Page 1
Massachusetts Community-Based Elder Care:
Characteristics, Care Support and the Future
A dissertation presented
By
Dorothy Marita Bausemer
to the
Law and Public Policy Program
In partial fulfillment of the requirements for the degree of
Doctor of Philosophy
Northeastern University
Boston, Massachusetts
December, 2012
Page 2
1
Massachusetts Community-Based Elder Care:
Characteristics, Care Support and the Future
by
Dorothy Marita Bausemer
Abstract of Dissertation
Submitted in partial fulfillment of the requirements for the degree of
Doctor of Philosophy in Law and Public Policy
in the Graduate School of Social Sciences and Humanities
of
Northeastern University
Boston, Massachusetts
December, 2012
Page 3
2
Abstract
The numbers of older adults continues to rise and is expected to soar to 19 million by the
year 2050. Economic and social realities make institutional long-term care increasingly
challenging and undesirable for many people. As a result, older adults are choosing to remain in
the community, often with help from informal caregivers. There are no known studies that have
looked at the demographic and clinical differences in the care giving populations; in particular,
there is limited information to help our understanding of the comparative needs of those
receiving informal care versus formal care services. Further, there is no known description of
how different regulations across states influence the number of individuals receiving informal
and formal caregiver services in the community. The present study aimed to fill these gaps in the
literature about those receiving care support services in the community. Such knowledge will
help inform needed public policies on how to best meet the needs of informal caregivers who
help keep older adults in the community.
Using details from more than 3,600 cases taken from the community based elder
Medicaid rolls in Massachusetts, this study aims were to: (1) describe the socio-demographic and
clinical characteristics among the recipients of care based on their caregiver populations, (2)
compare the differences between older adults receiving informal and/or formal care services, (3)
determine if receiving informal and/or formal services predicted functional status and eligibility
for a nursing facility, and (4) describe how regulatory changes in determining nursing facility
eligibility influences the numbers of individuals receiving informal and/or formal care services.
Page 4
3
This study found that the populations of older adults receiving informal and/or formal
services were similar across most socio-demographic and clinical variables except: continence,
and cognitive related variables. Persons in receipt of informal care supports and mixed supports
and services more frequently reported incontinence, difficulty with decision making, ability to
express themselves and be understood and have a diagnosis of Alzheimer’s disease. It was also
found that those receiving informal care support services only had lower functional status and
more likely to be eligible for a nursing facility than those receiving some formal services. There
were only minor differences between the numbers of older adults receiving informal and/or
formal caregiver services based on changes across states in regulations of eligibility for a nursing
facility. Such findings highlight that those receiving informal care support only may require
additional assistance since they are more functionally dependent and likely for nursing facility
eligibility than those receiving some formal caregiver services. As a result, the Commonwealth
of Massachusetts should explore approaches that can maximize support to those providing
informal care to offset the burden of providing care to older adults with lower functional status
than those receiving some formal caregiver services. A discussion is provided highlighting such
approaches and possible financial mechanisms/models to support the approaches. By providing
more support to informal caregivers, it is hoped that the state and federal policymakers can keep
older adults in the community and away from institutionalization, and reduce long-term care
costs.
Page 5
4
Acknowledgments
My journey through the dissertation process has been a wonderful learning experience,
full of peaks and valleys. I have gained invaluable insight and knowledge with respect to
research, and have been able to build upon the program curriculum because of the amazing
people I have had the pleasure to work with along the way. I would like to thank the members of
my committee for their patience in the face of many obstacles. Their feedback, support and
willingness to share their knowledge and time with me truly illustrate the quality of professors
found at Northeastern University.
I would like to express my deep appreciation and gratitude to my chair, Dr. Nathaniel
Rickles. His unwavering support, mentorship, patient guidance and friendship made difficult
situations palatable. Governor Michael Dukakis has been a constant support throughout my
entire program, offering his guidance and thought provoking suggestions through the years;
thank you does not say enough for his gracious encouragement. I would also like to express my
thanks to Dr. Carmen Sceppa, who has offered collegiality, supported my intellectual growth,
and has added humility to this process as an integral committee member.
I would also be remiss to not acknowledge and thank Dr. Edward Alan Miller, whose
consultation and support was immeasurable. He offered his subject matter expertise and
guidance, for which I am very thankful. In addition, I would like to acknowledge Karen
McClure, without her assistance all the pieces would not have come together quite so effectively.
My family has made innumerable sacrifices. I cannot begin to thank my husband and
parents for their continued support enough, through the struggles and tears, happiness and
Page 6
5
triumphs. As a person with a disability causing daily struggles, I needed to prove to myself and
demonstrate to my young children that regardless of limitations, goals and dreams can be
reached.
“Far away there in the sunshine are my highest aspirations. I may not reach them, but I
can look up and see their beauty, believe in them and try to follow where they lead.”
-Louisa May Alcott
The beauty of this aspiration has been met, now I can follow the sunshine to a brighter future.
Page 7
6
Table of Contents Abstract ......................................................................................................................................................... 2
Acknowledgments ......................................................................................................................................... 4
Chapter 1: Introduction .............................................................................................................................. 12
Chapter 2: Literature Review ...................................................................................................................... 18
Growing Population of Older Adults ...................................................................................................... 18
Institutionalization .................................................................................................................................. 20
Shift towards Non-institutionalized Care of Older Adults ...................................................................... 21
Policies and Laws related to Care of Older Adults ................................................................................. 22
Rebalancing Long-Term Care and Community Based Services ............................................................. 26
Informal Caregiver Support Role ............................................................................................................ 28
Characteristics that Impact Services: Formal and Informal .................................................................... 30
Background Characteristics ................................................................................................................ 31
Resources that Assist Elders ............................................................................................................... 33
Level of Disability and Need .............................................................................................................. 34
Health Care System: Valuation and Cost of Services ............................................................................. 34
Chapter 3: Theoretical Framework ............................................................................................................. 37
Andersen’s Behavioral Model of Health Service Utilization ................................................................. 38
Conceptual Framework of this Study: Modified Andersen’s Behavioral Model of Health Service
Utilization ............................................................................................................................................... 38
Population Characteristics ................................................................................................................. 40
Predisposing Characteristics ............................................................................................................... 41
Enabling Resources ............................................................................................................................. 41
Need .................................................................................................................................................... 43
Health behavior .................................................................................................................................. 43
Outcomes............................................................................................................................................. 43
Page 8
7
Environmental Variables ................................................................................................................... 44
Conclusion .............................................................................................................................................. 44
Chapter 4: Methods ..................................................................................................................................... 46
Data Sources ........................................................................................................................................... 49
Study Protocol and Sample Description ................................................................................................. 54
Dependent Variables and Analytic Methods .......................................................................................... 54
Dependent Variable 1: Care Group Categorizations ......................................................................... 55
Dependent Variable 2: Nursing Facility Eligibility ............................................................................ 56
Measurement of Independent/Control Variables .................................................................................... 58
Independent Variables ........................................................................................................................ 58
Control Variables ................................................................................................................................ 64
Data Analysis .......................................................................................................................................... 70
Aim A ................................................................................................................................................... 70
Aim B ................................................................................................................................................... 71
Aim C .................................................................................................................................................. 78
Aim D .................................................................................................................................................. 80
Conclusion .............................................................................................................................................. 82
Chapter 5: Findings ..................................................................................................................................... 84
Aim A ..................................................................................................................................................... 84
Sample characteristics ........................................................................................................................ 84
Results ................................................................................................................................................. 85
Aim B .................................................................................................................................................... 100
Factors that Predict Membership in Care Support Group Categories ............................................. 100
Results Summary for Aim B: Bivariate and Multivariate Analyses .................................................. 101
Aim C .................................................................................................................................................... 121
Missing Data and Univariate Outliers .............................................................................................. 122
Multicollinearity ............................................................................................................................... 122
Correlation Analysis ......................................................................................................................... 123
Page 9
8
Results ............................................................................................................................................... 125
Aim D ................................................................................................................................................... 130
Results ............................................................................................................................................... 131
Conclusion ............................................................................................................................................ 134
Chapter 6: Discussion and Policy Implications ........................................................................................ 136
Key Findings ......................................................................................................................................... 137
Recommendations ................................................................................................................................. 139
Problem Identification ...................................................................................................................... 139
Key Actor/Player Determination and Support .................................................................................. 139
Policy Suggestions ............................................................................................................................ 140
Funding Stream Options ................................................................................................................... 151
Limitations to the Study ........................................................................................................................ 155
Directions for Future Research ............................................................................................................. 157
Conclusion ............................................................................................................................................ 160
References……………………………………………………………………………………………….164
Appendix 1: Figure 1…………………………………………………………………………………….173
Appendix 2: Table 1……………………………………………………………………………………..174
Appendix 3: Data source………………………………………………………………………………...176
Appendix 4: Letter……………………….………………………………………………………………181
Page 10
9
List of Tables
Table 1. MDS–HC Variables used in this Study. .......................................................................... 52
Table 2. Nursing Facility Eligibility by State. .............................................................................. 58
Table 3. Frequency of Demographic Variables across Total Sample Population (N = 3,622)1 .. 86
Table 4. Frequency of Demographic Variables across Care Support Groups (N = 3,280–3,622)1
....................................................................................................................................................... 87
Table 5. Frequency of Variables Related to Clinical and Functional Needs across Total
Population Sample (N = 3,622)1 ................................................................................................... 88
Table 6. Frequency of Variables Related to Clinical and Functional Needs across Caregiver
Groups (N = 3,353–3,622)1 .......................................................................................................... 90
Table 7. Frequency of Cognitive and Psychiatric Health Variables across Total Sample
Population (N = 3,622)1 ............................................................................................................... 91
Table 8. Frequency of Variables Related to Cognitive and Psychiatric Health across the
Different Caregiver Groups (N = 3,353–3,622)1 ......................................................................... 93
Table 9. Frequency of Nursing Facility Eligibility across Total Sample Population (N = 3,622)1
....................................................................................................................................................... 94
Table 10. Frequency of Nursing Facility Eligibility across Caregiver Support Groups (N =
3,353–3,622)1 ................................................................................................................................ 95
Table 11. Frequency of Perception of Health Status across Total Sample Population (N =
3,622)1 ........................................................................................................................................... 95
Table 12. Frequency of Perception of Health Status across Caregiver Support Groups (N =
3,353–3,622)1 ................................................................................................................................ 96
Table 13. Frequency of Functional Status across Total Sample Population (N = 3,622)1 .......... 98
Table 14. Frequency of Functional Status across Caregiver Support Groups (N = 3,609–3,622)1
....................................................................................................................................................... 98
Table 15. Frequency of Cognitive Risks of Depression and Behavior across Total Sample
Population (N = 3,622)1 ............................................................................................................... 99
Table 16. Frequency of the Care Support Group by Cognitive Risks: Depression and Behavior
(N = 3,609–3,622)1 ..................................................................................................................... 100
Table 17. Bivariate Statistical Analysis Details for Hypotheses 1A to 1G (N = 3,622) ............. 102
Page 11
10
Table 18. Multivariate Statistical Analysis of Ethnicity1
and Care Support Group A (n = 3,450)
..................................................................................................................................................... 103
Table 19. Multivariate Statistical Analysis for Ethnicity1 and Care Support Group B (n = 3,450)
..................................................................................................................................................... 103
Table 20. Multivariate Statistical Analysis for Education1
and Care Support Group A (n = 3,541)
..................................................................................................................................................... 106
Table 21. Multivariate Statistical Analysis for Education1 and Care Support Group B (n =
3,450) .......................................................................................................................................... 106
Table 22. Multivariate Statistical Analysis for Primary Language1 and Care Support Group A (N
= 3,622) ...................................................................................................................................... 108
Table 23. Multivariate Statistical Analysis for Primary Language1 and Care Support Group B (n
= 3,450) ...................................................................................................................................... 108
Table 24. Multivariate Statistical Analysis for Marital Status1 and Care Support Group A (N =
3,622) .......................................................................................................................................... 110
Table 25. Multivariate Statistical Analysis for Marital Status1 and Care Support Group (n =
3,450) .......................................................................................................................................... 110
Table 26. Multivariate Statistical Analysis for Living Arrangement1 and Care Support Group A
(N =3,622) .................................................................................................................................. 112
Table 27. Multivariate Statistical Analysis for Living Arrangment1 and Care Support Group B (n
= 3,450) ...................................................................................................................................... 113
Table 28. Care Group Category A: ANOVA Analysis of Functional Status .............................. 117
Table 29. Care Group Category B: T-Test Analysis for Functional Status ............................... 117
Table 30. Multivariate Statistical Analysis for Nursing Facility Eligibility1 and Care Support
Group A (N = 3,622) ................................................................................................................... 121
Table 31. Multivariate Statistical Analysis for Nursing Facility Eligibility1 and Care Support
Group B (n = 3,450) ................................................................................................................... 121
Table 32. Impact of Care Support Group A Caregiver type on Nursing Facility Eligibility (n =
3,386) .......................................................................................................................................... 126
Table 33. Impact of Care Support Group A Caregiver Type on Nursing Facility Eligibility (n =
3,386) .......................................................................................................................................... 129
Table 34. Neighboring States Nursing Facility Eligibility (N = 3,609–3,622)1 ........................ 133
Page 12
11
List of Figures
Figure 1. Massachusetts Community Based Elder Care: Superimposed on Andersen's 1995
Healthcare Model. ........................................................................................................... Appendix 1
Figure 2. Functional status frequency histogram. ...................................................................... 115
Page 13
12
Chapter 1: Introduction
The present research is directly focused on the Massachusetts Medicaid population,
specifically those elders over 65 years of age living in the community. The Medicare and
Medicaid programs were enacted in 1965 (Social Security Amendments of 1965, 42 U.S.C. §
301 et seq.). At that time, the average life expectancy in the United States was 71.3 years
(http://www.cdc.gov/nchs/data/vsus/mort66_2a.pdf, pp. 5-1–5-8). Today, individuals aged 85
and above are the fastest growing demographic group in America. This trend is expected to
continue for at least the next 30years. In 2005, there were 4.2 million adults over 85 years old,
with a projected 5.7 million in 2010 and an estimated 19 million by 2050 per statistics compiled
by the United States Administration on Aging, based on the U.S. Census Bureau Data from 2008
(www.aoa.gov/aoaroot/aging_statistics/future_growth/docs/By_Race_and_Hispanic_origin_pers
ons_85_and_older.xls). This results in greater demand for health care for age-related changes,
chronic conditions, and disabilities, which proportionally increases functional dependence (A
Profile of Older Americans: 2011, found at
http://www.aoa.gov/aoaroot/aging_statistics/profile/2011/docs/2011profile.pdf ). Decision
makers at all government levels grapple with the question of how to provide improved services
to this growing population of older Americans, while facing profound financial constraints. As a
result of the growing number of older adults in need of assistance, the role of the informal
caregiver has gained prominence within healthcare services as a cost-efficient way to keep older
adults from more costly services and institutionalization.
Page 14
13
In response to a general lack of community based social supports and services for older
adults, the Older Americans Act of 1965 (42 U.S.C. § 3001) began to pave the way for programs
and services to assist older Americans. This Act has been amended several times since then to
enhance care giving, with support to informal caregivers and programming through formal
mechanisms. Prior to that time, and historically many elders when faced with health issues and
care needs were institutionalized. In the past 30 years there has been a move toward
deinstitutionalization and a move toward care of elders in the community. Much of
deinstitutionalization began with diagnosis or disease focused care shifts, initially those with
mental health problems. Much of the impetus for deinstitutionalization has been derived from the
Americans with Disabilities Act (ADA) of 1990 (28 CFR § 35.130). The move toward
deinstitutionalization received a landmark boost with the 1999 Supreme Court ruling in
Olmstead v. L.C (527 U.S. 581). In that decision, the Supreme Court ruled that states must
provide services, “in the most integrated setting appropriate to the needs of the qualified
individuals with disabilities.” The Olmstead Decision has been the most proactive and major
drive toward community based versus institutional long-term care services in recent years,
specifically for people with chronic care needs. Although there have been many published policy
reports that discuss outcomes and implementation concerns with regard to deinstitutionalization
and community re-entry, there is little empirical research on the topic.
Rebalancing long term care has encouraged many states to enhance community-based
options for disabled and older adults. To reduce long-term care spending and respect individual
preferences, states are employing new strategies. Some of the most common changes include
transitioning people from nursing homes back into the community, shortening institutional stays,
and re-allocating services and monies into community-based programming. In addition, for the
Page 15
14
past three decades, evolving federal policies have impelled changes in state policies because
many of these enhancements and programming changes are fueled by federal funding. Some of
the federal funding cannot be accessed unless the individual states comply with certain federal
provisions. Since1989, the inception of Home and Community-Based Waiver Services
(HCBWS) has significantly influenced the expansion of community-based programming by
states, because states could cover services without a waiver and approval by the Center for
Medicare and Medicaid Services (CMS). HCBWS are programs under Section 1915(c) of the
Social Security Act that allow states to cover services not typically part of Medicaid
programming to individuals who would otherwise need institutional care. HCBWS allow for
limited coverage areas within a state, unlike traditional Medicaid, which provides the same
coverage statewide, and they provide coverage to people with targeted needs, such as mental
health or brain injured individuals. HCBWS must demonstrate cost effectiveness and neutrality
compared to those living in institutional settings (http://aging.senate.gov/crs/medicaid17.pdf).
HCBWS allow states to offer community based alternatives for individuals who would otherwise
be institutionalized, with the eligibility criteria including meeting the same standards as
admission to a nursing facility. HCBWS allows states to be creative in design of community
programming, and services that are not traditionally Medicaid or medical services are covered
through the program.
The provisions of the Deficit Reduction Act of 2005 helped to increase growth, options,
and availability of HCBWS programs by increasing federal funding to community service
providers. The resultant programming options included adjustments in eligibility so that
individuals were not required to meet the same specific criteria for institutional care, allowing for
some freedom in eligibility standards. In addition, the Act expanded flexibility to offer services
Page 16
15
that included nonmedical social and supportive programs with the medical and nursing services
not generally allowed in Medicaid funded programs, which were now being supported by the
Centers for Medicare and Medicaid Services (http://aging.senate.gov/crs/medicaid17.pdf).
Of the many variables that influence community service allocation and long-term care,
the shortage of professional healthcare workers or formal caregivers to care for older adults
residing in the community appears to be most significant. This shortage limits availability of
services and may also decrease the quality of care provided. Informal caregivers may be the
group that can bridge the gap and provide the support and sustainability for health care provision
in this population.
Studies on informal caregivers usually address the burdens and stressors faced by these
caregivers when caring for older people and describe the socio-demographic characteristics
among those in caregiver populations. However, there is limited information about the actual
care tasks that informal caregivers perform and the reasons why some older adults remain in the
community and receive informal care services while others do not. A more detailed
understanding of the needs of the aging population will allow us to compare the services they
receive informally to those provided by formal caregivers.
To accomplish my study goals, I developed four study aims:
Aim A identifies the characteristics of Medicaid-eligible, community-dwelling older adults in the
Commonwealth of Massachusetts during the period studied. This population is associated with
higher use of formal long-term care services due to their financial situation and functional
characteristics, versus non-Medicaid individuals (Liao & Chelmow, 2007). This objective
Page 17
16
identified placement in particular care support groups. I examined care support membership in
two ways.
Care Group Category A includes: (1) no caregiver support or services (NoSS), (2)
informal caregiver support only (InfCS), (3) mixed care support—those with informal
supports and formal services—(MixS), and (4) and formal services (ForS).
Focusing exclusively on those receiving at least some support, Care Group Category B
includes: (1) informal caregiver support only (InfCS) and (2) receipt of at least some
formal services—both MixS and ForS—(SForS).
Aim B determines the variables predictive of membership in various care support groups across
selected independent variables.
Aim C examines the relationship between care support group and nursing facility eligibility
using current Massachusetts regulatory criteria.
Aim D describes the distribution of Medicaid nursing facility eligibility across care groups using
different eligibility criteria from states neighboring Massachusetts.
The present study also has a direct policy focus. As U.S. legislators strive to move the
nation toward universal healthcare, it is still uncertain what specific services will be covered and
what portion will constitute out-of-pocket expenditures for citizens. To remain fiscally prudent
in a troubled and struggling economy, it will be necessary to allocate resources to emphasize
systemic cost savings and support alternatives to traditional service provisions while preserving
competition and capitalistic incentives. In this context, and to answer the challenges brought by
limited resources and increased demand, exploration of mechanisms to provide services to this
Page 18
17
ever growing population and determination of which services are best to offer need to be
considered. One mechanism to support this population is informal caregivers, and I include
recommendations for state-level policymakers, as well as suggested legislative language that
supports informal caregivers. By providing additional support to informal caregivers, it may be
possible to keep the elderly in the community and in home settings for a longer time, thus
reducing healthcare costs to the Commonwealth of Massachusetts, the federal government, and
ultimately, to the American taxpayers. Although some of my recommendations may have
nationwide implications, because of the diversity of state implementation mechanisms, it may be
challenging to extrapolate these results to other states or to the national level without further
study.
The chapters that follow provide a detailed description of every step taken in this study,
as well as my findings and recommendations. Chapter 2 focuses on related literature and
historical events that impacted the development of this study and helped to define its scope.
Chapter 3 focuses on the theoretical framework of Andersen’s Healthcare Utilization Model and
how it provides the structure for this research. Chapter 4 outlines the research methodology,
including the databases and sampling, data collection, data analysis tools, and analytical methods
I employed to generate the findings. I share the findings in Chapter 5. Finally, Chapter 6
provides a discussion of the dissertation as a whole and provides direction for policy implications
and future research.
Page 19
18
Chapter 2: Literature Review
This study examines factors associated with the care of older people residing in the
community and how their needs are met, with a particular focus on the role of the informal
caregiver. Historically, “long-term care” has implied institutional care, usually in nursing homes
or rehabilitation centers. Increasingly, older individuals are expressing their desire to remain in
the community and to avoid institutional care. These people often depend upon the informal
services of relatives or friends to meet their needs. Because traditional health care delivery
systems focus almost exclusively on providing care in institutionalized settings or services by
paid care givers, there is little support available for patients or caregivers who struggle to remain
in the community. For numerous reasons, policy makers have recently moved to correct this
imbalance, both at the federal and state levels.
In this chapter, I examine the findings of the studies that have already been done to
chronicle this shift in emphasis from institutionally-based long-term care toward community-
based care of the elderly. I look at shifts in demographics that have overburdened the
institutional care system in recent years. I also examine political, economic, and sociological
considerations that influenced policy makers to consider revamping the health care delivery
system. In particular, because of its critical importance to the development of long-term care
policies, I consider the history of Medicaid and its influence of the decisions made by state
policy makers. Finally, I consider the literature that reports on the social and health needs of both
the elderly patients and their caregivers.
Growing Population of Older Adults
Page 20
19
As the population ages and health care continues to improve, people will likely live
longer and remain healthier. There is an expectation that over the next 20 years the number and
proportion of older adults in the United States will continue to grow, according to the U.S.
Census Bureau (www.census.gov). In 2000, there were a reported 35 million persons in the U.S.
over 65 years of age, a 12 % increase from 1990. That number is expected to double by 2030. At
this growth rate, people over the age of 65 will make up approximately 20 % of the population
by the year 2030. Chronic conditions may be better managed, but will still require monitoring
and may cause functional limitations that necessitate long-term care services. This anticipated
increased need for long-term care services poses a policy paradigm shift, to consider alternatives
to current programs and service allocations. In planning to provide the necessary expansion of
services, it is helpful for policy makers to understand the characteristics that determine how
those services are currently used.
Subsets of the aging population pose additional policy challenges. For example, the
number of people in the U.S. over age 85—deemed oldest of old—is increasing and these
individuals are the most likely to need and use long-term care services (Poon, 2005;
www.aoa.gov). It is also likely that as the oldest of old increases, there will be fewer informal
caregiver resources available. Caregivers of those oldest of old that are siblings or even children
of those very old, are also fairly old themselves and may have some physical limitations, health
related concerns, stamina issues, or even financial resource limits that younger caregivers do not
have.
Public policy makers have yet to track systematically the key factors that keep people in
the community or establish the linkages between formal and informal care. To date, their primary
Page 21
20
focus has been on the high cost of institutional care, such as nursing facilities and rehabilitation
hospitals. Policy makers understand the fiscal side of decisions to place people in the
community, and understand the societal push from family members and friends wishing for their
loved ones to be at home. Policy makers set budget limits, and Medicaid directors and program
managers have to allocate public funds. Quality of care and regulations for institutional settings
remains a priority for long-term care initiatives. In more recent times, the attention of
policymakers and legislators has moved toward improving community-based services. The focus
on deinstitutionalization and rebalancing long-term care has forced an increase in community
based services. The emphases within long-term care rebalancing efforts are: (a) to keep people
with long-term and chronic care needs in the community without sacrificing quality of care; and
(b) to make better use of the financial resources provided by Medicaid (Kassner, 2008; Miller,
Allen & Mor, 2009).
Institutionalization
Historically, the U.S. has focused attention on institutional settings. As changes in
funding and the health care system as a whole continue, and people become better consumers of
their own care and their families care, this is expected to continue to transform long-term care.
The internet and other sources of health information have blossomed over the past 20 years; as a
result, consumers are more educated and more willing to ask questions about alternatives for
long-term care. Studies over the last decade support the idea that older adults wish to age in
place—to remain in their homes and receive long-term care in the community. One study by
AARP in 2003 stated that 84% of people over the age of 50 wanted to age in place, with services
as needed (Gibson, 2003). Despite studies supporting consumer preference to remain in the
Page 22
21
community, there continues to be a tendency for states to have more monies allocated to
institutional settings by Medicaid (Roche, 2007; Gibson, 2003; www.ageinplace.org).
Shift towards Non-institutionalized Care of Older Adults
Throughout the last thirty years, advocates and policy makers have worked tirelessly to
develop effective and appropriate alternatives and options for people to remain in the community
and out of institutions. The strategy for older adults involved diversion and delay of
institutionalization, while the focus for disabled younger people has been on re-entry into the
community. There have been more attempts to transition nursing facility residents regardless of
age back into the community over the last few years, with some degree of success. Both of these
approaches demonstrate that some institutionalized people can have their needs met in alternative
settings and thrive in the community despite their disability, while others do improve while
institutionalized. This finding supports providing options and opportunities for older adults with
care needs. (Kassner, et al., 2008; Kasper & O’Malley, 2006)
In addition to efforts to keep people out of institutional settings, there has also been an
increased push to transition people who are institutionalized back into the community. This
interest is motivated by the desires of patients and their caregivers to return to the community as
well as the aim of policy makers to curb rising long-term care spending while enabling people to
live in the least restrictive environment. Recent policy decisions and court rulings have held that
the disabled and older adults have the right to receive services in community settings. These
rulings place pressure on state and federal governments to rethink the balance between
community and nursing facility services. Starting with the Americans with Disabilities Act
(ADA) in 1990 (28 C.F.R. 35.130), caring for people in the least restrictive environment has
Page 23
22
been a goal for both federal and state policy makers. As the ADA evolved and with the onset of
the Olmstead decision in 1999(527 U.S.C. . §581), there was an increased focus on rebalancing
long-term care and on helping elders and disabled people live in the community. The Olmstead
decision stated that people who can be cared for in the community should be able to remain
there, and that public monies should not be used to keep disabled people institutionalized. The
Olmstead decision declared that institutionalization was against the Americans with Disabilities
Act; because no age was specified, elders are included in these rulings. Incentives were given to
states, through the Olmstead decision and with dedicated allocations, to keep people who were
eligible for nursing facility-based care in the community and to discharge people from
institutional settings.
Policies and Laws related to Care of Older Adults
Deinstitutionalization has been driven by many federal mandates, laws and Acts. Most of
these laws were bred out of the necessity to regulate and insure safe and effective care and
support to people. Nursing facility care and treatment was brought to public attention in the
1960’s; and, in 1963 the U.S. Senate Special Committee on Aging began holding hearing and
issuing reports around the care. In 1967 the first set of regulatory standards were issued, which
paved the road toward legislation for other settings in the future. The Omnibus Budget
Reconciliation Act of 1980, 1987, and 1990 placed requirements on nursing facilities, focusing
on national minimum standards. (42 U.S.C. §9902(2); §483.25)(Harrington, Mullan & Currillo,
2004; Capitman, Leutz, Bishop & Caster, 2003). OBRA 1993 is better known as the Deficit
Reduction Act of 1993 (42 U.S.C. §1396) included tax increases and cuts in appropriations,
which led to the closure of many nursing facilities (H.R. 2264, 103d Congress; available at
Page 24
23
gpo.gov). To handle the flood of individuals who would lose their living and care arrangements
as a result of these closures, policy makers began to consider alternatives to institutionalization,
particularly examining options that enabled elders to remain in the community
The Americans with Disabilities Act of 1990 (28 C.F.R. §35.130) also had a profound
effect on deinstitutionalization, but its focus was more on the disabled population than elders.
The ADA expects providers in both community and facility settings to provide their consumers
with not only the services they need, but those services in a safe and appropriate manner. Within
the ADA there is another act, the Civil Rights of Institutionalized Persons Act (CRIPA) which
specifically deals with people who are institutionalized (42U.S.C. §1997), authorizing the U.S.
Attorney General to investigate reported incidents or suspected issues. (www.ada.gov)
The Balanced Budget Act of 1997 (42 U.S.C. § 395) caused spending reform for
Medicare and changes in payment methodologies to nursing facilities. (The Library of
Congress, Pub L. 105–33, 111 Stat. 251). As a result, many facilities closed, stimulating efforts
to move elders out of nursing facilities and into community-based programs.
Although not legislation, the Olmstead decision greatly influenced the way the
Americans with Disabilities Act was interpreted. In 1999, The Supreme Court ruled on Olmstead
v. LC, now known as the Olmstead decision, which stated that people should be cared for in the
least restrictive environment possible and that unnecessary institutionalization was a violation of
the ADA. Specifically, the Olmstead decision stated that it was a violation of the Americans with
Disabilities Act to institutionalize people unnecessarily, and that long-term care needs must be
met in the least restrictive environment possible. The New Freedom Initiative was issued in 2001
by President George W. Bush, calling for a comprehensive assessment of federal policies,
Page 25
24
programs, and regulations that are barriers to older adults and persons with disabilities to remain
in or be re-introduced into the community. The New Freedom Initiative assisted States to better
develop infrastructure to support the rebalancing of long-term care and the transitioning of
people out of nursing facilities. In the years following the New Freedom Initiative, the Centers
for Medicare and Medicaid Services (CMS) have awarded over 300 grants, totaling over $270
million dollars, to create systematic change to the long-term care system (Kassner, 2008;
www.hhs.gov/grants retrieved 7/15/2010; www.hhs.gov/newfreedom retrieved 7/15/2010).
The Olmstead decision influenced the movement of disabled people out of facilities and
into community-based, less restrictive independence. This act helped propel the community-
based services cause, striving toward community placement for those who could thrive in a
community setting. Although nursing facility census numbers have dropped slightly, there has
been massive growth in community-based services (Bovbjerg & Ullman, 2002). This
information poses a question: are people using community based services they would not
otherwise have used.
Part of the Deficit Reduction Act of 2005 (42 U.S.C. §1396) was an initiative known
popularly as Money Follows the Person. Congress broadly defined it as:
…elimination of barriers or mechanisms, whether in state law, the state Medicaid plan,
the state budget or otherwise, that prevent or restrict the flexible use of Medicaid funds to
enable Medicaid-eligible individuals to receive support for appropriate and necessary
long-term care services in the settings of their choice. (S.1932 Section 6071)
Page 26
25
The Deficit Reduction Act of 2005 also authorized a second wave of Money Follows the
Person demonstration grants, which provide enhanced Medicaid-matched funds to states to pay
for the first year of home and community-based services for long-stay individuals who transition
from a nursing facility to the qualified community residence. The DRA allows for additional
flexibility in deinstitutionalization and community re-entry for elders being discharged from
nursing facilities.
The Older Americans Act Amendments of 2006 (42 U.S.C. § 3001) expanded the family
caregiver definition to include qualified grandparents, adjusting the age from 60 to 55 years old,
and including adopted children. These amendments also identified priority to caregivers of
people with Alzheimer’s disease as well as for grandparents caring for children with severe
disabilities. Expansion of the formal definition of caregiver at the federal level allows for
additional people to be paid as formal caregivers under Medicaid programs that would otherwise
fall under the definition of an informal caregiver. This allows a person to be considered a
caregiver and employed, which may make significant differences in a person’s life.
(www.aoa.gov)
In early 2010, Congress passed the Patient Protection and Affordable Care Act, also
known as Obamacare or ACA. A key provision of that law is to develop comprehensive, cost-
effective medical coverage for all. As the law is implemented, determining which services to
cover, and to what extent, will be an ongoing challenge. Healthcare was a hot topic during the
recent Presidential election debates and was evidently a factor in the re-election of President
Obama, although support for the legislation is hardly universal. With the election outcome, it is
now safe to assume that Obamacare will not be repealed; however, there remains much to be
Page 27
26
worked out, not only pertaining to Obamacare, but also with regard to how Medicare and
Medicaid will be managed. As the government moves forward to implement Obamacare and
reform Medicare, some of the key questions yet to be answered include how states will manage
the insurance exchanges mandated by Obamacare, what conditions and procedures will be
covered and to what extent, and how Obamacare will interface with Medicare and Medicaid.
One aspect of coverage that must be evaluated is long-term care, and regardless of
political view, long-term care is here to stay. Currently about 69 % of long-term care
expenditures are covered by federal and state taxpayers (Feder, 2007). Healthcare alone
represents 17.6% of the gross domestic product within the U.S.
(http://healthreform.mckinsey.com/Home/Insights/Latest_thinking/Accounting_for_the_cost_of_
US_health_care.aspx). Although the Obamacare expands coverage, it also specifies greater
incentives to encourage states to increase access to home and community-based services through
the Medicaid program, provides training of family caregivers, and expands access to care
coordination for persons in need of long-term care services. (The Library of Congress Pub L. 11–
148, 124 Stat. 119)
Rebalancing Long-Term Care and Community Based Services
Rebalancing of long-term care has affected care giving services in the community,
increasing formal services available. Studies have attempted to identify whether the increase in
community-based formal services has affected the role of the informal caregiver and caused
replacement of formal care services for informal caregivers. Long-term care and rebalancing
toward community based options is not a new concept and research has been done for many
years. In 1982, the Department of Health and Human Services commissioned the Census Bureau
Page 28
27
for an evaluative study, and the Long-Term Care Study was completed. This study was a first
step in looking broadly at the needs and demographics of elders and disabled people. Since that
time, aspects of this study have been replicated or delved into more extensively, and because of
its well-formulated data analysis, this study forced policy makers to take notice of long-term care
needs.
When Medicaid was established in 1965, nursing facilities were viewed as the only
option for long-term care. As times changed, so did this ideal. From deinstitutionalization
through the legislative changes encouraging community programming, the focus of long-term
care has morphed from institutions to community-based services.
Home and Community-Based Services (HCBS) Waivers provide a mechanism for
Medicaid agencies to expand programs beyond traditional medically restricted federal Medicaid
reimbursement guidelines. HCBS waivers are a very diverse program group, allowing freedom to
states to create innovative options for Medicaid recipients in the community who meet nursing
facility clinical eligibility regulatory standards. For purposes of this study the focus is only on the
elderly, but HCBS waivers include programs for those with brain injury, mental retardation,
HIV/AIDS, and many other targeted populations. HCBS waivers were initiated through the
Omnibus Reconciliation Act (OBRA) 1981(42 U.S.C. . § 9902(2)), and funded in part 1915(c) of
the Social Security Act, allowing more freedom to include social, support, and non-medical
services. Each waiver program must be approved through the Centers for Medicare & Medicaid
Services (CMS).
There were 38 pieces of legislation introduced in the first session of the 110th
Congress
related to long-term care and/or the caregiver (not related to veterans), all of which have been
Page 29
28
reviewed. Long-term care is a salient and important issue; how best to meet the needs of elders
will continue to be an area of much growth and discussion. Much of the regulatory focus around
nursing facilities has been based on the quality of care provided. These concerns can be traced
back to the 1940s and 1950s, when there were low federal payments, unsanitary conditions, and
overcrowding. Community-based services and long-term care provisions have been based state
and federal legislation and regulatory standards, many of which are reviewed here because they
impact the study population.
Informal Caregiver Support Role
Informal care supports are non-paid caregivers, such as family, friends, and volunteers.
As an older adult has increased needs, a family member may begin to help with instrumental
activities of daily living (IADL), including money management, household maintenance, chores,
and shopping. As function declines, IADL assistance often transitions into help with activities of
daily living (ADL), such as bathing, grooming, dressing, eating, and continence care. At some
point it may become advisable for the elder and the informal caregiver to move in together, but
this is a tough decision for both parties. Nevertheless, according to the National Survey of
Families and Households (1992), 38% of informal care giving is an adult child assisting an
elderly parent. (Takamura,1998)
Informal care, although not an expenditure of Medicaid, can be a burden. Social cost and
resources need to be considered. Often, caregivers must restructure their lifestyle. Caregivers
may change or stop working to care for an older adult loved one. According to Tamakura (1998),
many caregivers for older adults are also:
Page 30
29
…balancing work and care of young children. Many caregivers are foregoing
opportunities for career advancement and salary increases, electing instead to modify
their work schedules so they can fulfill what they see to be their family obligation.
In 1981, Dorothy A. Miller coined the phrase the sandwich generation to describe these
middle-aged adults who must balance caring for their children, or even grandchildren, with
caring for their own parents. These individuals are in a difficult situation, often forced to make
life altering decisions or brutal compromises. Caregiver stress, including physical strain and
financial hardship, is a strong predictor of nursing facility entry, especially over time. (Spillman
& Long, 2009; Tamakura, 1998).
Many feel that informal supports should be encouraged, with complementary formal
supports as needed. Several of the empirical studies have focused on the relationship between
formal and informal care. These studies found no correlations between expanded formal
community-based services and withdrawal of informal sources. Withdrawal of informal supports
and the replacement of service needs by formal support would invariably increase cost. Informal
care can reduce medical expenditure only if it substitutes for formal care.
There must be an understanding of the relationship and co-dependency between informal
and formal supports, through service, care need, and costs. As Medicaid has offered and funded
more community-based services, both long-term care recipients and caregivers have learned of
and become more accepting of these services. It may be possible that with the abundance of new
community-based supports being offered, elders may be electing to stay in the community
longer. With longer community stays, there may also be increases in ADL limitations, in turn
increasing the burden on their caregivers, thus stimulating caregivers to urge utilization of formal
Page 31
30
care. Additionally living with others, has been shown to correlate with higher ADL limitations
even when they claim they have no caregiver, suggesting that that perhaps people help without
acknowledging what they are doing is informal care support, or that living with others gives
people with functional limitations encouragement to be more independent. (Engleston, Rudberg
& Brody, 1999; Graber, Liao & Buchanan, 2002).
Taking notice of the informal caregiver is a first step. Understanding what services the
informal and formal caregivers provide and how cost can be minimized is a policy maker’s
struggle. Policy makers require a clear understanding of everything that influences the care of the
individual in the community if they are to make policy decisions that meet the needs of the elders
and caregivers, while also protecting the public coffers.
Characteristics that Impact Services: Formal and Informal
Better understanding of the population being studied and their resource needs and
impacts on the healthcare system will drive policy decisions. Issues pertaining to the Medicaid
population and elders generally are inherently complex. Population characteristics and general
demographics will be used to describe the basic features of the elders being studied.
Many characteristics of the elder, their caregiver, their living arrangement, services
available and desire to seek assistance influence and impact service provision be it formal or
informal care. Review of these many interlocking aspects of an elder and their care needs offer a
better understanding of the holistic person. Much of this study focuses on a better understanding
of the population, through intense investigation of these variables. Once these variables are fully
understood, clearer policy recommendations can be made based on solid research and need.
Page 32
31
Background Characteristics
In this study I looked at the interplay between caregiver type, formal services and
informal supports, and various characteristics such as race and ethnicity, marital status, level of
education, primary language spoken, and living arrangement.
Ethnic differences and caregiving have been reviewed in numerous studies. Pinquart and
Sorensen (2005) completed a meta-analysis to integrate the results of over one hundred studies.
In this meta-analysis they found ethnic minorities were more likely to have an informal caregiver
and that an informal caregiver was more likely to feel they had a filial obligation to be a
caregiver, than those that were Caucasian. They also found that minority elders were more
functionally dependent than their Caucasian counterparts. These ethnic minority elders and
caregivers reported worsened physical health, with Hispanics and Asians reporting much
worsened depression than either African-Americans or Caucasians. This analysis supports the
idea that effective intervention must consider cultural differences among ethnic groups.
Several studies have explored the role of filial obligation in caregiving, both
independently and along with race and ethnicity. African-American and Caucasian middle-aged
adults typically are in the sandwich generation. Researchers have begun to explore the
differences in these specific caregivers and the roles they have; these studies reveal Caucasian
middle-aged adults provide more support to their grown children than African-American adults
do, but the African-American adults provide more support to their elders (Fingerman, et. al.,
2011).
Level of education has been negatively associated with informal care provision by adult
children or children-in-law, meaning those with higher levels of education are less likely to have
Page 33
32
adult children as caregivers. Byrne, et al., (2008) found that family members of more highly
educated parents have less incentive to spend time caring for them, suggesting that more highly
educated people rely more heavily on formal care. Another study found the likelihood of
receiving informal or formal care decreases with income and education: Johnson and Wiener
(2006) found that three-quarters of frail adults below the federal poverty level received paid or
unpaid care in 2002, versus less than half of those with incomes of 400% of the poverty level.
Types of care utilized has also been associated with levels of education, with higher levels of
schooling associated with formal and institutional care.
(http://www.ahrq.gov/research/ltcusers/ltcuse1.htm)
Communication is an important aspect of care for all people. Fostering relationships and
positive communication are key aspects of clinical care of elders. Language difficulties and
communication issues often cause strain on relationships. Much of the research is focused on
discrimination and quality of care around language barriers, versus the caregiver differences.
Much of the caregiver research is focused on frustration with an inability of the elder to express
needs effectively. Cultural competency is often an extension of language. (Spencer, Martin,
Bourgeault, & O’Shea, 2010).
Marital status is an interesting element to review as a variable associated with informal
care support. Spouses often do not see the care they provide as care, but as their responsibility.
Many of the studies look at differences between spouses and children caregivers or at ethnic
differences in spousal caregivers. One study (Feld, Dunkle & Schroepfer, 2004) found that ethnic
differences were related to marital status. For example, they found that unmarried Hispanic
elders were more likely than African-American or Caucasian elders married or unmarried to
Page 34
33
have a combined caregiver type of both formal and informal services. Married African -
American elders were more likely to have informal caregivers other than their spouse, when only
informal caregivers were present. This study, by Feld, S; Dunkle, R.E.; & Schroepfer, T., (2004)
did not investigate who those other caregivers were to the African-Americans, they may have
been outside help or even children. There have been studies investigating childless people being
at risk in older age for inadequate support. Larsson and Silverstein (2004) looked at whether
marital status and parenthood increased chances of staying at home in the community and
decreased institutionalization; they found that being a parent considerably increased the odds of
receiving informal care support at home.
Resources that Assist Elders
Engleston, Rudberg & Brody (1999) examined the relationship between prior living
arrangements and average ADL function upon nursing facility admission and found that older
Medicaid recipients who lived alone prior to entering nursing homes had better physical
functioning than those who lived with others. Another study by Egleston (1999) showed similar
results: those who lived alone prior to nursing facility entrance had fewer ADL deficits than
those who lived with others. These two studies suggest that elders who live alone function at a
higher level than those who live with others, being more independent. Additionally, the studies
found that those who lives alone prior to nursing facility admission were admitted with better
functioning. These studies suggest that elders that live with others stay longer in the community,
and one can assume that they may also be having care needs met by their housing companion.
Most studies look at informal or formal care services, or how one may substitute for the
other. One study by Pezzin, Kemper and Reschovsky (1994), started with the understanding that
Page 35
34
there are often other influences that cause decisions to use formal or informal care services and
supports. They specifically looked at living arrangement and substitution of informal for formal
services, and found only small effects.
Level of Disability and Need
ADL deficit and disability was reviewed by Li (2005) over a two-year period to
demonstrate whether people in the community remain stable. Li found that those who remain in
the community indeed remain maintain a stable level of ADL deficit, whereas those who are
institutionalized demonstrate rapid decline in ability to perform ADLs. Disability level has been
related to informal care use and likelihood of admission to a nursing facility; using only informal
care people had lower disability levels and increasing disability levels correlated with being in a
nursing home. (http://www.ahrq.gov/research/ltcusers/ltcuse1.htm). Those who demonstrate
increased functional disability for between three and six ADLs are less likely to only have
informal care services, whereas those with less disability may be in the community with informal
services only. This suggests that it is difficult to manage people with high levels of disability in
the community without informal supports. (Fhttp://www.ahrq.gov/research/ltcusers/ltcuse1.htm)
Health Care System: Valuation and Cost of Services
Public policies and managing the needs of elders that are utilizing public assistance, such
as Medicaid encompasses prudent spending and allocation of services. Takamura (1998), the
Assistant Secretary for Aging for the Department of Health and Human Services, estimated in
1998 that if informal caregivers were to be replaced by formal paid home care workers, the cost
would be between $45 billion and $94 billion dollars a year. In 2012 dollars, that cost range
would be more like $63 billion to $133 billion dollars a year, an 41.9% annual rate change
Page 36
35
(www.usinflationcalculator.com).Valuation of care services, including formal and informal care,
is difficult. Completing a time study and attaching an hourly wage may sound simple, but it does
not account for the elders acuity or psychological needs or the caring aspect. There are many
valuation articles that attempt to place monetary amounts on care, especially the informal care
support role. A 2009 study looked specifically at co-residential informal care (Mentzakis,
McNamee & Ryan, 2009). This study acknowledged the importance of informal care and
attempted to understand the factors that influence the likelihood and amount of informal care
provided. They were able to determine that co-residential informal care competes with other time
consuming activities within the household (and often within the family), and informal supports
substituted for formal services.
Health care costs are soaring. Current national estimates of nursing facility costs are
$62,532 to $70,912 per year; Massachusetts average rates are higher, between $93,429 and
$102,956 per year (Genworth Financial, 2006). Community-based living costs need to consider
housing, room and board, utilities (heat, electricity, gas, telephone), medical and nursing
services, care giving (which may include a home health aide and attendants), and meals on
wheels. Economic valuation of informal care has become a topic of much speculation and
research. Methods to be more inclusive on total costs and valuation of forgone activities are now
additionally being considered in valuation estimates (van den Berg, et al., 2006). These different
ways of looking at costs make it difficult to apply consistently. Some researchers feel that
informal care is invaluable and cannot be measured, whereas it provides the assistance older
adults need to remain at home, avoiding costly services and institutionalization (Dooley, Shaffer,
Lance & Williamson, 2005). A large portion of formal long-term care service is paid for through
Medicaid and Medicare. Estimate state that Medicaid finances roughly half (49%) of all long-
Page 37
36
term care expenses, Medicare covers about twenty (20%), and families and private insurance
finance the remaining 18% and 7% respectively (Feder, 2007). Nursing facility care can have a
high annual price tag, and due to that and a push toward deinstitutionalization, states are devising
increasingly creative measures to support people in community-based settings in an effort to
conserve resources and save money.
Page 38
37
Chapter 3: Theoretical Framework
This chapter addresses the conceptual framework for the study based upon Andersen’s
Healthcare Utilization Model (Andersen, 1968, 1995), specifically Andersen’s Behavioral Model
of Health Service. I describe the variables within the model, how they relate to the study, and
relevant literature.
I wanted a theoretical framework that would not only pull together the holistic aspects of
elders who receive either formal or informal services and support, but that also addressed their
needs and the variables that influence the choices they make. The policies developed by state and
federal legislators and lawmakers are the gateways to enhanced programs, but are also often the
pathways to the cracks in service allocations that people fall through. Economically, state and
federal programs are limited in the ability to deliver services to all. These limitations to services
also cause gaps in care for people just above the cut off for eligibility. Because of this, I wanted a
framework that would capture the needs of individuals, the care they receive, how that care is
provided, and the policies and laws that influence the care delivered.
I examined other theoretical frameworks as well as Andersen’s Behavioral Model of
Health Service. Except for the Andersen model, each theory and model that I reviewed met the
needs of the study in some aspects, but fell short in others. Most did not provide the necessary
data on an individual level: information about the demographics and living situations of each
person, as well as their care needs, the care options they have, the kind of care they receive, who
provides their care, external stressors, and all of the characteristics that influence their needs. I
considered the Disablement Process Model because it looks at disability and functional
Page 39
38
consequences; the model fit nicely with regard to functional deficit but did not address the policy
implications and the health care environment. I rejected the Stress Adaptation Model because it
did not fit all aspects of the study. This study has the broad outlook that mirrors Andersen’s
Behavioral model.
Andersen’s Behavioral Model of Health Service Utilization
In 1968, Andersen proposed a model that employed multivariate relationships to
comparatively analyze health care utilization and structure.
Since that first iteration, Andersen has adjusted the model to better meet the needs of
contemporary research, adding characteristics and expanding definitions. (Aday & Andersen,
1974; Andersen, 1995) The evolution of the model reflects a shift from the individual to an
interwoven focus on the combined Gestalt-like effects that an individual, the health care system,
policy change, and the external environment have on each other and the system as a whole.
For the reasons described above, the Andersen model provides a framework for the
multivariate analysis in this study. This model has been used extensively within health policy,
gerontology, and healthcare service allocation studies to drive studies of several aspects of
healthcare. For example, Renn (2005) used the model to evaluate efficiency of services. Couture,
Nguyen, Alvarado, Velasquez, & Zunzunegui, (2008) used the model in studies relating to
equity, health disparity, and access to services. Fuller-Thomson (2008) found the model useful to
identify factors associated with or that influence how people use health services.
Conceptual Framework of this Study: Modified Andersen’s Behavioral Model of
Health Service Utilization
Page 40
39
This study focuses on the characteristics of the population of adults over age 65 that
resided in the community from 1999–2004, received Massachusetts Medicaid, and requested
long-term care services. I analyzed factors that influenced the care and services provided to this
population. Figure 1, found in appendix 1, presents the conceptual model being used, as modified
from Andersen 1995, with the aims and hypotheses of this study superimposed.
Figure 1. Massachusetts Community Based Elder Care: Superimposed on Andersen’s
1995 Healthcare Model
This model theorizes that the changes in policy at both the state and federal level may
directly influence how health services are allocated and furthermore how recipients use those
services. My study takes this model and adds the study period and population to the theory. This
includes the expectation that sequential changes in health policy influence allocation of funding
and services, thus affecting the population. The model suggests that each aspect of the model
links back to the various parts, that better understanding of the population characteristics,
including predisposing, enabling and need characteristics, will help to drive outcomes data and
further influence the development of effective policies. Andersen’s model fit this study well
because it allows a rich understanding of the elder as a holistic person, encompassing the many
aspects that effect their care from their care needs, to the ;location of services and the service
availability. Additionally it allows for the understanding of the population and how the
environmental system and policy factors impact elders and outcomes.
Andersen’s model suggests that there are direct influences between the environment,
including the health care system, the population characteristics, and the outcomes. To understand
Page 41
40
how effective the model predictions are, the following evaluative elements must be understood
and incorporated into the analysis:
Understand how legislative policy changes occur, which is important because each aspect
influences outcome;
Recognize of the populations’ needs and utilization of services;
Determine availability of formal community-based services paid for through Medicaid
covered plans or other State or federally funded services, because if programs and
resources are not available placing and keeping elders in the community may be unsafe;
Realize the circumstances pertaining to informal care provided by unpaid caregivers,
because a better understanding of the supports provided will assist in creating appropriate
legislative language and programs of support; and
Recognize the circumstances under which individuals seek assistance, because programs
developed without this understanding are likely to fail.
My study focused on the thorough evaluation of population characteristics and resource use,
because through this you can better identify the direction for environmental impact and health
care system change and policy implication suggestions. Additionally, outcomes of elders care
and needs will be better understood once population characteristics are determined clearly.
Population Characteristics
I examined several variables relating to specific characteristics of the study population,
including certain attributes that may predispose individuals to seek health services (predisposing
characteristics), knowledge of what resources are available and how to access them (enabling
Page 42
41
resources), and need for services. I describe these variables in greater depth in the following
sections. The population in this study includes individuals who were at least 65 years of age
during the years studied, who received assistance from Massachusetts Medicaid, and who
requested long-term care services. These people may be affected by changes in the health care
system and the introduction of new health policies.
Andersen’s model assumes that there is a relationship among the population
characteristics, that the predisposing characteristics affect enabling resources, which in turn
affect need. When using his model to further understanding of preventive care, the relationships
he suggests and causality do make sense. In this study, the focus is on the factors that influence
cross relation of the factors and characteristics. For example, need may drive the enabling
resources, but not be the end focus.
Predisposing Characteristics
Predisposing characteristics are personal traits, behavioral tendencies, or circumstances
that may affect the use of health services. For example, in this study I have hypothesized that
people with a high level of education may use more preventive care and formal services, and
people who are married may use more informal care supports. Additionally, in this study I have
hypothesized that people from different ethnic backgrounds may differ in their use of informal
care supports. The predisposing characteristics I considered are ethnicity and race, gender,
marital status, primary language spoken, and level of education.
Enabling Resources
Page 43
42
Before individuals can make a decision to use health care services, most need to know
what services are available and must possess the means to access the services and pay for them.
This is the enabling resources component identified in the Andersen model. For example, to
obtain formal care services, an individual must either have health insurance that covers the
services or have access to personal income to pay for the services. Those who wish to use
informal care supports must have a network of individuals willing and able to provide informal
care. All of the subjects in this study were covered through Massachusetts Medicaid. To qualify
for Medicaid, income level must fall below a poverty level defined by the state. Socioeconomic,
access, quality of care, race and ethnicity, language, and disability are well documented
population characteristics for those at risk of health disparity, which places much of the
Medicaid population and this study group into an at-risk category. Health disparities are defined
legally in the Minority Health and Health Disparities Research and Education Act of 2000 (42
U.S.C. §202 et. seq.) this way:
A population is a health disparity population if there is a significant disparity in the
overall rate of disease incidence, prevalence, morbidity, mortality or survival rates in the
population as compared to the health status of the general population, and burden of
diseases and other adverse health conditions that exist among specific population groups
in the United States. (p.2498)
This study investigates the enabling resources used by the studied population, including
the formal services and the informal supports utilized. I also investigated the living arrangements
of each study participant, independently and combined with caregiver type.
Page 44
43
Need
No one seeks health care services unless they believe it is necessary. How these needs are
perceived by the individual may vary—they may be actively sick, they may have worrisome
symptoms, or they may simply have concerns about their health. This is the need component of
Andersen’s model within population characteristics. In his initial model, Andersen described
need as the “most immediate cause of health service use” (Andersen, 1968, p.17). In later
renditions of the model, he stated that need is determined by perceived need and evaluated need.
To determine need for purposes of this study, I used the evaluated level of function identified as
activities of daily living ability, including the ability to move in bed, transfer, ambulate in the
home, dress, eat, use the toilet, and perform hygiene and bathing activities. I considered
continence, as well as several neurological considerations including cognition, comprehension,
expression, and memory.
Health behavior
Andersen (1995) uses the term health behaviors to describe personal health practices and
use of health services. To determine health behavior trends within the study population, I
examined the services used by these individuals—older adults who reside in the community—
including informal care supports and formal services. This study investigated the differences
between these populations and their use of health services.
Outcomes
Andersen (1995) describes outcomes as perceived and evaluated health status and
consumer satisfaction after health care service delivery. Outcomes data was not looked at in this
Page 45
44
study, Outcomes data would evaluate satisfaction and whether needs are being met. A time
study would allow for this to be evaluated.
Environmental Variables
The environmental variables in Andersen’s model include the current health care system
and the external environment. This study population includes Massachusetts Medicaid recipients
over 65 years of age, and the healthcare system includes Medicaid. The external environment
includes the economy, the current state and federal policies that effect the study population, and
the availability of services and programs. New health care policies often arise from efforts by
lawmakers to meet a stated national goal, such as lowered health care costs or improved services.
Lawmakers also propose health care policies in response to perceived needs of their constituents.
In either case, policies are usually developed without associated resources and funds must be
found and allocated to accomplish the goals of the policy. Policy with resource allocation and
appropriation remain unenforceable. Healthcare policies affect all functions of health care
provision- from practice and care priorities, and typically involve several bureaucratic levels,
including State and federal lawmakers. In theory, health policy should provide the most efficient
and best possible health care provision.
Conclusion
Andersen’s model of health service utilization provides a theoretical structure for this
research study. It assists in identifying variables that delineate differences in characteristics
among the care support groups being analyzed. These include those with informal care supports,
no care supports, formal services only, and those with mixed services and support. As described
Page 46
45
in this chapter, Andersen’s model allows for description of variables for predisposing, enabling,
and need characteristics among people accessing health care services. The goal of this research is
to demonstrate variation within the care support groups in those characteristics that Andersen
describes, specifically among community-dwelling Medicaid-eligible older adults receiving care
and services.
Page 47
46
Chapter 4: Methods
As illustrated in my review of existing literature, there are significant changes occurring
in both the long-term health care needs of our country’s aging population and in our expectations
of the U.S. health care delivery system. As the population ages, increasing numbers of
Americans will require long-term care. Traditionally, such care was provided in institutional
settings, such as nursing or rehabilitation facilities. Institutional care is expensive and because
the bulk of it is billed to Medicare and Medicaid, these costs are borne by U.S. taxpayers. In the
current economic atmosphere, policy makers are looking for ways to pare down health costs
without negatively affecting the quality of services delivered to those who need them. At the
same time, significant numbers of older Americans are attempting to remain in the community as
they age. They do not want to spend their waning years in institutions, cut off from everything
that gives their lives meaning. Either on their own, or with formal or informal assistance, they
are struggling to maintain their freedom and independence, even as they require increasing levels
of care.
The purpose of this study is to provide information that policy makers can use to predict
how many older adults currently dwelling in the community will require long-term care, so they
can develop improved systems to enable these individuals to remain in the community as their
need for medical care increases. Any effort to keep people in the community must consider the
burdens placed on volunteer caregivers, such as family, friends, and neighbors. This study also
provides information about the levels of care that are currently be provided, both in institutional
settings and at home, to help policy makers devise ways to provide support to these volunteer
caretakers.
Page 48
47
I identified four study aims, to examine which characteristics of Medicaid-eligible
community-dwelling older adults are most associated with their eventual need for formal long-
term care. The setting for this research is the Commonwealth of Massachusetts. Massachusetts
has maintained an extensive archive of data about Medicaid-eligible long-term care applicants
since 1999. This database, known as the Minimum Data Set–Home Care, or MDS–HC, made it
possible to obtain the analyses necessary to develop program and policy recommendations based
on the population studied. This resource provided me with complete profiles of more than 3,600
individuals in my target population who applied for Medicaid in Massachusetts between 1999
and 2004.
Aim A: Identify the characteristics of Medicaid-eligible, community-dwelling, older adults in
Massachusetts during the period studied.
This population is associated with higher use of formal long-term care services than non-
Medicaid recipients, due to their financial situation and functional characteristics (Liao &
Chelmow, 2007).
Aim B: Determine variables predictive of membership in various care support groups.
I first divided the entire sample from MDS–HC into two Care Group Categories, for
further examination. Care Group Category A included individuals who met these criteria: (1) no
caregiver support or services (NoSS); (2) informal caregiver support only (InfCS); (3) a mixture
of informal support and formal services (MixS); and (4) formal services only (ForS).
To focus on those receiving at least some support, I established Care Group Category B,
which includes individuals who met one of these criteria: (1) informal caregiver support only
Page 49
48
(InfCS) or (2) at least some formal services—a combination of the MixS and ForS categories
above (SForS). The following hypotheses were specific to Aim B:
Hypothesis 1A (H1A): Hispanics and African–Americans would be more likely than
non-Hispanic whites to be in the InfCS group (and thus cared for by informal care givers)
than in the other care support groups.
Hypothesis 1B (H1B): People with lower levels of education would be more likely than
their more educated counterparts to be in the InfCS group (and thus cared for by informal
care givers) than in the other care support groups.
Hypothesis 1C (H1C): People with a primary language other than English would be more
likely than those whose primary language is English to be in the InfCS group (and thus
cared for by informal care givers) than in the other care support groups.
Hypothesis 1D (H1D): People who are married would be more likely than those who are
unmarried to be in the InfCS group (and thus cared for by informal caregivers) than in the
other care support groups.
Hypothesis 1E (H1E): People who live with others would be more likely than those who
live alone to be in the InfCS group (and thus cared for by informal caregivers) than those
other care support groups.
Hypothesis 1F (H1F): People with more deficits in activities of daily living (ADLs)
would be more likely than those with less deficits to be in the InfCS group (and thus
cared for by informal caregivers) than those in the other care support groups.
Page 50
49
Hypothesis 1G (H1G): People who meet nursing facility eligibility criteria would be
more likely than those who do not meet eligibility criteria to be in the InfCS group (and
thus cared for by informal caregivers) than those in the other care support groups.
Aim C: Examine the relationship between Care Support Group Category A (NoSS, InfCS, MixS,
and ForS), Care Support Group Category B (InfCS and SForS), and nursing facility eligibility in
Massachusetts, while controlling for all other independent variables in the data set.
The following hypothesis was specific to Aim C:
Hypothesis 2A (H2A): Those in the InfCS group would be more likely to meet eligibility
criteria for Massachusetts nursing facility admission as those in other care support
groups.
Aim D: Describe the distribution of Medicaid eligibility across Care Group Category A and
Care Group Category B using different eligibility criteria from states neighboring
Massachusetts.
Data Sources
For more than 20 years, those concerned with delivering care to aging and disabled
individuals have been assisted by a suite of data collection tools. One such set of tools was
developed by an international network of researchers calling itself “InterRAI”
(www.interrai.org). InterRAI is a non-profit group of researchers from over 30 countries with a
common goal, to ensure that clinical practice is evidence-based and policy development is
backed by solid research. The Resident Assessment Instrument (RAI) was created as a
Page 51
50
standardized assessment tool, in hopes of increasing quality of care. The Omnibus Reconciliation
Act of 1987 (OBRA 1987) (42 U.S.C. §139 br, 42 U.S.C. §1395i-3, 42 CFR 483) required
nursing facilities to complete a standardized tool, the RAI, which was incorporated in 1990 as
the Minimum Data Set–Resident Assessment Instrument (MDS–RAI) 2.0. This tool was created
by InterRAI. Subsequently, the Minimum Data Set, version 3.0 (MDS 3.0), was implemented in
2010 as an update to the former MDS–RAI 2.0.
The MDS-RAI serves multiple purposes. It facilitates the management of care in nursing
homes. It also provides a way for oversight bodies to assess the capabilities of long-term care
facilities while helping nursing home staff to identify health problems among residents. In
nursing facilities, the MDS–RAI is completed on admission, by day 14, and quarterly thereafter,
including annually and with significant change in health status, such as hospitalization (found at:
http://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-
Instruments/NursingHomeQualityInits/downloads/MDS20rai1202ch5.pdf) Additionally, many
states have created payment systems based on this tool. The Center for Medicare and Medicaid
Services (CMS), a division of the U.S. Department of Health and Human Services, requires
nursing facilities that accept Medicare and Medicaid payment to use MDS 3.0 to demonstrate the
need for services. Because all states must use it, MDS 3.0 is especially useful for conducting
comparative analyses across states.
The MDS 3.0 only targets institutionalized individuals. However, InterRAI developed a
companion tool, called the Minimum Data Set–Home Care (MDS–HC), to help manage and
assess the delivery of community services. The MDS–HC uses the same definitions and language
as the MDS 3.0, but is adapted for services delivered within the home and community (Morris et
Page 52
51
al., 1997). MDS–HC includes approximately 185 queries, which enables users to gather
comprehensive information about cognition, communication skills, mood and behavior, self-
care, medical care, physical help, and use of services. It is completed at standard assessment
intervals and with change in condition or service allocation.
In short, the MDS–HC provides a standardized way to compare services and usages
among states and between community-dwelling and nursing facility residents. It has been shown
to have 0.92 to 0.96 reliability (Morris et al., 1997). Because of this high reliability, the tool has
been adopted by many countries, including New Zealand, Australia, England, Canada, Hong
Kong, Japan, Italy, France, and Germany. It is also used by sixteen U.S. states, including
Massachusetts.
Massachusetts designated MDS–HC as the standard assessment tool for Medicaid clients
being evaluated for long-term care. These include adult day health, adult family care, group adult
foster care, assisted living, Program for All-Inclusive Care of the Elderly (PACE), and nursing
facility services, regardless of living arrangement. In the past several years, Massachusetts has
extended the use of the MDS–HC to assess state-funded services and programs within the
Executive Office of Elder Affairs, including home-based care, personal care workers, and
respite.
I chose to use the MDS–HC because it provides a wealth of information on the
community-dwelling Medicaid population. This includes the Medicaid elder population in need
of long-term care services; the population I focused on in this study. The data in the MDS–HC
came in an easy-to-use format, which allowed for relatively simple extraction of the sample’s
background, socio-demographic, clinical and functional need characteristics that relate to living
Page 53
52
arrangement, informal care use, and formal service utilization. For my study, I used the
following data variables from the MDS–HC, as shown in Table 1, which is found in Appendix 2.
Table 1. MDS–HC Variables used in this Study.
Of more than 185 data points available from the MDS–HC, I chose specific variables
based on clinical practice, policy direction, and items related to eligibility guidelines. I also
chose certain items to complement the extensive literature and research findings that exist on the
long-term care population.
Physical and cognitive deficits can generate the need for long-term care. MDS–HC
measures these deficits using indices for activities of daily living (ADL); specifically a modified
Katz Index of activities of daily living (ADLs) and the Lawton and Brody Instrumental Activities
of Daily Living (IADL) scale. Developed in 1963, the Katz Index of ADLs has gained
acceptence as an accurate measure of physical functioning, including eating, bathing, dressing,
toileting, transfering, and continence. Lawton and Brody developed the IADL scale in 1969,
which provides a more comprehensive and integrative evaluative analysis of daily activities,
including housework, laundry, shopping, medication, money management, and transportation.
The MDS uses the characteristics of both indices to measure ADL and IADL performance (Katz,
1963; Lawton & Brody, 1969; Morris et al., 1997).
The MDS–HC identifies the use of formal paid care services and informal care supports,
such as volunteers, family, and friends. An algorithm was created that combined living
arrangement, care provided, and care provider to tease out those who receive assistance from an
informal caregiver in performing ADLs. Additionally, this algorithm helped to determine those
Page 54
53
with formal caregiver support, allowing for the determination of each caregiver group. Appendix
3 provides a copy of the MDS–HC tool (Morris et al., 2001).
Because of the subjectivity inherent in the assessment and the desire to enhance
consistency, Massachusetts requires a registered nurse to administer and authenticate the
accuracy of the MDS–HC; however, a licensed social worker may complete subsections of the
assessment related to social setting and living arrangement. Assessors complete the assessment
process through a combination of direct interview and secondary source review, such as
caregiver response and medical record review, to corroborate and verify information. Extensive
standardized training is required for providers and those that administer the assessment tool, to
enhance consistency (per interview with Clinical Manager, MassHealth Office of Long-Term
Care).
Initial trainings were conducted by a registered nurse who participated in the
development of the MDS–HC. This individual then trained two team members, who conducted
additional trainings. Initial trainings were completed in 1998, by an InterRAI team member; in
2000, the registered nurse staff within the Office of Long-Term Care then continued the trainings
for all of the providers within the Commonwealth, insuring consistency in content delivery and
expectations of the data collection instrument.
The Commonwealth of Massachusetts has collected and electronically stored MDS–HC
data since 1999. For the present study, I used MDS–HC data from the years 1999–2004. During
that five-year period, 4,260 Medicaid recipients requested long-term care services. I obtained
permission to use the data from the Assistant Commissioner of Long-Term Care within the
Division of Medical Assistance (currently MassHealth). I also obtained permission from the
Page 55
54
Privacy Office and followed all procedures for data release. The document-granting permission
can be found in Appendix 4. In keeping with Personal Health Information requirements, the
MassHealth Privacy Office removed all personal identifying information, such as names, social
security numbers, Medicare and Medicaid numbers, provider names, and zip codes, before
releasing the information to me.
Study Protocol and Sample Description
This study focused on community-based Medicaid recipients aged 65 and above who
were screened for services during the first five years that the MDS–HC was in use as a
standardized assessment tool. Of the 4,260 Medicaid recipients screened between 1999 and
2004, the following cases were excluded from the analysis: individuals under the age of 65
(n = 337), those residing in a nursing home (n = 100), cases with incomplete assessments
(n = 82), duplicate cases (n = 22), and those with data entry errors (n = 97). The final total
sample was 3,622, 85% of the original sample.
Dependent Variables and Analytic Methods
In this study, I examined two dependent variables. The first dependent variable I looked
at was the type of care each participant received. I wanted to determine differences in
characteristics across care group populations, including (a) no care support versus informal care
only versus formal care only versus mixed formal/informal care; and (b) informal care only
versus at least some formal care. To determine whether type of care support was associated with
nursing home eligibility, I also examined whether each participant met the criteria necessary to
be deemed eligible for nursing home care. Nursing facility eligibility is the basic threshold for
Page 56
55
many clinical programs within Medicaid and other federally-funded programs. In Massachusetts,
this includes the state’s Nursing Facility (NF) Care, Program for All-Inclusive Care of the
Elderly (PACE), and Senior Care Options (SCO), as well as higher levels of reimbursement in
Adult Foster Care (AFC), Group Adult Foster Care (GAFC), and Adult Day Health (ADH).
Dependent Variable 1: Care Group Categorizations
I categorized each participant based on care supports, including informal caregiver
support (InfCS); formal care services (ForS), Mixed Services and Support (MixS), or no care
support or services (NoSS). I defined informal caregivers as unpaid family, friends, or volunteers
who assist participants with activities of daily living (ADLs), including bathing, grooming,
dressing, toileting, and eating. Formal care services are fee-for-service supports, such as home
health aides, visiting nurses, and adult day health, whether funded independently or through
Medicaid. MixS characterizes clients who receive both informal support and formal services.
NoSS applies when the client receives no support services of any kind.
The MDS–HC codes whether a person has informal care support services, as well as the
kind of assistance provided, including IADL and ADL care, advice, and emotional support. The
MDS–HC also codes formal service utilization.
Within the data set itself, I recoded each variable to identify care group, as a first step in
determining whether the participant had an informal caregiver who assisted with ADLs and
whether there was formal care support in place. The care groups were then separated into no
support, informal support, formal support, or mixed formal and informal support. Separate
indicator variables were created for each step, and then combined into an overall care support
Page 57
56
variable for purposes of analysis. This variable was coded as 1 to represent those receiving no
services from an informal or formal caregiver, 2 to represent those receiving informal care
supports only assisting with ADLs, 3 to represent mixed informal support and formal services,
and 4 to represent formal services.
To complete the analysis of care groups, I created two overarching categories: Care
Group Category A and Care Group Category B. Care Group Category A included NoSS, InfCS,
MixS, and ForS. Total sample for Care Group Category A is 3,622: NoSS (n = 269), InfCS
(n = 2,072), MixS (n = 1,127), and ForS (n = 154). Care Group Category B included the InfCS
group and a combined care group (SForS) consisting of MixS and ForS, which captures all
people with some degree of formal services. The total sample for Care Group Category B is
3,353: InfCS (n = 2,072), and SForS (n = 1,281). Those with no care support were excluded from
the analysis of Care Group Category B.
Dependent Variable 2: Nursing Facility Eligibility
I evaluated nursing home eligibility as an independent variable in Aim B and a dependent
variable in Aim C. The analyses for Aims A, B, and C were based on Massachusetts nursing
facility clinical eligibility regulations. Aim B within hypothesis 1G used nursing facility
eligibility as an independent variable, and Aim C used nursing facility eligibility as a dependent
variable. Each state establishes its own eligibility criteria for nursing facility services. In Aim D,
I compared the nursing facility criteria in neighboring states to those used in Massachusetts to
assess the policy implications of applying different eligibility criteria to the population studied.
Algorithms are reported in Table 2: they reference the regulations behind each nursing facility
eligibility criterion and the specific items in the MDS–HC used to determine eligibility. I
Page 58
57
excluded New York because that state’s criteria could not be applied to the MDS–HC data
elements for analysis. The percentages of people in each Care Support Group who met the
nursing facility eligibility criteria in each state are reported in the results. I defined the
categorical variable “nursing facility eligibility” as meets or does not meet (code 0 = not eligible,
1 = eligible) and determined eligibility in each case by referencing Massachusetts and the
regulatory guidelines for each state. Individuals who are eligible for nursing facility care are
more likely to need some sort of long-term assistance eventually, either in an institutional setting
or in the community, with formal or informal support (Luppa, et al., 2009).
Page 59
58
Table 2. Nursing Facility Eligibility by State.
State Reference Citation Eligibility Criteria
Massachusetts 130 CMR 456.409
Commonwealth of Massachusetts
Division of Medical Assistance,
Provider Manual Series
Part of MGL Ch. 118
Skilled service daily performed by a nurse
(examples include intravenous injections,
catheters, dressings) or need for assistance
with two ADLs and a nursing service at least
three times per week. ADLs include: bathing,
dressing, toileting, eating, mobility, and
transfers. Nursing service is a skilled service
or skill such as medication administration.
Rhode Island 0378.10
Rhode Island
Department of Human Services
Policy Manual
Part of RIGL 42-35
Nursing Facility level of care requires the
services of a nurse or rehabilitation
professional or assistance with activities of
daily living.
Connecticut 17b. -262-704
Ch. 7
Continuous skilled nursing services and need
for substantial assistance with hands-on care,
determined by the need for assistance with 3
or more of these 7 critical needs: bathing,
dressing, transferring, toileting,
eating/feeding, meal preparation, and
medication management.
Vermont Division of Disability and Aging
Services
A combination of the following: limited or
extensive assistance with ADLs, severe—
moderate cognitive impairment, daily skilled
nursing need, unstable medical need.
Measurement of Independent/Control Variables
I used relevant literature and practice to determine those clinical characteristics in the
MDS–HC that were most relevant in determining care support, service utilization, and nursing
home eligibility in the population studied. This study only captures a subset of the available data
from the MDS–HC.
Independent Variables
Page 60
59
Ethnicity
There were six categories for race and ethnicity in the MDS–HC: American Indian/
Alaskan Native (code = 1), Asian (code = 2), White or Caucasian (code = 5), Black or African–
American (code = 3), Hispanic or Latino (code = 5), and Native Hawaiian or other Pacific
Islander (code = 4). People may have self-identified in more than one category, which is
reported within the findings. Data were reported using four categories for Aim A: White or
Caucasian (non-Hispanic), Hispanic, African–American, and other, to match the hypothesis and
relevent literature. Hispanic refers to anyone who self-identified as Hispanic, whether they also
identified as White/Caucasian or African–American; White or Caucasian (non-Hispanic) is
anyone who self-identifies as White or Caucasian, but not Hispanic; African–Amercan is anyone
who self-identifies as African–American but not Hispanic. Because so few individuals in the
study group self-identified as Pacific Islander, Asian, or Native American, for the analysis of
Aim B, I categorized all of these as “other”. For Aim B, Hypothesis 1A, I looked to identify
differences in the larger ethnic categories of Caucasian, African–American, and Hispanic. The
reference category was Caucasian. For Aim C, I used this definition for a control variable and
completed the regression analysis. I included race because it has been strongly associated with
care giving differences. For example, several sources report that minority populations are more
likely to care for elders at home for longer periods of time (Friedman, et al., 2005; Akamigbo &
Wolinsky, 2006; Konetzka & Werner, 2009; Scharlach, Giunta, Chow & Lehning, 2008; Casado,
Vulpen & Davis, 2011).
Education
Page 61
60
The level of education of the individual or the caregiver is linked to decisions regarding
use of formal services (Williams & Dilworth–Anderson, 2002). Level of education is captured
within the MDS–HC as the highest level of education completed; the designation “N/A” means
the level of education was unknown or indeterminate.The MDS–HC reports education by the
highest level of schooling completed. The categories are as follows: no school, 8th
grade or less,
grades 9–11, high school, technical or trade school, some college, bachelor’s degree, or
graduate degree. For this study and the descriptive analysis (Aim A) of the sample, I combined
the education categories of no schooling, 8th
grade or less, and completion of grades 9–11 into a
single category I called less than high school graduate (code = 1). Code 2 included those who
were high school graduates, had attended technical/trade school, or who had some college. I
grouped those who completed a bachelor’s degree or higher into a single category, college
degree or higher (code = 3). For the analyses related to Aim B, I dichotomized education as high
school graduate or less (code = 0) and some college or college graduate (code = 1). The reference
category for education was high school graduate or less. I used this definition in Aim C for a
control variable in the regression analysis.
Language
A person’s inability to speak English in a predominantly English speaking long-term care
system is often linked to increased use of informal versus formal services (Kendig, et al., 2010;
Miller, 2010; Liand, et al., 2005). The MDS–HC evaluates a person’s primary language as
English, French, and Other. The tool does not require specification of other languages. For
purposes of this study, primary language was dichotomized into two categories: English (code =
Page 62
61
0) or other (code = 1). I used this definition in Aim C for a control variable in the regression
analysis.
Marital Status
Being married with the continuous support of a spouse, is linked to remaining in the
community and often predict less use of formal care supports. The MDS–HC allowed individuals
to self-identify their status, regardless of the laws in the individual state. The choices given were
never married, married, widowed, separated, divorced, and other. For this study, I recoded
marital status as a dichotomized variable: married (code = 1) or not married (code = 2). Marital
status not married included those never married, widowed, divorced, and separated. I used this
definition in Aim C for a control variable in the regression analysis.
Living Arrangement
Living arrangement encompasses location of residence and whether the participant lived
alone or with others. Living arrangment and with whom people live can affect how they use
formal long-term care supports (Kasper, Pezzin & Rice, 2010; Ritchie, Roth & Allman, 2011).
The MDS–HC collects information about where a person lived at the time of referral, including
private home with or without home care services, board and care, assisted living, and nursing
facility. The MDS–HC defines a person’s living arrangements as living alone, spouse only,
spouse and others, child (not spouse), others (not spouse or child), or group setting with non-
relative. Living arrangement was dichotomized as living alone (code = 0) or living with others
(code = 1). I used this definition in Aim C for a control variable in the regression analysis.
Functional Status
Page 63
62
It has been demonstrated that ability to complete self-care activities is an important
determinant of quality of life (Morris, et al., 2001, p. 130). Additionally, functional status, and
inability to complete self-care activities is a primary determinant of the need for long-term care
supports and services, either within an institutional setting or in a community-based formal
setting. I measured functional status as a continuous variable, representing a score on scales
measuring the client’s ability to complete Instrumental Activities of Daily Living (IADL) and
Activities of Daily Living (ADL) tasks. As mentioned earlier, the MDS–HC utilizes two well-
known and validated assessment tools with excellent psychometrics, known as the Katz ADL
scale (1963) and the Lawton and Brody IADL scale (1969).
Participants were rated on their ability to complete seven ADLs, as defined below:
mobility in bed (moving to and from lying position and positioning body in bed),
transfer (moving between surfaces such as from bed to chair or to standing),
locomotion in home (movement in home),
dressing the upper and lower body (dressing and undressing),
eating (taking in food by any method), and
toileting (using toilet and cleaning self).
The ability of the client to complete each ADL was rated on a scale of 0 to 6, defined in
this way:
0 = independent
Page 64
63
1 = set-up help only
2 = supervision
3 = limited assistance (client highly involved in activity)
4 = extensive assistance (client performed part of activity on own—50% or more of
subtasks— but needed weight-bearing or full performance by another for part of the
time)
5 = maximal assistance (client involved in activity but completed less than 50% of
subtasks)
6 = total dependence (full performance by another)
The maximum score was 42.
The MDS–HC also reports bathing and locomotion outside of the home. I did not include
these elements in this aspect of the analysis because they are not used to determine eligibility
within Massachusetts Medicaid. Older individuals do not typically bathe daily and for eligibility
for Massachusetts Medicaid programs, the task must occur daily. Locomotion outside of the
house is not utilized as eligibility and not captured, because individuals often may be limited in
ability outside of their homes but able to maneuver in their homes without difficulty.
The MDS–HC also used IADL as a continuous measure of functional status, as measured
by client performance of the following seven activities:
meal preparation (planning, cooking, assembling meals),
Page 65
64
ordinary housework (dishes, dusting, tidying),
managing finances (paying bills and balancing expenses),
managing medications (remembering to take medications and taking correct
medications),
phone use (making calls even with devices),
shopping (for food and household items, both selection and management), and
transportation (getting places).
Raters evaluated the extent to which clients could perform these IADLs according to
these definitions: 0 = independent, 1 = some help (some of the time), 2 = full help (help all of the
time), and 3 = by others. The maximum score was 21. I used this definition in Aim C for a
control variable in the regression analysis.
Control Variables
Clinical Need
Clinical needs are services that are normally provided by skilled nursing staff or
unlicensed, trained para-professionals under the direction of a nurse. Clinical needs are often
those services related to eligibility for medical programs, both state and federally funded. The
variables I chose to include in this study are directly related to services seen in long-term care
eligiblity guidelines in either institutional or community-based programs.
Page 66
65
I chose the clinical variable of bladder continence because this is often a care need that
causes a person to switch from informal care support to formal services. When a care need
becomes highly personal, informal caregivers are often reluctant to continue providing services
(Green & Ondrich, 1990; Baker & Bice, 1995; Fultz, et al., 2003). The MDS–HC codes bladder
continence as continent, continent with catheter, usually continent, occasionally incontinent,
frequently incontinent, and incontinent. I recoded the variable of bladder continence into three
categories; continent (combined continent and continent with catheter, code = 1), occasional
incontinence (combined usually continent and occasionally incontinent, code = 2), and
incontinent (combined frequently incontinent and incontinent, code = 3). When examined using a
series of dummy variables for the multivariate analysis in Aim B, I used continent as the
reference category. In Aim C, I used this definition for a control variable in the regression
analysis.
There can be numerous causes for medical non-compliance or non-adherence, including
practical considerations, such as an inability to acquire or pay for medications. However, this
situation can also signal changes or declines in medical, neurological, or psychological status,
such as memory deficits or psychological disease. (Gatti, et al., 2009; Elliott, 2009; Vik,
Maxwell & Hogan, 2004). Medication adherence was coded within the MDS–HC as always
compliant (code = 1), compliant 80% of the time (code = 2), compliant less than 80% of the time
(code = 3), and no medications prescribed (code = 4). When examined using a series of dummy
variables in the multivariate analyses in Aims B and C, I used medication adherence as the
reference category.
Page 67
66
Number of medications was identified as a continuous variable, between zero and nine. A
high number of medications being administered has been found to be associated with increased
risk of utilization of long-term care services and nursing facility placement (Luppa, et al., 2010;
Steiman & Hanlon, 2010). In Aim C, I used this definition for a control variable for the
regression analysis.
Cognitive and Psychological Need
I used five variables to describe a client’s cognitive and psychological need. The first and
second variables related to a client’s short-term and procedural memory. MDS–HC codes
memory as memory okay or memory problem. These items are based on direct interview, with
corroboration using secondary sources. Short-term memory looks at subject’s ability to recall
after five minutes, and procedural memory describes a subject’s ability to perform all or almost
all steps in a multi-task sequence without cues. Both short-term and procedural memory concerns
were coded as memory problem or memory okay (code 0 = memory ok, 1 = problem). (Burdick,
et al., 2005; Kendig, et al., 2010; Luppa, et al., 2010). In Aim C, I used this definition for a
control variable for the regression analysis.
The MDS–HC measures cognitive ability as the ability to make decisions about
organizing the day. Decision making and cognitive ability are often linked to dementia and
Alzheimer’s disease; impairment of these functions can increase the likelihood that an individual
will require long-term care services and institutionalization (Burdick, et al., 2005; Kendig, et al.,
2010; Luppa, et al., 2010). This was identified as no issues or independent (code = 0), modified
independence (code = 1), minimally impaired (code = 2), moderately impaired (code = 3), or
severely impaired (code = 4). For the descriptive analysis in Aim A, I used the coding on the
Page 68
67
MDS–HC, identified here with codes 0 through 4. For the data analysis in Aim B, I re-coded
them into dichotomized variables as follows: independent and modified independence with
making decisions (code = 0), and minimal, moderate, and severe impairment with decision
making (code = 1). In Aim C, I used this definition for a control variable in the regression
analysis.
Using the MDS–HC definitions and coding, I defined comprehension in terms of how the
person understood verbal information and categorized levels of comprehension as understands,
usually understands, often understands, and sometimes or rarely understands (coded as 0, 1, 2,
3, and 4 respectively). I defined expression, using the MDS-HC definitions and coding, as the
ability of the person to express information and content. The categories are understood, usually
understood, often understood, and sometimes or rarely understood (code 0, 1, 2, 3, and 4
respectively). In Aim C, I used this definition for a control variable in the regression analysis.
Neurological diagnoses are often linked to nursing facility care and also to clinical care
needs. Alzheimer’s disease and dementia are identified as present, present and monitored treated
by a home care professional, or not applicable within the MDS–HC. I manipulated the coding to
include code 0 = not applicable and code 1 = disease present (combining present and treated). In
Aim C, I used this definition for a control variable in the regression analysis. The MDS–HC
defines Alzheimer’s disease and dementia regardless of level of need or severity of illness. Other
neurological conditions are also captured in the MDS–HC, including head trauma, multiple
sclerosis, and Parkinsonism. Because there were so few respondents in these categories, I did not
analyze these neurological conditions in this study.
Page 69
68
Because depression is quite common among older individuals and is often linked to
formal care and increased risk for institutionalization, I included depression risk as an
independent variable in this study (Burdick, et al., 2005; Gruber–Baldinin, et al., 2005). The
MDS–HC combines seven symptoms that could indicate depression: feeling of sadness,
persistent anger with self or others, expression of unrealistic fears, repetitive health complaints,
repetitive anxious complaints or concerns, sad, pained or worried facial expressions, and
recurrent crying or tearfulness. Each symptom was coded as 0 = that the symptom was not
exhibited; 1 = the symptom was exhibited on one or two of the past three days or 2 = the
symptom was exhibited on each of the past three days. Of a possible combined score of 14
(seven symptoms with a maximum score of two per symptom), depression risk is identified by
the MDS–HC as any person with a score of over two. In Aim C, I used this definition for a
control variable for the regression analysis. This assessment and score allocation have shown
good interpretation for risk (Morris et al. 2001). I used the definitions and coding of the MDS–
HC, and identified depression risks as a continuous variable, with possible calculated values
from 0 to 2.
Behavior problems, whether associated or not to a psychological diagnosis or disease,
increase use of long-term care services and potential nursing home placement (Morris, et al.,
2007; Diwan, 1999; Casado, Vulpen & Davis, 2011). Behavioral issues are often linked to
institutionalization and nursing facility placement, because it is increasingly difficult for the
person to be cared for in the community. The MDS–HC identifies a behavior score measuring
the presence of five behaviors: wandering, verbally abusive, physically abusive, socially
inappropriate or disruptive behavior, and resistance to care. Behavior symptoms are coded as
0 = the behavior did not occur, 1 = the behavior occurred but was easily altered by the care
Page 70
69
provider and 2 = the behavior occurred and was not easily altered over the previous three days.
Behavior problems were reported as a continuous variable, with a possible calculated score of
0–10 (five behavior symptoms with a maximum score of two per symptom). The MDS–HC
identifies behavioral issues with a score of two or higher. In Aim C, I used this definition for a
control variable for the regression analysis.
Perception of Health Status
There are two variables pertaining to perception of health status. The first variable
concerned whether clients believed they were capable of increased independence (code 0 = false,
1 = true). The second variable related to the participants’ perceptions about the state of their
health. Specifically, participants were asked if they believed they were in poor health (code 0 =
false, 1 = true). In Aim C, I used this definition for a control variable in the regression analysis.
These two factors are important because they can predict the extent to which individuals will
participate in their care and how determined they may be to improve their well-being (Chen &
Thompson, 2010; Akamigbo & Wolinsky, 2006; Luppa, et al., 2010). These variables have been
shown to be important predictors for nursing facility eligibility.
Another control variable I evaluated for Aim C is trade-offs. Trade-offs are defined as a
decision not to purchase certain important goods or services because of limited funds. The
particular trade-offs cataloged in the MDS–HC were prescribed medications, sufficient or
necessary home heat, food, or home care, etc. The MDS–HC does not identify the specific trade-
off: it simply notes that when asked about trade-offs during the assessment, the participant
acknowledged a decision not to purchase a necessary item in exchange for another that is needed.
Page 71
70
I used the MDS–HC coding for this variable, which simply noted if the individual had made any
trade-offs: (code 0 = no, 1 = yes).
Data Analysis
Aim A
Identify the characteristics of community-dwelling, Medicaid-eligible, older adults in
Massachusetts who apply for long-term care services.
I performed a descriptive analysis that included basic demographics and clinical
characteristics, to give a picture of the population studied. This first step was an opportunity to
evaluate the sample as a whole and across both care groups. As part of this analysis, I used mean
and standard deviation to evaluate continuous variables, and studied categorical variables using
frequency and percentages. The total sample size for the analysis of Aim A was 3,622.
In this paper, I present continuous and categorical variables in tabular form. I provide
means and standard deviations for continuous variables, and frequencies and percentages for
categorical variables across each of the groups and categories. I also include a description of the
total sample. Descriptive statistics are provided on the following variables:
demographics—gender, ethnicity and race, level of education, language, marital status,
and living arrangement;
clinical and functional needs—bladder continence, medication adherence, number of
medications taken, activities of daily living (ADL) status, and instrumental activities of
daily living (IADL) status;
Page 72
71
nursing facility eligibility;
cognitive and psychiatric characteristics—short-term and procedural memory, cognitive
decision making, comprehension, expression, diagnosis of Alzheimer’s disease,
depression risk, and behavioral problems; and
perception of health status—belief of capability of increased independence, and belief of
poor health,
I present descriptive data for the total population, as well as for each of the Care Group
Categories: Category A (NoSS, InfCS, MixS, and ForS) and Category B (InfCS and SForS).
Aim B
Determine the differences in the characteristics of care group categories across selected
independent variables.
In Aim B, I examined whether there were significant differences in how the independent
and control variables were distributed across care groups, as described in hypotheses 1A–G.
Each of these hypotheses was chosen because the literature supports a relationship between these
independent variables and types of services needed (Friedman, et al., 2005; Akamigbo &
Wolinsky, 2006; Konetzka & Werner, 2009; Scharlach, Giunta, Chow & Lehning, 2008; Casado,
Vulpen & Davis, 2011; Williams & Dilworth–Anderson, 2002; Kendig, et al., 2010; Miller,
2010; Kasper, Pezzin & Rice, 2010; Ritchie, Roth & Allman, 2011; Hong, 2010; Luppa, et al.,
2009). The total sample used for analysis of Aim B was 3,609.
Care Group Category Analysis by Independent Variables:
Page 73
72
Hypotheses 1.A, 1.B, 1.C, 1.D, 1.E, 1.F, 1.G.
In analyzing data for Aim B, I examined differences in the hypothesized variables and the
characteristics of the Care Group Categories, including Care Group Category A (NoSS, InfCS,
MixS, and ForS) and Care Group Category B (InfCS and SForS). I present the results in a table
for the bivariate analysis, displaying the hypothesized variables and Care Group Categories A
and B. I provide separate tables to report the multivariate analyses, examining the relationship
between the hypothesized variables of ethnicity, education, language, marital status, living
arrangement, functional deficit, and nursing facility eligibility, and Care Group Category A and
Care Group Category B, controlling for other characteristics.
Preliminary Analysis
I first analyzed all data for Aim B, to insure that the numbers were accurate and the
sample had not been compromised. For Hypotheses 1A, 1B, 1C, 1D, 1E, and 1G, the
independent variables were categorical in nature and evaluated for normal distribution and
outliers. I assessed mean, expected, and observed frequencies for accuracy. For hypothesis 1F,
the independent variable was continuous in nature, so I analyzed for sample size and distribution,
mean, standard deviation, skewness, and kurtosis. I also completed preliminary analyses for the
multivariate analysis and included an evaluation for parametric assumptions and
multicollinearity, as well as an evaluation of outliers utilizing Mahalanobis distance.
Multicollinearity was analyzed to assess the correlation of the variables and to determine
whether a significant relationship existed among the variables. I also assessed assumptions about
the distribution of scores and relationships between the variables.
Page 74
73
Care Group Category A Bivariate Analysis
I analyzed the data for Care Group Category A (NoSS, InfCS, MixS, and ForS), for Aim
B using bivariate statistics including chi-square and ANOVA, as appropriate. Chi-square is used
to detect differences in means of categorical variables. ANOVA evaluates the differences
between the scores on the independent variables (i.e., their means) across subcategories of the
dependent variables (Pallent, 2007).
I completed bivariate analysis for each hypothesized variable, including ethnicity (H1A),
education (H1B), language (H1C), marital status (H1D), living arrangement (H1E), ADL
functional status (H1F), and nursing facility eligibility (H1G). Using chi-square analysis, I
reported the chi-square test value, degrees of freedom (df), and statistical significance using p
values. In interpreting the chi-square analysis output, I first checked assumptions for frequency
in each cell. I reviewed the significance, or p value, using a value of .05 or smaller to indicate
significance. ANOVA analyses were reported using the F-test statistic and p value. The F ratio is
a tool to determine if the variances in different sample groups are equal. If the F ratio is
significant, the hypotheses are rejected. For bivariate analysis, I ran individual analyses using
each care group. I analyzed each hypothesized relationship using the elements contained in Care
Group Category A: NoSS, InfCS, MixS, and ForS. Below, I provide the coding procedures used
to conduct the tests, the handling of missing data, and the specific tests conducted for each
hypothesis of Aim B. I also reported p values for each of the tests conducted. I have outlined the
details, by hypothesis, below.
Care Group Category B Bivariate Analysis
Page 75
74
I analyzed the data for Care Group Category B: InfCS and SForS, within Aim B using
bivariate statistics including chi-square and t-test. Chi-square was analyzed as outlined above, in
the same fashion as completed for Aim B, Care Group Category A. I used t-test to detect the
differences in the means of the categorical and continuous variables, as seen in Hypothesis 1F
(functional status: ADLs). These analyses enabled me to determine if there were any significant
differences by the hypothesized variable across the care group categories of InfCS and SForS.
For t-test, I have reported the mean, standard deviation, t value, degrees of freedom, and
significance (p value). I have outlined the details, by hypothesis, below.
Aim B: Specific analysis by hypothesis for Care Group Categories A and B.
Hypothesis 1A. Hispanics and African–Americans would be more likely than non-
Hispanic whites to be in the InfCS group (and thus cared for by informal care givers) than
in other care support groups.
I analyzed Hypothesis 1A using chi-square to determine if there were any significant
differences across ethnic categories and care groups. I dichotomized the variable of ethnicity in
this way: non-Hispanic whites (code = 0) and Hispanics and African–Americans (code = 1). I
excluded additional race and ethnic categories because of small sample size.
I found no cases with missing data in the data set. There were 3,622 responses from
clients received and after removing cases that were coded as additional racial and ethnicity types,
3,450 remained. The additional racial and ethnicity types identified included Asian, Pacific
Islander, and American Indian.
Page 76
75
Hypothesis 1B: People with lower levels of education would be more likely than their
more educated counterparts to be in the InfCS group (and thus cared for by informal care
givers) than in the other care support groups.
Using chi-square, I analyzed Hypothesis 1B to determine if there were any significant
differences across levels of education and care groups. The variable of education was
dichotomized: high school or less (code = 0), and some college or more (code = 1). I did not
evaluate specific levels of education. When completing the MDS–HC, if it was not possible to
determine the person’s level of education, the assessor completing the data tool left the item
blank. This occurred in 81 cases. As a result, of 3,622 responses received from clients, I retained
3,541.
Hypothesis 1C: People with a primary language other than English would be more
likely than those whose primary language is English to be in the InfCS group (and thus
cared for by informal care givers) than in the other care support groups.
I analyzed Hypothesis 1C using chi-square, to determine if differences existed across the
different care groups in the number of clients whose primary language is English. The variable of
language was dichotomized: English (code =1), and Other (code = 2).The Other category
included French, Spanish, and Other. The data set related to language was complete, with 3,622
responses received and retained.
Hypothesis 1D: Clients who are married would be more likely than those who are
unmarried to be in the InfCS group (and thus be cared for by informal care givers) than in
theother care support groups.
Page 77
76
Using chi-square, I analyzed Hypothesis 1D to determine if differences existed in the
number of clients who were married across the different care groups. The variable of marital
status was dichotomized: married (code = 1), and Other (code = 2). The Other category included
never married, divorced, separated, and widowed. Married also included those who considered
themselves married, even if only by common law. Among the 3,622 responses received, there
were 13 cases in which the marital status of the clients was not identified, leaving a study sample
of 3,609 for this category.
Hypothesis 1E: People who live with others would be more likely than those who live
alone to be in the InfCS group (and thus cared for by informal care givers) than those in
other care support groups.
I used chi-square to analyze Hypothesis 1E, to determine if differences existed in the
proportion of clients who live with others across the different care groups. Living arrangement
was dichotomized: living alone (code = 1) and living with others (code = 2). From a total of
3,622 responses received, 13 were missing data pertaining to living arrangement, leaving a study
sample of 3,609. The missing data resulted from miscoded entries and data entry error.
Hypothesis 1F: People with more deficits in activities of daily living (ADLs) would be
more likely than those with less deficits to be in the InfCS group (and thus cared for by
informal caregivers) than those in the other care support groups.
Hypothesis 1F was analyzed using ANOVA for Care Group Category A because it
includes four elements and ADL deficits is a continuous variable. I used t-test for Care Group
Category B. This enabled me to determine if there were any significant differences in the degree
Page 78
77
of functional status across the different care groups. Functional status scores were based on the
client’s level of ability to complete ADL tasks. There were no cases found with missing data in
the data set; I retained all 3,622 cases.
Hypothesis 1G: People who meet nursing facility eligibility criteria would be more
likely than those who do not meet eligibility criteria to be in the InfCS group (and thus
cared for by informal caregivers) than those in the other care support groups.
For Hypothesis 1G, I used Massachusetts Medicaid criteria to determine eligibility for
nursing facility services. I used chi-square analysis to determine if there were significant
differences in the number of clients who meet nursing facility eligibility criteria across care
groups. The variable of nursing facility eligibility was coded thus: not eligible (code = 0), and
eligible (code = 1). There were no missing cases in the data set: the data set was complete, with
3,622 responses reviewed and retained.
Care Group Categories A and B Multivariate Analysis
I conducted multivariate analyses to confirm bivariate relationships by controlling for
other factors likely to affect the dependent variables. To further evaluate the differences between
care group categories across the independent variables identified in Hypotheses 1A–G, I
completed a series of logistic regressions. For Care Group Category A, I ran four regressions
with the dependent variables, NoSS and a dummy variable, InfCS and a dummy variable, MixS
and a dummy variable, and ForS and a dummy variables. For Care Group Category B, I ran one
regression using the care groups of InfCS versus SForS, as the dependent variable, thus looking
at the likelihood of using some formal supports versus informal support only. Through such
Page 79
78
logistic regressions, I was able to control for the study’s focal variables (race/ethnicity, level of
education, language, marital status, living arrangement, functional status, and nursing facility
eligibility), as well as other independent variables identified in Table 1 (i.e., the control
variables).
I evaluated the output from the logistic regression to determine significance. Logistic
regression analysis reported Wald values, p values, odds ratios, and Beta values. The Wald test
indicates the importance of each variable in predicting an outcome (Pallent, 2007). Beta values
predict the probability of a variable being in a set category, and its value is positive or negative,
identifying directionality of the factor. The odds ratio shows the odds of a variable falling into a
specific category.
Aim C
Examine the relationship between those in Care Group Category A (NoSS, InfCS, MixS, and
ForS) and Care Group Category B (InfCS and SForS) and nursing facility eligibility in
Massachusetts, while controlling for all other independent variables in the data set.
Nursing facility eligibility is the basis for clinical determination for eligibility and
funding levels for many Medicaid and Massachusetts state-funded programs, including
community-based long-term care services. The goal of this aim was to determine if nursing
facility eligibility was associated with membership in care groups while controlling for other
variables within the study. Additionally, this aim helps determine which factors prove most
influential in determining nursing facility eligibility.
Nursing Facility Eligibility and Care Support Group B Membership (Hypothesis 2A)
Page 80
79
Using logistic regression, I analyzed the data to determine the relationships between care
groups and eligibility for nursing home placement in Massachusetts. I chose logistic regression
as the mechanism for analysis because it is a proven way to predict accurately and explain the
categorical dependent variable and the relative importance of each predictor variable.
Hypothesis 2A: Those in the InfCS group would be more likely to meet eligibility
requirements for Massachusetts nursing facility admission as those in the other care
support groups.
Hypothesis 2A was analyzed using logistic regression. The dependent variable was
nursing facility eligibility, which was coded as 1 for nursing facility eligible or 0 for not eligible.
I ran separate logistic regressions: one with Care Group Category A as an independent variable
and one with Care Group Category B as an independent variable. The independent variables of
Care Group Category A included a series of dummy variables: NoSS and InfCS, MixS and
InfCS, and ForS and InfCS. The independent variables of Care Group Category B included
InfCS (code = 1), and SForS (code = 0). Both regression models included several control
variables.
The categorical control variables included gender, race/ethnicity, level of education,
marital status, primary language, living arrangement, presence of Alzheimer’s disease, presence
of short-term memory issues, presence of procedural memory issues, and trade-offs. The
continuous control variables include number of behavioral problems, functional status with
activities of daily living, instrumental activities of daily living status, cognition/decision making
status, and depression risk.
Page 81
80
I tested the validity of the model in several ways. I ran omnibus tests of model coefficient
and the Hosmer–Lemeshow statistic to determine how well the model performed. To assess the
usefulness of the model, I used both the Cox and Snell R Square and Nagelkerke R Square,
which help to indicate the amount of variation in the dependent variable determined by the
model. To report the logistic regression, I used the Wald test, p value, Beta values, and odds
ratios. The importance of the predictor variable is identified in the Wald, the significance
identified by the p value demonstrates that with a value of less than .05 the associated variable
contributes significantly to the model. Beta values predict the extent to which an independent
variable is related to the dependent variable. The odds ratio shows the odds of falling into a
specific category.
I coded the initial data and identified reference categories as described in variable
specification. The preliminary analysis included an evaluation for parametric assumptions. To
explain unexplained cases, I evaluated outliers using Mahalanobis distance. Normality, linearity,
and homoscedasticity were also assessed, as previously discussed. R-squared values were
evaluated. I retained 3,609 cases for purposes of these analyses.
Aim D
Examine how adjustments in nursing facility eligibility criteria would change the percentage of
persons eligible for nursing facility services in each care support group, by evaluating the
eligibility criteria in neighboring states.
Nursing facility eligibility varies by state, and I chose to evaluate neighboring states to
determine if adjustments in eligibility would substantially change the populations who meet and
Page 82
81
do not meet eligibility. In Massachusetts, certain components of the eligibility requirements have
been revised since the data in this study were collected, but there has been no comprehensive
evaluation of the entire regulation in the last thirteen years. The eligibility criterion has not been
adjusted since 2004, although the focus on rebalancing long-term care has been on moving
people into the community for a number of years. Many other states have adjusted their clinical
eligibility criteria for nursing facility services, because the eligibility criteria are used as a
baseline for community-based waiver programs, thus supporting rebalancing and enhanced
community-based options. Massachusetts Medicaid eligibility has very few criteria on residence
and length of time in state to be eligible for services, whereas in other states people must
demonstrate proof of residence for a length of time (i.e. Rhode Island is six months). This lack of
residence requirement has encouraged people from other states to come to Massachusetts to get
care (Uccello, McCallum & Gallagher, 1996). To address this aim, I examined the nursing
facility eligibility criteria in states that border Massachusetts, including Vermont, New
Hampshire, Rhode Island, and Connecticut. New York requires a physician’s order to determine
nursing home eligibility, which is not a data element available on the MDS–HC. Because of that,
New York’s eligibility criteria are incompatible with this data set, so I did not include New York
in the study.
Hypothesis 3A: Those in the InfCS group would be more likely to meet eligibility for
nursing facility services when applying the eligibility criteria of neighboring states.
Because of the nature of this aim, there was no need to employ any inferential statistical
methods. Instead, I used descriptive statistics, based solely on the clinical eligibility criteria for
nursing facility services, to determine the percentage of those people in Care Group Category A
Page 83
82
and Care Group Category B who meet the nursing home eligibility criteria for each of the states I
studied. I did the same thing for the total sample. I did not evaluate Medicaid eligibility or
residency requirements for services. I arrayed the results in a table that includes the population as
a whole, as well as each Care Group Category (Category A = NoSS, InfCS, MixS, and ForS;
Category B = InfCS and SForS). Because the variables were categorical, for each of these three
groups, I reported sample sizes (n) and the percentage of individuals who were or were not
eligible for nursing home care in each of the states studied.
Conclusion
In this chapter, I provided detailed descriptions of the tools and techniques I used to
conduct this study. I described the research design, including the aims and hypotheses, the
MDS–HC as a tool, the sample, and the data processing and analysis. My discussion of the
MDS–HC was extensive, because it is important to understand both the power and the limits of
this unique database. Because I did not use all of the data available, I provided a detailed
description of my rationale for making the choices I made. I depended upon this data source for
all four aims of the study, including Aim A: identifying the characteristics of Medicaid-eligible,
community-dwelling, older adults in the State of Massachusetts during the period studied.
In many ways, the heart of the study lay in Aim B; identifying and analyzing the key
background and clinical variables of the population I studied—Medicaid-eligible, community-
dwelling older residents of Massachusetts—who may or may not have met the criteria for long-
term institutional care. To get at the most valuable information and provide meaningful analysis,
it was important for me to identify subsets of this large population, so I provided a detailed
explanation of my reasons for every sorting decision I made. The decision to present the analysis
Page 84
83
across care groups—Care Group Category A (NoSS, InfCS, MixS, and ForS) and Care Group
Category B (InfCS and SForS)—laid the groundwork for several valuable insights, but I also
chose to conduct a bivariate analysis, as described above. By doing so, I was able to evaluate the
likelihood that individuals with particular traits would require long-term care in the future. To
confirm the possibility of bivariate relationships, I also conducted multivariate analyses.
By introducing the regression analyses used in Aim C, I was able to investigate the
sometimes complex relationships among all of the variables I followed. This was essential
because the goal is to provide policy makers with powerful predictive data they can use to
enhance delivery of long-term care services to older people without consigning them to
institutional living. Finally, the descriptive analysis I employed in Aim D enabled me to provide
the kind of comparative data that decision makers can use to support the development of
innovative policies on behalf of older Americans and their caretakers.
Page 85
84
Chapter 5: Findings
The immediate purpose of these analyses was to provide data that will help decision
makers to understand the characteristics and needs of the elder population, thus enabling them to
develop or revise policies to enhance support for elders at home and in the community. Because
informal caregivers appear to comprise a significant percentage of the total care provided to
elders and because their contributions are not formally recognized through Medicare or
Medicaid, I focused on efforts to characterize the nature and extent of these contributions. My
ultimate goal is to inform policies that support efforts to rebalance long-term care away from
nursing homes toward home- and community-based service alternatives.
To draw conclusions from the sample population, I used inferential statistics, specifically
the Statistical Package for the Social Sciences (SPSS) version 19.0, to code and tabulate
collected assessments and to provide summarized values where applicable. In addition, I
analyzed several variables and relationships using appropriate frequency statistics, including chi-
square, analysis of variance (ANOVA), independent samples t-tests and logistic regression.
Before analyzing the hypotheses within the study, I ensured that the variables of interest met
appropriate statistical assumptions by screening them and employing appropriate data hygiene
measures. The findings are presented within tables and discussed in each aim presented below as
well.
Aim A
Sample characteristics
Page 86
85
The first aim of the study was to identify characteristics of Medicaid-eligible community-
dwelling older adults in the state of Massachusetts within the total sample. This descriptive
analysis included all of the variables collected for the study, including demographic and clinical
variables, and is presented across all Care Support Groups, including Care Group Category A:—
No Support Services (NoSS), Informal Care Support (InfCS), Mixed Care Support and Services
(MixSS), and Formal Support Services (ForS)—and Care Group Category B—Informal Care
Support (InfCS) and Receipt of Some Formal Services (SForS). This analysis allows for an
overall picture of the population being studied. The data in Tables 3–16 relate to this aim. Basic
demographics are presented in Tables 3 and 4. Clinical background information is presented
across Tables 5–16, which report the proportion of community-based participants, both overall
and across the different care support groups, with different clinical characteristics, reflecting a
potential need for support services.
Results
Tables 3 and 4 array the demographic characteristics of the sample, including gender,
ethnicity and race, level of education, primary language, marital status, and living arrangement.
Findings were similar across each of the care support groups examined. Table 3 describes these
characteristics across the total study population and Table 4 describes the same variables but also
provides a breakdown of distribution according to the different caregiver groups. The sample
was predominantly female (~ 75%), Caucasian (~ 83%), and English speaking (~ 76%). Slightly
more than half (~ 51%) had less than a high school diploma, 40% graduated high school and
only 8% had a college education. Approximately 65% of the study subjects identified themselves
as living with others, and only 20% were married.
Page 87
86
As demonstrated in Table 4, there was minimal variation in frequency or percentage of
respondents across Care Group Category A or Care Group Category B, with respect to gender,
ethnicity, level of education, or language.
Table 3. Frequency of Demographic Variables across Total Sample Population (N = 3,622)1
Total Sample
Variables % Total n
Gender Male 25.5% 920
Female 74.5% 2,689
Ethnicity2 Caucasian
(non-Hispanic)
83%
(2,992)
2,992
Hispanic 7.6% 275
African American 5.1% 183
Other 3 4.2% 153
Education Less than High
School
51.4%
1,819
High School 40.2% 1,424
College 8.4% 298
Language English 76% 2,753
Other 24% 869
Marital Status Married 20.2% 729
Other 79.8% 2,880
Living Arrangement Live Alone 34.9% 1,259
Other 65.1% 2,350 Footnotes: 1. Total population N = 3,622. Some items have missing cases due to the mechanism of data collection
and data entry error.; 2. Individuals can self-identify in more than one racial/ethnic group, and thus the percentage
may be over 100%.; 3. Other includes Native American, Asian, and Pacific Islander.
Page 88
87
Table 4. Frequency of Demographic Variables across Care Support Groups (N = 3,280–3,622)1
Care Group Category A Care Group Category B
Variables
NoSS2
InfCS2
MixS2
ForS2 Test
statistic InfCS
2 SForS
2 Test
statistic
% (N) %(N) %(N) %(N) %(N) %(N)
Gender
Male 26.8%
(72)
25.8%
(532)
24.6%
(276)
26%
(40)
.776 25.8%
(532)
24.8%
(316)
.646
Female 73.2%
(197)
74.2%
(1,533)
75.4%
(845)
74%
(114)
74.2%
(1,533)
75.2%
(959)
Ethnicity3
Caucasian
(non-Hispanic)
85.9%
(231)
82.9%
(1,713)
82.6%
(926)
79.3%
(122)
10.579 82.9%
(1,713)
82.4%
(1,048)
7.275
Hispanic 6.3%
(17)
7,8%
(161)
7.8%
(84)
8.5%
(13)
7.8%
(161)
7.6%
(97)
African-
American
3.7%
(10)
5.2%
(103)
5.2%
(61)
5.8%
(9)
5.2%
(103)
5.5%
(70)
Other 4
4.1%
(11)
4.1%
(85)
4.4%
(47)
6.5%
(10)
4.1%
(85)
4.5%
(57)
Education
< High School 48.7%
(131)
82.9%
(1,051)
82.6%
(559)
79.2%
(78)
2.688 82.9%
(1,051)
82.4%
(637)
2.585
High School 41.6%
(112)
7.8%
(817)
7.5%
(434)
8.5%
(61)
7.8%
(817)
7.6%
(495)
>College 6.7%
(18)
8.5%
(165)
9.4%
(102)
8.4%
(13)
8.5%
(165)
9.2%
(115)
Language
English 73%
(197)
76.6%
(1,587)
75.6%
(852)
76%
(117)
1.627 76.5%
(1,587)
75.6%
(969)
1.616
Other 26.7%
(72)
23.3%
(485)
24.4%
(275)
24%
(37)
23.5%
(485)
24.4%
(312)
Marital Status
Married 22.7%
(61)
20.1%
(415)
18.5%
(226)
17.5%
(27)
1.718 20.1%
(415)
19.8%
(253)
1.138
Other 77.3%
(208)
79.9%
(1,650)
79.8%
(895)
82.5%
(127)
79.9%
(1,650)
80.2%
(1,022)
Living
Arrangement
Live Alone 34.9%
(94)
33.8%
(699)
37.3%
(418)
31.2%
(48)
1.627 33.8%
(699)
36.5%
(466)
2.529
Other 65.1%
(175)
66.2%
(1,366)
62.7%
(703)
68.8%
(106)
66.2%
(1,366)
63.5%
(809)
Footnotes: 1. Total population N = 3,622. Some items have missing cases due to the mechanism of data collection
and data entry error. 2. Care Group Category A: NoSS = No Caregiver Support or Services, InfCS = Informal Care
Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal care supports and
Page 89
88
services), ForS = Formal Services Only. Care Group Category B: InfCS = Informal Care Support Only, SForS =
Receipt of at least some formal services (includes those with formal only and mixed services); 3.Individuals could
self-identify in more than one racial/ethnic group, and thus the percentage may be over 100%; 4. Other includes
Native American, Asian, and Pacific Islander
Table 5 describes the clinical aspects of the total study population. Table 6 describes the
same clinical variables, but provides a breakdown of distribution according to the different
caregiver groups. According to Table 5, a large proportion of people living in the community
(47.5%) had continence issues. Three-quarters of the population were adherent with medications,
although 7.8% adhered less than 80% of the time. Approximately 30.6% took nine or more
medications.
Table 5. Frequency of Variables Related to Clinical and Functional Needs across Total Population
Sample (N = 3,622)1
Total Sample
Variable % Total n
Bladder
Incontinence
Continent 52.5% 1,899
Occasionally incontinent 25.8% 933
Incontinent 21.7% 783
Medication
Adherence
Always adherent/no
medications 73.4% 2,652
Adherent > 80% 18.8% 681
Adherent < 80% 7.8% 282
Number of
Medications
0 4.7% 170
1–4 27.0% 976
5–8 37.6% 1,361
> 9 30.6% 1,108
Footnote: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error.
Page 90
89
Table 6 shows that the numbers of those reporting incontinence across care group
categories. There were statistically significant differences in the distribution of incontinence
across the care groups within Care Group Category A (p < .001). Here, incontinence was more
frequently reported in the MixSS (26.8%) than in all other care groups, including NoS, InfCS,
and ForS (14.9%, 19.9% and 19.5% respectively). There were statistically significant differences
in incontinence across Care Group Category B as well (p <.001). Here, incontinence was more
frequently seen across the SForS (25.9%) group than the InfCS (19.9%) group. There was
minimal variation across care groups with medication adherence. Table 6, however, does
demonstrate some variability in the number of medications taken across the care group
categories. There were significant differences in the number of medications taken across Care
Group Category A (p <.05), with those in the ForS group being somewhat more likely to take
five or more medications (76.5%) than those in the other groups (< 70%).
Page 91
90
Table 6. Frequency of Variables Related to Clinical and Functional Needs across Caregiver Groups (N
= 3,353–3,622)1
Care Group Category A Care Group Category B
Variable NoSS
2 InfCS
2
MixSS2 ForS
2 Test
Statistic3 InfCS
2 SForS
2 Test
Statistic3
% (N) % (N) % (N) % (N) % (N) % (N)
Bladder
Incontinence
Continent 63.2%
(170)
54.7%
(1,130)
47.1%
(530)
44.8%
(69)
12.588
***
54.7%
(1,131)
46.8%
(599)
18.469 ***
Occasionally
Incontinent
21.9%
(59)
25.4%
(526)
26.0%
(293)
35.7%
(55)
25.4%
(526)
27.2%
(348)
Incontinent 14.9%
(40)
19.9%
(411)
26.8%
(302)
19.5%
(30)
19.9%
(411)
25.9%
(332)
Medication
Adherence
Always Adherent/
No Medications
79.9%
(205)
73.3%
(1,515)
72.1%
(812)
77.9%
(120)
1.738
73.2%
(1,520)
72.9%
(934)
.330
Adherent > 80% 15.6%
(42)
19.1%
(394)
19.6%
(220)
16.2%
(25)
19.0%
(394)
19.1%
(245)
Adherent < 80% 8.2%
(22)
7.6%
(158)
8.3%
(93)
5.8%
(9)
7.6%
(158)
8%
(102)
Number of
Medications
None or 0 4.1%
(11)
4.9%
(102)
4.9%
(55)
1.3%
(2)
2.464
*
4.9%
(102)
4.4%
(57)
1.914
1-4 26.4%
(71)
28.1%
(580)
25.9%
(291)
22%
(34)
28.1%
(580)
25.4%
(326)
5-8 39.4%
(106)
37.1%
(768)
37.6%
(422)
42.1%
(65)
37.1%
(768)
37.9%
(487)
9 or more 30.1%
(81)
29.9%
(617)
31.7%
(357)
34.4
(53)
29.9%
(619)
32.1%
(411)
Footnotes: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error. 2. Care Group Category A: NoSS = No Caregiver Support or Services, InfCS = Informal Care
Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal care supports and
services), ForS = Formal Services Only; Care Group Category B: InfCS = Informal Care Support Only, SForS =
receipt of at least some formal services (includes those with formal only and mixed services); 3. Test statistics
included F statistics, significance, identified with p value: *p < .05. **p < .01. ***p < .001.
Cognitive and psychiatric health variables are displayed in Tables 7 and 8, including
short-term and procedural memory, cognitive decision making, comprehension, expression, and
Alzheimer’s disease diagnosis. Alzheimer’s disease diagnoses can vary substantially in
presentation and severity; this study, however, looked only at whether or not there was a
diagnosis made. Table 7 presents the cognitive and psychiatric health variables for the total study
Page 92
91
population. Table 8 shows the same variables as they were distributed across the different
caregiver groups. Across the population sample as a whole (Table 7), 50–60% of people residing
in the community had a memory problem, either short-term (60%), or procedural (53%).
Furthermore, 43% of people reported being minimally to moderately impaired with decision
making, though close to three-quarters (~73%) understood or usually understood information
shared with them. Twelve percent were diagnosed with Alzheimer’s disease.
Table 7. Frequency of Cognitive and Psychiatric Health Variables across Total Sample Population (N
= 3,622)1
Variables
Total Sample
% Total n
Memory:
short-term
memory ok 39.8% 1,440
memory problem 60.2% 2,175
Memory:
procedural
memory ok 46.9% 1,695
memory problem 53.1% 1,920
Cognition:
Decision
Making
no issues/independent 22% 796
modified independence 22.1%
799
minimally impaired 22.7% 821
moderately impaired 21% 758
severely impaired 12.2% 441
Comprehension Understands 40.9% 1,480
usually understands 31.8% 1,149
often understands 12.5% 451
sometimes understands 12.5% 452
rarely/never understands 2.3% 83
Expression Understood 47.7% 1,725
usually understood 29.3% 1,058
often understood 12% 435
sometimes understood 8.2% 296
rarely/never understood 2.8% 101
Alzheimer’s not applicable 88% 3,183
disease present 12% 432 Footnote: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error.
Page 93
92
Across both Care Group Categories A and B, there was minimal variation across type of
support with respect to short-term memory issues. However, cognitive decision making ability
varied somewhat across care groups, with respect to both Care Group Category A (p < .01)
and B (p < .001). Across Care Group Category A, high numbers of individuals were independent
with decision making ability in the NoSS and ForS care groups (26.4% and 26%) versus the
InfCS and MixS groups (21.3% and 21.7%). Across Care Group B, higher numbers of
individuals were independent (either fully or modified) in the SForS (22.3%) than in the InfCS
(21.3%) care group. There were no significant differences in the ability to understand others
(comprehension) across Care Group A and a small statistically significant difference across the
InfCS (39.7%) and SForS (41.9%) groups in Care Groups B (p < .05). There were statistically
significant differences in the ability to be understood in Care Group A (p < .01) but not Care
Group B. Here, 57.1% of those in the ForS group could express themselves and be understood as
compared to 50.6% in the NoSS group and ~ 46.0% in both the InfCS and MixSS groups. In
Care Group A, Alzheimer’s disease was significantly more frequent in the InfCS and MixSS
groups (13.7% and 11.5% respectively) than the ForS and NoSS groups (3.7% and 6.5%; p <
.001). Additionally, across Care Group Category B, the presence of an Alzheimer’s disease
diagnosis was more frequently seen across the InfCS (13.7%) group than the SForS (10.9%)
group (p < .001).
Page 94
93
Table 8. Frequency of Variables Related to Cognitive and Psychiatric Health across the Different
Caregiver Groups (N = 3,353–3,622)1
Care Group Category A Care Group Category B
Variable NoSS
2 InfCS
2 MixSS
2 ForS
2 Test
Statistic3 InfCS
2 SForS
2 Test
Statistic3
% (N) % (N) % (N) % (N) % (N) % (N)
Memory:
Short-Term
Memory Ok 43.5%
(117)
38.8%
(802)
40.6%
(457)
41.6%
(64)
2.987 38.8%
(804)
40.8%
(523)
2.948
Memory
Problem
56.5%
(152)
61.2%
(1,265)
59.4%
(668)
58.4%
(90)
61.2%
(1,268)
59.2%
(758)
Memory:
Procedural
Memory Ok 52.8%
(142)
46.0%
(951)
46.5%
(523)
51.3%
(79)
6.312 46.1%
(954)
47.2%
(604)
5.103
Memory
Problem
47.2%
(127)
54.0%
(1,116)
53.5%
(602)
48.7%
(75)
53.9%
(1,117)
52.8%
(677)
Cognition:
Decision
Making
No Issues-
Independent
26.4%
(71)
21.3%
(441)
21.7%
(244)
26.0%
(40)
3.748
**
21.4%
(444)
22.3%
(286)
4.685
***
Modified
Independence
22.7%
(61)
20.7%
(427)
24.5%
(276)
22.7%
(35)
20.6%
(427)
24.3%
(311)
Minimally
Impaired
21.6%
(58)
23.4%
(484)
21.5%
(242)
24.0%
(37)
23.5%
(486)
21.8%
(279)
Moderately
Impaired
21.2%
(57)
21.4%
(442)
20.4%
(229)
19.5%
(30) 21.3%
(442)
20.2%
(259)
Severely
Impaired
8.2%
(22)
13.2%
(273)
11.9%
(134)
7.8%
(12)
13.2%
(273)
11.4%
(146)
Comprehension Understands 46.8%
(126)
39.6%
(819)
40.9%
(460)
48.7%
(75)
2.897 39.7%
(822)
41.9%
(537)
3.551
*
Usually
Understands
30.9%
(83)
31.8%
(658)
32.4%
(364)
28.6%
(44)
32.4%
(659)
31.9%
(408)
Often
Understands
11.2%
(30)
12.9%
(266)
12.4%
(139)
10.4%
(16)
12.9%
(267)
12.1%
(155)
Sometimes
Understands
8.9%
(24)
13.4%
(278)
12.2%
(137)
8.4%
(13)
13.4%
(278)
11.7%
(150)
Rarely/Never
Understands
2.2%
(6)
2.2%
(46)
2.2%
(25)
3.9%
(6)
2.2%
(46)
2.4%
(31)
Expression Understood 50.6%
(136)
46.6%
(963)
47.8%
(538)
57.1%
(88)
3.748
**
46.7%
(976)
49%
(628)
.922
Usually
Understood
27.9%
(75)
29.7%
(613)
29.8%
(335)
22.7%
(35)
29.6%
(614)
28.9%
(370)
Often
Understood
11.9%
(32)
12.6%
(261)
11.3%
(127)
9.7%
(15)
12.6%
(261)
11.1%
(142)
Sometimes
Understood
7.4%
(20)
8.4%
(174)
8.4%
(94)
5.2%
(8) 8.4%
(174)
7.9%
(102)
Rarely/Never
Understood
2.2%
(6)
2.7%
(56)
2.8%
(31)
5.2%
(8)
2.7%
(56)
3%
(39)
Alzheimer’s Not Applicable 96.3%
(259)
86.3%
(1,784)
88.5%
(996)
93.5%
(144)
27.747
***
86.3%
(1,789)
89.1%
(1,142)
24.583
***
Disease Present 3.7%
(10)
13.7%
(283)
11.5%
(129)
6.5%
(10)
13.7%
(283)
10.9%
(139)
Footnotes: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error; 2. Care Group Category A: NoSS = No Caregiver Support or Services, InfCS = Informal Care
Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal care supports and
services), ForS = Formal Services Only; Care Group Category B: InfCS = Informal Care Support Only, SForS =
Page 95
94
Receipt of at least some formal services (includes those with formal only and mixed services); 3. Test statistics
included chi-square, and F statistics, significance, identified with p value: * p < .05. ** p < .01. *** p < .001.
Tables 9 and 10 display the descriptive findings for nursing facility eligibility across the
total population and the caregiver support groups. As noted in Table 9, 11% of the sample met
nursing facility eligibility criteria in the state of Massachusetts.
Table 9. Frequency of Nursing Facility Eligibility across Total Sample Population (N = 3,622)1
Total Sample
Variable % Total n
Nursing
Facility
Eligibility
Eligible 11.2% 406
Not Eligible 88.8% 3,216
Footnote: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error.
In Table 10, nursing facility eligibility is reported more frequently in the MixSS (14.8%)
and ForS (14.9%) versus the NoSS (9.3%) and InfCS (9.2%), across Care Group Category A (p
< .001). Similar results were obtained when using Care Group Category B, with a smaller
percentage of those in the InfCS (9.2%) than SForS (17.4%) group being eligible (p < .001).
Page 96
95
Table 10. Frequency of Nursing Facility Eligibility across Caregiver Support Groups (N = 3,353–
3,622)1
Care Group Category A Care Group Category B
Variable NoSS
2 InfCS
2 MixSS
2 ForS
2 Test
Statistic3 InfCS
2 SForS
2 Test
statistic3
% (N) % (N) % (N) % (N) % (N) % (N)
Nursing
Facility
Eligibility
Eligible
9.3%
(25)
9.2%
(191)
14.8%
(167)
14.9%
(23)
26.140
***
9.2%
(191)
17.4.%
(190)
26.139
***
Not Eligible
90.7%
(244)
90.8%
(1,881)
85.2%
(960)
85.1%
(131)
90.8%
(1,881)
82.6%
(1,091)
Footnotes: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error. 2. Care Group Category A: NoSS = No Caregiver Support or Services, InfCS = Informal Care
Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal care supports and
services), ForS = Formal Services Only; Care Group Category B: InfCS = Informal Care Support Only, SForS =
receipt of at least some formal services (includes those with formal only and mixed services). 3. Test statistics
included chi-square with degrees of freedom identified in parentheses, significance, identified with p value:
* p < .05. ** p < .01. *** p < .001.
Perception of health status is displayed in Tables 11 and 12 and includes beliefs of poor
health and beliefs of capability for increased independence. According to Table 11, just over
80% of people within the population did not feel that they are capable of increased
independence. However, only about 30% of the population sampled believed they were in poor
health.
Table 11. Frequency of Perception of Health Status across Total Sample Population (N = 3,622)1
Total Sample
Variable %
Total n
Believes Capable of
Increased Independence
False 81.9% 1,962
True 18.1% 653
Belief Poor Health False 70.6% 2,553
True 29.4% 1,062 Footnote: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error.
Page 97
96
Table 12 indicates that there were fewer individuals in the InfCS and MixSS groups of
Care Group Category A that believed they were capable of increased independence (17.3% and
16.4%, respectively) than those in the NoSS and ForS groups (26.8% and 25.3%, respectively;
p < .001). In contrast, across Care Group Category B, the InfCS group was 20% more likely to
believe they were capable of increased independence than the SForS group (57.8% v. 35.6%;
p < .001). Table 12 reveals no statistically significant differences in the belief in poor health
across either Care Group A or Care Group B, though 57.2% in the InfCS group reported being in
poor health, compared to 35.9% in the SForS group, a 22% point difference.
Table 12. Frequency of Perception of Health Status across Caregiver Support Groups (N = 3,353–
3,622)1
Care Group Category A Care Group Category B
NoSS2
InfCS2
MixSS2
ForS2 Test
Statistics3 InfCS
2 SForS
2 Test
Statistics3
Variable % (N) % (N) % (N) % (N) % (N) %(N)
Believes Capable
of Increased
Independence
False
73.2%
(197)
82.7%
(1,710)
83.6%
(940)
74.7%
(115)
21.945
***
57.8%
(1,714)
35.6%
(1,056)
14.829
***
True
26.8%
(72)
17.3%
(357)
16.4%
(185)
25.3%
(39)
54.7%
(358)
34.4%
(225)
Belief of Poor
Health
False
74.7%
(201)
70.4%
(1,455)
69.5%
(782)
74.7%
(115) 4.094
57.0%
(1,457)
35.2%
(899) 2.389
True
25.3%
(68)
29.6%
(612)
30.5%
(343)
25.3%
(39)
57.7%
(615)
35.9%
(382)
Footnotes: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error. 2. Care Group Category A: NoSS = No Caregiver Support or Services, InfCS = Informal Care
Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal care supports and
services), ForS = Formal Services Only; Care Group Category B: InfCS = Informal Care Support Only, SForS =
Page 98
97
Receipt of at least some formal services (includes those with formal only and mixed services). 3. Test statistics
include F statistic, significance, identified with p value: * p < .05. ** p < .01. *** p < .001.
Tables 13 and 14 present data regarding activities of daily living (ADL). ADL measures
functional status based on performance of activities such as bathing, hygiene, dressing, toilet use,
eating, transferring, and mobility. Table 13 shows that, on average, individuals within the
population studied needed help with at least 1.24 ADLs. There were statistically significant
differences across both Care Group A and B, however (both p < .001). In Table 14, within Care
Group Category A, NoSS had the lowest average ADL score (.91), while the MixSS group had
the highest score (1.53). Within Care Group Category B, the InfCS group had a slightly lower
ADL score (1.28) than the SForS Group (1.49).
Tables 13 and 14 also describe results related to instrumental activities of daily living
(IADLs). These are activities such as meal preparation, ordinary housework, managing finances,
managing medications, phone use, shopping, and transportation. Across the entire sample
population, as seen in Table 13, most individuals ranked as receiving some help to full help with
IADLS, with a mean IADL score of 2.66. There were statistically significant differences across
both Care Group A and B, however (both p < .001). According to Table 14, within Care Group
Category A, those people within the ForS group had significantly greater difficulty with IADLs.
A performance code of 3 indicates that the service was done by someone else. With an average
score of 2.96, the data shows that the ForS group was closer than the other care support groups to
needing others to perform their IADLs.. Within care group category B, those in receipt of some
degree of formal services were slightly more likely to have IADL difficulties than those in the
InfCS group (2.96 v. 2.64).
Page 99
98
Table 13. Frequency of Functional Status across Total Sample Population (N = 3,622)1
Variable Total Sample
Mean St. Dev.
Functional Status2
1.24 1.16
IADL Score3
2.66 1.77
Footnotes: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error. 2. Functional status was calculated as an average score, based on the following activities of
daily living: mobility in bed, transfer, locomotion in home, dressing the upper body and dressing the lower body,
and the level of assistance from 0 (independent) to 6 (total dependent). 3. IADL was calculated as an average score,
based on the following instrumental activities of daily living – meal preparation, ordinary housework, managing
finances, managing medications, phone use, shopping and transportation and the level of performance is coded as 0
(independent) to 3 (by others).
Table 14. Frequency of Functional Status across Caregiver Support Groups (N = 3,609–3,622)1
Footnotes: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error. 2. Care Group Category A: NoSS = No Caregiver Support or Services, InfCS = Informal Care
Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal care supports and
services), ForS = Formal Services Only; Care Group Category B: InfCS = Informal Care Support Only, SForS =
Receipt of at least some formal services (includes those with formal only and mixed services). 3. Functional status
was calculated as an average score, based on the following activities of daily living: mobility in bed, transfer,
locomotion in home, dressing the upper body and dressing the lower body, and the level of assistance from 0
(independent) to 6 (total Dependent). 4. IADL was calculated as an average score, based on the following
instrumental activities of daily living: meal preparation, ordinary housework, managing finances, managing
medications, phone use, shopping and transportation, and the level of performance is Coded as 0 (independent) to 3
(by others). 5. Test statistics include F statistics, significance, identified with p value:
* p < .05. ** p < .01. *** p < .001.
Tables 15 and 16 present findings related to depression risk and other behavioral
symptoms. Table 15 demonstrates the findings across the entire sample population. With a score
of 2.7, elders in this study tended to report at least two depressive symptoms, on average. There
were statistically significant differences across both Care Group Category A and B (both
Variable
Care Group Category A Care Group Category B
NoSS2
InfCS2
MixS2
ForS2
Test
Statistic5
InfCS2
SForS2
Test
Statistic5
Mean St.
Dev.
Mean St.
Dev.
Mean St.
Dev.
Mean St.
Dev.
Mean St.
Dev.
Mean St.
Dev.
Functional
Status3 .91 1.06 1.28 1.17 1.53 1.24 1.24 1.16
23.042
*** 1.28 1.17 1.49 1.24
30.55.4
***
IADL
Scores4 2.27 1.41 2.64 2.07 2.75 2.02 2.96 1.61
12.145
*** 2.64 1.35 2.77 1.26
16.484
***
Page 100
99
p<.001). In Care Group Category A, depression risk was higher, on average, in the InfCS group
(3.29) and NoSS group (2.77) than for those in the mixed or formal supports. Across Care Group
Category B, average depression risk was significantly higher in the InfCS group (3.29) than in
the SForS group (2.40).
Behavioral symptoms are also indicators of problems and potential increased need.
Behavior is measured by actual behaviors demonstrated, with score of “1” indicating that a
problematic behavior that was difficult to manage occurred. The total sample demonstrated
behavioral symptoms with a mean of 0.65, indicating that the population did have behaviors, but
that they were not unmanageable. There were statistically significant differences across both
Care Group A and B, however (both p < .001). Across Care Group Category A, behavioral
symptoms were more frequently seen, on average, in the InfCS group (.80) and the ForS group
(.71) than in the NoSS group (.63) or the MixS (.49) group. Within Care Group Category B,
behavioral symptoms were significantly more frequent, on average, in the InfCS group (.80) than
in the SForS group (.50).
Table 15. Frequency of Cognitive Risks of Depression and Behavior across Total Sample Population
(N = 3,622)1
Total Sample
Variable Mean St. Dev.
Depression Risk2
2.7 3.5
Behavior Problems3
0.6 1.4
Footnotes: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error. 2. Depression risk is based on a calculated score: when a respondent reports a mood problem
requiring intervention and has two or more symptoms. 3. Behavior problems are a risk score when an individual
exhibits an identified symptom.
Page 101
100
Table 16. Frequency of the Care Support Group by Cognitive Risks: Depression and Behavior (N =
3,609–3,622)1
Footnotes: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error. 2. Care Group Category A: NoSS = No Caregiver Support or Services, InfCS = Informal Care
Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal care supports and
services), ForS = Formal Services Only; Care Group Category B: InfCS = Informal Care Support Only, SForS =
Receipt of at least some formal services (includes those with formal only and mixed services). 3. Depression risk is
based on a calculated score, when someone has a mood problem requiring intervention and has two or more
Symptoms; 4. Behavior problems are a risk score when someone exhibits an identified symptom. 5. Test statistics
included chi-square with degrees of freedom identified in parentheses, significance, identified with p value:
* p < .05. ** p < .01. *** p < .001.
Aim B
Factors that Predict Membership in Care Support Group Categories
The second aim of the study was to identify differences across care support groups on
selected independent variables. I used inferential statistics to draw conclusions from the sample
population tested (both Care Group Categories A and B), initially through bivariate statistical
analysis and then multivariate statistical analysis. Here, I present the results from testing each
hypothesis.
Using chi-square, I initially analyzed each hypothesis with a categorical variable to assess
the relationship between the hypothesized variable and type of care support group. I chose chi-
square specifically to explore the proportion of cases within categories; and if differences existed
Variable
Care Group Category A2 Care Group Category B
2
NoSS2
InfCS2
MixS2
ForS2 Test
Statistics5 InfCS
2 SForS
2 Test
Statistics5
Mean St.
Dev. Mean
St.
Dev. Mean
St.
Dev. Mean
St.
Dev. Mean
St.
Dev. Mean
St.
Dev.
Depression
Risk3
2.77 3.63 3.29 3.91 2.41 3.28 2.32 3.15 15.733
***
3.29 3.91 2.40 3.26 23.566
***
Behavior
Problems4
.63 1.33 .80 1.54 .49 1.19 .71 1.64 11.404
***
.80 1.54 .52 1.26 15.455
***
Page 102
101
between the number of individuals who were in the hypothesized variable group (ethnicity,
education, marital status, language, living arrangement, and nursing facility eligibility) across
Care Group Category A (NoSS, InfCS, MixSS, ForS). I then repeated the chi-square analysis for
the hypothesized variables across Care Group Category B (InfCS and SForS). ANOVA and
independent samples t-tests were employed to examine the relationship between the continuous
variable analyzed-ADL functional status, as hypothesized in 1F, across Care Group Categories A
and B, respectively.
After employing bivariate analysis, I performed logistic regression to confirm bivariate
relationships while controlling for other factors likely to have an effect on the dependent
variables. Regressions were completed using each dependent variable within Care Group
Category A, NoSS and dummy variable, InfCS and dummy variable, MixSS and dummy
variable, and ForS and dummy variable; and repeated using each dependent variable with Care
Group Category B, InfCS and dummy variable, and SForS and dummy variable. Each
regression was completed across the hypothesized variables, while controlling for additional
predictor variables identified in table 1.
Results Summary for Aim B: Bivariate and Multivariate Analyses
Hypothesis 1A (H1A)
Hispanics and African–Americans would be more likely than non-Hispanic whites to be in the InfCS
group(and thus cared for by informal care givers) than in the other care support groups.
I analyzed Hypothesis 1A using chi-square and then logistic regression to discern any
differences in the proportion of Hispanics and African–Americans across care support groups.
Page 103
102
No cases with missing data were found in the data set. Thus, for H1A, 3,622 responses from
participants were received and 3,450 were retained; n = 3,450. The cases that were coded as
additional racial and ethnicity types within the data set, which included Asian, Pacific Islander,
and American Indian, were removed, identifying the 172 case difference.
Bivariate: Chi-Square Analysis of Hypothesis 1A
My analysis showed no significant difference in the number of Hispanics and African–
Americans across care support groups in Care Group Category A; χ2 (1, n = 3,450) p = .504. I
also found no significant difference in the number of Hispanics and African–Americans across
care support groups in Care Group Category B; χ2 (1, n = 3,450) p = .366. Table 17 provides the
results of the chi-square analysis. These results indicate that there is no significant difference
between the number of Hispanics and African–Americans across care support groups.
Table 17. Bivariate Statistical Analysis Details for Hypotheses 1A to 1G (N = 3,622)
Variable Care Group Category A Care Group Category B
Chi2/F (Df) P Chi
2/F (Df) P
H1A: Ethnicity 2.346 (3) .504 2.013 (2) .366
H1B: Education 2.171 (3) .538 2.070 (2) .355
H1C: Language 1.627 (3) .653 1.616 (2) .446
H1D: Marital Status 1.718 (3) .633 1.138 (2) .566
H1E: Living Arrangement 4.761 (3) .101 2.529 (2) .282
H1F: Functional Status 23.042 (3) *** 30.554 (2) ***
H1G: Nursing Facility Eligibility 26.140 (3) *** 26.139 (2) ***
Note: Statistical significance identified: * p < .05. ** p < .01. *** p < .001.
Page 104
103
Multivariate Analysis of Hypothesis 1A
After employing bivariate analysis, multivariate logistic regression was completed and
the analysis was completed across Care Group Categories A and B, tables 18 and 19 provide the
detailed results. The multivariate model reporting differences in the number of Hispanics and
African-Americans across Care Group Category A was not statistically significant, 2
(2, n =
3,450) = 1.107, p = .293. Similar results were seen between the number of Hispanics and
African-Americans and Care Group Category B, 2
(1, n = 3,450) = .968, p = .325. These results
indicate that there are no differences in the number of Hispanics and African–Americans across
care support groups at the multivariate level, and that the predictors could not reliably distinguish
between care group. Thus, the null hypothesis could not be rejected.
Table 18. Multivariate Statistical Analysis of Ethnicity1 and Care Support Group A (n = 3,450)
Predictor Beta SE Wald Df P
Odds Ratio
NoSS2 –.069 .082 .710 1 .4 .933
InfCS2
–.025 .041 .370 1 .543 .975
MixS2
.029 .044 .423 1 .516 1.028
ForS2
.095 .091 1.082 1 .298 1.10
Constant
.103 .098 1.104 1 .293 –.089
Test Test value (Df) P
Overall model
evaluation:
–2 Log
Likelihood
39.715 1.107(2) .293
Goodness of
Fit: Pearson 1.307(2) .520
Footnote: 1. Ethnicity is defined as Caucasian or other. 2. NoSS = No Caregiver Support or Services, InfCS =
Informal Care Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal
care supports and services), ForS= Formal Services Only.
Table 19. Multivariate Statistical Analysis for Ethnicity1 and Care Support Group B (n = 3,450)
Predictor Beta SE Wald Df P
Odds Ratio
Page 105
104
InfCS2
–.025 .041 .370 1 .543 .972
SForS2
.46 .042 1.159 1 .282 1.047
Constant
.097 .099 .961 1 .327 –.10
Test Test value 2(Df) P
Overall model
evaluation:
–2 Log
Likelihood
28.224 .968(1) .325
Goodness of
Fit: Pearson 1.128(1) .288
Footnote: 1. Ethnicity is defined as Caucasian or other. 2. InfCS = Informal Care Support only, SForS = Some
Formal Services.
Hypothesis 1B (H1B)
People with lower levels of education would be more likely than their more educated counterparts to be
in the InfCS group(and thus cared for by informal care givers) than in the other care support groups.
I used chi-square and logistic regression to test Hypothesis 1B for differences in the
number of those with lower levels of education across Care Group Categories A and B. I used
logistic regression to explore how the relationship of other care support groups affects the
relationship between the number of those with lower levels of education and care support group
membership. There were 81 cases with missing data in the data set, which meant that for H1B, I
retained 3,541 of the 3,622 responses received. The missing cases were related to the
mechanism of data collection.
Bivariate Chi-Square Analysis of Hypothesis 1B
There was no significant difference in the number of people having lower levels of
education within Care Group Category A; χ2 (1, n = 3,541) p = .538. Additionally, I found no
significant difference in the number of people having lower levels of education within Care
Page 106
105
Group Category B ; χ2 (1, n = 3,541) p = .355. Table 17 provides the results of the chi-square
analysis. These results indicate that there was no significant difference in the number of
individuals with lower level of education across care support groups.
Multivariate Analysis of Hypothesis 1B
After employing bivariate analysis, I performed logistic regression to control for the
effects of care support groups on the likelihood that the number of people with lower levels of
education would be associated with the informal care support group. The analysis was completed
across Care Group Categories A and B, and results are detailed in Tables 20 and 21. The
multivariate model reporting differences in the number of people with lower levels of education
across Care Group Category A was not statistically significant, 2
(1, n = 3,541) = 1.846, p
=.174. Similar results were seen between the number of individuals with lower levels of
education and Care Group Category B, 2(1, n = 3,540) = 2.011 , p = .156. These results indicate
that there were no differences in the number of people with lower levels of education across care
support groups at the multivariate level. Thus, the null hypothesis could not be rejected.
Page 107
106
Table 20. Multivariate Statistical Analysis for Education1 and Care Support Group A (n = 3,541)
Predictor Beta SE Wald Df P
Odds Ratio
NoSS2 .033 .101 .107 1 .744 1.034
InfCS2
.039 .053 .538 1 .463 1.040
MixS2
–.053 .057 .871 1 .351 .948
ForS2
–.015 .129 .013 1 .909 .985
Constant
–.160 .118 1.842 1 .175 –.282
Test Test value 2(Df) P
Overall model
evaluation:
–2 Log
Likelihood
37.659 1.846(1) .174
Goodness of
Fit: Pearson .330(2) .848
Footnote: 1. Level of Education is defined as high school graduate or less, or more than high school graduate. 2.
NoSS = No Caregiver Support or Services, InfCS = Informal Care Support only, MixSS = Mixed Care Support (i.e.:
those with a combination of informal and formal care supports and services), ForS = Formal Services Only.
Table 21. Multivariate Statistical Analysis for Education1 and Care Support Group B (n = 3,450)
Predictor Beta SE Wald Df P
Odds Ratio
InfCS2
.039 .053 .538 1 .463 1.040
SForS2
–.056 .055 .877 1 .346 .949
Constant
–.169 .119 2.011 1 1.56
Test Test value 2(Df) P
Overall model
evaluation:
–2 Log
Likelihood
26.461 2.011(1) .156
Goodness of
Fit: Pearson .080(1) .777
Footnote: 1. Level of Education is defined as high school graduate or less, or more than high school graduate. 2.
InfCS = Informal Care Support only, SForS = Some Formal Services.
Hypothesis 1C (H1C)
People with a primary language other than English would be more likely than those whose
primary language is English to be in the InfCS group (and thus cared for by informal care
givers) than in the other care support groups.
Page 108
107
Using chi-square and logistic regression, I analyzed Hypothesis 1C for any differences in
the proportion of those whose primary language is other than English across care support groups
for both Care Group Category A and B. None of the 3,622 cases were missing data (N = 3,622).
Bivariate chi-Square Analysis of Hypothesis 1C
I found no significant difference in the number of people having a primary language other
than English within Care Group Category A ; χ2 (1, N = 3,622) p = .653. Additionally, there was
no significant difference in the number of people having a primary language other than English
within Care Group Category B; χ2 (1, n = 3,450) p = .446. Table 17 provides the results of the
chi-square analysis. These results indicate that there was no significant difference in the number
of individuals with a primary language other than English across care support groups.
Multivariate Analysis of Hypothesis 1C
After employing bivariate analysis, logistic regression was completed, to control for the
effects of care support groups on the likelihood that primary language other than English would
be associated with the informal care support group membership, tables 22 and 23 detail the
results across Care Group Category A and B. People with lower levels of education across Care
Group Category A was not statistically significant, 2
(1, N = 3,622) = .003, p = .958. There
were similar results between the number of individuals with lower levels of education and Care
Group Category B, 2 (1, n = 3,450) = .002, p = .969. These results indicate that there were no
differences in the number of people with a primary language other than English across care
support groups at the multivariate level. Thus, the null hypothesis could not be rejected.
Page 109
108
Table 22. Multivariate Statistical Analysis for Primary Language1 and Care Support Group A (N =
3,622)
Predictor Beta SE Wald Df P
Odds Ratio
NoSS2 –.097 .150 .414 1 .520 .908
InfCS2
.11 .081 1.874 1 .174 1.117
MixS2
–.082 .087 .891 1 .345 .921
ForS2
–.067 .198 .114 1 .736 .935
Constant
.004 .076 .003 1 .958
Test Test value 2 (Df) P
Overall model
evaluation:
–2 Log
Likelihood
41.798 .003(1) .958
Goodness of
Fit: Pearson 1.598(2) .450
Footnote: 1. Primary Language is defined as English or Other. 2. NoSS = No Caregiver Support or Services, InfCS
= Informal Care Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal
care supports and services), ForS= Formal Services Only.
Table 23. Multivariate Statistical Analysis for Primary Language
1 and Care Support Group B (n =
3,450)
Predictor Beta SE Wald Df P
Odds Ratio
InfCS2
.11 .081 1.874 1 .174 1.117
SForS2
–.089 .085 1.110 1 .292 .915
Constant
.003 .077 .002 1 .969
Test Test value 2(Df) P
Overall model
evaluation:
–2 Log
Likelihood
29.989 .002(1) .969
Goodness of
Fit: Pearson 1.589(1) .208
Footnote: 1. Primary language is defined as English or other. 2. InfCS = Informal Care Support only, SForS = Some
Formal Services.
Hypothesis 1D (H1D)
People who are married would be more likely than those who are unmarried to be in the InfCS
group (and thus cared for by informal caregivers) than in the other care support groups.
Page 110
109
I analyzed Hypothesis 1D using chi-square and logistic regression to test if there were
any differences in the number of those married across care support groups, and then to explore
the predictors of the relationships for both Care Group Categories A and B. Thirteen cases had
missing data in the data set. Thus, for H1D I retained 3,609 of the 3,622 responses from
participants (n = 3,609). Marital status data was missing due to the mechanism of data entry and
miscoded entries.
Bivariate Chi-Square Analysis of Hypothesis 1D
There were no significant differences in marital status within Care Support Group A ; χ2
(1, n = 3,609) p = .633. Additionally, there were no significant differences in marital status
within Care Group Category B; χ2 (1, n = 3,609) p = .566. Table 17 provides the results of the
chi-square analysis. These results indicate that there was no significant difference in the number
of individuals who were married across care support groups.
Multivariate Analysis of Hypothesis 1D
After employing bivariate analysis, logistic regression was completed to determine
predictors on the likelihood of married persons associated with the informal care support group.
The detailed analysis was completed across Care Group Categories A and B, and can be found in
tables 24 and 25. The multivariate model reporting differences in the number of married people
across Care Group Category A was not statistically significant, 2
(1, N = 3,622) = .632, p =
.427. Similar results were seen between the number of married people and Care Group Category
B ,2
(1, n = 3,450) = .490, p = .484. These results indicate that there are no differences in the
Page 111
110
number of married people across care support groups at the multivariate level. Thus, the null
hypothesis could not be rejected.
Table 24. Multivariate Statistical Analysis for Marital Status1 and Care Support Group A (N = 3,622)
Predictor Beta SE Wald Df P
Odds Ratio
NoSS2 .112 .170 .432 1 .511 1.118
InfCS2
–.028 .091 .096 1 .756 .972
MixS2
.051 .098 .272 1 .602 1.052
ForS2
–.286 .233 1.510 1 .219 .751
Constant
–.065 .081 .632 1 .427 –.224
Test Test value 2 (Df) P
Overall model
evaluation:
–2 Log
Likelihood
40.806 .632(1) .427
Goodness of
Fit: Pearson 1.081(2) .582
Footnote: 1. Marital status is defined as married or other. 2. NoSS = No Caregiver Support or Services, InfCS =
Informal Care Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal
care supports and services), ForS = Formal Services Only.
Table 25. Multivariate Statistical Analysis for Marital Status1 and Care Support Group (n = 3,450)
Predictor Beta SE Wald Df P
Odds Ratio
InfCS2
–.028 .091 .096 1 .756 .972
SForS2
–.003 .095 .001 1 .971 .997
Constant
–.058 .082 .491 1 .483
Test Test value 2 (Df) P
Overall model
evaluation:
–2 Log
Likelihood
28.766 .490(1) .484
Goodness of
Fit: Pearson .626(1) .429
Footnote: 1. Marital status is defined as Married or other. 2. InfCS = Informal Care Support only, SForS = Some
Formal Services.
Page 112
111
Hypothesis 1E (H1E)
People who live with others would be more likely than those who live alone to be in the InfCS
group (and thus cared for by informal caregivers) than those other care support groups.
Using chi-square and logistic regression, I analyzed Hypothesis 1E to evaluate living
arrangement, and care giving group. I used chi-square to determine if there were differences in
the proportion of people living with others based on care support groups for both Care Group
Categories A and B. I then used logistic regression to explore how the relationship of other care
support groups affects and predicts that relationship. There were thirteen cases with missing data
in the data set. As a result, for H1E I retained 3,609 of the 3,622 responses received (n = 3,609).
Data from living arrangement was missing because of the mechanism of data entry and miscoded
entries.
Bivariate Chi-Square Analysis of Hypothesis 1E
There were no significant differences in the number of people living with others within
Care Support Group A; (1, n = 3,609) p = .190. Additionally, there were no significant
differences in number of individuals living with others within Care Group Category B; χ2 (1, n =
3,450) p = .282. Table 17 provides the results of the chi-square analysis. These results indicate
that there was no significant difference in the number of individuals who were living with others
across care support groups.
Multivariate Analysis of Hypothesis 1E
After employing bivariate analysis, I performed logistic regression. The findings are
detailed in tables 26 and 27. The multivariate model reporting differences in the number of
Page 113
112
individuals living with others across Care Group Category A was not statistically significant, 2
(1, n = 3,609) = 1.288, p = .257. Similar results were seen between the number of individuals
living with others and Care Group Category B, 2 (1, n = 3,450) = 1.820, p = .177. These results
indicate that there were no differences in the number of individuals living with others across care
support groups at the multivariate level. Thus, the null hypothesis could not be rejected.
Table 26. Multivariate Statistical Analysis for Living Arrangement1 and Care Support Group A (N
=3,622)
Predictor Beta SE Wald Df P3
Odds Ratio
NoSS2 .019 .142 .018 1 .892 1.019
InfCS2
–.141 .075 3.550 1 .059 .868
MixS2
.200 .080 6.253 1 .012** 1.222
ForS2
–.227 .185 1.507 1 .220 .797
Constant
.078 .068 1.287 1 .257
Test Test value 2(Df) P
Overall model
evaluation:
–2 Log
Likelihood
44.204 1.288(1) .257
Goodness of
Fit: Pearson 3.426(2) .180
Footnote: 1. Living Arrangement indicates that the person lives alone or with others. 2. NoSS = No Caregiver
Support or Services, InfCS = Informal Care Support only, MixSS = Mixed Care Support (i.e.: those with a
combination of informal and formal care supports and services), ForS = Formal Services Only. 3. Significance p
value: # p < .1. * p < .05. ** p < .01. *** p < .001.
Page 114
113
Table 27. Multivariate Statistical Analysis for Living Arrangment1 and Care Support Group B (n =
3,450)
Predictor Beta SE Wald Df P
Odds Ratio
InfCS2
–.141 .075 3.550 1 .059 .868
SForS2
.146 .088 3.553 1 .059 1.158
Constant
.093 .069 1.825 1 .177 .135
Test Test value 2 (Df) P
Overall model
evaluation:
–2 Log
Likelihood
29.474 1.820 (1) .177
Goodness of
Fit: Pearson .707(1) .401
Footnote: 1. Living Arrangement indicates that the person lives alone or with others. 2. InfCS = Informal Care
Support only, SForS = Some Formal Services.
Hypothesis 1F (H1F)
People with more deficits in activities of daily living (ADLs) would be more likely than those
with less deficits to be in the InfCS group (and thus cared for by informal caregivers) than those
in the other care support groups.
To analyze Hypothesis 1F, to identify differences in caregiver group type and functional
status I used analysis of variance (ANOVA) and independent samples t-tests. I employed
ANOVA to determine if differences existed between participant Care Group Category A
subgroups (NoSS, InfCS, MixS, ForS) and their average scores on functional status scale. The
caregiver types within Care Support Group A serve as the independent variable or factor for this
analysis. With Care Support Group B, the caregiver types of subgroups are Informal Support or
InfCS and Some Formal Services or SForS serve as the independent variable or factor for this
analysis, and independent samples t-tests were used for analysis.
Page 115
114
I then used Multivariate Regression to explore the relationship among the variables and
to address the impact of functional status on care support group. Functional status scores were
measured on response to mobility in bed, transfer, locomotion in home, dressing upper and lower
body, eating and toilet use. The respondents ADL scores were measured on a scale where 0 =
independent, 1 = set-up help only, 2 = supervision, 3 = limited assistance, 4 = extensive
assistance, 5 = maximal assistance,” and 6 = total dependence. I calculated a composite score
by summing scores and dividing by the number of ADL items. The dependent variable for the
question was care support group (Care Support Group). The participant’s Functional Score
(Functional Status) served as the dependent variable for H1F. There were no cases found with
missing data in the data set. Thus, for H1F I retained all of the 3,622 responses received (N =
3,622).
Tests of Normality and Univariate Outliers
Before analyzing Hypothesis 1F, I assessed basic parametric assumptions. That is, I
evaluated functional status, assumptions of normality, linearity, and homogeneity. The functional
status frequency histogram in Figure 4 provides visual evidence of normality. – I conducted a
test for univariate outliers and found none within the distributions.
Page 116
115
Figure 2. Functional status frequency histogram.
The normalized histogram indicates positive skewness (1.41) and some detectable
kurtosis (2.15). To test if this deviation from normality was significant, I calculated a z score
using the standard error of the skew (std. error skew = .041). Results indicated that participant
scores were not normally distributed; (skew = .353, z = 2.99, p < .001). Standard errors for both
skewness and kurtosis decrease with larger sample sizes, and the null hypothesis is likely to be
rejected even when there are only minor deviations from normality (Tabchnick and Fidell, 2007).
Although the histogram demonstrates skewness, this is an expected finding. The population is
being served in the community, where one would expect more people to be independent (lower
scores) than to be dependent with ADLs (with higher scores).
Page 117
116
Test of Homogeneity
To examine the assumption of homogeneity of variance, we conducted Levene’s test.
Homogeneity of variance is evaluated to determine if distributions are equal across the two levels
of the independent variable. Results from Levene’s test revealed that the distributions were equal
across groups, F = 6.529, p < .001. These results suggest that variances across groups are not
equally distributed. I employed Kruskal–Wallis to compare the scores on the continuous variable
for functional status. Examination of the descriptive statistics, Normalized Frequency Histogram,
and Levene’s test results conditionally confirm that the distributions meet parametric
assumptions.
ANOVA Analysis of Hypothesis 1F
I found a significant difference in functional status scores between caregiver types within
Care Support Group Category A. I conducted a one-way between-groups analysis of variance to
explore the relationship between caregiver type within Care Support Group A and functional
status, as measured by the score of functional status generated by the assessment tool. There was
a statistically significant difference in functional status among four caregiver types as shown in
Table 28.
The null hypothesis for Hypothesis 1F was rejected. These results indicate that functional
status does influence caregiver group as predicted by H1F, and there is a significant difference
between caregiver groups; that is, one group is more likely than another to have worse or better
functional status.
Page 118
117
Table 28. Care Group Category A: ANOVA Analysis of Functional Status
Sum of
Squares
Df Mean
Square
F P
Between Groups 97.0 3 32.3 23.04 .000
Within Groups 5,072 3,618 1.4
Total 5,169 3,621
T-Test Analysis of Hypothesis 1F
Within Care Support Group Category B, I found a significant difference in Functional
Status scores among caregiver types. The data violated the assumption of equal variance;
Levene’s test reported a significance of .009, identifying that the samples for InfCS and SForS
are not the same. As shown in Table 29, there was a significant difference between those in the
InfCS (M = 1.28, SD = 1.17) and SForS (M = 1.49, SD = 1.24); t (2596) = – 4.93, p < .001 (two
tailed). The magnitude in the differences in the means (mean difference = – .21, 95%,
Confidence Interval(CI): – 2.96 to – 1.28) suggests more individuals in the SForS group than the
InfCS group will have greater decline in functional status.
Table 29. Care Group Category B: T-Test Analysis for Functional Status
Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. T Df Sig. (2-
tailed)
Mean
Difference
Std. Error
Difference
95% Confidence
Interval of the
Difference
Lower Upper
Functional
Status
Equal
variances
assumed
6.77 .009 –4.99 3,351 .000 –.212 .042 –.295 –.129
Equal
variances not
assumed
–4.92 2,596 .000 –.212 .043 –.296 –.128
Page 119
118
Multivariate Analysis of Hypothesis 1F
After employing bivariate analysis, I used multivariate analysis to confirm, after
controlling for the influences of care support groups, the impact of care support groups on the
likelihood that functional status is associated with care group category. There was a statistically
significant difference between Care Group Category A and functional status, F (41,1), p < .001.
Additionally, there was a statistically significant difference between Care Group Category B and
functional status, F(58, 1), p < .001.
The null hypothesis for Hypothesis 1F was rejected. These results indicate that there is a
significant difference in functional status between both care group categories. These results
confirm functional status does influence care support group and caregiver type.
Hypothesis 1G (H1G)
People who meet nursing facility eligibility criteria would be more likely than those who do not
meet eligibility criteria to be in the InfCS group (and thus cared for by informal caregivers) than
those in the other care support groups.
I used chi-square and logistic regression to test Hypothesis 1G for any differences in the
proportion of people meeting nursing facility eligibility across care support groups, and then to
explore predictors of those relationships. Nursing facility eligibility was determined by applying
criteria identified in 130 CMR 456.409 of the Commonwealth of Massachusetts State Regulatory
Document for Medicaid Nursing Facilities. Participants were found to be eligible or not eligible.
Page 120
119
Findings details can be found in tables 30 and 31. Because there were no missing cases in the
data set, for H1G I retained all of the 3,622 responses received (N = 3,622).
Bivariate Chi-Square Analysis of Hypothesis 1G
A significant difference in nursing facility eligibility was apparent across groups within
Care Support Group A; χ2 (1, N = 3,622) p < .001***. Additionally, I found a significant
difference in nursing facility eligibility across groups within Care Group Category B; χ2 (1, n =
3,450) p < .001***. Table 17 provides the results of the chi-square analysis. As indicated by
Table 17, and the chi-square p value (< .001***), the null hypothesis for H1G was rejected. That
is, these results indicate that nursing facility eligibility does influence care support groups.
Multivariate Analysis of Hypothesis 1G
After employing bivariate analysis, I performed multivariate logistic regression across
Care Group Categories A and B. The multivariate model reporting differences in the number of
individuals who are eligible for nursing facility services across Care Group Category A was
statistically significant, 2
(1, N = 3,622) = 22.737, p = .001. Similar results were seen between
the number of individuals eligible for nursing facility services and Care Group Category B, 2
(1,
n = 3,450) =,23.364, p = .001. These results indicate that there are differences in the number of
individuals meeting nursing facility eligibility criteria across different care group categories at
the multivariate level. Thus, the null hypothesis was rejected.
As shown in Table 30, two of the independent variables of caregiver types made a unique
statistically significant contribution to nursing facility eligibility across Care Support Group A
(InfCS, p < .001***, and MixSS, p < .001***) when evaluated by multivariate logistic
Page 121
120
regression. The strongest predictor of nursing facility eligibility across Care Support Group A
was InfCS, recording an odds ratio of 1.547. With a positive beta value, this indicates that those
in the Informal Care Support caregiver group were almost 55% more likely not to meet
eligibility for nursing facility services than those in the other caregiver groups. Those in the
MixSS and ForS caregiver groups reported the lowest odds ratio, .662 and .673, with a negative
beta, being less likely to meet nursing facility eligibility.
As shown in Table 31, both of the independent variables of caregiver types made a
unique statistically significant contribution to nursing facility eligibility across Care Support
Group B (InfCS, p < .001***; and, SForS, p < .001***) when evaluated by multivariate logistic
regression. The strongest predictor of nursing facility eligibility across Care Support Group B
was InfCS, recording an odds ratio of 1.547. This indicates that those in the Informal Care
Support caregiver group were almost 55% less likely to meet eligibility for nursing facility
services than those in the SForS caregiver group.
The null hypothesis was rejected, supporting with advanced examination that nursing
facility eligibility is associated care support group and caregiver type.
Page 122
121
Table 30. Multivariate Statistical Analysis for Nursing Facility Eligibility1 and Care Support Group A
(N = 3,622)
Predictor Beta SE Wald Df P2
Odds Ratio
NoSS3 .029 .230 .016 1 .90 1.0
InfCS3
.436 .110 15.76 1 .001*** 1.547
MixS3
–.413 .113 13.39 1 .001*** .662
ForS3
–.397 .242 2.693 1 .101 .673
Constant
–.487 .101 23.074 1 .001
Test Test value (Df) P
2
Overall model
evaluation:
–2 Log
Likelihood
41.024 22.737(1) .001***
Goodness of
Fit: Pearson 2.643(2) .278
Footnote: 1. Nursing facility eligibility is defined as eligible (coded as 1) or not eligible (coded as 0). 2. Significance
p value: * p < .05. ** p < .01. *** p < .001. 3. NoSS= No Caregiver Support or Services, InfCS = Informal Care
Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal care supports and
services), ForS= Formal Services Only.
Table 31. Multivariate Statistical Analysis for Nursing Facility Eligibility1 and Care Support Group B
(n = 3,450)
Predictor Beta SE Wald Df P3
Odds Ratio
InfCS2
.436 .110 15.76 1 .001*** 1.547
SForS2
.468 .111 17.841 1 .001*** .626
Constant
–.503 .103 23.684 1 .001***
Test Test value 2(Df) P
3
Overall model
evaluation:
–2 Log
Likelihood
28.957 23.364(1) .001***
Goodness of
Fit: Pearson 2.042(2) .153
Footnote: 1. Nursing Facility Eligibility is defined as eligible (coded as 1) or not eligible (coded as 0). 2. InfCS =
Informal Care Support only, SForS = Some Formal Services. 3. Significance p value: * p < .05. ** p < .01. *** p
< .001.
Aim C
Page 123
122
Hypothesis 2A (H2A): Those in the InfCS group would be more likely to meet eligibility for
Massachusetts nursing facility admission as those in the other care support groups.
Missing Data and Univariate Outliers
No univariate outliers were found to exist within the distributions. For logistic regression,
the dependent variable must be dichotomous in nature; therefore, individuals who answered
‘none’ for Care Support Group were removed from analysis. The dependent variable became
InfCS group and SForS, which combined the Mixed and Formal care support groups. For H2A,
3,186 responses from participants were entered into the logistic regression model.
Before the H2A was analyzed, basic parametric assumptions were assessed. For the
criterion variable “Care Support Group,” assumptions of multicollinearity were evaluated.
Logistic regression does not make assumptions concerning the distribution of scores for the
predictor variables (Pallant, 2007).
Multicollinearity
I tested the assumptions of multicollinearity by calculating correlations between variables
and collinearity statistics (Tolerance and Variance Inflation Factor). Correlations between
criterion and predictor variables were low. Several of the correlations between predictor
variables did exceed the recommended critical limit of 0.7, including nursing facility eligibility,
gender, ethnicity, language, marital status, functional status, depression risk, trade-offs, and
diagnosis of Alzheimer’s or Dementia. Given the evidence provided, normality of the criterion
and predictor variables was conditionally affirmed. After examining the multicollinearity
Page 124
123
diagnostics and other descriptive statistics, the distributions are assumed to meet parametric
assumptions.
Correlation Analysis
I used a Pearson Correlation matrix to explore whether there were any significant
relationships across predictor variables. Preliminary analyses were performed to ensure no
violation of the assumptions of normality, linearity, and homoscedasticity. There was strength in
the following relationships between predictor variables of short-term and procedural memory,
expression, and comprehension (Pearson Correlations over .4). Additionally, there was a strong
relationship between IADL and ADL functional status (Pearson Correlation = .389). All other
predictor variables demonstrated weak relationships between the remaining variables. Because of
the relationship strengths, comprehension, expression, and procedural memory variables were
removed from the analysis. I kept short-term memory in the analysis because clinically short-
term memory issues lead to more problematic issues in the community. For example, a person
with short-term memory issues could easily get lost outside and not remember where they were,
they might forget to take important medications or complete treatments necessary for good
health, or they may be unable to complete tasks necessary to remain in the community.
Using logistic regression, I analyzed Hypothesis 2A, employing Massachusetts nursing
facility eligibility as the dependent variable. The independent variable was care support group
category, and the predictor or control variables included gender; level of education;
race/ethnicity; primary language; marital status; living arrangement; Alzheimer’s disease
diagnosis; dementia diagnosis; memory issues, including short-term memory; belief of poor
health and belief of increased independence; and medication compliance. The continuous
Page 125
124
predictor variables were ADL functional status, IADL functional status, incontinent status,
cognition/decision making, number of medications taken, behavioral symptoms, and depression
risk.
Many predictor variables made a statistically significant impact on nursing facility
eligibility including ethnicity, level of education, primary language, short-term memory, ADL
functional level, and instrumental activities of daily living. Tables 32 and 33 display the
multivariate analyses of Care Support Group Categories A and B.
As shown in Table 32, across Care Group Category A seven of the predictor variables
had a unique statistically significant contribution, including ethnicity (p = .046), level of
education (p < .001), primary language (p < .001), short-term memory (p = .004), ADL
functional status (p < .001), and instrumental activities of daily living (p = .067).The following
variables were significant predictors of people being eligible for nursing facility services: those
with a high school diploma or less and those more dependent with ADL and IADL services.
People who identified themselves with an ethnic or racial background other than Caucasian, a
primary language other than English, or with short-term memory problems were less likely to be
found eligible for nursing facility services.
As shown in Table 33, across Care Group Category B, seven of the predictor variables
had a unique statistically significant contribution, including ethnicity (p = .045), level of
education (p < .001), primary language (p < .001), short-term memory (p = .004), ADL
functional status (p < .001) and instrumental activities of daily living (p = .067). People with a
high school diploma or less and those more dependent with ADL and IADL services were
significant predictors of eligibility for nursing facility services. People who identified themselves
Page 126
125
with an ethnic or racial background other than Caucasian, a primary language other than English,
or with short-term memory problems were less likely to be found eligible for nursing facility
services.
Results
I performed logistic regression to determine the impact of informal caregiver support on
the likelihood that a person met or did not meet nursing facility eligibility, while controlling for
the many potential impacting variables. Initially evaluated for Care Group Category A, the full
model containing all predictors was statistically significant: 2
(25, n = 3,386) = 154.526, p <
.001***. The model as a whole explained between 4.5% (Cox and Snell R square) and 8.91%
(Nagelkerke R squared) of the variance in eligibility status, and correctly classified 89.1% of
cases.
Caregiver type demonstrated a statistically significant impact on nursing facility
eligibility, specifically with InfCS group. This demonstrates that those with informal caregivers
are less likely to meet nursing facility eligibility, with a p = .053. This demonstrates that the
hypothesis is rejected.
Page 127
126
Table 32. Impact of Care Support Group A Caregiver type on Nursing Facility Eligibility (n = 3,386)
Beta S.E. Wald Df p
Odds 95%
Confidence
Interval
Ratio Lower
Bound
Upper
Bound
Pred
ictor V
ariables
Gender .057 .133 .187 1 .665 17.4 –4.93 3.14
Ethnicity .358 .179 3.995 1 .046* 2.79 142.9 1.41
Level of Education –.590 .182 10.511 1 .001*** –1.69 –1.06 –4.29
Primary Language .565 .167 11.401 1 .001*** 1.77 4.22 1.2
Marital Status –.141 .160 .778 1 .378 -6.89 –2.19 5.78
Living Arrangement –.163 .129 1.596 1 .207 –6.15 –2.41 11.11
Activities of Daily
Living
Functional Status
–.333 .047 50.441 1 .001*** –3.003 –2.35 –4.16
Incontinence –.047 .076 .380 1 .538 –21.27 5.1 9.8
Belief in ability for
Increased
Independence
–.111 .148 .562 1 .453 -9.05 –2.5 5.6
Belief of Poor Health –.013 .130 .010 1 .920 76.9 –3.72 4.12
Behavior Problems .031 .045 .468 1 .494 32.79 –17.2 8.4
Depression Risk –.002 .017 .009 1 .925 –666.66 –28.5 31.25
Instrumental Activities
of
Daily Living
–.081 .045 3.274 1 .070* –12.35 –5.92 142.9
Memory: short-term .454 .158 8.292 1 .004** 2.2 –5.92 1.31
Trade-offs –.175 .337 .271 1 .603 –5.69 –2.02 2.06
Medication Adherence .133 .083 2.564 1 .109 7.55 1.19 3.4
Page 128
127
Alzheimer’s disease .221 .192 1.329 1 .249 4.5 –33.3 1.67
Dementia –.053 .142 .138 1 .711 –19.05 –6.45 4.42
Cognitive Decision
Making .021 .071 .083 1 .773 48.78 –3.02 6.25
NoSS .352 .329 1.145 1 .285 2.84 –15.6 1.0
InfCS .492 .254 3.735 1 .053* 2.03 –3.4 1.01
MixSS .051 .258 .038 1 .845 19.8 –2.2 1.8
ForS 0a 0.0 0.0 0 0.0
Constan
t
Nursing Facility
Eligibility: Eligible –2.039 .753 7.327 1 .007 –2.45 –.284 –1.776
Notes: p value: * p < 0.05. ** p < 0.01. *** p < 0.001.
Dependent variables coding: NF eligibility, 0 = eligible, 1 = not eligible.
Independent variables coding: Each Caregiver Type was run by type and a dummy variable 0 = caregiver type
NoSS, InfCS, MixSS, or ForS, and 1 = dummy variable of other category.
Predictor variables coding: Gender, 0 = male, 1 = female; Education, 0 = high school or less, 1 = some college or
college degree; Ethnicity, 0 = other, 1 = Caucasian; Language, 0 = English, 1 = other; Marital Status, 0 = married,
1= other; Living Arrangement, 0 = lives alone, 1 = lives with others; Incontinence, 0–3, 0 = continent, 1=
incontinent; Behavior issues,0–6, 0 = no behavior problems, 6 = multiple behaviors and behaviors not easily altered;
Depression Risk, 0–3, 0 = no risk, 3 = high risk; ADL Functional Status, 0–6, 0 = independent, 6 = total
dependence; IADL Status, 0–3, 0= independent, 3 = by others; Short-Term Memory, 0 = memory ok, 1 = memory
problem; Alzheimer’s Diagnosis, 0 = n/a, 1 = disease present; Dementia Diagnosis, 0 = n/a, 1 = disease present;
Cognitive Decision Making, 0 = no issues, 1 = decision making problems; Belief of Poor Health, 0 = false, 1 = true;
Belief of Increased Independence Ability, 0 = false,1 = true; Trade-Offs, 0 = no,1 = yes; Medication Adherence, 0 =
compliant > 80% of time, 1 = compliant < 80% of the time.
After analyzing the findings for Care Group Category A, I also looked at hypothesis 2A
as it related to Care Group Category B. The full model containing all predictors was statistically
significant: 2
(24, n = 3,386) = 154.487, p < .001***). The model as a whole explained
between 4.5% (Cox and Snell R square) and 8.9% (Nagelkerke R squared) of the variance in
eligibility status, and correctly classified 89.1% of cases.
Caregiver type demonstrated a statistically significant impact on nursing facility
eligibility, specifically with InfCS group p < .001***. This demonstrates that those with informal
Page 129
128
caregivers are less likely to meet nursing facility eligibility. This demonstrates that the
hypothesis has been rejected.
Page 130
129
Table 33. Impact of Care Support Group B Caregiver Type on Nursing Facility Eligibility (n = 3,386)
Beta S. E. Wald Df p Odds 95% C.I.
Ratio Lower Upper
Pred
ictor V
ariables
Gender .058 .133 .191 1 .662 17.2 –4.95 3.14
Ethnicity .359 .179 4.010 1 .045* 2.8 125 1.41
Level of Education –.590 .182 10.504 1 .001*** –1.7 –1.06 –4.3
Primary Language .565 .167 11.405 1 .001*** 1.8 4.22 1.1
Marital Status –.142 .160 .789 1 .374 –7.02 –2.2 8.31
Living Arrangement –.163 .129 1.612 1 .204 –6.1 –2.4 11.24
Activities of Daily
Living
Functional Status
–.332 .047 50.476 1 .001*** –3.01 –2.36 –4.16
Incontinence –.047 .076 .382 1 .536 –21.3 –5.1 9.8
Belief in ability for
Increased
Independence
–.112 .148 .578 1 .447 –8.89 –2.5 5.65
Belief of Poor
Health –.012 .130 .009 1 .924 –80 –3.7 2.04
Behavior Problems .030 .045 .455 1 .500 32.8 –17.24 8.4
Depression Risk –.002 .017 .009 1 .926 –666.6 –28.6 30.3
Instrumental
Activities of
Daily Living
–.082 .045 3.359 1 .067* –12.27 –5.92 166.1
Memory: short-
Term .454 .158 8.285 1 .004** 2.2 6.89 1.13
Trade-offs –.173 .337 .264 1 .608 –5.78 1.2 2.05
Page 131
130
Medication
adherence .133 .083 2.578 1 .108 7.49 –34.5 5.64
Alzheimer’s disease .223 .192 1.359 1 .244 4.47 –6.6 1.67
Dementia –.052 .142 .134 1 .715 –19.2 –3.03 4.42
Cognition decision
making .021 .071 .085 1 .771 48.78 –8.4 6.25
InfCS .447 .118 14.308 1 .001*** 2.23 4.63 1.47
SForS 0a . . 0 .
Constan
t
Nursing Facility
Eligibility:
Eligible
–2.085 .715 8.496 1 .004 –.479 –.287 –1.46
Notes: p value: * p < 0.05. ** p < 0.01. *** p < 0.001.
Dependent variables coding: NF eligibility, 0 = eligible, 1 = not eligible.
Independent variables coding: Each caregiver type was run by type and a dummy variable 0 = caregiver type NoSS,
InfCS, MixSS, or ForS, and 1 = dummy variable of other category.
Predictor variables coding: Gender, 0 = male, 1 = female; Education, 0 = high school or less, 1 = some college or
college degree; Ethnicity, 0 = other, 1 = Caucasian; Language, 0 = English, 1 = other; Marital Status, 0 = married,
1= other; Living Arrangement, 0 = lives alone, 1 = lives with others; Incontinence, 0–3, 0 = continent, 1=
incontinent; Behavior issues, 0–6, 0 = no behavior problems, 6 = multiple behaviors and behaviors not easily
altered; Depression Risk, 0–3, 0 = no risk, 3 = high risk; ADL Functional Status,0–6, 0 = independent, 6 = total
dependence; IADL Status, 0 – 3, 0 = independent, 3 = by others; Short-Term Memory, 0 = memory ok, 1 = memory
problem; Alzheimer’s Diagnosis, 0 = n/a, 1 = disease present; Dementia Diagnosis, 0 = n/a, 1 = disease present;
Cognitive Decision Making, 0 = no issues, 1 = decision making problems; Belief of Poor Health, 1 = false, 2 = true;
Belief of Increased Independence Ability, 0 = false,1 = true; Trade-Offs, 0 = no,1 = yes; Medication Adherence, 0 =
compliant > 80% of time, 1 = compliant < 80% of the time.
Aim D
The fourth and final aim, Aim D, was to examine if nursing facility eligibility criteria in
states that border Massachusetts would significantly change the proportion of people eligible for
nursing facility eligibility in the InfCS group. To do this, I used the eligibility criteria employed
by the states that border Massachusetts, including Vermont, New Hampshire, Rhode Island, and
Connecticut. The criteria in New York cannot be measured, so I did not evaluate them.
Page 132
131
Hypothesis 3A (H3A): Those in the InfCS group would be more likely to meet eligibility for
nursing facility services when applying criteria of neighboring states.
I employed frequency statistics to determine if differences existed between participant
care support groups and nursing facility eligibility using eligibility criteria from states
neighboring Massachusetts. Each nursing facility eligibility criterion, based on regulation and
laws, for Massachusetts, New Hampshire, Connecticut, Rhode Island, and Vermont, were
evaluated between two Care Group Categories and the individual caregiver groups within them:
Care Group Category A (InfCS, NoSS, ForS, and MixSS) and Care Group Category B (InfCS
and SForS). The criterion from each of the states was applied to the population sample to
determine the percentage of people who would meet or not meet eligibility if the alternate
criteria were utilized. Findings are detailed in Table 34.
Results
The neighboring states have differing criteria, and some of the eligibility criteria
demonstrate very different frequencies of individuals in caregiver types. Care Group Category A
and the MixS caregiver group had the widest variation in frequency of eligibility for nursing
facility services. The MixS group meets nursing facility eligibility criteria at a higher frequency
utilizing Connecticut standards. Approximately seventeen percent of this group met the
Connecticut, eligibility criteria, whereas only 14.8% meet the Massachusetts standards. The ForS
caregiver group had the smallest percentage eligible when utilizing criteria from Vermont
guidelines. For that group, only 0.1% could meet Vermont requirements, although 14.9% met
the criteria for Massachusetts. The variations in frequency demonstrate differences in eligibility
standards and strictness of guidelines. In states with more stringent eligibility criteria, for
Page 133
132
example Vermont, there were 5% more individuals eligible for nursing facility services using
Massachusetts criteria. In Connecticut, which appears less stringent, about 3.5% more
individuals were eligible for nursing facility admission in Connecticut than in Massachusetts.
Page 134
133
Table 34. Neighboring States Nursing Facility Eligibility (N = 3,609–3,622)1
Care Group Category A Care Group Category B
Variable NoSS
2 InfCS
2 MixS
2 ForS
2 Total InfCS
2 SForS
2 Total
% (n) % (n) % (n) % (n) N % (n) % (n) N
Massachusetts
Eligible 9.3%
(25)
9.2%
(191)
14.8%
(167)
14.9%
(23) 406
9.2%
(191)
17.4.%
(190) 381
Not
Eligible
90.7%
(244)
90.8%
(1,881)
85.2%
(960)
85.1%
(131) 3,216
90.8%
(1,881)
82.6%
(1,091) 2972
Rhode Island
Eligible 9.3%
(25) 9.2%
14.8%
(167)
14.9%
(23) 406
9.2%
(191)
17.4%
(190) 381
Not
Eligible
90.7%
(244)
90.8%
(191)
85.2%
(960)
85.1%
(131) 3,216
90.8%
(1,881)
82.6%
(1,091) 2,972
Connecticut
Eligible 3.8%
(10)
4.9%
(98)
17.3%
(166)
12.4%
(17) 291
4.9%
(98)
16.7%
(183) 381
Not
Eligible
96.2%
(259)
95.1%
(1974)
82.7%
(959)
87.6%
(137) 3,329
95.1%
(1,974)
83.3%
(1,096) 2,972
Vermont
Eligible 2.3%
(6)
4.4%
(87)
12.6%
(126)
0.1%
(14) 359
4.4%
(87)
12.3%
(140) 381
Not
Eligible
97.7%
(263)
95.6%
(1,985)
87.4%
(999)
99.9%
(140) 3,387
95.6%
(1,985)
87.7%
(1,139) 2,972
New
Hampshire
Eligible 3%
(8)
5.1%
(106)
8.5%
(96)
11%
(17) 227
3%
(8)
9.7%
(113) 381
Not
Eligible
97%
(261)
94.9%
(1,966)
91.5%
(1,031)
89%
(137) 3,395
97%
(261)
90.3%
(1,168) 2,972
Footnotes: 1. Total population N = 3,622, some items have missing cases due to the mechanism of data collection
and data entry error. 2. Care Group Category A: NoSS = No Caregiver Support or Services, InfCS = Informal Care
Support only, MixSS = Mixed Care Support (i.e.: those with a combination of informal and formal care supports and
services), ForS = Formal Services Only; Care Group Category B: InfCS = Informal Care Support Only, SForS =
receipt of at least some formal services (includes those with formal only and mixed services).
Page 135
134
Conclusion
Three hypotheses received full or mixed support: hypotheses 1E, 1F, and 1G. As
predicted by Hypothesis 1F, people with more deficits in activities of daily living were as likely
to be in the InfCS group as other care support groups. As predicted by hypothesis 1G, people
who met nursing facility eligibility criteria were as likely to be in the InfCS group as other care
support groups. Hypothesis 1E, related to living arrangement, was partially supported at the
multivariate level. When living arrangement (H1E) was analyzed at the multivariate level by
individual caregiver types, those people in the MixSS group were more likely to live alone.
Hypothesis 1E had predicted that people who lived with others would be more likely to be in the
InfCS than other care support groups.
Aim A described the sample characteristics and population as a whole and across the
caregiver groups. The population samples will be further discussed and compared in the
discussion chapter to the elders and Medicaid populations in Massachusetts. Aim B looked at the
factors that predicted membership in caregiver groups. Aim B demonstrated that Hypothesis 1A
related to ethnicity, 1B related to level of education, 1C related to language, and 1D related to
marital status, were all rejected and could not statistically support at the bivariate or multivariate
level a difference in the hypothesized variable across the care support groups. Aim C determined
if there was a relationship between caregiver type and eligibility for nursing facility services,
based on Massachusetts Medicaid criteria. Hypothesis 2A was rejected, demonstrating that those
with an informal caregiver are not as likely to meet nursing facility eligibility as those in the
other care groups. The final aim, Aim D, examined nursing facility eligibility criteria of
surrounding states and applied those criteria to this sample. This aim found that indeed
Page 136
135
differences do exist between the states’ criteria, and there is a need to review criteria and
residency requirements. I will discuss these findings and their implications on public policy in
Chapter 6.
Page 137
136
Chapter 6: Discussion and Policy Implications
Decision makers at every level are attempting to reduce health care costs in America
while improving the quality of life for seniors and others who require long-term care. Essential to
this effort is the ability to keep this population out of institutions for as long as possible, using
the logic that institutional care is more costly than home care and that remaining at home is
emotionally superior to confinement in even the best institution. My research elucidated key
factors that enable seniors in Massachusetts to remain in their homes as their need for care
increases. It also revealed limitations in the current support systems and suggested modifications
that can be made at both the health policy and implementation levels, to improve services to this
growing population of Americans. In this chapter, I expound upon the implications of these
findings and suggest steps that can be taken to remedy some of the most pressing problems.
Informal caregivers are an integral component of any effort to keep seniors out of
institutions, so I also examined the various roles they play and burdens they face in caring for
aging relatives and friends. I started from the premise that informal caregivers and informal care
support services would be substantial in comparison to formal care services for those living in
the community and would present a cost savings. Many elders who need support are being
provided for by friends and family. While this potentially saves countless state and federal
dollars, the expenses are borne by the caregivers, who are also bearing the physical and
emotional burdens of care. As discussed in Chapter 2, many studies have attempted to assign
monetary value to the services their caregivers provide. Examination of these informal caregivers
for valuation of services is one step toward better understanding of these services. I offer several
recommendations that are aimed at easing the burden on informal caregivers, thus helping them
Page 138
137
to keep elders out of institutional care. Some of these would require national legislation while
others pertain to changes at the state level, either in policy or implementation. I identify where I
believe actions should take place, where preliminary work would be necessary, and in some
cases I recommend funding strategies.
Key Findings
Many of the hypotheses focused on the differences between those being cared for by
informal caregiver supports versus those with formal service providers, or a combination of the
two. I used this focus to demonstrate the importance of the informal caregiver role as well as to
demonstrate the differences in the populations and to increase awareness and understanding of
how each population differs.
This study found that people with increased functional deficits, meaning more difficulty
completing tasks of daily living such as bathing, grooming, dressing, eating or ambulating, are
more likely to be cared for by a combined provider type, defined in Care Group Category A as
MixS and in Care Group Category B as SForS. Informal caregiver supports and formal care
services were found to be almost identical. This is an important finding because it demonstrates
that the informal caregiver is completing similar care tasks, at similar levels as those provided by
paid formal service providers. For many seniors, and certainly for everyone in this sample, paid
formal service providers are reimbursed via state and federal dollars, because these individuals
are Medicaid recipients. I found a significant relationship between an individual’s functional
status and caregiver type. The more functionally needy a person was, the more likely he or she
was to be found in the mixed care support and services group. This finding suggests that the
combination of formal services and informal supports assists those with more functional deficit,
Page 139
138
or need for assistance with activities of daily living, in the community. The combination may be
keeping elders with functional deficits in the community longer, further study of this mixed care
group population would be necessary to determine the possible reasons. These findings also
suggest that the combination of formal and informal services offers the respite and assistance that
the caregiver needs to be able to continue to provide the services effectively and for a longer
time. Further, respite and collaborative efforts on behalf of people in the community is likely the
best way to manage elders. I would recommend further study to address these questions.
Those who meet nursing facility eligibility criteria are more likely to be cared for in the
community by informal caregiver supports when the care giver categories were delineated as
mixed services, no services and formal care services. When the caregiver groups were
combined within Care Group Category B, interestingly the significance shifted to be equal for
both informal and some formal to be eligible for nursing facility eligibility. Nursing facility
eligibility in Massachusetts was used as the criterion for this hypothesis within the study, which
looks not only at ability to complete daily living tasks but also at nursing tasks and skills. This
becomes an interesting finding in light of how persons become eligible for services. Eligibility is
determined based on need for assistance with daily living, as well as the need for skilled nursing
services. In light of how the statistical significance changed, I assume that those in the InfCS
group are meeting nursing facility eligibility based on ADL need. This also suggests that those
in the MixS group have ADL need but not skilled needs. The addition of the formal services
care category to the analysis with care group category B added skilled nursing needs based
eligibility only to the analysis. This may suggest that supporting the informal caregiver with care
services of elders in the community is needed.
Page 140
139
Recommendations
Problem Identification
For health policy to move forward many factors must be in place. The critical first step is
to acknowledge that there is a problem to solve, and then to obtain general agreement among
political leaders that it is both important and feasible to find a solution. Strategic planning is
essential, to minimize obstacles and to time proposals to coincide with the agendas of key
decision makers. Policy makers and politicians often carefully plan when to introduce ideas and
concepts to match the state of the economy, the societal undertones, and the needs and demands
of constituents. ( Lipsky, M, 2010)
Key Actor/Player Determination and Support
Issues that affect elders often resonate strongly with voters. Everyone can relate to elders:
they may think of their own parents, grandparents, or other beloved older people or they may
simply understand that they, too, will one day be old and they do not want to imagine suffering
in any way. The media uses this as a mechanism to increase the saliency of topics during
elections, and also to encourage society to become emotionally tied to a topic and thus actively
involved. Despite the activities and influence of numerous elder networks at all levels, they do
not seem to reach many who are in need of assistance and advocacy. Elders and their caregivers
may need to be more proactive in identifying unmet needs. The state level decision makers also
must agree that eldercare is an issue, and they must identify resources to meet the needs. For
elders in Massachusetts, there are many resources that may meet their needs, specifically needs
relating to care and care giving. Most of these options are administered by the Executive Office
Page 141
140
of Elder Affairs and MassHealth, the Office of Long-Term Care. Other state agencies are
focused on specialty needs, such as those with brain injuries, developmental disabilities or
mental retardation, or they are agencies that deal with assistance with housing or food. Once the
Executive Office of Elder Affairs and MassHealth has determined an issue is important,
developing the policy and the next major hurdle to overcome is how to pay for the policy,
service, or concept.
Policy Suggestions
Now is the time to make changes in support of improved services to elders and further
deinstitutionalization of seniors who could continue to live and thrive in the community with
adequate support. State-level politicians and decision makers are already on board with the idea
that change must occur. Funding options have been identified. The policy suggestions must be
delineated.
Growth during a difficult economic time calls for creative strategies. A theme emerged in
the findings, not unexpected although not as salient as expected when the study began. Informal
caregivers play an important role that supports elders in many ways. The results of this study
found it statistically significantly that informal caregivers were more likely to support those with
functional deficits, those living with others, and those who met nursing facility eligibility over
those in other care support groups.
Looking Toward Alternative Programs
Many countries have already begun to implement in part or whole universal healthcare,
whether a socialist society or democratic one. With the onset of Obamacare, there will need to
Page 142
141
be creative mechanisms and implementation strategies for long-term care support. Looking
toward other countries and how they have handled long-term care and informal care supports
may give insight into the future in the United States.
Many European countries with socialist medical coverage began in the 1990’s to cover
informal care support services through cash payment to healthcare recipients, who can use the
cash to in turn pay their caregivers. This mechanism has occurred in Austria and Germany, since
1994 and 1996 respectively. (Saltman, Dubois & Chawla, 2006)
In the United States, through Medicaid a demonstration project dubbed “Cash and
Counseling” was implemented in 1998 as a mechanism for persons to purchase assistance with
routine daily living activities from sources other than formal home care service agencies with a
monthly allowance. The goal of “Cash and counseling” was increased control and satisfaction.
The demonstration project showed positive outcomes by caregivers and recipients alike. In 2009
the “Cash and counseling” state demonstration grants ended, yet 15 states continue these
programs- including Massachusetts. In Massachusetts this program is only utilized with the
Personal Care Attendant program, for persons with disabilities that can self direct their care.
According to the National Resource Center for Participant-Directed Services, cash and
counseling programs demonstrate positive outcomes, a decrease in unmet needs of participants,
and have found that they did not result in the misuse of Medicaid funds, but personal care costs
are higher. In light of these findings, prudent controls will need to be developed if similar wider
scale programs are to be developed. It may be an option to be considered on a wider scale to
elders and their caregivers. (http://www.cashandcounseling.org)
Encouraging Growth in Current Programming
Page 143
142
For those who live in a home, alone or with others, a support service can often be a
balancing mechanism that helps maintain elders successfully in the community, even when their
needs may make them eligible for nursing facility services. This study demonstrated that those
people being cared for solely by informal caregivers were as likely to meet nursing facility
clinical eligibility as those who had both informal caregivers and formal services. Further study
is needed to determine at what point formal services replace informal caregiver services, or at
what point an informal caregiver would be likely to turn to institutionalization. I can only assume
that combined services (informal care support and formal services), may assist in maintaining
elders in the community longer, by offering respite and other support to the caregiver. One such
service includes day programs. Senior centers and adult day health (ADH) programs are
excellent sources of support for community-dwelling elders.
Senior centers are purely social, but they often partner with local food pantries, meals-on-
wheels, or other supportive lunch programs, thus providing nutritional support as well. Many
centers have added other components in an effort to be of more assistance to their clients. ADH
programs began as mechanisms to help those with medical and clinical needs to remain in the
community. These programs were created to offer six hours of daily programming including a
snack and a hot lunch, nursing oversight and care coordination, and social programming care.
These programs are available throughout Massachusetts and accept MassHealth and private
payment for services.
Home- and Community-based Services (HCBS) Waivers, as described in earlier chapters,
offer a mechanism to expand programs beyond the original federal rules that restricted Medicaid
Page 144
143
to medically oriented services only. The expansion of ADH programs into more medical models
with additional services and oversight was an opportunity to help elders remain in the
community who may otherwise need to be institutionalized. This expansion came through the
Program for All Inclusive Care of the Elderly, or PACE.
PACE is a program that has been around in various forms for the past forty years, despite
being recently touted in the press (CBS News feed, July 2012) as a “new elderly care program”
that may contain healthcare costs. PACE attendees, similar to those in adult day health, attend a
day center where all medical and supportive services are provided. There is also a physician on-
site, and all case management and care coordination is provided. The PACE care model can be
traced to On Lok Senior Health services in San Francisco, California, which started in the early
1970s. The model began as an Adult Day Health, offering day services and care. A major
breakthrough included reimbursement through Medicaid. The program continued to expand, and
in 1978 and 1979 it began to provide services to those who were eligible for nursing facilities..
By the mid-eighties, the program was formally renamed PACE through federal legislation, and
became part of HCBWS operation. Through the 1990s PACE programs became operational in 24
states, and in 1997 the Balanced Budget Act established PACE as a provider type through
Medicare and Medicaid. In 2011, there were 82 operational PACE programs in 29 states; in
Massachusetts there are six PACE programs (www.npaonline.org, 2012).
Although PACE provides valuable support to many elders who meet clinical eligibility
for nursing facility services in the community, it does have limitations. A key drawback is
geography. In Massachusetts, for example, five of the six PACE programs are located around the
Boston urban area and one is located in Worcester, Massachusetts, only 45 miles west of Boston.
Page 145
144
More than two million Massachusetts residents live outside of the greater Boston and Worcester
areas, including thousands in such population centers as Springfield, Lowell, and Brockton, and
they receive no benefit from PACE (http://quickfacts.census.gov/qfd/states/25000.html, 2012).
I suggest supporting the development of additional PACE programs in areas outside of
the greater Boston area, to serve elders throughout the state. Increasing geographic diversity of
PACE programs is needed. Having the support and buy-in of the current PACE providers will
help streamline the addition of new sites. This process can start by scheduling a meeting with
each of the six PACE leadership persons about the strengths and difficulties they have faced.
State PACE program managers should partner with the current PACE providers, so that they can
act in a collaborative role as mentors to the new applicants and potential providers. The goal is
greater geographic diversity and there are significant population centers in Massachusetts that are
not be aided by a PACE site. A needs analysis may be helpful to determine which areas are in
highest demand and need. It will also be essential to develop a clearly defined Request for
Proposals (RFP) for interested ADH providers. Quality ADH programs that are not within the
geographic territories of the current PACE sites, that meet the RFP guidelines, must be
encouraged to respond to these solicitations. Incentives or partnerships for ADH providers to
move into a PACE model may be helpful to grow this program into other areas. PACE providers
reimbursements are higher than traditional ADH programs, which may promote incentive. There
may also be a mechanism to offer a one time assistance with capital improvements to offset the
costs of changing provider type, with a clause that the provider must continue to serve a
minimum number of Medicaid enrollees for a set length of time.
Page 146
145
Although some funding streams have been potentially identified, another limitation is the
financial ability of the Commonwealth to offer this service. Medicaid services are “a-la-carte.”
An elder who meets nursing facility eligibility may need PACE as well as therapy, personal care
services in the evening, nursing services on weekends or evenings, or assistance with medication
delivery. Depending on the level of care needs, and the support systems needed to maintain the
elder’s level of wellness in the community, it may be necessary to evaluate the feasibility and
cost effectiveness of remaining in the community versus moving to a nursing facility. In any
such decision making process, the desire of the individual must be considered. Although it may
be difficult to discuss, at some point it is necessary to determine the breaking point, for the
family, individual, and potentially the upper limits of support available through Medicaid.
Supporting the Informal Caregiver
I feel that there is more that should be considered. Currently, the Family Medical Leave
Act (FMLA) allows working caregivers rights, including flexibility and security around their
employment. There are inconsistencies in how FMLA is applied, and many private and public
employers make it difficult for people to utilize this benefit. Many informal caregivers are not
eligible for this, because they work part-time or have given up their jobs to take over the duty of
caregiver. One way to benefit informal caregivers is to improve their access to information about
FMLA and other legal recourses that may be available to them as they struggle to balance the
needs of the person they care for with their own needs. Some suggestions may include a
requirement to provide new employees with clear information about FMLA provisions when
they are hired, as well as at, say, an annual benefits fair, on the internet, at senior centers, at
hospitals, in doctors’ offices, and other appropriate sites. These sites can also provide
Page 147
146
information on other support and awareness topics. Caregiver advocacy would also be beneficial
and caregivers must have access to information about this aspect of support as well.
There are other ways to support caregivers, such as providing some sort of compensation
for their time and the expertise they have gained, or reimbursement for their out-of-pocket
expenses. They would also benefit from stronger assurances that they will not be penalized or
hindered in their employment because of lost days. Another intangible way to support caregivers
would be to institute some sort of official recognition or acknowledgement that this sort of
volunteer care giving counts as legitimate work experience. Aspects of these are in place, with
Family Medical Leave Act (FMLA), but it may be advisable to revisit and strengthen the
protections provided under FMLA.
Many people who choose to care for someone must forgo employment to do so. As a
result, they forfeit any Social Security credits they may have earned otherwise, thus thwarting
their own retirement security. Even if these individuals were not interested in a tax credit, they
might be able to benefit from an amendment to the Social Security Act, which would allow them
to accumulate Social Security credits according to some algorithm that converted care giving
hours into quarters of Social Security coverage. Before proposing such an amendment, it would
be essential to undertake both a feasibility study and a cost-benefit analysis. It would also be
necessary to create a mechanism to quantify the care services and to provide oversight.
With both tax breaks and the potential allocation of Social Security work credits for
informal care giving, a demonstration study on a small scale would be a mechanism for states to
determine the best methods to support quantification and auditing. By allowing smaller scale
Page 148
147
demonstration studies, creative and effective delivery models may be developed to address the
cultural, social, psychological, and medical needs of both elders and informal caregivers.
Enhancing the Positive Aspects of Current Programming
Case management has been identified as a major factor in helping people to remain in the
community and is a key factor in enabling PACE to succeed where traditional Adult Day Health
Centers often do not. Although case management occurs with ADH, it is enhanced and more
comprehensive in a PACE program. The case management approach demands that all members
of the care team, including the nurse, doctor(s), nutritionist, social worker, physical therapist,
homemaker, caretaker, and the individual, review the plan to insure that all of the client’s needs
are being met and all parties agree that the plan is appropriate. Case review and care planning
must be active, consistent, and focused. It is important to delve into the living arrangement
thoroughly and determine whether it is the safest and best option for all involved. Case
management needs full disclosure, especially from the caregiver and elder, and without a full and
accurate picture this information is unknown.
There are many positive aspects of long-term care programming that work and work well
within the Commonwealth. Enhancement of current programming, while insuring that the
programs are and continue to be of good quality, providing the needed services for elders is
imperative. Case management is an aspect of every program covered by Medicaid that services
Massachusetts elders in the community. Regulations for these programs identify case
management as part of the covered program. Case management is not an easy aspect of the care
giving role, because it expects cooperation in goal setting with the elder and their care providers,
Page 149
148
both formal and informal, while insuring cost effective care. Even the most skilled case
managers, with a difficult clientele, will have minimal effectiveness.
In addition to case management, two other related aspects of PACE appear to make an
enormous difference in positive outcomes for elders, as opposed those in traditional adult day
health (ADH) programs. The levels of both physician oversight and care coordination are greater
in PACE. This close contact and communication among the primary care physician and specialty
physicians is the secondary piece that drives PACE success. To insure the success of
community-based Medicaid programs that have case management mandates within their current
regulations, an important aspect that must be included is communication with physicians. The
current regulations for community-based Medicaid programs should be specifically evaluated for
the depth of the case management role, to determine if the role is significant enough to meet the
needs of the population. The role should have a holistic approach and include physician
communication and oversight.
Mandating case management services for populations with chronic illnesses that are not
covered by a formal services provider would potentially minimize hospitalizations and acute
exacerbations. In order to begin to see cost containment of services to high use elders, and those
who are unstable or without a single healthcare provider overseeing their care needs, this type of
mandate would work. If the system allowed for case managers to be assigned to elders who are
solely receiving informal support, many benefits could accrue. The case manager could help to
identify clinical needs before a crisis arises. Case managers might also be able to ascertain
whether caregivers need additional training in certain procedures or if they would benefit from
additional services, such as counseling or respite care. Case managers might even be the conduit
Page 150
149
by which caregivers learn of other support or educational opportunities. In any of these
capacities, case managers would be in a position to ensure that the needs of both caregivers and
elders are fully met. Although the costs would not be insubstantial, in the grand scheme, it would
be far less expensive than early institutionalization or avoidable emergency room visits. One way
to avoid overburdening case managers would be to require that an elder either have documented
functional deficits or meet the criteria for nursing facility eligibility to qualify for case
management services.
Emergencies do happen, and if there is an effective case manager involved during stable
conditions, the case manager could ensure that acute events, such as hospitalizations, go
smoothly. The case manager can ensure that medication and clinical records are up-to-date at the
time of admission, which may avoid unnecessary testing and confusion over treatment options.
In addition, the case manager can help make sure that the necessary coordination among clinical
service providers takes place and all parties have the same information. Another benefit to case
management in such instances might be to “introduce” the caregiver to the treatment team, thus
enhancing collaboration between the clinical team and the caregiver and avoiding the
communication gaps and subsequent mistrust that can often follow. If a case manager is called in
only after an acute event, the crisis model of care kicks in, often resulting in unrealistic goals and
expectations of caregivers including formal providers, and informal supports.
I have suggested this case management model for those without any formal care services.
The first step to implement such a model would be to determine which chronic medical condition
or conditions causes the highest cost, most acute hospitalizations, or most frequent emergency
visits. I would suggest a pilot initiative, with a case group and control, so that outcomes can be
Page 151
150
monitored at set time intervals. Using a population of seniors who receive informal support only,
case management services could be provided for the specific chronic medical condition(s)
chosen. At the end of the pilot program, if it is determined that these people do “better” and that
the program demonstrated a cost savings, then the case management program could be expanded.
People are often thrown into the caregiver role. A mother develops pneumonia. Her adult
son moves in temporarily to help her with cooking and cleaning, but she has a stroke and
suddenly the son, who never expected to be his mother’s caregiver, must do a lot more. His
mother recovered, but was left with impairments. She had told the hospital her son lived with
her, so they sent her home without services. Now he is alone, with no one to turn to and no idea
what to do. This is a familiar story. The labyrinth of Medicaid services and medical care is
difficult to navigate even for those who understand the system. It can quickly become
overwhelming to ailing seniors or their exhausted family members. People don’t know where to
begin, or who to ask. But what if Medicaid members were linked with a nurse coordinator, who
is knowledgeable about their case, their needs, and their history—someone who knows the
services available, understands the pitfalls they might encounter, and can guide them to the
support and services they need? At the very least, a nurse coordinator would help both the client
and the caregiver decrease their fears, avoid panic, and obtain the services they need. A nurse
coordinator might also decrease cost, by doing less than a case manager but assist with the
determination of needs and allocation of services. How does this differ from the case
management suggested above? Case management is specific to the disease. In the hypothetical
case of the mother and son described above, once the nurse coordinator was certain that the
patient and the caregiver were set up at home with what they needed, that might be it. If,
however, the mother required continued management or developed a chronic condition, then she
Page 152
151
might transition into case management. This approach would have an upfront cost, but I feel
with proper analysis, implementation, guidelines, and monitoring, the eventual outcome would
be a savings.
I suggest that each Medicaid client with a determining factor, whether a hospitalization
diagnostic code or chronic condition, have a nurse review the billing case within the Medicaid
system. The nurse would look at the bill and see that in this example, a Medicaid client was
treated for pneumonia then had an acute hospitalization for a stroke and was sent home. That
nurse could then make contact to insure that community services are not needed, and offer
information regarding programming and services available should they be needed. I feel that in
the long term, there would be a cost savings. Unfortunately, Medicaid is still used fraudulently as
insurance for those who do not report income. Additionally, because there are minimal controls
set in the Medicaid system, there are many who go from doctor to doctor to get medications and
treatments that are unnecessary and are of high cost, and through billing review these could be
identified and minimized.
Funding Stream Options
The current state of the economy is a major driver in policy implications and
recommendations, as is the planned implementation of universal healthcare. Finding a funding
stream is by far the most important task before proposing any additional services to meet the
needs of this study population. Regardless of political affiliation, most agree that medical care
and services should be available to the American people. The questions that remain are how
services should be allocated, who is covered, how plans should be developed, and how to pay for
services. Most discussion and argument surrounds eligibility, resource allocation, and services
Page 153
152
offered, especially when state funded programs like Medicaid and federally funded programs
like Medicare are discussed. Economic constraints limit all aspects of service delivery, from
enrollment to eligibility and allocation.
A few years ago, before the economy took a down turn, I would have suggested tax
breaks and tax incentives to assist in service allocation for for-profit organizations, tax breaks
and incentives do not assist non-profits directly. Today, I believe other mechanisms are
necessary. I recommend looking at the role of informal caregiver more like employment. I
recommend auditing current programs and reallocating funds to optimize rebalancing,
community support, and deinstitutionalization. I present these recommendations in depth below,
along with suggestions to enhance implementation and to ensure adequate funding streams.
In the long-term care MassHealth Medicaid units, staff members are available to review
the medical records and complete client interviews to validate care provided to MassHealth
members. By ensuring that the facility provided the same level of care that was billed to
Medicaid, these staff members recoup millions of dollars annually from nursing facility audits.
These same staff members are available and able to complete medical record review and client
interviews to validate care provided to MassHealth members in the community-based programs,
such as adult foster care, assisted living/group adult foster care, and adult day health providers.
Even with these staff completing audits two to four times per year in nursing facilities
throughout the state, a 10–12% error rate continues (Cutter, 2010). Community-based programs
have never been audited for recoupment purposes. Many times throughout early 2000 to 2005,
trial audits were completed for an impact analysis. The findings were astonishing, with a 35–
40% error rate in overbilling. I submit that the impact to begin an audit of the Medicaid budget
Page 154
153
would be minimal: the staff to complete the audits is already a line item in the budget. The
results of such an audit could provide the necessary evidence to support requests to increase
funding for community-based programs and services.
With the Affordable Care Act of 2010 (ACA or Obamacare), health homes have been
identified as a care option to improve care and contain costs. Through Obamacare, states can be
creative in design of community programming for development of health homes for elders, and
in the first two years that the program is operational it will be federally funded at 90%
reimbursement with a 10% state contribution and after two years will return to the state Medicaid
match (typically 50/50 federal/state match). ACA states that with these homes, Medicaid can
reimburse care management, care coordination, health promotion, transitional care, support to
individuals and families, and referral to community and social supports. Specific health home
ideas for Massachusetts are discussed below. The details of how this Act will be seen in future
healthcare related bills and the development of the details will be forthcoming, as President
Obama moves forward in his second term in office.
Education and Outreach
There are groups funded by the government, including Aging and Disability Resources
Centers, that offer support and information for elders including programming. These centers are
wonderful resources, but many do not know where they are or how to find them. People are
unaware of the programs available to them or what questions to ask or where to begin. Often
crisis and illness begin the path of inquiry, and caregivers are overwhelmed. Disease specific
Page 155
154
organizations are extraordinarily well marketed, publicized and funded. It may be feasible to
develop a partnership among such an organization and both state and federal governments, with
a dollar for dollar funding match. This collaboration might stimulate creative thinking that would
result in innovative programming with a more generous funding stream. Independent
organizations would need to be reviewed to insure that there is not a conflict of interest. This
type of partnership may also lead to broader discussions with stakeholders that have different
backgrounds and knowledge bases.
Although the overarching goal of outreach and increasing information availability is
altruistic, maintaining a database of available services of this magnitude may be difficult and
costly to maintain. Once such a database existed, it would still be necessary to market it properly,
to ensure that elders and caregivers would know about it and how to access the data. In
Massachusetts, there are currently 30 organizations for seniors, known collectively as area
agencies on aging (AAAs; http://www.seniorconnection.org/aaa_asap.htm). Although these
agencies have a generous amount of information available, including topics such as referral
sources, insurance coverage, health programs, respite services, and other services, many people
have no idea that these organizations exist, where they are located, or what they have to offer.
People may be aware there is a senior center in their town, but they are unaware of the plethora
of information available there, or that their local senior center can assist in linking them to more
in-depth information and referral sources of the AAA. These networked organizations are
advocacy-based agencies whose mission is “to help older persons…live with dignity and choices
in their homes and communities for as long as possible” (www.n4a.org). They were created in
part through the Older Americans Act of 1965 and various amendments, and they are funded
through a mix of local, state, and federal sources. Historically, outreach efforts by these agencies
Page 156
155
have been less than optimal, often hampered by language barriers, failure to identify and
properly target audiences, or choice of marketing manner. Another limitation is location: most of
the AAAs in Massachusetts are located in urban or suburban areas and do not adequately serve
those in the more rural areas. Decision makers might consider tying future financial support for
these entities to demands for better outreach with defined benchmarks to demonstrate successful
improvement. Benchmarks should be identified with timelines of three to five years because
improved outreach may stimulate increased usage of services, which would, at least over the
short haul, cause a rebound effect.
Limitations to the Study
The data I used in this study was between eight and thirteen years old. More recent data
may be available from the Commonwealth of Massachusetts, although there was a four to five
year period during which there were multiple methods of data input, which affected the
consistency of data and possibly compromised its accuracy. The sample from 2005 to 2010
would not be as accurate a picture of the community-based population as the one I used.
Between 2008 and 2010, all of the consistency and data issues were resolved. More recent data
can be requested. This was a convenience sample. Although there have been no significant
changes in eligibility requirements for long-term care or in the overall quality of care facilities in
that time and we have not experienced any monumental shifts in the health picture for
individuals in this population, it is still possible that the sample may not present a completely
accurate picture of the needs of today’s elders.
Page 157
156
The study was limited to data on individuals who received Medicaid benefits in
Massachusetts. Although the findings provide valuable insights for policy makers nationally and
in other states, it may not be possible to extrapolate all of the findings to other states. In addition,
the findings may not be generalizable across other populations, such as those who are privately
insured or who are not economically disadvantaged. Although the total sample was large, certain
subpopulation samples were small, which may limit the ability to generalize to other
subpopulations. The smallest population, those with no formal services in place, and no informal
supports (n = 269), may be more difficult to generalize to the total population.
The study did not evaluate outcomes data, an aspect of Anderson’s model and a piece that
is significant when looking at how people feel about the services and care they received. This
aspect was not evaluated because the data collection instrument used did not capture outcomes in
a time study manner and also did not capture effects of treatments. A qualitative piece would
have been helpful in substantiating the findings and in teasing out any additional influences that
were not captured from the data collection instrument. It may have been possible to demonstrate
outcomes using a population characteristic study; looking at people over time and assessing
functional improvements or declines in relation to services available.
The MDS–HC tool was a logical choice because it is the device used in Massachusetts to
evaluate eligibility for specific programs under Medicaid. However, the tool does have inherent
limitations. The information is self-reported, which immediately raises the possibility of recall
bias and subjectivity. Without strict guidance in how to interpret the guidelines, individuals may
not all have the same ideas about how to assess their functional levels. Self-reported data can
also be influenced by a desire to please, to provide the “right” answers, the answers that meet the
Page 158
157
perceived expectations of the assessor or the caregiver. Finally, the answers provided by
applicants can be affected by what they personally want: individuals who want to be seen as
strong and independent may overstate their capabilities, while others may portray themselves as
more needy than they actually are.
An additional limitation is related to the choice of statistical analyses. I could have used
other types of statistics, which may have shown more detailed or slightly different results. For
example, profile analysis is a statistical tool that is more commonly completed on psychological
research, but is starting to be used in sociological research as well. Profile analysis allows for
several measurements on a single dependent variable, to compare across groups to determine if
there is an effect.
In this study, I only looked at the community-based population because the focus was on
factors that influence individuals to remain in the community and on how informal and formal
care giving affects the types of services that community-based elders receive. A major factor was
to evaluate the population in the community that also met nursing facility eligibility. Meeting the
clinical eligibility regulations is the baseline for eligibility in many of the community-based
Medicaid programs, especially those that are qualified under the Home- and Community-based
Services Waiver 1115. If there had been similar information available for elders residing in
nursing homes, I would have been able to compare and create implementation strategies that
look not only at keeping people in the community, but also at assisting elders to transition from
nursing facilities back into the community.
Directions for Future Research
Page 159
158
Through this study, I elucidated numerous valuable insights into the factors that
contribute to success in providing for the needs of elders within the community. Informal
caregiver support, stable housing often in a cohabitative setting, and the role of coordinated
formal and informal care supports and services are some of the many aspects that contribute to
successful aging. However, a primary value of this study lies in the number of questions it raises.
A fundamental and difficult topic remains in economics, the cost of remaining in the
community versus moving into an institutionalized setting. A nursing facility is an inclusive care
setting: all services, care, housing, and medical needs, are wrapped into one cost per day to
Medicaid or the individual. In the community setting it is “a la carte”. Each need is a separate
cost to the individual or if it is a medical service, to Medicaid. Depending on the care needs,
community-based care can become quite costly. Additionally valuation of quality of care and
quality of life is so individual and difficult to place a dollar amount on, so a follow-up study on
cost of community-based care would be helpful.
Having a residence is more than just having a place to live. A home conveys a sense of
pride and a feeling of stability and safety. Where there are homes, there are usually neighbors
and together, these neighbors are a community and a support network. All of these characteristics
provide a foundation upon which to build a service network. In this study I found that caregiver
group and living arrangement were not correlated. There are difficulties discharging people from
acute hospital stays and nursing facilities when they do not have a housing option to return to. It
would be interesting to examine how frequently housing is the block to community-based
services for an elder being discharged either from a nursing facility or an acute care hospital.
Page 160
159
There remains a gap for those who are above the poverty limits; those who cannot meet
the requirements to receive Medicaid and other state and federal subsidized programs, but who
also do not have enough private funds to pay for necessary services. It is unknown how many
people fall into this classification of elders, and increasingly, it may capture all of us. Ironically,
these elders may struggle to meet their needs and may be more likely to forego healthcare,
medications, nutrition, home heating, and other basic needs than other individuals who receive
public assistance. A study to determine how large the population is, and what the needs are is a
place to begin. There may not be a service need, or it may be a hole too large to begin to patch.
Without a formal study and effective research elders may continue to suffer.
The data and findings from this study provide me with several avenues to consider. A
future research report could branch out beyond this study by repeating the study with comparable
states that use the MDS–HC as an assessment tool, or repeat the study using more recent data to
compare findings to this timeline. A study could be completed to compare the community
population to the nursing facility population, at specific points in service requests. The
assessment tools used in Massachusetts are comparable for community-based programs and
nursing facility services. Upon beginning a community-based long-term care program, providers
are expected to complete an accurate MDS–HC and, in nursing facilities, an MDS 3.0. These
data tools use the same language and definitions, and can be cross-referenced in many places and
thus analyzed fairly easily.
Few studies quantify nursing facility eligibility and other factors that influence eligibility.
Home and Community Based Services Waivers (HCBSW) based on 1115 by CMS are quantified
by nursing facility clinical eligibility. Taking items within this study and further examining
Page 161
160
characteristics that influence nursing facility eligibility would be future research that may guide
demonstration waivers for the Commonwealth, potentially identify new avenues for best
practice, and provide better management and care of elders in the community. Looking at elders
through both quantitative research and qualitative research methodologies would also allow for a
more comprehensive look at elders in the community to understand the intricacies of how they
manage.
Conclusion
I set out to gather and evaluate information that would be useful in developing policies to
improve the care provided to older adults living in the community and help decision makers in
their efforts to rebalance long-term care. I believe that the quality of care provided by informal
caregivers is vital in keeping elders in the community. Because of this, I focused on the role of
the informal caregiver. In particular, I wanted to know the differences between the formal and
informal care delivered, to better understand their characteristics. For example, what burdens are
placed upon informal caregivers and what do they need to succeed? In doing this, I hoped to be
able to devise recommendations for state level policies and legislative language to support these
caregivers, thereby helping elders to thrive while remaining in the community.
I determined that although my original hypothesis is correct—people who need help with
activities of daily living are just as likely to be found in the community, receiving informal care,
as they are formal care —the circumstances are not as clear-cut as they may at first appear. These
informal caregivers have taken on significant responsibilities, usually without adequate training.
They receive little or no financial support, even though they often have assumed significant
financial burdens. The amount of educational support and respite assistance is highly variable
Page 162
161
and correlates strongly with location and with an individual’s ability and motivation to ask
questions and request assistance. At least in Massachusetts, those who live in large urban areas
are much more likely to have access to adult day care and other services than those in the rest of
the state.
I also learned that many individuals in Massachusetts who are eligible for nursing home
care (in this study, 11.6%) choose to remain in the community instead. For some reason, the
individuals, who clearly demonstrate clinical eligibility for nursing facility care and meet the
criteria for 1115c waiver programs have chosen to remain at home. Although it was outside the
scope of this study to quantify the amount of money these decisions save the federal and state
government, given the current bed charges alone in long-term care facilities, it is a substantial
savings to the public. At the same time, it may be posing an unfair burden on the caregivers.
Another key lesson I learned was the importance of adequate and appropriate housing
options and support for housing outside of institutional facilities. The populations studied were
economically disadvantaged, and likely would also meet eligibility for elder housing. Elder
housing options are limited, and many have long wait lists. There are some Massachusetts
Medicaid programs that offer housing supplements through partnerships with Supplemental
Social Security programs. These programs again are quite limited in both number of available
housing units and geographic locations. Although assisted living residences have boomed in the
past several years throughout the country including Massachusetts, they are almost exclusively
private pay, and for those with any economic disadvantage out of the feasible financial range.
Page 163
162
The overarching goal was to then identify the similarities to those receiving care from
informal caregivers and supports and formal care services as well as the supports that are beyond
the scope of formal services, and determine how policies could be developed to support this role.
I was able to determine that there are many similarities. Informal caregivers are providing care to
individuals at similar levels of disabilities. People with informal caregivers are as likely to meet
nursing facility eligibility as those with formal care services.
Using what I learned, I formulated some recommendations and implementation ideas for
decision makers. Where possible, I identified particular steps to take and who should take them.
In some cases, I identified possible sources of funding or ways to cut or recoup costs, all while
providing better care for the nation’s fast expanding elderly community.
Rebalancing long-term care lies in the hands of many and everyone involved shares some
burden of responsibility for making the system work. True, there are bureaucratic obstacles to be
surmounted, but that tells only part of the story. Problem solving and effective forward motion
can only happen if all parties take responsibility for making it happen. This includes policy
makers, service providers, caregivers, and elders.
Service fragmentation and issues with continuity of care come not only from minimal
programming and lack of service availability, but also from not knowing what is available. Some
aspects of that blame lie with the elders, too proud to ask for help, too sick to care, too scared
that if they ask for help they will automatically be sent to a nursing facility. Some blame lies
with the caregiver the informal caregiver who thinks they can do it all- work full time, care for
their children and ill parent, and maintain life, afraid perhaps that if services start their role will
be less important. Caregivers, who may not have fully understood what they were getting into
Page 164
163
initially, must recognize that they need help and in some cases, they must be willing to pay for it.
Elders and their informal caregivers have responsibilities. They must be honest about what they
need and be willing to see and accept help. Some blame lies with the formal service provider, the
hospital that assumes that what is being communicated is true, or that the dapper gentleman
seems to have all his faculties as the caregiver and not ask more questions, only to find out two
weeks later on readmission that he is not the caretaker but the patient at home with Alzheimer’s
disease. The community-based formal service providers are often also the evaluators, who
determine service provision, and may base services on those only available through their
agencies, or may be unaware of alternative available services. Service providers must be able to
work with the caregivers and elders to provide effective and safe care, while maintaining clear
two-way communication. When considering changes to regulatory language, policies, and laws,
policy makers must actively involve all of the stakeholders, including providers, elders, and
caregivers, by soliciting and respecting their opinions and expertise. It is much easier to
implement new policies if the stakeholders have bought into the concept in advance. The
struggle with long-term care provision has a long road ahead, and this is only a small step
towards some resolution.
Page 165
164
References
Aday, L.A. & Andersen, R.M. (1974). “A framework for the Study of Access to Medical Care.”
Health Services Research, 9, 209-220.
Andersen, R.M. (1968). Behavioral Model of Families’ Use of Health Services. Research Series
No. 25. Chicago, IL: Center for Health Administration Studies, University of Chicago.
Andersen, R.M. (1995). Revisiting the Behavioral Model and Access to Medical Care: Does it
Matter? Journal of Health and Social Behavior, 36(3), 1-10.
Akamigbo, A.B. & Wolinsky, F.D. (2006). Reported Expectations for nursing home placement
and their role as risk factors for nursing home admissions. The Gerontologist, 46, 464-
473.
Al-Snih, S.; Markides, K.S.; Ray, L.A.; Freeman, J.L.: Ostir, G. .: & Goodwin, J.S. (2006).
Predictors of Healthcare Utilization among older Mexican Americans. Ethnicity and
Disease, 16(3), 640-646.
Angel, J.L., Angel, R.J., Aranda, M.P., & Miles, T.P. (2004). Can the family still cope? Social
support and health as determinants of nursing home use in the older Mexican-origin
populations. Journal of Aging and Health, 16(3), 338-354.
Arno, P.S., Levine, C. &Memmott, M.M. (1999). The economic value of informal care giving.
Health Affairs, The Policy Journal of the Health Sphere, 18(2), 182-8.
Babcock, E.D. & Watt, H. (2002). Keeping elders home: new lessons learned about supporting
frail elders in our communities. (Issue Brief No. 7) Boston, Ma: The Massachusetts
Health Policy Forum.
Baker, D.I. & Bice, T.W. (1995). The influences of urinary incontinence on publicly financed
home care services to low-income elderly people. The Gerontologist, 35(3), 360-369.
Baumgarten, M., Lebel, P., Laprise, H, Leclerc, C. & Quinn, C. (2002). Adults day care for the
frail elderly: outcomes, satisfaction, and cost. Journal of Aging and Health, 14(2), 237-
259.
Bellows, N.M. & Halpin, H.A. (2008). MDS-based state Medicaid reimbursement and the ADL-
decline quality indicator. The Gerontologist, 48(3), 324-329.
Bharucha, A.J., Pandav, R., Shen, C., Dodge, H.H., & Ganguli, M. (2004). Predictors of nursing
Page 166
165
facility admission: a 12-year epidemiological study in the United States. Journal of the
American Geriatrics Society, 52(3), 434-439.
Black, W., Brown, R., Cybulski, K. &Hall, J. (2003). A descriptive analysis of older adults in
Boston, Massachusetts. Robert Wood Johnson Foundation: Mathematica Policy
Research, Inc.
Bonsang, E. (2009). Does informal care from children to their elderly parents substitute for
formal care in Europe. Journal of Health Economics, 28(10), 143-154.
Bovbjerg, R.R. & Ullman, F.C. (2002). State related changes in health policy for low income
people in Massachusetts. The Urban Institute State Update, No. 17, March 2002
Bracke, P., Christiaens, W. & Wauterickx, N. (2008). The Pivotal Role of Women in Informal
Care. Journal of Family Issues, 29, 1348. DOI: 10.1177/0192513X08316115
Branch, L.G. (2001). Community long-term Care services: what works and what doesn’t. The
Gerontologist, 41(3), 305-306.
Burdick,D.J., Rosenblatt, A., Samus, Q.M., Steele, C., Baker, A., Harper, M., Mayer, L., Brandt,
J., Rabins, P. & Lyketsos, C.G.(2005) Predictors of Functional impairment in residents of
assisted living facilities: The Maryland Assisted Living Study. The Journals of
Gerontology Series A: Biological Sciences and Medical Sciences, 60(2), 258-264.
Burton, L, Kasper, J, Shore, A, Cagney, K, LaVeist, T, Cubbin, C & German, P. (1995). The
Structure of Informal Care: Are There Differences by Race? The Gerontologist, 35(6),
744-752. doi: 10.1093/geront/35.6.744
Cantrell, R. (2007). Common Care: Toward a More Acceptable Option of Care for the Frail
Elderly. Social Work in Public Health, 23(1). doi:10.1300/J523v23n01_04 61
Carmichael, F. & Charles, S. (2003). The opportunity cost of informal care: does gender matter?
Journal of Health Economics, 22(5), 781-803.
Casado, B.L., von Vulpen, K.S. & Davis, S.L. (2011). Unmet needs for home and community
based services among frail older Americans and their caregivers. Journal of Aging and
Health, 23, 529-553.
Centers for Medicare & Medicaid Services. History of Medicare and Medicaid. Retrieved
January 15, 2005, from http://www.cms.hhs.gov/about/history/searchr.asp
Chappell,N. & Hollander, M. (2001). Cost implications of formal supports. Canada: Health
Transition Fund.
Charles, K.K. & Sevak, P. (2005). Can family caregiving substitute for nursing homes? Journal
Page 167
166
of Health Economics, 24(6), 1174-1190.
Chen, X.M & Thompson, E.A. (2010). Understanding factors that influence success of home-
and- community based services in keeping older adults in community settings. Journal of
Aging and Health, 22(3), 267-291.
Cohen-Mansfield, J. &Wirtz, P. (2007). Characteristics of adult day care participants who enter
a nursing home. Psychology and Aging, 22(2), 354-360.
Couture, M.C., Nguyen, C.T., Alvarado, B.E., Velasquez, L.D. & Zunzunegui, M.V. (2008).
Inequalities in breast and cervical cancer screening among urban Mexican women.
Preventative Medicine, 47(5), 471-476.
Cutter, J. (2011) interview, clinical manager, Office of Long-Term Care, MassHealth
Dale,B., Saevareid, H.I., Kirkevold,M. & Solderhamm, O. (2008). Formal and informal care in
relation to activities of daily living and self-perceived health among older care-dependent
individuals in Norway. International Journal of Older People Nursing, 3(3), 194-203.
Dilworth-Anderson,P, Brummett, B.h., Goodwin, P, Williams, S.W., Williams, R.B., & Siegler,
I.C. (2005). Effect of Race on Cultural Justifications for Caregiving. Journals of
Gerontology: Series B, 60(5), S257-S262.
Diwan, S. (1999). Allocation of case management resources in long-term care: Predicting high
use of case management time. The Gerontologist, 39(5), 580-590.
Elliott, R. (2009). Non-adherence to medicines, no solved but solvable. Journal of Health
Services Research & Policy, 14, 58-61.
Farran, C.J., Miller, B.H., Kaufman, J.E. & Davis, L. (1997). Race, Finding Meaning, and
Caregiver Distress. Journal of Aging and Health, 9(3), 316-333.
Feld, S., Dunkle, R.E.,& Schroepfer, T. (2004). Race/Ethnicity and Marital Status in IADL
Caregiver Networks. Research on Aging, 26(5), 531-558.
Fingerman, K.L., VanderDrift, L.E., Dotterer, A.M., Birditt, K.S., & Zarit, S.H. (2011). Support
to Aging Parents and Grown Children in Black and White Families. The Gerontologist,
51(4), 441-452. doi: 10.1093/geront/gnq114
Friedman, S.M., Steinwachs, D.M., Rathouz, P.J., Burton, L.C.& Mukamel, D.B. (2005).
Characteristics predicting nursing home admission in the Program for All Inclusive Care
for the Elderly People. The Gerontologist,45(2)157-166.
Fultz, N.H., Burgio, K., Diokno, A.D., Kinchen, K.S., Obenchain, R. & Bump, R.C. (2003).
Page 168
167
Burden of stress urinary incontinence for community dwelling women. American journal
of Obstetrics and Gynecology, 189(5), 1275-1282.
Fuller-Thomson, R.M. (2008). Falling through the social safety net : food stamp use and non use
among older impoverished Americans. The Gerontologist, 48 (2), 235-244.
Gannon, B. & Davin, B. (2010). Use of formal and informal care services among older people in
Ireland and France. The European Journal of Health Economics: Health Economics In
Prevention and Care, 11(5), 499-511. ISSN: 1618-7601
Gatti, M.E., Jacobson, K.L., Gazmararian, J.A., Schmotzer, B. & Kripalani, S. (2009).
Relationships between beliefs about medications and adherence. American Journal of
Health-System Pharmacy, 66, 657-664.
Gaugler, J.E., Duval, S., Anderson, K.A. & Kane, R.L. (2007). Predicting nursing home
admission in the U.S.: a meta-analysis. BioMed Central/BMC Geriatrics, 7, 13.
Gaugler, J.E., Kane, R., Kane, R. & Newcomer, R. (2006). Predictors of institutionalization in
Latinos with dementia. Journal of Cross-Cultural Gerontology, 21(3-4), 139-155.
Gaugler, J.E., Leach, C.R., Clay, T. & Newcomer, R.C. (2004). Predictors of nursing home
placement in African Americans with dementia. Journal of the American Geriatrics
Society, 52(3), 445-452.
Gore, S. (2011). Innovations in Medicaid: Considerations for MassHealth. Prepared for-Center
for Health Care Strategies, Inc
Gosman-Hedstrom,G & Claesson, L. (2005). Gender perspective on informal care for elderly
people one year after acute stroke. Aging Clinical and Experimental Research, 17(6),
479-485.
Graber, D., Liao, J. & Buchanan, R.J. (2002). ADL Functioning Among Individuals Entering
Nursing Homes. Academy Health Services Annual Research Health Policy Meeting, 19:
13.
Grando, V.T., Mehr, D., Popejoy, L., Maas, M., Rantz, M, Wipke-Tevis, D.D. & Westhoff, R.
(2002). Why older adults with light care needs enter and remain in nursing homes.
Journal of Gerontological Nursing, 28(7), 47-53.
Green, J.L. & Ondrich, J.I. (1990). Risk factors for nursing home admissions and exits: A
discrete time hazard function approach. Journal of Gerontology, 45(6), S250-S258.
Hanaoka, C. & Norton, E.C. (2008). Informal and formal care for elderly persons: How adult
children's characteristics affect the use of formal care in Japan. Social Science &
Medicine, 67(6), 1002-1008.
Page 169
168
Hawes, C., Fries, B.E., James, M.L., & Guihan, M. (2007). Prospects and Pitfalls: Use of the
RAI-HC Assessment by the Department of Veterans Affairs for Home Care Clients. The
Gerontologist, 47, 378-387.
Hendrickson, L. & Kyzr-Sheeley, G. (2008). Determining Medicaid Nursing Home Eligibility: A
Survey of State Level of Care Assessment. Prepeared for Rutgers Center for State Health
Policy.
Herrera, A.P., Lee, J., Palos, G. Torres-Vigil, I.(2008). Cultural influences in the patterns of
long term care use among Mexican American family caregivers. Journal of Applied
Gerontology, 27(2), 141-165.
Hirdes, J. P., Fries, B.E., Morris, J.N., Ikegami, N., Zimmerman, D, Dalby, D.M., Aliaga, P.,
Hammer, S., & Jones, R. (2004). Home Care Quality Indicators (HCQIs) Based on the
MDS-HC. The Gerontologist, 44, 665-679.
Hirdes, J.P., Poss, J.W., & Curtin-Telegdi, N. (2008). The Method for Assigning Priority Levels
(MAPLe): A new decision-support system for allocating home care resources. BioMed
Central, 6:9. doi:10.1186/1741-7015-6-9
Hong,S. (2010). Understanding patterns of service utilization among informal caregivers of
community older adults. The Gerontologist,50(1), 87-99.
Johnson & Wiener. (2006). A Profile of Frail Older Americans and Their Caregivers.
Occasional Paper Number 8, The Retirement Project, Urban Institute.
Karlsson, S., Edberg, A. & Hallberg, I.R. (2008). Functional ability and health complaints among
older people with a combination of public and informal care vs. public care only.
Scandinavian Journal of Caring Sciences, 22(1), 136-148.
Kasper,J,D., Pezzin, L.E. & Rice, J.B. (2010). Stability and changes in living arrangements,
relationship to nursing home admission and timing of placement. Journals of
Gerontology Series B: Psychological Sciences and Social Sciences,65B(6),782-791
Katz, S., Ford, A.B., Moskowitz, R.W., Jackson, B.A.& Jaffe, M.W. (1963). Studies of illness in
the aged. The Index of ADL: a standardized measure of biological and psychosocial
function. JAMA,185(12),914-9.
Konetzka, R.T. & Werner, R.M. (2009). Review: Disparities in Long-term care Building equity
into market based reforms. Medical Care Research and Review, 66(5),491–521.
Lai & Surood (2010). Types and factor structure of barriers to utilization of health services
among aging south Asians in Calgary Canada, Candian Journal on Aging, 29(2), 249-
258.
Page 170
169
Liao, C.C. & Chelmow, T. (2007). Assessment of Disparity in Home Health Services among
Medicaid Elder Recipients. Annual meeting: American Public Health Association.
Larsson, K. & Silverstein, M. (2004). The effects of marital and parental status on informal
support and service utilization: A study of older Swedes living alone. Journal of Aging
Studies, 18, 231–244.
Lawton, M.P. & Brody, E.M. (1969). Assessment of Older People: Self-Maintaining and
Instrumental Activities of Daily Living. The Gerontologist, 9(3), 179-186.
Lee, T., Kovner, C.T., Mezey, M.D. & Ko, I. (2001). Factors influencing long-term homes care
utilization by the older population: implications for targeting. Public Health Nursing,
18(6), 443-449.
Li, L.W. (2005). Trajectories of ADL disability among community-dwelling frail older persons.
Research on Aging, 27(1), 56-79.
Li, L.W. & Fries, B.E. (2005). Elder Disability as an Explanation for Racial Differences in
Informal Home Care. The Gerontologist, 45(2), 206-215. doi: 10.1093/geront/45.2.206
Lipsky, M. (2010). Street-Level Bureaurocracy: Dilemmas of the Individual in public services.
Russell Sage: New York, NY.
Litwin, H. & Attias-Donfut, C. (2009). The inter-relationship between formal and informal care:
a study in France and Israel. Ageing &Society, 29(1), 71-91.
Long, S.K., Liu, K., Black, K., O’Keefe, J. & Molony, S. (2005). Getting by in the community:
lessons from frail elders. Journal of Aging and Social Policy, 17(1), 19-44.
Luppa, M., Luck, T, Weyerer, S, Konig, H., Brahler, E. & Riedel-Heller, S.G. (2010). Prediction
of institutionalization in the elderly. A systemic review. Age and Aging, 39(1), 31-38.
Marek, K.D., Popejoy, L., Petroski, G, Mehr, D., Rantz, M. & Lin, W. (2005). Clinical
outcomes of aging in place. Nursing Research, 54(3), 202-211.
Matsumoto, M. (2007). Predictors of institutionalization in elderly people living at home: The
impact of continence and commode use in rural Japan. Journal of Cross-Cultural
Gerontology, 22(4), 421-432.
Mentzakis, E., McNamee, P. & Ryan, M. (2009). Who cares and how much: exploring the
determinants of co-residential informal care. Review of Economics of the Household,
7(3), 283-303.
Merck Institute of Aging and Health. (2000). The state of aging and health in America.
Page 171
170
Gerontological Society of America.
Miller, B., McFall, S., Campbell, R.T. (1994). Changes in sources of community long term care
among African American and White frail older persons. Journal of Gerontology, 49(1),
S14-S24.
Miller, E.A. (2006). Explaining Incremental and Non-Incremental Change: Medicaid Nursing
Facility Reimbursement Policy, 1980-98. State Politics & Policy Quarterly 6.2: 117,
150,240.
Miller, E.A., Allen, S.M. & Mor, V. (2009). Commentary: Navigating the Labyrinth of Long-
Term Care: Shoring Up Informal Caregiving in a Home- and Community- Based World.
Journal of Aging & Social Policy, 21, 1-16.
Miller, E.A. & Rosenheck, R.A. (2006). Risk of nursing home admission in association with
mental illness nationally in the Department of Veterans Affairs. Medical Care, 44(4),
343-351.
Min, L., Yoon, W., Mariano, J., Wenger, N.S., Elliott, M.N., Kamberg, C., & Saliba, D. (2009).
The Vulnerable Elders- 13 survey predicts 5-year functional decline and mortality
outcomes in older ambulatory care patients. Journal of the American Geriatrics Society,
57(11), 2070-2076.
Morris, J.N.,Fries, B.E., Bernabei, R., Steel, K., Ikegami, N., Carpenter, I., Gilgen, R.,
DuPasquier, J., Frijters, D., Henrard, J., Hirdes, J.P. & Belleville-Taylor, P. (2001). RAI-
Home Care Assessment Manual for Version 2.0. Washington, D.C.: interRai
Corporation.
Noel-Miller, C. (2010). Spousal loss, Children, and the risk of Nursing Home Admission. The
Journals of Gerontology: Series B, 65B(3), 370-380.
Pinquart, M. & Sorensen, S (2005). Ethnic Differences in Stressors, Resources, and
Psychological Outcomes of Family Caregiving: A Meta-Analysis. The Gerontologist,
45(1), 90-106.
Publication Manual of the American Psychological Association (6th
ed.). (2010). Washington,
D.C.: American Psychological Association.
Ray, M. (2006). Informal care in the context of long-term marriage: the challenge to practice.
Practice, 18(2), 129-142.
Renn, N.L. (2005) An examination of factors that lead to successful aging in a community
dwelling elderly population: A follow-up analysis. Available through Proquest.
Ritchie,C.S., Roth, D.L., & Allman, R.M. (2011). Living with an aging parent, “it was a
Page 172
171
beautiful invitation”.
Saltman, R.B., Dubois, H.F.W., & Chawla, M. The impact of aging on long-term care in Europe
and some potential policy responses. International Journal of Health Services, 36(4),
719-746.
Scharlach,A.E., Giunta,N., Chow,J., & Lehning, A. (2008). Racial and ethnic variations in
caregiver service use. Journal of Aging and Health, 20(3), 326-346.
Schwab. T.C.,Leung, K., Gelb, E., Meng, Y. & Cohn,J. (2003). Home-and community-based
alternatives to nursing homes. Journal of Aging and Health, 15(2), 41-48).
Shugarman, L.R., Fries, B.E. & James, M. (1999). A comparison of home care clients and
nursing home residents: can community based care keep the elderly and disabled at
home? Home Health Care Services Quarterly, 18(1), 25-45.
Spillman, B.C. & Long, S.K. (2009). Does high caregiver stress predict nursing home entry?
Inquiry, 46(2), 140-161.
Spillman, B.C. & Pezzin, L.E. (2000). Potential and active family caregivers: changing
networks and the “sandwich generation”. The Milbank Quarterly, 78(3), 347-374.
Spruyette, N., van Audenhove, C. & Lammertyn, F. (2001). Predictors of institutionalization of
cognitively-impaired elderly cared for by their relatives. International Journal of
Geriatrics Psychiatry, 16(12), 1119-1128.
Tennstedt, S.L., Crawford, S.L., & McKinlay, J.B. (1993). Is Family Care on the Decline? A
Longitudinal Investigation of the Substitution of Formal Long-Term Care Services for
Informal Care. The Millbank Quarterly, 71(4). 601-624.
Tennstedt, S.L., Harrow, B. &Crawford, S.L. (1996). Informal care versus formal services:
changes in patterns of care over time. Journal of Aging and Social Policy. 7(3/4), 71-91.
Vik, A., Maxwell, C.J., & Hogan, D.B. (2004). Measurement, correlates and health outcomes of
medication adherence among seniors. The Annals of Pharmacology, 38(2), 303-312.
Von Bonsdorff, M., Rantanen, T., Laukkanene, P., Suutama, T. & Heikkinen, E. (2006).
Mobility Limitations and Cognitive Deficits as Predictors of Institutionalization among
Community-Dwelling Older People. Gerontology, 52(6), 359-365.
Williams, S.W. & Dilworth-Anderson,N. (2002). Systems of social support in families who care
for dependent African American Elders. The Gerontologist, 42(2), 224-236.
Wimo, A., Sjolund, B.M., Skoldunger, A., Johansson, L., Nordberg, G., & vonStrauss, E. (2011).
Page 173
172
Incremental patterns in the amount of informal and formal care among non-demented and
demented elderly persons results from a 3-year follow-up population-based study.
International Journal of Geriatric Psychiatry, 26(1), 56-64.
Page 174
173
Appendix 1
Figure 1. Application of Andersen’s Behavioral Health Model to Study
Appendix 2
Enabling Resources
living arrangement, informal care support, formal
services
Hypothesis 1E: Persons who live with others are more likely to be in the informal caregiver support only group than the other care support groups.
Need
bladder incontinence, memory status, cognitive
status, comprehension, expression, behavior
problems, functional status IADL and ADL
Hypothesis 1F: Persons with higher functional
deficit are as likely to be in the informal caregiver
support only group as those receiving some formal
services.
Health Care
System &
External
Environment
-Massachusetts
Medicaid
-Economy
-Federal and
State policies
Predisposing Characteristics
-gender, race/ethnicity, marital status, primary
language, level of education, living arrangement
Hypothesis 1A: Hispanics and African-
American persons would be more likely than
non- Hispanic whites to be in the InfCS group
than in the other care support groups.
Hypothesis 1B: Persons with lower levels of
education would be more likely than their more
educated counterparts to be in the InfCS than in
the other care support groups.
Hypothesis 1C: Persons with a primary
language other than English would be more likely
than those whose primary language is English to
be in the InfCS group than in the other care
support groups.
Hypothesis 1D: Persons who are married would
be more likely than those unmarried to be in the
InfCS group than in the other care support
groups.
Hypothesis 1E: Persons who live with others are
more likely to be in the InfCS group than in the
other care support groups.
Enabling Resources
-informal care support, formal services
All of the hypotheses relate to Enabling
Resources
Personal
Health
Practices &
Use of Health
Services
-trade offs,
medication
compliance,
resistance to
care, use of
formal
services,
limits going
outside,
isolation,
social
functioning,
treatments,
alcohol and
smoking,
medication
use,
consumption
Need
-bladder incontinence, memory status, cognitive
status, comprehension, expression, behavior
problems, functional status IADL and ADL
Hypothesis 1F: Persons with more deficits in
activities of daily living (ADL) would be more
likely than those with less deficits to be in the
InfCS group than other care support groups.
Perceived Health
Status
-self perceived
health status, client
feeling of poor
health
Evaluated Health
Status
-Nursing facility
eligibility,
depression risk,
functional status
Consumer
Satisfaction
Not being
evaluated in this
study.
Page 175
174
Appendix 2
Table 1. MDS-HC Variables used in this Study.
Variable Items from MDS–
HC being Used for this Study
Response Set Found in MDS–HC Reference
Item Location
on MDS–HC
Sample Characteristics
Gender
Race/ethnicity
Marital status
Language
Education
Living arrangement
Clinical and Functional Needs
Bladder incontinence
Medication adherence
Number of medications
Cognitive and Psychological
Needs
Memory status: short-term
Memory status: procedural
Cognitive status: ability
to make decisions about
the day
Comprehension
male or female
Native American, Asian, White, Black, Hispanic,
Pacific Islander
Married or other
English or other
less than high school graduate, high school
graduate or technical/trade school or some college,
college degree or greater lives alone or lives with others
continent, occasionally incontinence, incontinent
always compliant, compliant 80% of the time,
compliant less than 80% of the time, no
medications
0 to 9
memory ok or memory problem
memory ok or memory problem
independent, modified independence, minimally
impaired moderately impaired, severely impaired
understands, usually understands, often
understands, sometimes understands, rarely/never
BB. 1
BB. 3
BB. 4
BB. 5
BB. 6
CC. 6
I 1 a
Q. 4
Q. 1
B1a
B1b
B 2 a
C 3
Page 176
175
Expression
Diagnosis of Alzheimer’s
disease
Client feels he is in poor
health
Client believes capable
of increased independence
Functional status:
Activities of Daily Living
(ADL)
Instrumental Activities of
Daily Living (IADL) status
Cognitive Risks
Depression Risk
Behavior Problems
understands
understood, usually understood, often understood,
sometimes understood, rarely/never understood
not applicable or disease present
yes or no
yes or no
score
score
score
score
C 2
J. 1 g
K8a
H7b
H. 2 a, b, c, e,
f, g, h
H. 1 a, b, c, d,
e, f, g
E. 1 a,b,c,d, e,
f, g
E. 3 a,b,c,d,e
Note. Reference location on the MDS–HC tool, please see Appendix 3 for a copy of the tool.