University of Southern California Dissertations and Theses

Evaluation of the Care Advocate Program: Bridging managed care and home community -based services

(USC Thesis Other)

Evaluation of the Care Advocate Program: Bridging managed care and home community -based services

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content EVALUATION OF THE CARE ADVOCATE PROGRAM:
BRIDGING MANAGED CARE AND HOME AND COMMUNITY-BASED
SERVICES
by
George R. Shannon
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(GERONTOLOGY)
May 2004
Copyright 2004 George R. Shannon
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
UMI Number: 3140553
Copyright 2004 by
Shannon, George R.
All rights reserved.
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
UMI
UMI Microform 3140553
Copyright 2004 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
1 1
DEDICATION
I would like to dedicate this dissertation to my wife, Ellen, without whose
patience and understanding I could not have achieved what has been, for me, an
enlightening and humbling experience. I am certain that my parents would have been
very proud of this accomplishment. I owe much to them and I wish they were here to
share this with me. Certainly, their guidance eventually brought me to this place. I
am grateful to my daughters, Mary Ellen, Elizabeth, Margaret, and Catherine, whose
many accomplishments inspired me to persevere in the face of adversity.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Ill
ACKNOWLEDGEMENTS
Many persons have helped me through this dissertation and across seven years
in the Masters and Ph.D. programs at the Leonard Davis School of Gerontology. A
special thank you goes to the faculty, all of whom were incredibly kind, generous with
their time, and tolerant of my contentious approach to the learning experience. Donna
Benton presented the opportunity to participate in creating a study examining case
management skills and practices that served me well in all research projects that
followed. Thanks to the California Health care Foundation for funding the Care
Advocate Program and those wonderful annual meetings that have set the standards of
excellence in conference procedures. I would also like to thank the many individuals
at PacifiCare who were so willing to do whatever was required to complete the study.
In particular, Douglas Allen who created the CA Program and Lura Aheam who made
it work. Heartfelt thanks for your hard work and boundless energy.
Phoebe Leibig and Jon Pynoos were on my guidance committee. They have
provided both guidance and inspiration to me for as long as I have known them. A
special thank you goes to Judy Yip for her unending patience and for tolerating my
sense of humor. Sandy Atkins is a good ftiend, office “mom,” and one of the most
talented persons I have ever met. Thanks to comrades - Chris Kelley, Kristen Suthers,
Michele Maines, Ross Andel, Christy Matsuoka, Crystal Flynn-Longmire, Erica
Nielsen, and Aaron Hagedom for sharing this experience and being there whenever I
needed them. Thank you Betty Oswald - your organization skills and generosity
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
IV
proved invaluable to getting through quals. Jeff Hyde never let me down; “Thanks,
big fella’.”
Eileen Crimmins, Bob Myrtle, and Kate Wilber were the members of my
dissertation committee. Each of them made important contributions to this
dissertation. Eileen has the special gift of getting right to the heart of an issue and
does not “tolerate fools.” She extended herself many times to hasten my journey
through the Ph. D. program and I am very grateful. Her wisdom and sense of humor
are always apparent and I am grateful for many thoughtful insights. Bob and I were
bom within a couple of days of each other; Kate refers to us as the office twins.
Thank you. Bob, for being what I wish I could be - an intelligent, articulate, and
generous human being, whose knowledge of public policy and insights into problem
solving are awesome. Finally, Kate Wilber has mentored me, whether she knew it or
not, since I first took a class with her in the fall of 1997. She is one of those rare
teachers that can inspire a student to lift his or her consciousness to the next step - and
then the next step. I am grateful to have met Kate in this lifetime and I thank her for
accepting me into her world long enough to show me that this task could be
accomplished and that I could be more than I imagined.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
TABLE OF CONTENTS
Dedication............................................................................................................................ ii
Acknowledgements........................................................................................................... iii
List of Tables.....................................................................................................................vii
List of Figures.....................................................................................................................ix
Abstract.................................................................................................................................x
Chapter 1; The Care Advocate Program............................................................................1
Introduction..................................................................................................................... 1
Background..................................................................................................................... 1
The Care Advocate Model.............................................................................................6
Summary........................................................................................................................10
Contribution to the Literatiue......................................................................................11
Organization of Dissertation........................................................................................ 13
Chapter 2: Literature Review........................................................................................... 15
Introduction................................................................................................................... 15
Part 1: Health Care Programs for Older Americans.................................................. 16
Part 2: Long-Term Care and Home and Community-Based Services......................23
Part 3. Funding LTC/HCBS.........................................................................................31
Part 4: Care Advocate Program Outcomes.................................................................34
Summary.......................................................................................................................44
Chapter 3: Research Design and Methods.......................................................................45
Research Questions and Hypotheses..........................................................................45
Data Sources.................................................................................................................46
Research Design........................................................................................................... 48
Statistical Analysis....................................................................................................... 54
Study Limitations......................................................................................................... 59
Summary....................................................................................................................... 60
Chapter 4: Results..............................................................................................................61
Introduction...................................................................................................................61
Part 1: Findings Pre-Intervention to During Intervention..........................................61
Part 2: Findings Pre- to Post-Intervention..................................................................67
Part 3: Findings for Program Objectives....................................................................69
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
VI
Chapter 5: Propensity Scores............................................................................................ 79
Introduction...................................................................................................................79
Propensity Score Methods........................................................................................... 79
Propensity Score Results.............................................................................................. 81
Propensity Score Conclusions......................................................................................85
Chapter 6: Discussion........................................................................................................88
Part 1: Goals of the Intervention..................................................................................88
Part 2: Program Objectives..........................................................................................90
Part 3: Implications for health care organizations and considerations for future
studies...............................................................................................................92
References......................................................................................................................... 94
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
V ll
LIST OF TABLES
Table 1. Eight Categories of Referrals.............................................................................. 8
Table 2. Three Models of Care/Case Management (CM) by Orientation....................27
Table 3. Percent of Total Medicare Dollars Spent on Beneficiaries Aged 65+ in
Their Last Year of Life 1997-1999...................................................................38
Table 4. Average Medicare Spending by Type of Service and Decedent Status,
1997-1999.......................................................................................................... 39
Table 5. Algorithm to Determine Program Eligibility...................................................48
Table 6. Factor Matrix for Components of Satisfaction (N= 733)................................ 57
Table 7. Characteristics of ITT/Control Groups for 12 months Pre-Intervention.......62
Table 8. Differences in Member Retention During 12-Month Intervention................ 64
Table 9. Differences in Member Retention by CA Program and Control....................65
Table 10. Multinomial Logistic Regressions: Utilization Change Scores Pre- to
During Intervention............................................................................................66
Table 11. Logistic Regression: Influence of Satisfaction on Member Retention.......... 67
Table 12. Differences in Retention by Study Groups: ITT & CAl versus Control.......68
Table 13. Multinomial Logistic Regressions: Utilization Change Scores Pre- to
Post-Intervention................................................................................................71
Table 14. Baseline Comparison of 85+ Group and Rest of Target Population............. 72
Table 15. Logistic Regression of Eligibles Age 85+ vs. Rest of Study Population.......73
Table 16. Variations in Medical Group and Health Plan Services Utilization.............. 77
Table 17. Comparisons of Members Referred to Medical Groups or Not......................78
Table 18. Ten Diagnoses Predicting High Cost and Hospital Utilization......................80
Table 19. Baseline Comparisons of Intent-To-Treat and Propensity (PPS) Groups.....82
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
V lll
Table 20. Differences in Member Retention by Study Group: During Intervention.....83
Table 21. Differences in Member Retention by Study Group: Post-Intervention......... 83
Table 22. Multinomial Logistic Regressions: Utilization Change Scores Pre- to
During Intervention............................................................................................84
Table 23. Multinomial Logistic Regressions: Utilization Change Scores Pre- to
Post-Intervention................................................................................................86
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
IX
LIST OF FIGURES
Figure 1. Payments Sources: Medical Services for Medicare Beneficiaries, 1999......19
Figure 2. Health Care Expenditures for Medicare Beneficiaries, 1999........................31
Figure 3. Analytic Model: Linking Managed Care and HCBS..................................... 50
Figure 4. Satisfaction Pre-Intervention to After-Intervention: ITT/Control................ 63
Figure 5. Intervention Process..........................................................................................75
Figure 6. Variations in Non-Insured Home and Community-Based Services............. 76
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
X
ABSTRACT
The evaluation of the Care Advoeate Program tested the efficacy of linking
medical and community services. Member satisfaction, member retention, and
utilization of insured medical services were tracked for three 12-month periods (pre-,
during, and post-intervention). Using an algorithm designed to target frail, high-cost
users of Medicare insured health care services, the Care Advocate Program partnered
PacifiCare’s Secure Horizons and four of its medical groups with two social service
organizations.
PacifiCare members aged 69 to 96 years, receiving care from four PacifiCare-
contracted medical groups, were randomized to intent-to-treat (ITT, N==389) and
control (N=434) groups. The twelve-month intervention provided telephone
assessments, links to six types of home and community-based services, and monthly
follow-up contacts. In addition, care advocates (CAs) made referrals back to medical
groups and the health plan, when appropriate.
During the 12-month intervention period, ITT members died at a rate
significantly lower than control members, were significantly more likely to have
increased utilization of PCP and specialist services, and were significantly more likely
than control group members to have decreased hospital admissions during the
intervention. There were no significant changes in utilization patterns post-
intervention. Those who used referrals to medical groups during the intervention did
not use more physician services (PCP and specialist), but did use more hospital
services than those who did not use medical group referrals.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CHAPTER 1: THE CARE ADVOCATE PROGRAM
Introduction
The Care Advoeate Program was a demonstration program designed to
determine whether linking high-risk Medicare-eligible members of a major California-
based health maintenance organization to existing home and community-based
services (HCBS) would reduce utilization of costly, insured services, improve member
satisfaction, and increase member retention. Partners included a Medicare-risk HMO
(PacifiCare’s Secure Horizons), foiu* contracted medical groups (Cedars Sinai
Medical Network, Harriman Jones Medical Group, Health Care Partners/United
Physicians of South Bay, and Talbert Medical Group), two home and community-
based social service agencies [Jewish Family Service of Los Angeles (JFS) and
Jewish Family and Children’s Services of Long Beach/West Orange County (JFCS)],
and evaluators from the University of Southem California. Care advocates were three
masters-level case managers, trained and supervised by staff from the two social
service partners and housed within the social service agencies.
Background
The Problem
In the current U.S. health care system, efforts to meet the rapidly escalating
needs of an aging population have been challenged by market exigencies such as
retum-on-investment (ROI) and availability of health service resources. The
inevitable clash of cost-containment strategies with quality and access to care has
resulted in an increasingly unstable health care environment (Lesser and Ginsburg,
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2000; Liberman and Rotarius, 2001; Sandy, 2002). This has led to spiraling costs in
the provision of health care services to older Americans over the last two decades,
with no end in sight (Reuben et al., 2003; Spillman and Lubitz, 2000).
Medicare spending will likely double as a share of the economy by 2050 and
government nursing home expenditures could triple in the same period (Cutler, 1998;
Vladeck, 1997). By 2030, it is estimated that older Americans will comprise 20% of
the population and account for 50% of the medical costs (Rice, 1996). Between 1987
and 1995, for example. Medicare payments to health care providers increased between
four and five percent per year, adjusting for inflation. During the same period, the
gross domestic product increased by about 1.2%.
If current trends hold, the cost of health care in the year 2020 (in 1995 dollars)
will be approximately $25,000 per person compared with $9,200 in 1995 (Fuchs,
1998; Iglehart, 1999). Over the same period, (1995-2020), the population of the
United States, aged 65 or more, is expected to double (Centers for Medicare and
Medicaid, Services, 2002). Further, by 2050, the percentage of elders over 85 is
expected to triple (Siegel, 1993).
Experts in demography (Crimmins, Saito, and Reynolds, 1997) and heath care
service delivery (Kunkel and Applebaum, 1992) have expressed concerns that longer
lives may mean greater susceptibility to disease, increased comorbidity, greater
impairment, functional limitations, and disability in old age. Although age is not the
cause of illness and disease, there is, nevertheless, an association between age and the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
prevalence of disease and disability (Verbrugge, 1989; Crimmins, 1996; Crimmins
and Saito, 1993).
As the population ages, even if disability rates decline, and there are
indications that this may be occurring (Crimmins and Saito, 2001), there likely will be
more, in absolute numbers, older and sicker individuals than ever before. As our
ability to extend the lives of those with chronic illnesses improves, the sheer number
of frail elders using health care resources will increase dramatically. Ironically, in the
current system those who need health care services most, frail older adults and
disabled persons, face a number of serious challenges to receiving adequate care.
These challenges include overcoming fragmentation of service delivery mechanisms,
lack of adequate chronic and long-term care (LTC) services, and relatively few
preventive care programs (Boult et al., 2000; Kane, Kane, and Ladd, 1998). This
speaks to the heart of the Medicare problem; that is. Medicare is an acute-care model
lacking chronic and LTC capabilities, fraught with spiraling costs and concomitant
limitations in access to health care services (Fuchs, 1998; Kane, 1998).
The Promise of the Future
Medicare managed care
The integration of chronic and long-term care in Medicare managed care
systems remains problematical, despite changes in payment structure (e.g., capitated
payments) and innovative program models at the state level. There is some evidence
of success with integration efforts. Examples of integrated models that have achieved
some success with integrating health care for older adults include programs like the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Program for All-Inclusive Care for the Elderly (PACE), the Minnesota Long-Term
Care Project (MLTCP), the EverCare Program, and Social HMOs. These
demonstration programs provide cost-effective, integrated health care services to older
adults (Harrington and Newcomer, 1991) (Kane, 1997; Kane, 1998; Kane and Huck,
2000; Kane, Kane, and Finch, 1995). The PACE program, for example, has
successfully integrated acute, chronic, and long-term care, while improving overall
health status and quality of life for participants, targeting frail, nursing home eligible
individuals, most of whom qualify for both Medicare and Medicaid reimbursements.
The MLTCP targets a more diverse population, including not only nursing home
eligible frail elders, but also those who are “well” and living in the community. The
social HMO, designed to offer supplemental benefits to Medicare HMO chronic care
clients, emphasizes the provision of added health care benefits (prescription drugs, eye
glasses, etc.) to Medicare-eligible older adults (Harrington, Lynch, and Newcomer,
1993; Harrington and Newcomer, 1991).
It appears that the U. S. health care system may be evolving towards greater
utilization of HCBS. A recent study (Mollica, 2003)described efforts to shift the
emphasis of LTC spending from institutional to HCBS. Nursing facility occupancy
rates declined by about six percent from 1985 to 1996 (Strahan, 1997), supporting the
assertion that the balance is shifting toward greater reliance on HCBS. The drop in
nursing facility occupancy rates was attributed by Mollica to the overall effectiveness
of HCBS in providing personal services as alternatives to nursing homes for those
with functional impairments. In addition, modified state regulations have provided
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
incentives to test new models of care that encourage privacy, aging in place (Meiners,
1996), consumer-centered care (Stone, 2000) and targeting older adults at highest risk
for adverse, costly outcomes (Weissert, Hirth, Chemew, Diwan, and Kim, 2003).
Policy experts (Doty, 2002; Feldman and Kane, 2003) have expressed some
reservations about the cost-effectiveness HCBS programs. On the other hand, other
research (Boult, Boult, and Pacala, 1998; Shapiro and Taylor, 2002) (Stuck et al.,
1995; Wheatley, DeJong, and Sutton, 1997) suggests that timely HCBS interventions
can improve health related quality-of-life, cut hospitalization admissions and length of
stay, decrease the number emergency room visits, and reduce mortality. This is
particularly true for the 5-10% of high-risk older persons who incur an inordinately
high share of costs (Centers for Medicare & Medicaid Services, 2004). The success of
the demonstration programs in integrating long-term care and providing increased
benefits to Medicare populations suggests that improved health outcomes and greater
cost-effectiveness may be achievable goals (Centers for Medicare & Medicaid
Services, 1998; Meiners, 1996).
Citing the need for more cost-effective service delivery options, the Centers for
Medicare and Medieaid Services (CMS), the government agency charged with
administrating the Medicare and Medicaid programs, has increasingly foeused on
prospective payments for health care providers and capitated payments to Medicare
managed care organizations (MMCOs), rather than cost-based fee-for-service
reimbursements (Centers for Medicare & Medicaid Services, 2002; Scully, 2001).
Capitated payments offer incentives to cut costs. Medicare managed care offers the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
promise of increased cost-effectiveness and greater access to care by using its
organizational capabilities to integrate acute, chronic and long-term care services in
concert with current successes in monitoring service utilization and levels of health
status (Boult et al., 1998; Boult et al., 2001; Boult, 2000; Kane, 1998).
Consumer choice
The Care Advocate Program is a consumer choice model of care. In referring
HCBS services to members, the care advocates first ascertained which services
members preferred and incorporated these choices into individual care plans.
Consumer choice and autonomy for older adults in the U. S. health care system have
been topics of considerable research since the late 1980s (Davies and Ware, 1988;
Davis, Collins, Schoen, and Morris, 1995). Efforts to preserve autonomy and choice
in health care decision-making have helped move the continuum of health care from a
predominantly medical model to a more balanced medical-social model.
For Medicare managed care enrollees, consumer choice also means choice in
selecting health plans. Medicare beneficiaries have the right to change health plans or
return to traditional fee-for-service (FFS) Medicare during annual open enrollment
periods. Having the choice to exit or switch health plans presents an active process for
managed care members to express perceived dissatisfaction with providers or
treatment practiees.
The Care Advocate Model
The Care Advocate Program addressed the coordination of acute, chronic, and
long-term care services by bridging managed acute care to home and eommunity-
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
based services through an information and referral case management intervention.
Care advocates provided referral information and direct linkages to eight categories of
service. Table 1 displays the eight referral categories employed by the care advocates
(the first six are HCBS categories and related sub-categories, the last two are referrals
to medical groups and health plan).
Unfortunately, many older adults with unmet needs are unaware of the range of
community resources available to them (Aliotta, Clarke, and Paulman, 1998; Mebane,
2001). Research goals of the Care Advocate Program were to determine if systematic,
mutually agreed-upon case manager referrals to HCBS could positively influence
member satisfaction and member retention, while reducing high-cost insured service
use, such as hospital admissions, hospital days, and emergency room encounters.
Most referred home and community-based services were provided outside of
the Medicare managed care plan, which focuses on medical treatment. HCBS
concentrate on: 1) disease and injury prevention (nutrition programs, home
modifications, home safety, and adaptive equipment); 2) linkages to community
services (commercial vouchers, escort, and volunteer transport); and 3) supportive
services (case management, personal care services, chore services, day care, and
respite for caregivers) (Booth, 1997). The care advocates also assisted members in
finding sources of payment for the recommended HCBS in their care plans.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 1. Eight Categories of Referrals
1. In-Home Care
- Home Chores (Laundry, Shopping, Cleaning,
Meal Preparation, Medication Management)
- Personal Care (Bathing, Grooming, Toileting,
Incontinence Care, 24-Hour Supervision)
2. Nutrition
- Congregate Meals
- Grocery Delivery
- Home-Delivered Meals
- Nutritional Supplements
3. Home Safety
- Emergency Response System
- Alzheimer’s Association Safe Return
- Medic-Alert
- Telephone Reassurance
- Uninsured Durable Medical Equipment
- Bath Safety Equipment
- Rearranging Home To Improve Safety
- Outdoor Safety Equipment
4. Transportation
- Taxi Voucher
- Curb-to-Curb Transport
- Escort
- Volunteer Transport
- DMV Disability Placard
5. Adaptive Equipment
- Hearing Adaptive Phone
- Hearing Aids
- Incontinence Pads
- Visual Aids
- Wheelchair Pads
- Diabetic Supplies
- Home Therapy Supplies
6. Supportive Services
- Case Management
- Money Management
- Legal Referrals
- Advanced Directives
- Housing Referrals
- Adult Day Health Care
- Support Groups
- Advocacy Groups
- Community Groups
- Recreation
- Exercise
- Adult Protective Services
- Employment
- Volunteering
- Financial Assistance
- Counseling
- Dental Referrals
- Information & Referral
- Friendly Visiting
- Medicare
- Burial Info
- Medi-Cal Eligibility Assistance
7. Medical Services
- Primary Care Physician
- Case Management at Medical Group
- Medical Specialists
- Skilled Nursing Facility
- Hospice
- Respite Care
- Physical Therapy
- Occupational Therapy
8. Member Services
- Any Benefit-Related Questions.
- Advocacy Services
Goals of the Care Advocate Program
1. To increase member satisfaetion by linking older adults enrolled in a Medicare
managed care plan to community-based serviees, and provide referrals to needed
serviees with monthly follow-up calls.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2. To improve member retention by increasing member satisfaction.
3. To decrease overall utilization of costly medical services including: Primary care
physician visits, specialist visits, hospital admissions, hospital days, and
emergency room visits.
Objectives
1. To create a screening mechanism useful for identifying frail Medicare HMO
members.
2. To establish a linkage for members to access HCBS.
3. To create a practical database tool for making referrals, following up, and
collecting referral, compliance, and other data.
4. To demonstrate that linking medical and home and community-based services
can occur without major impact on medical group physicians or resources.
Structure
Three full-time, grant-funded care advocates were employed by and housed
within the HCBS social services agencies (JFS & JFCS). Each care advocate had a
master’s degree in social work or family counseling and prior experience working
with older adults. In addition to the care advocates, several individuals from
PacifiCare functioned as coordination managers. These included the principal
investigator (5% time) who provided overall direction, a program manager (50% time)
and a project coordinator (90% time).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
10
PacifiCare’s contribution was largely administrative; staff contributed about
20% of their time addressing research components of the project. For example, the
program manager served as a liaison to all partners, was responsible for project
oversight and budget management, and ensured that research procedures met federal
guidelines regarding anonymity and other legal protections. The project coordinator’s
duties included assisting with the budget and serving as the liaison to the care
advocates. She also maintained the projeet’s database and coordinated team meetings.
The evaluation team from the Andrus Gerontology Center of the University of
Southem Califomia (USC) provided consultation on study design principles and
recommended procedures for data analysis. (Wilber, Allen, Shannon, and Alongi,
2003).
Summary
The U. S. health care system faces many challenges in the coming decades.
These include a rapidly increasing demand for health care services, the need for
innovative chronic and long-term care models, increasing demand for eonsumer
choice, and the need for greater cost-effectiveness in the provision of health care
services to aged and disabled populations. Some practitioners and researchers in the
field of aging believe that the potential of Medicare managed care to improve cost-
containment and access to health care serviees holds promise for the future. Part o f
the solution to current fiscal problems may lie in shifting the focus of long-term care
from institutionalization to HCBS, when appropriate.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
11
The Care Advocate Program built on previous studies that sought to increase
the range of health care services and improve the health of frail older adult
populations, while reducing the costs associated with health outcomes (Bemabei et al.,
1998; Booth, 1997; Boult, Rassen, Rassen, Moore, and Robison, 2000; Dunn, Sohl-
Kreiger, and Marx, 2001; Stanton, Walizer, Graham, and Keppel, 2000). In addition,
it was predicated on the belief that improving access to HCBS would reduce
fragmentation in the delivery of health care services. It used the relatively simple
approach of bridging existing insured, case management services to non-insured,
home and community-based services.
The goals of the CA Program were to increase plan satisfaction, improve
member retention, and reduce in utilization of high cost, insured medical services by
offering information and referrals to HCBS. If the Care Advocate Program is proved
to be cost-effective and outcomes are positive, other HMOs may consider similar
programs. In that case, this demonstration program will have provided a framework
for useful collaboration between managed care home and community-based services,
one that could enhance health care practices and quality of life for frail, older adult
populations.
Contribution to the Literature
The Care Advocate Program attempted to find a workable approach to linking
the often-polarized interests and incentives of managed health care and HCBS.
Medicare managed care policies are driven by bottom-line considerations (optimum
care provided in the most cost-effective manner); whereas, HCBS focus on individual
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
12
gains through social programs to maintain individuals’ health and well being at
optimal levels of independence. Attempts to link the macro-level goals of managed
care with the micro-level goals of HCBS offer the possibility of providing greater
access to health care services more efficiently and with greater cost-effectiveness than
the current, fragmented system.
There is growing support for the notion that managed care holds promise for
the future for the U.S health care system. Current estimates are that by 2020, 5 million
more persons may require long-term care than did in 1990 (Administration on Aging,
2002) and there has been a gradual shift in chronic and long-term care from
institutions to home and eommunity-based services (Rantz, 2000). The Care Advocate
Program embraces these ideas by attempting to demonstrate that linking these diverse
organizational types, currently co-existing within our health care delivery system, is
both possible and advantageous.
Finally, the Care Advocate Program promoted consumer choice, a concept first
articulated by disabled persons who compellingly expressed their desire to choose
how, when, and where they were to receive health care and personal services
(Squillace, 2002). For older adults, consumer choice presents greater opportunities for
autonomy and self-determination in planning which health care services they are to
receive and where they are to live (Meiners, Mahoney, Shoop, and Squillace, 2002).
The consumer choice model applied in the Care Advocate Program encourages frail
older adults to voice their preferences for referrals to services they may need and
respects their choiee to use them or not.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
13
Organization of Dissertation
■ Chapter 1 has presented the background, problem statement, description of the
Care Advocate Program, and its relevance to health care research.
■ Chapter 2 is a review of the literature relating to the Medicare program.
Medicare Managed Care, LTC and HCBS, federal funding programs, and goals
of the Care Advocate Program. Chapter 2 has four parts:
1. A brief summary of the Medicare Program and the evolution of
Medicare health maintenance organizations (HMOs).
2. An examination of the roles of home and community-based services,
case management models, consumer-directed care, and consumer
choice in the provision of long-term care to disabled or frail
individuals.
3. A description of financing of long-term care and home and community-
based services, including Medicare, Medicaid, the Older Americans
Act, and Social Services Block Grants.
4. A review of literature relevant to Care Advocate Program outcomes
(member satisfaction, member retention, and utilization of insured
health care services).
■ Chapter 3 presents the research design and methods including:
1. Research questions and hypotheses
2. Data sources
3. Research design model
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
14
4. Statistical analyses
5. Study limitations
Chapter 4 presents the results of the study in three parts. Part 1 presents findings
from pre- to during intervention, Part 2 presents findings from pre- to post
intervention, and Part 3 presents findings for program objectives.
Chapter 5 presents a methodology to compare intent-to-treat and control groups.
This chapter delineates the rationale, methodology, results and conclusions drawn
from this approach.
Chapter 6 discusses study findings, lessons learned, and implications for health
care organizations and future research.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
15
CHAPTER 2: LITERATURE REVIEW
Introduction
Chapter 2 is a review of current literature: Part 1 is a brief summary of the
Medicare Program, recounting the evolution of Medicare managed care in the
twentieth century from physician group practices to independent practice associations
(IP As) and Medicare managed care organizations. Part 2 examines the roles of home
and community-based services, including case management models, consumer-
directed care, and the role of consumer choice in providing social support and care
planning to frail older adults, with the goal of minimizing fragmentation in the
delivery of health care services.
Part 3 describes the financing of long-term care (LTC) and home and
community-based services (HCBS) by federal health care programs for older adults
enacted by Congress and first signed into law by President Johnson as part of his
“Great Society.” These include Medicare, Medicaid, the Older Americans Act, and
Social Services Block Grants. Medicaid, the federal/state means-tested program
designed to assist pregnant women, adults in families with dependent children,
individuals with disabilities, and persons 65 or over, is not discussed at length because
only 2.8 % of those who received the Care Advocate Program intervention (N = 23)
qualified for the M edicaid (Medi-Cal) Program. Part 4 provides a review o f studies
that relate to program outcomes: satisfaction, retention, and utilization of insured
health care services.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
16
Part 1: Health Care Programs for Older Americans
Background
Only six decades ago, over half of the older adults in this country were poor.
Many depended almost wholly on their children for support. As late as 1965, when
Congress enacted Medicare, Title XVIII of the Social Security Act, only about half of
United States residents aged 65 or older had health insurance, spending an average of
19% of their income on health care. Yet, 97% of older Americans had health care
coverage hy 1970, and average out-of-pocket expenses had declined by 50%. Still,
the news was not all positive. At the same time, utilization of health care services and
related expenditures were in the steady upward spiral that has characterized the
Medicare program since its inception (Moon, 2001).
Those involved in conceiving and shaping the original Medicare legislation,
planned it to be only the first step towards making health care universally accessible to
older adults. This was to he offered, not as charity or “socialized medicine,” hut
rather, as an entitlement to Americans who had earned the right to health care
coverage during their working years. There were several key elements missing from
the final draft of the Medicare program that continue to negatively affect the
circumstances of older adults, particularly prescription drugs and chronic or long-term
care (Marmor, 1988). Over time, these omissions created a widening chasm between
the aims of the program and the needs of beneficiaries.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
17
The Medicare Program
In 1965, Congress enacted Medicare, Title XVIII of the Social Security Act, to
provide federally funded medical benefits, primarily acute care, to people aged 65 and
over and to certain disabled individuals, regardless of financial need. The Medicare
Program health care coverage divides into two distinct areas:
1. Medicare Part A applies to people with long-term disabilities and those eligible
for Social Security at age 65. It covers hospital, post-hospital, and other in
patient services through mandatory payroll tax contributions by employers,
employees, and the self-employed to the Hospital Insurance Trust Fund.
2. Medicare Part B, a voluntary benefit, is funded through general tax revenues,
monthly premiums and co-payments. It is modeled after traditional cost-based
indemnity insurance programs and covers physician services (in both hospital
and non-hospital settings) and services furnished by certain other medical
practitioners. It also covers clinical laboratory and other diagnostic tests,
durable medical equipment, most medical supplies, ambulance services,
prescription drugs that cannot be self-administered, certain self-administered
anticancer drugs, certain other therapy and health services, and blood not
supplied by Part A.
In addition. Medicare Part C, or Medicare+Choice (M+C), introduced in the
Balanced Budget Act of 1997 (BBA 97), presented new types of managed care
organizations (MCOs) as options to traditional, fee-for-service (FFS) Medicare risk-
contracted health maintenance organizations (HMOs). The choices available.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
18
including Medical Savings Accounts (MSAs), provider sponsored organizations
(PSOs), and private fee-for-service (FFS) plans, vary widely in their appeal to
prospective enrollees. While M-i-C created choice for beneficiaries, most of these
options have gotten off to a slow start (Centers for Medicare & Medicaid Services,
2002).
Many older adults require chronic or long-term care in addition to acute care.
Although it is a purchaser of some skilled nursing facility (SNF) care and home health
services. Medicare's interest in these services is to aid beneficiaries in recovering from
acute care episodes, not for chronic care. Medicare covers only the most intensive
level of skilled nursing facility (SNF) care, following a hospitalization stay of at least
three days. Care in a SNF is limited to 100 days, for one episode of care. Medicare
home health care is available only for rehabilitative or convalescent service prescribed
by a physician, not for on-going or custodial care. Medicare-covered services are
available only on an episodic basis and they are not for long-term care (Pierce, 1987).
Figure 1 depicts various funding sources for medical services in the U.S. health
care system. In 1999, Medicare (both FFS and Medicare MCOs) paid a little more
than half of the total cost of all medical care (53%) provided for Medicare
beneficiaries, or $5,043 per beneficiary. Out-of-pocket costs followed at 19% ($1825
per beneficiary (not including premiums for Medicare Part B or MCOs). Private
insurance paid 12%, or $1,161 per beneficiary.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
19
Figure 1. Payments Sources: Medical Services for Medicare Beneficiaries, 1999
Medicaid
12%
FVivate
Insurance
12%
Other
Out-of-
- Backet*
19%
Medicare
52%
Source: CMS, Office o f Research, Development, and Information (1999)
Traditional Fee-Far Service (FFS)
In traditional, fee-for-service Medicare, the patient is reimbursed or
indemnified for the cost of the health services rendered based on “usual and
customary” fees in the service area (Dudley and Luflt, 2001; Tabbush and Swanson,
1996). Under FFS, provider incentives are to encourage service utilization and
consumer incentives are to utilize available health services, regardless of efficacy or
expense. This is sometimes called a supplier-induced market, where sellers of goods
and services help create the demand for them (Tufts Managed Care Institute, 2001).
So long as costs were recaptured by reimbursements, providers promoted the latest
technologies and tests, thereby creating an expansionist environment that encouraged
the building o f new facilities to provide even more services. The inevitable outcome
of this cost structure was spiraling costs. The Medicare FFS payment structure lasted
for more than 25 years before market changes and rapidly increasing costs demanded a
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 0
new methodology for reimbursing health care providers (Tufts Managed Care
Institute, 2001) (Zelman and Berensen, 1998).
Medicare Managed Care
The notion of managed care can be traced all the way back to medieval
German guilds of craftsman and miners who paid fixed monthly fees, essentially
capitated rates, in return for the offered health services (Reagan, 1996). Moving ahead
to 1933, a small physician group practice provided medical care for 5,000 workers on
an aqueduct construction project in the United States. Workman’s compensation
insurance companies paid this group a set fee, derived from their premiums, for
accident cases. Workers paid five cents, as an out of pocket expense, for other medical
services. Five years later Henry Kaiser hired the same physician group to provide
prepaid medical care for workers building the Grand Coulee Dam (Firshein, 2000).
By 1965, when Congress enacted the Medicare Program, not only were pre
paid physician groups and Kaiser health care plans in existence, but also, the first
Independent Practice Association (IPA) model was introduced in California’s San
Joaquin Valley. Until the early 1970s, however, health maintenance organizations
(HMOs) did not exist. The Medicare provisions of the 1972 amendments to the Social
Security Act defined the term HMD and introduced the concept of Medicare HMD
enrollment for the provision of health care services (Tufts Managed Care Institute,
2000). These provisions transformed the traditional reimbursement scheme of fee-for-
service (FFS), creating a new payment mechanism to reimburse HMOs. This new
payment structure was termed capitation, a prepaid concept that was the precursor to
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
21
Medicare MCO full-risk contracts, as set in place by the Tax Equity and Fiscal
Responsibility Act of 1982.
TEFRA ‘ 82
The 1972 amendments to the Social Security Act paved the way for Medicare
demonstration projects offering prospective, in place of cost-based retrospective,
payments for insured services. Full risk-contracting options for MCOs were not
available, however, until the Tax Equity and Fiscal Responsibility Act of 1982
(TEFRA ‘82). Implementation took almost three years, until the publication of the
final regulations in January 1985. TEFRA ’82 proposed full risk-sharing contracts
between managed care organizations (MCOs) and Medicare. The costs of providing
services by contracting MCOs were set using the adjusted average per capita cost
(AAPCC) - the estimate of costs that would have been incurred for the same services
under FFS. The AAPCC was computed for each county of the United States.
Accordingly, Medicare MCO Plans were to receive 95 % of the AAPCC rates in
return for accepting the full risk for covered medical services (Centers for Medicare &
Medicaid Services, 2002; Luft, 1999).
Capitation
Capitation is a prospective payment mechanism that represents a projection of
how much it will cost to care for person or group of people over a given period of
time. An insurer and a provider of health care services agree to the capitation rate. A
provider such as physician, medical group, or hospital, signs a contract with a private
insurer or a Medicare MCO, not only to provide care but also to manage it. Under
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 2
such an arrangement, an MCO, typically an HMD accepts the risk associated with per-
member-per-month prepayments. If the agreed upon capitated fee is too low the MCO
loses, if capitated fee is too high the MCO wins; however, there is nothing in the
contract that stipulates who must absorb the risk. The assumption of capitation is that
MCOs will allocate care and resources so that, on average, it will take in more in
capitation payments than it will spend providing care. Since the capitation rates are
set below the amount the average FFS consumer spends, the Medicare program should
also do better. The potential for cost savings was not overlooked by public and private
insurers: Between 1987 and 1995, the number of Medicare beneficiaries whose health
care was paid for using capitated payments almost tripled (Berwick, 1996).
Risk Adjusted Payments
An unanticipated consequence of capitation was that MCOs actively sought
out and enrolled the healthiest benefieiaries (Gage, 1999). In January 2000, CMS
began a five-year phase-in of risk-adjusted payment methods for Medicare MCOs to
provide additional payments to MCOs for the health care needs of the sickest
beneficiaries. The Medicare-risk adjustment methodology, by adding case-mix
adjustors, promoted equal access to managed care organizations by giving MCOs and
other managed care providers greater incentive to enroll beneficiaries with chronic
health care needs (Centers for Medieare & Medicaid Services, 2002).
Contractual Arrangements
There are as many as three tiers of contract arrangements between Medicare,
MCOs, and health care providers. In the first tier. Medicare pays a capitated payment
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
23
to MCOs, which then assume 100% of the liahility or risk for the health care costs of
the patients enrolled in their plan. MCOs may pay physicians directly or payments
may be made to intervening entities like Independent Practitioner Associations (the
second tier). In three-tiered arrangements, MCOs may pass on some or all of the risk
to medical groups, which may, in turn, pass risk to their physicians. Alternatively,
these medical groups may compensate physicians through salaries or hy making fee-
for-service payments (Kerr, 1995).
Part 2: Long-Term Care and Home and Community-Based Services
Long-term care (LTC) may be defined as assistance given over a sustained
period to those who are experiencing chronic limitations in functioning because of a
poor health or a disability. LTC is not age-restricted; children and adults with
disabilities use LTC. In 1999, as many people over age 65 as under age 65 required
LTC services (Tilley, 2001). LTC services are delivered across multiple settings - in
hospitals, nursing facilities, residential facilities, day care centers, and private homes
(Kane, 1994; Kane, Kane, Ladd, and Veazie, 1998).
Home and community-based services complement medical services hy
providing, LTC services, including social support and care coordination, to persons
who have lost some capacity for self-care due to a chronic illness or condition
(Congressional Budget Office, 1995). Home and community-based services, in
addition to social support and care coordination, include personal care and assistance,
home health, adult day care, respite services, and assisted living facilities.
Institutional services (skilled nursing facilities and intermediate care facilities for the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
24
developmentally disabled) provide additional long-term care services to those who
require 24-hour custodial care (Mollica, 2003; Weiner, 2002).
It is important to note that most older adults are disability-free and do not
require long-term care. Still, in 1995 approximately 40 million people in the United
States were disabled, about 12.8 million of whom required LTC. Of these, 5.1 million
were severely disabled, that is, in need of intensive assistance with activities daily of
living (ADLs) and/or instrumental activities of daily living (lADLs) (Crimmins and
Saito, 2001; Crimmins et ah, 1997). Functional impairment can occur at any age, but
the likelihood of experiencing some functional disability increases with age. Chronic
conditions that may result in the need for LTC include heart disease, cancer, stroke,
chronic obstructive pulmonary disease (COPD), fractures, arthritis, diabetes,
hyperplasia of prostate, kidney disease, and hypertension (Blaum, Liang, and Liu,
1994; Blumberg, 1999; Haan, 1997; Long and Marshall, 1999).
Older persons prefer to “age in place,” that is, receive health care services
where they already live. Home and community-based services, either formal (paid) or
informal (fnends and family), are one means to accomplish this goal; however, to
assure access to needed health care services, frail older adults may need additional
support to overcome the fragmentation in our health care system (Mollica, 2003;
Shapiro and Taylor, 2002). Medicare covers only medical services, not long-term
supports, which are generally available only on a private pay or privately insured
basis, or through Medicaid for the low-income population. One way for Medicare to
improve service delivery to functionally impaired adults would be to fimd long-term
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
25
care programs that preserve autonomy and foster independence as long as possible; for
example through case management (Fox, Etheredge, and Jones, 1998; Mollica, 2003).
This could prove to be cost-effective to the federal government because HCBS,
particularly those services targeted to high-risk populations, can decrease utilization of
high-cost healthcare services (Landi, 1999; Stewart, Pearson, and Horowitz, 1998) and
significantly lower mortality rates (Boult et al., 1994; Shapiro and Taylor, 2002).
Case Management
Some, though not all, frail and disabled people require assistance to determine
what services they need, to arrange for those services, and to ensure that they are of
high quality and provided in a timely and reliable way. This is the function of case or
care management. “Case” management (medical model care and administrative
responsibilities) has been differentiated from “care” management (social model care
planning and coordination) (Scharlach, Giunta, Mills-Dick, and Taylor, 2001);
however, inasmuch as the literature frequently uses these terms interchangeably, to
avoid confusion this evaluation made no distinction between the terms. The term case
management first appeared in the social welfare and public health nursing literature in
the early 1970s, although the practice of care coordination and management of health
care services was anticipated as early as the 1860s with the establishment of the Board
of Charities in Massachusetts (Tahan, 1998). Case management may be defined as “a
collaborative process which assesses, plans, implements, coordinates, monitors, and
evaluates options and services to meet an individual’s health needs through
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
26
communication and available resources to promote quality, cost-effective outcomes”
(Andolina, 2001; Zander, 2002).
Although the essential functions of case management are widely accepted, the
goals of case management may differ greatly (Applehaum and Austin, 1990). Two
systems of case management may he distinguished: payer-side and provider-side.
Payer-side case management emphasizes eost-containment, that is, balancing cost of
service provision with quality of service. The primary function of payer-side case
managers is to avoid providing unnecessary services. Provider-side case management,
on the other hand, promotes consumer advocacy and balances cost and quality of care
at the direct service level (Andolina, 2001).
The goals of case management may he consumer and/or system oriented.
Consumer oriented case management focuses on improving functional capacity,
increasing or maintaining quality of life, and determining the appropriate care setting
for clients and improve access to the continuum of care, caregiver support, and bridge
the gap between institutional and community-based services. Systems oriented goals
aim at heightened cost-effectiveness and quality of care through greater system
efficiency, improved service accessibility, and increased consumer direction in health
care planning and implementation (Huber, 2001; Scharlach, 2001).
Figure 2 presents three models of case management. Case management may be
social, medical, or integrated in orientation. Social case management interventions
involve screening, assessment, care planning, providing information and referrals of
clients to medical and community-based health care services, and follow-up to ensure
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
27
utilization of referred services (Schifalacqua, Hook, O'Heam, and Schmidt, 2000;
Zander, 2002). Medical case management is for people with chronic conditions,
assisting them to adapt their lifestyles to their condition, ensuring ongoing symptom
monitoring, giving feedback to primary care providers, and making referrals to
community-based services. Trained nurses, mental health professionals, and social
workers provide most case management services, social workers, as in the Care
Advocate Program, more often serve as case managers in home and community-based
practice (Quinn, 1993; Scharlach et al., 2001; Scharlach, 2001; Tahan, 1998).
Table 2. Three Models of Care/Case Management (CM) by Orientation
Orientation Brokerage
Models
(Provider-side)
Managed Care
Models
(Payer-side)
Integrated
Models
Organizational Case manager not
affiliated
Links to HCBS agencies
Limited access to
medical providers
Case manager linked to
system service
providers
Administrative No risk unless capitated At-risk, linked to
capitated system
Cost-control and
gatekeeping
Partial-risk,
CM part o f team
decision-making
Care/Case Manager Assessment, care plan
Some brokering of
services, if needed
Assess, care plan,
implement and
monitor
Integrates all
components of CM
Consumer High functioning client
Information and referrals
Consumer direction and
choice model
Self-advocacy
CM facilitates,
monitors social
support and HCBS
High-risk client,
CM counsel,
implement and
monitor
Adapted from: (Scharlach et al., 2001)
Models of case management
Brokerage model
Characteristics of brokerage models include: 1) Linkages - case manager links
clients with HCBS, 2) Lower service intensity - clients often require less intensive
services, 3) Larger caseloads - as many as 60 or more, 4) No financial risk assumed by
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
28
case management agency 5) Clients are mostly higher functioning/lower risk of
needing institutional care (Huber, 2000; Scharlach, 2001).
Managed care model
In the managed care, or payer-side model, the responsibilities and financial
risks are increased: 1) Case managers are either gatekeepers or utilization review
managers; 2) Case management is frequently subcontracted and capitated, placing the
case management agency at financial risk for service provision; 3) Medical services
are usually referred back to health plan-contracted providers; 4) Social services are
frequently referred to community-based agencies, 5) Case managers may have higher
level of authority and sometimes have smaller caseloads, 30-60, than in brokerage
models (Huber, 2000; Scharlach, 2001).
Integrated model
The third model is the integrated model of case management. Integrated
models frequently employ a multidisciplinary team approach to case management: the
PACE Program is an example of an integrated model. The case manager is usually a
registered nurse or social worker, providing a wide span of functions, including
counseling and other psychosocial services. Characteristics of integrated models
include: 1) Intensive case management services for clients most at risk of
institutionalization; 2) Smallest caseloads, usually less than 30; 3) Case managers
work in a multidisciplinary team; 4) Financial risk is high, as the case management
program is often at risk for all social and medical long-term care services (Huber,
2000; Scharlach, 2001).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
29
Consumer-Directed Care
In consumer-directed and consumer choice models of service provision,
individuals have the primary authority to make choices that work best for them
regardless of the nature or extent of their disability or the source of payment for
services (The National Institute on Consumer-Directed Long-Term Services, 1996).
Younger disabled people and their advocates, who have fought to achieve more
dignity and control in their lives, began the movement for consumer choiee and
direction in the U.S. health care system (Benjamin, 2001; Meiners et al., 2002).
Consumer choice and autonomy among older adults in the U. S. health care system
have been topics of considerable research since the late 1980s (Davies and Ware,
1988; Davis et al., 1995). Efforts to preserve autonomy and choice in health eare
decision-making have helped move the continuum of health care from a
predominantly medical model to a more balanced medical-social model. Innovative
models of consumer choice and consumer-directed care emphasize confer many of the
decision-making responsibilities to those individuals most affected by care planning
(Benjamin, 2001).
The evolution of formal care for frail and disabled people from
institutionalization to home and community-based services has increasingly had the
effect of enabling the U.S. health care system to provide more individualized goods
and services, encouraging greater creativity in using funding sources, and engendering
greater flexibility in personal assistance and caregiving employment (Squillace, 2002).
Programs like cash and counseling (Stone, 2000), which provide cash payments for
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
30
covered services that are disbursed at the discretion of the consumer, give program
beneficiaries some control and possibly a stronger sense of autonomy in making health
care decisions that affect them. Currently, several states are considering a capitated
managed care method for funding the delivery of long-term care services, using a cash
and counseling approach where appropriate (Squillace, 2002).
For Medicare managed care enrollees, consumer direction means more than
participating in psychosocial and medical strategies for care planning, it also means
choice in selecting health plans. Currently, 64% of all Medicare beneficiaries have
access to Medicare managed care with the right to change health care plans or return
to traditional Medicare FFS during annual open enrollment periods. Having the
choice to exit or switch health plans is a process that managed care members may use
to express dissatisfaction with health plan administrative or care practices (Centers for
Medicare & Medicaid Services, 2002).
As part of the Medicaid Program’s Freedom of Choice Initiative, CMS has
announced plans sponsor pilot programs to serve as the entry point to long term care,
assisting older adults plan for long-term care and social supports and providing
information about service and support options. The initiative also promotes projects
that assist people in avoiding nursing home stays and retuming to the community after
periods of institutionalization.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
31
Part 3. Funding LTC/HCBS
This section briefly details federal and state funding streams, including local
programs that promote or provide HCBS to older Americans: Medicare and Medicaid,
the Older Americans Act, and Social Services Block Grants.
Medicare and Medicaid
The largest purchaser of LTC is the federal government, mainly through the
Medicaid program. In 1999, Medicare and Medicaid together accounted for just over
$50 billion of LTC expenditures for older adults, about 57 percent of the total cost
(Centers for Medicare & Medicaid Services, 2002). Although Medicare primarily
pays for primary and acute medical services. Medicare beneficiaries may also receive
limited LTC services (see Figure 2) through Medicare’s skilled nursing and home
health care benefits.
Figure 2. Health Care Expenditures for Medicare Beneficiaries, 1999
Hospice
1%
Dental
3%
hstitutionai
23%
Source: CMS, Office o f Research, Development, and Information (1999)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
32
The Congressional Budget Office (CEO) estimates that inflation-adjusted
expenditures for LTC for older adults will grow by 2.6% annually from 2000 to 2040.
At this predicted growth rate, those expenditures will reach $207 billion in 2020 and
$346 billion in 2040 (Congressional Budget Office, 1995). The Medicare and
Medicaid Programs are only a part of federal, state, and local programs that enable
older adults to age in place, that is, in their homes and communities.
Older Americans Act
The Medicare and Medicaid programs are only a part of federal, state, and
local programs that enable older adults to age in place, that is, in their homes and
communities. The Older Americans Act (OAA), enacted in 1965, established the
Administration on Aging (AoA) within what is now the Department of Health and
Human Services (DHHS). The AoA is an advocacy organization for older Americans
and acts as a clearinghouse for aging information (Binstock, 1991). As part of its
unique responsibility, the AoA works closely with its nationwide network of regional
offices and Area Agencies on Aging to plan, coordinate, and develop community level
systems of services to meet not only the needs of individual older persons but also
their caregivers. Title III of the OAA authorizes 57 state agencies on aging and 660
area agencies on aging (AAA) to act as advocates and coordinate programs to benefit
older adults. Over 27,000 service organizations provide a variety o f social support
services, including meals, day care, case management, information and assistance, and
home care services. Funding is based on a formula that takes into account state
population of individuals aged 60 or over (Administration on Aging, 2002).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
33
Social Services Block Grants
One of the tasks confronting social workers and others providing health care
and social services to community-dwelling older adults is to increase awareness of and
access to HCBS. Congress enacted Social Services Block Grants (SSBG) as Title XX
of the Social Security Act in 1965 to provide homemaking, chore services, home
management, and home health aide services and help families and individuals in the
community retain autonomy (Centers for Medicare & Medicaid Services, 2002).
How these block grants affect recipients depends on how the individual states
choose to disperse the funds. SEC. 2001 (42 U.S.C. 1397) provides that states may
have flexibility and self-determination in the use of social service grants by
consolidating federal assistance programs into a single grant. In order to receive their
allotted block grants, states must provide, within reasonable bounds, the following
services to children, the aged, the mentally retarded, the blind, the emotionally
disturbed, the physically handicapped, alcoholics, and drug addicts:
Childcare services
Protective services for children and adults
Services for children and adults in foster care
Services related to the management and maintenance of home
Day care services for adults
Transportation services
Family planning services
Training and related services
Employment services
Information, referral, and counseling services
Preparation and delivery of meals
Health support services (Department of Health and Human Services, 2003)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
34
Part 4: Care Advocate Program Outcomes
The Care Advocate Program’s express goals were to raise levels of member
satisfaction, increase member retention, and reduce utilization of bigb-cost insured
medical services; therefore, this section examines literature relevant to MCO member
satisfaction, member retention, and utilization of insured medical services.
Member Satisfaction
Federal and state regulatory agencies and health plan administrators need
information about member satisfaction. Member satisfaction questionnaires are used
to provide feedback on bow Medicare managed care members rate their health plan’s
performance and quality of care. In general, most who chose to be MMCO members
are healthy. Yearly surveys by managed care organizations reveal high levels of
satisfaction among enrollees (Allen and Rogers, 1997). Despite these positive
findings, as MCOs assume an even more important role in the US health care system
over the next two decades, many health care researchers express concern that managed
care will reduce services in the name of cost-effectiveness, which may negatively
influence outcomes for frail older members (Boult et al., 1998; Kane, 1994; Lamphere
and Rosenbach, 2000). Therefore, the current trend of emphasizing capitated.
Medicare managed care over traditional fee-for-service calls for special attention to
quality of care for older adults with chronic and long-term care needs.
Experts in health care research (Dellana and Glascoff, 2001; Ipsen et al., 2000;
Sinay, 2002; Tudor, Riley, and higher, 1998) have cautioned U.S. health care
policymakers to examine multiple aspects of satisfaction. These include not only
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
35
overall satisfaction with health plan service, but also satisfaction with health plan
processes, and satisfaction with quality of care. Studies (Allen and Rogers, 1997;
Carlson, Blustein, Fiorentino, and Prestianni, 2000; Druss, Schlesinger, Thomas, and
Allen, 2000) have found numerous factors that influence member satisfaction with
Medicare managed care plans, including socioeconomic status, payment structure
(MCO vs. fee-for-service), number and severity of chronic conditions, and patient
outcomes.
Member Retention
Member retention can be a key indicator of plan satisfaction (Carlson et al.,
2000). Several recent studies (Carlson et al., 2000; Cunningham and Kohn, 2000;
Druss et al., 2000; Ipsen et al., 2000; Tudor et al., 1998) have found that socio
demographic characteristics, chronic disease, and dissatisfaction with health plan are
associated with plan disenrollment. Yet, disenrollment rates tend to be higher among
healthy members who perceive problems with access to services or quality of care
than among those with chronic conditions.
Frailer members are less likely to disenroll from health plans despite high
levels of dissatisfaction. Studies show a 12% to 17% average population
disenrollment rate per year for those with chronic diseases versus a 25% to 29%
disenrollment rate for similarly dissatisfied, healthy members (Riley, Ingber, and
Tudor, 1997). In fact, those who do not voice their concems are more at-risk to be
underserved by their health plan and have greater unmet needs than those who raise
concems and threaten to leave (Allen and Rogers, 1997; Dmss et al., 2000;
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
36
Schlesinger, Druss, and Thomas, 1999; Schlesinger, Mitchell, and Elbel, 2002). A
recent study (Lied, Sheingold, Landon, Shaul, and Cleary, 2003) examined Medicare
managed care disenrollment patterns, finding a strong association between perceived
health plan satisfaction (using Health Plan Employer Data and Information Sets
[HEDIS] and Consumer Assessment of Health Plans [CAHPS] surveys) and voluntary
disenrollment.
Member retention is not a clearly differentiated issue; those who stay may do
so out of necessity or out of their inability or unwillingness to voice dissatisfaction
(Lied et al., 2003), at the same time, members may express high levels of satisfaction
with some components of the health plan, while expressing dissatisfaction with
coverage limits, eo-pays, or premiums. In fact, the threat of plan switching is an
effective strategy for members to influence MCO premium costs and type of services
provided (Buchmueller, 1997; Newhouse, 2000). Members are more likely to switch
plans or return to traditional fee-for-service if their utilization of medical services is
low and lower plan premiums offset the financial costs of switching (Cunningham and
Kohn, 2000; Riley et al., 1997; Ullman, Hill, Scheye, and Spoeri, 1997).
Sinee 1998 there have been many changes in the Medieare program,
eonfounding researchers’ ability to draw eonclusions from MMCO participation rates.
A large number of health plans have dropped out, there have been changes in plan
benefits, and enrollment has deelined (Bender, Lance, and Guess, 2003; Centers for
Medicare & Medicaid Services, 2002). By 2001, about 15% of current
Medicare+Choice plan members, more than 934,000 Medicare beneficiaries.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2>1
experienced lost or reduced coverage due to their MMCO choosing not to renew
Medicare+Choice contracts or reducing service areas. About 159,000 beneficiaries
had no choice but to return to FFS. Overall, access to at least one Medicare managed
care plan dropped from 74% in 1998 to 60.5% in 2002. From 2001 to 2002, monthly
premiums rose from $21.60 to $31.14 and beneficiaries receiving basic plan drug
coverage fell from 71.85 to 62.6% (Centers for Medicare & Medicaid Services, 2004).
Utilization of Health Services
Understanding costs and utilization pattems of insured health care services
and assessing levels of functioning and disability are vital to Medicare policy makers.
Medicare MCO administrators, and contracted health care providers (Inouye et al.,
1998; National Chronic Care Consortium, 2001; Portrait, Lindeboom, and Deeg,
2001). According to the Medicare Current Beneficiary Survey (MCBS, 1999), annual
use of physician services for Medicare-risk MCO members living in the community
was 89.4% and the rate of inpatient hospital use was 17.1% (Centers for Medicare &
Medicaid Services, 2004).
A common goal for researchers in health care service provision is to find cost-
effective interventions that will minimize, delay, or prevent decline in functioning,
disease, disability, or death. Current research suggests that to be cost-effective,
interventions should target particularly frail or vulnerable older adults, those
considered at-risk for adverse health outcomes and high expenditures (Sloan et al.,
2003; Vita, Terry, Hubert, and Fries, 1998).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
38
A major topic for researchers estimating health care expenditures is the cost of
care in the last year of life. Older adults who are close to death have much higher
costs than those of the same age that survive (Yang, Norton, and Steams, 2003).
Other predictors of high utilization and expenditures among Medicare beneficiaries
include socio-demographic characteristics (age and gender), chronic diseases such as
heart disease, diabetes, and hypertension, number of diagnoses, and prior utilization of
health care services (Fishman, Von Korff, Lozano, and Hecht, 1997; National Chronic
Care Consortium, 2001; Sherman and Reuben, 1998).
End o f life
In the year 2000, the six leading causes of deaths among older adults were
heart diseases (33%), cancer (22%), cerebrovascular diseases (8%), chronic
obstmctive pulmonary disease (6%), pneumonia and influenza (3%), and diabetes
(3%) (Sahyoun, 2001). Table 3 indicates that over one quarter of all Medicare
expenditures for those aged 65 and older were incurred for end-of-life care. The
overall mortality rate for Medicare beneficiaries in 1999 was about six percent;
whereas. Medicare spending for end-of-life care in 1999 approached 29% of total
Medicare expenditures. Clearly, a disproportionate amount of total Medicare
expenditures occurs in the last 12 months of life.
Table 3. Percent of Total Medicare Dollars Spent on Beneficiaries Aged 65+ in Their Last
Year of Life 1997-1999
Year 1997 1998 1999
Percent 27.6% 28.5% 28.8%
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
39
Source: (National Center fo r Health Care Statistics, 2004)
Table 4 reports differences in Medicare spending for the years 1997-1999
between those who survived and those who died. In 1999, expenditures were nearly
seven times greater for those who died during the year when compared to those who
survived.
Table 4. Average Medicare Spending by Type of Service and Decedent Status, 1997-1999
1997 1998 1999
Decedent Survivor Decedent Survivor Decedent Survivor
Total $24,921 $3,801 $24,902 $3,684 $24,856 $3,669
Inpatient 56.4% 43.0% 56.6% 44.6% 58.4% 45.3%
SNF 10.0% 5.8% 10.3% 6.3% 8.8% 5.3%
Home Care 6.6% 11.4% 5.5% 7.0% 4.4% 5.3%
Hospice 3.3% 0.1% 3.6% 0.1% 3.9% 0.2%
Outpatient 5.6% 10.4% 5.6% 10.6% 5.2% 10.3%
Physician 15.5% 26.4% 16.2% 28.5% 17.1% 30.5%
DME 2.5% 2.8% 2.3% 2.9% 2.2% 3.1%
Attempts to define frailty abound in the literature. Frailty may he interpreted
as a combination of three out of five symptoms: low strength, slow walking speed, low
physical activity, self-reported unintentional weight loss, and self-reported exhaustion
(Newman, 2001). Borst (2002) noted that many researchers consider frailty to he a
syndrome, generally undefined and complex; yet, he goes on to define fi*ailty as “a
state of muscular weakness and other secondary widely distributed losses in function
and structure that are usually initiated hy decreased levels of physical activity” (p.
M284). Fried, et al. (2001) called frailty a syndrome with increasing prevalence in
older age, part of a “cycle of frailty associated with declining energetics and reserve”
(p. M l47). Recent studies measured frailty using gender, number of ADL and lADL
dependencies, number of chronic conditions, number and type of diagnoses, prior
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
40
utilization of medical services, and self-reported health status (Borst, 2002; Newman
and Brach, 2001; Newman, 2001; Nourhashemi et al., 2001; Rigaud, 2001; Sherman
and Reuben, 1998; Yelin, Criswell, and Feigenbaum, 1996).
Screening and targeting
A primary requisite to providing HCBS to a frail population is identification
of need (Challis and Hughes, 2002). Targeting those most in need of health care
serviees (both medical and HCBS) is important to prevent them from “falling through
the cracks of the system” and, eventually, from becoming the highest users of medical
and social health care (National Chronic Care Consortium, 2001). Risk screening and
assessments by case managers are frequently employed to determine who needs health
care services and which services are appropriate (Challis and Hughes, 2002; Fries,
Shugarman, Morris, Simon, and James, 2002; Keeler et al., 1999; Pacala, Boult,
Urdangarin, and McCaffrey, 2003). Over the last decade both public and private
insurers have invested considerable time and resources in developing screening tools
to assess risk among specific sub-populations, particularly frail older adults living in
the community (Fries et al., 2002) (Reuben et al., 2002; Saliba et al., 2001).
The probability of repeated admissions (Pra) instrument uses self-reports of
consumer health status and past utilization patterns to determine the likelihood of
future health-related service use among Medicare managed care enrollees. After one
year, the likelihood of hospital admissions and other medical claims among older
adults were two and one-half times greater among those rated high-risk for utilization
using the Pra screening tool. The research team that developed the Pra also found that
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
41
frail older adults identified as at high risk may benefit from interventions aimed at
detecting or preventing onset of disease or disability (Pacala, Boult, Reed, and
Aliberti, 1997).
Reuben et al. (2002) developed a method to identify older adults at high risk
for hospitalization. Researchers used self-reports of hospitalizations in last year, sex,
health status, impairments in certain ADLS and lADLs, and lab tests as predictor
variables. They found the instrument to be an efficient and economical method of
identifying older adults at risk for high health care utilization. With respect to those
85 or older, they found that the screening instrument performed well in identifying
high users of health care when combined with other factors such as gender and prior
utilization of services.
The Care Advocate frailty algorithm uses pharmacy utilization as one of the
criteria to produce a score for determining program eligibility. Several studies have
used pharmacy data to predict utilization patterns and expenditures. Citing the
increasing availability of pharmaceutical data, researchers (Von Korff, Wagner, and
Saunders, 1993) tested the potential for using a chronic disease score (CDS)
determined from automated pharmacy records to predict health care service utilization.
They found that, where the data are uniform (e.g. within a managed care organization),
pharmacy data were an adequate predictor of utilization patterns. Follow-up studies
(Johnson, Kramer, Lin, Kowalsky, and Steiner, 2000; Tamers, 1999; Sales et al., 2003;
Sloan et al., 2003) confirmed that, subject to certain limitations, pharmacy-based risk
assessment tools are a valuable, low cost, and reliable tool for determining health
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
42
status and likely utilization patterns. Some potential disadvantages in using this
approach are: 1) Many non-drug therapies cannot be captured; 2) Beneficiaries with
better benefits or who are more compliant will generate higher scores, creating false
positives; and 3) There may be issues of prescriber reliability, with one physician
more likely to prescribe medications than another would.
Influence o f HCBS
Several studies have demonstrated the effectiveness of HCBS in improving
health status and functioning while reducing utilization of medical services among
community dwelling older adults. An in-home assessment of older adults receiving
“home visitors” (Hendriksen, Lund, and Stromgard, 1984), for example, resulted in
significant increases in the use of social services, and reductions in hospitalizations
and deaths, among an older adult population. These older adults were not determined
to be frail nor were they enrolled in a Medicare managed care organization; however,
the intervention was similar to the Care Advocate Program in that home visitors
provided assessment and referrals to community-hased services, but did not provide
services that were otherwise available.
In a randomized trial studying the effects of a senior center-based, self-
management and disability prevention program, researchers found significant cost
savings stemming from decreased hospital days for those in the intervention,
compared to those in the control group. In addition, they found greater stability in
physical functioning, with higher levels of physical activity among study group
members. Moreover, an overall reduction in the use of psychoactive medications was
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
43
observed among study group members (Leveille et al., 1998). Moreover, in a
randomized test of the effectiveness of nurse-practitioner administered, comprehensive
in-home assessments to frail older adults, researchers (Stuck et al., 1995) found a
significant reduction in the number of people who required assistance in performing
basic activities of daily living and a significant reduction in number of permanent
nursing home admissions.
Studies of homebound older people in Italy (Landi, 2001) found significant
reductions in number of hospitalizations and a cost savings of $1,260 per person after
the one-year ease management implementation of the Minimum Data Set for Home
Care (MDS-HC) geriatric assessment instrument. The MDS-HC is a standardized
assessment tool adopted around the world to ascertain client performance and
capacity. The goal of the assessment is to identify ftmctional, medical, and social
issues that may limit a client’s capacity for independence and then to develop a care
plan based on the results of the assessment, aiming to maximize, in the context of life
circumstances, a client’s quality of life. A recent study (Doty, 2002) sought to
determine whether implementing selected HCBS for a Medicaid nursing facility
certifiable population could be cost-effective. They found that, despite the presence of
a “woodwork” effect (increased availability of HCBS producing greater total
utilization of long-term care services, regardless of need), HCBS targeted to specific,
needy, sub-populations could achieve budget-neutrality, particularly in the context of a
managed LTC system.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
44
Summary
This review of the literature described the foundations of the Medicare
program and the evolution of Medicare managed care. The Medicare program
provides primary and acute care services to aged and disabled beneficiaries; however,
many older adults also require chronic and long-term care. Beneficiaries must look to
other sources to meet chronic and long-term care needs. Home and community-based
serviees, funded and supported by federal, state, and local programs, and also paid for
out of pocket, attempt to fill the gaps in the system. This chapter examined service
provision and funding sources for LTC and HCBS services for older adults, including
case management and consvuner choice models. It also presented literature relevant
to Care Advocate Program outcomes - satisfaction, retention, and utilization of health
care services - and the potential influence of HCBS on the utilization of insured
medical services.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
45
CHAPTER 3: RESEARCH DESIGN AND METHODS
This chapter will describe the research questions and hypotheses, data sources,
research design, statistical analysis, and study limitations.
Research Questions and Hypotheses
Research Question #1: Can a telephone case management intervention, consisting of
assessment, information, and referrals to HCBS, significantly improve Medicare MCO
member satisfaction?
Hypothesis I: Intent-to-treat members will report higher scores, on average,
than control members in three components of satisfaction (overall, plan, and
care) from pre-intervention to after 12 months of intervention.
Research Question #2: Can a telephone case management intervention, consisting of
assessment, information, and referrals to HCBS, significantly improve HMO member
retention?
Hypothesis II: There will be significant differences in member retention such
that the ITT group members will leave the plan at lower rates than those in the
control group during the 12-month intervention.
Hypothesis III: There will be significant differences in member retention such
that the ITT group members will leave the plan at lower rates than those in the
control group in the 12 months post-intervention.
Research question #3: Can a telephone case management intervention, consisting of
assessment, information, and referrals to HCBS, significantly decrease member
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
46
utilization of medical services, specifically primary care physician (PCP), specialist,
and emergency room encounters, hospital admissions, and hospital days?
Hypothesis IV: There will be significant differences in utilization of insured
services between intent-to-treat and control groups from 12 months pre
intervention to 12 months during intervention, such that those in the intent-to-
treat group will show decreased utilization of medical services, on average,
compared to those in the control group.
Hypothesis V : There will be significant differences in utilization of insured
services between intent-to-treat and control groups from 12 months pre
intervention t to 12 months post-intervention, such that those in the intent-to-
treat group will show decreased utilization of medical services, on average,
compared to those in the control group.
Data Sources
PacifiCare data analysts provided satisfaction survey and administrative
utilization and retention data. Data analysts from the four medical groups provided
separate administrative utilization data. Moreover, intervention data were derived
from care advocate records.
PacifiCare Data
a. The Care Advocate Program evaluation relied on PacifiCare/Secure Horizons
member satisfaction surveys taken in March 1999, in March 2000, and in July
2002. To maximize the number of study participants in the satisfaction analyses,
a decision was made to combine the March 1999 and March 2000 satisfaction
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
47
surveys. The combined surveys (N = 733, ITT = 160, control = 129) constituted
the pre-intervention satisfaction data. The post-intervention satisfaction survey
targeted study members, exclusively (N = 238, ITT = 135, control = 103).
b. PacifiCare/Secure Horizons provided disenrollment statistics from
administrative records for all 5,800 eligible members for the period July 1, 2000
to June 30, 2003. This period reflects the rolling enrollment, 12-montb pre
intervention, intervention, and post-intervention study periods.
Medical Group Data
Utilization of five categories of insured healthcare services (PCP encounters,
specialist encoimters, hospital admissions, hospital days, and ER encounters) was
obtained from administrative data provided by the Medical Informatics data section of
PacifiCare Health Plans and data analysts from Talbert, Cedars-Sinai, Harriman-Jones,
and Health Care Partners Medical Groups for all 5,800 eligible members.
Care Advocate Data
a. The standard Jewish Family Services (JFS) assessment tool was modified to fit
the parameters of the Care Advocate Program. The functions of the
assessment tool were to determine member mental and physical health and
identify specific unmet needs.
b. Referral and follow-up data were maintained to measure which services were
referred and which were “accepted only” or “utilized” one or more times.
A computer programmer hired by PacifiCare created the assessment and
referral database specifically for the Care Advocate Program. PacifiCare data
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
48
analysts, care advocates, supervisors from the two HCBS agencies, and USC data
analysts contributed to the final product.
Research Design
Eligible population
To be eligible for the program, individuals bad to be age 65 or older, receiving
care from one of four participating medical groups, and enrolled for a minimum of one
year in PacifiCare’s Medicare risk program. All members who achieved a score of
four or more (scale of 0 to 11) using a health care utilization algorithm developed to
qualify participants for the Care Advocate Program (See Table 5) were eligible.
Members age 85 or older automatically qualified for the study. Staff from PacifiCare
developed the algorithm based on a review of the literature and analyses of
preliminary utilization data and the PacifiCare Member Health Questionnaire (MHQ).
These analyses indicated a positive association between algorithm-generated past
utilization and future health care utilization.
Table 5. Algorithm to Determine Program Eligibility
Age 85+ = add 4
Hospitalizations in the past year
1 = add 1
2+ = add 2
Emergency room visits in past year
1 = add 1
2+ = add 2
Current medications from pharmacy records
1-2 = add 1
3-4 = add 2
5+ = add 3
Score o f 4 or higher required fo r study eligibility
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
49
Excluded population
Certain older adults enrolled in PacifiCare/Secure Horizons were excluded
from eligibility for the Care Advocate Program:
■ Nursing home residents
■ Those enrolled in similar studies
Model o f Analysis
Study Group
Figure 3 presents the research model of the Care Advocate Program. The
eligible population (N = 5,800) was randomized to study (N = 2,976) and control (N =
2,814) groups. Statistical power analysis revealed that, using a significance criteria of
0.05, a sample size of 390 would be necessary for a power of 0.80, using non-
directional (two-tailed) tests (Cohen, 1988). Randomly selected study group members
received a letter inviting them to participate. The response rate to this letter was low
(14%). A focus group of older adults evaluated the letter content and formatting and
offered suggestions to improve readability and understanding. These suggestions were
incorporated into a second letter and the response rate almost doubled (26%). All
members responding to the invitation were contacted by care advocates to set up
assessment dates and were tracked for satisfaction, retention, and utilization outcomes
for as long as they remained in the health plan (see intent-to-treat analysis).
Of the 389 members in the original ITT group, 118 disenrolled from the study
before assessment (eight died, 14 left the plan, and 96 declined to participate further).
Care Advocates assessed the remaining 271 ITT members. Four members died after
assessment, but before any referrals were made and 16 other members disenrolled
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
50
from the health plan between assessment and intervention, leaving 251 members who
received the information and referral to HCBS intervention.
Figure 3. Analytic Model: Linking Managed Care and HCBS
Baseline
Characteristics
Stuc^Grixps Intervertion
OutOOfTBS
During &Fbst-
Irtervention
Denagraphics
Gender
(N=6800)
Satisfaction
O/e^l
Plan
care
(N=733)
Fteterton
LeflR ai
D ed(
Ative
(N=5800)
Udlizaticn
PCPEnoourters
Specialist
Enoourters
Hospital
A trissions
Hospital Days
B^Erxxxrters
(N=68(X))
Irtert-toTreat
Groip
(N = 389)
Ded (1^ 8)
Left Han (N= 14)
ReUjsed tN=96i
Totd (N=118
Derrographics
figs
Gender
Qontnol
Qoup
(N=434)
C^AMocate
A sessed
(N=271)
Irfomnationand
Referral to
HCBS
Ded (N= 4)
Left Han i t a S )
Tolal (N=20)
Axept Ffeferrals
(N=251)
Utilize Referrals
(N=164)
SatislAtion
CVerall
Han
care
Retention
Left Ran
Ded
A tive
Lliiization
PCPBTCOLiiters
Specialist
Bnoointers
Hospital
A hissions
Hospital Days
ERBnoourters
Control Group
Four hundred and thirty-four control group members were selected from the
original control group using a random number generation process and were matched to
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
51
ITT group members by termination dates to insure active study status when the
intervention commenced.
Outcomes
To determine whether the Care Advocate Program met goals designated in the
grant proposal, the following change variables were created:
1. Change from baseline to post-intervention for satisfaction (overall, plan, and
care)
2. Change in retention (left, died, and active in plan)
3. Change in utilization patterns (PCP encounters, specialist encounters, hospital
admissions, hospital days, and emergency room encounters) from 12 months
baseline (pre-intervention) to 12 months during the intervention and from the
12 months baseline to 12 months post-intervention.
Instruments and Procedures
Assessment Instrument
The care advocates used an 83-question assessment instrument, which
measured socio-demographic characteristics (age, gender, marital status, education,
living arrangement, income, primary language), current medical conditions, service
utilization over the past year, and current pharmacy use. It employed a brief cognitive
(name, date, birth date, age) screening method modeled after the Mini M ental State
Examination (MMSE), a validated, widely accepted tool for assessing cognitive
mental status (Folstein, Folstein, and McHugh, 1975) and the Katz Index of
Independence in Activities of Daily Living (ADLs) to assess performance in bathing.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
52
dressing, toileting, transferring, and continence (Katz, Ford, Moskowitz, Jackson, and
Jaffe, 1963). Lawton’s Instrumental Activities of Daily Living (lADLs) measured
participants’ need for assistance with cooking, cleaning, shopping, money
management, transportation, use of telephone, and medication administration (Lawton
and Brody, 1969). Care advocates also asked about alcohol use, smoking habits, and
self-reported quality of life. Since consumer preferences are an important component
of the study, care advocates asked participants what services interested them and
incorporated those suggestions into their recommendations.
Those not able to respond by phone at the initial assessment received a one
time home visit. Criteria for a home visit were that the client was unable to provide
information to a care advocate or to utilize/access a referral from a care advocate due
to physical, cognitive, or emotional disabilities. The care advocate made this
determination based on participant responses to the assessment questionnaire.
Problems that triggered home encounters included hearing loss, emotional/mental
illness, significant dementia, or any medical conditions that precluded telephone
assessment. Twenty two percent of those contacted (N = 60) met these criteria and
received an in-home assessment and 11 monthly follow up visits.
Satisfaction Surveys
The PacifiCare member satisfaction surveys employed in this study followed
Consumer Assessment of Health Plans (CAHPS) guidelines established by CMS.
Since 1999, completion of this survey has been a requirement for accreditation by the
National Committee of Quality Assurance (NCQA). The original CAHPS survey was
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
53
developed for the Health Care Financing Administration (HCFA), now CMS. It is the
most comprehensive and widely used health care survey in the U.S. (Hays et al., 1999;
Schnaier, 1999; Solomon, Zaslavsky, Landon, and Cleary, 2002). CAHPS data are a
validated, standardized measure of consumer experiences with health plan
administration, care, and practices (Schnaier, 1999; Solomon et al., 2002; Zaslavsky,
Shaul, Zaborski, Cioffi, and Cleary, 2002). Data are collected once a year, according
to CMS guidelines, and adjusted for case-mix differences among plans to control for
socio-demographic and health characteristics (Zaslavsky et al., 2002).
Baseline pre-intervention consumer satisfaction used survey results taken in March
1999 and March 2000. These surveys combined included 733 responders from the
5,800 eligibles. Two hundred and eighty nine responders (34.8%) were ITT [N= 160
(55.4%)] or control [N = 129 (44.6%)] group members. PacifiCare targeted the post
intervention satisfaction survey to ITT and control members, contacting members in
the first three months following the conclusion of the Care Advocate Program
intervention. Two hundred and seventy-five members responded to the satisfaction
survey. Of these, 37 surveys were incomplete and excluded from the analysis. The
remaining 238 responders were as follows: ITT - 135 (56.7%) and control = 103
(43.3%).
Missing Data
There were missing data in the post-intervention satisfaction surveys of ITT
and control members. To determine if missing data significantly threatened internal
validity, a dummy variable was created (missing or not) to evaluate predictors of
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
54
missing variables in the satisfaction measures. Data are considered missing
completely at random (MCAR) if their being missing is independent of all other data
(Sheiner, 2002). Logistic regression analysis revealed no predictors of “missingness”
among the satisfaction variables. Since no predictors of missing variables were found,
the potential bias created when imputing mean values for missing values was
minimized (Petkova and Teresi, 2002). Therefore, the statistical mean of each
variable was imputed for all missing values. The number of values for each variable
with missing data did not exceed 10% of the total number of observations: Imputing
fewer than 10% of the total number of observations is considered acceptable (Little,
2002).
Statistical Analysis
Intent-to-Treat Analysis
In clinical trials, participants refuse treatment or drop out of the study for
various reasons: used up henefits, moved out of area, and death, for example. Intent-
to-treat (ITT) analysis takes the view that all treatment group members should be
analyzed regardless of whether they received the treatment or not. Intent-to-treat
analysis uses last observation carried forward, imputation models, or mixed effects
models. These analyses assume that the subjects remain “on-treatment” throughout the
observation period. This approach is useful in understanding the effect o f the treatment
in subjects who stay on treatment, those who receive partial treatment, or those who
receive no treatment. To verify this approach, utilization outcomes of those who
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
55
received any treatment were compared with those in the ITT group who remained with
the health plan but did not receive treatment (Gilden, 1997; Houck, 2003).
Focus Group Analysis
Two focus group meetings were held. The first was with older adults,
volunteers at the Andrus Gerontology Center of the University of Califomia (USC).
This focus group was organized approximately three months after the first group of
participants were assessed, when the response rate to the first letter of invitation was
considered too low to attraet enough members to meet the initial sample size estimates
to meet the study requirements of a power of .80; a = .05. A second focus group, this
time with the care advocates, was organized approximately one month after the
conclusion of the Care Advocate Program study. A trained member of the USC
evaluation team facilitated both focus groups. The purpose of the focus groups was to
explore and eonfirm subjective impressions, first from the Andrus Volunteers
regarding the first letter of invitation to the CA Program, and second, to gain insights
from the eare advocates regarding their experiences in the CA Program (Kerschner,
1992; Stewart and Shamdasani, 1998)
Statistical Programs
The SAS System for Windows (8.02) was employed to convert data reeeived
from PacifiCare and the four medical groups from claims-level utilization to person-
level utilization, to ascertain individual member start dates, and to determine
individual pre-, during, and post-intervention study periods. Determining 12-month
study periods was complicated by the rolling enrollment status over a two-year period.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
56
Once claims-level data were converted to person-level data, they were analyzed using
the SPSS (Version 10.01) statistical program to generate descriptive statistics and
perform all other statistical analyses.
Eligibility Algorithm
To confirm the validity of the PacifiCare frailty algorithm, t-tests and chi-
squares were performed to determine baseline differences between those eligible
solely because they were 85 or older and the rest of the eligible population. Logistic
regression analysis determined the likelihood of the presence of certain diagnoses,
service utilization, and study status by 85+ group or rest of the eligible population,
eontrolling for demographics, using post-intervention data.
Baseline Characteristics
Chi-square tests determined significant differences in admission characteristics
for discrete variables. T-tests determined significant differences for continuous
variables where the distribution was normal. Where the distributions were skewed, as
is fi"equently the case with count data (PCP encounters, specialist encounters, hospital
admissions, hospital days, and ER encounters), Mann-Whitney nonparametrie tests
were employed to test the null hypothesis that two independent samples are from the
same population (SPSS, 1999).
Care Advocate Program Outcomes
Satisfaction
Cronbach’s Alpha measured the reliability of the 12 component variables from
the satisfaction instrument (a = .8839). Exploratory factor analysis reduced the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
57
thirteen satisfaction variables to three principal components: 1) overall satisfaction, 2)
satisfaction with health plan, and 3) satisfaction with care. Table 6 shows the factor
matrix from the pre-intervention satisfaction survey to these three components of
satisfaction. Both pre-intervention and post-intervention satisfaction survey questions
were the same.
Table 6. Factor Matrix for Components of Satisfaction (N= 733)
Conponent*
Measure 1 2 3
QST2 Satisfied? -0.091 0.003 0.953
QST3 Confidence? -0.166 -0.186 0.927
QST5A Trust? 0.879 0.139 -0.136
QST5B SH concerned? 0.890 -0.003 -0.061
QST5C Help stay healthy? 0.816 -0.128 -0.050
QST5D Good value? 0.737 -0.017 -0.258
QST5G Keeps its Promises? 0.845 0.080 -0.137
QST5H Communicates Directfy & Honestfy? 0.857 -0.078 -0.095,
Q STl 3 A Treat you with courtesy & respect -0.001 0.902 -0.150
QST13B Listen caretulfy? -0.012 0.839 -0.136
QST13C Explain well? -0.034 0.926 -0.087
QST13E Spend enough time.'' 0.038 0.875 0.030
'Extraction Method: Princqial Conponent Anab'sis.
Since survey teams specifically targeted them, pre-intervention (N = 289) and
post-intervention (N = 238) satisfaction survey responders were all members of the
ITT/control groups; yet, only 125 (ITT = 78; control = 47) answered both surveys.
This sample size was insufficient for a = .05 significance and .80 power criteria
(Cohen, 1988). Therefore, only group comparisons were analyzed from pre- to post
intervention. Mean satisfaction scores were compared at both pre and post
intervention for each measure (overall, plan, and care satisfaction). T-tests were used
to determine significant differences between groups at pre- and post-intervention.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
58
Model 1: Overall satisfaction mean value at post-intervention compared to overall
satisfaction mean value at pre-intervention
Model 2: Plan satisfaction mean value at post-intervention compared to plan
satisfaction mean value at pre-intervention
Model 3: Care satisfaction mean value at post-intervention compared to care
satisfaction mean value at pre-intervention
Retention
Three outcomes of member retention were identified - left plan, died, and
active in plan. Member retention criteria were determined by assigning a start date for
each participant and comparing it to any assessment date and study or plan termination
date to determine the 12-montb study period (pre-intervention, during intervention,
and post-intervention). Frequencies of the three outcomes were examined at time one
(12 months pre-intervention), time two (completion of the intervention) and time three
(12 months post-intervention). T-tests were performed to determine if there were
significant differences by study group (ITT or control) and between those who
actually received the Care Advocate intervention (N=271) and those in the control
group (N=434). The effects of post-intervention satisfaction (overall, plan, and care)
on plan disenrollment in the following 12 months were estimated using logistic
regressions. Regression models controlled for study group (ITT/control), number of
diagnoses post-intervention, and utilization (use/not use) of medical services (PCP,
specialist, hospital, and ER).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
59
Utilization
Use of insured medical services was measured as change from pre-intervention
to during or post-intervention. Change scores were determined by subtracting pre
intervention scores from during-intervention scores and pre-intervention scores from
post-intervention scores. Change score frequencies were examined and categories of
change (increased = 1, decreased = 2, and no change = 3) were determined using plus
or minus one standard deviation as cut-off points (Ware, Bayliss, and Rogers, 1996).
Multinomial logistic regression models compared change (A) in utilization patterns by
ITT and control group, controlling for months in study (to control for changes due to
disenrollment or death), and died within the period (to control for end of life
utilization) as follows:
Model 1: PCP A = f (months in study -i- died in period + study group - H e)
Model 2: Specialist A = / (months in study - i- died in period + study group + e)
Model 3: Hospital Days A = f (months in study + died in period + study group + e)
Model 4: Hospital Admits A = f (months in study + died in period + study group + e)
Model 5: ERA=f (months in study + died in period + study group + e)
Study Limitations
Selection bias may have been introduced because it was not possible to
randomize participants after they opted into the program since only those in the ITT
group were offered the intervention. To address this problem, a second comparison
group was created using a propensity score approach (See chapter 5). In addition, 733
eligible members responded to the pre-intervention surveys, of which 289 became
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
60
study members. Post-intervention, 238 members responded to the satisfaction survey,
but only 125 responded at both pre- and post intervention. The small sample size led
to the statistical power limitations for within-group analyses (Cohen, 1988);
consequently, only between-group data were analyzed.
Moreover, each medical group had its own data entry and coding procedures.
This lack of uniformity made merging data files cumbersome and limited service types
to the five presented. Where possible, efforts were made to validate medical group
data with data from PacifiCare. Finally, the demonstration reflected the experiences
of members from only one managed care plan, in one area of the country. Although
the involvement of four distinct medical groups contracted to provide medical services
to PacifiCare/Secure Horizon plan members may have improved the external validity,
results were not stratified by medical groups.
Summary
Chapter 3 has reviewed sources of data and described the research design,
including the model of analysis, study sample, selection criteria, assessment and
satisfaction survey instruments and study procedures. In addition, methods of
statistical analysis and study limitations were presented.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
61
CHAPTER 4: RESULTS
Introduction
This chapter is organized into three parts: Part 1 presents baseline
characteristics for study groups and study results by research hypothesis from pre
intervention to during intervention (or after the intervention, in the case of satisfaction
survey results). The characteristics include differences in member satisfaction
(overall, plan, and care satisfaction), member retention (active in plan, died, left plan),
and utilization of insured medical services (PCP encounters, specialist encounters,
hospital admissions, hospital days, and emergency room encounters). Part 2 presents
findings by hypothesis from 12-months pre-intervention to 12-months post
intervention. Part 3 lists the four objectives of the Care Advocate Program and
presents findings relevant to each.
Part 1: Findings Pre-Intervention to During Intervention
Baseline Characteristics
Table 7 shows that there were no statistically significant differences in age,
gender, or health status between ITT and control groups at baseline. Utilization of
health care services was measured in two ways: use of services a binary variable (used
or not) and number of encounters (count data). Primary care physician use over the
previous 12 months was significantly higher (97.2% vs. 91.7%) for the ITT group
compared to the control group (x^ = 11.387; p < .001) and significantly higher for
number of PCP encounters (z-score = -1.967; p <05). The differences in PCP
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
62
utilization, while significant, were not substantial enough to reject the null hypothesis
that the study groups (ITT/control) came from the same population.
Table 7. Characteristics of ITT/Control Groups for 12 months Pre-Intervention
ITT
(N=389)
Control
(N ^ 3 4 ) Chi. Sq. t-test Signiflcance
Demographics
Age 83.04(7.16) 83.67 (7.43) -1.225 0.221
Gender (Female) 250 (64.3%) 284 (66.6%) 0.123 0.770
Health status
Number of diagnoses 4.07 (3.13) 3.88 (3.12) 1.580 0.115
Heart 188 (48.3%) 195 (50.1%) 0.007 0.934
Cancer 72(18.5%) 61 (14.1%) 3.003 0.088
Cerebrovascular 58 (14.9%) 50 (12.9%) 1.519 0.220
Pneumonia & COPD 119(30.6%) 114(29.3%) 2.326 0.140
Fractures 32 (8.2%) 29 (7.5%) 0.341 0.601
Osteoarthritis 107 (27.5%) 126 (32.4%) 2.661 0.107
Diabetes 77 (19.8%) 69 (17.7%) 0.246 0.657
Prostate 33 (8.5%) 27 (6.9%) 1.234 0.164
Kidney 11 (2.8%) 13 (3.3%) 0.226 0.634
Hypertension 245 (63.0%) 253 (65.0%) 2.475 0.117
Utilization
PCP use 378 (97.2%) 398 (91.7%) 11.387 0.001
Specialist use 334 (85.9%) 357 (82.3%) 1.978 0.160
Hospital use 126 (32.4%) 142 (32.7%) 0.010 0.941
ERuse 100 (25.7%) 133 (30.6%) 2.465
0.122
Mann-Whitney U Tests Rank Rank Z-score
PCP count 429.18
396.60
-1.967 0.049
Specialist count 422.79
402.33 -1.239 0.216
Hospital admiss. count 409.17 414.54 -0.390 0.696
Hospital days count 407.76
415.80
-0.592 0.696
ER count 400.41
422.39 -1.673 0.094
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
63
Hypotheses
Hypothesis I: Intent-to-treat members will report higher scores, on average,
than control members in three components of satisfaction (overall, plan, and care)
from pre-intervention to after the 12-month intervention.
Figure 4 indicates high levels of satisfaction for all three components of
satisfaction (overall satisfaction = OverSat, 5-point scale; plan satisfaction = PlanSat,
10-point scale; and care satisfaction = CareSat, 4-point scale). There were no
significant differences between ITT/control groups in overall, plan, or care satisfaction
at pre- or post-intervention (pre-intervention group N=289; post-intervention group
N=238).
Figure 4. Satisfaction Pre-Intervention to After-Intervention; ITT/Control
ITT OverSat
ControlOverSat
ITT PlanSat
Control PlanSat
ITT CareSat
Control CareSat
8.88
.66
2.73
3.79
3.85
3.85
0.00 2.00 4.00 6.00 .00 10.00
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
64
Hypothesis II: There will be significant differences in member retention such
that the ITT group members will leave the plan at lower rates than those in the control
group during the 12-montb intervention.
Table 8 presents t-tests for differences in member retention (active, deceased,
or left health plan) by study group. Comparing the intent to treat (ITT) with the
control group, there were significant differences in the number who died (ITT = 26,
Control = 66). ITT members left the plan at a about same rate as control members
(ITT = 13.6% vs. control = 10.8%).
Table 8. Differences in Member Retention During 12-Month Intervention
ITT
(N = 389
Control
(N =434) t-value p-value
Active 310 (79.7%) 321 (74.0% ) 1.942 0.052
Deceased 26 (6.7%) 66(15.2% ) -3.905 0.000
Left 53 (13.6% ) 47 (10.8% ) 1.225 0.221
Table 9 indicates results for those who received the Care Advocate Program
intervention. Once assessed, members received 11 monthly calls (or, as appropriate,
visits) from their Care Advocate. During the 12-montb intervention period 212
intervention group members remained active, 18 members died, and 39 left the health
plan. CA group members bad significantly lower mortality levels from the control
group (CA= 18, control = 66; p-value = .001). There were no significant differences
between groups in the proportion of members who left the plan.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
65
Table 9. Differences in Member Retention by CA Program and Control
CA
(N = 271)
Control
(N =434) t-value p-value
A ctive 214 (79.0% ) 321 (74.0% ) 1.283 0.200
D eceased 18 (6.6%) 66(15.2% ) -3.271 0.001
Left 39 (14.4% ) 47 (10.8% ) 1.544 0.123
Hypothesis IV: There will be significant differences in utilization of insured
services between intent-to-treat and control groups fi*om 12 months pre-intervention to
12 months during intervention, such that those in the intent-to-treat group will show
decreased utilization of medical services, on average, compared to those in the control
group.
To determine if the Care Advocate intervention affected utilization patterns
from 12 months pre-intervention to 12 months during intervention, multinomial
logistic regressions of utilization change (PGP encounters, specialist encounters,
hospital admissions, hospital days, and ER encounters) were performed. Outcomes
studied were the likelihood of increased or decreased service use, compared to no
change (referent), controlling for months in the plan and death during the intervention.
Table 10 indicates that ITT members were 74% (0/R =1.741) more likely to
have increased utilization of PCP services, 74% (0/R = 1.736) more likely to have
increased utilization of specialist services, and 43% (0/R = .572) less likely have
increased admissions to the hospital, compared to control members, controlling for
months in the plan, and died. The referent is no change in utilization.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CD
■ D
O
Q .
C
o
CD
Q .
■ D
CD
C/)
W
o'
o
o
Table 10. Multinomial Logistic Regressions: Utilization Change Scores Pre- to During Intervention
o
o
■ D
c q '
o
o
CD
■ D
O
Q .
c
a
o
o
■ O
o
CD
Q .
■ D
CD
( / )
( / )
Dependent Variables: ITT/Control Group' (N=823)
Decreased
Utilization
Independent
Variables
PCP
Change
(N=94)
beta p-value O/R
Specialist
Change
(N=107)
beta p-value O/R
Hosp. Admission
Change
(N=66)
beta p-value O/R
Hospital Days
Change
(N=52)
beta p-value O/R
ER Encounters
Change
(N=63)
beta p-value O/R
Intercept 0.153 0.746 -0.241 0.602 -0.902 0.114 -1.909 0.006 -1.161 0.000
Months In -0.073 0.000 0.929 -0.050 0.000 0.951 -0.045 0.011 0.956 -0.027 0.198 0.973 -0.037 0.930 1.002
Died 0.307 0.348 1.360 0.088 0.793 1.093 0.363 0.369 1.438 0.788 0.084 2.199 0.220 0.100 3.404
ITT/Control -0.116 0.622 0.890 -0.268 0.219 0.765 -0.462 0.090 0.630 -0.129 0.665 0.879 -0.439 0.113 0.645
Increased PCP Specialist Hosp. Admission Hospital Days ER Encounters
Utilization Change Change Change Change Change
Independent (N=95) (N=69) (N=62) (N=48) (N=59)
Variables
beta p-value O/R beta p-value O/R beta p-value O/R beta p-value O/R beta p-value O/R
Intercept -2.240 0.000 -3.175 0.000 -2.037 0.004 -2.562 0.001 -2.714 0.000
Months In 0.003 0.854 1.003 0.020 0.367 1.020 -0.009 0.667 0.991 -0.006 0.809 0.994 0.002 0.930 1.002
Died -1.014 0.115 0.353 0.056 0.924 1.058 0.839 0.067 2.308 1.143 0.022 3.138 1.225 0.010 3.404
ITT/Control 0.555 0.014 1.741 0.552 0.035 1.736 -0.559 0.048 0.572 -0.404 0.203 0.668 -0.043 0.876 0.957
^Likelihood of increased/decreased utilization from pre- to post-intervention, no change (referent).
o^
o^
67
Part 2: Findings Pre- to Post-Intervention
This section discusses findings from retention and utilization data analyses in
the 12 months post intervention. Satisfaction was measured before and after the 12-
month intervention. No satisfaction data are available for the 12 months post
intervention; however, a study goal was to determine if increased member satisfaction
could improve member retention. Table 11 shows the influence of satisfaction after
the intervention on member retention post-intervention, controlling for study group,
number of diagnoses post-intervention, and utilization post-intervention (PCP,
specialist, hospital, and ER - these are binary, use/not use variables). Those with
lower overall satisfaction scores after the intervention were 58% more likely to leave
the plan post-intervention (b=-.86S; p-value = .006; 0/R = .420). Similarly, those with
lower scores on plan satisfaction were 69% more likely to leave the plan post
intervention (b = -1.170; p-value = .000; 0/R = .310).
Table 11. Logistic Regression: Influence of Satisfaction on Member Retention
Left Post-Intervention
Model 1
Overall Satisfaction
beta p-value 0/R
Model 2
Plan Satisfaction
beta p-value O/R
Model 3
Care Satisfaction
beta p-value O/R
Constant -12.286 0.751 -12.897 0.835 -15.980 0.680
ITT/Control 1.006 0.140 2.735 1.060 0.141 2.886 1.121 0.094 3.068
Number of Diagnoses -0.054 0.624 0.947 -0.072 0.537 0.930 -0.091 0.395 0.913
PCP 6.002 0.831 404.420 7.011 0.873 1108.290 6.425 0.817 617.275
Specialist 6.203 0.816 494.420 6.981 0.873 1076.450 6.464 0.810 641.476
Hospital 1.820 0.012 6.174 2.087 0.004 8.057 1.918 0.004 6.806
ER 0.397 0.587 0.972 -0.450 0.542 0.638 -0.575 0.408 0.563
OverSat -0.868 0.006 0.420
PlanSat -1.170 0.000 0.310
CareSat -0.049 0.905 0.952
Nagelkerke
-2 Log likelihood
0.244
91.953
0.319
84.287
0.167
99.479
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
6 8
Hypothesis III: There will be significant differences in member retention such
that the ITT group members will leave the plan at lower rates than those in the control
group in the 12 months post- intervention.
Table 12 presents t-tests for differences in member retention (active, deceased,
left) by study group (ITT versus control) in the 12 months post-intervention. There
were no significant differences between ITT and control groups in number died or left
the health plan. Those assessed and active in the plan post-intervention, the CA group,
left the plan in significantly greater numbers than did control group members (CA=
17%, control = 10.6%; t-value = 2.360; p = .019).
Table 12. Differences in Retention by Study Groups: ITT & CA’ versus Control
ITT vs. Control 12 Months Post-Intervention
ITT Control
(N = 310 (N =321) t-value p-value
Active 234 (75.5%) 246 (73.0%) 1.075 0.283
Deceased 30 (9.7%) 41 (12.8%) -0.884 0.377
Left 46 (14.8%) 34 (10.6%) 1.983 0.054
CA vs. Control 12 Months Post-Intervention
CA Control
(N = 212) (N=321) t-value p-value
Active 157 (74.1%) 246 (76.6%) 0.386 0.700
Deceased 19 (9.0%) 41 (12.8%) -1.127 0.260
Left 36(17.0%) 34 (10.6%) 2.360 0.019
‘ CA group = members assessed
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
69
Hypothesis V : There will be significant differences in utilization of insured
services between intent-to-treat and control groups from 12 months pre-intervention to
12 months post-intervention, such that those in the intent-to-treat group will show
decreased utilization of medical services, on average, compared to those in the control
group.
Table 13 shows the results of multinomial logistic regressions indicating that
there were no significant differences between the ITT and control groups in the
likelihood of utilization change (increased or decreased) post-intervention, controlling
for number of months in the plan and whether the member died post-intervention,
compared to no change (referent).
Part 3: Findings for Program Objectives
Objective 1: Create a screening mechanism to identify frail Medicare HMD members
Data analysts from PacifiCare identified a frail population within Secure
Horizons using the point-system algorithm. The literature on fi*ailty and screening
mechanisms to identify older adults most likely to use costly health care services is
considerable (Borst, 2002; Newman, 2001; Nourhashemi et al., 2001; Rigaud, 2001;
Sherman and Reuben, 1998; Yelin et al., 1996). The algorithm automatically qualified
members aged 85 or older for the Care Advocate Program. In other words, the
algorithm assumed that age 85-1- is a proxy for need.
To test this assumption. Table 14 presents baseline characteristics of those who
were aged 85 or older (85-)- group), who qualified by age only, compared to the rest of
the eligible population, who qualified based on utilization criteria. The 85-i- group
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
70
differed significantly from the rest of the eligible group in number of diagnoses, in
specific diagnoses of heart conditions, cancer, cerebrovascular diseases, pneumonia
and chronic obstructive pulmonary disease, osteoarthritis, diabetes, kidney disease,
and hypertension, and in use of health care services (PCP, specialist, hospital, and ER)
at baseline (12 months pre-intervention).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CD
■ D
O
Q .
C
o
CD
Q .
■ D
CD
C/)
W
o'
o
o
Table 13. Multinomial Logistic Regressions: Utilization Change Scores Pre- to Post-Intervention
o
o
■ D
c q '
o
O ’
o
CD
■ D
O
Q .
c
a
o
Q
■ o
o
CD
Q .
■ D
CD
( / )
W
o'
o
Dependent Variables: ITT/Control Gronp^ (N=631)
DECREASE
D Utilization
Independent
Variables
PCP
Change
N = 27
beta p-value O/R
Specialist
Change
N = 89
beta p-value O/R
Hosp. Admission
Change
N = 35
beta p-value O/R
Hospital Days
Change
N = 23
heta p-value O/R
ER Encounters
Change
N =41
beta p-valne O/R
Intercept-11.313 0.003 -11.025 0.000 -13.560 0.000 -10.857 0.000 -7.924 0.000
Months In 0.236 0.025 1.267 0.260 0.000 2.338 0.298 0.002 1.347 0.202 0.013 1.224 0.146 0.004 1.157
Died 1.113 0.125 3.042 0.849 0.081 1.297 2.669 0.000 14.425 2.994 0.000 19.961 1.859 0.000 6.418
ITT/Control -0.245 0.543 0.783 0.214 0.358 1.238 -0.009 0.981 0.991 -0.068 0.878 0.934 -0.039 0.905 0.962
INCREASED PCP Specialist Hosp. Admission Hospital Days ER Encounters
Utilization Change Change Change Change Change
Independent N =119 N =81 N = 87 N = 44 N = 73
Variables beta p-value O/R beta p-value O/R heta p-value O/R heta p-value O/R beta p-valne O/R
Intercept 1.115 0.001 -0.006 0.985 0.091 0.800 -0.813 0.476 -0.491 0.215
Months In -0.110 0.000 0.896 -0.078 0.000 0.924 -0.074 0.000 0.928 -0.074 0.017 0.928 -0.057 0.000 0.944
Died 0.384 0.264 1.468 0.697 0.053 2.008 0.087 0.836 1.092 0.778 0.467 2.176 0.440 0.277 1.552
ITT/Control
0,323 0.134 1.381 0.054 0.822 1.056 -0.188 0.426 0.828 -0.079 0.805 0.924 -0.405 0.116 0.667
^Likelihood o f increased/decreased utilization from pre- to post-intervention, no change (referent).
72
Table 14. Baseline Comparison of 85+ Group and Rest of Target Population
12 Months Pre-Intervention
Age 85+
Only
(n = 796)
Rest
(n = 5004)
t-test Significance
2-tailed
Demographics
Age 88.85 (3.48) 82.16 (7.50) 27.47 0.000
Gender (Female) 637 (64.3%) 3060 (63.6%) 0.18 0.671
Health status
Number of diagnoses 2.13 (2.65) 4.13 (3.48) -17.105 0.000
Heart 273 (27.5%) 2356 (49.0%) 152.46 0.000
Cancer 110(11.1%) 715 (14.9%) 9.56 0.001
Cerebrovascular 85 (8.6%) 744 (15.5%) 31.88 0.000
Pneumonia & COPD 167(16.9%) 1392 (28.9%) 61.15 0.000
Fractures 74 (7.5%) 338 (7.0%) 0.24 0.635
Osteoarthritis 180(18.2%) 1044 (21.7%) 6.21 0.013
Diabetes 42 (4.2%) 974 (20.3%) 145.87 0.000
Prostate 58 (5.9%) 357 (7.4%) -1.75 0.081
Kidney 16 (1.6%) 182 (3.8%) 11.74 0.000
Hypertension 319 (32.2%) 2960 (61.6)% 288.27 0.000
Utilization
PCP use 723 (73.0%) 4492 (93.4%) 378.97 0.000
Specialist use 587(59.2%) 3927(81.7%) 239.48 0.000
Hospital use 229(23.1%) 1843 (38.3%) 82.85 0.000
ERuse 209(21.1%) 1587(33.0%) 54.53 0.000
Table 15 compares those who were included only because they were aged 85
or over (85+ group) with the rest of the population in the 12 months post-intervention.
Those in the 85+ group were more likely to be male than the rest of the eligible
population and were significantly 30% less likely to have been diagnosed with heart
disease (b = -0.362, 0/R = 0.696), 60% less likely to have been diagnosed with
diabetes (b = -0.903, 0/R = 0.405), and 50% less likely to have been diagnosed with
hypertension (b = -0.691, O/R = 0.501). There were no signifieant differences in the
likelihood of medical service utilization and those in the 85+ group were 38% (b = -
0.484, 0/R = 0.616) less likely to leave the plan and 43% (b = -0.555, 0/R = 0.574)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
73
less likely to die in the 12 months post intervention than the rest of the eligible
population.
Table 15. Logistic Regression of Eligibles Age 85+ vs. Rest of Study Population
12 Mouths Post-Iuterveutiou
beta p-value Odds Ratio
Constant -14.799 0.000 0.000
Demographics
Age 0.165 0.000 1.179
Female
-0.510 0.000 0.600
Health status
Number of diagnoses -0.073 0.121 0.930
Heart -0.362 0.018 0.696
Cancer 0.123 0.487 1.131
Cerebrovascular -0.261 0.227 0.770
Pneumonia & COPD -0.211 0.180 0.810
Fractures 0.078 0.736 1.082
Osteoarthritis 0.055 0.734 1.056
Diabetes -0.903 0.000 0.405
Prostate 0.379 0.088 1.461
Kidney 0.202 0.568 1.224
Hypertension -0.691 0.000 0.501
Utilization
PCP use 0.127 0.475 1.135
Specialist use -0.197 0.202 0.821
Hospital use -0.009 0.957 0.991
ERuse -0.147
0.326 0.864
Study Status
Left -0.484
0.006 0.616
Deceased -0.555
0.006 0.574
Nagelkerke Pseudo R ^
0.255
Objective 2: Establish a linkage for seniors to access HCBS
The Care Advocate intervention was a consumer-directed, prevention-focused,
telephone contact over a 12-month period. Care advocates were three Master’s level
social workers who worked full time as employees of the social service agencies.
Care advocates completed a comprehensive psychosocial/functional assessment to
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
74
determine short-term, immediate care needs and recommend supportive services for
future use. Members obtained resource assistance spanning the continuum of long
term care service delivery.
Care advocates developed relationships with a variety of community providers
as well as medical case management staff within the medical groups to ensure the
efficient delivery of services and attempt to minimize costs. Care advocates also
discussed the range of long-term care services available, asked about member
preferences, and encouraged members to consider long range planning. Taking both
consumer preferences and care advocate observations into account, intervention
members were offered information about and direct linkages to HCBS.
Figure 5 presents the six steps of the intervention process, taken from
statements made at the care advocate focus group meeting, conducted after the study
was concluded. Members who requested to contact resources themselves received
necessary information over the phone. For others, the care advocate offered to link
members to resources. Members received a letter detailing the resources discussed
and information about follow up procedures.
Members received a follow up call within one week of the assessment process
and monthly follow up calls for the next 11 months to monitor referrals and provide
additional information as needed. They were free to contact the care advocate at any
time with questions, concerns, and crisis issues, and for further consultation. In the
tenth and eleventh months, CAs reminded members that the intervention was nearly
over. Upon completion of the intervention phase, members received additional
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
75
referrals to ensure appropriate, ethical termination of the participant-care advocate
relationship.
Figure 5. Intervention Process
Step 1
Client returns informed consent to HMO
HMO sends client demographic information to CA
CA calls client to discuss study and answer questions.
Step 2
CA completes phone assessment (client or designated
family member)
CA determines whether client meets home visit criteria
CA discusses HCBS resources and provides appropriate
referrals ________________________
Step 3
CA sends letter to client listing referred HCBS resources
CA sends memo to primary care physician reporting
outcome of assessment
CA calls client after one week to follow up____________
Step 4
CA calls client on monthly basis (for next 11 months)
a) completes follow up assessment
b) monitors HCBS utilization
c) provides additional appropriate service resources
Step 5
(Two months before termination date)
CA reminds client of impending study termination and
addresses any concerns
(One month before termination date)
CA reminds client of termination date and addresses
additional or continuing concerns___________________
Step 6
(At last monthly follow up call)
CA discusses termination and final referrals with client
and completes closing questionnaire
(After final follow-up call)
CA sends termination letter to client with list of HCBS
resources for future reference
Figure 6 depicts patterns of service utilization. Among those in the ITT group
(N=389), 118 members were excluded. These members died (N = 8), left the plan (N
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
76
= 14), or declined to continue (N = 96). The remaining 271 members received a
telephone assessment. After assessment, but before the information and referral
intervention, 20 members either died (N = 4) or left the plan (N = 16). In all, 251
members received HCBS referrals over the 12-montb intervention period.
During the 12-montb intervention, care advocates made 3,323 referrals, 782 of
which were accepted (23.5%). For study purposes, acceptance of an HCBS referral
meant the member or a representative accepted the information, whether for
immediate or future utilization. All members (or their representatives) who
participated in the information and referral intervention accepted at least one referral.
Utilization of a referral to HCBS meant that the member or a representative actually
contacted the service at least once. Of those who accepted a referral, 87 (34.7%) did
not contact the referral source but took the information for future reference only. The
remaining 164 members used at least one service referral. Supportive services were
accepted by 216 of intervention members (86%); but, once accepted, other services
(e.g., adaptive equipment, nutrition, and transportation) received higher utilization
rates.
Figure 6. Variations in Non-Insured Home and Community-Based Services
ITT (N = 390) -
Assessed (N = 271)
Accepted (N = 251)
Info only (N = 87)
Used HCBS (N = 164)
HCBS categories Accept Use
Supportive services 216(86.1%) 110(21.3%)
Home safety 146 (58.2%) 43 (29.5%)
Transportation 120(47.8%) 43 (35.8%)
Nutrition 111(44.2%) 43 (38.7%)
In-home care 108 (43.0%) 23 (21.3%)
Adaptive equip.________ 81(32.3%) 37(45.7%)
Totals 782 (23.5%) 299 (38.2%)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
77
At the care advocate focus group meeting, care advocates observed that HCBS
supportive services requiring higher out-of-pocket expenditures were less likely to be
used than referrals requiring less out-of-pocket expense (or none at all), such as
adaptive equipment, nutrition, or transportation
Table 16 presents acceptance and utilization rates of referrals back to the
medical groups and to the health plan. Care advocates referred 133 members back to
their medical groups (17% of first referrals). This occurred when requested by
members. Members utilized 74% of medical group referrals.
Table 16. Variations in Medical Group and Health Plan Services Utilization
_______ In-Plan Referrals Accepted____________ Used____________
Medical Group 115 (45.8%) 85 (73.9%)
Health Plan 18_ _ (7.0%)_________13 (12.2%)___________
Totals 133 (17.0%) 98 (73.7%)
Objective 3: Create database for referrals, follow-up, and other data.
PacifiCare programmers developed a database tool specifically for the Care
Advocate Program, using suggestions from all partners. Care Advocates used the
newly created database to track member responses and to input assessment and
referral information over the 12-month intervention period. At the post-study focus
group meeting, CAs described some early technical problems; but, overall, found the
database to he a valuable tool for tracking referrals, member compliance, and follow
up.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
78
Objective 4: Impact on medical group physicians or resources
CAs informed medical groups which patients were involved in the study; but
no physicians were contacted directly. As noted above, CAs referred members hack to
medical groups. Table 17 compares the demographics, number of diagnoses and
service use during the 12 months of the intervention for who used referrals to medical
groups (N=85) and those who did not (N=186). Chi-square-tests revealed no
significant differences between groups in PCP use {')(2 = .155; p = .694) or use of
specialist services (^2 = 1.890; p = .169). There were significant differences in use of
hospital services ()(2 = 6.152; p = .013), but not ER services (x2 = 3.269; p = .071).
These findings would suggest that referrals back to medical groups did not increase
physician service use; but hospital service use was significantly lower for those who
used referrals to medical groups.
Table 17. Comparisons of Members Referred to Medical Groups or Not
12 Months During-Intervention
Used Not Used
(N = 85) (N=186) t-value p-value
Age 81.20 (7.92) 82.99 (7.92) -1.936 0.054
Female 58 (68.2%) 125 (67.2%) 0.028 0.866
# Diagnoses 2.60 (3.65) 2.34 (6.65) 1.373 0.171
PCP Use 83 (97.6%) 180 (96.8%) 0.155 0.694
Specialist Use 77 (90.6%) 157 (84.4%) 1.890 0.169
Hospital Use 32 (37.6%) 43 (23.1%) 6.152 0.013
ERUse 28 (32.9%) 42 (22.6%) 3.269 0.071
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
79
CHAPTER 5: PROPENSITY SCORES
Introduction
PacifiCare data analysts performed a randomization of eligible members into
study and control groups before invitations to participate in the Care Advocate
Program were sent out. Sending invitations to both control and study group members
was not an option for PacifiCare. Lura Aheam, Care Advocate Program manager, at a
meeting of program partners, stated that from PacifiCare’s perspective, sending
invitations to all eligible members and subsequently randomizing members into study
and control groups was unacceptable because following that protocol would result in
control group members being denied the intervention after they accepted the
invitation. PacifiCare administrators believed that the process of inviting members to
participate in the study and then denying them the intervention would negatively affect
member satisfaction.
In a randomized control trial, all members have an equal chance to participate
in the intervention. That was not true for this study; this omission opened the door to
the possibility of selection bias. Bias refers to systematic differences in the baseline
characteristics between control and treatment groups (Braitman and Rosenbaum,
2002). To confirm the findings of the original study, a retrospective study of the ITT
group versus a propensity score comparison group (PPS) was performed.
Propensity Score Methods
The propensity score process matched the observed covariates of the ITT
group and the comparison group (called the PPS group for propensity score) to
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
80
compensate for the possibility of selection bias. The propensity score is the predicted
probability of receiving the treatment given the observed covariates. It does not
control for unobserved covariates; however, construction of propensity scores
minimizes selection bias by controlling for all observed covariates in a parsimonious
model of analysis (Braitman, 2002; Joffe, 1999).
The observed baseline characteristics (covariates) are age, gender, diagnoses
(dummy variables into ten diagnostic categories selected for their relationship to high
cost and hospitalization), and utilization (dummy variables for PCP, specialist, ER,
and hospital use). Table 18 presents ten selected diagnoses leading to high costs and
hospital utilization (Blaum et al., 1994; Blumberg, 1999; Haan, 1997; Long and
Marshall, 1999).
Table 18. Ten Diagnoses Predicting High Cost and Hospital Utilization
Diagnosis ICD-9 codes
■ Heart disease 391-392, 393-398, 402,404,410-416,420-429
■ Malignant neoplasms 140-208, 230-234
■ Cerebrovascular diseases 430-440
■ Pneiunonia and COPD 465-466, 480-486, 490-492
■ Fractures 800-829
■ Osteoarthritis 715
■ Diabetes mellitus 250
■ Hyperplasia of prostate 600
■ Kidney disease 584-585
■ Hypertension 401-405
With ITT analysis, propensity scores are particularly useful to adjust for cases
of noncompliance or attrition (Petkova and Teresi, 2002). Propensity scores, derived
from logistic regressions, use observed characteristics of the treatment group as
predictors. The logistic regression to determine predicted probability scores was as
follows:
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
81
ITT/Rest of sample = age, gender, number of diagnoses, medical group, PCP
use, specialist use, hospital use, ER use, diagnoses (dummy variables for ten
selected diagnoses).
The basic model was:
e(X) - (prob(Z-l \X) where, e(X) is the propensity score, Z = ITT (ITT = 1;
rest of sample = 0), andX= covariates (Joffe, 1999).
Propensity Score Results
Baseline Characteristics
Table 19 indicates significant differences between groups in diagnosis of
cancer (ITT = 18.5% and Propensity = 12.9%; = 4.705; p = .030) and number of
specialist encounters (Z-score = -2.366; p = .018). Significant differences between
groups in diagnoses of cancer were unexpected, since cancer diagnoses were included
in the propensity score process and were controlled. Number of specialist visits
(specialist count) was not included in the propensity score process, only specialist use
(or not). No other variables were determined to differ significantly at baseline (12
months pre-intervention).
Member Satisfaction
Only ITT and control members were targeted for the after-intervention
satisfaction survey; therefore, no member satisfaction statistics were available for the
PPS group and no comparisons pre- to after the intervention were available.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
82
Table 19. Baseline Comparisons of Intent-To-Treat and Propensity (PPS) Groups
ITT
(n = 389)
PPS
(n = 434) Chi. Sq. t-test p-value
Demographics
Age 83.04 (7.16) 82.67 (7.43) 0.380 0.704
Gender (Female) 250 (64.3%) 284 (65.4%) 0.460 0.498
Health status
Number of diagnoses 4.07 (3.13) 3.71 (3.38) 0.838
Heart 188 (48.3%) 195 (50.1%) 0.252 0.616
Cancer 72(18.5%) 50 (12.9%) 4.705 0.030
Cerebrovascular 58 (14.9%) 50 (12.9%) 0.688 0.407
Pneumonia & COPD 119(30.6%) 114 (29.3%) 0.153 0.696
Fractures 32 (8.2%) 29 (7.5%) 0.160 0.689
Osteoarthritis 107 (27.5%) 126 (32.4%) 2.212 0.137
Diabetes 77 (19.8%) 69(17.7%) 0.540 0.463
Prostate 33 (8.5%) 27 (6.9%) 0.650 0.420
Kidney 11 (2.8%) 13 (3.3%) 0.172 0.678
Hypertension 245 (63.0%) 253 (65.0%) 0.357 0.550
Utilization
PCP use 378 (97.2%) 398 (91.7%) 0.959 0.327
Specialist use 334 (85.9%) 357 (82.3%) 0.639 0.424
Hospital use
126 (32.4%) 142 (32.7%)
1.762 0.184
ERuse 100 (25.7%) 133 (30.6%) 0.171 0.679
Mann-Whitney U Tests Rank Rank Z-score p-value
PCP coimt 429.18 396.6 -1.418 0.156
Specialist coimt 422.79 402.33 -2.366 0.018
Hospital admiss. count 409.17 414.54 -1.245 0.213
Hospital days coimt 407.76 415.8 -0.739 0.460
ER count 400.41 422.39 -0.182 0.856
Member Retention
Table 20 presents t-tests for differences in member retention (active, deceased,
left) by study group (ITT versus PPS) during the 12-month intervention. Between the
ITT and PPS groups there were significant differences in death rate (ITT = = 26 versus
PPS = 50; t-value = -2.910; p = .004) during the study period. ITT members left the
plan at a about the same rate as control members (ITT = 53, PPS = 56; t-value = -.309;
p-value .757).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
83
Table 20. Differences in Member Retention by Study Group: During Intervention
During Intervention
ITT PSS
(N = 389) (N =389) t-value p-value
Active 310 (79.7%) 283 (72.8%) 2.278 0.023
Deceasd 26 (6.7%) 50 (12.9%) -2.910 0.004
Left 53 (13.6%) 56 (14.4%) -0.309 0.757
Table 21 presents t-tests for differences in member retention (active, deceased,
left) by study group (ITT versus PPS) in the 12-month post-intervention period. There
were no significant differences among those who died or left the plan by study group
(ITT versus PPS) in the post-intervention period.
Table 21. Differences in Member Retention by Study Group: Post-Intervention
12 Months Post-Intervention
ITT
(N = 310)
PPS
(N =283) t-value p-value
Active 234 (46.0%) 207 (46.9%) 1.956 0.051
Deceasd 30 (7.7%) 32 (8.2%) -0.264 0.792
Left 46(11.8%) 44(11.3%) 0.224 0.823
Utilization o f Health Care Services
Table 22 presents multinomial logistic regressions of utilization change (PCP,
specialist, hospital, and ER) from 12 months pre-intervention to during the 12-month
intervention on ITT/PPS groups, controlling for months in study, and death during the
intervention. Intent-to-treat members were 58.5% more likely to have increased PCP
service utilization than PPS members. There were no other significant differences
between ITT/PPS study groups in increased or decreased utilization, compared to no
change in utilization.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CD
■ D
O
Q .
C
o
CD
Q .
■ D
CD
C/)
W
o'
o
o
5
CD
O
O
■ D
c q '
o
O ’
o
CD
■ D
O
Q .
c
a
o
Q
■ D
O
CD
Q .
■ D
CD
( / )
( / )
Table 22. Multinomial Logistic Regressions; Utilization Change Scores Pre- to During Intervention
Dependent Variables: ITT/Control Group* (N=778)
DECREASED
Utilization
Independent
Variables
beta
PCP
Change
N = 87
P-
value O/R
Specialist
Change
N = 90
P-
beta value O/R
Hosp. Admission
Change
N = 48
P-
beta value O/R
Hospital Days
Change
N = 47
P-
beta value O/R
ER Encounters
Change
N = 50
P-
beta value O/R
Intercept 0.394 0.349 -0.818 0.072 -1.595 0.007 -2.167 0.001 -1.385 0.015
Months In -0.085 0.000 0.919 -0.038 0.008 0.963 -0.041 0.030 0.960 -0.014 0.248 0.976 -0.012 0.019 0.958
Died 0.133 0.695 1.143 0.155 0.683 1.167 0.508 0.276 1.662 1.064 0.025 2.897 0.213 0.651 1.237
ITT/PPS -0.031 0.898 0.969 -0.043 0.854 0.958 0.103 0.737 1.108 0.007 0.998 1.001 -0.066 0.825 0.936
INCREASED PCP Specialist Hosp. Admission Hospital Days ER Encounters
Utilization Change Change Change Change Change
Independent N = 94 N = 74 N = 49 N = 36 N = 49
Variables
P- P- P- P- P-
beta value O/R beta value O/R beta value O/R beta value O/R beta value O/R
Intercept -2.527 0.000 -3.468 0.000 -2.547 0.000 -2.810 0.001 -3.301 0.000
Months In 0.015 0.424 1.015 0.036 0.101 1.037 -0.047 0.829 0.995 -0.015 0.557 0.985 0.014 0.554 1.014
Died -0.437 0.471 0.646 0.705 0.204 2.024 1.083 0.031 2.954 1.458 0.007 4.229 1.135 0.039 3.112
ITT/PPS 0.461 0.042 1.585 0.290 0.246 1.336 -0.212 0.484 0.809 0.047 0.894 1.048 0.186 0.535 1.205
^Likelihood of increased/decreased utilization from pre- to post-intervention, no change (referent).
00
85
Table 23 presents multinomial logistic regressions of utilization change (PCP,
specialist, hospital, and ER use/not use) from 12 months pre-intervention to 12 months
post-intervention comparing study groups, controlling for months in study, and died
post-intervention. There were no significant differences in the likelihood of increased
or decreased utilization patterns by study group compared to the likelihood of no
change in utilization.
Propensity Score Conclusions
At baseline, the propensity group differed from the ITT group in number of
diagnoses of cancer (x^ = 4.705; p = .030) and number of specialist encounters (Z-
score =-2.366; p = .018). Similarly, the ITT and control groups differed in number of
diagnoses of cancer, but these differences were not significant (x^ = 3.003; p = .088).
Further, there were differences between ITT and PPS groups in number of specialist
encounters (Z-score = -2.366; p = .018) and in PCP use (x^ = 11.387; p < .001).
Overall, although there were significant differences at baseline between groups, these
differences were not considered substantive.
The ITT group did show substantive differences in mortality rates
(significantly lower) during the intervention from both control (t-value = -3.905; p =
.000) and PPS groups (t-value = -2.910; p = .004). The effect was lost when
comparing study groups (IIT/control or ITT/PPS) in mortality rates in the 12 months
post-intervention. This suggests the possibility that the Care Advocate intervention,
similarly to effects reported in other studies (Boult et al., 1994; Shapiro and Taylor,
2002), may have influenced mortality rates while it was ongoing.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CD
■ D
O
Q .
C
o
CD
Q .
■ D
CD
C/)
W
o'
o
o
3
CD
O
O
■ D
c q '
o
O ’
o
CD
■ D
O
Q .
c
a
o
Q
■ o
o
CD
Q .
■ D
CD
( / )
( / )
Table 23. Multinomial Logistic Regressions: Utilization Change Scores Pre- to Post-Intervention
Dependent Variables: ITT/Control Group* (N=593)
DECREASED
Utilization
Independent
Variables
PCP
Change
N = 24
P-
beta value O/R
Specialist
Change
N = 86
P-
beta vaiue O/R
Hosp. Admission
Change
N = 36
P-
beta value O/R
Hospital Days
Change
N = 25
P-
heta value O/R
ER Encounters
Change
N = 44
P-
beta value O/R
Intercept -11.211 0.055 -11.983 0.000 -10.017 0.000 -11.009 0.001 -7.462 0.000
Months In 0.323 0.185 1.199 0.287 0.000 1.332 0.202 0.006 1.223 0.218 0.018 1.243 0.139 0.004 1.157
Died 0.629 0.591 1.594 1.201 0.009 3.324 2.361 0.000 10.596 2.579 0.000 13.186 -0.279 0.000 5.766
Control ITT -0.130 0.722 0.861 0.206 0.382 1.229 -0.143 0.689 0.867 -0.313 0.467 0.731 -0.039 0.386 0.757
INCREASED PCP Specialist Hosp. Admission Hospital Days ER Encounters
Utilization Change Change Change Change Change
Independent N = 47 N = 36 N = 37 N = 20 N = 32
Variables P- P- P- P- P-
beta value O/R beta value O/R beta value O/R beta value O/R beta value O/R
Intercept 0.332 0.089 -0.391 0.290 -0.452 0.219 -1.155 0.014 -0.446 0.231
Months In -0.092 0.000 0.854 -0.075 0.000 0.928 -0.070 0.000 0.933 -0.069 0.000 0.933 -0.065 0.000 0.937
Died 0.334 0.858 1.085 0.862 0.031 2.369 0.586 0.177 1.797 0.631 0.258 1.879 -0.724 0.326 0.485
Control ITT 0.634 0.078 1.798 0.307 0.164 1.360 0.169 0.512 1.184 0.139 0.685 1.149 -0.184 0.501 0.832
^Likelihood o f increased/decreased utilization from pre- to post-intervention, no change (referent).
00
o^
87
Unlike ITT/control group comparisons, there were no differences between ITT
and propensity groups in the likelihood of increased hospitalizations. Less likelihood
of ITT members to have increased hospitalizations was potentially significant since
that would have supported the Care Advocate Program goal of reducing utilization of
costly medical services. This lack of support diminishes the importance of the earlier
finding.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
8 8
CHAPTER 6: DISCUSSION
Chapter 6 discusses the findings of the Care Advocate Program. These
findings are in three parts: Part 1 describes the success of the program in answering
the three research questions that address the goals of the study (increased member
satisfaction, greater member retention, and decreased utilization of costly healthcare
services). Significant findings include reduced likelihood of hospitalization and lower
mortality rates for the ITT group during the intervention. Part 2 describes Care
Advocate Program objectives, which were the products and outcomes envisioned by
the program’s originators (create a useful screening mechanism, provide greater access
to HCBS, develop a replicable database, and maintain a minimal impact on physicians
or resources). All of these objectives were met. Part 3 is a discussion of the
implications for health care organizations and considerations for future studies.
Part 1: Goals of the Intervention
Summary o f Study Findings
Satisfaction:
Satisfaction levels among HMO members were high at baseline and did not
change significantly in either the treatment or control group from pre-intervention
baseline to after the intervention.
Member Retention
Comparisons of ITT and control groups indicated no significant differences in
the proportion of participants who left the health plan. Expectations were that
providing members with greater access to HCBS would increase member satisfaction.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
89
which would in turn increase member retention. Although there were no significant
differences in retention (i.e. proportion who left the plan) between ITT and control
group members, logistic regression analysis (see Table 11) revealed that members
with lower scores in overall satisfaction after the intervention were 58% more likely to
leave the plan, and those with lower scores in plan satisfaction after the intervention
were 69% more likely to leave the plan. These findings give substance to the notion
that improving levels of satisfaction may lead to increased retention rates.
A key finding was that ITT members had significantly lower death rates than
control members. Although the likelihood of dying during the intervention was
significantly greater for those in the control group, there were no significant
differences in the likelihood of death in the 12 months post-intervention. This
suggests that the intervention may have affected mortality rates while it was ongoing,
but not after the intervention was completed.
Utilization of Insured Services
Increased access to and utilization of uninsured HCBS were expected to reduce
utilization of insured high cost, medical services. At the end of the intervention
period, insured service utilization results were mixed: Those in the ITT group had
increased use of physician services (both primary care and specialists) and a
significantly lower likelihood of hospital admission. Since hospital use is a very high
cost service, the finding of decreased likelihood of hospitalization for ITT members is
an important outcome. That said, ITT/PPS results did not support this finding and
further research with a larger study sample may be necessary to substantiate this
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
90
outcome. There were no significant differences in the likelihood of utilization change
for any insured medical services in the 12-month post-intervention period.
Part 2: Program Objectives
Summary o f Study Findings
Objective l:1o create a screening mechanism to identify frail Medicare HMO
members
An algorithm identified frail older adult members of a Medicare-risk HMO
who would benefit from the care advocate intervention. The algorithm included aged
85 or older as an automatic eligibility criterion and also scored different levels of
inpatient hospital, ER, and pharmacy use. Prior utilization of health services was a
valuable predictor, whereas age 85 or older was less effective in identifying those who
needed home and community-based services.
It is important to note that, although the likelihood of increased utilization of
health care service increases with age (Reuben et al., 2002; Seeman, 2002), the use of
age 85 or older as sole criterion for frailty is questionable. The Michigan Choice
screening algorithm, for example, based eligibility for publicly funded HCBS on 30
client characteristics derived from the MDS-HC. Age, alone, was not a factor. Age is
most often used as a control variable (either continuous or categorical), along with
other socio-demographic variables, in multivariate regressions to determine predictors
of future decline, utilization of services, or death (Saliba et al., 2001).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
91
Objective 2: To establish a linkage for seniors to access HCBS
Of the 271 members who received assessments, 251 accepted at least one
HCBS referral. The remaining 20 members either left the plan or died between
assessment and the information and referral intervention. Everyone who received the
information and referral intervention accepted at least one referral over the 12-month
intervention period. The utilization rate of accepted referrals was almost 39%.
Supportive services were the most frequently accepted HCBS referral resource. Once
accepted, however, some other services (e.g., adaptive equipment, nutrition and
transportation) had higher rates of utilization. Some participants accepted the referrals
in case they needed the information later.
The care advocate’s role was to assess needs and present available HCBS
options; changing behavior patterns was not an objective of the program. Some
participants used referrals immediately; others set the information aside for future use.
For those who used referrals after the study was over or may use those referrals at
some time in the future, the care advocates were still an important link. Care
advocates encouraged and supported members to make effective decisions within a
consumer choice model (Wilber et al., 2003).
Objective 3: To create a practical database tool for making referrals, foliowing-
up, and collecting referral, compliance, and other data.
The creation of a suitable database to track assessments, referrals, and follow-
up data was accomplished over time and with several false starts. It was not until the
final year of the study that care advocates were able to enter assessment, referral, and
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
92
follow-up data into the new system At the care advocates’ focus group meeting, they
commented that, although early efforts were fraught with technical problems, the final
database proved to be both workable and efficient.
Objective 4: Demonstrate that linking medical and HCBS can occur without
major impact on medical group physicians or resources.
Medical group physicians were notified that patients were involved in the
study, but no physicians were contacted directly by care advocates. Referrals to
medical groups (N = 115; 42.5%) occurred when members inquired about medical
services. Referred medical services were utilized about 74% of the time. Comparisons
of insured medical service utilization between those who used medical group referrals
and others in the intervention group (whether referred to medical groups or not)
indicated no significant differences in physician service use (PCP or specialist). At
the care advocate focus group meeting, care advocates recommended that, if the
program were replicated, the program model should include greater involvement of
PCPs, including the development of clinical assessment guidelines for referral into the
program and placement of care advocates in physicians’ offices for some part of each
month.
Part 3: Implications for health care organizations and considerations for
future studies
Several important findings should encourage replication of the Care Advocate
Program. The collaborative nature of the Care Advocate Program holds a promise for
integrated service delivery in the care of frail or disabled older adults by
demonstrating that providers and insurers with disparate processes and goals could
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
93
work together in delivering health care services. The finding that mortality rates were
significantly lower for ITT members during the intervention suggests the importance
of social contact; active support that keeps frail older adults engaged in life - if only
through a telephone case management link. In addition, the likelihood of increased
hospital admissions was significantly less for ITT members than for control members.
This finding, while not verified by propensity score analyses, provides some incentive
to develop integrated health systems to improve the health and well being of frail and
disabled older adults while avoiding some of the human and financial cost of
avoidable hospitalizations.
Moreover, the Care Advocate Program tested the efficacy of the PacifiCare
frailty algorithm, currently implemented in PacifiCare’s targeted case management
program. The finding that referrals back to medical groups did not increase utilization
of physician services but did increase hospital use may indicate that those in need of
more intense, acute care may have sought medical group referrals. The consumer
choice model seems to have been effective for these older adults.
Future research should focus more explicitly on whether a consumer
direction/choice model, with periodic assessments, monthly contacts, and improved
access to both medical services and HCBS can significantly influence the wellness and
health outcomes of targeted, high-risk, older persons over sustained periods.
Successfully targeting frail, older adults who are willing to accept behavioral change,
willing to become involved in their own care planning, and ready to accept appropriate
referrals, may be the next step in bridging the medical and social service chasm.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
94
REFEP^NCES
Administration on Aging. (2002). History o f the Older Americans Act. Retrieved July,
2003, from http://www.aoa.gov/about/legbudg/oaa/legbudg oaa.asp
Aliotta, S. L., Clarke, J., & Paulman, T. F. (1998). Case management assessment and
planning for high-risk Medicare HMO members. Care Management Journal,
4(2), 86-91.
Alkema, G. E., Shannon, G. R., & Wilber, K. H. (2003). Using interagency
collaboration to serve older adults with chronic care needs: The care advocate
program. Family & Community Health, 25(3), 221-229.
Allen, H. M., Jr., & Rogers, W. H. (1997). The consumer health plan value survey:
'Ko\m6.\SNO.{commQn\\. Health Affairs, 75(4), 156-66.
Andolina, K., Zander, K., & Vogenberg, F. (2001). Case management: An overview of
challenges and opportunities. Hospital Pharmacy, 35(1), 15, 16, 19, 20, 22-24,
27.
Bender, R. H., Lance, T. X., & Guess, L. L. (2003). Including disenrollees in CAHPS
managed care health plan assessment reporting. Health Care Finance Review,
25(1), 67-79.
Benjamin, A. E. M., R. E. (2001). Age, consumer direction, and outcomes of
supportive services at home. Gerontologist, 41{5), 632-42.
Bemabei, R., Landi, F., Gambassi, G., Sgadari, A., Zuccala, G., Mor, V., et al. (1998).
Randomised trial of impact of model of integrated care and case management
for older people living in the community. British Medical Journal, 375(7141),
1348-51.
Binstock, R. H. (1991). From the great society to the aging society: 25 years of the
Older Americans Act. Generations, XV(3), 11-18.
Blaum, C. S., Liang, J., & Liu, X. (1994). The relationship o f chronie diseases and
health status to the health services utilization of older Americans. Journal o f
the American Geriatrics Society, 42(10), 1087-93.
Blumberg, M. S. (1999). Risk adjustment for Medicare. New England Journal o f
Medicine, 340(19), 1514-5.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
95
Booth, M., Fralich, J., & Saucier, P. (1997). Integration o f Acute and Long-Term Care
for Dually Eligible Beneficiaries Through Managed Care (MMIP Technical
Assistance Paper No. 1): University of Maryland Center on Aging.
Borst, W. M. (2002). A Conceptual Framework of Frailty: A Review. Gerontologist,
57A{5), M283-M288.
Boult, C., Boult, L., Murphy, C., Ebbitt, B., Luptak, M., & Kane, R. L. (1994). A
controlled trial of outpatient geriatric evaluation and management. Journal o f
the American Geriatrics Society, 42(5), 465-70.
Boult, C., Boult, L., & Pacala, J. T. (1998). Systems of care for older populations of
the future. Journal o f the American Geriatrics Society, 46(4), 499-505.
Boult, C., Boult, L. B., Morishita, L., Dowd, B., Kane, R. L., & Urdangarin, C. F.
(2001). A Randomized Clinical Trial of Outpatient Geriatric Evaluation and
Management. Journal o f the American Geriatrics Society, 49(4), 351-359.
Boult, C., Rassen, J., Rassen, A., Moore, R. J., & Robison, S. (2000). The effect of
case management on the costs of health care for enrollees in Medicare Plus
Choice plans: A randomized trial. Journal o f the American Geriatrics Society,
48(8), 996-1001.
Boult, C., Rassen, J., Rassen, A., Moore, R. J., & Robison, S. (2000). The effect of
case management on the costs of health care for enrollees in Medicare Plus
Choice plans: A randomized trial. Journal o f the American Geriatrics Society,
48(8), 996-1001.
Braitman, L. E. R., P. R. (2002). Rare Outcomes, Common Treatments: Analytic
Strategies Using Propensity Scores. Annals o f Internal Medicine, 137(8), 693-
695.
Buchmueller, T. C. F., P.J. (1997). The effect of price on switching among health
plans. Journal o f Health Economics, 16(2), 231-247.
Carlson, M. J., Blustein, J., Fiorentino, N., & Prestianni, F. (2000). Socioeconomic
status and dissatisfaction among HMO enrollees. Medical Care, 38(5), 508-16.
Centers for Medicare & Medicaid Services. (1998, July). Evaluation o f the Program
for All-Inclusive Care fo r the Elderly (PACE) Demonstration: The Impact on
Participant Outcomes. Retrieved January 18, 2004 from
http://www.cms.hhs.gov/researchers/report/1998/irvine.pdf.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
96
Centers for Medicare & Medicaid Services. (2002). Statistics, data, and research
information, from http://cms.hhs.gon/researchers/statsdata.asp
Centers for Medicare & Medicaid Services. (2004). What Health Care Services do
Medicare Beneficiaries Receive? Retrieved January 15, 2004, from
http://www.cms.hhs.gov/mcbs/MCBSsrc/1999/99cbc3c.pdf
Challis, D., & Hughes, J. (2002). Frail old people at the margins of care: Some recent
research findings. British Journal o f Psychiatry, 180,126-30.
Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (3rd ed.).
Hillside: Lawrence Earlbaum Associates.
Congressional Budget Office. (1995). Report to Congress. Retrieved May, 2001, from
www.cbo.gov
Crimmins, E. M., & Saito, Y. (2001). Trends in healthy life expectancy in the United
States, 1970-1990: Gender, racial, and educational differences. Social Science
& Medicine, 52(11), 1629-41.
Crimmins, E. M., Saito, Y., & Reynolds, S. L. (1997). Further evidence on recent
trends in the prevalence and incidence of disability among older Americans
from two sources: The LSOA and the NHIS. Journals o f Gerontology:
Psychological Sciences & Social Sciences, 52(2), S59-71.
Cunningham, P. J., & Kohn, L. (2000). Health plan switching: Choice or
circumstmce? Health Affairs, 19(3), 158-64.
Cutler, D. M. S., L. (1998). Demographics and medical care spending: Standard and
non-standard effects (NBER Working paper No. No. 6866). Cambridge:
National Bureau of Economic Research.
Davies, A. R., & Ware, J. E., Jr. (1988). Involving consumers in quality of care
assessment. Health Affairs, 7(1), 33-48.
Davis, K., Collins, K. S., Schoen, C., & Morris, C. (1995). Choice matters: Enrollees'
views of their health plans, [comment]. Health Affairs, 14(2), 99-112.
Dellana, S. A., & Glascoff, D. W. (2001). The impact of health insurance plan type on
satisfaction with health care. Health Care Management Review, 26(2), 33-46.
Department of Health and Human Services. (2003). Social Services Block Grants.
Retrieved June 23, 2003, 2003, from
http://www.acf.dhhs.gov/programs/ocs/ssbg/docs/overv.htm
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
97
Doty, P. (2002). Cost-Effectiveness o f Home and Community-Based Long-Term Care
Services. Washington, D.C.: U. S. Department of Health and Human Services.
Druss, B. G., Schlesinger, M., Thomas, T., & Allen, H. (2000). Chronic illness and
plan satisfaction under managed care. Health Affairs, 19(1), 203-9.
Dudley, R. A., & Luft, H. S. (2001). Managed care in transition. New England Journal
o f Medicine, 344(\A), 1087-92.
Dunn, S. A., Sohl-Kreiger, R., & Marx, S. (2001). Geriatric case management in an
integrated care system. Journal o f Nursing Administration, 31(1), 60-2.
Feldman, P. H., & Kane, R. L. (2003). Strengthening research to improve the practice
and management of long-term care. Milbank Quarterly, 81(2), 179-220,171.
Firshein, J. S. & Sandy, L. G. (2000). The changing approach to managed care. In
Isaacs, S. L. & Knickman (Eds.), To Improve Health and Health Care, 2001
(Chapter 4). Princeton, NJ: The Robert Wood Johnson Foundation.
Fishman, P., Von Korff, M., Lozano, P., & Hecht, J. (1997). Chronic care costs in
managed care. Health Affairs (Millwood), 16(3), 239-47.
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). "Mini-mental state". A
practical method for grading the cognitive state of patients for the clinician.
lournal o f Psychiatric Research, 12(3), 189-98.
Fox, P. D., Etheredge, L., & Jones, S. B. (1998). Addressing the needs of chronically
ill persons under Medicare. Health Affairs (Millwood), 17(2), 144-51.
Fries, B. E., Shugarman, L. R., Morris, J. N., Simon, S. E., & James, M. (2002). A
screening system for Michigan's home- and community-based long-term care
programs. Gerontologist, 42(A), 462-74.
Fuchs, V. R. (1998). Ethics and economics. Antagonists or allies in making health
policy? Western Journal o f Medicine, 765(3), 213-6.
Gage, B. (1999). Impact of the BBA on post-acute utilization. Health Care Financing
Review, 20(A), 103-26.
Gilden, D. (1997). The medical merry-go-round: Drugs for 1998 are new but not
novel. General Medical Health Care Treatment Issues, 72(1), 18-25.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
98
Haan, M. N., Selby, J. V., Quesenberry, C. P., Schmittdiel, J. A., Fireman, B. A., &
Rice, D. P. (1997). The Impact of Aging and Chronic Disease on use of
Hospital and Outpatient Services in a large HMO: 1971 - 1991. Journal o f the
American Geriatrics Society, 45(6), 661-61A.
Harrington, C., Lynch, M., & Newcomer, R. J. (1993). Medical services in social
health maintenance organizations, [see comments.]. Gerontologist, 33(6), 790-
800.
Harrington, C., & Newcomer, R. J. (1991). Social health maintenance organizations'
service use and costs, 1985-89. Health Care Financing Review, 12(f), 37-52.
Hays, R. D., Shaul, J. A., Williams, V. S., Luhalin, J. S., Harris-Kojetin, L. D.,
Sweeny, S. F., et al. (1999). Psychometric properties of the CAHPS 1.0 survey
measures. Consumer Assessment of Health Plans Study. Medical Care, 37(3
Suppl), MS22-31.
Hendriksen, C., Lund, E., & Stromgard, E. (1984). Consequences of assessment and
intervention among elderly people: A three year randomised controlled trial.
British Medical Journal (Clinical Research Edition), 289(6A51), 1522-4.
Houck, P. R., Mazumdar, S., Liu, K. S., Reynolds, C. F. (2003). A Pragmaticlntent-
to-Treat Analysis Using SAS®. Retrieved august 14, 2003, 2003, from MD
http://www.nesug.org/Proceedings/nesiig00/st/st9008.pdf
Huher, D. L. (2000). The Diversity of Case Management Models. Case Management:
Managing the Process o f Patient Care, 5((6)), 248-255.
Huher, D. L., Hall, J. A., & Vaughn, T. (2001). The Dose of Case Management
Intervention. Case Management, 6(3), 119-126.
Iglehart, J. K. (1999). The American health care system— Medicare. New England
Journal o f Medicine, 340(A), 327-32.
Inouye, S. K., Peduzzi, P. N., Robison, J. T., Hughes, J. S., Horwitz, R. 1., & Concato,
J. (1998). Importance of functional measures in predicting mortality among
older hospitalized patients. 27P(15), 1187-93.
Ipsen, S. K., Foshinder, D., Williams, M., Wamick, M., Lertwachara, K., & Paita, L.
M. (2000). Satisfaction with managed care. Journal o f Nursing Care Quality,
75(1), 12-21.
Joffe, M. M. R., P. R. (1999). Invited Commentary: Propensity Scores. American
Journal o f Epidemiology, 150(A), 327-333.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
99
Johnson, M. F., Kramer, A. M., Lin, M. K., Kowalsky, J. C., & Steiner, J. F. (2000).
Outcomes of older persons receiving rehabilitation for medical and surgical
conditions compared with hip fracture and stroke. Journal o f the American
Geriatrics Society, 48{\\), 1389-1397.
Kane, R. L. (1994). Long-term care in the LFnited States: Problems and promise. In T.
Marmor, T. Smeeding & V. Greene (Eds.), Economic Security and
IntergenerationalJustice. Washington, D.C: The Urban Institute Press.
Kane, R. L. (1997). PACE: A model of integrated acute and long-term care. Hospital
Practice (Off Ed), 32(4), 23-4, 27, 30.
Kane, R. L. (1998). Managed care as a vehicle for delivering more effective chronic
care for older persons. Journal o f the American Geriatrics Society, 46(S),
1034-1039.
Kane, R. L., & Huck, S. (2000). The implementation of the EverCare demonstration
project. Journal o f the American Geriatrics Society, 48(2), 218-223.
Kane, R. L., Kane, R. A., & Finch, M. D. (1995). Once and future SHMOs.
Gerontologist, 35(3), 294-5.
Kane, R. L., Kane, R. A., Ladd, R. C., & Veazie, W. N. (1998). Variation in state
spending for long-term care: Factors associated with more balanced systems.
Journal o f Health Politics, Policy & Law, 23(2), 363-90.
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, 914-9.
Keeler, E. B., Robalino, D. A., Frank, J. C., Hirsch, S. H., Maly, R. C., & Reuben, D.
B. (1999). Cost-effectiveness of outpatient geriatric assessment with an
intervention to increase adherence. Medical Care, 37(12), 1199-206.
Kerr, E. A. (1995). Managed Care and Capitation in California: How do Physicians at
Financial Risk Control their own Utilization? Annals o f Internal Medicine,
123(1), 500-504.
Kerschner, H. K. (1992). Focus Groups: Three Organizing Steps for Education,
Design, and Delivery (2nd ed.). Washington, D. C.: The American Association
of International Aging.
Kunkel, S. R., & Applebaum, R. A. (1992). Estimating the prevalence of long-term
disability for an aging society. Journal o f Gerontology, 47(5), S253-60.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
100
Lamers, L. M. (1999). Pharmacy costs groups: A risk-adjuster for capitation payments
based on the use of prescribed drugs. Medical Care, 37(S), 824-30.
Lamphere, J. A., & Rosenbach, M. L. (2000). Promises imfulfilled; Implementation of
expanded coverage for the elderly poor. Health Services Research, 35(1 Pt 2),
207-17.
Landi, F., Gambassi, G., Pola, R., Tabaccanti, S., Cavinato, T., Carbonin, P. U., &
Bemabei, R. (1999). Impact of integrated home care services on hospital use.
Journal o f the American Geriatrics Society, 1430-4.
Landi, F., Onder, G., Russo, A., Tabaccanti, S., Rollo, R., Federici, S., Tua, E., Cesari,
M., & Bemabei, R. (2001). A new model of integrated home care for the
elderly: Impact on hospital use. Journal o f Clinical Epidemiology, 54(9), 968-
70.
Lawton, M. P., & Brody, E. M. (1969). Assessment of older people: Self-maintaining
and instramental activities of daily living. Gerontologist, 9(3), 179-86.
Leigh, J. P., Richardson, N., Beck, R., Kerr, C., Harrington, H., Parcell, C. L., et al.
(1992). Randomized controlled study of a retiree health promotion program.
The Bank of American Study. Archives o f Internal Medicine, 752(6), 1201-6.
Lesser, C. S., & Ginsburg, P. B. (2000). Update on the nation's health care system:
1997-1999. Health Affairs, 79(6), 206-16.
Leveille, S. G., Wagner, E. H., Davis, C., Grotbaus, L., Wallace, J., LoGerfo, M., et al.
(1998). Preventing disability and managing chronic illness in frail older adults:
A randomized trial of a community-based partnership with primary care.
Journal o f the American Geriatrics Society, 4(5(10), 1191-8.
Liberman, A., & Rotarius, T. M. (2001). Marketing in today's health care
environment. Health Care Manager, 79(4), 23-8.
Lied, T. R., Sbeingold, S. H., Landon, B. E., Shaul, J. A., & Cleary, P. D. (2003).
Beneficiary reported experience and voluntary disenrollment in Medicare
managed care. Health Care Financing Review, 25(1), 55-66.
Little, R. J., & Rubin, D. B. (2002). Statistical analysis with missing data (2 ed.).
Hoboken, NJ: Wiley-Interscience.
Long, M. J., & Marshall, B. S. (1999). The relationship between self-assessed health
status, mortality, service use, and cost in a managed care setting. Health Care
Management Review, 24(4), 20-7.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
101
Luft, H. S. (1999). Why are Physicians So Upset About Managed Care? Journal o f
Health Politics, Policy & Law, 24(5), 957-966.
Malmgren, J. A., Koepsell, T. D., Martin, D. P., Diehr, P., & LaCroix, A. Z. (1999).
Mortality, health services use, and health behavior in a cohort of well older
adults. Journal o f the American Geriatrics Society, 47(\), 51-9.
Marmor, T. R. (1988). Reflections on Medicare. Journal o f Medical Philosophy,
13(\), 5-29.
Mebane, F. (2001). Want to understand how Americans viewed long-term care in
1998? Start with media coverage. Gerontologist, 41(\), 24-33.
Meiners, M. R. (1996). The Financing and Organization o f Long-Term Care.
Baltimore: The Johns Hopkins University Press.
Meiners, M. R., Mahoney, K. J., Shoop, D. M., & Squillace, M. R. (2002). Consumer
direction in managed long-term care: An exploratory survey of practices and
perceptions. Gerontologist, 42(1), 32-38.
Mollica, R. (2003). Coordinating Services Across the Continuum of Health, Housing,
and Supportive Services. Journal o f Aging and Health, 75(1), 165-188.
Moon, M. (2001). Medicare. New England Journal o f Medicine, 544(12), 928-31.
National Center for Health Care Statistics. (2004). Medicare expenditures: Last year
o f life. Retrieved January, 15, 2004, from
http://www.cdc.gov/nchs/express.htm
National Chronic Care Consortium, N. (2001). Targeting Beneficiaries Who Are Most
At-Risk. Bloomington: Robert Wood Johnson Foundation (Medicare/Medicaid
Integration Project).
Newhouse, J. P. (2000). Switching health plans to obtain drug coverage, [comment].
JAMA, 255(16), 2161-2.
Newman, A. B., & Brach, J. S. (2001). Gender gap in longevity and disability in older
persons. Epidemiological Review, 25(2), 343-50.
Newman, A. B., Gottdiener, J. S., McBumie, M. A., Hirsch, C. H., Kop, W. J., Tracy,
R., Walston, J. D., & Fried, L. P. (2001). Associations of Subclinical
Cardiovascular Disease with Frailty. The Journals o f Gerontology: Medical
Sciences, 56A(3), M158-M166.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
102
Nourhashemi, F., Andrieu, S., Gillette-Guyonnet, S., Vellas, B., Albarede, J. L., &
Grandjean, H. (2001). Instrumental activities of daily living as a potential
marker of frailty: A study of 7364 community-dwelling elderly women (the
EPIDOS study). Journals o f Gerontology: Medical Sciences, 56(1), M448-53.
Pacala, J. T., Boult, C., Reed, R. L., & Aliberti, E. (1997). Predictive Validity of the P-
ra Instrument Among Older Recipients of Managed Care. Journal o f the
American Geriatrics Society, 45(5), 614-617.
Pacala, J. T., Boult, €., Urdangarin, C., & McCaffrey, D. (2003). Using self-reported
data to predict expenditures for the health care of older people. Journal o f the
American Geriatrics Society, 57(5), 609-14.
Petkova, E., & Teresi, J. (2002). Some statistical issues in the analyses of data from
longitudinal studies of elderly chronic care populations. Psychosomatic
Medicine, 64(3), 531-47.
Portrait, F., Lindeboom, M., & Deeg, D. (2001). Life expectancies in specific health
states: Results from a joint model of health status and mortality of older
persons. Demography, 38(4), 525-36.
Quinn, J. (1993). Successful Case Management in Long-Term Care (4th ed.). New
York: Springer Publishing Company.
Rantz, M. J., Marek, K. D., & Zwygart-Stauffacher, M. (2000). The Future of Long-
Term Care for the Chronically 1 1 1 . Nursing Administration Quarterly, 25(1),
51-58.
Reuben, D. B., Keeler, E., Seeman, T. E., Sewall, A., Hirsch, S. H., & Guralnik, J. M.
(2002). Development of a method to identify seniors at high risk for high
hospital utilization. Medical Care, 40(9), 782-93.
Reuben, D. B., Keeler, E., Seeman, T. E., Sewall, A., Hirsch, S. H., & Guralnik, J. M.
(2003). Identification of risk for high hospital use: Cost comparisons of four
strategies and performance across subgroups. Journal o f the American
Geriatrics Society (Vol. 51, pp. 615-20).
Rice, D. P. (1996). Beneficiary profile: Yesterday, today, and tomorrow. Health Care
Financing Review, 18(2), 23-46.
Rigaud, A. S. F., B. (2001). Hypertension in Older Adults. Journals o f Gerontology:
Medical Sciences, 56A(4), M217-M225.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
103
Riley, G. F., Ingber, M. J., & Tudor, C. G. (1997). Disenrollment of Medicare
beneficiaries from HMOs. Health Affairs, 16(5), 117-24.
Sahyoun, N. R., Lentzner, H., Hoyert, D., & Robinson, K. N. (2001). National Vital
Statistics Reports. Retrieved January 5, 2004, from
http://www.cdc. gov/nchs/data/agingtrends/01 death.pdf
Sales, A. E., Liu, C. F., Sloan, K. L., Malkin, J., Fishman, P. A., Rosen, A. K., et al.
(2003). Predicting costs of care using a pharmacy-based measure risk
adjustment in a veteran population. Medical Care, 41{6), 753-60.
Saliba, D., Elliott, M., Rubenstein, L. Z., Solomon, D. H., Young, R. T., Kamberg, C.
J., et al. (2001). The Vulnerable Elders Survey: A tool for identifying
vulnerable older people in the commimity. Journal o f the American Geriatrics
Society, 49{12), 1691-9.
Sandy, L. G. (2002). Homeostasis without reserve-the risk of health system collapse.
New England Journal o f Medicine. Online, 347(24), 1971-1975.
Scharlach, A., Giunta, N., Mills-Dick, K., & Taylor, T. (2001). An overview o f current
case management programs and evaluation methods for long-term care
integration (PowerPoint): UC, Berkeley.
Scharlach, A. E., Giunta, N., Mills-Dick, K. (2001). Case Management in Long-Term
Care Integration: An Overview o f Current Programs and Evaluations (Report
to the Califomia Center for Long-Term Care Integration). Berkeley: University
of Califomia, Berkeley.
Schifalacqua, M., Hook, M., O'Heam, P., & Schmidt, M. (2000). Coordinating the
care of the chronically ill in a world of managed care. Nursing Administration
Quarterly, 24(3), 12-20.
Schlesinger, M., Druss, B., & Thomas, T. (1999). No exit? The effect of health status
on dissatisfaction and disenrollment from health plans. Health Services
Research, 34(2), 547-76.
Schlesinger, M., Mitchell, S., & Elbel, B. (2002). Voices unheard: Barriers to
expressing dissatisfaction to health plans. Milbank Quarterly, 80(4), 709-55.
Schnaier, J. A., Sweeney, S. F., Williams, V. S. L., Kosiak, B., Lubalin, J. S., Hays, R.
D., & Harris-Kojetin, L. (1999). Special Issues in the CAHPS Survey of
Managed Care Beneficiaries. Medical Care, 37(3), MS69-MS78.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
104
Scully, T. (2001). National Academy o f Social Insurance. Paper presented at the "LTC
and Medicare Policy: Can We Improve the Continuity of Care?" Washington,
D.C.
Seeman, T. E. C., X. (2002). Risk and Protective Factors for Physical Functioning in
Older Adults With and Without Chronic Conditions: MacArthur Studies of
Successful Aging. Journals o f Gerontology: Social Sciences, 57B(3), S135-
S144.
Shapiro, A., & Taylor, M. (2002). Effects of a community-based early intervention
program on the subjective well-being, institutionalization, and mortality of
low-income elders. Gerontologist, 42(3), 334-41.
Sheiner, L. B. (2002). Is intent-to-treat analysis always (ever) enough? British Journal
o f Clinical Pharmacology, 54(2), 203-11.
Sherman, S. E., & Reuben, D. (1998). Measures of functional status in community-
dwelling elders. Journal o f General Internal Medicine, 73(12), 817-23.
Sinay, T. (2002). Access to quality health services: Determinants of access. Journal o f
Health Care Finance, 28(4), 58-68.
Sloan, K. L., Sales, A. E., Liu, C. F., Fishman, P., Nichol, P., Suzuki, N. T., et al.
(2003). Construction and characteristics of the RxRisk-V: A VA-adapted
pharmacy-based case-mix instrument. Medical Care, 41(6), 761-74.
Solomon, L. S., Zaslavsky, A. M., Landon, B. E., & Cleary, P. D. (2002). Variation in
patient-reported quality among health care organizations. Health Care
Financing Review, 23(4), 85-100.
Spillman, B. C., & Lubitz, J. (2000). The Effect of Longevity on Spending for Acute
and Long-Term Care, New England Journal o f Medicine (Vol. 342, pp. 1409-
1415).
Squillace, M. R., Firman, J. P. (2002). The Myths and Realities o f Consumer-Directed
Services for Older Persons. Retrieved June, 2003, 2003
Stanton, M. P., Walizer, E. M., Graham, J. I., & Keppel, L. (2000). Case management:
A case study. Nursing Case Management, 5(1), 37-45.
Stewart, D. W., & Shamdasani, P. N. (1998). Focus Group Research: Exploration and
Discovery. In L. B. D. J. Rog (Ed.), Handbook o f Applied Social Research
Methods (pp. 580). Thousand Oaks: Sage Publications, Inc.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
105
Stewart, S., Pearson, S., & Horowitz, J. D. (1998). Effects of a home-based
intervention among patients with congestive heart failure discharged from
acute hospital care. Archives o f Internal Medicine, /55(10), 1067-72.
Stone, R. I. (2000). Long-Term Care for the Elderly with Disabilities: Current Policy,
Emerging Trends, and Implications for the Twenty-First Century. Retrieved
January 5, 2004 from www.milbank.org/0008stone/
Strahan, G. W. (1997). An overview o f nursing homes and their current residents:
Data from the 1995 national nursing home survey (No. 280). Washington, DC:
National Center for Health Statistics, U.S. Department of Health and Human
Services.
Stuck, A. E., Aronow, H. U., Steiner, A., Alessi, C. A., Bula, C. J., Gold, M. N., et al.
(1995). A trial of annual in-home comprehensive geriatric assessments for
elderly people living in the community. New England Journal o f Medicine,
355(18), 1184-9.
Tabbush, V., & Swanson, G. (1996). Changing paradigms in medical payment.
Archives o f Internal Medicine, 756(4), 357-60.
Tahan, H. A. (1998). Case management: A heritage more than a century old. Nursing
Case Management, 5(2), 55-60; quiz 61-2.
Tilley, J., Goldenson, S., & Kasten, J. (2001). Long-term care: Consumers, providers,
and financing. Washington, D.C.: Urban Institute.
Tudor, C. G., Riley, G., & Ingber, M. (1998). Satisfaction with care: Do Medicare
HMOs make a difference? Health Affairs, 77(2), 165-76.
Tufts Managed Care Institute. (2001). Health care trends for 2001. Tufts.
Ullman, R., Hill, J. W., Scheye, E. C., & Spoeri, R. K. (1997). Satisfaction and choice:
A view from the plans. Health Affairs, 16(3), 209-17.
Vita, A. J., Terry, R. B., Hubert, H. B., & Fries, J. F. (1998). Aging, health risks, and
cumulative disability. New England Journal o f Medicine, 555(15), 1035-41.
Vladeck, B. D. (1997). The long-term future o f Medicare. Washington, D. C.:
Congressional Sub-Committee on Aging.
Ware, J. E., Bayliss, M. S., & Rogers, W. H. (1996). Differences in 4-year health
outcomes for elderly and poor, chronically ill patients treated in HMO and Fee-
for-Services Systems. 5/4M4, 276(13), 1039-1047.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
106
Weiner, J. M., Tilley, J., & Alecxih, L. M. (2002). Home and Community-Based
Services in Seven States. Health Care Financing Review, 23(3), 89-114.
Weissert, W. G. (1985). Seven reasons why it is so difficult to make community-based
long-term care cost-effective. Health Services Research, 20{A), 423-33.
Wetzler, H. P., & Cruess, D. F. (1985). Self-reported physical health practices and
health care utilization: Findings from the National Health Interview Survey.
American Journal o f Public Health, 75(11), 1329-30.
Wheatley, B., DeJong, G., & Sutton, J. P. (1997). Managed care and the
transformation of the medical rehabilitation industry. Health Care
Management Review, 22(3), 25-39.
Wilber, K. H., Allen, D., Shannon, G. R., & Alongi, S. (2003). Partnering managed
care and community-based services for frail elders: The care advocate
program. Journal o f the American Geriatrics Society, 57(6), 807-12.
Yang, Z., Norton, E. C., & Steams, S. C. (2003). Longevity and health care
expenditures: The real reasons older people spend more. Journals o f
Gerontology: Psychological Sciences and Social Sciences, 55(1), S2-10.
Yelin, E. H., Criswell, L. A., & Feigenbaum, P. G. (1996). Health care utilization and
outcomes among persons with rheumatoid arthritis in fee-for-service and
prepaid group practice settings. JAMA, 275(13), 1048-53.
Zander, K. (2002). Nursing Case Management in the 21st Century: Intervening Where
Margin Meets Mission. Nursing Administration Quarterly, 25(5), 58-67.
Zaslavsky, A. M., Shaul, J. A., Zaborski, L. B., Cioffi, M. J., & Cleary, P. D. (2002).
Combining health plan performance indicators into simpler composite
measures. 7/eaM Care Financing Review, 23(4), 101-15.
Zelman, W. A., & Berensen, R. A. (1998). The managed care blues and how to cure
them. Washington, D. C.: Georgetown University Press.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Gender and the bump: An investigation of the reminiscence effect in the Long Beach Longitudinal Study

PDF

An analysis of health risk selection and quality of care under Medicare fee -for -service and Medicare managed care health care systems

PDF

Dual eligible beneficiaries and managed care: Ensuring consumer protection through state Medicaid contracts

PDF

Heart disease among middle -aged and elderly persons in the United States: Trends and a multistate model

PDF

Assessing the cost implications of combined pharmacotherapy in the long term management of asthma: Theory and application of methods to control selection bias

PDF

Communication patterns between case management and contracted medical providers within the Social HMO

PDF

Effects of a formulary expansion of the use of SSRIs and health care services by depressed patients in the California Medicaid program

PDF

Filial expectations and social exchange patterns among older Taiwanese parents and their adult children

PDF

Using network perspective to examine the organization of community -based elder care systems across four communities

PDF

Elder abuse crisis intervention techniques and their effectiveness

PDF

Education and intelligence test scores: Predictors of dementia?

PDF

Translating research into practice: a community-based medication management intervention

PDF

A study of employee health plan choice and medical cost: Panel data probit regression and sample selection model

PDF

A methodology to identify high -risk patients with diabetes in the California Medicaid populations (Medi -Cal)

PDF

Impact of housing arrangements on social support and health status among Chinese American elderly

PDF

Health care finance reform and market failure at King/Drew Medical Center, Los Angeles: An analysis of pediatric social outcomes in an urban public safety -net hospital

PDF

Demographic diversity, team process, and team performance: Assessing moderator effects of cognitive conflict management practices and task interdependence

PDF

Service evaluation and quality improvement of a geriatric evaluation and management program

PDF

An analysis of postacute treatment and outcome differences between Medicare fee-for-service and managed care

PDF

Assessment of prognostic comorbidity in hospital outcomes research: Is there a role for outpatient pharmacy data?

Asset Metadata

Creator Shannon, George R. (author)

Core Title Evaluation of the Care Advocate Program: Bridging managed care and home community -based services

School Graduate School

Degree Doctor of Philosophy

Degree Program Gerontology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag health sciences, health care management,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Crimmins, Eileen M. (committee member), Myrtle, Robert (committee member), Wilber, Kathleen (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-534729

Unique identifier UC11340919

Identifier 3140553.pdf (filename),usctheses-c16-534729 (legacy record id)

Legacy Identifier 3140553.pdf

Dmrecord 534729

Document Type Dissertation

Rights Shannon, George R.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA