FACTORS AFFECTING HOSPICE CARE USE
AMONG LONG-TERM CARE FACILITY RESIDENTS IN CANADA
by
Beibei Xiong
B.M., B.A., Jilin University, 2017

THESIS SUBMITTED IN PARTICIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER IN SCIENCE
IN
HEALTH SCIENCES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
April 2019
© Beibei Xiong, 2019

Abstract
Hospice care can improve quality of life for persons nearing end of life. Little is known about
hospice care practices in long-term care facilities (LTCFs) in Canada. This thesis included
185,715 residents in LTCFs in Canada in 2015 and followed their death records to 2016 to
examine the characteristics of residents who received hospice care and those who did not but
may have benefitted from it. Univariate, bivariate and multivariate analyses were used
depending on the variable type. Results show the actual use of hospice care in LTCFs is very
low in Canada (i.e. less than 3%). Residents who received hospice care had more severe and
complex clinical needs than those who did not. Findings suggest several possible barriers to
hospice use in the LTCF population including ageism, rurality, and disease diagnoses.
Immediate action is needed to provide improved access to, and utilization of, hospice care in
LTCFs in Canada.

ii

TABLE OF CONTENTS
ABSTRACT............................................................................................................................. II
TABLE OF CONTENTS ..................................................................................................... III
LIST OF TABLES .................................................................................................................. V
LIST OF FIGURES .............................................................................................................. VI
ACKNOWLEDGEMENT .................................................................................................. VII
CHAPTER ONE: INTRODUCTION .................................................................................... 1
Definition of Hospice Care ................................................................................................ 1
History of Hospice Care .................................................................................................... 3
Long-Term Care Facility (LTCF) in Canada ................................................................. 6
Hospice Care in LTCFs in Canada .................................................................................. 8
Problem Statement and Research Questions .................................................................. 9
Conceptual Framework ................................................................................................... 11
Operational Definition ..................................................................................................... 13
CHAPTER TWO: LITERATURE REVIEW..................................................................... 16
Literature Search ............................................................................................................. 16
Database selection. ...................................................................................................... 16
Search strategies. ......................................................................................................... 16
Literature selection and search outcome...................................................................... 17
Literature Analysis .......................................................................................................... 18
Characteristics of included studies. ............................................................................. 18
Prevalence of hospice use in LTCFs. .......................................................................... 19
Length of hospice use in LTCFs. ................................................................................. 20
Hospice use in LTCFs versus in non-LCTFs. ............................................................. 21
Hospice use among residents in LTCFs. ..................................................................... 22
Hospice use among end-stage residents in LTCFs. ..................................................... 26
Discussion ......................................................................................................................... 30
CHAPTER THREE: METHODS ........................................................................................ 32
Study Design ..................................................................................................................... 32
Data Sources ..................................................................................................................... 32
Sample. ........................................................................................................................ 32
Data collection. ............................................................................................................ 33
Measures ........................................................................................................................... 34
Predisposing characteristics. ........................................................................................ 34
Enabling characteristics. .............................................................................................. 34
Need characteristics. .................................................................................................... 34
Outcomes. .................................................................................................................... 44
Data Analysis .................................................................................................................... 44
Data cleaning. .............................................................................................................. 45
Univariate analyses. ..................................................................................................... 45
iii

Bivariate analyses. ....................................................................................................... 46
Multivariate analyse..................................................................................................... 46
Ethical Considerations .................................................................................................... 46
CHAPTER FOUR: HOSPICE USE AMONG RESIDENTS IN LONG-TERM CARE
FACILITIES IN CANADA .................................................................................................. 47
Introduction ...................................................................................................................... 48
Conceptual framework................................................................................................. 50
Methods ............................................................................................................................ 51
Study design and data sources. .................................................................................... 51
Measures. ..................................................................................................................... 51
Statistical analyses. ...................................................................................................... 53
Ethical considerations. ................................................................................................. 54
Results ............................................................................................................................... 54
Predisposing characteristics. ........................................................................................ 55
Enabling characteristics ............................................................................................... 55
Need characteristics. .................................................................................................... 55
Predictive models......................................................................................................... 56
Discussion ......................................................................................................................... 57
Strengths and limitations. ............................................................................................ 62
Conclusion ........................................................................................................................ 63
CHAPTER FIVE: AN EXAMINATION OF THE CHARACTERISTICS OF
RESIDENTS WHO RECEIVE VERSUS DO NOT RECEIVE HOSPICE CARE IN
LONG-TERM CARE FACILITIES IN CANADA ............................................................ 74
Introduction ...................................................................................................................... 76
Conceptual framework................................................................................................. 78
Methods ............................................................................................................................ 78
Study design and data sources. .................................................................................... 78
Measures. ..................................................................................................................... 79
Statistical analyses. ...................................................................................................... 81
Ethical considerations. ................................................................................................. 82
Results ............................................................................................................................... 82
Study sample characteristics. ....................................................................................... 82
Differences among four groups. .................................................................................. 85
Multivariate results. ..................................................................................................... 86
Discussion ......................................................................................................................... 88
Strengths and limitations. ............................................................................................ 94
Conclusion ........................................................................................................................ 95
CHAPTER SIX: SUMMARY AND SIGNIFICANCE .................................................... 105
REFERENCES .................................................................................................................... 111
APPENDIX A ....................................................................................................................... 123
APPENDIX B ....................................................................................................................... 136

iv

LIST OF TABLES
Table 1. Keyword Search Used for All Databases .................................................................. 16
Table 2. Subject Heading Search Used for PubMed (Medline), CINAHL (EBSCO), and
PsycINFO (EBSCO) ............................................................................................... 17
Table 3. List of Predisposing, Enabling, and Need Characteristics Study Variables .............. 35
Table 4. Selected List of interRAI Outcome Scales Common to the RAI-MDS 2.0
Instruments with Description and Ranges .............................................................. 39
Table 5. Variables Used to Measure the Outcomes of Residents ............................................ 40
Table 6. Predisposing, Enabling, and Need Characteristics by Status of Hospice Use of
Residents (N=185,715) and End-Stage Residents (N=12,684) in Long-Term Care
Facilities Across Canada in 2015 ........................................................................... 65
Table 7. Logistic Regression Model Predicting Hospice Use Among Residents in Long-Term
Care Facilities in Canada in 2015 ........................................................................... 68
Table 8. Final Logistic Regression Model Predicting Hospice Use Among Residents in LongTerm Care Facilities in Canada in 2015 ................................................................. 70
Table 9. Predisposing, Enabling, and Need Characteristics by One-Year Survival
Survivorship of Hospice Use Among Residents in Long-Term Care Facilities in
Canada at Baseline in 2015 (N=185,715) ............................................................... 96
Table 10. Multinomial Logistic Regression of Predisposing, Enabling, and Need
Characteristics on One Year Survivorship of Hospice Use Among Residents in
Long-Term Care Facilities in Canada in 2015 with Residents Who Did Not
Receive Hospice Care and Were Alive One Year Following the Last Assessment
as Reference (Odds Ratios and 95% Confidence Intervals). ................................ 100

v

LIST OF FIGURES
Figure 1. The Role of Hospice Palliative Care During Illness .................................................. 2
Figure 2. The Conceptual Framework of This Thesis: A Modified Andersen and Newman’s
Behavioral Model for Health Services Utilization (1973)...................................... 15
Figure 3. Flow Diagram of Article Selection Process ............................................................. 18
Figure 4. Selected Common Diagnoses by Status of Hospice Use of Residents (N=185,715)
and End-Stage Residents (N=12,684) in Long-Term Care Facilities in Canada in
2015 ........................................................................................................................ 72
Figure 5. Number of Clinical Protocol Assessments Triggered by Status of Hospice Use
Among Residents (N=185,715) and End-Stage Residents (N=12,684) in LongTerm Care Facilities in Canada in 2015 ................................................................. 73
Figure 6. Survival Analysis by Status of Hospice Use of Residents in Long-Term Care
Facilities in Canada in 2015 (N=185,715) ........................................................... 102
Figure 7. Selected Common Diagnoses by Status of Hospice Use of Residents in Long-Term
Care Facilities in Canada in 2015 (N=185,715) ................................................... 103
Figure 8. Number of Clinical Protocol Assessments Triggered by One-Year Survivorship of
Hospice Use Among Residents in Long-Term Care Facilities in Canada in 2015
(N=185,715) ......................................................................................................... 104

vi

Acknowledgement
I would like to express my deepest appreciation to my thesis supervisor, Dr. Shannon
Freeman, for being such a wonderful and supportive supervisor. I really appreciate how she
has always been so friendly and supportive of all of my efforts. I become more confident and
creative because she saw the potential in me. Her knowledge, expertise, perseverance,
patience, and encouragement were vital for the development and accomplishment of my
thesis. Her encouragement, attention to details, and critiques have immensely improved my
thesis and challenged me to develop as a researcher.
I would like to express my great gratitude to the members of my thesis committee, Dr.
Davina Banner-Lukaris and Dr. Lina Spirgiene for their depth of knowledge, insight and
guidance. Their support and mentorship have meant so much. I would like to sincerely thank
Dr. Tammy Klassen-Ross for assessment for my thesis proposal and serving as the chair of
my thesis defense. I am honored to have had Dr. Annie Duchesne as the external examiner of
my thesis.
I am grateful to have received financial support from the University of Northern
British and Mitacs towards completion of my thesis program. Thank the Canadian Institute
for Health Information for proving the study data.
Thank you to my parents for being my biggest supporters and always encouraging me
to pursue my dreams. Thank you to my boyfriend Xiangyi Zeng for all his love and support.
Big thanks to everyone who has helped me with my thesis work.

vii

Chapter One: Introduction
This chapter introduces the definition of hospice care, what is involved in hospice
care, who provides hospice care, and the history of hospice care. This chapter also introduces
the definition of long-term care facilities (LTCFs) and gives an overview of the background
of hospice care in LTCFs in Canada. This chapter also presents the research questions and
the conceptual framework that was used to guide the work described in this thesis.
Definition of Hospice Care
According to the Canadian Hospice and Palliative Care Association (CHPCA)
(2016a), hospice care (also known as palliative care in Canada) is defined as “a special kind
of health care for individuals and families who are living with a life-limiting illness that is
usually at an advanced stage” (para. 2). The goal of hospice care is “to provide comfort and
dignity for the person living with the illness as well as the best quality of life for both this
person and his or her family” (CHPCA, 2016a, para. 2). Hospice care involves a) pain
management; b) symptom management, such as anxiety and depression; c) social,
psychological, spiritual and emotional support; and d) caregiver support (CHPCA, 2016b). It
is provided by a diverse team, which may include physicians, nurses, pharmacists, social
workers, trained volunteers, home care workers, bereavement support workers, and informal
caregivers such as family members (Health Canada, 2016). Hospice care can reduce or
relieve physical and psychological symptoms, provide comfort and dignity for the person
living with the illness as well as the best quality of life for both this person and his or her
family (CHPCA, 2016a).
The term palliative care is often used interchangeably with hospice care
internationally, however, the two words have slightly different meanings (Hui et al., 2013).
The word hospice is the original word that is used to describe St. Christopher’s Hospice, the
1

first modern hospice, which opened in 1967 in the UK (Stoddard, 1978). The term palliative
care emerged from Canada in 1974, which was initially used to refer to the opening of a
hospice care unit in the hospital (Sheehan & Forman, 1996). The term has been
predominantly used in Canada since then. However, the terms hospice care and palliative
care have become more synonymous in both Canada and the UK over the last 20 years
(Hutchings, 2002). The Canadian Hospice Palliative Care Association has added the word
hospice to its name and it also clearly pointed out that “In Canada, palliative care and hospice
care are used to refer to the same thing - this specific approach to care.” (CHPCA, 2016c,
para. 1). In Canada, hospice palliative care starts at the beginning of the life-threatening
illness which aims to relieve suffering and improve quality of life, although the focus of this
stage is mainly on therapy to treat or manage the illness (CHPCA, 2013) (Figure 1).
Commonly, the treatments and care provided across the course of the life-threatening illness
may fluctuate according to the person’s and family’s issues and needs, goals of care and/or
treatment priorities (CHPCA, 2013).
Figure 1. The Role of Hospice Palliative Care During Illness

Source: “A model to guide hospice palliative care: Based on national principles and Norms of Practice” by
Canadian Hospice Palliative Care Association, 2013, retrieved from
http://www.chpca.net/media/319547/norms-of-practice-eng-web.pdf

2

In the U.S., hospice care is defined as a special kind of health care that is “provided
for a person with a terminal illness whose doctor believes he or she has 6 months or less to
live if the illness runs its natural course” (National Institute on Aging (NIA), 2016, para. 9).
Like hospice care in Canada, hospice care in the U.S. provides comprehensive comfort care
as well as support for the family through a hospice team, however, unlike hospice care in
Canada, hospice care in the U.S. is provided only when attempts to cure the person’s illness
are stopped (NIA, 2016). Palliative care is different from hospice care in the U.S. “Palliative
care extends the principles of hospice care to a broader population that could benefit from
receiving this type of supportive care earlier in their illness or disease process” (Ferrell &
Coyle, 2010, p. 19). The NIA (2016) asserts that “palliative care can be provided along with
curative treatment and may begin at the time of diagnosis” (NIA, 2016, para. 6). Attention
should be paid to the meaning of hospice care and palliative care, because they are different
in Canada and the U.S.
History of Hospice Care
The first hospice may date back as far as the 4th century in Europe when Christians
provided care to the sick and destitute (CHPCA, 2016d). In the early 14th century, the order
of the Knights Hospitaller of St. John of Jerusalem opened the first hospice in Rhodes (which
belongs to Greece now) to provide care for the ill and dying as well as refuge for travelers
and pilgrims (Connor, 1998). Hospices thrived in the Middle Ages (5th-15th century), but
languished in 16th century as religious orders became dispersed (Robbins, 1983). Hospice
was revived, by the Daughters of Charity of Saint Vincent de Paul, in the 17th century in
France (Connor, 1998). France continued to see the development of hospice; the hospice of
L'Association des Dames du Calvaire founded by Jeanne Garnier, opened in 1843, followed
by six other hospices before 1900 (Lewis, 2007). In the meantime, hospices also developed in
3

other countries. In the UK, the articles published in the Lancet and the British Medical
Journal put forward the need of terminally ill for good care and sanitary conditions in the
middle of the 19th century (Lewis, 2007). The first hospice in the UK, the Friedenheim, was
founded by Davidson in 1885 (Glasgow End of Life Studies Group, 2014). The Friedenheim
initially had eight beds, but in 1892 it could provide accommodation for up to 35 patients
(Glasgow End of Life Studies Group, 2014). Four more hospices were established in the UK
by 1905 (Lewis, 2007). Australia also experienced the active development of hospice
facilities, with well-known hospices including the Home for Incurables in Adelaide (1879),
the Home of Peace (1902), and the Anglican House of Peace for the Dying in Sydney (1907)
(Lewis, 2007). In New York City, St. Rose's Hospice was founded by the Servants for Relief
of Incurable Cancer in 1899 and soon expanded with six locations in other cities (Lewis,
2007). The more influential early developer of hospice was the Irish Religious Sisters of
Charity, who opened Our Lady's Hospice in Harold's Cross, Dublin, Ireland in 1879 (Lewis,
2007) and who expanded internationally with the Sacred Heart Hospice for the Dying opened
in Sydney in 1890 (Lewis, 2007), St Joseph's Hospice opening in London in 1905 (Connor,
1998; Foley & Herbert, 2002), and hospices in Melbourne and New South Wales followed in
the 1930s (Lewis, 2007).
The first modern hospice, the St. Christopher's Hospice, was created by Dame Cicely
Saunders in the UK in 1967 (Stoddard, 1978). She was the first to apply hospice care to
specialized care for terminally ill patients. She also introduced the idea of specialized care for
the dying to the US and promoted the foundation of the first modern hospice in the US,
which was established in New Haven, Connecticut in 1974 (National Hospice and Palliative
Care Association, 2016; The World Federation of Right to Die Association, 2016).

4

In Canada, modern hospice care movement began with the creation of palliative care
units in a hospital setting (CHPCA, 2016d). As the word hospice was already used in France
to refer to nursing homes, the term palliative care ward was used to refer to hospice care
units in a hospital setting in Canada (Duffy, 2005; Feldberg, Ladd-Taylorm, & Li, 2003). The
term palliative care has been in predominant use in Canada since then. In 1974, the first
palliative care unit opened at Winnipeg's St. Boniface General Hospital (CHPCA, 2016d).
Within a matter of weeks, the second palliative care ward in Canada opened at the Royal
Victoria Hospital in Montreal, Quebec (CHPCA, 2016d). Subsequently, hospice palliative
care programs developed as divisions of larger organizations or agencies (CHPCA, 2016d).
The first university institute for research and education in palliative care was created at the
University of Ottawa in 1983 (CHPCA, 2016d). Volunteer-based hospice societies and
organizations began to develop soon after (CHPCA, 2016d). In November 1991, the
Canadian Palliative Care Foundation (renamed as Canadian Palliative Care Association in
2001) was formally established as a national charitable organization (CHPCA, 2016d). In
1992, the Ministry of Health (later renamed the Ministry of Health and Long-Term Care in
1999) provided funding for the purpose of increasing palliative care in communities as
opposed to hospitals in Ontario (DeMiglio, Dykeman, Williams, & Kelley, 2012). In 1995,
Community Care Access Centers in Ontario were developed, which “enforced the
momentum towards delivering care within the community and increasing opportunities for
patients to remain at home even at the end of life” (DeMiglio et al, 2012, p. 113). In the same
year, the report “Of Life and Death - Final Report” from the Special Senate Committee on
Euthanasia and Assisted Suicide examined the issues regarding assisted suicide and
euthanasia in response to the request for the right to die case of Sue Rodriguez (Special
Senate Committee on Euthanasia and Assisted Suicide, 1995). This report put forward that
5

palliative care, pain control, sedation, and advance directives should be considered before
assisted suicide and euthanasia were considered (Chapter 2). In this report, the term palliative
care was used and was defined as “care aimed at alleviating suffering - physical, emotional,
psychosocial, or spiritual - rather than curing” (Chapter 2). This report also demonstrated that
“people need better support during the dying process and in dealing with the circumstances
surrounding death” and that “palliative care could address many of these needs and may
serve as alternatives to assisted suicide and euthanasia” (Chapter 3). In 2000, the Standing
Senate Committee on Social Affairs, Science and Technology released a landmark report
entitled “Quality End-Of-Life Care: The Right of Every Canadian” to examine and report
upon developments since the 1995 report “Of Life and Death - Final Report” (CHPCA,
2016d; Standing Senate Committee on Social Affairs, Science and Technology, 2000). In
2001, Senator Carstairs was appointed as Minister with Special Responsibility for Palliative
Care and the Secretariat on Palliative and End-of-Life Care at Health Canada was created
(CHPCA, 2016d). Since then, more attention on policy, research, and education on hospice
care was paid in Canada. (CHPCA, 2016d).
Long-Term Care Facilities (LTCFs) in Canada
According to Canadian Institute of Health Information (CIHI), a LTCF is defined as a
care institution that “serves diverse populations who need access to 24-hour nursing care,
personal care and other therapeutic and support services that are not provided through home
care programs, in retirement homes, in the persons own home or in a shared private home”
(Freeman et al., 2017, p. 2). Long-term care facilities are not publicly-insured services under
the Canada Health Act but is partially insured as extended health care services (McGregor, &
Ronald, 2011; Canadian Healthcare Association [CHA], 2009). They are governed by
provincial and territorial legislation; jurisdictions offer a different range of services and cost
6

coverage across Canada (Health Canada, 2004). As a result, there is lack of consistency in
what LTCFs are called and how LTCFs are governed (McGregor, & Ronald, 2011). The U.S.
commonly describes these facilities as “nursing homes” (McGregor, & Ronald, 2011). In
Canada, different terms are used across and within provinces and territories, such as “longterm care facilities”, “long-term care homes”, “nursing homes”, “personal care homes”,
“special care homes” (Alberta Health Services, 2019; McGregor, & Ronald, 2011). In most
provinces, LTCFs are publicly funded; however, there is also an income-tested user-pay
funding component for residential long-term care (McGregor, & Ronald, 2011). In some
provinces, residents pay for accommodation and/or personal care expenses and these expense
rates vary across Canada (CHA, 2009). Furthermore, while the majority of LTCFs are
publicly funded, service delivery is provided by a mix of public (government-owned), private
not-for-profit, and private for-profit providers (Centre for Health Services and Policy
Research, 2015; McGregor, & Ronald, 2011).
Hospice care can be provided in a variety of settings, such as private homes,
hospitals, LTCFs, and hospice facilities and “although hospitals are designed to address
severe and urgent needs, they may not be the best location for comfortable end-of-life care”
(Health Canada, 2016, para. 8). Long-term care facilities may be a good location for end-oflife care. In LTCFs, hospice care can benefit residents without family and/or caregivers and
residents who need constant monitoring by professional staff (Infeld, Crum, &
Koshuta,1990). Providing hospice services in LTCFs can be cost-efficient compared to
providing hospice care at private homes and in hospice facilities because the basic services
such as housekeeping and central supplies can be shared between hospice facilities and
LTCFs and be provided by LTCF staff (Infeld et al., 1990). Having a hospice program in
LTCFs also has benefits for both hospice facilities and LTCFs. Hospice facilities may have
7

greater revenues by increasing their resident volume, utilizing staff more efficiently,
overlapping basic services, and increasing the average length of stay (Petrisek & Mor, 1999).
Long-term care facilities who enroll residents in hospice care programs can increase their
competitiveness of the market by promising to provide hospice care to residents who are
nearing the end of life, reduce in-house staff time while providing these special services, and
also benefit from the knowledge of hospice staff (Castle, 1999; Petrisek & Mor, 1999).
Hospice Care in LTCFs in Canada
According to Statistics Canada, there were about 267,000 deaths in 2016 in Canada,
about 80% of whom were aged 65 years and older (Statistics Canada, 2018). Population
growth forecasts predicted that the number of deaths in Canada would increase annually by
24% to 330,000 by 2026 and 40% to more than 375,000 by 2036 (Statistics Canada, 2015a).
It was estimated that about 10% of deaths occur suddenly, thus, 90% of decedents may have
benefitted from hospice care (Carstairs, 2010). However, hospice care was only available to
5% of the deceased in 1995 (Carstairs, 2010) and 16% to 30% in 2008 (CHPCA, 2014). A
survey showed 34% of Canadians did not know to whom hospice services were available to
(CHPCA, 2014). Thus, it is important to make hospice care available to more Canadians as
well as to increase Canadians’ awareness of hospice care.
In addition to the increasing number of deaths, Canada is also facing an ageing
population. According to Statistics Canada (2017a), about one in six of Canadians were 65
years old and older in 2016. It was estimated that about one in four of Canadians will be 65
years old and older by 2031 (Carstairs, 2010). The increasing ageing population and
improvements in life expectancy have caused an increase in the number of older adults (aged
65 years and older) with chronic and complex diseases (Carstairs, 2010). Almost four in five
of older adults have at least one chronic disease (Carstairs, 2010). Many older adults suffer
8

serious cognitive and physical impairments and have unstable and complex health needs
(Canadian Union of Public Employees, 2009; Carstairs, 2010). Thus, there is an increasing
need for formal care services, such as LTCFs, to provide care to these persons (Jayaraman &
Joseph, 2013).
In Canada, approximately 150,000 residents live in LTCFs (Statistics Canada,
2015b). Most of these residents stay in LTCFs until death, making the LTCF a common place
of death for frail residents with chronic and complex diseases in Canada (Carstairs, 2010;
Jayaraman & Joseph, 2013). It has been estimated that about one in four of Canadians die in
LTCFs each year and this number is expected to rise steadily (Brink & Kelley, 2015). It was
predicted that about 39% of Canadians would die in LTCFs by 2020 (Jayaraman & Joseph,
2013). For those aged 65 years and older, nearly 30% die in LTCFs every year (Motiwala,
Croxford, Guerriere, & Coyte, 2006). For those over the age of 85 years, LTCFs are the most
common place of death, with approximately 50% of all deaths occurring there (Motiwala et
al., 2006). Given the increasing number of deaths that occur in LTCFs, it is important to
examine hospice care practices in LTCFs in Canada.
Problem Statement and Research Questions
While a large proportion of Canadians die in LTCFs each year, most LTCFs lack a
formalized hospice care program (Quality Palliative Care in Long Term Care Alliance,
2011). Research has emerged studying hospice care across various settings (e.g., private
homes, hospitals, and hospice facilities), however, a particular gap in knowledge exists
regarding hospice care use in the long-term care setting, particularly across Canada (Ersek &
Carpenter, 2013). Little is known about what the residents in LTCFs in Canada who receive
hospice care look like and what factors affect hospice use among residents in LTCFs in
Canada. Hospice care may also not be used appropriately in LTCFs. The general eligibility
9

for hospice care program in Canada is a) the person has a specific prognosis (often six
months or less), b) a decision has been made to focus on comfort rather than cure, and c)
resuscitation will not be used when the illness brings a natural death (Canadian Virtual
Hospice, 2019). However, some persons may be referred to hospice programs in the last days
of life (Han, Tiggle, & Remsburg, 2008; Lun, 2007), while others may have lived more than
six months after enrollment in hospice (Harris et al, 2014). It is hard to recognize terminal
stages when the person has several medical problems but no specific terminal diagnosis. Due
to limited research, it is unclear who receive hospice care but survive an unexpected length of
time (i.e. more than six months) in LTCFs and the related characteristics. It is also unclear
who may have benefited from hospice care but did not receive it and the related
characteristics.
The aim of the work described in this thesis was to examine the characteristics of
residents who received hospice care in LTCFs in Canada and to explore factors that could
predict hospice use in the LTCF population in Canada. Specifically, the work described in
this thesis was guided by the following two main research questions:
(a) What are the characteristics of residents who received hospice care in LTCFs in
Canada versus those who did not by end-stage status and by one-year survivorship?
(b) What variables can predict hospice use and one-year survivorship of hospice use
among LTCF residents in Canada?
The characteristics of residents who received hospice care in LTCFs versus those who
did not by end-stage status and the predictors of hospice use among LTCFs in Canada will be
examined in Chapter Four. The characteristics of, and one-year survivorship among,
residents who received hospice care in LTCFs versus those who did not and predictors of
one-year survivorship of hospice use among LTCF residents in Canada will be examined in
10

Chapter Five. To the knowledge of the author, the work described in this thesis was the first
to quantitatively assess and compare the hospice use among LTCF residents across Canada.
Understanding these characteristics may help LTCFs and hospice care providers to better
understand hospice care practices in LTCFs in Canada, provide evidence for needs of change
or improvements in care in the long-term care setting, and identify ways to implement
services and interventions that can improve access to, and utilization of, hospice care in
LTCFs in Canada.
Conceptual Framework
Andersen and Newman’s (1973) behavioral model for health services utilization was
used to guide the work described in this thesis. This behaviour model was initially developed
in the late 1960s, which was designed to help “to understand why family use health services,
to define and measure equitable access to health care, and to develop policies to promote
equitable access” (Andersen, 1995, p1). The initial model used the family as the unit of
analysis. Because it was difficult to examine the family characteristics, the initial model was
then modified to use the individual as the unit of analysis. The modified model was designed
to assist in explaining and predicting demographic and societal determinants for utilization of
health services and developing policies to promote equitable access to these services
(Andersen & Newman, 1973). The behavioral model proposed that the utilization of health
services was a result of three components of population characteristics: (a) the predisposition
of the person to use services (predisposing characteristics), (b) the person’s ability to secure
the services (enabling characteristics), and (c) the person’s illness level (need characteristics)
(Andersen and Newman, 1973).
Predisposing characteristics. Predisposing characteristics (i.e. demographic
characteristics, social structure, and health beliefs) are identified as social-demographic
11

characteristics that influence or bias the use of health care services (Andersen and Newman,
1973). Demographic characteristics such as age and gender represent biological imperatives
which suggest the likelihood that a person will need the health services. The social structure
such as education and occupation shows the person’s status in the community, the person’s
ability, and the resources available to the person to help them cope with the presenting
problems. Health beliefs such as values and attitudes may influence the person’s subsequent
perceptions of need and utilization of health services.
Enabling characteristics. Enabling characteristics (i.e. family resources and
community characteristics) reflect the person’s ability to access health services (Andersen
and Newman, 1973). As applied using the Andersen and Newman’s (1973) model, both
community and personal enabling resources must be present for the utilization of health
services to take place.
Need characteristics. Need characteristics are the conditions related to the health
level (perceived or evaluated) that are the reasons why health services are needed (Andersen
and Newman, 1973). The evaluated needs stand for professional judgement about the
person’s health status and the person’s need for health services.
The outcome of the Andersen and Newman’s (1973) model is utilization of health
services. This model suggests that “people’s use of health services is a function of their
predisposition to use services, factors which enable or impede use, and their need for care”
(Andersen, 1995, p. 1).
Andersen (1995) also described the measurements of four types of access to health
services: potential access, realized access, equitable access, and inequitable access. Potential
access is defined as the presence of enabling resources. More enabling resources relate to the
increasing likelihood of using health services. Realized access is the actual use of the health
12

services. Equitable access occurs “when demographic and need characteristics account for
most of the variance in utilization”, while inequitable access occurs “when the social
structure, health beliefs, and enabling characteristics determine who gets health services”
(Andersen, 1995, p. 4-5).
Operational Definition
Figure 2 illustrates a modified version of Andersen and Newman’s behavioral model
that was used for the conceptual framework of the work described in this thesis. It should be
noted that family resources were not included in the model, because the secondary data used
in the work described in this thesis does not provide variables that could measure family
resources.
Predisposing characteristics. Predisposing characteristics are the resident’s
characteristics that would influence or bias susceptibility for health or illness and one's
attitude toward obtaining health services (Andersen, 1995). The work described in this thesis
included demographic characteristics, social structure, and health beliefs. Variables that were
used to examine demographic characteristics are age, sex, and marital status. Variables that
were used to measure social structure are education and language. Variables that were used to
measure health beliefs are do-not-resuscitate (DNR) order and do-no-hospitalize (DNH)
order.
Enabling Characteristics. Enabling characteristics reflect the resident’s ability to
access health services. In the work described in this thesis, all participants were LTCF
residents, which means that they were able to access long-term care services. The work
described in this thesis included community characteristics that could reflect the resident’s
ability to access hospice care services. Variables that were used to examine the community
characteristics are geographic region, urban-rural status, and Quintile of Adjusted Income per
13

Person Equivalent (QAIPPE) of the LTCF. According to Statistics Canada (2017b), “an
urban area was defined as having a population of at least 1,000 and a density of 400 or more
people per square kilometre and all areas outside urban areas was defined as rural areas”.
Urban areas and rural areas cover the entire nation together (Statistics Canada, 2017b). The
QAIPPE is a simple area-based socioeconomic measure of neighbourhood income quintile
for the facility providing care to the resident (Public Health Ontario, 2013).
Need characteristics. Need characteristics represent the resident’s need to access to
health services. The work described in this thesis included perceived and evaluated clinical
characteristics. The variables that were used to measure the clinical characteristics are disease
diagnoses, cognitive function, communication function, mood and behavior, physical
function, continence, pain, skin condition, nutritional status, health stability, and health
complexity of the LTCF resident.
The outcome of the behavior model is utilization of hospice care. Utilization of
hospice care is defined as LTCF residents being in a hospice care program for the
management of terminal diseases (CIHI, 2015).

14

Figure 2. The Conceptual Framework of This Thesis: A Modified Andersen and Newman’s
Behavioral Model for Health Services Utilization (1973)

Adapted from “Societal and individual determinants of medical care utilization in the United States” by R. M.
Andersen and J. F. Newman, 1973, Milbank Memorial Fund Quarterly, 51, 107.

The modified version of Andersen and Newman’s (1973) behavioral model was used
in the work described in this thesis to assist in evaluating hospice care practices in LTCFs in
Canada and understanding and predicting social-demographic and health determinants for
utilization of hospice care services.

15

Chapter Two: Literature Review
In this chapter, the literature that is related to the characteristics of residents who
receive hospice care in LTCFs and the predictors of hospice use among residents in LTCFs is
reviewed and discussed.
Literature Search
Database selection. Literature was identified through a search of electronic
databases. PubMed (Medline), CINAHL (EBSCO), Academic Search Complete (EBSCO),
PsycINFO (EBSCO), Web of Science, Cochrane Reviews (Ovid), and Scopus were chosen
for this review.
Search strategies. The three key concepts of the work described in this thesis are
hospice care, LTCF, and characteristic. Search strategies were created for individual
databases based on these three concepts. These strategies involved various combinations of
keywords (Table 1) and subject headings (Table 2). Subject headings were only available for
PubMed (Medline), CINAHL (EBSCO), and PsycINFO (EBSCO).
Table 1. Keyword Search Used for All Databases
Key Concept
Keywords
Hospice/palliative care
Hospice care OR palliative care OR end of life care OR
terminal care
Long-term care facility
(Long-term care OR residential OR nursing OR personal
care OR convalescent OR continuing care OR assisted living
OR for the aged) AND (facility OR facilities OR home OR
homes)
Characteristic
Characteristic* OR factor* OR predict* OR determinant*
OR influence* OR reason* OR cause*

16

Table 2. Subject Heading Search Used for PubMed (Medline), CINAHL (EBSCO), and
PsycINFO (EBSCO)
Database
PubMed
CINAHL
PsycINFO
Key Concept
(Medline)
(EBSCO)
(EBSCO)
Hospice/palliative care
Hospice and
Hospice and
Hospice OR
palliative care
palliative nursing
palliative care
nursing OR
OR hospice care
hospice care OR
OR palliative care
palliative care OR
OR terminal care
terminal care
Long-term care facility Long-term care OR Long-term care
Long-term care OR
residential
OR residential
residential care
facilities OR
facilities OR
institutions OR
assisted living
residential care
assisted living OR
facilities OR
OR assisted living nursing homes
homes for the aged OR nursing homes
OR nursing homes OR skilled nursing
OR skilled nursing facilities
facilities
Literature selection and search outcome. The full searches were done on March 22,
2019. No time limit was applied to the databases. Searches were limited to published articles
in English and Chinese languages because of the cost and time involved in translating the
literature. Articles in which children (human beings aged below the age of 19 years) were the
study population were excluded. Articles in which long-term care was provided through
home care programs, in retirement homes, in the person’s own home or in a shared private
home were excluded. Literature related to factors associated with hospice use among LTCF
residents were included. Through the systematic literature search, a total of 3,656 nonduplicate articles were identified, of which 3,550 were excluded during the title and abstract
screening stage. The full text of the remaining 106 articles were reviewed, of which 77 were
excluded after full text screen and 3 more were excluded during data extraction. Finally, 26
studies are included for this review (Figure 3).

17

Figure 3. Flow Diagram of Article Selection Process

Literature Analysis
Characteristics of included studies. Of the 26 included studies, the earliest study
was published in 1990 (Infeld et al., 1990) and the most recent studies were published in
2018 (Nilsen, Wallerstedt, Behm, & Ahlström, 2018; Rodriquez & Boerner, 2018). Two
studies were conducted in Sweden (Rodriquez & Boerner, 2018; Smedbäck, Öhlén, Årestedt,
Alvariza, Fürst, & Håkanson, 2017), one was an international study (Hjermstad et al, 2016),
and the others were all carried out in the U.S. The international study recruited thirty
facilities including two in Alberta, Canada, however, analyses in this study were not stratified
by country (Hjermstad et al, 2016). As a result, this literature review was mostly based on
American studies. All the included articles were read critically and study details were
extracted and summarized thematically based on the analyses of the study findings using the
Matrix Method (Garrard, 2011) (Appendix A). The themes include prevalence of hospice use
in LTCFs, length of hospice use in LTCFs, hospice use in LTCFs versus in non-LCTFs,

18

hospice use among residents in LTCFs, and hospice use among end-stage residents in
LTCFs.
Prevalence of hospice use in LTCFs. Hospice care in LTCFs is provided through a
collaboration with hospice facilities, specially trained LTCF staff, or separate specialty units
focusing specifically on providing hospice care (Vulpen, 2013). The use of hospice care
among LTCF residents increased over time (Gozalo & Miller, 2007; Infeld et al., 1990;
Petrisek & Mor, 1999). Infeld et al.’s (1990) study shows there was an increase in the number
of annual admissions in the LTCF examined in this study from 1978 to 1988. Since 1989,
hospice services in LTCFs have been reimbursable under Medicare and Medicaid, which led
to a rapid expansion of hospice care in LTCFs in the U.S (Petrisek & Mor, 1999). About 8%
of hospice recipients lived in LTCFs in 1989, and this proportion rose to 17% in 1995
(Petrisek & Mor, 1999). Petrisek and Mor (1999) found about 30% of all Medicare/Medicaid
certified LTCFs in the U.S. reported having at least one hospice resident in 1995-1997.
Although the number of hospice users steadily increased each year, hospice services
were still underutilized (Chapin, Gordon, Landry, & Rachlin, 2007). Gozalo and Miller
(2007) found 8% of end-stage residents in nursing homes were enrolled in hospice during
1995 and 1997 in five states (Kansa, Maine, New York, Ohio, South Dakota) of the U.S. Han
et al. (2008) referenced the statistics from National Hospice and Palliative Care Organization
that “less than 5% of those who died in nursing homes in 1998 received hospice care” (p.
478). About one third of the residents in nursing homes who died before the end of the year
2000 received hospice care in Michigan (Lin, 2010). In 2006, the rate of hospice use was
39% for all urban decedents in nursing homes in the U.S. (Lepore, Miller, & Gozalo, 2011).
The international study referenced the statistics from World Health Organization in 2014 that
“only 1 in 10 of those who need palliative care receive it” (Hjermstad et al., 2016, p. 1). In
19

Hjermstad et al.’s study, palliative care referred to care that is “available early in the disease
trajectory and may be delivered alongside potentially curative treatments” (p. 1). Hospice use
in LTCFs from a global and national perspective in recent years has not been extensively
examined in the included studies. Because the rate of hospice use in LTCFs may vary
substantially across states and across facilities (Zheng, Mukamel, Caprio, & TemkinGreener, 2013; Petrisek & Mor, 1999), the finding from included studies may not be
generable to other populations.
Length of hospice use in LTCFs. Hospice users in the U.S. were thought to have
less than six-month survival (NIA, 2016). Many residents in the U.S. nursing homes were
enrolled in hospice in the last few days of their life, while other residents who were eligible
for hospice at the admission lived more than six months after hospice enrollments (Huskamp,
Stevenson, Grabowski, Brennan & Keating, 2010; Unroe et al, 2013). Infeld et al. (1990)
found about 35% of residents stayed less than six days, most stayed one day, and the average
length of stay decreased from about 29 days in 1978 to 12 days in 1988. An investigation of a
large regional hospice facility which served five New England states revealed 25% of 13,479
hospice users were enrolled for 5 or fewer days, 10% were enrolled for 240 or more days,
and the median length of hospice stay in LTCFs was 17 days in 2001-2008 (Huskamp et al.,
2010). However, a study of 1,452 hospice users aged 65 years or over in nursing homes in
Indianapolis showed the mean length of hospice stay was 114 days, about 20% of whom had
hospice stays longer than six months in 1999-2009 (Unroe et al, 2013). Concerns about
longer stays have been emphasized because “long hospice stay is associated with a lower
intensity of services provided by hospice (providers), raising the issue of the degree of value
added to the usual care of nursing home patients” (Unroe et al, 2014, p. 193).

20

Hospice use in LTCFs versus in non-LCTFs. Of the 26 included articles, three
studies compared hospice care in LTCFs with hospice care in other settings and revealed
residents who received hospice care in LTCFs were different from residents who received
hospice care in other settings (Han et al., 2008; Hjermstad et al., 2016; Unroe et al., 2014).
Han et al. (2008) compared 617 deceased residents who received hospice care in nursing
homes to 1,170 deceased residents who received hospice care at homes in the U.S. and found
compared to those who received hospice care at home, residents who received hospice care in
nursing homes were more likely to be 85 years and older (45% vs. 17%) and widowed (49%
vs. 29%) and have dementia (10% vs. 2%), noncancer primary diagnoses (66% vs. 36%),
heart disease (23% vs. 11%), and difficulties in controlling bowels (74% vs. 34%) and
bladder (62% vs. 29%). This study also revealed residents who received hospice care in
nursing homes received more dietary/nutrition service (77% vs. 13%), medication
management (92% vs. 64%), and physician services (94% vs. 28%) than home hospice
patients. Likewise, a study of 3,771 hospice users aged 65 years or over in Indianapolis was
conducted by Unroe et al. (2014) to address this issue with hospice patients classified into
four groups: in nursing homes, not in nursing homes, crossover patients using hospice in both
settings, and “near-transition” patients. This study revealed residents who received hospice
care in nursing homes were more likely to be old and female and have dementia, arthritis,
and diabetes compared to residents who received hospice care in nursing homes in other
settings. Hjermstad et al. (2016) conducted an international study of 1,698 residents aged 18
years and older from 12 countries and found that patients in hospice/nursing homes category
were more likely to be older, have poorer performance status, report more tiredness, and have
shorter survival compared with patients who received hospice care at hospitals. However,
this study combined hospice facilities, hospice in nursing homes and hospice at private
21

homes in one category (hospice/nursing homes category) and did not examine the differences
between hospice facilities, hospice in nursing homes and hospice at private homes. The
characteristics of residents in hospice/nursing homes category may not be generalizable to
hospice in nursing homes. Although this study defined palliative care, they did not define
hospice care and ignored the differences in definitions of palliative care and hospice care in
different countries.
Hospice use among residents in LTCFs. Half of the included studies (n=13/26)
examined hospice use among the general LTCF population. Infeld et al. (1990) conducted a
descriptive longitudinal study using administrative records from 1978 to 1988. This study
revealed there was a wide variation in age over time and the mean age of hospice patients did
not changed dramatically over time ranging from 57 years old in 1978 to 63 years old in
1988. This study also revealed there was a decline in proportion of females (82% in 1978 vs.
43% in 1988), Caucasian (82% in 1978 vs. 51% in 1988), married (77% in 1978 vs. 54% in
1988) and a diagnosis of cancer (100% in 1978 vs. 75% in 1988) over time. Huskamp et al.
(2010) examined 13,479 hospice residents from 394 LTCFs from five New England states
from 2001 to 2008 and found two thirds of the residents were female, three quarters were
aged 81-90 years, two thirds were unmarried, four fifths were Caucasian and their common
diagnosis was dementia (35%), cancer (19%), terminal debility (17%), and cardiovascular
disease (9%). Unroe et al. (2013) identified 1,452 hospice residents aged 65 years and over in
Indianapolis from 1999 to 2008 and found the mean age of the residents were 81(+8.2) years,
62% of them were female, 40% were African-American, 18% had severe cognitive
impairment (a CPS of 5 or 6), their most common diagnoses were dementia (75%) and
cancer (56%). This study also examined the primary hospice diagnosis of the residents and
revealed cancer was the most common primary hospice diagnosis followed by dementia and
22

the proportion of residents with cancer as primary hospice diagnosis decreased over time
(60% in 1999 vs. 24% in 2008). There was no comparison group in these three studies,
therefore, the causes of changes in resident characteristics were unknown.
The other 10 studies compared the characteristics of residents who received hospice
care in LTCFs with those who did not in LTCFs. Andersen and Newman’s (1973) theoretical
framework was used to guide and organize the following description about the findings for
these studies.
Predisposing characteristics. Predisposing characteristics that were examined among
the 10 studies included age, sex/gender, marital status, race/ethnicity, education, and advance
directives.
Age. Two studies revealed residents who received hospice care and those who did not
were similar in the value of mean age (Kiely, Givens, Shaffer, Teno, & Mitchell, 2010;
Kwak, Haley, & Chiriboga, 2008). Lepore et al. (2011) found hospice residents were more
likely to be aged 75 years and less. However, Vulpen (2013) found older age could predict
utilization of end-of-life services.
Sex/gender. The majority of studies showed residents who received hospice care were
more likely to be female (Lepore et al., 2011; Kwak et al., 2008; Zheng, Mukamel, Caprio,
Cai, & Temkin-Greener, 2011) with an odds ratio ranging from 1.09 (95% CI [1.02, 1.17]) in
Kwak et al.’s (2008) study to 1.22 (95% CI [1.20, 1.24]) in Lepore et al.’s (2011) study.
However, Kiely et al. (2010) found there was no difference in proportion of females among
hospice and non-hospice residents (83% vs. 85%).
Marital status. Only one of the ten studies examined marital status and found hospice
residents were more likely to be married (OR=1.16, 95% CI [1.08, 1.25]) compared to nonhospice residents (Kwak et al., 2008).
23

Race/ethnicity. Kwak et al.’s (2008) found that for non-Hispanic residents, those who
received hospice care were more likely to be Caucasians (OR=1.25, 95% CI [1.16, 1.37]).
Kiely et al. (2010) conducted a study of 323 residents with advanced dementia and found
non-white residents were more likely to receive hospice care. Frahm, Brown, and Hyer
(2012) found that compared to Caucasian residents, Asian residents were less likely to
receive hospice care (OR=0.59, 95%CI [0.51, 0.68]) whereas Hispanic residents were more
likely to receive hospice care (OR=1.21, 95%CI [1.12, 1.31]).
Education. Kwak et al. (2008) found hospice and non-hospice residents were similar
in years of education (11.0 vs. 10.7). Lepore et al. (2011) also found there was no significant
difference in the proportion of finishing high school or over between hospice and nonhospice residents (60% vs. 59%).
Advance directives. Kiely et al. (2010) found hospice and non-hospice residents were
similar in prevalence of do-not-hospitalize orders (68% vs. 64%). Lepore et al. (2011), Zheng
et al. (2011), and Vulpen (2013) found hospice residents were more likely to have do-nothospitalize orders and do-not-resuscitate orders compared to non-hospice residents.
Enabling characteristics. Enabling characteristics that were examined among the 10
studies were urban-rural status, income, for-profit ownership, and attitude.
Urban-rural status. Two of the 10 studies examined the relationship between urbanrural status and hospice use and both found nursing homes in rural counties were less likely
to have hospice services (Kwak, 2008; Petrisek & Mor, 1999).The definition of urban/rural
area was not clearly defined or consistent in these two studies. Petrisek and Mor’s (1999)
study used metropolitan/rural classification based on Area Resource File without clarifying
the exact definition, while Kwak et al.’s (2008) study used the Rural–Urban Commuting

24

Areas code developed for the State of Florida to recode the urban-rural area without
elucidating the exact definition.
Income. Castle (1999) found average income in the county (defined as the average
income of all persons in the county) (adjusted OR=0.54, 95% CI [0.22, 1.32]), number of
hospital beds in the county (adjusted OR = 0.41, 95% CI [0.2, 0.97]), and market competition
(measured by Herfindahl index) (adjusted OR=1.43, 95% CI [1.02, 2.40]) were associated
with the provision of hospice use in nursing homes.
For-profit ownership. Castle (1999) found for-profit ownership was associated with
higher probability of having hospice unites in nursing homes (adjusted OR= 1.2, p < 0.1), so
was and Petrisek and Mor’s (1999) study. Castle (1999) also found facility size (adjusted OR
= 1.3, p < 0.05) and having other special care units (adjusted OR = 4.75, p < 0.001) were all
significant predictors of having hospice unites in nursing homes.
Attitude. Three studies examined how nursing home staff beliefs and attitudes
influence hospice utilization (Hanson, Sengupta, & Slubicki, 2005; Nilsen et al., 2018;
Rodriquez & Boerner, 2018). These studies revealed that nursing home staff who agreed that
hospice improves quality of care and proactively approached families about hospice had
higher rates of hospice use in their facilities.
Need characteristics. Need characteristics that were examined among the 10 studies
were categorized into diseases and clinical symptoms.
Diseases. Among those that examined cancer (Kwak et al., 2008; Lepore et al., 2011;
Vulpen, 2013; Zheng et al., 2011), all studies found cancer was a predictor of hospice use
with an odds ratio of 1.96 (95% CI [1.81, 2.13]) in Zheng et al.’s (2011) study and 3.46 (95%
CI [3.09, 3.87]) in Kwak et al.’s (2008) study. Lepore et al. (2011) found that residents with
Alzheimer’s disease (13% vs. 8%) and dementia other than Alzheimer’s (24% vs. 18%) in
25

90% of Medicare/Medicaid certified urban middle-sized LTCFs (30-500 residents) were
more likely to receive hospice care, while residents with congestive heart failure (26% vs.
31%) and stoke (21% vs. 22%) were less likely to receive hospice care. Dementia (OR=1.71,
95% CI [1.58, 1.84]) and depression (OR=1.11, 95%CI [1.05, 1.18) were also found to be
predictors of hospice use in Kwak et al.’s (2008) and Zheng et al. (2011) study, respectively.
Clinical symptoms. Kiely et al. (2010) found hospice residents were more likely to
have eating problems (53% vs. 15%) compared to those who did not received hospice care.
Zheng et al. (2011) found cognitive (CPS, OR=1.03, 95%CI [1.01, 1.05]) and physical
(OR=1.02, 95%CI [1.01, 1.02) impairment, severe pain (OR=1.48, 95%CI [1.35, 1.63), and
ulcers (OR=1.16, 95%CI [1.08, 1.25]) were predictors of hospice care use. Vulpen (2013)
also found residents with indicators of emotional distress (depression, sadness, or anxiety)
were more likely to receive hospice care among residents in LTCFs. A retrospective study of
270 patients aged 65 years and older residing in two LTCFs in Saint Louis found frailty was
associated with hospice enrollment/mortality (adjusted OR=3.96, 95%CI [1.44- 10.87])
(Kaehr, Pape, Malmstrom, & Morley, 2016).
Hospice use among end-stage residents in LTCFs. An end-stage disease is defined
as having an expected survival of less than six months (Buchanan, Choi, Wang, & Ju, 2004;
Parker-Oliver, Porock, Zweig, Rantz, & Petroski, 2003). A diagnosis of end-stage disease
was found to be a strong predictors of hospice care use (OR=7.07, 95% CI [6.42, 7.79])
(Zheng et al., 2011). Eight studies examined the characteristics of end-stage residents in
LTCFs, seven of them also examined factors associated with hospice use among end-stage
residents in LTCFs (Buchanan et al., 2004; Chapin et al., 2007; Gozalo & Miller, 2007; Lin,
2010; Monroe, Carter, Feldt, Dietrich, & Cowan, 2013; Parker-Oliver et al., 2003; Zheng et
al., 2013). Smedbäck et al. (2017) conducted a study of 49,172 residents in Sweden who
26

were aged 60 and older, received palliative care (hospice care), and were at their final week
of life. This study found hospice end-stage residents who received hospice care had a mean
age of 86.7 (+7.4) years, 62% were women, the most common underlying causes of death
were circulatory diseases (42%) and dementia (23%), and the most prevalent symptoms were
pain (59%), rattles (42%), and anxiety (33%). Andersen and Newman’s (1973) theoretical
framework is used to guide and organize the following description about the findings for the
remaining seven studies which examined factors associated with hospice use among endstage residents in LTCFs.
Predisposing characteristics. Predisposing characteristics that were examined among
the seven studies were age, sex/gender, marital status, race/ethnicity, education, and advance
directives.
Age. Three studies found hospice and non-hospice end-stage residents were similar in
the value of mean age (Lin, 2010; Monroe et al., 2013; Zheng et al., 2013; Parker-Oliver et
al., 2003). Two studies found hospice end-stage residents were more likely to be younger
(Chapin et al., 2007; Gozalo & Miller, 2007), while another study found hospice end-stage
residents were more likely to be older (Buchanan et al., 2004).
Sex/gender. The majority of studies found hospice end-stage residents were more
likely to be female compared to non-hospice end-stage residents (Buchanan et al., 2004;
Gozalo & Miller, 2007; Lin, 2010; Monroe et al., 2013; Zheng et al., 2013). One study found
the two groups were similar in the proportion of females (60% vs. 59%) (Chapin et al.,
2007).
Marital status. One study found hospice end-stage residents were more likely to be
widowed (45% vs. 37%) (Buchanan et al., 2004), while one study found hospice end-stage
residents were more likely to be married (22% vs. 20%, OR=1.13, 95%CI [1.08, 1.18])
27

(Gozalo & Miller, 2007). One study found the two groups were similar in the proportion of
being married (23% vs. 21%) (Zheng et al., 2013).
Race/ethnicity. Most studies found the two groups were similar in the proportion of
Caucasians (Chapin et al., 2007; Lin, 2010; Monroe et al., 2013), while Gozalo & Miller
(2007) found hospice end-stage residents were more likely to be Caucasians (OR=1.11,
95%CI [1.03, 1.20]).
Advance directives. Buchanan et al. (2004) and Gozalo & Miller (2007) found
hospice end-stage residents were more likely to have do-not-hospitalize orders and do-notresuscitate orders. Having do-not-resuscitate orders was also found to be more likely among
hospice end-stage residents in Parker-Oliver et al.’s (2003) study (87% vs. 66%) and Zheng
et al.’s (2013) study (80% vs. 68%).
There was lack of consistency in the association between age, sex, marital status,
race, advance directive and hospice use (Appendix A). The inconsistent findings may be due
to different study populations, sample sizes and may also be affected by the impact of
geographical variations.
Enabling characteristics. One of the 8 studies examined the urban-rural status
(Chapin et al., 2007). Chapin et al. (2007) found nursing homes in rural counties were less
likely to have hospice beneficiaries (41% vs. 24%) and hospice end-stage residents were
more likely to live in an urban county compared to non-hospice end-stage residents (68% vs.
58%). In Chapin et al.’s (2007) study, an urban county was defined as a county with
population of 200,000 or more.
Need characteristics. Need factors that were examined among the seven studies were
categorized into diseases and clinical symptoms.
Diseases. Among those that examined cancer, all studies found end-stage residents
28

with cancer were more likely to receive hospice care: 67% vs. 49% in Parker-Oliver et al.’s
(2003) study, 59% vs. 52% in Buchanan et al.’s (2004) study, 47% vs. 22% in Gozalo and
Miller’s (2007) study, 41% vs. 20% in Chapin et al.’s (2007) study, and 20% vs. 14% in
Zheng et al.’s (2013) study. As for dementia, Buchanan et al. (2004) and Zheng et al. (2013)
found hospice end-stage residents were more likely to have dementia, while Gozalo and
Miller (2007) found hospice end-stage residents were less likely to have dementia (33% vs.
48%). The majority studies found end-stage residents with congestive heart failure were less
likely to receive hospice care: 14% vs. 28% in Parker-Oliver et al.’s (2003) study, 31% vs.
34% in Gozalo and Miller’s (2007) study, and 26% vs. 31% in Zheng et al.’s (2013) study.
Buchanan et al. (2004) found end-stage residents with congestive heart failure were more
likely to receive hospice care (21% vs. 20%), however, the clinical difference was small.
Renal failure was found to be less likely among hospice end-stage residents in Parker-Oliver
et al.’s (2003) study (7% vs. 16%) and Buchanan et al.’s (2004) study (7% vs. 15%).
Buchanan et al. (2004) also found hospice residents were less likely to have a diagnosis of
hypertension (36% vs. 39%), diabetes (19% vs. 23%, p<0.01), or stoke (12.6% vs. 13.0%).
However, Zheng et al. (2013) found the two groups were similar in prevalence of diabetes
(29% vs. 32%, p=0.138). The large sample size (18,211) in Buchanan et al.’s (2004) study
may cause the statistical significances even if the clinical differences were small.
Clinical symptoms. Some studies found the two groups were similar in cognitive
impairment (Parker-Oliver et al., 2003) and physical impairment (Chapin et al., 2007; ParkerOliver et al., 2003). Some studies found hospice end-stage residents were less likely to have
cognitive impairment (Gozalo and Miller, 2007; Monroe et al., 2013; Zheng et al., 2013) and
physical impairment (Gozalo and Miller, 2007). However, Buchanan et al. (2004) found
hospice end-stage residents were more likely to have greater cognitive impairment (a mean
29

CPS score of 3.1 vs. 3.4) and physical impairment (a mean ADL score of 3.7 vs. 3.8).
Besides, Parker-Oliver et al. (2003) found the two groups were similar in depressive
symptoms (a mean DRS score of 1.9 vs. 1.5), prevalence of incontinence (34% vs. 38%),
skin ulcers (25% vs. 27%), and dehydration (5% vs. 6%). Buchanan et al. (2004) found
hospice residents were more likely to have bladder incontinence (45% vs. 42%), bowel
incontinence (58% vs. 52%), and pain symptoms (74% vs. 66%).
Discussion
The studies included in this review provide a comprehensive description of the
utilization of hospice care in LTCFs, or rather in LTCFs in the U.S., which includes the
characteristics of residents who received hospice care, the trends of the characteristics, the
differences in the characteristics between residents who received hospice care and residents
who did not receive hospice care in nursing homes, the differences in the characteristics
between residents who received hospice care in nursing homes and residents who received
hospice care not in nursing homes, and the predictors of hospice use among the overall LTCF
population and among the end-stage LTCF population. In contrast, knowledge about hospice
use in Canadian LTCFs is limited, which highlights the need for studies on hospice use in
Canadian LTCFs.
The use of an appropriate theoretical framework to guide the conduct of the end-oflife research in the reviewed studies was not common: seven out of 26 of the included articles
described a theoretical framework in their studies (Han et al., 2008; Kwak et al, 2008; Lepore
et al., 2011; Lin, 2010; Nilsen et al., 2018; Unroe et al, 2013; Vulpen, 2013). Except two
studies which only used data from interviews, the other 24 studies all used secondary data
and about half of them (n=13/24) used data from the Minimum Data Set (MDS) assessment.

30

Some variations in the prevalence of hospice care use among studies were found in
the included articles. The variances may be due to differences in resident demographic and
clinical characteristics, local health care resources, health care providers’ practices, the
geographic development of hospice, and community customs and culture (Parker-Oliver et
al., 2003). Some studies reported the predictors of the hospice use, however, there is a lack of
consistency among the findings. These inconsistent findings may result from different study
designs, study populations, sample sizes, and hospice agency ownership and may also be
affected by the impact of geographical variations (Han et al., 2008). Further research should
use national-level data to address this issue and be aware of the potential impact caused by
differences within the study population and geographical variations.

31

Chapter Three: Methods
This chapter introduces the methods employed to answer the two research questions:
(a) What are the characteristics of residents who receive hospice care in LTCFs versus those
who did not by end-stage status and by one-year survivorship? (b) What variables can predict
hospice use and one-year survivorship of hospice use among LTCF residents?
Study Design
The work described in this thesis consists of a cross-sectional study (Chapter Four)
and a retrospective cohort study (Chapter Five) using health administrative data across
Canada to explore the factors that may predict the utilization of hospice care among residents
in LTCFs.
Data Sources
The work described in this thesis used routinely collected data from the Continuing
Care Reporting System (CCRS) databases (CIHI, 2018a) and Discharge Abstract Database
(DAD) (CIHI, 2018b) from CIHI. The CCRS database contains standardized, resident
demographic characteristics, clinical and functional characteristics, special treatments and
programs, admissions and discharges information on residents receiving 24-hour continuing
care services in LTCFs in Canada by using the RAI-MDS 2.0 (CIHI, 2018a). The DAD
contains data on separations from acute inpatient institutions and day surgery, chronic,
rehabilitation and psychiatric institutions (CIHI, 2018b).
Sample. The sample was drawn from residents aged 19 years or older residing in
LTCFs in Canada during Jan. 1st, 2015 and Dec 31st, 2015. During that period, all residents
who had a RAI-MDS 2.0 assessment were included. The work described in this thesis
included 185,715 residents from 1,449 LTCFs across Canada. This cohort was followed to
Dec 31st, 2016 to tract one-year survivorship. The date when the residents died during the
32

follow-up, which is between Jan 1st, 2015 and Dec 31th, 2016, was obtained from the DAD
and from the Discharge form from the RAI-MDS 2.0. The most recent RAI-MDS 2.0
assessments in 2015 and the date of death up to the end of 2016 were linked at the individual
level. Extra 5.2% residents who were discharged from LTCFs after the assessment and did
not have death records in LTCFs were tracked through the data linkage.
Data collection. Data was collected by the RAI-MDS 2.0, which was mainly
completed by trained registered nurses and multidisciplinary teams and included information
from the residents and their family (when available) and chart records. Full assessments are
required for each resident at admission, upon significant changes in status, and at least
annually (CIHI, 2015). Residents are also assessed quarterly on a subset of the full
assessment. Among the 185,715 assessments obtained from CIHI, 17.4% were admission
assessments, 22.0% were annual full assessments, 5.3% were significant change in status full
assessment, and 55.3% were quarterly review assessments.
The completed RAI-MDS 2.0 was submitted electronically on a quarterly basis to
CIHI (CIHI, 2017). The Canadian Institute for Health Information has strict quality control
over the data; submissions that do not comply with CIHI specifications are returned until the
specifications are met (CIHI, 2017). The RAI-MDS 2.0 has undergone significant reliability
and validity testing and it has been confirmed that the RAI-MDS 2.0 has both high reliability
and high validity (CIHI, 2018c). Howes et al. (1995) found the general reliability of MDS
items and sections were adequate for research purpose: 63% of the items achieved an average
intra-class correlation of 0.6 or higher; 12 of 18 MDS sections achieved an excellent average
reliability of 0.6 or higher.

33

Measures
Variables utilized in the work described in this thesis were selected based on the RAIMDS 2.0 questionnaire. Categorizations of variables were based on the previous literature
(Vulpen, 2013; Bainbridge, Seow, Sussman, & Pond, 2015; Krishnan, Thompson, &
McClement, 2017) and were adjusted depending on their frequency distribution. Based on
Andersen and Newman’s (1973) behavioral model, three determinants (predisposing
characteristics, enabling characteristics, and need characteristics) of hospice use in LTCF
residents were examined in the work described in this thesis.
Predisposing characteristics. The predisposing factors reflect resident socialdemographic characteristics. The work described in this thesis included age, sex, marital
status, language, and education (CIHI, 2015) (Table 3).
Enabling characteristics. The enabling factors reflect the resident’s ability to access
health services. The work described in this thesis included urban-rural status, geographic
region, and the QAIPPE of the facility (CIHI, 2015) (Table 3). The QAIPPE is a simple areabased socioeconomic measure of neighbourhood income quintile for the facility providing
care to the resident (Public Health Ontario, 2013).
Need characteristics. The need factors represent the resident’s need to access health
services. The work described in this thesis included disease diagnoses, cognitive function,
communication function, mood and behavior, physical function, continence, health
conditions, nutritional status, and outcome scales of the LTCF residents (CIHI, 2015) (Table
3).

34

Table 3. List of Predisposing, Enabling, and Need Characteristics Study Variables
Variables

Descriptions

Determinants 1: Predisposing Characteristics
Age
Age of the resident at the time of assessment

Sex

Sex of the resident

Marital Status

Marital status of the resident

Language

Primary language of the resident

Education

The highest completed education of the resident

Advanced Directives for Not Whether the resident’s advance directive for
Resuscitating
resuscitation is in place
Advanced Directives for Not Whether the resident’s advance directive for not
Hospitalizing
hospitalization is in place
Determinants 2: Enabling Characteristics
Urban-rural Status
Whether the facility is located in an urban area or in a
rural area
Geographic Region
Province or territory of the facility

35

Categorizations

19-64 = 1
65-74 = 2
75-84 = 3
85-94 = 4
≥ 95 = 5
Male = 0
Female = 1
Never Married = 1
Married = 2
Widowed = 3
Separated = 4
Divorced = 5
English = 1
French = 2
Other =3
8th Grade or Less = 1
9th Grade to High School = 2
Technical or Trade School or College = 3
Bachelor’s or Graduate’s Degree = 4
Not in Place = 1
In Place =2
Not in Place = 1
In Place =2
Urban = 1
Rural = 2
Alberta = 1
British Columbia = 2
Manitoba = 3
Ontario = 4
Other Provinces = 5

Variables
Quintile of Adjusted Income
per Person Equivalent
(QAIPPE)

Descriptions
Neighbourhood income quintile for the facility providing
care to the resident

Categorizations
Lowest QAIPPE = 1
Low QAIPPE =2
Medium QAIPPE = 3
High QAIPPE = 4
Highest QAIPPE = 5

Determinants 3: Need Characteristics
Disease Diagnosis
Disease Diagnosis
Cognition
Change in Cognitive Status
Cognitive Performance
Scale (CPS) (CIHI, 2018d)
Communication
Change in Communication
or Hearing
Mood and Behavior Patterns
Depression Rating Scale
(DRS) (CIHI, 2018d)
Change in the Mood
Aggressive Behavior Scale
(ABS) (CIHI, 2018d)
Change in Behavior
Symptoms

Current active diagnosis in the last 90 days

No = 0
Yes = 1

Whether resident’s cognitive status, skills, or abilities
have changed compared to the status of 90 days ago (or
since last assessment if less than 90 days)
Cognition status of the resident

Improved = 0
No Change = 1
Deteriorated = 2
No-Mild Cognitive Impairment (CPS = 0-1) = 1
Moderate Cognitive Impairment (CPS = 2-4) = 2
Severe Cognitive Impairment (CPS = 5-6) = 3

Whether resident’s ability to express, understand, or hear
information has changed compared to the status of 90
days ago (or since last assessment if less than 90 days

Improved = 0
No Change = 1
Deteriorated = 2

Whether the resident showed symptoms of depression,
sadness, or anxiety in last 30 days

No Depressive Symptom (DRS = 0) = 1
Some Depressive Symptom (DRS = 1-2) = 2
Moderate Depressive Disorder (DRS = 3-5) = 3
Severe Depressive Disorder (DRS = 6-14) = 4
Improved = 0
No Change = 1
Deteriorated = 2
No Signs of Aggression (ABS = 0) = 1
Mild- Moderate Aggression (ABS = 1-4) = 2
Severe Aggression (ABS = 5-12) = 4
Improved = 0
No Change = 1
Deteriorated = 2

Whether mood status has changed as compared to the
status of 90 days ago (or since last assessment if less than
90 days)
Aggressive behavior in the last seven days
Whether behavior status has changed compared to the
status of 90 days ago (or since last assessment if less than
90 days)

36

Variables
Psychosocial Well-Being
Index of Social Engagement
(ISE) (CIHI, 2018d)
Physical Function
Change in Activities of
Daily Living (ADL)
Function
Activities of Daily Living
Hierarchy Scale (ADL-H)
(CIHI, 2018d)
Continence
Bladder Continence

Descriptions

Categorizations

The resident’s sense of initiative and social involvement
within the facility

No-Low Social Engagement (ISE = 0-1) = 1
Moderate Social Engagement (ISE = 2-4) = 2
High Social Engagement (ISE = 5-6) = 3

Whether resident’s ADL self-performance status has
changed compared to the status of 90 days ago (or since
last assessment if less than 90 days)
Disablement process in the last seven days

Improved = 0
No Change = 1
Deteriorated = 2
No Functional Impairment (ADL-H = 0) = 1
Mild Functional Impairment ADL-H = 1-2) = 2
Moderate Functional Impairment (ADL-H = 3-4) = 3
Severe Functional Impairment (ADL-H = 5-6) = 4

Resident’s pattern of bladder continence (control) over
the last 14 days

Continent = 0
Usually Continent = 1
Occasionally Incontinent = 2
Frequently Incontinent =3
Incontinent = 4
Continent = 0
Usually Continent = 1
Occasionally Incontinent = 2
Frequently Incontinent =3
Incontinent = 4
Improved = 0
No Change = 1
Deteriorated = 2

Bowel Continence

Resident’s pattern of Bowel continence (control) over the
last 14 days

Change in urinary
continence

Whether resident’s urinary continence has changed
compared to the status of 90 days ago (or since last
assessment if less than 90 days)

Health Conditions
Pain Symptom Frequency

The frequency with which resident complains or shows
evidence of pain

Pain Symptom Intensity

The intensity of pain

Pain Scale (CIHI, 2018d)

The presence and intensity of pain in the last seven days

37

No Pain = 0
Pain Less Than Daily = 1
Pain Daily = 2
Mild Pain = 1
Moderate Pain = 2
Times When Pain is Horrible Excruciating = 3
No Pain (Pain Scale = 0) = 1
Mild-Moderate Pain (Pain Scale = 1-2) = 2
Severe Pain (Pain Scale = 3) = 3

Variables
End-Stage Disease
Nutritional Status
Weight Loss
Leave Food Uneaten
Skin Conditions
Pressure Ulcer Risk Scale
(PURS) (CIHI, 2018d)

Health Instability
Changes in Health, EndStage Disease, and Signs
and Symptoms (CHESS)
(CIHI, 2018d)
Health Complexity
interRAI Clinical
Assessment Protocols
(CAPs) (CIHI, 2008)

Descriptions
Whether the resident has 6 months or less to live

Categorizations
No = 0
Yes = 1

5% or more weight loss in the last 30 days or 10% or
more in last 180 days
Leaving 25% or more of food uneaten at most meals in
last 7 days

Not Having Weight Loss = 0
Having Weight Loss = 1
Not Having Nutritional Problems = 0
Having Nutritional Problems = 1

The risk for developing pressure ulcer

No Pressure Ulcer Risk (PURS = 0)
Mild Pressure Ulcer Risk (PURS = 1-2)
Moderate Pressure Ulcer Risk (PURS = 3)
High Pressure Ulcer Risk (PURS = 4-5)
Very High Pressure Ulcer Risk (PURS = 6-8)

Whether the resident has frailty and health instability and
is at risk of serious decline

No Indication of Health Instability (CHESS = 0)
Mild Health Instability (CHESS = 1-2)
Moderate-Severe Health Instability (CHESS = 3-5)

Number of interRAI CAPs trigged

0=1
1-5 = 2
6-10 = 3
≥11 = 4

38

Table 4. Selected List of interRAI Outcome Scales Common to the RAI-MDS 2.0 Instruments with Description and Ranges
Outcome Scales
Cognitive Performance Scale
(CPS)

Description
Measures cognitive status of the residents
• Validated against Mini-Mental State Examination
(MMSE) (R2 = 0.81)

Depression Rating Scale
(DRS)

Clinical screener for depression
• Validated against Hamilton Depression Rating Scale (r =
0.69)
Measures aggressive behavior
• Validated against Cohen Mansfield Agitation Inventory
(CMAI) (r = 0.72)
Measures residents' quality of life and level of social involvement
in activities within the long-term care facility.
• showed reasonable internal consistency (Cronbach's
alpha = 0.72) and an acceptable construct validity
(Bentler's Fit Index > 0.98)
Measures the level of impairment in the disablement process of
early, mid, and late loss of activities of daily living.
• Validated against Barthel Index (R2 = 0.74)
Measures pain presence and intensity
• Validated against the Visual Analogue Scale
(κ = 0.71)
Measures pressure ulcer risk
• Verified to be a good predictor of pressure ulcer risk (c
statistics = 0.708) and validated against the Braden Scale
(r = 0.66)
Measure of health instability as a clinical outcome and predictor
of mortality
• Verified to be a good predictor of mortality
(p <0.0001)

Aggressive Behavior Scale
(ABS)
Index of Social Engagement
(ISE) Scale

Activities of Daily Living
Hierarchy Scale (ADL-H)
Pain Scale
Pressure Ulcer Risk Scale
(PURS)
Changes in Health End Stage
Signs and Symptoms
(CHESS) Scale

39

Range
Range from 0 to 6
• 0 indicates no cognitive impairment
(cognitively intact)
• 6 indicates severe cognitive impairment
Range from 0 to 14
• Score greater than 3 indicates possible
depression
Range from 0 to 12
• 0 indicates no signs of aggression
• 12 indicates very severe aggression
Range from 0 to 6
• 0 indicates no/low social engagement
• 6 indicates high social engagement
Range from 0 to 6
• 0 indicates no functional impairment
• 6 indicates severe functional impairment
Range from 0 to 4
• 0 indicates no pain
• 4 indicates severe pain
Range from 0 to 8
• 0 indicates no risk for pressure ulcers
• 8 indicates severe risk for pressure ulcers
Range from 0 to 5
• 0 indicates no health instability
• 5 indicates very high health instability

Table 5. Variables Used to Measure the Outcomes of Residents
Outcome
Use of Hospice Care

One-Year Survivorship

Descriptions
Whether the LTCF residents received hospice care in last
14 days
Whether the residents died or were alive one year
following the last assessment in 2015

40

Categorizations
No = 0
Yes = 1
Died within one year (1-365) = 1
Alive one year following the last assessment (≥ 366) = 2

Bladder continence is coded into five categories (Table 3) (CIHI, 2015). Continent is
defined as complete control; Usually Continent is defined as bladder incontinent episodes
occur once a week or less; Occasionally Incontinent is defined as bladder incontinent
episodes occur two or more times a week but not daily; Frequently Incontinent is defined as
bladder incontinent episodes tend to occur daily, but some control is present (e.g., on day
shift); and Incontinent is defined as having inadequate control and bladder incontinent
episodes occur multiple times daily.
Bowel continence is coded into five categories. Continent is defined as complete
control; Usually Continent is defined as bowel incontinent episodes occur less than once a
week; Occasionally Incontinent is defined as bowel incontinent episodes occur once a week;
Frequently Incontinent is defined as bowel incontinent episodes occur two to three times per
week; and Incontinent is defined as having inadequate control and bowel incontinent is all (or
almost all) of the time.
Disease Diagnoses include cancer, congestive heart failure, artery heart disease,
cerebrovascular accident (stroke), dementia other than Alzheimer’s disease, Alzheimer’s
disease, rental failure, emphysema (CIHI, 2015). Each of these disease diagnoses is coded as
No = 0, Yes = 1 (Table 3).
Outcome scales. The RAI-MDS 2.0 has seven outcome scales: the Cognitive
Performance Scale (CPS) (Morris et al., 1994), the Depression Rating Scale (DRS) (Burrows,
Morris, Simon, Hirdes, & Phillips, 2000), the Index of Social Engagement (ISE) (Mor et al,
1995), the Activities of Daily Living Self-Performance Hierarchy Scale (ADL) (Morris et al.,
1994), the Changes in Health, End-Stage Disease, and Signs and Symptoms (CHESS)
(Hirdes, Frijters, & Teare, 2003), the Aggressive Behavior Scale (ABS) (Perlman & Hirdes,
2008), the Pain Scale (Fries, Simon, Morris, Flodstrom, & Bookstein, 2001), and the Pressure
41

Ulcer Risk Scale (PURS) (Poss et al., 2010) (CIHI, 2018d). The RAI-MDS 2.0 outcome
scales are embedded in the assessments, which are standardized composite variables made
out of several individual items. The categorizations of these outcome scales were based on
the RAI-MDS 2.0 Outcome Scales Reference Guide and previous studies (CIHI, 2018d;
Hirdes, Mitchell, Maxwell, & White, 2011) (Table 3–4).
The CPS consists of five cognitive performance scale items with scores ranging from
0 (cognitively intact) to 6 (very severe impairment) and describes the cognitive status of a
resident. The CPS has been validated against the Mini-Mental State Examination (MMSE) in
the identification of cognitive impairment with the sensitivity of 94% and the specificity of
94% (Hartmaier et al., 1995). The CPS has also shown good reliability against the MMSE
(R2 = 0.81) (Landi et al., 2000).
The DRS, which is made out of seven depression rating scale items ranging from 0
to14 with a score of 3 or more representing potential depression, is used as a clinical screener
for identifying whether the resident showed symptoms of depression, sadness, or anxiety in
last 30 days. The DRS has been validated against the Hamilton Depression Rating Scale (r =
0.69) (Burrows et al., 2000).
The ABS, which consists of four behaviour scale items ranging from 0 (no signs of
aggression) to12 (very severe aggression), provides a measure of aggressive behavior in the
last seven days. The ABS has shown good reliability (Cronbach's alpha = 0.80) and was
significantly correlated with Cohen Mansfield Agitation Inventory (CMAI) data (r = 0.72, p
< 0.001) (Perlman & Hirdes, 2008).
The ISE, made from six social engagement items, ranges from 0 to 6 with higher
scores indicating eater levels of engagement and is used to describe the resident’s sense of
initiative and social involvement within the facility. The ISE has shown an approximately
42

good reliability (Cronbach's alpha = 0.72) (Lou, Chi, Kwan, & Leung, 2013) and an
acceptable construct validity (Bentler's Fit Index > 0.98) (Mor et al, 1995).
The ADL-H, which is made out of four physical function items ranging from 0 to 6
with higher scores indicating more severe functional impairment, reflects the disablement
process in the last seven days by grouping ADL performance levels into a discrete stage of
loss. The ADL-H has also shown acceptable reliability against the Barthel Index (R2 = 0.74)
(Landi et al., 2000).
The pain scale, which consists of two pain scale items ranging from 0 (no pain) to 3
(severe pain), summarizes the presence and intensity of pain in the last seven days. The pain
scale has been validated against the Visual Analogue Scale showing strong inter-rater
reliability (κ = 0.71) (Fries et al., 2001).
The PURS consists of seven pressure ulcer risk scale items ranging from 0 to 8 with
higher scores indicating higher risk for pressure ulcers. The PURS has been verified to be a
good predictor of new pressure ulcers (c statistics = 0.708) (Poss et al., 2010) and has been
validated against the Braden Scale (Spearman’s r = -0.66, p < 0.01) (Carreau, Niezgoda,
Trainor, Parent, & Woodbury, 2014).
The CHESS consists of nine health instability items ranging from 0 to 5 with higher
scores indicating more severe health instability and detects frailty and health instability and is
designed to identify persons at risk of serious decline. Hirdes et al. (2003) found that CHESS
was a strong predictor of mortality (p <0.0001) and provided a useful new MDS-based test to
measure instability in health as a clinical outcome.
interRAI Clinical Assessment Protocols (CAPs). The interRAI CAPs (embedded
within the RAI-MDS 2.0) focus on “a person’s function and quality of life, assessing the
person’s needs, strengths and preferences” (CIHI, 2008, p. 1). There are 19 CAPs, two for
43

functional performance (activities of daily living and physical restraints), five for cognition
and mental health (cognitive loss, delirium, communication, mood, behaviour), two for social
life (activities, social relationship), and ten for clinical issues (falls, pain, pressure ulcer,
cardio-respiratory conditions, undernutrition, dehydration, feeding tube, appropriate
medications, urinary incontinence, bowel conditions) (CIHI, 2008). Trigged CAPs are used
to identify person-specific need and inform clinicians of priority issues requiring
improvement (CIHI, 2008; Freeman et al., 2014). In the work described in this thesis, the
number of triggered CAPs was used to identify complexity of clinical needs (Freeman et al.,
2014).
Outcomes. The outcome variables were use of hospice care and one-year
survivorship of hospice use following the last assessment (Table 5).
Data Analysis
Analyses were conducted using Statistical Analysis Software (SAS) for Windows
Version 9.4, Cary, NC: SAS Institute Inc. Descriptive statistics, t-test and chi-square test
were used depending on variable type to examine and compare predisposing, enabling, and
need characteristics of residents in LTCFs. Logistic regression models were constructed to
examine which variables were prognostic for use of hospice care and one-year survivorship
following the last assessment. All statistical tests were based on two-sided probability and an
alpha of 0.05 or less were used to indicate statistical significance (Tabachnick & Fidell,
2012).
The central limit theorem indicates that the sampling distribution of the sampling
means approaches a normal distribution as the sample size gets larger (over 30), no matter
what the shape of the population distribution is (Pett, 2015). Required sample size (N)
depends on the desired power, alpha level, number of predictors, and expected effect sizes.
44

The formula commonly used is N ≥ 50 + 8M for testing the multiple correlations and N ≥ 104
+ M for testing individual predictors with M being the number of independent variables
(Tabachnick & Fidell, 2012). The large sample size in the work described in this thesis met
the criteria of at least 234 samples to test both multiple correlations and individual predictors
with a medium-size relationship between the independent variables and the dependent
variables (i.e., hospice use and one-year survivorship of hospice use following the last
assessment) (α = 0.05 and β = 0.20).
Data cleaning. Data screening included issues about the accuracy of the data file,
missing data, detecting outliers, correlations between variables and assumptions the
descriptive analysis, t-test, chi-square test, and logistical regression analyses were based on
(Tabachnick & Fidell, 2012). Screening for accuracy involved examination of univariate
descriptive statistics (mean, range and standard deviation) and graphic representations
(histograms and scatterplots of the variables). Dealing with the problem of missing data were
based on the amount and distribution of missing data. Plots and standardized scores were
used to detect and evaluate the univariate outliers and Mahalanobis distance with a
significance level of 0.001 were examined to detect the multivariate outliers (Tabachnick &
Fidell, 2012). Normality of the variables were assessed using kurtosis and skewness statistics
and histograms. Assumptions of multicollinearity was assessed by variance inflation factors
(VIF), tolerance, and condition index. If VIF value exceeded 4.0, tolerance was less than 0.2,
or condition index exceeded 30.0, multicollinearity was thought to be present (Hair et al.,
2010). Assumptions of linearity in the logit were assessed using SAS Interactive Data
Analysis.
Univariate analyses. Univariate analyses including frequency, mean, standard
deviation, and percentage were conducted on the variables from resident predisposing,
45

enabling, and need characteristics. These analyses were utilized to describe the characteristics
of the residents and facilities participating in the work described in this thesis.
Bivariate analyses. T-test and chi-square tests (depending on the variable type) were
conducted to compare the characteristics of residents who received hospice care versus those
who did not receive hospice care. Comparisons of categorical variables were done using the
chi-square test, and comparison of continuous variables were done using t-tests.
Multivariate analyse. With mixed independent variables and a dichotomous
independent variable, binary logistical regression analyse were chosen to explore the possible
factors that predict the use of hospice services within the LTCFs. Binary logistic regression
analyse were conducted with all predictors entering the equation simultaneously because
there are no specific hypotheses about the order or importance of predictor variables
(Tabachnick & Fidell, 2012). This method allows evaluation of the contribution made by
each predictor variable over and above that of the other predictor variables. In other words,
each predictor variable is evaluated as if it entered the equation last.
Multinomial logistic regression analyses were conducted to explain how the
residents’ characteristics influence their one-year survivorship of hospice use with residents
who did not receive hospice care and were alive one year following their last assessment as
the reference group. This method allows prediction of the probabilities of different possible
outcomes of a categorically distributed outcome variable, given a set of predictor variables.
Complete case analyses were applied for predictive modelling.
Ethical Considerations
Data access is granted and monitored by CIHI. Data is de-identified prior to analysis.
Ethics approvals for this project was not required. This was confirmed by the University of
Northern British Columbia’s Research Ethics Board (Appendix B).
46

Chapter Four: Hospice Use Among Residents in Long-Term Care Facilities in Canada
Objective: The aim of the study introduced in this chapter is to describe the characteristics of
residents who receive hospice care, to compare the characteristics of residents who receive
hospice care with those who do not receive hospice care, to compare the characteristics of
residents who are noted as end-stage and receive hospice care with those who are noted as
end-stage but do not receive hospice care, and to build a predictive model of hospice use
among residents in LTCFs in Canada.
Methods: The last RAI-MDS 2.0 assessments of all unique residents (N = 185,715)
submitted to CIHI in 2015 in Canada were included. At stage 1, descriptive univariate
analyses were conducted to describe the study population. At stage 2, bivariate analyses were
performed to compare residents in hospice care with other residents not in hospice care. At
stage 3, multivariate analyses (binary logistic regression analyses) were conducted to build a
predictive model of hospice use among residents in LTCFs in Canada.
Results: Of all residents, 2.7% received hospice care (n = 4,973) and 6.8% were profiled as
having an end-stage disease (n = 12,684). Residents who received hospice care had more
severe and complex clinical needs, most were noted as end-stage (89.5%), and had severe
physical impairment (ADL-H ≥ 5, 74.3%), mild to severe pain (Pain Scale ≥ 1, 76.0%), and
moderate to severe health instability (CHESS ≥ 3, 82.9%).
Conclusion: Only a small proportion of residents in LTCFs received hospice care. Residents
who received hospice care had more severe and complex clinical needs compared to those
who did not. The findings may help LTCFs administrators, hospice care providers, and
policy makers to identify ways to implement services and interventions that can improve
access to, and utilization of, hospice care in LTCFs in Canada.

47

Introduction
It is well known that the Canadian population is ageing. About one in five Canadians
were 65 years or older in 2016 (Statistics Canada, 2017a). It was estimated that about one in
four Canadians will be 65 years or older by 2031 (Carstairs, 2010). The increasing ageing
population and improvements in life expectancy have resulted in an increase in the number of
older adults (aged 65 years and older) with chronic and complex diseases (Carstairs, 2010).
Many older adults experience impaired mobility, suffer serious cognitive and physical
impairments, and have unstable and complex health needs (Canadian Union of Public
Employees, 2009; Carstairs, 2010). Thus, there is an increasing need for formal care services,
such as LTCFs, to provide care to these people (Hirdes et al., 2011; Jayaraman & Joseph,
2013).
In Canada, a LTCF is a care institution that “serves diverse populations who need
access to 24-hour nursing care, personal care and other therapeutic and support services that
are not provided through home care programs, in retirement homes, in the persons own home
or in a shared private home” (Freeman et al., 2017, p. 2). Long-term care facilities in Canada
are governed by provincial and territorial legislation (Health Canada, 2004; McGregor &
Ronald, 2011), consequently, there is variability in provision of services and cost coverage
among LTCFs across provinces and territories (Health Canada, 2004). Long-term care
facilities in Canada are partially insured as extended health care service, and service delivery
is provided by a mix of public not-for-profit (government-owned), private not-for-profit, and
private for-profit providers (CHA, 2009; Centre for Health Services and Policy Research,
2015; McGregor, & Ronald, 2011). In recent years, the length of resident stay has dropped in
many provinces and territories largely due to the enhancement of community support
services, the expansion of home care services, and advances in technology (CHA, 2009). As
48

a result, admission to LTCFs is often delayed until the persons are nearing the end of their
lives (CHA, 2009).
In Canada, approximately 150,000 residents live in LTCFs (Statistics Canada, 2015a).
Most of these residents stay in LTCFs until death, making the LTCF a common place of
death for frail residents with chronic and complex diseases in Canada (Carstairs, 2010;
Hirdes et al., 2011; Jayaraman & Joseph, 2013). It has been estimated that about one in four
Canadians die in LTCFs each year and this number is expected to rise steadily (Brink &
Kelley, 2015). It is predicted that about 39% of Canadians will die in LTCFs by 2020
(Jayaraman & Joseph, 2013). However, resource allocations do not always correspond to the
highly complex needs of residents in LTCFs (Hirdes et al., 2011). Given the increasing
number of deaths that occur in LTCFs, it is important to examine hospice care practices in
LTCFs.
According to CHPCA (2016a), hospice care is defined as “a special kind of health
care for individuals and families who are living with a life-limiting illness that is usually at
an advanced stage” (para. 2). Hospice care can reduce or relieve physical and psychological
symptoms, provide comfort and dignity for the person living with the illness as well as the
best quality of life for both this person and his or her family (CHPCA, 2016a). While a large
proportion of Canadians die in LTCFs each year, most LTCFs lack a formalized hospice care
program (Quality Palliative Care in Long Term Care Alliance, 2011). Research has emerged
studying hospice care across various settings (e.g., at hospitals and at private homes),
however, a particular gap in knowledge exists regarding hospice care use in LTCF setting,
particularly across Canada (Erel, Marcus, & Dekeyser-Ganz, 2017; Ersek & Carpenter, 2013;
Thompson, Bott, Boyle, Gajewski, & Tilden, 2010). To the knowledge of the author, there is
no national Canadian study addressing this issue. As such, this study utilized a large national
49

data set to describe residents who received hospice care in LTCFs and explore factors that
can predict hospice use in LTCFs across Canada. This study was guided by the following
four main questions: (a) What are the characteristics of residents who receive hospice care in
LTCFs? (b) What are the differences in characteristics between residents who receive
hospice care and who do not receive hospice care in LTCFs? (c) What are the differences in
characteristics between end-stage residents who receive hospice care and who do not receive
hospice care in LTCFs? and (d) What variables can predict hospice use among LTCF
residents?
Conceptual framework. Andersen and Newman’s (1973) behavioral model has been
widely used for studies on health services utilization (Lo & Fulda, 2008; Li, Nong, Wei,
Feng, & Luo, 2016). This model assists in explaining and predicting demographic and
societal determinants for utilization of health services and developing policies to promote
equitable access to these services (Anderson & Newman, 1973). The behavioral model
proposed that the utilization of health services was a result of three components of population
characteristics: (a) the predisposition of the person to use services (predisposing
characteristics), (b) the person’s ability to secure the services (enabling resources), and (c)
the person’s illness level (needs) (Andersen and Newman, 1973).
Anderson (1995) proposed four types of access to health services: potential access,
realized access, equitable access, and inequitable access. Potential access is defined as the
presence of enabling resources. More enabling resources relate to the increasing likelihood of
using health services. Realized access is the actual use of the health services. Equitable
access occurs “when demographic and need factors account for most of the variance in
utilization”, while inequitable access occurs “when the social structure, health beliefs, and
enabling resources determine who gets health services” (Anderson, 1995, p. 4-5).
50

This model was also used in this study to assist in understanding and predicting
social-demographic and health determinants for utilization of hospice care services and
evaluating hospice care practices in LTCFs in Canada.
Methods
Study design and data sources. This population-based cross-sectional study used
health administrative data from CIHI. The Canadian Institute for Health Information has
strict quality control over the data; submissions that do not comply with CIHI specifications
are returned until the specifications are met (CIHI, 2017). The last RAI-MDS 2.0
assessments of all unique residents (N=185,715) submitted to CIHI in 2015 throughout
Canada were included in this study.
The RAI-MDS 2.0 was completed by trained registered nurses and multidisciplinary
teams through direct observation over all shifts prior to the RAI-MDS 2.0 assessments and
also included information from the resident and their family (when available) and chart
records (CIHI, 2017). Completed assessments were submitted electronically on a quarterly
basis to CIHI (CIHI, 2017). Full assessments are required for each resident at admission,
upon significant changes in status, and at least annually (CIHI, 2015). Residents are also
assessed quarterly on a subset of the full assessment. Among the 185,715 assessments
obtained from CIHI, 17.4% were admission assessments, 22.0% were annual full
assessments, 5.3% were significant change in status full assessment, and 55.3% were
quarterly review assessments.
Measures. The RAI-MDS 2.0 contains standardized and comprehensive information
of residents receiving 24-hour continuing care services in LTCFs in Canada. Based on
Andersen and Newman’s (1973) behavioral model, three determinants of population
characteristics for hospice use were examined in this study. The three determinants included
51

predisposing characteristics (e.g. age, gender, and marital status), enabling characteristics
(e.g. urban-rural status, neighborhood income quantile), and need characteristics (e.g.
cognition, mood, behavior, and physical function) (Figure 2). The RAI-MDS 2.0 also
includes information on special treatments and programs (e.g. hospice care), admissions (e.g.
admission date and type), and discharges (e.g. discharge date and type).
The resident is profiled as having an end-stage disease when his/her physician
believes that the resident has only six or fewer months to live (CIHI, 2015). This judgement
is required to be substantiated by a well-documented disease diagnosis and deteriorating
clinical course (CIHI, 2015). The RAI-MDS 2.0 defines hospice care as a program for
terminally ill persons where services are necessary for the palliation and management of
terminal illness and related conditions.
Outcome Scales. There were eight widely used outcome scales that were derived
from the items in the RAI-MDS 2.0: the Cognitive Performance Scale (CPS) (Morris et al.,
1994), the Depression Rating Scale (DRS) (Burrows et al., 2000), the Index of Social
Engagement (ISE) (Mor et al, 1995), the Activities of Daily Living Self-Performance
Hierarchy Scale (ADL) (Morris et al., 1994), the Changes in Health, End-Stage Disease, and
Signs and Symptoms (CHESS) (Hirdes et al., 2003), the Aggressive Behavior Scale (ABS)
(Perlman & Hirdes, 2008), the Pain Scale (Fries et al., 2001), and the Pressure Ulcer Risk
Scale (PURS) (Poss et al., 2010) (CIHI, 2018d). The categorizations of these outcome scales
are based on the RAI-MDS 2.0 Outcome Scales Reference Guide and previous studies (CIHI,
2018d; Hirdes et al., 2011) (Table 4).
Clinical Assessment Protocols (CAPs). The CAPs (embedded within the RAI-MDS
2.0) focus on “a person’s function and quality of life, assessing the person’s needs, strengths
and preferences” (CIHI, 2008, p. 1). There are 19 CAPs, two for functional performance
52

(activities of daily living and physical restraints), five for cognition and mental health
(cognitive loss, delirium, communication, mood, behaviour), two for social life (activities,
social relationship), and ten for clinical issues (falls, pain, pressure ulcer, cardio-respiratory
conditions, undernutrition, dehydration, feeding tube, appropriate medications, urinary
incontinence, bowel conditions) (CIHI, 2008). Trigged CAPs are used to identify personspecific need and inform clinicians of priority issues requiring improvement (CIHI, 2008;
Freeman et al., 2014).
Statistical analyses. Analyses were conducted using Statistical Analysis Software
(SAS institute, Inc., Cary, NC). At stage 1, descriptive univariate analyses including means,
medians, standard deviations, and percentages were used to describe the characteristics of
residents and LTCFs participating in the study and the residents who received hospice care in
LTCFs. At stage 2, bivariate analyses including t-test and chi-square test were performed
depending on variable type to test for statistically significant differences in predisposing,
enabling, and need characteristics between residents in hospice care to other residents not in
hospice care and between end-stage residents in hospice care to other end-stage residents not
in hospice care. At stage 3, multivariate analyses including binary logistic regression
analyses and ROC analysis were conducted to build a predictive model of hospice use among
residents in LTCFs in Canada. Residents who had missing values were removed during
predictive model building. The variables that were tested in the need factors model were
chosen based on the correlation issues between variables and multicollinearity tests. To avoid
correlation issues between variables, variables that were included in the outcome scales were
not considered for the predictive model building. Multicollinearity was measured by variance
inflation factors (VIF), tolerance, and condition index. If VIF value exceeded 4.0, tolerance
was less than 0.2, or condition index exceeded 30.0, multicollinearity was thought to be
53

present (Hair et al., 2010). All statistical tests will be based on two-sided probability and an
alpha of 0.05 or less was used to indicate statistical significance (Tabachnick & Fidell, 2012).
Ethical considerations. Data access was granted and monitored by CIHI. Data was
de-identified prior to analysis. Ethics approvals for this project was not required, which was
confirmed by the University of Northern British Columbia’s Research Ethics Board
(Appendix B).
Results
This study included 185,715 LTCF residents in Canada in 2015. Residents in LTCFs
in Canada had a mean age of 83.0 years (±11.4 years). Most residents were aged over 85
years (54.3%), were female (65.8%), and lived in an urban LTCF (85.2%). Of all residents,
2.7% received hospice care (n = 4,973) and 6.8% were profiled as having an end-stage
disease (n = 12,684). Of all end-stage residents, 35.1% received hospice care (n =
4,449/12,684). Among residents who died in 2015,11.8% of residents received hospice care
at the time of last assessment (n = 4009/33,908).
More than half of residents who received hospice care were aged below 85 years
(57.1%) and female (58.2%) (Table 6). The majority of residents lived in an urban LTCF
(93.9%). Most residents had a diagnosis of cancer (57.5%), were profiled as having an endstage disease (89.5%), and had moderate to severe cognitive impairment (CPS ≥ 2, 78.1%),
severe physical impairment (ADL-H ≥ 5, 74.3%), mild to severe pain (Pain Scale ≥ 1,
75.9%), moderate to very high pressure ulcer risk (PURS ≥ 3, 83.7%), and moderate to
severe health instability (CHESS ≥3, 82.9%).
Significant differences in predisposing, enabling, and need characteristics between
residents in hospice care to other residents not in hospice care and between end-stage
residents in hospice care to other end-stage residents not in hospice care are presented in
54

Table 6 and Figure 4–5.
Predisposing characteristics. Compared to residents who did not receive hospice
care, residents who received hospice care had a higher percentage of being younger than 75
years old (30.2% vs. 18.3%), male (41.8% vs. 34.0%), and married (35.9% vs. 23.7%), and
had education higher than high school (31.7% vs. 27.4%), a DNR order (89.8% vs. 80.4%),
and a DNH order (62.2% vs. 35.0%). However, a lower percentage of end-stage residents
who received hospice care had a DNR order (89.3% vs. 93.3%) compared to those who did
not receive hospice care.
Enabling characteristics. A higher percentage of residents who received hospice
care than those who did not lived in an urban LTCF (93.9% vs. 85.0%), high-income area
(QAIPPE ≥ 4, 38.1% vs. 34.1%), and Ontario (93.3% vs. 61.9%). However, a slightly lower
percentage of end-stage residents who received hospice care lived in a high-income area
(QAIPPE ≥ 4, 37.5% vs. 39.2%), compared to those who did not receive hospice care.
Need characteristics. Compared to residents who did not receive hospice care,
residents who received hospice care had a much higher percentage of having a diagnosis of
cancer (57.3% vs. 9.4%) but a lower percentage of having a diagnosis of dementia (27.1% vs.
49.1%) (Figure 4). Residents who received hospice care had more severe and complex
clinical needs than those who did not with regard to severe cognitive impairment (CPS ≥ 5,
33.9% vs. 21.9%), potential depressive disorder (DRS ≥ 3, 33.3% vs. 28.6%), severe physical
impairment (ADL-H ≥ 5, 74.3% vs. 33.6%), bowel incontinence (52.7% vs. 29.7%), pain
(Pain Scale ≥ 1, 75.9% vs. 38.5%), high to very high pressure ulcer risk (PURS ≥ 4, 56.0%
vs. 15.9%), moderate to severe health instability (CHESS ≥ 3, 82.9% vs. 10.4%), and the
number of CAPs triggered (a mean of 6.6 vs. 5.4). A lower percentage of residents who
received hospice care had severe aggressive behaviors (ABS ≥ 5, 4.6% vs. 8.2%) and high
55

social engagement (ISE ≥ 5, 9.0% vs. 20.5%) compared to residents who did not receive
hospice care.
Unlike differences between hospice users and non-users among all residents, a lower
percentage of end-stage residents who received hospice care than those who did not had
severe cognitive impairment (CPS ≥ 5, 35.1% vs. 44.3%), potential depressive disorder (DRS
≥ 3, 33.0% vs. 36.0%), bowel incontinence (53.2% vs. 56.6%), and bladder incontinence
(37.0% vs. 53.9%). Besides, end-stage residents who received hospice care had a higher
percentage of having high social engagement (ISE ≥ 5, 8.5% vs. 6.8%) compared to those
who did not receive hospice care.
Predictive models. The Model 1 that contained predisposing factors (i.e. age, sex,
marital status, language, education, DNR order, and DNH order) was significantly different
from an intercept-only model, Wald χ2(7, N = 95,667) = 582.0, p < 0.0001, AIC = 14,683.5,
R2max-rescaled = 0.0503, c statistics = 0.694 (Table 7). The Model 2 run with only enabling
factors (i.e. province, rural-urban status, and QAIPPE) against an intercept-only model was
statistically significant, Wald χ2(3, N = 184,557) = 1,883.3, p < 0.0001, AIC = 42,320.9,
R2max-rescaled = 0.0818, c statistics = 0.726. The Model 3 that included selected need factors
(i.e. diagnosis of dementia but not Alzheimer’s disease, diagnosis of cancer, CPS, DRS,
ABS, ISE, ADL_H, Pain Scale, and CHESS) against an intercept-only model was
statistically significant, Wald χ2(9, N = 184,411) = 11,132.1, p < 0.0001, AIC = 25,063.6,
R2max-rescaled = 0.4664, c statistics = 0.943.
The Model 4, the final model, which included predisposing, enabling, and need
factors (i.e. age, rural-urban status, QAIPPE, diagnosis of dementia but not Alzheimer’s
disease, diagnosis of cancer, ABS, ISE, ADL_H, Pain Scale, and CHESS), against an
intercept-only model was statistically significant, Wald χ2(10, N = 138,048) = 6779.2, p <
56

0.0001, AIC = 15,819.2, R2max-rescaled = 0.4472, c statistics = 0.944 (Table 8). The results of
Hosmer-Lemeshow test (χ2= 6.6, df = 8, p = 0.58) indicated a good model fit.
The ROC analysis showed there was no significant difference between the
predisposing factors model (Model 1) and the enabling factors model (Model 2) (p=0.75).
The need factors model (Model 3) was stronger than both the predisposing factors model
(Model 1) (p<0.001) and the enabling factors model (Model 2) (p<0.001), but was not as
strong as the final model (Model 4) (p=0.003).
The final predictive model showed living in an urban LTCF and high-income area
and having a diagnosis of cancer, severe functional impairment, mild to severe pain, and mild
to severe health instability significantly increased the likelihood of receiving hospice care
(OR > 1, p < 0.05; Table 8), while being older, having a diagnosis of dementia (not
Alzheimer’s disease), signs of aggression, and moderate to high social engagement were less
likely to receive hospice care (OR < 1, p <0.05).
Discussion
Although LTCFs have become a major location of death, hospice care was highly
underutilized among dying LTCF residents. Less than 3% of residents received hospice care
in LTCFs. Among those who were profiled as having an end-stage disease, about one third
received hospice care. According to the Canadian Virtual Hospice (2019), the general
hospice eligibility criteria in Canada are “1) a specific prognosis (often six months or less), 2)
a decision to focus on comfort rather than cure, and 3) an acceptance that resuscitation will
not be used when the illness brings a natural death”. This study found the 91.9% of end-stage
residents had a DNR order indicating about nine in ten of residents may be eligible to hospice
care. The underutilization of hospice care is substantial and requires immediate attention.

57

Policy makers should consider to allocate more resources and investments in hospice care
services in LTCFs.
Residents in LTCFs were often under-recognized as having a prognosis of six months
or less, as it is hard to identify terminal stages when the person has several medical problems
but no specific terminal diagnosis (CHPCA, 2009). This study found among those who died
with a DNR order in place, one in nine received hospice care before death. Hospice use rate
among residents who may have benefitted from receiving hospice care was much lower than
those who were profiled as having end-stage disease. This suggests that accuracy of endstage diagnosis and effectiveness of hospice eligibility criteria were poor. The current criteria
of referral to hospice care may limit the benefits to residents in terms of improvements in
end-of-life care, particularly for residents with several medical problems but no specific
terminal diagnosis (Bennett, Ziegler, Allsop, Daniel, & Hurlow, 2016). Standardized
assessment of terminal status or more effective hospice eligibility criteria with less
dependency on a specific life expectancy of six months or less is needed to improve
accessibility to, and utilization of, hospice care resources.
Residents who received hospice care were more likely to have a diagnosis of cancer,
which was consistent with previous studies (Buchanan et al., 2004; Chapin et al., 2007;
Gozalo & Miller, 2007; Kwak et al., 2008; Lepore et al., 2011; Parker-Oliver et al., 2003;
Zheng et al., 2011; Zheng et al.; 2013; Vulpen, 2013). However, residents with dementia
were less likely to access hospice care. This may be partly because of difficulty in prediction
of life expectancy for persons with dementia and lack of recognition of the terminal nature of
dementia, lack of advance directives in place among persons with dementia, and lack of
recognition of benefits from hospice care for persons with dementia (Erel et al., 2017;
Mitchell et al., 2010; Sachs, Shega, & Cox-Hayley, 2004). It is important for health care
58

providers and policy-makers to improve recognition of dementia as a terminal disease and
increase understanding of the role of hospice services for persons with dementia.
This study also found residents who received hospice care had more severe and
complex clinical needs, which was consistent with previous studies (Vulpen, 2013; Zheng et
al., 2011) and highlights a demand on LTCFs to provide complex care. Efforts should be
made to improve access to hospice care among residents with dementia or other progressive
chronic diseases with severe and complex clinical needs, so these residents can have
improved access to hospice care as residents with end-stage cancer. Lynn (2004) suggests
that “people who would benefit from hospice palliative and end-of-life care now fall into
three categories: a) People who have an overwhelming illness and go through a rapid,
predictable decline to death (e.g., end-stage cancers); b) People who have a chronic illness
who may live for many years but may at any time experience some exacerbation of
symptoms that leads to sudden death (e.g., people with congestive heart failure, chronic
obstructive pulmonary disease, cirrhosis, kidney failure); and c) People who experience
increasing frailty and neurological disorders (e.g., people with Alzheimer’s Disease) and
dwindle over time” (CHPCA, 2009), which may be good guidance to help identify people
who would benefit from hospice care.
Pain was common among LTCF residents, especially among those with end-stage
disease. Although end-stage residents who received hospice care had the same intense pain
compared to those who did not receive hospice care, their pain were more frequent than those
who did not receive hospice care. However, residents who received hospice care had the
same frequent and intense pain no matter whether were noted as having end-stage disease
indicating the pain may be part of the reason why residents without end-stage diagnosis
received hospice care. As this is a cross-sectional study, casual effect between hospice use
59

and pain cannot be determined. It was not clear whether the same intensity of pain among
end-stage residents was a result of effective pain control from hospice care or long-term care
or a result of inadequate pain assessment for persons at end of life (Herr, Coyne, McCaffery,
Manworren, & Merkel, 2011).
Interestingly, this study indicates end-stage residents who received hospice care had
less severe health conditions in some aspects than those who did not receive hospice care.
Compared to those who did not received hospice care, they were less likely to have low
social engagement, severe cognitive impairment, potential depressive disorder, and bowel
and bladder incontinence. End-stage residents who received hospice care had deteriorated
changes in cognitive function, mood, and urinary continence in last three months. Therefore,
it can be inferred that end-stage residents who received hospice care had less severe cognitive
impairment, lower depressive symptoms, and better bowel and bladder control. More
frequent pain, as discussed earlier, may be one of the reasons as residents with severe
cognitive impairment may not be able to self report pain (Herr et al., 2011; Miu, & Chan,
2014) and thus were more likely to receive hospice care. Stigma attached to residents with
severe cognitive impairment and potential depressive disorder may be part of the reason for
this. Cognitive impairment is often mistaken as an untreatable natural process of aging
(Graham et al., 2003; Lebowitz & Niederehe, 1992) and residents with cognitive impairment
may not benefit from hospice care (Garand, Lingler, Conner, & Dew, 2009). Stigma
associated with the diagnostic labels of depressive disorder and cognitive impairment can
have a negative impact on interpersonal relationships, interactions with the LTCF staff, and
attitudes about hospice service utilization.
The majority of residents triggered at least one CAP, while those who received
hospice care triggered more CAPs. The CAPs “identify areas in which a resident has a higher
60

than expected rate of decline, an increased potential to improve, and symptoms that could be
alleviated if a problem was addressed” (interRAI, 2017). Evidence of CAP triggering may be
useful to predict health complexity (Freeman et al., 2014). The high number of CAPs
triggered among residents who received hospice care reflects high levels of complex clinical
needs within this group. Clinical assessment protocols triggered at high rates such as
delirium, pressure ulcer, bowel incontinence, and pain warrant increased attention for the
majority of end-stage residents. A previous study found treatment of symptoms for residents
who are at end of life can still improve their quality of life (Lorenz et al., 2008).
Consideration of triggered CAPs in this study provide evidence to inform a collaborative
decision making process from LTCF staff and hospice providers on whether (or not) and how
issues raised by the CAPs should be addressed in the plan of care.
This study found the utilization of hospice services was a result of the predisposition
of the residents to use hospice services, the resident’s ability to secure the hospice services
and the resident’s needs to use hospice services, which were consistent with prior research
(Gozalo & Miller, 2007; Han et al., 2008; Zheng et al., 2013). The realized access (the actual
use of the health services) to hospice care is low in LTCFs. Although residents in an urban
LTCF and higher income neighborhood are more likely to receive hospice care, need factors
showed the strongest predictive value on hospice use indicating all LTCF resident may have
poor access to hospice care services. As the population ages, there will be increasing need to
provide hospice care for the growing number of residents in LTCFs. Since access to hospice
care in LTCFs is poor, there is a great and immediate need to improve hospice care for the
LTCF population.
This study found CHESS was a good predictor of hospice use in LTCFs, consistent
with previous findings that CHESS was a strong predictor of death (Hirdes et al., 2003). The
61

predictive model developed in this study was superior than CHESS alone, as clinical needs
were not the only determinants of hospice use. The variables in the predictive model
examined in this study were common in the suite of interRAI assessment instruments. As a
result, persons receiving care in different service settings can be compared and evaluated
directly with equivalent measures (Hirdes et al., 2003) in regard to hospice admission
eligibility. Further investigation of hospice use in different settings such as the person’s
home, hospitals, and hospice facilities is needed and may help hospice providers identify the
unique needs of LTCF residents enrolled in hospice care and tailor the care to be provided
within the LTCF population.
This study found 10.7% of end-stage residents who had a DNR orders received
hospice care. It would be interesting to see further study investigate the characteristics of
residents who receive hospice care but do not meet hospice eligibility criteria and residents
who receive hospice care for an unexpected long-term period, which will inform better
resources allocation for hospice care services. As accuracy of end-stage diagnosis continues
to be challenging and criteria for hospice eligibility are restrictive, further longitudinal study
is needed to explore standardized assessment of terminal status or more effective hospice
eligibility criteria with less dependency on six-month prognosis. Reasons why end-stage
residents with higher social engagement, less severe cognitive impairment, and lower
depressive symptoms are more likely to receive hospice care and why end-stage residents
have the same intense pain no matter whether they receive hospice care need further
investigation.
Strengths and limitations. This is the first Canadian study to use the RAI-MDS 2.0
data to examine the LTCF population and a subpopulation of LTCF residents who receive
hospice care at the national level. The Andersen and Newman’s (1973) behavioral model for
62

health service utilization used to guide this study, which helped to better understand hospice
use among the LTCF population in Canada. This study used a mix of admission assessments,
quarterly assessments, and annual full assessments, so the study population covered all stages
in LTCFs in Canada.
This study also has several limitations. First, while the CCRS database contains full
coverage for Ontario, British Columbia, Alberta, Yukon, Saskatchewan, and Newfoundland
and Labrador, it has partial coverage for Manitoba, New Brunswick and Nova Scotia, and
does not cover Quebec, Prince Edward Island, the Northwest Territories, and Nunavut.
Second, residents who received continuing care services in hospitals were mixed in the
sample and cannot be singled out, therefore, the study population in this study also include
individuals receiving continuing care services at 107 hospital-based complex continuing care
facilities in Ontario and Manitoba in addition to 1,342 LTCFs across Canada (CIHI, 2016).
Third, hospice use may be underestimated as some resident who did not receive hospice care
at the time of their last assessment in 2015 may have received hospice care after the
assessment.
Conclusion
The underutilization of hospice care in LTCFs is substantial and requires immediate
attention. Residents who received hospice care exhibited more severe and complex clinical
needs, such as pain, severe physical impairment, and high health instability than those who
did not receive hospice care. End-stage residents who received hospice care exhibited high
social engagement and were less likely to have severe cognitive impairment, depressive
symptoms, bowel and bladder incontinence than those who did not receive hospice care.
Further study is needed to identify the underlying reason. Further investigation of
standardized assessment of terminal status is also needed as accuracy of end-stage diagnosis
63

continues to be challenging and criteria for hospice eligibility are narrow. Special attention
should be paid to improve access to hospice care among residents with dementia or other
progressive chronic diseases with severe and complex clinical needs.

64

Table 6. Predisposing, Enabling, and Need Characteristics by Status of Hospice Use of Residents (N=185,715) and End-Stage
Residents (N=12,684) in Long-Term Care Facilities in Canada in 2015
All Residents (N=185,715)

Variables

Determinants 1: Predisposing Characteristics
Age Group (in years)
19-64
65-74
75-84
85-94
95+
Female Sex
Marital Status
Never Married
Married
Widowed
Separated/Divorced
Primary Language Spoken
English
French
Other
Education
8th Grade or Less
9th Grade-High School
Technical/Trade School/College
Bachelor's or Higher Degree
Advance Directives-Do Not Resuscitate
Advance Directives-Do Not Hospitalize
Determinants 2: Enabling Characteristics
Province
Alberta
British Columbia
Ontario
Other Provinces
Urban Facility
Facility Neighbourhood Income Quintile
Lowest (QAIPPE = 1)
Low (QAIPPE = 2)
Medium (QAIPPE = 3)
High (QAIPPE = 4)
Highest (QAIPPE = 5)

End-Stage Residents (N=12,684)

All residents
(N=185,715)

Did not receive
hospice care
(97.3%,
n=180,742)

Received
hospice care
(2.7%,
n=4,973)

7.6 (14,055)
11.0 (20,509)
27.2 (50,483)
43.9 (81,435)
10.4 (19,233)
65.8 (122,145)

7.4 (13,364)
10.9 (19,701)
27.1 (49,137)
44.1 (79,736)
10.4 (18,804)
66.1 (119,252)

13.9 (691)
16.3 (808)
27.1 (1,346)
34.3 (1,699)
8.6 (4,29)
58.2 (2,893)

10.3 (16,647)
27.5 (44,546)
51.6 (83,583)
10.7 (17,354)

10.3 (16,198)
27.1 (42,759)
51.9 (81,749)
10.7 (16,963)

10.1 (449)
40.1 (1,787)
41.1 (1,834)
8.8 (391)

83.6 (155,309)
2.5 (4,560)
13.9 (25,846)

83.6 (151,039)
2.4 (4,357)
14.0 (25,346)

85.9 (4,270)
4.1 (203)
10.1 (500)

28.5 (33,527)
44.0 (36,767)
16.7 (19,666)
10.7 (12,596)
89.8 (137,001)
35.7 (62,244)

28.6 (33,035)
44.1 (50,950)
16.7 (19,312)
10.7 (12,348)
80.4 (137,001)
35.0 (59,463)

25.9 (492)
42.3 (756)
18.6 (354)
13.1 (248)
89.8 (4,071)
62.2 (2,781)

10.0 (18,505)
16.3 (30,288)
62.8 (116,552)
11.0 (20,370)
85.2 (158,246)

10.1 (18,333)
16.7 (30,189)
61.9 (111,910)
11.2 (20,319)
85.0 (153,575)

3.5 (172)
2.0 (99)
93.3 (4,642)
0.8 (60)
93.9 (4,671)

26.1 (48,213)
19.1 (35,280)
20.4 (37,655)
18.9 (3,4912)
15.4 (28,497)

26.4 (47,439)
18.9 (33,917)
20.5 (36,726)
18.9 (33,905)
15.3 (27,613)

15.6 (774)
27.5 (1,363)
18.7 (929)
20.3 (1,007)
17.8 (884)

65

P
Value

<0.0001

<0.0001
<0.0001

<0.0001
<0.001

<0.0001
<0.0001
<0.0001

<0.0001
<0.0001

All residents
at end stage
(N=12,684)

Did not receive
hospice care
(64.9%,
n=8,235)

Received
hospice care
(35.1%,
n=4,449)

8.1 (1,028)
11.5 (1,455)
25.6 (3,245)
42.8 (5,432)
12.0 (1,524)
61.4 (7,782)

4.9 (399)
8.7 (718)
24.6 (2,025)
47.8 (3,932)
14.1 (1,161)
63.1 (5,192)

14.1 (629)
16.6 (737)
27.4 (1,220)
33.7 (1,500)
8.2 (363)
58.2 (2,590)

7.9 (903)
33.5 (3,812)
50.1 (5,703)
8.6 (976)

6.8 (501)
29.5 (2,186)
55.3 (4,091)
8.4 (623)

10.1 (402)
40.7 (1,626)
40.4 (1,612)
8.8 (353)

87.0 (11,039)
2.8 (350)
10.2 (1,295)

87.4 (7,193)
2.1 (170)
10.6 (872)

86.5 (3,846)
4.1 (180)
9.5 (423)

26.1 (1,736)
44.6 (2,969)
17.8 (1,185)
11.6 (773)
91.9 (10,885)
60.6 (7,111)

26.3 (1,318)
45.2 (2,268)
17.4 (875)
11.1 (559)
93.3 (7,239)
59.7 (4,599)

25.4 (418)
42.7 (701)
18.9 (310)
13.0 (214)
89.3 (3,647)
62.4 (2,512)

6.1 (770)
5.8 (734)
84.0 (10,653)
4.2 (527)
87.3 (11,074)

7.4 (613)
8.0 (660)
78.5 (6,464)
6.1 (498)
83.7 (6,892)

3.5 (157)
1.7 (74)
94.2 (4,189)
0.7 (29)
94.0 (4,182)

19.8 (2,497)
21.9 (2,761)
19.8 (2,507)
22.1 (2,790)
16.5 (2,084)

22.2 (1,819)
18.1 (1,486)
20.6 (1,686)
23.6 (1,936)
15.6 (1,277)

15.2 (678)
28.8 (1,275)
18.5 (821)
19.3 (854)
18.2 (807)

P
Value

<0.0001

<0.0001
<0.0001

<0.0001

0.06

<0.0001
<0.01
<0.0001

<0.0001
<0.0001

Determinants 3: Need Characteristics
Change in Cognitive Status
Improved
No Change
Deteriorated
Cognition
No-Mild Cognitive Impairment (CPS = 0-1)
Moderate Cognitive Impairment (CPS = 2-4)
Severe Cognitive Impairment (CPS = 5-6)
Change in Communication
Improved
No Change
Deteriorated
Change in Mood
Improved
No Change
Deteriorated
Depression
No Depressive Symptoms (DRS = 0)
Some Depressive Symptoms (DRS = 1-2)
Mild Depressive Disorder (DRS = 3-5)
Moderate-Severe Depressive Disorder (DRS = 6-14)
Change in Behavior Symptom
Improved
No Change
Deteriorated
Aggressive Behavior
No Signs of Aggression (ABS = 0)
Mild-Moderate Aggression (ABS = 1-4)
Severe Aggression (ABS = 5-12)
Social Engagement
No-Low Social Engagement (ISE = 0-1)
Moderate Social Engagement (ISE = 2-4)
High Social Engagement (ISE = 5-6)
Change in ADL Function
Improved
No Change
Deteriorated
Physical Function
No Functional Impairment (ADL-H = 0)
Mild Functional Impairment (ADL-H = 1-2)
Moderate Functional Impairment (ADL-H = 3-4)
Severe Functional Impairment (ADL-H = 5-6)

1.8 (3,349)
85.9 (159,103)
12.3 (22,683)

1.8 (3,294)
87.1 (157,072)
11.1 (19,980)

1.2 (55)
42.4 (2,031)
56.4 (2,703)

22.3 (41,329)
55.5 (103,097)
22.2 (41,289)

22.3 (40,241)
55.8 (100,900)
21.9 (39,601)

21.9 (1,088)
44.2 (2,197)
33.9 (1,688)

0.7 (1,212)
91.9 (170,042)
7.5 (13,881)

0.7 (1,190)
92.6 (167,044)
6.7 (12,112)

0.5 (22)
62.6 (2,998)
36.9 (1,769)

5.1 (9,465)
82.4 (152,576)
12.5 (23,094)

5.2 (9,290)
83.2 (150,097)
11.6 (20,959)

3.7 (175)
51.8 (2,479)
44.6 (2,135)

41.0 (75,838)
30.3 (56,163)
19.8 (36,669)
8.9 (16,465)

41.2 (74,204)
30.3 (54,599)
19.7 (35,493)
8.9 (16,050)

34.1 (1,634)
32.7 (1,564)
24.6 (1,176)
8.7 (415)

4.4 (8,181)
85.0 (157,437)
10.5 (19,517)

4.4 (7,973)
85.5 (154,164)
10.1 (18,209)

4.3 (208)
68.3 (3,273)
27.3 (1,308)

58.9 (108,962)
33.0 (61,091)
8.2 (15,082)

58.7 (10,5939)
33.0 (59,546)
8.2 (14,861)

63.1 (3,023)
32.3 (1,545)
4.6 (221)

27.3 (50,606)
52.5 (97,572)
20.2 (37,537)

26.5 (47,948)
53.0 (95,706)
20.5 (37,088)

53.5 (2,658)
37.5 (1,866)
9.0 (449)

4.5 (8,382)
74.6 (138,594)
20.9 (38,739)

4.6 (8,314)
76.1 (137,480)
19.3 (34,948)

1.4 (68)
22.4 (1,114)
76.2 (3,791)

4.2 (7,774)
16.7 (31,052)
44.4 (82,424)
34.7 (64,465)

4.3 (7,744)
17.0 (30,766)
45.1 (81,461)
33.6 (60,771)

0.6 (30)
5.8 (286)
19.4 (963)
74.3 (3,694)

66

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

1.0 (120)
52.6 (6,473)
46.2 (5,712)

0.9 (74)
59.4 (4,775)
39.7 (3,188)

1.1 (46)
39.8 (1,698)
59.1 (2,524)

16.0 (2,031)
42.9 (5,442)
41.1 (5,211)

13.1 (1,080)
42.6 (3,507)
44.3 (3,648)

21.4 (951)
43.5 (1,935)
35.1 (1,563)

0.4 (47)
66.1 (8,139)
32.5 (4,119)

0.3 (27)
69.0 (5,549)
30.6 (2,461)

0.5 (20)
60.6 (2,590)
38.9 (1,658)

4.7 (576)
60.3 (7,417)
35.0 (4,312)

5.2 (420)
65.7 (5,279)
29.1 (2,338)

3.7 (156)
50.1 (2,138)
46.3 (1,974)

31.7 (3,902)
33.3 (4,103)
23.7 (2,911)
11.3 (1,389)

30.7 (2,467)
33.3 (2,679)
23.1 (1,858)
12.9 (1,033)

33.6 (1,435)
33.4 (1,424)
24.7 (1,053)
8.3 (356)

5.6 (685)
70.0 (8,618)
24.4 (3,002)

6.3 (503)
71.3 (5,726)
22.5 (1,808)

4.3 (183)
65.0 (2,892)
28.0 (1,194)

56.3 (6,926)
36.2 (4,458)
7.5 (921)

52.8 (4,242)
38.2 (3,072)
9.0 (723)

62.9 (2,684)
32.5 (1,386)
4.6 (198)

54.6 (6,919)
38.1 (4,833)
7.4 (932)

54.7 (4,503)
38.6 (3,176)
6.8 (556)

54.3 (2,416)
37.2 (1,657)
8.5 (376)

1.3 (163)
32.7 (4,147)
66.0 (8,374)

1.4 (118)
39.6 (3,261)
59.0 (4,856)

1.0 (45)
19.9 (886)
79.1 (3,518)

0.6 (78)
5.3 (671)
23.4 (2,969)
70.7 (8,966)

0.7 (56)
5.2 (431)
26.0 (2,138)
68.1 (5,610)

0.5 (22)
5.4 (240)
18.7 (831)
75.4 (3,356)

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

Bowel Continence
Continent
Usually Continent
Occasionally continent
Frequently Continent
Incontinent
Bladder Continence
Continent
Usually Continent
Occasionally continent
Frequently Continent
Incontinent
Change in Urinary Continence
Improved
No Change
Deteriorated
Pain Symptoms Frequency
No Pain
Pain Less Than Daily
Pain Daily
Pain Symptoms Intensity
Mild Pain
Moderate Pain
Severe Pain
Pain
No Pain (Pain Scale = 0)
Mild-Moderate Pain (Pain Scale = 1-2)
Severe Pain (Pain Scale = 3)
Weight Loss
Leaves Food Uneaten
Pressure Ulcer Risk
No Pressure Ulcer Risk (PURS = 0)
Mild Pressure Ulcer Risk (PURS = 1-2)
Moderate Pressure Ulcer Risk (PURS = 3)
High Pressure Ulcer Risk (PURS = 4-5)
Very High Pressure Ulcer Risk (PURS = 6-8)
Health Instability
No Indication of Health Instability (CHESS = 0)
Mild Health Instability (CHESS = 1-2)
Moderate-Severe Health Instability (CHESS = 3-5)

34.9 (64,768)
11.8 (21,915)
8.0 (14,876)
15.0 (27,873)
30.3 (27,873)

35.3 (63,824)
11.9 (21,457)
8.1 (14,586)
15.1 (27,213)
29.7 (53,662)

19.0 (944)
9.2 (458)
5.8 (290)
13.3 (660)
52.7 (2,621)

20.7 (38,469)
8.6 (15,941)
9.9 (18,317)
24.5 (45,548)
36.3 (67,440)

20.3 (36,709)
8.6 (15,614)
9.9 (17,944)
24.8 (44,890)
36.3 (65,585)

35.4 (1,760)
6.6 (327)
7.5 (373)
13.2 (658)
37.3 (1,855)

2.8 (5,199)
85.8 (159,362)
11.4 (21,154)

2.8 (4,975)
86.7 (156,737)
10.5 (19,030)

4.5 (224)
52.8 (2,625)
42.7 (2,124)

60.5 (112,389)
24.8 (45,980)
14.7 (27,346)

61.5 (111,193)
24.4 (44,093)
14.1 (25,456)

24.1 (1,196)
37.9 (1,887)
38.0 (1,890)

45.7 (33,523)
46.1 (33,791)
8.2 (6,012)

46.8 (32,515)
45.5 (31,661)
7.7 (5,373)

26.7 (1,008)
56.4 (2,130)
16.9 (639)

60.5 (112,389)
37.2 (68,997)
2.3 (4,329)
11.0 (18,973)
33.3 (61,836)

61.5 (111,193)
36.4 (65,759)
2.1 (3,790)
10.4 (17,557)
32.2 (58,161)

24.1 (1,196)
65.1 (3,238)
10.8 (539)
28.5 (1,416)
73.9 (3,675)

19.0 (34,861)
33.1 (60,827)
31.0 (57,005)
15.4 (28,401)
1.6 (2,847)

19.4 (34,721)
33.6 (60,112)
31.1 (55,682)
14.5 (25,949)
1.4 (2,524)

2.8 (140)
14.4 (715)
26.7 (1,323)
49.5 (2,452)
6.5 (323)

40.7 (75,639)
47.0 (87,227)
12.3 (22,849)

41.8 (75,549)
47.8 (86,467)
10.4 (18,726)

1.8 (90)
15.3 (760)
82.9 (4,123)

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001
<0.0001
<0.0001

<0.0001

16.1 (2,042)
8.0 (1,019)
5.4 (689)
15.0 (1,904)
55.4 (7,030)

15.8 (1,210)
7.3 (603)
5.5 (449)
15.9 (1,309)
56.6 (4,664)

18.7 (832)
9.4 (416)
5.4 (240)
13.4 (595)
53.2 (2,366)

22.8 (2,893)
5.4 (678)
6.9 (873)
17.0 (2,153)
48.0 (6,087)

15.8 (1,298)
4.7 (387)
6.5 (531)
19.2 (1,579)
53.9 (4,440)

35.9 (1,595)
6.5 (291)
7.7 (342)
12.9 (574)
37.0 (1,647)

3.3 (414)
61.3 (5,494)
35.5 (4,499)

2.7 (224)
66.7 (5,494)
30.6 (2,517)

4.3 (190)
51.2 (2,277)
44.6 (1,982)

36.7 (4,652)
31.6 (4,005)
31.8 (4,027)

44.3 (3,646)
27.7 (2,282)
28.0 (2,307)

22.6 (1,006)
38.7 (1,723)
38.7 (1,720)

27.5 (2,208)
55.7 (4,474)
16.8 (1,350)

28.2 (1,295)
55.1 (2,527)
16.7 (767)

26.5 (913)
56.6 (1,947)
16.9 (583)

36.7 (4,652)
54.7 (6,935)
8.7 (1,097)
31.7 (3,424)
71.4 (9,057)

44.3 (3,646)
48.4 (3,985)
7.3 (604)
28.0 (2,131)
68.7 (5,654)

22.6 (1,006)
66.3 (2,950)
11.1 (493)
40.3 (1,293)
76.5 (3,403)

2.8 (351)
14.7 (1,851)
30.8 (3,892)
45.6 (5,760)
6.2 (778)

3.0 (246)
15.1 (1,235)
33.3 (2,728)
42.7 (3,503)
6.0 (488)

2.4 (105)
13.9 (616)
26.3 (1,164)
50.9 (2,257)
6.5 (290)

0.0
22.3 (2,826)
77.7 (9,858)

0.0
27.8 (2,293)
72.2 (5,942)

0.0
12.0 (533)
88.0 (3,916)

<0.0001

<0.0001

<0.0001

<0.0001

0.23

<0.0001

<0.0001
<0.0001
<0.0001

<0.0001

Note. Other provinces include Manitoba, Saskatchewan, Newfoundland and Labrador, Nova Scotia, Yukon.
QAIPPE denotes Quintile of Adjusted Income per Person Equivalent; CPS denotes Cognitive Performance Scale; DRS denotes Depression Rating Scale; ABS
denotes Aggressive Behavior Scale; ISE denotes Index of Social Engagement; ADL-H denotes Activities of Daily Living Hierarchy; PURS denotes Pressure
Ulcer Risk Scale; CHESS denotes Changes in Health, End-Stage Disease, Signs, and Symptoms Scale.

67

Table 7. Logistic Regression Model Predicting Hospice Use Among Residents in Long-Term Care Facilities in Canada in 2015
Variables

OR (95% CI)

p-Value

0.73 (0.57-0.93)
0.54 (0.43-0.67)
0.45 (0.36-0.57)
0.54 (0.42-0.70)

<0.01
<0.0001
<0.0001
<0.0001

0.81 (0.72-0.91)

<0.001

0.72 (0.59-0.89)
0.91 (0.80-1.04)
0.59 (0.48-0.72)

<0.01
0.18
<0.0001

1.81 (1.39-2.34)
1.19 (1.01-1.39))

<.0001
0.04

1.00 (0.88-1.14)
1.29 (1.10-1.50)
1.19 (1.00-1.43)

0.99
<0.01
0.06

4.20 (3.19-5.54)

<0.0001

2.47 (2.21-2.76)

<0.0001

2.37 (2.11-2.67)

<0.0001

2.28 (2.09-2.50)
1.33 (1.21-1.47)
1.55 (1.41-1.70)
1.65 (1.50-1.82)

<0.0001
<0.0001
<0.0001
<0.0001

0.24 (0.21-0.28)
0.08 (0.06-0.10)
0.08 (0.06-0.11)

<0.0001
<0.0001
<0.0001

Model 1: Predisposing Factor Model (c statistics = 0.694)
Age Group (in years, ref = 19-64)
65-74
75-84
85-94
95+
Sex (ref = Male)
Female
Marital Status (ref = Married)
Never Married
Widowed
Separated/Divorced
Language (ref = English)
French
Other
Education (ref = 8th Grade or Less)
9th Grade-High School
Technical/Trade School/College
Bachelor's or Higher Degree
Do Not Resuscitate Order (ref = Not in Place)
In Place
Do Not Hospitalize Order (ref = Not in Place)
In Place
Model 2: Enabling Factor Model (c statistics = 0.726)
Urban vs. Rural
Facility Neighbourhood Income Quintile (ref = Lowest [QAIPPE = 1])
Low (QAIPPE = 2)
Medium (QAIPPE = 3)
High (QAIPPE = 4)
Highest (QAIPPE = 5)
Province (ref = Ontario)
Alberta
British Columbia
Other Provinces

68

Model 3: Need Factor Model (c statistics = 0.943)
Dementia Not Alzheimer’s (ref = No)
Yes
0.51 (0.48-0.56)
<0.0001
Cancer (ref = No)
Yes
6.94 (6.47-7.45)
<0.0001
Cognition (ref = No-Mild Cognitive Impairment [CPS = 0-1])
Moderate Cognitive Impairment (CPS = 2-4)
0.88 (0.80-0.97)
<0.01
Severe Cognitive Impairment (CPS = 5-6)
1.17 (1.04-1.31)
<0.01
Depression (ref = No Depressive Symptoms [DRS = 0])
Some Depressive Symptoms (DRS = 1-2)
0.83 (0.76-0.90)
<0.0001
Mild Depressive Disorder (DRS = 3-5)
0.88 (0.90-0.97)
0.01
Moderate-Severe Depressive Disorder (DRS = 6-14)
0.58 (0.51-0.67)
<0.0001
Aggressive Behavior (ref = No Signs of Aggression [ABS = 0])
Mild to Moderate Aggression (ABS = 1-4)
0.83 (0.77-0.90)
<0.0001
Severe Aggression (ABS = 5-12)
0.50 (0.43-0.59)
<0.0001
Social Engagement (ref = No to Low Social Engagement [ISE = 0-1])
Moderate Social Engagement (ISE = 2-4)
0.66 (0.61-0.71)
<0.0001
High Social Engagement (ISE = 5-6)
0.51 (0.45-0.58)
<0.0001
Physical Function (ref = No Functional Impairment [ADL-H = 0])
Mild Functional Impairment (ADL-H = 1-2)
1.15 (0.77-1.71)
0.50
Moderate Functional Impairment (ADL-H = 3-4)
1.16 (0.79-1.71)
0.46
Severe Functional Impairment (ADL-H = 5-6)
2.77 (1.88-4.07)
<0.0001
Pain (ref = No Pain [Pain Scale = 0])
Mild-Moderate Pain (Pain Scale = 1-2)
2.05 (2.05-1.89)
<0.0001
Severe Pain (Pain Scale = 3)
2.26 (1.97-2.59)
<0.0001
Health Instability (ref = No Indication of Health Instability [CHESS = 0])
Mild Health Instability (CHESS = 1-2)
5.74 (4.60-7.15)
<0.0001
Moderate-Severe Health Instability (CHESS = 3-5)
75.39 (60.86-93.40)
<0.0001
Note. Other provinces include Manitoba, Saskatchewan, Newfoundland and Labrador, Nova Scotia, Yukon.
QAIPPE denotes Quintile of Adjusted Income per Person Equivalent; CPS denotes Cognitive Performance Scale; DRS denotes Depression Rating Scale; ABS
denotes Aggressive Behavior Scale; ISE denotes Index of Social Engagement; ADL-H denotes Activities of Daily Living Hierarchy; PURS denotes Pressure
Ulcer Risk Scale; CHESS denotes Changes in Health, End-Stage Disease, Signs, and Symptoms Scale.

69

Table 8. Final Logistic Regression Model Predicting Hospice Use Among Residents in Long-Term Care Facilities in Canada in 2015
B

Variables

OR (95% CI)

p-Value

Model 4 (Final Model): Predisposing, Enabling, and Need Factor Model (c statistics = 0.944)
Intercept
-6.70
<0.0001
Age Group (ref=19-64)
65-74
-0.22
0.80 (0.66-0.98)
0.03
75-84
-0.60
0.55 (0.45-0.66)
<0.0001
85-94
-0.80
0.45 (0.37-0.54)
<0.0001
95+
-0.72
0.49 (0.39-0.61)
<0.0001
Urban-Rural Status (ref = rural)
Urban
0.94
2.55 (2.24-2.90)
<0.0001
Facility Neighbourhood Income Quintile (ref = Lowest [QAIPPE = 1])
Low (QAIPPE = 2)
0.62
1.86 (1.67-2.07)
<0.0001
Medium (QAIPPE = 3)
0.16
1.17 (1.05-1.32)
<0.01
High (QAIPPE = 4)
0.59
1.81 (1.62-2.03)
<0.0001
Highest (QAIPPE = 5)
0.48
1.61 (1.44-1.81)
<0.0001
Dementia Not Alzheimer’s (ref = No)
Yes
-0.50
0.61 (0.56-0.66)
<0.0001
Cancer (ref = No)
Yes
1.87
6.48 (6.03-6.96)
<0.0001
Aggressive Behavior (ref = No Signs of Aggression [ABS = 0])
Mild to Moderate Aggression (ABS = 1-4)
-0.26
0.77 (0.71-0.83)
<0.0001
Severe Aggression (ABS = 5-12)
-0.85
0.43 (0.37-0.50)
<0.0001
Social Engagement (ref = No-Low Social Engagement [ISE = 0-1])
Moderate Social Engagement (ISE = 2-4)
-0.50
0.60 (0.56-0.65)
<0.0001
High Social Engagement (ISE = 5-6)
-0.83
0.44 (0.38-0.50)
<0.0001
Physical Function (ref = No Functional Impairment [ADL-H = 0])
Mild Functional Impairment (ADL-H = 1-2)
0.12
1.12 (0.75-1.69)
0.57
Moderate Functional Impairment (ADL-H = 3-4)
0.12
1.13 (0.76-1.67)
0.55
Severe Functional Impairment (ADL-H = 5-6)
1.01
2.76 (1.87-4.07)
<0.0001
Pain (ref = No Pain [Pain Scale = 0])
Mild-Moderate Pain (Pain Scale = 1-2)
0.65
1.92 (1.77-2.08)
<0.0001
Severe Pain (Pain Scale = 3)
0.68
1.98 (1.72-2.27)
<0.0001
Health Instability (ref = No Indication of Health Instability [CHESS = 0])
Mild Health Instability (CHESS = 1-2)
1.79
5.99 (4.90-7.47)
<0.0001
Moderate-Severe Health Instability (CHESS = 3-5)
4.33
76.31 (61.52-94.64)
<0.0001
Note. QAIPPE denotes Quintile of Adjusted Income per Person Equivalent; ABS denotes Aggressive Behavior Scale; ISE denotes Index of Social Engagement;

70

ADL-H denotes Activities of Daily Living Hierarchy; PURS denotes Pressure Ulcer Risk Scale; CHESS denotes Changes in Health, End-Stage Disease, Signs,
and Symptoms Scale.

71

Figure 4. Selected Common Diagnoses by Status of Hospice Use of Residents (N=185,715)
and End-Stage Residents (N=12,684) in Long-Term Care Facilities in Canada in 2015
All residents (N=185,715)
Cancer
Renal Failure

Arterial Heart Disease
Congestive Heart Failure
Alzheimer’s Disease
Emphysema
Cerebrovascular Accident
Diabetes Mellitus
Dementia Not Alzheimer’s Disease
0.0%

25.0%

Received hospice care

50.0%

75.0%

100.0%

75.0%

100.0%

Did not receive hospice care

End-stage residents (N=12,684)
Cancer
Renal Failure
Arterial Heart Disease
Alzheimer’s Disease
Congestive Heart Failure
Cerebrovascular Accident
Emphysema
Diabetes Mellitus

Dementia Not Alzheimer’s Disease
0.0%
Received hospice care

25.0%

50.0%

Did not received hospice care

72

Figure 5. Number of Clinical Protocol Assessments Triggered by Status of Hospice Use
Among Residents (N=185,715) and End-Stage Residents (N=12,684) in Long-Term Care
Facilities in Canada in 2015
All residents (N=185,715)

25.0%

20.0%

15.0%

10.0%

5.0%

0.0%

0

1

2

3

4

5

6

7

8

Received hospice care

9

10

11

12

13

14

15

16

14

15

16

Did not receive hospice care

End-stage residents (N=12,684)
25.0%

20.0%

15.0%

10.0%

5.0%

0.0%

0

1

2

3

4

5

6

7

Received hospice care

8

9

10

11

12

13

Did not receive hospice care

73

Chapter Five: An Examination of the Characteristics of Residents Who Receive Versus
Do Not Receive Hospice Care in Long-Term Care Facilities in Canada
Objective: Hospice care is designed for persons in the final phase of a terminal illness. Some
persons who did not meet the hospice eligibility received hospice care, while many persons
who may benefit from hospice care did not receive it. The aim of the study introduced in this
chapter is to examine the characteristics of, and one-year survivorship among, residents who
received hospice care versus those who did not receive hospice care in LTCFs in Canada.
Methods: This retrospective cohort study used linked health administrative data from the
CCRS and the DAD. All persons residing in LTCFs between Jan. 1st, 2015 and Dec 31st,
2015 assessed with the RAI-MDS 2.0 in CCRS database were included in this study and their
death records were linked up to Dec 31th, 2016. Bivariate and multivariate analyses were
conducted to examine differences in characteristics of residents in four groups: a) did not
receive hospice care and were alive one year following the last assessment, b) did not receive
hospice care and died within one year of assessment, c) received hospice care and were alive
one year following the last assessment, and d) received hospice care and died within one year
of assessment.
Results: The assessed hospice care rate in LTCFs is very low (i.e., less than 3%), while one
in five of the residents died within three months of the assessment. Residents who received
hospice care and died within one year of assessment had more severe and complex health
conditions than residents in the other groups. Compared to those who did not receive hospice
care but died within one year of assessment, residents who received hospice care and were
alive one year following the last assessment were younger (a mean age of 79.4 [+13.5] years
vs. 86.5 [+9.2]), had a higher percentage of living in an urban LTCF (93.2% vs. 82.6%),
higher percentage of having a diagnosis of cancer (50.7% vs. 12.9%) but a lower percentage
74

of having a diagnosis of dementia (30.2% vs. 54.5%), and exhibited more severe acute
clinical needs.
Conclusions: The actual use of hospice care among LTCF residents is very poor in Canada.
This study indicates several possible barriers to hospice use in the LTCF population
including ageism, rurality, and a diagnosis of dementia. Improving understanding of
characteristics of residents who did not access but may have benefitted from hospice care
may help LTCFs administrators, hospice care providers, and policy makers to make plans
and take actions to improve hospice accessibility in this target group.

75

Introduction
According to CHPCA, hospice care is defined as “a special kind of health care for
individuals and families who are living with a life-limiting illness that is usually at an
advanced stage” (CHPCA, 2016a, para. 2). Hospice care can reduce or relieve physical and
psychological symptoms, provide comfort and dignity for the person living with the illness as
well as the best quality of life for both this person and his or her family (CHPCA, 2016a). In
Canadian version of the RAI-MDS 2.0 assessment, hospice care is defined as “a program for
terminally ill persons where services are necessary for the palliation and management of
terminal illness and related conditions” (CIHI, 2015). The general eligibility for hospice care
program in Canada is a) the person has a specific prognosis (often six months or less), b) a
decision has been made to focus on comfort rather than cure, and c) resuscitation will not be
used when the illness brings a natural death (Canadian Virtual Hospice, 2019). However, it is
hard to recognize terminal stages when the person has several medical problems but no
specific terminal diagnosis. Hospice care in Canada has traditionally been offered only in the
last weeks or months of life (Langille, 2013). Late referrals limit the ability of health systems
to reach maximum potential for reduction or relief of suffering and healthcare cost
containment (Hawley, 2014).
Hospice care can be provided in a variety of settings, such as homes, hospitals,
LTCFs, and hospices (Health Canada, 2016). “Although hospitals are designed to address
severe and urgent needs, they may not be the best location for comfortable end-of-life care”
(Health Canada, 2016, para. 8). Providing hospice services in LTCFs is cost-efficient
compared to providing hospice care at private homes and in hospice facilities (Infeld et al.,
1990). For example, hospices may have greater revenues by increasing their resident volume,
utilizing staff more efficiently, overlapping basic services, and increasing the average length
76

of stay (Petrisek & Mor, 1999). The basic services such as environmental services,
housekeeping, and central supplies can be shared by hospice facilities and LTCFs. Long-term
care facilities who enroll residents in a hospice care program can increase their
competitiveness of the market by promising to provide hospice care to residents who are
nearing the end of life, reduce in-house staff time while providing these special services and
also benefit from the knowledge of hospice staff (Castle, 1999; Petrisek & Mor, 1999).
In Canada, a LTCF is a care institution that serves diverse populations who need
access to 24-hour nursing care, personal care and other therapeutic and support services that
are not provided elsewhere (Freeman et al., 2017) and is a common place of death for
Canadians (Carstairs, 2010; Hirdes et al., 2011; Jayaraman & Joseph, 2013). Long-term care
facilities are not included under the Canada Health Act (McGregor, & Ronald, 2011). It is
governed by provincial and territorial legislation, and jurisdictions offer a different range of
services and cost coverage across Canada (Health Canada, 2004). While the majority of
LTCFs are publicly funded, service delivery is provided by a mix of public (governmentowned), private not-for-profit, and private for-profit providers (Centre for Health Services
and Policy Research, 2015; McGregor, & Ronald, 2011).
While a large proportion of Canadians die in LTCFs each year, most LTCFs lack a
formalized hospice care program (Quality Palliative Care in Long Term Care Alliance, 2011;
Brink & Kelley, 2015). As indicated in Chapter Four, some residents who did not meet the
hospice eligibility received hospice care, while many residents who may benefit from hospice
care did not receive it. However, there is limited research examining hospice use in LTCFs in
Canada (Ersek & Carpenter, 2013). It remains unknown what the residents who receive
hospice care but survive an unexpected length of time (i.e. more than six months) look like. It
also remains to be seen what the residents who may benefit from hospice care but did not
77

receive it look like. As such, this study aimed to examine the characteristics of, and one-year
survivorship among, residents who received hospice care versus those who did not receive
hospice care in LTCFs in Canada. This study was guided by the following two main
questions: (a) What are the characteristics of LTCF residents in Canada who receive versus
who do not receive hospice care by their one-year survivorship? and (b) What variables can
predict one-year survivorship of hospice use among LTCF residents?
Conceptual framework. To better understand hospice services utilization in LTCFs
in Canada, Andersen and Newman’s (1973) behavioral model was used to guide this study.
This model was designed to assist in explaining and predicting demographic and societal
determinants for utilization of health services and identifying access disparities and other
barriers to these services (Anderson & Newman, 1973). This behavioral model was widely
used for studies on health services utilization (Lo & Fulda, 2008; Li et al., 2016). The
behavioral model indicates that the utilization of health services was a result of three
components of population characteristics: (a) the predisposition of the person to use services
(predisposing characteristics), (b) the person’s ability to secure the services (enabling
resources), and (c) the person’s illness level (needs) (Andersen and Newman, 1973). This
model was used in this study to assist in evaluating the current hospice care practices in
LTCFs in Canada and understanding how the residents’ predisposing, enabling, and need
characteristics influence their utilization of hospice care services.
Methods
Study design and data sources. This is a population-based cohort study using health
administrative data from the CCRS and the DAD from CIHI. The Canadian Institute for
Health Information is an independent and not-for-profit organization that provides essential
information on Canada’s health systems and the health of Canadians. The Canadian Institute
78

for Health Information has a comprehensive and high-standard Data and Information Quality
Program to ensure the data can be trusted by stakeholders (CIHI, 2017; CIHI, 2018a). All
persons residing in LTCFs who had a RAI-MDS 2.0 assessment in the CCRS database
between Jan. 1 st, 2015 and Dec 31st, 2015 were included in this study. Their death records
were linked through the RAI-MDS 2.0 discharge form and the DAD up to Dec 31th, 2016.
The RAI-MDS 2.0 was completed by trained clinical professionals (including
registered nurses, social workers, physicians) through direct observation over all shifts prior
to the assessments and also included information from the resident and his/her family (when
available) and chart records (CIHI, 2017). The RAI-MDS 2.0 contains standardized and
comprehensive information of residents receiving 24-hour continuing care services in LTCFs
in Canada. The full assessment of the RAI-MDS 2.0 are required for each resident at
admission, upon significant changes in status, and within one year of the last full assessment
(User’s Manual, 2015). Residents are also assessed quarterly on a subset of the full
assessment. The Discharge Abstract Database captures information on hospital discharges
(including deaths, sign-outs and transfers) directly from acute care facilities or from their
respective health/regional authority or ministry/department of health in all provinces and
territories except Quebec (CIHI, 2018b).
Measures. Based on Andersen and Newman’s (1973) behavioral model, three
determinants (predisposing characteristics, enabling characteristics, and need characteristics)
of hospice use in LTCF residents were examined in this study (Figure 2).
Predisposing characteristics. Age was measured in years and also categorized into
five age groups. Other predisposing characteristics included gender, marital status, language,
education, and advance directives (do-not-resuscitate [DNR] and do-not-hospitalize [DNH]
orders).
79

Enabling characteristics. This study included province, urban-rural status, and
QAIPPE. According to Statistics Canada (2017b), “an urban area was defined as having a
population of at least 1,000 and a density of 400 or more people per square kilometre and all
areas outside urban areas was defined as rural areas”. The QAIPPE is a simple area-based
socioeconomic measure of neighbourhood income quintile for the facility providing care to
the resident (Public Health Ontario, 2013).
Need characteristics. This study included disease diagnoses, cognitive function,
communication function, mood and behavior, physical function, continence, pain, skin
condition, nutritional status, health stability, and complexity of clinical needs.
Outcome scales. In addition to individual item in the RAI-MDS 2.0, a variety of risk
and clinical summary outcome scales that embedded in the RAI-MDS 2.0 were also used to
examine need characteristics. The outcome scales included the Cognitive Performance Scale
(CPS) (Morris et al., 1994), the Depression Rating Scale (DRS) (Burrows et al., 2000), the
Index of Social Engagement (ISE) (Mor et al, 1995), the Activities of Daily Living SelfPerformance Hierarchy Scale (ADL) (Morris et al., 1994), the Changes in Health, End-Stage
Disease, and Signs and Symptoms (CHESS) (Hirdes et al., 2003), the Aggressive Behavior
Scale (ABS) (Perlman & Hirdes, 2008), the Pain Scale (Fries et al., 2001), and the Pressure
Ulcer Risk Scale (PURS) (Poss et al., 2010) (CIHI, 2018d) (Table 4).
Clinical Assessment Protocols (CAPs). The CAPs (embedded in the RAI-MDS 2.0)
were also used to assess need characteristics. The CAPs “identify areas in which a resident
has a higher than expected rate of decline, an increased potential to improve, and symptoms
that could be alleviated if a problem was addressed” (interRAI, 2017). There are 19 CAPs
including activities of daily living, physical restraints, cognitive loss, delirium,
communication, mood, behaviour, activities, social relationship, falls, pain, pressure ulcer,
80

cardio-respiratory conditions, undernutrition, dehydration, feeding tube, appropriate
medications, urinary incontinence, and bowel conditions (CIHI, 2008). The number of
triggered CAPs was used to identify complexity of clinical needs in this study (Freeman et
al., 2014).
Statistical analyses. Analyses were conducted using SAS software for Windows
Version 9.4, Cary, NC: SAS institute, Inc. At stage 1, univariate analyses including means,
medians, standard deviations, and percentages were used to describe the characteristics of
residents. At stage 2, bivariate analyses including t-tests for continuous variables and chisquare test for categorical variables were performed to test the statistically significant
differences in characteristics of residents in four groups: a) did not receive hospice care and
were alive one year following their last assessment, b) did not receive hospice care and died
within one year of assessment, c) received hospice care and were alive one year following
their last assessment, and d) received hospice care and died within one year. At stage 3,
multinomial logistic regression analysis was conducted to explain how the residents’
characteristics influence their hospice use with residents who did not receive hospice care
and were alive one year following their last assessment as the reference group.
Variables that had more than 5% of missing values were not included in the
multinomial logistic regression analysis (Tabachnick & Fidell, 2012). For variables with less
than 5% of missing values, complete case analysis was applied for the predictive modelling.
To reduce multicollinearity, variables that were already in the outcome scales were not
considered for the predictive model building. Multicollinearity was measured by variance
inflation factors (VIF), tolerance, and condition index. Multicollinearity was thought to be
present when VIF value exceeded 4.0, tolerance was less than 0.2, or condition index
exceeded 30.0, (Hair et al., 2010). All statistical tests were based on two-sided probability
81

and an alpha of 0.05 or less was used to indicate statistical significance (Tabachnick & Fidell,
2012).
Ethical considerations. Data access was granted and monitored by CIHI. Data is deidentified prior to data analysis. Ethics approvals for this project was not required. This
decision was confirmed by the University of Northern British Columbia’s Research Ethics
Board (Appendix B).
Results
Study sample characteristics. There were 185,715 unique residents living in LTCFs
assessed with the RAI-MDS 2.0 in Canada in 2015. Residents in LTCFs in Canada had a
mean age of 83.0 years (±11.4 years). More than half of the residents were aged over 85
years (54.3%), and widowed (51.6%). Most were female (65.8%), lived in an urban LTCF
(85.2%) and Ontario (62.8%), and had a primary language of English (83.6%) (Table 9). As
shown in Figure 6, 30.9% of the residents died within one year of assessment (N =
57,398/185,715), most of whom died within three months (65.5%, N = 37,602/57,398). Of all
185,715 residents living in LTCFs in Canada in 2015, only 2.7% of the residents received
hospice care (n=4,973/185,715). Of those who receive hospice care, 88.9% died within one
year of assessment (n = 4,417/4,973), while about 10.1% were still alive one year following
their last assessment (n = 556/4,973). Of all residents, 28.5% did not receive hospice care and
died within one year (n = 52,981/185,715).
Did not receive hospice care and alive. Residents who did not receive hospice care
and were alive one year following their last assessment (n = 127,761) had a mean age of 81.7
(+11.8) years; about half were aged over 85 years (49.5%). About half of the residents in this
group were widowed (49.7%), two thirds were female (66.8%), three in four had DNR order
(76.6%), and less than one third had DNH order (30.7%). Most of these residents lived in an
82

urban LTCF (86.0%) (Table 9). A diagnosis of dementia (46.8%) was more common than a
diagnosis of cancer (7.9%) in this group (Figure 7). The majority of the residents in this
group exhibited no to moderate cognitive impairment (CPS ≤ 4, 82.5%), no to mild
depressive symptoms (DRS ≤ 2, 72.7%), no signs of aggression (ABS = 0, 60.8%), moderate
to high social engagement (ISE = 2-6, 78.6%), mild to moderate physical impairment (ADLH = 1-4, 73.1%), bowel incontinence (24.6%), bladder incontinence (31.2%), no pain (Pain
Scale = 0, 63.5%), no to moderate pressure risk (PURS ≤ 3, 89.1%), and no to mild health
instability (CHESS ≤ 2, 94.5%) (Table 9). Less than half of the residents in this group
triggered more than five CAPs (44.7%) (Figure 8); the top three triggered CAPs were
activities of daily living (84.3%), urinary incontinence (82.5%), and mood (56.7%).
Did not receive hospice care and died. Residents who did not receive hospice care
but died within one year of assessment (n = 52,981) had a mean age of 79.4 (+13.5) years;
two thirds were aged over 85 years (66.5%). More than half of them were widowed (57.4%),
less than two thirds were female (64.2%), most had DNR order (89.4%), nearly half had
DNH order (45.4%), and most lived in an urban LTCF (82.6%) (Table 9). Dementia was a
common diagnosis in this group (54.5%), but cancer was not (12.9%) (Figure 7). About one
third of the residents in this group exhibited severe cognitive impairment (CPS ≥ 5, 32.4%)
and possible depressive disorder (DRS ≥ 3, 31.7%), barely half had signs of aggression (ABS
≥ 1, 46.3%), more than one third had no to low social engagement (ISE ≤ 1, 39.0%), half had
severe physical impairment (ADL-H ≥ 5, 50.0%), barely half had completely bowel (42.0%)
incontinence and completely bladder (48.7%) incontinence, less than half had mild to severe
pain (Pain Scale ≥ 1, 43.2%), one in four had high to very high pressure risk (PURS ≥ 4,
27.9%), and one in five had moderate to severe health instability (CHESS ≥ 3, 21.9%) (Table
9). More than half of the residents in this group had more than five CAPs triggered (52.2%)
83

(Figure 8); the top three triggered CAPs were urinary incontinence (77.0%), mood (58.3%),
and social relationship (52.3%).
Received hospice care and alive. Residents who received hospice care and were alive
one year following their last assessment (n = 556) had a mean age of 86.5 (+9.2) years; less
than half were aged over 85 years (43.3%). Less than half of the residents in this group were
widowed (44.0%), most had a DNR order (92.5%), more than half had a DNH order (58.0%),
and most lived in an urban LTCF (93.2%) (Table 9). A diagnosis of cancer (50.7%) was
more common than a diagnosis of dementia (30.2%) in this group (Figure 7). About less than
one fourth of the residents in this group exhibited severe cognitive impairment (CPS ≥ 5,
23.0%), less than one third had depressive disorder (DRS ≥ 3, 29.7 %), less than one third
had signs of aggression (ABS ≥ 1, 31.7%), more than one third had no to low social
engagement (ISE ≤ 1, 36.90%), more than half had severe physical impairment (ADL-H ≥ 5,
51.6%), two in five had complete bowel incontinence (41.9%), one third had complete
bladder incontinence (33.3%), about two thirds had mild to severe pain (Pain Scale ≥ 1,
64.2%), about one third had high to very high pressure risk (PURS ≥ 4, 32.6%), and about
half had moderate to severe health instability (CHESS ≥ 3, 50.5%) (Table 9). More than half
of them had more than five CAPs triggered (53.0%) (Figure 8), and the top three triggered
CAPs were activities of daily living (68.6%), urinary incontinence (67.6%), and mood
(63.6%).
Received hospice care and died. Residents who received hospice care and died within
one year of assessment (n = 4,417) had a mean age of 79.7 (+12.7) years, and less than half
of them were aged 85 years and older (42.7%). More than one third of the residents in this
group were widowed (40.7%), more than half were female (57.9%), most had DNR order
(89.4%), less than two thirds had DNH order (62.7%), and most lived in an urban LTCF
84

(94.0%) (Table 9). More than half had a diagnosis of cancer (58.4%) and about one in five
had a diagnosis of dementia (26.7%). More than one third of the residents in this group
exhibited severe cognitive impairment (CPS ≥ 5, 35.3%) and possible depressive disorder
(DRS ≥ 3, 33.7%), more than one third had signs of aggression (ABS ≥ 1, 37.5%), more than
half had no to low social engagement (ISE ≤ 1, 55.5%), more than two thirds had severe
physical impairment (ADL-H ≥ 5, 77.1%), more than half had completely bowel
incontinence (54.1%), more than one third had completely bladder incontinence (37.8%),
more than two thirds had mild to severe pain (Pain Scale ≥ 1, 77.4%), more than half had
high pressure ulcer risk (PURS ≥ 4, 59.0%), and most had moderate to severe health
instability (CHESS ≥ 3, 87.0%) (Table 9). Two thirds of these residents triggered more than
five CAPs (63.9%) (Figure 8), and the top three triggered CAPs were urinary incontinence
(64.7%), mood (64.1%), and social relationship (48.6%).
Differences among four groups. Compared to residents who did not receive hospice
care and alive one year following their last assessment, residents who did not receive hospice
care but died within one year of assessment (n = 556) were older (66.5% aged over 85 years
vs. 49.5%), had a higher percentage of being widowed (57.4% vs. 49.7%), having a DNR
order (89.4% vs. 76.6%), and having a DNH order (45.4% vs. 30.7%) (Table 9). Similar to
those who did not receive hospice care and were alive one year following their last
assessment, dementia was a common diagnosis in this group (54.5%) but cancer was not
(12.9%) (Figure 7). Residents who did not receive hospice care but died within one year of
assessment had more severe and complex clinical needs than those who did not receive
hospice care and were alive one year following their last assessment (Table 9; Figure 8).
Compared to residents who did not receive hospice care and alive one year following
their last assessment, residents who received hospice care and were alive one year following
85

their last assessment were younger (43.3% aged over 75 years vs. 49.5%), had a lower
percentage of being widowed (44.0% vs. 49.7%), but a higher percentage of having DNR
order (92.5% vs. 76.6%), and having a DNH order (58.0% vs. 30.7%) and living in an urban
LTCF (93.2%) (Table 9). Cancer was a common diagnosis in this group (50.7%), while
dementia was less common (30.2%) (Figure 7). Residents who received hospice care and
were alive one year following their last assessment also had more severe and complex
clinical needs than those who did not receive hospice care and alive one year following their
last assessment. This group exhibited more severe acute clinical needs than those who did not
receive hospice care but died within one year of assessment with regard to presence of pain
(Pain Scale ≥ 1, 64.2% vs. 43.2%), high to very high pressure risk (PURS ≥ 4, 32.6% vs.
27.9%), and moderate to severe health instability (CHESS ≥ 3, 50.5% vs. 21.9%), but had
less chronic clinical needs in terms of severe cognitive impairment (CPS ≥ 5; 23.0% vs.
32.4%) and bladder incontinence (33.3% vs. 48.7%) (Table 9).
Compared to residents who did not receive hospice care and were alive one year
following their last assessment, residents who received hospice care and died within one year
of assessment were younger (42.7% aged over 75 years vs. 49.5%), had a lower percentage
of being widowed (40.7% vs. 49.7%), but a higher percentage of having DNR order (89.4%
vs. 76.6%), having a DNH order (62.7% vs. 30.7%), and living in an urban LTCF (94.0%)
(Table 9). Compared to other three groups, this group had and the highest percentage of
cancer diagnosis (58.4%) and the most severe and complex clinical conditions with respect to
cognition, depression, social engagement, physical function, bowel control, pain, pressure
ulcer risk, health instability and highest complexity (Table 9, Figure 7; Figure 8).
Multivariate results. Table 10 shows the predictors of hospice use status identified
in the multinomial logistic regression analysis. Compared to residents who did not receive
86

hospice care and were alive one year following their last assessment, residents who did not
receive hospice care and died within one year of the assessment were less like to live in an
urban LTCF (OR = 0.78, 95% CI = 0.75-0.80), and were more likely to be older (OR range
from 1.60 [95% CI = 1.50-1.72] among 65-74 age group to 5.45 [95% CI = 5.11-5.82]
among 95+ age group) and have moderate (OR = 1.21, 95% CI = 1.11-2.33) to severe (OR =
1.83, 95% CI = 1.67-2.01) physical impairment, have mild-moderate (OR = 1.13, 95% CI =
1.10-1.16) to severe (OR = 1.39, 95% CI = 1.27-1.53) pain, pressure ulcer risk (OR range
from 1.38 [95% CI = 1.31-1.44] in mild pressure ulcer risk category to 2.64 [95% CI = 2.362.97] in very high pressure ulcer risk category) and mild (OR = 1.55, 95% CI = 1.50-1.59) to
moderate-severe (OR = 3.53, 95% CI = 3.34-3.71) health instability.
Compared to residents who did not receive hospice care and were alive one year
following their last assessment, residents who received hospice care and were alive one year
following their last assessment were more likely to be aged between 75 to 84 (OR = 0.63,
95% CI = 0.47-0.84) and 85 to 94 (OR = 0.66, 95% CI = 0.50-0.88) years, live in urban
LTCF (OR = 2.04, 95% CI = 1.46-2.84) and have a diagnosis of cancer (OR = 7.98, 95% CI
= 6.68-9.54), moderate (OR = 2.07, 95% CI = 1.25-3.44) to high (OR = 2.18, 95% CI = 1.273.75) pressure ulcer risk, and mild (OR = 2.83, 95% CI = 2.01-3.98) to severe (OR = 17.00,
95% CI = 15.60-24.88) health instability.
Compared to residents who did not receive hospice care and were alive one year
following their last assessment, residents who received hospice care and died within one year
of the assessment were more likely to be 95 years and older (OR = 1.31, 95% CI = 1.101.56), have a diagnosis of cancer (OR = 8.96, 95% CI = 8.27-9.91), and severe physical
impairment (OR = 5.38, 95% CI = 2.12-13.64), and were also more likely have mildmoderate (OR = 2.03, 95% CI = 1.82-2.28) to severe (OR = 2.68, 95% CI = 2.19-3.27) pain,
87

moderate to very high pressure ulcer risk (OR ranges from 2.18 [95% CI = 1.45-3.29] in
moderate pressure ulcer risk category to 3.50 [95% CI = 2.24-4.38] in very high pressure
ulcer risk category), and mild (OR = 10.40, 95% CI = 6.67-16.22) to moderate-severe (OR =
205.37, 95% CI = 132.09-319.30) health instability.
Discussion
The assessed use rate of hospice care in LTCFs was very low (i.e., less than 3%),
while one in five of the residents died within three months of the assessment and one in three
died within one year of the assessment. Most deaths occurred three months after the
assessment among those who received hospice care and among those who did not receive
hospice care. Among those who did not receive hospice care, more than one fourth of
residents died within one year the assessment. This indicates one in four of residents in
LTCFs who had potential to benefit from hospice care may not have received it. The actual
use of hospice care among residents in LTCFs is very poor in Canada, which indicates urgent
needs for immediate action to improve hospice care utilization in LTCFs in Canada.
This study is one of the first to really quantify and compare those who received
hospice care and those who did not with proximity to death. The findings of this study
indicate how the residents’ predisposing, enabling, and need characteristics influence their
utilization of hospice care services, which helps to identify access disparities and other
barriers to hospice use in LTCFs in Canada (Andersen and Newman, 1973). The results
indicate there are big differences in residents’ predisposing, enabling, and need
characteristics among the four groups. Younger age, living in an urban LTCF, having a
diagnosis of cancer, and having more severe, complex, and acute clinical needs significantly
increased the likelihood of hospice use among residents in LTCFs in Canada.

88

This study indicates residents who were older were more likely to die without hospice
care, which reveals evidence of ageism in relation to equitable access to hospice care in
LTCFs. Evidence of inequalities in access to end-of-life care, particularly between age
groups, has been reported in the United States, the United Kingdom, and Australia
(Addington-Hall, 2000; Burt & Raine, 2006; Rosenwax & McNamara, 2006). Age
inequalities in access to end-of-life care was also found in a study of cancer patients in the
province of Nova Scotia, Canada (Burge, Lawson, Johnston, & Grunfeld, 2008). It is well
known that the Canadian population is ageing, thus, failure to recognize the role of ageism in
relation to accessibility to hospice care may pose severe consequences. Ageism affects the
way in which services at the end of life are often designed without reference to older LTCF
residents’ needs (Gardiner, Cobb, Gott, & Ingleton, 2011; Gott, Ibrahim, & Binstock, 2011).
The tendency to give greater value to youth over old age and attribute negative characteristics
to older residents may influence older residents’ expectations and experiences of hospice care
(Burge, Lawson, Johnston, & Grunfeld, 2008; Gott et al., 2011). Older LTCF residents’
needs should be valued during policy development with regard to improving equitable access
to hospice care within LTCFs.
About nine in ten of residents in LTCFs had a DNR order compared to one in three
had a DNH order. Residents who received hospice care had a higher percentage of having a
DNR or DNH order in place than those who did not. The casual effect between having a
DNR order and receiving hospice care cannot be determined in this study. Having a DNR
order in place may lead to higher probability of referral to hospice care. Having severe and
complex health conditions may result in more discussion of advance care planning and higher
probability of having a DNR in place to be eligible for hospice referral. However, among
those who received hospice care, around 10% did not have a DNR order in place. There is
89

still room to improve the use of advance directives among LTCF residents who receive
hospice care. Among those who received hospice care, a large percentage of residents did not
have a DNH in place. It remains unknow why residents who received hospice care still
wanted to be hospitalized when possible and this requires further study.
The results of this study show living in an urban LTCF were more likely to receive
hospice care regardless of their one-year survivorship and less likely to die without hospice
care. Studies have demonstrated that in addition to the ability to recognize terminal stage, for
LTCF residents, access to hospice may be more influenced by the facility and its location
than by the residents’ treatment preferences (Zerzan, Stearns, & Hanson, 2000; Welch,
Miller, Martin, & Nanda, 2008). Contributors of the rural-urban difference include lack of
resources and funding (e.g. lack of equipment, low salaries, and lack of specialist
geriatricians) and limited access to hospice services in rural LTCFs, compared with urban
LTCFs (Virnig, Hartman, Moscovice, & Carlin, 2006; Kelley, Sletmoen, Williams, Nadin, &
Puiras, 2012; Kaasalainen et al., 2012). As hospice requires physician referral, physician
shortage and high physician turnover in rural LTCFs may create barriers of hospice referral
and lower the threshold for transferring residents to hospitals, especially when LTCFs and
hospitals are co-located or located within a relatively short distance of each other (TemkinGreener, Zheng, & Mukamel, 2012; Kaasalainen et al., 2012). Urban areas are more likely to
have a larger number of hospice providers and closer proximity of those providers to LTCFs,
thus decreasing potential barriers to hospice utilization (Temkin-Greener et al., 2012).
Recently, there are some studies about hospice palliative care practices in private home of
rural communities in Canada (Kaasalainen et al., 2012; Whitfield, 2018), however, it remains
unclear how hospice care works in rural LTCFs (DeMiglio et al. 2012). The issue of LTCF-

90

based urban-rural disparity needs further investigation to find ways to improve access to, and
provision of, hospice care services in rural LTCFs.
Although residents with a diagnosis of cancer were more likely to die without hospice
care, they were also more likely to receive hospice care regardless of their one-year
survivorship. However, residents with a diagnosis of Dementia Not Alzheimer’s were less
likely to receive hospice care regardless of their one-year survivorship and more likely to die
without hospice care. A report from CIHI also revealed that few seniors living with dementia
in Canada receive hospice palliative care and this is more prevalent among residents in
LTCFs (CIHI, 2019). Residents with dementia were relatively under-served by hospice,
partly because of challenges in prediction of life expectancy for persons with dementia and
lack of recognition of the terminal nature of dementia, lack of advance directives in place
among persons with dementia, and short of perceived benefits to the persons with dementia
(Erel et al., 2017; Mitchell et al., 2010; Sachs et al., 2004; Kiely, 2010, Mitchell, Kiely,
Hamel, 2004). It is important for health care providers and policy-makers to improve
recognition of dementia as a terminal disease and better understand the role of hospice
services for persons with dementia so these residents can have improved access to hospice
care as residents with end-stage cancer.
The clinical characteristics of residents who did not receive hospice care and were
alive one year following the assessment confirmed general expectations that this group had
less severe and complex health conditions than other groups. This study indicates that
residents who received hospice care and died within one year of assessment had the most
severe and complex clinical needs among the four groups. Residents who received hospice
care and were alive one year following their last assessment exhibited more severe acute
clinical needs (i.e. pain, high pressure risk, and health instability) and had less chronic
91

clinical needs (i.e. cognitive impairment) than those who did not receive hospice care but
died within one year of assessment. More attention should be paid to residents with chronic
clinical needs (i.e. cognitive impairment) when considering hospice referral.
Residents with more frequent and intense pain were more likely to receive hospice
care regardless of their one-year survivorship. Due to the limitation related to the study
design, it was not clear whether pain is one of the reasons for hospice referral or residents
who received hospice care were more likely to develop pain. Many studies revealed hospice
care is often targeted to dying residents with higher levels of reported pain as hospice care in
LTCFs can lead to better pain assessment and management for dying LTCF residents (Miller,
Mor, & Teno, 2003; Baer & Hanson, 2000; Miller, Mor, Wu, Gozalo, & Lapane, 2002; Wu,
Miller, Lapane, & Gozalo, 2003). Hospice care providers should recognize that residents
with cognitive impairment who may not be able to self report pain (Herr et al., 2011; Miu, &
Chan, 2014). Data from this study indicates residents who did not receive hospice care and
died within one year reported less pain than those who received hospice care. Part of the
reason may be good control of pain from LTCF staff. A previous study found LTCF staff
seem well positioned to control pain for residents whose deaths were expected (Munn et al.,
2006). More possible reasons may be under-detected pain among LTCF resident without
hospice care, especially among those with cognitive impairment (Herr et al., 2011; Marx,
2005; Miu, & Chan, 2014). On the other hand, many studies indicate that hospice positively
affects and improves the assessment of symptoms including pain (Miller, Mor, & Teno,
2003; Baer & Hanson, 2000; Miller, Mor, Wu, Gozalo, & Lapane, 2002; Wu, Miller, Lapane,
& Gozalo, 2003). Efforts from both LTCF staff and hospice providers are needed to have
pain better controlled for residents who received hospice care.

92

Most residents triggered more than one CAP, while those who received hospice care
and died within one year of assessment had more CAPs triggered. The high number of CAPs
triggered among residents who received hospice care and died within one year of assessment
reflects high levels of complex clinical needs within this group. Among those who died
within one year of assessment, urinary incontinence, mood, and social relationship were the
top three area in which the residents had a higher than expected rate of decline, an increased
potential to improve, and symptoms that could be alleviated if problems are addressed.
Among those who were still alive one year following the assessment, activities of daily
living, urinary incontinence, and mood were the top three areas in which the residents had a
higher than expected rate of decline and may have benefited from alleviated symptoms if
problems were addressed. The CAPs triggered at high rates such as urinary incontinence and
mood among all residents warrant increased attention for the majority of LTCF residents.
Consideration of specific triggered CAPs in this study provide evidence to support care plan
related decisions based on residents’ needs.
Although residents with health instability were more likely to die with hospice care,
they were also more likely to receive hospice care regardless of their one-year survivorship.
Studies have found CHESS was a good predictor of mortality for hospitalized patients and
persons with neurological conditions (Hirdes et al., 2003). However, this study revealed
hospice care was offered to a large proportion of residents who did not die within one year
but had severe health instability. This indicates that while CHESS may be a good predictor of
mortality, it may not be a good predictor of receipt of hospice care. The predictive model
developed in this study (p. 73) were superior than CHESS alone, as clinical needs were not
the only determinants of hospice use.

93

Strengths and limitations. The major strength of this study is that it is the first
Canadian study to use the RAI-MDS 2.0 data to examine hospice use among LTCF residents
stratified by their one-year survivorship. Andersen and Newman’s (1973) behaviors model
for health services utilization was used to help understand hospice care access and utilization
in LTCFs in Canada. This study included three population characteristics components of the
framework to explore their association with health services utilization. In addition, this study
used a mix of admission assessments (17.4%), significant change in status full assessment
(5.3%), annual full assessments (22.0%), and quarterly assessments (55.3%). Thus, the study
population covered all stage of residents living in LTCFs. Furthermore, this study linked the
DAD to the CCRS to track an extra 5.2% residents who were discharged from LTCFs and
did not have death records in LTCFs.
This study also has several limitations. First, this study showed one-year survivorship
of hospice use among residents living in LTCFs in Canada at the time of last assessment in
2015. The hospice use rate among LTCF residents at death cannot be identified, as some
residents who did not receive hospice care at the time of last assessment in 2015 may have
received hospice care after the assessment. Second, this study revealed substantial differences
in hospice use among provinces. However, the differences among all provinces in Canada
cannot be explicitly examined, as the CCRS database does not contain full coverage of
LTCFs in all provinces of Canada. Third, although the study revealed numerous associations
between resident characteristics and hospice use, causality could not be determined due to
study design. Longitudinal study of differences of resident characteristics in hospice use is
recommended to better understand the association between resident characteristics and
hospice use. Fourth, complete case analyses were used in the predictive modeling. To reduce
potential for biased and inefficient estimates, variables with a large proportion of missing
94

values, such as DNR order (5.8%), marital status (12.7%) and education (36.7%) were
removed from predictive modeling.
Conclusion
Residents who received hospice care and died within one year of assessment
exhibited more severe and complex clinical needs than those who did not receive hospice
care and those who receive hospice care and were alive one year following their last
assessment. Residents who received hospice care and were alive one year following their last
assessment exhibited more severe acute clinical needs (i.e. pain, high pressure risk, and
health instability) and had less chronic clinical needs (i.e. cognitive impairment, depression,
and low social engagement) than those who did not receive hospice care but died within one
year of assessment. This study indicates several possible barriers to hospice use in the LTCF
population including ageism, rurality, and a diagnosis of dementia. As Andersen and
Newman (1973) indicates, all these factors come together to explain and inform health care
utilization. Immediate action is needed to address inequality in care at the end of life for the
LTCF population and provide improved access to high quality hospice care in LTCFs in
Canada.

95

Table 9. Predisposing, Enabling, and Need Characteristics by One-Year Survival Survivorship of Hospice Use Among Residents in

Long-Term Care Facilities in Canada at Baseline in 2015 (N=185,715)
Variables
Determinants 1: Predisposing Characteristics
Age Group (in years)
19-64
65-74
75-84
85-94
95+
Female Sex
Marital Status
Never Married
Married
Widowed
Separated
Divorced
Primary Language Spoken
English
French
Education
8th Grade or Less
9th Grade-High School
Technical/Trade School/College
Bachelor's or Higher Degree
Advance Directives
Do Not Resuscitate
Do Not Hospitalize
Determinants 2: Enabling Characteristics
Province
Alberta
British Columbia
Ontario
Other Provinces

Total
Population
N=185,715

Did not receive hospice care
Alive after
Died within
one year,
one year,
68.8%,
28.5%,
N= 127,761
N= 52,981

Received hospice care
Alive after
Died within
one year,
one year,
0.3%,
2.4%,
N= 556
N=4,417

7.6 (14,055)
11.0 (20,509)
27.2 (50,483)
43.9 (81,435)
10.4 (19,233)
65.8 (122,145)

9.2 (11,790)
12.7 (161,66)
28.6 (36,490)
41.4 (52,922)
8.1 (10,393)
66.8 (85,289)

3.0 (1,574)
6.7 (3,535)
23.9 (12,647)
50.6 (26,814)
15.9 (8,411)
64.2 (33,963)

15.1 (84)
17.1 (95)
28.6 (136)
33.6 (187)
9.7 (54)
60.3 (335)

13.7 (607)
16.1 (713)
23. 9 (1,210)
34.2 (1,512)
8.5 (375)
57.9 (2,558)

10.3 (16,647)
27.5 (44,546)
51.6 (83,583)
3.1 (5,017)
7.6 (12,337)
83.6 (155,309)
2.5 (4,560)
13.9 (25,846)

11.6 (13,129)
27.1 (30,637)
49.7 (56,137)
3.3 (3,683)
8.4 (9478)
82.5 (105,345)
2.5 (3,245)
15.0 (19,171)

6.9 (3,069)
27.2 (12,122)
57.4 (25,612)
2.8 (1,238)
5.8 (2,564)
86.3 (45,694)
2.1 (1,112)
11.7 (6,175)

8.3 (42)
36.9 (187)
44.0 (233)
2.2 (11)
8.7 (44)
83.6 (465)
4.9 (27)
11.5 (64)

10.3 (407)
40.5 (1,600)
40.7 (1,611)
2.2 (85)
6.4 (251)
86.1 (3,805)
4.0 (176)
9.9 (436)

28.5 (33,527)
44.0 (36,767)
16.7 (19,666)
10.7 (12,596)

28.6 (23,158)
43.9 (35,522)
16.8 (13,608)
10.7 (8,694)

28.5 (9,877)
44.5 (15,428)
16.5 (5,704
10.5 (3,654)

28.1 (62)
37.1 (82)
22.6 (50)
12.2 (27)

25.6 (430)
43.2 (725)
18.1 (304)
13.2 (221)

89.8 (137,001)
35.7 (62,244)

76.6 (92,034)
30.7 (36,710)

89.4 (44,946)
45.4 (22,753)

92.5 (457)
58.0 (282)

89.4 (3,614)
62.7 (2,499)

10.0 (18,505)
16.3 (30,288)
62.8 (116,552)
11.0 (20,370)

8.8 (11,236)
15.3 (19,494)
65.8 (84,114)
10.2 (12,917)

13.4 (7,097)
20.2 (10,695)
52.5 (27,796)
14.1 (7,393)

1.3 (7)
3.2 (18)
92.8 (516)
2.7 (15)

3.7 (165)
1.8 (81)
93.4 (4,126)
1.1 (45)

96

p-value

<0.0001

<0.0001
<0.0001

<0.0001
<0.0001

<0.0001

<0.0001

Urban Facility
Facility Neighbourhood Income Quintile
Lowest (QAIPPE = 1)
Low (QAIPPE = 2)
Medium (QAIPPE = 3)
High (QAIPPE = 4)
Highest (QAIPPE = 5)
Determinants 3: Need Characteristics
Change in Cognitive Status
No Change
Improved
Deteriorated
Cognition
No-Mild Cognitive Impairment (CPS = 0-1)
Moderate Cognitive Impairment (CPS = 2-4)
Severe Cognitive Impairment (CPS = 5-6)
Change in Communication
No Change
Improved
Deteriorated
Change in Mood
No Change
Improved
Deteriorated
Depression
No Depressive Symptoms (DRS = 0)
Some Depressive Symptoms (DRS = 1-2)
Mild Depressive Disorder (DRS = 3-5)
Moderate-Severe Depressive Disorder (DRS = 6-14)
Change in Behavior Symptom
No Change
Improved
Deteriorated
Aggressive Behavior
No Signs of Aggression (ABS = 0)
Mild to Moderate Aggression (ABS = 1-4)
Severe Aggression (ABS = 5-12)

85.2 (158,246)

86.0 (109,860)

82.6 (43,715)

93.2 (518)

94..0 (4,153)

26.1 (48,213)
19.1 (35,280)
20.4 (37,655)
18.9 (3,4912)
15.4 (28,497)

26.4 (33,511)
18.9 (24,016)
20.4 (25,833)
19.1 (24,214)
15.2 (19,293)

26.4 (13,928)
18.8 (9,901)
20.7 (10,893)
18.4 (9,691)
15.8 (8,320)

15.5 (86)
25.2 (140)
18.2 (101)
29.0 (161)
12.1 (67)

15.6 (688)
27.8 (1,233)
18.8 (828)
19.2 (846)
18.6 (817)

<0.0001
<0.0001

<0.0001
85.9 (159,103)
1.8 (3,349)
12.3 (22,683)

90.2 (115,121)
2.0 (2,515)
7.8 (9,985)

79.6 (41,951)
1.5 (779)
19.0 (9,995)

69.2 (384)
2.2 (12)
28.7 (159)

38.9 (1,647)
1.0 (43)
60.1 (2,544)

22.3 (41,329)
55.5 (103,097)
22.2 (41,289)

25.6 (32,689)
56.9 (72,660)
17.5 (22,412)

14.3 (7,552)
53.3 (28,240)
32.4 (17,189)

29.1 (162)
47.8 (266)
23.0 (128)

21.0 (926)
43.7 (1,931)
35.3 (1,560)

91.9 (170,042)
0.7 (1,212)
7.5 (13,881)

95.1 (121,307)
0.7 (918)
4.2 (5,396)

86.3 (45,737)
0.5 (272)
12.7 (6,716)

83.4 (462)
1.3 (7)
15.3 (85)

59.9 (2,535)
0.4 (15)
39.8 (1,684)

82.4 (152,576)
5.1 (9,465)
12.5 (23,094)

85.0 (108,505)
5.3 (6,772)
9.7 (12,344)

78.9 (41,592)
4.8 (2,518)
16.3 (8,615)

66.2 (368)
4.7 (26)
29.0 (161)

49.9 (2,111)
3.5 (149)
46.6 (1,974)

41.0 (75,838)
30.3 (56,163)
19.8 (36,669)
8.9 (16,465)

43.2 (55,161)
29.5 (37,633)
19.1 (24,364)
8.2 (10,463)

36.1 (19,043)
32.2 (16,996)
21.1 (11,129)
10.6 (5,587)

41.6 (231)
28.7 (159)
23.4 (130)
6.3 (35)

33.1 (1,403)
33.2 (1,405)
24.7 (1,046)
9.0 (380)

85.0 (157,437)
4.4 (8,181)
10.5 (19,517)

87.3 (111,368)
4.3 (5,487)
8.4 (10,766)

81.2 (43,796)
4.7 (2,486)
14.1 (7,443)

78.2 (434)
6.0 (33)
15.9 (88)

67.1 (2,839)
4.1 (175)
28.8 (1,220)

58.9 (108,962)
33.0 (61,091)
8.2 (15,082)

60.8 (77,594)
31.6 (40,337)
7.6 (9,690)

53.8 (28,345)
36.4 (19,209)
9.8 (5,171)

68.2 (379)
27.9 (155)
3.8 (21)

62.5 (2,644)
32.8 (1,390)
4.7 (200)

97

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

Social Engagement
No to Low Social Engagement (ISE = 0-1)
Moderate Social Engagement (ISE = 2-4)
High Social Engagement (ISE = 5-6)
Change in Physical Function
No Change
Improved
Deteriorated
Physical Function
No Functional Impairment (ADL-H = 0)
Mild Functional Impairment (ADL-H = 1-2)
Moderate Functional Impairment (ADL-H = 3-4)
Severe Functional Impairment (ADL-H = 5-6)
Bowel Continence
Continent
Usually Continent
Occasionally continent
Frequently Continent
Incontinent
Bladder Continence
Continent
Usually Continent
Occasionally continent
Frequently Continent
Incontinent
Change in Urinary Continence
No Change
Improved
Deteriorated
Pain Symptoms Frequency
No Pain
Pain Less Than Daily
Pain Daily
Pain Symptoms Intensity
Mild Pain
Moderate Pain
Severe Pain

27.3 (50,606)
52.5 (97,572)
20.2 (37,537)

21.4 (27,303)
54.4 (69,529)
24.2 (24.2)

39.0 (20,645)
49.4 (26,177)
11.6 (6,159)

36.9 (205)
46.0 (256)
17.1 (95)

55.5 (2,453)
36.5 (1,610)
8.0 (354)

74.6 (138,594)
4.5 (8,382)
20.9 (38,739)

79.2 (101,127)
5.6 (7,130)
15.3 (19,504)

68.6 (68.6)
2.2 (1,184)
29.2 (15,444)

46.8 (260)
5.9 (33)
47.3 (263)

19.3 (854)
0.8 (35)
79.9 (3,528)

4.2 (7,774)
16.7 (31,052)
44.4 (82,424)
34.7 (64,465)

5.3 (6,817)
20.3 (25,930)
47.5 (60,706)
26.9 (34,308)

1.8 (927)
9.1 (4,836)
39.2 (20,755)
50.0 (26,463)

3.6 (20)
15.1 (84)
29.7 (165)
51.6 (287)

0.2 (10)
4.6 (202)
18.1 (798)
77.1 (3,407)

34.9 (64,768)
11.8 (21,915)
8.0 (14,876)
15.0 (27,873)
30.3 (27,873)

40.9 (52,279)
12.3 (15,712)
8.0 (10,161)
14.2 (18,177)
24.6 (31,432)

21.8 (11,545)
10.8 (5,745)
8.4 (4,425)
17.1 (9,036)
42.0 (22,230)

31.7 (176)
10.1 (56)
5.2 (29)
11.2 (62)
41.9 (233)

17.4 (768)
9.1 (402)
5.9 (261)
13.5 (598)
54.1 (2,388)

20.7 (38,469)
8.6 (15,941)
9.9 (18,317)
24.5 (45,548)
36.3 (67,440)

22.9 (29,197)
9.6 (12,304)
10.8 (13,795)
25.6 (32,656)
31.2 (39,809)

14.2 (7,512)
6.3 (3,310)
7.8 (4,149)
23.1 (12,234)
48.7 (25,776)

38.5 (214)
8.1 (45)
7.0 (39)
13.1 (73)
33.3 (185)

35.0 (1,546)
6.4 (262)
7.6 (334)
13.2 (585)
37.8 (1,670)

85.8 (159,362)
2.8 (5,199)
11.4 (21,154)

88.2 (112,711)
3.1 (3,892)
8.7 (11,158)

83.1 (44,026)
2.0 (1,083)
14.9 (7,872)

71.2 (396)
5.2 (29)
23.6 (131)

50.5 (2,229)
4.4 (195)
45.1 (1,993)

60.5 (112,389)
24.8 (45,980)
14.7 (27,346)

63.5 (81,083)
23.8 (30,391)
12.8 (16,287)

56.8 (30,110)
25.9 (13,702)
17.3 (9,169)

35.8 (199)
34.2 (190)
30.0 (167)

22.6 (997)
38.4 (1,697)
39.0 (1,723)

45.7 (33,523)
46.1 (33,791)
8.2 (6,012)

48.6 (22,702)
44.5 (20,781)
6.8 (3,195)

42.9 (9.813)
47.6 (10,880)
9.5 (2,178)

32.2 (115)
54.6 (195)
13.2 (47)

26.1 (893)
56.6 (1,935)
17.3 (593)

98

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

<0.0001

Pain
<0.0001
No Pain (Pain Scale = 0)
60.5 (112,389)
63.5 (81,083)
56.8 (30,110)
35.8 (199)
22.6 (997)
Mild-Moderate Pain (Pain Scale = 1-2)
37.2 (68,997)
34.9 (44,540)
40.1 (21,219)
57.2 (318)
66.1 (2,920)
Severe Pain (Pain Scale = 3)
2.3 (4,329)
1.7 (2,138)
3.1 (1,652)
7.0 (39)
11.3 (500)
Weight Loss
11.0 (18,973)
7.3 (8,719)
17.7 (8,838)
28.3 (128)
40.4 (1,288)
<0.0001
Leaves Food Uneaten
33.3 (61,836)
26.0 (33,250)
47.0 (24,911)
50.9 (283)
76.8 (3,392)
<0.0001
Pressure Ulcer Risk
<0.0001
No Pressure Ulcer Risk (PURS = 0)
19.0 (34,861)
23.8 (30,126)
8.8 (4,595)
9.2 (51)
2.0 (89)
Mild Pressure Ulcer Risk (PURS = 1-2)
33.1 (60,827)
36.5 (46,254)
26.5 (13,858)
27.9 (154)
12.8 (561)
Moderate Pressure Ulcer Risk (PURS = 3)
31.0 (57,005)
28.8 (36,406)
36.8 (19,276)
30.4 (168)
26.3 (1,155)
High Pressure Ulcer Risk (PURS = 4-5)
15.4 (28,401)
10.1 (12,797)
25.1 (13,152)
30.2 (167)
51.9 (2,285)
Very High Pressure Ulcer Risk (PURS = 6-8)
1.6 (2,847)
0.8 (1,041)
2.8 (1,483)
2.4 (13)
7.1 (310)
Health Instability
<0.0001
No Indication of Health Instability (CHESS = 0)
40.7 (75,639)
48.1 (61,422)
26.7 (14,127)
10.4 (58)
0.7 (32)
Mild Health Instability (CHESS = 1-2)
47.0 (87,227)
46.4 (59,239)
51.4 (27,228)
39.0 (217)
12.3 (12.3)
Moderate-Severe Health Instability (CHESS = 3-5)
12.3 (22,849)
5.6 (7,100)
21.9 (11,626)
50.5 (281)
87.0 (3,842)
Number of CAPs Triggered
<0.0001
0
0.7 (1,228)
0.6 (763)
0.8 (434)
1.4 (8)
0.5 (23)
1-5
51.9 (96,344)
54.7 (69,938)
46.4 (24,579)
45.5 (253)
35.6 (1,574)
6-10
45.5 (84,412)
43.2 (55,197)
49.8 (26,387)
49.8 (277)
57.8 (2,551)
11-16
1.6 (3,009)
1.1(1,456)
2.4 (1,266)
3.2 (18)
6.1 (269)
Note. QAIPPE denotes Quintile of Adjusted Income per Person Equivalent; CPS denotes Cognitive Performance Scale; DRS denotes Depression Rating Scale;
ABS denotes Aggressive Behavior Scale; ISE denotes Index of Social Engagement; ADL-H denotes Activities of Daily Living Hierarchy; PURS denotes
Pressure Ulcer Risk Scale; CHESS denotes Changes in Health, End-Stage Disease, Signs, and Symptoms Scale.

99

Table 10. Multinomial Logistic Regression of Predisposing, Enabling, and Need Characteristics on One Year Survivorship of Hospice Use

Among Residents in Long-Term Care Facilities in Canada in 2015 with Residents Who Did Not Receive Hospice Care and Were
Alive One Year Following the Last Assessment as the Reference Group (Odds Ratios and 95% Confidence Intervals).
Variables
Age Group (ref=19-64)
65-74
75-84
85-94
95+
Urban-Rural Status (ref = Rural)
Urban
Dementia Not Alzheimer’s (ref = No)
Yes
Cancer (ref = No)
Yes
Cognition (ref = No-Mild Cognitive Impairment [CPS = 0-1])
Moderate Cognitive Impairment (CPS = 2-4)
Severe Cognitive Impairment (CPS = 5-6)
Depression (ref = No Depressive Symptoms [DRS = 0])
Some Depressive Symptoms (DRS = 1-2)
Mild Depressive Disorder (DRS = 3-5)
Moderate-Severe Depressive Disorder (DRS = 6-14)
Aggressive Behavior (ref = No Signs of Aggression [ABS = 0])
Mild to Moderate Aggression (ABS = 1-4)
Severe Aggression (ABS = 5-12)
Social Engagement (ref = No-Low Social Engagement [ISE = 0-1])
Moderate Social Engagement (ISE = 2-4)
High Social Engagement (ISE = 5-6)
Physical Function (ref = No Functional Impairment [ADL-H = 0])
Mild Functional Impairment (ADL-H = 1-2)
Moderate Functional Impairment (ADL-H = 3-4)
Severe Functional Impairment (ADL-H = 5-6)
Pain (ref = No Pain [Pain Scale = 0])
Mild-Moderate Pain (Pain Scale = 1-2)
Severe Pain (Pain Scale = 3)

Did not receive hospice
Died within one year,
28.5%,
N= 52,981

Received hospice care
Alive after one year,
0.3%,
N= 556

Died within one year,
2.4%,
N=4,417

1.60 (1.50-1.72)****
2.39 (2.25-2.54)****
3.46 (3.26-3.67)****
5.45 (5.11-5.82)****

0.86 (0.64-1.17)
0.63 (0.47-0.84)**
0.66 (0.50-0.88)**
1.06 (0.74-1.53)

0.93 (0.79-1.09)
0.86 (0.74-0.99)*
0.89 (0.78-1.02)
1.31 (1.10-1.56)**

0.78 (0.75-0.80)****

2.04 (1.46-2.84)**

2.17 (1.89-2.50)****

1.04 (1.01-1.07)*

0.71 (0.57-0.87)*

0.59 (0.54-0.65)****

1.68 (1.62-1.74)****

7.98 (6.68-9.54)****

8.96 (8.27-9.71)****

1.10 (1.06-1.15)****
1.33 (1.27-1.40)****

1.08 (0.81-1.45)
1.41 (0.97-2.06)

0.91 (0.78-1.05)
1.27 (1.06-1.52)**

1.04 (1.00-1.07)*
1.01 (0.97-1.05)
1.03 (0.98-1.09)

0.87 (0.67-1.14)
1.18 (0.87-1.61)
0.74 (0.46-1.18)

0.81 (0.71-0.92)***
0.89 (0.77-1.02)
0.72 (0.60-0.87)***

0.95 (0.92-0.98)**
0.90 (0.86-0.95)***

0.66 (0.51-0.85)**
0.36 (0.21-0.62)***

0.80 (0.71-0.89)****
0.49 (0.39-0.60)****

0.74 (0.72-0.76)****
0.54 (0.52-0.57)****

0.67 (0.52-0.86)**
0.63 (0.45-0.90)*

0.52 (0.46-0.58)****
0.39 (0.33-0.48)****

1.05 (0.96-1.15)
1.21 (1.11-1.33)****
1.83 (1.67-2.01)****

0.74 (0.38-1.45)
0.48 (0.25-0.95)*
0.86 (0.43-1.72)

1.85 (0.72-4.73)
1.84 (0.73-4.65)
5.38 (2.12-13.64)***

1.13 (1.10-1.16)****
1.39 (1.27-1.53)****

1.29 (1.02-1.63)*
1.33 (0.79-2.24)

2.03 (1.82-2.28)****
2.68 (2.19-3.27)****

100

Pressure Ulcer Risk (ref = No Pressure Ulcer Risk [PURS = 0])
Mild Pressure Ulcer Risk (PURS = 1-2)
Moderate Pressure Ulcer Risk (PURS = 3)
High Pressure Ulcer Risk (PURS = 4-5)
Very High Pressure Ulcer Risk (PURS = 6-8)
Health Instability (ref = No Indication of Health Instability [CHESS = 0])
Mild Health Instability (CHESS = 1-2)
Moderate-Severe Health Instability (CHESS = 3-5)

1.38 (1.31-1.44)****
1.69 (1.60-1.78)****
2.30 (2.17-2.45)****
2.64 (2.36-2.97)****

1.33 (0.84-2.12)
2.07 (1.25-3.44)**
2.18 (1.27-3.75)**
1.21 (0.50-2.92)

1.23 (0.82-1.84)
2.18 (1.45-3.29)***
3.05 (2.02-4.61)****
3.50 (2.24-4.38)****

1.55 (1.50-1.59)****
3.53 (3.34-3.71)****

2.83 (2.01-3.98)****
17.00 (15.60-24.88)****

10.40 (6.67-16.22)****
205.37 (132.09-319.30)****

*p <0 .05, **p <0 .01, ***p < 0.001, ****p <0 0.001.
Note. Wald χ2(12, N = 182,858) = 33,066.2, p < 0.0001, AIC = 212,328.1, classification rate=73.4%
CPS denotes Cognitive Performance Scale; DRS denotes Depression Rating Scale; ABS denotes Aggressive Behavior Scale; ISE denotes Index of Social
Engagement; ADL-H denotes Activities of Daily Living Hierarchy; PURS denotes Pressure Ulcer Risk Scale; CHESS denotes Changes in Health, End-Stage
Disease, Signs, and Symptoms Scale.

101

Figure 6. Survival Analysis by Status of Hospice Use of Residents in Long-Term Care
Facilities in Canada in 2015 (N=185,715)
Did not receive hospice care

Received hospice care

Time to Deaths (Days)

102

Figure 7. Selected Common Diagnoses by Status of Hospice Use of Residents in Long-Term Care Facilities in Canada in 2015
(N=185,715)
100.0%

Did not reveive hospice care and alive
Did not reveive hospice care and died
Reveived hospice care and alive
Reveived hospice care and died

80.0%

60.0%

40.0%

20.0%

0.0%

Dementia Not Diabetes Mellitus Cerebrovascular
Accident
Alzheimer’s

Emphysema

Alzheimer’s

103

Congestive Heart Arterial Heart
Failure
Disease

Renal Failure

Cancer

Figure 8. Number of Clinical Protocol Assessments Triggered by One-Year Survivorship of Hospice Use Among Residents in LongTerm Care Facilities in Canada in 2015 (N=185,715)

104

Chapter Six: Summary and Significance
The work described in this thesis was completed to describe the characteristics of
residents who received hospice care and to explore predictors of hospice use among residents
in LTCFs in Canada. The work described in this thesis was guided by two main research
questions: (a) What are the characteristics of residents who receive hospice care in LTCFs
versus those who did not by end-stage status and by one-year survivorship? (b) What
variables can predict hospice use and one-year survivorship of hospice use among LTCF
residents?
Findings from the work described in this thesis show one in nine of LTCF residents
who died in 2015 received hospice care in Canadian LTCFs. Previous studies reveal less than
5% of residents who died in nursing homes in 1998 received hospice care, one third of the
LTCF residents who died before 2000 received hospice care in Michigan, compared to 39%
of urban decedents received hospice care in nursing homes in the U.S. (Han et al., 2008;
Lepore, Miller, & Gozalo, 2011; Lin, 2010). Although evidences show the use of hospice
care among LTCF residents has increased over time (Gozalo & Miller, 2007; Infeld et al.,
1990; Petrisek & Mor, 1999), hospice use rate in Canadian LTCFs in 2015 is even lower than
the hospice use rate in the US LTCFs ten years ago.
Residents who received hospice care in Canadian LTCFs had severe and complex
clinical needs, most were noted as end-stage, and had severe physical impairment, pain, and
high health instability. Residents who received hospice care had more severe and complex
clinical needs compared to those who did not. However, compared to those who did not
receive hospice care, end-stage residents who received hospice care exhibited higher social
engagement and were less likely to have severe cognitive impairment, depressive symptoms,
bowel and bladder incontinence. Residents who received hospice care and died within one
105

year of assessment had more severe and complex health conditions than other residents. A
predictive model was developed in the cross-sectional study, which were superior than
CHESS alone, as clinical needs were not the only determinants of hospice use. Younger age;
living in an urban LTCF; having a diagnosis of cancer; and having more severe, complex,
and acute clinical needs significantly increased the likelihood of hospice use among residents
in LTCFs in Canada. Ageism, rurality, and a diagnosis of dementia may be the barriers to
accessibility to hospice care among the LTCF population in Canada.
The work described in this thesis has several strengths. First, it was the first to use the
RAI-MDS 2.0 data to examine a subpopulation of persons in LTCFs in Canada who receive
hospice care at a national level. The RAI-MDS 2.0, a standardized resident assessment
instrument filled by trained staff, contains comprehensive and detailed information on
resident demographic, clinical and functional characteristics, special treatments and
programs, admissions and discharges and are widely used worldwide. Second, the work
described in this thesis was population-based and had a large sample size, so the selection
bias was less likely to occur and the power of the work described in this thesis was large.
Third, Andersen and Newman’s (1973) behavioral model for health service utilization was
used to guide the work described in this thesis, which helps to better understand hospice use
among the LTCF population in Canada. Fourth, the work described in this thesis included
three population characteristics components of the framework to explore their association
with health services utilization. The work described in this thesis was also one of the first to
not only explore the resident characteristics affecting utilization of health services, but also to
consider the characteristics of LTCFs. Fifth, the work described in this thesis used a mix of
admission assessments, quarterly assessments, and annual full assessments, so the study
population covered all stages in LTCFs in Canada. Last but not the least, the retrospective
106

cohort study linked the DAD to the CCRS to track an extra 5.2% residents who were
discharged from LTCFs and did not have death records in LTCFs.
The work described in this thesis also has several limitations. First, while the CCRS
database contains full coverage for Ontario, British Columbia, Alberta, Yukon,
Saskatchewan, and Newfoundland and Labrador, it has partial coverage for Manitoba, New
Brunswick and Nova Scotia, and does not cover some provinces such as Quebec and some
territories such as Northwest Territories. Second, residents who received continuing care
services in hospitals were mixed in the sample and cannot be singled out, therefore, the study
population in the work described in this thesis also include individuals receiving continuing
care services at 107 hospital-based complex continuing care facilities in Ontario and
Manitoba in addition to 1,342 LTCFs across Canada (CIHI, 2016). Third, as applied using
the Andersen and Newman’s (1973) model, enabling resources must be present for the
utilization of hospice care to take place. In the work described in this thesis, facilities that
have or do not have hospice care available could not be identified or measured due to the
limitation of the data. Consequently, the analyses were based on all LTCFs that reported to
CIHI without limiting to the LTCFs that have hospice care available. Andersen (1995) also
mentioned “we are limited more by the feasibility and costs of developing such measures
than by conceptual limitations”. Fourth, previous studies identified several factors associated
with hospice use in LTCFs, such as race or ethnicity, administrative altitudes, and LTCF
ownership (Castle, 1999; Monroe et al., 2013; Nilsen et al., 2018). The CCRS database does
not have or provide information on these variables, so the potential factors affecting hospice
use could not be explored in the work described in this thesis. Fifth, hospice use may be
underestimated as some resident who did not receive hospice care at the time of their last
assessment in 2015 may have received hospice care after the assessment.
107

Despite limitations, the work described in this thesis provides several implications for
LTCFs and policy makers. As the population ages, there will be increasing need to provide
hospice care for the growing number of residents in LTCFs. The underutilization of hospice
care is substantial and requires immediate attention. Policy makers should consider to
allocate more resources and investments in hospice care services in LTCFs. Several possible
barriers to hospice use in the LTCF population may exist including ageism, rurality, and a
diagnosis of dementia. Policy makers and LTCF administrator should consider to implement
policies and practices that could remove barriers to referral and use of hospice care services
in LTCFs. Special attention should be paid to improve access to hospice care among
residents with dementia or other progressive chronic diseases with severe and complex
clinical needs. Among those who received hospice care, around 10% did not have a DNR
order in place. There is still room to improve the use of advance directives among LTCF
residents who received hospice care. Pain was prevalent among LTCFs, especially among
those who received hospice care. Efforts from both LTCF staff and hospice providers are
needed to have pain better controlled for residents who received hospice care.
The majority of residents who received hospice care triggered more than three CAPs,
which reflects high levels of complex clinical needs within this group. A previous study
found treatment of symptoms for residents who are at end of life can still improve their
quality of life (Lorenz et al., 2008). Consideration of triggered CAPs in the work described in
this thesis provide evidence to inform a collaborative decision making process from LTCF
staff and hospice providers on whether (or not) and how issues raised by the CAPs should be
addressed in the plan of care. The work described in this thesis could be foundational to
future policy initiatives mandating end-of-life training for LTCF staff and hospice providers.

108

The work described in this thesis has several implications for researchers. Finding
from the work described in this thesis indicate that accuracy of end-stage diagnosis and
effectiveness of hospice eligibility criteria were restrictive, which may limit the benefits to
residents in terms of improvements in end-of-life care, particularly for residents with several
medical problems but no specific terminal diagnosis (Bennett, Ziegler, Allsop, Daniel, &
Hurlow, 2016). Further longitudinal study about standardized assessment of terminal status
or more effective hospice eligibility criteria with less dependency on a specific life
expectancy of six months or less is needed to help make better use of limited hospice care
resources. Among those who received hospice care, a large percentage of residents did not
have a DNH in place. It remains unknow why residents who received hospice care still
wanted to be hospitalized when possible and this requires further study. The work described
in this thesis indicates end-stage residents who received hospice care did have less severe
cognitive impairment, lower depressive symptoms, and better bowel and bladder control.
Reasons why end-stage residents with higher social engagement, less severe cognitive
impairment, and lower depressive symptoms are more likely to receive hospice care need
further investigation. Additionally, more research needs to be conducted focusing further on
the effect of enabling characteristics on hospice access among LTCF residents. Differences in
hospice use between hospital-based continuing care facilities and LTCFs and differences in
hospice use among LTCFs that have hospice care available need further investigation. This
study revealed that differences in hospice use among provinces may exist. Future study is
needed to examine differences in hospice use among all provinces in Canada. It remains
unclear how hospice care works in rural LTCFs in Canada. There is a need for further
research on the issue of LTCF-based urban-rural disparity in order to find ways to improve
access to, and provision of, hospice care services in rural LTCFs.
109

In summary, the work described in this thesis reveals that a very small proportion of
residents in LTCFs are using hospice care, which provides evidence for needs of
improvements in accessibility to hospice care in the long-term care setting. Findings suggest
several possible barriers to hospice use in the LTCF population including ageism, rurality,
and disease diagnoses. Immediate action is needed to provide improved access to, and
utilization of, hospice care in LTCFs in Canada. Special attention should be paid to improve
access to hospice care among residents with dementia or other progressive chronic diseases
with severe and complex clinical needs. Further investigation of standardized assessment of
terminal status is needed as accuracy of end-stage diagnosis continues to be challenging and
criteria for hospice eligibility are narrow.

110

References
Addington-Hall, J. (2000). Which terminally ill cancer patients in the United Kingdom
receive care from community specialist palliative care nurses? Journal of Advanced
Nursing, 32, 799–806.
Alberta Health services. (2019). Long Term Care. Retrieved from
https://www.albertahealthservices.ca/cc/Page15491.aspx
Andersen, R. M. (1995). Revisiting the behavioral model and access to medical care: Does it
really matter? Journal of Health and Social Behavior, 36(1), 1–10.
Andersen, R. M., & Newman, J. F. (1973). Societal and individual determinants of medical
care utilization in the United States. Milbank Memorial Fund Quarterly, 51, 95–124.
Baer, W. M, & Hanson, L. C. (2000). Families’ perception of the added value of hospice in
the nursing home. Journal of the American Geriatrics Society, 48, 879-882.
Bainbridge, D., Seow, H., Sussman, J., & Pond, G. (2015). Factors associated with acute care
use among nursing home residents dying of cancer: A population-based study.
International Journal of Palliative Nursing, 21(7), 349–356.
Bennett, M. I., Ziegler, L., Allsop, M., Daniel, S., & Hurlow, A. (2016). What determines
duration of palliative care before death for patients with advanced disease? A
retrospective cohort study of community and hospital palliative care provision in a
large UK city. BMJ Open, 6(12), e012576. doi:10.1136/bmjopen-2016-012576
Brink, P. & Kelley, M. L. (2015). Death in long-term care: A brief report examining factors
associated with death within 31 days of assessment. Palliative Care: Research and
Treatment, 1(9), 1–5. doi: 10.4137/PCRT.S20347.
Buchanan, R. J., Choi, M., Wang, S., & Ju, H. (2004). End-of-life care in nursing homes:
Residents in hospice compared to other end-stage residents, Journal of Palliative
Medicine, 7(2), 221–232.
Burge, F. I., Lawson, B. J., Johnston, G. M., & Grunfeld, E. (2008). A population-based
study of age inequalities in access to palliative care among cancer patients. Medical
care, 46(12), 1203–1211
Burrows, A. B., Morris, J. N., Simon, S. E., Hirdes, J. P., & Phillips, C. D. (2000).
Development of an MDS-based depression rating scale for use in nursing homes. Age
and Ageing, 29, 165–172.
Burt, J., & Raine, R. (2006). The effect of age on referral to and use of specialist palliative
care services in adult cancer patients: a systematic review. Age and Ageing, 35, 469–
476.

111

Canadian Healthcare Association. (2009). New directions for facility-based long term care.
Retrieved from http://www.healthcarecan.ca/wpcontent/themes/camyno/assets/document/PolicyDocs/2009/External/EN/NewDirectio
nsLTC_EN.pdf
Canadian Hospice Palliative Care Association (CHPCA). (2009). Caring for Canadians at
end of life: A strategic plan for hospice, palliative and end‐of‐life care in Canada to
2015. Retrieved from
http://www.chpca.net/media/7562/chpca_strategic_plan_2010_2015.pdf
CHPCA. (2013). A model to guide hospice palliative care: Based on national principles and
norms of practice. Retrieved from http://www.chpca.net/media/319547/norms-ofpractice-eng-web.pdf
CHPCA. (2014). Fact sheet – hospice palliative care in Canada. Retrieved from
http://www.chpca.net/media/330558/Fact_Sheet_HPC_in_Canada%20Spring%20201
4%20Final.pdf
CHPCA. (2016a). What is palliative care? Retrieved from http://www.chpca.net/familycaregivers/faqs.aspx
CHPCA. (2016b). What is involved in palliative care? Retrieved from
http://www.chpca.net/family-caregivers/faqs.aspx
CHPCA. (2016c). Is there a difference between palliative care and hospice care? Retrieved
from http://www.chpca.net/family-caregivers/faqs.aspx
CHPCA. (2016d). History. Retrieved from http://www.chpca.net/about-us/history.aspx
Canadian Institute for Health Information (CIHI). (2008). interRAI Clinical Assessment
Protocols (CAPs)—For use with interRAI’s community and long-term care
assessment instruments. Ottawa, ON: CIHI.
CIHI. (2015). Resident Assessment Instrument (RAI) MDS 2.0 User's Manual, Canadian
version. Retrieved from https://secure.cihi.ca/estore/productSeries.htm?pc=PCC127
CIHI. (2016). Profile of residents in residential and hospital-based continuing care 20152016. Retrieved from
https://www.cihi.ca/sites/default/files/document/ccrs_quick_stats_2015-2016_enweb.xlsx
CIHI. (2017). Continuing Care Reporting System Data Submission User Manual. Retrieved
from https://secure.cihi.ca/free_products/CCRS-DataSubManual-2017-2018-EN.pdf
CIHI. (2018a). Continuing Care Metadata. Retrieved from
https://www.cihi.ca/en/continuing-care-metadata

112

CIHI. (2018b). Discharge Abstract Database Metadata. Retrieved from
https://www.cihi.ca/en/discharge-abstract-database-metadata
CIHI. (2018c). Continuing Care Reporting System (CCRS) Data Quality Documentation.
Retrieved from https://secure.cihi.ca/estore/productSeries.htm?pc=PCC570
CIHI. (2018d). RAI-MDS 2.0 Outcome Scales Reference Guide. Retrieved from
https://secure.cihi.ca/estore/productSeries.htm?locale=en&pc=PCC1203
CIHI. (2019). Seniors living with dementia in Canada facing gap in palliative care.
Retrieved from https://www.cihi.ca/en/dementia-in-canada/spotlight-on-dementiaissues/palliative-and-end-of-life-care
Canadian Union of Public Employees (CUPE). (2009). Residential long-term care in
Canada: Our vision for better seniors care. Retrieved from
https://cupe.ca/residential-long-term-care-canada-our-vision-better-seniors-care
Canadian Virtual Hospice. (2019). How is eligibility for palliative care decided? What
procedure is followed? Retrieved from
http://www.virtualhospice.ca/en_US/Main+Site+Navigation/Home/Support/Support/
Asked+and+Answered/Palliative+Care/How+is+eligibility+for+palliative+care+deci
ded_+What+procedure+is+followed_.aspx
Carreau, L., Niezgoda, H., Trainor, A., Parent, M., & Woodbury, M. G. (2014). Pilot study
compares scores of the resident assessment instrument Minimum Data Set Version
2.0 (MDS 2.0) Pressure Ulcer Risk Scale with the Braden Pressure Ulcer Risk
Assessment for patients in complex continuing care. Advances in Skin and Wound
Care, 28(1), 28–33.
Carstairs, S. (2010). Raising the bar: A roadmap for the future of palliative care in Canada.
Retrieved from http://www.chpca.net/media/7859/Raising_the_Bar_June_2010.pdf
Castle, N. G. (1999). Hospice and nursing homes. Journal of Health and Social Policy, 11(2),
1–16. doi: 10.1300/J045v11n02_01
Centre for Health Services and Policy Research. (2015). Long-term care - evidence and
perspectives on funding healthcare in Canada. Retrieved from
http://healthcarefunding.ca/long-term-care/
Chapin, R. K., Gordon, T., Landry, S., & Rachlin, R. (2007). Hospice use by older adults
knocking on the door of the nursing facility: Implications for social work practice.
Journal of Social Work End Life and Palliative Care, 3(2), 19–38. doi:
10.1300/J457v03n02_03
Connor, S. (1998). Hospice: Practice, pitfalls, and promise. Washington, DC: Taylor &
Francis.

113

DeMiglio, L., Dykeman, S., Williams, A., & Kelley, M. L. (2012). Evolution of palliative
care in Ontario: The impact of geography, funding, and advocacy. Journal of Rural
and Community Development, 7(4), 109–121.
Duffy, A. (2005). A moral force: The story of Dr. Balfour Mount. Retrieved from
https://web.archive.org/web/20061215112019/http:/www.canada.com/ottawacitizen/s
tory.html?id=896d005a-fedd-4f50-a2d9-83a95fc56464
Erel, M., Marcus, E. L., & Dekeyser-Ganz, F. (2017). Barriers to palliative care for advanced
dementia: a scoping review. Annals of Palliative Medicine, 6(4), 365–379. doi:
10.21037/apm.2017.06.13.
Ersek, M., & Carpenter, J. G. (2013). Geriatric Palliative Care in Long-Term Care Settings
with a Focus on Nursing Homes. Journal of Palliative Medicine, 16(10), 1180–1187.
Feldberg, G. D., Ladd-Taylor M., & Li, A. (2003). Women, health and nation: Canada and
the United States since 1945. Montreal, QC: McGill-Queen's University Press.
Ferrell, B. R., & Coyle, N. (2010). Oxford textbook of palliative nursing. New York, NY:
Oxford University Press.
Foley, K. M., & Herbert, H. (2002). The case against assisted suicide: For the right to endof-life care. Baltimore, MD: Johns Hopkins University Press.
Frahm, K. A., Brown, L. M., Hyer, K. (2012). Racial disparities in end-of-life planning and
services for deceased nursing home residents. Journal of American Medical Directors
Association, 13(9), 819.e7–819.e11. doi: 10.1016/j.jamda.2012.07.021.
Freeman, S., Bishop, K., Spirgiene, L., Koopmans, E., Bothelo, F. C., Fyfe, T., … MacLeod,
M. (2017). Factors affecting residents transition from long-term care facilities to the
community: A scoping review. BMC Health Services Research, 17(1), 689. doi:
10.1186/s12913-017-2571-y
Freeman, S., Hirdes, J. P., Stolee, P., Garcia, J., Smith, T. R., Steel, K. & Morris, J. N.
(2014). Care planning needs of palliative home care clients: Development of the
interRAI palliative care assessment clinical assessment protocols (CAPs). BMC
Palliative Care, 13, 58
Fries, B. E., Simon, S. E., Morris, J. N., Flodstrom, C., & Bookstein, F. L. (2001). Pain in the
U.S. nursing homes: Validating a pain scale for the Minimum Data Set. The
Gerontologist, 41(2), 173–179.
Garand, L., Lingler, J. H., Conner, K. O., & Dew, M. A. (2009). Diagnostic labels, stigma,
and participation in research related to dementia and mild cognitive impairment.
Research in Gerontological Nursing, 2(2), 112–121.

114

Gardiner, C., Cobb, M., Gott, M., & Ingleton, C. (2011). Barriers to providing palliative care
for older people in acute hospitals. Age and Ageing, 40(2), 233–8. doi:
10.1093/ageing/afq172.
Garrard, J. (2011). Health sciences literature review made easy. The matrix method (3rd ed).
Burlington, MA: Jones & Bartlett.
Glasgow End of Life Studies Group. (2014). Women pioneers in 19th century hospice care.
Retrieved from http://endoflifestudies.academicblogs.co.uk/women-pioneers-in-19thcentury-hospice-care/
Gott, M., Ibrahim, A. M., & Binstock, R. H. (2011). The disadvantaged dying: Ageing,
ageism, and palliative care provision for older people in the UK. In M. Gott, & C.
Ingleton (Eds.), Living with ageing and dying. Oxford, UK: Oxford University Press.
Gozalo, P. L., & Miller, S. C. (2007). Hospice enrollment and evaluation of its causal effect
on hospitalization of dying nursing home patients. Health Services Research, 42(2),
587–610. doi: 10.1111/j.1475-6773.2006.00623.x
Graham, N., Lindesay, J., Katona, C., Bertolote, J. M., Camus, V., Copeland, J.
R., …Wancata, J. (2003). Reducing stigma and discrimination against older people
with mental disorders: A technical consensus statement. International Journal of
Geriatric Psychiatry, 18, 670–678.
Han B., Tiggle, R. B., & Remsburg, R. E. (2008). Characteristics of patients receiving
hospice care at home versus in nursing homes: Results from the National Home and
Hospice Care Survey and the National Nursing Home Survey. American Journal of
Hospice and Palliative Medicine, 24(6), 479–486. doi: 10.1177/1049909107305654
Hanson, L. C., Sengupta, S., & Slubicki, M. (2005). Journal of Palliative Medicine, 8(6),
1207–1212. doi: 10.1089/jpm.2005.8.1207
Harris, P. S., Stalam, T., Ache, K. A., Harrold, J. E., Craig, T., Teno, J., … Casarett, D.
(2014). Can hospices predict which patients will die within six months? Journal of
Palliative Medicine, 17(8), 894–898. doi: 10.1089/jpm.2013.0631
Hartmaier, S. L., Sloane, P. D., Guess, H. A., Koch, G. G., Mitchell, C. M., & Phillips, C. D.
(1995). Validation of the Minimum Data Set Cognitive Performance Scale:
Agreement with the Mini-Mental State Examination. The Journals of Gerontology.
Series A, Biological Sciences and Medical Sciences, 50(2), M128–133.
Hawes, C., Morris, J. N., Phillips, C. D., Mor, V., Fries, B. E., & Nonemaker, S. (1995).
Reliability estimates for the Minimum Data Set for nursing home resident assessment
and care screening (MDS). The Gerontologist, 35(2), 172–178.
Hawley, P. H. (2014). The bow tie model of 21st century palliative care. Journal of Pain and
Symptom Management, 47(1), e2–e5.

115

Health Canada. (2004). Long-term facilities-based care. Retrieved from
https://www.canada.ca/en/health-canada/services/home-continuing-care/long-termfacilities-based-care.html
Health Canada. (2016). Palliative care. Retrieved from https://www.canada.ca/en/healthcanada/services/palliative-care.html
Herr, K., Coyne, P. J., McCaffery, M., Manworren, R., & Merkel, S. (2011). Pain assessment
in the patient unable to self-report: Position statement with clinical practice
recommendations. Pain Management Nursing, 12(4), 230–50. doi:
10.1016/j.pmn.2011.10.002.
Hirdes, J. P., Frijters, D. H., & Teare, G. F. (2003). The MDS‐CHESS Scale: A new measure
to predict mortality in institutionalized older people. Journal of the American
Geriatrics Society, 51(1), 96–100.
Hirdes, J. P., Mitchell, L., Maxwell C. J., & White, N. (2011). Beyond the ‘iron lungs of
gerontology’: Using evidence to shape the future of nursing homes in Canada.
Canadian Journal of Aging, 30(3), 371–390. doi: 10.1017/S0714980811000304.
Hjermstad, M. J., Aass, N., Aielli, F., Bennett, M., Brunelli, C., Caraceni, A., …Kaasa, S.
(2016). Characteristics of the case mix, organisation and delivery in cancer palliative
care: A challenge for good-quality research. BMJ Supportive and Palliative Care.
Advance online publication. doi: 10.1136/bmjspcare-2015-000997
Hui, D., De La Cruz, M., Mori, M., Parsons, H.A., Kwon, J.H., Torres-Vigil, I., …Bruera, E.
(2013). Concepts and definitions for "supportive care," "best supportive care,"
"palliative care," and "hospice care" in the published literature, dictionaries, and
textbooks. Supportive Care in Cancer, 21(3), 659–685.
Huskamp, H.A., Stevenson, D.G., Grabowski, D.C., Brennan, E., & Keating, N.L. (2010).
Long and short hospice stays among nursing home residents at the end of life. Journal
of Palliative Medicine,13(8), 957–964. doi: 10.1089/jpm.2009.0387
Hutchings, D. (2002). Parallels in practice: Palliative nursing practice and Parse's theory of
human becoming. American Journal of Hospice and Palliative Medicine, 19(6), 408–
414. doi: 10.1177/104990910201900612
Infeld, D. L., Crum, G. E., & Koshuta, M. A. (1990). Characteristics of patients in a longterm care hospice setting. The Hospice Journal, 6(4), 81–104. doi: 10.1080/0742969X.1990.11882685
interRAI. (2017). Understanding LTCF Clinical Assessment Protocols (CAPs). Retrieved
from https://www.interrai.co.nz/assets/Documents/ESS-Information-forAssessors/Understanding-CAPs.pdf
Jayaraman, J., & Joseph, K.S. (2013). Determinants of place of death: A population-based
retrospective cohort study. BMC Palliative Care, 12(19).
116

Kaasalainen, S., Brazil, K., Williams, A., Wilson, D., Willison, K., Marshall, D., &
Taniguchi, A. (2012). Barriers and enablers to providing palliative care in rural
communities: A nursing perspective. The Journal of Rural and Community
Development, 7(4), 4–19.
Kaehr, E. W., Pape, L. C., Malmstrom, T. K., Morley, J. E. (2016). FRAIL-NH Predicts
Outcomes in Long Term Care. Journal of Nutrition Health and Aging, 20(2), 192–
198. doi: 10.1007/s12603-016-0682-5.
Kelley, M. L., Sletmoen, W., Williams, A. M., Nadin, S., & Puiras, T. (2012). Integrating
policy, research, and community development: A case study of developing rural
palliative care. In J. C. Kulig & A. M. Williams (Eds.), Health in Rural Canada (pp.
219-238). Vancouver, BC: UBC Press.
Kiely, D. K., Givens. J. L., Shaffer, M. L., Teno, J. M., & Mitchell, S. L. (2010). Hospice use
and outcomes in nursing home residents with advanced dementia. Journal of the
American Geriatrics Society, 58(12), 2284–2291. doi: 10.1111/j.15325415.2010.03185.x.
Krishnan, P., Thompson, G., & McClement, S. (2017). Predicting hospital transfers among
nursing home residents in the last months of life. International Journal of Palliative
Nursing, 23(11), 535–542.
Kwak, J., Haley, W. E., & Chiriboga, D. A. (2008). Racial differences in hospice use and inhospital death among Medicare and Medicaid dual-eligible nursing home residents.
The Gerontologist, 48(1), 32–41. doi: 10.1093/geront/48.1.32
Landi, F., Tua, E., Onder, G., Carrara, B., Sgadari, A., Rinaldi, C., …Bernabei, R. (2000).
Minimum Data Set for home care: A valid instrument to assess frail older people
living in the community. Medical Care, 38(12), 1184–1190.
Langille, 2013. Rights of passage: Integrating palliative care. Retrieved from
https://www.canadian-nurse.com/en/articles/issues/2013/november-2013/rights-ofpassage-integrating-palliative-care
Lebowitz, B. D., & Niederehe, G. (1992). Concepts and issues in mental health and aging. In:
J. E. Birren, R. B. Sloane, G. D Cohen (2 nd ed.), Handbook of mental health and
aging (pp. 3-26). San Diego, CA: Academic Press.
Lepore, M. J., Miller, S. C., & Gozalo, P. (2011). Hospice use among urban Black and White
U.S. nursing home decedents in 2006. The Gerontologist, 51(2), 251–260. doi:
10.1093/geront/gnq093
Lewis, M. J. (2007). Medicine and care of the dying: A modern history. New York, NY:
Oxford University Press.

117

Li, Y. N., Nong, D. X., Wei, B., Feng, Q. M., & Luo, H. Y. (2016). The impact of
predisposing, enabling, and need factors in utilization of health services among rural
residents in Guangxi, China. BMC health services research, 16(1), 592.
doi:10.1186/s12913-016-1825-4
Lin, C. (2010). Nursing home organizational characteristics and utilization of cancer-related
medical services (Doctoral dissertation). Retrieved from
https://scholarscompass.vcu.edu/etd/?utm_source=scholarscompass.vcu.edu%2Fetd%
2F2043&utm_medium=PDF&utm_campaign=PDFCoverPages
Lo, K. M., & Fulda, K. G. (2008). Impact of predisposing, enabling, and need factors in
accessing preventive medical care among U.S. children: Results of the national
survey of children's health. Osteopathic Medicine and Primary Care, 2, 12.
doi:10.1186/1750-4732-2-12
Lorenz, K. A., Lynn, J., Dy, S. M, Shugarman, L. R, Wilkinson A, Mularski, R. A., …,
Shekelle, P. G. (2008). Evidence for improving palliative care at the end of life: A
systematic review. Annals of Internal Medicine, 148(2), 147–159.
Lou, V. W., Chi, I., Kwan, C. W., & Leung, A. Y. (2013). Trajectories of social engagement
and depressive symptoms among long-term care facility residents in Hong Kong. Age
and Ageing, 42, 215–222.
Lun, C. (2007). The use of four care directives and hospice care in elderly nursing home
residents at admission (Doctoral dissertation). Retrieved from https://search-proquestcom.prxy.lib.unbc.ca/docview/304779648?accountid=14601
Lynn, J. (2004) Sick to Death and Not Going to Take It Anymore! Reforming Health Care for
the Last Years of Life. California: University of California Press.
Marx, T. L. (2005). Partnering with hospice to improve pain management in the nursing
home setting. The Journal of the American Osteopathic Association, 105(3), S22–
S26.
McCarty, C. E. (2011). The impact of hospice and dementia special care units on end-of-life
care for individuals with dementia (Doctoral dissertation). Retrieved from https://search.proquest.com/docview/909889237?pq-origsite=gscholar
McGregor, M. J. & Ronald, L. A. (2011). Residential long-term care for Canadian seniors:
Nonprofit, for-profit or does it matter? Montreal: Institute for Research on Public
Policy.
Miller, S. C., Mor, V., & Teno, J. (2003). Hospice enrollment and pain assessment and
management in nursing homes. Journal of Pain and Symptom Management, 26, 791–
799.

118

Miller, S. C, Mor, V., Wu, N., Gozalo, P., & Lapane, K. (2002). Does receipt of hospice care
in nursing homes improve the management of pain at the end of life? Journal of the
American Geriatrics Society, 50(3), 507–515.
Mitchell, S., Kiely, D., Hamel, M. (2004). Dying with advanced dementia in the nursing
home. Archives of internal medicine American Medical Association,164, 321–326
Mitchell, S. L., Miller, S. C., Teno, J. M., Kiely, D. K., Davis, R. B., & Shaffer, M. L.
(2010). Prediction of 6-month survival of nursing home residents with advanced
dementia using ADEPT vs. hospice eligibility guidelines. Journal of the American
Medical Association, 304(17), 1929–1935.
Miu, D. K.Y., & Chan, K. C. (2014). Under-detection of pain in elderly nursing home
residents with moderate to severe dementia. Journal of Clinical Gerontology and
Geriatrics, 5(1), 23–27.
Monroe, T. B., Carter, M. A., Feldt, K. S., Dietrich, M. S., & Cowan, R. L. (2013). Pain and
hospice care in nursing home residents with dementia and terminal cancer. Geriatrics
and Gerontology International, 13(4), 1018–25. doi: 10.1111/ggi.12049.
Mor, V., Branco, K., Fleishman, J., Hawes, C., Phillips, C., Morris, J., & Fries, B. (1995).
The structure of social engagement among nursing home residents. Journals of
Gerontology: Psychological Sciences, 50B(1), P1–P8.
Morris, J. N., Fries, B. E., Mehr, D. R., Hawes, C., Phillips, C., Mor, V., & Lipsitz, L. A.,
(1994). MDS cognitive performance scale. Journals of Gerontology, 49(4), M174–
M182.
Motiwala, S. S., Croxford, R., Guerriere, D. N., & Coyte, P. C. (2006). Predictors of place of
death for seniors in Ontario: A population-based cohort analysis. Canadian Journal
on Aging, 25(4), 363–371.
Munn, J. C., Hanson, L. C., Zimmerman, S. , Sloane, P. D. & Mitchell, C. M. (2006), Is
hospice associated with improved end‐of‐life care in nursing homes and assisted
living facilities? Journal of the American Geriatrics Society, 54, 490–495.
doi:10.1111/j.1532-5415.2005.00636.x
National Hospice and Palliative Care Association. (2016). History of hospice care. Retrieved
from https://www.nhpco.org/history-hospice-care
National Institute on Aging (NIH). (2016). What are palliative care and hospice care?
Retrieved from https://www.nia.nih.gov/health/what-are-palliative-care-and-hospicecare
Nilsen, P., Wallerstedt, B., Behm, L., & Ahlström, G. (2018). Towards evidence-based
palliative care in nursing homes in Sweden: A qualitative study informed by the
organizational readiness to change theory. Implementation science, 13(1), 1.
doi:10.1186/s13012-017-0699-0
119

Public Health Ontario. (2013). Summary measures of socioeconomic inequalities in health.
Toronto, ON: Queen’s Printer for Ontario.
Parker-Oliver, D., Porock, D., Zweig, S., Rantz, M., & Petroski, G. F. (2003). Hospice and
nonhospice nursing home residents. Journal of Palliative Medicine, 6(1), 69–75. doi:
10.1089/10966210360510136
Perlman, C. M., & Hirdes, J. P. (2008). The Aggressive Behavior Scale: A new scale to
measure aggression based on the Minimum Data Set. The American Geriatrics
Society, 56(12), 2298–2303.
Pett, M. A. (2015). Nonparametric statistics for health care research: Statistics for small
samples and unusual distributions (2nd ed). Singapore, SG: Sage.
Poss, J. W., Murphy, K. M., Woodbury, M. G., Orsted, H. L., Stevenson, K., Williams, G.
D., …Hirdes, J. P. (2010). Development of the interRAI Pressure Ulcer Risk Scale
(PURS) for use in long-term care and home care settings. BMC geriatrics, 10, 67.
doi:10.1186/1471-2318-10-67
Quality Palliative Care in Long-Term Care Alliance. (2011). Long-term care homes:
Hospices of the future. Retrieved from
http://www.palliativealliance.ca/assets/files/OLTCA_final1.pdf
Robbins, J. (1983). Caring for the dying patient and the family. London, LDN: Harper Row.
Rodriquez, J., & Boerner, K. (2018). Social and organizational practices that influence
hospice utilization in nursing homes. Journal of Aging Studies, 46, 76–81. doi:
10.1016/j.jaging.2018.06.004.
Rosenwax, L. K., & McNamara, B. A. (2006). Who receives specialist palliative care in
Western Australia–and who misses out. Journal of Palliative Medicine, 20, 439–445.
Sachs, G. A., Shega, J. W., & Cox-Hayley, D. (2004). Barriers to excellent end-of-life care
for patients with dementia. Journal of general internal medicine, 19(10), 1057–1063.
Sheehan, D. C., & Forman, W. B. (1996). Hospice and palliative care: Concepts and
practice. Boston, MA: Jones and Bartlett.
Smedbäck, J., Öhlén, J., Årestedt, K., Alvariza, A., Fürst, C., & Håkanson, C. (2017).
Palliative care during the final week of life of older people in nursing homes: A
register-based study. Palliative and Supportive Care, 15(4), 417–424.
Special Senate Committee on Euthanasia and Assisted Suicide. (1995). Of Life and Death Final Report. Retrieved from
https://sencanada.ca/content/sen/committee/351/euth/rep/lad-e.htm

120

Standing Senate Committee on Social Affairs, Science and Technology. (2000). Quality endof-life care: The right of every Canadian. Retrieved from
https://sencanada.ca/content/sen/committee/362/upda/rep/repfinjun00-e.htm
Statistics Canada. (2015a). Population projections for Canada (2013 to 2063), provinces and
territories (2013 to 2038). Retrieved from
https://www150.statcan.gc.ca/n1/en/pub/91-520-x/91-520-x2014001eng.pdf?st=5N6pJ2ap
Statistics Canada. (2015b). Long-term care facilities survey – 2013. Retrieved from
http://www23.statcan.gc.ca/imdb/p3Instr.pl?Function=assembleInstr&Item_Id=22107
6
Statistics Canada. (2017a). Census profile, 2016 census. Retrieved from
https://www12.statcan.gc.ca/census-recensement/2016/dp-pd/prof/index.cfm?Lang=E
Statistics Canada. (2017b). Population centre and rural area classification 2016. Retrieved
from https://www.statcan.gc.ca/eng/subjects/standard/pcrac/2016/introduction
Statistics Canada. (2018). Leading causes of death, total population, by age group. Retrieved
from https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1310039401
Stoddard, S. (1978). The hospice movement: A better way of caring for the dying. New York,
NY: Stein and Day.
Tabachnick B. G., & Fidell L. S. (2012). Using multivariate statistics (6th ed). New Jersey,
NJ: Pearson Education.
Temkin-Greener, H., Zheng, N. T., & Mukamel, D. B. (2012). Rural-urban differences in
end-of-life nursing home care: Facility and environmental factors. The Gerontologist,
52(3), 335–344.
The World Federation of Right to Die Association. (2016). Hospice description & history.
Retrieved from https://www.worldrtd.net/hospice-description-history
Thompson, S., Bott, M., Boyle, D., Gajewski, B., & Tilden, V. P. (2010). A measure of
palliative care in nursing homes. Journal of pain and symptom management, 41(1),
57–67.
Unroe, K. T., Sachs, G. A., Dennis, M. E., Hickman, S. E., Stump, T. E., Tu, W., & Callahan,
C. M. (2014). Hospice use among nursing home and non-nursing home patients.
Journal of General Internal Medicine, 30(2), 193–198. doi: 10.1007/s11606-0143080-x
Unroe, K. T., Sachs, G. A., Hickman, S. E., Stump, T. E., Tu, W., & Callahan, C. M. (2013).
Hospice use among nursing home patients. Journal of American Medical Directors
Association, 14(4), 254–259.

121

Virnig, B. A., Hartman, L., Moscovice, I., & Carlin, B. (2006). Access to home-based
hospice care for rural populations: Identification of areas lacking service. Journal of
Palliative Medicine, 9, 1292–1299.
Vulpen K. V. (2013). End-of-life care in the U.S. nursing homes: Resident and facility
factors that predict presence and use of services (Doctoral dissertation). Retrieved
from https://archive.hshsl.umaryland.edu/handle/10713/2778.
Welch, L. C., Miller, S. C., Martin, E. W., & Nanda, A. (2008). Referral and timing of
referral to hospice care in nursing homes: The significant role of staff members. The
Gerontologist, 48, 477–484.
Whitfield, K. Y. (2018). A case study exploring the implications of one Alberta rural
community’s experience with planning their own hospice care. The Journal of Rural
and Community Development, 13(1), 1–12
Whittemore, R., & Knafl, K. (2005). The integrative review: Updated methodology. Journal
of Advanced Nursing, 52(5), 546–553. doi:10.1111/j.1365-2648.2005.0362.x
Wu N, Miller SC, Lapane K, Gozalo P. (2003). The problem of assessment bias when
measuring the hospice effect on nursing home residents’ pain. Journal of Pain and
Symptom Management, 26, 998–1009.
Zerzan, J., Stearns, S., & Hanson, L. (2000). Access to palliative care and hospice in nursing
homes. Journal of the American Medical Association, 284, 2489–2494.
Zheng, N. T., Mukamel, D. B., Caprio, T. V., & Temkin-Greener, H. (2013). Hospice
utilization in nursing homes: Association with facility end-of-life care practices. The
Gerontologist, 53(5), 817–827. doi: 10.1093/geront/gns153
Zheng, N. T., Mukamel, D. B., Caprio, T. V., Cai, S., & Temkin-Greener, H. (2011). Racial
disparities in in-hospital death and hospice use among nursing home residents at the
end-of-life. Medical Care, 49(11), 992–998. doi: 10.1097/MLR.0b013e318236384e

122

Appendix A
Summary of Studies Associated with Hospice Use in Long-term Care Facilities
Author
Purpose
Participants
Methods
Infeld, D. L., To examine the
697 residents
• Descriptive longitudinal
Crum, G. E., characteristics of
in a LTCF
study
& Koshuta,
residents in a long• Administrative records
M. A. (1990) term-care based
(1978-1988)
hospice setting
• Descriptive analysis
from 1978 to1988

Castle, N. G.
(1999)

To examine
14,646
nursing homes with nursing homes
hospice units and
examine their
determinants and
compare nursing
homes with
hospice units to
nursing homes
with/without other
special care units

Findings Relevant to This Thesis
The characteristics of hospice use in a LTCF
from 1978 to1988:
• Number of annual admissions increased over
time
• Length of stay decreased from about 29 days
in 1978 to 12 days in 1988
• Mean age did not changed dramatically (57
years old in 1978 vs. 63 years old in 1988)
• Proportion of females (82% in 1978 vs. 43%
in 1988), whites (82% in 1978 vs. 51% in
1988), married (77% in 1978 vs. 54% in
1988) and a diagnosis of cancer (100% in
1978 vs. 75% in 1988) decreased over time

• Retrospective cohort study Predictors of having hospice unites in nursing
homes:
• Medicare/Medicaid
• Average income in the county (defined as the
Automated Certification
average income of all persons in the county)
Survey, Area Resource File
(adjusted OR=1.43, p<0.05), number of
(1996)
hospital beds in the county (adjusted OR =
• Descriptive analysis,
0.41, p < 0.01), and market competition
logistic regression
(measured by Herfindahl index) (adjusted OR
= 1.43, p < 0.05)
• For-profit ownership (adjusted OR= 1.2, p <
0.1), facility size (adjusted OR = 1.3, p <
0.05), and having other special care units
(adjusted OR = 4.75, p < 0.001)

123

Author
Petrisek, A.
C. &
Mor, V.
(1999)

Purpose
To examine the
distribution of
hospice
beneficiaries in
nursing homes

Participants
16,945
nursing homes

Methods
• Retrospective cohort study
• Provider of Service file,
Online Survey and
Certification of Automated
Records, Area Resource
File
• Chi-square statistics,
analysis of variance,
multinomial logistic
regression

Findings Relevant to This Thesis
The higher the percentage of residents
receiving the hospice benefit
• The more likely the nursing home was forprofit, belonged to a chain, lacked full-time
physician coverage
• The less likely the nursing home was in rural
counties

ParkerOliver, D.,
Porock, D.,
Zweig, S.,
Rantz, M., &
Petroski, G.
F. (2003)

To compare
hospice residents in
nursing homes with
non-hospice
residents with endstage disease in
nursing homes

363 newly
admitted
residents with
end-stage
disease: 171
received
hospice care,
192 did not
receive
hospice care

• Descriptive study
• Minimum Data Set (1999)
• Descriptive analysis

Hospice and non-hospice end-stage residents
were similar in
• Mean age (76.2 ±12.5 vs. 76.0 ±11.1);
• Cognitive performance (a mean CPS score of
2.2 vs. 2.4), physical abilities (a mean ADL
score of 24.8 vs. 25.8), depressive symptoms
(a mean DPS score of 1.9 vs. 1.5), prevalence
of incontinence (34% vs. 38%), skin ulcers
(25% vs. 27%), and dehydration (5% vs. 6%).
Compared to non-hospice end-stage residents,
hospice residents were
• More likely to have do-not-resuscitate orders
(87% vs. 66%), cancer (67% vs. 49%), and
moderate (70% vs. 59%) or severe (26% vs.
19%) pain
• Less likely to have congestive heart failure
(14% vs. 28%) and renal failure (7% vs.
16%)

124

Author
Buchanan,
Choi, Wang,
& Ju. (2004)

Purpose
To compare
residents in hospice
care at admission
to the nursing
facility to end stage
residents not in
hospice at
admission

Participants
18,211newly
admitted
residents with
end-stage
disease:
10,656
received
hospice care;
7,555 did not
receive
hospice care.

Methods
• Descriptive study
• Minimum Data Set (2000)
• Descriptive analysis

Findings Relevant to This Thesis
Hospice and non-hospice end-stage residents
were similar in
• Depression (20% vs. 21%) and severe
cognitive function (15% vs. 16%)
Compared to non-hospice residents, hospice
end-stage residents were more likely to
• Be older (a mean age of 76.3 vs. 71.8),
female (59% vs. 54%), and widowed (45%
vs. 37%) and have do-not-hospitalize (14%
vs. 8%) order and do-not-resuscitate order
(84% vs. 64%)
• Have a diagnosis of cancer (59% vs. 52%),
congestive heart failure (21% vs. 20%), and
dementia other than Alzheimer’s disease
(17% vs. 15.0%)
• Have greater cognitive impairment (a mean
CPS score of 2.6 vs. 2.3), physical
impairment (a mean ADL long scale score of
19.0 vs. 17.8), bladder incontinence (45% vs.
42%), bowel incontinence (58% vs. 52%),
and pain symptoms (74% vs. 66%)
Compared to non-hospice residents, hospice
end-stage residents were less likely to
• Have a diagnosis of hypertension (36% vs.
39%), diabetes (19% vs. 23%), stoke (12.6%
vs. 13.0%), and renal failure (7% vs. 15%)

125

Author
Hanson, L.
C., Sengupta,
S., &
Slubicki, M.
(2005).

Purpose
To describe
facilitators and
barriers to hospice
in nursing homes
and to test the
relationship
between nursing
home
administrators'
attitudes and
hospice use.

Participants
Methods
241 nursing
• Cross-sectional design
home
• Self-developed survey,
administrators,
questionnaire;
74 hospice
• Chi-square test, t-tests
administrators

Findings Relevant to This Thesis
Compared to hospice administrators, nursing
home administrators were less likely to agree
that hospice improves the quality of care for
dying residents (70% vs. 99%).

Chapin, R.
K., Gordon,
T., Landry,
S., &
Rachlin, R.
(2007)

To examine
hospice use among
hospice eligible
adults and barriers
to hospice use from
the perspectives of
the Client
Assessment,
Referral, and
Evaluation
assessors and
surviving family
members

400 Kansas
end-stage
older adults,
359 Client
Assessment,
Referral, and
Evaluation
assessors, 18
surviving
family
members

Hospice and non-hospice end-stage residents
were similar in
• Proportion of females (60% vs. 59%), whites
(96% vs. 95%), and competition rate for the
do-not-resuscitate order (59% vs. 57%);
• Physical abilities (a mean ADL score of 17.0
vs. 16.6)

• Explanatory sequential
design
• Client Assessment,
Referral, and Evaluation
assessment (1999 or 2000);
Hospice Standard Analytic
File; Minimum Data Set;
Outcome and Assessment
Information Set; Selfadministered survey
(2004); Interview
• Descriptive analysis,
independent sample t-test,
logistic regression analysis

126

Among nursing home administrators with a
hospice contract, those who agreed that hospice
improves quality of care had higher rates of
hospice use in their facilities (73% vs. 42%).

Compared to non-hospice residents, hospice
end-stage residents were more likely to
• Be younger (a mean age of 83 vs. 85),
• Have cancer (41% vs. 20%) and fewer
multiple diagnoses (53% vs. 62%).
Nursing homes in rural counties were less
likely to have hospice beneficiaries (41% vs.
24%) and hospice residents were more likely to
live in urban county compared to non-hospice
end-stage residents (68% vs. 58%).

Author
Purpose
Gozalo, P. L., To examine factors
& Miller, S.
associated with
C. (2007)
hospice use by
hospice eligible
nursing home
residents and
evaluate the causal
effect of hospice on
hospitalization

Participants
183,742 endstage residents
in five states
(Kansa,
Maine, New
York, Ohio,
South Dakota)
who died
during 1995
and 1997

Methods
• Retrospective cohort study
• Minimum Data Set,
Medicare claims and
denominator files, Provider
of Service file, Area
Resource File (1995–1997)
• Descriptive analysis,
Inverse Probability of
Treatment weighting
method, logistic regression

Findings Relevant to This Thesis
Compared to non-hospice residents, hospice
end-stage residents were more likely to
• Be married (OR=1.13, 95%CI [1.08, 1.18])
and have do-not-hospitalize (OR=1.34,
95%CI [1.21, 1.49]) and do-not-resuscitate
orders (OR=1.30, 95%CI [1.24, 1.36]);
• Have cancer (47% vs. 22%).

Han, B.,
Tiggle, R. B.,
& Remsburg,
R. E. (2008)

1170 deceased
home hospice
patients, 617
deceased
nursing home
hospice
patients

• Comparative descriptive
design
• National Home and
Hospice Care Survey
(1998), National Nursing
Home Survey (1997,1999)
• Weighted percentage, ttests, Bonferroni
adjustments

Compared to residents receiving hospice care at
homes, residents receiving hospice care in
nursing homes were more likely to
• Be 85 years and older (45% vs. 17%) and
widowed (49% vs. 29%)
• Have dementia (10% vs. 2%), noncancer
primary diagnoses (66% vs. 36%), heart
disease (23% vs. 11%), and bowels (74% vs.
34%) and bladder (62% vs. 29%)
incontinence
• Receive physician services (94% vs. 28%)

To assess
differences in
characteristics of
patients receiving
hospice care at
home versus in
Nursing homes

127

Compared to non-hospice residents, hospice
residents were less likely to
• Be male (OR=0.80, 95%CI [0.77, 0.84]),
older (OR=0.99, 95%CI [0.98, 0.99]) and
non-whites (OR=0.90, 95%CI [0.83, 0.97])
• Have congestive heart failure (31% vs. 34%)
and dementia (33% vs. 48%), greater
cognitive impairment (a mean CPS score of
3.1 vs. 3.4) and physical impairment (a mean
ADL score of 3.7 vs. 3.8)

Author
Kwak, J.,
Haley, W. E.,
& Chiriboga,
D. A. (2008)

Huskamp, H.
A.
Stevenson,
D. G.
Grabowski,
D. C.
Brennan, E.
Keating, N.
L. (2010)

Purpose
To investigate the
role of race in
predicting the
likelihood of using
hospice and inhospital death
among Medicaid
and Medicare
insured residents in
nursing homes

Participants
30,765 aged
65 and older,
non-Hispanic,
Medicaid and
Medicare
insured
residents who
died in 2000,
2001, or 2002

Methods
• Retrospective cohort study
• Medicaid Long-Term Care
Analysis file, Medicaid
eligibility and claims
records, hospice claims
file, death certificates
• Descriptive analysis,
Pearson’s correlations and
point-biserial correlation,
logistic regression

Findings Relevant to This Thesis
Hospice and non-hospice residents were similar
in
• Mean age (86.0 vs. 85.8) and years of
education (11.0 vs. 10.7)
• Prevalence of heart disease (33% vs. 38%)

To identify
characteristics of
nursing homes and
residents associated
with particularly
long or short
hospice stays

13,479
hospice
residents in
394 facilities

• Observational study
• Medicare Provider of
Service file, Online Survey
and Certification
Automated Records, and
Minimum Data Set (2001–
2008)
• Descriptive analysis,
logistic regression

Characteristics of hospice users in nursing
homes:
• Two thirds of the residents were female, three
quarters were aged 81-90 years, two thirds
were unmarried, and four fifths were whites
• Common diagnosis was dementia (35%),
cancer (19%), terminal debility (17%), and
cardiovascular disease (9%)

128

Compared to non-hospice residents, hospice
residents were more likely to
• Be female (OR=1.09, 95% CI [1.02, 1.17]),
whites (OR=1.25, 95% CI [1.16, 1.37]), and
married (OR=1.16, 95% CI [1.08, 1.25]),
• Reside in urban area (OR=1.36, 95% CI
[1.25, 1.48])
• Have cancer (OR=3.46, 95% CI [3.09, 1.87])
and dementia (OR=1.71, 95% CI [1.58,
1.84]).

Author
Kiely, D. K.,
Givens, J. L.,
Shaffer, M.
L., Teno, J.
M., &
Mitchell, S.
L. (2010).

Purpose
To examine the
association
between hospice
use and the
treatment of pain
and dyspnea and
unmet needs during
the last 7 days of
life

Participants
Methods
323 residents
• Prospective cohort study
with advanced • Minimum Data Set (2003dementia in 22
2007), interviews
Boston-area
• Descriptive analysis,
nursing homes
logistic regression, general
estimated equations

Lin, C.
(2010).

To examine the
relationships
among nursing
home
organizational
characteristics and
cancer-related
medical service
utilization
(including hospice
use)

1,183
Medicaid and
Medicare
insured
residents
diagnosed
with cancer in
396 nursing
homes

Findings Relevant to This Thesis
Hospice and non-hospice residents were similar
in
• Mean age (85.0 vs. 86.0), proportion of
females (83% vs. 85%), prevalence of do-nothospitalize orders (68% vs. 64%)
• Pain frequency (11 vs. 9 days/month).
Compared to other residents in nursing homes,
those who received hospice care were more
likely to
• Be non-whites (23% vs. 9%),
• Have eating problems (53% vs. 15%),
• Have an estimated survival less than 6
months (15% vs. 2%).

• Cross-sectional design
• Michigan Tumor Registry,
Medicare denominator file,
Medicare/ Medicaid claim
files, Medicaid eligibility
files, Online Survey and
Certification Automated
Records, Area Resource
File (1996-2000)
• Descriptive analysis, chisquare test, simple student
t-tests, logistic regression

129

Hospice and non-hospice end-stage residents
were similar in
• Age (p>0.05), proportion of whites (p>0.05)
• Cancer site (p>0.05), cancer stage (p>0.05),
and comorbidity (p>0.05).
Compared to non-hospice residents, hospice
end-stage residents were more likely to
• Be female (OR=1.65, 95%CI [1.51, 2.36]),
• Reside in a county that had greater number of
general internal medicine physicians
(OR=1.02, 95%CI [1.01, 1.02]).

Author
Lepore, M.
J., Miller, S.
C., &
Gozalo, P.
(2011)

Zheng, N. T.,
Mukamel, D.
B., Caprio, T.
V., Cai, S., &
TemkinGreener, H.
(2011)

Purpose
To identify factors
associated with
hospice use among
urban Black and
White decedents in
nursing homes

To examine
whether the
observed racial
disparities in inhospital death and
hospice use among
residents in nursing
homes were due to
within-facility or
cross-facility
variations

Participants
288,202 urban
residents aged
over 65 died
in 2006

49,048
residents in
555
New York
State nursing
homes died
between 2005
and 2007

Methods
• Cross-sectional study
• Minimum Data Set,
Medicare files, Online
Survey, Online Survey and
Certification Automated
Records database; Provider
of Service files, Area
Resource File (2006)
• Descriptive analyses,
multilevel logistic
regression

Findings Relevant to This Thesis
Compared to non-hospice residents, hospice
residents were more likely to
• Be aged 75 years and less (OR<1), be female
(OR=1.22, 95%CI [1.20, 1.24]), have do-notresuscitate order (OR=1.83, 95%CI [1.79,
1.87]), and have do-not-hospitalize orders
(OR=1.47, 95%CI [1.42, 1.52]);
• Have cancer (12% vs. 6%), Alzheimer’s
disease (13% vs. 8%), and dementia other
than Alzheimer’s (24% vs. 18%)

• Cross-sectional study
• Minimum Data Set,
Medicare inpatient claim
and hospice claim
• Descriptive analyses,
hierarchical logistic
regression mode, randomeffects model

Compared to non-hospice residents, hospice
residents were more likely to
• Be female (OR=1.15, 95%CI [1.07, 1.23),
have do-not-resuscitate orders (OR=1.51,
95%CI [1.40, 1.63]) and do-not-hospitalize
orders (OR=1.24, 95%CI [1.11, 1.38])
• Have an end-stage diagnosis (OR=7.07,
95%CI [6.42, 7.79]), cancer (OR=1.96,
95%CI [1.81, 2.13), depression (OR=1.11,
95%CI [1.05, 1.18), cognitional (CPS,
OR=1.03, 95%CI [1.01, 1.05]) and physical
(OR=1.02, 95%CI [1.01, 1.02) impairment,
severe pain (OR=1.48, 95%CI [1.35, 1.63),
and ulcers (OR=1.16, 95%CI [1.08, 1.25)

130

Compared to non-hospice residents, hospice
residents were less likely to
● Have congestive heart failure (26% vs. 31%)
and stoke (21% vs. 22%).

Author
Frahm, K. A.,
Brown, L.
M., Hyer, K.
(2012).

Purpose
To examine the
relationship
between race and
advance directives,
hospice services,
and hospitalization
at the end of life
among deceased
nursing home
residents

Participants
183,841
residents aged
65 years and
older who
died in 2007

Methods
• Prospective cohort study
• Minimum Data Set (2007)
• Descriptive analysis, linear
regression models, general
linear model, logistic and
poisson regressions

Findings Relevant to This Thesis
Compared to white residents, Asian residents
were less likely to receive hospice care
(OR=0.59, 95%CI [0.51, 0.68]) whereas
Hispanic residents were more likely to receive
hospice care (OR=1.21, 95%CI [1.12, 1.31]).

Monroe, T.
B., Carter, M.
A., Feldt, K.
S., Dietrich,
M. S., &
Cowan, R. L.
(2013)

To examine the
association
between hospice
enrolment,
dementia severity
and pain among
residents in nursing
homes

55 residents
• Cross-sectional study
with dementia • Minimum Data Set within
who died from
90 days of life
cancer
• Descriptive analysis,
independent t-tests, Mann–
Whitney tests, Kruskal–
Wallis tests, chi-squaretests, multiple logistic
regression

Hospice and non-hospice end-stage residents
were similar in
• Mean age (85 vs. 87), proportion of females
(56% vs. 53%) and whites (76% vs. 67%)
• Dementia type (p<0.05).

Unroe, K. T.,
Sachs, G. A.,
Hickman, S.
E., Stump, T.
E., Tu, W., &
Callahan, C.
M. (2013)

To describe trends
in hospice use,
hospice utilization
patterns, and
factors associated
with very long
hospice stays (>6
months)

1,452 urban
• Retrospective study
hospice
• Medicare/Medicaid claims,
residents who
Minimum Data Set (1999 are aged 65
2008)
years or over
• Descriptive analysis, chiin Indianapolis
square test, t-tests, logistic
regression

Characteristics of hospice users in nursing
homes are
• Mean age of the residents were 81(+8.2)
years old, 62% were female, 40% were
African-American
• 18% had severe cognitive impairment (a CPS
of 5 or 6), most common diagnosis was
dementia (75%) and cancer (56%).

131

Compared to non-hospice end-stage residents,
hospice residents were less likely to
• Have cognitive impairment (a mean CPS
score of 3.0 vs. 4.4).

Author
Vulpen K. V.
(2013)

Purpose
To explore the
resident and
organizational
factors that may
serve as catalysts
and barriers to
provide quality
end-of-life care in
nursing homes in
the U.S.

Participants
13,419
residents in
1174 facilities

Methods
• Cross-sectional design
• National Nursing Home
Survey (2004)
• Descriptive analysis, chisquare test, logistic
regression

Findings Relevant to This Thesis
“Factors specific to the nursing home resident
that predicted utilization of end-of-life services
included
• Older age, having an advanced directive
• Decreased bed mobility, reports of pain,
indicators of emotional distress (depression,
sadness, or anxiety), and specific diagnoses
(neoplasms and respiratory diseases)” (p. 7).

Zheng, N. T.,
Mukamel, D.
B., Caprio, T.
V., &
TemkinGreener, H.
(2013)

To examine the
relationship
between facility
self-reported endof-life practices
and residents’
hospice utilization

4,540 endstage residents
in 290 New
York State
nursing homes

• Retrospective cohort study
• End-of-Life Survey,
Minimum Data Set,
Medicare Hospice Claims,
Online Survey and
Certification Automated
Records, and Area
Resource File (2007)
• Descriptive analysis, ttests, logistic regression

Hospice and non-hospice end-stage residents
were similar in
• Mean age (87 vs. 86), proportion of married
(23% vs. 21%)
• Prevalence of diabetes (29% vs. 32%)
Compared to non-hospice end-stage residents,
hospice residents were more likely to
• Be female (70% vs. 65%) and have do-notresuscitate orders (80% vs. 68%)
• Dementia (46% vs. 43%), cancer (20% vs.
14%), and depression (51% vs. 42%).
Compared to non-hospice end-stage residents,
hospice residents were less likely to
● Have congestive heart failure (26% vs. 31%),
chronic obstructive pulmonary disease (16%
vs. 23%), and sever cognitive impairment (26%
vs. 29%).

132

Author
Unroe, K. T.,
Sachs, G. A.,
Dennis, M.
E., Hickman,
S. E., Stump,
T. E., Tu, W.,
& Callahan,
C. M. (2014)

Purpose
To describe
characteristics and
utilization of
hospice among
patients in nursing
homes and patients
not in nursing
homes

Participants
3,771 hospice
patients aged
over 65 years
from one
safety net
health system

Methods
• Comparative descriptive
study
• Medicare/Medicaid
Claims, Minimum Data Set
(1999–2008)
• Descriptive analysis, ttests, chi-square test, Wald
test

Findings Relevant to This Thesis
Compared to hospice residents not in nursing
homes, hospice residents in nursing homes
were more likely to
• Be old and female
• Have dementia, arthritis, and diabetes

Hjermstad,
M. J., Aass,
N., Aielli, F.,
Bennett, M.,
Brunelli, C.,
Caraceni,
A., …Kaasa,
S. (2016)

To describe
palliative care
organisation and
services in
participating
centres and
characteristics of
patients in
palliative care
programmes

1698 Patients
aged18 and
older at 30
centres in 12
countries
including 24
hospitals, 4
hospices, 1
nursing home,
1 home-care
service

• Descriptive study
• Single web-based survey;
self-reported symptoms;
medical data
• Descriptive statistics, oneway analysis of variance
with Bonferroni
adjustments, chi-square
test

Compared to hospice residents in hospitals,
hospice residents in nursing homes were more
likely to
● Be old (a mean age of 73 vs. 64 years);
● Have poorer functional performance and
report more tiredness.

Kaehr, E. W.,
Pape, L. C.,
Malmstrom,
T. K.,
Morley, J. E.
(2016).

To investigate the
predictive validity
of FRAIL-NH
frailty screening
tool and frailty
index (FI) for 6month adverse
health outcomes

270 patients
ages ≥ 65
years old
residing in
two long-term
care facilities
in Saint Louis

• Retrospective study
• Minimum Data Set and
chart review from JuneDecember 2014
• Descriptive statistics,
ANOVA, chi-square,
logistic regression

The FRAIL-NH frailty was the strongest
predictor of 6-month mortality (adjusted
OR=3.36; 95%CI [1.26-8.98]) and the frailty
index (FI) was a more modest predictor
(adjusted OR=2.28, 95%CI [1.01, 5.15 ]).

133

Author
Smedbäck, J.,
Öhlén, J.,
Årestedt, K.,
Alvariza, A.,
Fürst, C., &
Håkanson, C.
(2017)

Purpose
“To explore the
presence of
symptoms,
symptom relief,
and other key
aspects of
palliative care
during the final
week of life among
older people
residing in nursing
homes” (p. 417).

Participants
49,172
registered
individuals
aged 60 and
older who
died in
nursing homes

Methods
• Descriptive study
• Swedish Palliative Care
Register (2011-2012)
• Descriptive analysis

Nilsen, P.,
Wallerstedt,
B., Behm, L.,
& Ahlström,
G. (2018).

“To identify
barriers and
facilitators to
implementing
evidence-based
palliative care in
nursing homes” (p.
1).

20 managers
• Qualitative study
from 20
• Semi-structured interview
nursing homes
(2015-2016)
in two
• Conventional con- tent
municipalities
analysis
and 2
managers
responsible for
all elderly care
in each
municipality

134

Findings Relevant to This Thesis
Characteristics of hospice users in nursing
homes are
• Mean age was 86.7 years old (+7.4) and 62%
were women
• The most common underlying causes of death
were circulatory diseases (42%) and dementia
(23%), and the most prevalent symptom was
pain (59%), rattles (42%), and anxiety (33%).

Four factors constituted barriers:
• “Staff’s beliefs in their capabilities to face
dying residents, their attitudes to changes at
work as well as the resources and time
required” (p. 1).
One factor constituted facilitator:
• Leadership
Five factors functioned as either facilitators or
barriers
• Staff’s competence and confidence,
motivation, and attitudes to work in general,
as well as the managers’ plans and decisional
latitude concerning efforts to develop
evidence-based palliative care.

Author
Purpose
Rodriquez, J., To examine “how
& Boerner,
staff beliefs and
K. (2018).
work routines
influenced hospice
utilization in two
nursing homes” (p.
76)

Participants
21 staff in two
nursing homes
in
Massachusetts

Methods
• Qualitative study
• Interviews
• Thematic analysis

Findings Relevant to This Thesis
“Staff identified barriers to hospice including
families who saw hospice as giving up and
gaps in the reimbursement system” (p. 76)
“At the high- hospice nursing home (76%),
staff said hospice care extended beyond what
they provided on their own” (p. 76)
“At the low- hospice nursing home (24%), an
influential group said hospice was essentially
the same as their own end-of-life care and
therefore needlessly duplicative” (p. 76)
“Staff at the high-hospice nursing home
proactively approached families about hospice,
whereas staff at the low-hospice nursing home
took a reactive approach, getting hospice when
families asked for it” (p. 76)

135

Appendix B

136