PAIN AND FATIGUE SEVERITY AND FREQUENCY:
CONGRUENCE BETWEEN CANCER PATIENTS AND THEIR FAMILY CAREGIVERS
By
Chia Tai Hung

A DISSERTATION
Submitted to Michigan State University
In Partial Fulfillment of the requirements
For the degree of
DOCTOR OF PHILOSOPHY
Nursing
2012

ABSTRACT
PAIN AND FATIGUE SEVERITY AND FREQUENCY:
CONGRUENCE BETWEEN CANCER PATIENTS AND THEIR FAMILY CAREGIVERS

By
Chia Tai Hung
Family caregivers are asked to provide patient care, specifically assisting symptom
management, at home, but their reliability of symptom management and assessment for patients
is challenged because of the low agreement between patients’ and caregivers’ symptom reports.
There is a gap in the knowledge in that a lack of a theory or model has been used to illustrate the
scenario of a low level of congruence of symptom perception between patients and caregivers.
The study developed the framework of Empathic Accuracy of Symptom Perception (EASP)
based on Davis’s Empathy Model and Ickes’s Empathic Accuracy. Given’s Family Care Model
was used to support empathic accuracy in the context of the caregiving environment. EASP was
conceptually defined as the similarities in the cognitive appraisal of patients’ cancer symptom
experiences between family caregivers and patients. Empathetic accuracy was operationally
defined as congruence of pain and fatigue severity and frequency. The specific aim of this
research was to examine how caregivers’ emotionality (caregiver depression) and caregivers’
threats of caregiving (caregiver burden) predict EASP (congruence of pain and fatigue severity
and frequency), when taking into account contextual factors of caregivers (age, sex, education,
relationship to patients, observation of the total number of patients’ symptoms, and living
arrangement).
This is a cross-sectional, descriptive study adopting the data from the study entitled
Family Home Care for Cancer- A Community-Based Model (# R01Ca-79280). A total of 225

cancer patient-caregiver dyads were examined regarding their reports on severity and frequency
of patients’ pain and fatigue in the seven days prior to the baseline interview.
The results show that congruence of pain and fatigue severity and frequency between
patients and caregivers was at a fair-to-moderate level (ICC = 0.3 ~ 0.5). The Odds Ratio (OR)
of caregivers’ congruence versus noncongruence of pain severity was decreased by a higher
score of caregiver depressive symptoms, a greater number of patient symptoms, and caregivers
living with patients, while congruence was increased by caregiver self-esteem. The OR of
caregivers’ underestimation versus congruence of pain frequency was decreased by 46% when
one point of schedule interruption increased (95% CI of exp [-0.62] = 0.4 - 0.83). The OR of
caregivers’ underestimation versus congruence of fatigue severity was increased 10% by
caregivers perceiving one more patient symptom (95% CI of exp [-1.02] = 0.81 - 1.01). The OR
of caregivers’ underestimation versus congruence of fatigue frequency was decreased by 23 %
when caregivers recognized one more patient symptom (95% CI of exp [-0.14] = 0.77 - 0.97).
Developing the EASP framework can enrich knowledge of family caregivers’ empathy related to
symptom management. Clinical care providers need to acknowledge that caregivers’ awareness
of patient symptoms and their congruence of symptom perception may be a sign of caregiver
burden and caregiver depressive symptoms; therefore, caregiver burden and depression may be
detected early. Nursing research is encouraged to integrate the EASP framework that can
develop the approaches for caregivers’ empathic accuracy and also to strengthen the
generalization of this framework. Health policies are needed to advocate for family caregivers as
partners in the health care system to ensure that care resources are available and the coordination
for patients and their families occurs.

DEDICATION
This work is dedicated to my parents who were family caregivers for my grandfather who
passed away because of non-small cell lung cancer. It is also dedicated to all the cancer patients
and their family members whom I took care of and communicated with. Their experiences have
inspired me to enter into the research area of family cancer care.

iv

ACKNOWLEDGEMENTS
I would like to sincerely thank the following individuals who have supported me in my
studies in the nursing doctoral program at Michigan State University: Dr. Linda Spence, PhD,
RN; Dr. Denise Saint Arnault, PhD, RN; and Dr. Barbara Given, PhD, RN, FAAN. Their
guidance, inspiration, mentorship, expertise and patience as Major Professors has been essential
for the completion of this study. I would also like to thank Dr. Charles W. Given, PhD, and Dr.
Paula Sherwood, PhD, RN, for their support, inspiration, and expertise as my guidance
committee; and thank you to Dr. Alexander von Eye, PhD, and Mei You, MS, for their statistical
expertise. My sincere thanks goes to Dr. Gwen Wyatt, PhD, RN, for her mentorship and
encouragement when I worked for her as an assistant for nursing research consultation in
Sparrow Hospital at Lansing, Mich.; and thank you to Dr. Yeur Hur Lai, PhD, RN, in Taiwan for
her inspiration and support to my nursing research in cancer care. I would like to thank the many
faculty members in the Michigan State University College of Nursing who have helped me in my
studies.
Most importantly, immeasurable thanks go to my parents, my sister and her family, and
my brother and his family for their long-term and long-distance support and encouragement. A
special thank you to Sandra Lee Spoelstra, PhD, RN; Karen Burritt, PhD, RN; Lori HoughtonRahrig, PhD, RN; Melodee Vanden Bosch, PhD, RN; Kari Wade, PhD, RN; Elisabeth Byma,
PhD, RN; Denise Soltow, PhD, RN; Arunee Chaiyarit, PhD, RN; Saowaluk Netchang, PhD, RN;
Phanarat Chenchop, PhDc, RN; and Dawn, Frambes, MS, RN, for sharing their doctoral
experience, valuable friendship, and generosity with me and for supporting my studies on my
journey.

v

I would also like to thank my Taiwanese friends at MSU, Chu-yin Yeh, PhD; Chi Chang,
MS; Grace Chen, PhD; Christine Shyu, PhD; I Chun Wu, PhD; I-Chien Chen, MS; Chia Ying
Huang, DMA; Wen ling Chang, MS; and Yoko Yeh, MS, for strengthening my physical health
and for sharing life experiences, support, and friendship with me. A special thank you to longdistance friends in Taiwan: ChienTsun Chen, Hui Chun Lin, Yu Shuang Lin, I Shan Tsai, Li
Hsueh Chen, Shu Chin Fang, Hui Ju Yang, Wen Jing Hung, Yi Yu Chen, Yu Lin Shue, and Kai
Hsien Fang for their long-term support and encouragement. I am thankful for the friendship in
the Taiwanese bible study cell group which helped me to keep faith in God and to learn God’s
love, righteousness, and wisdom in my spiritual life.
I also appreciated the funding support from the Sigma Theta Tau International Honor
Society of Nursing Small Grant (#RC100577), College of Nursing, and Graduate School for my
current research.

vi

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. xi
LIST OF FIGURES ......................................................................................................... xiii
CHAPTER 1
OVERVIEW ........................................................................................................................1
Purpose.....................................................................................................................1
Background ..............................................................................................................1
Cancer caregiver population ........................................................................1
Congruent levels of symptom perception as a measure of empathic
accuracy .......................................................................................................4
Factors related to empathic accuracy (congruence of symptom
perception) ...................................................................................................5
Specific Aims .........................................................................................................10
Hypotheses .................................................................................................11
Significance for Nursing Science...........................................................................11
New knowledge for nursing .......................................................................11
New strategies for nursing research ...........................................................12
Implementation in cancer care ...................................................................14
Limitations of the present study.................................................................15
Purposes of Chapters..............................................................................................16
CHAPTER 2
THEORETICAL FRAMEWORK AND RATIONALE ...................................................17
Philosophical Underpinnings for the Concept of Empathic Accuracy ..................18
Empathic accuracy .....................................................................................20
Empathy model ..........................................................................................23
Middle-Range Nursing Care Model.......................................................................28
Conceptual Framework of EASP ...........................................................................33
Specific aim ...............................................................................................38
Hypothesis..................................................................................................38
CHAPTER 3
LITERATURE REVIEW ..................................................................................................40
Family Caregivers of Cancer Patients....................................................................40
Cancer Care at Home .............................................................................................42
Congruent Levels of Symptom Perception as a Measure of Empathic Accuracy .44
Congruence of pain and fatigue severity and frequency ............................44
Cancer-related symptoms...........................................................................46
Cancer pain ................................................................................................47
Cancer-related fatigue ................................................................................48
vii

Symptom characteristics ............................................................................49
Relationship Among Caregiver Emotionality, Caregiver Threats of Caregiving,
and Congruence of Symptom Perception ..............................................................51
Caregiver emotionality...............................................................................51
Threats of caregiving .................................................................................52
Caregiver Demographics as the Underlying Factors for Caregiver Emotionality
and Threats of Caregiving ......................................................................................55
Sex and age ................................................................................................55
Education ...................................................................................................56
Patient symptoms .......................................................................................56
Caregiver Demographics Related to Congruent Symptom Perception..................57
Caregiver age .............................................................................................57
Caregiver sex .............................................................................................58
Caregiver education ...................................................................................59
Caregivers’ relationship to patients ...........................................................60
Living arrangements ..................................................................................60
Methods to Measure Congruent Levels of Symptom Perception Between
Cancer Patients and Family Caregivers .................................................................63
Subtraction method ....................................................................................63
Statistical value ..........................................................................................65
CHAPTER 4
METHODOLOGY ............................................................................................................67
Study Design ..........................................................................................................68
Parent study design ....................................................................................68
Design of present study..............................................................................69
Sample....................................................................................................................70
Power analysis ...........................................................................................71
Measurements ........................................................................................................74
Demographic characteristics ......................................................................74
Caregiver perception ..................................................................................75
Caregiver depression ..................................................................................75
Caregiver burden ........................................................................................77
Symptom ....................................................................................................79
Congruence ................................................................................................80
Human Subjects Protection ....................................................................................82
Data Management ..................................................................................................84
Merging data files ......................................................................................84
Data recode ................................................................................................84
Data Analysis .........................................................................................................85
Descriptive analysis of sample...................................................................85
ICC .............................................................................................................85
Pearson correlation.....................................................................................87
GLMs .........................................................................................................87
Loglinear regression...................................................................................88

viii

CHAPTER 5
RESULTS
....................................................................................................................92
Sample Size and Power ..........................................................................................92
Descriptive Analysis of Participants’ Demographic Characteristics .....................95
Age .............................................................................................................95
Sex..............................................................................................................95
Education ...................................................................................................95
Race............................................................................................................95
Marital status ..............................................................................................96
Employment ...............................................................................................96
Living arrangement ....................................................................................96
Cancer diagnosis ........................................................................................96
Comorbidity ...............................................................................................96
Hours of providing symptom management................................................97
Caregivers’ emotionality............................................................................97
Caregivers’ threats of caregiving ...............................................................97
Differences in Caregiver Scores by Their Demographics ...................................100
Sex............................................................................................................100
Education .................................................................................................101
Relationship to patient .............................................................................101
Living arrangement ..................................................................................101
Description of Congruent Levels of Pain and Fatigue Severity and Frequency ..102
ICC ...........................................................................................................102
Subtraction ...............................................................................................103
Differences of Caregiver Characteristics by Three Congruent Levels of Pain
and Fatigue Severity and Frequency ....................................................................107
Congruence of pain severity ....................................................................108
Congruence of fatigue severity ................................................................108
Congruence of pain frequency .................................................................109
Congruence of fatigue frequency .............................................................109
Results of Testing Two Research Hypotheses .....................................................114
Pearson’s correlation ................................................................................115
The t Test for Differences of Caregivers’ Characteristic Between Caregiver
Whose CESD ≥16 and CESD<16. .......................................................................116
GLM .........................................................................................................118
Factors of caregiver depressive symptoms ..............................................118
Factors of caregiver burden .....................................................................119
Multinomial Loglinear Regression ......................................................................120
Hypothesis 1.........................................................................................................121
Hypothesis 2.........................................................................................................122
Congruence of pain severity ....................................................................122
Congruence of fatigue severity ................................................................125
Congruence of pain frequency .................................................................126
Congruence of fatigue frequency .............................................................128
Conclusion ...........................................................................................................129
ix

CHAPTER 6
DISCUSSION AND IMPLICATIONS ...........................................................................132
Effect Size ............................................................................................................132
Congruent Levels of Pain and Fatigue Severity and Frequency ..........................133
Hypothesis 1: Caregiver Depression and Caregiver Burden Relevant to
Congruence Model ...............................................................................................134
Hypothesis 2: Caregiver Characteristics Relevant to Congruence ......................139
Living arrangement ..................................................................................140
Total number of patient symptoms ..........................................................141
The Fitness of Theoretical Framework of EASP .................................................143
Limitations ...........................................................................................................145
Summary ..............................................................................................................146
Implications..........................................................................................................147
For knowledge in nursing science............................................................147
For nursing clinical practice.....................................................................148
For future research ...................................................................................149
For health policy ......................................................................................151
Conclusion ...........................................................................................................153
APPENDICES .................................................................................................................155
APPENDIX A: Instruments .................................................................................156
APPENDIX B: Data Analysis Flow Chart ..........................................................167
APPENDIX C: IRB Approval Letters .................................................................169
REFERENCES ................................................................................................................174

x

LIST OF TABLES

Table 1

Instruments Used to Measure Caregiver Depressive Symptoms,
Caregiver Burden and Cancer Symptom Perception of Patient
and Caregivers ..........................................................................................83

Table 2

Variables and Their Measurements Used for Data Analysis .....................86

Table 3

Linear Regression in Testing Hypothesis ..................................................91

Table 4

Cancer Patients’ Demographic Characteristics (N = 225) .........................93

Table 5

Caregivers’ Demographic Characteristics (N = 225).................................94

Table 6

Differences in Caregivers’ Depression, Feelings of Burden, Age,
Hours of Symptom Management, and Total Number of Patient
Symptoms by Sex, Education, Relationship to Patient, and Living
Arrangement (t test) ...................................................................................98

Table 7

Means with SD And ICCs for Examining Congruent Levels of Pain
Severity, Fatigue Severity, Pain Frequency, and Fatigue Frequency
Between Caregivers and Patients .............................................................104

Table 8

The Substation for Examining Congruent Levels of Pain Severity,
Fatigue Severity, Pain Frequency, and Fatigue Frequency Between
Caregivers and Patients (N = 225) ...........................................................105

Table 9

Differences Among Three groups—Underestimation, Congruence,
and Overestimation in Testing Characteristics (binomial variables) of
Caregiver (Chi-Square) ............................................................................110

Table 10

Differences Among Three Groups—Underestimation, Congruence,
and Overestimation in Testing Characteristics (continual variables) of
Caregiver (ANOVA) ................................................................................111

Table 11

Pearson’s Correlation Coefficient for Caregiver Characteristics and
Caregiver Perception of Pain and Fatigue Severity and Frequency ........116

Table 12

Differences of Caregiver Demographics, Caregiver Scores of Patients
Pain and Fatigue Severity and Frequency, Congruence of Pain and
Fatigue Severity and Frequency between CESD ≥ 16 and CESD <16
(t test, N = 225) ........................................................................................117

xi

Table A1

Sociodemographic Information for Cancer Patients ................................156

Table A2

Sociodemographic Information for Family Caregivers ...............................159

Table A3

Patient symptoms ......................................................................................161

Table A4

CES-D Scale ............................................................................................163

Table A5

Caregiver Reaction Assessment Instrument (CRA) ...............................165

xii

LIST OF FIGURES

Figure 1

Theoretical framework of Empathic Accuracy of Symptom
Perception (EASP) .....................................................................................35

Figure 2

Analysis model for predicting factors and empathic accuracy in the
theoretical framework, Empathic Accuracy of Symptom Perception ......90

xiii

CHAPTER 1
Overview
Purpose
The purpose of this research is to examine how caregivers’ emotionality and perceived
threats of caregiving impact empathic accuracy of symptom perception (EASP) between
caregivers and their cancer patients. This research also explores empathic accuracy of patients’
symptom perception between family caregivers and cancer patients, which the primary
investigator operationally defined in this study as congruence of symptom perceptions. This
study begins to create a theoretical framework examining the relationships among caregivers’
emotionality, caregivers’ threats of caregiving, and empathic accuracy while considering the
effect of caregivers’ characteristics.
The purpose of this chapter is to provide a brief overview of the cancer caregiver
population and care demands related to cancer symptoms, of empathic accuracy (congruence) of
symptom perception in the study, and of the factors influencing empathic accuracy (congruence)
of symptom perception between caregivers and their cancer patients. The specific aims and
hypotheses of this study are introduced at the end of this chapter. Likewise, this study illustrates
how the conceptual framework of EASP and these findings may contribute to new nursing
knowledge, to new research perspectives, and to the implementation of the framework of EASP
for cancer family care, all of which demonstrate this study’s significance for nursing science.
Background
Cancer caregiver population. Family caregivers generally are family members who
provide care for patients without medical training or payments (Family Caregiving Alliance,
2011). Family caregivers of patients with cancer provide vital and continued physical and

1

emotional support for patients, helping them to endure the cancer experience at home. According
to the National Alliance for Cancer and the American Association of Retired Persons [NAC and
AARP] (2009), more than 65.7 million informal caregivers take care of their family members,
who have had diagnoses varying across all chronic conditions in the United States. In the general
caregiver population, caregivers are predominantly female (66%) and are 48 years of age on
average. In addition, 7% of adult family caregivers are providing care for their adult family
members with cancer. In cancer families, involvement of family caregivers is essential for
optimal treatment and treatment compliance, continuity of care, and social support (Glajchen,
2004). Most family caregivers of adult cancer patients are spouses, partners, or adult children.
Relatives, close friends, or neighbors may take this role when the primary family caregiver is not
available (ACS, 2010; Mintz, 2007; Paul, Kurtz, & Mentzer, 2007).
For patients with illnesses or disabilities, the majority of care demands are daily life
activities and instrumental activities (56%). Their caregivers spend 20 hours per week on average
providing care, and 13% of caregivers provide 40 hours or more. Seventy-four percent of
caregivers have continued working while managing caregiving responsibilities. On average, the
length of caregiving is 4.3 years (NAC & AARP, 2009). According to the American Cancer
Society [ACS] (2010), cancer families provide about 80% of home-care services. Most
caregivers whose patients were diagnosed with cancer spent more than 15 hours per week
providing care at home in the first year (Van Houtven, Ramsey, Hornbrook, Atienza, & van Ryn,
2010; van Ryn et al., 2011). Kim, Baker, Spillers, and Wellisch (2006) also found that cancer
and dementia family caregivers spend approximately 30 hours per week providing patient care at
home. Performing cancer care at home is time-consuming, and time demands increase along with
patients’ changing dependency needs.

2

Cancer is a life-threatening disease: Malignant neoplasm has been the second leading
cause of death in the United States for five years, according to the Centers for Disease Control
and Prevention mortality report (CDC, 2009), despite the declining mortality in the cancer
population (Jemal, Center, DeSantis, & Ward, 2010). Typically, when the cancer-related
symptoms appear, symptom management becomes a major part of family caregivers’
responsibilities. However, cancer patients and their families are usually not familiar with their
dynamic characteristics during the course of treatment. (ACS, 2010; Fan, Filipczak, & Chow,
2007). Caregivers are required to pay more attention to new care demands related to patients’
symptoms. van Ryn et al. (2011) focused on the caregivers who provided at least 50% of
informal care for patients who were diagnosed with lung and colorectal cancer, and found that
family caregivers provided cancer-specific clinical care, including monitoring side effects of
treatments (68%), managing symptoms (47%), administering medication (34%), making
decisions for contacting doctors (30%) and for providing medicine (29%), and changing wound
dressing (19%). Regarding the reports of van Ryn et al., symptom management is the most
prevalent caregiving activity aside from assisting with patients’ daily activities. Therefore, most
care requirements for caregivers occur during patients’ in the treatment phase. In this
circumstance, caregivers become not only proxies for cancer patients in reporting cancer
patients’ symptoms to formal care providers and in assisting with symptom management at
home, but also partners with professional health care providers.
The experience of cancer can introduce various threats into the family. The threat of
death from cancer is perceived not only by patients, but also by most family caregivers. Family
caregivers may also image threats from providing care because care demands can interrupt
caregivers’ lives, which has been recognized as caregiver burden (Grunfeld et al., 2004).

3

Perceiving patients’ symptom burdens in order to provide symptom management for patients can
impact caregivers’ own lives (Palos et al., 2011). Caregivers have to readjust their lives for those
caregiving situations. In addition, perceiving patients’ symptoms and assisting with symptom
management are acknowledged as ongoing care demands placed on caregivers. Caregivers may
project a threat from providing patient care to their living quality. There may be a relationship
between caregivers’ accurate perceptions of patients’ cancer symptoms and their perceptions of
the threats of caregiving. Developing the conceptual framework of EASP can contribute to
supporting nursing research in symptom management.
Congruent levels of symptom perception as a measure of empathic accuracy. This
study develops a conceptual framework of EASP (see Figure 1) using theoretical support from
three theories: empathic accuracy (Ickes, 1993), the empathy model (M. H. Davis, 1996), and the
family care model (FCM; B. Given, 1995). The concept of empathic accuracy refers to the
“perceivers’ capability to accurately infer the actors’ thoughts and feelings” (Ickes, 1993). The
present research, inspired by Ickes’s definition of empathic accuracy, examines caregivers’
accurate perceptions of patients’ cancer symptoms while caregivers use their cognitive appraisals
to be accurate in perceiving patients’ symptoms. Therefore, in this study, empathic accuracy of
symptom perception is conceptually defined as family caregivers’ and patients’ similarities in the
cognitive appraisal of patients’ cancer symptom experiences. Furthermore, empathic accuracy
has a parallel definition as “interpersonal accuracy.” According to the empathy model (M. H.
Davis, 1996), having “interpersonal accuracy” is one of the results from empathic processes that
can determine perceivers’ behavior toward other persons. Having interpersonal accuracy is
referred to as the largely cognitive outcomes through processing cognitive and affective
empathy.

4

Ickes (1993) suggested that a straightforward way to measure empathic accuracy is to
compare the actor’s actual experience with that of the perceiver’s corresponding inferred
perceptions of him or her , EASP is operationally defined as the “congruence of symptom
perception,” which measures the caregiving situation where family caregivers and patients have
similar perceptions of patients’ symptoms, specifically focusing on severity and frequency of
pain and fatigue.
Several studies measuring congruence between caregivers and care receivers have found
that caregivers’ reports of cancer-related symptoms were low in congruence (Kurtz, Kurtz,
Given, & Given, 1996, 2006; C. Lin, 2001; McPherson, Wilson, Lobchuk, & Brajtman, 2008; Tu
& Chiou, 2007; Wennman-Larsen, Tishelman, Wengstrom, & Gustavsson, 2007). In a few
studies on congruence of symptom perception, researchers have reported the specific caregiver
characteristics– such as caregiver burden, male caregivers, and older caregiver–which are
correlated with low congruence (Broberger, Tishelman, & von Essen, 2005; McPherson, et al.,
2008; Wennman-Larsen, et al., 2007).
Factors related to empathic accuracy (congruence of symptom perception).
Theoretically, a perceiver will be less able to empathize with the target person and also feel less
empathic accuracy if the individual perceives a threat in his or her situation, such that perceiving
threats may decrease empathic accuracy (Ickes, 1993; M. H. Davis, 1996). For family care
situations, the threats of caregiving in the present study are conceptually defined as caregivers
perceiving threats from providing care. Indeed, some studies have reported that caregivers with
feelings of burden have lower likelihood of congruent perception of patients’ symptoms
(McPherson, et al., 2008; Wennman-Larsen, et al., 2007). Because caregiving is demanding and
impacts caregivers’ lives, some caregivers may perceive these caregiving circumstances as

5

threats. Glajchen (2004) emphasized that cancer caregivers often operate within caregiving
situations and without preparation. Given, Sherwood, and Given (2011) reinforced the necessity
of caregiver involvement in patient care after patients received cancer treatment because
symptom management can be a core responsibility of care through treatment and into to posttreatment. Caregivers may feel burdened when their lives are continually interrupted because
they are providing care to meet patients’ demands. In that manner, a caregiver may perceive this
burdensome caregiving situation as a threat. In the present study, the threats from providing care
that caregivers perceived are represented by caregiver burden.
Caregiver emotionality. Empathy may be influenced by emotionality and perceivers’
negative emotionality, such as sadness and anger, which can lessen their abilities for cognitive
empathy (M. H. Davis, 2006; Eisenberg & Eggum, 2009; Eisenberg, Eggum, & Di Giunta,
2010). In the current study, caregiver emotionality is defined as caregivers’ mood expressions,
such as anxiety and depression. Depression is frequently recognized in the family caregiver
population as an indicator of caregivers’ poor mental health resulting from long-term, stressful
caregiving circumstances in which caregivers have not developed sufficient knowledge or
strategies to cope with dynamic care demands (National Family Caregivers Association, 2010;
Pellegrino et al., 2010; Rivera, 2009). Therefore, in the current study, emotionality is referred to
as caregiver depression. Primary investigator hypothesizes that increasing caregivers’ depression
levels may decrease their ability to accurately perceive patients’ symptoms.
Furthermore, feelings of burden that are related to providing care may generate negative
emotions, such as depression, in caregivers. Studies found that a high correlation ( r= 0.3- 0.6)
between caregivers’ reports of caregiving burden and depression (Clark & Diamond, 2010;
Grunfeld, et al., 2004; Y. Kim, Duberstein, Sorensen, & Larson, 2005; Kurtz, Kurtz, Given, &

6

Given, 2004). In this regard, caregivers’ emotionality and threats of caregiving may be
covariants and may have an effect on caregivers’ EASP. Therefore, it raises the critical question
of whether caregivers’ emotionality (depression) can predict low empathic accuracy (congruence
of symptom perception) for nursing science. In the meantime, this study presumes that the
congruence of symptom perception is an important factor in determining how caregivers provide
appropriate symptom management. Exploring how caregivers’ emotionality (depression) and
their perception of threats (caregiver burden) are related to caregivers’ empathic accuracy
(congruence of symptom perception) may help to understand the phenomenon of caregivers’
EASP in nursing science and may also help nursing practice in assisting caregivers to perform
symptom management for patients. The issue has not been tested in previous research, but this
study will fill that gap.
Caregiver demographics. Caregivers’ demographics are related to caregiver depression
and caregiver burden (Family Caregiving Alliance, 2009). Caregivers’ demographics may
indirectly affect EASP. Previous studies have shown that caregivers who are older and female
tend to have perceptions about symptoms that are congruent with patients’ perceptions
(Broberger, et al., 2005; Wennman-Larsen, et al., 2007). However, the effects of these
characteristics on congruent levels of symptom perception were inconsistent between studies.
Caregivers with low education have correlated with noncongruence, which may be because
education is related to the ability of understanding and receiving information (C. Lin, 2001;
Yeşilbalkan & Okgün, 2010). While laypersons might suspect that spousal caregivers would be
most likely to have congruence when compared with nonspouses, this has not been supported by
research findings (Broberger, et al., 2005; McPherson, et al., 2008; Wennman-Larsen, et al.,
2007). Therefore, this research will examine the impact of caregivers’ age, sex, and relationship

7

to patients on empathic accuracy (congruence of symptom perception).
Cancer-related symptoms. The current study uses pain and fatigue to represent cancerrelated symptoms. Although different cancer diagnoses may generate different symptoms, pain
and fatigue are the most common symptoms of cancer and cancer treatments (National Institutes
of Health State-of-the-Science, 2004). Cancer patients commonly report pain as the most
unpleasant symptom, and unrelieved pain interrupts a patient’s quality of life. The
multidimensional nature of pain can make effective pain management more difficult (DelgadoGuay & Bruera, 2008; McGuire, 1995; van den Beuken-van Everdingen et al., 2007).
Furthermore, fatigue is the most prevalent symptom in patients’ daily lives and is related to
receiving aggressive treatment and/or cancer malignancy (Bower, 2007; Diaz et al., 2008; Ryan
et al., 2007). In the studies that identified symptom clusters, pain and fatigue commonly shared
the same related symptoms, such as depression and sleeping problems (Donovan & Jacobsen,
2007; Fan, Hadi, & Chow, 2007; B. A. Given, Given, Sikorskii, & Hadar, 2007; E. Kim et al.,
2009), and these related symptoms are commonly identified as the ones that most influence
patients’ physical functioning (B. Given, Given, Azzouz, & Stommel, 2001; B. A. Given, et al.,
2007; Miaskowski, 2010). Therefore, pain and fatigue should be a priority to be measured for
empathic accuracy of symptom perception.
In addition, the present study examines symptom severity and frequency. Symptom
experiences are multidimensional and consist of the following characteristics: distress, which is
highly correlated to intensity (severity); occurrence, which includes duration, prevalence, and
frequency; and quality, which includes interference (Armstrong, 2003; Dodd et al., 2001).
Severity and distress have been commonly measured to represent symptom burden, because the
levels of severity and distress have the most interference with a patient’s quality of life; but

8

symptom severity is suggested to positively determine the patient’s distressful perception of
symptoms (H. J. Kim & Abraham, 2008; Kirkova et al., 2010). Some studies have also reported
that an increase in frequency and severity of symptoms can impact a patient’s performance of
daily physical functioning and worsen a patient’s quality of life (Lutz et al., 2001; Mercadante,
Casuccio, & Fulfaro, 2000). Meanwhile, severity and frequency are convenient measurable
symptom characteristics for caregivers to use at home when assessing patients’ symptom burden.
In the present study, EASP will be measured by comparing cancer patients’ and caregivers’
perception of pain and fatigue severity and frequency.
Furthermore, congruence of symptom perception has been suggested as a guide for
caregivers in providing appropriate symptom management. Unrelieved pain and fatigue can limit
patients’ physical functioning and lower patients’ quality of life (B. A. Given, et al., 2007;
Miaskowski, 2010), which may indicate that EASP is important for patients’ quality of life.
Silveira, C. Given, B. A. Given, Rosland, and Piette (2010) have published the only study
examining the relationship between congruence of symptom severity and frequency and health
outcomes of patients and caregivers. They reported that patients whose caregivers’ symptom
perceptions were congruent with the patients’ self-reports were more likely to have better quality
of life. Their findings may imply that helping caregivers to achieve congruent perceptions of
patients’ symptoms may theoretically improve patients’ quality of life. The gaps influencing the
development of empathic accuracy in the family care environment are (a) prior studies have not
oriented by theoretical perspectives, (b) prior studies have also not examined any of the
relationships among EASP, appropriate symptom management, and the impact patients’ quality
of life; (c) EASP in family care has been neither described in the caregiving theories or
supported by nursing theoretical perspectives; and (d) it has not been tested empirically.

9

Therefore, a long-term goal of this research is to build on the EASP conceptual framework to
design studies that can enrich the understanding of empathic accuracy in cancer family care.
This present study, however, will focus only on exploring EASP (congruence of pain and
fatigue severity and frequency), and how caregiver emotionality and caregiver perceptions of
threats of caregiving predict EASP, as depicted in the conceptual framework of EASP.
Specific Aims
The purpose of this research is to examine how caregivers’ emotionality and perceived
threats of caregiving impact EASP between caregivers and their cancer patients. This research
will also explore empathic accuracy of patients’ cancer symptom perceptions. As stated above,
caregivers’ empathic accuracy is operationally defined as congruence of symptom perception,
severity, and frequency of pain and fatigue between cancer patients and family caregivers. The
congruent levels of pain and fatigue severity and frequency will be measured in this study by
examining the value of the difference between patients’ scores and caregivers’ scores of patients’
pain and fatigue severity and frequency. If the score difference is close to zero, indicating similar
scores between patients and caregivers, then congruence of pain and fatigue and frequency
between patients and caregivers is high, and vice versa. Noncongruence is also categorized as
overestimation when a caregiver’s score is higher than a patient’s and underestimation when a
caregiver’s score is lower than a patient’s. Emotionality is operationally defined as caregivers’
depressive symptoms, and perception of threats related to caregiving is represented by caregiver
burden.
This study used a secondary data analysis for 225 pairs of cancer patients and family
caregivers, adopted from the parent study entitled Family Home Care for Cancer—A
Community-Based Model (# R01Ca-79280). The specific aim of this research is to examine how

10

caregivers’ emotionality (caregiver depression) and caregivers’ threats of caregiving (caregiver
burden) predict EASP (congruence of pain and fatigue severity and frequency), when taking into
account contextual factors of caregivers–including caregivers’ age, sex, education, relationship
to patients, and total number of patients’ symptoms that caregivers observe, as well as living
arrangement–as they can influence caregivers’ emotionality and threats resulting from
caregiving.
Hypotheses. The two hypotheses for this study are:
1. Increased caregiver depression and caregiver burden will be related to lower congruent
values of pain and fatigue severity and frequency.
2. Caregivers who are female, older, patients’ spouses, better educated, perceiving a
lower number of patients’ symptoms, and living with patients will have greater congruence of
pain and fatigue severity and frequency than will caregivers, who are male, younger, nonspouses
of patients, less educated, perceiving a higher number of symptoms, and not living with patients.
Significance for Nursing Science
This conceptual framework of EASP has the potential to make a significant contribution
to nursing science in the areas of theoretical nursing knowledge, new strategies for research, and
implementation of cancer care. The framework of EASP depicts that both emotionality and
threats of caregiving can influence caregivers’ perceptions of patients’ symptom experiences,
thereby affecting the congruent levels of symptom perception between patients and caregivers.
New knowledge for nursing. The EASP framework has been developed by applying the
theoretical perspectives of Ickes’s empathic accuracy in psychological therapy, Davis’s empathy
model in psychology, and Given’s family care model in nursing science. Most prior studies,
aimed at identifying discrepant symptom perceptions between cancer patients and caregivers,

11

have not oriented themselves with theories or models, but mainly have addressed the
unreliability of caregivers’ perception of patients’ symptoms. This focus kept most of these
studies at the level of description for caregivers’ accurate perception of patients’ symptoms.
Therefore, the present study will extend the science by synthesizing empathy theories in
caregiving research.
The EASP framework has been established to understand the caregiving situation in the
family. M. A. Davis (1994) described empathy as people’s general temperament, and Ickes
(1993) described empathic accuracy as the nature of everyday mind reading, so that being
empathic naturally happens in people’s daily lives. The perspective of this research is that
caregivers’ empathy may lead caregivers to be more likely to accurately perceive patients’
symptoms because, today, most cancer care has shifted to home and provided by family
caregivers; caregivers’ empathy may encourage them to make an effort to understand patients’
symptom burdens. Patients’ symptom burden often refers to their physical, psychological, social,
and spiritual problems related to cancer and anticancer treatments. Testing this set of
relationships is the focus of this study, and sets the stage for future work aimed at examining
empathic accuracy of patients’ symptom experience.
Although the perspective of empathic accuracy came from Ickes, Ickes’s own study
applied this conceptualization of empathic accuracy in order to differentiate various statuses of
relationships–such as dating relationship, spousal relationship, and friendship–between the
perceiver and the target person. Few studies have used the concept of empathic accuracy in the
symptom perception of caregiving dyads. More research in the area of empathic accuracy within
the relationship of caregiving dyads will extend the science of empathy.
New strategies for nursing research. This study has developed the framework of EASP

12

to conceptualize empathic accuracy that includes the influence of the factors of emotionality and
perceived threats of caregiving. At least 30 previous studies have shown congruence or
noncongruence of cancer patients’ symptoms by comparing patients’ self-reports and caregivers’
report about patients’ symptoms; however, congruence of symptom perception has not been
conceptually defined. The framework of EASP allows researchers to develop hypotheses that can
account for variations in EASP between cancer patients and their caregivers. The present study
also defines congruent levels of symptom perception as a measure of EASP in a caregiving
situation, in which family caregivers and patients have similar perceptions of patients’
symptoms, specifically focusing on pain and fatigue severity and frequency.
Previous studies also commonly have identified low congruent levels of symptom
perception and often have concluded that the reliability of caregivers’ reports of patients’
symptoms may be questionable. Very few of the previous studies that tested family caregivers’
accurate reports of patients’ symptoms have extended their investigations to identify how
caregivers’ characteristics, such as emotionality and perceived threats of caregiving, may pose
risks to empathic accuracy. The present study will test the predictive relationship between
congruence of symptom perception, caregivers’ emotionality, and threats of caregiving. It also
will consider the influence of caregiver demographics and has the potential to generate
knowledge of empathy accuracy in family care.
Furthermore, the EASP framework shows that there is a logical sequence between
empathic accuracy and patients’ health outcomes. These sequential relationships between
empathic accuracy and outcomes of patients and caregivers have been tested in the subsequent
studies. Nursing research can begin to establish research evidence for the relationship between
empathic accuracy and caregivers’ helping behaviors. Establishing the evidence for empathic

13

accuracy of symptom perception can enrich family cancer care research.
Implementation in cancer care. The EASP framework takes into account the
contributions of the affective components of the empathic process to the cognitive outcome of
empathic accuracy. This conceptualization acknowledges the fact that caregivers use both
cognitive and affective empathic processes to provide daily care for patients, including symptom
management and assistance with daily activities. Examining the hypotheses in the present study
may support existing research that has emphasized the negative impact of caregiver burden and
caregiver depression on cancer caregiving. Several studies have investigated the different
strategies to reduce the risk of caregiver burden and/or caregiver depression for caregivers
(Honea et al., 2008; Rivera, 2009). However, the present study presumes that both caregiver
burden and caregiver depression not only may represent caregivers’ health outcomes, but also are
important factors that influence caregivers’ accurate perceptions of cancer patients’ symptom
experiences.
The EASP framework shows that accurately perceiving patients’ symptoms is an
essential caregiver ability that is presumed to lead caregivers to provide appropriate symptom
management. The EASP framework may help nurses to recognize family caregivers’
characteristics, such as caregivers with depression and burden, and to take into account caregiver
demographics in order to deliver effective nursing interventions for patients and caregiver in
expressing symptoms. Caregivers with greater empathic accuracy of patients’ symptom
experience may help patients alleviate symptom burdens by determining appropriate symptom
management. Additionally, empathic accuracy may help caregivers to feel more confident in
symptom management, thereby decreasing their burdened feelings from providing care. In
conclusion, improving empathic accuracy may benefit cancer patients and their family

14

caregivers.
Limitations of the Present Study. Several limitations exist in the present study. First,
the study is a cross-sectional study using baseline data. Caregiver emotionality, threats of
caregiving, and empathic accuracy may change because of the dynamic nature of symptoms and
caregivers’ perception of caregiving. A longitudinal study should be conducted to understand the
dynamic relationship among those three concepts. Second, this study focuses on patients who
have been newly diagnosed with cancer and who are receiving chemotherapy for the first time.
The results may not be generalized to other cancer patients who are not newly diagnosed, are not
the first-time recipients of chemotherapy, or have received anticancer treatments different from
chemotherapy. Third, this study focuses on identifying how caregivers experience depression
and how caregiving burden can impact their ability of empathic accuracy of patients’ symptoms.
Further research will test the effect of EASP on caregivers’ helping behavior and both patients’
and caregivers’ outcomes. Fourth, this study only selects negative caregiving characteristics
(caregiver depression and caregiver burden) to predict low empathic accuracy. There is a lack of
knowledge on whether or not caregivers who have positive perceptions about providing care can
enhance EASP. Fifth, this present study does not test the entirety of the empathic process as
defined by Davis and others in the framework of EASP. Davis’s empathy model has
demonstrated that both cognitive and affective empathy occur simultaneously, which may imply
that it is difficult to determine which empathic outcomes were produced from one process but
not from the other. For example, Lobchuk et al (2002) identified five cues (patient-oriented,
caregiver-oriented, generalized other, feeling distanced, and other) for different types of
perceptive processes that may link with congruence of symptom perception. However, when the
studies used those five cues in developing interventions to guide caregivers’ perspective-taking

15

process, congruence did not improve (Lobchuk & Degner, 2002; Lobchuk, Degner, Chateau, &
Hewitt, 2006). Further study about caregivers’ empathic accuracy is needed to examine empathic
processes within verbal and nonverbal communication, and consequences and influencing factors
of empathic processes in order to develop effective interventions for specific caregiver
populations to improve EASP.
Purposes of Chapters
Chapter 2 establishes the model of EASP using theoretical literature in different fields to
identify and explore the factors impacting caregivers’ EASP (congruence) within a caregiving
situation. Chapter 3 reviews the relationships among the concepts within the EASP model by
examining the state of science, providing the background for the conceptual framework of the
study, including the factors influencing congruent levels of symptom perception and the
expected outcomes generated by the concept. The methodology in chapter 4 describes the design
of the present study using secondary data and includes a brief introduction of the parent study.
Chapter 5 shows the result of analyses regarding the two hypotheses. The discussion of
congruence and the discrepancy between the results of the present study and previous studies
will be illustrated in Chapter 6.

16

CHAPTER 2
Theoretical Framework and Rationale
This chapter will establish a conceptual framework for the phenomenon that occurs when
family caregivers accurately perceive cancer patients’ symptom experiences. The phenomenon
describing “the observer’s accurate perception toward the other person” has been conceptually
illustrated by Ickes (1993) as the concept of empathic accuracy (p.558). Accurate perception has
been termed as congruence in the prior literature. Because congruence is understood as the
quality or levels of agreeing or corresponding, congruent viewpoint occurs when two or more
people or groups achieve the same opinion or perspective through understanding, negotiating, or
communicating (M. A. Davis, 2009; Lobchuk, 2006).
Perception has been described as a capability of understanding by receiving stimulation
from the environment and also has been defined as an observation of receiving impressions of
the sense or of the mind; synonyms include awareness, sense, and recognition (Bunting, 1988).
Armstrong (2003) has described symptom perception as the ability to understand symptom
experiences that are the perception of the frequency, intensity (severity), distress, and meaning
occurring when symptoms are produced and expressed.
Symptom perception is defined in this research as the cognitive activity of recognition of
the physical and psycho-social-spiritual state of another person’s symptom experiences,
including symptom severity, symptom frequency, duration of symptoms, and personal belief
system. In this research, the cognitive ability of symptom perception includes the dimensions of
severity and frequency. In sum, a caregiver perceiving a patient’s symptoms is meant as the
caregiver’s ability to recognize the patient’s symptom experiences.
When patient care is provided at home, a family caregiver may live in a caregiving
17

circumstance that is distinctively different from a noncaregivers’ living situation. EASP between
patients and caregivers has been presumed to be influenced by their characteristics. Given’s
middle-range model, the Family Care Model (FCM; B. Given, 1995), has been applied to
discover the patients’ and caregivers’ factors of EASP in the home-based care. The FCM
describes the caregiving circumstances within the nature of the care demands for patients, such
as symptom management and assistance for daily life activities, and how these demands can
influence patients’ quality of life. The FCM was applied to support the importance of empathic
accuracy (congruent symptom perception) in caregiving situations.
Therefore, theoretical concepts were examined and derived from three theories, including
Ickes’s (1997) empathic accuracy, Davis’s (1994) empathy model, and B. Given’s (1995) family
care model. A conceptual framework was also established for family caregivers’ perceptions of
patients’ symptoms. The following section will use theoretical literature in psychology and
nursing to define the concept of empathic accuracy and explore the factors that support or inhibit
empathic accuracy.
Philosophical Underpinnings for the Concept of Empathic Accuracy
Perceiving the other’s experience has been illustrated and argued to be sympathy or
empathy. Sympathy is the shared feeling that results when a person observes another person in an
emotional state, such that sympathy has a greater component of passive and emotional reactions.
Empathy is an attempt to understand others’ feelings rather than a sharing of them, such that
empathy can be accomplished by viewing and interpreting other people’s actions, movements,
and physical cues. Therefore, empathy is the general attempt to get inside the other in an
intellectual capacity (M. H. Davis, 1994; Stueber, 2008). Philosophically, E. Husserl described
the phenomenon of empathy as putting one’s self into another one’s shoes, whereby people make

18

conscious attributions of intentional acts to other subjects. Husserl further has explained that this
process involves a perceiver who observes the target person’s situation and projects his or her
image of himself or herself in the same situation in order to achieve an understanding of the
target person’s feelings. Philosophically, Husserl has stated this empathic experience is an
intersubjective one (Beyer, 2011). In sum, a caregivers perceiving a patient’s symptom
experience is an expression of empathy.
In psychology, perceiving another person’s feelings or thoughts has been referred to as
empathy (M. H. Davis, 1996; Ickes, 1993). Although sympathy may also be inferred to perceive
others feelings or thoughts, it simply shows sorrow for another person’s distress, whereas
empathy is to understand within cognitive perspectives (Stueber, 2008). Decety and Jackson
(2004) also have defined empathy as “a sense of similarity in feelings experienced by the self
and the other, without confusion between the two individuals” (p. 71). Decety and Ickes (2009)
have stated that empathy consists of affective and cognitive constructs. In addition, empathy
includes fully communicating that experience back to the other in an objectively explicit manner
(M. A. Davis, 2009; Giacalone et al., 2008).
In the early studies of empathy research, the cognitive activity known as role-taking and
its products were examined. Expanding on the cognitive definition of empathy, Hogan (1969, p.
308) has defined empathy as “the act of constructing for oneself another person's mental state.”
(as cited in M. H. Davis, 1996, p. 17). Hogan’s definition restricts empathy to the cognitive
realm and places the focus on both the outcome and the process. This is the line of research from
which this present conceptual model will be developed. Below, the cognitive ability of empathy
and the related concept of empathic accuracy will be described and synthesized into this present
research. While this research is about caregivers’ perceptions of cancer symptom experiences, I

19

will begin my conceptualization of empathic accuracy with definitions proposed by Ickes and
Davis. Then, I will reposition the definitions to focus on perceptions of cancer experiences.
Empathic accuracy. Empathic accuracy has been defined as “a perceiver or an observer
is able to accurately infer the specific content of another person’s thoughts and feelings” (Ickes,
1993, p. 388). Empathic accuracy has been used in the study of interpersonal perception in
European psychology and philosophy in the late 1940s (Ickes, 1997). M. H. Davis (2006) has
described interpersonal accuracy as the successful estimation of other people’s thoughts,
feelings, and characteristics that largely results from cognitive empathic processes. Ickes has
explained that the ability to have empathic accuracy reflects a person’s capability of complex
psychological inferences consisting of observation, memory, knowledge, and reason, which
combine to yield insights into the thoughts and feelings to others (Ickes, 1997). From Ickes’s
perspective, empathic accuracy is a product of cognitive empathic processes. Ickes et al.’s
research on empathic accuracy has assumed that every perceiver has a distinct ability for
empathy, so that the determination of the empathic accuracy depends only on the different
characteristics of perceivers. Empathetically accurate perceivers are those who are consistently
good at “reading” other people’s thoughts and feelings, while empathically inaccurate perceivers
are those who are consistently poor at reading other people’s thoughts and feelings.
Furthermore, Ickes’s perspective of empathic accuracy has focused on the accuracy of
dyadic interpretation that the observer can recognize and perceive the target person’s thoughts
and feelings (Ickes, 1997). Ickes (1993) has reviewed previous studies and found that empathic
accuracy can be seen as including four dimensions: (a) perceivers’ accuracy in judging others’
personality traits by relying on inter-rater consensus; (b) dyad members’ accurate perception or
understanding of each other’s attitudes, and values that compare partner perspectives within the

20

dyad; (c) perceivers’ accuracy or affective sensitivity in inferring emotional states of the target
persons; and (d) perceivers’ empathetic accuracy related to their abilities to infer the specific
content of another person’s thoughts and feelings. These empathic accuracy dimensions are
cognitive abilities. The present research uses this perspective of empathic accuracy to examine
family caregivers’ perception of patients’ symptoms.
Ickes’s research has measured empathic accuracy directly without presumption about the
processes that underlie a perceiver’s empathy (Ickes, 2009), which may imply that each
perceiver is assumed to have empathy for the target person. Ickes and his colleagues’ previous
studies found that relationship of the dyad is a major factor that can influence the levels of
accuracy. A close relationship between the observer and the target person also has been
considered a factor that influences empathic accuracy. The observers who have a close
relationship with the target person—for example, a dyad of a partner and close friend—are more
likely to accurately recognize the changes of the target person’s behavior than the stranger. The
dyads with close relationships may have better communication, which improves the perceiver’s
ability of accurate recognition (Ickes, 2007). That may suggest that caregivers’ relationships to
patients, such as spouse and nonspouse, may influence the level of accuracy. It may also suggest
that caregivers who live with patients may have better communication, so that caregivers are
more likely to have empathic accuracy. Therefore, relationship to patient will be considered as a
personal factor influencing empathic accuracy in the present study.
Likewise, individuals’ emotional statuses, such as depression and anxiety, may partially
block their affective sensitivity toward the others’ feelings (Decety & Jackson, 2004; Decety &
Meyer, 2008; Gadassi, Mor, & Rafaeli, 2011). For example, a study examining spouses
experiencing moderate levels of depression and perceiving their partners’ emotion has found that

21

spouses who were depressed were less likely to accurately perceive their partners’ emotions
(Tranmer et al., 2003). In that case, the perceiver’s emotionality can influence that individual’s
own affective ability and can also impact cognitive ability to interpret the target person’s
experience. Based on these results, caregivers’ relationship to patients and emotionality can
influence caregivers’ empathic accuracy. The present study will use the findings of the studies
based on Ickes’s empathic accuracy model to support the theoretical hypothesis that caregivers’
depression can be a negative factor influencing them to be unable to accurately perceive patients’
symptoms.
Ickes (1997) has argued that another central impediment to empathic accuracy is when
perceivers feel highly threatened. Perceiving threats is specifically related to concerns by the
perceiver about the consequences that would likely result from accurately inferring the partner’s
thoughts and feelings. Ickes’s previous studies have shown that when a dyad is in highly
threatening circumstances, the members tend to have low empathic accuracy, or they become
less congruent (Ickes, 1997). Although Ickes and Simpson’s studies have focused on
relationship-threatening situations that lead to inaccurate empathy (Simpson, Orina, & Ickes,
2003), Preston and de Waal (2002) have addressed the issue that perceiving threats in any
circumstances can make empathy unwieldy. Therefore, empathic accuracy may decrease as a
result of perceived threats in certain situations. For example, a caregiver may feel threatened
when a loved one receives a cancer diagnosis, especially because cancer is the second-leading
cause of death in United States. Family caregivers may also feel threatened if they have very
high stress loads in dealing with complex cancer situations, or if they do not have adequate
support for their caregiving activities. The issue of threat in the conceptual framework of the
EASP (Figure 1) focuses on caregivers’ threats from their stressful caregiving lives. The

22

theoretical hypothesis that threat of caregiving can decrease empathic accuracy will be examined
in this research.
Ickes’s (1993) review also has shown that the empathy process should be understood as
having three components: empathic understanding (accurately inferring the thoughts and feelings
of another person), empathic expression (accurately expressing inferred thoughts and feelings
that match the actual experience of the other person), and empathic communication (the dialogic
and dialectical aspect of the empathy process, in which communicative relationships can help in
developing empathic understanding). Ickes has presumed that the relationship among these three
components is circular, and empathic accuracy is a theoretical term close to empathic
understanding and is beneath this circular empathic process. However, in this research, only the
cognitive aspect of empathic understanding will be examined. M. H. Davis (1994) has expanded
on these processes and also has developed the empathy model, which links personal
characteristics, situations, processes, and outcomes to empathic accuracy. In Davis’s perspective
(1994 & 2006), empathic accuracy is a cognitive outcome of complex empathic processes.
Empathy model. The empathy model (M. H. Davis, 1994, 2006) has described empathy
as a process occurring when a person perceives other people, which generates the perceiver’s
behaviors toward the target person. There are four constructs in the model: antecedence, process,
intrapersonal outcomes, and interpersonal outcomes. When a perceiver observes the target
person in a certain situation, the perceiver’s cognitive and affective empathic processes are
activated to generate two levels of outcomes in which the proximal outcomes (intrapersonal
outcomes) are cognitive and affective reactions, and the distant outcomes (interpersonal
outcomes) are advanced performances of social-relationship behaviors. The present study
understands empathic accuracy of patients’ symptom perceptions to be a cognitive intrapersonal

23

outcome. In the following section, I will examine the construct of intrapersonal outcomes,
specifically focusing on interpersonal accuracy, because interpersonal accuracy has been used by
Davis as a parallel term for empathic accuracy.
Davis (1994) has defined interpersonal accuracy as the successful estimation of other
people’s thoughts, feeling, and characteristics, and has described it as one of the cognitive
intrapersonal reactions that largely result from the perceiver’s role-taking processes. Other
cognitive responses include attributional judgment and cognitive representations, which are
linked with empathy-related processes of observing the actors’ behavior and perspective-taking,
respectively. In particular, the perspective-taking process is the observer’s cognitive
representation of the target person’s self (M. H. Davis, 1994). Later, Davis used empathic
accuracy as a parallel to his concept of interpersonal accuracy and explained that empathic
accuracy is a consequence of an observer using perspective-taking in order to have accurate
insights into the target person’s thoughts and feelings (M. H. Davis, 2006). Both role-taking and
perspective-taking are the attempts of one individual to understand another by imagining the
other’s perspectives (M. H. Davis, 1994).
In other words, empathic accuracy is interpersonal accuracy, in which the observer
cognitively takes the target person’s perspective. According to M. H. Davis’s perspective of
empathic accuracy to caregiving situations, caregivers may use the intrapersonal ability of
cognitively taking the cancer patients’ perspective about their symptoms, so that they can
accurately estimate patients’ symptoms. Therefore, empathic accuracy of symptom perception is
related to both sides of patients’ and caregivers’ symptom perception. The communication
between caregivers and patients is important in order to understand or accurately perceive
patients’ symptoms. The importance of communication has been reflected in Davis’s model,

24

which addresses the language-mediated association in the cognitive empathic process. Therefore,
in the framework of EASP, the communication or negotiation between caregivers and patients
will be considered.
In addition, intrapersonal outcomes of empathic processes in Davis’s model include
affective outcomes (Davis, 2006). Affective outcomes are perceivers’ emotional expressions in
responding to their experiences of the target persons. The most common example is an infant’s
mimicry (M. H. Davis, 1994; 2006). Feelings of compassion and feelings of discomfort for the
target person’s experience has been referred to as emotional concern and personal distress, both
of which may be associated with sympathy (M. H. Davis, 1994). Affective outcomes in
caregiving situations may refer to those caregivers who have depressive symptoms or negative
emotions after providing care.
The construct of antecedents of empathic processes in Davis’s empathy model consists of
personal factors and situational factors (M. H. Davis, 1994; 2006). Perceivers’ personal
characteristics are related to their emotionality, whereby perceivers tend to become upset or
distressed in the face of threatening or annoying situations. The strength of these situations can
evoke an emotional response from the observer (M. H. Davis, 2006). Davis (2006) also has
stressed that the explanation of empathy within a specific situation must examine features of
caregiving environments and of the individuals involved. Affect intensity, a dimension of
individual differences, characterizes perceivers’ tendencies to experience strong emotional
reactions. High affect-intensity perceivers respond with more anger and become more depressed
by frustrating stimuli, but also can respond more joyfully to pleasant situations (M. H. Davis,
1996). Given that this emotionality may be important in the caregiving circumstance, caregivers’
depression may be construed as a factor that may influence affect intensity. This present research

25

will examine the personal factors of emotionality (measured as caregiver depression) in the
empathy process.
The construct of empathic process is important in determining perceivers’ reactions when
they observe target persons’ performances. The empathic process that affects the perceiver’s
viewpoint in Davis’s model (1996) includes noncognitive process (e.g., mimicry) and cognitive
process (verbal and nonverbal communication). Applying this perspective of empathy processes
to caregivers’ perceptions of patients’ symptoms, the present study will examine EASP that
infers an intrapersonal outcome from complex empathic processes between the caregiving and
care-receiving dyad. Although Davis (2006) has pointed out that most cognitive empathic
outcomes result from cognitive empathic processes, it is difficult to ignore the contribution of
noncognitive processes in empathic outcomes. In other words, both cognitive and noncognitive
empathic processes occur simultaneously. In the EASP framework, I will address
communication and negotiation to represent empathic processes. Because the present study aims
to identify the impact of caregivers’ characteristics on EASP, empathic process issues will be
examined in future research.
The last construct of importance in Davis’s empathy model is the interpersonal outcomes
of the observers’ behaviors toward the target persons. Being empathic can increase helping
behaviors and decrease aggressive behaviors, resulting in cooperation, active support for others,
and improved communication; as a consequence, those helping behaviors can improve social
relationships in observers and target persons (M. H. Davis, 2006). Regarding the phenomenon of
congruence of symptom perception, caregivers who are empathic with patients’ symptoms may
have empathic accuracy (congruent perception of patients’ symptoms), such that they may
engage in helping behaviors, such as symptom management or daily activities assistance, for

26

patients. Therefore, patients can generally benefit from caregivers’ empathic accuracy (congruent
perceptions). However, M. H. Davis also has indicated that aggressive behavior by caregivers
sometimes occurs even in cases where high empathic accuracy is displayed. Consequently, the
perceivers’ empathy or empathic accuracy does not always guide them to perform helping
behaviors. In other words, EASP in the home environment does not guarantee that family
caregivers will successfully provide symptom management. However, caregivers who did not
have empathy or EASP will be unable to carry out an appropriate symptom management
programs to help patients.
In sum, the theoretical conceptualization of empathic accuracy relevant to this current
research includes the family caregiver’s relationship to the patient, caregivers’ emotionality, and
the perception of threats in the caregiving situation. From Davis’s perspective, empathic
accuracy is an intrapersonal outcome of elements of the empathy process. People have abilities
to comprehend other people in their daily life and specific circumstances. In turn, these
experiences can encourage perceivers to perform helpful behaviors or social behaviors.
Perceivers’ emotional states and perceptions of threats around them is an important antecedent in
empathic accuracy. In caregiving situations, family caregivers who have empathic accuracy
(congruent symptom perceptions) with their patients may act to relieve those symptoms with
helping behaviors; these behaviors can be understood within the interpersonal outcomes of
Davis’s model. Again, empathic accuracy, the theoretical concept in this research, will use
congruence of symptom perception as an operational definition. The full, complex, and dynamic
empathic processes that take place in caregiving situations will be examined in future research.
In the current study, empathic accuracy is a cognitive act of the caregivers in perceiving
patients’ symptom experiences through accurate recognition and assessment, so that patients’

27

symptom experiences can be understood. Congruence of symptom perception is used to
represent a measure of the caregiver’s empathic accuracy. A straightforward way to measure
empathic accuracy, which has been suggested by Ickes (1993), is to compare the content of the
target person’s actual thoughts and feelings with the content of the perceiver’s corresponding
inferred thoughts and feelings. EASP in this present study is operationally defined as the
congruent levels of symptom perception, which is a measure of a caregiving situation in which
family caregivers and patients have similar perceptions of patients’ symptoms, specifically
focusing on pain and fatigue severity and frequency. Less congruent symptom perceptions when
comparing patients’ and caregivers’ symptoms scores are differentiated as overestimation and
underestimation. Symptom experiences are subjective experiences in which the patient’s own
symptom perceptions are always used as the referents. Therefore, overestimation occurs when
caregivers’ symptom scores are higher than patients’, and underestimation occurs when
caregivers’ symptoms scores are lower than patients’.
The theoretical model providing empathy research and the conceptualization of empathic
accuracy used in this study illustrate that this research study will focus on the caregiver’s
perception of the patient’s symptoms and will examine the influences of the caregiver’s
emotionality and perception of threats. In the next section, B Given’s middle-range FCM is used
to contribute to the development of the conceptual model of EASP.
Middle-Range Nursing Care Model
The FCM has been developed by B. Given (1995) for families of patients with chronic
illnesses who are living in the community. In general, the FCM describes the antecedents and
consequences of supporting patients’ care demands and requirements for family care. Care
demands are patients’ needs that include usual daily care, such as bathing and dressing;

28

instrumental daily care, such as transportation and shopping; and symptom management.
Antecedents are the characteristics of both patients and caregivers–such as age, sex, and
socioeconomic status (SES)–as well as the patients’ disease status and social supports.
Consequences of family caregivers involved in patient care fall into health outcomes for both
patients and family caregivers (B. Given, 1995).
Typically, cancer symptom management has been the most common care demand for
family caregivers who provide care at home (van Ryn, et al., 2011). In the present research, the
FCM will be applied to support the framework of EASP (congruent symptom perception) to
identify consequences of empathic accuracy in conducting symptom management and
maintaining patient health. Because symptom management in the FCM is one of the care
requirements that can generate caregiver burden, which may influence patients’ health outcomes,
caregivers’ EASP will be regarded here as an influencing factor to determine appropriate
symptom management. EASP may fit in the construct of care demands as a source of
information that family caregivers can use to provide symptom management.
Other sources for caregivers include formal support (e.g., visiting nurses) informal social
support, friends, or secondary family caregivers (B. Given, 1995). Therefore, EASP is important
for the family caregivers of cancer patients, and congruence of symptom perception may
contribute to patients’ quality of life. The present study will examine caregivers’empathic
accuracy (congruence of symptom perception), and will specifically measure
caregivers’empathic accuracy of patients’ pain and fatigue severity and frequency, because pain
and fatigue are the most burdensome symptoms for cancer patients in their daily lives.
Furthermore, the FCM conceptualizes caregiver burden as a potential consequence of a
caregiving situation with multiple care demands and requirements. B. Given’s research (1995)

29

has hypothesized that the strength of care demands and requirements depends on patients’
physical, functional, and psychological capacity, which is directly correlated with patients’
physical and functional performances and also causes increased loads of care demands placed
upon family caregivers (B. Given, 1995). Once the patient’s capacity has deteriorated, family
members become more involved in patients’ lives to provide care, such as symptom management
and assistance of daily life activities. As described in B. Given’s model, care demands, including
symptom management, can be considered threatening caregiving situations for caregivers,
because supporting care demands can interfere with caregivers’ lives, causing caregiver burden.
In this regard, caregiver burden can be considered a threat to caregivers, which may potentially
decrease caregiver’s empathic accuracy with patients. Caregiver burden describes caregivers’
negative subjective perceptions about their healthy strength, their support from other family
members, the interruption of their social lives, and their changes in employment status. The
description of caregivers’ responses to providing care in B. Given’s perspective has included a
postive aspect—self-esteem in providing care. Those five caregiver reactions have resulted from
dealing with care demands that include symptom management (B. Given, 1995).
Caregivers’ feelings of burden may also be related to the concept of compassion fatigue.
The term compassion fatigue was first used by Johnson (1992) to describe a situation whereby
health care providers had either become apathetic or experienced helplessness and anger in
response to the stress they felt in witnessing patients go through devastating illness or trauma.
Because family caregivers replace a part of the health care providers’ role in patient care,
compassion fatigue may occur in family caregivers. However, compassion fatigue may only
describe extremely negative emotional reactions resulting from caregiving, and not every family
caregiver experiences extreme burden. Typically, some family caregivers can derive positive

30

meaning from caregiving, although they feel caregiver burden (B. Given, 1995; Y. Kim, Schulz,
& Carver, 2007). Therefore, caregivers who have caregiver burden are more likely to have
compassion fatigue, so they may be unlikely to perceive patients’ symptoms accurately. Both
Ickes and Davis have argued that there are situations that can influence the accuracy of empathic
perception. Threatening situations are hypothesized to decrease empathic accuracy (Ickes, 1997).
The strength of the situation in Davis’s (1994) model is assumed to influence the empathy
process. Because cancer care is a source of stress for caregivers and the families of cancer
patients, and given that caregiver burden is one threat of a caregiving situation, this research
hypothesizes that a threat of caregiver burden can decrease accurate perceptions of cancer
patients’ symptoms. The concept of caregiver burden is critical for reflecting caregivers’ abilities
to manage patient symptoms. However, the caregivers who already provide symptom
management for patients may perceive threats. Caregivers may be worried about their abilities to
dealing with patients’ variable symptoms and also afraid of additional changes in their lives
when caregiving becomes an endless situation. This threatening feeling may reflect in their
caregiver reactions; in particulary, caregiver burden. Therefore, the present research
conceptualizes caregiver burden as a threat to the caregiver and also assumes this threatening
feeling is an impediment to caregivers’ accurate empathy or congurent perceptions.
Additionally, the FCM has shown that patients’ and caregivers’ individual characteristics
influence how care requirements and demands are experienced and dealt with, which also
indirectly determine patients’ outcomes. Caregivers, especially female family members and
spouses, often take responsibility for those care requirements. Caregivers who live with patients
are more likely to report a higher level of caregiver depression and caregiver burden (B. Given,
1995). Caregivers who live with patients, such spouses, may be more familiar with patients’ cues

31

about symptoms than caregivers who do not live with patients. The links between social support
and care demands and requirements have indicated that caregivers can obtain the assistance to
deal with care demands, such that families who have broader social networks may have access to
a wider range of relatives, friends, and formal health care providers to support care demands and
requirements, thereby lightening a caregiver’s burden of care.
In conclusion, regarding the empathy model, the concept of empathic accuracy, and the
FCM, EASP between cancer patients and their family caregivers describes a caregiving situation
in which family caregivers are capable of peceiving patients’ symptoms and accurately assess
patients’ symptom experiences. Empathic accuracy may guide caregivers to provide appropriate
symptom management. These three theories in the current study will be applied to support the
theoretical framework of EASP.
Ickes’s empathic accuracy has demonstrated that similar perceptions of patient symptoms
between patients and caregivers (empathic accuracy) is influenced by caregivers’ perceptions of
threat in the caregiving situations. Davis’s empathy model proposes that empathic accuracy, as
congruence of symptom perception, is a proximal cognitive outcome of empathy processes
influenced by emotionality and the threats of the situations. Emotionality may be represented by
caregivers’ depression as an antecedent, such that caregivers’ depression may decrease empathic
accuracy and may lead to inappropriate perceptions of patients’ symptoms. Briefly, these two
theories can support the hypothesized relationships among emotionality, threat of caregiving, and
empathic accuracy; therefore, this research hypothesizes that caregiver depression and caregiver
burden may decrease empathic accuracy.
The FCM has supported the importance of symptom perception for symptom
management and has established the rationale for hypothesizing the relationship between

32

accurately perceiving symptoms and patients’ quality of life. The perspective from FCM can
help intergrate the concept of EASP into caregiving research to enrich an understanding of the
caregiving situation, thereby advancing the science of family cancer care.
In addition, regarding Davis’s empathy model, two important elements should be
considered in the present conceptual framework of EASP. The first is communication (verbal
and nonverbal), which is an essential empathic process as an intersubjective interaction beween
the perceiver and the actor that can affect caregivers’ perception of cancer patients’ symptoms.
The second is the concept of helping behaviors, which are the distant consequence of having
empathic perception. However, communication and helping behaviors, related to the changes
within the dyadic relationship, in family cancer care are beyond the scope of this current study.
This study will focus on identifying the relationships among caregivers’ emotionality, threats,
and empathic accuracy (congruence) of pain and fatigue severity and frequency. Therefore,
communication and helping behaviors are merely added to the conceptual framework of EASP
for a complete depiction of the processes and variables.
Conceptual Framework of EASP
For the present study, the EASP conceptual framework represents an aspect of empathy
that may be relevant to the caregiving situation for cancer patients and their caregivers (Figure
1). Three theories orient the model with concrete theoretical support oriented: empathic
accuracy, the empathy model, and the FCM. The EASP depicts the influence of caregivers’
emotionality (depression) and threats of caregiving (caregiver burden) on the main concept of
empathic accuracy (the congruence of pain and fatigue severity and frequency) while considering
the influences from patients’ and caregivers’ demographics. The EASP framework also shows
the expected outcomes of EASP in caregivers’ helping behaviors and the quality of life of

33

patients and caregivers.
The theoretical concepts being examined in this study are the impact of caregivers’
emotionality and threats of caregiving on EASP. Caregivers–instead of physically feeling the
same symptom experience that patients do–can cognitively appraise patients’ symptom
experiences. The EASP framework depicts that both emotionality and threats of caregiving can
influence caregivers’ perceptions of patients’ symptoms experiences, thereby affecting the
congruence levels of symptom perception between patients and caregivers. Therefore, EASP in
this study, which is inspired by Ickes’s and Davis’s perspectives of empathic accuracy, is
conceptually defined as the similarity of family caregivers’ cognitive appraisals of their patients’
cancer symptom experiences. Symptom experience, is described as a cognitive symptom
perception, including severity, time (duration and frequency), distress, and quality (Armstrong,
2003). The current study will operationally examine severity and frequency of pain and fatigue
to represent symptom perception.
Based on the operational definition of empathic accuracy as congruent levels of symptom
perception, empathic accuracy is obtained from both patients and caregivers. In this regard, the
characteristics of caregivers and patients influencing the EASP should not be ignored. The
contextual factors in the EASP framework include patients’ and caregivers’ demographics, such
as age and sex, as well as the caregivers’ relationship to the patient, living arrangement, and
awareness of the total number of patients’ symptoms. However, because the present study will
focus on examining how caregivers’ emotionality and threats of caregiving can impact empathic
accuracy, caregivers’ demographic characteristics, which are expected to affect emotionality and

34

Theoretical concept &
Contextual factors

Predicting factors

Patient
characteristics:
Age
Sex

Outcomes

Empathic accuracy
Congruence of pain and fatigue
severity and frequency

Caregiver
characteristics:
Age
Sex
Relationship to
patient
Education
Number of patient’s
symptoms
Living arrangement

Emotionality
Caregiver
depression

Threats of caregiving
Caregiver burden

Caregiver
helping
behavior

Patient
QOL

Symptom
management

Caregiver
QOL

Empathy process
Communication and negotiation
Figure 1. Theoretical framework of Empathic Accuracy of Symptom Perception (EASP). QOL = quality of life. The relationships
among the concepts in the boxes with double-line are tested. The concepts in the gray boxes were not tested in the current research.

35

threats of caregivers, will be taken into account, but patients’ factors will not be included in the
analysis.
In the current study, caregiver emotionality is defined as caregivers’ mood expression,
such as anxiety and depression. Depression is frequently recognized in the family caregiver
population when caregivers have insufficient knowledge and insufficient support for care
(National Family Caregivers Association, 2010). Accordingly, caregivers’ depression, perhaps,
leads caregivers to believe that patients are suffering from pain or fatigue; depressed caregivers
may be more likely to rate a higher score of severity and frequency for their patients’ symptoms
than the patients’ self-rating of symptoms, thereby decreasing empathy accuracy (congruence of
symptom perception).
The current study describes the threats of caregiving as perceived by caregivers. Threat
has often been described as an indication of imminent harm, danger, or pain. In a family, cancer
diagnosis is a threat because family caregivers believe that they are losing a patient, and they
may also believe that patients must be suffering from cancer. Threats of caregiving to caregivers
can refer to negative expectations of life challenges resulting from patients’ cancer and cancer
treatment. Some cancer dyads have revealed that they were unprepared for symptom
management, and didn’t know how to cope with the unpredictable nature of life (Harden et al.,
2002; Pickett, Barg, & Lynch, 2001). Caregiver appraisal of providing care has been interpreted
by using the following terms: threat, benign, benefit, and challenge (Bakas & Burgener, 2002;
Oberst, Thomas, Gass, & Ward, 1989). Threats of caregiving do not describe actual physical
harm or danger to caregivers, but infer the negative perception of potential life challenges.
Caregiver burden has been described as caregivers’ negative reactions from the
imbalances that are created in their social, psychological, and economic living situations,

36

because providing care has interrupted these aspects of their lives (B. Given, Given, &
Sherwood, 2012). Often, caregivers may imagine that there will be more unknown or difficult
patient care in the future, especially when patients are undergoing anticancer treatment.
Caregiver burden may be a reflection of the fact that caregivers are perceiving threats of
caregiving.
Therefore, caregiver depression will be used to represent emotionality, and caregivers’
feelings of burden from providing care will represent the perceptions of threats in caregiving
environments. Both emotionality and threats are theoretically understood as factors that may
diminish empathic accuracy within the empathic processes surrounding caregivers’ perceptions
of the patient.
Contextual factors of the EASP refer to dyadic characteristics. According to the FCM,
both patients’ and caregivers’ demographic characteristics and patients’ health conditions can
influence care demands related to health outcomes of patients and caregivers. Both Davis’s and
Ickes’s theories on empathic accuracy have emphasized the perceiver’s personal factors
influencing empathic process, but have not included the target person’s characteristics. The
present study will examine caregivers’ demographic characteristics (age, sex, education,
relationship to patients, living arrangement, and number of patient symptoms) and then test their
relationships with caregiver emotionality and threats of caregiving. It will also examine the
covariant effect of those concepts on empathic accuracy.
Based on the empathy model and FCM, the researcher expects that the conceptual
framework of EASP can generate caregivers’ helping behavior, such as approaching symptom
management, and improve patients’ and caregivers’ quality of life as well. Providing symptom
management may be a helping behavior, which is one of the interpersonal outcomes of empathy

37

processes in Davis’s empathy model. Patients’ and caregivers’ quality of life is the outcome of
providing for care demands in Given’s FCM. Based on the EASP conceptual framework,
increasing congruence of symptom perception is assumed to generate more appropriate symptom
management and improve quality of life for patients and caregivers. The concept of symptom
management is broad and beyond the scope of this study. Symptom management is the care
given to improve quality of life of patients who have a serious or life-threatening disease, and
includes any strategies used to prevent and treat the symptoms related to disease, side effects
related to anticancer treatments, and any psychological, social, or spiritual problems related to
disease and treatment (Cella, Wagner, et al., 2007; Dodd, et al., 2001; Fu, LeMone, & McDaniel,
2004). The data used for this current study does not include sufficient information to differentiate
between caregivers’ behaviors of symptom management for patients. Further study will evaluate
types of caregivers’ helping behavior regarding symptom management to support this
framework. The present study will focus on predicting the relationship between caregiver
emotionality and EASP, and between caregivers’ perception of threats of caregiving and EASP.
Specific aim. The specific aim of this research is to examine how caregivers’
emotionality (caregiver depression) and caregivers’ threats of caregiving (caregiver burden)
predict EASP (congruence of pain and fatigue severity and frequency), when taking into account
contextual factors of caregivers–including caregiver’s age, sex, education, relationship to
patients, observation of the total number of patients’ symptoms, and living arrangement.
Hypothesis. There are two hypotheses for this study:
1. Increased caregiver depression and caregiver burden will be related to lower congruent
values of pain and fatigue severity and frequency.
2. Caregivers who are female, older, patients’ spouses, better educated, perceiving a

38

lower number of patients’ symptoms, and living with patients will have greater congruence of
pain and fatigue severity and frequency than the caregivers who are male, younger, nonspouses
of patients, less educated, perceiving a higher number of symptoms, and not living with patients.
Therefore, this study will examine the predicted relationship that caregivers’ emotionality
and threats of caregiving impact EASP with and without contextual factors of caregivers’
characteristics.

39

CHAPTER 3
Literature Review
This chapter will review literature to support the framework of EASP. It will begin by
introducing the population of family caregivers of cancer patients and describes research about
cancer patients’ care demands specific to cancer and its symptom management. Next, literature
that supports the relationships among the main concept, factors, and outcomes in the theoretical
framework of EASP will be discussed. Specifically, the literature reviewed on the following
pages will address (a) the family caregivers of cancer patients, and at-home cancer care related to
pain and fatigue; (b) EASP and its relationship with caregivers’ emotionality and threats of
caregiving; and (c) caregivers’ demographic characteristics that are relevant to caregivers’
emotionality, threats of caregiving, and EASP. The final section in this chapter will illustrate the
methods that are designed to identify congruent symptom perception.
Family Caregivers of Cancer Patients
Family caregivers generally are the family members who are unpaid care providers for
patients, and who have no medical training (Family Caregiving Alliance, 2011). According to the
NAC and AARP (2009), approximate 65.7 million informal caregivers in the United States take
care of patients who have had diagnoses across all chronic condition, and 7 % of adult family
caregivers were providing care for an adult family member diagnosed with cancer. Also, among
those adult caregivers, they are predominantly female (66%), the average age caregivers is 48,
and most are married and living with their partners (62%). In addition to this, 74% of adult
caregivers continue to work while they manage caregiving responsibilities. Families provide
about 80% of at-home care services. Caregivers spend 20 hours per week, on average, providing
care for their ill family members, and 13% provide weekly care for 40 hours or more. The

40

average length of caregiving is 4.3 years (NAC & AARP, 2009).
In addition, caregivers who live with patients provide three times as many hours of care
per week, and primary caregivers spend more hours giving care than nonprimary caregivers.
Female caregivers spend more time providing care than do males, and females also handle most
of the time-consuming and difficult care demands (NAC and AARP, 2009; Reinhard, Given,
Petlick, & Bemis, 2008). The research reports on family caregivers of adult cancer patients have
shown that most of these family caregivers are spouses, partners, or adult children. Relatives,
close friends, or neighbors may take this role when family is not available (ACS, 2010; Mintz,
2007; Paul, et al., 2007).
According to the American Cancer Society (ACS, 2010) annual report, the mortality from
all cancers in the United States has declined by 60% between 1990 and 2006. The 5-year
survival rate from cancer was 68% between 1999 and 2006, which is higher than the survival
rate between 1975 and 1986 (Jemal, Siegel, Xu, & Ward, 2010). Today’s cancer population has a
longer life expectancy. Additionally, a shorter period of hospitalization for treatments is a current
medical trend that is shifting cancer care from hospitals to families. In sum, the cancer
population is growing in the community. Because of shorter periods of cancer treatment in
hospitals and longer life expectancies for cancer patients, it is necessary to have a larger number
of family caregivers who can cope with the increasing number of cancer survivors. For
caregivers of cancer patients, symptom management is a major care demand when they provide
care for patients at home, second in frequency only to assisting with patients’ daily activities.
Caregivers’ perceptions of patients’ symptom will be an important influencing factor in deciding
how to provide symptom management.

41

Cancer Care at Home
Family caregivers of patients with cancer provide vital and continued physical and
emotional support for patients, helping them to get through the cancer experience. In cancer
families, involvement of family caregivers is essential for optimal treatment of cancer patients–
ensuring treatments compliance, continuity of care, and social support (Glajchen, 2004;
Reinhard, et al., 2008). The B. Given et al. study (2004) examined 152 dyads of cancer patients
and their family caregivers, and found that most cancer care demands provided by family
members are related to assisting patients with routine activities of daily life. A majority of care
demands that require patients to request assistance from family are daily life activities and
instrumental activities (56%), including helping feed, dress, and bathe patients, as well as
arranging schedules, managing insurance, and providing transportation (NAC & AARP, 2009).
When patients are more dependent, performing those daily care demands becomes more timeconsuming and intensive.
Meanwhile, family caregivers are expected to assess and report information regarding
patients’ symptoms, to identify side effects of patients’ cancer treatments, and to administer
medication for symptom management. Family caregivers are expected to provide emotional
support for patients as well (Glajchen, 2004; Reinhard, et al., 2008). Despite a decline in cancer
mortality in a report of the National Cancer Institute (Jemal, Siegel, et al., 2010), malignant
neoplasm is the second leading cause of death next to the heart disease, which has been the first
cause of death in the last decade (CDC, 2009). Additionally, cancer symptoms usually change
during the time of receiving treatment and post treatment (Fan, Filipczak, et al., 2007). In
particular, the changes of cancer symptoms may generate new care demands related to

42

symptoms. These uncertain situations may require more attention from caregivers, who monitor
the changes of symptoms and provide symptom management. Therefore, caregivers who have
been providing patient care may perceive more interruption to the family because of patients’
care demands.
Nowadays, cancer patients are often discharged to their homes when they just receive the
anti-cancer treatment and may still have treatment-related symptoms. Those patients need their
family caregivers to remain with them and support their daily activities at home. Family
caregivers provide patient care at home as the care partners in the current health care systems.
The present study explores congruent levels of cancer patients’ pain and fatigue severity and
frequency between patient and caregivers. However, because this study only uses the one-time
measure of information at the first interview from the parent study, there will not be sufficient
information to examine the changes of symptoms.
Generally, family caregivers are not as familiar with clinical cancer care, especially
symptom assessment and management, as they are with usual daily care. However, since the
cancer treatments can be provided in the clinics, cancer care demands become a major part of
caregivers’ responsibilities (ACS, 2010). van Ryn, et al (2011) recruited 677 caregivers who
provided at least 50% of informal, at-home care for patients diagnosed with lung and colorectal
cancer, and then found that those family caregivers provided most of the cancer-specific care,
including monitoring the side effects of treatments (68%), assisting with symptom management
(47%), administering medication (34%), making decisions for contacting doctors (30%) and for
using medications (29%), and changing wound dressings (19%). In their findings, symptom
management was the most prevalent caregiving activity that provided by caregivers, except for
assisting with patients’ daily activities. In conclusion, caregivers’ perceiving patients’ symptoms

43

and assisting with symptom management are ongoing care demands in patients’ cancer
trajectories. In these circumstances, family caregivers who provide care for cancer patients
become not only advocates in assisting with symptom management and reporting cancer
patients’ symptoms to formal care providers, but also partners with professional health providers.
Therefore, caregivers’ accurate perceptions of patients’ symptoms may be important for
managing symptoms in patients.
Congruent Levels of Symptom Perception as a Measure of Empathic Accuracy
Congruence of symptom perception between patients and caregivers is the operational
definition of empathic accuracy in the present study. Congruence in each dyad is measured by
comparing the scores of a patient’s and a caregiver’s perception of patient symptoms. Greater
similarity in symptom scores between patients and caregivers indicates greater empathic
accuracy (congruence) for the dyad.
Congruence of pain and fatigue severity and frequency. Generally, previous studies
examined congruence of symptoms perception to identify whether caregivers can have a accurate
symptoms perception to guide them in providing symptom management. Congruence in previous
studies is a level of agreement, in which less agreement statistically means non-congruence.
Congruent perception of multiple symptoms is measured by correlation coefficient in previous
studies ranged widely from fair to moderate (Broberger, et al., 2005; Higginson & Gao, 2008;
Kurtz, et al., 1996; Lobchuk & Degner, 2002; Wennman-Larsen, et al., 2007; Yeşilbalkan &
Okgün, 2010). Without specific operational definition of congruence, a fair-to moderate level of
agreement is recognized as non-congruence. Previous studies also found that caregivers tended
to overestimate most patients’ symptom distress and severity, but some symptoms, such as dry
mouth, constipation, and insomnia, were underestimated by caregivers (Broberger, et al., 2005;

44

Kurtz, et al., 1996; Lobchuk & Degner, 2002; McPherson, et al., 2008; Molassiotis, Zheng,
Denton-Cardew, Swindell, & Brunton, 2010; Wennman-Larsen, et al., 2007).
Researchers interested in pain experiences between cancer patients and caregivers have
found that most caregivers tend to overestimate patients’ pain severity (Greig, Chow, Bovett,
Abdolell, & Gillies, 2005; C. Lin, 2001; Miaskowski, Zimmer, Barrett, Dibble, & Wallhagen,
1997). Some studies have evaluated pain frequency, and also have found that caregivers tend to
overrate the occurrence of pain (Broberger, et al., 2005; Lobchuk & Degner, 2002; McPherson,
et al., 2008). Other studies which measured multiple symptoms also have reported that caregivers
commonly overestimate frequency of patients’ symptoms (Higginson & Gao, 2008; McPherson,
et al., 2008; Molassiotis, Zheng, et al., 2010; Silveira, Given, Given, Rosland, & Piette, 2010;
Wennman-Larsen, et al., 2007; Yeşilbalkan & Okgün, 2010).
Furthermore, In those studies on congruence of multiple symptoms, fatigue–but not pain–
and its related symptoms, such as lack of energy and weakness, have been identified as the most
prevalent, and may also be the most severe, symptoms by both patient and caregivers (Higginson
& Gao, 2008; McPherson, et al., 2008; Wennman-Larsen, et al., 2007). Vogelzang et al.’s
research (1997) measured both family caregivers’ and cancer patients’ perceptions of cancerrelated fatigue. They found that caregivers overestimated the frequency of patients’ daily fatigue,
but they did not identify congruence of symptom frequency within the dyad. Stone and Ream
(2003) focused on congruence of cancer-related fatigue in 1,370 dyads of patients and their
family caregivers in the United Kingdom. Their study showed that 71% of the 454 responding
caregivers recognized patients’ fatigue, while 56% of 576 responding patients reported fatigue
during the same time period.
Family caregivers also had a higher percentage than patients in reporting patients’ other

45

fatigue-related feelings and the effects of patients’ fatigue on different aspects of patients’ lives.
Other studies measuring multiple symptoms also have reported a higher score of fatigue from
caregivers (McPherson, et al., 2008; Molassiotis et al., 2010; Silveira, et al., 2010). These results
also were extended over the long term in longitudinal studies (Hinton, 1996; Molassiotis, Zheng,
et al., 2010; Silveira, et al., 2010). Briefly, family caregivers of cancer patients tend to
overestimate severity and distress of patients’ pain and fatigue. Based on the findings of either
low correlation coefficient or the non-zero score differences between caregivers and patient’s
symptom scores, the researchers typically provided a conclusion that caregivers’ symptom
perception is not reliable. However, in spite of this conclusion, caregivers are still required to
assist symptom management and provide patient care.
Cancer-related symptoms. Cancer-related symptoms have commonly impacted
patients’ performance of physical functioning as well as quality of life (Butt et al., 2008;
Delgado-Guay & Bruera, 2008; Miaskowski et al., 2006; Pud et al., 2008). Symptoms have been
referred to as the physical and psychological indicators of the changes in normal functioning that
are experienced by an individual who has an illness (Lenz, Pugh, Milligan, Gift, & Suppe, 1997).
Cancer–related symptoms, such as pain and fatigue, may occur before the cancer diagnosis is
confirmed, and may also alter when patients receive cancer treatments or when the progression
of cancer worsens health conditions. As patients get weaker, family caregivers usually take on
the patients’ care in its entirety, including assisting with patients’ daily activities, instrumental
daily activity, and symptom management. In order to help patients to manage cancer-related
symptom and side effects of cancer treatment, caregivers have to continually adjust their lives for
caregiving situations. Dynamic characteristics of cancer-related symptoms, along with treatments
in the cancer trajectory often create unfamiliar demands for cancer patients and their family

46

caregivers (Fan, Filipczak, et al., 2007). The National Institute of Health [NIH] (2004) statement
for symptom management in the cancer population also reported that although different cancer
sites may generate different symptoms, pain and fatigue are the most common symptoms of
cancer and cancer treatments.
Cancer pain. Cancer pain has been the focus of most cancer research on symptom
management. Pain is the most common and feared symptom of cancer by Americans (ACS,
2010). Cancer patients report pain as the most unpleasant and disruptive symptom because
unrelieved pain can interrupt patients’ daily lives and limit their capacities to performing
physical functioning activities (Vallerand, Templin, Hasenau, & Riley-Doucet, 2007).
Specifically, because of its multidimensional nature, pain can be triggered or worsened by other
psychological and physical symptoms, such as anxiety and sleep problems. Therefore, it is
complicated to alleviate pain distress (Delgado-Guay & Bruera, 2008; McGuire, 1995; van den
Beuken-van Everdingen, et al., 2007). Regarding the recognition of multidimensionality of
symptoms, several symptom instruments have been developed to evaluate multiple dimensions
of multiple symptoms, such as the M. D. Anderson Symptom Inventory [MDASI] and the
Memorial Symptom Assessment Scale [MSAS] which are used to evaluate symptom severity,
interference, and frequency (Cleeland, 2007; Portenoy et al., 1994).
In addition, cancer patients in the late stage of cancer, especially breast cancer and lung
cancer, have a high incidence of bone metastasis. Bone metastasis means that the tumor cells
spread into bone and replace the bone tissues. As the bones are replaced with tumors, patients
often feel bone pain because nerve endings in and around the bone send pain signals to the brain.
Therefore, cancer patients with bone metastasis are usually suffering more from pain than they
were in the earliest stages of their disease. Pain management has been in the spotlight of the

47

cancer research on symptoms, since pain experiences are typically uncomfortable for cancer
patients. The investigators who were interested in congruent levels of symptom perception
between cancer patients and their caregivers typically measured patients’ pain severity and
distress (Lobchuk, et al., 2006; McPherson, et al., 2008; Molassiotis, Zheng, et al., 2010;
Silveira, et al., 2010).
Cancer-related fatigue. Cancer-related fatigue is another multidimensional symptom.
Fatigue has been recognized as one of the most prevalent symptoms for cancer patients who
receive treatment (Donovan et al., 2004b; Goedendorp, Gielissen, Verhagen, Peters, &
Bleijenberg, 2008; Hagelin, Wengstrom, Ahsberg, & Furst, 2009; Hoffman, Given, von Eye,
Gift, & Given, 2007). According to a National Cancer Institute report, cancer-related fatigue
persists at some level for months or years. Some cancer survivors report long-term complaints of
fatigue (Baker, Denniston, Smith, & West, 2005; Bower, 2007). Cancer patients often report
fatigue as a major obstacle to maintaining normal daily activities and quality of life, especially
for patients who receive cancer treatments (Bower, 2007; Byar, Berger, Bakken, & Cetak, 2006;
Diaz, et al., 2008; Ryan, et al., 2007; Stasi, Abriani, Beccaglia, Terzoli, & Amadori, 2003). In
addition, fatigue can impact all areas of life function, including mood, physical function, and
social interaction (Curt et al., 2000; Donovan et al., 2004a; B. Given et al., 2002). For family
caregivers who have been involved in patients’ lives, patients’ fatigue may indirectly impact
caregivers’ lives as well (Passik & Kirsh, 2005; Vogelzang, et al., 1997).
Furthermore, cancer patients often report pain and fatigue at the same time. Both pain and
fatigue practically have been identified as the core of symptom clusters (Fan, Hadi, et al., 2007;
B. A. Given, et al., 2007; Hoffman, et al., 2007; E. Kim, et al., 2009). Fatigue often coexists with
insomnia and depression, and these symptoms are also associated with pain (Donovan &

48

Jacobsen, 2007; Hoffman, et al., 2007; E. Kim, et al., 2009; Redeker, Lev, & Ruggiero, 2000).
As mentioned above, pain is the most unpleasant symptom, and fatigue is recognized as the most
prevalent symptom by both patients and caregivers. Pain and fatigue are the symptoms that most
commonly interfere with patients’ physical function (Cella, Yount, et al., 2007; B. Given, Given,
et al., 2001). Patients who are suffering from both pain and fatigue may have less capability to
perform physical functioning, such that they have to depend more on family caregivers for
assistance in their daily activities. Patients’ dependence indicates caregiving around these
symptoms (Miaskowski, 2010; P. C. Stone & Minton, 2008), which may generate more care
demands for caregivers. It is worthwhile to prioritize pain and fatigue before attempting to
understand overall cancer-related symptoms.
Symptom characteristics. Symptom experience, in Dodd et al. (2001), is described as
the individual’s perception of the symptoms, evaluation of symptom meaning, and subsequent
responses. Later, Armstrong (2003) defined symptom experience as the perception of the
frequency, intensity, distress, and meaning occurring as symptoms are produced. As symptom
characteristics including distress (severity and/or distress), occurrence (prevalence, duration, and
frequency), and interference are counted, the researcher and caregivers can translate subjective
symptom feelings into cognitive perception. Severity and distress commonly have been
measured to represent symptom burden, because the high levels of severity and distress create
the most interference with patients’ quality of life. Kirkova et al. (2006) reviewed the symptom
instruments used to measure multiple symptoms, and they found that among those symptom
dimensions, symptom prevalence and severity were often used as the main determinant for
symptom burden, while frequency and duration were less frequently reported or measured.
Symptom severity has been suggested to positively determine the patients’ distressful

49

perception of symptoms (H. J. Kim & Abraham, 2008; Kirkova, et al., 2010). Therefore, severity
often has simply been measured for symptoms. Mercadante et al. (2000) showed that 370 cancer
patients reported worse functional performance when they had increasingly frequent and intense
(severe) symptoms. Lutz et al. (2001) investigated the relationship between symptom experience
and quality of life among 69 patients who were diagnosed with lung cancer, and found that the
patients who had frequent and severe symptoms reported worse quality of life. Both Mercadante
et al.’s and Lutz et al.’s findings have suggested that an increase in both frequency and severity
of symptoms can impact cancer patients’ performance of daily physical functioning and worsen
patients’ quality of life (Lutz, et al., 2001; Mercadante, et al., 2000). Meanwhile, family
caregivers are named as informal caregivers, because they do not receive formal training for
symptom management as professional caregivers do. It is easy for caregivers to assess severity
and frequency of symptoms at home to represent patients’ symptoms burden.
In summary, measuring the severity and frequency of patients’ pain and fatigue is a
priority for symptom management, especially for the patients who are receiving chemotherapy
(Molassiotis et al., 2009; Sigurdardottir & Haugen, 2008; Soltow, Given, & Given, 2010).
Because caregivers play a major role in assisting patients in appropriate symptom management,
having empathic accuracy of the severity and frequency of pain and fatigue between patients and
caregivers is critical. However, because caregivers have low congruent perception and also
commonly overestimate cancer patients’ symptoms, caregivers have been considered as
unreliable proxies of patients’ symptoms (Chau et al., 2010; Greig, et al., 2005; Kirou-Mauro,
Harris, Sinclair, Selby, & Chow, 2007). Therefore, it is essential to understand the congruence of
symptom perception. In the present study, EASP will be measured by comparing cancer patients’
and caregivers’ perception of pain and fatigue severity and frequency. In this study, the EASP

50

framework will examine the three congruence levels of symptom perception between patients
and caregivers for pain and fatigue severity and frequency, in terms of overestimation,
congruence, and underestimation, separately, to identify factors influencing congruence.
Relationship Among Caregiver Emotionality, Caregiver Threats of Caregiving, and
Congruence of Symptom Perception
Caregiver emotionality. Emotionality consists of positive and negative affect and is
measured or represented by emotional intensity (Eisenberg et al., 1994). Caregiver emotionality
is often described with negative mood components in which depression is commonly
emphasized. The Family Caregiver Alliance (Family Caregiving Alliance, 2009) has reported
that caregivers with depressive symptoms may isolate themselves, and lose interest in people.
Lee, Brennan, and Daly (2001) also emphasized that increasing caregivers’ depression can
prevent them from perceiving the benefits of providing care, and may also decrease caregivers’
empathy. Redinbaugh and her colleagues (2002) reported that caregivers are more likely to
overestimate patients’ pain when they have more negative feelings about patients’ pain, when
they believe that patients are suffering from great pain, and when they need to provide numerous
pain medications for pain relief. Tang, et al. (2006) examined 170 Taiwanese family caregivers
of cancer patients at the end of life, and found that the caregivers who were more aware of
patients’ disease experiences and symptoms reported higher level of depression. Lobchuck and
Vorauer (2003) also have suggested that the overestimation of patient worrying may be
influenced by caregivers’ depressive symptoms and anticipating grief. Accordingly, caregivers
who have negative emotionality may be more likely to inaccurately perceive patients’ symptoms.
In contrast to the description above about the relationship between caregiver depression
and congruence of symptom perception, Kurtz et al. (1996) found that caregiver depression was

51

not correlated with congruence of symptom frequency among 216 dyads of caregivers and
cancer patients. Shega, Hougham, Stocking, Cox-Hayley, and Sachs (2004) evaluated 115 dyads
of family caregivers and community dwelling persons with dementia and found that caregivers’
depression was correlated with congruence of patients’ pain severity. Meanwhile, Hung, Pickard,
Witt, and Lambert (2007) found that depressed caregivers of stroke patients were more likely to
underestimate patients’ pain experience but caregivers with pain experience tended to
overestimate patients’ pain. These inconsistent results may have occurred because the studies
dealt with different diseases and different symptoms, or perhaps these patients’ pain had been
treated before they returned home. The present study will use caregiver depression to represent
emotionality in order to examine how caregiver depression predicts empathic accuracy of pain
and fatigue severity and frequency.
Threats of caregiving. According to the EASP framework for this study, for a family
care situation, the threats of caregiving are conceptually defined as caregivers perceiving threats
from providing care because supporting patients’ care demands can interrupt caregivers’ lives,
which has been recognized as caregiver burden (Grunfeld, et al., 2004). Generally, threats of
caregiving do not describe actual physical harm or danger to caregivers, but infer the negative
perception of potential life challenges. Cancer not only brings the threat of death to the ill family
members, but also generates various threats in the family. Family caregivers may feel empathic
with patients’ symptom burdens, and also feel threats from providing patient care. Supporting
care demands can consume caregivers’ time and interrupt caregivers’ lives, which has been
recognized as caregiver burden (Grunfeld, et al., 2004). B. Given, et al (2011) reinforced the
necessity of caregiver involvement in patient care after patients received cancer treatments,
because symptom management can be a core care demand from treatment to post treatment.

52

Glajchen (2004) emphasized that cancer caregivers often operate within caregiving situations
and without preparation. Caregivers may feel burdened when their lives are continually
interrupted to provide care to meet patients’ demands. Caregiver burden is described as
caregivers’ negative reactions from the imbalances that are created in their social, psychological,
and economic living consequences, because providing care has interrupted these aspects of their
lives (B. Given, et al., 2012). Therefore, caregivers may perceive the threats of providing care
that impact their own lives, because caregivers have to readjust their lives for those caregiving
situations.
From these scholars, the investigator can derive a equationtion whereby caregiver burden
can be described as caregivers’ negative anticipations about how their lives will be impacted
from providing care. Caregiver burden was defined as a multidimensional bio-psychosocial
reaction resulting from an imbalance in care demands relative to the caregiver’s personal life, in
terms of physical and emotional health, social role, financial recourses, and formal care resources
(B. Given, Kozachik, Collins, DeVoss, & Given, 2001). Montgomery, Gonyea, and Hooyman
(1985) categorized caregiver burden in two aspects: the subjective aspect describes perception of
self-distress from care circumstances, and the objective aspect refers to the numerous kinds of
care demands and numbers of caring hours.
Vitaliano and Katon (2006) showed that caregivers with high levels of caregiver burden
are less willing to provide care. Caregivers with greater burden may not notice patients’ feelings
and may avoid involvement in patient care; lower congruent symptom perception is not
surprising. Indeed, some studies have reported that caregivers with feelings of burden have a
lower likelihood of congruent perception of patients’ symptoms (McPherson, et al., 2008;
Wennman-Larsen, et al., 2007). Lobchuk et al. (2006) also reported that higher caregiver burden

53

might lead caregivers to be less likely to agree with patients’ self-reports of symptom experience.
Wennman-Larson et al. (2007) found low congruence of multiple cancer symptoms among 52
dyads of patients and their family caregivers. Three aspects of caregiver burden, including a lack
of family support, a negative impact on health, and low self-esteem, were positively correlated
with discrepant ratings of patients’ emotional functioning. Wennman-Larson et al. (2007)
interpreted that the low congruence as resulting from caregivers’ stressful situations was because
they experienced a low level of caregiving esteem and lack of family support or worsening
health.
In addition, caregiver depression and caregiver burden often have been correlated, and
caregiver burden has been identified as predictor of caregiver depression (Clark & Diamond,
2010 ; Grunfeld, et al., 2004; Y. Kim, et al., 2005; Kurtz, et al., 2004). Clark and Diamond
(2010) examined 112 caregivers of elderly people and found that caregivers’ feelings of burden,
which were measured using Zarit Burden Inventory (ZBI), were correlated with caregivers’
depression, which were measured by using the Center of Epidemiological Study Depression
scale (CESD). Similarly, four subscale of caregiver burden measured by the Cancer Reaction
Assessment was moderately correlated to caregiver depression measured by the CESD (range =
0.34 - 0.57), except for caregiver self-esteem which was negatively correlated to depression (r =
- 0.15 - 0.2; (C. W. Given et al., 1992; Sherwood, Given, Given, & von Eye, 2005). Grunfeld et
al. (2004) examined 89 dyads of caregivers and patients with breast cancer and found that
caregiver burden (measured by ZBI) was the major predictor of caregivers’ psychological
distress (depression and anxiety). Kurtz et al. (2004) analyzed 351 caregivers’ reports on the
relationship between their caregiver burdens and their physical health and depression; they found
that higher scores on “schedule interruption” and “family abandonment” predicted higher scores

54

of caregiver depression and lower scores of caregivers’ physical health. Therefore, caregivers
with a higher burden may have a higher likelihood of being depressed. Likewise, higher levels of
depression in a caregiver may imply the existence of co-occurring caregiver burden. Therefore,
both emotionality and threats can coexist to influence congruence of symptom perception. For
example, McPherson et al. (2008) examined 66 dyads of cancer patients and their caregivers.
They identified that depression and anxiety can predict discrepancy of psychological symptoms;
meanwhile, greater caregiver burden can predict discrepancy of physical symptoms. Miaskowski
et al. (1997) compared caregivers’ characteristics between the congruent group and the
noncongruent group when they examined the congruent pain intensity among 78 cancer patientcaregiver dyads. They found that caregivers in the noncongruent group were more likely to
report having greater caregiver strain than the caregivers in the congruent group. Caregiver
strain, along with caregiver burden and depression, has been the major characteristic that
negatively correlates to family caregivers’ health and physical functioning (Honea, et al., 2008).
Therefore, caregivers who have negative emotionality and who perceive threats may have
decreased accuracy when perceiving patients’ symptoms.
Caregiver Demographics as the Underlying Factors for Caregiver Emotionality and
Threats of Caregiving
Sex and age. Caregivers’ demographics have been related to depression and caregiver
burden. Ferrario, Vitaliano, Zotti, Galante, and Fornara (2003) reported that caregivers over the
age of 50 have higher depression scores than those who are younger. Female caregivers are
reported to have a higher likelihood of depression than male caregivers (B. Fletcher, Dodd,
Schumacher, & Miaskowski, 2008; Rivera, 2009). This finding on female caregivers’ depression
has been identified as being related to caregiver burden in some studies (B. Given et al., 2004;

55

Phillips, Gallagher, Hunt, Der, & Carroll, 2009; Pinquart & Sorensen, 2006). Fletcher et al.
(2008) specifically reported that younger female caregivers whose patients have prostate cancer
and who report low quality of life are more likely to report higher levels of depression, have
physical symptoms such as morning fatigue and sleeping disorders. Rhee et al. (2008) studied 60
caregivers of cancer patients and reported that depression was positively correlated with several
caregiver characteristics, including being female, being a spouse, having poor health, feeling
burdened, having poor adaptation, and demonstrating low functioning, as well as taking care of
the patients with low functioning performance. Dumont, Turgeon, & Allard, et al. (2006)
recruited 212 advanced-cancer patients with and their caregivers to examine caregivers’
psychological distress, including depression. They found that the caregivers who had a greater
caregiver burden and were younger females were significantly more likely to have higher levels
of psychological distress.
Education. In addition, education has been identified as being related to depression.
Previous aggregate-level studies have reported that people with a higher education have less
likelihood of being depressed (Kurtz & Brown, 1985; Kurtz, et al., 2004; Kurtz, Martinmorgan,
& Graff, 1987). Thus, a higher educational level may protect against the symptoms of
depression. In the caregiver population, education also has been associated with fewer symptoms
of depression. Family caregivers with higher education have also shown lower caregiver burden
(Barrentine, Street, & Kurtz, 1984; Cameron, Franche, Cheung, & Stewart, 2002; Y. Kim, et al.,
2005). The reason may be that education is related to the ability to identify resources, such that
caregivers who have a higher degree of education may be able to identify more social resources
to support their caregiving, so that they have less caregiver burden.
Patient symptoms. Caregivers’ demographic characteristics, which are described above,

56

can influence caregivers’ experience of burden and depression, but the number of patient
symptoms that caregivers observe also can be a risk factor. B. Given et al. (2004) examined 152
caregivers of cancer patients and found that caregiver reports of higher total number of patient
symptoms is positively correlated to caregivers’ depressive symptoms. Specifically, B. Given et
al. (2004) also found that most caregivers who identified seven or more patient symptoms
reported depression scores greater than 16. Therefore, caregivers who observe more patient
symptoms may have a risk of being more depressed.
Overall, caregivers’ demographics–such as female gender, younger age, higher level of
education, and being a spouse–and the total number of patient symptoms that the caregiver
observed are caregiver characteristics which may influence caregiver depression and burden.
Therefore, in this study, these caregiver characteristics may have an indirect effect on
congruence of symptom perception and will be used as control variables.
Caregiver Demographics Related to Congruent Symptom Perception
The relationship between caregivers’ demographic characteristics–age, sex and
relationship with the care receiver–and congruence of symptom perceptions is recognized in a
few studies (Broberger, et al., 2005; McPherson, et al., 2008; Wennman-Larsen, et al., 2007).
Other studies have identified these caregiver characteristics in the comparison of two groups of
symptom congruence or discrepancy (Kurtz, et al., 1996; C. Lin, 2001; Miaskowski, et al., 1997;
Silveira, et al., 2010).
Caregiver age. Two studies which measured congruence of pain severity between cancer
patients and caregivers found that caregivers in the noncongruence group were slightly older
than those in the congruence group, and that older caregivers tended to overestimate patients’
pain severity (C. Lin, 2001; Miaskowski, et al., 1997). However, the age differences were not

57

statistically significant. Other studies measuring multiple symptoms reported no differences in
caregivers’ age between congruence and noncongruence groups (Molassiotis, Zheng, et al.,
2010; Silveira, et al., 2010). Furthermore, caregivers of different ages might be more aware of
some symptoms, but not others. For example, Kurtz et al. (1996) compared cancer patients’
family caregivers among three age ranges. They reported that younger caregivers (age < 50)
exhibited the most discrepancy with patients in pain occurrence, while middle-aged caregivers
(age = 50 - 64) were more discrepant about constipation symptoms, and older caregivers (age ≥
65) were more discrepant with nausea. In this regard, caregivers in different age ranges may pay
attention to different patient symptoms.
Caregivers’ sex. Caregivers’ sex also has an unclear contribution to congruence.
Generally, in past studies, female caregivers have been more likely to be congruent with patients,
possibly because women are socialized more than men to focus on caregiving (Berg & Woods,
2009). Gadassi et al. (2011) reported that women’s depressive symptoms were specifically
associated with lower levels of empathic accuracy for negative feelings. They concluded that
depressive symptoms may have a greater impact on interpersonal perception of relationships for
women than for men. Perceiving other people’s feeling is related to individuals’ interpersonal
attributes and skills, such as empathy, sympathy and compassion. Female caregivers may be
more able to perceive others’ emotions and therefore may understand patients better than male
caregivers.
In some studies, female caregivers had a higher likelihood of having congruent symptom
perception with patients than did male caregivers (Kurtz, et al., 1996; Resnizky & Bentur, 2007).
McPherson et al. (2008) also found that being a male caregiver is a major factor in determining
the discrepancy of physical symptoms and psychological symptoms. However, the opposite has

58

been reported in other studies. Higginson and Gao’s (2008) report has shown that female
caregivers have a higher likelihood of noncongruence for patients’ pain symptoms than did male
caregivers. Wennman-Larsen et al. (2007) reported that female caregivers with male patients are
significantly more likely to be noncongruent for symptoms of emotional function. Additionally,
other studies have reported no effect of caregivers’ sex on congruence of symptoms (Silveira, et
al., 2010; Wennman-Larsen, et al., 2007). Overall, studies report inconsistent relationships
between female and male caregivers’ congruence, finding that females can be more congruent
with, but also can overestimate, patients’ symptoms when compared with male caregivers.
Caregiver education. Because caregiver education influences caregivers’ depression and
feelings of burden, it may also be a factor that influences congruence of symptom perception.
Prior studies have shown that caregivers whose reports were not congruent with patients’ had
less education than those who were congruent (Kurtz, et al., 1996; C. Lin, 2001; McPherson, et
al., 2008; Miaskowski, et al., 1997; Resnizky & Bentur, 2007). Silveira et al. (2010) found that
caregivers who overestimated patients’ symptoms were less likely to be high school educated
than those who accurately reported and those who underestimated patients’ symptoms.
Wennman-Larsen et al (2007) found that the caregivers who overestimated cancer patients’
symptoms were less likely to have post-secondary education compared to those who were
congruent or those who underestimated symptoms. Silveira et al. (2010) have pointed out that
education is related to literacy, which can contribute to the abilities of understanding and
communication. Schieman and Turner (2001) stressed that education can improve an individual’s
ability to think, symbolize and communicate, which can improve one’s perception of other
people’s pain. Therefore, it is reasonable that caregivers with higher education levels may have
more accuracy in congruent symptom perception. Less education, using a cutoff point of 12

59

years of education, can predict low congruence that may be related to caregivers’ understanding
of symptoms and symptoms management. The present study will include caregiver education as
a control factor in the EASP model.
Caregivers’ relationship to patients. Stinson and Ickes (1992) tested the differences in
empathic accuracy between male friends and male strangers, and they found that friends have
better empathic accuracy for interpreting the target persons’ response than strangers do. In this
regard, they believed that the perceiver-target relationship is important in empathic accuracy. In
general, most family caregivers of adult cancer patients are spouses or partners. According to the
report from the Family Caregiver Alliance, primary family caregivers, in general, are often
patients’ spouses. In most studies that recruited cancer patients and their family caregivers,
spouse caregivers were the majority of caregiver population (ACS, 2010). Spousal caregivers are
assumed to understand the patients’ behaviors more than nonspousal caregivers. However, no
findings have reported that having a spousal relationship is a significant factor affecting
congruence of symptom perception for caregivers. As we have seen, spousal relationships can be
a factor in caregiver depression and feelings of burden. This study will use caregivers’ social role
with patients as a covariant factor which may have an indirect effect on congruence of symptom
perception.
Living arrangement. Because spousal caregivers and patients are more likely to live
together, it is a common expectation that spousal caregivers have greater congruent levels of
symptom perception than nonspousal caregivers. A few studies have shown that caregivers who
lived with cancer patients were more likely to have congruent reports of patients’ quality of life
or general health (Magaziner, Simonsick, Kashner, & Hebel, 1988; O'Brien & Francis, 1988;
Sneeuw et al., 1997). However, Resnizky and Bentur (2007) identified that living with the

60

patient is a factor for having greater agreement of patient symptoms, when comparing caregivers
who lived with and who did not live with cancer patients, regardless of whether they were
spouses. Living arrangement may be also related to the caregivers’ experience about patients’
behavior and life style. Therefore, the current study will examine living with patient as a
caregiver’s characteristic and test its effect on EASP.
Briefly, based on the findings from previous studies that identified the dyadic factors
influencing congruence, caregiver emotionality and threats of providing care may predict
congruence, and some caregivers’ and patients’ demographics have unclear effects on
congruence. Those findings support the need to examine these factors in the model of congruent
symptom perception in this research. Furthermore, caregivers’ demographic characteristics have
associations with caregivers’ depression and feelings of burden. In this study, these variables will
be controlled in order to detect the main effect of caregivers’ emotionality and threats on
congruence of pain and fatigue severity and frequency between cancer patients and their
caregivers.
In sum, caregiver depression and caregiver burden in the previous studies mainly can be
identified as common negative perceptions of providing care. These caregivers’ negative
perceptions may also influence caregivers’ perception of patients’ symptoms and decrease
congruence of symptom perception. Regarding the findings about congruent symptom perception
in the prior studies, two issues concerning these hypothesized relationships have to be
emphasized for nursing science and nursing practice: (a) prior studies did not use a theoryoriented concept for congruence of symptom perception and (b) the congruent levels of symptom
perception may be determined by a combination of cognitive and affective empathic processes.
First of all, most of the previous studies did not apply theories when illuminating the

61

complexities of congruent symptom perception. One notable exception is Lobchuk (2006), who
used a concept of perspective-taking to describe how caregivers perceive patients’ symptoms.
With respect to the Davis’s empathy model (1994), an empathic process consists of noncognitive
and cognitive components, and these two elements respectively generate affective outcomes,
such as empathic concern and anger, and non-affective outcomes, such as accuracy and
judgments. Thus, Lobchuk’s (2002, 2006) concept of perspective-taking describes empathic
process to approach congruence. The present study will apply Ickes’s concept of empathic
accuracy, in which caregivers will be assumed to be empathic with patients’ symptom
experiences. The present study has selected with caution the theories to support the concept of
empathic accuracy in describing the phenomenon of congruent levels of symptom perception in
caregiving situations.
Lobchuk et al. (2006) conducted studies based on the concept of perspective-taking to
examine how caregivers perceive patients’ symptom perceptions and to identify caregivers’
perceiving processes by various cues. The results in their preliminary study (2006) have shown
that there is no significant difference in caregivers’ accuracy in perceiving patients’ symptoms
after caregivers are guided by different cues. The reason may be because cognitive and affective
empathic processes occur simultaneously. Although Davis’s empathy model categorized two
kinds of empathic processes, cognitive and affective, it is difficult to ask caregivers to choose
one kind of empathic process but ignore the other. Therefore, by using Davis’s model,
caregivers’ empathy can be measured by the outcome congruence or helping behaviors. In
addition, Lobchuk et al. (2006) suggested that improving patient-caregiver communication may
decrease the possibility of discrepant symptom perception. Instead of testing empathic process,
the present study will operate within the understanding of the inseparable empathic processes

62

when identifying how the factors, in terms of caregivers’ emotionality and the perception of
threats, influence congruent levels of symptom perception in caregiving environments.
Methods to Measure Congruent Levels of Symptom Perception Between Cancer Patients
and Family Caregivers
Congruent level of symptom perception describes the agreement between caregivers and
patients when scoring patients’ symptoms and patients’ reports of their symptoms as a reference.
Discrepancy describes situations where caregivers do not rate the patients’ symptoms the same
as the patients’ do. Among the reviewed studies, congruent levels of symptom perception was
identified by subtracting caregivers’ reports from patients’ reports (C. Lin, 2001; Miaskowski, et
al., 1997), and also by using statistics, including reliability analyses and correlation coefficients
(Lobchuk & Degner, 2002; McPherson, et al., 2008; Molassiotis, Zheng, et al., 2010; Silveira, et
al., 2010), to test agreements between patients and caregivers. The following section introduces
the subtraction method and statistical strategies used in previous studies to obtain congruence of
symptom perception between cancer patients and their family caregivers. The present study will
use the subtraction method and the intraclass correlation coefficient (ICC) to evaluate
congruence of pain and fatigue severity and frequency.
Subtraction method. The subtraction method is a simple and convenient strategy to
analyze the distance between patients’ and family caregivers’ scores, such that scores that are
closer to zero indicate greater the congruence of symptom perception. For example, Lobchuk et
al. (1997) and Resnizky and Bentur (2007) reported the numbers and percentages for dyads in
five score distances from -2 to 2 to differentiate dyadic congruency levels for multiple
symptoms. The exact same score, indicating perfect congruence, is represented by the zero
distance.

63

Although zero difference between patients’ and caregivers’ scores (perfect congruence) is
expected, having perfectly congruent perceptions for each of the patients’ symptoms may be an
impossible request for caregivers. Some studies have established congruence with a certain range
to distinguish congruent and discrepant groups (C. Lin, 2001; Miaskowski et al., 1997). C. Lin
(2001) defined congruent perception of cancer pain by a difference of less than 1 between
caregivers’ and patients’ reports on pain now, using an 11-point Likert-type scale on an item
from the Brief Pain Inventory (BPI). The scale ranged from 0 (no pain) to 11 (most severe pain).
Miaskowski et al. (1997) defined congruence as a range of score difference between caregivers’
and patients’ rating (± 10 mm) by using a 100-millimeter visual analog scale of pain severity. In
both studies, when the distance between symptom scores fells in the established range, there was
congruency; when the distances between scores fell outside of the congruent range was
discrepancy. In this regard, it may be practicable to establish a range of congruence at either plus
or minus1 point (C. Lin, 2001) or at plus or minus 10 mm (Miaskowski et al., 1997) for clinical
judgments.
Furthermore, the non-zero score distance between patients and their family caregivers,
and the score difference located outside of the congruent range have been referred as
noncongruence and discrepancy. The discrepancy can fall into two types: overestimation (higher)
and underestimation (lower), to illustrate caregivers’ over patients’ scores respectively (Hauser et
al., 2006; Kurtz, et al., 2006; McMillan & Moody, 2003; McPherson, et al., 2008; Molassiotis,
Zheng, et al., 2010; Silveira, et al., 2010). Therefore, subtraction method can show three
congruent levels: overestimation, congruence, and underestimation by each dyad. Previous
studies often used average score differences of caregivers’ over patients’ scores for
overestimation of cancer patients’ symptoms (McMillan et al. 2003; Molassiotis et al. 2010).

64

However, those studies show a typical congruent symptom perception at group-level. The fact is
that some caregivers can accurately report patients’ symptoms, while others overestimate and
underestimate symptom level. The present study reported the number of dyads in three congruent
levels.
Statistical value. Statistical methods, including the paired t-test, independent t-test, ICC,
Kappa, Pearson’s correlation, Spearmen’s correlation, and Lin’s concordance coefficient (LCC),
can also be used to evaluate agreement in the levels of congruent symptom perception between
family caregivers and cancer patients. In this regard, statistical congruence represents the grouplevel congruence, but not individual dyadic congruence.
Pearson’s correlation, ICC, and LCC are used to test relationships between two or more
raters’ scores. The correlation, which ranges from 0 to 1, represents the similar patterns between
patients’ and caregivers’ scores. Higher correlation indicates a greater similarity in the pattern
between caregivers’ and patients’ scores. ICC interprets agreement based on five levels: 0 to0.2
as poor, 0.3 to 0.4 as fair, 0.5 to 0.6 as moderate, 0.7 to0.8 as strong, and > 0.8 almost perfect
agreement (Cicchetti, 2001; Portney & Watkins, 2000). LCC represents the agreement between
two raters’ scores and uses the Lin’s correlation line passing through the origin and when it has a
slope of 1.0 (L. Lin, Hedayat, & Wu, 2007; L. Lin & Torbeck, 1998). The Kappa statistic shows
the ratio of percentages for two or more raters’ categorized answers for the same questions. High
agreement is represented by a higher ratio of the same answer between caregivers and patients.
In sum, both subtraction and statistics (reliability analyses and correlation coefficients),
are necessarily applied in the congruence studies. Subtraction is a basic strategy to show
differences between patients’ and caregivers’ perceptions of symptom scores. Statistical methods
are advanced strategies used to identify congruent values by comparing entire patterns of

65

symptom scores. For understanding group-level congruence, using subtraction will allow the
researcher to compare mean score differences, and using statistics can identify similar patterns of
symptoms between patients and caregivers. The present study will use subtraction to identify the
individual congruent values of each dyad and test the number of caregivers who overestimate,
congruent, and underestimate patients’ symptoms. This study also will use statistical strategies to
identify the group-value dyadic congruence.

66

CHAPTER 4
Methodology
The purpose of this research was to examine how caregivers’ emotionality and threats of
caregiving impact their empathic accuracy of symptoms perception (EASP) between them and
their cancer patients. The specific aim of this research is to examine how caregivers’
emotionality (caregiver depression) and caregivers’ threats of caregiving (caregiver burden)
predict EASP (congruence of pain and fatigue severity and frequency), when taking into account
contextual factors of caregivers, including caregivers’ age, sex, education, and relationship to
patients, as well as total number of patients’ symptoms that caregivers observe and living
arrangement. These caregiver characteristics can influence caregivers’ emotionality and threats
of caregiving, both of which influence congruence of pain and fatigue severity and frequency.
This study used secondary data analysis with 225 pairs of cancer patients and family
caregivers. Two hypotheses for this study were:
1) Increased caregiver depression and caregiver burden will be related to lower congruent
values of pain and fatigue severity and frequency.
2) Caregivers who are female, older, patients’ spouse, better educated, perceiving a lower
number of patient symptoms, and living with patients will have a greater congruence of pain and
fatigue severity and frequency than the caregivers, who are male, younger non-spouse of
patients, less educated, perceiving a higher number of symptoms, and not living with patients.
Therefore, this study examined the predicted relationship that caregivers’ emotionality
and threats of caregiving impact EASP with and without contextual factors of caregivers’
characteristics. These predictive relationships among three main theoretical concepts,
emotionality, threats of caregivers, and EASP, tested the main parts of the theoretical framework

67

of EASP. This study did not test the effect of empathic accuracy on caregivers’ helping behavior
and quality of life of patients and caregivers.
Study Design
This study was a secondary analysis of a de-identified dataset to examine the relationship
between empathic accuracy of the severity and frequency of pain and fatigue, caregivers’
emotionality, and caregivers’ threats of caregiving as depicted in the conceptual framework of
EASP. The parent study for this project is entitled “The Family Home Care for Cancer—A
Community-Based Model (# R01Ca-79280)”, and was funded by National Institute of Health.
This data was selected because it collected information about the perception of patients’
symptoms from both patients and caregivers, which allowed the current study to examine
congruent pain and fatigue severity and frequency as the operational concept of EASP between
patients and caregivers. The present study tested the factors of caregivers that affect congruent
symptom perception between patients and caregivers, and the relationship of congruence of pain
and fatigue severity and frequency with caregivers’ depression and caregiver burden.
Parent study design. The parent study was a longitudinal experimental study with a
randomized controlled trial that guided patients and caregivers to manage cancer symptoms. The
study was conducted from 1998 to 2003. Consented dyads of patients and their family caregivers
received a baseline interview and were then randomly assigned into two groups to receive
interventions based on the Symptom Management Toolkit. Patients in one group received the
“Patient Intervention for Management of Symptoms and Support (PIMSS)” which was delivered
by a trained cancer nurse; the patients in the other group received the self-management attention
control (SMAC) delivered by a non-nurse coach. The PIMSS was a stepped-approach cognitive
behavioral intervention that taught disease acquired skills to patients who were receiving cancer

68

treatments. The parent study also investigated patients’ behaviors and emotion and also involved
their family caregivers in managing patients’ pain, fatigue, and other commonly occurring cancer
symptoms, including constipation, dyspnea, anorexia, inability to concentrate, dry mouth,
nausea, cough, emotional distress, and insomnia. Briefly, in the PIMSS group, nurses offered
symptom-specific advice about self-care available though a website. The content of the PIMSS
included problem solving and self-care management of the symptoms, the impact on physical
role and social function, and emotional distress. In the SMAC group, the coach did not offer
specific advice, but referred patients to a written symptom management guide containing the
same content as the website. Patients were expected to apply the strategies on their own.
The nurse and coach contacted participants by six phone calls in eight weeks (Weeks 1,
2, 3, 4, 6, & 8). The family caregivers were called at Weeks 1, 4 and 8 to discuss concerning
symptoms and were referred to the symptom management guide for caregiving strategies. The
purpose of contacting family caregivers was to assist them to better manage patients’ problems.
Two follow-up interviews for data collection were conducted at Week 10 and 16 via
phone contact. The content of interviews included patients’ symptom experiences, depressive
symptoms, mastery of care, and health-related quality of life, as well as caregivers’ caregiving
experiences. The duration of participation in the parent study was 16 weeks. Both patients and
caregivers were expected to develop the strategies for communication, obtaining support,
acquiring information, carrying out self-care behaviors, and coping with emotional distress (C.
W. Given et al., 2004).
Design of present study. The present study was a secondary analysis of a de-identified
data set, and was a cross-sectional study. Specifically, the current study only used the data from
dyadic participants’ first interview from the parent study. The parent study was a randomized

69

clinical trial with cognitive-behavioral intervention, and the intervention had been tested to
efficiently improve patients’ cancer symptoms experience (C. W. Given, et al., 2004; Sikorskii et
al., 2007; Sikorskii, Given, Given, Jeon, & McCorkle, 2006). Because the patients in the PIMSS
and the SMAC groups received guidance for symptom management, their symptoms may have
improved at the two follow-up interviews. The patients in the PIMSS also had learned strategies
to communicate with family caregivers about their symptoms. In this regard, the patients at the
two follow-up interviews may have expressed their symptoms more precisely than they did at the
first interview. Therefore, the present study only examined the data of patients and caregivers at
the first interview, in order to avoid the potential effects of the cognitive-behavior intervention
on the patients’ symptom expressions in the second and third interview.
The relationship among caregiver burden, caregiver depressive symptoms, and
congruence of pain and fatigue severity and frequency between cancer patients and their family
caregivers was the primary process being investigated. Congruent levels of symptom perception
was operationally defined as the difference between patients’ scores and caregivers’ score on
patients’ pain and fatigue severity and frequency, in which the scores of pain and fatigue severity
and frequency were measured by using the modified MD Anderson’s symptom inventory. When
the score difference was close to zero indicates similar scores between patients and caregivers,
and congruence of pain and fatigue severity and frequency between patients and caregivers was
high; and vice versa. Furthermore, this study examined the relationships among caregivers’
emotionality, threats and congruence.
Sample
The parent study recruited participating dyads of patients and caregivers from six
healthcare systems in Michigan. Inclusion criteria for the patients in the parent study were: (a)

70

being age 21 or older; (b) having a new diagnosis of solid tumors, such as lung cancer, breast
cancer or non-Hodgkin lymphoma, which have been recurrent or have metastasis; (c) having
underwent the first course of cancer chemotherapy and completed no more than the first two
cycles before the first interview; (d) understand and speak English; (e) being able to
communicate by phone; (f) having experienced at least a two-point score on his or her pain
severity in the last 24 hours; and (g) having a recognized primary family caregiver. Both patients
and their caregivers were at least 21 years old, cognitively intact, English-speaking, able to speak
and hear for telephone interviews, and willing to take part in the study. Patients may have had
new or recurrent diseases. The parent study excluded patients who were diagnosed with
hematological malignancy, received bone marrow or stem cell transplant, and received hospice
care, or had diagnosed emotional or psychological disorders because these patients had different
needs for symptom management from those patients who were diagnosed with solid tumors and
the patients who received cancer treatments (C. W. Given, et al., 2004).
The present study used the same inclusion and exclusion criteria as the parent study. In
addition, this study selected all of the dyadic participants where both the patient and family
caregivers reported severity and frequency of patients’ pain and fatigue in the last seven days at
the first interview in the parent study. The parent study at baseline interview had a total number
of 225 dyadic participants who met these criteria, which was the sample size for the present
study.
Power analysis. Statistical power analysis examines the relationship among the four
elements involved in statistical inference: sample size (N), significance criterion (α), population
effect size (ES), and statistical power (1-β) (Cohen, 1992). Effect size is often used in practice to
estimate a likelihood of having a clinical significant difference between two treatment groups or

71

having any other numerical-comparison contrasts when the study has a specific sample size. A
higher effect size indicates a higher likelihood that the study can have a clinical significant
difference.(Faraone, 2008; Ferguson, 2009; Kelley, 2007). Cohen (1988) suggested that for
regression and correlation, an effect size at least 0.02 is small, at least 0.15 is moderate, and at
least 0.35 is large. The parent study used the results of 15-item symptom severity as a primary
outcome to examine the effect of intervention, and targeted 122 patients per study arm and
detected a similar effect size of approximately 0.35 between the two groups after weeks 10 and
20, with at least 80% power at 5% level of significance (C. W. Given, et al., 2004). Therefore, a
collected sample size of 122 in each study arm of the parent study would be sufficient to show
the significant effect differences between the experimental group and control group.
In the present study, which was conducted as a cross-sectional study with a fixed sample
size of 225, the statistical sample power was examined by following three questions:
1. What sample size is expected for the current study, if the study assumed a moderate
effect size (which is 0.15 and represents 0.36 Pearson’s correlation between dependent
variable and independent variable for simple linear regression modeling), alpha is 0.05,
and sample power is 80% within the 8 predictors (age, sex, education, relationship to
patients, living arrangement, number of observed patient symptoms, depression, and
caregiver burden), and congruence of symptoms will be an outcome variable?
2. How much sample power will result based on the fixed sample size of 225 pairs, if the
study assumed that median effect size is 0.15, alpha is 0.05, and the number of
predictors is seven, and congruency of symptoms will be an outcome variable?
3. How much effect size could this model capture by given the sample size of 225 pairs,
if the study assumed that power is 80%, alpha equals 0.05, and the number of predictor

72

of seven?
The present study operated the G*power software to examine statistical power analysis.
The G*power performs high-precision statistical power analyses for the most common statistical
tests: t-test, F-tests, and chi-square tests in behavioral research. (Erdfelder, Faul, & Buchner,
1996). The G*power has been used in many other disciplines, including biology, genetics,
ecology, pharmacology, and medical research (Faul, Erdfelder, Lang, & Buchner, 2007).
The main research concern in the present study was to detect any correlation between
congruent symptom perception variables. The G*power setting used F-test with linear multiple
regression procedure with null hypothesis of squared multiple correlation coefficient deviated
from zero. Squared multiple correlation coefficient ranges (R2) from 0 to 1 and represents how
much variance of the dependent variable (here congruence of symptom perception) could be
explained by the independent variables (Shieh, 2008). Thus, the deviation of R2 from zero means
that the independent variables can explain 0% of the variance of the dependent variable, so that
the dependent variable cannot be predicted by the independent variables.
For Question 1, the sample size would be at least 55 when this study has standard sample
power as 80%, alpha as 0.05, and effect size as 0.15. For Question 2, the sample power of
predicting a significant factor is 99.9% when the study has the samples size as 225, alpha as
0.05, effect size as 0.15. These two results show that the present study had a sufficient sample
size and sample power when the study has a moderate effect size.
For Question 3, the result by used G*power showed that the desired effect size 0.06 was
minimum, with a sample size as 225, power as 80%, alpha as 0.05, and seven predictors. A
cross-sectional, descriptive study with a fixed sample size and a small effect size, like the current
study, would be less likely to find a significant difference or correlation coefficient among the

73

variables.
Measurements
The present study tested the EASP conceptual framework, which focused on congruent
symptom perception of pain and fatigue severity and frequency. The theoretical concepts in this
study were the caregivers’ emotionality, threats of caregiving, and EASP. Based on two
hypotheses, caregivers’ emotionality and caregivers’ threats of caregiving are assumed to predict
congruence of pain and fatigue severity and frequency between cancer patients and their family
caregivers. Both patients and caregivers’ demographics, caregivers’ depression and burden and
congruent levels of symptom perception will be described for the profile of the dyadic samples
(see Table 1 & Table 2).
Demographic characteristics. Demographic characteristics of cancer patients and
caregivers including patients’ and caregivers’ age, sex, education, caregivers’ relationship to
patients, and living arrangement, which were considered based on the results of reviewed studies.
Age of patients and caregivers was based on the date of birth at the time of collecting data and
was a continuous variable. Sex, Education, Relationship to patients, and Living arrangement
were categorized as binomial variables. Sex was naturally coded in male and female. Caregivers’
education was split into two levels: the caregivers who attended college or some technical
training, and those who had no formal education and who were educated but did not attend
college. The categorical variable ‘relationship to patients’ focusing on caregivers’ role to
patients, but not quality of relationship, was coded 1 for spousal caregivers and 0 for non-spousal
caregivers. Caregivers’ Living arrangement was coded in caregivers living together with patients
and those living in separately. Other patients’ clinical characteristics also included the site of
cancer, stage of cancer, metastasis and recurrence of cancer.

74

When the analysis was processed the test of the general linear modeling for two
hypotheses, only caregivers’ demographic variables were treated as control variables. These
caregiver variables were reported in the previous studies to be associated with caregivers’
emotionality, caregiver burden, and congruence of symptom perception, and included caregivers’
age, sex, education, relationship to the patient, and living arrangement, as well as the total
number of patients’ symptoms that caregivers observed.
Caregiver perception. Caregiver perceptions-caregivers’ emotionality and their
perception of threats, in the theoretical framework have been hypothesized to affect congruent
levels of symptom perception. Emotionality has been operationally defined as caregivers’
depressive symptoms. Threats have been understood to be features related to care demands, for
example, caregiver burden. Caregiver burden has been operationally defined as caregivers’
subjective reactions about how providing patient care impacts them. For the current study, the
CESD) was used to measure caregivers’ depression and the Caregiver Reaction Assessment
(CRA) was used to measure caregiver burden. The modified MD Anderson Symptom Inventory
(MDASI) tested pain and fatigue severity and frequency in order to measure congruence.
Caregiver depression. The CESD (van Dam & Earleywine, 2010) consists of 20 items
which was developed by Radloff (1977) to evaluate the frequency of depressive symptoms
experienced in the preceding week. Each item is scored on a four-point scale (0-3) to represent
the range of frequency of experiencing depressive symptoms in the past week, in which 0
represents not at all or less than one day, 1 is for 1-2 days, and 2 is for 3-4 days, and 3 is for 5-7
days. The total score of the CESD has a range from 0 to 60 with higher scores indicating greater
depressive symptoms. The reliability, validity and factor structure have been tested in the general
population (Radloff, 1977). The inter-item and item-scale correlations were higher in the patient

75

sample (0.9) than in the general population (0.85). Lower test-retest reliability (r < 0.5) may
result from the life events that occurred in the measured time interval; therefore, the CESD is
also sensitive to the occurrence of life events. The validity has been obtained by comparing the
CESD score of the sample patient to the individual’s score on the Positive Affect Scale; the
correlation was highly negative (Radloff, 1977). By comparing the general population with
patients who had a depression diagnosis, researchers have identified the cutoff score of 16 or
higher as the clinical criterion for having depressive symptoms (Radloff, 1977; Schein &
Koenig, 1997). The score of 16 has shown a high sensitivity (100%) for clinical depressive
symptoms in different populations (Dozeman et al., 2010; Li & Hicks, 2010).
The CESD scale has been used to detect caregivers’ depressive symptoms with good
reliability and validity. Cronbach’s alpha was approximately 0.9 in the studies that used the
CESD scale to measure caregivers’ depressive symptoms (Y. Kim, et al., 2005; Sherwood et al.,
2007). Kim, Kashy, and Evans (2007) found that 25% of the caregivers of patients with
colon/rectal cancer diagnosed within the past 2 month had moderate depressive symptoms with
CESD scores of 16 or above (see Table 1).
In this study, the score of CESD represented the caregiver depressive symptoms and was
tested as an independent variable in the congruent model of pain and fatigue severity and
frequency when examining the two hypotheses. Caregivers were also categorized in two groups
by a cut-off point score on the CESD of greater than or equal 16. The categorization allowed the
researcher to examine the differences of caregiver demographic characteristics, hours of
providing symptom management, total number of patient symptoms, caregiver burden,
caregivers’ reports of patients’ pain and fatigue severity and frequency, and congruence of pain
and fatigue severity and frequency.

76

Caregiver burden. The CRA was developed to measure types of caregivers’ responses
and changes in the reaction over time (C. W. Given, et al., 1992). The CRA has 24 items falling
into five dimensions: self-esteem, family abandonment, schedule interruption, health burden, and
financial burden. Self-Esteem subscale (seven items) assesses whether or not caregiving is
enjoyable and rewarding, or whether it cause resentment. Family Abandonment subscale (five
items) measures the extent to which families support and work together with caregivers, and also
asks caregivers’ perception of being abandoned by other family members. Schedule Interruption
subscale (five items) assesses the degree to which caregiving interrupts caregivers’ usual
activities, causes the elimination of some activities, and interferes with relaxation times. Health
Burden subscale (four items) measures caregivers’ physical capability and energy to provide
care. The last subscale, Financial Burden (three items) looks at the adequacy, the difficulty, and
the strain of financial situation on the caregivers and the family (C. W. Given, et al., 1992;
Stommel, Wang, Given, & Given, 1992).
Each item is scored on a 5-point scale, from 1 (strongly agree) to 5 (strongly disagree).
Scores for items in the Self-Esteem subscale need to be reversely coded when a total score of the
CRA is calculated. A higher total score of the CRA refers to higher burden. The inter-scale
correlations have shown a fair independence (0.02- 0.45), which means that each subscale can be
independently used for measurement (C. W. Given, et al., 1992). The test-retest reliability of the
five subscales ranges from 0.62 to 0.86 over 12 months (Persson, Wennman-Larsen, Sundin, &
Gustavsson, 2008). The construct validity was tested by comparing with the patients’ scores for
Activities of Daily Life (ADL) and the CESD. Both the ADL and CESD have been shown to be
negatively correlated to caregiver self-esteem (-0.11 with ADL, and -0.23 with CESD), and
positively correlated to burden of finances, lack of family support, health and schedule

77

interruption (r = 0.34 to 0.57). Researchers have also found the higher correlations between the
ADL and the subscale “schedule interruption” (r= 0.52), and between the CESD scale and Health
Burden subscale (B. Given & Given, 1992; C. W. Given, et al., 1992).
Cronbach’s alpha scores have shown a high degree of reliability for the five subscales
(0.62 to 0.90). The homogeneity of the items in each subscale of the CRA has been found to
represent high internal consistency (C. W. Given, et al., 1992; Nijboer, Triemstra, Tempelaar,
Sanderman, & van den Bos, 1999). Using the CRA in face-to-face interviews has been
demonstrated to take 10 minutes or less (Nijboer, et al. 1999). There have been no studies
reporting the CRA’s sensitivity and specificity, possibly because it may be difficult to find a
comparison group of family caregivers who are not experiencing caregiving stress, which would
be needed to determine a clear cutoff score. The ceiling/floor effects of the CRA scale have also
not been tested, but the 5-point level scale for each item may give the range of the score needed
to decrease the possibility of ceiling/floor effects (see Table 1).
Although the CRA consists of five subscales, each of them can be independently used to
measure theses five characteristics of caregivers’ reactions of caregiving (C. W. Given, et al.,
1992). The present study used only three subscales from the CRA—Schedule Interruption,
Family Abandonment, and Self Esteem, to evaluate caregiver burden (Table 2). Both Schedule
Interruption and Family Abandonment depicted negative caregiver reactions that directly
resulted from care involvement, including symptom management. The Self Esteem subscale
represents positive caregiver reaction that may directly contribute to a caregiver’s perception of
helping patients. These three caregiver reactions may be recognized as soon as caregivers start
dealing with care demands. Sherwood et al. (2005) stated that these three caregiver reaction
subscales were thought to represent the portions of caregiver burden that would be more affected

78

by the exogenous factors, such as the number of caregiving demands.
Although the items of Health Burden subscale and Financial Burden subscale are relevant
to providing care, caregivers’ health and financial issue are not impacted by caregiving
involvement. Caregivers’ health in physical capability and energy in performing caregiving may
also be influenced by their pre-existing health conditions. The Finance Burden subscale
examines whether there is a financial difficulties resulting from providing care for caregivers and
the families, so that financial burden may only influence caregivers and family to pay for
utilizing social supports and to maintain households. Therefore, current study selected three
subscales—Schedule Interruption, Family Abandonment, and Self Esteem, to measure caregiver
burden in three separated domains and did not apply these two subscales in representing
caregiver burden (See Table 1).
Symptom. The MDASI (Cleeland et al., 2000) was modified in the parent study to
measure cancer patients’ symptom severity and frequency, so that the present study used them to
represent the severity and frequency of pain and fatigue in the present study. The parent study
selected the 15 most prevalent symptoms of patients during the time when patients received
active cancer treatments. The MDASI has demonstrated an internal consistency of 0.85. The
construct validity has shown similar factor loadings between two outpatient samples. The
MDASI has been also sensitive to detect high scores of symptom severity and interference for
patients with poor performance and for patients undergoing active treatments (Cleeland, et al.,
2000; Soltow, et al., 2010).
In the parent study, both patients and caregivers report patients’ pain and fatigue severity
and frequency which were measured by using the modified MDASI. Symptom severity was
examined by separately asking the patients and their caregivers to rate patients’ average pain and

79

fatigue severity within the past seven days by using an 11-point scale (0-10), where 10 indicated
the worst pain and fatigue severity. The MDASI was also modified to measure frequency of the
same 15 symptom, as perceived by the patients and caregivers. Symptom frequency examined
how many days that patients felt pain and fatigue in the past week. The frequency was recorded
by using a 8-point scale (0-7), in which 0 meant to no feelings of pain or fatigue on any day, and
7 meant feelings of pain and fatigue happening daily in the last seven days (Appendix A).
Therefore, for severity of both pain and fatigue, greater score indicated the more severe pain and
fatigue. Likewise, for frequency of patients’ pain and fatigue, which ranged from 0 to 7, greater
scores indicated the more frequently feelings of pain and fatigue. The present study tested pain
and fatigue severity and frequency between patients’ and caregivers’ symptom reports to
represent congruence of symptom perception.
Congruence. Congruence of pain and fatigue severity and frequency has been
operationally defined as the congruence of symptom perception, which is a measure of a
caregiving situation in which family caregivers and patients have similar perceptions of patients’
symptoms, specifically focusing on pain and fatigue severity and frequency. In the current
research, congruence of pain and fatigue severity and frequency was obtained by comparing
caregivers’ perception of patients’ pain and fatigue severity and frequency with patients’ selfreports, thereby, examining family caregivers’ accurate perception of patients’ symptoms as a
measure of empathic accuracy. The comparison for congruent symptom perception will be
presented by statistical and mathematical methods (see Table 2).
The mathematical methods for obtaining congruence of pain and fatigue severity and
frequency were to subtract caregivers’ reports of patients’ symptoms from patients’ self-reports,
calculated as the value of the difference between patients’ scores and caregivers’ score on

80

patients’ pain and fatigue severity and frequency, with patients’ reports as reference. When the
value of the difference was close to zero, indicating similar scores between patients and
caregivers, this also indicated high congruence of pain and fatigue severity and frequency
between patients and caregivers; and vice versa.
The symptom score difference via the mathematical approach was recalculated into two
forms to represent the congruent level, in which one was the absolute value, and the other
encompassed three levels of original natural scores of each symptom variable. The absolute
value of the distance between patients’ and caregivers’ score of patients’ pain and fatigue
severity and frequency represent congruent severity ranged from 0 to 10, and congruent
frequency ranged from 0 to 7. An absolute value of the difference close to zero designates
similar scores between patients and caregivers, indicated that the congruent degree of pain and
fatigue severity between patients and caregivers was high. The absolute value of congruent
symptom perception was used to test the two hypotheses.
The original natural score differences of these four symptom variables: pain and fatigue
severity and frequency were categorized into three levels: overestimation, congruence, and
underestimation. Operationally, overestimation describes that caregivers rate a higher score for
patients’ symptoms than patients (the scores differences was greater than 1); underestimation
describes that caregivers rate a lower score for patients’ symptoms than patients (the score
difference was less than -1), and congruence depicts caregivers gave a similar score of symptom
characters as patient do (the score differed in the range from -1 to 1). Therefore, each symptom
characteristic had three congruency categories.
The statistical method measured congruence, which was scored by continuous variables
of pain and fatigue severity and frequency, representing the levels of agreement between patients

81

and caregivers. The agreement was tested by the ICC (see description in the data analysis
section).
Meanwhile, the MDASI was used in the current study to measure the number of patients’
symptoms that the caregivers perceived. There were 15 total symptoms in the MDASI. The
variable for the number of patients’ symptoms was obtained by counting the symptoms that
caregivers had reported in the parent study, creating range of 0 to 15 for the total number
patients’ symptoms that caregivers perceived. However, the inclusion criteria for the dyad
selection in the present study stated that both caregivers and patients must at least report patients’
pain and fatigue severity and frequency, so each dyad had at least 2 points for the number of
symptoms at the first interview (see Table 2).
Human Subjects Protection
The current study obtained two approvals: the approval of Institutional Review Board
(IRB) from Michigan State University, and the approval of dataset utilization from the principle
investigator of the Family Home Care for cancer. Because this data has been de-identified, the
IRB application was deemed by the IRB as not pertaining to Human Subjects research and has
been exempted. The variables were selected based on the conceptual framework (Figure 1).
The current study obtained the approval of utilizing the data set from the principle investigators
of the parent study.

82

Table 1
Instruments Used to Measure Caregiver Depressive Symptoms, Caregiver Burden and Cancer Symptom Perception of Patients and
Caregivers
Measurements

Authors

Subscale

Items

Score

Psychometrics

CESD

Radloff
1977

N/A

20

4 levels: 0-3
0= 0 day,
1= 1-2 days,
2 = 3-4 days, &
3 = 5-7 days

Cronbach’ α =0.9
Inter-item &
Item-scale correlation
= 0.9

CRA

Given, et
al. 1992

Health impact (4)
Family abandonment (5)
Schedule interruption (5)
Financial burden (3)
& Self-esteem (7)

24

1-5 points (strongly
agree to strongly
disagree)

Cronbach’ α 0.62- 0.9 for 5 subscales
Health, α= 0.80
Self-esteem, α= 0.90
Schedule, α= 0.82
Abandonment α= 0.85
Finance, α= 0.81
Construct validity
0.23- 0.57 for 5 subscales

MDASI

Cleeland, Present Severity
19
0-10 (severity)
Internal consistency
et al.
Interference with life
= 0.85
2000
Note: The center for Epidemiologic studies depression (CESD), Caregiver Reaction Assessment (CRA), The MD Anderson Symptom
Inventory (MDASI)

83

Data Management
Merging data files. Because the congruent levels of symptom perception use both
patients’ and caregivers’ information at the same time, these two datasets were merged into one
in order to identify the dyadic sample. After the researchers defined the missing values, the
datasets from the cancer patient side and the family caregiver side in the parent study were
merged by the same subject ID number. Each dyad had the same research ID number for the
patient and his or her family caregiver. The procedure of merging was as follows: (a) sort the
research ID number from small to large for patient and caregiver datasets; (b) based on the same
ID number, the variables of two datasets were merged into one dyad’s information, and (c) the
variables were selected for the present study based on the theoretical framework (see Figure 1).
Additionally, selection of the dyads for the present study was based on the inclusion and
exclusion criteria (see Sample section in this chapter).
Data recode. In order to identify the factors for congruence of pain and fatigue severity
and frequency, this study has to obtain congruent levels of pain and fatigue severity and
frequency between caregivers and cancer patients. At the first step, we merged the patients’ and
caregivers’ information into one data set regarding their same research identification number.
The second step was to reorganize caregivers’ demographic characteristics, including sex,
relationship to patients, education, and living arrangement. We recoded female caregiver as ‘1’
and male caregivers as ‘0’. The original data of variable ‘Relationship to patients’ was coded for
spouse caregivers as ‘1’ and non-spouse caregivers as ‘0.’ Caregivers in the parent study who
completed college education are recoded as ‘1’and other caregiver who had secondary education
or less are recoded as ‘0’. Caregivers’ living arrangement was recorded ‘1’ if they lived with
patients when they were interviewed and ‘0’, if they did not live together. The literature

84

recommended that CESD score of 16 or higher may indicate the potential of having clinical
depressive symptoms. This study also recoded caregivers’ depression score that CESD score is
16 or higher as ‘1’ and CESD score is lower than 16 as ‘0,’ to examine the differences between
two CESD groups in binomial caregiver demographics, total number of patients’ symptoms,
caregiver burden, caregivers’ perception of pain and fatigue severity and frequency, and
congruence of pain and fatigue severity and frequency
Data Analysis
The current study used the conceptual framework in order to understand the predictors of
empathic accuracy of pain and fatigue severity and frequency between cancer patients and their
family caregivers. The Statistical Package for the Social Sciences (SPSS) version 18 (IBM, Inc.
Armonk, NY) was used for descriptive analysis, Pearson’s correlation, independent t-test, chisquare, ANOVA, the generalized linear models (GLM), and multinomial loglinear regression.
Descriptive analysis of sample. The data were examined according to the characteristics
of all variables. The continuous variables (e.g. age and pain severity), were represented by mean,
standard deviation, range, and confidence intervals. The categorical variables (e.g., sex, level of
education, relationship to patient, and cancer site) were shown by the percentage and frequency
(see Table 2).
ICC. ICC was used to identify the overall levels of congruence of pain and fatigue
perception. ICC is a statistical method that has been used to measure the correlation between
pairs of observations that do not have obvious order, and it has also been used to measure the
agreement between assessors, rather than between two methods (White & van den Broek, 2004).
A class of variables refers to metric (scale) and the variance of variables. When researchers are
interested in the relationship among variables of a common class, they can look for the ICC. The

85

Table 2
Variables and Their Measurements Used for Data Analysis
Theoretical domain

Measurement

Variables

Score

Statistical analysis

Control variable
Caregiver age
Caregiver sex
Caregiver education level
Caregiver relationship to patient
Living arrangement
Number of patients’ symptoms

≥ 21
0,1
0.1
0,1
0,1
0-15

Descriptive analysis

CESD

Caregiver depression

0-60

Descriptive analysis

Threat

CRA

Family Abandonment (5)
Schedule Interruption (5)
Self-Esteem (7)

5-25
5-25
7-35

Descriptive analysis

Congruent level
of symptom
perception

MDASI

Pain severity in the last 7 days
Fatigue severity in the last 7
days

0-10

Descriptive analysis
(1) Caregiver’s score - patients’ score
(2) Intraclass correlation coefficient (ICC)

Theoretical variable
Emotionality

Note: A total of 17 items were chosen in the CRA for three subscales: Family Abandonment, Schedule Interruption, and Self-Esteem
(see Appendix A: Caregiver Reaction Assessment Instrument). CESD = The Center for Epidemiologic studies depression; CRA =
Caregiver Reaction Assessment; MDSAI = The MD Anderson Symptom Inventory.

86

ICC represents the ratio of each symptom score that divides the rated variances of each dyad of
patients and caregivers by overall variance (McGraw 1996). Statistically significant results
indicated meaningful congruence between patients and caregivers. The following guidelines in
ICC were used for interpretation of the level of similarity or agreement: 0.81-1.00 is excellent
agreement; 0.61-0.80 is good agreement; 0.41-0.60 is moderate agreement; ≤ 0.40 is poor to fair
agreement (Cicchetti, 2001; Kramer & Feinstein, 1981). The ICC values between patients and
caregivers’ scores by the statistical method represents congruence levels of symptom perception,
which means that a higher ICC values represented higher levels of agreement between patients
and caregivers’ scores, which also indicated higher levels of congruence.
Pearson correlation. Pearson correlation was used to examine the relationship strength
among the variables, in order to choose variables for the tests of the GLMs. The Pearson
correlation is a measure of the strength of the relation between two or more variables. Using
Pearson’s correlation assumes that two variables are measured on at least interval scales, and
determines the extent to which the values of two variables are proportional to each other.
Correlations can range from 1.0 to -1.0 where a value close to 0 represents no relationship
between the variables; the value of -1 refers negative to a correlation, and 1 refers to a positive
correlation. However, the correlation only proves the strength of the relationship between
variables, and cannot show cause and effect (Bewick, Cheek, & Ball, 2003; Hill & Lewicki,
2011). The contextual factors (caregivers’ characteristics) were re-coded to 0 and 1 as dummy
variables in order to score the correlation between congruence and those contextual factors.
GLMs. General linear regression which solves the equation fits a straight line to a
number of points. The line represents the correlation between the dependent variable and the
independent variables corrected by the ratio of standard deviations of these variables. GLMs

87

were used to measure the regression of depressive symptoms and caregiver burden regarding the
caregiver demographics as independent variables.
Loglinear regression. In order to test this study’s hypotheses, the researcher used the
categorized three congruent levels: overestimation, congruence, and underestimation as
dependent variables. Because the dependent variable is categorical, loglinear regression was used
to test the analysis model (see Figure 2). Loglinear regression is a multiple regression, which is
typically used to test models when the dependent variables are categorical or dichotomous. The
factors (independent variables) used in the loglinear regression can be continual variables or
categorical variables. The multinomial loglinear regression identified caregiver characteristics
which predicted congruence/non-congruence of pain and fatigue severity and frequency.
Multinomial logistic regression is used to model nominal outcome variables, in which the log
odds of the outcomes are modeled as a linear combination of the predictor variables. Exp
(parameter) or exp (β) represents the odds ratios for the predictors. They are the exponentiation
of the coefficients.
For hypothesis 1, the dependent variable was three congruent levels of pain and fatigue
frequency and severity, and independent variables are caregiver depression and caregiver burden.
For hypothesis 2, caregiver demographic characteristics were tested to identify whether they are
covariates of caregiver depressive symptoms and caregiver burden the independent variables or
they are also the predictor of these three congruent levels. Caregiver demographics included age,
sex, education level, relationship to patient, living arrangement and the total number of patients’
symptoms that were perceived. These two hypotheses generated eight equations. The first four
equations tested Hypothesis 1 that increasing caregiver depression and caregiver burden can
decrease congruence. The other four equations tested Hypothesis 2 that considering caregivers’

88

demographic characteristics as confounding variables with depression and burden affects
congruence (see Table 3).

89

Contextual factors

Caregiver characteristics:
Age
Sex
Relationship to patient
Education
Living arrangement
Number of patient
symptoms

Predicting factors

Theoretical concept

Emotionality
Caregiver depression

Empathic accuracy of
symptom perception
Congruence of:
Pain severity
Fatigue severity
Pain frequency
Fatigue frequency

Threats of caregiving
Caregiver burden
Hypothesis 1

Hypothesis 2
Figure 2. Analysis model for predicting factors and empathic accuracy in the theoretical framework, Empathic Accuracy of Symptom
Perception. The boxes with dotted line are the two hypotheses.

90

Table 3
Linear Regression in Testing Hypothesis
Equation

Symptom
perception

Congruence of pain severity = a1 + b11 caregiver depression + b12 caregiver burden + b13 caregiver

PS

depression * caregiver burden+ b14 contextual variables + e1
Hypothesis 1

Congruence of pain frequency = a2 + b21 caregiver depression + b22 caregiver burden + b23 caregiver

PF

depression * caregiver burden+ b24 contextual variables + e2
Congruence of fatigue severity = a3 + b31 caregiver depression + b32 caregiver burden + b33 caregiver

FS

depression * caregiver burden+ b34 contextual variables + e3.
Congruence of fatigue frequency = a4 + b41 caregiver depression + b42 caregiver burden + b43 caregiver

FF

depression * caregiver burden+ b44 contextual variables + e4.
Congruence of pain severity = a5 + b51 caregiver depression + b52 caregiver burden + b53 caregiver

PS

Hypothesis 2

depression * caregiver burden + e5
Congruence of pain frequency = a6 + b61 caregiver depression + b62 caregiver burden + b63 caregiver

PF

depression * caregiver burden + e6
Congruence of fatigue severity = a7 + b71 caregiver depression + b72 caregiver burden + b73 caregiver

FS

depression * caregiver burden + e7
Congruence of fatigue frequency = a8 + b81 caregiver depression + b82 caregiver burden + b83 caregiver

FF

depression * caregiver burden + e8
Note. a=incept, b= parameter, e= error, PS = pain severity, FS= fatigue severity, PF= pain frequency, & FF= fatigue frequency

91

Chapter 5
Results
This descriptive secondary data analysis examined how caregivers’ emotionality
(caregiver depressive symptoms) and caregivers’ threats of caregiving (caregiver burden) predict
EASP (congruence of pain and fatigue severity and frequency). Based on the theoretical
framework of EASP, the contextual factors, including caregivers’ age, sex, education,
relationship to patients, the total number of patients’ symptoms that caregivers observe, and
living arrangement were covariants of caregivers’ emotionality and threats of caregiving. The
covariants were hypothesized to influence congruence of pain and fatigue severity and
frequency. In this chapter, the analysis model (Figure 2) will be tested, as will the study’s two
hypotheses: (1) that increased caregiver depressive symptoms and caregiver burden will be
related to lower congruent values of pain and fatigue severity and frequency, and (2) that
caregivers who are female, middle aged or older, a spouse of the cancer patients, have higher
education, perceive lower numbers of patients’ symptoms, and who live with patients will have
greater congruence of pain and fatigue severity and frequency.
Sample Size and Power
The present study only selected 225 dyads of cancer patients and their identified family
caregivers who both reported patients’ pain and fatigue severity and frequency. This sample size
contributed an effect size of 0.06. Effect size refers to the possibility of significant variance
between the groups. According to the criteria of effect size for the regression model provided by
Cohen (1988), the effect size of 0.06 is close to the minimum of at least 0.02. This small effect
size illustrates that a significant relationship among the variables is less likely and that a low
likelihood of identifying a significant relationship is expected.

92

Table 4
Cancer Patients’ Demographic Characteristics (N = 225)
Demographic characteristic
Age
Comorbidity
Average total # of patient symptoms
Age

Sex
Education

Race

Marital status

Employment

Comorbidity

Cancer site

Cancer stage
Recurrent
Metastasis

<45
45 ≤ age <65
≥65
Female
No education or primary school
High school
College
College graduate
Caucasian
African American
Other
Single (never married, widowed,
separated/divorced)
Married & partner
Currently work for pay (Yes
/No)
Retired (if not working for pay)
0 comorbidity
1 comorbidity
At least 2
Breast cancer
Colon
Lung (non-small
/small cell)
Genitourinary
Gastrointestinal
Gynecological
Others
Early
Late
Yes
Yes

93

Mean (SD)
58.35 (11.6)
2.10
(1.6)
8.32
(3.1)
N (%)
23 (10.2)
133 (59.1)
69 (30.6)
134 (59.6)
3
(1.3)
77 (34.3)
107 (47.5)
38 (16.9)
201 (90.1)
17
(7.6)
5
(2.1)
51 (22.7)
173

(77.2)

58
167
100
41
49
135
55
16
54
14
23
16
12
32
18
201
63
139

(25.8)
74.2)
(60.6)
(18.2)
(21.8)
(60.0)
(24.4)
(7.1)
(24.0)
6.2)
(10.2)
(7.1)
(5.3)
(14.2)
(8.0)
(89.3)
(28.0)
(61.8)

Table 5
Caregivers’ Demographic Characteristics (N = 225)
Demographic characteristic

Mean (SD)
54.8
(13.1)
1.7
(1.5)
8.0
(3.1)
10.6
(5.3)
11.9
(7.8)
N (%)
48
(21.3)
125
(55.5)
52
(23.1)
135
(60)
161
(71.6)
19
(8.4)
36
(20)
1
(0.4)
75
(33.3)
111
(49)
38
(16.9)
207
(92.0)
12
(5.3)
4
(2.7)

Age
Comorbidity
Average total # of patient symptoms
Hours of symptom management
Caregiver depressive symptoms (CESD)
Age

<45
45 ≤ age <65
≥65
Sex
Female
Relationship to patient
Spouse
Adult child
Other
Education
No education or primary school
High school
College
College graduate
Race
Caucasian
African American
Other
Marital status
Single (never married, widowed,
36
(16.1)
separated/divorced)
Married & partner
184
(83.7)
Employment
Currently work for pay (Yes)
130
(57.8)
Retired (if not working) for pay
66
(29.3)
Living arrangement
Together
189
(84)
Comorbidity
0 comorbidity
49
(21.8)
1 comorbidity
71
(31.6)
At least 2
105
(46.7)
Hours of symptom
At least 4 hours
88
(52.1)
management
≥ 11 hours
46
(54.7)
Caregiver depressive
≥12
103
(45.8)
symptoms (CESD)
≥16
56
(24.9)
Caregiver burden (CRA) Schedule interruption
2.9
( 0.8)
Self-esteem
4.5
( 0.5)
Family support
1.7
( 0.6)
Note. CESD = The center for Epidemiologic studies depression; CRA = Caregiver Reaction
Assessment

94

Descriptive Analysis of Participants’ Demographic Characteristics
A total of 225 dyadic cancer patients and their caregivers were included (see Tables 4 and
5). This study used descriptive analysis to examine the dyadic participants’ demographic
characteristics, including patients’ and caregivers’ age, sex, education, and comorbidities, hours
caregivers spent providing symptom management, caregivers’ living arrangements, patients’
cancer diagnoses, cancer stages, as well as caregivers’ emotionality (depressive symptoms) and
threats of caregiving (caregiver burden—family abandonment and schedule interruption, and
positive caregiver reaction of self-esteem) in the conceptual framework of EASP (see Tables 4
and 5). In the tables, the continuous and ordinal variables are represented by their means and
standard deviations, while the categorical and nominal variables are represented by their total
numbers and percentages.
Age. Overall, patients were older than their family caregivers. The average age of these
cancer patients was 58.35 (SD ± 11.59), and of caregivers was 54.80 (SD ± 13.14). There were
125 (55.5%) caregivers and 133 (59.1%)patients in the age range of 45 to 65. Participants whose
age was 65 or older were 52 (23.1 %) for caregivers and 69 (30.6%) for all cancer patients.
Sex. This study had 134 (59.6 %) female patients and 135 (60.0 %) female caregivers.
There were 79 (48.0%) female caregivers among the 161(71.0%) spousal caregivers.
Education. More than half of caregivers and patients at least had a college degree. There
were 111 (49.0 %) caregivers and 107 (47.5%) patients having a college degree. Seventy-five
(33.3%) caregivers and 77 (34.3%) patients had only a secondary education.
Race. Among 225 dyads, 207 (92.0 %) caregivers and 201 (90.1%) patients were
Caucasian. There were 12 (5.3%) caregivers and 17 (7.6%) patients who were African American.
The remaining minority categories, included Mexican (Latina), Asian, and Native American,

95

were represented by 4 (2.7%) caregivers, and 5 (2.1%) patients.
Marital status. A majority of patients and caregivers were married. Fifty-one (22.7%)
patients and 36 (16.1%) caregivers were single (never married, widowed, separated, and
divorced).
Employment. There were 58 (25.8%) employed cancer patients. Among 167
unemployed cancer patients, 100 of them (44.4% of 225 cancer patients and 60.6 % of 167
unemployed cancer patients) were retired. Among 225 family caregivers, 130 (57.8%) were
employed and 95 (42.2%) were unemployed at the time of interview. There were 66 unemployed
caregivers (29.3% of 225 caregivers, 69.5% of 95 unemployed caregivers) that were retired.
Living arrangement. There were 189 (84%) of family caregivers living with patients at
the time of interview. All spousal caregivers (N = 161) lived with patients (not shown in the
table).
Cancer diagnosis. Among 225 cancer patients, 68 (30.2 %) were diagnosed with lung
cancer, which included 54 (24%) patients diagnosed with non-small cell lung cancer and 14
(6.2%) patients diagnosed with small cell lung cancer. Fifty-five (24.4%) patients were
diagnosed with breast cancer. Thirty-nine (14.2%) patients had cancer related to the
gastrointestinal system (colon and other gastrointestinal organs cancer). There were 201 (89.3%)
patients at the late stage of cancer when received the cancer diagnosis, which is greater than
those who were at the early- staged cancer. Sixty three (28%) patients were recurrent and 139
(61.8%) patients were metastatic at the time of diagnosis.
Comorbidity. The average number of comorbidities reported by patients’ was 2.1 (SD ±
1.61) and by caregivers was 1.71 (SD ± 1.52). The number of caregivers who reported no
comorbidity was 49 (21.8%), who had one comorbidity was 71 (31.6%), and who had at least

96

two comorbidities was 105 (46.7%). All comorbidites were reported for the time they took care
of cancer patients. More than half of the patients (n = 135, 60.0%) reported that they had at least
two comorbidities at the time of interview.
Hours of providing symptom management. The hours of providing symptom
management measured how many hours that caregivers spent providing symptom management
for patients in the week before the interview. The average hours of symptom management in
prior week was 10.61 (SD ± 15.27) hours. Eighty-nine percent of family caregivers who spent at
least 4 hours in the last week providing symptom management, and 46 % caregivers spent more
than 11 hours in the last week providing symptom management.
Caregivers’ emotionality. The investigator used the CESD, which has 20 items, to
examine the individual depressive symptoms. The total score ranged from 0 for no depressive
symptoms to 60 for the worst depression. A cut-point score of 16 was used to determine whether
or not a person had clinical depressive symptoms (CESD ≥ 16). An average CESD score for
caregiver depressive symptoms was 11.89 (SD ± 7.83). One-hundred-three (22.3%) of 225
caregivers reported their depressive symptoms equal to or greater than 12, while 46 (54.7%) of
them reported their depression scores equal to or greater than 16.
Caregivers’ threats of caregiving. Caregivers’ threats of caregiving were represented
via caregiver burden. Caregiver burden was reported by using three subscales of the CRA:
family abandonment, schedule interruption, and self-esteem. Each item is scored by using a 5point scale ranging from 1 (strongly disagree) to 5 (strongly agree). Schedule interruption and
Family abandonment are used to evaluate caregivers for their negative responding to the
caregiving situations. Both have 5 items, so that each has a score range from 5 to 25.

97

Table 6
Differences in Caregivers’ Depression, Feelings of Burden, Age, Hours of Symptom Management, and Total Number of Patient
Symptoms by Sex, Education, Relationship to Patient, and Living Arrangement (t test)
Caregiver Characteristic
Age
Depression
Schedule interruption
Family abandonment
Self-esteem
Total number of symptoms
Hours of providing symptom management
Age
Depression
Schedule interruption
Family abandonment
Self-esteem
Total number of symptoms
Hours of providing symptom management
Age
Depression
Schedule interruption
Family abandonment
Self-esteem
Total number of symptoms
Hours of providing symptom management
Age
Depression

Mean (SD)
Mean (SD)
Male (N = 90, 40%)
Female (N = 135, 60%)
55.63 (12.76)
54.24 (13.39)
10.95
(7.48)
12.52
(8.02)
2.94
(0.68)
2.80
(0.84)
1.70
(0.62)
1.72
(0.65)
4.54
(0.41)
4.47
(0.49)
7.92
(3.07)
8.11
(3.16)
11.70 (16.07)
9.84 (14.71)
≤ High school (N = 76, 33.8%)
≥ College (N = 149, 66.2%)
58.17 (12.06)
53.08 (13.36)
13.48
(8.64)
11.08
(7.28)
2.80
(0.82)
2.88
(0.76)
1.69
(0.65)
1.72
(0.64)
4.46
(0.54)
4.52
(0.42)
7.96
(3.33)
8.07
(3.02)
17.10 (21.25)
7.88 (10.91)
Non-spouse (N = 55, 28.4%)
Spouse (N = 161, 71.6%)
46.12 (14.11)
58.03 (11.02)
14.29
(9.73)
11.13
(6.97)
2.79
(0.82)
2.88
(0.76)
2.03
(0.68)
1.58
(0.57)
4.45
(0.58)
4.52
(0.41)
8.45
(3.51)
7.91
(3.00)
10.90 (14.06)
10.96 (16.02)
Do not live together (N = 29, 16%) Live together (N = 189, 84%)
43.21 (13.64)
56.82 (12.04)
15.44
(9.36)
11.45
(7.50)

98

df

t

223
223
223
223
223
223
167

-0.78
-1.47
-1.31
0.19
-1.09
0.44
-0.78

223
223
223
223
223
223
167

-2.79 *
-2.20 *
0.78
0.33
0.86
0.25
-3.71 *

214
214
214
214
214
214
160

6.42 *
2.60
0.77
4.74 *
1.00
11.11
-0.19

216
216

5.57 *
-2.57

Table 6 (cont’d)
Caregiver Characteristic
Schedule interruption
Family abandonment
Self-esteem
Total number of symptoms
Hours of providing symptom management
*p < .05

Mean (SD)
Male (N = 90, 40%)
2.76
(0.84)
2.07
(0.68)
4.44
(0.53)
8.03
(3.57)
9.10
(7.53)

99

Mean (SD)
Female (N = 135, 60%)
2.87
(0.76)
1.64
(0.61)
4.51
(0.45)
8.06
(3.08)
11.16 (16.26)

df
216
216
216
216
161

t
0.72
-3.51 *
0.75
0.04
0.56

A higher score resulting from the scales for schedule interruption and family abandonment
indicates a higher level of caregiver burden. Self-esteem is a subscale examining positive
reactions of caregivers. It consists of 7 items and a score range from 7 to 35, in which a higher
score refers to a higher level of self-esteem. In this study, an average score for schedule
interruption was 2.86 (SD ± 0.78), family abandonment was 1.71 (SD ± 0.64), and self-esteem
was 4.50 (SD ± 0.46).
Differences in Caregiver Scores by Their Demographics
By using the independent t test, the current study tested the differences of caregivers’ age,
depression, three types of caregiver burden, hours of providing symptom management, and the
total number of patient symptom regarding caregivers’ binomial demographic characteristics: sex
(female vs. male), education (< college vs. ≥college), spouse (spouse vs. non-spouse), and living
arrangement (live together with patients vs. live separately). See Table 6 for all results.
Sex. There were no significant differences between female and male caregivers in their
ages, depressive symptoms, caregiver burden scores, total number of patient symptoms, and
hours of symptom management between (see Table 6). Male caregivers reported a higher score
of schedule interruption (2.94 ± 0.68) than female caregivers (2.80 ± 0.84). Male caregivers also
spent more hours providing symptom management (11.70 ± 16.07) than female caregivers (9.84
± 14.71). The average age of female caregivers (54.24 ± 13.39) was younger than that of male
caregivers (55.63 ± 12.76). Female caregivers in this study reported higher levels of depression
(12.52 ± 8.02) and family abandonment (1.72 ± 0.65), but lower scores of self-esteem (4.47
±0.49) than males. These female caregivers also recognized a greater number of patients’
symptoms (8.11 ± 3.16) and spent fewer hours (9.84 ± 14.71) in providing symptom
management as compared to male caregivers in the last seven days.

100

Education. Caregivers who did not have college degrees were younger and reported
significantly higher scores of depressive symptoms than the caregivers who had college or
advanced degrees. The former caregivers spend significantly more hours for symptom
management than the latter (see Table 6). In comparisons between two education groups of
caregivers, there were no significant differences in caregiver reactions and the total number of
patient symptoms. The caregivers without college degrees had lower scores of schedule
interruption, family abandonment, and self-esteem than the caregivers who had higher education
degrees. They also reported a lower number of patient symptoms.
Relationship to patient. Spouse caregivers, compared to non-spouse caregivers, were
significantly older and reported lower scores of depressive symptoms and family abandonment
(Table 6). Spouse caregivers also reported higher scores of schedule interruption and self-esteem.
The spouse caregivers spent more time providing symptom management for patients, although
they recognized fewer symptoms. Spouse caregivers may feel more obligated to patient health
than non-spouse caregivers; therefore, they may spend more time providing symptom
management and report higher scores on schedule interruption than non-spouse caregivers.
Meanwhile, spouse caregivers have higher scores on self-esteem, which could imply that taking
care of patients is valued as an accomplishment for them.
Living arrangement. Caregivers who were living with patients spent significantly fewer
hours supporting patients with symptom management than the caregivers who did not live with
patients (Table 6). The caregivers living with patients also were older and reported lower scores
of depression and family abandonment, but had higher scores of schedule interruption and selfesteem. All spouse caregivers in the current study lived with their patients. For caregivers living
with patients, one can infer that their lives revolve around their patients; therefore, it is possible

101

that they spend more time in providing symptom management and thus have higher scores of
schedule interruption.
Description of Congruent Levels of Pain and Fatigue Severity and Frequency
In this study, congruence of symptom perception was represented by comparing the score
differences of pain and fatigue severity and frequency between patients and caregivers. Pain and
fatigue severity was measured on an 11-point scale (0-10), and pain and fatigue frequency was
measured on an 8-point scale (0-7). Both patients and caregivers had to score the severity of
patient pain and fatigue in the last seven days and list the number of days in last seven that
patients felt pain and fatigue. Overall, both patients’ and caregivers’ scores on pain severity and
frequency were lower than fatigue severity and frequency. Patient-reported scores of pain
severity (mean = 2.74, SD ± 2.69) and pain frequency (mean = 2.89, SD ± 2.85), were lower than
caregivers’ ratings for pain severity (mean = 3.41, SD ± 2.89) and pain frequency (mean = 3.33,
SD ± 2.96). Patients’ reports on their fatigue severity (mean = 5.15, SD ± 2.47) and fatigue
frequency (mean = 5.10, SD ± 2.23) were higher than their own reports on pain severity and
frequency. Patients’ scores on fatigue severity and frequency were lower than caregivers’ ratings
of fatigue severity (mean = 5.68, SD ± 2.35) and fatigue frequency (mean = 5.47, SD ± 2.14).
Therefore, average of scores of pain and fatigue severity and frequency from caregivers were
higher than those from patients. Regarding the MDASI used to measure 15 patient symptoms,
the average total number of patient symptoms was eight for both patients’ (mean = 8.32, SD ±
3.12) and caregivers’ (mean = 8.03, SD ± 3.09) repots (Table 4 and 5). Both patients and
caregivers reported the same number of patient symptoms
ICC. The ICC was used to examine the agreement (congruence) of pain and fatigue
severity and frequency between cancer patients and their family caregivers. The ICC is used to

102

measure the reliability of measurements or ratings by identifying the similarity between two
raters by testing correlation strength between the two raters’ scores, which determined the level
of agreement (congruence). The higher the value of the ICC, the greater was the congruence.
ICC criteria range from poor (< 0.2), fair (0.21 - 0.40), moderate (0.41 - 0.60), substantial (0.61 0.80), to excellent or perfect (0.81 - 1.0). ICC will approach 1.0 when there is no variance within
the targets (Cicchetti, 2001; Kramer & Feinstein, 1981).
Four models of congruence of pain and fatigue severity and frequency had ICCs lower
than 0.5. The highest congruence level was pain severity (ICC = 0.46), and the lowest was
fatigue frequency (ICC = 0.26). Furthermore, the congruent level of pain severity and frequency
was greater than the level of fatigue severity and frequency (Table 7). According to the ICC
criteria above, this study had fair to moderate levels of ICC for four congruencies. Members of
cancer dyads may pay more attention to pain, because pain may be more observable than fatigue.
The congruence of perceiving pain severity and frequency, therefore, is better than that of
fatigue.
Subtraction. The score differences of pain and fatigue severity and frequency between
cancer patients and their family caregivers was calculated by subtracting patients’ scores from
caregivers’. The original value of score differences ranged from -10 to 10 for severity and from 7 to 7 for frequency. To calculate the absolute value, the original value was re-coded so that the
negative values became as positive. For example, -10 was re-coded as 10, and -5 was re-coded as
5. By using the original value, the score closest to the center (zero) indicated better congruence.
Using the absolute value of congruencies, the smaller score indicated higher congruence.
Therefore, 0 score difference of both original and absolute values indicates perfect congruence.
Among these 225 dyads, there were 60 (27%) cancer dyads in perfect congruence on pain

103

Table 7
Means with SD and ICCs for Examining CongruentLlevels of Pain Severity, Fatigue Severity,
Pain Frequency, and Fatigue Frequency Between Caregivers And Patients
Measurement
Total number of dyads
Patient score
Caregiver score
Intraclass Correlation
Coefficient (ICC)
*p < 0.05,

Pain severity
Mean (SD)
N= 222

Fatigue severity
Mean (SD)
225

Pain frequency
Mean (SD)
224

Fatigue frequency
Mean (SD)
225

2.74 (2.69)
3.41 (2.89)

5.15 (2.47)
5.68 (2.35)

2.89 (2.85)
3.33 (2.96)

5.10 (2.23)
5.47 (2.14)

0.46*

0.38*

0.43*

0.26*

104

Table 8
The Substation for Examining Congruent Levels of Pain Severity, Fatigue Severity, Pain
Frequency, and Fatigue Frequency Between Caregivers and Patients (N = 225)
Measurement

Fatigue severity
(-10 ~10)
N
225

Pain frequency
(-7~7)
N
224

Fatigue frequency
(-7~7)
N
225

Mean (SD)

Total number of dyads

Pain severity
(-10 ~10)
N
222

Mean (SD)

Mean (SD)

Mean (SD)

Original value of
score difference

0.68 (3.32)

0.53 (2.98)

0.43 (3.50)

0.37 (2.84)

Absolute value of
score difference

2.51 (2.28)

2.32 (1.94)

2.50 (2.48)

1.97 (2.07)

N (%)
50 (22.5)
87 (39.2)
85 (38.3)

N (%)
49 (21.8)
93 (41.3)
83 (36.9)

N (%)
53 (23.7)
100 (44.6)
71 (31.7)

N (%)
44 (19.5)
110 (48.9)
71 (31.6)

Underestimate
Congruence
Overestimate

Perfect congruence
60 (27.0)
36 (16.0)
76 (33.9)
83 (36.9)
(0 score difference)
1 score difference
27 (12.2)
57 (25.3)
24 (10.7)
27 (12.0)
2 score difference
31 (14.0)
44 (19.6)
28 (12.5)
38 (16.9)
>2 score difference
105 (46.8)
88 (39.1)
97 (42.9)
67 (34.2)
Note. Score difference = caregivers-patients; Underestimate = score difference < -1; Congruence
= score difference -1~1; Overestimate = score difference >1.
*p < 0.05

105

severity, 36 (16%) in congruence on fatigue severity, 76 (33.9%) in congruence on pain
frequency, and 83 (36.9%) in congruence on fatigue frequency (Table 8).
The original value of score difference of pain and fatigue severity and frequency was
operationally counted in three congruent levels: underestimation, congruence, and
overestimation so that each symptom characteristic had three congruent categories.
Underestimation was a score difference less than -1, and overestimation was a score difference
greater than 1, while congruence was a score difference ranging from -1 to 1. Table 7 shows that
there are more cancer dyads that were categorized in the congruence group for four symptom
variables than there were in the two noncongruence groups. Forty to 50% cancer dyads were
congruent for those four symptom variables, whereas 30 to 40% of caregivers overestimated
these symptoms. The findings about pain and fatigue severity and frequency are described
below:
1. The number of caregivers who were congruent or overestimated the patients’ symptoms
(score difference ≥ -1) was greater than the number of caregivers who underestimated
patient symptoms (score difference < -1).
2. Among three congruent levels of pain and fatigue severity and frequency, the group of
congruence (score difference -1~ 1) had the greatest number of cancer dyads, and the group
of underestimation (score difference < -1) had the fewest.
3. In comparing the number of cancer dyads in three congruent groups for pain and fatigue,
the percentage of cancer dyads in the group of congruence (-1~1) of fatigue severity and
frequency was greater than that of pain severity and frequency. Fatigue, however, had a
lower ICC level than pain.
4. In comparing the number of cancer dyads for severity and frequency among three

106

congruent levels (underestimation, congruence, and overestimation), the number of cancer
dyads in the congruence on frequency of both pain and fatigue was greater than the number
in the congruence on severity of both pain and fatigue. Likewise, the groups of
overestimation of pain and fatigue severity had a greater number of cancer dyads than the
groups of overestimation of pain and fatigue frequency.
In summary, the ICC levels and subtraction showed the different results in determining
which symptom variables had greater congruence. The ICC showed that congruence of pain is
higher than that of fatigue, while subtraction showed that congruence on fatigue had a greater
number of dyads than did congruence on pain. Subtraction showed congruence by counting the
score differences of each dyad as individuals, while the ICC operated on data structured as
groups, not individuals. In this study, ICC presented the congruent levels of symptoms of entire
samples, and subtraction showed the sizes of caregiver-patient score differences for each
symptom characteristic.
The distributions of congruence of pain and fatigue severity and frequency showed that
cancer dyads with zero score difference were greater than the dyads without zero score
differences. This study used three congruent levels (underestimation, congruence, and
overestimation) of four symptom variables (pain severity, fatigue severity, pain frequency, and
fatigue frequency) as dependent variables. The multinomial log-linear regression was used to test
the factors of caregivers for three congruencies of four symptoms variables.
Differences of Caregiver Characteristics by Three Congruent Levels of Pain and Fatigue
Severity and Frequency
A Chi-square and an ANOVA was used to examine the differences in caregivers’
characteristics among these three congruent levels of pain and fatigue frequency and severity

107

(Table 9 and 10). Overall, caregivers’ sex, education, relationship to patient (spouse or nonspouse), and living arrangement were not significantly different among the three congruent
levels (see Table 9). However, other caregiver characteristics age, depressive symptoms, three
caregiver reactions, perceiving total number of patient symptoms, and hours of providing
symptoms management were significantly different among three congruent levels of four
symptom variables.
Congruence of pain severity. Comparing the three congruent levels of pain severity (see
Table 10), significant differences between congruence and overestimation were found in selfesteem, family abandonment, and total number of patient symptoms; the only significant
difference between underestimation and overestimation was found in total number of patient
symptoms (see Table 10). Caregivers in the congruent group were the youngest and reported the
highest score of self-esteem, and they also reported the lowest scores on the CESD, schedule
interruption, family abandonment, the total number of patient symptoms, and hours of providing
symptom management, when compared to the other two noncongruent groups. Caregivers in the
overestimation group were the oldest and reported the greatest number of patient symptoms and
most hours providing symptom management.
Congruence of fatigue severity. Comparing the three congruent levels of fatigue
severity, there was only a significant difference between the underestimation group and the
overestimation group in the total number of patient symptoms, in which caregivers in the
overestimation group reported higher scores on patients’ total symptoms than did the caregivers
in the underestimation group (see Table 10). The caregivers in the congruence group also
reported the highest scores on self-esteem and the greatest number of hours providing symptom
management. Caregivers in the overestimation group for fatigue severity were the oldest and

108

reported the highest scores of family abandonment, schedule interruption, and total number of
patient symptoms. Caregivers in the underestimation group for fatigue severity reported the
highest scores of depressive symptoms.
Congruence of pain frequency. Comparing the three congruent levels of pain
frequency, a significant difference was found only in caregivers’ family abandonment between
congruence and underestimation groups, in which caregivers in the congruence group reported a
higher average score of family abandonment than those in the underestimation group (see Table
10). Caregivers in the congruence group were older and reported higher scores of CESD,
schedule interruption, self-esteem, and family abandonment. They also spent the most hours
providing symptom management. Caregivers in the overestimation group reported the greatest
number of patient symptoms.
Congruence of fatigue frequency. Comparing the three congruent levels of fatigue
frequency, the only significant difference was the total number of patient symptoms between
congruence and underestimation, in which caregivers in the congruence group of fatigue
frequency perceived a greater number of patient symptoms than did the caregivers in the
underestimation group (see Table 10). Caregivers in the overestimation group were the oldest and
reported the highest scores of self-esteem. Caregivers in the congruence group had higher scores
of CESD, schedule interruption, and family abandonment, and they reported more hours of
providing symptom management than did the caregivers in the other two groups.

109

Table 9
Differences Among Three groups—Underestimation, Congruence, and Overestimation in Testing
Characteristics (binomial variables) of Caregiver (Chi-Square)

Caregiver characteristic
Pain severity
Sex

Female
Male
Education
≥ College
≤ High school
Relationship
Spouse
to patient
Nonspouse
Living
Together
arrangement
Not together

Underestimation Congruence Overestimation
N (%)
N (%)
N (%)
df bg/wg
50 (22.5)
87 (39.6)
85 (38.3)
33 (66.0)
50 (57.5)
50 (58.8)
2/222
17 (19.1)
37 (42.5)
35 (42.1)
33 (66.0)
60 (69.0)
55 (64.7)
2/222
17 (34.0)
27 (31.0)
30 (35.3)
33 (70.2)
63 (73.3)
63 (78.8)
2/222
14 (29.8)
23 (26.7)
17 (21.3)
39 (81.3)
71 (82.6)
76 (93.8)
2/222
9 (18.8)
15 (17.4)
5 ( 6.2)
49 (21.8)
27 (55.1)
22 (44.9)
29 (59.2)
20 (40.8)
33 (70.2)
14 (29.8)
41 (85.4)
7 (14.6)

93 (41.3)
62 (66.7)
31 (33.3)
67 (72.0)
26 (28.0)
64 (71.1)
26 (28.9)
76 (83.5)
15 (16.5)

83 (36.9)
46 (55.4)
37 (44.6)
53 (63.9)
30 (36.1)
64 (81.0)
15 (19.0)
72 (91.1)
7 ( 8.9)

Female
Male
Education
≥ College
≤ High school
Relationship
Spouse
to patients
Nonspouse
Living
Together
arrangement
Not together

53 (23.7)
33 (62.3)
20 (37.7)
37 (69.8)
16 (30.2)
35 (67.3)
17 (32.7)
41 (78.8)
11 (21.2)

100 (44.6)
60 (60.0)
40 (40.0)
64 (64.0)
36 (36.0)
73 (76.0)
23 (24.0)
86 (88.7)
11 (11.3)

71 (31.7)
41 (57.7)
30 (42.3)
48 (67.6)
23 (32.4)
53 (79.1)
14 (20.9)
61 (89.7)
7 (10.3)

Fatigue frequency
Sex
Female
Male
Education
≥ College
≤ High school
Relationship
Spouse
to patients
Nonspouse

44 (19.6)
26 (59.1)
18 (40.9)
29 (65.9)
15 (34.1)
29 (69.0)
13 (31.0)

110 (48.9)
70 (63.6)
40 (36.4)
73 (66.4)
37 (33.6)
76 (71.0)
31 (29.0)

71 (31.5)
39 (54.9)
32 (45.1)
47 (66.2)
24 (33.8)
56 (83.6)
11 (16.4)

Fatigue severity
Sex

Female
Male
Education
≥ College
≤ High school
Relationship
Spouse
to patient
Nonspouse
Living
Together
arrangement
Not together

Pain frequency
Sex

110

2

X

1.03
0.36
1.29
6.00

2/222

2.94

2/222

2.70

2/222

2.76

2/222

2.22

2/222

0.26

2/222

0.58

2/222

2.29

2/222

3.62

2/222

1.38

2/222

0.00

2/222

4.25

Table 9 (cont’d)
Underestimation Congruence Overestimation
2
Caregiver characteristic
N (%)
N (%)
N (%)
df bg/wg
X
Together
36 (85.7)
92 (84.4)
61 (91.0)
2/222
1.63
Living
Not together
6 (14.3)
17 (15.6)
6 ( 9.0)
arrangement
Note. df = degree of freedom; bg/wg = between group/within group.
* p< 0.05,

111

Table 10
Differences Among Three Groups—Underestimation, Congruence, and Overestimation-in
Testing Characteristics (continual variables) of Caregiver (ANOVA)
Underestimation
Characteristic
Pain severity
Age
CESD
Schedule interruption
a
Self-esteem
Family abandonment

a

a,b

Total # of symptoms
Hours of providing
symptom management
Fatigue severity
Age
CESD
Schedule interruption
Self-esteem
Family abandonment
b
Total # of symptoms
Hours of symptom
management
Pain frequency
Age
CESD
Schedule interruptionc
Self-esteem
Family abandonment
Total # of symptoms
Hours of symptom
management
Fatigue frequency
Age
CESD
Schedule interruption
Self-esteem
Family abandonment

Congruence

Overestimation

Mean (SD)

Mean (SD)

Mean (SD)

55.30 (14.36)
13.70 ( 9.19)
2.88 ( 0.84)

55.48 (12.97)
10.50 ( 6.52)
2.84 ( 0.76)

53.88 (12.59)
12.23 ( 8.14)
2.85 ( 0.78)

4.50 ( 0.47)

4.58 ( 0.42)

1.77 ( 0.66)

1.56 ( 0.62)

df bg/wg

F
0.36
2.78
0.03

4.40 ( 0.48)

2/219
2/219
2/219
2/219

1.83 ( 0.63)

2/219

4.08*
8.59*

3.16*

7.68 ( 3.31)

7.19 ( 3.24)

9.06 ( 2.57)

2/219

11.02 (15.09)

10.03 (16.67)

11.08 (14.35)

2/219

0.89

53.93 (13.86)
12.30 ( 8.34)
2.80 ( 0.67)
4.46 ( 0.56)
1.68 ( 0.55)

54.07 (13.14)
11.45 ( 7.32)
2.80 ( 0.83)
4.51 ( 0.44)
1.66 ( 0.63)

56.12(12.74)
12.14 ( 8.13)
2.95 ( 0.79)
4.49 ( 0.43)
1.78 ( 0.70)

0.66
0.25
0.94
0.19
0.93

7.00 ( 5.27)

7.98 ( 2.99)

8.71 ( 3.05)

2/222
2/222
2/222
2/222
2/222
2/222

6.80 ( 7.33)

12.03 (18.48)

10.79 (13.73)

2/222

1.29

55.82 (
12.72 (
2.99 (
4.54 (
1.73 (
7.95 (

2.08)
7.91)
0.76)
0.42)
0.70)
3.12)

53.70 (13.64)
12.15 ( 8.22)
2.83 ( 0.80)
4.43 ( 0.52)
1.72 ( 0.62)
8.53 ( 2.20)

2/221
2/221
2/221
2/221
2/221
2/221

0.61
2.12
0.18
1.10
4.01*
1.55

9.10 (16.89)

10.61 ( 7.61)

8.53 (10.06)

2/221

0.76

52.63 (11.26)
10.20 ( 5.90)
2.71 ( 0.75)
4.45 ( 0.46)
1.70 ( 0.67)

54.42 (14.03)
12.85 ( 8.30)
2.91 ( 0.84)
4.49 ( 0.46)
1.75 ( 0.63)

56.72 (13.13)
11.45 ( 8.00)
2.86 ( 0.71)
4.53 ( 0.46)
1.65 ( 0.64)

2/222
2/222
2/222
2/222
2/222

1.40
1.98
1.06
0.37
0.56

54.12 (14.42)
10.02 ( 7.00)
2.62 ( 0.75)
4.50 ( 0.47)
1.66 ( 0.55)
7.56 ( 3.33)

112

4.81*

Table 10 (cont’d)
Underestimation

Overestimation

Mean (SD)

Characteristic

Congruence
Mean (SD)

Mean (SD)

c

F
df bg/wg
2/222
3.43*

Total # of symptoms
7.18 ( 3.79)
8.54 ( 3.13)
7.77 ( 3.46)
Hours of symptom
6.16 ( 9.06)
12.74 (17.57)
9.87 (13.88) 2/222
2.16
management
Note. df = degree of freedom; bg = between group; wg = within group
a
b
c
Congruence versus Overestimation. Underestimation versus Overestimation. Underestimation
versus Congruence.
* p< 0.05.

113

In summary, most caregivers in the congruence group of these four symptom variables
had a higher average score of self-esteem and spent more hours on symptom management than
did the caregivers in the other two noncongruence groups. Caregivers in congruence of pain
severity and fatigue severity reported the fewest depression symptoms, whereas caregivers in the
congruence groups of pain and fatigue frequency had the most depressive symptoms. Caregivers
who were categorized in the overestimation group for pain severity and fatigue severity reported
higher levels of family abandonment and greater numbers of patient symptoms. Caregivers in the
underestimation groups of pain frequency and fatigue frequency tended to have lower scores on
CESD, schedule interruption, family abandonment, and total number of patient symptoms.
Results of Testing Two Research Hypotheses
This study had two hypotheses:
1) Increased caregiver depressive symptoms and caregiver burden will be related to
lower congruent values of pain and fatigue severity and frequency.
2) Caregivers who are female, older, patients’ spouses, better educated, perceiving a
lower number of patient symptoms, and living with patients will have greater
congruence of pain and fatigue severity and frequency.
The dependent variables were four congruencies of pain severity, pain frequency, fatigue
severity, and fatigue frequency. Independent variables included caregiver depressive symptoms
(CESD), caregiver burden of schedule interruption, family abandonment, self-esteem (CRA), and
six caregiver demographics: age, sex, education, relationship to patients, living arrangement, and
total number of patient symptoms. We used the Independent t test (t) and Pearson’s correlation
(r) to identify the significant relationships among the independent variables. The GLM was used
to determine the factors of caregiver depressive symptoms and caregiver burden for the four

114

regression models of congruence of symptom perception. Multinomial log-linear regression was
employed to test the model of congruent pain and fatigue severity.
Pearson’s correlation. Pearson’s correlation tests the relationship between two
continuous variables. This study used Pearson’s correlation to test the relationships between
caregiver depressive symptoms, caregiver age, total number of patient symptoms, and perception
of patients’ pain and fatigue, and the relationships between caregiver burden and these caregiver
characteristics (see Table 11). The correlations between caregiver depressive symptoms and
caregiver characteristics including age, total number of patient symptoms, and their perception of
pain and fatigue severity and frequency were lower than 0.20. The correlations between
caregiver burden and the same caregiver characteristics were lower than 0.30, except for the
correlation between family abandonment and total number of patient symptoms (r = 0.68, p >
.05).
Table 10 shows that caregivers who had higher scores of depressive symptoms and
caregiver burden were more likely to report higher scores of patients’ symptom severity and
frequency. The score of caregiver depressive symptoms was significantly correlated with
caregivers’ perception of patients’ fatigue severity and frequency. Caregivers’ schedule
interruption was significantly correlated with caregivers’ perception of patients’ pain and fatigue
severity and frequency. Family abandonment was significantly correlated to caregivers’
perception of patient pain severity and pain frequency and fatigue severity. Caregiver selfesteem, however, had no significant correlation with these caregiver characteristics.

115

Table 11
Pearson’s Correlation Coefficient for Caregiver Characteristics and Caregiver Perception of
Pain and Fatigue Severity and Frequency

Caregiver characteristics
Caregiver age
Total number of symptom
management
Caregiver perception
Pain severity
Fatigue severity
Pain frequency
Fatigue frequency
* p < 0.05. ** p < .001

CESD

Schedule
interruption

Family
abandonment

-0.29**

- 0.17**

0.22**

0.22**

0.26**

0.00

0.08
0.17*
0.11
0.17**

0.14*
0.25**
0.14*
0.19**

0.21**
0.18**
0.15**
0.05

0.03
- 0.02
- 0.03
- 0.02

0.68

Self-esteem
- 0.29**

The t Test for Differences of Caregivers’ Characteristic Between Caregiver Whose CESD
≥16 and CESD<16.
Because the CESD score of 16 was used to determine screening clinical depressive
symptoms, caregivers’ CESD scores were separated into two groups (CESD ≥16 and CESD
<16). This study used the t test and cross tabulation to identify the differences between two
CESD groups for caregiver characteristics, including caregiver age, hours of symptom
management, total number of patient symptoms, and caregiver burden of schedule interruption
and family abandonment (Table 12). Caregivers who had a risk of clinical depressive symptoms
(CESD ≥16 ), as compared to the caregivers who had depression scores lower than 16, were
younger, spent double hours in providing symptom management in the last 7 days, recognized a
greater number of patients’ symptoms, and also reported a higher score on caregivers burden
(schedule interruption and family abandonment). However, four binomial caregiver
characteristics (sex, education, relationship to patient, and living arrangement) did not have
significant differences between two depression groups.

116

Table 12
Differences of Caregiver Demographics, Caregiver Scores of Patients Pain and Fatigue Severity
and Frequency, Congruence of Pain and Fatigue Severity and Frequency between CESD ≥ 16
and CESD <16 (t test, N = 225)
Caregiver characteristic

Age
Hours of symptom management
Total # of symptoms
Caregiver burden
Schedule interruption
Family abandonment
Self-esteem
Caregiver score
Pain severity
Pain frequency
Fatigue severity
Fatigue frequency
Congruence
Pain severity
Pain frequency
Fatigue severity
Fatigue frequency

CESD ≥ 16
N (%)
56 (24.89)

CESD <16
N (%)
169 (75.11)

Mean (SD)
49.59 (15.14)
1.18 ( 0.81)
8.92 ( 2.92)

df

t

Mean (SD)
56.53 (11.96)
0.55 ( 0.81)
7.74 ( 3.14)

223
223
223

-3.51*
3.68*
2.45*

3.34 ( 0.67)
1.99 ( 0.79)
4.46 ( 0.50)

2.70 ( 0.76)
1.62 ( 0.56)
4.51 ( 0.45)

223
223
223

5.65*
3.84*
-0.63

3.53 (
3.62 (
6.16 (
5.84 (

2.88)
3.04)
2.36)
1.81)

3.37 (
3.23 (
5.53 (
5.36 (

2.90)
2.94)
2.34)
2.23)

223
223
223
223

0.37
0.86
1.76
1.47

2.87 (
2.64 (
2.23 (
1.62 (

2.41)
2.55)
1.80)
1.87)

2.38 (
2.45 (
2.35 (
2.09 (

2.23)
2.47)
1.99)
2.13)

223
223
223
223

1.39
0.50
- 0.41
- 1.45

2

X
14.16*

1

1.92

1

1.77

1

3.58

1

1.52

N (% of total)
Age

Sex
Education
Relationship to
patients
Living arrangement

<45
45≤ age <65
≥65
Female
Male
≤High school
≥College
Spouse

22 ( 9.8)
25 (11.1)
9 ( 4.0)
18 ( 8.0)
38 (16.9)
23 (10.2)
33 (14.7)
19 ( 8.8)

26 (11.6)
100 (44.4)
43 (19.1)
72 (32.0)
97 (43.1)
53 (23.6)
116 (51.6)
36 (16.7)

Non-spouse
Together
Separate

35 (16.2)
10 ( 4.6)
45 (20.6)

126 (58.3)
19 ( 8.7)
144 (66.1)

2

N (% of total)

* p < .05

117

GLM. The present study examined the predictors of caregiver depression and caregiver
burden by using GLM. This examination was conducted to determine if there was the
multicollinearity of the context factors (caregiver age, sex, education, relationship to patient,
living arrangement, and total number of patient symptoms) and caregiver depression and
caregiver burden for the four symptom congruence models. Multicollinearity is a statistical
phenomenon in which two or more predictive variables in a multiple regression model are highly
correlated. In this situation the coefficient estimates may change erratically in response to small
changes in the model or the data.
Factors of caregiver depressive symptoms. Not only were caregivers’ demographic
characteristics (caregiver age, sex, education, relationship to patient, living arrangement, and
total number of patient symptoms) used to determine the predictor of caregivers’ depressive
symptoms, but also caregiver burden (schedule interruption, family abandonment, and selfesteem) was included in the regression test of depression because caregiver burden had been
identified as a possible component of caregiver depression. The equations (1) below shows that
the factors of caregiver depressive symptoms included caregiver education, age, family
abandonment, and schedule interruption. An increase in caregiver depressive symptoms was
predicted for caregivers who were younger, had college or advanced degrees, and reported
higher scores of family abandonment and schedule interruption.

Caregiver depressive symptom = 7.12 + 2.86* education + (-0.16) caregiver age
+ 1.97* caregiver family abandonment +
3.14*schedule interruption

118

(1)

Factors of caregiver burden. This analysis model considered binomial caregiver
demographics (age, sex, education, relationship to patient, living arrangement, and the total
number of patient symptoms) as predictors of caregiver schedule interruption, family
abandonment, and self-esteem.
Caregiver schedule interruption was predicted by one variable: the total number of patient
symptoms. The equations (2) below shows that a higher score of schedule interruption was only
directly predicted by an increase in the total number of patients’ symptoms. Therefore,
caregivers who perceived a greater number of patient symptoms were more likely to report
schedule interruption.

Schedule interruption = 2.42 + 0.05 * total number of patient symptoms

(2)

Family abandonment was predicted by caregiver age, relationship to patient, and the total
number of patient symptoms. The equations (3) below shows that a higher score on family
abandonment is predicted by a non-spouse caregiver, an increase in caregiver age, and an
increased number of patient symptoms. Therefore, caregivers who were patients’ spouses, were
younger, and recognized more patient symptoms were more likely to report a higher score of
family abandonment.

Family abandonment= 1.69 + 0.33* spouse + (- 0.08) * caregiver age +
0.05 * total number of patient symptoms

This study found that binomial caregiver demographic characteristics and the total

119

(3)

number of patient symptoms did not predict self-esteem. As seen in Table 6, there was also no
significant difference in self-esteem between binomial caregiver characteristics.
In summary, caregivers who were younger, a spouse of the patient, educated with a
college or advanced degree, and reported a larger number of their patients’ symptoms were more
likely to report a higher score of caregiver depression and caregiver burden. Caregivers’
characteristics were tested as confounding variables of caregiver depression and caregiver
burden in the multinomial log-linear regression that examined whether or not caregivers’
characteristics directly influenced the three levels of congruence of pain and fatigue severity and
frequency.
Multinomial Loglinear Regression
The investigator study used the multinomial loglinear regression because the data
displayed non-normality and because the investigator categorized each outcome variable
(congruence of pain and fatigue severity and frequency) with three congruent levels. The
investigator tested the main effect of independent variables— caregiver depressive symptoms,
caregiver burden and caregiver demographic characteristics—using the four congruence models
shown below. The investigator also tested interactions of main effects and their confounding
variables on these four models. The positive relationship between cancer depressive symptoms
and caregiver burden has been found in previous studies (Daly, Douglas, Lipson, & Foley, 2009;
Grunfeld, et al., 2004; Sherwood et al., 2006). Given et al. (1992) have reported the interrelationships among three caregiver reactions in developing the CRA instrument for caregivers
who provide patient care. The present study found that caregiver education, age, and burden can
influence caregiver depressive symptoms. Caregivers’ age, relationship to patients and total
number of patient symptoms can predict caregiver burden. If caregiver depressive symptoms and

120

caregiver burden, and their factors cannot directly impact congruence of pain and fatigue severity
and frequency, the multicollinearity has to be considered regarding the associations among these
caregiver characteristics.
Hypothesis 1
This study examined the effect of caregiver depressive symptoms and three caregiver
reactions—family abandonment, schedule interruption, and self-esteem—as independent factors
to explain three levels: overestimation, congruence, and underestimation of pain and fatigue
severity and frequency.
When one examines the main effect of those independent factors, the findings show that
congruence of pain severity was influenced by depressive symptoms and self-esteem, and
congruence of pain frequency was influenced by schedule interruption. Congruence of fatigue
severity and frequency were not influenced by depressive symptoms or caregiver reactions.
Using the congruence group of pain severity as a reference, two significant results were
found for the Odds Ratios (OR) of overestimation versus congruence and underestimation versus
congruence groups. The OR of the overestimation group versus congruence group of pain
severity could have increased by increasing the scores of family abandonment and family and
caregiver depressive symptoms, and simultaneously decreasing the score of self-esteem. The OR
of underestimation compared with congruence could have increased if the score of CESD
increased and the score of self-esteem decreased.
Schedule interruption was found to significantly influence the congruence of pain
frequency, where the OR of congruence versus underestimation of pain frequency increased if
the score of schedule interruption also increased. Caregiver depressive symptoms and caregiver
reactions did not contribute effect for the congruence of pain frequency.

121

The values of R square in these regression examinations for these congruence models of
pain and fatigue severity and frequency, however, were significant. This means that these models
cannot be explained by either caregiver depressive symptoms or three caregiver reactions alone.
Other factors determined congruent levels of pain and fatigue severity and frequency besides
caregiver depressive symptoms and caregiver burden.
Hypothesis 2
In Hypothesis 2, the caregiver demographics—age, sex, education, relationship to
patients, living arrangement, and caregivers’ perception of total number of patient symptoms
were included and tested as factors of congruence of pain and fatigue severity and frequency.
Exp (parameter) or exp (β) represents the odds ratios for the predictors. They are the
exponentiation of the coefficients.
Congruence of pain severity. This study tested the main effect of the independent
variables on congruence of pain severity (the equations 4-7). The equations (4) and (5) used the
congruence group as reference to compare with two noncongruence groups while the equations
(6) and (7) used the underestimation group as reference compared with congruence and
overestimation.
The equations (4) and (5) showed that total number of patient symptoms, living
arrangement, caregiver depressive symptoms, and caregivers’ self-esteem can influence this
model. Caregivers were more likely to be not congruent with patients’ pain severity than be
congruent with patients’ pain severity if caregivers reported a greater number of patient
symptoms, a higher score of caregiver depressive symptoms, and a lower score of self-esteem,
and if the caregiver lived with patient.
Both the equations (4) and (5) illustrated that caregivers were more likely to display

122

congruence rather than noncongruence (overestimation and underestimation) for patients’ pain
severity when they felt fewer depressive symptoms and more self-esteem, recognized fewer
patient symptoms, and did not live with patients. The effects of caregiver depressive symptoms
and caregiver self-esteem on congruence can support the hypothesis that increasing caregiver
depression can decrease congruence of symptom perception. Family abandonment and schedule
interruption, however, did not contribute a significant effect to the model of congruence of pain
severity.

Log [p(Overestimation of PS / Congruence of PS)] =

(4)

2.71 + 0.02*CESD+ (-1.01) * Self-esteem + 0.20 * total number of
patient symptoms + exp (-1.49) * not living with patient

Log [p(Underestimation of PS / Congruence of PS)] =
0.61 + 0.06* CESD + ( -0.44) * Self-esteem + 0.20* total number of patient

(5)

symptoms + exp (-0.22) * not living with patient

Meanwhile, the findings of the equations (4) and (5) showed that caregivers who did not
live with patients were more likely to be congruent with patients’ pain severity, and caregivers
who perceived greater patient symptoms were more likely to be noncongruent with patients.
These two caregiver characteristics are in opposition to the hypothesis that caregivers perceiving
more symptoms or living with patients have greater congruent symptom perception.

123

Log [p(Overestimation of PS / Underestimation of PS)]=

(6)

(-0.44)+ (-0.03)* CESD + (-0.18)* total number of patient symptoms +
exp (-1.22) not live with patient

Log [p(Congruence of PS / Underestimation of PS)]=

(7)

1.36+ (-0.06)*CESD + ( -0.02)* total number of patient symptoms + exp
(0.15) not live with patient

The equations (6) and (7) used the group of caregiver underestimation of pain severity as
the reference to find the factors. In the equation (6), when comparing two noncongruent groups
of pain severity, a higher OR of caregivers who overestimate patients’ severity rather than
underestimate was be predicted by a lower score of caregiver depressive symptoms and fewer
patient symptoms, and also by the caregivers living with patients. In the equation (7), an increase
in the OR of caregivers that are congruent versus those that underestimate patients’ pain severity
was predicted by a lower score of caregiver depressive symptoms and the total number of
patients’ symptoms, and also by the caregivers not living with patients. Therefore, the equation
(6) and (7) showed that caregiver depression, the total number of patient symptoms, and
caregivers’ living arrangement are important factors influencing caregivers to accurately
perceive patients’ pain severity.
Among these four pain severity equations (4-7) caregiver depressive symptoms and
caregiver self-esteem were main factors for congruence of pain severity, and caregiver living
arrangement and total number of patient symptoms contributed to the main effect for this model.
Caregivers who had fewer depressive symptoms were more likely to be congruent with patients’

124

pain severity. This means caregivers could more accurately perceive patients’ symptoms if they
had fewer depressive symptoms. A greater number of patient symptoms and living together
firmly contributed to decreasing the likelihood of congruent pain severity. When the model used
different categories as references, the effect of self-esteem was only significant in the equations
PS 1and PS 2, which used congruence as the reference. Caregiver self-esteem only produced an
effect for caregivers who had congruence greater than non-congruence.
Congruence of fatigue severity. This study tested how caregiver characteristics
determined three levels of congruence of fatigue severity and then obtained three equations (810). The results showed that the total number of patient symptoms was the only factor
determining whether the dyads of caregivers and patients had congruence or noncongruence for
fatigue severity. The equation (8) showed that the OR of caregivers who overestimated versus
the caregiver who were congruent with patients’ fatigue severity could increase by 20% when
caregivers perceived one more patient symptom than before ( 95% Confidence Interval [CI] of
exp (0.08) = 0.98 - 1.20). The equation (9) showed that the OR of underestimation rather than
congruence for patients’ fatigue severity decreased by 10% when caregivers perceived one more
patient symptom than before (95% CI of exp (-1.02) = 0.81-1.01). The equation (10) showed that
the OR of overestimation versus underestimation increased by 20% when caregivers perceived
one more patient symptoms than before (95% CI of exp (0.18) = 1.06 - 1.31). Only the equation
(9) showed a significant effect of the total number of patient symptoms on congruence of fatigue
severity, which indicated that when caregivers became increasingly aware of the total number of
patients’ symptoms, they were more likely to be congruent with patient fatigue severity rather
than underestimate it.

125

Log [p (Overestimation/ Congruence of FS)] =

(8)

(p > .05)

(9)

(p < .05)

(10)

(p > .05)

(-0.78) + 0.08 * total number of patient symptoms

Log [p (Underestimation/ Congruence of FS)] =
0.12 + (-0.10) * total number of patient symptoms

Log [p (Overestimation/ Underestimation of FS)] =
(-0.90) + 0.18 * total number of patient symptoms

Regarding Hypothesis 2, the association between caregiver characteristics and caregiver
burden was considered as a factor in the test. This study also found that the interaction between
the total number of patient symptoms and family abandonment could contribute to the model of
fatigue severity. As the study examined the predictors of family abandonment above, the total
number of patient symptoms was a predictor of family abandonment. Therefore, the reason that
family abandonment did not directly affect this congruence model, perhaps, was because of the
multicollnearity. The total number of the patient’s symptoms may have been more important
than the caregiver’s family abandonment in affecting congruence of fatigue severity.
Congruence of pain frequency. Based on finding in the equations (11-13) to test the
main effect and interaction of the factors, schedule interruption was the only factor influencing
the model of congruence of pain frequency. The equation (11) showed that the OR of caregivers
who overestimated patients’ pain frequency compared to those who were congruent with pain
frequency decreased by 24% when one point of schedule interruption increased (95% CI of exp
(-0.27) = 0.51 - 1.13). The equation (12) showed that the OR of caregivers who underestimated

126

rather than were congruent with patients’ pain frequency decreased by 46% when one point of
schedule interruption increased (95% CI of exp (-0.62) = 0.4 - 0.83). The equation (13) shows
that the OR of caregivers who overestimated versus underestimated patients’ pain frequency
increased by 42% when one point of schedule interruption increased (95% CI of exp (0.35) =
0.89- 2.25). Specifically, only the equation (12) showed a significant effect of schedule
interruption on congruence.

Log [p (Overestimation/ Congruence of PF)] =

(11) (P > .05)

0.45 + (-0.27) * schedule interruption

Log [p (Underestimation/ Congruence of PF)] =

(12) (P < .05)

1.11 + (-0.62) * schedule interruption

Log [p (Overestimation/ Underestimation of PF)] =

(13) (P > .05)

(-0.66) + 0.35 * schedule interruption

This study also tested the effect of the interaction between schedule interruption and its
confounding variables (spouse relationship to patients, caregiver depressive symptoms, and
caregiver feelings of abandonment and self-esteem) on the model of congruent pain frequency.
Schedule interruption was the only factor contributing a significant effect to the model, which
was consistent with the test for the main caregiver characteristic effect in the model of congruent
pain frequency.
In sum, caregivers who felt more schedule interruption were more likely to have an

127

accurate report of patients’ pain frequency; this finding, however, did not support the first
hypothesis. This study also found that schedule interruption from care involvement (including
symptom management) was predicted by the total number of patient symptoms, which may
indicate that caregivers must be aware of symptom frequency. Therefore, caregivers who
reported more schedule interruption could be more congruent on patients’ pain frequency.
Congruence of fatigue frequency. After testing both the main effect and interaction,
total number of symptoms was the only variable which influenced the congruence of fatigue
frequency. The equation (14) showed that the OR of caregiver who overestimated fatigue
frequency compared to a congruent caregiver decreased by18% when caregivers perceived one
more patient symptom than before (exp (-0.08), 95% CI of exp (β) = 0.84 - 1.02). The equation
(15) showed that the OR of caregivers who underestimated patient fatigue frequency rather than
being congruent decreased by 23 % when caregivers recognized one more patient symptom than
before (exp (-0.14), 95% CI of exp (β) = 0.77 - 0.97). The equation (16) showed that the OR of
caregivers who overestimated fatigue frequency rather than underestimated it increased by 6%
when caregivers perceived one more patient symptom than before (exp (0.06), 95% CI of exp
(β)= 0.94- 1.20). Caregivers who recognized more patient symptoms were more likely to have a
congruent report of patients’ fatigue frequency. Still, only the equation (15), by comparing the
underestimation group with the congruence group, showed a significant effect on the total
number of patient symptoms to congruence of fatigue frequency. Briefly, an increase in the total
number of patients’ symptoms may imply that caregivers become increasingly aware of patients’
symptoms, and that they are more likely to be congruent than to underestimate patients’ fatigue
frequency.

128

Log [p (Overestimation/ Congruence of FF)] =

(14) (P > .05)

0.23 + (-0.08) * total number of patient symptoms

Log [p (Underestimation/ Congruence of FF)] =

(15) (P < .05)

0.21 + (-0.14) * total number of patient symptoms

Log [p (Overestimation/ Underestimation of FF)] =

(16) (P > .05)

0.22 + 0.06 * total number of patient symptoms

Conclusion
In this study, caregivers, on average, were 52.8 (SD± 13.1), spouses of cancer patients ( n
= 161, 71.6%), and Caucasian (n = 207, 92%). Also, 149 (65.9 %) of the caregivers had a college
degree (Table 2), and 189 (84%) lived with their patients, which accounted for all the spouse
caregivers. Comparing the mean scores of pain and fatigue severity and frequency between
cancer patients and caregivers, caregivers overestimated these four symptom variables. The ICCs
of pain and fatigue severity and frequency were from low to moderate. The ICCs of pain and
fatigue severity between cancer patients and their family caregivers were lower than 0.5, and the
ICCs for fatigue severity and frequency were lower than 0.4 (Table 7). The ICC score for pain
and fatigue severity was higher than pain and fatigue frequency. Because the ICC criteria has
stated that a score of 0.3 to 0.4 indicates fair congruence and scores of 0.5 to 0.6 indicates
moderate congruence, the 225 cancer dyads have a fair to moderate level of congruence.
This study categorized each score difference of four symptom variables (pain and fatigue
severity and frequency) between cancer patients and caregivers into three levels of congruence:

129

congruence (-1~1), underestimation (< -1), and overestimation (>1). More than 40% of
caregivers were categorized in the group of congruence for these four symptoms variables, and
that was higher than the other two noncongruent groups. Less than 25 % of caregivers (n = 53)
were categorized in the underestimation group for these four symptoms. In brief, most caregivers
were congruent with patient symptoms or overestimated them.
The first and second hypotheses were supported when the results showed that an increase
in depressive symptoms and a decrease in self-esteem predicted less congruence of pain severity.
However, a higher score of schedule interruption predicted more congruence of pain frequency
rather than noncongruence. Neither caregiver depression nor caregiver burden affected the model
of fatigue severity and fatigue frequency, which is inconsistent with the first hypothesis.
The second hypothesis included caregiver demographics (age, sex, education, and
relationship to patients), but demographics did not have a significant effect on congruence.
However, two caregiver characteristics not found in previous studies (the total number of patient
symptoms and caregivers’ living arrangement) were identified for congruence. However, in
contrast to this hypothesis, increasing the total number of patients’ symptoms can independently
raise the OR of caregivers’ congruence rather than noncongruence for pain severity, fatigue
severity, and fatigue frequency. Also, caregivers who did not live with their patients had a
greater likelihood of being congruent with patients’ pain severity, but this caregiver
characteristic did not impact congruence of pain frequency, fatigue severity, or fatigue
frequency.
Overall, caregiver depressive symptoms and caregiver self-esteem were identified as the
two factors which affect congruence of pain severity, as hypothesized. Caregivers’ demographic
characteristics–including age, sex, and education–did not affect the congruence of pain and

130

fatigue severity and frequency as the covariants of caregiver depression and caregiver burden.
Therefore, because the finding on the relationships among caregiver depression, caregiver
burden, and caregiver characteristics did not completely support the two hypotheses, the EASP
model needs to be adjusted.

131

CHAPTER 6
Discussion and Implications
The purpose of the current study was to examine the concept of EASP in caregiving
situations and to identify how this concept can be influenced by caregiver characteristics.
Specific aims were to examine whether caregiver depression and caregiver burden can decrease
congruence of pain and fatigue severity and frequency, and how caregiver demographic
characteristics affect this concept. The findings of this study were compared to prior studies for
the similarities and differences of congruent levels and factors of congruence of cancer
symptoms between patients and caregivers. This chapter presents those comparisons and the
implications of the conceptual framework of EASP and of the results of this study for clinical
practice, nursing research, and health policy.
Effect Size
The power analysis in Chapter 4 showed that the sample size of 225 dyads of cancer
patients and family caregivers was adequate. Within a fixed sample size of 225 dyads of cancer
patients and their family caregivers, this study obtained an effect size of 0.06, standard sample
power of 80%, and significance of 0.05. Effect size of 0.06 in a regression model is higher than
the minimum effect size of 0.02 and lower than moderate effect size of 0.15, according to
Cohen’s (1988) suggestion for "conventional" values of effect size. This study used four
symptom score differences (pain severity, fatigue severity, pain frequency, and fatigue
frequency) between cancer patients and their family caregivers as dependent variables, and eight
independent variables (caregiver’s age, sex, education, relationship to patient, living
arrangement, total number of patient symptoms, caregiver depression, and caregiver burden).
Effect size addresses the magnitude of the effect under the alternate hypothesis between groups,

132

or of the relationship between variables. A small effect size, therefore, indicated that this study
would be less likely to find a statistically significant relationship between the dependent
variables and those eight independent variables.
In spite of the small effect size, this study found that caregiver depression, caregiver selfesteem, total number of patients’ symptoms and caregivers’ living arrangement had significant
effects on congruence of pain severity; caregivers’ schedule interruption can influence
congruence of pain frequency. Therefore, caregivers’ emotionality and perceived threats of
caregiving environment can influence congruence of symptom perception between cancer
patients and their caregivers. These two factors impact only pain severity and frequency but not
fatigue severity and frequency, while the total number of patient symptoms that caregivers
perceive can impact congruence of fatigue severity and fatigue frequency.
Congruent Levels of Pain and Fatigue Severity and Frequency
This study obtained a fair-to-moderate level (0.3 - 0.5) of congruence for pain and fatigue
severity and frequency between cancer patients and their family caregivers. This has been a
typical finding: that family caregivers and patients had low agreement for patients’ symptoms,
and caregivers overestimated their cancer patients’ symptoms (McPherson, et al., 2008; Resnizky
& Bentur, 2007; Silveira, et al., 2010; Wennman-Larsen, et al., 2007).
Meanwhile, the ICC level of cancer pain was greater than that of fatigue, which means
that congruence between caregivers and patients for patients’ pain is better than for patients’
fatigue. Caregivers may pay more attention to patients’ pain responses than to patients’ fatigue,
because pain management has been emphasized in nursing lectures to caregivers. Compared to
cancer fatigue, patient responses to cancer pain may also be easier for caregivers to recognize.
Cancer patients may express cancer pain more often because they want to lessen unpleasant

133

feelings, especially when cancer pain can limit patients’ capabilities for daily activities. Cancer
patients, however, may have become accustomed to feeling fatigue, so that they don’t
specifically point out these feelings to family caregivers (Wennman-Larsen, et al., 2007).
This study, additionally, categorized the dyads’ congruence into three levels (groups),
and found that the number of caregivers who were congruent with patients’ pain and fatigue
severity and frequency was greater than those who overestimated and underestimated it.
Therefore, although the agreement level was low, caregivers were capable of accurately
perceiving patients’ symptoms within the range of core difference from -1 to 1. Also, the number
of caregivers who overestimated patient symptoms was greater than those who underestimated
them. In that case, most caregivers tended to believe that patients were suffering from symptoms
burden, which was supported by previous studies that also found that patients’ symptoms were
overestimated by their family caregivers (Broberger, et al., 2005; Greig, et al., 2005; Lobchuk &
Degner, 2002; Redinbaugh, et al., 2002).
Therefore, the current study reveals that many caregivers of cancer patients can
accurately perceive patients’ symptoms, and that some may overestimate patients’ symptoms.
Another important finding is that caregivers who either were congruent with or overestimated
patient symptoms tended to provide symptom management, whereas caregivers who
underestimated patients’ symptoms delayed symptom management or provided insufficient
symptom management.
Hypothesis 1: Caregiver Depression and Caregiver Burden Relevant to Congruence Model
Table 9 shows that among three congruent levels (underestimation, congruence, and
overestimation) of four symptom variables (pain and fatigue severity and frequency), caregivers
who were in the congruence groups of pain severity and fatigue severity reported the lowest

134

scores of caregiver reactions, and caregivers who were in the congruence groups of pain
frequency and fatigue frequency reported the highest scores of three caregiver reactions. Also,
the caregivers in the two congruence groups of pain and fatigue severity had the lowest scores of
depressive symptoms, and those in the underestimation group reported the highest scores of
depression, in regards to CESD score differences. However, the caregivers in the congruence
groups of pain and fatigue frequency had the highest score of depressive symptoms and three
caregiver reactions than the other two noncongruence groups (Table 9).
Hypothesis 1 tested whether increasing caregiver depression and caregiver burden can
decrease congruence of symptom perception between cancer patients and their family caregivers.
Three results show that (a) increasing caregiver depression decreased congruence of symptom
perception, but this study showed that decrease only for pain severity; (b) increasing caregiver
self-esteem also increased the likelihood of congruence of pain severity; (c) increasing schedule
interruption generated a higher likelihood of congruence on pain frequency; (d) family
abandonment did not generate effects on any congruence of these four symptom variables; and
(e) neither caregiver depression nor caregiver burden impacted the three congruent levels of
fatigue frequency and fatigue severity. Previous studies have suggested that higher caregiver
burden and caregiver emotions can diminish caregivers’ empathic accuracy (congruence) of
patients’ symptoms (Higginson & Gao, 2008; McPherson, et al., 2008; Resnizky & Bentur,
2007; Wennman-Larsen, et al., 2007).
For the result “a,” Miaskowski et al. (1997), who used the CESD to measure caregiver
depression, found higher scores reported by caregivers who were not congruent with patients’
pain severity. McPherson et al. (2008) used the Profile of Mood States, which evaluates
individuals’ affective status, such as depression, anxiety, and sadness. They found that caregiver

135

mood can predict the discrepancy of psychological cancer symptom. The current study extended
the original findings to show that caregivers in the underestimation group reported higher scores
than those in the overestimation group of pain severity. Accordingly, caregivers who were at risk
of depression may have had one of two behaviors toward patient symptoms and themselves. If
caregivers had depressive symptoms and projected them to patients’ symptoms, they might
highly overestimate patients’ pain severity. If caregivers’ depression led them to be self-centered
(Bieling & Alden, 2001), caregivers would mind their own symptoms more than they would take
care of patients’. In that case, they would be likely to underestimate patients’ pain severity.
For the result “b,” self-esteem enhanced the possibility of congruence of pain severity.
Wennman-Larsen et al. (2007) also suggested that caregivers who have greater self-esteem are
more likely to have congruence with patients on fatigue severity, but not on pain. The Wennman
also used fatigue for recognizing caregivers’ congruence because it was the most frequently
occurring symptom. Therefore, improving caregivers’ self-esteem for providing care may
increase their accurate perceptions of patients’ symptom severity.
The result “c” shows that caregivers with more schedule interruption tended to be
congruent with patients’ pain frequency. In contrast with the results of a and b, the result c is
opposite from the hypothesis. This finding was identified in the previous studies. However,
caregivers who have been experiencing schedule interruption are assumed to spend time
involved in the patient’s life and care; accordingly, caregivers can notice or hear patients
expressing symptoms. Among all symptoms, cancer pain is more often expressed by patients,
although fatigue is the most commonly occurring symptom. In this regard, that might be why
caregiver schedule interruption can contribute to congruence of pain frequency.
The current study found that different caregiver reactions were associated with different

136

congruencies of symptoms and influenced pain but not fatigue. The reason for that may be
because dealing with different symptoms is related to different caregiver reactions. Persson et al
(2008) used the perspectives of stress theories and illustrated that these five dimensions of CRA
are related to different stressors of care. Caregivers who lack resources for caregiving often
reported family abandonment, while those who have an amount of resources reported schedule
interruption, and those who feel self-achievement on providing care reported self-esteem. In this
regard, when perceiving one patient symptom, caregivers may generate the care demands for this
symptom that are different from the other symptoms. That results in generating different
caregivers’ reactions. For example, accurately perceiving patients’ pain severity may depend on
caregivers’ self–esteem for pain evaluation and management. In another words, caregivers who
have higher self-esteem are more likely to accurately perceive patients’ pain severity. Accurately
perceiving patients’ pain frequency may indicate caregivers pay attention on the occurrence of
pain, which also means that caregivers spend time observing patient’s symptoms, so that they are
more likely to report schedule interruption.
Caregivers carefully paid attention to patients’ symptoms, especially cancer pain, in this
study so that they could accurately recognize pain occurrence and provide pain management in
time. However, that required caregivers to constantly ignore their own life events or rearrange
their life schedules in order to cope with patients’ symptom management. Therefore, they may
have felt their schedules were significantly interrupted (schedule interruption). In sum, greater
congruencies of these symptoms may indicate that caregivers had to involve themselves in
providing symptom management for patients’ pain and fatigue, and that also imbued greater
caregiving reactions. It is possible that congruent perception of patients’ pain severity and pain
frequency indicated that caregivers participated more in patients’ pain management, and that also

137

reflects the circumstance that those caregivers dealing with patients’ pain were more likely to
have an increase in caregiver burden.
The reason for the different results between the current study and the prior studies on the
effect of caregiver burden on congruence may be the measurements for caregiver burden and
different counting strategies for the outcome variable, congruence. The present study used three
subscales of the CRA–Family Abandonment, Schedule Interruption, and Self-Esteem–to detect
the influence of each caregiver reaction on these four congruencies of symptoms. Higginson and
Gao (2008) used the Zarit Burden Interview to measure caregiver burden and examine the
severity of pain and related palliative symptoms for congruence.
In turn, although two studies (Wennman-Larsen, et al., 2007 and McPherson, et al. 2008)
used the Caregiver Reaction Assessment (CRA) to measure caregiver burden as did the present
study, the congruence of patients’ variables affected by caregiver reactions were different.
Wennman-Larsen et al. (2007) reported that family abandonment and health impact predicted
only congruencies of emotional function and quality of life, while McPherson, et al. (2008)
obtained an average of the five subscales (schedule interruption, family abandonment, health
impact, financial impact, and self–esteem) together to predict congruence of psychological
symptoms. Therefore, those findings support Hypothesis 1: Caregiver burden can decrease
caregivers’ ability to be congruent with patients’ perception on symptoms and quality of life. It
might also indicate that accurately perceiving patients’ responses to their cancer may give rise to
different caregiver reactions.
For the results “d” and “e,” the prior studies lack information for family abandonment
and fatigue perception. Family abandonment was not specified as a factor for congruence of
symptom perception, but it was identified for the congruence of patients’ quality of life

138

(Wennman-Larsen et al., 2007). The Wennman-Larsen study also found that patients’ sex and
age can determine the discrepancy of fatigue severity, but that caregivers’ characteristics cannot.
Many variables can trigger fatigue in cancer patients, for example, cancer treatments and cancer
itself, that lead patients to commonly feel fatigue more than pain. Also, cancer patients often
perceive fatigue earlier than they perceive pain, so they may tolerate fatigue without expressing
it. That might be the reason for Wenmman-Larsen et al.’s (2007) result. That might also be why
caregiver depression and caregiver burden in the current study did not affect congruence of
fatigue severity and fatigue frequency.
In addition, the current study is a descriptive study with a cross-sectional design. The
predictive relationship from one variable to anther found in a cross-sectional descriptive study
may simply explain a correlation, which means that the predictive relationship can be reversible.
In the current study, by applying the previous statement, caregiver depressive symptoms, selfesteem, and schedule interruption can also be impacted by caregivers’ congruence of patients’
symptoms. Caregivers who can accurately perceive patients’ pain severity are more likely to
have fewer depressive symptoms and higher self-esteem that the caregivers who cannot;
caregivers who can accurately perceive symptom frequency are more likely to have more
feelings of schedule interruption. In this regard, congruence of patients’ pain severity may be a
marker of caregivers who are less depressed and have a growing self-esteem, while congruence
of pain frequency may indicate schedule interruption. A longitudinal study would be valuable to
examine the variation of the relationship between caregiver burden and congruence of patient
symptoms perception in the caregiver’s long-term care experience.
Hypothesis 2: Caregiver Characteristics Relevant to Congruence
Hypothesis 2 tested how caregivers’ demographic characteristics affected the congruence

139

of symptom perception if these characteristics were also tested with caregiver depression and
caregiver burden in the regression models. Still, caregiver depressive symptoms, caregiver selfesteem, and schedule interruption had the same contribution to congruence as the finding for
Hypothesis 1 above, even though the caregiver demographic characteristics were included in the
model. Caregivers who had fewer caregiver depressive symptoms were more likely to have
congruence of pain severity, rather than noncongruence; in addition, they also are more likely to
have overestimation or congruence, rather than underestimation. Caregivers who had more selfesteem tended to have congruence of pain severity rather than noncongruence. Schedule
interruption still was shown as a single factor for congruence of pain frequency.
The examination of Hypothesis 2 identified that both caregivers’ living arrangement and
total number of patients’ symptoms can affect congruence of pain severity. Caregivers who lived
with patients were more likely to not be congruent with patients’ pain severity. The total number
of patient symptoms can influence congruence of fatigue severity and fatigue frequency. The
fewer patient symptoms that a caregiver perceives, the greater likelihood of having congruence
on pain severity, fatigue severity.
Living arrangement. Living arrangement was described as caregivers who lived with
cancer patients and those who did not. The variable only contributed to the model of congruence
of pain severity, but it did not affect the other three congruent models of symptoms. Resnizky
and Benur (2002) emphasized that a caregiver’s living arrangement is a way to learn about a
caregiver’s involvement in patient care, and their findings have shown that caregivers who live
with their cancer patients can more accurately report patients’ symptoms than could the
caregivers who did not live with patients. However, the findings in the present study are opposite
from the findings of Resnizky et al. (2002). In the present study, the equation (4) and (5) show

140

that caregivers who did not live with patients were more likely to have congruence of patients’
pain severity, and the equation (6) shows that caregivers who did not live with patients were also
more likely to underestimate patients’ pain severity rather than to overestimate it. Regarding the
equations (4) and (5), those caregivers who did not live with patients may have provided
symptom assessment every time they visited patients. They could have more accurately
perceived patients’ symptoms; in this regard, they may have been more likely to objectively
assess patients’ reactions from symptoms. The equation (6) may imply that if the caregivers who
did not live with patients did not provide symptom assessment when they visited patients, they
would have underestimated, but not overestimated, patient symptoms. Therefore, symptom
assessment may be an important approach to accurately perceive patient symptoms. Further
study is needed to retest the effect of living arraignment and caregivers’ symptom assessment to
determine the congruence of symptom perception.
Total number of patient symptoms. The total number of patient symptoms contributed
to the models of congruence of pain severity, fatigue severity, and fatigue frequency. Those
results are supported by the conceptual framework of patient suffering and caregiver compassion
developed by Schulz et al.’s (2007) conceptual model for patient suffering. They (2007)
determined that patient suffering from symptom distress is the unique and independent factor of
caregiver health and is mediated by caregiver compassion and helping behavior. However, this
variable and the findings have not been investigated in the most prior studies on cancer.
Furthermore, the influences of the total number of patient symptoms on pain perception
and on fatigue perception were in opposition. Increasing the awareness of total number of patient
symptoms was directly related to the decrease in possibility of caregivers’ congruence of pain
severity, but it was also related to the increase in likelihood of congruence of fatigue severity and

141

fatigue frequency (see Chapter 5, Multinomial Log-Linear Regression).
Caregivers who observe more patient symptoms can generate two responses on pain
severity. Some caregivers may have witnessed patients suffering from multiple symptoms; they
may view these experiences as a sign and overestimate the patients’ pain severity. Others may
not focus on dealing with patients’ pain severity because they were distracted by other patient
symptoms; in this regard, these caregivers are more likely to give inaccurate perceptions for
patients’ pain severity. Therefore, caregivers recognizing greater numbers of patient symptoms
are more likely to be noncongruent with patients on pain severity.
In turn and in contrast to pain, caregivers who observe greater numbers of patient
symptoms can generate more accurate perceptions on patient’s fatigue but not underestimate it.
A greater number of patient symptoms makes an impression on family caregivers that patients
feel fatigue because their energy is being consumed by experiencing those symptoms. Caregivers
who recognize more patient symptoms, thus, may pay more attention to how often patients’
fatigue occurs. In that case, caregivers tend to more accurately perceive patients’ fatigue but not
to underestimate it.
Furthermore, in the equations of pain severity, congruence levels were determined by
caregiver depressive symptoms, caregiver self-esteem, the total number of patient symptoms and
living arrangement. Caregivers’ awareness of total number of patients’ symptoms was positively
correlated to schedule interruption and family abandonment. A greater number of patient
symptoms usually generate more care demands related to symptom management. In that case,
caregivers have to spend more time or require support from other family members to deal with
patients’ symptoms. In turn, a greater number of patient symptoms can generate a higher
likelihood for caregivers to feel family abandonment.

142

Therefore, according to the results of two hypotheses, accurately perceiving pain,
especially pain severity, was determined more by caregivers’ characteristics than by fatigue. That
may imply that caregivers paid more attention to patients’ pain than to patients’ fatigue. Cancer
pain is often the focus in cancer symptom management. Cancer patients, furthermore, are able to
more specifically describe pain in location, timing, and frequency than they are able to describe
fatigue. Meanwhile, the previous section also described a possibility of a reversible relationship
between caregiver depression, caregiver burden, and congruence of symptom perception
regarding the cross-sectional descriptive study. In that case, caregivers’ perception on patients’
cancer pain, compared to patients’ fatigue, can have more influence on caregivers’ depression,
self-esteem, and schedule interruption. Besides, caregivers’ attention to total number of patients’
symptoms is a vital factor in perceiving fatigue. Observing patients’ fatigue may project an
image to caregivers that patients are having many symptoms that cause fatigue.
The Fitness of Theoretical Framework of EASP
The framework of EASP was developed based on three theories. The factors were
selected and supported by the findings in the previous studies related to congruence of cancer
symptoms. However, the framework is still immature. First, cancer symptoms are many, but this
study selected severity and frequency of pain and fatigue to represent symptom perception.
Whether the factor total number of patient symptoms can influence congruence of other
symptoms is unknown. Besides, the current study shows that the factors which influenced
congruence of pain perception were different from those which influenced congruence of fatigue
perception. Using symptoms other than pain and fatigue to represent congruence of symptom
perception may reveal different caregiver characteristics which can influence congruence. Also,
the factors—caregivers’ schedule interruption, caregivers’ awareness of number of patient

143

symptoms, and their living arrangements—did not following the hypotheses. Those findings
indicate there is lack of knowledge of how caregivers project their own perceptions on patients’
symptoms.
Next, patients’ and caregivers’ characteristics were considered in the contextual factors of
the framework to influence the concept of EASP because the congruence of symptom perception
was obtained by comparing the symptom reports from both patients and caregivers. However,
the current study only selected caregiver characteristics to predict the empathic accuracy. A
complete examination needs to include patients’ characteristics as contextual factors influencing
the concept of EASP.
Furthermore, caregivers’ psychological perceptions are restricted to their negative
perspectives as influences on their perceptions of patient symptoms. The current study only used
caregiver depressive symptoms to represent caregiver emotionality and used caregiver burden to
represent threats of caregiving that the caregiver perceived. Emotionality includes several kinds
of emotional expression such as anxiety and sadness. Meanwhile, there is little information about
the threats of caregiving from the caregivers’ perspective. The threats of caregiving that
caregivers perceived can be more precisely represented by caregivers’ fear and worry, and those
threats of caregiving can result not only from patients’ symptoms, but also from household and
financial hardship.
With respect to the findings above, the EASP framework is still in development. The
framework needs to be tested and revised, so that it can reveal the details of the relationship
between cancer patients’ and caregivers’ characteristics on congruence of symptom perception.
The whole framework also includes two outcomes—caregivers’ helping behavior and quality of
life of patients and caregivers–resulting from EASP. These relationships were not tested in the

144

current study.
Limitations
Regarding the discussion above, the present study has the following limitations:
1. This study was a cross-sectional study. Caregiver emotionality, threats from providing
care, and empathic accuracy may change because of the dynamic nature of symptoms and
caregivers’ perceptions of caregiving. A longitudinal study should be conducted to
understand the dynamic relationship among those three concepts and their contribution
on the health outcomes of both patients and caregivers.
2. This study focused on patients who were newly diagnosed with cancer and who were
receiving chemotherapy for the first-time. The result may not be generalized to other
cancer patients who are not newly diagnosed, who are not the first-time recipients of
chemotherapy and who are not receiving chemotherapy. Also, caregivers of those
patients who receive follow-up treatments may have different caregiving experiences in
recognizing patients’ symptoms, which may improve congruence of symptom perception.
3. This study used caregiver depressive symptoms to represent caregiver emotionality, and
it only showed a major effect on the congruence of pain severity. Caregiver emotionality
is described as a family caregiver’s emotional reactions toward providing cancer care,
including other individual emotions, such as worry, fear, and anxiety. In order to support
the current framework of EASP, further studies need to apply an emotion scale, for
example, the Profile of Mood States, which consists of six subscales for mood status,
including depression, fatigue, anxiety- tension, anger-hostility, vigor-activity, and
confusion-bewilderment. If this is done, caregivers’ emotionality will be evaluated more
thoroughly in order to test the effect of each emotional dimension on caregivers’ EASP.

145

4. Caregivers’ empathic accuracy is more likely to go through the cognitive process for
patients’ pain and fatigue because these symptom perceptions were represented by scores.
Davis’s empathy model shows that both cognitive and affective empathy occur
simultaneously, which may imply that it is difficult to determine which empathic
outcomes were produced from one process but not from the other. Further study of
caregivers’ empathic accuracy will specify cognitive and affective empathic processes
and their consequences and influencing factors in order to develop effective interventions
for specific caregiver populations that will improve their EASP.
Summary
The present study identified that family caregivers of cancer patients had fair-to-moderate
congruence of patients’ pain and fatigue severity and frequency, which was similar to the
congruence levels of symptom perception in prior studies. In spite of a small effect size, this
study shows that some factors significantly influence congruence of pain and fatigue severity and
frequency. The hypotheses were partially supported because of the fact and that caregiver
depression and caregiver self-esteem influenced congruence of pain severity, a result that
supports Hypothesis 1 of the framework. However, schedule interruption, living arrangement,
and total number of patient symptoms were in opposition with the hypotheses. Schedule
interruption can improve the accuracy of caregivers’ perceptions of patient symptoms. Living
with patients decreases the possibility of caregivers being congruent with patients on pain
severity. Caregivers’ awareness of total number of patient symptoms is a new and important
finding; it is the only variable which can influence congruence of all four symptom variables.
Caregivers who recognize a greater number of patient symptoms tend to be noncongruent with
patients’ pain severity, but they tend to be congruent with patients’ pain frequency, fatigue

146

severity, and fatigue frequency.
Because those findings did not completely fit the hypotheses, it is necessary to develop
more studies to examine the associations among the concepts and factors in the EASP
framework, including remaining questions. For example, how are caregivers’ helping behavior
between congruence and noncongruence groups of pain and fatigue severity and frequency
related to congruence of symptom perception? How are the patients’ and caregivers’ quality of
life related to congruence of symptom perception? Further studies also need to apply other
symptoms when examining this framework to identify the factors influencing EASP. The
communication between patients and family caregivers will be important in order for caregivers
to have accurate perception of patient symptoms. Answering those questions about the
association between the concepts related to EASP will have an important contribution for the
understanding of cancer family care.
Implications
For knowledge in nursing science. The gaps in the prior studies aimed at identifying
congruent symptom perceptions between cancer patients and caregivers have been a lack of
oriented theories or models. The studies mainly emphasized the unreliability of caregivers’
perception of patients’ symptoms because of low levels of congruence of symptom perception.
Approximately 30 studies showed noncongruence of cancer patients’ symptoms by comparing
patients’ self-reports and caregivers’ reports about patients’ symptoms; however, congruence of
symptom perception has not been conceptually defined. The current study has developed the
EASP framework by applying the theoretical perspectives of Ickes’ Empathic Accuracy in
psychological therapy and Davis’ Empathy model in psychology, and has engaged these two
perspectives of empathy with Given’s Family Care Model in nursing research. Davis (1994)

147

described empathy as people’s general temperament and Ickes (1993) described empathic
accuracy as the nature of everyday mind reading, so that being empathic naturally happens in
people’s daily lives. In addition, the conceptualization of empathic accuracy in Ickes’ study was
used to differentiate various relationships between the perceiver and the target person. Few
studies have used the concept of EASP for caregiving dyads.
For nursing clinical practice. Specifically, because caregiver burden and caregiver
depression are typical negative perceptions that result from caregiving, several studies have
investigated different strategies to reduce the risk of these two perceptions for caregivers (Honea,
et al., 2008; Rivera, 2009). The examination of the EASP framework in the present study
revealed that both caregiver burden and caregiver depression not only may represent caregivers’
outcomes, but also are important factors associated with caregivers’ accurate perception of
cancer patients’ symptoms. This conceptualization of the EASP that caregivers use both
cognitive and affective empathic processes in symptom management for patients can be extended
to assist daily activities.
The EASP framework shows that the greater empathic accuracy for patients’ symptom
experience, the higher likelihood for caregiver to help patients to alleviate symptom burden, in
that caregivers can better determine appropriate symptom management. The EASP framework
may drive nurses’ attention toward recognizing family caregivers’ characteristics, such as
caregiver depressive symptoms and caregiver burden, and caregiver demographics as well.
Therefore, nurses can deliver effective nursing interventions for caregivers to decrease their own
depressive symptoms and caregiver burden in order to to deal with patient symptoms.
Eventually, improving EASP can benefit at patients’ and caregivers’ quality of life.
Results of this study, in turn, showed that caregiver depression, caregiver schedule

148

interruption and self-esteem can impact congruence of pain severity and frequency. Because this
study only used data information collected at baseline interview, this result may reveal a point of
view that congruence of symptom perception may have indicated caregivers were less depressed
and had higher self-esteem but experienced schedule interruption for care involvement as well.
Therefore, congruence and overestimation may be a sign of caregiver burden. Clinical nurses
may utilize this result for early detection of caregiver depression and caregiver burden. When
nurses find that caregivers are congruent with or overestimate patients’ symptoms perceptions,
nurses need to be aware of caregiver depressive symptoms and caregiver burden.
For future research. This study has developed the framework of EASP to conceptualize
empathic accuracy for nursing research that includes the influence of the factors of emotionality
and perceived threats of caregiving. However, the EASP framework is still in development, and
needs further study to support the relationships among the concepts. Future research can be
conducted:
(1) To develop the EASP framework.
The framework of EASP can guide the investigator to develop hypotheses that can
account for variations in EASP between cancer patients and their caregivers. Although
approximately 30 studies identified family caregivers’ reports of patients’ symptoms, few
extended their investigations to identify how caregiver characteristics, such as emotionality and
perceived threats of caregiving, may point out the risk factors of empathic accuracy. The present
study tested the predictive relationship between congruence of symptom perception, caregivers’
emotionality and threats to caregiving. This study also considered the influence on caregiver
demographics, and has the potential to generate knowledge of empathy accuracy in family care.
Furthermore, the EASP framework shows that there is a sequential correlation between

149

empathic accuracy and patients’ health outcomes. These sequential relationships between
empathic accuracy and outcomes of patients and caregivers should be tested in the subsequent
studies. Nursing research can begin to establish evidenced research to examine the relationship
between empathic accuracy and caregivers’ helping behaviors, enriching family cancer care
research.
Based on the theoretical framework of EASP, although it has not been tested yet, the
investigator assumes that communication and negotiation between patients and family caregivers
and between health care systems and families can influence the empathic accuracy. The Schulz
(2007) study also pointed out that communication can modulate the effect of patient symptom
burden on caregiver physical and psychological health. Further research must be developed to
test how caregivers use verbal and nonverbal communication with patients, as well as
communication with health care teams, in order to promote patient care. Because caregiver
health is positively correlated with patient health, intervention developed to alleviate patients’
symptom burden will help improve caregivers’ health.
(2) To design interventions to support the empathic accuracy.
Nursing researchers focused on cancer care also have suggested that interventions should
be delivered for both patients and caregivers to advocate patients’ self-management and
caregivers’ training, which will be beneficial for their health and quality of life (Northouse, et al,
2010 & Selviria, C Given, and B. Given, et al., 2012). The Schulz (2007) study also suggested
that research should develop interventions to decrease patient suffering and improve caregiver
compassion to contribute to caregiver health, based on its conceptual framework which
illustrates the relationship between patient suffering and caregiver compassion.
Hopkinson et al. (2012) reviewed the studies which conducted interventions for cancer
patients and their family caregivers to identify effects of interventions, delivering to the patient150

caregiver pairs and separately delivering to patients and their caregivers to improve health
outcomes of cancer patients. They found that couple-based interventions may assist emotional
support and improve dyadic coping. Therefore, future research regarding the EASP framework
should collect dyadic participants of cancer patients and caregivers in order to test the effect of
interventions related to alleviating patient symptom burden and caregiver burden.
3) To extend the concept EASP from cancer patients and their family caregivers to the
patients who are diagnosed with diseases other than cancer.
The EASP framework is in development and focuses on the caregiving circumstance of
cancer symptom perception between patient and caregivers. Pain and fatigue severity and
frequency were selected because they are typical symptoms in the cancer population. On the
other hand, different patient populations have their own typical symptoms which may need more
assistance from family caregivers. For example, patients diagnosed with Alzheimer’s disease
need more family caregiver assistance with their daily activities when they have memory
retardation and physical disabilities. In that case, EASP is a typical phenomenon in the
caregiving circumstance. Given the statement above about the different symptoms between two
diseases, the factors of the concept EASP will be different. The same determinants of the concept
EASP among the diseases will indicate the generalization of the EASP framework. Therefore,
further studies should utilize this framework in examining the different patient populations to
strengthen the generalization.
For health policy. Caregiver health issues receive little attention in public health, social
service policy and research because these areas focus on improving the health of populations
who have illnesses and who are disabled. Family caregivers are an invisible and potential patient
population because their health is impacted by patients’ symptom burden and care demands.

151

The EASP framework (Figure 1) emphasizes that caregivers’ EASP is influenced by their
psychological perceptions, modulates their helping behaviors in assisting patient symptom
management, and can consequently impact their own and patients’ health. Schulz and his
colleagues (2007) have pointed out that patient suffering and caregiver compassion often impact
family caregiver health. Caregivers who are more involved in caregiving performance generate
more caregiver burden and caregiver depression (Palos et al., 2010; Sherwood, et al., 2006). The
document of the Institute of Medicine (IOM) illustrates that 90% of long-term care is provided
by family caregivers because the health care systems encourage minimum hospital days, and life
expectancies are longer than before. Symptom management occupies most of the caregivers’
time (von Ray et al., 2011). Family caregivers still report that they are not prepared to provide
patient care. Few of them receive the training or have sufficient resource for providing patient
care.
The IOM document “Retooling for An Aging America: Building the health care
workforce” (2008) suggests that the health care workforce must include every individual who
provides patient care, including formal health care providers, paid care workers, and informal
caregivers (family members and friends), as well as patients themselves. This document also
suggests that all of these individuals must have the essential resources, knowledge, and tools that
are required to provide high-quality health care. In sum, the focus of these two documents is the
necessity of supporting adequate knowledge related to patient care for family caregivers to
ensure high quality of care for patients.
Therefore, although health policy may not directly improve caregiver burden and
caregiver depression, it might change the clinical environment in providing care resources for
family caregivers. Health policy may also focus on advocating for caregiver health in the health

152

care system because caregivers’ empathic accuracy in providing symptom management can
alleviate patient symptom burden, and may also decrease caregiver depression. Congress should
assign the IOM to conduct an aggregate level study to understand the difficulties and factors that
influence caregivers who provide symptom management in home-based environments.
Moreover, family caregivers are the main providers for most patient care at home. The National
Family Caregivers Association’s (NFCA, 2008-2009) has recommended establishing familycentered health care, and regulating comprehensive assessments of home-based caregiving
situations to determine what family assistance caregivers may require. Therefore, the health
policy related to supporting entire family care will be encouraged as well.
Conclusion
The EASP framework was developed with the inspiration of psychological perspectives
of empathy and empathic accuracy in the caregiving circumstance to displays the factors and
outcomes of caregivers and patients related to congruence of patient symptoms. Some caregivers
were able to be congruent with patients’ pain and fatigue perception, although some of them
typically overestimated those symptoms. Caregivers who have depressive symptoms and
caregiver burden may be more aware of patients’ pain than fatigue, because their depressive
symptoms and caregiver reactions related to patient care can influence their accurate perception
of patients’ pain. Caregivers’ living with patients may influence caregivers’ objective perception
of patients’ symptoms. Caregivers who perceive a greater number of patients’ symptoms tend to
accurately perceive patients’ pain and fatigue rather than to underestimate these symptoms.
This study can compensate for the gaps in the past and extend nursing science by
engaging psychology field (empathy theories) with caregiving research to provide a new
perspective on caregiver congruence. Because the EASP framework was in early development

153

and the investigator only examined one part of this framework, an essential task is to completely
test the EASP framework. Meanwhile, testing the EASP in different ill populations and their
family members will improve the generalization of this framework. More research in the area of
empathic accuracy within the relationship of caregiving dyads will extend the science of
empathy of family caregivers.

154

APPENDICES

155

APPENDIX A
Instruments

156

Table A1
Patient Demographics
1. Gender of patient

Male (1)

2. What is your birthday (write in):

Female (2)

____ /___ /____
month/ day/ year

3. What is your highest level of education
completesd? (circle one)

1) No formal education
2) Completed grade school
3) Completed some high school
4) Completed high school
5) Compelted some college or technical
training
6) Completed college
7) Complete graduate/professional degree
(post baccalaureate degree)

4. What is your race or ethnic
background?

1) Caucasion/white
2) Afraican American/Black
3) Mexican American/Hispanic/Chicano
4) Native American/Alaskan
5) Oriental/Ascian/Pacific Islander
6) Other (Specify): ______________

5. What is your current marital status?

1) Never married
2) Married
3) Divorce/Separated
4) Widowed
5) Living together

157

Table A1 (cont’d)
6. Including yourself, how many people
live in your house?_____________

a. How many younger than 13 years old?______
b.How many between the ages of 13 and 17 ?
______
c. How many 18 years or older?_________

7. Site of Cancer

1. Breast
2. Colon
3. Lung (non-small)
4. Lung (small cell)
5. GU
6. Gastrointestinal
7. Gynecological
8. Pancreas
9. Non-Hodgkin’s
10. Menothelioma
11. Other

8. Stage of Cancer

1. Early
2. Late

9. Recurrence of Cancer

1. Yes
2. No

10. Metastatic Cancer

1. Yes
2. No

158

Table A2
Caregiver Demographics
1

Gender of patient

(1)Male

2.

What is your birthday (write in):

(2)Female

____ /___ /___________
month/ day/ year

3.

What is your relationship to ________________________, (patient’s name) who has
cancer? You are his/her…. (check one.)
1) Spouse
2) Parent

8) Neice/nephew

3) Daughter/ son

9) Aunt/uncle

4) Daughter/ son-in- law

10) Other relative ____________________

5) Sister/brother

11) Friend or companion with hwom you live

6) Sister/brother-in-law
4.

7) Granddaguther/son

12) Other non-relative____________________

What is your highest level of
education completesd? (circle one)

1) No formal education
2) Completed grade school
3) Completed some high school
4) Completed high school
5) Compelted some college or technical training
6) Completed college
7) Complete graduate/professional degree (post
baccalaureate degree)

159

Table A2 (cont’d)
Caregiver Demographics
5.

What is your race or ethnic
background?

1) Caucasion/white
2) Afraican American/Black
3) Mexican American/Hispanic/Chicano
4) Native American/Alaskan
5) Oriental/Ascian/Pacific Islander
6) Other (Specify): ______________

6. What is your current marital status?

1) Never married
2) Married
3) Divorce/Separated
4) Widowed
5) Living together

7. Including yourself, how many people
live in your house?_____________

a. How many younger than 13 years old?______
b. How many between the ages of 13 and 17?
_______
c. How many 18 years or older?_________

8.

Did you move into _____________________’s home or did ____________ move into
your home so that you could assist him/her with his/her cancer care? (check one)
(1) Caregiver moved
(2) Patient moved
If a move occurred, what was the date of the move? (write in): ____/_______/__________
Moth/ day / year (4-digit needed)

160

Table A 3
Patient symtpoms
Patient version (cut)
1 During the past 7 days, on how many days did you experience pain?
0

1

2

3

4

5

6

Not present

7
Number of days

2 On a scale of 0 is not present to 10 is the worst it could be, how severe in pain (for you) in the
past 7 dyas?
0

1

2

3

4

5

6

7

8

9

Not present

10

Worst possible

3 During the past 7 days, on how many days did you experience fatigue?
0

1

2

3

4

5

6

Not present

7
Number of days

4 On a scale of 0 is not present to 10 is the worst it could be, how severe in fatigue (for you) in
the past 7 dyas?
0

1

2

3

4

5

Not present

6

7

8

9

10

Worst possible

5. The total number of Patients’ symptoms was reported by patients? ________________

161

Table A3 (cont’d)
Patient symtpoms
Caregiver version (cut)
1 During the past 7 days, on how many days did you experience pain?
0

1

2

3

4

5

6

Not present

7
Number of days

2 On a scale of 0 is not present to 10 is the worst it could be, how severe in pain (for you) in the
past 7 dyas?
0

1

2

3

4

5

6

7

8

9

Not present

10

Worst possible

3 During the past 7 days, on how many days did you experience fatigue?
0

1

2

3

4

5

6

Not present

7
Number of days

4 On a scale of 0 is not present to 10 is the worst it could be, how severe in fatigue (for you) in
the past 7 dyas?
0

1

2

3

4

5

Not present

6

7

8

9

10

Worst possible

5 The total number of Patients’ symptoms was reported by CG? ________________

162

CES-D SCALE
Below is a list of the ways you might have felt or behaved. How often have you felt this way
during the past week? Please circle the answer that fits your situation best.
Table A4
CES-D Scale
Items

Rarely or
none of
the time
(<1 day)

Some or
little of
the time
(1-2
days)

Moderate
amount of
time
(3-4 days)

Most or
all of
the time
(5-7
days)

1

I was bothered by things that usually
don’t bother me.

0

1

2

3

2

I did not feel like eating; my appetite was
poor.

0

1

2

3

3

I felt that I could not shake off the blues
even with help from my family and
friends.

0

1

2

3

4

I felt that I was just as good as other
people.

0

1

2

3

5

I had trouble keeping my mind on what I
was doing.

0

1

2

3

6

I felt depressed.

0

1

2

3

7

I felt that everything I did was an effort.

0

1

2

3

8

I felt hopeful about the future.

0

1

2

3

9

I thought my life had been a failure.

0

1

2

3

10 I felt fearful.

0

1

2

3

11 My sleep was restless.

0

1

2

3

12 I was happy.

0

1

2

3

13 I talked less than usual.

0

1

2

3

14 I felt lonely.

0

1

2

3

15 People were unfriendly.

0

1

2

3

16 I enjoyed life.

0

1

2

3

17 I had crying spells.

0

1

2

3

18 I felt sad.

0

1

2

3

163

Table A4 (cont’d)

19 I felt that people dislike me.

0

Some or
little of
the time
(1-2
days)
1

20 I could not get “going.”

0

1

Items

Rarely or
none of
the time
(<1 day)

164

Moderate
amount of
time

2

Most or
all of
the time
(5-7
days)
3

2

3

(3-4 days)

FCRP Research Instruments
Caregiver Reaction Assessment Instrument (CRA)
Table A5
Caregiver Reaction Assessment Instrument (CRA)

Items

Strongly Disagree Neither Agree Strongly
disagree
agree nor
agree
disagree

1.

I feel privileged to care for (patients’
name)
.

1

2

3

4

5

2.

Others have dumped caring for ___
onto me

1

2

3

4

5

3.

My activities are centered around care
for ___.

1

2

3

4

5

4.

It is very difficult to get help from my
family in taking care of ___.

1

2

3

4

5

5.

*I resent having to take care of ___.

1

2

3

4

5

6.

I have to stop in the middle of work

1

2

3

4

5

7.

I really want to care for ___.

1

2

3

4

5

8.

I visit family and friends less since I
have been caring for ___.

1

2

3

4

5

9.

I will never be able to do enough
caregiving to repay ___.

1

2

3

4

5

10. *My family works together at caring
for ___.

1

2

3

4

5

11. I have eliminated things from my
schedule since caring for ___.

1

2

3

4

5

12. Since caring for ___, I feel my family
has abandoned me.

1

2

3

4

5

13.

Caring for ___ makes me feel
good.

1

2

3

4

5

14. The constant interruptions make it
difficult to find time for relaxation.

1

2

3

4

5

15.

1

2

3

4

5

16. My family (brothers, sisters, and
children) left me alone to care for ___.

1

2

3

4

5

17. I enjoy caring for ___.

1

2

3

4

5

Caring for ___ is important to me.

165

Table A5 (con’d)

Note: * These questions are to be reversely;
Two subscales did not be presented in this study: the health impact and the financial burden.
Items of the Schedule Interruption: 3, 6, 8, 11, & 14; Items of the Family abandonment: 2, 4, 10,
12, & 16 ; Items of the Self-esteem:1, 5, 7, 9, 13, 15, &17
Additional information about this tool can be found at: Given, C.W., Given, B., Stommel,
M., Collins, C., King, S., & Franklin, S. (1992). The caregiver reaction assessment (CRA) for
caregivers to persons with chronic physical and mental impairments. Research in Nursing and
Health, 15(4), 271-283.
Note: No further information is available on this tool beyond what is included in the article
referenced above.

166

Appendix B
Data Analysis Flow Chart

167

Data Analysis Flow Chart
Merged two data sets- cancer patients and their
caregivers by their shared same research ID
Selected both patients and caregivers reported
patients’ pain and fatigue severity and frequency

225 Dyads of cancer patients and family

Recode variables:
Sex: male 0, female ‘1’ & male ‘0’
Education: ≥college ‘1’ & < college ‘0’
Spouse: spouse ‘1’ & non-spouse ‘0’
Living arrangement: together ‘1’ & separate ‘0’
CESD: ≥16 ‘1’& < 16 ‘0’
ANOVA: Test the differences of age, total number of patients’ symptoms, CESD, schedule
interruption, family abandonment, self-esteem among Underestimation, Congruence, &
Overestimation.
Chi-square: Test the difference of caregiver sex (male/female), education (college or not),
relationship to patients (spouse or non-spouse), living arrangement (together/separate)
1) By Underestimation, Congruence, Overestimation.
2) By 2 CESD levels-- ≥16 and < 16
T test: Caregiver age, total number of patients’ symptoms, CESD, schedule interruption, family
abandonment, self-esteem BY binominal factors: caregiver sex (male/female), education
(college or not), relationship to patients (spouse/ non-spouse), living arrangement
(together/separate)
Pearson’s correlation: CESD, caregiver burden, and caregivers’ report of patients’ symptom
severity, and congruent pain severity, caregiver age, and total number of patients’
symptoms
GLM: Regression model of caregiver burden
Loglinear regression: Regressions model of caregiver depression (CESD ≥16 and < 16)

Multinomial log-linear regression:
Congruence models of pain severity
Congruence models of fatigue severity
Congruence model of pain frequency
Congruence model of fatigue frequency

168

APPENDIX C
IRB Approval Letters

169

Renewal
Application
Approval
November 16, 2009
To: Charles GIVEN
B510 W. Fee Hall
MSU
Re: IRB# 08-963 Category: EXPEDITED 2-5
Renewal Approval Date: November 13, 2009
Project Expiration Date: November 12, 2010
Title: Concordance between caregivers and recipients perceive cancer-related symptoms impacts
on caregiver burden.
The Institutional Review Board has completed their review of your project. I am pleased to
advise you that the renewal has been approved.
The review by the committee has found that your renewal is consistent with the continued
protection of the rights and welfare of human subjects, and meets the requirements of MSU's
Federal Wide Assurance and the Federal Guidelines (45 CFR 46 and 21 CFR Part 50). The
protection of human subjects in research is a partnership between the IRB and the investigators.
We look forward to working with you as we both fulfill our responsibilities.
Renewals: IRB approval is valid until the expiration date listed above. If you are continuing your
project, you must submit an Application for Renewal application at least one month before
expiration. If the project is completed, please submit an Application for Permanent Closure.
Revisions: The IRB must review any changes in the project, prior to initiation of the change.
Please submit an Application for Revision to have your changes reviewed. If changes are made
at the time of renewal, please include an Application for Revision with the renewal application.
Problems: If issues should arise during the conduct of the research, such as unanticipated
problems, adverse events, or any problem that may increase the risk to the human subjects, notify
the IRB office promptly. Forms are available to report these issues.
Please use the IRB number listed above on any forms submitted which relate to this project, or
on any correspondence with the IRB office.
Good luck in your research. If we can be of further assistance, please contact us at 517-355-2180
or via email at IRB@msu.edu. Thank you for your cooperation.

Sincerely,

170

Ashir Kumar, M.D.
BIRB Chair
c: Chia Tai Hung
1537 Spartan Village, Apt.B
East Lansing
MI

171

Closure of
Protocol
October 18, 2011
To: Denise Saint Arnault
B-510 West Fee Hall
Re: IRB# 08-963 Category: EXPEDITED 2-5
Project Expiration Date: November 10, 2011
Title: Caregiver-patient congruent symptom perception of pain and fatigue {CGA 118534)
Thank you for submitting a closure form for your project. The Institutional Review Board has
accepted the information you have provided and your protocol has been officially closed.
Closure: Closure indicates that all research activities (including data collection and analysis) are
complete and have stopped. If your database is completely de-identified, you can continue to
perform data analysis, as the research no longer meets the definition of human subject research.
Guidance on de-identified data: http://www.humanresearch.msu.edu/regs/regs_index.htm
Storage of Archived Data: The confidentiality of identifiable information must be maintained
even after the research is complete. If you are maintaining data with identifiers, it is your
responsibility to continue to maintain the privacy and confidentiality of the subjects and to store
and protect the data in accordance with university and federal policy.
When individuals involved in research projects at Michigan State University leave the
University, they may take copies of research data for project on which they have worked.
The PI must, however, retain original data at Michigan State University.
If a PI leaves Michigan State University, and a project is to be moved to another institution,
ownership of the data may be transferred with the approval of the Vice President for
Research and Graduate Studies and with written agreement from the PI's new institution
that guarantees 1) its acceptance of custodial responsibilities for the data and 2) MSU's
access to the data should that become necessary.
IRB records relating to research must be retained for at least 3 years after research is
completed or longer as required by law. 45 CFR 46.115(b), 21 CFR 56.109(f). This includes
documents under the investigator's) control, such as the signed informed consent forms and
approval letters. To satisfy IRB requirements, IRB records for this project must be retained until
at least 3 years. If your research is subject to other laws or requirements that impose a longer
record retention period, records must be retained for the longer period.
Future Use of Archived Data: If the data is de-identified, further use of this data by students or
colleagues does not require IRB approval. If your stored data contains personal identifiers, future
use may require IRB approval - please contact the IRB office for advice when needed.

172

Problems: While your research is complete and closed, you may still encounter unanticipated
problems, adverse events, or other problems that may increase the risk to the human subjects. In
this event please notify the IRB office promptly - forms are available to report these issues.
Please use the IRB number listed above on any forms submitted which relate to this project, or
on any correspondence with the IRB office.
Thank you for your cooperation. If we can be of further assistance, please contact us at 517-3552180 or via email at IRB@msu.edu.
Sincerely,
Ashir Kumar, M.D.
BIRB Chair

c: Chia Tai Hung, Charles Given

173

REFERENCES

174

REFERENECES

ACS. (2010). Caregivers Retrieved 9/21, 2011, from
http://www.cancer.org/Treatment/Caregivers/index
Armstrong, T. S. (2003). Symptoms experience: a concept analysis. Oncol Nurs Forum, 30(4),
601-606. doi: 10.1188/03.ONF.601-606
Bakas, T., & Burgener, S. C. (2002). Predictors of emotional distress, general health, and
caregiving outcomes in family caregivers of stroke survivors. Topics in stroke
rehabilitation, 9(1), 34-45.
Baker, F., Denniston, M., Smith, T., & West, M. M. (2005). Adult cancer survivors: how are
they faring? [Research Support, Non-U.S. Gov't]. Cancer, 104(11 Suppl), 2565-2576.
doi: 10.1002/cncr.21488
Barrentine, W. L., Street, J. E., & Kurtz, M. E. (1984). Postemergence Control of Red Rice
(Oryza-Sativa). Weed Science, 32(6), 832-834.
Berg, J. A., & Woods, N. F. (2009). Global women's health: a spotlight on caregiving. Nurs Clin
North Am, 44(3), 375-384. doi: S0029-6465(09)00036-X [pii]
10.1016/j.cnur.2009.06.003
Bewick, V., Cheek, L., & Ball, J. (2003). Statistics review 7: Correlation and regression.
[Review]. Critical care, 7(6), 451- 459. doi: 10.1186/cc2401
Beyer, C. (2011, May 10, 2011). Edmund Husserl. The Stanford Encyclopedia of Philosophy
Summer 2011. from Retrieved Auguest 18, 2011
http://plato.stanford.edu/archives/sum2011/entries/husserl/.
Bieling, P. J., & Alden, L. E. (2001). Sociotropy, Autonomy, and the Interpersonal Model of
Depression: An Integration. Cognitive Therapy and Research, 25(2), 167-184. doi:
10.1023/a:1026491108540
Bower, J. E. (2007). Cancer-related fatigue: links with inflammation in cancer patients and
survivors. Brain Behav Immun, 21(7), 863-871. doi: S0889-1591(07)00077-3
[pii]10.1016/j.bbi.2007.03.013
Broberger, E., Tishelman, C., & von Essen, L. (2005). Discrepancies and similarities in how
patients with lung cancer and their professional and family caregivers assess symptom
occurrence and symptom distress. Journal of pain and symptom management, 29(6), 572583.
Bunting, S. (1988). The concept of perception in selected nursing theories. Nursing Science
Quarterly, 1(4), 168-174. doi: 10.1177/089431848800100410

175

Butt, Z., Wagner, L. I., Beaumont, J. L., Paice, J. A., Straus, J. L., Peterman, A. H., . . . Cella, D.
(2008). Longitudinal screening and management of fatigue, pain, and emotional distress
associated with cancer therapy. Support Care Cancer, 16(2), 151-159. doi:
10.1007/s00520-007-0291-2
Byar, K. L., Berger, A. M., Bakken, S. L., & Cetak, M. A. (2006). Impact of adjuvant breast
cancer chemotherapy on fatigue, other symptoms, and quality of life. Oncology Nursing
Forum, 33(1), E18-26.
Cameron, J. I., Franche, R. L., Cheung, A. M., & Stewart, D. E. (2002). Lifestyle interference
and emotional distress in family caregivers of advanced cancer patients. [Research
Support, Non-U.S. Gov't]. Cancer, 94(2), 521-527. doi: 10.1002/cncr.10212
CDC. (2009). Deaths: Final Data for 2009.”. National Vital Statistics Reports, 60(3).
Cella, D., Wagner, L., Cashy, J., Hensing, T. A., Yount, S., & Lilenbaum, R. C. (2007). Should
health-related quality of life be measured in cancer symptom management clinical trials?
Lessons learned using the functional assessment of cancer therapy. J Natl Cancer Inst
Monogr(37), 53-60. doi: 10.1093/jncimonographs/lgm009
Cella, D., Yount, S., Rothrock, N., Gershon, R., Cook, K., Reeve, B., . . . Rose, M. (2007). The
Patient-Reported Outcomes Measurement Information System (PROMIS): progress of an
NIH Roadmap cooperative group during its first two years. Med Care, 45(5 Suppl 1), S3S11. doi: 10.1097/01.mlr.0000258615.42478.5500005650-200705001-00002 [pii]
Chau, P. H., Kwok, T., Woo, J., Chan, F., Hui, E., & Chan, K. C. (2010). Disagreement in
preference for residential care between family caregivers and elders is greater among
cognitively impaired elders group than cognitively intact elders group. Int J Geriatr
Psychiatry, 25(1), 46-54. doi: 10.1002/gps.2296
Cicchetti, D. V. (2001). Methodological commentary the precision of reliability and validity
estimates re-visited: distinguishing between clinical and statistical significance of sample
size requirements J Clin Exp Neuropsychol, 23(5), 695-700. doi:
10.1076/jcen.23.5.695.1249
Clark, M. C., & Diamond, P. M. (2010). Depression in family caregivers of elders: a theoretical
model of caregiver burden, sociotropy, and autonomy. Res Nurs Health, 33(1), 20-34.
doi: 10.1002/nur.20358
Cleeland, C. S. (2007). Symptom burden: multiple symptoms and their impact as patientreported outcomes. J Natl Cancer Inst Monogr(37), 16-21. doi: 2007/37/16
[pii]10.1093/jncimonographs/lgm005
Cleeland, C. S., Mendoza, T. R., Wang, X. S., Chou, C., Harle, M. T., Morrissey, M., &
Engstrom, M. C. (2000). Assessing symptom distress in cancer patients: the M.D.
Anderson Symptom Inventory. Cancer, 89(7), 1634-1646. doi: 10.1002/10970142(20001001)89:7<1634::AID-CNCR29>3.0.CO;2-V

176

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (2nd ed.). New Jersey:
Lawrence Erlbaum Associates,.
Curt, G. A., Breitbart, W., Cella, D., Groopman, J. E., Horning, S. J., Itri, L. M., . . . Vogelzang,
N. J. (2000). Impact of cancer-related fatigue on the lives of patients: new findings from
the Fatigue Coalition. Oncologist, 5(5), 353-360.
Daly, B. J., Douglas, S., Lipson, A., & Foley, H. (2009). Needs of Older Caregivers of Patients
with Advanced Cancer. Journal of the American Geriatrics Society, 57, s293-s295. doi:
10.1111/j.1532-5415.2009.02516.x
Davis, M. A. (2009). A perspective on cultivating clinical empathy. Complementary Therapist in
Clinincal Practice, 15(2), 76-79.
Davis, M. H. (1994). Empathy: A social psychological approach. Madison, WI: Brown &
Benchmark Publishers.
Davis, M. H. (1996). Empathy: A social psychological approach. Madison, WI: Brown &
Benchmark.
Davis, M. H. (2006). Empathy. In J. E. Stets & J. H. Turner. (Eds.), Handbook of the Sociology
of Emotions (pp. 443-466). New York, NY: Springer.
Decety, J., & Ickes, W. (2009). The social neuroscience of empathy. Cambridge, MA: The MIT
Press.
Decety, J., & Jackson, P. L. (2004). The functional architecture of human empathy. Behavioral
and Cognitive Neuroscience Reviews, 3, 71-100.
Decety, J., & Meyer, M. (2008). From emotion resonance to empathic understanding: a social
developmental neuroscience account. Development and Psychopathology, 20(4), 10531080. doi: 10.1017/S0954579408000503
Delgado-Guay, M. O., & Bruera, E. (2008). Management of pain in the older person with cancer.
Oncology (Williston Park), 22(1), 56-61.
Diaz, N., Menjon, S., Rolfo, C., Garcia-Alonso, P., Carulla, J., Magro, A., . . . Gasquet, J. A.
(2008). Patients' perception of cancer-related fatigue: results of a survey to assess the
impact on their everyday life. Clinical & translational oncology : official publication of
the Federation of Spanish Oncology Societies and of the National Cancer Institute of
Mexico, 10(11), 753-757. doi: CLAT108 [pii]
Dodd, M., Janson, S., Facione, N., Faucett, J., Froelicher, E., Humphreys, J., . . . Rankin, S.
(2001). Advancing the science of symptom management. Journal of Advanced Nursing,
33(5), 668-676.
Donovan, K. A., & Jacobsen, P. B. (2007). Fatigue, depression, and insomnia: evidence for a
symptom cluster in cancer. Semin Oncol Nurs, 23(2), 127-135. doi:

177

S07492081(07)00014-9 [pii]10.1016/j.soncn.2007.01.004
Donovan, K. A., Jacobsen, P. B., Andrykowski, M. A., Winters, E. M., Balducci, L., Malik, U., .
. . McGrath, P. (2004a). Course of fatigue in women receiving chemotherapy and/or
radiotherapy for early stage breast cancer. J Pain Symptom Manage, 28(4), 373-380. doi:
S0885-3924(04)00295-7 [pii]10.1016/j.jpainsymman.2004.01.012
Donovan, K. A., Jacobsen, P. B., Andrykowski, M. A., Winters, E. M., Balducci, L., Malik, U., .
. . McGrath, P. (2004b). Course of fatigue in women receiving chemotherapy and/or
radiotherapy for early stage breast cancer. [Clinical Trial Comparative Study Controlled
Clinical Trial Research Support, U.S. Gov't, P.H.S.]. Journal of pain and symptom
management, 28(4), 373-380. doi: 10.1016/j.jpainsymman.2004.01.012
Dozeman, E., van Schaik, D. J., van Marwijk, H. W., Stek, M. L., van der Horst, H. E., &
Beekman, A. T. (2010). The center for epidemiological studies depression scale (CES-D)
is an adequate screening instrument for depressive and anxiety disorders in a very old
population living in residential homes. Int J Geriatr Psychiatry. doi: 10.1002/gps.2519
Dumont, S., Turgeon, J., Allard, P., Gagnon, P., Charbonneau, C., & Vézina, L. (2006). Caring
for a Loved One with Advanced Cancer: Determinants of Psychological Distress in
Family Caregivers Journal of Palliative Medicine, 9(4), 912-921. doi:
10.1089/jpm.2006.9.912.
Eisenberg, N., & Eggum, N. D. (2009). Ch. 6 Empathic responding: Sympathy and personal
distress. In J. Decety & W. Ickes (Eds.), The social neuroscience of empathy (pp. 71-84).
Cambridge, Massachusetts: MIT.
Eisenberg, N., Eggum, N. D., & Di Giunta, L. (2010). Empathy-related Responding:
Associations with Prosocial Behavior, Aggression, and Intergroup Relations. Social
issues and policy review, 4(1), 143-180. doi: 10.1111/j.1751-2409.2010.01020.x
Eisenberg, N., Fabes, R. A., Murphy, B., Karbon, M., Maszk, P., Smith, M., . . . Suh, K. (1994).
The relations of emotionality and regulation to dispositional and situational empathyrelated responding. J Pers Soc Psychol, 66(4), 776-797.
Erdfelder, E., Faul, F., & Buchner, A. (1996). GPOWER: A general power analysis program.
Behavior Research Methods, Instruments, & Computers, 28(1), 1-11.
Family Caregiving Alliance. (2009). Caregiving, from
http://www.caregiver.org/caregiver/jsp/publications.jsp?nodeid=345
Family Caregiving Alliance. (2011). Selected Caregiver Statistics Retrieved 9/21, 2011, from
http://www.caregiver.org/caregiver/jsp/content_node.jsp?nodeid=439
Fan, G., Filipczak, L., & Chow, E. (2007). Symptom clusters in cancer patients: a review of the
literature. Curr Oncol, 14(5), 173-179.
Fan, G., Hadi, S., & Chow, E. (2007). Symptom clusters in patients with advanced-stage cancer

178

referred for palliative radiation therapy in an outpatient setting. Support Cancer Ther,
4(3), 157-162. doi: VKR4074816667206 [pii]10.3816/SCT.2007.n.010
Faraone, S. V. (2008). Understanding Effect Size: How It's Measured and What It Means
Retrieved February 12, 2012, from http://www.medscape.org/viewarticle/569729
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: a flexible statistical
power analysis program for the social, behavioral, and biomedical sciences. Behav Res
Methods, 39(2), 175-191.
Ferguson, C. J. (2009). An Effect Size Primer: A Guide for Clinicians and Researchers.
Professional Psychology-Research and Practice, 40(5), 532-538. doi: Doi
10.1037/A0015808
Ferrario, S. R., Vitaliano, P., Zotti, A. M., Galante, E., & Fornara, R. (2003). Alzheimer's
disease: usefulness of the Family Strain Questionnaire and the Screen for Caregiver
Burden in the study of caregiving-related problems. Int J Geriatr Psychiatry, 18(12),
1110-1114.
Fletcher, B., Dodd, M., Schumacher, K., & Miaskowski, C. (2008). Symptom experience of
family caregivers of patients with cancer.
Fletcher, B. S., Paul, S. M., Dodd, M. J., Schumacher, K., West, C., Cooper, B., . . . Miaskowski,
C. A. (2008). Prevalence, severity, and impact of symptoms on female family caregivers
of patients at the initiation of radiation therapy for prostate cancer. Journal of Clinical
Oncology, 26(4), 599-605. doi: 10.1200/Jco.2007.12.2838
Fu, M. R., LeMone, P., & McDaniel, R. W. (2004). An integrated approach to an analysis of
symptom management in patients with cancer. [Review]. Oncology Nursing Forum,
31(1), 65-70. doi: 10.1188/04.ONF.65-70
Gadassi, R., Mor, N., & Rafaeli, E. (2011). Depression and empathic accuracy in couples: an
interpersonal model of gender differences in depression. Psychological Science, 22(8),
1033-1041. doi: 10.1177/0956797611414728
Giacalone, A., Talamini, R., Spina, M., Fratino, L., Spazzapan, S., & Tirelli, U. (2008). Can the
caregiver replace his/her elderly cancer patient in the physician-patient line of
communication? Supportive Care in Cancer, 16(10), 1157-1162. doi: 10.1007/s00520008-0406-4
Given, B. (1995). Believing and dreaming to improve cancer care. Oncology Nursing Forum,
22(6), 929-940.
Given, B., Given, C., Azzouz, F., & Stommel, M. (2001). Physical functioning of elderly cancer
patients prior to diagnosis and following initial treatment. Nurs Res, 50(4), 222-232.
Given, B., Given, C., & Sherwood, P. (2012). Caregiver burden In In A. Naeim, D. B. Reube &
P. A. Ganz (Eds.), Management of cancer in the older patient (pp. 241-248).

179

Philadelphia, PA: Elsevier.
Given, B., & Given, C. W. (1992). Patient and family caregiver reaction to new and recurrent
breast cancer. J Am Med Womens Assoc, 47(5), 201-206, 212.
Given, B., Given, C. W., McCorkle, R., Kozachik, S., Cimprich, B., Rahbar, M. H., & Wojcik,
C. (2002). Pain and fatigue management: results of a nursing randomized clinical trial.
Oncol Nurs Forum, 29(6), 949-956.
Given, B., Kozachik, S., Collins, C., DeVoss, D., & Given, C. W. (2001). Caregiver role strain. .
In K. C. B. M. Maas, M. Hardy, T. Tripp-Reimer, M. Titler & J. Specht (Ed.), Nursing
Care of Older Adults: Diagnoses, Outcomes & Interventions (pp. 679-695). St. Louis:
MS: Mosby, Inc.
Given, B., Wyatt, G., Given, C., Sherwood, P., Gift, A., DeVoss, D., & Rahbar, M. (2004).
Burden and depression among caregivers of patients with cancer at the end of life. Oncol
Nurs Forum, 31(6), 1105-1117.
Given, B. A., Given, C. W., Sikorskii, A., & Hadar, N. (2007). Symptom clusters and physical
function for patients receiving chemotherapy. Semin Oncol Nurs, 23(2), 121-126.
Given, B. A., Sherwood, P., & Given, C. W. (2011). Support for caregivers of cancer patients:
transition after active treatment. Cancer Epidemiol Biomarkers Prev, 20(10), 2015-2021.
doi: 20/10/2015 [pii]10.1158/1055-9965.EPI-11-0611
Given, C. W., Given, B., Rahbar, M., Jeon, S., McCorkle, R., Cimprich, B., . . . Bowie, E.
(2004). Effect of a cognitive behavioral intervention on reducing symptom severity
during chemotherapy. J Clin Oncol, 22(3), 507-516. doi: 10.1200/JCO.2004.01.241
Given, C. W., Given, B., Stommel, M., Collins, C., King, S., & Franklin, S. (1992). The
caregiver reaction assessment (CRA) for caregivers to persons with chronic physical and
mental impairments. Res Nurs Health, 15(4), 271-283. doi: 10.1002/nur.4770150406
Glajchen, M. (2004). The emerging role and needs of family caregivers in cancer care. J Support
Oncol, 2(2), 145-155.
Goedendorp, M. M., Gielissen, M. F. M., Verhagen, C. A. H., Peters, M., & Bleijenberg, G.
(2008). Severe fatigue and related factors in cancer patients before the initiation of
treatment.
Greig, M. L., Chow, E., Bovett, G., Abdolell, M., & Gillies, C. (2005). Level of concordance
between proxy and cancer patient ratings in brief painInventory. Support Cancer Ther,
3(1), 54-58. doi: PJ775P6T1474866T [pii]10.3816/SCT.2005.n.025
Grunfeld, E., Coyle, D., Whelan, T., Clinch, J., Reyno, L., Earle, C. C., . . . Glossop, R. (2004).
Family caregiver burden: results of a longitudinal study of breast cancer patients and their
principal caregivers. CMAJ, 170(12), 1795-1801.

180

Hagelin, C. L., Wengstrom, Y., Ahsberg, E., & Furst, C. J. (2009). Fatigue dimensions in
patients with advanced cancer in relation to time of survival and quality of life. Palliat
Med, 23(2), 171-178. doi: 0269216308098794 [pii]10.1177/0269216308098794
Harden, J., Schafenacker, A., Northouse, L., Mood, D., Smith, D., Pienta, K., & al., e. (2002).
Couples' experiences with prostate cancer: focus group research. Oncology Nursing
Forum, 29(4), 701-709. doi: 10.1188/02.ONF.701-709
Hauser, J. M., Chang, C. H., Alpert, H., Baldwin, D., Emanuel, E. J., & Emanuel, L. (2006).
Who's caring for whom? Differing perspectives between seriously ill patients and their
family caregivers. Am J Hosp Palliat Care, 23(2), 105-112.
Higginson, I. J., & Gao, W. (2008). Caregiver assessment of patients with advanced cancer:
concordance with patients, effect of burden and positivity. Health and Quality of Life
Outcomes, 6(1), 42. doi: 10.1186/1477-7525-6-42
Hill, T., & Lewicki, P. (2011). Basic Statistics: Correlations. Electronic Statistic Textbook, from
Retrieved http://www.statsoft.com/textbook/basic-statistics/#Correlations
Hinton, J. (1996). How reliable are relatives' retrospective reports of terminal illness? Patients
and relatives' accounts compared. Soc Sci Med, 43(8), 1229-1236. doi:
0277953695004378 [pii]
Hoffman, A., Given, B. A., von Eye, A., Gift, A. G., & Given, C. (2007). Relationships among
pain, fatigue, insomnia, and gender in persons with lung cancer. Oncol Nurs Forum,
34(4), 785-792. doi: N0P1524836267717 [pii]10.1188/07.ONF.785-792
Honea, N. J., Brintnall, R., Given, B., Sherwood, P., Colao, D. B., Somers, S. C., & Northouse,
L. L. (2008). Putting Evidence into Practice: nursing assessment and interventions to
reduce family caregiver strain and burden. Clin J Oncol Nurs, 12(3), 507-516. doi:
J448565361481228 [pii]10.1188/08.CJON.507-516
Hopkinson, J. B., Brown, J. C., Okamoto, I., & Addington-Hall, J. M. (2012). The effectiveness
of patient-family carer (couple) intervention for the management of symptoms and other
health-related problems in people affected by cancer: a systematic literature search and
narrative review. [Research Support, Non-U.S. Gov't]. Journal of pain and symptom
management, 43(1), 111-142. doi: 10.1016/j.jpainsymman.2011.03.013
Hung, S.-Y., Pickard, A. S., Witt, W. P., & Lambert, B. L. (2007). Pain and depression in
caregivers affected their perception of pain in stroke patients. Journal of Clinical
Epidemiology, 60(9), 963-970. doi: 10.1016/j.jclinepi.2006.12.010
Ickes, W. (1993). Empathic accuracy. Journal of Personality, 61, 587-610. doi: 10.1111/j.14676494.1993.tb00783.x
Ickes, W. (1997). Empathic accuracy New York, NY: Guilford
Ickes, W. (2009). Empathic accuracy : Its links to clinical, cognitve, developmental, social, and

181

physiological psychology. In J. Decety & W. Ickes (Eds.), The social neuroscience of
empathy (pp. 57-70). Cambridge, MA: The MIT Press.
Jemal, A., Center, M. M., DeSantis, C., & Ward, E. M. (2010). Global patterns of cancer
incidence and mortality rates and trends. Cancer Epidemiol Biomarkers Prev, 19(8),
1893-1907. doi: 1055-9965.EPI-10-0437 [pii]10.1158/1055-9965.EPI-10-0437
Jemal, A., Siegel, R., Xu, J., & Ward, E. (2010). Cancer statistics, 2010. CA Cancer J Clin,
60(5), 277-300. doi: caac.20073 [pii]10.3322/caac.20073
Joinson, C. (1992). Coping with compassion fatigue. Nursing,, 22(4), 116-122.
Kelley, K. (2007). Confidence intervals for standardized effect sizes: Theory, application, and
implementation. Journal of Statistical Software, 20(8), 1-24.
Kim, E., Jahan, T., Aouizerat, B., Dodd, M., Cooper, B., Paul, S., . . . Wara, W. (2009). Changes
in symptom clusters in patients undergoing radiation therapy. Supportive Care in Cancer,
17(11), 1383-1391. doi: 10.1007/s00520-009-0595-5Open Access
Kim, H. J., & Abraham, I. L. (2008). Statistical approaches to modeling symptom clusters in
cancer patients. Cancer nursing, 31(5), E1-10. doi:
10.1097/01.NCC.0000305757.58615.c8
Kim, Y., Baker, F., Spillers, R. L., & Wellisch, D. K. (2006). Psychological adjustment of cancer
caregivers with multiple roles. Psychooncology, 15(9), 795-804. doi: 10.1002/pon.1013
Kim, Y., Duberstein, P. R., Sorensen, S., & Larson, M. R. (2005). Levels of depressive
symptoms in spouses of people with lung cancer: effects of personality, social support,
and caregiving burden. Psychosomatics, 46(2), 123-130.
Kim, Y., Kashy, D. A., & Evans, T. V. (2007). Age and attachment style impact stress and
depressive symptoms among caregivers: a prospective investigation. J Cancer Surviv,
1(1), 35-43.
Kim, Y., Schulz, R., & Carver, C. S. (2007). Benefit-finding in the cancer caregiving experience.
Psychosomatic Medicine, 69(3), 283-291. doi: 10.1097/PSY.0b013e3180417cf4
Kirkova, J., Davis, M. P., Walsh, D., Tiernan, E., O'Leary, N., LeGrand, S. B., . . . Russell, K. M.
(2006). Cancer symptom assessment instruments: a systematic review. J Clin Oncol,
24(9), 1459-1473. doi: 24/9/1459 [pii]10.1200/JCO.2005.02.8332
Kirkova, J., Walsh, D., Rybicki, L., Davis, M. P., Aktas, A., Tao, J., & Homsi, J. (2010).
Symptom severity and distress in advanced cancer. Palliat Med, 24(3), 330-339.
Kirou-Mauro, A., Harris, K., Sinclair, E., Selby, D., & Chow, E. (2007). Are Family Proxies a
Valid Source of Information About Cancer Patients' Quality of Life at the End-of-Life? A
Literature Review. Journal Of Cancer Pain & Symptom Palliation, 2(2), 23-33.

182

Kramer, M. S., & Feinstein, A. R. (1981). Clinical biostatistics. LIV. The biostatistics of
concordance. [Comparative Study]. Clinical pharmacology and therapeutics, 29(1), 111123.
Kurtz, M., & Brown, M. A. (1985). Postemergence Herbicides and Trifluralin for Control of
Rhizome Johnsongrass in Cotton. Mississippi Agricultural & Forestry Experiment
Station Bulletin(934), 1-6.
Kurtz, M., Kurtz, J., Given, C., & Given, B. (1996). Concordance of cancer patient and caregiver
symptom reports. Cancer Pract, 4(4), 185-190.
Kurtz, M., Kurtz, J., Given, C., & Given, B. (2004). Depression and physical health among
family caregivers of geriatric patients with cancer--a longitudinal view. Med Sci Monit,
10(8), CR447-456. doi: 4846 [pii]
Kurtz, M., Kurtz, J., Given, C., & Given, B. (2006). Effects of a symptom control intervention on
utilization of health care services among cancer patients. [Randomized Controlled Trial
Research Support, N.I.H., Extramural
Research Support, Non-U.S. Gov't]. Medical science monitor : international medical journal of
experimental and clinical research, 12(7), CR319-324.
Kurtz, M., Martinmorgan, D., & Graff, R. J. (1987). Recognition of the Beta-2 Microglobulin-B
Molecule by a Ctl Clone. Journal of Immunology, 138(1), 87-90.
Lee, H. S., Brennan, P. F., & Daly, B. J. (2001). Relationship of empathy to appraisal,
depression, life satisfaction, and physical health in informal caregivers of older adults.
Res Nurs Health, 24(1), 44-56. doi: 10.1002/1098-240X(200102)24:1<44::AIDNUR1006>3.0.CO;2-S [pii]
Lenz, E. R., Pugh, L. C., Milligan, R. A., Gift, A., & Suppe, F. (1997). The middle-range theory
of unpleasant symptoms: an update. ANS Adv Nurs Sci, 19(3), 14-27.
Li, Z., & Hicks, M. H. (2010). The CES-D in Chinese American women: construct validity,
diagnostic validity for major depression, and cultural response bias. Psychiatry Res,
175(3), 227-232. doi: 10.1016/j.psychres.2009.03.007
Lin, C. (2001). Congruity of cancer pain perceptions between Taiwanese patients and family
caregivers: relationship to patients' concerns about reporting pain and using analgesics. J
Pain Symptom Manage, 21(1), 18-26.
Lin, L., Hedayat, A. S., & Wu, W. (2007). A unified approach for assessing agreement for
continuous and categorical data. J Biopharm Stat, 17(4), 629-652. doi: 780214499
[pii]10.1080/10543400701376498
Lin, L., & Torbeck, L. D. (1998). Coefficient of accuracy and concordance correlation
coefficient: new statistics for methods comparison. PDA J Pharm Sci Technol, 52(2), 55-

183

59.
Lobchuk, M. (2006). Concept analysis of perspective-taking: Meeting informal caregiver needs
for communication competence and accurate perception. Journal of Advanced Nursing,
54(3), 330-341. doi: 10.1111/j.1365-2648.2006.03815.x
Lobchuk, M., & Degner, L. F. (2002). Symptom experiences: perceptual accuracy between
advanced-stage cancer patients and family caregivers in the home care setting. J Clin
Oncol, 20(16), 3495-3507.
Lobchuk, M., Degner, L. F., Chateau, D., & Hewitt, D. (2006). Promoting enhanced patient and
family caregiver congruence on lung cancer symptom experiences. Oncol Nurs Forum,
33(2), 273-282. doi: 10.1188/06.ONF.273-282
Lobchuk, M., & Vorauer, J. D. (2003). Family caregiver perspective-taking and accuracy in
estimating cancer patient symptom experiences. Soc Sci Med, 57(12), 2379-2384. doi:
S0277953603001321 [pii]
Lutz, S., Norrell, R., Bertucio, C., Kachnic, L., Johnson, C., Arthur, D., . . . Palardy, G. (2001).
Symptom frequency and severity in patients with metastatic or locally recurrent lung
cancer: a prospective study using the Lung Cancer Symptom Scale in a community
hospital. Journal of Palliative Medicine, 4(2), 157-165.
Magaziner, J., Simonsick, E. M., Kashner, T. M., & Hebel, J. R. (1988). Patient-proxy response
comparability on measures of patient health and functional status. [Comparative Study
Research Support, U.S. Gov't, Non-P.H.S.]. Journal of Clinical Epidemiology, 41(11), 10651074.
McGuire, D. B. (1995). Chapter 1: The multiple dimensions of cancer pain: A framework for
assessment and management. In D. B. McGuire, C. H. Yarbro & B. Ferrell (Eds.),
Cancer pain management (pp. 1-13): Jones and Bartlett.
McMillan, S. C., & Moody, L. E. (2003). Hospice patient and caregiver congruence in reporting
patients' symptom intensity. Cancer Nurs, 26(2), 113-118. doi: 00002820-20030400000004 [pii]
McPherson, C. J., Wilson, K. G., Lobchuk, M. M., & Brajtman, S. (2008). Family caregivers'
assessment of symptoms in patients with advanced cancer: concordance with patients and
factors affecting accuracy. J Pain Symptom Manage, 35(1), 70-82. doi:
j.jpainsymman.2007.02.038
Mercadante, S., Casuccio, A., & Fulfaro, F. (2000). The course of symptom frequency and
intensity in advanced cancer patients followed at home. [Clinical Trial]. Journal of pain
and symptom management, 20(2), 104-112.
Miaskowski, C. (2010). Outcome measures to evaluate the effectiveness of pain management in
older adults with cancer. Oncol Nurs Forum, 37 Suppl, 27-32. doi:X11875W3484TT5M9

184

[pii]10.1188/10.ONF.S1.27-32
Miaskowski, C., Cooper, B. A., Paul, S. M., Dodd, M., Lee, K., Aouizerat, B. E., . . . Bank, A.
(2006). Subgroups of patients with cancer with different symptom experiences and
quality-of-life outcomes: a cluster analysis. Oncol Nurs Forum, 33(5), E79-89. doi:
10.1188/06.ONF.E79-E89
Miaskowski, C., Zimmer, E. F., Barrett, K. M., Dibble, S. L., & Wallhagen, M. (1997).
Differences in patients' and family caregivers' perceptions of the pain experience
influence patient and caregiver outcomes. Pain, 72(1-2), 217-226. doi:
doi:10.1016/S0304-3959(97)00037-7
Mintz, S. (2007). Who are America's Family Caregivers? , from 9/21/11 Retrieved
https://www.thefamilycaregiver.org/who_are_family_caregivers/
Molassiotis, A., Brearley, S., Saunders, M., Craven, O., Wardley, A., Farrell, C., . . . Luker, K.
(2009). Effectiveness of a home care nursing program in the symptom management of
patients with colorectal and breast cancer receiving oral chemotherapy: a randomized,
controlled trial. J Clin Oncol, 27(36), 6191-6198. doi: JCO.2008.20.6755
[pii]10.1200/JCO.2008.20.6755
Molassiotis, A., Wilson, B., Brunton, L., Chaudhary, H., Gattamaneni, R., & McBain, C. (2010).
Symptom experience in patients with primary brain tumours: a longitudinal exploratory
study. Eur J Oncol Nurs, 14(5), 410-416. doi: S1462-3889(10)00054-2
[pii]10.1016/j.ejon.2010.03.001
Molassiotis, A., Zheng, Y., Denton-Cardew, L., Swindell, R., & Brunton, L. (2010). Symptoms
experienced by cancer patients during the first year from diagnosis: patient and informal
caregiver ratings and agreement. Palliat Support Care, 8(3), 313-324. doi:
S1478951510000118 [pii]10.1017/S1478951510000118
Montgomery, R. J. V., Gonyea, J. G., & Hooyman, N. R. (1985). Caregiving and the experience
of subjective and objective burden. Family Relations, 34, 19–26.
NAC and AARP. (2009). Caregiving in the U.S., 2009. Retrieved from
http://www.caregiving.org/data/Caregiving_in_the_US_2009_full_report.pdf
National Family Caregivers Association. (2010). Caregiver depression Retrieved September 21,
2011, from Retrieved
https://www.thefamilycaregiver.org/improving_caregiving/depression.cfm
National Institutes of Health State-of-the-Science, P. (2004). National Institutes of Health Stateof-the-Science Conference Statement: Symptom Management in Cancer: Pain,
Depression, and Fatigue, July 15–17, 2002. JNCI Monographs, 2004(32), 9-16. doi:
10.1093/jncimonographs/djg014
Nijboer, C., Triemstra, M., Tempelaar, R., Sanderman, R., & van den Bos, G. A. (1999).
Measuring both negative and positive reactions to giving care to cancer patients:

185

psychometric qualities of the Caregiver Reaction Assessment (CRA). Soc Sci Med, 48(9),
1259-1269.
O'Brien, J., & Francis, A. (1988). The use of next-of-kin to estimate pain in cancer patients.
Pain, 35(2), 171-178. doi: 10.1016/0304-3959(88)90224-2
Oberst, M. T., Thomas, S. E., Gass, K. A., & Ward, S. E. (1989). Caregiving demands and
appraisal of stress among family caregivers. Cancer Nursing, 12(4), 209-215.
Palos, G. R., Mendoza, T. R., Liao, K. P., Anderson, K. O., Garcia-Gonzalez, A., Hahn, K., . . .
Cleeland, C. S. (2010). Caregiver symptom burden: The risk of caring for an underserved
patient with advanced cancer. Cancer. doi: 10.1002/cncr.25695
Palos, G. R., Mendoza, T. R., Liao, K. P., Anderson, K. O., Garcia-Gonzalez, A., Hahn, K., . . .
Cleeland, C. S. (2011). Caregiver symptom burden: the risk of caring for an underserved
patient with advanced cancer. [Research Support, N.I.H., Extramural]. Cancer, 117(5),
1070-1079. doi: 10.1002/cncr.25695
Passik, S. D., & Kirsh, K. L. (2005). A pilot examination of the impact of cancer patients' fatigue
on their spousal caregivers. [Research Support, Non-U.S. Gov't]. Palliative & Supportive
Care, 3(4), 273-279.
Paul, S., Kurtz, M., & Mentzer, S. J. (2007). Osteomalacia associated with a fibroblast growth
factor-23 secreting chest wall tumor. [Case Reports]. The Journal of thoracic and
cardiovascular surgery, 134(3), 803-805. doi: 10.1016/j.jtcvs.2007.05.025
Pellegrino, R., Formica, V., Portarena, I., Mariotti, S., Grenga, I., Del Monte, G., & Roselli, M.
(2010). Caregiver distress in the early phases of cancer. Anticancer Res, 30(11), 46574663. doi: 30/11/4657 [pii]
Persson, C., Wennman-Larsen, A., Sundin, K., & Gustavsson, P. (2008). Assessing informal
caregivers' experiences: a qualitative and psychometric evaluation of the Caregiver
Reaction Assessment Scale. Eur J Cancer Care (Engl), 17(2), 189-199. doi: ECC833
[pii]10.1111/j.1365-2354.2007.00833.x
Phillips, A. C., Gallagher, S., Hunt, K., Der, G., & Carroll, D. (2009). Symptoms of depression
in non-routine caregivers: the role of caregiver strain and burden. Br J Clin Psychol,
48(Pt 4), 335-346. doi: bjcp714 [pii]10.1348/014466508X397142
Pickett, M., Barg, F. K., & Lynch, M. P. (2001). Development of a home-based family caregiver
cancer education program. The Hospice Journal, 15(4), 19-40.
Pinquart, M., & Sorensen, S. (2006). Gender differences in caregiver stressors, social resources,
and health: an updated meta-analysis. J Gerontol B Psychol Sci Soc Sci, 61(1), P33-45.
doi: 61/1/P33 [pii]
Portenoy, R. K., Thaler, H. T., Kornblith, A. B., Lepore, J. M., Friedlander-Klar, H., Coyle, N., .
. . et al. (1994). Symptom prevalence, characteristics and distress in a cancer population.

186

Qual Life Res, 3(3), 183-189.
Portney, L. G., & Watkins, M. P. (2000). Foundations of clinical research Applications to
practice. . New Jersey: Prentice Hall Inc. .
Preston, S. D., & de Waal, F. B. (2002). Empathy: Its ultimate and proximate bases. Behav Brain
Sci, 25(1), 1-20; discussion 20-71.
Pud, D., Ben Ami, S., Cooper, B. A., Aouizerat, B. E., Cohen, D., Radiano, R., . . . Miaskowski,
C. (2008). The symptom experience of oncology outpatients has a different impact on
quality-of-life outcomes. Journal of pain and symptom management, 35(2), 162-170. doi:
DOI 10.1016/j.jpainsymman.2007.03.010
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general
population. Applied Psychological Measurement, 1(3).
Redeker, N. S., Lev, E. L., & Ruggiero, J. (2000). Insomnia, fatigue, anxiety, depression, and
quality of life of cancer patients undergoing chemotherapy. Research and Theory for
Nursing Practice, 14(4), 275-290.
Redinbaugh, E. M., Baum, A., DeMoss, C., Fello, M., & Arnold, R. (2002). Factors associated
with the accuracy of family caregiver estimates of patient pain. J Pain Symptom Manage,
23(1), 31-38.
Reinhard, S. C., Given, B., Petlick, N. H., & Bemis, A. (2008). Chapter 14. Supporting Family
Caregivers in Providing Care. In R. G. Hughes (Ed.), Patient Safety and Quality: An
Evidence-Based Handbook for Nurses (Vol. 3): Agency for Healthcare Research and
Quality, U.S. Dept. of Health and Human Services,.
Resnizky, S., & Bentur, N. (2007). Can family caregivers of terminally ill patients be a reliable
source of information about the severity of patient symptoms? American Journal of
Hospice and Palliative Medicine, 23(6), 447.
Rhee, Y. S., Yun, Y. H., Park, S., Shin, D. O., Lee, K. M., Yoo, H. J., . . . Kim, N. S. (2008).
Depression in family caregivers of cancer patients: the feeling of burden as a predictor of
depression. [Research Support, Non-U.S. Gov't]. Journal of clinical oncology : official
journal of the American Society of Clinical Oncology, 26(36), 5890-5895. doi:
10.1200/JCO.2007.15.3957
Rivera, H. R. (2009). Depression symptoms in cancer caregivers. Clinical Journal of Oncology
Nursing, 13(2), 195-202. doi: 10.1188/09.CJON.195.202
Ryan, J. L., Carroll, J. K., Ryan, E. P., Mustian, K. M., Fiscella, K., & Morrow, G. R. (2007).
Mechanisms of cancer-related fatigue. Oncologist, 12 Suppl 1, 22-34. doi: 12/suppl_1/22
[pii]10.1634/theoncologist.12-S1-22
Schein, R. L., & Koenig, H. G. (1997). The Center for Epidemiological Studies-Depression
(CES-D) Scale: assessment of depression in the medically ill elderly. Int J Geriatr

187

Psychiatry, 12(4), 436-446. doi: 10.1002/(SICI)1099-1166(199704)12:4<436::AIDGPS499>3.0.CO;2-M
Schieman, S., & Turner, H. A. (2001). "When Feeling Other People's Pain Hurts": The Influence
of Psychosocial Resources on the Association between Self-Reported Empathy and
Depressive Symptoms. Social Psychology Quarterly, 64(4), 376-389.
Schulz, R., Hebert, R. S., Dew, M. A., Brown, S. L., Scheier, M. F., Beach, S. R., . . . Nichols, L.
(2007). Patient suffering and caregiver compassion: new opportunities for research,
practice, and policy. Gerontologist, 47(1), 4-13. doi: 47/1/4 [pii]
Shega, J. W., Hougham, G. W., Stocking, C. B., Cox-Hayley, D., & Sachs, G. A. (2004). Pain in
community-dwelling persons with dementia: frequency, intensity, and congruence
between patient and caregiver report. J Pain Symptom Manage, 28(6), 585-592.
Sherwood, P., Given, B., Given, C., Schiffman, R. F., Murman, D. L., Lovely, M., . . . Remer, S.
(2006). Predictors of distress in caregivers of persons with a primary malignant brain
tumor. Res Nurs Health, 29(2), 105-120.
Sherwood, P., Given, B., Given, C., Schiffman, R. F., Murman, D. L., von Eye, A., . . . Remer, S.
(2007). The influence of caregiver mastery on depressive symptoms. J Nurs Scholarsh,
39(3), 249-255. doi: 10.1111/j.1547-5069.2007.00176.x
Sherwood, P., Given, C., Given, B., & von Eye, A. (2005). Caregiver burden and depressive
symptoms: analysis of common outcomes in caregivers of elderly patients. J Aging
Health, 17(2), 125-147.
Shieh, G. (2008). Improved Shrinkage Estimation of Squared Multiple Correlation Coefficient
and Squared Cross-Validity Coefficient. Organizational Research Methods, 11(2), 387407. doi: 10.1177/1094428106292901
Sigurdardottir, K. R., & Haugen, D. F. (2008). Prevalence of distressing symptoms in
hospitalised patients on medical wards: A cross-sectional study. BMC Palliative Care,
7(1), 16.
Sikorskii, A., Given, C., Given, B., Jeon, S., Decker, V., Decker, D., . . . McCorkle, R. (2007).
Symptom management for cancer patients: a trial comparing two multimodal
interventions. J Pain Symptom Manage, 34(3), 253-264.
Sikorskii, A., Given, C., Given, B., Jeon, S., & McCorkle, R. (2006). Testing the effects of
treatment complications on a cognitive-behavioral intervention for reducing symptom
severity. J Pain Symptom Manage, 32(2), 129-139.
Silveira, M. J., Given, C. W., Given, B., Rosland, A. M., & Piette, J. D. (2010). Patient-caregiver
concordance in symptom assessment and improvement in outcomes for patients
undergoing cancer chemotherapy. Chronic Illn, 6(1), 46-56.
Simpson, J. A., Orina, M. M., & Ickes, W. (2003). When accuracy hurts, and when it helps: a test

188

of the empathic accuracy model in marital interactions. Journal of Personality and
Soccial Psychology, 85(5), 881-893. doi: 10.1037/0022-3514.85.5.881
Sneeuw, K. C., Aaronson, N. K., Osoba, D., Muller, M. J., Hsu, M. A., Yung, W. K., . . .
Newlands, E. S. (1997). The use of significant others as proxy raters of the quality of life
of patients with brain cancer. Med Care, 35(5), 490-506.
Soltow, D., Given, B., & Given, C. (2010). Relationship between age and symptoms of pain and
fatigue in adults undergoing treatment for cancer. Cancer Nurs, 33(4), 296-303. doi:
10.1097/NCC.0b013e3181ce5a1a
Stasi, R., Abriani, L., Beccaglia, P., Terzoli, E., & Amadori, S. (2003). Cancer-related fatigue.
Cancer, 98(9), 1786-1801. doi: 10.1002/cncr.11742
Stinson, L., & Ickes, W. (1992). Empathic accuracy in the interactions of male friends versus
male strangers. Journal of personality and social psychology, 62(5), 787-797.
Stommel, M., Wang, S., Given, C. W., & Given, B. (1992). Confirmatory factor analysis (CFA)
as a method to assess measurement equivalence. Res Nurs Health, 15(5), 399-405.
Stone, P., Ream, E., Richardson, A., Thomas, H., Andrews, P., Campbell, P., . . . Young, A.
(2003). Cancer-related fatigue--A difference of opinion? Results of a multicentre survey
of healthcare professionals, patients and caregivers (Vol. 12, pp. 20-20-27). United
Kingdom: Wiley-Blackwell Publishing Ltd.
Stone, P. C., & Minton, O. (2008). Cancer-related fatigue. Eur J Cancer, 44(8), 1097-1104. doi:
S0959-8049(08)00148-2 [pii]10.1016/j.ejca.2008.02.037
Stueber, K. (2008). Empathy. The Stanford Encyclopedia of Philosophy Fall 2008. Retrieved
September 28, 2011, from http://plato.stanford.edu/entries/empathy/
Tang, S. T., Liu, T. W., Lai, M. S., Liu, L. N., Chen, C. H., & Koong, S. L. (2006). Congruence
of knowledge, experiences, and preferences for disclosure of diagnosis and prognosis
between terminally-ill cancer patients and their family caregivers in Taiwan. Cancer
Invest, 24(4), 360-366. doi: H70927XH56X5337H [pii]10.1080/07357900600705284
Tranmer, J. E., Heyland, D., Dudgeon, D., Groll, D., Squires-Graham, M., & Coulson, K. (2003).
Measuring the symptom experience of seriously Ill cancer and noncancer hospitalized
patients near the end of life with the Memorial Symptom Assessment Scale. Journal of
Pain and Symptom Management, 25(5), 420-429. doi: 10.1016/s0885-3924(03)00074-5
Tu, M. S., & Chiou, C. P. (2007). Perceptual consistency of pain and quality of life between
hospice cancer patients and family caregivers: a pilot study. Int J Clin Pract, 61(10),
1686-1691.
Vallerand, A. H., Templin, T., Hasenau, S. M., & Riley-Doucet, C. (2007). Factors that affect
functional status in patients with cancer-related pain. Pain, 132(1-2), 82-90. doi: S03043959(07)00049-8 [pii]10.1016/j.pain.2007.01.029

189

van Dam, N. T., & Earleywine, M. (2010). Validation of the Center for Epidemiologic Studies
Depression Scale-Revised (CESD-R): Pragmatic depression assessment in the general
population. Psychiatry Res. doi: 10.1016/j.psychres.2010.08.018
van den Beuken-van Everdingen, M. H., de Rijke, J. M., Kessels, A. G., Schouten, H. C., van
Kleef, M., & Patijn, J. (2007). Prevalence of pain in patients with cancer: a systematic
review of the past 40 years. Ann Oncol, 18(9), 1437-1449. doi: mdm056
[pii]10.1093/annonc/mdm056
Van Houtven, C. H., Ramsey, S. D., Hornbrook, M. C., Atienza, A. A., & van Ryn, M. (2010).
Economic burden for informal caregivers of lung and colorectal cancer patients.
Oncologist, 15(8), 883-893. doi: theoncologist.2010-0005
[pii]10.1634/theoncologist.2010-0005
van Ryn, M., Sanders, S., Kahn, K., van Houtven, C., Griffin, J. M., Martin, M., & al., e. (2011).
Objective burden, resources, and other stressors among informal cancer caregivers: a
hidden quality issue? Psycho-Oncology, 20(1), 44-52. doi: 10.1002/pon.1703
Vitaliano, P. P., & Katon, W. J. (2006). Effects of Stress on Family Caregivers: Recognition and
Management.(psychiatrists support to the patients and their caregivers)(Statistical data).
Psychiatric Times 23(7), 24.
Vogelzang, N. J., Breitbart, W., Cella, D., Curt, G. A., Groopman, J. E., Horning, S. J., . . .
Portenoy, R. K. (1997). Patient, caregiver, and oncologist perceptions of cancer-related
fatigue: results of a tripart assessment survey. The Fatigue Coalition. Semin Hematol,
34(3 Suppl 2), 4-12.
Wennman-Larsen, A., Tishelman, C., Wengstrom, Y., & Gustavsson, P. (2007). Factors
influencing agreement in symptom ratings by lung cancer patients and their significant
others. J Pain Symptom Manage, 33(2), 146-155.
White, S. A., & van den Broek, N. R. (2004). Methods for assessing reliability and validity for a
measurement tool: a case study and critique using the WHO haemoglobin colour scale.
Stat Med, 23(10), 1603-1619. doi: 10.1002/sim.1804
Yeşilbalkan, Ö. U., & Okgün, A. (2010). Patients' self reports and caregivers' perception of
symptoms in Turkish cancer patients. Eur J Oncol Nurs, 14(2), 119-124. doi: S14623889(09)00103-3 [pii]10.1016/j.ejon.2009.08.002

190