision-

making process: the decision whether to make a change and if so, what change should be made. For

example, the researchers contend that adult children increase the probability of change to parental

households (early detection of need for greater care). However, once the household is destabilized, the

presence of adult children (options for co-residency) reduces the odds of institutionalization. The

researchers also note that while marital status reduces the probability of change, once the change is

underway married older adults are at greater risk of institutionalization. The latter issue could be due to

either overwhelming caregiver burden or to the greater challenge of making alternative living

arrangements for two persons as compared to one. The researchers recommend that when examining

predictors of institutionalization it is critical to look at factors that led up to a change in living

arrangements, independent from factors that predict a particular type of change.

Hays et al., (2003) also found that different factors exerted distinct effects on various

residential outcomes. For example, smaller household size and impaired social interaction did not

significantly predict risk of household expansion but did predict greater odds of institutionalization.

Finlayson (2002) found similar varied effects which challenge the assumption that moves to

progressively more restrictive settings are associated with increasing or decreasing values of a

predictor. For example, Finlayson (2002) found that a decline in social support was a significant

predictor of home care but not of nursing home care. Studies such as these suggest that much of the

literature may be oversimplifying the complexity of factors that predict residential outcomes and speak


59

to the need for professional judgement. The implications for the development of appropriate

assessments are that a single set of predictors will not always effectively discriminate between

appropriate residential outcomes.

Another limiting factor is that some predictors are more easily quantified than others.

Therefore, they appear more frequently in the literature and limit the understanding of the effects of

variables that are more difficult to measure. Miller and Weissert’s (2000) review of the literature

concludes with a notation of the serious skew toward measuring easily quantifiable factors. The

authors, drawing on the Andersen framework, note that enabling and predisposing factors are

examined far less frequently than need-based factors, particularly those that are scale-based measures

such as ADL and cognitive assessments using the SPMSQ and MMSE. Consequently, the authors

recommend that future studies attend to the more neglected community-wide factors such as housing,

market and policy resources that may be important explanatory mechanisms for determining residential

outcomes.

At the same time, a difficulty with the assessment of market, community, and policy

environments is that a multitude of policies intersects in the provision of community and institutional

long-term care. Further, as Diez Roux (2004) notes, the influence of contextual variables is often

apparent primarily through their effect on micro level factors such as health or socio-demographic

indicators. The author suggests that this hinders understanding of the direct pathways of influence for

macro level factors but that quantitative RCTs and/or qualitative methodologies may help clarify the

interaction of micro and macro level factors in determining residential outcomes. This researcher

concludes that despite the challenges of examining the effects of macro factors, a more crucial question

is how much evidence is required before health policy planning includes interventions aimed at

housing, community, and policy environments, as well as interventions targeted to individual factors.


60

Other limitations in the research include the lack of attention to duration of nursing home stays,

the lack of investigation of policy interventions to reduce over and under care, and the neglect of

quality of life measures. Few studies measure the length of stay in and/or exits from nursing homes.

The research indicates that nursing homes are being used for rehabilitation purposes to offset costs for

hospital care which, as Kavanagh (1993) proposes, skew the data on nursing home placement.

Although studies have documented the extent of over and under care, the literature tends to focus on

reducing errors in either direction via improved assessment tools but neglects macro policy options

(e.g. careful planning, monitoring and development of new nursing home beds and expanded home

care programs). A flaw in the research highlighted by Miller and Weissert (2000) is that quality of life

outcomes such as care recipient and caregiver satisfaction, or the degree of self-determination are

rarely evaluated in the research. What is more, much of the focus on the consequences of providing

personal care has been on the impact of caregiving on the caregiver, rather than the predictors of

institutionalization as related to caregiving.

Further, the research seldom questions the impact of residential outcomes on the health and

well-being of older adults and misses an important opportunity to understand what settings convey

benefits to particular groups. In one of the few studies that posed this type of research question, Marek

et al., (2005), compared clinical outcomes of older adults in a community-based long-term care

program to matched older adults in institutional-based long-term care. Using a quasi-experimental

methodology, with a matched intervention and control, they compared various measures of health and

well-being at 6-month intervals over a 30-month period. The older adults in the community-based

intervention program, called Aging in Place (AIP), had significantly better mental health, functional,

and cognitive status, as well as better bladder control than the nursing home group. In absence of more

research about the consequences of one residential setting over another and for what groups, the


61

literature will only be able to tell us that a predictor is associated with a particular outcome but not why

this is the case. Further, without understanding the pathways by which a predictor influences an

outcome, it will be very difficult to construct interventions to support older adults in the least

restrictive settings possible.

From a methodological perspective, many of the studies are subject to problems. The prediction

of change, based on single cross-sectional baseline measures of selected factors is one of the more

serious issues. The disadvantage with this type of analyses is that the factors can be unstable and are

known to evolve over time. Few of the studies untangle the many steps in the decision-making process,

information is not collected at time of institutionalization or about service usage prior to placement and

the effect of different time intervals on placement are ignored. In addition, a number of the studies do

not have random assignment or random sampling and different measures of the predictors are used,

making cross study comparisons impossible. As a result, the function of a number of predictors of

institutionalization remains inconclusive.

6.2 Summary: Predisposing Characteristics

The importance of social support in determining institutionalization is modest at best and its

effects may mediated by health predictors. Nevertheless, Beland and colleagues suggest that though

social support interventions are not typically employed in health and social service settings, they are, in

fact, amenable to policy and program interventions (Beland et al., 2004). Further, Russell and

colleagues note that while social support interventions may delay moves to more restrictive settings, it

may be better to remove barriers and enhance existing relationships rather than try to foster new

relationships (Russell et al., 1997).


62

It would also appear that social support has different effects for different groups of older

adults. For example, those residing in poor neighbourhoods may be more reliant on family networks

(Beland et al., 2004) or those who are at the highest risk of institutionalization due to isolation may

experience the greatest benefits from interventions aimed at extending their community tenure

(Eloniemi-Sulkava et al., 2005).

Finally, the inconsistent and often contradictory results associated with informal

supports/caregiving highlight the need for more careful, fine-grained analysis of the mechanisms by

which caregiving impacts residential outcomes. Although there is significant public and private interest

in “aging in place,” as Wiles (in Andrews and Phillips, 2003) cautions, it should not be presumed to be

an ideal goal without examining what must be in place and what is transformed in the process of

achieving this goal. As homes increasingly become sites for the provision and consumption of care, it

is critical that health planners recognize and accommodate the fact that these resources are typically

provided by women.

6.3 Summary: Enabling Characteristics

Of the enabling characteristics, the most prevalent predictors of residential outcomes are

household or individual income and the supply of nursing home beds available. The results for income

are mixed and appear to be better understood through more subjective measures, such as perceived

adequacy of income, rather than absolute measures of dollar value or income range as determined by

various poverty measures. Homeownership has a moderate protective effect on moves to more

restrictive settings but it is unclear whether the pathway of the effect is through place attachment to

housing and/or neighbourhood or through its function as a proxy for wealth. Although there is an


63

emergent literature (especially grey literature) and a great deal of “best practice” documentation of the

role of various human and structural supports (supportive housing, home care, home modifications and

assistive technologies) in sustaining older adults in less restrictive settings, research directly evaluating

the effect of various housing characteristics on residential outcomes for older adults is limited.

While expenditures on the balance of healthcare determine the availability of community- and

home-based care versus institutional care, currently the only clear predictor of this balance and its

effect on residential outcomes is the supply of nursing home beds (Burr et al., 2005; Hoerger et al.,

1996; Miller and Weissert, 2000). As Coyte and colleagues contend, greater spending on institutional

care and consequently greater supply does translate into greater risk of nursing home placement (Coyte

et al, 2002).

6.4 Summary: Need Characteristics

The needs of older persons are complex and depend on the circumstances of the older adult.

For example, urinary incontinence may be a risk factor in rural areas and not in urban areas where

there is likely to be more preventive services. IADLs or instrumental functioning may be more

important for those living alone in the community in public housing than those living with a caregiver

where activities of daily living are more important (ADLs). Using Miller and Weissert’s (2000) article

as a guide, there are only a few strong predictors among the “need” characteristics. The most striking

finding is the importance of cognitive function in predicting institutionalization. In 25 of the 33 articles

reviewed, lower cognitive function was a predictor of institutionalization. Nevertheless, only in one-

half of the ten articles was dementia/Alzheimer’s Disease a predictor of institutionalization. Lower

ADL and IADL were both strong predictors of institutionalization but did depend on the context in

which the older person lived. Digestive system diseases, proved to have significant associations with


64

institutionalization. The research on co-morbidity is inconclusive. Although there is general acceptance

that self-reported health ratings are reliable measures of health, replication of studies of this factor are

required. All of the studies reviewed suffered from a number of methodological issues but it is

important to note that there was a consistency in the findings.

6.5 Summary: Use Characteristics

The most significant “use” predictor of institutional residence is prior admission to a nursing

home (Akyan, 2002; Akamigbo & Wolinsky, 2006; Miller and Weissert, 2000). However, it is unclear

whether the explanatory effect is due to greater endorsement or acceptance by an older adult because

of a previous placement or a bias of service providers/discharge planners to move someone to an

institutional setting because it has been done successfully in the past. Formal help, sometimes referred

to as “paid helpers,” has been widely evaluated with inconsistent results (Gaugler et al., 2007; Miller

and Weissert, 2000) which may be due to the fact, that the measure covers such a broad spectrum of

support (e.g. personal support workers to meal delivery services).

Overall, this review affirms that that older adults with lower levels of informal support (as

measured by absence of a spouse or other adults in the household and having fewer adult children), and

less economic resources (as measured by an absence of housing assets and lower income) are at

increased risk of moves into more restrictive settings. The presence of cognitive impairment or limits

to ADLs and IADLS, as measured by scale thresholds, is significantly associated with living in more

restrictive settings. Other factors that are consistently and significantly associated with higher rates of

institutionalization are prior admissions to a nursing home or hospital and residing in areas with a

greater supply of nursing home beds.


65

7. REFERENCES

Aditya, B. S., Sharma, J. C., Allen, S. C., & Vassallo, M. (2003). Predictors of a nursing home

placement from a non-acute geriatric hospital. Clinical Rehabilitation, 17(1), 108-113.

Aguero-Torres, H., von Strauss, E., Viitanen, M., Winblad, B., & Fratiglioni, L. (2001).

Institutionalization in the elderly: The role of chronic diseases and dementia. Cross-sectional

and longitudinal data from a population-based study. Journal of Clinical Epidemiology, 54(8),

795-801.

Akamigbo, A. B., & Wolinsky, F. D. (2006). Reported expectations for nursing home placement

among older adults and their role as risk factors for nursing home admissions. Gerontologist,

46(4), 464-473.

Alarcon, T., Barcena, A., Gonzalez-Montalvo, J. I., Penalosa, C., & Salgado, A. (1999). Factors

predictive of outcome on admission to an acute geriatric ward. Age and Ageing, 28(5), 429-432.

Andersen, R. M. (1995). Revisiting the behavioural model and access to medical care: Does it matter?

Journal of Health and Social Behaviour, 36(1), 1-10.

Andrews, G. J., & Phillips, D. R. (2005). Ageing and place: Perspectives, policy, practice. London:

Routledge.

Arber, S., & Ginn, J. (1991). Gender and later life: A sociological analysis of resources and

constraints. London: Sage.

Argimon, J. M., Limon, E., Vila, J., & Cabezas, C. (2005). Health-related quality-of-life of caregivers

as a predictor of nursing-home placement of patients with dementia. Alzheimer Disease and

Associated Disorders, 19(1), 41-44.


66

Aykan, H. (2002). Do state Medicaid policies affect the risk of nursing home entry among the elderly?

- Evidence from the AHEAD study.

Beland, F., Bergman, H., Johri, M., Kone, A., Wolfson, C., & Zunzunegui, M. V. (2004). Social

networks and self-rated health in two French-speaking Canadian community dwelling

populations over 65. Social Science and Medicine, 58(10), 2069-2081.

Beland, F., Bergman, H., Lebel, P., Clarfield, A. M., Tousignant, P., Contandriopoulos, A. P., et al.

(2006). A system of integrated care for older persons with disabilities in Canada: Results from

a randomized controlled trial. Journal of Gerontology, 61 (4), 367-373.

Berthelot, J.-M., Martel, L., Legare, J., Trottier, H., & Houle, C. J. (2000). Living at home or in an

institution: What makes the difference for seniors? [1996/97 data]. Health Reports, 11(4), 49.

Black, B. S., Rabins, P. V., & German, P. S. (1999). Predictors of nursing home placement among

elderly public housing residents. The Gerontologist, 39(5), 559-568.

Bound, J. (1991). Self-reported versus objective measures of health in retirement models. Journal of

Human Resources, 26(1), 106±138.

Boyd, M. (1991). Immigration and living arrnagements:Elderly women in Canada. International

Migration Review, 25, 4-7.

Burr, J. A., & Mutchler, J. E. (1999). Race and ethnic variation in norms of filial responsibility among

older persons. Journal of Marriage and the Family, 61(3), 674-687.

Burr, J. A., & Mutchler, J. E. (2007). Residential independence among older persons: Community and

individual factors. Population Research and Policy Review, 26(1), 85-101.

Burr, J. A., Mutchler, J. E., & Warren, J. P. (2005). State commitment to home and community-based

services: Effects on independent living for older unmarried women. Journal of Aging and

Social Policy, 17(1), 1-18.


67

Campbell, S. E., Seymour, D. G., & Primrose, W. R. (2004). Systematic literature review of factors

affecting outcome in older medical patients admitted to hospital. Age and Ageing, 33(2), 110-

115.

Canadian Research Network for Care in the Community. (2006). In focus: Social Sciences and

Humanities Research Council of Canada and Ryerson University.

Care Planning Partners. (1999). The care survey. Toronto: Care Planning Partners.

Chappell, N., McDonald, P.L. & Stones, M. (2007). Aging in Comtemporary Canada 2nd Edition.

Toronto: Pearson, prentice Hall.

Chappell, N.L., Havens, B., Hollander, M.J., Miller, J., & McWilliam, C. (2004). Comparative Costs of Home Care and Residential Care. The Gerontologist, 44(3), 389-400.

Coward, R. T., Horne, C., & Peek, C. W. (1995). Predicting nursing home admissions among

incontinent older adults: A comparison of residential differences across six years. The

Gerontologist, 35(6), 732-743.

Cox, C. B. (2005). Community care for an aging society: Issues, policies, and services. New York:

Springer Pub. Co.

Coyte, P. C., Laporte, A., Baranek, P., & Croson, W. S. (2002). Forecasting facility and in-home long-

term care for the elderly in Ontario: The impact of improving health and changing preferences.

. Retrieved February 14, 2006. from http://www.news.utoronto.ca/bin3/020606a.asp

Dang, T. T., Antolin, P., & Oxley, H. (2001). Fiscal implications of ageing: Projections of age-related

spending: OECD.

Davey, A., Femia, E. E., Zarit, S. H., Shea, D. G., Sundstrom, G., Berg, S., et al. (2005). Life on the

edge: Patterns of formal and informal help to older adults in the United States and Sweden.

Journals of Gerontology Series B-Psychological Sciences and Social Sciences, 60(5), S281-

S288.


68

Deeg, D. J. H., & Thomase, G. C. F. j. (2005). Discrepancies between personal income and

neighbourhood status: Effects on physical and mental health. European Journal of Ageing,

2(2), 98-108.

Diez Roux, A. (2004). Estimating neighbourhood health effects: The challenges of causal inference in

a complex world. Social science & medicine, 58, 1953-1960.

Division of Aging and Seniors. (2006). Healthy aging in Canada: A new vision, a vital investment,

from evidence to action - a background paper. Retrieved July 18, 2007. from http://www.phac-

aspc.gc.ca/seniors-aines/pubs/haging_newvision/pdf/vision-rpt_e.pdf.

Eloniemi-Sulkava, U., Notkola, I. L., Hentinen, M., L., K. S., Sivenius, J., & Sulkava, R. j. (2001).

Effects of supporting community-living demented patients and their caregivers: A randomized

trial. Journal of the American Geriatrics Society, 49(10), 1282-1287.

Finlayson, M. (2002). Changes predicting long-term care use among the oldest-old. The Gerontologist,

42(4), 443.

Freedman, V. A. (1996). Family structure and the risk of nursing home admission. Journals of

Gerontology Series B-Psychological Sciences and Social Sciences, 51(2), S61-S69.

Friedman, S. M., Steinwachs, D. M., & Rathouz, P. J. (2005). Characteristics predicting nursing home

admission in the program of all-inclusive care for elderly people. The Gerontologist, 45(2),

157-166.

Gaugler, J., Duval, S., Anderson, K., & Kane, R. J. (2007). Predicting nursing home admission in the

U.S.: A meta-analysis. BMC Geriatrics, 7(13).

Gee, E. M (1999). Ethnic identity among foreign-born Chinese Canadian elders. Canadian Journal on

Aging. 18(4) 415-429.


69

Gelberg, L., Andersen, R. M., & Leake, D. (2000). The behavioural model for vulnerable populations:

Applications to medical care use and outcomes for homeless people. Health Services Research

34(6), 1273-1302.

Gibler, K. M., Lumpkin, J. R., & Moschis, G. P. (1997). Mature consumer awareness and attitudes

toward retirement housing and long-term care alternatives. Journal of Consumer Affairs, 31(1),

113-138.

Grando, V. T., Mehr, D., Popejoy, L., Maas, M., Rantz, M., Wipke-Tevis, D. D., et al. (2002). Why

older adults with light care needs enter and remain in nursing homes. Journal of gerontological

nursing, 28(7), 47-53.

Greene, V. L., Lovely, M. E., & Ondrich, J. I. (1993). Do community-based, long-term-care services

reduce nursing home use? A transition probability analysis. The Journal of Human Resources,

28(2), 297.

Gutman, G. M., Stark, A., Donald, A., & Beattie, B. L. (2001). Contributions of self-reported health

ratings to predicting frailty, institutionalization, and death over a 5-year period. International

Psychogeriatrics, 13(Suppl1), 223-231.

Hancock, R., Arthur, A., Jagger, C., & Matthews, R. (2002). Effect of older people's economic

resources on care home entry under the United Kingdom’s long-term care financing system.

Journals of Gerontology: Series B: Psychological Sciences and Social Sciences, 57B (5), S285-

S293.

Harris, Y., & Cooper, J. K. (2006). Depressive symptoms in older people predict nursing home

admission. Journal of the American Geriatrics Society, 54(4), 593-597.

Hays, J., Pieper, C., & Purser, J. (2003). Competing risk of household expansion or institutionalization

in late life. The Journals of Gerontology, 58B (1), S11.


70

Hebert, R., Dubois, M. F., Wolfson, C., Chambers, L., & Cohen, C. (2001). Factors associated with

long-term institutionalization of older people with dementia: Data from the Canadian study of

health and aging. Journals of Gerontology Series A-Biological Sciences and Medical Sciences,

56(11), M693-M699.

Hebert, R., Durand, P. J., Dubuc, N., Tourigny, A., & The PRISMA Group. (2003). PRISMA: A new

model of integrated service delivery for the frail older people in Canada. International Journal of

Integrated Care, 3(18).

Heumann, L. (2004). Assisted Living for Lower Income and Frail Older Persons from the Housing and

Built Environment Perspective. Journal of Housing for the Elderly, 18(3/7), 165-177.

Hoerger, T. J., Picone, G. A., & Sloan, F. A. j. (1996). Public subsidies, private provision of care and

living arrangements of the elderly. Review of Economics and Statistics, 78(3), 428-440.

Holland, C., & Peace, S. M. j. (2001). Inclusive housing in an ageing society: Innovative approaches.

Bristol: Policy Press.

Hope, T., Keene, J., Gedling, K., Fairburn, C. G., & Jacoby, R. (1998). Predictors of

institutionalization for people with dementia living at home with a carer. International journal

of geriatric psychiatry, 13(10), 682-690.

Johnson, R. J., & Wolinsky, F. D. (1996). Use of community-based long-term care services by older

adults. Journal of Aging and Health, 8(4), 512.

Kamo, Y. & Zhou. M. (1994). Living arrangements of elderly Chinese and Japanese in the United

States. Journal of Marriage and the Family, 56, 544-558.

Kersting, R. C. (2001). Impact of social support, diversity, and poverty on nursing home utilization in a

nationally representative sample of older Americans. Social Work in Health Care, 33(2), 67-87.

Kritz, M., Gurak, D. & Chen, L. (2000). Elderly immigrants:Their composition and living

arrangements. Journal of Sociology and Social Welfare, 27(1) 85-114.

Lachs, Mark S., Williams, C. S. O'Brien, S. & Pillemer, K. (2002). Adult Protection Service Use and

nursing home placement. The Gerontologist, 42(6) 734-739.


71

Laditka, S. B. (1998). Modeling lifetime nursing home use under assumptions of better health.

Journals of Gerontology Series B-Psychological Sciences and Social Sciences, 53(4), S177-

S187.

Lo Sasso, A. T., & Johnson, R. W. (2002). Does informal care from adult children reduce nursing

home admissions for the elderly? Inquiry-the Journal of Health Care Organization Provision

and Financing, 39(3), 279-297.

Lord, S. R. (1994). Predictors of nursing home placement and mortality of residents in intermediate

care. Age and Ageing, 23(6), 499-504.

Lubben, J., Blozik, E., Gillmann, G., Iliffe, S., Renteln, v., Kruse, W., et al. (2006). Performance of an

abbreviated version of the Lubben social network scale among three European community-

dwelling older adult populations. The Gerontologist, 46(4), 503-513.

Lum, J., Ruff, S., & Williams, A. P. (2005). When home is community: Community support services

and the well-being of seniors in supportive and social housing.

Mayo, N. E., Nadeau, L., Levesque, L., Miller, S., Poissant, L., & Tamblyn, R. (2005). Does the

addition of functional status indicators to case-mix adjustment indices improve prediction of

hospitalization, institutionalization, and death in the elderly? Medical care, 43(12), 1194-1202.

McCallum, J., Simons, L., Simons, J. & Friedlander, Y. (2005). Patterns and predictors of nursing

home placement over 14 years: Dubbo study of elderly Australians, Australasian Journal on

Ageing (Vol. 24, pp. 169): United Kingdom: Blackwell Publishing.

McCallum, J., Simons, L. A., Simons, J., & Friedlander, Y. (2005). Patterns and predictors of nursing

home placement over 14 years: Dubbo study of elderly Australians. Australasian Journal on

Ageing, 24(3), 169-173.


72

McDonald, L., Sussman, T. & Donahue, P. (2007). When bad things happen to good people: The

economic consequences of retiring to caregive. SEDAP, Paper #202, McMaster. University:

Hamilton.

McDonald, L. (2006). The Welfare of Women and Retirement. In Leroy Stone (Ed.). New Frontiers of

Research about Retirement and Other Later-Life Transitions. Ottawa: Statistics Canada,

Canada, Institute de la statistique du Québec.

Miller, E. A., & Weissert, W. G. (2000). Predicting elderly people's risk for nursing home placement,

hospitalization, functional impairment, and mortality: A synthesis. Medical Care Research and

Review, 57(3), 259-297.

Moen, P. (2001). The gendered life course. In R. H. Binstock & L. K. George (Eds.), Handbook of

aging and the social sciences (5th ed., pp. 179-196). San Diego: Academic Press.

Mollica, R. (2003). Coordinating services across the continuum of health, housing, and supportive

services. Journal of aging and health, 15(1), 165-188.

Motiwala, S. S., Croxford, R., Guerriere, D. N., & Coyte, P. C. (2007). Predictors of place of death for

seniors in Ontario: A population-based cohort analysis. Canadian Journal on Aging / La Revue

Canadienne du vieillissement, 25(4), 363-371.

Mutchler, J. B., & Burr, J. (2003). Living arrangements among older persons. Research on Aging,

25(6), 531-558.

National Advisory Council on Aging. (1999). Statistical snapshots. Retrieved July 18, 2007 from

http://www.hc-sc.gc.ca/seniors-aines/index.htm.

O’Hara, A., & Day, S. (2001). Olmstead and supportive housing: A vision for the future. New Jersey:

Center for Health Care Strategies, Inc.


73

Draft comprehensive and integral international convention on the protection and promotion of the

rights and dignity of persons with disabilities, (2007).

Opoku, A., Maudlin, T., Sweaney, A., Batchtel, D., Atiles, J., & Eaves, C. (2006). Predisposing,

enabling and need predictors of community versus institutional long-term care. Journal of

Housing for the Elderly, 20(1/2), 133-153.

Ouslander, J. G. (1990). Urinary incontinence in nursing homes. Journal of the American Geriatrics

Society, 38, 289-291.

Ouslander, J. G, Kane, R. L., & Abass, I. B. (1982). Urinary incontinence in elderly nurisng home

patients. Journal of the American Medical Association. 248, 1194-1198.

Public Health Agency of Canada. (2005). Integrated approach to chronic disease. Retrieved July 18,

2007. from http://www.phac-aspc.gc.ca/ccdpc-cpcmc.

Robson, W. (2001). Will the baby boomers bust the health budget? Toronto: C. D. Howe Institute.

Russell, D. W., Cutrona, C. E., & Wallace, R. B. (1997). Loneliness and nursing home admission

among rural older adults. Psychology and Aging, 12(4), 574-589.

Sachs-Ericsson, N., Schatschneider, C., & Blazer, D. G. (2006). Perception of unmet basic needs as a

predictor of physical functioning among community-dwelling older adults. Journal of aging

and health, 18(6), 852-868.

Sherman, S. L., & Combs, E. R. (1997). Characteristics related to elderly persons' perceived difficulty

of remaining in their current homes. Family and Consumer Sciences Research Journal, 26(1),

59-74.

Shields, M., & Martel, L. (2006). Healthy living among seniors. Retrieved July 18, 2007 from

Spector, W. D., Reschovsky, J. D., & Cohen, J. W. (1996). Appropriate placement of nursing-home

residents in lower levels of care. Milbank Quarterly, 74(1), 139-&.


74

St John, P. D., & Montgomery, P. R. (2006). Depressive symptoms in older people predict nursing

home admission. Journal of the American Geriatrics Society, 54(11), 1794-1795.

St John, P. D., Montgomery, P. R., Kristjansson, B., & McDowell, I. (2002). Cognitive scores, even

within the normal range, predict death and institutionalization. Age and Ageing, 31(5), 373-378.

Statistics Canada. (2003a). Cycle 16 of the general social survey: Caring for an aging society.

Retrieved August 18, 2007. from

Statistics Canada. (2003b). Finances in the golden years: 1999. Retrieved July 18, 2007 from

http://www.statcan.ca/Daily/English/031117/d031117b.htm.

Statistics Canada. (2005). Population projections for Canada, provinces and territories 2005-2031.

Retrieved July 18, 2007 from http://www.statcan.ca/english/freepub/91-520-

XIE/00105/bfront1.htm.

Statistics Canada. (2007a). Age and sex, 2006 counts for both sexes, for Canada, provinces and

territories Retrieved July 18, 2007. from

http://www12.statcan.ca/english/census06/data/highlights/agesex/pages/Page.cfm?Lang=E&Ge

o=PR&Code=01&Table=1&Data=Count&Sex=1&StartRec=1&Sort=2&Display=Page.

Statistics Canada. (2007b). Portrait of the Canadian population in 2006, by age and sex: Findings.

Retrieved July 18, 2007 from

http://www12.statcan.ca/english/census06/analysis/agesex/index.cfm.

Statistics Canada. (2007c). Residential care facilities 2004/2005. Retrieved July 18, 2007. from

Stern, Y., Tang, M.-X., Albert, M. S., Brandt, J., Jacobs, D. M., Bell, K., et al. (1997). Predicting time

to nursing home care and death in individuals with Alzheimer disease. JAMA: Journal of the

American Medical Association, 277(10), 806-812, (807p.).


75

Strain, L. A., Blandford, A. A., Mitchell, L. A., & Hawranik, P. G. (2003). Cognitively impaired older

adults: Risk profiles for institutionalization. International Psychogeriatrics, 15(4), 351-366.

Tabbarah, M., Silverstein, M., & Seeman, T. (2000). A health and demographic profile of non-

institutionalized older Americans residing in environments with home modifications. Journal

of aging and health, 12(2), 204-228.

Thomas, K. & Wister, A. (1984). Living arrangements of older women: The ethnic dimensions.

Journal of Marriage and the Family, 46, 301-311.

Tomita, M. (2007). Use of currently available smart home technology by frail elders: Process and

outcomes, Topics in geriatric rehabilitation (Vol. 23, pp. 24). Gaithersburg, Md.: Aspen

Systems Corp.

Toronto. (2007a). Backgrounder - release of 2006 census results - marital status, families, households

and dwelling characteristics. Retrieved from

Toronto. (2007b). Backgrounder - release of 2006 census results: Age and sex population counts.

Retrieved from

Toronto Central Local Health Integration Network. (2007). Integration priorities of the 2007-2010

integrated health services plan. Toronto: Toronto Central Local Health Integration Network.

Toronto Central Local Health Integration Network (LHIN). (2007). 2007-2010 integrated health

service plan: Appendix j seniors' services.

Weeks, L., Branton, O., & Nilsson, T. (2005). The influence of the family on the future housing

preferences of seniors in Canada. Housing, Care and Support, 8(2), 29.

Wentzel, C., Rose, H., & Rockwood, K. (2001). Measurement of the influence of the physical

environment on adverse health outcomes: Technical report from the Canadian study of health

and aging. International Psychogeriatrics, 13(S1), 215-221.


76

Westat, I. (1996). Evaluation of the hope for the elderly independence demonstration: Second interim

report, 130.

Wilcox, J., Jones, B., & Alldrick, D. (1995). Identifying the support needs of people with dementia and

older people with mental illness on a joint community team: A preliminary report. Journal of

Mental Health, 4(2), 157-163.

Williams, P., Paul, J., Devitt-Wilson, A., & Kuluski, K. (2007). Workshop Presentation to: Ontario

Community Support Association Annual Conference Great Ideas: Fresh Approaches to Care

in the Community. Toronto, October 16, 2007.

World Health Organization. (2007). Global Age-friendly Cities: A Guide. Retrieved on October 11,

2007 from

http://www.who.int/ageing/publications/Global_age_friendly_cities_Guide_English.pdf

Zubritsky, C., Mullahy, M., Allen, M., & Alfano, E. (2006). The state of the Olmstead decision and the

impact of consumer participation in planning. American Journal of Psychiatric Rehabilitation,

9(2), 131 - 143.


77

8. APPENDICES

8.1 Appendix A: TABLES

Table 1

Definitions of Accommodation and Care Settings

Accommodation

Government Funding

Services Client Groups

Boarding/Lodging Homes No room/board

minimal supervision socially disadvantaged people

Home for Special Care Provincial care/supervision adults with developmental handicaps, adults with mental illness

Supportive Housing Provincial care/supervision

seniors, adults with physical disabilities, or mental health problems, people living with HIV/AIDS

Retirement Homes No care/supervision predominantly seniors (average age 82)

Long-term Care Facilities Provincial care/supervision therapies

predominantly seniors (average age 85)

Emergency Hostels/ Women’s’ Shelter

Municipal Provincial

counselling, standards, care management

women & children in transition from abusive situations homeless people people on public assistance

Source: Government of Ontario, 2005


78

Table 2

Living Arrangements, Ontario, 1996, by Gender (65+, 65-74, 75-84, 85+)

*Special care homes include chronic care hospitals, long-term care settings, and residences for seniors that provide minimal assistance and supervision for seniors who are independent in most activities of daily living.

Table 3

Living arrangements of seniors aged 65 and over by sex and age group, 2001

Source: Statistics Canada, General Social Survey, Cycle 16, 2002

Age Gender In Private Households

Hospitals Special Care Homes

Religious Institutions

Total in Institutions

65+ Men 95.9% 0.5% 3.4% 0.1% 4.1%

65+ Women 91.8% 0.4% 7.5% 0.25% 8.2%

65+ Total 93.5% 0.5% 5.8% 0.1% 6.4%

85+ Men 76.9% 1.5% 21.4% 0.1% 23.1%

85+ Women 62.8% 1.5% 35.3% 0.5% 37.2%

85+ Total 66.8% 1.5% 31.4% 0.3% 33.2%


79

Table 4

Number of Equations (N) Reporting Positive Significant (+), Negative Significant (–), and Non-significant (NS) Associations among Predictors and

Adverse Outcomes


80


81


82

8.2 Appendix B: FIGURES

Figure 1

Andersen’s Behavioral Model

Source: Adaptation from Gelberg, Andersen, and Leake (2000).


83

Figure 2

Population Projections for Canada, Provinces, and Territories, 2005-2056

Source: Statistics Canada, 2005


84

Figure 3

Age of Toronto’s Senior Population


85

Figure 4

Percentage of Canadians in Good Health, by Age Group, Household Population, Aged 65 and

Over

(Shields & Martel, 2006)

Aging in Place: Literature Review for Bridgepoint and Toronto Central LHIN

Documents