University of Southern California Dissertations and Theses

Investigating the effectiveness of a social work intervention on reducing hospital readmissions among older adults

(USC Thesis Other)

Investigating the effectiveness of a social work intervention on reducing hospital readmissions among older adults

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INVESTIGATING THE EFFECTIVENESS OF A SOCIAL WORK
INTERVENTION ON REDUCING HOSPITAL READMISSIONS AMONG OLDER
ADULTS

by

Alexis Marie Coulourides Kogan

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)

August 12, 2014

ii

DEDICATION
I dedicate this dissertation to my family, who sacrificed in many ways
during my tenure in this program and previously during my young, formative
years to give me the opportunities to pursue my dreams. This dissertation is truly
a team effort.
To my loving husband, Michael: Your endless support, encouragement,
and take-out Thai food were the fuel to my efforts, especially when my motivation
and enthusiasm abated. You are a wonderful provider that made my escapades
in higher education possible while supporting a home and family life that I could
only dream of. You keep me rooted in reality, focused, and always looking ahead
to the future; you make the journey worthwhile.
To my baby daughter, Maya: You make everything better in this world;
everything. You sacrificed unknowingly over the last 17 months and I thank you
for your unconditional love, endless amount of kisses, and extraordinary sleep
habits from just a few weeks of age; a girl after her mother’s heart. I love you and
hope that I inspire you to fulfill your highest potential in your endeavors.
To my parents, Leo and Christina, and my sister, Katrina: I cannot thank
you all enough for believing in me and trusting me to make decisions and be a
strong, independent woman, even when those decisions took me to New
Orleans, Israel, and beyond. You always impressed upon me the importance of
education and I thank you for supporting me through this journey, setting the
foundation, and affording me the tools to be successful beyond my

iii

undergraduate years. You recognized and nurtured my potential in areas I never
thought I could (or should) pursue, and as a parent, I only hope that I can do the
same for my child; it is truly a gift. Thank you for your hours of babysitting which
not only enriched Maya’s life (and hopefully your own, probably minus the
diapers…), but provided hours of guilt-free work time for me: priceless. The love,
encouragement, and support from each of you throughout my four years in this
program and 29 years of life have made me the person I am today and I hope
you are proud.
To my grandparents—Bob and Nancy (Papa and Nana) & Ann (Yaya)—
and to Aunt Kiki, my great aunt: While each of you may not have been able to
completely embrace your own aging and trajectories at times, you certainly
helped plant the seeds for my interest in aging and opened my eyes to the
systematic features of seeking and utilizing healthcare and related services faced
by older adults living in the community. Each of you also instilled upon me a
strong work ethic, love for family, and commitment to perseverance, and for that,
I thank you.

iv

ACKNOWLEDGEMENTS
I would like to thank my family and friends for their love and support
throughout my time at USC and pursuit of higher education; you make it all
possible.
Many heartfelt thanks to my dissertation committee members Dr. Kathleen
Wilber and Dr. Michael Nichol for their guidance, constructive feedback, thought-
provoking discussions, challenges presented, and support during the
development of this dissertation and other projects. I would also like to
specifically thank my faculty mentor and dissertation committee chair, Dr. Susan
Enguidanos, for her endless encouragement, advocacy, and tireless commitment
to the mentor role. She generously provided me with training opportunities that
allowed me to be successful throughout my time at USC and to develop
personally and professionally. I cannot thank her enough for her friendship,
warmth, and compassion over the years, but especially when my need to
balance work and (a growing) family became prominent. I feel lucky every day to
have had the privilege to work and train with such a productive, innovative, and
significant researcher and friend.
I would also like to thank the Enguidanos Lab and USC Gerontology
student colleagues, past and present, for their camaraderie. Specifically, Maria
Siciliano is a primary reason for my being in the PhD program in the first place,
and I cannot thank you enough for your encouragement, humor, and friendship;

v

you truly are an inspiration. Also, many thanks to Dr. Zach Gassoumis for his
generous assistance and statistical expertise with revising analyses.
Finally, I would like to thank the administrators and Huntington Memorial
Hospital and Senior Care Network involved in this study and each of my SWIFT
study interns that provided hundreds of hours of patient screening, enrollment,
data collection, and data entry over two years. It was such a pleasure working
with each of you and the struggles, laughter, and successes we shared together
are close to my heart. You all contributed to making this study and dissertation
possible and for that, I thank you: Kristie Wang, Sara McCleskey, Tameka
Brown, Ruth Barber, Rachel Piperno, Pavitra Anand, Patricia Kriek, Michelle
Pastrano, Crystal Bettenhausen-Bubulka, Kate Salvino, Kate Shanovich, Jennifer
Stone, Alice Liang, Nadia Akaweih, Lisa Saldana, Srestha Ghosh, Robert Weise,
Pamela Ramirez, and Pitcha Ratanawong

vi

TABLE OF CONTENTS

DEDICATION ........................................................................................................ii

ACKNOWLEDGEMENTS .....................................................................................iv

ABSTRACT .......................................................................................................... x

CHAPTER 1: REVIEW OF THE LITERATURE AND PROBLEM ......................... 1
Background ....................................................................................................... 1
The Problem and Significance .......................................................................... 1
State of the Problem.......................................................................................... 6
Addressing Gaps: Need and Opportunity for Research .................................. 11
Dissertation Research and Purpose ................................................................ 16
Organization of the Dissertation ...................................................................... 17

CHAPTER 2: METHODS ................................................................................... 18
Research Questions & Hypotheses ................................................................ 18
Conceptual Model ........................................................................................... 19
Study Design ................................................................................................... 23
Measures ........................................................................................................ 30
Analysis ........................................................................................................... 37

CHAPTER 3: OVERALL DEMOGRAPHIC AND HEALTH CHARACTERICTICS
........................................................................................................................... 39
Screening Results ........................................................................................... 39
Sample Description ......................................................................................... 40

vii

CHAPTER 4: PREDICTORS OF 30-DAY READMISSION AMONG AT-RISK
OLDER ADULTS ................................................................................................ 46
Introduction ..................................................................................................... 46
Methods .......................................................................................................... 46
Results ............................................................................................................ 47
Discussion ....................................................................................................... 54

CHAPTER 5: INVESTIGATING THE IMPACT OF UNMET SOCIAL SERVICE
NEEDS ON 30-DAY READMISSIONS AMONG AT-RISK OLDER ADULTS ..... 58
Introduction ..................................................................................................... 58
Methods .......................................................................................................... 59
Results ............................................................................................................ 60
Discussion ....................................................................................................... 66

CHAPTER 6: CHARACTERISTICS AND RISK FACTORS ASSOCIATED WITH
PATIENTS THAT DECLINE A HOME TRANSITIONS INTERVENTION ........... 69
Introduction ..................................................................................................... 69
Methods .......................................................................................................... 70
Results ............................................................................................................ 70
Discussion ....................................................................................................... 79

CHAPTER 7: CONCLUSIONS, LIMITATIONS, AND IMPLICATIONS FOR
POLICY AND PRACTICE ................................................................................... 82
Implications ..................................................................................................... 82
Limitations ....................................................................................................... 87
Conclusion ...................................................................................................... 88

BIBLIOGRAPHY ................................................................................................. 89

viii

LIST OF TABLES

Table 1: Overall Demographic Characteristics (n=181) ...................................... 42

Table 2: Overall Medical Conditions, Need Factors, and Health Behaviors
(n=181) ............................................................................................................... 44

Table 3: Bivariate Analyses for 30-day Hospital Readmissions (n=181) ............ 48

Table 4a: Multivariate Predictors of 30-day Hospital Readmission Accounting for
Study Group (n=181) .......................................................................................... 52

Table 4b: Multivariate Predictors of 30-day Hospital Readmission Accounting for
Study Intensity (n=181) ....................................................................................... 53

Table 5: Bivariate Analyses for 30-day Hospital Readmissions and Unmet Social
Service Needs(n=181) ........................................................................................ 61

Table 6a: Multivariate Predictors of 30-day Hospital Readmission Accounting for
Number of Unmet Social Service Needs (n=181) ............................................... 64

Table 6b: Multivariate Predictors of 30-day Hospital Readmission Accounting for
Unmet Social Service Needs in Key Areas (n=181) ........................................... 65

Table 7: Intervention Patient Demographic Characteristics (n=90) .................... 71

Table 8: Bivariate Analyses among Intervention Opt-Outs (n=90) ...................... 74

Table 9: Multivariate Predictors of Intervention Opt-out (n=90) .......................... 76

Table 10: Bivariate Analyses among Intervention Group Patients Readmitted
within 30-days ..................................................................................................... 77

Table 11: Multivariate Predictors of 30-day Hospital Readmission among
Intervention Group (n=90) .................................................................................. 80

ix

LIST OF FIGURES

Figure 1. Andersen’s Behavioral Model of Health Service Utilization—Part 4 .... 22

Figure 2. Measures ............................................................................................. 31

Figure 3. Flow Chart of Patient Screening, Enrollment, and Randomization ...... 40

x

ABSTRACT
Hospital readmissions among older adults has emerged as a significant
social problem associated with high risk for poor patient outcomes, fragmented
care, and exorbitant health care costs. In 2012, new penalties outlined in the
Affordable Care Act began to levy fines, increasing annually, on hospitals that
have high 30-day readmission rates. With the responsibility of reducing
readmissions placed solely on hospitals, many institutions have implemented
programs and services aimed at reducing readmissions. One such program is
the Social Work Intervention Focused on Transitions (SWIFT) intervention.
This dissertation aimed to investigate the effectiveness of the SWIFT
intervention on reducing 30-day hospital recidivism among at-risk older adults; a
randomized control, pilot study conducted at a large, non-profit, urban community
hospital in the Los Angeles area. Participants included cognitively intact, English-
speaking, older adults aged 65 years or more, living within a specified service
area with expected discharge back to the community, and identified as at “high-
risk” for readmission.
The intervention consisted of a maximum of two in-home visits (first in-
home assessment conduced within 48 hours after discharge) and up to four
telephone follow-up calls (maximum of six contacts). Intervention activities
included psychosocial assessment, home safety evaluation, medications
reconciliation, problem solving therapy, health goal setting, scheduling physician
follow-up appointments, and home and community based service referrals.

xi

Andersen’s Behavioral Model of Health Service Utilization was used to
guide this research. Primary data analysis among 181 randomized participants
revealed that the SWIFT social work-driven intervention was not effective at
reducing 30-day hospital recidivism among at-risk older adults. Instead, patient-
level factors such as predisposing characteristics (Caucasian race), need
(diagnosed with cancer, high acuity, need for supportive community services,
and potentially unmet food prep/shopping needs), and health behaviors (refusing
the home intervention and having had a prior inpatient stay in the previous six
months) emerged as predictors of short-term hospital readmissions.
The data presented here hold significant implications for policy and
practice. They contribute to the literature by identifying specific patient-level
factors that contribute to 30-day rehospitalization and complement a growing
body of research that suggests the primary factors driving readmissions are
patient-specific and are beyond the control of hospitals. Additionally, recent
investigations into the characteristics and number of hospitals being penalized by
the Centers for Medicare and Medicaid Services (CMS) for high readmission
rates reveal that few hospitals are escaping fines; including the SWIFT clinical
site despite a low readmission rate as compared to similar institutions. Holding
hospitals accountable for readmissions that are largely driven by factors out of
their control seems inappropriate and warrants reconsideration and further
investigation.

1

CHAPTER 1. REVIEW OF THE LITERATURE AND PROBLEM
Background
Advances in technology, modern medicine, and public health interventions
during the 21
st
century have led to reductions in death from acute disease but
have resulted in long-term management of chronic conditions well into older
adulthood. The Centers for Disease Control and Prevention defines chronic
conditions as prolonged illnesses that are not resolved spontaneously and are
rarely cured completely (CDC, 2009), and include conditions such as diabetes,
hypertension, chronic obstructive pulmonary disease (COPD), and congestive
heart failure (CHF), among others. The prevalence of Americans with chronic
conditions has been estimated at more than 145 million individuals, equating to
nearly half of our nation’s total population (CMS, 2012). Specifically among the
older adult segment of the population receiving Medicare benefits, more than half
of beneficiaries have five or more chronic conditions, with advancing age
corresponding to a greater number of diagnoses (CMS, 2012).
The Problem and Significance
Chronically ill individuals face greater challenges managing their health
and have been found to experience frequent changes in health status, resulting
in multiple transitions between care settings (namely, hospital and home) and
healthcare providers (primary care and specialist physician offices) (Robert
Wood Johnson Foundation, 2013). Transitions have been identified as vulnerable
exchange points or “handoffs” in care that are associated with increased risk for

2

hospital readmissions (Anderson, Helms, Hanson, & DeVilder, 1999; Coleman,
2003; Forster, Murff, Peterson, Gandhi, & Bates, 2003; Forster et al., 2004;
Friedman & Basu, 2004), medication errors (Coleman, Mahoney, & Parry, 2005;
Cornish et al., 2005), lapses in care and safety (Moore, Wisnivesky, Williams, &
McGinn, 2003), poor satisfaction with care (Coleman & Williams, 2007), unmet
needs (Naylor, Aiken, Kurtzman, Olds, & Hirschman, 2011), and subsequently
high rates of additional, costly health service use which could have potentially
been avoided (Cornish et al., 2005; Friedman & Basu, 2004). Along these lines,
poor care transitions have a direct, negative impact on patients and their
caregivers as it is well documented that these individuals often struggle with
inadequate support and guidance from the health care system, unmet needs,
lack of preparation for the self-management role, lack of knowledge to coordinate
care, and subpar skills at transition points (Harrison & Verhoef, 2002; Naylor et
al., 2011; van Walraven, Mamdani, Fang, & Austin, 2004; Weaver, Perloff, &
Waters, 1998).
Markers of a Poor Transition
Problems and risks associated with poor transitions are largely due to
fragmented care and poor communication. Among the early, foundational
research in this area, multiple studies identified communication between hospital
providers and patients/caregivers, and hospital providers and patients’ primary
care provider, as problematic (Coleman, 2003; Forster et al., 2003; Moore et al.,
2003), inadequate (Coleman, 2003; Forster et al., 2003; Moore et al., 2003), and
associated with a 25% increased risk of readmission (van Walraven, Seth,

3

Austin, & Laupacis, 2002). This is especially critical with hospital-to-home
discharges where patients and their caregivers are about to resume a self-care
and disease management role that may have increased in complexity since the
hospitalization.
Public Health Relevance and Older Adults
Hospitalized older adults may experience multiple transitions among
different care settings. A national evidence report of care transitions for
hospitalized older adults found that 28% were discharged to another institution,
14% discharged to home with home health, 4% died, and the remaining 54% had
routine discharges (AHRQ, 2004). Older adults are at highest risk for adverse
transitions and readmissions due to the increasing severity and number of
chronic health conditions, complexity of disease self-management behaviors,
greater number of prescription medications, and multiple medical appointments
coinciding with reduced independence (such as ceasing to drive), functional
decline, and greater dependence on caregivers. One study investigating the
prevalence of older adult care transitions within 30 days of hospital discharge
found 61% of patients had one transition, 18% had two transitions, 9% had three
transitions, 4% had four or more, and 8% died (Coleman, Min, Chomiak, &
Kramer, 2004).
Hospital readmissions among older adults are often avoidable and have
emerged as a significant social problem, with a recent study finding 20% of
hospitalized Medicare beneficiaries readmitted within 30 days and 34%
readmitted within 90 days of the index hospitalization (Jencks, Williams, &

4

Coleman, 2009). This national phenomenon of increased risk associated with
frequently transitioning between hospital, home, nursing home, and other care
settings represents skyrocketing healthcare costs and burden on the health care
system and patients alike, with estimated costs that have increased from 15
billion dollars in 2006 to a 2012 estimate of 26 billion dollars in annual Medicare
expenditures (CMS, 2012; MedPAC, 2007). Hospitals and resources in states
and counties with a high density of older adult residents such as California’s Los
Angeles County and Florida are especially burdened by this growing problem.
For example, specifically in California, which has found 30-day readmissions
among older adults to be as high as 32.7% (Allaudeen, Schnipper, Orav,
Wachter, & Vidyarthi, 2011), reducing avoidable hospitalizations would save
Medicare and Medi-Cal 227 million dollars in a single day (Grant & California
Discharge Planning Collaborative, 2011). Care transitions intervention studies
among older adult patients discharged from the hospital have found significant
cost savings among intervention recipients (Coleman, Parry, Chalmers, & Min,
2006; Naylor et al., 1999) and have thus captured the attention of policy makers
and health service organizations as an opportune area for action. As the
healthcare system confronts the aging of the baby boom generation, models of
care that reduce the use of high cost acute care are critical (Stone, Dawson, &
Harahan, 2003).
Patient and Caregiver Impact
The system-wide financial burden of poor transitions among older adults is
significant and so is the burden on older adult patients and their caregivers.

5

Numerous quantitative and qualitative studies have been conducted to explore
the impact of transitions between care settings on older adult patients and
caregivers and to delve into the problems they experience. Overall, these studies
found that patients received incomplete or erroneous information, poor
preparation, conflicting advice from providers, medication errors, poor
communication, inadequate administrative communication, and lack of follow up
care during and after transitioning out of the hospital (Coleman, 2003; Coleman,
Smith, Raha, & Min, 2005; Enguidanos, Cherin, & Brumley, 2005; Kravitz et al.,
1994; Moore et al., 2003; Morrill & Barreuther, 1988; Sobel, Medina-Walpole, &
Katz, 2004; Weaver et al., 1998). More specifically, Coleman et al. (2002)
conducted a qualitative study employing focus groups to elicit the key domains
older persons and their caregivers identified regarding their experience with care
transitions. Participants expressed confusion about their medication regimen and
frustration with healthcare providers who developed a care plan based on
provider convenience rather than their own. In addition, many were concerned
about their lack of ability to contact their healthcare provider and manage their
medical condition once at home. Correspondingly, an ethnographic study
conducted by Magilvy and Congdon (2000) revealed that patients and families
went into crisis mode upon admission to the hospital, attributed to minimal
planning done prior to the hospitalization. Crisis mode was found to be
compounded by surprises with health status or treatment and limited knowledge
of local resources, and exacerbated by inconsistent discharge planning that
disrupted the transition. Therefore, patients and caregivers were left ill equipped

6

to fully manage the patient’s condition upon discharge, seek timely supportive
services and follow-up, and make informed decisions about care which can lead
to subsequent health service use and rehospitalization. In a separate study of
post-discharge older adults, the primary challenge faced by patients was
identified as the effects associated with their medical conditions: specifically pain,
tiredness, loss of mobility, other progression to illness, and medical issues that
were not addressed in discharge plans (Grimmer, Moss, & Falco, 2004). Older
adult patients discharged to home in a frail status with increased dependence on
caregiver assistance for symptom and disease management are vulnerable to
poor transitions and shifts in health status requiring health service utilization. This
is largely due to patient’s and caregiver’s lack of knowledge, training, and
experience around self-management behaviors that typically increase in intensity
after a hospitalization. Therefore, perhaps not surprisingly, another study found a
lack of documented patient or family disease education in medical records to be
associated with increased hospital admission among older adults (Marcantonio et
al., 1999).

State of the Problem
Hospital readmissions is such a compelling economic, safety, and health
care quality and efficiency concern for patients, their family members/caregivers,
and acute health care institutions that it was targeted by the Affordable Care Act
(ACA) in section number 3026 Hospital Readmissions Reduction Program
(Patient protection and affordable care act, pub. L,. 2010). New penalties outlined

7

in the ACA began levying fines in 2012 equating to one-percent of total Medicare
billing among hospitals that have high 30-day readmission rates. Moreover,
penalties increased to two-percent in 2013 and are currently at three-percent,
leaving hospitals desperate for solutions. With increasing penalties arising from
the ACA readmission policy, hospitals are scrambling to determine which
patients constitute the high-risk pool that should be targeted for transitional care
services and resources. The Centers for Medicare and Medicaid Services [CMS
(through ACA Section 3026)] has invested half-a-trillion dollars over five years
through grants for Community-based Care Transitions Program implementation
in hospitals to improve the transitional care of beneficiaries at high risk for
readmissions. The success of this policy imperative to reduce readmissions has
great potential to positively impact patient care and health care spending across
the nation.
Current efforts to improve transitional care, reduce 30-day readmissions,
and identify older adults at highest risk for readmission have included the
development of algorithms and interventions (Allaudeen et al., 2011; Donze,
Aujesky, Williams, & Schnipper, 2013; Gruneir et al., 2011). Researchers have
developed various prediction tools, lists of clinical patient characteristics, and
algorithms to calculate risk—the most popular being the LACE model (Gruneir et
al., 2011), an acronym for the factors considered to be predictive of readmission:
length of hospital stay, acuity on admission, comorbidity, and emergency
department visits—although with variable success (Allaudeen et al., 2011).
However, the limited success with prediction tools may be attributed to the

8

inclusion of general, administrative variables in calculations that are readily
available in standard medical service use data. Sole use of administrative data
excludes other factors, such as information pertaining to the home situation and
support that may delve into the circumstances around a hospitalization because
they are missing from medical charts.
Interventions aimed at reducing 30-day hospital readmissions can be
categorized into the following three domains identified by Hansen and colleagues
in their meta-analysis: pre-discharge interventions, post-discharge interventions,
and bridging interventions (Hansen, Young, Hinami, Leung, & Williams, 2011).
Pre-discharge interventions largely include improved efforts around patient
education, discharge planning activities, reconciliation of patients’ medication list,
and scheduling follow-up physician appointment(s) prior to the patient being
discharged. These interventions employ an isolated pre-discharge intervention
and, with the exception of Evans and Hendrick’s (1993) study (a randomized
control trial), are cross-sectional and do not include any follow-up contact with
the patient once they leave the hospital. Results of the impact of U.S.-based pre-
discharge intervention on reducing older adult hospital readmissions are mixed
where 37.5% of studies found significant reductions in 30-day rehospitalization
rates (Einstadter, Cebul, & Franta, 1996; Evans & Hendricks, 1993; Lucas,
1998), and 62.5% did not (Gaft et al., 2010; Kramer et al., 2007; McPhee, Frank,
Lewis, Bush, & Smith, 1983; O’Dell & Kucukarsian, 2005; Schneider, Hornberger,
Booker, Davis, & Kralicek, 1993).

9

The second type of intervention aimed at reducing older adult 30-day
hospital recidivism –post-discharge interventions—initiate contact with patients
after they are discharged from the hospital. Study activities largely focus on
facilitating communication with patients and primary care providers, conducting
follow-up telephone contacts, providing patients with a hotline for questions and
referrals, and an in-home visit. Similar to the results of pre-discharge
interventions, post-discharge intervention results among randomized control
trials, quasi-experimental designs, cohort studies, and non-controlled pre-post
designs are mixed. Just over half (54.5%) of these studies found significant
reductions in 30-day rehospitalization rates for older adult patients (Coleman,
Parry, Chalmers, Min, 2006; Creason, 2001; Harrison, Harra, Pope, Young, &
Rula, 2011; Hernandez et al., 2010; Sharma, Kuo, Freeman, Zhang, & Goodwin,
2010) while 45.5% of post-discharge interventions did not (Balaban, Weissman,
Samuel, & Woolhandler, 2008; Bostrom, Caldwell, McGuire, & Everson, 1996;
Dudas, Bookwalter, Kerr, & Pantilat, 2001; Hess et al., 2010; Misky, Wald, &
Coleman, 2010).
The final type of readmission-reducing intervention identified by Hansen
and colleagues is interventions that bridge the care transition. These
interventions are characterized by combining two or more elements from pre-
discharge and post-discharge interventions, such as in-hospital contact/planning
and follow-up contact(s) at home, and are longitudinal. Half of bridging
interventions identified by Hansen et al. (2011) are randomized control trials and
their impact on reducing short-term older adult hospital recidivism has been more

10

consistent: 75.0% of bridge interventions observed significant reductions in
rehospitalization rates among study participants (Ahmed, Thornton, Perry,
Allman, & DeLong, 2004; Anderson, Deepak, Amoateng-Adjepong, & Zarich,
2005; Creason, 2001; Jack et al., 2009b; Koehler et al., 2009; Naylor et al., 1999)
while 25.0% did not (Dedhia et al., 2009; Parry, Min, Chugh, Chalmers, &
Coleman, 2009). Due to the more consistent outcomes with bridge interventions
as compared to isolated pre-discharge and post-discharge interventions,
multifaceted bridge interventions have emerged as a prevailing type of
transitional care.
Currently, there are three primary, bridging care transitions interventions
that are being replicated across the United States and Canada: the Care
Transitions Intervention (CTI) (Coleman et al., 2006), the Transitional Care Model
(TCM) (Naylor et al., 1999), and Project Re-Engineered Discharge (Project RED)
(Jack et al., 2009a; ProjectRED, 2014). The foundation of each of these
interventions consists of four key components (coined “pillars” by Coleman and
colleagues) that include medication reconciliation/self-management, patient-
centered approach/patient activation/personal health record, follow-up/post
discharge care organization, and identification of “red flags”/preventing decline.
Each of these interventions begins on the inpatient side at index hospitalization
and follows patients for a specified number of in-home and/or telephone contacts
when they are discharged to the community. Registered nurses or nurse
practitioners specially trained and certified to perform the enhanced transitional
care activities pertaining to the particular transition intervention (Transition Coach

11

for CTI; Transitional Care Nurse with TCM; and Discharge Advocate with Project
RED) conduct all patient contacts. Results from these three randomized control
trials revealed that the interventions were significantly associated with reductions
in 30-day rehospitalizations among their high-risk, older adult samples (Coleman
et al.: 8.3% vs. 11.9%, p=0.48; Naylor et al.: 10% vs. 23%, p=0.04; Jack et al.:
hospital utilization rate 0.341 vs. 0.451 visits per person per month, p=0.009).
Addressing Gaps: Need and Opportunity for Research
These models demonstrate that providing assistance to older adults
during and after discharge from acute care can reduce rehospitalizations and
increase patients’ quality of life. Another positive outcome is that these models
have enrolled patients discharged with and without home health, therefore,
demonstrating the need for transitional support irrespective of receipt of home
health. However, while the efforts of these three interventions have been
documented as successful at reducing 30-day hospital readmissions (Coleman et
al., 2006; Jack et al., 2009a; Naylor et al., 1999), several important weakness
exist. Namely, each of these models primarily has a medical focus and employs
high level medical personnel to perform the transitional care activities. High level
medical personnel, such as advanced practice nurses and nurse practitioners,
are highly skilled professionals in short supply nationally, making them difficult to
locate, recruit, and retain. Because nurses with advanced degrees represent only
about 4% of all registered nurses, there are insufficient numbers to meet the
needs of the medical system (HRSA, 2013). In addition, the use of highly trained
professional nurses increases the costs of the proposed interventions

12

significantly and reduces the likelihood that these models will be widely
replicated. Review of 2014 Los Angeles County salaries for advanced practice
nurses and nurse practitioners indicated that advanced practice nurses average
$80,000 annually and nurse practitioners average $99,000, compared to $64,000
for licensed clinical social workers (Indeed.com, 2014a, Indeed.com, 2014b). In
addition, because nursing models focus on a medical approach, such as
connecting patients with appropriate medical services, most exclude community
and home based services, long-term care planning, and broader social and
psychosocial issues (Corley & Mauksch, 1993).
Other shortcomings of these three popular transitions interventions are
that they largely fail to include patient social service needs and available
community-based supports in their activities, and they do not address the long-
term needs that inevitably continue to arise among chronically-ill older adults. In
spite of their (short lived, 30-day) success, these interventions are temporary
solutions to problems that manifest longitudinally and will reoccur, leading to
subsequent health service use and care transitions, potentially at shorter and
shorter intervals. Notwithstanding, their use of more highly paid, medically trained
personnel (nurses and nurse practitioners) to perform transitional care duties that
are an extension of inpatient physical medicine is not only costly to health care
institutions but should include patients’ social service needs; an area specialized
in by social workers.

13

Opportunities for Social Work
Social work practice is increasingly recognized as an important
component of health care (Dziegielewski, 1998; Volland & Keepnews, 2006).
Social work uniquely addresses the client’s biopsychosocial status and the social
system in which care is delivered (NASW, 1992). Older adults transitioning from
one care setting to another who need assistance in accessing medical services
may also need supportive services and community-based resources to remain
living safely at home. Nurses may consider non-medical transition assistance
outside their scope of practice and lack experience in this area. Social workers
specialize in providing psychosocial assessment and work with family systems
and fragmented health and social service systems to address multifaceted client
needs (Atkinson & Nelson, 1995; Geron, Andrews, & Kuhn, 2005; Rosen &
Teeson, 2001; Scharlach, Simon, & Dal Santo, 2002; Vourlekis, Gelfand, &
Greene, 1992). Over the last several decades, social workers have played
prominent roles as care managers or care coordinators in programs and services
for older adults (Geron & Chassler, 1994; Geron, 2000a; Geron, 2000b).
Empirical studies employing social workers in key roles have found them
effective in reducing emergency room visits and hospital admissions (Claiborne,
2003), reducing length of hospital stay and nursing home placement (Nickolaus,
Specht-Leible, Bach, Oster, & Schlierf, 1999), and reducing overall costs of care
(Nickolaus et al., 1999; Williams, Williams, Zimmer, Hall, & Podgorski, 1987).
Telephone social work interventions are effective in reducing medical
service use following emergency room visits and hospitalizations (Monsuez,

14

Fergelot, Papon, & Le Gall, 1993; Shannon, Wilber, & Allen, 2006), reducing
hospital days among high users of health care (Shannon et al., 2006), linking
patients with their primary care physicians following emergency room visits
(Kallis, Gonzalez del Rey, Ruddy, & Schubert, 1999), and providing caregiver
support (Albert, Im, Brenner, Smith, & Waxman, 2002). Furthermore, social work
telephone interventions have been linked to reduced mortality among high risk
older adults (Alkema, Wilber, Shannon, & Allen, 2007). Telephone care
management is also effective in reducing depression using cognitive behavioral
approaches (Simon, Ludman, Tutty, Operskalski, & Von Korff, 2004).
Social work skills and areas of expertise cross the continuum of care and
position the social worker as an ideal provider of transitions services. Because
the overall goal of a social work intervention is to restore or maintain independent
functioning of the client to the fullest extent possible (NASW, 1992), professional
social work values support the specific needs of older adults during transition.
Given that many hospitals, community-based care management, and social
service providers employ social workers, a social work transitional care model
could build on existing structures for replication. Finally, the growing shortage of
qualified health professionals will likely continue to pressure the health care
system to create new models of collaborative care to meet the health and
psychosocial needs of older adults (Buerhaus, 2002; Johnson, Billingsley, &
Costa, 2006; Sumner & Townsend-Rocchiccioli, 2003).
A new bridging intervention introduced by Rush University, aptly named
The Bridge Model (Altfeld et al., 2012; Fabbre, Buffington, Altfeld, Shier, &

15

Golden, 2011; Rush University Medical Center, 2013), utilizes a social worker in
a transitional care coordinating role. The social work care coordinator uses
medical records to identify older adult patients as at risk for readmission, meets
with patients to brainstorm and pinpoint unmet needs, and arranges services
prior to hospital discharge. Two days post discharge, patients are contacted via
telephone by the social work care coordinator where new problems are identified
and addressed in addition to determining patient understanding of discharge
instructions, understanding of medications, scheduling follow-up medical
appointments, and caregiver burden issues, among others. While this
intervention may appear to potentially address the shortcomings of the three,
primary bridge care transitions interventions described above, a recent
publication reported that The Bridge Model intervention was not effective in
reducing 30-day readmissions (Altfeld et al., 2012). Additionally, while follow-up
contact is provided in this intervention, it is limited to telephone contact only and
does not include an in-home assessment or medication safety reconciliation. This
could result in important environmental and social factors such as home safety,
support, and resources being overlooked and unreported by patients who many
not recognize them as problematic. Along these lines, Bridge only recruits
patients that are discharged with a prescription for home health visits and these
providers may also discount social service deficits as they are trained to focus on
the specific physical health problem(s) the patient was discharged with.

16

Dissertation Research and Purpose
The overall purpose of this dissertation is to determine the impact of social
workers on reducing older adult 30-day hospital readmission rates using a
randomized controlled trial of a bridging, in-home social work intervention so that
resources and efforts can be targeted to better meet patients’ needs and ACA
goals. This study builds on existing bridging interventions and is drawn from
several observations. First, as previously discussed, transitions between care
settings create elevated risk for poor outcomes and for readmission among older
adults leaving the hospital for home largely due to fragmented care and poor
communication. Next, while few studies exist that test methods to improve
transitions, those available are largely medically focused, using a nurse or
advanced practice nurse in their approach. Although evidence exists to support
the effectiveness of these models, few have been replicated and none have been
integrated into standard health care practice. This may be attributed to several
factors including the availability of the needed staff, the lack of existing structures
to support these roles, and the costs of implementing these interventions. Finally,
a social work driven intervention may provide a replicable mechanism for
bridging medical care, addressing psychosocial needs as well as medical needs,
and improving linkages with community services while reducing care duplication.
Important information derived from achieving the specific study aims can
inform future interventions and health policy around the area of older adult
hospital readmissions, specifically concerning resource allocation for successful
efforts to reduce older adult hospital readmissions. Andersen’s Behavioral Model

17

of Health Service Utilization is used to guide this research and answer the aims
and research hypotheses through appropriate, multivariate analyses.
Organization of the Dissertation
This dissertation is organized into seven chapters:
Chapter 1 is a review of the literature and overall problem guiding this
dissertation research. Backgrounds, explanation of the problem and its
significance, current understanding and action toward the problem, gaps in
practice, and opportunities for research are discussed.
Chapter 2 describes, in depth, the methodology used in this study
including research questions and hypotheses, guiding conceptual model, study
design, study groups, and measures used in analyses. This chapter concludes
with a detailed explanation of the analytical methods used in each chapter and
each analysis.
Chapter 3 is the first of four results chapters and provides a description of
the overall study sample.
Chapter 4 presents the results for the study question pertaining to the
overall effectiveness of the SWIFT study to reduce 30-day hospital readmissions.
Implications of the results are discussed.
Chapter 5 presents the results for the study question regarding the impact
of unmet social service needs on 30-day hospital readmissions. Implications of
the results are discussed.

18

Chapter 6 presents the results for the study question investigating the
impact of declining the in-home intervention on 30-day hospital readmissions.
Factors predicting intervention opt-out are also presented. Implications of the
results are discussed.
Chapter 7 recapitulates the key, overall findings of this dissertation,
discusses how these findings contribute to the literature and knowledge on
hospital readmissions among older adults, identifies potential limitations of the
research, and offers key implications.

18

CHAPTER 2: METHODS

This chapter lists the research questions and hypotheses, and describes
the conceptual model, study design, sample, measures, and analytic strategy for
chapters 3-6.
Research Questions & Hypotheses
Research Questions
1. Can the SWIFT intervention reduce 30 day readmission rates as
compared to usual care?
2. What impact does unmet social service needs have on readmissions?
3. What characteristics are associated with intervention opt-out?
4. What impact does intervention opt-out have on hospital readmissions?
Hypotheses
The overall purpose of this study is to determine the impact of a social
work intervention on hospital readmission rates among older adults transitioning
from hospital to home. Each of the research questions addresses the overall
study objective and explores specific elements of need and behaviors and how
they relate to 30-day readmissions, unmet social service needs, and SWIFT
intervention opt-out.
1. Subjects randomized to the SWIFT intervention group will be less likely
to be readmitted to the hospital within 30-days following enrollment as
compared to those in the usual care group.

19

2. Subjects with a greater number of unmet social service needs at 10-
days post discharge will be more likely to be readmitted to the hospital
within 30-days than those with fewer unmet social service needs at 10-
days post discharge.
3. Subjects randomized to the SWIFT intervention group that opt-out of
the intervention will be more likely to be readmitted to the hospital
within 30-days as compared to those that receive the intervention.

Conceptual Model
The conceptual model guiding this study is Andersen’s Behavioral Model
of Health Service Utilization (Andersen, 1968) which has been widely used to
explain health service use among older adults (Azuero, Allen, Kvale, Azuero, &
Parmelee, 2013; Heider et al., 2014; Robinson, Shugrue, Fortinsky, & Gruman,
2014). The health service utilization focus of this study is on hospital
readmissions among a sample of older adults identified as at high risk for
readmission. Andersen’s Model contains a prominent behavioral element that
arguably makes it more applicable to studying outcomes such as hospital
readmissions because health service utilization can be affected both directly and
indirectly by patient behaviors. Another relevant model is The Chronic Care
Model (Wagner, Austin, & Von Korff, 1996) that depicts a productive interplay
between patients, providers, and the health care system to influence health
outcomes and activate patients. The Chronic Care Model emphasizes that quality
care for chronic conditions can be delivered through an integrated system

20

consisting of various elements (Coleman, Austin, Brach, & Wagner, 2009;
Gabbay, Bailit, Mauger, Wagner, & Siminerio, 2011; Wagner et al., 1996;
Wagner et al., 2001) (relating to the community, patient support, and health
system design and support) and serves as a viable framework for this study,
however, we selected Andersen’s Model as it has a better fit due to the
importance of behavioral elements with acute service utilization. Andersen’s
model differs from the Chronic Care Model in that it highlights, specifies, and
takes into account the various patient-level characteristics and behaviors that
ultimately inform patients’ “productive interactions” in the health care environment
(i.e. predisposing characteristics, enabling resources, need, and health
behaviors).
The most recent, fourth phase of Andersen’s Model (Andersen, 1996;
Andersen, 1995) shown in Figure 1 posits that health service use is predicted by
an interaction of environmental factors, population characteristics, and health
behaviors. In terms of hospital readmissions, this means that subsequent
hospital utilization is influenced by hospital-level and environmental factors,
patient-level population factors, and health behaviors. In this dissertation, the
environmental factor is the hospital-to-home discharge for community-dwelling
older adults, and study subjects with this expected discharge course were
targeted for enrollment. This environmental factor innately included the status of
the patient at enrollment. Predisposing characteristics are typically patient socio-
demographic variables that represent an individual’s inherent tendency to utilize
health services, and include factors such as age, gender, and ethnicity

21

(Andersen & Newman, 1973). Enabling resources are factors that either
contribute to an individual’s ability to procure resources or serve as a barrier to
procuring resources. Examples include marital status, income, education level,
having a caregiver, and social service utilization. Enabling resources often are
most susceptible to manipulation through policy and research interventions such
as the SWIFT study and manipulations can influence health service utilization.
For example, theoretically, a social work driven care transitions intervention such
as SWIFT can impact hospital readmission rates by aligning patients with
services to meet social and physical health needs. Need factors represents an
individual’s most immediate reasons for utilizing health services and includes
both formally diagnosed and patient-perceived health conditions, symptoms, and
self-management abilities (Andersen & Newman, 1973). Need is also an
indication of condition acuity in light of unmet social service needs to address
physical health conditions. Health behaviors, such as accepting transitional care
support services, having an advance directive, disease self-management
behaviors, and medication adherence, among others, is the final factor
associated with predicting health service utilization in Andersen’s Model. This
model elucidates the importance of looking at patient-level factors including
social service needs and health behavior while considering the healthcare
environment/hospital factors when evaluating hospital readmissions.

22

Figure 1. Andersen’s Behavioral Model of Health Service Utilization—Part 4

23

Study Design
This dissertation is a primary analysis of the SWIFT randomized control
pilot study conducted at a large, non-profit hospital located in the Los Angeles
area between February 2011 and September 2013. The study was approved by
the Institutional Review Boards (IRB) at both the hospital study site and the
academic research institution executing the study (the University of Southern
California).
Study Site
Huntington Memorial Hospital (HMH) is a 625-bed, non-profit, community
hospital in Pasadena, California that serves a diverse patient population in terms
of age, race, ethnicity, and socio-economic status. HMH is a level two trauma
center (the only one in the San Gabriel Valley area) and offers a full spectrum of
acute, inpatient services in addition to outpatient, preventative care to patients.
On average, HMH treats approximately 10,000 older adult patients (aged 65 or
more) annually, 44% of whom are 80 years old or more. To meet the unique
needs of their prominent older adult population, in 1984 HMH developed the
Huntington Senior Care Network (HSCN) that provides information and referrals
to community-based services, case management (fee-for-service and Medicaid
waiver), community education initiatives, and support services for 23,000 older
adults and their caregivers. The primary goal of HSCN is to help older adults live
safely and independently in their homes for as long as possible. The social
workers who conducted the intervention were recruited from HSCN and trained in
the intervention protocol.

24

Sample Size
The primary outcome for this dissertation is hospital readmission after
index hospital discharge and enrollment in SWIFT. Based on the results of a pilot
study conducted by SWIFT investigators, previous related research in hospital-
to-home care transitions (Naylor et al., 1999), and a power analysis (utilizing
nQuery Advisor, version6), it is estimated that a sample of 445 patients meeting
the eligibility criteria is needed to obtain sufficient power to detect study group
differences. Out of the 445 eligible patients, approximately 40% can be expected
to refuse participation, leaving 267 for enrollment and randomization, and a
subsequent attrition rate of 25% (final estimated sample of n=200). To maintain
80% of power at a two-tailed alpha level of 0.05, 200 eligible patients out of the
445 are needed to complete all follow-up points during the six-month study
period. This target sample size would yield very strong power to detect a small
effect size (Cohen’s d=0.2-0.5 or f=0.1-0.25) (Cohen, 1988) and group
differences in readmission rates and duration of time between hospital discharge
and readmission.
Eligibility/Ineligibility
The study eligibility criteria included English-speaking, cognitively intact,
community-dwelling, older adults age 65 or older living within a specified service
area of Los Angeles County (identified prior to study initiation by the researchers
and study personnel), and identified as “at-risk” for subsequent readmission by
meeting at least one of the following:

25

 Age 75 or older
 Prescribed five or more prescription medications
 Had one or more inpatient hospitalization or emergency department
visit in the prior six months
Exclusion criteria. Patients diagnosed with end-stage renal disease,
long-term nursing facility residents, and hospice recipients were not eligible for
the study due to care variation and/or existing supports for these individuals.
Recruitment
Two methods of recruitment were used: case finding and direct referral.
Daily hospital census reports were reviewed (excluding the Intensive/Critical
Care Unit) to identify older adult patients and electronic medical records were
used to determine preliminary eligibility (number of medications, number of
inpatient hospitalizations or emergency department visits in previous six months,
zip code, type of residence, and primary language). Direct referrals were made
by a social worker conducting rounds in the nursing units. Patients meeting initial
eligibility criteria were approached at hospital bedside by master’s level student
research assistants who administered the Short Portable Mental Status
Questionnaire (SPMSQ) (Pfeiffer, 1975) to establish mental competency (as
determined by a score of five or more out of 10 possible). Eligible patients were
invited to participate and asked to sign an informed consent and HIPAA
Authorization documents.

26

Study Groups
SWIFT Intervention
The SWIFT intervention was developed using key elements identified in
previous hospital-to-home transitions interventions by Naylor (Naylor et al., 1999)
and Coleman’s “pillars” (Coleman, Parry, Chalmers, Min, 2006). These vital
areas include medication reconciliation and assessing patent medication self-
management abilities, patient activation through a patient-centered approach and
development of a personal health record, post-discharge follow-up and
organization of care, and identification of “red flags” to prevent decline. SWIFT
also expanded on these elements, in a true social work approach emphasizing
the strengths of the discipline, to facilitate linkage to home and community-based
services (HCBS) and engage patients in problem-solving therapy (PST). PST is
an evidence-based, cognitive-behavioral therapy widely used among older adults
in primary care and facilitates the application and acceptance of problem-solving
attitudes and skills among patients (D’Zurilla & Goldfried, 1971). PST has been
effectively administered by social workers (Enguidanos, Coulourides Kogan,
Keefe, Geron, & Katz, 2011; Geron & Keefe, 2006; Unutzer et al., 2001) and
found to significantly enhance the effectiveness of depression treatment (Arean,
Hegel, Vannoy, Fan, & Unutzer, 2008; Unutzer, Katon, & Callahan, 2003;
Unutzer et al., 2002; Unutzer, Patrick, Marmon, Simon, & Katon, 2002) and to aid
older adult patients in addressing key social health, physical health, and wellness
issues (Enguidanos et al., 2011).

27

SWIFT is a combination of previous transitional care research, social work
attributes, and theoretical underpinnings. According to Andersen’s Behavioral
Model of Health Service Utilization (the conceptual framework of this
dissertation), enabling resources, need factors, and health behaviors can
contribute to health service use (Andersen, 1995). SWIFT intervention activities,
specifically those in the psychosocial realm, inherently assessed these areas and
highlighted opportunity for intervention. For example, SWIFT assessed family
and caregiver support, social service needs, health equipment needs, disease
and medication self-management abilities, and provided PST patient-lead
coaching accordingly.
The SWIFT intervention consisted of two important segments carried out
by the study social worker: in-home contact and telephone follow-up. Patients
randomized to the intervention arm of the study received a maximum of two in-
home visits and up to four telephone contacts; a maximum of six contacts with
the social worker in total. An intervention checklist was developed to guide
intervention activities and the decision-making process. This included a bulleted
list of specific activities around each of the “pillars” established in previous
transitions interventions and those elements foundational to SWIFT.
Home Visits
The purpose of the first home visit was to conduct an initial assessment
and develop and implement a plan of care. Activities performed by the social
worker during this visit were guided by the intervention checklist and included a

28

psychosocial evaluation, home safety check, medication inventory for
reconciliation, review of hospital discharge instructions, health goal setting and
problem solving, coaching around scheduling follow-up physician appointments,
and home and community based service referrals.
The psychosocial evaluation was developed specifically for SWIFT by
study investigators and areas consisted of the following: presence of caregiver
and family support, current alignment with a primary care physician, current
smoking and drinking behaviors, use of mobility-assisting devises, recent
incidence of falling and fear of falling, pain level, efficacy of medication controlling
pain, challenges around taking and acquiring prescribed medication(s), and a
room-by-room home safety evaluation guided by a check-list from a nationally
recognized, evidenced-based falls coalition (Fall Prevention Center of
Excellence, 2014). A copy of patients’ 9-item Patient Health Questionnaire (PHQ-
9) depression screening taken at baseline was also given to the social worker
prior to the home visit to guide study activities around depression, such as
engaging in PST, when applicable. These areas are commonly included in the
psychosocial assessments utilized by other medical institutions and
organizations.
During the first home visit, the social worker also took inventory of all
prescribed and over-the-counter medications, vitamins, and supplements that the
patient was taking. Patients were also queried on the purpose, dosage, and
administration specifications of each of their medications to assess their level of
knowledge and medications management behaviors/compliance. The social

29

worker’s handwritten list of each patient’s medications was later entered into the
evidence-based home medication management software, HomeMeds (Partners
in Care Foundation, 2013), for reconciliation and automated alerts regarding
potential drug interactions, errors (i.e. duplicate therapies), and safety concerns.
In the event of an alert, the SWIFT study social worker promptly notified the
patient and their primary care physician of the potential problem and served as a
liaison between the two parties until the matter was resolved. Two important
elements of the social worker’s study-related activities were coaching and
problem-solving because of the focus on patient-centered training, progression,
and empowerment. The didactic coaching approach was used when reviewing
hospital discharge instructions, scheduling follow-up physician appointments, and
when making referrals to community-based services. Patients and caregivers
reviewed these materials together with the social worker and developed a plan
for prioritization and action. Similarly, health goal setting was also a patient-lead,
collaborative process and the social worker engaged in PST with the patient
when necessary.
The second in-home SWIFT intervention visit was conducted when
problems identified at the initial home visit were not sufficiently resolved or were
extensive enough that telephone contact would not be adequate. Patients and
caregivers sufficiently supported in the home with resources, exhibiting
knowledge and ability to execute disease self-management behaviors and
instructions, and follow-through with the care plan and heath goals, were not

30

provided a second in-home visit; instead subsequent contacts were conducted
via telephone.
Telephone Follow-Up
In addition to the in-home visit(s), SWIFT intervention patients received up
to four telephone contacts from the social worker. Telephone contacts were used
to follow-up on issues identified at the home visit, discuss outcomes from
physician visits, review established health goals (drafting new ones, when
applicable), determine success of linkage to community-based services, and
problem-solve around new issues.
Usual Care
Patients randomized to the usual care group received standard care
provided by the medical facility, hospital social workers, and discharge planners,
as well as standard care provided by their primary care physicians. These
participants also received a modest modification to usual care practices which
was a reassurance call following hospital discharge conducted by the doctoral-
level study coordinator.
Measures
Data were collected at enrollment, 30-days, and six months following
enrollment via several different mechanisms: electronic health records, in-person
at hospital bedside, and by telephone. Figure 2 illustrates the multitude of
patient-level variables and the timeline in which they were collected.

31

Figure 2. Measures

Screening Data
Secondary screening data were collected from the hospital electronic
database and through patient survey. Data on patient age, prior emergency
department visit(s), prior hospitalization(s), presence of advance directive,
number of medications, and admission and discharge dates (to calculate length
of stay) were collected using the hospital’s electronic health records (MediTec) to
determine study eligibility.
Similarly, the SPMSQ was administered and scored in-person at bedside
according to eligibility criteria (Pfeiffer, 1975). The authors of the tool specify a

32

cutoff score of five, where five or more correctly answered questions out of 10
indicates a patient is cognitively intact. Therefore, patients that correctly
answered five or more out of 10 possible questions were included in the study.
Baseline and Follow-Up
Predisposing and Enabling Factors (Demographics). Patient
demographics were collected at baseline only and included gender, race, marital
status, living situation, education, and income from patient self-reports at hospital
bedside. Additional, patient-level characteristics collected at baseline included
presence of a caregiver and cohabitation arrangement.
Need Factors. Each of the need factors were collected in-person via self-
reports with patients at bedside, at baseline, and specific variables included of
chronic disease diagnosis, discharge home without physician-prescribed
assistance (i.e. home, self-care), being depressed, unmet social service needs,
and intervention intensity.
Chronic disease diagnosis information was derived from a self-reported
inventory of 10, non-mutually exclusive diseases consisting of the following:
asthma, dementia/Alzheimer’s, depression/psychosis, diabetes, cancer,
congestive heart failure (CHF), chronic obstructive pulmonary disease
(COPD)/lung disease, heart disease, hypertension, and stroke. During analysis,
asthma and COPD/lung disease were combined to form a “respiratory disease”
category, and CHF and heart disease were collapsed into “cardiac disease.”

33

Physician prescription of home health services or self-care information
was obtained from the discharge summary in patient’s electronic health records
at discharge.
Patient depression was measured using the nine-item Patient Health
Questionnaire (Kroenke & Spitzer, 2002), which is documented to have high
reliability and validity, and has been widely used to detect depression among
older adults (Kroenke, Spitzer, & Williams, 2001; Lin et al., 2003). Possible
scores range from zero to 27; representing a scale of no depression (0) to severe
depression (27). Incidence of depression in the present study was indicated by a
score of 10 or more at baseline. This threshold score was used because it
represents the cutoff for moderate depression established by measurement
creators (Kroenke & Spitzer, 2002).
Pain was collected via the Number Rating Scale (NRS) where patients are
asked to rate their pain in the last 24-hours on a scale ranging from zero to 10
(zero indicating no pain and 10 being the worst pain possible).
Patients’ physical functioning ability was captured by using questions from
the 36-item Short Form Health Survey (SF-36) (Tarlov et al., 1989; Ware &
Sherbourne, 1992). In an effort to reduce burden to patients, we elected to
abbreviate the SF-36, however, we did not use the shorter 12-item version
because of its inclusion of emotional aspects and self-rated health on functioning.
These elements were collected elsewhere in the SWIFT survey. Eight questions
were selected and include different everyday physical activities ranging from

34

vigorous to low intensity. Questions focus on a patient’s ability to walk (one block
or several blocks), climb stairs (one flight or several flights), bend and kneel,
carry groceries, perform moderate intensity activities (vacuum, play golf, move a
table, or bowling), and perform vigorous activities (run, lift heavy objects, engage
in strenuous sports). Patients are asked whether or not they are physically limited
in performing the activity (yes or no). If they are limited, patients are further asked
to indicate if they are limited a lot of limited a little. Therefore, all possible
responses are: not limited, limited a little, and limited a lot, and were coded as
zero, one, and two, respectively. A composite score of overall physical
functioning was calculated where possible scores ranged from zero to 16 with a
higher value indicating greater physical limitations.
Indicators of high patient acuity, also utilized in other model of transitional
care research (Coleman, Parry, Chalmers, Min, 2006; Naylor et al., 1999) were
the three secondary eligibility criteria: aged 75 years or more, taking five or more
prescription medications, and had a previous inpatient stay in the prior six-
months. A variable for patient acuity was calculated by adding up the total
number of secondary criteria elements that each patient met. Patients that met all
three of the secondary eligibility criteria were identified as having highest acuity.
Social service needs and utilization information were collected using an
inventory of 18 common social services that included: adult day health care,
caregiver support group, case management, chore services/handyman,
companion/friendly visitor, congregate meals, home delivered meals, home
health/nursing, homemaker services, information/referral, legal services,

35

personal care, psychological counseling, respite care, shopping, transportation
services, and other. At baseline, patients were asked to indicate which of these
services they need and which ones they are currently using. At 10-day follow-up
(via telephone), patients were asked which of these services they are currently
using. Social service needs were calculated from participants’ reports of needing
a particular social service at baseline and their reports of not using that specific
service at 10-day follow-up. From this calculation, a total unmet social service
needs variable was created. From the original list of 18 services, four services
were dropped because they were deemed to have a low response rate and low
importance or impact on 30-day hospital recidivism: caregiver support group,
companion/friendly visitor, legal services, and respite care. Therefore, the total
number of unmet needs ranged from zero to 14. Additionally, the 14 different
social services were organized into five service areas as follows, and recorded
as a dichotomous variable representing unmet need in that area:
 Outside help: adult day health care, case management,
information/referrals, legal services, psychological counseling, and
physical therapy
1
.
 Household help: chore services/handyman and homemaker
services.
 Food help: congregate meals, home-delivered meals, and
shopping.
 Personal help: home health nursing and personal care.

1
Physical therapy was coded out of the open-ended responses for “other” social service based on high
frequency of responses

36

 Transportation help: transportation
The need measure—intervention intensity—consisted of the total number
of study-related contacts the participant had with the study social worker.
Possible values for this variable ranged from zero (for patients randomized to
usual care and for patients that opted-out of the intervention) to six (the
maximum number of intervention contacts allowed by the study protocol).
Health Behaviors. The health behaviors analyzed in this dissertation are
the presence of an advance directive, having had a previous inpatient stay in the
prior six months, and intervention opt-out. Information on patients’ advance
directive status was obtained from electronic medical records at baseline which
was indicated in the chart as has advance directive, does not have an advance
directive, or no information. Similarly, for previous inpatient stay in the prior six
months, electronic medical records were also reviewed at baseline and the
information was collected. This variable was later recoded from a count variable
into a dichotomous variable because it was severely skewed (1=had at least one
inpatient stay in the previous six months, or 0=no prior stay in the previous six
months). For intervention opt-out, this information was collected from participants
randomized to the intervention study group that declined the intervention (home-
visit and telephone follow-up) although agreed to remain in the overall study. A
dichotomous variable was created to capture intervention opt-out versus
intervention recipient.

37

Health Service Use. MediTec electronic health records were used to
determine if study subjects had been readmitted to HMH or had visited the HMH
emergency department during the six-month study period. In this dissertation, a
readmission is defined as an unplanned, all-cause, inpatient admission to HMH
within 30-days of discharge from index hospitalization.
Analysis
Hypothesis 1
1. Subjects randomized to the SWIFT intervention group will be less likely to
be readmitted to the hospital within 30-days following enrollment as
compared to those in the usual care group.
The dependent variable for this analysis is 30-day readmission. Use of count
variables in health service use data have been identified as problematic because
they violate two of the basic assumptions (normality and homoscedasticity)
essential to multivariate analysis such as linear regression (Diehr, Yanez, Ash,
Hornbrook, & Lin, 1999). For this reason and the limited instances of multiple
readmissions within 30-days of index discharge, 30-day hospital readmission
was coded as a dichotomous variable. Next, chi-square, t-tests, and correlations
were used to determine any significant associations between having a 30-day
readmission and study randomization assignment, intervention intensity,
predisposing characteristics, enabling factors, need factors, and health
behaviors. Results from correlations also ensured avoidance of common variable
constructs and informed the final logistic regression model by revealing any

38

multicolinearity between variables (as their inclusion would make the regression
unstable and should be excluded). Two stepwise logistic regressions were
performed to identify: 1) Predictors of 30-day readmission among study groups,
and 2) Predictors of 30-day readmission accounting for intervention intensity.
Hypothesis 2
2. Subjects with a greater number of unmet social service needs at 10-
days post discharge will be more likely to be readmitted to the hospital
within 30-days than those with fewer unmet social service needs at 10-
days post discharge.
To test this research hypothesis, the dependent variable in all analyses was
having a 30-day hospital readmission. Chi-square tests were performed to
analyze differences between readmitted patients and unmet social service needs
in each of the five key areas. Additionally, a t-test was conducted to analyze any
differences in average number of unmet social service needs between
readmitted and non-readmitted patients. Two stepwise logistic regressions were
performed to identify characteristics that predicted 30-day readmission
accounting for need factors and health behaviors. Predisposing characteristics
and enabling factors were not included in the multivariate models for several
reasons: the sample size for this analysis was reduced from 181 to 109 due to
missing data and since the bivariate results did not reveal any significant
associations with 30-day readmissions, these variables were excluded to make
the models the most parsimonious. The source of missing data were the unmet

39

social service needs measures collected at 10-day post-discharge follow-up,
where capturing patient responses that close to hospitalization proved very
difficult due to slow patient recovery, needed rest, or ill health; 39.8% of
participants did not complete a 10-day follow-up survey. In order to account for
missing data on the key unmet social service need variables (total unmet needs,
unmet outside help need, unmet household help need, unmet food
prep/shopping need, unmet personal help need, and unmet transportation need),
the regression analyses were conducted using Mplus version 6.11 statistical
software. Mplus allows for the inclusion of complete data in the dataset by using
a full-information maximum likelihood (FIML) as opposed to other methods of
handling missingness, such as imputation. Model A included a covariate for total
number of unmet social service needs, and Model B included dichotomous
covariates for an unmet need in each of the five service areas (outside help,
household help, personal help, and transportation help). Results of the bivariate
analyses, the conceptual model, and previous research guided inclusion of
variables in each regression to make them the most parsimonious.
Hypothesis 3
3. Subjects randomized to the SWIFT intervention group that opt-out of the
intervention will be more likely to be readmitted to the hospital within 30-
days as compared to those that receive the intervention.
To test this research hypothesis, data were split descriptively to analyze the
intervention group only and make two comparisons: 1) between those that opted-

40

out of the intervention to those that received the intervention, and 2) intervention
patients that had a 30-day readmission with those intervention patients that were
not readmitted within 30-days. Descriptive statistics were used to describe the
characteristics of this new sub-sample. Next, chi-square and t-tests were
performed to analyze differences between the patients that opted-out of the
intervention versus intervention recipients, and readmitted versus on-readmitted
intervention patients on predisposing demographic characteristics, need factors
(disease diagnosis and discharged home without assistance), and health
behaviors (presence of an advance directive, and 30-day readmission). Two
stepwise logistic regressions were performed to identify characteristics that
predicted opting-out of the SWIFT intervention and 30-day readmission. Results
of the bivariate analyses, the conceptual model, and previous research guided
inclusion of variables in each regression to make them the most parsimonious.
All analyses were performed with the statistical software SPSS version 18 or
Mplus version 6.11 (Chapter 5, only).

39

CHAPTER 3: OVERALL DEMOGRAPHIC AND HEALTH CHARACTERICTICS

This chapter reports the results from participant screening in addition to
descriptive, demographic analyses for the overall study sample.
Screening Results
Inpatient hospital census reports were used to screen for potential study
participants from February 2011 to March 2013. A total of 1,035 patients were
found to meet the eligibility criteria for inclusion in the SWIFT study and were
visited at hospital bedside by a master’s level research assistant and provided an
explanation of the study and invitation to participate. Out of the 1,035 eligible
patients, 26.5% refused to participate for a variety of reasons. The three most
common reasons for refusal were: 1) services perceived as not needed/does not
need “help”/not a “sick” person, 2) already has a caregiver at home, and 3) does
not want anyone coming into their home. Additionally, about half (53.5%) of
eligible patients were discharged prior to making a decision; meaning that
between the time that they were screened, found to meet the eligibility criteria,
visited at bedside, and often revisited (due to patient or provider request, patient
out of room for testing or procedure, posted contact precautions placard
prohibiting entry by study personnel, etc.), the patient was discharged. Twenty-
percent of all eligible patients enrolled in the SWIFT study and a small fraction of
enrollees (2.6%) were dropped from the study after being discharged to a skilled
nursing home where they remained for more than three weeks (the established
SWIFT threshold for study inclusion). Out of the 207 consenting and enrolled

40

participants, 181 were randomized to the SWIFT intervention (n=90) or the usual
care (n=91) arm. See Figure 3.

Figure 3. Flow Chart of Patient Screening, Enrollment, and Randomization

Sample Description
Predisposing Characteristics and Enabling Resources
Overall, the SWIFT study sample contained 181 older adults identified as
at high-risk for hospital readmission. Half of the sample (51.4%) were male and
half (50.8%) were aged 80 years or more (mean=78.8, sd=8.3). The racial
composition of the sample was diverse, with 60.9% self-identifying as Caucasian,
20.7% African American, 8.9% Latino, 5.0% Other, 3.4% Asian/Pacific Islander,
and 1.1% Native American. Less than half of participants were married (44.9%),

41

followed closely by widowed (30.3%), single (15.2%), and divorced (9.6%). The
sample was highly educated with the vast majority (91.9%) having graduated
from high school, 21.1% of which had a graduate degree or doctorate degree. Of
the 9.1% that did not graduate from high school, most (60.0%) completed some
high school and 40.0% completed eighth grade or less. Although the sample was
quite educated, the annual income level among reporting participants was
divergent: 13.3% reported earning less than $10,000 annually and 12.2% earned
$50,000 or more. However, a little more than half (53.6%) the sample declined to
provide this information. The vast majority (89.9%) of the sample reported living
in their own home or apartment and 68.8% indicated that they live with another
person (i.e. spouse, significant other, child, paid caregiver, or other arrangement
that was a combination of spouse and adult child or spouse and paid caregiver).
Nearly a third reported living alone. A little over half (57.0%) of participants
reported having a caregiver, and this individual was found to most commonly be
a spouse, adult child, other individual (i.e. neighbor, roommate, other family
member), paid caregiver, or significant other, respectively.
Randomization was successful, meaning that the predisposing
characteristics and enabling resources (demographics) of the usual care group
did not significantly differ from those in the intervention group. See Table 1.

42

Table 1
Overall Demographic Characteristics (n=181)

Frequency (%)

Usual Care Intervention Overall P-value
n=91 n=90 n=181
Age 79.2 ± 8.8 78.4 ± 7.8 78.8 ± 8.3 0.083
Age group (n=181)

62-79 years 41 (45.1) 48 (53.3) 89 (49.2)
0.270

80+ years 50 (54.9) 42 (46.7) 92 (50.8)
Gender (n=181)

Male 42 (46.1) 51 (56.7) 93 (51.4)
0.160
Female 49 (53.9) 39 (43.3) 88 (48.9)
Ethnicity (n=179)

African American 23 (25.6) 14 (15.7) 37 (20.7)
0.517

Caucasian 52 (57.8) 57 (64.0) 109 (60.9)

Latino 7 (7.8) 9 (10.1) 16 (8.9)

Native American 1 (1.1) 1 (1.1) 2 (1.1)

Asian/PI 4 (4.4) 2 (2.3) 6 (3.4)
Other 3 (3.3) 6 (6.7) 9 (5.0)
Marital status (n=178)

Married 35 (38.5) 45 (51.7) 80 (44.9)
0.061

Single 11 (12.1) 16 (18.4) 27 (15.2)

Divorced 10 (11.0) 7 (8.0) 17 (9.6)
Widowed 35 (38.5) 19 (21.8) 54 (30.3)
Highest education (n=163)

8th grade or less 1 (1.2) 5 (6.3) 6 (3.7)
0.435

9th-11th grade 4 (4.8) 5 (6.3) 9 (5.5)

High school graduate 17 (20.5) 13 (16.3) 30 (18.4)

Some college 27 (32.5) 20 (25.0) 47 (28.8)

College graduate 17 (20.5) 19 (23.8) 36 (22.1)

Graduate degree 10 (12.0) 14 (17.5) 24 (14.7)
Doctoral degree 7 (8.4) 4 (5.0) 11 (6.7)
Annual income (n=181)

Under $10,000 15 (16.5) 9 (10.0) 24 (13.3)
0.447

$10,000 - $19,000 9 (9.9) 10 (11.1) 19 (10.5)

$20,000 - $29,000 9 (9.9) 5 (5.6) 14 (7.7)

43

$30,000 - $39,000 3 (3.3) 1 (1.1) 4 (2.2)

$40,000- $49,000 0 (0.0) 1 (1.1) 1 (0.6)

$50,000 or more 11 (12.1) 11 (12.2) 22 (12.2)
Refused to reply 44 (48.3) 53 (58.9) 97 (53.6)
Living situation (n=178)

Own house/apartment 82 (90.1) 78 (96.7) 160 (89.9)
0.866

Living in family
member's home 7 (7.7) 6 (6.9) 13 (7.3)
Other 2 (2.2) 3 (3.4) 5 (2.8)
Who do you live with? (n=178)

Alone 35 (38.5) 20 (23.5) 55 (31.3)
0.180

Spouse/partner 31 (34.1) 43 (50.6) 74 (42.0)

Child 14 (15.4) 13 (15.3) 27 (15.3)

Paid caregiver 2 (2.2) 1 (1.2) 3 (1.7)
Other 9 (9.9) 8 (9.4) 17 (9.7)
Who is your primary caregiver? (n=177)

No one/self 43 (47.3) 33 (38.4) 76 (42.9)
0.267

Spouse 19 (20.9) 24 (27.9) 43 (24.3)

Significant other 0 (0.0) 4 (4.7) 4 (2.3)

Child 11 (12.1) 11 (12.8) 22 (12.4)

Paid caregiver 9 (9.9) 6 (7.0) 15 (8.5)
Other 9 (9.9) 8 (9.3) 17 (9.6)
*p<.05; **p<.01; ***p<.001

Need Factors and Health Behaviors
The medical need factors and health behaviors of the SWIFT sample
illustrate the overall health status of study participants and are presented in Table
2. At screening, SWIFT participants were identified as at high-risk for
readmission and this was most commonly due to polypharmacy [prescribed five
or more medications (77.3%; mean medications=7.8, sd=3.8)], followed by
advanced age [75 years or more (67.4%)] and prior hospitalization or emergency
department visit in the six months preceding study enrollment (37.0%). On
average, participants were diagnosed with 6.1 (sd=3.5) different medical
conditions. The prevalence of common medical conditions among the sample

44

included: cardiac disease (65.0%), hypertension (64.5%), respiratory disease
(34.3%), depression (26.9%), cancer (25.3%), diabetes (23.7%), stroke (9.9%),
and dementia/Alzheimer’s disease (4.1%). Just over half (53.0%) the sample had
an advance directive at the time of study enrollment, and most were discharged
home without physician-prescribed assistance such as home health or brief
placement in a skilled care facility (74.4%). Overall, 11.0% of patients were
readmitted to HMH within 30-days of discharge. On average, SWIFT participants
reported mild to moderate pain (mean=3.8 out of 10, sd=3.4), had high cognitive
functioning (mean=9.13 out of 10, sd=1.2), and reported mild to moderate
limitations with physical functioning (mean=8.6 out of 16, sd=5.1). Results of
depression screening revealed that, on average, usual care patients’ ratings
corresponded with minor depression (mean=4.3, sd=3.9) and intervention
patients’ ratings indicated mild depression (mean=6.2, sd=5.1). This difference
was highly significant for the PHQ-9 depression screening (p<.001) and self-
reports of having depression (p<.001). All other medical conditions, need factors,
and health behaviors did not significantly differ between study groups.
Table 2.
Overall Medical Conditions, Need Factors, and Health Behaviors (n=181)

Frequency (%)

Usual Care Intervention P-value
Chi-square n=91 n=90
Health Behaviors

Has advance directive

Yes 50 (54.9) 46 (51.1) 0.605
30-day readmission

Yes 11 (12.2) 9 (10.0) 0.417
Need Factors

Cancer 24 (26.7) 20 (23.8) 0.665

45

Cardiac disease 58 (64.4) 55 (65.5) 0.887

Dementia/Alzheimer's 3 (3.3) 4 (4.8) 0.632

Depression 14 (15.6) 32 (38.1) <.001***

Diabetes 18 (20.0) 23 (27.4) 0.252

Hypertension 57 (64.0) 54 (65.1) 0.889

Respiratory disease 30 (33.7) 29 (34.9) 0.865
Stroke 7 (7.8) 10 (11.9) 0.360
Aged 75 years or more

Yes 62 (68.1) 60 (66.7) 0.833
Takes 5 or more
medications
Yes 66 (80.5) 74 (89.1) 0.120
Had inpatient stay or ED visit in prior 6 months

Yes 35 (41.2) 32 (39.0) 0.780
Discharged home with self-care

Yes 61 (70.9) 67 (77.9) 0.294
T-tests
Mean (sd) P-value
# Daily medications 7.61 (3.8) 8.00 (3.8) 0.510
# Health conditions 6.35 (4.1) 5.88 (2.9) 0.400
Depression score 4.33 (3.9) 6.15 (5.1) <.001***
Physical functioning 7.67 (5.12) 7.18 (5.08) 0.819
Cognition 9.08 (1.2) 9.18 (1.1) 0.573
Pain 3.54 (3.4) 4.10 (3.3) 0.290
*p<.05; **p<.01; ***p<.001

In summary, the SWIFT sample is comprised of cognitively intact, well-
educated, fairly racially diverse, advanced-age older adults with multiple
comorbidities and taking multiple prescription medications. Randomization was
successful in terms of all predisposing characteristics, enabling factors, and
health behaviors; only one need factor, depression, was significantly different
between study groups.

46

CHAPTER 4: PREDICTORS OF 30-DAY READMISSION AMONG AT-RISK
OLDER ADULTS

Introduction
This chapter is an analysis of the complete study sample and aimed to
determine if the SWIFT intervention was effective in reducing 30-day readmission
rates as compared to those randomized to usual care. Hospital readmissions is
such a compelling economic, safety, and health care quality and efficiency
concern for patients, their family members/caregivers, and acute health care
institutions that it was targeted by the ACA (section number 3026) and levies
financial penalties on hospitals with above average readmission rates among
their Medicare patients. Hospitals, such as HMH, are desperate to find answers
in order to avoid costly penalties and improve transitional care for their older
adult patients, and many have implemented transitional care interventions such
as SWIFT. Determining the effectiveness of these interventions can better inform
other hospitals, policy makers, and CMS.
Methods
In this chapter, we hypothesized that study subjects randomized to the
SWIFT intervention would be less likely to be readmitted to the hospital within 30-
days as compared to those that receive usual care.
The specific methods for this chapter are described in detail in chapter 2. In
summary, chi-square and t-tests were performed to analyze differences between
the two study groups on predisposing characteristics, enabling factors, medical
conditions, need factors, and health behaviors. Finally, two stepwise logistic

47

regressions were performed to identify characteristics associated with 30-day
readmission. The only difference between the two regressions was how
randomization was accounted for in the model: Regression A had a dichotomous
intervention versus usual care independent variable loaded; Regression B had a
count variable for the number of study-related patient-social worker contacts to
account for intervention intensity. Results of bivariate analyses, the conceptual
model, and previous research guided inclusion of variables in the regressions to
make them the most parsimonious. For example, variables for cognition, physical
functioning, and pain were not included in these analyses.
Results
Demographics. The univariate and bivariate results for the predisposing
characteristics, enabling factors, medical conditions, need factors, and health
behaviors among the study sample are reported at length in Chapter 2.
In summary, the SWIFT sample was comprised of cognitively intact, well-
educated, fairly racially diverse, advanced-age older adults with multiple
comorbidities and taking multiple prescription medications. Randomization was
successful in terms of all predisposing characteristics, enabling factors, and
health behaviors; only one need factor, depression, was significantly different
between study groups where the intervention group reported higher depression
than usual care participants.

48

Bivariate results. Bivariate analyses of those with and without a 30-day
readmission revealed that participants with a 30-day readmission were
significantly more likely to be diagnosed with cancer (χ
2
=10.560, p<0.001), have
high acuity (meeting all three of the criteria for high-risk; x
2
=10.560, p=0.005),
and had a prior inpatient stay in the six months preceding their current, index
hospitalization (x
2
=6.179, p=0.013). Additionally, a trend was observed for
increased 30-day readmissions among Caucasians (χ
2
=2.977, p=0.084),
although this finding did not reach statistical significance. No significant
associations were found with 30-day readmissions and
predisposing/demographic characteristics, diagnosis with a cardiac or respiratory
disease, discharge home without physician-ordered assistance, depression, or
having an advance directive. Neither study randomization assignment nor study
intensity (i.e. number of study-related contacts with the social worker) was
significantly associated with 30-day readmissions. See Table 3.
Table 3
Bivariate Analyses for 30-day Hospital Readmissions (n=181)

Frequency (%)

30-day
Readmission
(n=20)
Not Readmitted
w/in 30-days
(n=161)
Overall
Sample
Group
(n=180) χ
2
P-value
Age 80.15 ± 7.9 78.62 ± 8.3 78.8 ± 8.3 0.778 0.438
Male

Yes 12 (60.0) 80 (50.0) 92 (51.1)
0.711 0.399
No 8 (40.0) 80 (50.0) 88 (48.9)
Caucasian

Yes 15 (78.9) 93 (58.5) 108 (60.7)
2.977 0.084
No 4 (21.1) 66 (41.5) 70 (39.3)
Married

49

Yes 11 (55.0) 69 (43.7) 80 (44.9)
0.921 0.337
No 9 (45.0) 89 (56.3) 98 (55.1)
College grad+

Yes 11 (55.0) 60 (42.0) 71 (43.6)
1.214 0.271
No 9 (45.0) 83 (58.0) 92 (56.4)
Respiratory condition

Yes 7 (35.0) 52 (34.2) 59 (34.3)
0.005 0.944
No 13 (65.0) 100 (65.8) 113 (65.7)
Cardiac condition

Yes 14 (70.0) 99 (64.3) 113 (64.9)
0.254 0.614
No 6 (30.0) 55 (35.7) 61 (35.1)
Cancer

Yes 11 (55.0) 33 (21.4) 44 (25.3)
10.560 <.001***
No 9 (45.0) 121 (78.6) 130 (74.7)
Home, self-care (no services)

Yes 16 (80.0) 112 (73.7) 128 (74.4)
0.370 0.543
No 4 (20.0) 40 (26.3) 44 (25.6)
Moderate depression (10+)

Yes 2 (10.5) 21 (14.4) 23 (13.9)
0.209 0.648
No 17 (89.5) 125 (85.6) 142 (86.1)
Highest acuity

Yes 8 (44.4) 25 (16.8) 33 (19.8)
7.753 0.005**
No 10 (55.6) 124 (83.2) 134 (80.2)
Presence of advance directive

Yes 12 (60.0) 84 (52.2) 96 (53.0)
0.437 0.508
No 8 (40.0) 77 (47.8) 85 (47.0)
Had previous inpatient stay

Yes 12 (63.2) 53 (34.0) 65 (37.1)
6.179 0.013*
No 7 (36.8) 103 (66.0) 110 (62.9)
Intervention group

Yes 9 (45.0) 81 (50.3) 90 (49.7)
0.201 0.654
No 11 (55.0) 80 (49.7) 91 (50.3)
Number of
contacts 0.65 ± 1.2 1.12 ± 1.7 1.07 ± 1.7
-1.161 0.247
*p<.05; **p<.01; ***p<.001

Multivariate results. Two stepwise logistic regressions were performed to
investigate: 1) the effect of the intervention on 30-day readmission, and 2) the

50

effect of intervention intensity on 30-day readmission. The variable for high
patient acuity was not included in multivariate analyses because it was nearly
identical to the previous inpatient stay variable, causing redundancies and over-
measurement. Results revealed that study randomization assignment did not
significantly predict 30-day hospital readmission (see Table 4a). Additionally,
when investigating study intensity in the model, results were unchanged, and
although the direction of the intervention intensity covariate was negative, the
effect was not significant. See Table 4b. The null hypothesis is accepted and
results indicate that the SWIFT intervention was not effective at reducing 30-day
readmission rates for the intervention group overall and for those patients that
received a higher intensity intervention.
While controlling for other confounding variables such as predisposing
characteristics, need, and health behaviors, Caucasian race, a cancer diagnosis,
and having had an inpatient stay in the previous six-months significantly
predicted 30-day readmission among the study sample (p=0.032, p=0.030, and
p=0.033, respectively). Caucasians and patients with cancer were found to be
five times more likely to be readmitted within 30-days of discharge (Caucasian:
OR=5.572, 95% CI 1.155-26.874; cancer: OR=5.495, 95% CI 1.183-25.519) and
those with a pervious inpatient stay were found to be nearly four times more
likely to be readmitted within 30-days of discharge (OR=3.713, 95% CI 1.133-
12.389). Other factors informed by the conceptual model such as additional
predisposing characteristics, enabling factors, need factors (including other

51

disease diagnosis), and health behaviors did not significantly predict 30-day
hospital readmission. See Tables 4a and 4b.

52

Table 4a
Multivariate Predictors of 30-day Hospital Readmission Accounting for Study Group (n=181)

Model 1

Model 2

Model 3

Model 4
OR SE p-value

OR SE p-value

OR SE p-value

OR SE p-value
Predisposing & Enablers

Age 1.000 0.000 0.722

1.000 0.000 0.520

1.000 0.000 0.555

1.000 0.000 0.486
Male 1.259 0.582 0.692

1.224 0.606 0.739

1.270 0.604 0.692

1.242 0.634 0.733
Caucasian 3.736 0.710 0.063

4.202 0.734 0.050*

4.628 0.743 0.039*

5.572 0.803 0.032*
Married 1.250 0.577 0.699

1.087 0.608 0.891

1.099 0.625 0.879

1.105 0.649 0.878
College grad or more 0.751 0.590 0.627

0.520 0.634 0.302

0.564 0.631 0.363

0.484 0.671 0.279
Intervention group 0.644 0.537 0.412

0.676 0.561 0.485

0.632 0.601 0.445

0.690 0.624 0.553
Need: Diagnosis

Respiratory condition

0.471 0.640 0.239

0.468 0.644 0.239

0.362 0.693 0.143
Cardiac condition

0.301 0.772 0.119

0.324 0.784 0.151

0.314 0.815 0.155
Cancer

6.128 0.767 0.018*

5.956 0.775 0.021*

5.495 0.783 0.030*
Need

Home, self-care

2.084 0.718 0.307

2.694 0.768 0.197
Moderate depression

1.074 0.911 0.938

0.763 0.951 0.776
Health Behaviors

Has advance directive

0.682 0.612 0.532
Had previous inpatient stay

3.713 0.615 0.033*
Model Summary

-2Log likelihood 99.185

91.325

90.177

85.238

R
2
0.074

0.172

0.186

0.245

Hosmer and Lemeshow χ
2
11.548

5.602

9.123

3.405
*p<.05; **p<.01; ***p<.001

53

Table 4b
Multivariate Predictors of 30-day Hospital Readmission Accounting for Study Intensity (n=181)

Model 1

Model 2

Model 3

Model 4
OR SE p-value

OR SE p-value

OR SE p-value

OR SE p-value
Predisposing & Enablers

Age 1.000 0.000 0.682

1.000 0.000 0.496

1.000 0.000 0.574

1.000 0.000 0.508
Male 1.217 0.585 0.737

1.186 0.610 0.779

1.202 0.606 0.761

1.178 0.635 0.796
Caucasian 3.697 0.708 0.065

4.064 0.727 0.050*

4.495 0.741 0.043*

5.535 0.810 0.035*
Married 1.221 0.579 0.731

1.072 0.612 0.909

1.100 0.633 0.880

1.130 0.657 0.852
College grad or more 0.748 0.590 0.622

0.509 0.636 0.288

0.543 0.628 0.331

0.464 0.664 0.247
Diagnosis

Respiratory condition

0.491 0.634 0.261

0.474 0.642 0.244

0.350 0.699 0.133
Cardiac condition

0.293 0.771 0.111

0.324 0.785 0.151

0.325 0.815 0.167
Cancer

6.344 0.766 0.016*

6.132 0.774 0.019*

5.444 0.782 0.030*
Need

Home, self-care

2.002 0.714 0.331

2.735 0.777 0.195
Moderate depression

0.834 0.869 0.835

0.598 0.920 0.577
Number of contacts

0.624 0.854 0.580

0.527 0.881 0.466
Health Behaviors

Has advance directive

0.688 0.611 0.540
Had previous inpatient stay

3.981 0.618 0.025*
Model Summary

-2Log likelihood 99.868

91.819

90.444

85.015

R
2
0.065

0.166

0.183

0.248

Hosmer and Lemeshow χ
2
2.364

6.748

11.011

8.035
*p<.05; **p<.01; ***p<.001

54

Discussion
This chapter aimed to determine if the SWIFT intervention was effective at
reducing 30-day readmission rates among participants. Findings revealed that
the intervention did not significantly reduce 30-day readmissions and that
Caucasians, patients diagnosed with cancer, and patients with a previous
inpatients stay were more likely to be readmitted within 30-days of discharge.
These results suggest two things: 1) the SWIFT intervention may not be effective
at lowering short-term hospital recidivism among at-risk older adults, and 2) a
social worker may not be an effective provider in a transitional care role to reduce
rehospitalizations. Currently, only one study implementing a telephone-based
social worker-lead intervention (no-home contact) addressing 30-day
rehospitalization rates is available and their intervention was also ineffective at
reducing 30-day hospital recidivism (Altfeld et al., 2010). However, two
randomized control trials including social workers as part of an intervention team
(i.e. social worker and pharmacist, and social worker and discharge planner)
tested the impact of care transitions interventions among a high-risk Medicare
population, and did find significant reductions in 30-day readmission rates among
participants (Evans & Hendricks, 1993; Koehler et al., 2009). These results may
be attributed to the team approach taken to implement interventions with
common aspects of SWIFT, and considering the results of the present study,
may further suggest that employing social work in isolation may not be effective
at reducing short-term hospital recidivism. More investigation into social worker-

55

lead care transitions interventions is needed to better understand the effect of
social workers in transitional care roles.
Although we hypothesized that the SWIFT intervention would reduce 30-
day readmissions, this lack of significant effect is not uncommon with transitional
care interventions lead by other providers, such as specially trained registered
nurses, discharge planners, and pharmacists/pharmacy personnel. For example,
since 1993 when Evans and Hendricks (1993) conducted the first randomized
control trial evaluating the effectiveness of a discharge planning intervention on
hospital readmission rates among a high-risk patient group, eight additional
randomized control trials have been conducted in the U.S. and published in this
area (Balaban et al., 2008; Coleman, Parry, Chalmers, Min, 2006; Dudas et al.,
2001; Jack et al, 2009b; Koehler et al., 2009; Naylor et al., 1999; Parry et al.,
2009; Rainville, 1999) with 25-percent also finding no effect on short-term
readmission rates. Additionally, Hansen and colleagues (2011) conclude in their
recent systematic review of interventions aimed at reducing 30-day
rehospitalizations that no single-dimension intervention implemented alone has
been regularly associated with reductions in short-term readmission rates. These
findings, along with those of the present study, suggest that hospital
readmissions are a complex area that likely includes a multitude of patient-level,
environmental, and organizational factors.
Another important consideration for analyzing the impact of care
transitions interventions on reducing 30-day hospital recidivism is the overall
hospital 30-day readmission rate for older adults at the clinical study site. Among

56

the SWIFT study sample, 10% of participants were readmitted within 30-days of
discharge (see Chapter 3). This rate is much lower compared to that of the other
two randomized controlled trials employing a social worker on the intervention
team: Evans and Hendricks (1993) observed a 22% 30-day readmission rate for
the non-study control group, and Koehler et al. (2009) observed a rate of 38.1%
among their control group. Therefore, hospitals with high 30-day readmission
rates, such as these two comparable trials, may have a greater opportunity to
find differences between study groups. It is also possible that the SWIFT clinical
study site (Huntington Memorial Hospital) is already doing a good job with their
usual care procedures to limit 30-day hospital recidivism among at-risk older
adults to 11% and that the subsequent readmission rate cannot be prevented. A
growing body of literature has revealed that hospital readmission rates are
largely driven by patient-level factors that are out of the hospital’s control (Joynt
& Jha, 2012) and penalizing the institutions is inappropriate and unfair. The
present study found Caucasian patients and patients diagnosed with cancer to
be significantly more likely to be readmitted within 30-days. Other studies have
also identified patient-level factors such as socioeconomic status (Joynt, Orav, &
Jha, 2011) and illness severity (Rathmore et al., 2003) to predict short-term
readmissions and explain 56.2% of the variation in 30-day readmissions (Singh,
Lin, Kuo, Nattinger, & Goodwin, 2013), although none corroborated our results
specifically.

57

In a recent study investigating 30-day readmission rates among Medicare
beneficiaries without implementing an intervention, African American patients
were found to be readmitted at higher rates (24.8% vs. 22.6%) than Caucasian
patients, and African American patients discharged from minority-serving
hospitals were readmitted at an even higher rate (26.4%) (Joynt et al., 2011).
This finding is opposite of what we found with SWIFT and may be related to the
fact that researchers only analyzed Medicare patients hospitalized for
pneumonia, congestive heart failure, or myocardial infarction and did not include
all-cause hospitalization for older adults as included in SWIFT.
Correspondingly, further investigation into the potential role of a cancer
diagnosis on 30-day hospital readmissions among high-risk older adults is
warranted. We found only one study that investigated 30-day rehospitalization
rates among older adult patients with cancer and researchers specifically
targeted patients with thyroid cancer (Tuggle, Park, Roman, Udelsman, & Sosa,
2010). While the readmission rate was low (8%) among their sample, Tuggle and
colleagues found factors such as advanced disease stage, increased
comorbidity, and longer index length-of-stay to predict unplanned, 30-day
readmission (Tuggle et al., 2010). According to the conceptual model of the
present study, the findings from Tuggle and colleagues represent more need-
based factors than predisposing characteristics but it still highlights that
readmissions may be influenced by factors outside of the hospital’s control. None
of the nine randomized control care transitions interventions reported racial or

58

specific disease diagnosis differences between those with and without 30-day
hospital recidivism.
Now that the ACA penalties have been in place for just over one year,
researchers have begun to investigate the hospital characteristics associated
with higher penalties. Large hospital (400 or more beds), major teaching
hospitals, and safety-net hospitals have each been found to be more highly
penalized under the ACA Hospital Readmissions Reduction Program than small,
non-teaching, non-safety-net hospitals (Joynt & Jha, 2013). The reasons behind
the variable readmission rates and subsequent penalties among these different
types of institutions is unknown but research suggests that case mix (including
predisposing and enabling factors and need) and socioeconomic factors are
important (Joynt et al., 2011; Joynt & Jha, 2013; Rathmore et al., 2003). For
example, more seriously ill patients might utilize large, teaching hospitals
because they are commonly known for their enhanced capabilities to provide
sophisticated diagnostic services and treatment to patients (American Hospital
Association, 2009) and sicker patients are known to utilize acute services more
often and experience readmissions (Joynt et al., 2011; Joynt & Jha, 2012).
Additionally, high readmission rates have been found to be associated with lower
mortality rates among heart failure patients (Gorodeski, Starling, & Blackstone,
2010), suggesting that these hospitals are able to keep sicker patients alive, who
in turn, are more likely to utilize acute care (Joynt & Jha, 2012). The study site for
the SWIFT study is both a large hospital and a teaching hospital and therefore,
readmission rates may be more of a function of the patient population served

59

than quality of care, although further investigation is needed to fully understand
this area.

58

CHAPTER 5: INVESTIGATING THE IMPACT OF UNMET SOCIAL SERVICE
NEEDS ON 30-DAY READMISSIONS AMONG AT-RISK OLDER ADULTS

Introduction
This chapter aimed to determine the impact of unmet social service needs
on 30-day hospital readmissions among the high-risk, older adult SWIFT study
sample. New enhanced transitional care interventions targeting at-risk older
adults, such as the Bridge Model (Altfeld et al., 2012; Fabbre et al., 2011; Rush
University Medical Center, 2013) and SWIFT, that employ social workers are
placing increased importance on the social service needs of patients. As
transitional care research begins to delve into the patient-level factors driving
readmissions, consideration for non-medical patient-level factors, including social
service needs such as caregiver assistance, meal preparation/delivery,
homemaker services, and transportation assistance is needed. For example, a
recent study conducted by Feigenbaum et al. (2012) examined factors
contributing to all-cause 30-day hospital readmissions among older adults from a
managed care organization, and identified unaddressed psychosocial needs as
the primary transitional care/coordination factor contributing to potentially
preventable readmissions.
Determining the impact of unmet social service needs on 30-day hospital
readmissions is important because if confirmed to be a contributing factor, it
could better inform hospitals, policy makers, and CMS of the types of needed
community-based resources to aid patients, and the necessary hospital
personnel to facilitate linkage and referral. Moreover, providing a social solution

59

as a means to address a physical health problem would undoubtedly prove more
cost effective as compared to subsequent, potentially avoidable acute health
service use and even transitions to skilled nursing facilities.
Methods
In this chapter, we hypothesized that participants with a greater number of
unmet social service needs at 10-days post discharge will be more likely to be
readmitted to the hospital within 30-days than participants with fewer unmet
social service needs at 10-days post discharge.
The specific methods for this chapter are described in detail in chapter 2.
In summary, chi-square and t-tests were performed to analyze differences
between SWIFT study participants that did and did not have a 30-day
rehospitalization and two different types of need variables: 1) a count variable of
the total number of unmet social service needs, and 2) five dichotomous
variables that indicated unmet need in five different types of social service areas
(outside help, household help, food help, personal help, and transportation help).
Finally, two stepwise logistic regressions were performed using Mplus (due to
missingness in the unmet social service need data) to identify characteristics
associated with 30-day readmission. The difference between the two regressions
was how unmet social service needs were accounted for in the model:
Regression A included a count variable for the number of reported unmet needs
to account for need severity; Regression B included five separate dichotomous
variables that represent five common areas of social service needs (outside,
household, food, personal care, and transportation help needs). Results of

60

bivariate analyses, the conceptual model, and previous research guided inclusion
of variables in the regressions to make them the most parsimonious. For
example, variables for cognition, physical functioning, and pain were not included
in these analyses.
Results
Demographics. The univariate and bivariate results for the predisposing
characteristics, enabling factors, need factors, and health behaviors among the
study sample are reported at length in Chapter 2. Additionally, the bivariate
results for associations of predisposing characteristics, enabling factors, need
factors, and health behaviors with 30-day hospital recidivism are reported in
Chapter 4.
In summary, the SWIFT sample was comprised of cognitively intact, well-
educated, fairly racially diverse, advanced-age older adults with multiple
comorbidities and taking multiple prescription medications. Randomization was
successful in terms of all predisposing characteristics, enabling factors, and
health behaviors; only one need factor, depression, was significantly different
between study groups. A significant association with having a cancer diagnosis
and 30-day readmission was observed, but no other predisposing characteristics,
enabling factors, medical diagnoses, need factors, or health behaviors were
associated with short-term hospital readmission.

61

Bivariate results. Bivariate analyses of those with and without a 30 day
readmission revealed a strong trend for participants readmitted within 30-days of
discharge and having an unmet need in the area of outside help (x
2
=3.511,
p=0.061). No significant relationships were found with 30-day readmissions and
greater number of unmet social service needs overall, or any of the other four
unmet social service need areas (household help, food help, personal help, or
transportation). See Table 5.
Table 5
Bivariate Analyses for 30-day Hospital Readmissions and Unmet Social Service
Needs(n=181)

Frequency (%)

30-day
Readmission
(n=14)
Not Readmitted
w/in 30-days
(n=95)
Overall
(n=109) χ2 P-value
Total unmet needs

2.21 ± 2.89 1.59 ± 1.87
1.67 ± 2.02
0.080 0.283
Area 1: Outside help

Yes 8 (57.1) 30 (31.6) 38 (34.9)
3.511 0.061
†

No 6 (42.9) 65 (68.4) 71 (65.1)
Area 2: Household help

Yes 5 (35.7) 30 (31.6) 35 (32.1)
0.096 0.757
No 9 (64.3) 65 (68.4) 74 (67.9)
Area 3: Food help

Yes 5 (35.7) 16 (16.8) 21 (19.3)
2.794 0.095
No 9 (64.3) 79 (83.2) 88 (80.7)
Area 4: Personal help

Yes 2 (14.3) 20 (21.1) 22 (20.2)
0.347 0.556
No 12 (85.7) 75 (78.9) 87 (79.8)
Area 5: Transportation help

Yes 3 (21.4) 19 (20.0) 22 (20.2)
0.015 0.901
No 11 (78.6) 76 (80.0) 87 (79.8)
*p<.05; **p<.01; ***p<.001
†= tr e nd

62

Multivariate results. Two stepwise logistic regressions were performed to
identify: 1) the impact of number of unmet social service needs on 30-day
readmission, and 2) relationship between each of the five social service needs
areas on 30-day readmission. Results of Model 1 revealed that controlling for
other confounding variables such as need and health behaviors, total unmet
social service needs did not significantly predict 30-day hospital readmission.
However, being diagnosed with cancer, high acuity, or having had a prior
inpatient stay in the previous six-months did significantly predicted 30-day
readmission among the study sample (p=0.026, p=0.020, and p=0.030,
respectively). Individuals diagnosed with cancer and individuals meeting the
criteria for high acuity were each found to be just over five times more likely to be
readmitted within 30-days of discharge (cancer: OR=5.419, 95% CI 1.227-
23.943; high acuity: OR=5.286, 95% CI 1.307-21.381). Additionally, those with a
prior inpatient stay in the previous six-months were four times more likely to be
readmitted within 30-days (OR=4.121, 95% CI 1.150-14.761). Other factors
informed by the conceptual model such as need factors, other disease diagnosis,
health behaviors, and study randomization assignment did not significantly
predict 30-day hospital readmission. See Table 6a.
When investigating unmet social service needs by key areas (Model 2),
results revealed a significant association with deficit in outside help and 30-day
readmission (p=0.008), and a trend for food prep/shopping need predicting 30-
day readmission (p=0.076). Participants with an unmet need for outside
help/community-based services were nearly eight times more likely to be

63

readmitted within 30-days of discharge (OR=7.988, 95% CI 1.702-37.504).
Additionally, consistent with what we found previously, a cancer diagnosis, high
acuity, and having had a prior inpatient stay in the previous six-months
significantly predicted 30-day hospital recidivism (p=0.039, p=0.026, and
p=0.006, respectively). Participants diagnosed with cancer were more than four
times more likely to experience 30-day hospital recidivism (OR=4.759, 95% CI
1.081-20.942) than those without cancer, and those identified as the most acute
were more than seven times more likely to be readmitted (OR=7.396, 95% CI
1.267-43.164). Yet the highest odds of 30-day hospital recidivism were the
patients that had had a prior inpatient stay in the previous six-months: these
individuals were more than eight times more likely to be readmitted (OR=8.372,
95% CI 1.822-38.470). Other factors informed by the conceptual model such as
other disease diagnoses, need factors including other key social service areas,
health behaviors, and study randomization assignment did not significantly
predict 30-day hospital readmission. See Table 6b.

64

Table 6a
Multivariate Predictors of 30-day Hospital Readmission Accounting for Number of Unmet Social Service Needs (n=181)

Model 1

Model 2

Model 3
OR 95% CI p-value

OR 95% CI p-value

OR 95% CI
p-
value
SWIFT intervention 0.857 0.326-2.257 0.755

0.704 0.235-2.11 0.530

0.739 0.23-2.371 0.611
Diagnosis

Respiratory 0.970 0.356-2.642 0.954

1.002 0.348-2.882 0.997

0.851 0.283-2.556 0.774
Cardiac condition 0.419 0.101-1.745 0.232

0.503 0.116-2.184 0.359

0.471 0.098-2.261 0.348
Cancer 7.308 1.894-28.203 0.004**

6.693 1.691-26.494 0.007**

5.419 1.227-23.943 0.026*
Need

Home, self-care

1.315 0.399-4.339 0.652

1.511 0.429-5.319 0.520
Moderate depression

0.806 0.17-3.82 0.786

0.643 0.151-2.73 0.549
Total unmet needs

1.122 0.849-1.482 0.420

1.047 0.773-1.419 0.768
High acuity

3.796 1.124-12.822 0.032*

5.286 1.307-21.381 0.020*
Health Behaviors

Has advance directive

1.076 0.354-3.268 0.897
Had prior inpatient stay

4.121 1.15-14.761 0.030*
Model Summary

Wald χ²

10.802

21.638

24.053

df

4

8

10

p-Value

0.029*

0.006**

0.008**
*p<.05; **p<.01; ***p<.001

65

Table 6b
Multivariate Predictors of 30-day Hospital Readmission Accounting for Unmet Social Service Needs in Key Areas (n=181)

Model 1

Model 2

Model 3
OR 95% CI p-value

OR 95% CI p-value

OR 95% CI p-value
SWIFT intervention 0.857 0.326-2.257 0.755

0.394 0.106-1.466 0.165

0.359 0.091-1.413 0.143
Diagnosis

Respiratory 0.970 0.356-2.642 0.954

1.150 0.315-4.194 0.832

0.835 0.2-3.493 0.805
Cardiac condition 0.419 0.101-1.745 0.232

0.587 0.126-2.742 0.499

0.567 0.096-3.359 0.531
Cancer 7.308 1.894-28.203 0.004**

6.341 1.608-25.003 0.008**

4.759 1.081-20.942 0.039*
Need

Home, self-care

1.353 0.353-5.189 0.660

1.369 0.396-4.734 0.620
Moderate depression

1.113 0.191-6.495 0.905

0.920 0.19-4.45 0.918
Need outside help

5.078 0.958-26.922 0.056
†

7.988 1.702-37.504 0.008**
Need household help

0.546 0.137-2.183 0.392

0.416 0.073-2.381 0.324
Need food help

3.133 0.661-14.854 0.150

4.688 0.852-25.795 0.076
†

Need personal help

0.358 0.047-2.752 0.324

0.253 0.044-1.445 0.122
Need transportation help

1.051 0.209-5.286 0.952

0.630 0.116-3.413 0.592
High acuity

4.389 1.164-16.542 0.029*

7.396 1.267-43.164 0.026*
Health Behaviors

Has advance directive

1.132 0.32-4 0.847
Had prior inpatient stay

8.373 1.822-38.47 0.006**
Model Summary

Wald χ²

10.802

29.548

18.760

df

4

12

12

p-Value

0.029*

0.003*

0.014*
*p<.05; **p<.01; ***p<.001
† = tren d

66

Discussion
This chapter aimed to determine the impact of unmet social service needs
on 30-day hospital readmissions. Findings suggest that participants’ overall
number of unmet social service needs were not related to 30-day readmission
among our sample. This effect may be due to a lack of power given the small
study sample size due to missing data, and further investigation into this area is
warranted. To our knowledge, this is the first study to investigate the impact of
different types of unmet social service needs beyond caregiver support on 30-
day hospital recidivism and we identified an unmet need in supportive community
services (outside help) to predict readmission along with a trend for unmet need
for food prep/shopping support. Current approaches to transitional care focus on
a medical approach and do not include social aspects: neither any of the
currently published randomized control trials on care transitions interventions, nor
any of the three replicated evidence-based, multifaceted care transitions
interventions include social service needs in intervention activities or analytical
methods. Moreover, at the crux of intervention activities is the development of a
patient care plan (“patient health record,” “personal health record,” “patient
discharge form”) to support and educate the patient in the hospital and at home,
and commonly includes health diagnosis information, medication instructions,
follow-up appointments, and recommended outpatient workups, among others,
with little consideration for social service issues as a barrier to self-management
activities. For example, a patient’s lack of supportive community services/outside
help (such as adult day health care, case management, referrals, psychological

67

counseling, and physical therapy) and unmet food/nutritional needs could inhibit
their ability to adequately manage their health condition(s), particularly during a
post-hospital recuperative period. Intervention efforts and invested resources
cannot truly enable patients if barriers exist.
Previous telephone-based, social work interventions targeting older adults
have been effective at reducing medical service use following emergency room
visits and hospitalizations (Monsuez et al., 1993); Shannon et al., 2006, reducing
hospital days among high health service utilizers (Shannon et al., 2006), linking
patients with their primary care physicians following emergency room visits
(Kallis et al., 1999), and providing caregiver support (Albert et al., 2002). These
accomplishments are attributed to combining social and medial aspects of
medicine in non-intensive interventions. Furthermore, a recent, multi-center study
of a managed care group examined factors contributing to all-cause 30-day
hospital readmissions among older adults and identified unaddressed
psychological and social needs as the primary transitional care/coordination
factor contributing to 73% of potentially preventable readmissions (Feigenbaum
et al., 2012). However, examples of unaddressed psychosocial needs were not
reported in the publication, highlighting a need for more specific information in
this area.
New reports, analyses, and commentaries on 30-day readmissions
acknowledge the important role patient-level and community factors (i.e.
availability of community-based resources) play on hospital recidivism (Joynt &
Jha, 2012; Robert Wood Johnson Foundation, 2013). Furthermore, they suggest

68

that an evidence-based, holistic approach to quality improvement in transitional
care (i.e. including a medical, social, spiritual, and emotional focus) is potentially
most likely to achieve better patient care at lower costs while appeasing policy
makers, health service organizations, health care providers, and patients (Joynt
& Jha, 2013). Results of these previous studies and the present study highlight
the potential importance of incorporating patients’ social service needs,
particularly supportive community service help and food needs, in interventions
targeting 30-day rehospitalization rates among at-risk older adults. Investigation
into this unexamined area is needed especially since timely primary care follow-
up after an inpatient hospitalization has been widely identified as impacting
readmission (Coleman, Parry, Chalmers, Min, 2006; Coleman & Williams, 2007;
Jack et al., 2009b; Joynt et al., 2011; Naylor et al., 2011; Naylor et al., 1999).
Patient characteristics, such as being diagnosed with cancer and having
had an inpatient stay in the prior six-months, found to be significantly associated
with 30-day hospital recidivism presented in Chapter four were also corroborated
here when investigating the impact of unmet social service needs.

69

CHAPTER 6: CHARACTERISTICS AND RISK FACTORS ASSOCIATED WITH
PATIENTS THAT DECLINE A HOME TRANSITIONS INTERVENTION

Introduction
This chapter is an analysis of patients randomized to the intervention arm
of the SWIFT pilot study and aimed to identify the characteristics and risk factors
associated with opting-out of a social work driven transition intervention. With
new penalties arising from the ACA readmission policy, hospitals are scrambling
to determine which patients constitute the high-risk pool that should be targeted
for transitional care services and resources. The Centers for Medicare and
Medicaid Services [CMS (through ACA Section 3026)] has invested half-a-trillion
dollars over five years through grants for Community-based Care Transitions
Program implementation in hospitals to improve the transitional care of
beneficiaries at high risk for readmissions. Increased knowledge of the factors
associated with intervention refusal can help identify those patients that appear
resistant to interventions. Since hospitals are being held accountable, this
information can highlight patient groups that may need to be targeted with other
means to reduce readmissions, such as education and intervention in the primary
care setting, and would better inform CMS, potential grant allocation, and other
hospitals.

70

Methods
In this chapter, we hypothesized that study subjects randomized to the
SWIFT intervention group that opt-out of the intervention will be more likely to be
readmitted to the hospital within 30-days as compared to those that receive the
intervention.
The specific methods for this chapter are described in detail in chapter 2. In
summary, descriptive statistics were used to describe the characteristics of this
sub-sample used in this chapter. Next, chi-square and t-tests were performed to
analyze differences between the patients that opted-out of the intervention and
intervention recipients on predisposing characteristics, enabling factors, need
factors (medical diagnosis and physician-prescribed home health services), and
health behaviors (presence of an advance directive and 30-day readmission).
Finally, two stepwise logistic regressions were performed to identify
characteristics associated with opting-out of the SWIFT intervention and 30-day
readmission. Results of bivariate analyses, the conceptual model, and previous
research guided inclusion of variables in the regressions to make them the most
parsimonious given the small sample size.
Results
Demographics
Overall, 90 participants were randomized to the SWIFT intervention group
which consisted of mostly Caucasian (63.6%) males (56.1%) living in their own
house or apartment (89.7%). The average age of participants was 78.4 years
(SD=7.8). Educational attainment was high, where the vast majority (87.6%)

71

completed high school or beyond; 46.3% held a bachelor’s degree or higher.
However, annual income was very diverse as 26.7% reported earning $20,000 to
$29,000 or less and 10% reported earning $50,000 or more, although the
majority of participants (61.1%) declined to provide this information. Half (50.6%)
of participants indicated that they lived with their spouse or partner, followed by
living alone (23.5%), living with their adult child (15.3%), or having some other
type of living arrangement (9.4%) which was most commonly living with their
spouse and adult child. Participants’ primary caregiver type was varied but was
largely no one/self (38.4%), spouse (27.9%), or adult child (12.8%). The
demographic characteristics of intervention opt-outs did not significantly differ
from those of the intervention recipients (see Table 7).
Table 7
Intervention Patient Demographic Characteristics (n=90)

Frequency (%)

Received
Intervention
(n=62)
Refused
Intervention
(n=28)
Overall
Intervention
Group
(n=90) P-value
Age 78.3 ± 8.2 78.4 ± 7.1 78.4 ± 7.8 0.963
Gender

Male 32 (52.5) 18 (64.3) 50 (56.1)
0.208
Female 29 (47.5) 10 (35.7) 30 (43.8)
Highest education

8th grade or less 4 (7.0) 1 (4.3) 5 (6.3)
0.692

9th-11th grade 3 (5.3) 2 (8.7) 5 (6.3)

High school graduate 10 (17.5) 3 (13.0) 13 (16.3)

Some college 15 (26.3) 5 (21.7) 20 (25.0)

College graduate 13 (22.8) 6 (26.1) 19 (23.8)

Graduate degree 8 (14.0)) 6 (26.1) 14 (17.5)
Doctoral degree 4 (7.0) 0 (0) 4 (5.0)

72

Ethnicity

African American 11 (18.0) 3 (11.1) 14 (15.9)
0.736

Caucasian 38 (62.3) 18 (66.7) 56 (63.6)

Latino 5 (8.2) 4 (14.8) 9 (10.2)

Native American 1 (1.6) 0 (0) 1 (1.1)

Asian/PI 1 (1.6) 1 (3.7) 2 (2.3)
Other 5 (8.2) 1 (3.7) 6 (6.8)
Marital status

Married 29 (49.2) 16 (57.1) 45 (51.7)
0.622

Single 10 (16.9) 6 (21.4) 16 (18.4)

Divorced 6 (10.2) 1 (3.6) 7 (8.0)
Widowed 14 (23.7) 5 (17.9) 19 (21.8)
Living situation

Own
house/apartment 51 (86.4) 27 (96.4) 78 (89.7)
0.317

Living in family
member's home 5 (8.5) 1 (3.6) 6 (6.9)
Other 3 (5.1) 0 (0) 3 (3.4)
Who do you live with?

Alone 13 (22.4) 7 (25.9) 20 (23.5)
0.300

Spouse/partner 27 (46.6) 16 (59.3) 43 (50.6)

Child 9 (15.5) 4 (14.8) 13 (15.3)

Paid caregiver 1 (1.7) 0 (0) 1 (1.2)
Other 8 (13.8) 0 (0) 8 (9.4)
Who is your primary caregiver?

No one/self 24 (40.7) 9 (33.3) 33 (38.4)
0.907

Spouse 16 (27.1) 8 (29.6) 24 (27.9)

Significant other 3 (5.1) 1 (3.7) 4 (4.7)

Child 6 (10.2) 5 (18.5) 11 (12.8)

Paid caregiver 4 (6.8) 2 (7.4) 6 (7.0)
Other 6 (10.2) 2 (7.4) 8 (9.3)
Annual income

Under $10,000 8 (12.9) 1 (3.6) 9 (10.0)
0.529

$10,000 - $19,000 7 (11.3) 3 (10.7) 10 (11.1)

$20,000 - $29,000 4 (6.5) 1 (3.6) 5 (5.6)

$30,000 - $39,000 1 (3.7) 0 (0) 1 (1.1)

$40,000- $49,000 0 (0) 1 (3.6) 1 (1.1)

$50,000 or more 6 (9.7) 3 (10.7) 9 (10.0)
Refused to reply 36 (55.9) 19 (67.8) 55 (61.1)
*p<.05; **p<.01; ***p<.001

73

Intervention Opt-out
Descriptive statistics. Overall, 52.2% of participants reporting having a
respiratory condition, 62.5% had a cardiac condition, and 29.2% reporting having
some type of cancer. Nearly two-thirds of the sample was found to have an
advance directive (63.0%) in place and the majority of patients were discharged
home without any assistance (77.9%) such as home health services or a brief
stay at a skilled nursing facility. Ten-percent of participants were readmitted to
the hospital within 30-days of index discharge. Out of the 90 participants
randomized to the intervention group, 28 opted-out (31.1%) while 62 (68.9%)
received the intervention.
Bivariate results. Bivariate analyses revealed that as compared to
intervention recipients, participants that opted-out of the intervention were
significantly more likely to be diagnosed with a respiratory condition (χ
2
=4.157,
p=0.039). Additionally, a trend was observed for a larger proportion of
intervention opt-outs to have an advance directive in place (χ
2
=2.169, p=0.107),
although this finding did not reach statistical significance. No significant
associations were found with opting-out of the intervention and
predisposing/demographic characteristics, diagnosis with a cardiac disease or
cancer, or discharge home without physician-ordered assistance. See Table 8.

74

Table 8.
Bivariate Analyses among Intervention Opt-Outs (n=90)

Frequency (%)

Received
Intervention
(n=62)
Refused
Intervention
(n=28)
Overall
Intervention
Group (n=90) χ
2
P-value
Age 78.3 ± 7.12 78.4 ± 8.10
78.4 ± 7.6
t= -.059 0.953
Male

Yes 32 (51.6) 18 (66.7) 50 (56.2)
1.731 0.139
No 30 (48.4) 9 (33.3) 39 (43.8)
Caucasian

Yes 39 (62.9) 17 (65.4) 56 (63.6)
0.049 0.512
No 23 (37.1) 9 (34.6) 32 (34.6)
Married

Yes 29 (48.3) 16 (59.3) 45 (51.7)
0.890 0.239
No 31 (51.7) 11 (40.7) 42 (48.3)
Respiratory condition

Yes 17 (28.3) 12 (52.2) 29 (34.9)
4.157 0.039*
No 43 (71.7) 11 (47.8) 54 (65.1)
Cardiac condition

Yes 40 (66.7) 15 (62.5) 55 (65.5)
0.132 0.452
No 20 (33.3) 9 (37.5) 29 (34.5)
Cancer

Yes 13 (21.7) 7 (29.2) 20 (23.8)
0.532 0.322
No 47 (78.3) 17 (70.8) 64 (76.2)
Home, self-care (no services)

Yes 45 (73.8) 22 (88.0) 67 (77.9)
2.086 0.149
No 16 (26.2) 3 (12.0) 19 (22.1)
Presence of advance directive

Yes 29 (46.0) 17 (63.0) 46 (51.1)
2.168 0.107
†

No 34 (54.0) 10 (37.0) 44 (48.9)
*p<.05; **p<.01; ***p<.001
†
=trend

75

Multivariate results. A stepwise, binary logistic regression was
performed to identify the characteristics that predicted opting-out of the SWIFT
intervention. Results revealed that the odds of opting-out of the SWIFT
intervention were significantly higher for participants with respiratory
conditions, where opt-outs were three times more likely to have a respiratory
condition than intervention recipients (OR=3.584, 95% CI 1.159-11.080,
p=.027). Other covariates such as a diagnosis of cancer or cardiac disease,
having an advance directive, or being discharged home without services were
found to not significantly predict intervention opt-out. See Table 9.

76

Table 9
Multivariate Predictors of Intervention Opt-out (n=90)

Model 1

Model 2

Model 3
Diagnosis OR SE P OR SE P OR SE P
Respiratory condition 2.969 0.529 0.040*

3.105 0.545 0.038*

3.584 0.576 0.027*
Cardiac condition 0.761 0.616 0.658

0.741 0.627 0.632

0.850 0.645 0.801
Cancer 1.552 0.656 0.503 1.630 0.677 0.470 0.846 0.790 0.833
Need

Home, self-care 3.842 0.820 0.102 3.993 0.866 0.110
Health Behaviors
Has advance
directive

1.719 0.570 0.342
30-day readmission 6.238 0.868 0.035*
Model Summary

-2Log likelihood 87.353

84.006

78.906

R
2
0.084

0.141

0.222
Hosmer and
Lemeshow χ
2
2.567

8.378 1.857
*p<.05; **p<.01; ***p<.001

77

30-day Readmission
Bivariate results. Significant associations were found between
intervention patients that were readmitted to the hospital within 30-day of index
discharge and diagnosis with cancer (χ
2
=10.206, p<.001). A strong trend for
associations between 30-day readmission and opting-out of the SWIFT
intervention was also identified but it did not reach statistical significance
(χ
2
=3.110, p=0.078). Neither predisposing/demographic characteristics, nor other
disease diagnoses (respiratory or cardiac disease), need factors (discharge
home without assistance), or behavioral factors (presence of advance directive)
among intervention patients were associated with 30-day readmission. See Table
10.

Table 10
Bivariate Analyses among Intervention Group Patients Readmitted within 30-
days

Frequency (%)

30-day
Readmission
(n=9)
Not Readmitted
w/in 30-days
(n=81)
Overall
Intervention
Group (n=90)
χ
2
P-value
Age 76.11 ± 6.451 78.63 ± 7.921
78.4 ± 7.6
t= -.917 0.362
Male

Yes 6 (66.7) 44 (55.0) 50 (56.2)
0.447 0.504
No 3 (33.3) 36 (45.0) 39 (43.8)
Caucasian

Yes 7 (87.5) 49 (61.3) 56 (63.6)
2.166 0.141
No 1 (12.5) 31 (38.8) 32 (36.4)
Married

Yes 6 (66.7) 39 (50.0) 45 (51.7)
0.898 0.343
No 3 (33.3) 39 (50.0) 42 (48.3)
Respiratory condition

78

Yes 3 (33.3) 26 (35.1) 29 (34.6)
0.011 0.915
No 6 (66.7) 48 (64.9) 54 (65.1)
Cardiac condition

Yes 7 (77.8) 48 (64.0) 55 (65.5)
0.675 0.411
No 2 (22.2) 27 (36.0) 29 (34.5)
Cancer

Yes 6 (66.7) 14 (18.7) 20 (23.8)
10.206
<
0.001**
* No 3 (33.3) 61 (81.3) 64 (76.2)
Home, self-care (no services)

Yes 7 (77.8) 60 (77.9) 67 (77.9)
0.000 0.992
No 2 (22.2) 17 (22.1) 19 (22.1)
Presence of advance directive

Yes 4 (44.4) 42 (51.9) 46 (51.1)
0.178 0.673
No 5 (55.6) 39 (48.1) 44 (48.9)
Opted-out of intervention

Yes 5 (55.6) 22 (27.2) 27 (30.0)
3.110 0.078
†

No 4 (44.4) 59 (72.8) 63 (70.0)
*p<.05; **p<.01; ***p<.001
†
=trend

Multivariate results. A second stepwise logistic regression was
performed with the dichotomous dependent variable of being readmitted to the
hospital within 30-days of index discharge. Results revealed that while controlling
for other confounding variables such as predisposing characteristics, need, and
health behaviors, a cancer diagnosis and opting-out of the SWIFT intervention
predicted 30-day hospital readmission (p=.014 and p=.040, respectively).
Intervention participants diagnosed with cancer were found to be 18 times more
likely to be readmitted within 30-days of discharge (OR=18.717, 95% CI 1.829-
191.416), and those that opted-out of the home intervention were found to be six
times more likely to be readmitted within 30-days of discharge (OR=6.474, 95%

79

CI 1.089-38.488). Other factors of the conceptual model such as need and health
behavior did not significantly predict 30-day hospital readmission. See Table 11.

80

Table 11
Multivariate Predictors of 30-day Hospital Readmission among Intervention Group (n=90)

Model 1

Model 2

Model 3
Diagnosis OR SE P OR SE P OR SE P
Respiratory condition 0.897 0.807 0.893

0.887 0.811 0.883

0.431 0.994 0.397
Cardiac condition 0.354 1.254 0.408

0.357 1.255 0.411

0.439 1.310 0.530
Cancer 15.748 1.129 0.015* 15.865 1.130 0.014* 18.717 1.186 0.014*
Need

Home, self-care 1.182 0.920 0.856 1.329 1.066 0.790
Health Behaviors
Has advance
directive

0.499 0.836 0.406
30-day readmission 6.474 0.910 0.040*
Model Summary

-2Log likelihood 46.864

46.831

41.988

R
2
0.220

0.220

0.324
Hosmer and
Lemeshow χ
2
2.195

2.542 4.498
*p<.05; **p<.01; ***p<.001

79

Discussion
This chapter aimed to identify the characteristics and risk factors
associated with opting-out of a social work driven transition intervention. Findings
suggest that some at-risk patients may not be receptive to in-home transitions
interventions, with nearly a third of patients opting out of a home visit after
consenting to participate in the study. In the present study, participants that
opted-out of the intervention were significantly more likely to have a respiratory
condition. Literature on older adult intervention opt-outs and medical diagnosis is
sparse, however, a recent study examining recruitment of hospitalized Medicare
patients for behavioral research found that patients who reported a perceived
inability to control important life domains, had low expectations of recovery, or
reported confusion with the researcher’s questions were significantly less likely to
consent to the research (Voss et al., 2013. The authors suggest that stress, self-
expectations for recovery, and health literacy are potential influences on older
adults’ decision to participate in behavioral research. Although participants in the
present study did originally consent to the research and later opted-out of the
home intervention aspect of the study, the constructs that Voss et al. describe
may be particularly prevalent among older adults with respiratory conditions. For
example, chronic obstructive pulmonary disease (COPD) is a common
respiratory condition found to be most prevalent among older adults aged 65-74
years (CDC, 2013) and was the third leading cause of death in 2011(CDC,
2013). Exacerbations of COPD can be significant events among patients with this
condition that would cause them to be hospitalized and require any number of

80

inpatient interventions, each which are considered to have fatal risks (Chenna &
Mannino, 2010; Mannino & Braman, 2007). The constructs of perceived stress,
recovery, and health literacy offered by Voss et al. (2013) could be impacted by
the severity of COPD exacerbations and hospital course of care, and later
translate to intervention opt-outs. Another possible explanation is that patients
with a respiratory disease may feel uncomfortable and opt-out of home visits due
to environmental factors, such as smoking or presence of a smoker in the
residence, that the patient may not want the social worker to see. Further
research among older COPD and other respiratory condition patients is needed,
albeit warranted.
We also found significantly higher odds of 30-day readmission among
patients opting-out of the intervention and having a cancer diagnosis. This finding
suggests that the timing of intervention initiation is important because the
patients you lose after discharge are most likely to return within the 30-day ACA
penalty zone. A previous study by Welch and colleagues (2009) targeting a
similar, high-risk Medicare population as SWIFT found opposite results; in their
randomized control trial of the voluntary Medication Therapy Management (MTM)
program by the Centers for Medicare and Medicaid Services (CMS) for Medicare
Part D patients, they found that intervention opt-outs were less likely to have a
hospitalization as compared to intervention recipients (Welch, Delate, Chester, &
Stubbings, 2009).Conflicting results in these studies highlight other important
findings that show variability in 30-day hospitalization rates are largely driven by
the composition of the patient population they serve, access to care, and

81

availability of community resources (Joynt et al., 2011; Joynt & Jha, 2012). With
patient-level and community factors impacting readmission rates the most,
penalizing hospitals for aspects they cannot control seems misguided. Moreover,
many hospitals have undertaken efforts to improve transitional care provided to
patients, but some interventions have been found to be associated with an
increase in readmission rates, believed to be caused by increases in access to
care and patient satisfaction (Joynt & Jha, 2012; Weinberger, Oddone, &
Henderson, 1996). The conceptual model guiding the present study depicts this
interaction of environmental factors, population characteristics, health behaviors,
and outcomes, explains the influence on health service utilization, and can inform
policy makers of more fair metrics for performance evaluation that take patient-
level factors into account.
Finally, a trend was observed where intervention opts-outs were more
likely to have an advance directive in place as compared to intervention
recipients. Although not significant, this finding is contrary to that of Enguidanos
and colleagues (2012) who found seriously ill patients without an advance
directive were more likely to have a 30-day readmission. Other studies have
identified higher levels of spousal support and family functioning to be
significantly associated with increased odds of advance care planning (Boerner,
Carr, & Moorman, 2013), thus, for our current study, those opting out may have
had higher levels of family support and therefore did not perceive a need for the
intervention.

82

CHAPTER 7: CONCLUSIONS, LIMITATIONS, AND IMPLICATIONS FOR
POLICY AND PRACTICE

Implications
This dissertation investigated effectiveness of an innovative social
work intervention on reducing 30-day hospital readmissions among older adults.
Primary findings from this dissertation hold significant implications for policy and
practice, where 30-day hospital recidivism was driven by patient predisposing
characteristics (Caucasian race), need (being diagnosed with cancer, high acuity,
unmet need for supportive community resources/outside help, and food
prep/shopping help), and health behaviors (intervention opt-out and having had a
prior inpatient stay in the previous six-months), and not influenced by the SWIFT
social work intervention.
Practice
As care transitions interventions are proliferating to help hospitals reduce
readmission rates in light of financial penalties, several important factors from
SWIFT may better inform institutions implementing interventions. The SWIFT
intervention may not be effective at reducing rehospitalizations for reasons
related to the limited social work-lead intervention, patient-level factors of the
target population, or variables specific to the clinical study site.
Implementing a social work intervention in isolation limited to two in-home
visits and four telephone contacts may not be sufficient to reduce short-term
hospital recidivism among at-risk older adults. Contrary to previous studies

83

finding reduced acute health service use from employing social workers in non-
invasive, telephone-based follow-up interventions (Shannon et al., 2006; Kallis et
al., 1996; Monsuez et al., 1993), social workers may not be the right providers to
conduct both in-home and phone-based follow-up. No other published
randomized controlled trials have tested a social work intervention in isolating,
however, recent literature suggests that holistic, multidisciplinary efforts aimed at
reducing older adult hospital recidivism are most effective (Joynt & Jha, 2012)
where a social worker might be included as a part of a transitional care team.
The SWIFT intervention also may not have been effective at reducing 30-
day hospital recidivism due to the specific patient population targeted for this
study: an at-risk sample that was more highly educated than other samples from
care transitions interventions. However, findings from the present study
complement a growing body of literature suggesting that 30-day hospital
readmission rates may be influenced by patient-level factors uncontrollable by
hospitals and interventions, such as race, disease diagnosis, and condition
severity. For example, a recent study by Singh and colleagues (2014) found that
as much as 56.2% of the variation in 30-day hospital recidivism is explained by
patient characteristics. Although no other studies have corroborated our results
by identifying relationships between Caucasian race and diagnosis with cancer
with 30-day hospital recidivism, our findings may point to a group of patients,
already identified as at-risk for readmission, whose condition may be more
serious and management extends beyond the scope of self-management and an
in-home social work intervention. Previous research has also found that more

84

seriously ill patients experience 30-day rehospitalizations at end-of-life (Donze,
Lipsitz, & Schnipper, 2014; Setoguchi, Warner, Stevenson, & Schneeweiss,
2007) and therefore, may represent health utilization behaviors that are
unavoidable or mutable by care transitions interventions. Additionally, our
findings of higher readmission rates among patients that opted-out of the
intervention also supports the literature on personal factors driving hospital
recidivism and identifies patients that may not be amenable to interventions.
Clearly, some at-risk patients may not be receptive to in-home transitions
interventions yet hospitals are being held accountable for patient actions.
Another patient-level factor—unmet supportive community resources and
food needs—also represents an area that hospitals have little control over. To
our knowledge, this is the first study to investigate the impact of different types of
unmet social service needs, including community-based services and food,
beyond caregiver support on 30-day hospital recidivism. Current approaches to
transitions care focus on a medical approach and do not include social aspects.
Moreover, they largely include patient education and planning for follow-up
behaviors (i.e. filling prescriptions, outpatient tests, follow-up with primary care
and specialist physicians, etc.) that require social means such as in-home
support to accomplish recommendations and avoid subsequent physical health
problems. Further investigation into the impact of unmet supportive community
service and food needs on 30-day hospital recidivism is need to confirm the
results of the present study and make a more compelling case for potential
action, however, possible implications for policy and practice exist. Namely, more

85

information in this area could better inform hospitals and community-based
services on the potential expansion of services and better inform policy makers
regarding resource allocation for new and existing programs. Social service
means, such as supportive community resources, as a solution to solving
physical health problems could represent an area of significant cost savings if
found to reduce 30-day readmission and warrants further investigation.
Considering hospital-level variables specific to that clinical site is also
important when investigating the effectiveness of interventions such as SWIFT as
they may contribute to outcomes. For example, the 30-day read readmission rate
among at-risk older adults at the same clinical site as SWIFT in 2007 was found
to be 10.8% in a previous study conducted by Navarro et al. (2012) and was
11.0% in the present study. Previous literature has shown that care transitions
interventions implemented at hospitals with much higher readmission rates
[ranging from 22.0% to 38.1% (Evans & Hendricks, 1993; Jack et al., 2009;
Koehler et al., 2009)] as compared to Huntington Memorial Hospital are effective
at reducing 30-day hospital recidivism. Findings may be attributed to a greater
opportunity presented by these institutions to make a difference. Comparably, the
low, maintained 11% readmission rate at Huntington Memorial Hospital may
indicate that this institution is already doing a good job at addressing short-term
hospital recidivism among their at-risk Medicare beneficiaries and that the
readmissions that do occur are due to patient-level factors beyond their control.
However, in spite of the low readmission rate, in the fall of 2013, Huntington
Memorial Hospital was penalized 0.25% of their Medicare billing for the previous

86

business quarter. Recent research shows that in 2013, two-thirds of the nation’s
hospitals were levied penalties up to one-percent of their Medicare billing for high
30-day readmission rates (Joynt & Jha, 2013a); Huntington Memorial Hospital
being one of these. Hospital characteristics such as large size (400 beds or
more) and practice as a teaching hospital are associated with higher penalties
(Joynt & Jha, 2013a) yet more and more research, including the present study,
points to patient-level factors driving short-term hospital recidivism.
Policy
Emerging research on the types of institutions more highly penalized by
the ACA Hospital Readmission Reduction Program is also a critical piece that
supports evidence indicating that factors driving 30-day readmissions are largely
out of the hospital’s control. Current policies have charged hospitals with the task
of reducing 30-day readmission rates, however, these results suggest that it may
not be reasonable to place this burden solely on hospitals. Policies should
potentially be re-evaluated for their appropriateness, fairness, and efficacy at
addressing reductions in 30-day hospitalization rates among older Medicare
beneficiaries. More research is needed to confirm these findings and better
understand patient characteristics associated with 30-day hospital recidivism
among at-risk older adults.

87

Limitations
Results of the present study may be limited in several important ways.
First, the sample size in sub-sample analyses between intervention opt-outs and
intervention recipients may weaken the statistical power to detect differences
between study groups, although, parsimonious models were analyzed to account
for this. Also, study participants were recruited from a single, large, non-profit,
urban, Los Angeles area hospital and while the diversity of Los Angeles is
notable, results may not be generalizable to other areas. Along these lines, the
average educational attainment and annual income among participants was
considerably higher than those typically found in studies targeting an at-risk
Medicare population for acute health service utilization. Due to the nature of the
pilot study, only cognitively intact, English-speaking older adults were eligible to
participate and, therefore, this sample may not be representative of hospitalized
older adult patients and recidivism. Finally, hospital readmissions within 30-days
were only collected at Huntington Memorial Hospital and do not capture any
readmissions at other hospitals. Use of CMS data to account for other possible
readmissions at other institutions could strengthen this study, however, Hansen
et al. (2011) report in their systematic review of other care transitions
interventions targeting at-risk older adults that the vast majority of studies
similarly relied on single-site readmission data.

88

Conclusion
This dissertation investigated the effectiveness of a new and innovative
approach to transitional care by utilizing a social work-driven intervention to
reduce 30-day rehospitalizations among at-risk older adults. The intervention was
not effective, however, key findings contribute to the literature in this area and
complements emerging research suggesting that not all hospital readmissions
are avoidable; that they are largely driven by patient-level factors outside of the
hospital’s control such as predisposing characteristics, need, and health
behaviors. Placing the burden of readmission rates and associated penalties
solely on hospitals may be inappropriate and warrants reconsideration.

89

BIBLIOGRAPHY

Ahmed, A., Thornton, P., Perry, G. L., Allman, R. M., & DeLong, J. F. (2004).
Impact of atrial fibrillation on mortality and readmission in older adults
hospitalized with heart failure. European Journal of Heart Failure, 6, 421-6.
AHRQ. (2004). Literacy and health outcomes. Retrieved June 28, 2013, from
http://www.ahrq.gov/clinic/tp/littp.htm
Albert, S., Im, A., Brenner, L., Smith, M., & Waxman, R. (2002). Effect of social
work liasion program on family caregivers to people with brain injury.
Journal of Head Trauma Rehabilitation, 17(2), 175-189.
Alkema, G. E., Wilber, K. H., Shannon, G. R., & Allen, D. (2007). Reduced
mortality: The unexpected impact of a telephone-based care management
intervention for older adults in managed care. Health Services Research,
42(4), 1632-50.
Allaudeen, N., Schnipper, J.,L., Orav, E. J., Wachter, R. M., & Vidyarthi, A., R.
(2011). Inability of providers to predict unplanned readmissions. Journal of
General Internal Medicine, 26(7), 771-6.
Altfeld, S. J., Shier, G. E., Rooney, M., Johnson, T. J., Golden, R. L., Karavolos,
K., . . . Perry, A. J. (2012). Effects of an enhanced discharge planning
intervention for hospitalized older adults: A randomized trial. The
Gerontologist, 53(3), 430-440.

90

American Hospital Association. (2009). Teaching hospitals: Their impact on
patients and the future health care workforce. ( No. September).
Andersen, R. M. (1968). Behavioral model of families' use of health services
(Health Series No. 25 ed.). Chicago, Il: Center for Health Administrative
Services.
Andersen, R. M. (1996). Revisiting the behavioral model and access to medical
care: Does it matter? Journal of Health Society and Behavior, 36 (March), 1-
10.
Andersen, R. M. (1995). Revisiting the behavioral model and access to medical
care: Does it matter? Journal of Health and Social Behavior, 36(1), 1-10.
Andersen, R. M., & Newman, J. F. (1973). Societal and individual determinants
of medical care utilization in the united states. Milbank Memorial Fund
Quarterly Journal, 51, 95-124.
Anderson, C., Deepak, B. V., Amoateng-Adjepong, Y., & Zarich, S. (2005).
Benefits of comprehensive impatient education and discharge planning
combined with outpatient support in elderly patients with congestive heart
failure. Congestive Heart Failure, 11, 315-21.
Anderson, M. A., Helms, L. B., Hanson, K. S., & DeVilder, N. W. (1999).
Unplanned hospital readmissions: A home care perspective. Nursing
Research, 48(6), 299-307.

91

Arean, P., Hegel, M., Vannoy, S., Fan, M., & Unuzter, J. (2008). Effectiveness of
problem-solving therapy for older, primary care patients with depression:
Results from the IMPACT project. The Gerontologist, 48(3), 311-323.
Atkinson, V., & Nelson, G. (1995). Adult protective services. In G. Nelson, A.
Eller, D. Streets & M. Morse (Eds.), The field of adult services: Social work
practice and administration (pp. 215-230). Washington D.C.: NASW Press.
Azuero, C., Allen, R. S., Kvale, E., Azuero, A., & Parmelee, P. (2013).
Determinants of psychology service utilization in a palliative care outpatient
population. Psycho-Oncology, Nov 28
Balaban, R. B., Weissman, J. S., Samuel, P. A., & Woolhandler, S. (2008).
Redefining and redesigning hospital discharge to enhance patient care: A
randomized controlled study. Journal of General Internal Medicine, 23(8),
1228-33.
Boerner, K., Carr, D., & Moorman, S. (2013). Family relationships and advance
care planning: Do supportive and critical relations encourage or hinder
planning? Journals of Gerontology Series B: Psychological Sciences and
Social Sciences, 68(2), 246-256.
Bostrom, J., Caldwell, J., McGuire, K., & Everson, D. (1996). Telephone follow-
up after discharge from the hospital: Does it make a difference? Applied
Nursing Research, 9, 47-52.

92

Buerhaus, P. I. (2002). Shortages of hospital registered nurses: Causes and
perspectives on public and private sector actions. Nursing Outlook, 50(1), 4-
6.
Centers for Disease Control and Prevention. (2009). The power of prevention:
Chronic disease...the public health challenge of the 21st century. (No.
CS201478).
Centers for Disease Control and Prevention. (April 25, 2013). Chronic obstructive
pulmonary disease (COPD). Retrieved October 13, 2013, from
http://www.cdc.gov/copd/
Centers for Medicare & Medicaid Services. (2012). Community-based care
transitions program. Retrieved July 15, 2013, from
http://innovation.cms.gov/initiatives/CCTP/?itemID=CMS1239313
Centers for Medicare and Medicaid Services (CMS). (April, 2012). Historical
overview of national health expenditures. Retrieved June 26, 2012, from
http://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-
Trends-and-
Reports/NationalHealthExpendData/index.html?redirect=/NationalHealthExp
endData/
Chenna, P. R., & Mannino, D. M. (2010). Outcomes of severe COPD
exacerbations requiring hospitalization. Seminars in Respiratory and Critical
Care Medicine, 31(3), 286-94.

93

Claiborne, N. (2003). Efficiency of a care coordination model: A randomized
study with stroke patients. Research on Social Work Practice, 16(1), 57-66.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).
New York: Academic Press.
Coleman EA, Parry C, Chalmers S,Min S. (2006). The care transitions
intervention: Results of a randomized controlled trial. Archives of Internal
Medicine, 166(17), 1822-1828.
Coleman, E. A., & Williams, M. V. (2007). Executing high-quality care transitions:
A call to do it right. Journal of Hospital Medicine, 2(5), 287-290.
Coleman, E. A. (2003). Falling through the cracks: Challenges and opportunities
for improving transitional care for persons with continuous complex care
needs. Journal of the American Geriatrics Society, 51(4), 549-555.
Coleman, E. A., Mahoney, E., & Parry, C. (2005). Assessing the quality of
preparation for posthospital care from the patient's perspective: The care
transitions measure. Medical Care, 43(3), 246-55.
Coleman, E. A., Min, S. J., Chomiak, A., & Kramer, A. M. (2004). Posthospital
care transitions: Patterns, complications, and risk identification. Health
Services Research, 39(5), 1449-1465.
Coleman, E. A., Parry, C., Chalmers, S., & Min, S. J. (2006). The care transitions
intervention: Results of a randomized controlled trial. Archives of Internal
Medicine, 166(17), 1822-1828.

94

Coleman, E. A., Smith, J. D., Frank, J. C., Eilertsen, T. B., Thiare, J. N., &
Kramer, A. M. (2002). Development and testing of a measure designed to
assess the quality of care transitions. International Journal of Integrated
Care, 2, e02.
Coleman, E. A., Smith, J. D., Raha, D., & Min, S. J. (2005). Posthospital
medication discrepancies: Prevalence and contributing factors. Archives of
Internal Medicine, 165(16), 1842-1847.
Coleman, K., Austin, B. T., Brach, C., & Wagner, E. H. (2009). Evidence on the
chronic care model in the new millennium. Health Affairs, 28(1), 75-85.
Corley, M. C., & Mauksch, H. O. (1993). The nurse's multiple commitments.
Journal of Professional Nursing, 9(2), 116-122.
Cornish, P. L., Knowles, S. R., Marchesano, R., Tam, V., Shadowitz, S., Juurlink,
D. N., & Etchells, E. E. (2005). Unintended medication discrepancies at the
time of hospital admission. Archives of Internal Medicine, 165, 424-429.
Creason, H. (2001). Congestive heart failure telemanagement clinic. Lippincotts
Case Management, 6, 146-56.
Dedhia, P., Kravet, S., Bulger, J., Hinson, T., Sridharan, A., Kolodner, K., . .
.Howell, E. (2009). A quality improvement intervention to facilitate the
transition of older adults from three hospitals back to their homes. Journal of
the American Geriatrics Society, 57, 1540-6.

95

Diehr, P., Yanez, D., Ash, A., Hornbrook, M., & Lin, D. Y. (1999). Methods for
analyzing health care utilization and costs. Annual Review of Public Health,
20, 125-144.
Donze, J., Aujesky, D., Williams, D., & Schnipper, J. L. (2013). Potentially
avoidable 30-day hospial readmissions in medical patients. JAMA: Internal
Medicine, 173(8), 632-638.
Donze, J., Lipsitz, S., & Schnipper, J. L. (2014). Risk factors for potentially
avoidable readmissions due to end-of-life care issues. Journal of Hospital
Medicine, 000, 000-000.
Dudas, V., Bookwalter, T., Kerr, K. M., & Pantilat, S. Z. (2001). The impact of
follow-up telephone calls to patients after hospitalization. American Journal
of Medicine, 111(9B), 26S-30S.
Dziegielewski, S. (1998). The changing face of healthcare social work:
Professional practice in the era of managed care. New York: Springer
Publishing Company.
D'Zurilla, T. J., & Goldfried, M. R. (1971). Problem solving and social behavior
modification. Journal of Abnormal Psychology, 78, 107-126.
Einstadter, D., Cebul, R. D., & Franta, P. R. (1996). Effect of a nurse case
manager on postdischarge follow-up. Journal of General Internal Medicine,
11(11), 684-688.

96

Enguidanos, S., Cherin, D., & Brumley, R. (2005). Home-based palliative care
study: Site of death, and costs of medical care for patients with congestive
heart failure, chronic obstructive pulmonary disease, and cancer. Journal of
Social Work in End of Life & Palliative Care, 1(3), 37-56.
Enguidanos, S., Coulourides Kogan, A., Keefe, B., Geron, S., & Katz, L. (2011).
Patient-centered approach to builiding problem-solving skills among older
primary care patients: Problems identified and resolved. Journal of
Gerontological Social Work, 54, 276-291.
Enguidanos, S., Vesper, & Lorenz, K. A. (2012). 30 day readmissions among
seriously ill older adults. Journal of Palliative Medicine, 15(12), 1356-61.
Evans, R. L., & Hendricks, R. D. (1993). Evaluating hospital discharge planning:
A randomized clinical trial. Medical Care, 31(4), 358-370.
Fabbre, V. D., Buffington, A. S., Altfeld, S. J., Shier, G. E., & Golden, R. L.
(2011). Social work and transitions of care: Observations from an
intervention for older adults. Journal of Gerontological Social Work, 54(6),
615-626.
Fall Prevention Center of Excellence. (2014). Information for researchers and
educators. Retrieved December 12, 2013, from http://stopfalls.org/wp-
content/uploads/2013/09/ProgramExpansion-HomeAssessmentTool1.pdf;
www.stopfalls.org

97

Feigenbaum, P., Neuwrith, E., Trowbridge, L., Teplitsky, S., Barnes, C., Fireman,
E., . . . Bellows, J. (2012). Factors contributing to all-cause 30-day
readmissions: A structured case series across 18 hospitals. Medical Care,
00(00), 000-000.
Forster, A. J., Murff, H. J., Peterson, J. F., Gandhi, T. K., & Bates, D. W. (2003).
The incidence and severity of adverse events affecting patients after
discharge from the hospital. Annals of Internal Medicine, 138(3), 161-167.
Forster, A. J., Clark, H. D., Menard, A., Dupuis, N., Chernish, R., Chandok, N., . .
. van Walraven, C. (2004). Adverse events among medical patients after
discharge from hospital. CMAJ, 170(3), 345-349.
Friedman, B., & Basu, J. (2004). The rate and cost of hospital readmissions for
preventable conditions. Medical Care Research and Review, 61(2), 225-240.
Gabbay, R. A., Bailit, M. H., Mauger, D. T., Wagner, E. H., & Siminerio, L. (2011).
Multipayer patient-centered medical home implementation guided by the
chronic care model. The Joint Commission Journal on Quality and Patient
Safety, 37(6), 265-273.
Gafft, C. A., McDonald, F. S., Ruud, K. L., Liesinger, J. T., Johnson, M. G., &
Naessens, J. M. (2010). Effect of hospital follow-up appointment on clinical
event outcomes and mortality. Archives of Internal Medicine, 170, 955-60.

98

Geron, S. (2000a). Care management in the united states. In M. White, & R.
Applebaum (Eds.), Key issues in case management around the globe. San
Francisco, CA: American Society on Aging.
Geron, S. (2000b). Measuring the quality and success of care management. In
M. White, & R. Applebaum (Eds.), Key issues in case management around
the globe. San Francisco, CA: American Society on Aging.
Geron, S., Andrews, C., & Kuhn, K. (2005). Infusing aging skills into the social
work practice community: A new look at strategies for continuing
professional education. Families in Society: The Journal of Contemporary
Social Services, 86(3), 431-440.
Geron, S., & Chassler, D. (1994). Guidelines for case management practive
across the long-term care continuum. Bristol, CT: The Robert Wood Johnson
Foundation.
Geron, S. M., & Keefe, B. (2006). Moving evidence-based interventions to
populations: A case study using social workers in primary care. Home Health
Care Services Quarterly, 25(1-2), 95-113.
Gorodeski, E. Z., Starling, R. C., & Blackstone, E. H. (2010). Are all readmissions
bad readmissions? New England Journal of Medicine, 363(3), 297-298.
Grant, D., & California Discharge Planning Collaborative. (2011). What a
difference a day makes: Part I summary report. San Francisco, CA:
California Discharge Planning Collaborative.

99

Grimmer, K., Moss, J., & Falco, J. (2004). Experiences of elderly patients
regarding independent community living after discharge from hospital: A
longitudinal study. International Journal for Quality in Health Care, 16(6),
465-472.
Gruneir, A., Dhalla, I. A., van Walraven, C., Fischer, H. D., Camacho, X.,
Rochon, P. A., & Anderson, G. M. (2011). Unplanned readmissions after
hospital discharge among patients identified as being at high risk for
readmission using a validated predictive algorithm. Open Medicine, 5(2),
e104-e111.
Hansen, L. O., Young, R. S., Hinami, K., Leung, A., & Williams, M. V. (2011).
Interventions to reduce 30-day rehospitalization: A systematic review. Annals
of Internal Medicine, 155(8), 520-528.
Harrison, A., & Verhoef, M. (2002). Understanding coordination of care from the
consumer's perspective in a regional health system. Health Services
Research, 37, 1031-1054.
Harrison, P. L., Harra, P. A., Pope, J. E., Young, M. C., & Rula, E. Y. (2011). The
impact of postdischarge telephonic follow-up on hospital readmissions.
Population Health Management, 14, 27-32.
Heider, D., Matschinger, H., Muller, H., Saum, K., Quinzler, R., Haefeli, W. E., . .
. Konig, H. (2014). Health care costs in the eldergly in germany: An analysis
applying andersen's behavioral model of health care utilization. BMC Health
Services Research, 14(71), 1-12.

100

Hernandez, A. F., Greiner, M. A., Fonarow, G. C., Hammill, B. G., Heidenreich,
P. A., Yancy, C. W., . . . Curtis, L. H. (2010). Relationship between early
physician follow-up and 30-day redmission among medicar beneficiaries
hospitalized for heart failure. Journal of the American Medical Association,
303, 1716-22.
Hess, D. R., Tokarczyk, A., O'Malley, M., Gavaghan, S., Sullivan, J., & Schmidt,
U. (2010). The value of adding a verbal report to written handoffs on early
readmisison following prolonged respiratory failure. Chest, 138, 1475-9.
HRSA. (2013). The U.S. nursing workforce: Trends in supply and education. (No.
April).
Indeed.com. (2014a). Advanced practice nurse salary in los angeles, CA.
Retrieved April 3, 2014, from
http://www.indeed.com/salary?q1=advanced+practice+nurse&l1=Los+Angel
es%2C+CA&tm=1
Indeed.com. (2014b). Nurse practitioner salary in los angeles, CA. Retrieved
April 3, 2014, from
http://www.indeed.com/salary?q1=nurse+practitioner&l1=Los+Angeles%2C+
CA&tm=1
Jack, B. W., Chetty, V. K., Anthony, D., Greenwald, J. L., Sanchez, G. M.,
Johnson, A. E., . . . Culpepper, L. (2009a). A reeingineered hospital
discharge program to decrease rehospitalization: A randomized trial. Annals
of Internal Medicine, 150(3), 178-187.

101

Jack, B. W., Chetty, V. K., Anthony, D., Greenwald, J. L., Sanchez, G. M.,
Johnson, A. E., . . . Culpepper, L. (2009b). A reengineered hospital
discharge program to decrease RehospitalizationA randomized trial. Annals
of Internal Medicine, 150(3), 178-187.
Jencks, S. F., Williams, M. V., & Coleman, E. A. (2009). Rehospitalizations
among patients in the medicare fee-for-service program. New England
Journal of Medicine, 360(14), 1418-1428.
Johnson, J. E., Billingsley, M. C., & Costa, L. L. (2006). Xtreme nursing and the
nurse shortage. Nursing Outlook, 54(5), 294-299.
Joynt, K., & Jha, A. K. (2013). A path forward on medicare readmissions. New
England Journal of Medicine, 368, 1175-1177.
Joynt, K. E., & Jha, A. K. (2013). Research letter: Characteristics of hospitals
receiving penalties under the hospital readmissions reduction program.
Journal of the American Medical Association, 23(30), 342-343.
Joynt, K. E., Orav, E. J., & Jha, A. K. (2011). Thrity-day readmission rates for
medicare beneficiaries by race and site of care. Journal of the American
Medical Association, 305, 675-81.
Joynt, K. E., & Jha, A. K. (2012). Thirty-day readmissions — truth and
consequences. New England Journal of Medicine, 366(15), 1366-1369.

102

Kallis, J. M., Gonzalez del Rey, J. A., Ruddy, R. M., & Schubert, C. J. (1999).
Linking emergency department patients to primary care: Phone vs face-to-
face interaction. Academic Emergency Medicine, 6(8), 817-822.
Koehler, B. E., Richter, K. M., Youngblood, L., Cohen, B. A., Prengler, I. D.,
Cheng, D., & Masica, A. L. (2009). Reduction of 30-day postdischarge
hospital readmissions or emergency department (ED) visit rates in high-risk
elderly medical patients through delivery of a targeted care bundle. Journal
of Hospital Medicine, 4, 211-218.
Kramer, J. S., Hopkins, P. J., Rosendale, J. C., Garrelts, J. C., Hale, L. S., Neter,
T. M., . . . Haneke, R. D. (2007). Implementation of an electronic system for
medication reconciliation. American Journal of Health-System Pharmacy, 64,
404-22.
Kravitz, R. L., Reuben, D. B., Davis, J. M., Mitchell, A., Hemmerling, K., Kingston,
R. S., & Siu, A. L. (1994). Geriatric home assessment after hospital
discharge. Journal of the American Geriatric Society, 42(12), 1229-1234.
Kroenke, K., & Spitzer, R. L. (2002). The PHQ-9: A new depression and
diagnostic severity measure. Psychiatric Annals, 32(9), 1-7.
Kroenke, K., Spitzer, R. L., & Williams, J. B. W. (2001). The PHQ-9: Validity of a
brief depression severity measure. Journal of General Internal Medicine, 16,
606-613.

103

Lin, E. H. B., Katon, W., Von Korff, M., Tang, L., Williams, J. W.,Jr, Kroenke, K., .
. . Unutzer, J. (2003). Effect of improving depression care on pain and
functional outcomes among older adults with arthritis: A randomized
controlled trial. Journal of the American Medical Association, 290(18), 2428-
2429.
Lucas, K. S. (1998). Outcomes evaluation of a pharmacist discharge medication
teaching service. American Journal of Health-System Pharmacy, 55, S32-5.
Magilvy, J. K., & Congdon, J. G. (2000). The crisis nature of health care
transitions for rural older adults. Public Health Nursing, 17(5), 336-345.
Mannino, D. M., & Braman, S. (2007). The epidemiology and economics of
chronic obstructive pulmonary disease. Proceedings of the American
Thoracic Society, 4, 502-506.
Marcantonio, E. R., McKean, S., Goldfinger, M., Kleefield, S., Yurkofsky, M., &
Brennan, T. A. (1999). Factors associated with unplanned hospital
readmission among patients 65 years of age and older in a medicare
managed care plan. The American Journal of Medicine, 107(1), 13-17.
McPhee, S. J., Frank, D. H., Lewis, C., Bush, D. E., & Smith, C. R. (1983).
Influence of a "discharge interview" on patient knowledge, compliance, and
functional status after hospitalization. Medical Care, 21, 755-67.
MedPAC. (2007). Report to the congress: Promoting greater efficiency in
medicine. Washington DC: Medicare Payment Advisory Commission.

104

Misky, G. J., Wald, H. L., & Coleman, E. A. (2010). Post-hospitalization
transitions: Examining the effects of timing of primary care provider follow-
up. Journal of Hospital Medicine, 5, 392-7.
Monsuez, J., Fergelot, H., Papon, B., & Le Gall, J. (1993). Early social
intervention in the emergency department. European Journal of Medicine,
2(8), 489-492.
Moore, C., Wisnivesky, J., Williams, S., & McGinn, T. (2003). Medical errors
related to discontinuity of care from an inpatient to an outpatient setting.
Journal of General Internal Medicine, 18(8), 646-651.
Morrill, G. B., & Barreuther, C. (1988). Screening discharge prescriptions.
American Journal of Hospital Pharmacy, 45(9), 1904-1905.
NASW. (1992). National association of social workers (NASW) standards for
social work case management. Washington, D.C.: NASW.
Navarro, A., Enguidanos, S., & Wilber, K. H. (2012). Identifying risk of hospital
readmission among Medicare aged patients: an approach using routinely
collected data. Home Health Care Service Quarterly. 31:2, 181-195.
Naylor, M. D., Aiken, L. H., Kurtzman, E. T., Olds, D. M., & Hirschman, K. B.
(2011). The importance of transitional care in achieving health reform. Health
Affairs, 30(4), 746-754.
Naylor, M., Brooten, D., Campbell, R., Jacobsen, B. S., Mezey, M. D., Pauly, M.
V., & Schwartz, J. S. (1999). Comprehensive discharge planning and home

105

follow-up of hospitalized elders. Journal of the American Medical
Association, 281(7), 613-620.
Nickolaus, T., Specht-Leible, N., Bach, M., Oster, P., & Schlierf, F. (1999). A
randomized trial of comprehensive geriatric assessment and home
intervention in the care of hospitalized patients. Age and Ageing, 28(6), 543-
550.
O'Dell, K. M., & Kucukarslan, S. N. (2005). Impact of the clinical pharmacist on
readmission in patients with acute coronary syndrome. Annals of
Pharmacotherapy, 39, 1423-7.
Parry, C., Min, S., Chugh, A., Chalmers, S., & Coleman, E. A. (2009). Further
application of the care transitions intervention: Results of a randomized
controlled trial conducted in a fee-for-service setting. Home Health Care
Services Quarterly, 28, 84-99.
Partners in Care Foundation. (2013). Community-based medications
management. Retrieved December 12, 2013, from
https://www.picf.org/landing_pages/23,3.html;
https://homemeds.homemeds.org/
Patient protection and affordable care act, pub. L. 111-148, 2702, 124 Stat. 119,
318-319Cong. (2010).
Pfeiffer, E. (1975). A short portable mental status questionnaire for the
assessment of organic brain deficit in elderly patients. Supportive Care in

106

Cancer : Official Journal of the Multinational Association of Supportive Care
in Cancer, 23(10), 433-441.
Project RED. (2014). Project RED (re-engineered discharge): Toolkit. Retrieved
January 2, 2014, from http://www.bu.edu/fammed/projectred/toolkit.html
Rainville, E. C. (1999). Impact of pharmacist interventions on hospital
readmissions for heart failure. American Journal of Health-System
Pharmacy, 56, 1339-42.
Rathore, S. S., Foody, J. M., Yang, Y., Smith, G. L., Herrin, J., Masoudi, F. A., . ..
Krumholtz, H. M. (2003). Race, quality of care, and outcomes of elderly
patients hospitalized with heart failure. Journal of the American Medical
Association, 289(19), 2517-2524.
Robert Wood Johnson Foundation. (2013). The revolving door: A report on US
hospital readmissions. RWJF.
Robinson, J., Shugrue, N., Fortinsky, R. H., & Gruman, C. (2014). Long-term
supports and service planning for the future: Implications from a statewide
survey of baby boomers and older adults. The Gerontologist, 54(2), 297-313.
Rosen, A., & Teeson, M. (2001). Does case management work? the evidence
and the abuse of evidence-based medicine. The Australian and New
Zealand Journal of Psychiatry, 35(6), 731-746.
Rush University Medical Center. (2013). Enhanced discharge planning program.
Retrieved, January 2, 2014,

107

fromfile:///C:/Users/Michael/Downloads/ITCC_Bridge%20Model_Summary%
20(1).pdf; http://www.rush.edu/rumc/page-1298330101593.html
Scharlach, A., Simon, J., & Dal Santo, T. (2002). Aging of california's baby
boomers: Implications for state policies, programs, and personnel (CPRC
report). Berkeley, CA: Center for Advanced Services.
Schneider, J. K., Hornberger, S., Booker, J., Davis, A., & Kralicek, R. (1993). A
medication discharge planning program: Measuring the effect on
readmissions. Clinical Nursing Research, 2, 41-53.
Setoguchi, S., Warner Stevenson, L., Schneeweiss, S. (2007). Repeated
hospitalizations predict mortality in the community population with heart
failure. American Heart Journal, 154, 260-6.
Shannon, G. R., Wilber, K., & Allen, D. (2006). Reductions in costly healthcare
service utilization: Findings from the care advocate program. Journal of the
American Geriatric Society, 54, 1102-1107.
Sharma, G., Kuo, Y. F., Freeman, J. L., Zhang, D. D., & Goodwin, J. S. (2010).
Outpatient follow-up visit and 30-day emergency department visit and
readmission in patients hospitalized for chronic obstructive pulmonary
disease. Archives of Internal Medicine, 170, 1664-70.
Simon, G. E., Ludman, E. J., Tutty, S., Operskalski, B., & Von Korff, M. (2004).
Telephone psychotherapy and telephone care management for primary care

108

patient starting antidepressant treatment: A randomized controlled trial.
Journal of the American Medical Association, 292(8), 935-942.
Singh, S., Lin, Y., Kuo, Y., Nattinger, A. B., & Goodwin, J. S. (2013). Variation in
the risk of readmission among hospitals: The relative contribution of patient,
hospital and inpatient provider characteristics. Journal of General Internal
Medicine, 29(4), 572-8.
Sobel, H. R., Medina-Walpole, A. M., & Katz, P. R. (2004). Information transfer
between levels of care: A pilot study. Journal of the American Geriatrics
Society, 52(Supplement 1), S97.
Stone, R., Dawson, S. & Harahan, M. (2003). Why workforce development
should be part of the long-term care quality debate. Retrieved December 1,
2004, from
http://phinational.org/sites/phinational.org/files/clearinghouse/LTC%20Quality
%20Debate%20Paper1.pdf
Sumner, J., & Townsend-Rocchiccioli, J. (2003). Why are nurses leaving
nursing? Nursing Administration Quarterly, 27(2), 164-171.
Tarlov, A. R., Ware, J. E. J., Greenfield, F., Nelson, E. C., Perrin, E., & Zubkoff,
M. (1989). The medical outcomes study. an application of methods for
monitoring the results of medical care. Journal of the American Medical
Association, 262(7), 925-930.

109

Tuggle, C. T., Park, L. S., Roman, S., Udelsman, R., & Sosa, J. A. (2010).
Rehospitalization among elderly patients with thyroid cancer after
thyroidectomy are prevalent and costly. Annals of Surgical Oncology, 17,
2816-2823.
Unutzer, J., Katon, W., & Callahan, C. (2003). Depression treatment in a sample
of 1,801 depressed older adults in primary care. Journal of the American
Geriatrics Society, 51, 505-514.
Unutzer, J., Katon, W., Callahan, C. M., Williams, J. W. J., Hunkeler, E., Harpole,
L., . . . IMPACT Investigators. Improving Mood-Promoting Access to
Collaborative Treatment. (2002). Collaborative care management of late-life
depression in the primary care setting: A randomized controlled trial. Journal
of the American Medical Association, 288(22), 2836-45.
Unutzer, J., Katon, W., Williams, J. W. J., Callahan, C. M., Harpole, L., Hunkeler,
E. M., . . . Langston, C. A. (2001). Improving primary care for depression in
late life: The design of a multicenter randomized trial. Medical Care, 39(8),
785-99.
Unutzer, J., Patrick, D. L., Marmon, T., Simon, G. E., & Katon, W. J. (2002).
Depressive symptoms and mortality in a prospective study of 2,558 older
adults. American Journal of Geriatric Psychiatry, 10(5), 521-30.
van Walraven, C., Mamdani, M., Fang, J., & Austin, P. C. (2004). Continuity of
care and patient outcomes after hospital discharge. Journal of General
Internal Medicine, 19, 624-631.

110

van Walraven, C., Seth, R., Austin, P. C., & Laupacis, A. (2002). Effect of
discharge summary availability during post-discharge visits on hospital
readmission. Journal of General Internal Medicine, 17(3), 186-192.
Volland, P., & Keepnews, D. (2006). Generalized and specialized hospitals. In B.
Berkman (Ed.), Handbook of social work in health and aging. Oxford: Oxford
University Press.
Voss, R., Gavenstein, S., Baier, R., Butterfield, K., Epstein-Lubow, G., Shamji,
H., & Gardner, B. (2013). Recruiting hospitalized patients for research: How
do participants differ from eligible nonparticipants? Journal of Hospital
Medicine, 8(4), 204-14.
Vourlekis, B. S., Gelfand, D. E., & Greene, R. R. (1992). Psychosocial needs and
care in nursing homes: Comparison of views of social workers and home
administrators. Gerontologist, 32(1), 113-119.
Wagner, E. H., Austin, B. T., Davis, C., Hindmarsh, M., Schaefer, J., & Bonomi,
A. E. (2001). Improving chronic illness care: Translating evidence into action;
interventions that encourage people to acquire self-management skills are
essential to chronic illness care. Health Aff (Millwood), 20(6), 64-78.
Wagner, E. H., Austin, B. T., & Von Korff, M. (1996). Improving outcomes in
chronic illness. Managed Care Quarterly, 4(2), 12-25.

111

Ware, J. E. J. & Sherbourne, C. D. (1992). The MOS 36-item short-form health
survey (SF-36). Conceptual framework and item selection. Medical Care,
30(6), 473-483.
Weaver, F. M., Perloff, L., & Waters, T. (1998). Patients' and caregivers'
transition from hospital to home: Needs and recommendations. Home Health
Care Service Quarterly, 17(3), 27-48.
Weinberger, M., Oddone, E. Z., & Henderson, W. G. (1996). Does increased
access to primary care reduce hospital readmissions? New England Journal
of Medicine, 334, 1441-7.
Welch, E. K., Delate, T., Chester, E. A., & Stubbings, T. (2009). Assessment of
the impact of medication therapy management delivered to home-based
medicare beneficiaries. Annals of Pharmacotherapy, 43(4), 603-10.
Williams, M., Williams, T., Zimmer, J., Hall, W., & Podgorski, C. (1987). How
does the team approach to outpatient geriatric evaluation compare with
traditional care: A report of a randomized controlled trial. Journal of the
American Geriatrics Society, 35(12), 1071-1078.

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Exploring end-of-life care experiences and disparities among Hispanic populations in the United States

PDF

Palliative care: breaking the knowledge barrier among community-based older adults

PDF

Fall related injuries among older adults in the Los Angeles region

PDF

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

PDF

Using mixed methods to identify the characteristics of older fraud victims

PDF

Thinking generatively versus acting generatively: exploring the associations of generative self-concept and generative activity with cognitive function among older adults

PDF

New directions for family caregiver interventions

PDF

What is success? an exploration of person-centered intervention of complex elder mistreatment

PDF

Translating two physical activity programs for older adults into home- and community-based settings: "Active Start" and "Healthy Moves for Aging Well"

PDF

Planning care with the patient in the room: a patient-focused approach to reducing heart failure readmissions

PDF

The effect of body weight on mortality: different countries and age groups

PDF

Forensic markers of physical elder abuse in medical and community contexts: implications for criminal justice interventions

PDF

Intergenerational support between grandparents and grandchildren in rural China and its effect on the psychological well-being of older adults

PDF

Falls among older adults: characteristics of fallers, predictors of falls, and the impact of falls on health care and long-term care utilization

PDF

Improving hospital readmission rates: the benefit of a transitional case management program

PDF

Nursing home transitions: a new framework for understanding preferences, barriers and outcomes

PDF

Mechanisms of stress effects on learning and decision making in younger and older adults

PDF

Older adults with visual impairments: the role of health dimensions in predicting falls

PDF

Racial disparities in advance care planning and directives completion and end-of-life outcomes in the United States

PDF

Examination of the long-term psychosocial and functioning effects of a healthy living, life management behavior intervention for older adults

Asset Metadata

Creator Kogan, Alexis Marie Coulourides (author)

Core Title Investigating the effectiveness of a social work intervention on reducing hospital readmissions among older adults

School Leonard Davis School of Gerontology

Degree Doctor of Philosophy

Degree Program Gerontology

Publication Date 08/06/2014

Defense Date 04/22/2014

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

Tag care transitions,hospital readmissions,OAI-PMH Harvest,older adults,randomized control trial,Social Work

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Enguídanos, Susan M. (committee chair), Nichol, Michael B. (committee member), Wilber, Kathleen H. (committee member)

Creator Email acoulour@usc.edu,acoulourides@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-456017

Unique identifier UC11287054

Identifier etd-KoganAlexi-2782.pdf (filename),usctheses-c3-456017 (legacy record id)

Legacy Identifier etd-KoganAlexi-2782.pdf

Dmrecord 456017

Document Type Dissertation

Format application/pdf (imt)

Rights Kogan, Alexis Marie Coulourides

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 a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA