Citation
Caregiver-patient relationship and stress response in multiple myeloma

Material Information

Title:
Caregiver-patient relationship and stress response in multiple myeloma
Creator:
Madore, Shannon Laura ( author )
Language:
English
Physical Description:
1 electronic file (167 pages) : ;

Subjects

Subjects / Keywords:
Cancer -- Patients -- Care ( lcsh )
Caregivers -- Counseling of ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Caregiving for a loved one with a chronic illness is a multifaceted experience that can be deleterious on overall physical and mental health, but can also be meaningful and rewarding. Previous research with caregivers has noted that caregivers of a family member diagnosed with cancer often experience a decline in physical health and psychological distress. However, there is limited research on caregivers of multiple myeloma patients. The primary goal of this cross sectional study of 51 caregivers of partners with multiple myeloma was to increase our understanding of the relationship between caregivers’ neuroendocrine and immune responses (oxytocin receptor genotype, C-reactive protein, Interleukin-6, Tumor necrosis factor-a), psychological processes (depression, negative affect, illness related distress), patient-caregiver relationship characteristics (relationship quality, attachment style), caregiver health behaviors (sleep, diet, physical activity) and quality of life during the caregiving experience. Fifty- one caregivers, (mean age of 62.9 (SD = 7.85) and 77percent female) were recruited from the Colorado Blood Cancer Institute (CBCI) at Presbyterian/St. Luke's Medical Center. Participants were included if they were in a romantic relationship and living with someone who was currently receiving or had recently received treatment for multiple myeloma at CBCI. Data was collected at CBCI via a psychosocial survey and a blood draw. Unfortunately, there were concerns regarding the validity of some of the biological measures that prevented the inclusion of oxytocin receptor genotype and TNF-a in the overall analysis. Bivariate correlational analyses and multiple regression analyses revealed four key findings; 1) physical activity levels were inversely related to level of plasma inflammation (CRP) after controlling for caregiver health variables, age, caregiver perception of disease severity, and length of illness time; 2) sleep quality was negatively associated with caregiver distress after controlling for perception of illness severity and length of illness time; 3) depression was inversely related to caregiver quality of life after controlling for perception of illness severity and length of illness time; and 4) depression was negatively associated with relationship quality after controlling for length of illness time and perception of illness severity. These results demonstrate a cross-sectional association between health behaviors (sleep and physical activity), distress, quality of life, and markers of immune response in those caring for a loved one with multiple myeloma. Future studies should examine the longitudinal relationship of these variables in order to understand the causal nature of these associations. The clinical implications of this study emphasize the importance of assessing levels of distress and health behaviors in caregivers of chronically ill partners, and providing resources and interventions to assist caregivers in maintaining their emotional and physical health with the potential for increasing long-term resilience.
Thesis:
Thesis (Ph.D) - University of Colorado Denver.
Bibliography:
Includes bibliographic references
System Details:
System requirements: Adobe Reader.
General Note:
Department of Psychology
Statement of Responsibility:
by Shannon Laura Madore.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
944457592 ( OCLC )
ocn944457592
Classification:
LD1193.L645 2015d M34 ( lcc )

Downloads

This item has the following downloads:


Full Text
CAREGIVER-PATIENT RELATIONSHIP AND STRESS RESPONSE IN MULTIPLE
MYELOMA
by
SHANNON LAURA MADORE
B.A., University of Colorado Boulder, 2008
M.A., University of Colorado Denver, 2012
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Clinical Health Psychology
2015


This thesis for the Doctor of Philosophy degree by
Shannon Madore
has been approved for the
Clinical Health Psychology Program
by
Kristin Kilboum, Advisor
Beth Allen, Chair
Teri Simoneau
Dave Albeck
October 6, 2015
11


Madore, Shannon Laura, (Ph.D., Clinical Health Psychology)
Caregiver-Patient Relationship and Stress Response in Multiple Myeloma
Thesis directed by Associate Professor, Kristin Kilbourn
ABSTRACT
Caregiving for a loved one with a chronic illness is a multifaceted experience that can be
deleterious on overall physical and mental health, but can also be meaningful and
rewarding. Previous research with caregivers has noted that caregivers of a family member
diagnosed with cancer often experience a decline in physical health and psychological
distress. However, there is limited research on caregivers of multiple myeloma patients.
The primary goal of this cross sectional study of 51 caregivers of partners with multiple
myeloma was to increase our understanding of the relationship between caregivers
neuroendocrine and immune responses (oxytocin receptor genotype, C-reactive protein,
Interleukin-6, Tumor necrosis factor-a), psychological processes (depression, negative
affect, illness related distress), patient-caregiver relationship characteristics (relationship
quality, attachment style), caregiver health behaviors (sleep, diet, physical activity) and
quality of life during the caregiving experience. Fifty- one caregivers, (mean age of 62.9
(SD = 7.85) and 77% female) were recruited from the Colorado Blood Cancer Institute
(CBCI) at Presbyterian/St. Luke's Medical Center. Participants were included if they were
in a romantic relationship and living with someone who was currently receiving or had
recently received treatment for multiple myeloma at CBCI. Data was collected at CBCI via
a psychosocial survey and a blood draw. Unfortunately, there were concerns regarding the
validity of some of the biological measures that prevented the inclusion of oxytocin
receptor genotype and TNF-a in the overall analysis. Bivariate correlational analyses and


multiple regression analyses revealed four key findings; 1) physical activity levels were
inversely related to level of plasma inflammation (CRP) after controlling for caregiver
health variables, age, caregiver perception of disease severity, and length of illness time;
2) sleep quality was negatively associated with caregiver distress after controlling for
perception of illness severity and length of illness time; 3) depression was inversely related
to caregiver quality of life after controlling for perception of illness severity and length of
illness time; and 4) depression was negatively associated with relationship quality after
controlling for length of illness time and perception of illness severity. These results
demonstrate a cross-sectional association between health behaviors (sleep and physical
activity), distress, quality of life, and markers of immune response in those caring for a
loved one with multiple myeloma. Future studies should examine the longitudinal
relationship of these variables in order to understand the causal nature of these associations.
The clinical implications of this study emphasize the importance of assessing levels of
distress and health behaviors in caregivers of chronically ill partners, and providing
resources and interventions to assist caregivers in maintaining their emotional and physical
health with the potential for increasing long-term resilience.
The form and content of this abstract are approved. I recommend its publication.
Approved: Kristin M. Kilboum
IV


ACKNOWLEDGMENTS
This study was funded by the Colorado Clinical Sciences and Translational Institute
TL1 Training Grant: TL1 TR001081. This study would have not been possible without the
help of numerous faculty and staff. I would like to thank Dr. Kristin Kilboum for her
dedicated mentorship, support, and guidance required to develop, obtain funding for,
implement and complete this project. I would also like to thank Dr. Teri Simoneau for her
guidance in the development and implementation of this project, specifically for her
dedication to the recruitment of subjects. I would like to thank my committee chair, Dr.
Elizabeth Allen, for her guidance and support with statistical analysis, the selection and
interpretation of measures, and her thoughtful comments and suggestions to help improve
the written document. I would like to thank Dr. Dave Albeck for his assistance and support
with the biological data. In addition, this project would not have been possible without the
support of the CBCI staff, specifically the psychosocial team who assisted with
recruitment, the phlebotomists at CBCI for drawing the blood, and Tammy Robles, who
shared her lab space and equipment with me. I would also like to thank Dr. Mary Coussons
- Read, who was an instrumental part of the initial development of the project and who
guided me in learning the ELISA technique. Dr. Celia Sladek, Professor of Neuroscience
and director of the CCTSI training program, who taught me the Radioimmunoassay
Technique, allowed me to use her laboratory, and encouraged me through the process of
obtaining funding for my project. I would like to thank Dr. Christopher Phi el, Professor of
Integrative Biology, and his lab, who allowed me to utilize their laboratory space to run
the ELISAs, SNP-Assays, and isolate the DNA. In addition, I would like to thank Josh
Fowler, who assisted in teaching me DNA isolation techniques and SNP assay techniques
v


and helped me trouble shoot through the OTXR genotyping challenges. I would like to
thank Jean Quispe, who assisted me with recruitment and data collection while I was on
maternity leave. I would like to thank Elizabeth Berkholm, who assisted me with date entry.
I would also like to thank Kimberly Hill, for her administrative assistance in ordering
research supplies. I would like to thank the other UCD faculty who supported my learning
and growth as a clinical researcher, without which this project would have not been
possible. Finally, I would like to thank my supportive family for their dedicated assistance
through this process!
vi


TABLE OF CONTENTS
CHAPTER
I. BACKGROUND AM) SIGNIFICANCE............................................... 1
Caregivers of Multiple Myeloma Patients.................................1
Stem Cell Transplantation and Caregiving................................4
Romantic Relationships, Health, and Caregiving..........................5
Oxytocin................................................................9
Inflammation, Health, Caregiving, and Romantic Relationships.......... 12
Conceptual Model.......................................................14
Specific Aims and Hypotheses...........................................15
Specific Aims and Hypotheses in the Context of Conceptual Model........19
II. METHOD...................................................................22
Study Setting..........................................................22
Participants...........................................................23
Recruitment and Enrollment Procedures..................................24
Study Design...........................................................27
Data Collection Procedures.............................................27
Data Analysis..........................................................36
III. RESULTS..................................................................41
Recruitment Accrual and Attrition......................................41
Sociodemographic Characteristics of the Participant Sample.............41
Patient Illness Characteristics........................................43
Caregiving Variables for the Participant Sample........................46
Health Related Variables for the Participant Sample....................48
vii


Additional Health Related Variables for the Participants who Completed Blood
Draw...................................................................50
Psychosocial Characteristics of the Participant Sample.................51
Caregiver Neuroendocrine Variables.....................................53
Results of Inferential Statistical Analyses for each of the Study Aims.55
IV. DISCUSSION................................................................73
Overview...............................................................73
Limitations............................................................88
Future Research Directions.............................................90
CCTSI EDUCATIONAL AND TRAINING EXPERIENCES....................................92
REFERENCES....................................................................99
APPENDICES...................................................................120
A: Consent Forms......................................................120
B: Initial and Follow Up Recruitment Letters Patient................128
C: Initial and Follow Up Recruitment Letters Caregiver..............130
D: Letter to Psychosocial Team to Request Recruitment
Assisstance...........................................................132
E: Caregiver Informational Flyer......................................133
F: Pre-Blood Draw Screen..............................................134
G: Caregiver Psychosocial Survey......................................137
viii


LIST OF TABLES
TABLE
1. Key Variables and Sources of Data.............................................35
2. Tests of Normality and Internal Consistency for
Variables Used in Inferential Statistical Analyses...............................36
3. Participant Demographic Variables.............................................42
4. Caregiver Reported Patient Illness Characteristics............................44
5. Caregiver Level of Involvement and Health Behaviors...........................48
6. Caregiver Health Related Variables and Neuroendocrine Data....................50
7. Caregiver Scores on Psychosocial Measures.....................................52
8. Group Differences for Pro-Inflammatory Markers................................53
9. Correlation Table for
Caregiver Distress and Inflammatory Marker Variables.............................56
10. Summary of Multiple Regression Analysis for
Caregiver Distress Variables Predicting CRP......................................58
11. Summary of Multiple Regression Analysis for
Caregiver Distress Variables Predicting IL-6.....................................58
12. Correlation Table of
Caregiver Health Variables and Inflammatory Markers..............................60
13. Summary of Multiple Regression Analysis for Caregiver Health Behavior Variables
Predicting Inflammatory Markers..................................................62
14. Correlation Table for Caregiver Pro-Inflammatory Markers and Quality of Life.63
15. Summary of Multiple Regression Analysis for
Pro-Inflammatory Markers Predicting Caregiver Quality of Life....................64
16. Correlation Table of Caregiver Psychological Processes and Health Variables..65
17. Summary of Multiple Regression Analysis
for Health Behaviors Predicting Caregiver Distress...............................68
ix


18. Correlation Table for Caregiver Distress and Quality of Life
69
19. Summary of Multiple Regression Analysis
for Caregiver Distress Predicting Caregiver Quality of Life......................70
20. Summary of Multiple Regressions for Mediation Model.........................71
21. Comparative Data for Psychosocial Measures across Various Studies...........78
x


LIST OF FIGURES
FIGURE
1. Overall Conceptual Model to Guide Hypotheses and Analyses.....................15
2. Conceptual Model Reflecting Only the Variables Included in the Final Analyses.21
3. Types of Treatment Received by Patient........................................46
4. Caregiver Perception of Provision of Adequate Care............................47
5. Caregiver Reported Health and Diet Quality....................................49
6. Mediation Model...............................................................72
xi


CHAPTER I
BACKGROUND AND SIGNIFICANCE
Caregivers of Multiple Myeloma Patients
Multiple myeloma (MM) is an incurable hematologic cancer. It is currently the
second most common blood cancer and in 2015 there was an estimated 26,850 new cases
diagnosed (The American Cancer Society, 2015). Treatments for MM are aimed at
mitigating symptoms, improving quality of life and increasing longevity. Recent
improvements in treatment have led to increases in long-term survivorship with MM
patients surviving, on average, up to 11 years after initial diagnosis from about 4.5 years
in 2003 (Child et al., 2003). For patients diagnosed with multiple myeloma who are under
the age of 65, the current standard of treatment includes an autologous (ones own) stem
cell transplantation1 (SCT) with high dose melphalan (Child et al., 2003). In autologous
transplantation, ones own hematopoietic stem cells are harvested from peripheral blood
and then frozen for storage and later use. Next, high dose chemotherapy and/or radiation is
administered to an individual to eradicate the cancerous blood cells, but with side effects
of compromising the immune system. After chemotherapy or radiation is complete, the
harvested cells are thawed and returned to the patient. This process can be done on an
inpatient or outpatient basis with follow up by the oncologist who is overseeing the
1 Blood and bone marrow transplant and stem cell transplant will be used interchangeably in the subsequent
sections.
1


procedure. Whether a patient will undergo transplant depends on a number of factors
including disease progression and stage, response to chemotherapies, other medical
comorbidities, age, psychosocial factors, and financial considerations. For those patients
who do not receive a SCT, a combination of Bortezomib, Thalidomide or Lenalidomide,
chemotherapy, radiation therapy, and corticosteroids is standard (Palumbo et al., 2014).
Unfortunately, the treatment and survivorship trajectory for MM is very
unpredictable; it is often characterized by illness relapse followed by additional treatment,
which sometimes entails second or even third autologous transplants and in fewer cases
allogeneic (donor cell) transplant (Greipp et al., 2005). Approximately 65% of the
treatments for MM occur on an outpatient basis, resulting in patients and their caregivers
spending a considerable amount of time managing the side-effects of treatment at home.
These side-effects can include bone pain, neuropathy, fatigue, loss of appetite, severe
diarrhea and vomiting (Molassiotis, Wilson, Blair, Howe & Cavet, 2011). This is not only
difficult for the patient, but can be extremely stressful for the caregiver who is trying to
support and care for their loved one while managing their own self-care routines. However,
little is known about the impact that the multitude of treatment related stressors coupled
with longer, yet unpredictable survivorship trajectory has on the caregiver. Therefore,
partnered caregivers of MM patients are a unique and important population to study.
Compared to other cancer populations, there are relatively few studies examining
psychosocial adjustment in caregivers of MM patients. A study of 93 partners of MM
patients found that one third of partners reported unmet supportive care needs (e.g.
existential issues, sexuality issues, chronic symptom management, self and family care),
almost half (48.8%) reported signs of anxiety, and 13.6% reported signs of depression
2


(Molassiotis et al., 2011). A qualitative study, conducted by the same group of researchers,
found informal caregivers (88% spouses) were providing practical and emotional support
to patients almost exclusively, often by neglecting their own needs (Molassiotis et al.,
2011). No gender differences were noted in this study. Furthermore, informal caring often
led to disruption of caregivers daily life, heightened illness burden and difficulties coping
with caregiver-related stressors. Both patients and caregivers reported significant fears and
uncertainty about the future; for example, one caregiver described MM as a 'time bomb'
(Molassiotis et al., 2011). In a case study of a couples journey with MM, a wife of a MM
patient stated we decided early on that Ian would fight the disease and I would take care
of everything else. Neither knew what we were taking on (Coon, McBride-Wilson, &
Coleman, 2007). Another study supported the notion that MM caregivers are a unique
group compared to many other cancer caregivers because of the psychosocial fatigue
associated with the accumulation of years of medical caregiving (Zabora et al., 2015).
These studies consistently illustrate the psychosocial toll of caregiving for a loved one with
MM although few studies have specifically examined the specific emotional, social and
practical challenges faced by MM patients and their loved ones (Dahan & Auerbach, 2006).
Despite the limited research assessing and describing psychosocial reaction in caregivers
of MM patients, overall, studies have found that caregivers of cancer patients experience
higher psychological distress than the patients to whom they provide care (Kim, Carver,
Rocha-Lima, & Schaffer, 2011; Soothill, Morris, Thomas, Harman, Francis, Mclllmurray,
2003). Caregivers of cancer patients also report lower quality of life and are at increased
risk for medical comorbidities compared to non-caregivers (Vitaliano, Zhang, & Scanlan,
2003). Additionally, high rates of psychosocial distress and physical exhaustion may lead
3


to caregiver burnout resulting in decreased caregiver involvement in patient needs (Marks,
Lambert, & Choi, 2002; Pinquart & Sorensen, 2003).
Stem Cell Transplantation and Caregiving
The transplantation process is laden with frequent fluctuations in medical status,
repeated invasive medical procedures, possibility of death, prolonged hospitalization
(Foxall & Gaston-Johansson, 1996), as well as uncertainty, disappointment, dislocation
from home and friends, worries about children, and loss of job (Lesko, 1994). Often
before the transplantation process begins, patients need to make plans regarding their
family, home, finances, pets, and employment for the duration of their transplant
procedure, which can last weeks to months. During transplant, caregivers experience a
number of practical problems such as significant time constraints and financial burden
(Meehan, Fitzmaurice, Root, Kimits, Patchett, & Hill, 2006). In addition, caregivers
experience a disruption of daily roles (Chow & Coyle, 2011), decreased martial
satisfaction, and increased levels of anxiety and depression both six months and one year
following stem cell transplantation (Langer, Abrams, & Syijala, 2003). The physiological
and psychosocial impact of SCT on informal caregivers has received little attention,
although the literature suggests that caregivers may experience immunological and
psychological changes (Futterman, Wellisch, Zighelboim, LunaRaines, & Weiner, 1996)
and heightened caregiver burden (Foxall & Gaston-Johansson, 1996). SCT is a profound
experience shared by the entire family of the patient. The uncertain and multifaceted
trajectory of the SCT procedure is distressing and burdensome on the caregiver from a
biopsychosocial perspective and it can have a profound impact on caregiver quality of
life.
4


Romantic Relationships, Health, and Caregiving
Romantic relationships and health. Overall, being in a long term romantic
relationship has been linked to better health and well-being. Relationship functioning,
which encompasses overall satisfaction, communication behaviors, companionship,
decision making, and cognitions (attributions and expectations) is bidirectionally related to
health habits including sleep (Troxel, 2010) diet, and exercise. Cardiovascular (Berkman,
1995), endocrine, and immune responses have also been linked to relationship functioning
(see Kiecolt- Glaser & Newton, 2001 and Robles & Kiecolt-Glaser, 2003 for extensive
reviews). For example, women who reported moderate to severe marital strain were three
times more likely to experience a coronary event after controlling for demographic and
disease status variables (Orth-Gomer, Wamala, Horsten, Schenck-Gustafsson, &
Schneiderman, 2000). Furthermore, the quality of and attachment style within a romantic
relationship, as well as in the context of caregiving, has been linked to differences in the
psychological reaction of and care provided by the significant other (Kim & Carver, 2007;
Kim, Deci, Carver, & Kasser, 2008; Morse, Shaffer, Williamson, Dooley, & Schulz, 2012).
The sections below will review the research on attachment, relationship quality and the
experience of caregiving.
Attachment. Attachment is defined as ones orientation to close relationships.
Attachment theory provides a framework from which to conceptualize relationship
dynamics, which are shaped from early experiences and impact our view of ourselves and
others in the context of close relationships. From infancy, humans have an attachment
system that is activated by stress or threat, in order to maintain a sense of security (Bowlby,
1973). Since Bowlby, various working models have been developed to conceptualize adult
5


human attachment. In 1987, Hazan and Shaver described three types of attachment styles:
1) secure, which is defined as a sense of closeness and easy reliance on others; 2) anxious-
ambivalent. which is defined as a desire for intimacy plus insecurity about others
responses; 3) avoidant, which is defined as independence, distance from others and
discomfort with closeness. In 1991, Bartholomew and Horowitz created a four group model
to describe attachment styles, in which patterns of attachment are defined using
combinations of ones self-image (positive or negative) and image of others (positive or
negative). This model describes four attachment styles: 1) secure- positive thoughts about
self and others; 2) dismissive-avoidant- positive thoughts about self, negative thoughts
about others; 3) anxious-preoccupied- negative feelings about self, positive feeling about
others; and 4) fearful-avoidant- negative thoughts about self and others. In adults,
attachment security (comfort with closeness and interdependence) is the foundation for
effective caregiving relationship because it allows individuals to be more attentive to
partners needs. A significant amount of research has been conducted evaluating how adult
attachment style impacts the experience of and reaction to the demands of being a caregiver
for a spouse with cancer (Kim & Carver, 2007; Kim et al., 2008), or a spouse with cognitive
disabilities (Morse et al., 2012).
In 1997, Carver created the Measure of Attachment Qualities (MAQ) theoretically
based on Hazan and Shavers model of attachment (1987). The MAQ uses separate
subscales to assess secure attachment tendencies, avoidant tendencies, and two subscales
to assess anxious-ambivalent patterns. These anxious-ambivalent subscales include the
Ambivalence Merger, which is reflective of an unmet desire for more closeness, and
Ambivalence Worry, which is reflective of concern about ones partners feelings being
6


insincere or not being as strong as ones feelings toward his or her partner. Development
of the MAQ is described in more detail in method section.
Using this measure to assess the role of attachment and frequency with which
various care tasks were provided (e.g. emotional, instrumental, tangible), Kim and Carver
(2007) studied a group of spousal caregivers of cancer patients. They found that securely
attached wives provided more frequent emotional care (e.g. boost his/her mood when
he/she felt low) and anxiously attached wives provided higher levels of tangible care
(e.g. providing financial help) for spouses with cancer. Securely attached spouses
reported less difficulty in providing care and reported less burden. This difference may be
explained by more securely attached spouses feeling more able to meet the emotional
needs of an ill spouse due to their perceived interdependence and positive view of self
and others. In another study focusing on adult attachment, psychological well-being, and
motivation for caregiving for a spouse with cancer, attachment security was positively
associated with autonomous motives (e.g. I provided care because it was
important/something I valued deeply) and finding benefit in caregiving (Kim et al.,
2008). Again these differences in perception of the caregiving experience reflect the
general caregiver self-esteem and differences in views of relationships with others as
either negative or positive.
In a non-cancer population, securely attached spouses offered more comfort and
reassurance while avoidantly attached spouses displayed more anger or blaming in
response to situational stresses (Feeney, & Collins, 2001). Fearful spouses, defined as those
having negative thoughts about self and others, were less likely to report using problem-
focused coping and more likely to state that the caregiving relationship led to increased
7


marital conflict. These findings suggest that attachment style not only impacts patient-
caregiver interactions, but also influences how the caregiver copes with the stress of caring
for a loved one with a health condition. Decreased use of adaptive coping may represent
one mechanism that increases caregivers risk of developing mental and physical health
issues. Along those lines, among children with elderly, frail or demented parents, an
insecure attachment style has been linked to increased psychological distress (Crispi,
Schiaffino, & Berman, 1997) and is a predictor of depression (Besser & Priel, 2005;
Camelley, Pietromonaco, & Jaffe, 1994). Secure attachment has been linked to lower levels
of caregiver stress (Cripsi et al., 1997). Again this elucidates the important relationship
between attachment style and mental health while caring for a loved one with a chronic
illness. Overall, these findings suggest that caregivers who report a secure attachment to
their partner may experience less distress and better overall health than caregivers who do
not report a secure attachment. The present study will elucidate the role that secure
attachment plays with regard to neuroendocrine response in caregivers of MM patients.
Relationship quality. Relationship quality has also been found to be an important
predictor of psychological and physiological responses to stressful life events. The
definition of relationship or marital quality varies from study to study depending on the
measures used, but generally encompasses satisfaction, agreement on values and life goals,
overall happiness, adjustment, reliance on one another, and exchange of ideas, or a
combination thereof. One study examining neural response to the threat of an electric shock
found that higher levels of self-reported marital quality predicted attenuation of activation
in the neural systems supporting emotional and behavioral threat responses when women
held their spouses hand (Coan, Schaefer, & Davidson, 2006). Higher marital quality has
8


also been associated with faster recovery from illness and a lower rate of mortality
following the diagnosis of a life threatening illness (Robles & Kiecolt-Glaser, 2003).
Lastly, marital quality also predicted hospital stay following coronary artery bypass
surgery for women (Kulik & Mahler, 2006). However, there have been no studies to date
that have assessed relationship quality in the context of partners facing MM. Therefore,
relationship quality is an important variable to include in our study because there is limited
research on how it is related to mood, quality of life, and health in the MM caregiver
population.
Oxytocin
The neuropeptide oxytocin and its receptor are expressed widely in the central and
peripheral nervous system, suggesting a number of endocrine and paracrine roles
(MacDonald & MacDonald, 2010; Tom & Assinder, 2010). Oxytocin has been shown to
modulate social behavior such that intranasally administered oxytocin has been lined to
increased eye contact, improved social memory, and less social anxiety (see MacDonald
& MacDonald, 2010 for a review). Another important facet of the oxytocin literature that
is relevant to this study is that it has been found to improve ones ability to manage stressful
events (Amico, Cai, & Vollmer, 2008; Ditzen, Schaer, Gabriel, Bodenmann, Ehlert, &
Heinrichs, 2009; Heinrichs, Baumgartner, Kirschbaum, & Ehlert, 2003; Light, Grewen, &
Amico, 2005; Neumann, 2002), and help to establish or maintain social attachments
(Bucheim et al., 2009). Oxytocin levels have also been negatively associated with
depression and anxiety (Scantamburlo et al., 2007), and positively associated with displays
of romantic love (Gonzaga, Turner, Keltner, Campos, & Altemus, 2006), affectionate
behaviors between partners (Gutkowska, Jankowski, Lambert, Mukaddam-Daher, Zingg,
9


& McCann, 1997), and trust (Kosfeld, Heinrichs, Zak, Fischbacher, & Fehr, 2005).
Oxytocin has been linked to a variety of health outcomes such as cancer (Reversi, Cassoni,
& Chini, 2005) and cardiovascular functioning (Gimpl & Fanreholtz, 2001; Gutkowska et
al., 1997; Petersson, 2002). In addition, increased levels of oxytocin are related to lower
blood pressure and heart rate (Light et al., 2005).
Recent developments in the field of molecular genetics have also linked single
nucleotide polymorphisms in the oxytocin receptor (OXTR rs 53576) to differences in
attachment style (Chen, Barth, Johnson, Gotlib, & Johnson, 2011), emotion regulation
(Kim, Sherman, Mojaverian, Sasaki, Park, et al., 2011), and parenting behavior
(Bakermans-Kranenburg, & van Ijzendoom, 2008). In terms of behavioral phenotype,
individuals with the GG phenotype of OXTR rs53576 exhibit more prosocial temperament
(Tost et al., 2010), more sensitive parenting behavior (Bakermans-Kranenburg & van
IJzendoom, 2008), and greater sensitivity to infant crying (Riem, Pieper, Out, Bakermans-
Kranenburg, & van Ijzendoom, in press). GG homozygotes also report being less lonely
(Lucht et al., 2009) and possess greater empathic accuracy (Rodrigues, Saslow, Garcia,
John, & Keltner, 2009) relative to those with the AA genotype. Those with the AG
genotype fall between the two homozygous genotypes. Although there is literature linking
prosocial behavior and parental caregiving behaviors to oxytocin genotype receptor, there
is no research to date examining genotypical differences in the context of caregiving for a
significant other.
Polymorphisms within the OXTR impact the production, transport, and metabolism
of oxytocin, thus creating a systemic impact, influencing behavior, mood and health. Thus
10


it is important to briefly review literature related to oxytocins role in attachment, physical
contact, as well as stress and health.
Attachment and physical contact. Results from one study found that more
frequent partner hugs and higher oxytocin levels are linked to lower blood pressure and
heart rate in premenopausal women (Light et al., 2005). These findings suggest that
supportive marital interactions may promote greater oxytocin production therefore
modulating stress and immune responses.
Health and stress. Increases in oxytocin levels have been associated with lower
blood pressure (Grewen, Girdler, Amico, & Light, 2005). An interesting study
investigating the relationships between marital behavior, oxytocin, and wound healing,
found that higher levels of oxytocin were associated with more positive communication
and quicker wound healing (Gouin et al., 2010). Also, circulating level of plasma oxytocin
was found to be associated with HIV viral status in low-income minority women such that
higher oxytocin was associated with lower viral titers (Fekete et al., 2011). Compared to
no treatment controls, individuals who received intranasal oxytocin showed a reduced
cortisol response in response to a public speaking task (Quirin, Kuhl, & Duesing, 2011).
After a social stressor, men who were administered intranasal oxytocin, and accompanied
by a friend, showed decreased cortisol levels, higher ratings of calmness, and lower anxiety
compared to controls (Heinrichs et al., 2003). In addition, a number of studies have
suggested that oxytocin may mediate or moderate the relationship between social support
and health outcomes (Carter, 1998; Robles & Kiecolt-Glaser, 2003). Finally, intranasally
administered oxytocin has been associated with decreased levels of pro-inflammatory
cytokines including TNF- a, IL-6, and VEG-F in response to a bacterial endotoxin. Despite
11


its link to stress modulation and attachment, there are no studies to date examining oxytocin
genotype in couples experiencing chronic stress. Including oxytocin as variable of interest
in caregiver studies can elucidate the relationship between attachment style, relationship
quality and health in individuals who are caring for a significant other with a serious illness.
Limitations to oxytocin research. There are some limitations to conducting
oxytocin genotyping in research. Research has shown that our environment can interact in
complex ways changing our genetic expression (Kumsta, Hummel, Chen, & Heinrichs,
2013). Despite this, there is still clear evidence that differences in genotypes are related to
changes in behavior, affect, and interactions with others.
Inflammation, Health, Caregiving, and Romantic Relationships
Inflammation is a vital part of the human immune response to acute infection. Pro-
inflammatory cytokines signal immune cells to an injury site to help clear viral and
bacterial pathogens (Kiecolt-Glaser, Gouin, & Hantsoo, 2010). However, chronic
inflammation is problematic. Elevated levels of pro-inflammatory cytokines (IL-6,TNF-
a, IL1-P) and C-Reactive Protein (CRP) have been linked to a number of medical
conditions including cardiovascular disease, Type-II Diabetes, certain cancers
(Gruenewald, Seeman, Ryff, Karlamangala, & Singler, 2006; Papanicolaou, Wilder,
Manolagas & Chrousos, 1998; Pradhan, Manson, Rifai, Buring, & Ridker, 2001), arthritis,
osteoporosis, and Alzheimers Disease (Ershler & Keller, 2000). C reactive protein
(CRP) is a protein in the plasma that can signal systemic inflammation. Although CRP is
not routinely measured in clinical settings, it is commonly used as a marker of
inflammation in studies examining risk for various illnesses. IL-6, CRP and TNF- a have
also been associated with negative mood (Wright, Strike, Brydon, & Steptoe, 2005) and
12


recent research has indicated that this relationship is bidirectional (Dantzer, O'Connor,
Freund, Johnson, & Kelley, 2008). Sleep disruption, often seen with caregivers, can
increase the expression of pro-inflammatory cytokines (IL-6 and TNF-a), which may
promote tumorigenesis by inhibiting DNA repair through the generation of reactive oxygen
species, inducing DNA damage (Antoni et al., 2006). Finally, inflammation levels are
impacted by other health behaviors such as exercise, diet, chronic health problems,
medications, adiposity, alcohol, and smoking (Kiecolt-Glaser et al., 2010).
Psychological stress can also evoke pro-inflammatory cytokine production in the
absence of infection or injury (Kiecolt-Glaser, Gouin, & Hantsoo, 2010). Studies of
caregivers have found higher levels of inflammatory markers compared to non-caregivers
(Kiecolt-Glaser, Preacher, MacCallum, Atkinson, Malarkey, & Glaser, 2003; von Kanel et
al., 2006). One study in particular found that the average level of IL-6 was four times higher
in caregivers than in non-caregivers (Kiecolt-Glaser et al., 2003). It has been suggested that
chronic inflammation may represent a biological mechanism by which the stress and
burden of caregiving leads to declining health status and increased risk for morbidity and
mortality (Maier, 2003) especially in older caregivers by way of frailty syndrome (von
Kanel, Kudielka, Preckel, Hanebuth, & Fischer, 2006).
Romantic relationships and inflammation. Attachment, social support, and
relationship quality have been associated with immune responses. Specifically, IL-6 levels
during an acute marital argument show an 11% increase in individuals with high
attachment avoidance and a 6% decrease in individuals with low attachment avoidance
(Gouin et al., 2009). Furthermore, more frequent hostile marital interaction was associated
with a 40% slower rate of blister wound healing and lower wound site production of IL-6,
13


TNF-a, IL-ip compared to low-hostile couples (Kiecolt-Glaser et al., 2005), indicating a
slowed local immune response. In a different sample of couples engaging in conflict
discussion, those who used more cognitive words (e.g. words of reasoning, insight, and
thinking) had smaller increases and lower levels of IL-6 and TNF-a 24 hours after the
discussion compared to those who used fewer cognitive words (Graham et al., 2009).
Among older adults, married men had lower CRP levels than divorced men, married
women and unmarried women which may help to explain the finding that men appear to
experience greater health benefit from marriage (Sbarra, 2009). The role of inflammatory
cytokines in the decline of caregiver health is a complex process which may be impacted
by various psychological (e.g. attachment, depression), demographic (e.g. age, gender),
and behavioral variables. Thus, assessing caregiver inflammation levels to help elucidate
the relationship between inflammation, mood, health and quality of life among partnered
caregivers of MM patients will add a psychoneuroimmunological facet to the MM
caregiver literature.
Conceptual Model
Based on the reviewed literature, an initial conceptual model, presented in Figure
1, was developed to help guide the overall specific aims and hypothesis described in the
following section. As illustrated in the model, during the process of caregiving, an
individual may experience psychological, behavioral and biological reactions that are
inter-related. Specifically, each individual has a unique set of sociodemographic,
caregiving-specific, and health-specific factors that may influence a caregivers
psychological and biological response while caring for a loved one. Furthermore,
psychological processes such as depression, and relationship factors such as marital
14


quality have been linked to immunologic and endocrine functioning. Finally, while the
relationship between psychological processes and quality of life is well established in the
literature, little is known about the relationship between neuroendocrine and immune
response with quality of life and likely a bidirectional relationship exists. It should be
noted that while not all of the relationships outlined in the model were directly assessed,
this model helped guide the development of the hypothesis and the planned analysis.
Participant Characteristics
Sociodemographics
Age
Caregiving Variables
Length of illness time
Perception of illness
severity
Health Variables
Diet quality
Physical activity
Sleep quality
BMI
Psychological Processes
Depression
Negative affect
Illness related distress
Relationship
Characteristics
Neuroendocrine
Regulation
Oxytocin Receptor
Genotype
Immune
Relationship quality Response
Attachment IL-6, TNF-a, CRP
Figure I. Overall conceptual model to guide hypotheses and analyses. Note: TNF- a and
oxytocin genotype were not included in final analyses.
Specific Aims and Hypotheses
The purpose of this study was to better understand the psychosocial and
physiological response of caregivers whos loved one has a diagnosis of MM. Specifically,
15


this project aimed to increase our understanding of the relationship between caregiver
psychosocial variables (depression, illness-related distress, affect), patient-caregiver
relationship characteristics (relationship quality, attachment style), neuroendocrine and
immune responses (oxytocin receptor genotype and pro-inflammatory markers), and
quality of life, in significant others who are caring for MM patients. While the literature
suggests associations among the variables of interest, mainly among caregivers of loved
ones with other types of cancer, no study to date has examined these variables in a sample
of individuals caring for a significant other with MM. The experience of caring for a loved
one with MM is somewhat unique given the chronic nature of the illness and the uncertain
trajectory. Caregivers face both chronic (e.g., living with MM for many years) and acute
stressors (e.g., chemotherapy treatments, bone marrow transplants) that likely impact
psychosocial distress, immune and endocrine function, relationship quality, and quality of
life. The following aims guided this cross-sectional study to better understand how these
factors are associated with one another within this population:
Aim 1. Evaluate the relationships between caregiver oxytocin genotype,
psychological processes, and caregiver relationship characteristics in the context of caring
for a loved one with MM.
Hypothesis 1.1: Caregivers with GG oxytocin genotype will report a higher number
of positive caregiver-patient relationship characteristics (secure attachment style and high
relationship quality).
Hypothesis 1.2: Caregivers with GG oxytocin genotype will report lower levels of
distress (depression, negative affect, illness-related distress).
16


Aim 2. Evaluate the relationships between caregiver pro-inflammatory markers,
psychological processes and health variables.
Hypothesis 2.1: Caregivers who report higher levels of distress (depression,
negative affect, illness-related distress) will have higher levels of plasma pro-inflammatory
markers.
Hypothesis 2.2: Caregiver distress (depression, negative affect, and illness-related
distress) will significantly predict the level of caregiver pro-inflammatory markers while
controlling for caregiver age, length of illness time, and caregiver perception of illness
severity. Specifically, within those models, depression, negative affect, and illness-related
distress will be significant individual predictors of the level of pro-inflammatory markers
above and beyond the control variables.
Hypothesis 2.3: Caregivers reporting better health behavior indicators, as measured
by sleep quality, diet quality, and physical activity levels, will have lower levels of plasma
pro-inflammatory markers.
Hypothesis 2.4: Caregiver self-reported health behavior indicators (sleep quality,
diet quality, and level of physical activity) will together predict level of caregiver pro-
inflammatory markers (IL-6, TNF a, CRP), while controlling for age, length of illness
time, and BMI. Within that model, it is hypothesized that sleep quality, diet quality, and
level of physical activity will each significantly predict level of caregiver inflammatory
markers above and beyond the control variables.
Aim 3. Evaluate the association between caregiver biomarkers and caregiver
quality of life.
17


Hypothesis 3.1: Levels of plasma pro-inflammatory markers will be inversely
related to quality of life.
Hypothesis 3.2: Levels of plasma pro-inflammatory markers (IL-6, CRP, TNF -a)
will predict caregiver quality of life, while controlling for age, length of illness time, and
caregiver perception of illness severity. Within those models it is hypothesized that pro-
inflammatory markers will be significant individual predictors of caregiver QoL above and
beyond the control variables.
Hypothesis 3.3: Quality of life will be significantly associated in the positive
direction with OXTR GG phenotype.
Aim 4. Evaluate the relationship between caregiver health behavior indicators as
measured by sleep quality, physical activity and self-reported quality of diet and
psychological processes (depression, affect, illness-related distress).
Hypotheses 4.1: Quality of sleep, quality of diet and level of physical activity will
be inversely related to level of depression, negative affect and illness-related distress.
Hypothesis 4.2: Sleep quality, diet quality, and level of physical activity together
will predict caregiver distress (depression, affect, illness-related distress) after controlling
for length of illness time and perception of illness severity. Within that model it is
hypothesized that sleep quality, diet quality and level of physical activity will be significant
as individual predictors of distress (depression, affect, illness-related distress) above and
beyond the control variables.
Aim 5. Evaluate the relationships between caregiver distress and caregiver quality
of life.
18


Hypothesis 5.1: Level of illness-related distress, level of depression, and negative
affect will be inversely related to caregiver QoL.
Hypothesis 5.2: Caregiver distress (illness-related distress, depression, and
negative affect) will predict caregiver quality of life while controlling for perception of
illness severity and length of illness time. Within that model, measures of caregiver distress
(illness-related distress, depression, and negative affect) will be significant as individual
predictors above and beyond the control variables.
Exploratory Aim 6. Examine the relationship between caregiver-patient
relationship quality, caregiver distress and caregiver quality of life in the context of caring
for a partner with MM. Determine if relationship quality is a partial mediator of the
relationship between caregiver level of depression and caregiver quality of life (see
conceptual model and explanation below).
Specific Aims and Hypotheses in the Context of Conceptual Model
The updated model, displayed in Figure 2 below, reflects only the variables that were
included in the final analyses. The top boxes that represent sociodemographic variables,
health variables (e.g. level of physical activity, diet quality, and sleep quality) and
caregiving variables (e.g. length of illness time and perception of illness severity) are
suggested by the literature to be related to caregiver psychological processes, immune
response, and caregiver QoL, and are thus included as control variables in many of the
aims and analyses. Specifically, Aim 1 was designed to evaluate the association between
OXTR genotype (representing neuroendocrine regulation) and caregiver psychological
processes, distress (depression, affect, and illness-related distress) and relationship
characteristics (secure attachment and relationship quality). Unfortunately, due to
19


difficulties with the storage and purification of DNA, the OXTR genotype data was
invalid and was not included in the study and therefore is not included in the updated
model presented in Figure 2. Aim 2 will examine the relationship between inflammatory
markers and health variables as well as the relationship between inflammatory markers
and psychological variables. We expect that positive health behaviors (e.g. healthful
diet and regular physical activity) will be significantly inversely related to inflammation
and that distress will be significantly positively related to inflammation. Aim 3 will
focus on understanding the strength of the relationship between immune response (as
measured by pro-inflammatory markers IL-6 and CRP) and caregiver QoL. Aim 4 will
examine psychological processes, specifically caregiver distress (illness-related distress,
depression, and negative affect) that are impacted by sociodemographic variables and
health behaviors. Aim 5 will evaluate the degree to which caregiver psychological
processes and distress directly impact quality of life. While there is overlap in the
constructs of distress and quality of life, quality of life is a multifaceted construct that
encompasses more global experiences such as perceived burden, life interruption,
communication, positive benefit finding, spirituality and psychological adjustment.
Finally, exploratory Aim 6 will elucidate the underlying mechanism by which
depression is associated with quality of life. While, it is clear that many types of
mediation and moderation among these variables are possible, the pathway of
relationship quality influencing depression, which in turn impacts quality of life, is just
one of many possible pathways. This particular one is explored in order to highlight the
possible importance of relationship functioning even when one is in a caregiver role.
20


Therefore, the relationship quality between the caregiver and the MM patient is an
important factor to consider with regard to caregiver quality of life.
Participant Characteristics
Sociodemographics
Age
Caregiving Variables
Length of illness time
Perception of illness
severity
Health Variables
Diet quality
Physical activity
Sleep quality
BMI
Aim 4

Aim 2
Depression
Negative affect
Illness related distress
Aim 2

Aim 6
Aim 3
Figure 2. Conceptual model reflecting only the variables that were included in final
analyses. Arrows reflect which relationships were directly assessed in the each of the
study aims. Note: Due to difficulties with the storage and purification of DNA, the
OXTR genotype data was invalid and was not included in the study and therefore is not
included in the updated model presented in Figure 2.
21


CHAPTER II
METHOD
Study Setting
The Colorado Blood Cancer Institute (CBCI) is a large regional transplant center for
Colorado and surrounding states. Transplanting over 240 patients per year, the program
performs allogeneic and autologous transplants for adult patients.
The transplant program is part of the Colorado Blood Cancer Institute (CBCI) at
Presbyterian/St. Luke's Medical Center (PSLMC). PSLMC is a large medical center,
licensed for 680 beds, serving Denver, Rocky Mountain, and Great Plains regions. There
are three inpatient units at PSLMC dedicated to stem cell transplantation. A fourth inpatient
unit, adult oncology, provides care for patients with hematologic and solid tumor
cancers. CBCI, the outpatient treatment center for transplant patients, is in a professional
building attached to the hospital. Retrieved from
http://www.bloodcancerinstitute.com/).
Utilizing a multidisciplinary team approach, CBCI provides comprehensive care to
transplant patients and their family members. CBCI is representative of other transplant
programs in the country. The large psychosocial team at CBCI makes it an attractive
research site. In addition, there are large-scale NIH-funded research studies that have been
done at this site and are currently being conducted by psychosocial research personnel at
this site. Thus, there is a high level of support for psychosocial research both within the
psychosocial team and from the institution (T. Simoneau, personal communication, June
4, 2011).
22


The ethnic and racial diversity in the SCT population treated at CBCI reflects the
general diversity in the Colorado region: White 69%, Hispanic/Latino about 13%, Black
or African American 8%, Asian 5%, American Indian/Alaska Native 5%. In 2011, a
total of 242 transplants were conducted, and 77 were multiple myeloma patients (T.
Simoneau, personal communication, May, 15 2012). Furthermore, a database of over 200
current and previous multiple myeloma patients who have been treated at CBCI exists was
utilized for recruitment purposes. Recruitment procedures are discussed in more detail
below.
Participants
The study population consisted of spousal/partnered caregivers of patients
diagnosed with multiple myeloma (MM) who were currently being treated or had
received treatment within the last two years at the Colorado Blood Cancer Institute at
Presbyterian/St. Luke's Medical Center (PSLMC). Eligibility criteria for study enrollment
included the following: 1) Caregiver must be co-habitating and involved in a romantic
relationship with the MM patient; 2) Patient who is being cared for has a diagnosis of
MM and is currently or has received treatment at CBCI; 3) Partner must be in an active
caregiving role for the patient diagnosed with MM; 4) Caregiver must be able to read and
understand English; and 5) Caregiver must be 18 years of age or older. This study was
limited to English speaking caregivers because many of the study instruments are
validated only in English.
Caregivers were excluded if they had a medical, psychological or cognitive condition
that would interfere with the ability to consent and/or participate in the study. Competence
for study participation was evaluated by potential participants ability to explain to study
23


personnel the goals of the study, requirements of study participation, and potential risks
and benefits. This study did not enroll individuals who required proxy consent. Caregivers
were not excluded if they did not consent to blood draws. Preliminary eligibility was
assessed by a member of the psychosocial team, who was familiar with the caregivers
relationship status, and recruitment efforts were focused on those caregivers who met the
above mentioned eligibility criteria.
Recruitment and Enrollment Procedures
After obtaining approval to conduct the study through the Colorado Multiple
Institutional Review Board (COMIRB) at the University of Colorado Denver and
Presbyterian St. Lukes Medical Center, partnered caregivers of patients who have been
diagnosed with MM were recruited from PSLMC in two manners 1) In person, during the
pre-transplant psychosocial intake and 2) Via a mail out (recruitment letters can be found
in Appendix B). In person recruitment primarily occurred during the pre-transplant
psychosocial intake, however on occasion patients and caregivers who had already
received treatment at CBCI were approached. A psychologist, psychology fellow, social
worker, or psychology practicum student from the Psychosocial Oncology Department at
PSLMC provided potential participants with an informational sheet describing the research
study and provided a brief verbal description of the study. The informational sheet included
the study coordinator (Shannon Madores) contact information. If a caregiver was
interested in receiving more information about the study, the HTP A A A Recruitment form
was completed to procure the participants contact information. Shannon Madore then
contacted the caregiver to provide more detailed information about the study and confirm
that the caregiver met participation eligibility criteria. If the caregiver was still interested
24


in participating, a study team member set up a time for him/her to come to the CBCI clinic
for a onetime 30-45 minute visit.
The mail out recruitment method consisted of three waves of letters, each wave
included one initial contact letter and one follow-up letter (included in Appendix B and C)
sent out to CBCI patients in the MM database. Participants were provided study team
coordinators contact information and a brief description of the study. If interested,
participants contacted the study coordinator to set up a time to come to the CBCI clinic for
a onetime 30-45 minute appointment. In general, to control for circadian variability,
appointment times occurred in a standardized block of time (9am -1pm) during the week
for 15 months. In an effort to standardize behaviors that may impact neuroendocrine
functioning, participants were instructed to not eat a meal at least two hours prior to the
appointment, as well as to cease tobacco, caffeine and alcohol use the morning of the
appointment. Other data that may impact neuroendocrine functioning (e.g. BMI, certain
medications, presence of acute of chronic illness) were collected from the caregiver prior
to the blood draw.
Upon arrival to the clinic, the caregiver was consented. In order to ensure
confidentiality and privacy, the study was explained in a private clinic room. During the
consent procedure, a study team member reviewed the consent and HIPAA B forms
thoroughly. The participant was informed that participation is completely voluntary. After
reviewing the consent form thoroughly, a study team member inquired whether the
participant had questions. If the caregiver had no questions, he/she was asked to explain
the purpose of the study to assess understanding. If the participant was fully aware of the
25


details of the study and had no further questions, he/she signed and dated the consent form.
He/she was then given a copy of the consent form.
After the consent process, the caregiver was given a brief pre-blood draw
questionnaire (located in Appendix D) to complete. The caregiver was then escorted to the
blood drawing area and the blood draw was conducted by a trained phlebotomist. The
caregiver was then given the written psychosocial survey to complete. A member of the
study team processed the blood samples at the CBCI clinic and immediately froze samples
(processing and storage delineated more below). The participants were given the choice to
1) complete the survey at CBCI and return it to a study team member OR 2) complete the
survey at home and mail it back in a self-addressed envelope that was provided by a study
team member upon request. We also obtained permission to give participants two reminder
calls, should they not return the survey.
A random study identification number was stamped on all pages of the survey as
well as the blood collection tubes. The first page of the survey contained general
instructions and blanks for name, address, and telephone number. Once surveys were
collected, the first page, with all identifying information, was removed. The study team
maintained a separate tracking document that contained the identifying information of the
study participants with the corresponding random study identification numbers. Finally,
the first page of the survey (with the identifying information) and the survey were stored
in separate, locked file cabinets that could only be accessed by members of the study team.
26


Study Design
The research study was a cross-sectional descriptive design, utilizing validated
psychosocial assessments and blood draws with partnered caregivers of MM patients.
Data Collection Procedures
The pre-blood draw survey required approximately 5-10 minutes to complete and
the psychosocial survey required about 30 minutes for the caregiver to complete. The blood
drawing procedure for the caregiver required approximately five minutes. Table 1 provides
an overview of the key areas assessed and the respective measures utilized. The titles of
the measures were not included in the caregiver psychosocial survey packet in order to
reduce response and social desirability biases.
Caregiver pre-blood draw screen. Literature suggests that many health factors
and behaviors may temporarily impact neuroendocrine functioning (Kiecolt-Glaser et al.,
2010). The pre-blood draw screen (located in Appendix F) inquired about acute and
chronic illness, medications, height and weight to calculate body mass index, mental
health conditions, and nicotine use. These variables were taken into consideration when
analyzing the neuroendocrine samples and will be described more in the data analysis
section.
Caregiver psychological processes. Caregiver psychosocial processes were
assessed in the domains of depression, negative affect, and illness-related distress.
Depression: The Center for Epidemiological Studies Depression Scale (CES-D).
The CES-D is a 20-item Likert scale developed to measure depressive symptoms in the
general population (Raldoff, 1977). The range of scores is 0-60, with higher scores
indicating greater depression. A score >_16 indicates a clinically significant level of
27


depression. The CES-D has been shown to have adequate reliability for use as a measure
of depressive symptoms in older adults (Hertzog, Van Alstine, Usala, & Hultsch, 1990)
and has been widely used in dementia caregiving research (Pinquart & Sorensen, 2003)
and other caregiver studies (Kim et al., 2008; Kim et al., 2011). The CES-D was scored
by first reversing four items (4, 8, 16, and 12) and then summing all of the responses as
suggested by scale developers (Raldoff, 1977). Mean scores are presented below to
describe the sample and CES-D total (summed) score was used as a continuous variable in
inferential statistical analyses.
Negative Affect: Positive and negative affect scale (PANAS). The PANAS is a 20
question self-report measure that was used to assess positive and negative aspects of affect
(Watson, Clark, & Tellegen, 1988). The PANAS asked participants to rate the extent to
which they feel certain emotions from one (very slightly or not at all) to five
(extremely) and yielded a sum for both positive and negative emotions. This scale has
been validated in non-clinical samples (Crawford & Henry, 2004). This scale has also been
used in marital studies (Graham et al., 2009; Kiecolt-Glaser et al., 2005), and studies
assessing neuroendocrine responses and affect (Taylor, Gonzaga, Klein, Hu, Greendale, &
Seeman, 2006). The PANAS was scored by summing all of the ten items that loaded onto
the Negative and Positive Affect subscales to yield a possible score of 50 on each subscale.
While means of the summed scores for all participants on both sub scales were summarized
to describe our sample, only negative affect subscale was used in inferential statistical
analyses.
Illness-related distress: Impact of events scale revised (IES-R). The IES-R is a
22-item self-report measure that assessed subjective distress caused by traumatic events
28


(Weiss & Marmar, 1997). For the purposes of this study, the partners MM illness was
conceptualized as a traumatic event. Respondents were asked to indicate how much they
were distressed or bothered by their loved ones illness during the past seven days by each
"difficulty" listed. Items were rated on a five-point scale ranging from zero ("not at all") to
four ("extremely"). The IES-R yields a total score (ranging from 0 to 88), which was the
continuous variable used in our inferential statistical models. In addition, subscale scores
can also be calculated by using the mean item response for the questions that loaded onto
the Intrusion, Avoidance, and Hyperarousal subscales (range 0 4). It is a revised version
of the older version, the 15-item IES (Horowitz, Wilner, & Alvarez, 1979). The IES-R
contains seven additional items related to the hyperarousal symptoms of PTSD, which were
not included in the original IES (Weiss & Marmar, 1997).
Caregiver- patient relationship characteristics. Patient-caregiver relationship
characteristics were assessed in the domains of relationship quality and attachment from
the caregivers perspective only.
Relationship Quality: Social relationships index (SRI). The SRI was used to assess
individuals perceptions of how positive (e.g. how helpful when needing advice,
understanding, or a favor) or negative (e.g. how upsetting/unpredictable when needing
advice, understanding or a favor) their significant other is in the context of asking for
advice, understanding, or a favor (Campo et al., 2009). This is a five question self-report
measure which respondents rate perception of partners as one (not at all) to six
(extremely) positive, upsetting, unpredictable, and conflictual. The SRI has been
validated for use in health studies. This scale yields three scores calculated by the mean
item response for relational positivity, relational negativity, and importance of relationship.
29


This data was used as descriptive data to further understand our samples perception of
how their significant other responds to requests for advice, understanding or a favor. This
scale was not used for inferential statistics.
Quality Marriage Index (QMI). The QMI is a six-item self-report scale designed
to measure marital satisfaction by inquiring about the stability and strength of the marital
relationship. Participants choose answers from a six-point Likert Scale (strongly agree to
strongly disagree). An additional question asked participants to rate the degree of their
overall happiness within the marriage, ranging from one = very happy to seven = very
unhappy. A total score is calculated by summing all of the question responses. Scores on
this scale were used as continuous variables for inferential statistics. Higher scores on
the QMI indicate higher marital satisfaction, whereas lower scores corresponded with
lower satisfaction (Norton, 1983). The wording of this measure in the caregiver
assessment packet was changed to relationship instead of marriage to be more
inclusive of all types of relationships.
Secure Attachment: Measure of attachment qualities (MAQ). The MAQ is a
measure of adult attachment patterns (Carver, 1997). Respondents were asked to rate their
degree to which they agree with various statements on a four-point Likert Scale ranging
from one (not at all) to four (extremely). The MAQ has four separate scales to assess
secure attachment tendencies (e.g. It feels relaxing and good to be close to someone),
avoidant tendencies (e.g. I get uncomfortable when my partner wants to get close), and
two scales reflecting aspects of the anxious-ambivalent pattern, one reflecting the desire
for merger called ambivalence-merger (e.g. I have trouble getting my partner to be as
close as I want them to be) and worry of abandonment called ambivalence-worry (e.g. I
30


often worry my partner doesnt really love me). The wording of the questions in the
caregiver assessment packet was changed to be specific to the caregivers feeling toward
only their significant other. This measure was initially included to determine level of secure
attachment for use in inferential analyses. However, because attachment was only specified
in Hypothesis 1.1, which was ultimately not tested due to difficulty obtaining oxytocin data
(described in more detail below), secure attachment and the three other MAQ subscale
totals were used only as descriptive information.
Caregiver quality of life: Caregiver Quality of Life Index-Cancer (CQOLC). The
CQOLC is a quality of life instrument designed to measure family functioning, perceived
burden, psychological adaptation, and psychological morbidity in family caregivers of
persons with cancer. The 35-item scale can be taken in ten minutes and has been used in
clinical and other settings (Weitzner, Jacobsen, Wagner, Friedland & Cox, 1999). A total
score is yielded for this scale by reverse scoring items (4, 10, 12, 16, 22, 27, 28, and 34)
and then summing all of the responses. This total score was used as a continuous variable
in the interferential statistical analyses.
Caregiver neuroendocrine regulation, immune functioning, and oxytocin
genotype. Blood was collected via blood draws (approx, volume = 20mL) conducted by
trained PSLMC phlebotomists. Samples were centrifuged at 3500rpm directly after
collection for ten minutes in a small clinical centrifuge and aliquots were collected using a
sterile Pasteur pipet into cryovials at the CBCI clinic. Materials for blood processing (e.g.
clinical centrifuge, pipettes, cryovials) were stored in a plastic Tupperware and brought to
the clinic for each participant. Samples were transported on ice and stored at -70 C in
the biobehavioral laboratory in the UCD Department of Psychology. Samples were frozen
31


until the study was completed and all samples were run together to avoid problems with
assay drift and interassay variability.
Oxytocin Receptor Genotyping. In order to genotype the participants for the OXTR
gene (oxytocin receptor gene) DNA was first isolated in the plasma sample. Once the DNA
was ready for genotyping, DNA was removed from the isolated sample and placed in a
separate polymerase chain reaction, or PCR, tube. PCR is a biochemical technology in
molecular biology used to amplify a single or few copies of a piece of DNA across several
orders of magnitude, generating thousands to millions of copies of a particular DNA
sequence. In this case, we amplified the OXTR gene DNA sequence. Unfortunately, the
SNP assay procedure used to identify the OXTR genotype yielded no results. Upon
consulting with Dr. Phi el and others who have expertise in this area, it appears that the
DNA concentrations from the samples were too low for detection in the SNP assay
procedure. This could have been due to either contamination during the DNA purification
procedure or degraded DNA due to collection and storage methods. Methods to trouble
shoot this challenge are discussed more in the Future Directions section.
Inflammatory markers (CRP, TNF- a, IL-6). Commercially available enzyme-
linked immunosorbent assay (ELISA) kits (R & D Systems) were used to measure levels
of CRP, IL-6 and TNF-a in the circulation. The concentration obtained represents the
overall level of systemic inflammation (both chronic and acute). Undiluted plasma samples
were tested in duplicate and according to the directions provided by the manufacturer for
TNF- a and CRP, while diluted plasma samples were used for the IL-6 ELISA per the
manuals instructions. These procedures suggest that samples are read at an optical density
at 450 nm using an automatic microplate reader (LabSystems MultiSkan). The amount of
32


inflammatory marker in each sample was determined using the standard curve generated
with each assay according to the manufacturers instructions. ELISA kits from the same
manufacturers lot were used for all assays for all measures. These assays show minimal
variability between the standard curves (less than 6% variability) in our laboratory (M.
Coussons-Read, personal communication, September 21, 2011). The mean of the
duplicates were used as the unit of analysis for statistical evaluation of these data. Results
from ELISAs were deemed usable based on the standard cure results that were yielded (C.
Phiel, (Professor of Integrative Biology, whose lab I ran these data in), personal
communication, January 2015). However, the ELISA for TNF- a yielded below detection
results, therefore TNF- a was not included in subsequent analyses. Reasons for this will be
discussed more in the Discussion Limitations section.
Caregiver characteristics. Caregiver characteristics were assessed using self- report
questions inquiring about sociodemographic information, caregiving variables, health
behaviors, and sleep.
Demographics. Sociodemographic variables were assessed via self-report questions
asking about age, gender, income, race, education level, occupation, and distance to travel
to the clinic. Age was used as a continuous control variable in the regression analyses. The
other variables were summarized to describe the sample.
Caregiving variables. Caregiving variables were assessed using self-report questions
asking about length of time being a caregiver, time since MM diagnosis, types of support
provided by the caregiver (e.g. physical, emotional, practical), level of caregiver
involvement, and perceived severity of patient illness. Perceived illness severity and length
33


of illness time were included as control variables in the regression models. The other
variables were used to describe the sample.
Health behaviors. Caregiver health behaviors were assessed using self-report
questions asking about diet, physical activity habits, alcohol, caffeine, medications, chronic
illness/mental health diagnoses. Amount of weekly physical activity and self-reported diet
quality were used as continuous variables within the inferential statistical analyses. Body
mass index (BMI) was used as a control variable in some of the regression models.
Medications and presence of chronic illnesses were entered as categorical variables to
determine whether or not differences existed with level of pro-inflammatory markers
between groups. Other health behavior variables were provided as descriptive information.
Sleep. Sleep was assessed using the Pittsburg Sleep Quality Index (PSQI). The
PSQI is a 19-item self-report measure that has been used to measure quality of sleep in a
clinical population. The instrument assesses sleep disturbances and quality during a one-
month duration. Seven component scores consist of sleep latency, subjective sleep
quality, sleep efficiency habits, sleep duration, sleep disturbances, daytime dysfunction,
and use of sleep medications (Buysse, Reynolds, Monk, Berman & Kupfer, 1988). A total
score was also calculated by summing these seven component scores for a range of 0-21,
where scores greater than five are indicative of poor sleep quality and scores less than
or equal to five are indicative of good sleep quality. The total score was used as a
continuous variable for the inferential statistical analyses and was included as an
indicator of health behaviors.
34


Table 1.
Key variables and sources of data
Key Variables Sources of Data
Caregiver Psychological Processes
Depression Center for Epidemiological Studies- Depression (CES-D)
General Affect Positive and Negative Affect Scale (PANAS)
Impact of Events (illness specific) Impact of Events Scale-Revised (IES-R)
Caregiver-Patient Relationship Characteristics
Relationship Quality Social Relationships Inventory (SRI) Quality Marriage Index (QMI)
Attachment Style Measure of Attachment Qualities (MAQ)
Caregiver Quality of Life
Quality of Life (QoL) Caregiver Quality of Life Index-Cancer (CQOLC)
Caregiver Neuroendocrine Regulation and Immune Response
Oxytocin genotype Inflammation (CRP, TNF-a, IL-6) Blood plasma (20mL) Blood plasma (20mL)
Caregiver Characteristics
Demographics Caregiving variables Health behaviors Sleep Age, gender, income, race, education Length of illness time, amount of time spent with patient, level of caregiving involvement, perceived illness severity Alcohol, smoking, diet, physical activity Pittsburg Sleep Quality Index (PSQI)
Patient Variables
Demographics Medical Household income Caregiver reported stage of disease, treatment side effects, length of illness
Note: TNF- a and oxytocin genotype were not included in the final analyses.
35


Data Analysis
Scale Reliability. Before conducting planned inferential statistics, a psychometric
assessment of the collected data was conducted to confirm internal consistency of the
measurement tools used in the analyses. All of the measures were determined to have
adequate internal consistency as indicated by a Cronbachs alpha greater than 0.7. The
values for the measures are as follows: CES-D (a = .89), QMI (a = .95), CQOLC (a =
.92), PANAS-Negative Affect (a = .86), PSQI (a = .78), IES-R Avoidance (a = .71),
IES-R Intrusion (a = .83), IES-R Hyperarousal (a = .77) and IES-R Total (a = .83).
Overview of scale reliability can be found in Table 2.
Table 2.
Tests of normality and internal consistency for variables used in inferential statistical
analyses_________________________________________________________________________
Variable (Scale used) Mean (SD) Media n Skewness (Std. Error) Kurtosis (Std. Error) Internal Consistency (Cronbachs alpha) Outliers removed (Z-score)
Sleep Quality 5.86 5.50 0.69 -0.51 0.78 none
(PSQI total) (3.95) (0.37) (0.73)
CRP 1.04 0.80 1.1 0.54 n/a none
(.83) (0.37) (0.72)
IL-6 0.10 .10 0.83 0.23 n/a 1
(0.03) (0.35) (0.69) (z = 4.52)
Relationship 31.21 33.5 -1.29 1.01 0.95 none
Quality (7.75) (0.37) (0.73)
(QMI total)
Illness-related 18.90 15.5 0.98 .75 0.83 none
Distress (12.61) (0.34) (0.67)
(IES-R total)
Quality of Life 52.33 50.50 0.58 -.23 0.92 none
(CQOLC) (21.61) (0.37) (0.73)
Negative Affect 18.64 17.00 0.88 0.68 0.86 none
(PANAS (6.88) (0.37) (0.73)
Negative)
Depression 12.94 10.00 .99 .70 0.89 none
(CES-D) (9.46) (0.34) (0.66)
36


Variable (Scale used) Mean (SD) Media n Skewness (Std. Error) Kurtosi s (Std. Error) Internal Consistency (Cronbachs alpha) Outliers removed (Z-score)
Physical Activity 269.21 180 1.00 0.71 n/a 1 (z = 8.27)
(219.2) (.37) (0.72)
Health Quality 1.78 2.0 .18 (.34) -.49
(.62) (.67)
Diet Quality 2.05 2.0 .32 -.21 n/a none
(.75) (.34) (.67)
Perception of 1.8 2.0 n/a none
Illness Severity (.82) .88 (.34) .42 (.67)
Length of Illness 23.63 15 1.67 2.96 n/a 2 (z =4.56,
Time (23.32) (.34) (.67) 6.23)
* Scores reflect data once outlier(s) were flagged and evaluated using a Z- score of +/-
2.58. Z-scores of skewness and kurtosis were determined with an absolute value of >
2.58.
*Cronbachs alpha was not calculated for single item measures.
Quantitative analysis. Data was first examined for outliers using box plots,
histograms, skewness, and kurtosis. Outliers were removed in four cases (one participant
for hours of physical activity; two participants for years of caregiving time) after
determining that the data was inaccurate when taken into context. For example, one
participant reported exercising 12 hours per day. While she had a highly physical job as a
rancher, this level of activity significantly skewed the data and had a z score of 8.27. In
addition, 1 participant IL-6 had a z-score of 4.25 and thus was removed for the inferential
analyses. An overview of central tendencies for variables included in inferential statistical
analyses can be found in Table 2. Furthermore, no transformations were needed for data
analysis.
Scores on psychosocial assessments and plasma level of pro-inflammatory makers
were summarized using a variety of exploratory analyses in SPSS (e.g. frequency, box
plots, histograms, measures of central tendencies, etc.) to examine and describe the data.
37


All of the data was tested for all other parametric assumptions of each respective
statistical test before analyses were conducted. Linearity, meaning that the data are
related in a linear fashion, is an assumption for bivariate correlations as well as a linear
multiple regression. This was assessed via plotting the data. For multiple regressions the
assumptions of normally distributed errors, and homoscedasticity (the variance of error
is the same across all levels of the IV, this was assessed graphically) were assessed (Field,
2005). Additional assumptions include: no multicollinearity (no two predictor variables
should correlate perfectly), independent errors (errors of each observation should be
uncorrelated as measured by the Durbin-Watson test, which should be between one and
three), and independence of outcome variable value from the predictor variables,
meaning that each value of an outcome variable comes from a separate entity. In addition,
for multiple regression, there is an assumption that no external variable (e.g. a variable
that has not been included in the regression model) is correlated perfectly with any of the
predictor variables. If this were the case, the model would not be reliable because other
variables exist that can predict the DV just as well. According to Field, in multiple
regression equations, predictors do not need to be normally distributed (2005).
Missing data. As expected there was minimal missing data. For psychosocial
measures with one item missing, mean imputation either from the subscale on which the
missing data was from (if there was a subscale) or the total sum was used. This was the
case on the CQOLC and CES-D. On the CQOLC, missing values were observed to be
most prominent on the positively phrased items on questions such as: I am satisfied with
my sex life; Family communication has increased. If more than one item was missing the
38


participant was list wise deleted from that particular analysis, such was the case for
participant 15 on the CQOLC on the analyses that included QoL (Aims 3, 4, and 5).
Statistical Power. Statistical power is the probability that a statistical test will
reject the null hypotheses, when the null hypothesis is actually false (Miles & Shelvin,
2001). Three factors that impact power include: 1) statistical significance criterion used
(e.g. a = 0.05); 2) magnitude of the effect of a given construct in a given sample; and 3)
sample size. Based on calculations using G*Power (Erdfelder, Faul, & Buchner, 1996), the
correlational analyses, a = 0.05, power = 0.80, would require n = 21 to detect a large effect
(r = 0.50), n = 64 to detect a medium effect (r = 0.30), and n = 614 to detect a small effect.
If a = 0.10 and power = 0.80, one would need n = 47 to detect a medium effect (r = 0.30)
and n = 448 to detect a small effect. Based on the power issue, an a = 0.10 would enable
us to find a medium or large effect significant. For the multiple regression analyses a =
0.05, power = 0.80, two predictors would need n= 31, to detect a large effect (f2 = 0.35)
and n = 68 to detect a medium effect (f2 = .15). One predictor variable would require 55
people to detect a medium effect (f2 = 0.15). Based on the Baron & Kenny model of
mediation, one to two predictors are needed per regression equation to determine a
mediational relationship (Baron & Kenny, 1986). Based on the sample for each statistical
test, which ranged from 50-44 (specific n for each analysis is included in Tables 9-13),
there is a (1- estimated power) probability that a Type II error occurred (failure to detect
significant relationship when one exists between two variables or a false negative). Post
hoc power which is based on estimated effect size, statistical tests used (e.g. bivariate
correlation or multiple regression), alpha (.05), and number of predictors for the regression
models will be discussed within each section that results are reported.
39


CHAPTER III
RESULTS
Recruitment Accrual and Attrition
Sixty-two percent (n = 32) of the participants were recruited from the mail out
recruitment process and 37% (n = 19) of the participants were recruited in person at CBCI.
The response rate from the mail outs was 27%, out of a total of 120 individuals who were
each sent one initial and one follow up letter. We were not able to track data about how
many people were approached about the study in-person versus how many people agreed
to participate. This was due to records not being kept about which caregivers declined or
consented by the multiple CBCI psychosocial staff, including social workers, post-doctoral
fellows, and practicum students, who were involved in recruitment efforts. In addition,
there were three individuals who consented but did not complete the survey. Out of those
three individuals, two did not return reminder phone calls and the other participant stated
that he was too busy to complete the survey. A total of 46 caregivers participated in the
blood draw portion of the study. The reasons for non-participation included the participants
stating that they had insufficient time and participants reporting that they did not want to
participate in the blood draw because of past difficulties with blood draws.
Sociodemographic Characteristics of the Participant Sample
The final sample included 51 caregivers, with a mean age of 62.9 (SD = 7.85).
Seventy- seven percent (n = 39) were female and 23% (n = 12) were male. Regarding
ethnicity, 88% (n = 45) identified as White, 6% (n = 3) identified as Latino/Hispanic, 2%
(n=l) identified as Asian/Pacific Islander, and the other 2% (n = 1) identified as Multi-
Ethnic. Regarding family structure, 91% (n = 50) reported that they were married and one
40


caregiver reported being in a committed relationship with a partner of the opposite sex.
The mean length of time in the relationship was 413.04 months (SD = 159.68) or 34.4
years. The majority of participants (65%, n = 33) had one to three children. However,
only 14% (n = 7) of participants had children who were living in the home at the time the
study was conducted. Regarding level of education, 10% (n = 5) completed less than or
equal to 12th grade, 25% (n = 13) completed some college or an associate degree, and
65% (n = 33) completed college/advanced degree. Sixteen participants (31%) were
currently employed full time, seventeen (33%) were employed part- time, sixteen (31%)
participants were retired, and one (2%) was on temporary medical leave from work.
Twenty percent of participants (n = 10) reported that their household income decreased
after their partner was diagnosed with multiple myeloma. Participant demographic
information is summarized in Table 3 below.
Table 3.
Participant demographic variables__________________________________________
Demographic Variables Frequency Mean (SD) Range
________________________________________(%)________________________________
Gender Female 39 (77%)
Male 12 (23%)
Ethnicity White 45 (88%)
Latino/Hispanic 3 (6%)
Other 3 (6%)
Family Structure
Marri ed/partnered 51 (100%)
Family Income (Before Cancer)
S0-50K 5 (10%)
S50-99K 20 (41%)
>$100K 24 (49%)
41


Demographic Variables Frequency (%) Mean (SD) Range
Family Income (Current)
$0-50K 9(18%)
$50-99K 20 (41%)
>$100K 20 (41%)
Education Level
<12th grade 5 (10%)
Some college/associates degree 13 (25%)
College/advanced degree 33 (65%)
Employment Status
Employed full time 16(31%)
Employed part-time/Homemaker 17 (33%)
Retired 16(31%)
Number of Children
0 1 (2%)
1-3 33 (65%)
4+ 10 (20%)
Children Living in the Home
Yes 7 (14%)
No 38 (75%)
Age (Years) 62.90 (7.85) 37-76
Length of Time in Relationship 413.04 90-708
(Months) (159.68)
* All percentages are based off a total N of 51.
Patient Illness Characteristics
Commonly reported treatment side effects included GI distress (nausea, diarrhea,
vomiting, and constipation), fatigue, and neuropathy. Less frequently reported side effects
included hair loss, change in appetite, rash, insomnia, body aches, weight changes, and
fever. While 39% (n = 20) of caregivers did not report changes in their significant others
personality or mood, 53% (n = 27) reported noticing negative changes such as increased
irritability, anxiety, depression, less optimistic, cognitive changes, withdrawal,
and more self-focused Many caregivers attributed increased irritability and aggression
to the corticosteroids (dexamethasone) that were prescribed for their significant other.
Interestingly, 8% (n = 4) of caregivers reported positive personality and mood changes in
42


their partner such that they were more thoughtful, more mellow, and more
introspective. Caregiver reported patient illness characteristics are summarized in Table
4.
Table 4.
Caregiver reported patient illness characteristics______________________________________
Variable______________________________% (n)_______________Median (Range)________________
Length of illness time (months) 14.5 (1-102)
Perceived severity of diagnosis
Very severe 41% (21)
Somewhat severe 43% (22)
Mildly severe or 12% (6)
Not at all severe 4% (2)
Stage
I 18% (10)
II 11% (6)
III 24% (13)
IV 9% (5)
Unknown or not provided 31% (17)
Exact Diagnosis Highly variable responses (e.g. kappa light chain, lambda light chain)
Treatments undergone*
Chemotherapy (Revlamid, Velcade) 91% (50)
SCT (1 or more) 80% (44)
Radiation 26% (14)
Dexamethasone 38% (21)
Treatment related side effects
Fatigue 44% (24)
GI distress 38% (21)
Neuropathy 22% (12)
Experience of physical pain
All the time 22% (11)
Most of the time 12% (6)
Some of the time 43% (22)
None of the time 22% (11)
Personality and mood changes experienced
Negative 53% (n = 27)
Positive 8% (n = 4)
No changes 39% (n = 20)
*Note: Some patients underwent multiple conjunct treatments. Figure 3 provides
information on the frequencies of various combinations of treatment modality.
43


Regarding patient illness characteristics, caregivers reported a median length of
illness time of 14.5 months with a range from 4-105 months. Caregivers reported a number
of various symptoms experienced by their significant other prior to receiving a multiple
myeloma diagnosis including orthopedic pain (back, hip, shoulder), elevations on blood
labs drawn at primary care physician (e.g. anemia, protein, white count), shortness of
breath, fatigue, blood in stools, and broken bones, which generally resulted in the patient
receiving an MRI or CT scan that revealed bone lesions and a subsequent multiple
myeloma diagnosis. Regarding the stage and specific MM diagnosis received, there was a
large amount of variability in reporting of stage and diagnosis. 31% of caregivers did not
know or did not provide staging information. Of those participants who did report stage,
24% (n = 13) of the patients were diagnosed with Stage III, followed by Stage I (18%, n =
10), Stage II (11%, n = 6) and Stage IV (9%, n = 5). Reasons for the variability of responses
regarding exact diagnosis are included in the discussion section. Finally, regarding
treatment modalities, 8% (n = 4) of caregiver reported that their significant other received
only chemotherapy, 4% (n = 2) received chemotherapy and radiation, 64% (n = 32)
received chemotherapy and 1-2 stem cell transplants (autologous and or allogeneic are both
included) and the remaining 24% (n = 12) received a combination of chemotherapy,
radiation, and one or more stem cell transplants. The two most commonly reported
chemotherapies utilized were Revlimid and Velcade. Other types of adjunct treatments or
procedures reported included kyphoplasty, back surgery, dialysis, and apheresis or blood
transfusions. Table 3 and Figure 3 provide an overview of the treatment modalities. Finally,
96% of caregiver reported that their significant other experienced some treatment related
44


side effect. The most common side effects were fatigue (44%, n = 24), GI distress, which
included diarrhea, nausea, vomiting, constipation (38%, n = 21), and neuropathy (22%, n
= 12). Other reported side effects included hair loss, change in appetite, rash, insomnia,
body aches, weight changes, and fever.
Rad + Chemo
Figure 3. Types of treatment received by patient
Caregiving Variables for the Participant Sample
Caregivers reported a median time of 16 months serving as a caregiver (range = 1
- 358). Forty-nine percent (n = 25) of caregivers reported providing an adequate level of
care, all of the time. An overview of caregiver response to provision of adequate care
is provided in Figure 4.
45


All the time Most of the Some of the None of the
time time time
Figure 4. Caregiver perception of provision of adequate care.
Specifically, 47% (n = 24) of caregivers reported conducting physical caregiving
tasks such as helping their loved one bathe, eat, dress, spending a median of 120 minutes
per day (range = 0 720) for a median time of 11.5 months (range = 0 108). The entire
samples (n = 51) of caregivers reported providing emotional support to their loved one such
as talking to them about their illness and boosting their partners spirits, for an average of
60 minutes per day (range = 2- 709) for a median time of 11 months (range = 1 109).
Finally, 92% (n = 46) of caregivers reported providing practical care such as medication
management, household chores, transportation, and financial support for an average of 240
minutes per day (range = 0 660) over a median time of 11 months (range = 0 105). A
summary of the caregiving variables can be found below in Table 5.
46


Table 5.
Caregiver level of involvement and health behaviors
% (n) Median (range)
Time as a caregiver 16 (1 358)
(months)
Physical Tasks 47% (24)
Daily (minutes) 120 (0-720)
Total duration (months) 11.5 (0-108)
Emotional Tasks 100% (51)
Daily (minutes) 60 (2-707)
Total duration (months) 11 (1 109)
Practical Tasks 92% (46)
Daily minutes 240 (0-660)
Total duration (months) 11 (0-105)
Tobacco (cigarettes per day)
Yes 5 (10%) 10(1- 11)
No 42 (82%)
Mean (SD)
Physical Activity (minutes per week)
Yes 46 (90%) 315 (SD = 324.59)
No 2 (4%)
Alcohol (drinks per week)
Yes 29 (57%) 1.9 (SD = 2.02)
No 22 (43%)
Caffeine (drinks per day)
Yes 46 (90%) 2.10 (SD = 1.48)
No 5 (10%)
Sleep Quality (PSQI)
Poor 25 (49%) 5.86 (SD = 3.95)
Good 26 (51%)
* All percentages are based off a total N of 51.
Health Related Variables for the Participant Sample
Regarding overall quality of health, 33% (n = 17) of caregivers reported being in
excellent health, 57% (n= 29) reported being in good health, and 10% (n = 5) reported
47


being in fair health. Similarly, 26% (n = 13) of caregivers reported having an excellent
diet, 51% (n = 26) reported having a good diet, 20% (n = 10) reported having a fair
diet and 3% (n = 2) report a poor diet. All responses are summarized in Figure 5.
Health
Diet
Figure 5. Caregiver reported health and diet quality.
In terms of health behaviors, three caregivers reported having only one meal per
day, 18 caregivers reported having 2-2.5 meals per day, 24 reported consuming three
meals per day and six caregivers consumed more than three meals per day. Specifically, 47
caregivers consumed at least one serving of leafy vegetables, regular vegetables, and whole
grain per day, 46 caregivers consumed at least one serving of fruit and dairy per day, and
48 caregivers consumed at least one serving of meat per day.
The majority (80%, n = 41) of caregivers reported engaging in recreational
activities, which ranged from hobbies, physical activity, entertainment, and social
experiences. In addition, 46 of the caregivers (90%) indicated that they participate in
moderate physical activity for more than 10 minutes per day, with a median of 5.5 days per
48


week (range = 1-7) for a mean of 315 minutes per week (SD = 324.59). Regarding alcohol
use, 57% (n = 29) of participants endorsed consuming alcohol, with an average of 1.9
drinks per week (SD = 2.02). In addition, 90% (n = 46) of participants reported consuming
one or more caffeinated beverages daily and 10% (n = 5) of participants endorsed currently
smoking tobacco with a median of 10 cigarettes per day (range = 1-11) for a median of 15
years (range = 40). Regarding quality of sleep, scores on the PSQI indicated that 51% (n =
26) of participants experience good sleep quality, while the remainder experience poor
sleep quality. A summary of the caregiver health behaviors is presented in Table 4.
Additional Health Related Variables for Participants who Completed Blood Draw
For participants who took part in the blood draw portion of the study, the mean
height of participants was 64.85 inches (SD = 4.60), the average weight was 161.53
pounds (SD = 5.47), and the average BMI was 27.11 (SD = 6.23). Regarding BMI, 23
participants (45%) were in the Normal Range (BMI = 18.5-24.99). However, the
majority (n = 32, 70%) of caregivers reported having a chronic health condition, the most
common of which were cancer (n = 9, 20%), hypertension (n = 7, 15%), hypothyroidism
(n = 5, 11%) heart disease (n = 5, 11%) and sleep apnea (n = 4, 9%), with 26% of
participants with two or more chronic health conditions. Regarding the presence of
mental illness, 11% (n = 5) caregivers endorsed a Major Depressive Disorder diagnosis
and 4% (n = 2) endorsed a Generalized Anxiety Disorder diagnosis. Five caregivers
(11%) reported being on a psychotropic medication (e.g. Zoloft, Wellbutrin). This
information is presented in Table 6 below.
49


Table 6.
Caregiver health related variables and neuroendocrine data
Variable (%) Frequency Mean (SD) Range
Body Mass Index (BMI)
Underweight (< 18.5) 0 (0%)
Normal (18.5-24.99) 23
(45%)
Overweight (25-29.99) 10
(20%)
Obese Class I (30-34.99) 10
(20%)
Obese Class II (35-39.99) 3 (6%)
Obese Class III (> 40) 2 (4%)
Presence of Acute Illness 2 (4%)
Presence of Chronic Illness 30 (65%)
Presence of Psychiatric Illness 7(15%)
Height (Inches) 64.85 (4.60) 54-73
Weight (Pounds) 161.53 (5.47) 110-260
BMI 27.11 (6.23) 19.22-46.48
CRP plasma concentration (ng/mL) 1.08 (0.80) 0.21-3.24
IL-6 plasma concentration (pg/mL ) 0.13 (0.17) 0.07-0.28
*BMI ranges are per the International Classification of BMI (World Health Organization, 2014).
* Percentages are calculated out of 46.
Psychosocial Characteristics of the Participant Sample
Psychological processes. Caregiver psychosocial processes were assessed in the
domains of depression, general affect, and distress associated with the illness of the loved
one. A summary of the descriptive statistics for the measures can be found in Table 7.
Depression. On the Center for Epidemiological Studies Depression Scale (CES-D),
participants had a mean score of 13.19 (SD = 9.56), indicating that on average, caregivers
were below clinical levels of depression. However, sixteen participants (31%) were above
the clinical cut off score of 16, indicating clinically significant levels of depression.
Affect On the Positive and Negative Affect Scale (PANAS) caregivers reported a
mean positive affect score of 33.13 (SD = 7.75) compared to negative affect (M= 18.64,
50


SD = 6.88). Scores were out of a possible 50 with higher scores indicating higher levels of
positive or negative affect.
Illness-related distress. The mean level of distress on the Impact of Events Scale-
Revised was 19.35 (SD = 12.67) out of a total possible score of 88, where higher scores
indicate more illness-related distress. The subscale scores, which were generated using the
item mean for each subscale, were the following: illness-related intrusive thoughts (M =
1.01, SD = 0.75), hyperarousal (M = 0.84, SD = 0.70), and avoidance (M = 0.79, SD =
0.56).
Quality of Life. On the (CQOLC-C), caregivers had a mean score of 52.33 (SD =
21.61) out of a potential total of 140 where higher scores indicate lower quality of life.
Caregiver- patient relationship characteristics. Patient-caregiver relationship
characteristics were assessed via the caregivers perspective within the domains of
relationship quality and attachment.
Relationship Quality. On the Social Relationships Inventory (SRI) caregivers
scored the following: relational positivity (M= 4.29, SD = 1.40) and relational negativity
(M = 2.34, SD = 1.12). The majority of caregivers (78%) reported that their significant
other was extremely important to them (M=5.76, SD = 0.52). On another measure of
relationship quality, the QMI, caregivers reported an average of 31.21 (SD = 7.75) out of
a possible 36, indicating on average caregivers perceived their relationships to be strong,
stable, and happy.
Attachment. On the Measure of Attachment Quality caregivers scored the
following on the subscales: Secure (M= 10.48, SD = 2.00), Avoidant (M = 7.83, SD =
51


3.22), Ambivalence-Merger (M= 5.1, SD = 2.30), and Ambivalence Worry (M= 3.54,
SD = 1.21). Higher scores on each sub scale indicate a higher level of that attachment style.
Table 7.
Caregiver scores on psychosocial measures
Scale/Construct Mean (SD) Interpretation
Depression (CES-D) 13.19 (9.56) Higher scores indicate more depression. Scores above 16 indicate clinical levels of depression (16, 31%).
Affect (PANAS)
Positive 33.13 (7.75) Higher scores indicate more positive affect.
Negative 18.64 (6.88) Higher scores indicate more negative affect.
Illness-related Distress (IES-R) 19.35 (12.67) Higher scores indicate more illness-related distress
Intrusive Thoughts 1.01 (0.75) Out of 4. Higher scores indicate more severe sx.
Hyperarousal 0.84 (0.70) Out of 4. Higher scores indicate more severe sx.
Avoidance 0.79 (0.56) Out of 4. Higher scores indicate more severe sx.
Importance of partner 5.76 (0.52) Out of a possible 6. Partner is extremely important.
Relational Negativity 2.34 (1.12) Higher scores indicate more negativity, out of 6.
Relational Positivity 4.29 (1.40) Higher scores indicate more positivity, out of 6.
Overall Quality (QMI) 31.21 (7.75) Out of a possible 36. Higher scores indicate stronger relationship.
Attachment Style (MAQ)
Secure 10.48 (2.00) Out of 12. Higher indicates more secure attachment.
Avoidant 7.83 (3.22) Out of 20. Higher indicates more avoidant attachment.
Ambi val ence-W orry 3.54 (1.21) Out of 12. Higher indicates more A-W attachment.
Ambivalence-Merger 5.1 (2.30) Out of 12. Higher indicates more A-M attachment.
Quality of life (CQOLC- 52.33 (21.61) Out of a total of 140 where higher scores equal
C) lower quality of life.
*Caregiver Quality Of Life-Cancer (CQOLC); Impact of Events-Revised (IES-R); Social
Relationships Inventory (SRI); Measure of Attachment Quality (MAQ); Quality of
Marriage Index (QMI); Positive and Negative Affect Scale (PANAS); Center for
Epidemiological Studies Depression Scale (CES-D)
*Note: Additional interpretative information for scales is located in the Discussion
Section.
52


Caregiver Neuroendocrine Variables
A total of 46 caregivers participated in the blood draw portion of the study. There
were no significant differences between demographic variables for total sample compared
to demographic variables for those who also participated in the additional neuroendocrine
portion of the study (however, as noted, power was very low to detect differences). For the
neuroendocrine data, t-tests were utilized and no significant differences were found in
pro-inflammatory marker levels between participants on medication compared to no
medication, CRP: t(44) = -.30,/) >.05; IL-6: t(44) = -1.29,/) >.05; Participants with a
BMI of above 25 compared to aBMIbelow 24.99, CRP: t{44) = -1.78,/) >.05; IL-6: t{44)
= -1.86,p >.05; Nor participants with a chronic illnesses compared to those with no
chronic illness: CRP: t(44) = 1.92,p >.05; IL-6 : t(44) = 1.37,/) >.05. Therefore we were
able to group all individuals together in subsequent analyses. These findings are
presented in Table 8 below.
Table 8.
Group differences for pro-inflammatory markers (CRP and IL-6)
N Mean (SD) t df
CRP
Medication
Yes 30 1.05 (.74) -.30 44
No 16 1.12 (.94)
BMI
<24.99 21 .85 (.79) -1.78 44
>25 25 1.26 (.77)
53


N Mean (SD) t df
Chronic Illness
Yes 30 1.23 (.81) 1.92 44
No 16 .77 (.58)
IL-6
Medication
Yes 30 .09 (.01) -1.29 44
No 16 .10 (.02)
BMI
<24.99 21 .09 (.01) -1.86 44
>25 25 .10 (.02)
Chronic Illness
Yes 30 .10 (.02) 1.37 44
No 16 .09 (.01)
Plasma levels of IL-6 were found to be an average of 0.13 ng/pl (SD = 0.17) and
plasma levels of CRP was 1.08 (SD = 0.80). TNF-a results were below detection level
which may signify a problem with the preparation of the samples or in the laboratory
procedures. As a result, we were unable to include the TNF-a results in the additional
analysis. Also as mentioned above, difficulties purifying the DNA impeded us from
obtaining the OXTR genotype information. The results for the other biomarker data are
presented in Table 6.
Results of Inferential Statistical Analyses for Each of the Study Aims
Aim 1: Evaluate the relationships between caregiver oxytocin genotype,
psychological processes, and caregiver relationship characteristics in the context of caring
for a loved one with MM.
54


Hypothesis 1.1: Caregivers with GG oxytocin genotype will report a higher number
of positive caregiver-patient relationship characteristics (secure attachment style, high
relationship quality).
Hypothesis 1.2: Caregivers with GG oxytocin genotype will report lower levels of
distress (depression, negative affect, illness-related distress).
Due to methodological difficulties with the SNP assay and the DNA purification
procedures, OXTR genotype data was not obtained and Aim 1 was unable to be completed
for this study.
Aim 2. Evaluate the relationships between caregiver pro-inflammatory markers,
psychological processes and health variables.
Hypothesis 2.1: Caregivers who report higher levels of distress (depression,
negative affect, illness-related distress) will have higher plasma pro-inflammatory markers.
In order to evaluate the relationship between levels of pro-inflammatory cytokines (CRP
and IL-6) and distress (depression, negative affect, and illness-related distress), bivariate
correlations were conducted. CRP was not significantly associated with depression (r =
.06,p= .68), negative affect (r= .l$,p = .44), nor illness-related distress (r = .11, p = .48).
Similarly, IL-6 was not significantly associated with depression (r = -.21, p = 0.18),
negative affect (r = -.05,p = .76), nor illness-related distress (r = -.04,p = .80). Regarding
the power of these analyses, G*POWER (Erdfelder, Faul, & Buchner, 1996) estimated that
the power of the correlational analysis examining IL-6 and depression was .28, indicating
that there is a .72 probability that a Type II error occurred assuming that a relationship
exists between IL-6 and depression in the sample. Estimated effect size or strength of the
correlational relationship, can be interpreted using the absolute value of the Pearsons r
55


coefficient for the correlational analyses. For all of the above correlational analyses the
effect size ranged from .21 (small) to .04 (nil). Overview of these correlational
relationships can be found in Table 9.
Table 9.
Correlation table for caregiver distress and inflammatory marker variables
Negative affect Depression Illness-related distress CRP IL-6 Age
Depression .80**
(n) (50)
Illness- .65** 70**
related (50) (50)
distress (n)
CRP .18 .06 .11
(n) (44) (44) (44)
IL-6 -.05 -.21 -.04 29**
(n) (44) (44) (44) (46)
Age -.08 -.18 -.01 -.12 .08
(n) (51) (50) (50) (44) (44)
Length of -.09 -.14 -.07 -.12 -.13 .13
illness time (49) (48) (48) (42) (42) (49)
(n)
p<.01, p< .05 (two-tailed)
Hypothesis 2.2: Caregiver distress (depression, negative affect, and illness-related
distress) will significantly predict the level of caregiver pro-inflammatory markers while
controlling for caregiver age, length of illness time, and caregiver perception of illness
severity. Specifically, within those models, depression, negative affect, and illness-related
distress will be significant individual predictors of the level of pro-inflammatory markers
above and beyond the control variables.
In order to determine whether caregiver distress (depression, negative affect, and
illness-related distress) predicted caregiver pro-inflammatory markers, six separate linear
regression models were conducted with depression, negative affect, and illness-related
56


distress each predicting CRP and IL-6 separately, controlling for length of illness time,
caregiver perception of illness severity, and age. Length of illness time and caregiver
perception of illness severity were both single item questions and were included in the
model as control variables, to help control for variation in the sample due to caregivers
being at different points in their caregiver trajectory. In addition, older age has been shown
to be related to higher systemic levels of inflammation, so this variable was included in the
model to account for variability in our sample so that we can understand the unique
contribution that the predictor variables, caregiver distress (depression, illness-related
distress and negative affect), had on the change in the dependent variable, IL-6 and CRP.
None of the models were significant and the regression coefficients are therefore not
included in the text, but can be found in Tables 10 and 11. In addition, the Tables contain
the standardized beta ((f) coefficient, which is a standardized measure of the relationship
between the individual predictor and the DV within the regression model. Standardized
beta fi utilizes standard deviations, making its value directly comparable across different
models and across difference variables within models and thus give a better insight into the
importance of a predictor (Field, 2005). The value is therefore a measure of the individual
contribution (effect size) for a single predictor in a regression model (Field, 2005). In
addition, Cohens f, which is also in the tables, was used as a measure of effect or strength
of a full regression model. Power analysis for the full models revealed the following:
Model 1 (IL-6 predicted by depression, controlling for length of illness, perception of
illness severity, and age): Power = .41 (Total n = 42, 4 predictors,a = .05, f = .14); Model
2 (IL-6 predicted by negative affect, controlling for length of illness time, perception of
illness severity, and age): Power = .26 (Total n = 42, 4 predictors, a =.05, f = .09); Model
57


3 (IL-6 predicted by illness-related distress, controlling for length of illness time,
perception of illness severity, and age): Power = .24 (Total n = 42, 4 predictors, a = .05, f2
= .08). An overview of all of these findings can be found in Tables 10 and 11.
Table 10.
Summary of multiple regression analysis for caregiver distress variables predicting CRP
(n = 42)_______________________________________________________________________________
Variable B SEB fi R 2 Cohens./2
Model 1 Control: Length of illness time (months) -.00 .01 -.09 .02 .03
Control: Perception of illness severity .01 .16 .02
Control: Age -.01 .02 -.07
Depression Model 2 .01 .02 .05 .07 .08
Control: Length of -.00 .00 -.08
illness time (months) Control: Perception of .08 .17 .08
illness severity Control: Age -.00 .02 -.02
Negative affect Model 3 .03 .02 .23 .03 .05
Control: Length of -.00 .01 -.07
illness time (months) Control: Perception of .02 .16 .02
illness severity Control: Age -.01 .02 -.08
Illness-related distress .01 .01 .12
Table 11.
Summary of multiple regression analysis for caregiver distress variables predicting IL-6 (n = 42)
Variable B SEB p R2 Cohens f
Model 1 Control: Time of -.00 .00 .12 .14 -.23
illness (months) Control: Perception -.03 .03 -.12
of illness severity Control: Age .00 .00 -.00
Depression -.01 .00 -.32
58


Variable B SEB P R 2 Cohens./2
Model 2 Control: Time of -.00 .00 -.18 .08 .09
illness (months) Control: Perception of illness severity -.03 .04 -.15
Control: Age .00 .00 .04
Negative affect Model 3 -.01 .01 -.24 .07 .08
Control: Time of .00 .00 -.21
illness (months) Control: Perception of illness severity -.02 .03 -.10
Control: Age .00 .00 .08
Illness-related distress .00 .00 -.20
Hypothesis 2.3: Caregivers reporting better health behavior indicators, as measured
by sleep quality, diet quality, and physical activity levels, will have lower levels of plasma
pro-inflammatory markers.
The relationships between health indicators (sleep quality, level of physical activity, and
diet quality) and pro-inflammatory markers (CRP and IL-6) were assessed using bivariate
correlations.
CRP was not significantly related to sleep quality (r = .03,p = .84). However, CRP
was significantly associated with level of physical activity (r = -.41, p < 01) and self-
reported diet quality (r = -.27, P < .05). IL-6 trended towards correlational significance
with sleep quality (r = -.22, p = .15) and was significantly associated with the level of
physical activity (r = -.40, p < .01). IL-6 and self-reported quality of diet were not
associated. Many of the r values indicated a small to medium strength of association as
indicated by r values ranging from 1-.3 and ideally would need a sample of 65 or more to
detect a significant relationship based on the a priori power analysis. These analysis
59


included 42-46 participants as denoted in Table 12, meaning that these tests were likely
underpowered to detect a significant relationship between the identified variables. For
example, using G*Power it is estimated that the power of the analysis of the relationship
between IL-6 and sleep quality, which included a sample of 43, a = .05 and an estimated
effect size of .22, was .30. This suggests that there is a .70 probability that a Type II error
occurred given a small relationship exists between those two variables. An overview of
these findings can be found in Table 12.
Table 12.
Correlation table of caregiver health variables and inflammatory markers
CRP IL-6 Sleep quality Physical activity Diet BMI
Age -.12 .05 -.12 -.33* -.16 -.13
(n) (44) (44) (50) (49) (51) (46)
CRP 29** .03 _ 4t** -27* 27*
(n) (46) (43) (42) (44) (46)
IL-6 -.22 _ 4Q** -.16 .24
(n) (43) (42) (44) (46)
Sleep quality -.11 39** -.10
(n) (49) (50) (45)
Physical .07 -.10
activity (n) (49) (44)
Diet .11
(n) (46)
BMI
(n)
**p<.01, *p< .05 (two-tailed)
Note: BMI and age were included in the regression analyses as controls.
Hypothesis 2.4: Caregiver self-reported health behavior indicators (sleep quality,
diet quality, and physical activity level) will together predict level of caregiver pro-
inflammatory markers (IL-6, CRP), while controlling for age, length of illness time, and
body mass index. Within that model, it was hypothesized that sleep quality, diet quality,
60


and level of physical activity would be significant individual predictors of level of
caregiver pro-inflammatory markers above and beyond the control variables.
These relationships were assessed with two linear regression models, controlling
for age, BMI, and length of illness time. BMI and age have been suggested in the
literature to impact level of inflammation (Kiecolt-Glaser et al., 2010) so these variables
were included in the model to account for variability in our sample so that we can
understand the unique contribution that the predictor variables, quality of sleep, quality of
diet, and level of physical activity, had on the change in the dependent variables, IL-6 and
CRP. The full Model 1 that predicted CRP, indicated that 34% of the variance in CRP
can be accounted for by diet quality, sleep quality, age, hours of physical activity per
week, length of illness time and BMI together (R2 = 0.34, F (7, 32) = 2.40, p < 0.05).
Specifically of interest, the unique contribution within the model accounted for by hours
of physical activity per week, was significant (fi = -0.47, p > .01). Sleep quality and diet
quality were both not significant in this model as individual predictors. The overall
strength of the Model 1 which predicted CRP, was Cohen sf2= .52. The full Model 2 that
predicted IL-6, indicated that 21% of the variance in IL-6 can be accounted for by diet
quality, sleep quality, age, length of illness time, hours of physical activity and BMI
together (R2 = 0.21, F (7, 32) = 1.22, p = 0.32). None of the individual predictors, nor the
full models were significant. The overall strength of Model 2, which predicted IL-6, was
Cohen sf2 = .27. Using an effect size estimated at .27, a sample of 40 participants that
were included in this analysis, 6 predictors, and an a = .05, the post hoc power for Model
2 was .62. Overview of these regression findings can be found in Table 13.
61


Table 13.
Summary of multiple regression analysis for caregiver health behavior variables
predicting inflammatory markers (n = 40)
Variable B SEB P R2 Cohens./2
Model 1 (CRP) Control: Length of illness time -.04 .01 -.09 .34 .52
Control: Age .00 .02 .09
Control: BMI .03 .02 .26
Sleep quality -.01 .04 -.06
Physical activity .08 .02 _ 47**
Diet quality Model 2 (IL-6) .22 .19 .20 .21 .27
Control: Length of illness time -.02 .01 -.19
Control: BMI .01 .01 .18
Control: Age .01 .01 .18
Sleep quality -.01 .01 .18
Physical activity .01 .05 -.27
Diet quality .02 .05 .09
**p<. 01 (two-tailed)
Aim 3. Evaluate the association between caregiver biomarkers and caregiver
quality of life.
Hypothesis 3.1: Levels of plasma pro-inflammatory markers will be inversely related to
quality of life.
The relationship between pro-inflammatory markers and caregiver quality of life
was assessed with two bivariate correlations, which revealed that quality of life was not
significantly related to CRP (r = .17, p = .28) nor IL-6 (r = -.18,/?= .24). A sample of 43
was included in both analyses. Therefore, the power for these analyses was estimated at
.29 and .32, respectively. An overview of these relationships, including variables that are
62


included in the regression model for this aim (described more below) can be found in Table
14.
Table 14.
Correlation table for caregiver pro-inflammatory markers and quality of life
IL-6 Perception of illness severity Age Length of illness time Quality of life
CRP (n) .39* .04 (44) -.12 -.12(42) .17 (43)
(46) (44)
IL-6 (n) .02 (44) .05 -.02 (42) -.18 (43)
(44)
Perception of -.22 .14 (49) .01 (50)
illness severity (51)
(n)
Age (n) .13 (49) -.28* (50)
Length of illness .13 (48)
time (n)
*p<.05 (two-tailed)
Hypothesis 3.2: Levels of pro-inflammatory markers (IL-6, CRP) will predict
caregiver quality of life, while controlling for age, length of illness time, and caregiver
perception of illness severity. Within those models it is hypothesized that pro-inflammatory
markers will be significant individual predictors of caregiver QoL above and beyond the
control variables.
In addition, a multiple regression analysis indicated that together IL-6, perception
of illness severity, age, and length of illness time accounted for 19% of the variance in
caregiver quality of life (R2 = 0.19, F (4, 36) = 2.21 ,p = 0.09). While this finding was not
significant, it trended towards significance. IL-6 was not found to be significant as an
individual predictor in the model, however age was found to be a significant predictor of
quality of life within this model (fi = -.42 ,p< .01). Similarly, a second model that included
63


CRP as the continuous predictor variable, and perception of illness severity, age, and length
of illness time as control variables, indicated that together those variables accounted for
19% of variance in caregiver quality of life (R2 = 0.19, F (4, 36) = 2.17, p = 0.09), where
age was also significant as an individual predictor of caregiver quality of life in this model
(fi = -Al,p< .05).
The overall strength of the Model 1 (which included IL-6), was Cohens/^ .23
and Model 2 (which included CRP) was Cohens,/2 = .23. Using an estimated effect size
of .23, a total sample of 40 participants, 4 predictors and an a = .05, post hoc estimated
power is estimated at .60 for both models. Overview of results can be found in Table 15.
Table 15.
Summary of multiple regression analysis for pro-inflammatory markers predicting
caregiver quality of life (n = 40)
Variable B SEE fi R 2 Cohens,/2
Model 1 Control: Time of -.01 .01 -.13 .19 .23
illness (months) Control: Perception of illness severity -4.52 3.96 -.18
Control: Age -1.10 .41 -.42*
IL-6 Model 2 Control: Time of -220.02 193.10 -.17 .19 .23
illness (months) Control: Perception of illness severity -4.58 3.93 -.19
Control: Age -1.08 .42 -.41*
CRP 4.39 4.07 .16
*p<.05 (two-tailed)
Hypothesis 3.3: Quality of life will be significantly associated in the positive
direction with OXTR GG phenotype.
64


This aim was not able to be evaluated due to issues with the DNA purification and SNP
assay processes.
Aim 4. Evaluate the relationship between caregiver health behavior indicators as
measured by sleep quality, level of physical activity, self-reported diet quality and
psychological processes (depression, negative affect, illness-related distress).
Hypotheses 4.1: Quality of sleep, quality of diet, and level of physical activity will
be inversely related to level of depression, negative affect and illness-related distress.
In order to examine the relationship between caregiver health behavior indicators and
psychological processes, bivariate correlations were utilized. Results indicated that sleep
quality was significantly associated with depression (r = .56, p < .01), negative affect (r =
37 ,p < .01) and illness-related distress (r = .59, p < .01). In addition, self-reported diet
quality was significantly associated with depression (r = .32, p < .05) and illness-related
distress (r = .30, p < .05). Again, it should be noted that negative affect, depression, and
illness-related distress were conceptualized to measure one underlying construct (caregiver
distress) and are therefore significantly intercorrelated. Correlational values can be located
in Table 16. However, level of physical activity was not significantly related to any of the
three measures of caregiver distress/psychological processes (depression, negative affect,
and illness-related distress). Sample size for these analyses ranged from 48-50 participants.
An estimated power for the relationship between physical activity and depression is .50,
indicating that there is a .5 probability that a Type II error occurred.
65


Table 16.
Correlation table of caregiver psychological processes and health variables.
Negative Affect Illness- related distress Physical activity Sleep quality Diet quality Perception of illness severity
Depression .80** 70** -.28 (48) .56** .32* -.03
(n) (50) (50) (49) (50) (50)
Negative .65** -.06 (49) 27** .25 -.09
Affect (50) (50) (51) (49)
(n)
Illness-related -.18 (48) .59** .30* .04
distress (n) (49) (50) (50)
Physical -.02 .15 .30*
activity (49) (50) (49)
(n)
Sleep quality 39** .16
(n) (50) (50)
Diet quality .30*
(n) (51)
p<.01, p< .05 (two-tailed)
Note: Poorer sleep quality is indicated by higher scores on PSQI.
Hypothesis 4.2: Sleep quality, diet quality and level of physical activity together
will predict caregiver distress (depression, affect, illness-related distress) after controlling
for length of illness time and perception of illness severity. Within that model it is
hypothesized that sleep quality, diet quality and physical activity will be significant as
individual predictors of distress (depression, affect, illness-related distress) above and
beyond the control variables.
Three separate multiple regression models were conducted predicting depression,
illness-related distress, and negative affect by sleep quality, diet quality and level of
physical activity, controlling for length of illness time and perception of illness severity. In
Model 1, which predicted depression, 38% of the variance in depression was accounted for
by sleep quality, diet quality, level of physical activity, length of illness time and perception
66


of illness severity (R2 = 0.38, F (5, 40) = 4.86, /K0.001). Within Model 1, better sleep
quality was significantly associated with depression scores (fi = 0.52,/KO.01). For Model
2, which predicted negative affect, 30% of the variance in negative affect was accounted
for by sleep quality, diet quality and level physical activity, length of illness time and
perception of illness severity (R2= 0.30, F (5, 41) = 3.56, p < 0.01). Again, sleep quality
was significantly associated with negative affect scores (fi = 0.49, /K0.01). Finally in
Model 3, which predicted illness-related distress, 39% of the variance in illness-related
distress was accounted for by sleep quality, physical activity, length of illness time and
perception of illness severity (R2 = 0.39, F (4, 42) = 6.82, p < 0.01). Within Model 3, better
sleep quality was also significantly associated (as an individual predictor) with illness-
related distress level (fi = 0.63, p < .01). An overview of all regression findings can be
found in Table 17. However, it should be noted that illness-related distress, negative affect
and depression were all highly intercorrelated, meaning that they are measuring the
underlying construct of caregiver distress and therefore the findings of the three
aforementioned regression models are all related. The estimated effect or strength of
relationship for the overall model (Cohens/)) and the individual predictors (fi) can be
found in Table 17. Post hoc power analyses were not calculated for these tests; it can be
assumed that because we had significant findings that we had adequate power.
67


Table 17.
Summary of multiple regression analysis for health behaviors predicting caregiver
distress (n = 46)
Variable B SEB fi R2 Cohens./2
Model 1 .38 .61
(DV: Depression) Control: Length of -.05 .05 -.13
illness time (months) Control: Perception of illness severity -.91 1.68 -.06
Sleep quality 1.19 .32 .52*
Physical activity .01 .01 .08
Diet quality Model 2 2.30 1.78 .18 .30 .43
(DV: Negative Affect) Control: Length of -.01 .04 -.02
illness time (months) Control: Perception of illness severity -1.92 1.03 -.27
Sleep quality .57 .21 .39*
Physical activity .01 .01 .14
Diet quality Model 3 2.10 1.17 .26 .40 .67
(DV: Illness-related distress) Control: Length of -.03 .07 -.05
illness time (months) Control: Perception of illness severity -1.46 2.16 -.09
Sleep quality 1.80 .40 .60*
Physical activity .01 .01 .07
*p<. 01
Note: Poorer sleep quality is indicated by higher scores on PSQI.
Aim 5. Evaluate the relationships between caregiver distress and caregiver quality
of life.
Hypothesis 5.1: Level of illness-related distress, level of depression, and negative
affect will be inversely related to caregiver QoL.
68


The relationship between illness-related distress, depression, and negative affect to
caregiver quality of life was assessed using bivariate correlations. Caregiver quality of life
was significantly related to depression (r = .78, p < .01), negative affect (r = .69, p < .01)
and illness-related distress (r = .65, p < .01). Note that higher scores on caregiver quality
of life measure indicated a lower quality of life, so QoL and measures of caregiver distress
are inversely related. These results are displayed in Table 18. Post hoc power analyses were
not calculated for these tests; it can be assumed that because we had significant findings
that we had adequate power to detect the relationships.
Table 18
Correlation table for caregiver distress and quality of life
Negative Illness-related Length of illness time QOL
affect distress
Depression .80** 70 ** -.14
(n) (50) (50) (48) (49)
Negative affect .65** -.09 09**
(50) (49) (50)
Illness-related -.07 .65**
distress (48) (50)
Severity .13 .01
perception (49) (50)
Length of illness .13
time (48)
**p<.01, *p< .05 (two-tailed)
Hypothesis 5.2: Caregiver distress (illness-related distress, depression, and
negative affect) will predict caregiver quality of life while controlling for perception of
illness severity and length of illness time. Within that model, measures of caregiver
distress (illness-related distress, depression, and negative affect) will be significant as
individual predictors above and beyond the control variables.
69


A linear regression model was employed including caregiver distress (illness-
related distress, depression, and negative affect) as predictors of caregiver quality of life
while controlling for perception of illness severity and length of illness time. The full model
was significant and accounted for 71% of the variance in caregiver quality of life (R2 =
0.71, F (5, 41) = 20.19, p < 0.01). Specifically within the model, higher levels of depression
significantly predicted lower levels of caregiver quality of life (fi = 0.54, p < .01). In
addition, longer time of illness was also significantly associated with decreased quality of
life (fi = 0.24,/) < 01). Negative affect and illness-related distress were not significant as
individual predictors in this model. An overview of findings is presented in Table 19. Post
hoc power analyses were not calculated for this test because there were significant findings.
Table 19
Summary of multiple regression analysis for caregiver distress predicting caregiver
quality of life (n = 46)
Variable B SEB fi R 2 Cohens./2
Model 1 Control: Time of illness .23 .08 .24* .71 2.45
(months) Control: Perception of 1.13 2.39 .04
illness severity Negative affect .64 .51 .20
Depression 1.28 .37 .54*
Impact of illness .30 .21 .17
*p< 01
Exploratory Aim 6 Examine the relationship between caregiver-patient
relationship quality and caregiver distress and caregiver quality of life in the context of
caring for a partner with MM. Determine if relationship quality is a partial mediator of
caregiver level of depression and caregiver quality of life.
70


A mediation analysis was conducted using SPSS to assess whether the effect of
depression (X) on quality of life (Y) was partially mediated by relationship quality (M).
This model involved 3 steps, all of which controlled for length of illness time and
perception of illness severity. First, a linear regression determined that there was a
significant association between the independent variable, depression, and the dependent
variable, quality of life (fi = .82, p < 0.01), while controlling for perception of illness
severity and length of illness time. A second linear regression found a significant
relationship between the mediator variable, relationship quality (assessed using the QMI
total score) and the independent variable depression (fi =-.31 ,P< o .05). A third linear
regression assessed the association of the mediator variable (relationship quality) and
quality of life while controlling for the independent variable depression. Relationship
quality was not a significant predictor of QOL when controlling for the independent
variable (depression) (fi = -.13, p = .18), indicating that the assumptions of mediation (per
Baron and Kenny) were not met. Of note, as an individual predictor, the inclusion of
relationship quality reduced the association between depression and QOL very little (fi =
78 ,p < 0.01). A summary of all of the regression analyses statistics for each step can be
located in Table 20. The mediation figure is presented in Figure 6.
Table 20
Summary of multiple regressions for mediation model (n = 45)
Variable B SEB fi R 2 Cohen sf2
Step I (DV = Quality of life) Control: Time of illness .243 .09 .25* .68 2.13
(months) Control: Perception of .05 2.34 .00
illness severity Depression 1.93 .21 .82*
Variable B SEB fi R 2 Cohens./2
71


Step 2 (DV= Relationship quality) .20 .25
Control: Time of illness -.10 (months) .04 -.37*
Control: Perception of .40 illness severity .97 .06
Depression -.19 .08 -.31*
Step 3 (D V= Quality of life) .71 2.45
Control: Time of illness .22 (months) .09 .23*
Control: Perception of -.48 illness severity 2.30 -.02
Relationship quality -.48 .36 -.13
Depression 1.82 .21
*p<.05, **p<.01
Depression
6 = .82**
Quality of
Life
6 =
Relationship
Quality
-.13
Depression
6 = .78**
Quality of
Life
* p < 0.05, **p<.01
Figure 6: Mediation Model
72


CHAPTER IV
DISCUSSION
Overview
Caregivers of multiple myeloma patients face many challenges. The primary goal
of this cross sectional study was to increase the understanding of the relationship between
caregivers immune variables (CRP, IL-6), psychological variables (depression, negative
affect, illness-related distress), patient-caregiver relationship characteristics in the context
of caregiving (relationship quality, attachment style), caregiver health behaviors (sleep
quality, diet, physical activity) and quality of life during this multifaceted and stressful
experience of caring for a loved one with multiple myeloma.
Participant Characteristics
In the following section, many comparisons are made between the current sample
and other samples. Terms such as higher and lower are used frequently. These are
not formal statistical comparisons; rather, they are informal appraisals of relative levels.
Caregiver Health Characteristics. Regarding chronic illness, 65% of participants
in this study were managing some type of chronic health condition (e.g. hypertension,
cancer, hyperthyroidism) and 26% were managing two or more chronic illnesses. In
comparison, the CDC estimated that 49% of persons (18-65) are affected by a chronic
disease or condition (Ward, Schiller, & Goodman, 2014). The higher rates in our sample
are consistent with studies that suggest caregivers are at increased risk for medical
comorbidities although the cross-sectional nature of our study makes it impossible to assess
if these comorbidities developed over the course of caregiving (Schultz & Beach, 1999;
73


Vitaliano, Zhang, & Scanlan, 2003). Interestingly, regarding physical activity behavior,
67% of our sample reported meeting the CDC requirement of 150 minutes of moderate
physical activity or exercise per week, which is a higher percentage than the estimated
49.6% of adults who met the physical activity guidelines in 2012
(http://www.cdc.gov/nchs/fastats/exercise.htm). One possible reason for this difference is
that Colorado typically has higher than the average national level of physical activity
among older adults (Center for Disease Control, 2015). Regarding other health related
variables, the majority (80%) reported being in good or excellent health. In addition,
71% reported having a good to excellent diet. Interestingly, despite higher than average
level of physical activity and positive perception of health and diet quality, 30% of our
sample met World Health Organization criteria for obesity (World Health Organization,
2014). Our sample was lower than the estimated national prevalence in adults older than
60 (35.4%) and higher compared to Colorado, which has a 20-25% obesity rate (Center for
Disease Control, 2015). Regarding sleep quality, about half of our sample met criteria on
the PSQI for sleep difficulty, which was lower compared to another study of caregivers of
stem cell transplant (SCT) patients in the acute phase of an allogeneic transplant, who
found that all of the caregivers met the criteria for sleep disruption (Simoneau et al., 2013).
Caregiving and Illness Characteristics. Consistent with the literature on family
caregivers of MM patients, caregivers in our sample reported a high amount of daily time
dedicated to caregiving for their partners with MM (Molassiotis et al., 2011). Specifically,
our sample reported a median of 3 hours dedicated to physical caregiving activities, 2
hours dedicated to practical activities, and 1 hour dedicated to emotional activities daily.
Many also reported a decrease in income following their partners multiple myeloma
74


diagnosis, which has been reported in other studies of SCT recipients (Meehan et al.,
2006; Chow & Coyle, 2011).
Regarding staging for multiple myeloma, there is a wide range of heterogeneity
found in the reported diagnosis/staging of our study sample. This may be due to the
existence of various systems that help predict prognosis that are consistently being
refined and updated as MM patients have longer rates of survival (Bataille, Annnweiler,
& Beauchet, 2013). The two most common systems used are called the Durie-Salmon
System and the International Staging System for Multiple Myeloma, which are based on
different diagnostic factors. Lack of accuracy and specificity among caregiver report of
their significant others exact diagnoses may be a reflection of the overall diagnostic
system and/or caregiver specific issues. Also multiple myeloma tends to be a chronic,
long standing illness for which patients receive different types of treatments over the
course of their life. Tracking specifics about stage and diagnosis can be difficult given
that it may fluctuate over time. In addition, the caregiver receives a large amount of
complex medical information over a long period of time and may potentially not have
the cognitive resources to remember everything. This phenomenon is likely exacerbated
by caregiver distress level.
Caregiver Psychological Variables. The prevalence of clinical levels of
depression was 33% in our sample, which is consistent with other similar caregiver
studies (Mosher et al., 2013; Langer et al., 2003; Lambert et al., 2013), lower than levels
of depression compared to end-of-life cancer caregivers (Given et al., 2004) and
caregivers of SCT recipients (Laudenslager et al., 2015; Simoneau et al., 2013), and
higher levels compared to normative data as measured in community-dwelling older
75


adults (Lewinsohn, Seeley, Roberts, & Allen, 1997). Within the multiple myeloma
literature, our sample reported higher levels of depression (31%) than the 13.6% cited in
another sample of MM caregivers (Molassiotis et al., 2011).
Regarding positive and negative affect, our sample was observed to have lower
levels of negative affect compared to caregivers of Multiple Sclerosis patients and similar
levels of negative affect compared to a group of health care professionals (Bassi et al.,
2014). Our sample was also observed to have lower negative affect levels on average
compared to a study of parents of pediatric cancer patients (Hexem, Miller, Carroll,
Faerber, & Feudtner, 2013). The complexity of emotions that accompanies a child
receiving cancer treatment is likely more distressing than having a significant other in
his/her early sixties diagnosed with and treated for cancer. Specifically, when couples
age together there is an expectation that health problems will arise later in the relationship
and later in life. On the contrary, it is rarer, and therefore arguably more distressing to
have a child experiencing a life-threatening illness as a parent. One study to date was
located that included negative affect scores (PANAS) among caregivers of stem cell
transplant recipients. Comparatively, our sample had slightly lower scores on the
negative affect subscale. However there was only a three point difference in the average
scores among participants in the two studies, which we would expect given the similarity
in the two samples. Means and brief description of comparative samples can are
summarized in Table 21.
Regarding illness-related distress, as measured by the IES-R, our study had lower
levels of illness-related distress compared to a study with caregivers of allogeneic SCT
recipients (Laudenslager et al., 2015). Also parents of pediatric oncology patients,
76


specifically those who are about to receive a SCT, had higher levels of illness-related
distress compared to our sample (Virtue et al., 2014). However, our sample was found to
have higher levels of illness-related distress compared to a sample of undergraduate
students who had witnessed at least one traumatic event (defined as the DSM-IV criteria)
in their lifetime (Adkins, Weathers, McDevitt-Murphy & Daniels, 2008). One hypothesis
for the observed difference is that our sample is likely re-experiencing illness-related
traumatic events with each challenging side effect and medical treatment, potentially
placing them at a higher risk for distress. Again means and sample demographic
information is presented in Table 21.
Regarding relationship quality, our sample reported lower relationship quality
compared to husbands of breast cancer patients (Boeding, Pukay-Martin, Porter, Kirby,
Gremore, et al., 2014) and similar levels of relationship quality compared to a sample of
underinsured primary care patients who are not caregivers (Woods & Denton, 2014).
Compared to breast cancer, MM is characterized as a chronic and uncertain illness
trajectory. The chronicity of challenges faced by MM caregivers may negatively impact
marital quality contributing to the observed difference in marital quality compared to
couples facing breast cancer. Finally, the group of primary care patients were thought to
be a normative sample. However, although this sample was not actively caregiving, it
is likely that they were facing economic challenges that may have negatively impacted
marital quality. As mentioned in the literature review, no studies were identified that
included marital satisfaction in partnered caregivers of MM patients.
Regarding attachment to significant other, scores on the MAQ suggest that our
sample had higher levels of secure attachment on average compared to another sample
77


of cancer caregivers (Kim et al., 2008) and healthy undergraduates (Carver, 1997). Table
21 displays comparative data among studies.
Finally, regarding quality of life, our sample reported higher quality of life
compared to spousal caregivers of lung, breast, and prostate cancer patients (Weitzner,
McMillan, & Jacobsen, 1999; Weitzner et al., 1999). While the samples included in these
studies are demographically similar (e.g. race, level of education, length of illness time,
gender, etc.), the biggest difference is that about 40% of the caregivers were caring for a
partner at the end of life, indicating end-of-life challenges that may negatively impact
quality of life. Average scores and comparative data can be located in Table 21. No
studies were found that used the CQOLC in a MM sample, however, within the MM
population, a study that used a different measure of QoL, the European Organization for
Research and Treatment of Cancer- Quality of Life Scale, found that caregivers
experienced moderately low quality life (Molassiotis et al., 2011). Our sample also
reported moderate quality of life, indicating that caregivers of MM patients consistently
report lower QoL compared to non-caregiving samples (Vitaliano, Zhang, & Scanlan,
2003).
Table 21.
Comparative data for psychosocial measures across various studies
Measure Sample Score Demographic data Reference
CES-D (depression) End-of-life cancer 14.87 75% 55 and up Given et al.,
caregivers (9.1) 87% Female 2004
Community-dwelling 8.33 88% Spouses 63.9 (7.9) yrs Lewinsohn et
older adults (6.84) 58% Female al., 1997
14% above 16 77% married
78


Measure Sample Score Demographic data Reference
SCT caregivers 15.2(9.6) 45% above 16 52.2(11.3) yrs 77% female 75% married Simoneau et al., 2013
MM caregivers 13.2(9.6) 31% above 16 62.9 (7.85) yrs 77% female 100% married Present study
PANAS (negative affect)
SCT caregivers 21.5 (7.8) 56.5 (11.8) yrs 62% female 78% spouses Kessler et al, 2014
MS family caregivers 21.2(7.2) 46.4 (11.7) yrs 59% female 80% married Bassi et al., 2014
Health professionals 18.3 (5.4) 40.9 (8.8) 73% female 69% married Bassi et al., 2014
MM caregivers 18.6(6.7) 62.9 (7.85) yrs 77% female 100% married Present study
IES-R (impact related distress)
Parents of children 31.3 37.4 (8.1) yrs Virtue et al.,
receiving SCT (22.5) 88% mothers 69% married 2014
Trauma exposed 16.5 19.4 (1.6) Adkins et al.,
undergraduates STC caregivers (16.8) 30.7 53% female 53.5 yrs 76% female 70% spouses 2008
MM caregivers 19.4 (12.7) 62.9 (7.85) yrs 77% female 100% married Present study
QMI (relationship quality)
Husbands of breast cancer patients 38.7(7.1) No age provided 100% male 100% married Boeding et al., 2014
Underinsured primary care clinic patients 32.0(10) 46 (12.2) yrs 81% female 100% married Woods & Denton, 2014
MM caregivers 31.2(7.8) 62.9 (7.85) yrs 77% female 100% married Present study
79


Measure Sample Score Demographic data
MAQ (secure attachment)
Cancer caregivers 3.69 (0.5) 56.5 (10.6) 50.5% female 100% married Kim et al., 2008
Undergraduates 3.57 (.49) 56% female Age and marital status not reported Carver, 1997
MM caregivers 10.5 (2.0) 62.9 (7.85) yrs 77% female 100% married Present study
CQOLC (quality of life)
Family caregivers of various cancer patients 96.2 60 yrs 60% female 80% spouses Weitzner et al., 1999
Family caregivers of cancer patients 93.2 59 yrs 67% female 70% spouses Weitzner et al., 1999
MM caregivers 52.3 (21.6) 62.9 (7.85) yrs 77% female 100% married Present study
*Note: Comparative sample for IES-R difficult to find as many of the extant caregiving
studies used the original IES.
Caregiver Neuroendocrine Variables. Levels of pro-inflammatory cytokines
were similar in this study compared to other studies involving caregivers of cancer
patients (Futterman et al., 1996, Rohleder et al., 2009) and lower than in other stressed
caregiver populations, such as parents of children with Autism or Attention Deficit
Hyperactivity Disorder (Lovell, Moss, & Wetherell, 2012). Regarding TNF- a, this is an
inducible protein that is typically elevated in individuals with chronic and acute health
conditions (Gruenewald et al., 2006; Papanicolaou et al., 1998; Pradhan et al., 2001).
Furthermore, basal levels are typically below detection rate for commercially available
ELISA kits (lOpg/ml). Surprisingly, in our sample, despite many individuals actively
managing chronic illnesses the TNF a levels were below detection. TNF a is an
80


inducible protein that is generated during an immune response. Lack of detection could
signify a problem with the preparation of the samples or in the laboratory procedures
however it was determined that the ELISA kit was valid as signified by the presence of an
accurate standard curve.
Discussion of Results Related to Study Aims
Many of the study hypotheses were supported and are consistent with the literature.
First, within Aim 2, this study found that number of hours participants engaged in physical
activity each week was significantly negatively related to levels of plasma CRP and IL-6
which are indicators of systemic inflammation. This relationship is consistent with the
literature that has noted that people who are physically active have lower CRP compared
to less active counterparts (Kasapis & Thompson, 2005; Kiecolt-Glaser et al., 2010).
Furthermore, this important finding supports the critical role that regular physical activity
may have on physical health, including lowering risk for negative health outcomes such as
cardiovascular disease, cancer, type II diabetes, functional decline and arthritis, all of
which are associated with inflammation (Kiecolt-Glaser, McGuire, Robles, & Glaser,
2002). Specifically, chronic inflammation influences tumor promotion, damages healthy
cells, and inhibits angiogenesis (Kiecolt-Glaser et al., 2010). In addition, our study also
found that CRP and BMI were positively associated, suggesting that those who are heavier
have more systemic inflammation. This association is also consistent with the literature
(Kiecolt-Glaser et al., 2010). In addition, as hypothesized, CRP was significantly predicted
by caregiver reported quality of diet, sleep quality, age, length of illness time, hours of
physical activity, and BMI together. Within that model physical activity was a significant
predictor and was inversely related to CRP.
81


With regard to Aim 4, the study results indicated that that sleep quality was
significantly inversely associated with levels of depression, negative affect, and illness-
related distress both in bivariate correlational analyses and in linear regression models that
included length of illness time and perception of disease severity as control variables.
These findings are consistent with previous literature suggesting that depression is
negatively related to sleep quality (Antoni et al., 2006; Van Moffaert, 1994). In addition,
diet quality was inversely related to depression and illness-related distress in the
correlational analyses. Finally, as hypothesized, positive health behavior indicators (sleep
quality, diet quality, and level of physical activity) significantly predicted caregiver distress
while controlling for age, length of illness time, and caregiver perception of disease
severity.
With regard to Aim 5, quality of life was inversely related to negative affect, illness-
related distress, and depression. Furthermore, a regression model including caregiver level
of distress (depression, illness-related distress, and negative affect), controlling for
perception of illness severity and length of illness time, also significantly predicted levels
of caregiver quality of life. Depression was a significant individual predictor above and
beyond all of the other measures of distress, length of illness time, and caregiver perception
of illness severity. In addition, longer length of illness time was also a significantly
associated with decreased quality of life, in the regression model. The studies examining
the nature of the relationship between length of illness and quality of life have identified
fear of recurrence, social isolation, and marital stress as some factors that contribute to
lower QoL in middle to long term survivorship (defined as 3.5 years after diagnosis) (Lewis
& Deal, 1995; Mellon, Norhouse, & Weiss, 2006). In addition, depression has been linked
82


to lower quality of life in breast cancer patients undergoing autologous stem cell transplants
(Gaston-Johansson, Lachica, Fall-Dickson, & Kennedy, 2004) and in cardiothoracic
transplant patients (Myaskovsky et al., 2012). As patients diagnosed with multiple
myeloma live longer, these findings indicate that it will be especially necessary to monitor
and/or screen caregivers for factors that may negatively impact quality of life and provide
interventions accordingly.
With regard to Aim 6, this study found that within two regression models, depression
is negatively associated with relationship quality and quality of life, while controlling for
perception of illness severity and length of illness time. However, the full mediation model,
which hypothesized that relationship quality is the mechanism by which depression is
related to quality of life, was not found to be significant in our sample. Rather, the data
suggest that depression has a direct relationship with quality of life that is not mediated by
relationship quality.
The significant relationship between variables measuring similar constructs
supported the construct validity of some of the measures. All measures of caregiver distress
(depression, negative affect, illness-related distress) were positively associated confirming
that those measures were measuring a related underlying construct as we conceptualized.
This was also the case for the inflammatory markers (CRP and IL-6) which were measuring
an element of the caregivers immune response (systemic inflammation). Also caregiver
self-reported quality of diet and sleep quality were positively associated. Again, these
variables were all conceptualized as measures of caregivers health/health behaviors.
Not all of the hypotheses of the study were supported. For example, contrary to what
we hypothesized, quality of life was not significantly associated with the level of pro-
83


inflammatory markers. One reason for this finding may be that quality of life is a diffuse
and multifaceted construct that cannot be accurately represented by a biological correlate.
In addition, we did not find a significant relationship between inflammatory cytokines and
distress (negative affect, depression, and illness-related distress) despite the fact that a link
between IL-6 levels and depression has been supported extensively in the literature,
specifically in older people living in the community (Dentino et al., 1999) and adults with
untreated Major Depressive Disorder compared to healthy controls (Rawdin et al., 2013;
Wright et al., 2005). The lack of significant findings in our sample could also be explained
by an underpowered analysis which increases the likelihood that a Type II error occurred,
meaning that a significant relationship was not detected in our sample although one may
exist.
Clinical Implications
Research suggests that there are often unmet mental health care needs among MM
caregivers (Molassiotis et al., 2011). For example, in our sample, despite the prevalence
of clinical levels of depression, only 11% reported receiving a diagnosis of major
depression, and only 9% (n = 4) reported being treated pharmacologically, suggesting a
disparity in obtaining care when caring for others. While this study did not assess whether
our sample was receiving psychotherapy, many of the caregivers in our sample had met
with a licensed clinical psychologist once during their loved ones treatment at CBCI for
an intake assessment. Furthermore the treatment of patients at CBCI includes the option
of individual and family sessions, as well as caregiver support groups run by mental
health care professionals (e.g. licensed clinical psychologists, post-doctoral fellows, and
social workers). Despite this service, caregivers have been observed to under utilize
84


supportive services (T. Simoneau, personal communication, July 6, 2015) likely due to
limited time. Caregivers and patients were also often referred to community mental health
providers if a need was determined. While the model of including psychological support
during treatment is likely helpful in addressing acute mental health concerns, it is unclear
at CBCI whether caregivers maintain continuous mental health care after their significant
other has completed treatment. Generally, a disparity in accessing mental health care for
caregivers has been found in other studies (Santin, Treanor, Mills, & Donnelly, 2014) and
has been linked to barriers such as inadequate knowledge of mental health services,
inadequate time, stigma and avoidance of mental health care, a desire to manage
emotional concerns independently and inadequate finances (Mosher, Given, & Ostroff,
2015). Despite limitations to access, there have been recently published frameworks of
linking psychosocial care to MM patients and their families, that recognize the unique
challenges faced by MM patient and their families (Kurtin, Lilleby, & Spong, 2013;
Zabora et al., 2015). These frameworks recommended regular distress screening and
early intervention (e.g. around time of diagnosis) with a focus on psycho-education,
Cognitive Behavioral Therapy, caregiver self-care, and patient-caregiver
communication.
While our cross-sectional study does not allow us to infer causal relationships, we
would hypothesize that based on the results, the use of multi-modal interventions that focus
on healthy behaviors and mood may improve psychological adjustment and potentially
increase long-term resilience in caregiving populations. Specifically, the results suggested
that sleep hygiene, sleep quality, and physical activity would be ideal priority targets in
interventions aimed at improving the health and mood of caregivers. In addition,
85


interventions aimed at decreasing level of caregiver depression may simultaneously
improve caregiver quality of life. Finally, these findings speak to the importance of
providing psychosocial screening and treatment to caregivers to mitigate the negative
impact that caregiver distress has on their quality of life even in the context of a close,
supportive relationship with a partner diagnosed with MM.
Limitations
There were a number of limitations to this study. These include the low number of
participants resulting in underpowered statistical analyses, variable disease progression
trajectories leading to heterogeneous caregiver experiences, the selectivity of the
caregiver sample, the methodological challenges associated with the preparation and
analysis of the biological markers, and the limitations of a cross-sectional design.
As mentioned throughout this document, the low number of subjects impacted the
ability to detect significant findings within our sample. Although the original goal of the
study was to recruit 65 participants, which would have allowed us to detect a medium
effect, we were unable to obtain that number due to slower than anticipated recruitment.
As a result, the study may have been unable to detect significant relationships between
some of the variables. While statistical techniques such as effect sizes and tentative
language about trends towards significance were utilized, weaker relationships among
variables may not have been detected due to lack of power. Future research could aim to
replicate these findings in a larger sample.
A second limitation was the potential for selection bias related to participants who
responded to the recruitment efforts. For example, the majority of the sample was White
(88%); African Americans and Latino/as were underrepresented in our sample, based on
86


the diversity at PSLMC and in the state of Colorado. In addition, the participants who
replied to the mail out or agreed to participate during the in person recruitment were a
self-selected sample with sociodemographic variables including high income/education
level and high level of physical activity, and thus may not be completely representative
of all caregivers served at CBCI. Thus, these findings may not generalize to all MM
caregivers of all ethnic/racial backgrounds.
A third limitation is that caregivers were caretaking for loved ones at various points
of disease trajectory and for variable amounts of time (e.g. length of caregiving). The
differences in the time of caregiving and the severity of the partners illness could
potentially impact a number of psychosocial and health outcomes. On the one hand, those
who have served as caregivers for an extended period of time may report higher levels of
stress. On the other hand, some individuals who have experienced the illness longer may
have an attenuated stress response, as they have been exposed to illness-related stressors
for longer. It is also possible that participants who are new caregivers may report higher
levels of distress given their close proximity to diagnosis and issues associated with their
adjustment to caregiver activities. In addition, patients experienced a wide range of
treatments and side effects, some more severe than others, also impacting caregiver
adjustment. While some of this variability was accounted for by including caregiver
perception of illness severity and time since diagnosis as control variables when assessing
certain relationships, it is impossible to capture all of the caregiving variables that may
have impacted stress and psychosocial adjustment of the participants.
A fourth limitation involved the collection and analysis of the biological data. For
the majority of the participants we employed standardized collection procedures (e.g.
87


standard sitting time before blood draw, standard collection time (4 hour time window in
morning), asking participants to cease nicotine, caffeine, food, and alcohol 3 hours prior
to blood draw) and using data from the pre-blood draw screen to assist us in controlling
for variables or excluding data that may cause variation in inflammation levels. Six
individuals were outside of the 9am 1pm time window, due to scheduling challenges,
however these participants were not observed to be outliers. Also, of those who opted to
return the survey via mail, the majority of caregivers returned the survey in less than a
week. This difference in timing was not observed to impact the biological data.
In addition, based on the nature of collecting data in older populations, many of our
participants had chronic illnesses and were on medications. While medications such as
antidepressants, systemic and respiratory steroids, nonsteroidal anti-inflammatories,
estrogens, immunomodulators, immunosuppressants, antirheumatic medications,
chemotherapy and other anti-cancer medications, diabetes medications, anti-hypertensives
have been shown to impact inflammation, psychoneuroimmunology researchers recognize
that an older adult who is not on one or multiple medications is rare (Kiecolt-Glaser et al.,
2010). Thus it is suggested that statistical tests be utilized to determine whether significant
differences exist between participants who are on medications or who are suffering from
chronic health conditions and variables of interest. This was conducted on our sample
and no significant differences were found with regard to inflammatory cytokine levels
and presence of chronic medical diagnoses, BMI, and medication. While the downside is
that the relationships between inflammation and psychosocial factors may be confounded
by medications and chronic illnesses, the level of chronic illness in our sample is
representative of the caregivers and does make these findings more generalizable to the
88


typical caregiver with regard to presence of health conditions. Also, as mentioned in the
methods section, the SNP assay procedure used to identify the OXTR genotype yielded
no results. There are a few methods to remedy this situation including: repurifying the
DNA as well as adding a larger amount of the current DNA sample into the SNP assay.
This could be the focus of future research, which is discussed more below. Regarding the
TNF- a, it is unclear whether our below detection results were due to methodological
issues or not using a sensitive enough assay to detect the TNF-a. One way to address this
question is to use a higher sensitivity assay, known as an AlphaLISA. However, in
general we would expect that TNF- a would have similar patterns of relationships as CRP
and IL-6 did to the psychological processes and health behavior indicators.
Finally, the cross-sectional design is problematic in that it does not allow one to
infer causal relationships between the variables. However this design was the most
feasible given resources, time restraints and reducing the burden on the participants. In
addition, there was not a control group of non-caregivers with which to compare the
descriptive data. Thus, when possible data from similar studies of caregivers which
included controls were used to yield comparisons (see Table 21). Ideally, a control group
consisting of individuals matched for sociodemographic factors such as age, gender,
relationship status and SES as well as health variables such as level of physical activity
and BMI would enable us to look at differences between psychological distress, immune
functioning, and quality of life in our samples of caregivers.
Future Research Directions
As mentioned in the introduction, there is a strong link between marital/relational
quality and health/well-being. Future research should evaluate the relationship between
89


Full Text

PAGE 1

CAREGIVER PATIENT RELATIONSHIP AND STRESS RESPONSE IN MULTIPLE MYELOMA by SHANNON LAURA MADORE B.A. University of Colorado Boulder 2008 M.A., University of Colorado Denver, 2012 A thesis submitted to the Faculty of the Graduate School of the Univ ersity of Colorado in partial fulfillment of the requirements for the degree of Doctor o f Philosophy Clinical Health Psychology 2015

PAGE 2

ii This thesis for the Doctor of Philosophy degree by Shannon Madore has been approved for the Clinical Health Psychology Program by Kristin Kilbourn, Advisor Beth Allen, Chair Teri Simoneau Dave Albeck October 6, 2015

PAGE 3

iii Madore, Shannon Laura, (Ph.D. Clinical Health Psychology ) Caregiver Patient Relationship and Stress Response in Multiple Myeloma Thesis directed by Associate Professor, Kristin Kilbourn ABSTRACT Caregiving for a loved one with a chronic illness is a multifaceted experience that can be deleterious on overall physical and mental health, but can also be meaningful and rewarding. Previous research with c aregivers has noted that caregivers of a family member diagnosed with cancer often experience a decline in physical health and psychological distress However, there is limited research on caregivers of multiple myeloma patients. The primary goal of this c ross sectional study of 51 caregivers of partners with multiple myeloma neuroendocrine and immune responses (oxytocin receptor genotype, C reactive protein I nterleukin 6, Tumor necr osis factor (depression, negative affect, illness related distress ), patient caregiver relationship characteristics (relationship quality, attachment style), caregiver health behaviors (sleep, diet, physical activity ) and qualit y of life during th e caregiving experience. Fifty one caregivers, (mean age of 62.9 ( SD = 7.85) and 77% female) were recruited from the Colorado Blood Cancer Institute (CBCI) at Presbyterian/St. Luke's Medical Center Participants were included if they w ere in a romantic relationship and living with someone who was currently receiving or had recently received treatment for multiple myeloma at CBCI. Data was collected at CBCI via a psychosocial survey and a blood draw. Unfortunately, there were concern s re garding the validity of some of the biological measures that prevented the inclusion of oxytocin receptor genotype and TNF in the overall analysis. Bivariate correlational analyses and

PAGE 4

iv multiple regression analyses revealed four key finding s ; 1) physical activity levels were inversely related to level of plasma inflammation (CRP) after controlling for caregiver health variables, age, caregiver perception of disease severity and length of illness time ; 2) sleep quality was negatively associated with careg iver distress after controlling for perception of illness severity and length of illness time ; 3) depression was inversely related to caregiver quality of life after controlling for perception of illness severity and length of illness time ; and 4) depress ion wa s negatively associ ated with relationship quality after controlling for length of illness time and perception of illness severity These results demonstrate a cross sectional association between health behaviors (sleep and physical activity ), distres s, quality of life and markers of immune response in those caring for a loved one with multiple myeloma Future studies should examine the longitudinal relationship of these variables in order to understand the causal nature of these associations. The cl inical implications of this study emphasize the importance of assessing levels of distress and health behaviors in caregivers of chronically ill partners, and providing resources and interventions to assist caregivers in maintaining their emotional and phy sical health with the potential for increasing long term resilience The form and content of this abstract are approved. I recommend its publication. Approved: Kristin M. Kilbourn

PAGE 5

v ACKNOWLEDGMENTS This study was funded by the Colorado Clinical Scienc es and Translational Institute TL1 Training Grant: TL1 TR001081 This study would have not been possible without the help of numerous faculty and staff. I would like to thank Dr. Kristin Kilbourn for her dedicated mentorship, support, and guidance required to develop, obtain funding for, implement and complete this project. I would also like to thank Dr. Teri Simoneau for her guidance in the development and implementation of this project, specifically for her dedication to the recruitment of subjects. I wo uld like to thank my committee chair, Dr. Elizabeth Allen, for her guidance and support with statistical analysis the selection and interpretation of measures and her thoughtful comments and suggestions to help improve the written document I would like to thank Dr. Dave Albeck for his assistance and support with the biological data In addition, this project would not have been possible without the support of the CBCI staff, specifically the psychosocial team who assisted with recruitment, the phlebotomi sts at CBCI for drawing the blood, and Tammy Robles, who shared her lab space and equipment with me. I would also like to thank Dr. Mary Coussons Read, who was an instrumental part of the initial development of the project and who guided me in learning the ELISA technique. Dr. Celia Sladek, Professor of Neuroscience and director of the CCTSI training program, who taught me the Radioimmunoassay Technique, allowed me to use her laboratory, and encouraged me through the process of obtaining funding for my p roject. I would like to thank Dr. Christopher Phiel, Professor of Integrative Biology, and his lab, who allowed me to utilize their laboratory space to run the ELISAs, SNP Assays, and isolate the DNA. In addition, I would like to thank Josh Fowler, who ass isted in teaching me DNA isolation techniques and SNP assay techniques

PAGE 6

vi and helped me trouble shoot through the OTXR genotyping challenges. I would like to thank Jean Quispe, who assisted me with recruitment and data collection while I was on maternity l eave. I would like to thank Elizabeth Berkholm, who assisted me with date entry. I would also like to thank Kimberly Hill, for her administrative assistance in ordering research supplies. I would like to thank the other UCD faculty who supported my learnin g and growth as a clinical researcher, without which this project would have not been possible. Finally, I would like to thank my supportive family for their dedicated assistance through this process!

PAGE 7

vii TABLE OF CONTENTS CHAPTER I. BACKGRO UND AND SIGNIFICANCE ................................ ................................ ..... 1 Caregivers of Multiple Myeloma Patients ................................ .............................. 1 Stem Cell .... .. 4 Romantic Relation ships, Health, and Caregiving ................................ ................... 5 Oxytocin ................................ ................................ ................................ .................. 9 Inflammation, Health, Caregiving, and Romantic Relation .. .. 12 Conceptual Model. Specific Aims and Hypothe ses. .. Specific Aims and Hypotheses in II. METHOD ................................ ................................ ................................ .................... 22 Study Sett ing ................................ ................................ ................................ ......... 22 Participants ................................ ................................ ................................ ............ 23 Recruitment an d Enrollment Procedures ................................ .............................. 24 Study Design ................................ ................................ ................................ ......... 27 Data Collection Procedures ................................ ................................ ................... 27 .. III. RESULTS ................................ ................................ ................................ .................. 41 Recr uitment Accrual and Attrition ................................ ................................ ........ 41 Sociodemographic Characterist ics of the Part icipant Sample .............................. 41 Pa tient Illness Characteristics ................................ ................................ ............... 43 Caregiving Variab les for the Participant Sample ................................ .................. 46 Health Related Variabl es for the Participant Sample ................................ ........... 48

PAGE 8

viii Additional Health Related Variables for the Participants who Completed Bl ood Psychosocial Characteris 51 Car egiver Neuroendocrine 53 Results of Inferential Statistical Analys es for each of the Study Aims ................ 55 IV. DISCUS SION ................................ ................................ ................................ ............. 73 Overview ................................ ................................ ................................ ............... 73 Limit ations ................................ ................................ ................................ ............ 88 Future Research Direction s ................................ ................................ ................... 90 CCTSI EDUCATIONAL AND TRAINING EXPERIENCES ................................ ........ 92 REFERENCES ................................ ................................ ................................ ................. 99 APPENDICES ................................ ................................ ................................ ................ 120 A: C onsent Forms ................................ ................................ ............................... 120 B: Initial and Follow Up R e cruitment Letters Pati en t ................................ ...... 128 C: Initial and Follow Up Rec ruitment Letters Caregiver ................................ 130 D: Letter to Psychosocial Team to Request Recruitment Assisstance .132 E: Caregiver Informational Flyer ................................ ................................ ....... 133 F: Pre Blood Draw Screen ................................ ................................ ................. 134 G: Caregiver Psychosocial Survey ................................ ................................ ..... 137

PAGE 9

ix LIST OF TABLES TABLE 1. Key V ariables and Sources of Data. ................................ ................................ ............. 35 2. Tests of Normality and Internal Consistency for 36 3. Part icipant Demographic Variables ................................ ................................ .............. 42 4. Caregiver Reported Patient Illness Characterist ics ................................ ...................... 44 5. Caregiver Level of Involvement and Health Behaviors ................................ ............... 48 50 7. Caregiver Scores on Psychosocial Measures ................................ ................................ 52 8. Group Differences for Pro Inflammatory Markers ................................ ....................... 53 9. Correla tion Table for Caregiver Distress and Inflammatory Mark 10. Summary of Multiple Regression Analysis for Caregiver Distress Variables Predict 11. Summary of Multiple Regression Analysis for Caregiver Distress Variables Predicti ng IL 12. Correlation Table of Caregiver Health Variables and Inf lammatory 13. Summary of Multiple Regression Analysis for Caregiver Health Behavior Variables Predi cting Inflammatory 14. Correlation Table for Caregiver Pro Inflammatory Markers and Quality of Life 15. Summary of Multiple Regression Analysis for Pro I nflammatory Markers Predict 64 16. Correlation Table of Caregiver Psychological Processes and Health Variables 17. Summary of Multiple Regression Analysis for Health Behaviors Predicting Caregiver Distress

PAGE 10

x 18. Correlation Table for Caregiver Distress an d Q uality 19. Summary of Multiple Regression A nalysis for Caregiver Distress Predicting Caregiver Quality of L ife 20. Summary of Multiple R egressions for M edi ation M odel 21. Comparative Data for Ps ychosocial M easure

PAGE 11

xi LIST OF FIGURES FIGURE 1. Overall Conceptua l Model to Guide Hypotheses and 2. Conceptual Model Reflecting Only the Varia bles Incl uded in the Final Analyses 3. Types of Treatment Received by Patient 4. Caregiver Perc 5. Caregiver Reported Health and Diet Quality 6. Mediati on Model

PAGE 12

1 C HAPTER I BACKGROUND AND SIGNIFICANCE Caregivers of Multiple Myeloma Patients Multiple myeloma (MM) is an incurable hematologic cancer. It is currently the second most common blood cancer and in 2015 the re was an estimated 26, 8 5 0 new cases diagnosed ( The American Cancer Society, 201 5) Treatments for MM are aimed at mitigating symptoms, improving quality of life and increasing longevity. Recent improvements in treatment have led to increases in long term survivorship with MM patients surviving, on average, up to 11 years after initial diagnosis from about 4.5 year s in 2003 (Child et al., 2003). Fo r patients diagnosed with multiple myeloma who are under the age of 65, the current standard of treatment includes an a cell transplantation 1 (SCT) with high dose melphalan (Child et al 2003) In autologous transplantation, one and then f rozen for storage and later use. Next, high dose chemotherapy and/or radiation is administered to an individual to eradicate the cancerous blood cells, but with side effects of c ompromising the immune system. After chemotherapy or radiation is complete, the harvested cells are thawed and returned to the patient. This process can be done on an inpatient or outpatient basis with follo w up by the oncologist who is overseeing the 1 Blood and bone marrow transplant an d stem cell transplant will be used interchangeably in the subsequent sections

PAGE 13

2 proc edure Whether a patient will undergo transplant depends on a number of factors including disease progression and stage, response to chemotherapies, other medical comorbidities, age, psychosocial factors and financial considerations For those patients w ho do not receive a S C T a combination of Bortezomib Thalidomide or Lenalidomide, chemotherapy, radiation therapy, and corticosteroids is standard ( Palumbo et al., 2014 ). Unfortunately, the treatment and survivorship trajectory for MM is very unpredictab le; it is often characterized by illness relapse followed by additional treatment, which sometimes entails second or even third autologous transplants and in fewer cases allogeneic (donor cell) transplant (Greipp et al., 2005) Approximately 65% of the treatments for MM occur on an outpatient basis, resulting in patients and their caregivers spending a considerable amount of time managing the side effect s of treatment at home These side effects can include bone pain, neurop athy, fatigue, loss of appetite, seve re diarrhea and vomiting (Molassiotis, Wilson, Blair, Howe & Cavet, 2011). This is not only difficult for the patient, but can be extremely stressful for the caregiver who is trying to support and care for their loved o ne while managing their own self care routines. However, little is known about the impact that the multitude of treatment related stressors coupled with longer, yet unpredictable survivorship trajectory has on the caregiver. Therefore, partnered caregiver s of MM patients are a unique and important population to study. Compared to other cancer populations, there are relatively few studies examining A study of 93 partners of MM patients found that on e third of partners reported unmet supportive care needs ( e.g. existential issues, sexuality issues, chronic symptom management, self and family care), almost half (48.8%) reported signs of anxiety, and 13.6% reported signs of depression

PAGE 14

3 ( Molassiotis et al ., 2011) A qualitative study, conducted by the same group of researchers, found informal caregivers (88% spouses) were providing practical and emotional support to patients almost exclusively, often by neglecting their own needs ( Mola ssiotis et al., 2011 ). No gender differences were noted in this study. Furthermore, informal caring often with caregiver related stressors. Both patients and caregivers reported sig nificant fears and uncertainty about the future; for example, one caregiver described MM as a 'time bomb' ( Molassiotis et al., 2011). In a disease and I would take care Coon, McBride Wilson, & Coleman, 2007 ). Another study supported the notion that MM caregivers a re a unique group compared to many other cancer caregivers because of th e fatigue associated with the accumulation of years of medical caregiving (Zabora et al. 2015). These studies consistently illustrate the psychosocial toll of caregiving for a loved one with MM although few studies have specifically examin ed the specific emotional social and practical challenges faced by MM patients and their loved ones (Dahan & Auerbach, 2006) Despite the limited research assessing and describing psychosocial re action in caregivers of MM patients, overall, studies have found that caregivers of cancer patients experience higher psychological distress than the patients to whom they provide care (Kim, Carver, Rocha Lima, & Schaffer, 2011; Soothill, Morris, Thomas, H arman, Francis, McIllmurray, 2003 ). Caregivers of cancer patients also report lower quality of life and are at increased risk for medical comorbidities compared to non caregivers (Vitaliano, Zhang & Scanlan, 2003). Additionally, high rates of psychosocial distress and physical exhaustion may lead

PAGE 15

4 to caregiver burnout resulting in decreased caregiver involvement i n patient needs (Marks, Lambert, & Choi, 2002; Pinquart & Sorensen, 2003). Stem Cell Transplantation and Caregiving The transplantation process i s laden with frequent fluctuations in medical status, repeated invasive medical procedures, possibility of death, prolonged hospitalization ( Foxall & Gaston Johansson 1996), as well as uncertainty, disappointment, dislocation from home and friends, worrie s about children, and loss of job (Lesko, 1994). Often before the transplantation process begins, patients need to make plans regarding their family, home, finances, pets, and employment for the duration of their transplant procedure, which can last weeks to months During transplant, caregivers experience a number of practical problems such as significant time constraints and financial burden (Meehan, Fitzmaurice, Root, Kimit s, Patchett, & Hill, 2006). In addition, caregivers experience a disruption of dai ly roles ( Chow & Coyle, 2011 ), decreased martial satisfaction, and increased levels of anxiety and depression both six months and one year following stem cell transplantation ( Langer, Abrams, & Syrjala 2003 ). The physiological and psychosocial impact of S CT on informal caregivers has received little attention although the literature suggests that caregivers may experience immunological and psychological changes ( Futterman, Wellisch, Zighelboim, LunaRaines, & Weiner, 1996 ) and heightened caregiver burden ( Foxall & Gaston Johansson, 1996). SCT is a profound experience shared by the entire family of the patient. The uncertain and multifaceted trajectory of the SCT procedure is distressi ng and burdensome on the caregiver from a biopsychosocial perspective and it can have a profound impact on caregiver quality of life.

PAGE 16

5 Romantic Relationships Health and Caregiving Romantic r elationships and h ealth Overall, being in a long term romantic relationship has been linked to better health and well being. Relationship functioning, which encompasses overall satisfaction, communication behaviors, companionship, decision making, and cognitions (attributions and expectations) is bidirectionally related to health habits including sleep (Troxel, 2010) diet, and exercise. Car diovascular (Berkman, 1995) endocrine, and immune responses have also been linked to relationship functioning (see Kiecolt Glaser & Newton, 2001 and Robles & Kiecolt Glaser, 2003 for extensive reviews). For example, women who reported moderate to severe marital strain were three times more likely to experience a coronary event after controlling for demographic and disease status variables (Orth Gomer, Wamala, Horsten, Schenck Gustafsson, & Schneiderman, 2000). Furthermore, the quality of and attachment style within a romantic relationship as well as in the context of caregiving, has been linked to differences in the psychological reaction of and care provided by the significant other ( Kim & Carver, 2007; Kim, Deci, Carver, & Kasser, 2008; Morse, Shaffer Williamson, Dooley, & Schulz, 2012 ). The sections below will review the research on attachment relationship quality and the experience of caregiving. Attachment Attachment theory provi des a framework from which to conceptualize relationship dynamics, which are shaped from early experiences and impact our view of ourselves and others in the context of close relationships. From infancy, humans have an attachment system that is activated by stress or threat, in order to maintain a sense of security (Bowlby, 1973 ). Since Bowlby various working models have been developed to conceptualize adult

PAGE 17

6 human attachment. In 1987, Hazan and Shaver described three types of attachment styles: 1) secure which is defined as a sense of closeness and easy reliance on others; 2) anxious ambivalent responses; 3) avoidant which is defined as independence, distance from others and discomf ort with closeness. In 1991, Bartholomew and Horowitz created a four group model to describe attachment styles, in which patterns of attachment are defined using image (positive or negative) and image of others (positive or negat ive). This model describes four attachment styles: 1) secure positive thoughts about self and others; 2) dismissive avoidant positive thoughts about self, negative thoughts about others; 3) anxious preoccupied negative feelings about self, positive feel ing about others; and 4) fearful avoidant negative thoughts about self and others. In adults, attachment security (comfort with closeness and interdependence) is the foundation for effective caregiving relationship because it allows individuals to be more attentive to A significant amount of research has been conducted evaluating how adult attachment style impacts the experience of and reaction to the demands of being a caregiver for a spouse with cancer (Kim & Carver, 2007; Kim et al., 20 08), or a spouse wit h cognitive disabilities (Morse et al., 2012). In 1997, Carver created the Measure of Attachment Qualities (MAQ) theoretically (1987). The MAQ uses separate sub scales to assess secure a ttachment tendencies, avoidant tendencies, and two sub scales to assess anxious ambivalent pattern s These anxious ambivalent sub scales include the Ambivalence M erger, which is reflective of an unmet desire for more closeness, and Ambivalence W orry whi ch is reflective of concern about s being

PAGE 18

7 insincere o r not being as strong as Development of the MAQ is described in more detail in method section. Using this measure to assess the role of at tachment and frequency with which various care tasks were provided ( e.g. emotional, instrumental, tangible) Kim and Car ver (2007) studied a group of spousal caregivers of cancer patients They found that securely attached wives provided more frequent emo tional care ( e.g. ( e.g. providing financial help) for spouses with cancer Securely attached spouses reported less difficulty in providing care and reported less burden This difference may be explained by more securely attached spouses feeling more able to meet the emotional needs of an ill spouse due to their perceived interdependence and positive view of self and others. In another study focus ing on adult attachment, psychological well being, and motivation for caregiving for a spouse with cancer attachment security was positively associated with autonomous motives ( e.g. finding benefit in caregiving (Kim et al., 2008). Again these differences in perception of the caregiving experience reflect the general caregiver self esteem and differences in views of relationships with others as either negative or positive. In a non c ancer population, secure ly attached spouses offered more comfort and reassurance while avoidantly attached spouses displayed more anger or blaming i n response to situational stresses (Feeney, & Collins, 2001). Fearful spouses, defined as those having negat ive thoughts about self and others, were less likely to report using problem focused coping and more likely to state that the caregiving relationship le d to increased

PAGE 19

8 marital conflict. These findings suggest that attachment style not only impacts patient c aregiver interactions, but also influences how the caregiver copes with the stress of caring for a loved one with a health condition. Decreased use of adaptive coping may represent developing mental and phys ical health issues. Along those lines, among children with elde rly, frail or demented parents, an insecure attachment style has been linked to increased psychological distress ( Crispi, Schiaffino, & Berman, 1997 ) and is a predictor of depression ( Besser & Priel, 2005; Carnelley, Pietromonaco, & Jaffe, 1994 ) S ecure attachment has been linked to lower levels of caregiver stress (Cripsi et al., 1997) Again this elucidates the important relationship between attachment style and me n tal health while caring for a loved one with a chronic illness Overall, these findings suggest that caregiver s who report a secure attachment to their partner may experience less distress and better overall health than caregivers who do not report a secure attachment Th e present st udy will elucidate the role that secure attachment plays with regard to neuroendocrine response in caregivers of MM patients Relationship q uality. Relationship quality has also been found to be an important predictor of psychological and physiological re sponses to stressful life events The definition of relationship or marital quality varies from study to study depending on the measures used, but generally encompasses satisfaction, agreement on values and life goals, overall happiness, adjustment, relian ce on one another, and exchange of ideas, or a combination thereof. One study examining neural response to the threat of an electric shock found that higher levels of self reported marital quality predicted attenuation of activation in the neural systems s upporting emotional and behavioral threat responses when women Coan, Schaefer, & Davidson, 2006 ). Higher marital quality has

PAGE 20

9 also been associated with faster recovery from illness and a lower rate of mortality following the diagno sis of a life threatening illness (Robles & Kiecolt Glaser, 2003). Lastly, marital quality also predicted hospital stay following coronary artery bypass surgery for women (Kulik & Mahler, 2006). However, there have been no studies to date that have assess e d relationship quality in the con text of partners facing MM. Therefore, r elationship quality is an important variable to include in our study because there is limited research on how it is related to mood, qu ality of life, and health in the MM caregiver po pulation Oxytocin The neuropeptide oxytocin and its receptor are expressed widely in the central and peripheral nervous system, suggesting a number of endocrine and paracrine roles ( MacDonald & MacDonald, 2010; Tom & Assinder, 2010 ). Oxytocin has been sho wn to modulate social behavior such that intranasally administered oxytocin has been lined to increased eye conta ct, improved social memory, and less social anxiety (see MacDonald & Mac Donald, 2010 for a review). Another important facet of the oxytocin lit erature that is relevant to this study is that it has b een found to events ( Amico, Cai, & Vollmer, 2008; Ditzen Schaer, Gabriel, Bodenmann, Ehlert, & Heinrichs 2009; Heinrichs, Baumgartner, Kirschbaum, & Ehlert, 2003; Light, Grewen, & Amico, 2005; Neumann, 2002 ), and help to establish or maint ain social attachments (Bucheim et al., 2009). Oxytocin levels have also been negatively associated with depression and anxiety ( Scantamburlo et al., 2007 ), and positively a ssociated with displays of romantic love ( Gonzaga, Turner, Keltner, Campos, & Altemus, 2006 ), affectionate behaviors between partners (Gutkowska, Jankowski, Lambert, Mukaddam Daher, Zingg,

PAGE 21

10 & McCann, 1997), and trust ( Kosfeld, Heinrichs, Zak, Fischbacher, & Fehr, 2005 ). Oxytocin has been linked to a variety of health outcomes such as cancer ( Reversi, Cassoni, & Chini, 2005 ) and cardiovascular functioning (Gimpl & Fanreholtz, 2001; Gutkowska et al., 1997; Petersson, 2002) In addition, increased levels of oxytocin are related to lower blood pressure and heart rate (Light et al., 2005). Recent developments in the field of molecular genetics hav e also linked single nucleotide polymorphisms in the oxytocin receptor (OXTR rs 53576) to differences in attachment style (Chen, Barth, Johnson, Gotlib, & Johnson, 201 1 ), emotion regulation (Kim, Sherman, Mojaverian, Sasaki, Park, et al., 2011), and parent ing behavior (Bakermans Kranenburg, & van Ijzendoorn, 2008). In terms of behavioral phenotype, individuals with the GG phenotype of OXTR rs53576 exhibit more prosocial temperament (Tost et al., 2010), more sensitive parenting behavior (Bakermans Kranenburg & van IJzendoorn, 2008), and greater sensitivity to infant crying (Riem, Pieper, Out, Bakermans Kranenburg, & van IJzendoorn, in press). GG homozygotes also report being less lonely (Lucht et al., 2009) and possess greater empathic accuracy (Rodrigues, S aslow, Garcia, John, & Keltner, 2009) relative to those with the AA genotype. Those with the AG genotype fall between the two homozygous genotypes. Although there is literature linking prosocial be havior and parental caregiving behaviors to oxytocin genoty pe receptor, there is n o research to date examining g e notypical differences in the context of caregiving for a significant other. Polymorphisms within the OXTR impact the production, transport, and metabolism of oxytocin, thus creating a systemic impact i nfluencing behavior, mood and health Thus

PAGE 22

11 it contact, as well as stress and health. Attachment and physical contact Results from one study found that more frequ ent partner hugs and higher oxytocin levels are linked to lower blood pressure and heart rate in premenopausal women (Light et al., 2005). These findings suggest that supportive marital interactions may promote greater oxytocin production therefore modulat ing stress and immune responses. Health and stress. Increases in oxytocin levels have been associated with lower blood pressure ( Grewen, Girdler, Amico, & Light, 2005 ). An interesting study investigating the relationships between marital behavior, oxyto cin, and wound healing, found that higher levels of oxytocin were associated with more positive communication an d quicker wound healing (Gouin et al., 2010). Also, circulating level of plasma oxytocin was found to be associated with HIV viral status in low income minority women such that higher oxytocin was associated w ith lower viral titers (Fekete et al., 2011). Compared to no treatment controls, individuals who received intranasal oxytocin showed a reduced cortisol response in response to a public speaki ng task ( Quirin, Kuhl, & Duesing, 2011 ). After a social stressor, men who were administered intranasal oxytocin, and accompanied by a friend, showed decreased cortisol levels, higher ratings of calmness, and lower anxiety compared to controls (Heinrichs et al., 2003). In addition, a number of studies have suggested that oxytocin may mediate or moderate the relationship between social support and health outcomes (Carter, 1998; Robles & Kiecolt Glaser, 2003). Finally, intranasally administered oxytocin has be en associated with decreased levels of pro inflammatory cytokines including TNF 6, and VEG F in response to a bacterial endotoxin. Despite

PAGE 23

12 its link to stress modulation and attachment, there are no studies to date examining oxytocin genotype in couples experiencing chronic stress. Including oxytocin as variable of interest in caregiver studies can elucidate the relationship between attachment style, relationship quality and health in individuals who are caring for a significant other with a serious illness. Limitations to oxytocin research There are so me limitations to conducting oxytocin genotyping in research. Research has shown that our environment can interact in complex ways changing our genetic expression (Kumsta, Hummel, Chen, & Heinrichs, 2013). Despite this, there is still clear evidence that d ifferences in g enotypes are related to changes in behavior, affect, and interactions with others. Inflammation, Health, Caregiving, and Romantic Relationships Inflammation is a vital part of the human immune response to acute infection. Pro inflammatory c ytokines signal immu ne cells to an injury site to help clear viral and bacterial pathogens (Kiecolt Glaser, Gouin, & Hantsoo, 2010). However, chronic inflammation is problematic. Elevated levels of pro inflammatory cytokines (IL 6,TNF Reactive Protein (CRP) have been linked to a number of medical conditions including cardiovascular disease, Type II Diabetes, certain cancers (Gruenewald, Seeman, Ryff, Karlamangala, & Singler, 2006; Papanicolaou, Wilder, Manolagas & Chrousos, 1998; Pradhan, Manson, Rifai, Bur ing, & Ridker, 2001), arthritis, reactive protein (CRP) is a protein in the plasma that can signal systemic inflammation. Although CRP is not routinely measured in clinical settings, it is commonly used as a marker of inflammation in studies examining risk for various illnesses. IL 6, CRP and TNF also been associated with negative mood (Wright, Strike, Brydon, & Steptoe, 2005 ) and

PAGE 24

13 recent research has indicated that this relations hip is bidirectional ( Dantzer, O'Connor, Freund, Johnson, & Kelley, 2008 ) Sleep disruption, often seen with caregivers, can increase the expression of pro inflammatory cytokines (IL 6 and TNF promote tumorigene sis by inhibiting DNA repair th rough the generation of reactive oxygen species, inducing DNA damage ( Antoni et al., 2006 ). Finally, inflammation levels are impacted by other health behavior s such as exercise, diet, chronic health problems, medications, adiposity, alcoh ol, and smoking (K iecolt Glaser et al., 2010). Psychological stress can also evoke pro inflammatory cytokine production in the absence of infection or injury (Kiecolt Glaser, Gouin, & Hantsoo, 2010). Studies of caregivers have found higher levels of inflammatory markers c ompared to non caregivers ( Kiecolt Glaser Preacher, MacCallum, Atkinson, Malarkey, & Glaser, 2003; von Kanel et al., 2006). One study in particular found that the average level of IL 6 was four times higher in caregivers than in non caregivers (Kiecolt Gl aser et al., 2003). It has been suggested that chronic inflammation may represent a biological mechanism by which the stress and burden of caregiving leads to declining health status and increased risk for morbidity and mortality (Maier, 2003) especially i n older caregivers by way of frailty syndrome (von Kanel, Kud ielka, Preckel, Hanebuth, & Fischer, 2006) Romantic relationships and i nflammation Attachment, social support, and relationship quality have been associated with immune responses. Specifically, IL 6 levels during an acute marital argument show an 11% increase in individuals with high attachment avoidance and a 6% decrease in individuals with l ow attachment avoidance (Gouin et al., 2009). Furthermore, more frequent hostile marital interaction was associated with a 40% slower rate of blister wound healin g and lower wound site production of IL 6,

PAGE 25

14 TNF hostile couples (Kiecolt Glaser et al., 2005), indicating a slowed local immune response. In a different sample of couples engaging in conflict discussion, those who used more cogniti ve words ( e.g. words of reasoning, insight, and thinking) had smaller increases and lower levels of IL 6 and TNF discussion compared to those who used fewer cognitive words (Graham et al., 2009). Among older adults, married men had lo wer CRP levels than divorced men, married women and unmarried women which may help to explain the finding that men appear to experience greater health benefit from marriage (Sbarra, 2009). The role of inflammatory cytokines in the decline of caregiver heal th is a complex process which may be impacted by various psychological ( e.g. attachment, depression), demographic ( e.g. age, gender), and behavioral variables. Thus, assessing caregiver inflammation level s to help elucidate the relationship bet ween inflamm ation, mood, health and quality of life among partnered caregivers of MM patients will add a psychoneuroimmunological facet to the MM caregiver literature Conceptual Model Based on the reviewed literature, an initial conceptual model, presented in Figure 1, w as developed to help guide the overall specific aims and hypothesis described in the following section. As illustrated in the model, d uring the process of caregiving, an individual may experience psychological, behavioral and biological reactions that are inter related. Specifically, each individual has a unique set of sociodemographic caregiving specific and health specific factors that may influence a ps ychological and biological response while caring for a loved one Furthermore psycho logical processes such as depression and relationship factors such as marital

PAGE 26

15 quality have been linked to immunologic and endocrine functioning. Finally, while the relationship betw een psychological processes and quality of life is well established in the literature, little is known about the relationship between neuroendocrine and immune response with quality of life and likely a bidirectional relationship exists It should be noted that while not all of the relationships outlined in the model were direct ly assessed, this model helped guide the development of the hypothesis and the planned analysis. Figure I. Overall conceptual model to guide hypotheses and analyses. Note: TNF oxytocin genotype were no t included in final analyses. Specific Aims and Hypotheses The purpose of this study was to better understand the psychosocial and MM Specifica lly,

PAGE 27

16 this project aimed to increase our understanding of the relationship between caregiver psyc hosocial variables (depression, illness related distress affect), patient caregiver relationship characteristics (relationship quality, attachment style), neur oendocrine and immune responses (oxytocin receptor genotype and pro inflammatory markers), and quality of life, in significant others who are caring for MM patients. While the literature suggests associations among the variables of interest mainly among c aregivers of loved ones with other types of cancer no study to date has examined these variables in a sample of individuals caring for a significant other with MM. The experience of caring for a loved one with MM is somewhat unique given the chronic natur e of the illness and the uncertain trajectory C aregiver s face both chronic (e.g., living with MM for many years) and acute stressors (e.g., chemotherapy treatments, bone marrow transplants) that likely impact psychosocial distress, immune and endocrine f unction, relationship quality, and quality of life. The following aims guided th is cross sectional study to better understand how these factors are associated with one another within this population : Aim 1. E valuate the relationship s between caregiver oxy tocin genotype, psychological processes and caregiver relationship characteristics in the context of caring for a loved one with MM. Hypothesis 1.1: Caregivers with GG oxytocin genotype will report a higher number of positive caregiver patient relationshi p characteristics (secure attachment style and high relationship quality). Hypothesis 1.2: Caregivers with GG oxytocin genotype will report lower levels of distress (depression, negative affect, illness related distress ).

PAGE 28

17 Aim 2. E valuate the relationship s between caregiver pro in flammatory markers, psychological processes and health variables Hypothesis 2.1: Caregivers who report higher levels of distress ( depression, negative affect, illness related distress ) will have higher levels of plas ma pro inflam matory markers Hypothesis 2.2: Caregiver distress (depression, negative affect, and illness related distress ) will significantly predict the level of caregiver pro inflammatory markers while controlling for caregiver age length of illness time and care giver perception of illness severity Specifically, within those models, depression, negative affect, and illness related distress will be significant individual predictors of the level of pro inflammatory markers above and beyond the control variables Hy pothesis 2.3 : Caregivers reporting better health behavior indicators, as measured by sleep quality diet quality and physical activity levels, will have lower levels of plasma pro inflammatory markers. Hypothesis 2.4: Caregiver self reported health behav ior indicators (sleep quality, diet quality, and level of physical activity) will together predict level of caregiver pro inflammatory markers (IL 6, TNF time, and BMI. Within that model, it is hypothesized that sleep quality, diet quality and level of physical activity will each significantly predict level of caregiver inflammatory markers above and beyond the control variables. Aim 3. E valuate the association between caregiver biomarkers and caregiver quality of life.

PAGE 29

18 Hypothesis 3.1: Levels of plasma pro inflammatory markers w ill be inversely related to quality of life Hypothesis 3.2: Level s of plasma pro inflammat ory markers (IL 6, CRP, TNF will predict caregiver quality of life while controlling for age, length of illness time and caregiver perception of illness severity Within those models it is hypothesized that pro inflammatory markers will be significan t individual predictors of caregiver Q o L above a nd beyond the control variables. Hypothesis 3.3 : Quality of life will be significantly associated in the positive direction with OXTR GG phenotype. Aim 4. E valuate the relationship between caregiver health b ehavior indicators as measured by sleep quality physical activity and self reported quality of diet and psychological processes (depression, affect, illness related distress ). Hypotheses 4.1: Quality of sleep quality of diet and level of physical acti vity will be inversely related to level of depression, negative affect and illness related distress Hypothesis 4.2: Sleep quality diet quality, and level of physical activity together will predict c aregiver distress (depression, affect, illness related distress ) after controlling for length of illness time and perception of illness severity. Within that model it is hypothesized that sleep quality diet quality and level of physical activity will be significant as individual predictors of distress ( depres sion, affect, illness related distress ) above and beyond the control variables Aim 5. E valuate the relationships between caregiver distress and caregiver quality of life.

PAGE 30

19 Hypothesis 5.1: Level of illness related distress level of depression, and negativ e affect will be inversely related to caregiver Qo L Hypothesis 5.2: Caregiver distress ( illness related distress, depression, and negative affect) will predict c aregiver quality of life while controlling for perception of illness severity and length of i llness time Within that model, measures of caregiver distress ( illness related distress depression, and negative affect) will be significant as individual predictors above and beyond the control variables Exploratory Aim 6. Examine the relationship bet ween caregiver patient relationship quality caregiver distress and caregiver quality of life in the context of caring for a partner with MM. Determine if relationship quality is a p artial mediator of the relationship between caregiver level of depression and caregiver quality of life (see conceptual model and explanation below). Specific Aims and Hypotheses in the Context of Conceptual Model T he updated model, displayed in Figure 2 below, reflects only the variables that were included in the final anal yses. T he top boxes that represent sociodemographic variables, health variab l es (e.g. level of physical activity diet quality and sleep quality ) and caregiving variables (e.g. length of illness time and perception of illness severity) are suggested by th e literature to be related to caregiver psychological processes, i mmune response, and caregiver Qo L, and are thus included as control variables in many of the aims and analyses. Specifically, Aim 1 w as designed to evaluate t he association between OXTR gen otype (representing neuroendocrine regulation) and caregiver psychological processes distress (depression, affect, and illness related distress ) and relationship characteristics (secure attachment and relationship quality). Unfortunately, d ue to

PAGE 31

20 difficult ies with the storage and purification of DNA, the OXTR genotype data was invalid and was not included i n the study and therefore is not included in the updated model presented in Figure 2. Aim 2 will examine the relationship between inflammat ory markers an d health variables as well as the relationship between inflammatory markers and psychological variables. We expect that positive health behaviors (e.g. healthful diet and regular physical activity) will be significantly inversely related to inflammation an d that distress will be significantly positively related to inflammation. Aim 3 will focus on understanding the strength of the relationship between immune response (as measured b y pro inflammatory markers IL 6 and CRP) and caregiver Qo L. Aim 4 will examin e psychological processes, specifically caregiver distress ( illness related distress depr ession, and negative affect) that are impacted by sociodemographic variables and health behaviors Aim 5 will evaluate the degree to which caregiver psychological pro cesses and distress directly impact quality of life. While there is overlap in the constructs of distress and quality of life, quality of life is a multifaceted construct that encompasses more global experiences such as perceived burden, life interruption communication, positive benefit finding, spiritu ality and psychological adjustment. Finally, e xploratory Aim 6 will elucid ate the underlying mechanism by which depression is associated with quality of life While it is clear that many types of mediation and moderation amo ng these variables are possible, t he pathway of relationship quality influencing depression which in turn impacts quality of life, is just one of many possible pathways. This particular one is explored in order to highlight the possible importance of relationship functioning even when one is in a caregiver role

PAGE 32

21 Therefore, the relationship quality betwee n the caregiver and the MM patient is an important factor to consider w ith regard to caregiver quality of life Figure 2. Conceptu al model reflecting only the variables that were included in final analyses. Arrows reflect which relationships were directly assessed in the each of the study aims. Note: Due to difficulties with the storage and purification of DNA, the OXTR genotype dat a was invalid and was not included in the study and therefore is not included in the updated model presented in Figure 2

PAGE 33

22 CHAPTER II METHOD Study S etting The Colorado Blood Cancer Institute (CBCI) is a large regional transplant center for Colorado and s urrounding states. Transplanting over 240 patients per year, the program performs allogeneic and autologous transplants for adult patients. The transplant program is part of the Colorado Blood Cancer Institute (CBCI) at Presbyterian/St. Luke's Medical Center (PSLMC). PSLMC is a large medical center, licensed for 680 beds, serving Denver, Rocky Mountain, and Great Plains regions. There are three inpatient units at PSLMC dedicated to stem cell transplantation. A fourth inpatient unit, adult oncology, prov ides care for patients with hematologic and solid tumor cancers. CBCI, the outpatient treatment center for transplant patients, is in a professional building attached to the hospital. Retrieved from http://www.bloodcancerinstitute.com/). Utilizing a multidisciplinary team approach, CBCI provides comprehensive care to transplant patients and their family members. CBCI is representative of other transplant programs in the country. The large psychosocial team at CBCI makes it an attractive research si te. In addition, there are large scale NIH funded research studies that have been done at this site and are currently being conducted by psychosocial research personnel at this site. Thus, there is a high level of support for psychosocial research both wi thin the psychosocial team and from the institution (T. Simoneau, personal communication, June 4, 2011).

PAGE 34

23 The ethnic and racial diversity in the SCT population treated at CBCI reflects the general diversity in the Colorado region: White 69%, Hispanic/Lat ino about 13%, Black or African American 8%, Asian 5%, American Indian/Alaska Native 5%. In 2011, a total of 242 transplants were conducted, and 77 were multiple myeloma patients (T. Simoneau, personal communication, May, 15 2012). Furthermore, a d atabase of over 200 current and previous multiple myeloma patients who have been treated at CBCI exists was utilized for recruitment purposes. Recruitment procedures are discussed in more detail below. Participants The study population consisted of spousa l/partnered caregivers of patients diagnosed with multiple myeloma (MM) who were currently being treated or had received treatment within the last two years at the Colorado Blood Cancer Institute at Presbyterian/St. Luke's Medical Center (PSLMC) Eligibili ty criteria for study enrollment included the following: 1) Caregiver must be co habitating and involved in a romantic relationship with the MM patient; 2) Patient who is being cared for has a diagnosis of MM and is currently or has received treatment at C BCI; 3) Partner must be in an active caregiving role for the patient diagnosed with MM; 4) Caregiver must be able to read and understand English; and 5) Caregiver must be 18 years of age or older. This study was limited to English speaking caregivers becau se many of the study instruments are validated only in English. Caregivers were excluded if they had a medical, psychological or cognitive condition that would interfere with the ability to consent and/or participate in the study. Competence for study pa

PAGE 35

24 personnel the goals of the study, requirements of study participation, and potential risks and benefits. This study did not enroll individuals who required proxy consent. Car egi vers were not excluded if they did not consent to blood draws. Preliminary eligibility was rela tionship status, and recruitment efforts were focused on those caregivers who met the above mentioned eligibility criteria. R ecruitment and Enrollment Procedures After obtaining approval to conduct the study through the Colorado Multiple Institutional Review Board (COMIRB) at the University of Colorado Denver and Presbyterian S Medical Center partnered caregivers of patients who have been diagnosed with MM were recruited from PSLMC in two manners 1) In person, during the pre transplant psychosocial intake and 2) Via a mail out (recruitment letters can be found in Appen dix B). In person recruitment primarily occurred during the pre transplant psychosocial intake, however on occasion patients and caregivers who had already received treatment at CBCI were approached. A ps ychologist, psychology fellow, social worker, or psy chology practicum student from the Psychosocial Oncology Department at PSL MC provided potential participants with an informational sheet describing the research study and provided a brief verbal description of the study. The informational sheet included th interested in receiving more information about the study, the HIPAA A Recruitment form contac ted the caregiver to provide more detailed information about the study and confirm that the caregiver met participation eligibility criteria. If the caregiver was still interested

PAGE 36

25 in participating, a study team member set up a time for him/her to come to the CBCI clinic for a onetime 30 45 minute visit. The mail out recruitment method consisted of three waves of letters, each wave included one initial contact letter and one follow up letter (included in Appendix B and C) sent out to CBCI patients in the M M database. Participants were provided study team participants contacted the study coordinator to set up a time to come to the CBCI clinic for a onetime 30 45 minute app ointment. In general, to control for circadian variability, appointment times occurred in a standardized block of time (9am 1pm) during the week for 15 months. In an effort to standardize behaviors that may impact neuroendocrine functioning, participants were instruc ted to not eat a meal at least two hours prior to the appointment, as well as to cease tobacco, caffeine and alcohol use the morning of the appointment. Other data that may impact neuroendocrine functioning (e.g. BMI, certain medications, pres ence of acute of chronic illness) were collected from the caregiver prior to the blood draw. Upon arrival to the clinic, the caregiver was consented. In order to ensure confidentiality and privacy, the study was explained in a private clinic room. During the consent procedure, a study team member reviewed the consent and HIPAA B forms thoroughly. The participant was informed that participation is completely voluntary. After reviewing the consent form thoroughly, a study team member inquired whether the par ticipant had questions. If the caregiver had no questions, he/she was asked to explain the purpose of the study to assess understanding. If the participant was fully aware of the

PAGE 37

26 details of the study and had no further questions, he/she signed and dated the consent form. He/she was then given a copy of the consent form. After the consent process, the caregiver was given a brief pre blood draw questionnaire (located in Appendix D) to complete. The caregiver was then escorted to the blood drawing area an d the blood draw was conducted by a trained phlebotomist. The caregiver was then given the written psychosocial survey to complete. A member of the study team processed the blood samples at the CBCI clinic and immediately froze samples (pr ocessing and stor age delineated more below). The participants were given the choice to 1) complete the survey at CBCI and return it to a study team member OR 2) complete the survey at home and mail it back in a self addressed envelope that was provided by a study team memb er upon request. We also obtained p ermission to give participants two reminder calls, should they not return the survey. A random study identification number was stamped on all pages of the survey as well as the blood collection tubes. The first page of th e survey contained general instructions and blanks for name, address, and telephone number. Once surveys were collected, the first page, with all identifying information, was removed. The study team maintained a separate tracking document that contained th e identifying information of the study participants with the corresponding random study identification numbers. Finally, the first page of the survey (with the identifying information) and the survey were stored in separate, locked file cabinets that could only be accessed by members of the study team.

PAGE 38

27 Study Design The research study was a cross sectional descriptive design, utilizing validated psychosocial assessments and blood draws with partnered caregivers of MM patients Data Collection Procedure s The pre blood draw survey required approximately 5 10 minutes to complete and the psychosocial survey required about 30 minutes for the caregiver to complete. The blood drawing procedure for the caregi ver required approximately five minutes. Table 1 pr ovides an overview of the key areas assessed and the respective measures utilized. The titles of the measures were not included in the caregiver psychosocial survey packet in order to reduce response and social desirability biases. Caregiver pre blood draw screen. Literature suggests that many health factors and behaviors may temporarily impact neuroendocrine functioning (Kiecolt Glaser et al., 2010). The pre blood draw screen (l ocated in Appendix F ) inquired about acute and chronic illness, medications, he ight and weight to calculate body mass index, mental health conditions, and nicotine use. These variables were taken into consideration when analyzing the neuroendocrine samples and will be described more in the data analysis section. Caregiver psychologi cal processes. Caregiver psychosocial processes were assessed in th e domains of depression, negative affect, and illness related distress. Depression: The Center for Epidemiological Studies Depression Scale (CES D) The CES D is a 20 item Likert scale de veloped to measure depressive symptoms in the general population (Raldoff, 1977). The range of scores is 0 60, with higher scores indicating greater depression. A score > 16 indicates a clinically significant level of

PAGE 39

28 depression. The CES D has been shown to have adequate reliability for use as a measure of depressive symptoms in older adults (Hertzog, Van Alstine, Usala, & Hultsch, 1990) and has been widely used in dementia caregiving research (Pinquart & Sorensen, 2003) and other caregiver studies (Kim e t al., 2008; Kim et al., 2011). The CES D was scored by first reversing four items (4, 8, 16, and 12) and then summing all of the responses as suggested by scale developers (Raldoff, 1977). Mean scores are presented below to describe the sample and CES D total (summed) score was used as a continuous variable in inferential statistical analyses. Negative Affect: Positive and negative affect scale (PANAS). The PANAS is a 20 question self report measure that was used to assess positive and negative aspects o f affect ( Watson, Clark, & Tellegen, 1988 ). The PANAS asked participants to rate the extent to cale has been validated in non clinical samples ( Crawford & Henry, 2004 ). This scale has also been used in marital studies (Graham et al., 2009; Kiecolt Glaser et al., 2005), and studies assessing neuroendocrine responses and affect (Taylor, Gonzaga, Klein Hu, Greendale, & Seeman, 2006). The PANAS was scored by summing all of the ten items that loaded onto the Negative and Positive Affect subscales to yield a possible score of 50 on each subscale. While means of the summed scores for all participants on bo th subscales were summarized to describe our sample, only negative affect subscale was used in inferential statistical analyses. Illness related distress : Impact of events scale revised (IES R). The IES R is a 22 item self report measure that assessed subjective distress caused by traumatic events

PAGE 40

29 ( Weiss & Marmar, 1997 conceptualized as a traumatic event. Respondents were asked to indicate how much they were distressed or bothered by their l "difficulty" listed. Items were rated on a five point scale ranging from zero ("not at all") to four ("extremely"). The IES R yields a total score (ranging from 0 to 88), which was the continuous variab le used in our inferential statistical models. In addition, subscale scores can also be calculated by using the mean item response for the questions that loaded onto the Intrusion, Avoidance, and Hyperarousal subscales (range 0 4). It is a revised versio n of the older version, the 15 item IES ( Horowitz, Wilner, & Alvarez, 1979 ). The IES R contains seven additional items related to the hyperarousal symptoms of PTSD, which were not included in the original IES ( Weiss & Marmar, 1997 ). Caregiver patient re lationship characteristics Patient caregiver relationship characteristics were assessed in the domains of relationship quality and attachment from Relationship Quality: Social relationships index (SRI). The SRI was used to assess understanding, or a favor) or negative (e.g. how upsetting/unpredictable when needing advice, understanding or a favor) their significant other is in the context of asking for advice, understanding, or a favor ( Campo et al., 2009 ). This is a five question self report SRI has been validated for use in health studies. This scale yields three scores calculated by the mean item response for relational positivity, relational negativity, and importance of relationship.

PAGE 41

30 This data was used as descriptive data to further unde how their significant other responds to request s for advice, understanding or a favor. This scale was not used for inferential statistics. Quality Marriage Index (QMI) The QMI is a six item self report scale designed to m easure marital satisfaction by inquiring about the stability and strength of the marital relationship. Participants choose answers from a six point Likert Scale (strongly agree to strongly disagree). An additional question asked participants to rate the d egree of their overall happiness within the marriage, ranging from one = very happy to seven = very unhappy A total score is calculated by summing all of the question responses. Scores on this scale were used as continuous variables for inferential statis tics. Higher scores on the QMI indicate higher marital satisfaction, whereas lower scores corresponded with lower satisfaction (Norton, 1983). The wording of this measure in the caregiver inclusive of all types of relationships. Secure Attachment: Measure of attachment qualities (MAQ). The MAQ is a measure of adult attachment patterns ( Carver, 1997) Respondents were asked to rate their degree to which they agree with vario us statements on a four point Likert Scale ranging from one (not at all) to four (extremely). The MAQ has four separate scales to assess two scales reflecting aspects of the anxious ambivalent pattern, one reflecting the desire for merger called ambivalence close as I want th

PAGE 42

31 only their s ignificant other. This measure was initially included to determine level of secure attachment for use in inferential analyses. However, because attachment was only specified in Hypothesis 1.1, which was ultimately not tested due to difficulty obtaining oxy tocin data (described in more detail below), secure attachment and the three other MAQ subscale totals were used only as descriptive information. Caregiver quality of life : Caregiver Quality of Life Index Cancer (CQOLC). The CQOLC is a quality of life ins trument designed to measure family functioning, perceived burden, psychological adaptation, and psychological morbidity in family caregivers of persons with cancer. The 35 item sc ale can be taken in ten minutes and has been used in clinical and other setti ngs ( Weitzner, Jacobsen, Wagner, Friedland & Cox, 1999 ). A total score is yielded for this scale by reverse scoring items (4, 10, 12, 16, 22, 27, 28, and 34) and then summing all of the responses. This total score was used as a continuous variable in the interferential statistical analyses. Caregiver neuroendocrine regulation, immune functioning, and oxytocin genotype. Blood was collected via blood draws (approx. volume = 20mL) conducted by trained PSL MC phlebotomists. Samples were centrifuged at 3500rpm directly after collection for ten minutes in a small clinical centrifuge and aliquots were collected using a sterile Pasteur pipet into cryovials at the CBCI clinic. Materials for blood processing (e.g. clinical centrifuge, pipettes cryovials) were store d in a plastic Tupperware and brought to the clinic for each participant. Samples were transported on ice and stored at 70 o C in the biobehavioral laboratory in the UCD Department of Psychology. Samples were frozen

PAGE 43

32 until the study was completed and all samples were run together to avoid problems with assay drift and interassay variability. Oxytocin Receptor Genotyping. In order to genotype the participants for the OXTR gene (oxytocin receptor gene) DNA was first isolated in the plasma sample. Once the DNA was ready for genotyping, DNA was removed from the isolated sample and placed in a separate polymerase chain reaction, or PCR, tube. PCR is a biochemical technology in molecular biology used to amplify a single or few copies of a piece of DNA across s everal orders of magnitude, generating thousands to millions of copies of a particular DNA sequence. In this case, we amplified the OXTR gene DNA sequence. Unfortunately, the SNP assay procedure used to identify the OXTR genotype yielded no results. Upon c onsulting with Dr. Phiel and others who have expertise in this area, it appears that the DNA concentrations from the samples were too low for detection in the SNP assay procedure. This could have been due to either contamination during the DNA purification procedure or degraded DNA due to collection and storage methods. Methods to trouble shoot this challenge are discussed more in the Future Directions section. Inflammatory markers (CRP, TNF 6). Commercially available enzyme linked immunosorbent ass ay (ELISA) kits (R & D Systems) were used to measure levels of CRP, IL 6 and TNF The concentration obtained represents the overall level of systemic inflammation (both chronic and acute). Undiluted plasma samples were tested in dupli cate and according to the directions provided by the manufacturer for TNF 6 ELISA per the at 450 nm u sing an automatic microplate reader (LabSystems MultiSkan). The amount of

PAGE 44

33 inflammatory marker in each sample was determined using the standard curve generated manufactur These assays show minimal variability between the standard curves (less than 6% variability) in our laboratory (M. Coussons Read, personal communication, September 21, 2011). The mean of the duplicates w ere used as the unit of analysis for statistical evaluation of these data. Results from ELISAs were deemed usable based on the standard cure results that were yielded (C. Phiel, (Professor of Integrative Biology, whose lab I ran these data in), personal co mmunication, January 2015). However, the ELISA for TNF results, therefore TNF discussed more in the Discussion Limitations section. Caregiver characteristics Caregiver characteristics were assessed using self repo rt questions inquiring about sociodemographic information, caregiving variables, health behaviors, and sleep. Demographics. Sociodemographic variables were assessed via self report questions asking about age, gender, income, race, education level, occupat ion, and distance to travel to the clinic. Age was used as a continuous control variable in the regression analyses. The other variables were summarized to describe the sample. Caregiving variables. Caregiving variables were assessed using self report que stions asking about length of time being a caregiver, time since MM diagnosis, types of support provided by the caregiver (e.g. physical, emotional, practical) level of caregiver involvement, and perceived severity of patient illness. Perceived illness se verity and length

PAGE 45

34 of illness time were included as control variables in the regression models. The other variables were used to describe the sample. Health behaviors. Caregiver health behaviors were assessed using self report questions asking about diet, physical activity habits, alcohol, caffeine, medications, chronic illness/mental health diagnoses. Amount of weekly physical activity and self reported diet quality were used as continuous variables within the inferential statistical analyses. Body mass in dex (BMI) was used as a control variable in some of the regression models. Medications and presence of chronic illnesses were entered as categorical variables to determine whether or not differences existed with level of pro inflammatory markers between gr oups. Other health behavior variables were provided as descriptive information. Sleep Sleep was assessed using the Pittsburg Sleep Quality Index (PSQI). The PSQI is a 19 item self report measure that has been used to measure quality of sleep in a clinica l population. The instrument assesses sleep disturbances and quality during a one month duration. Seven component scores consist of sleep latency, subjective sleep quality, sleep efficiency habits, sleep duration, sleep disturbances, daytime dysfunction, a nd use of sleep medications ( Buysse, Reynolds, Monk, Berman & Kupfer, 1988 ). A total score was also calculated by summing these seven component scores for a range of 0 21, s than continuous variable for the inferential statistical analyses and was included as an indicator of health behaviors.

PAGE 46

35 Table 1. Key variables and sources of dat a Key Variables Sources of Data Caregiver Psychological Processes Depression Center for Epidemiological Studies Depression (CES D) General Affect Positive and Negative Affect Scale (PANAS) Impact of Events (illness specific) Impact of Events Scale Re vised (IES R) Caregiver Patient Relationship Characteristics Relationship Quality Social Relationships Inventory (SRI) Quality Marriage Index (QMI) Attachment Style Measure of Attachment Qualities (MAQ) Caregiver Quality of Life Quality of Life (Qo L) Caregiver Quality of Life Index Cancer (CQOLC) Caregiver Neuroendocrine Regulation and Immune Response Oxytocin genotype Blood plasma (20mL) Inflammation (CRP, TNF 6) Blood plasma (20mL) Caregiver Characteristics Demographics Age, gender, i ncome, race, education Caregiving variables Length of illness time amount of time spent with patient, level of caregiving involvement, perceived illness severity Health behaviors Alcohol, smoking, diet, physical activity Sleep Pittsburg Sleep Quality Index (PSQI) Patient Variables Demographics Household income Medical Caregiver reported stage of disease, treatment side effects, length of illness Note: TNF

PAGE 47

36 Data Analysis Scale Reliability. Before conducting planned inferential statistics, a psychometric assessment of the collected data was conducted to confirm internal consistency of the measurement tools used in the analyses. All of the measures were determined to have values for the measures are as follows: CES D ( = .89), QMI ( = .95), CQOLC ( = .92), PANAS Negative Affect ( = .86), PSQI ( = .78), IES R Avoidance ( = .71), IES R Intrusion ( = .83), IES R Hyperarousal ( = .77) and IES R Total ( = .83). Overview of scale reliability can be found in T able 2 Table 2. Tests of nor mality and internal consistency for variables used in inferential statistical analyses Variable (Scale used) Mean ( SD ) Media n Skewness (Std. Error) Kurtosis (Std. Error) Internal Consistency alpha) Outliers removed (Z score) Sleep Quality (PS QI total) 5.86 (3.95) 5.50 0.69 (0.37) 0.51 (0.73) 0.78 none CRP 1.04 (.83) 0.80 1.1 (0.37) 0.54 (0.72) n/a none IL 6 0.10 (0.03) .10 0.83 (0.35) 0.23 (0.69) n/a 1 (z = 4.52) Relationship Quality (QMI total) 31.21 (7.75) 33.5 1.29 (0.37) 1.01 (0.73 ) 0.95 none Illness related Distress (IES R total) 18.90 (12.61) 15.5 0.98 (0.34) .75 (0.67) 0.83 none Quality of Life (CQOLC) 52.33 (21.61) 50.50 0.58 (0.37) .23 (0.73) 0.92 none Negative Affect (PANAS Negative) 18.64 (6.88) 17.00 0.88 (0.37) 0.68 (0 .73) 0.86 none Depression (CES D) 12.94 (9.46) 10.00 .99 (0.34) .70 (0.66) 0.89 none

PAGE 48

37 Variable (Scale used) Mean ( SD ) Media n Skewness (Std. Error) Kurtosi s (Std. Error) Internal Consistency alpha) Outliers removed (Z score) Physical Activi ty 269.21 (219.2) 180 1.00 (.37) 0.71 (0.72) n/a 1 (z = 8.27) Health Quality 1.78 (.62) 2.0 .18 (.34) .49 (.67) Diet Quality 2.05 (.75) 2.0 .32 (.34) .21 (.67) n/a none Perception of Illness Severity Length of Illness T ime 1.8 (.82) 23.63 (23.32) 2.0 15 .88 (.34) 1.67 (.34) .42 (.67) 2.96 (.67) n/a n/a none 2 (z =4.56, 6.23) *Scores reflect data once outlier(s) were flagged and evaluated using a Z score of +/ 2.58. Z scores of skewness and kurtosis were determined with an absolute value of > 2.58. Quantitative analysis Data was first examined for outliers using box plots, histograms, skewness, and kurtosis. Outliers were removed in four cases (one participant for hours of phy sical activity ; two participants for years of caregiving time) after determining that the data was inaccurate when taken into context. For example, one participant reported exercising 12 hours per day. While she had a highly physical job as a rancher, this level of activity significantly skewed the data and had a z score of 8.27. In addition, 1 participant IL 6 had a z score of 4.25 and thus was removed for the inferential analyses. An overview of central tendencies for variables included in inferential st atistical analyses can be found in Table 2 Furthermore, no transformations were needed for data analysis. Scores on psychosocial assessments and plasma level of pro inflammatory makers were summarized using a variety of exploratory analyses in SPSS (e. g. frequency, box plots, histograms, measures of central tendencies, etc.) to examine and describe the data.

PAGE 49

38 All of the data was tested for all other parametric assumptions of each respective statistical test before analyses were conducted. Linearity, mean ing that the data are related in a linear fashion, is an assumption for bivariate correlations as well as a linear multiple regression. This was assessed via plotting the data. For multiple regressions the assumptions of normally distributed errors, and ho moscedasticity (the variance of error is the same across all levels of the IV, this was assessed graphically) were assessed (Field, 2005). Additional assumptions include: no multicollinearity (no two predictor variables should correlate perfectly), indepen dent errors (errors of each observation should be uncorrelated as measured by the Durbin Watson test, which should be between one and three), and independence of outcome variable value from the predictor variables, meaning that each value of an outcome var iable comes from a separate entity. In addition, for multiple regression, there is an assumption that no external variable (e.g. a variable that has not been included in the regression model) is correlated perfectly with any of the predictor variables. If this were the case, the model would not be reliable because other variables exist that can predict the DV just as well. According to Field, in multiple regression equations, predictors do not need to be normally distributed (2005). Missing data As expec ted there was minimal missing data. For psychosocial measures with one item missing, mean imputation either from the subscale on which the missing data was from (if there was a subscale) or the total sum was used. This was the case on the CQOLC and CES D. On the CQOLC, missing values were observed to be most prominent on the positively phrased items on questions such as I am satisfied with my sex life; If more than one item was missing the

PAGE 50

39 participant was list wise del eted from that particular analysis, such was the case for participant 15 on the CQOLC on the analyses that included Qo L (Aims 3, 4, and 5). Statistical Power. Statistical power is the probability that a statistical test will reject the null hypotheses, wh en the null hypothesis is actually false (Miles & Shelvin, 2001). Three factors that impact power include: 1) statistical significance criterion used (e.g. = 0.05); 2) magnitude of the effect of a given construct in a given sample; and 3) sample size. Based on calculations using G*Power (Erdfelder, Faul, & Buchner, 1996), the correlational analyses, = 0.05, power = 0.80, would require n = 21 to detect a l arge effect ( r = 0.50), n = 64 to detect a medium effect ( r = 0.30), and n = 614 to detect a small effect. power = 0.80, one would need n = 47 to detect a medium effect (r = 0.30) and n = 448 to detect a small effect. Based on the power iss us to find a medium or large effect significant 0.05, power = 0.80, two predictors would need n= 31, to detect a large effect (f = 0.35) and n = 68 to detect a medium effect (f = .15) One predictor variable would require 55 people to detect a medium effect (f = 0.15). Based on the Baron & Kenny model of mediation, one to two predictors are needed per regression equation to determine a mediational relationship (Baron & Kenny, 1986). B ased on the sample for each statistical test, which ranged from 50 44 (specific n for each analysis is included in Tables 9 13), there is a (1 estimated power) probability that a Type II error occurred (failure to detect significant relationship when one exists between two variables or a false negative). Post hoc power which is based on estimated effect size, statistical tests used (e.g. bivariate correlation or multiple regression), alpha (.05), and number of predictors for the regression models will be d iscussed within each sect ion that results are reported.

PAGE 51

40 CHAPTER III RESULTS Recruitment Accrual and Attrition Sixty two percent (n = 32) of the participants were recruited from the mail out recruitment process and 37% (n = 19) of the participants were recruited in person at CBCI. The response rate from the mail outs was 27%, out of a total of 120 individuals who were each sent one initial and one follow up letter. We were not able to track data about how many people were approached about the study in p erson versus how many people agreed to participate. This was due to records not being kept about which caregivers declined or consented by the multiple CBCI psychosocial staff, including social workers, post doctoral fellows, and practicum students, who w ere involved in recruitment efforts. In addition, there were three individuals who consented but did not complete the survey. Out of those three individuals, two did not return reminder phone calls and the other participant stated that he was too busy to c omplete the survey. A total of 46 caregivers participated in the blood draw portion of the study. The reasons for non participation included the participants stating that they had insufficient time and participants reporting that they did not want to parti cipate in the blood draw because of past difficulties with blood draws. Sociodemographic Characteristics of the Participant Sample The final sample included 51 caregivers, with a mean age of 62.9 ( SD = 7.85). Seventy seven percent (n = 39) were female a nd 23% (n = 12) were male. Regarding ethnicity, 88% (n = 45) identified as White, 6% (n = 3) identified as Latino/Hispanic, 2% (n=1) identified as Asian/Pacific Islander, and the other 2% (n = 1) identified as Multi Ethnic. Regarding family structure, 91% (n = 50) reported that they were married and one

PAGE 52

41 caregiver reported being in a committed relationship with a partner of the opposite sex. The mean length of time in the relationship was 413.04 months ( SD = 159.68) or 34.4 years. The majority of participant s (65%, n = 33) had one to three children. However, only 14% (n = 7) of participants had children who were living in the home at the time the study was conducted. Regarding level of education, 10% (n = 5) completed less than or equal to 12 th grade, 25% (n = 13) completed some college or an associate degree, and 65% (n = 33) completed college/advanced degree. Sixteen participants (31%) were currently employed full time, seventeen (33%) were employed part time, sixteen (31%) participants were retired, and o ne (2%) was on temporary medical leave from work. Twenty percent of participants (n = 10) reported that their household income decreased after their partner was diagnosed with multiple myeloma. Participant demographic info rmation is summarized in Table 3 b elow. Table 3 P articipant demographic v ariables Demographic Variables Frequency (%) Mean ( SD ) Range Gender Female 39 (77%) Male 12 (23%) Ethnicity White 45 (88%) Latino/Hispanic 3 (6%) Other 3 (6%) Family Structure Married/partnered 51 (100%) Family Income (Before Cancer) $0 50K 5 (10%) $50 9 9K 20 (41%) 24 (49%)

PAGE 53

42 Demographic Variables Frequency (%) Mean ( SD ) Range Family Income (Current) $0 50K 9 (18%) $50 99K 20 (41%) 20 (41%) Education Level th grade 5 (10%) Some college/associates degree 13 (25%) College/advanced degree 33 (65%) Employment Status Employed full time 16 (31%) Employed part time/Homemaker 17 (33%) Retired 16 (31%) Number o f C hildren 0 1 (2%) 1 3 33 (65%) 4+ 10 (20%) Children Living in the Home Yes 7 (14%) No 38 (75%) Age (Years) 62.90 (7.85) 37 76 Length of T ime in Relationship (Months) 413.04 (159.68) 90 708 *All percentages are based off a total N of 51. Patient Illness Characteristics Commonly report ed treatme nt side effe cts included GI distress ( nausea, diarrhea, vomiting, and constipation ) fatigue, and neuropathy L ess frequently reported side effects included hair loss, change in appetite, rash, insomnia, body aches, weight changes, and fever. While 39% (n = 20) of car and to the corticosteroids (dexamethasone) that were prescribed for their significant other. Interestingly, 8% (n = 4) of caregivers reported positive personality and mood changes in

PAGE 54

43 t istics are summarized in Table 4 Table 4 Caregiver reported p atient illness c haracteristics Variable % (n) Median (Range) Length of illness time (months) 14.5 (1 102) Perceived severity of diagnosis Very severe 41% (21) Somewhat severe 43% (22) Mildly severe or 12% (6) Not at all severe 4% (2) Stage 18% (10) 11% (6) 24 % (13) 9% (5) 31% (17) I II III IV Unknown or not provided Exact Diagnosis Highly variable responses (e.g. kappa light chain, lambda light chain) Treatments undergone* Chemotherapy (Revlamid, Velcade) 91% (50) SCT (1 or more) 80% (44) Radiation 26% (14) Dexamethasone 38% (21) Treatment related side effects Fatigue 44% (24) GI distress 38% (21) Neuropathy 22% (12) Experience of physical pain All the time 22% (11) Most of the time 12% (6) Some of the time 43% (22 ) None of the time 22% (11) Personality and mood changes experienced Negative 53% (n = 27) Positive 8% (n = 4) No changes 39% (n = 20) *Note: Some patients underwent multiple conjunct treatments. Figure 3 provides information on the frequencies of various combinations of treatment modality.

PAGE 55

44 Regarding patient illness characteristics, caregivers reported a median length of illness time of 14.5 months with a range from 4 105 months. Caregivers reported a number of various symptoms experienced by t heir significant other prior to receiving a multiple myeloma diagnosis including orthopedic pain (back, hip, shoulder), elevations on blood labs drawn at primary care physician (e.g. anemia, protein, white count), shortness of breath, fatigue, blood in sto ols, and broken bones, which generally resulted in the patient receiving an MRI or CT scan that revealed bone lesions and a subsequent multiple myeloma diagnosis. Regarding the stage and specific MM diagnosis received, there was a large amount of variabili ty in reporting of stage and diagnosis. 31% of caregivers did not know or did not provide staging information. Of those participants who did report stage, 24% (n = 13) of the patients were diagnosed with Stage III, followed by Stage I (18%, n = 10), Stage II (11%, n = 6) and Stage IV (9%, n = 5). Reasons for the variability of responses regarding exact diagnosis are included in the discussion section. Finally, regarding treatment modalities, 8% (n = 4) of caregiver reported that their significant other rec eived only chemotherapy, 4% (n = 2) received chemotherapy and radiation, 64% (n = 32) received chemotherapy and 1 2 stem cell transplants (autologous and or allogeneic are both included) and the remaining 24% (n = 12) received a combination of chemotherapy radiation, and one or more stem cell transplants. The two most commonly reported chemotherapies utilized were Revlimid and Velcade. Other types of adjunct treatments or procedures reported included kyphoplasty, back surgery, dialysis, and apheresis or b lood transfusions. Table 3 and Figure 3 provide an overview of the treatment modalities. Finally, 96% of caregiver reported that their significant other experienced some treatment related

PAGE 56

45 side effect. The most common side effects were fatigue (44%, n = 24) GI distress, which included diarrhea, nausea, vomiting, constipation (38%, n = 21), and neuropathy (22%, n = 12). Other reported side effects included hair loss, change in appetite, rash, insomnia, body aches, weight changes, and fever. Figure 3 Types of treatment r eceived by patient Caregiving Variables for the Participant Sample Caregivers reported a median time of 16 months serving as a caregiver (range = 1 358). Forty c of is provided in Figure 4.

PAGE 57

46 Figure 4. Caregiver perception of provision of adequate c are Specifically, 47% (n = 24) of caregivers re ported conducting physical caregi ving tasks such as helping their loved on e bathe, eat, dress, spending a median of 120 minutes per day (range = 0 720) for a median time of 11.5 months (range = 0 108). The entire samples (n = 51) of caregivers reported providing emotional support to their loved one such 60 minutes per day (range = 2 709) for a median time of 11 months (range = 1 109). Finally, 92% (n = 46) of caregivers reported providin g practical care such as medication management, household chores, transportation, and financial support for an average of 240 minutes per day (ran ge = 0 660) over a median time of 11 months (range = 0 105). A summary of the caregiving variabl es can be found below in Table 5

PAGE 58

47 Table 5 Caregiver level of involvement and health b ehaviors % (n) Median (range) Time as a caregiver (months) 16 (1 358) Physical Tasks 47% (24) Daily (minutes) 120 (0 720) Total duration (months) 11.5 (0 108) Emotional Tasks 100% (51) Daily (minutes) 60 (2 707) Total duration (months) 11 (1 109) Practical Tasks 92% (46) Daily minutes 240 (0 660) Total duration (months) 11 (0 105) Tobacco (cigar ettes per day) Yes 5 (10%) 10 (1 11) No 42 (82%) Mean ( SD ) Physical Activity (m inutes per week) 315 ( SD = 324.59) Yes 46 (90%) No 2 (4%) Alcohol (drinks per week) Yes 29 (57 %) 1.9 ( SD = 2.02) No 22 (43%) Caffeine (drinks per day) Yes 46 (90%) 2.10 ( SD = 1.48) No 5 (10%) Sleep Quality (PSQI) Poor 25 (49%) 5.86 ( SD = 3.95) Good 26 (51%) All percentages are based off a total N of 51. Health Related Variables for the Participant Sample Regarding overall quality of health, 33% (n = 17) of caregivers reported being in

PAGE 59

48 Figure 5. Caregiver reported health and diet q uality In terms of health behaviors, three caregivers reported having only one meal per day, 18 caregivers reported having 2 2.5 meals per day, 24 reported consuming three meals per day and six caregivers consumed more than three meals per day. Speci fically, 47 caregivers consumed at least one serving of leafy vegetables, regular vegetables, and whole grain per day, 46 caregivers consumed at least one serving of fruit and dairy per day, and 48 caregivers consumed at least one serving of meat per day. The majority (80%, n = 41) of caregivers reported engaging in recreational activities, which ranged from hobbies, physical activity, entertainment, and social experiences. In addition, 46 of the caregivers (90%) indicated that they participate in moderate physical activity for more than 10 minutes per day, with a median of 5.5 days per

PAGE 60

49 week (range = 1 7) for a mean of 315 minutes per week ( SD = 324.59). Regarding alcohol use, 57% (n = 29) of participants endorsed consuming alcohol, with an average of 1.9 d rinks per week ( SD = 2.02). In addition, 90% (n = 46) of participants reported consuming one or more caffeinated beverages daily and 10% (n = 5) of participants endorsed currently smoking tobacco with a median of 10 cigarettes per day (range = 1 11) for a median of 15 years (range = 40). Regarding quality of sleep, scores on the PSQI indicated that 51% (n = presented in Table 4. Additional Health Related Variables for Participants who Completed Blood Draw For participants who took part in the blood draw portion of the study, the mean height of participants was 64.85 inches ( SD = 4.60), the average weight was 161.53 pounds ( SD = 5.47), and the average BMI was 27.11 ( SD = 6.23). Regarding BMI, 23 participants (45%) were in the Normal Range (BMI = 18.5 24.99). However, the majority (n = 32, 70%) of caregivers reported having a chronic health condition, the most common of which were cancer (n = 9, 20%), hypertension (n = 7, 15%), hypothyroidism (n = 5, 11%) heart disease (n = 5, 11%) and sleep apnea (n = 4, 9%), with 26% of participants with two or more chronic health conditions Regarding the presence of mental illness, 11% (n = 5) caregivers endorsed a Major Depressive Disorder diagnosis and 4% (n = 2) endorsed a Generalized Anxiety Disorder diagnosis. Five caregivers (11%) reported being on a psychotropic medication (e.g. Zoloft, Wellbutrin). This inf ormation is presented in Table 6 below.

PAGE 61

5 0 Table 6 Caregiver health related variables and neuroendocrine d ata Variable Frequency (%) Mean (SD) Range Body Mass Index (BMI) Underweight (< 18.5) 0 (0%) Normal (18.5 24.99) 23 (45%) Overweight (25 29.99) 10 (20%) Obese Class I (30 34.99) 10 (20%) Obese Class II (35 39.99) 3 (6%) Obese Class III ( 40) 2 (4%) Presence of Acute Illness 2 (4%) Presence of Chronic Illness 30 (65%) Presence of Psychiatric Illness 7 (15%) Height (Inches) Weight (Pounds) BMI 64.85 (4.60) 161.53 (5.47) 27.11 (6.23) 54 73 110 260 19.22 46.48 CRP p lasma concentration (ng/mL) IL 6 plasma concentration (pg/mL ) 1.08 (0.80) 0.13 (0.17) 0.21 3.24 0.07 0.28 *BMI ranges are per the International Classification of BMI (World Health Organization, 2014). *Percentages are calculated out of 46. Psychosocia l Characteristics of the Participant Sample Psychological processes Caregiver psychosocial processes were assessed in the domains of depression, general affect, and distress associated with the illness of the loved one. A summary of the descriptive stati stics for the measures can be found in Table 7 Depression On the Center for Epidemiological Studies Depression Scale (CES D), participants had a mean score of 13.19 ( SD = 9.56), indicating that on average, caregivers were below clinical levels of depres sion. However, sixteen participants (31%) were above the clinical cut off score of 16, indicating clinically significant levels of depression. Affect. On the Positive and Negative Affect Scale (PANAS) caregivers reported a mean positive affect score of 33 .13 ( SD = 7.75) compared to negative affect ( M = 18.64,

PAGE 62

51 SD = 6.88). Scores were out of a possible 50 with higher scores indicating higher levels of positive or negative affect. Illness related distress The mean level of distress on the Impact of Events Sc ale Revised was 19.35 ( SD = 12.67) out of a total possible score of 88, where higher scores indicate more illness related distress. The subscale scores, which were generated using the item mean for each subscale, were the following: illness related intrusi ve thoughts ( M = 1.01, SD = 0.75), hyperarousal ( M = 0.84, SD = 0.70), and avoidance ( M = 0.79, SD = 0.56). Quality of Life On the (CQOLC C), caregivers had a mean score of 52.33 ( SD = 21.61) out of a potential total of 140 where higher scores indicate l ower quality of life. Caregiver patient relationship characteristics Patient caregiver relationship relationship quality and attachment. Relationship Quality On the S ocial Relationships Inventory (SRI) caregivers scored the following: relational positivity ( M = 4.29, SD = 1.40) and relational negativity ( M = 2.34, SD = 1.12). The majority of caregivers (78%) reported that their significant M =5.76, SD = 0.52). On another measure of relationship quality, the QMI, caregivers reported an average of 31.21 ( SD = 7.75) out of a possible 36, indicating on average caregivers perceived their relationships to be strong, stable, and happy. Attachment. On the Measure of Attachment Quality caregivers scored the following on the subscales: Secure ( M = 10.48, SD = 2.00), Avoidant ( M = 7.83, SD =

PAGE 63

52 3.22), Ambivalence Merger ( M = 5.1, SD = 2.30), and Ambivalence Worry ( M = 3.54, SD = 1.21). Highe r scores on each subscale indicate a higher level of that attachment style. Table 7 Caregiver scores on psychosocial m easures Scale/Construct Mean ( SD ) Interpretation Depression (CES D) 13.19 (9.56) Higher scores indicate more depression. Scores abo ve 16 indicate clinical levels of depression (16, 31%). Affect (PANAS) Positive 33.13 (7.75) Higher scores indicate more positive affect. Negative 18.64 (6.88) Higher scores indicate more negative affect. Illness related Distress (IES R) 19.35 (12.6 7) Higher scores indicate more illness related distress Intrusive Thoughts 1.01 (0.75) Out of 4. Higher scores indicate more severe sx. Hyperarousal 0.84 (0.70) Out of 4. Higher scores indicate more severe sx. Avoidance 0.79 (0.56) Out of 4. Higher sc ores indicate more severe sx. Importance of partner 5.76 (0.52) Relational Negativity 2.34 (1.12) Higher scores indicate more negativity, out of 6. Relational Positivity 4.29 (1.40) Higher scores i ndicate more positivity, out of 6. Overall Quality (QMI) 31.21 (7.75) Out of a possible 36. Higher scores indicate stronger relationship. Attachment Style (MAQ) Secure 10.48 (2.00) Out of 12. Higher indicates more secure attachment. Avoidant 7.83 (3 .22) Out of 20. Higher indicates more avoidant attachment. Ambivalence Worry 3.54 (1.21) Out of 12. Higher indicates more A W attachment. Ambivalence Merger 5.1 (2.30) Out of 12. Higher indicates more A M attachment. Quality of life (CQOLC C) 52.33 (21. 61) Out of a total of 140 where higher scores equal lower quality of life. *Caregiver Quality Of Life Cancer (CQOLC); Impact of Events Revised (IES R); Social Relationships Inventory (SRI); Measure of Attachment Quality (MAQ); Quality of Marriage Index (Q MI); Positive and Negative Affect Scale (PANAS); Center for Epidemiological Studies Depression Scale (CES D) *Note: Additional interpretative information for scales is located in the Discussion Section.

PAGE 64

53 Caregiver Neuroendocrine Variables A total of 46 caregivers participated in the blood draw portion of the study. There were no significant differences between demographic variables for total sample compared to demographic variables for those who also participated in the additional neuroendocrine portio n of the study (however, as noted, power was very low to detect differences) For the neuroendocrine data, t tests were utilized and no significant differences were found in pro inflammatory marker levels between participants on medication compared to no m edication, CRP: t (44) = .30, p >.05; IL 6: t (44) = 1.29, p >.05; Participants with a BMI of above 25 compared to a BMI below 24.99, CRP: t (44) = 1.78, p >.05; IL 6: t (44) = 1.86, p >.05; Nor participants with a chronic illnesses compared to those with no chronic illness: CRP: t (44) = 1.92, p >.05 ; IL 6 : t (44) = 1.37, p >.05. Therefore we were able to group all individuals together in subsequent analyses. These findings are p resented in Table 8 below. Table 8 Group differences for pro inflammatory m ark ers (CRP and IL 6) N Mean (SD) t df CRP Medication Yes 30 1.05 (.74) .30 44 No 16 1.12 (.94) BMI <24.99 21 .85 (.79) 1.78 44 >25 25 1.26 (.77)

PAGE 65

54 N Mean (SD) t df Chronic Illness Yes 30 1.23 (.81) 1.92 44 No 16 .77 (.58) IL 6 Medication Yes 30 .09 (.01) 1.29 44 No 16 .10 (.02) BMI <24.99 21 .09 (.01) 1.86 44 >25 25 .10 (.02) Chronic Illness Yes 30 .10 (.02) 1.37 44 No 16 .09 (.01) Plasma levels of IL 6 were found to be an average of 0.1 3 ng/l ( SD = 0.17) and plasma levels of CRP was 1.08 ( SD = 0.80). TNF which may signify a problem with the preparation of the samples or in the laboratory procedures. As a result, we were unable to include the TNF sults in the additional analysis. Also as mentioned above, difficulties purifying the DNA impeded us from obtaining the OXTR genotype information. The results for the other biomarker data are presented in Table 6 Results of Inferential Statistical Analyse s for Each of the Study Aims Aim 1: Evaluate the relationships between caregiver oxytocin genotype, psychological processes, and caregiver relationship characteristics in the context of caring for a loved one with MM.

PAGE 66

55 Hypothesis 1.1: Caregivers with GG oxy tocin genotype will report a higher number of positive caregiver patient relationship characteristics (secure attachment style, high relationship quality). Hypothesis 1.2 : Caregivers with GG oxytocin genotype will report lower levels of distress (depressi on, negative affect, illness related distress). Due to methodological difficulties with the SNP assay and the DNA purification procedures, OXTR genotype data was not obtained and Aim 1 was unable to be completed for this study. Aim 2. Evaluate the relatio nships between caregiver pro inflammatory markers, psychological processes and health variables. Hypothesis 2.1: Caregivers who report higher levels of distress (depression, negative affect, illness related distress) will have higher plas ma pro inflammator y markers In order to evaluate the relationship between levels of pro inflammatory cytokines (CRP and IL 6) and distress ( depression, negative affect, and illness related distress), bivariate correlations were conducted. CRP was not significantly associa ted with depression ( r = .06, p = .68), negative affect ( r = .18, p = .44), nor illness related distress ( r = .11, p = .48). Similarly, IL 6 was not significantly associated with depression ( r = .21, p = 0.18), negative affect ( r = .05, p = .76), nor ill ness related distress ( r = .04, p = .80). Regarding the power of these analyses, G*POWER ( Erdfelder, Faul, & Buchner, 1996 ) estimated that the power of the correlational analysis examining IL 6 and depression was .28, indicating that there is a .72 probab ility that a Type II error occurred assuming that a relationship exists between IL 6 and depression in the sample. Estimated effect size or strength of the r

PAGE 67

56 coefficie nt for the correlational analyses. F or all of the above correlational analyses the effect size ranged from .21 (small) to .04 (nil) Overview of these correlational relationships can be found in Table 9. Table 9. Correlation table for caregiver distress and inflammatory marker variables Negative a ffect Depression Illness related d ist r ess CRP IL 6 Age Depression (n) .80** (50) Illness related d istress (n) .65** (50) .70** (50) CRP (n) .18 (44) .06 (44) .11 (44) IL 6 (n) .05 (44) .2 1 (44) .04 (44) .39** (46) Age (n) .08 (51) .18 (50) .01 (50) .12 (44) .08 (44) Length of illness time (n) .09 (49) .14 (48) .07 (48) .12 (42) .13 (42) .13 (49) ** p <.01, p < .05 (two tailed) Hypothesis 2.2 : Caregiver distress (depress ion, negative affect, and illness related distress) will significantly predict the level of caregiver pro inflammatory markers while controlling for caregiver age, length of illness time, and caregiver perception of illness severity. Specifically, within t hose models, depression, negative affect, and illness related distress will be significant individual predictors of the level of pro inflammatory markers above and beyond the control variables. In order to determine whether caregiver distress (depression negative affect, and illness related distress ) predicted caregiver pro inflammatory markers, six separate linear regression models were conducted with depression, negative affect, and illness related

PAGE 68

57 distress each predicting CRP and IL 6 separately, cont rolling for length of illness time caregiver perception of illness severity, and age. Length of illness time and caregiver perception of illness severity were both single item questions and were included in the model as control variable s to help control for variation in the sample due to caregivers being at different points in their caregiver trajectory. In addition, older age has been shown to be related to higher systemic levels of inflammation, so this variable was included in the model to account for variability in our sample so that we can understand the unique contribution that the predictor variables, caregiver distress (depression, illness related distress and negative affect), had on the change in the dependent variable, IL 6 and CRP. None of th e models were significant and the regression coefficients are therefore not included in the text, but can be found in Tables 10 and 11 In addition, the Tables contain the standardized beta ( ) coefficient, which is a standardized measure of the relationship between the individual predictor and the DV within the regression model. Standardized beta utilizes standard deviations, making its value directly comparable across different models and across difference variables within models and thus give a better insight into the importance of a predictor (Field, 2005). The value is therefore a measure of the i ndividual contribution (effe ct size) for a single predictor in a regression model (Field, 2005) In f which is also in the tables, was used as a measure of effect or strength of a full regression model. Power analysis for the full models revealed the following: Model 1 (IL 6 predicted by depression, controlling for length of illness, perception of illness severity, and age): Power = .41 f = .14); Model 2 (IL 6 predicted by negative affect controlling for length of illness time perception of illness severity f = .09); Model

PAGE 69

58 3 (IL 6 pred icted by illness related distress, controlling for length of illness time perception of illness severity f = .08). An overview of all of these findings can be found in Table s 10 and 11. Table 1 0. Summary of multiple regression analysis for caregiver distress variables predicting CRP (n = 42) Variable B SE B R f Model 1 .02 .03 Control: Length of illness time (months) .00 .01 .09 Control: Perception of illness severity .01 .16 .02 Control: Age .01 .02 .07 Depression .01 .02 .05 Model 2 .07 .08 Control: Length of illness time (months) .00 .00 .08 Control: Perception of illness severity .08 .17 .08 Control: Age .00 .02 .02 Negative affect .03 02 .23 Model 3 .03 .05 Control: Length of illness time (months) .00 .01 .07 Control: Perception of illness severity .02 .16 .02 Control: Age .01 .02 .08 Illness related distress .01 .01 .12 Table 11. Summary of multiple regressio n analysis for caregiver distress variables predicting IL 6 (n = 42) Variable B SE B R f Model 1 .12 .14 Control: Time of illness (months) .00 .00 .23 Control: Perception of illness severity .03 .0 3 .12 Control: Age .00 .00 .00 Depression .01 .00 .32

PAGE 70

59 Variable B SE B R f Model 2 .08 .09 Control: Time of illness (months) .00 .00 .18 Control: Perception of illness severity .03 .04 .15 Control: Age .00 .00 .04 Negative affect .01 .01 .24 Model 3 .07 .08 Control: Time of illness (months) .00 .00 .21 Control: Perception of il lness severity .02 .03 .10 Control: Age .00 .00 .08 Illness related distress .00 .00 .20 Hypothesis 2.3: Caregivers reporting better health behavi or indicators, as measured by sleep quality, diet quality, and physical activity levels, will have lower levels of plasma pro inflammatory markers. The relationships between health indicators (sleep quality, level of physical activity and diet quality ) a nd pro inflammatory markers (CRP and IL 6) were assessed using bivariate correlations. CRP was not significantly related to sleep quality ( r = .03, p = .84). However, CRP was significantly associated with level of physical activity ( r = .41, p <.01) and self reported diet quality ( r = .27, P < .05). IL 6 trended towards correlational significance with sleep quality ( r = .22, p = .15) and was signific antly associated with the level of physical activity ( r = .40, p < .01). IL 6 and self reported quality of diet were not associated. Many of the r values indicated a small to medium strength of association as indicated by r values ranging from .1 .3 and ideally would need a sample of 65 or more to detect a significant relationship based on the a priori powe r analysis. These analysis

PAGE 71

60 included 42 46 participants as denoted in Table 12, meaning that these tests were likely underpowered to detect a significant relationship between the identified variables. For example, using G*Power it is estimated that the powe r of the analysis of the relationship between IL and an estimated effect size of .22, was .30. This suggests that there is a .70 probability that a Type II error occurred given a small relationshi p exists between those two variables. An overview of these findings can be found in Table 12. Table 12. Correlation table of caregiver health variables and inflammatory markers CRP IL 6 Sleep q uality Physical a ctivity Diet BMI Age (n) .12 (44) .05 ( 44) .12 (50) .33* (49) .16 (51) .13 (46) CRP (n) .39** (46) .03 (43) .41** (42) .27* (44) .27* (46) IL 6 (n) .22 (43) .40** (42) .16 (44) .24 (46) Sleep quality (n) .11 (49) .39** (50) .10 (45) Physi cal a ctivity (n) .07 (49) .10 (44) Diet (n) .11 (46) BMI (n) ** p <.01, p < .05 (two tailed) Note: BMI and age were included in the regression analyses as controls. Hypothesis 2.4 : Caregiver self reported health behavior indicators (sleep quality, diet quality, a nd physical activit y level ) will together predict level of caregiver pro inflammatory markers (IL 6, CRP), while controlling for age, length of illness time and body mass index Within that model, it wa s hypothesized that sleep quality, diet quality,

PAGE 72

61 and level of physical activity would be significant individual predictors of level of caregiver pro inflammatory markers above and beyond the control variables. These relationships were assessed with two linear regression models, controlling for age BMI, and length of illness time BMI and age have been suggested in the literature to impact level of inflammation (Kiecolt Glaser et al., 2010) so these variables were included in the model to account for variability in our sample so that we can understand the uni que contribution that the predictor variables, quality of sleep, quality of diet and level of physical activity had on the change in the dependent variables, IL 6 and CRP. The full Model 1 that predicted CRP, indicated that 34% of the variance in CRP can be accounted for by diet quality sleep quality, age, hours of physical activity per week length of illness time and BMI together ( R 2 = 0.34, F (7, 32) = 2.40, p < 0.05). Specifically of interest, the unique contribution within the model accounted for by hours of physical activity per week, was significant ( = 0.47, p > .01). Sleep quality and diet quality were both not significant in this model as individual predictors. The overall f= .52. The f ull Model 2 that predicted IL 6, indicated that 21% of the variance in IL 6 can be accounted for by diet quality sleep quality, age, length of illness time hours of physical activity and BMI together ( R 2 = 0.21, F (7, 32) = 1.22, p = 0.32). None of the i ndividual predictors, nor the full models were significant. The overall strength of Model 2, which predicted IL 6, was f = .27 Using an effect size estimated at .27, a sample of 40 participants that w ere included in this analysis, 6 predictors a st hoc power for Model 2 was .62 Overview of these regression findings can be found in Table 13.

PAGE 73

62 Table 13. Summary of multiple regression analysis for caregiver health behavior variables predicting inflammatory markers (n = 40) Var iable B SE B R f Model 1 (CRP) .34 .52 Control: Length of illness time .04 .01 .09 Control: Age .00 .02 .09 Control: BMI .03 .02 .26 Sleep q uality .01 .04 .06 Physical a ctivity .08 .02 .47** Diet quality .22 .19 .20 Model 2 (IL 6) .21 .27 Control: Length of illness time .02 .01 .19 Control: BMI .01 .01 .18 Control: Age .01 .01 .18 Sleep q uality .01 .01 .18 Physical a ctivity .01 .05 .27 Diet quality .02 .05 .09 ** p <.01 (two tailed) Aim 3 E valuate the association between caregiver biomarkers and caregiver quality of life. Hypothesis 3.1 : Levels of plasma pro inflammatory markers will be inversely related to quality of life The relationship between pro inflammatory markers and caregiver quality of life was assessed with two bivariate correlations, which revealed that quality of life was not significantly related to CRP ( r = .17, p = .28) nor IL 6 ( r = .18, p = .24). A sample of 43 was included in both analyses. Therefore, the power for these analyses was estimated at .29 and .32, respectively. An overview of these relationships, including variables that are

PAGE 74

63 included in the regression model for this aim (described more below) can be found in Table 14. Table 14. Correlation table for ca regiver pro inflammatory markers and quality of life IL 6 Perception of illness severity Age Length of illness time Quality of life CRP (n) .39* (46) .04 (44) .12 (44) .12 (42) .17 (43) I L 6 (n) .02 (44) .05 (44) .02 (42) .18 (43) Perception of i llness severity (n) .22 (51) .14 (49) .01 (50) Age (n) .13 (49) .28* (50) Length of illness time (n) .13 (48) p <.05 (two tailed) Hypothesis 3.2 : Levels of pro inflammat ory markers (IL 6, CRP ) will predict caregiver quality of life, while c ontrolling for age, length of illness time, and caregiver perception of illness severity. Within those models it is hypothesized that pro inflammatory markers will be significant individual predictors of caregiver QoL above and beyond the control variables In addition, a multiple regression analysis indicated that together IL 6, perception of illness severity age, and length of illness time accounted for 19% of the variance in caregiver quality of life ( R 2 = 0.19, F (4, 36) = 2.21, p = 0.09). While this finding was not significant, it trended towards signific ance. IL 6 was not found to be significant as an individual predictor in the model, however age was found to be a significant predictor of quality of life within this model ( = .42, p < .01). Simila rly, a se cond model that included

PAGE 75

64 CRP as the continuous predictor variable, and perception of illness severity age, and length of illness time as control variables, indicated that together those variables accounted for 19% of variance in caregiver quality of life ( R 2 = 0.19, F (4, 36) = 2.17, p = 0.09), where age was also significant as an individual predictor of caregiver quality of life in this model ( = .41, p < .05). The overall strength of the Model 1 (which included IL f= .23 f = .23. Using an estimated effect size post hoc estimated power is estimated at .60 for both models. Overview of results can be found in Table 15. Table 15. Summary of multiple regression analysis for pro inflammatory markers predicting caregiver quality of life (n = 40) Variable B SE B R f Model 1 .19 .23 Control: Time of illness (months) .01 .01 .13 Control: Perception of illness severity 4.52 3.96 .18 Control: Age 1.10 .41 .42* IL 6 220.02 193.10 .17 Model 2 .19 .23 Control: Time of illness (mont hs) Control: Perception of illness severity 4.58 3.93 .19 Control: Age 1.08 .42 .41* CRP 4.39 4.07 .16 p <.05 (two tailed) Hypothesis 3.3 : Quality of life will be significantly associated in the positive direction with OXTR GG phenotyp e.

PAGE 76

65 This aim was not able to be evaluated due to issues wit h the DNA purification and SNP a ssay process es Aim 4 E valuate the relationship between caregiver health behavior indicators as measured by sleep quality, level of physical activity, self reported diet quality and psychological processes (depression, negative affect, illness related distress). Hypotheses 4.1: Quality of sleep quality of diet, and level of physical activity will be inversely related to level of depression, negative affect and illnes s related distress. In order to examine the relationship be tween caregiver health behavior indicators and psychological processes, bivariate correlations were utilized. Results indicated that sleep quality was significant ly associated with depression ( r = .56, p < .01), negative affect ( r = .37, p < .01) and illness related distress ( r = .59, p < .01). In addition, self reported diet quality was significantly associated with depression ( r = .32, p < .05 ) and illness related distress ( r = .30, p < .05). Aga in, it should be noted that negative affect, depression, and illness related distress were conceptualized to measure one underlying construct (caregiver distress) and are therefore significantly intercorrelated. Correlational values can be located in Table 16. However, level of physical activity was not significantly related to any of the three measures of caregiver distress/psychological processes (depression, negative affect, and illness related distress). Sample size for these analyses ranged from 48 50 participants. An estimated power for the relationship between physical activity and depression is .50, indicating that there is a .5 probability that a Type II error occurred.

PAGE 77

66 Table 16. Correlation table of caregiver psychological processes and health variables. Negative Affect Illness related d istress Physical a ctivity Sleep q uality Diet quality Perception of illness severity Depression (n) .80** (50) .70** (50) .28 (48) .56** (49) 32* (50) .03 (50) Negative Affect (n) .65** (50 ) .06 (49) .37** (50) 25 (51) .09 (49) Illness related d istress (n) .18 (48) .59** (49) 30* (50) .04 (50) Physical a ctivity (n) .02 (49) .1 5 ( 50 ) .30* (49) Sleep q uality (n) .3 9** (50) .16 (50) Diet quality (n) 3 0 (51) ** p <.0 1, p < .05 (two tailed) Note: Poorer sleep quality is indicated by higher scores on PSQI. Hypothesis 4.2: Sleep quality diet quality and level of physical activity together will predict caregiver distress (depression, affect, illness related distress) after controlling for length of illness time and perception of illness severity. Within that model it is hypothesized that sleep quality diet quality and physical activity will be significant as individual predictors of distress (depression, affect, illne ss related distress) above and beyond the control variables. Three separate multiple regression models were conducted predicting depression, illness related distress, and negative affect by sleep quality, diet quality and level of physical activity, contr olling for length of illness time and perception of illness severity. In Model 1, which predicted depression, 3 8 % of the variance in depression was accounted for by sleep quality, diet quality, level of physical activity, length of illness time and percept ion

PAGE 78

67 of illness severity ( R 2 = 0.3 8 F ( 5 4 0 ) = 4.86 p <0.001). Within Model 1, better sleep quality was significantly associated with depression scores ( = 0.5 2 p <0.01). For Model 2, which predicted negative affect, 30 % of the variance in negative affect was accounted for by sleep quality, diet quality and level physical activity, length of illness time and perception of illness severity (R 2 = 0.30 F ( 5, 41) = 3.56 p < 0.01). Again, sleep quality was significantly associated with negative affect scores ( = 0.49, p <0.01). Finally in Model 3, which predicted illness related distress, 39% of the variance in illness related distress was accounted for by s leep quality, physical activity, length of illness time and perception of illness severity ( R 2 = 0.39, F (4, 42) = 6.82, p < 0.01). Within Model 3, better sleep quality was also significantly associated (as an individual predictor) with illness related dis tress level ( = 0.63, p < .01). An overview of all regression findings can be found in Table 17. However, it should be noted that illness related distress, negative affect and depression were all highly intercorrelated, meaning that they are measuring the underlying c onstruct of caregiver distress and therefore the findings of the three aforementioned regression models are all related. The estimated effect or strength of f) and the individual predictors ( ) can be found in T able 17. Post hoc power analyses were not calculated for these tests; it can be assumed that because we had significant findi ngs that we had adequate power.

PAGE 79

68 Table 17. Summary of multiple regression analysis for health behaviors predicting caregiver distre ss (n = 46) Variable B SE B R f Model 1 (DV: Depression) .3 8 61 Control: Length of illness time (months) .05 .05 .13 Control: Perception of illness severity .91 1.68 .06 Sleep quality 1.19 .32 .52* Physical activity .01 .01 .08 Diet quality 2.30 1.78 .18 Model 2 (DV: Negative Affect) .30 .43 Control: Length of illness time (months) .01 .04 .02 Control: Perception of illness severity 1.92 1.03 .27 Sleep quality .57 .21 .39* Physical activity .01 .01 .14 Diet quality 2.10 1.17 .26 Model 3 (DV: Illness related distress) .40 .67 Control: Length of illness time (months) .03 .07 .05 Control: Perception of illness severity 1.46 2.16 .09 Sleep quality 1.80 .40 .60* Physical a ctivity .01 .01 .07 p <.01 Note: Poorer sleep quality is indicated by higher scores on PSQI. Aim 5. E valuate the relationships between caregiver distress and caregiver quality of life. Hypothesis 5.1: Level of illness related distress, level of depre ssion, and negative affect will be inversely related to caregiver QoL.

PAGE 80

69 The relationship between illness related distress, depression, and negative affect to caregiver quality of life was assessed using bivariate correlations. Caregiver quality of life was significantly related to depression ( r = .78, p < .01), negative affect ( r = .69, p < .01) and illness related distress ( r = .65, p < 01 ). Note that higher scores on caregiver quality of life measure indicated a lower quality of life, so QoL and measures of caregiver distress are inversely related. The se results are displayed in Table 18. Post hoc power analyses were not calculated for these tests; it can be assumed that because we had significant find ings that we had adequate power to detect the relation ships. Table 18 Correlation table for caregiver distress and quality of life Negative a ffect Illness related distress Length of illness time QOL Depression (n) .80** (50) .70 ** (50) .14 (48) .78** (49) Negative a ffect .65** (50) .09 (49) .69** (50) Illness related d istress .07 (48) .65** (50) Severity perception .13 (49) .01 (50) Length of illness time .13 (48) ** p <.01, p < .05 (two tailed) Hypothesis 5.2: Caregiver distress (illness related distress, depression, and negative affect) will predict caregiver quality of life while controlling for perception of illness severity and length of illness time Within that model, measures of caregiver distress (illness related distress, depression, and negative affect) will be significant as individual predictors above and beyond the control variable s

PAGE 81

70 A linear regression model was employed including caregiver distress (illness related distress, depression, and negative affect) as predictors of caregiver quality of life while c ontrolling for per ception of illness severity and length of illness time The full model was significant and accounted for 71% of the variance in caregiver quality of life ( R 2 = 0.71, F (5, 41) = 20.19, p < 0.01). Specifically within the model, higher lev els of depression significantly predicted lower levels of caregiver quality of life ( = 0.54, p < .01). In addition, longer time of illness was also significantly associated with decreased quality of life ( = 0.24, p <.01). Negative affect and illness related distress were not significant as individual predictors in this model. An overvi ew of findings is presented in Table 19. Post hoc power analyses wer e not calculated for this test because there were significant findings. Table 19 Summary of multiple regression analysis for caregiver distress predicting caregiver quality of life (n = 4 6) Variable B SE B R f Model 1 .71 2.45 Control: Time of illness (months) .23 .08 .24* Control: Perception of illness severity 1.13 2.39 .04 Negative affect .64 .51 .20 Depression 1.28 .37 .54* Impact of illness .30 .21 .17 *p<.01 Explorato ry Aim 6 Examine the relationship between caregiver patient relationship quality and caregiver distress and caregiver quality of life in the context of caring for a partner with MM. Determine if relationship quality is a partial mediator of caregiver lev el of depression and caregiver quality of life.

PAGE 82

71 A mediation analysis was conducted using SPSS to assess whether the effect of depression (X) on quality of life (Y) was partially mediated by relationship quality (M). This model involved 3 steps, all of w hich controlled for length of illness time and perception of illness severity First, a linear regression determined that there was a significant association between the independent variable, depression, and the dependent variable, quality of life ( = .82 p < 0.01), while controlling for perception of illness severity and length of illness time A second linear regression found a significant relationship between the mediator variable, relationship quality (assessed using the QMI total score) and the indep endent variable depression ( = .31, p < 0.05). A third linear regression assessed the association of the mediator variable (relationship quality) and quality of life while controlling for the independent variable depression. R elationship quality was not a significant predictor of QO L when controlling for the independent variable (depression) ( = .13, p = .18), indicating that the assumptions of mediation (per Baron and Kenny) were not met Of note, a s an individual predictor, the inclusion of relationship quality reduced the associ ation between depression and QOL very little ( = .78, p < 0.01) A summary of all of the regression analyses statistics for each step can be located in Table 20. The mediation figure is presented in Figure 6. Table 20 Summary of multiple regressions f or mediation model (n = 45) Variable B SE B R f Step I (DV = Quality of life) .68 2.13 Control: Time of illness (months) .243 .09 .25* Control: Perception of illness severity .05 2.34 .00 Depression 1.93 .21 .82* Variable B SE B R f

PAGE 83

72 Step 2 (DV= R elationship quality) .20 .25 Control: Time of illness (months) .10 .04 .37* Control: Perception of illness severity .40 .97 .06 Depression .19 .08 .31* Step 3 (DV= Quality of life) .71 2.45 Co ntrol: Time of illness (months) .22 .09 .23* Control: Perception of illness severity .48 2.30 .02 Relationship quality .48 .36 .13 Depression 1.82 .21 .78** p <.05, ** p <.01 Figure 6: Mediation Model

PAGE 84

73 CHAPTER IV DISCUS SION Overview Caregivers of multiple myeloma patients face many challenges. The primary goal of this cross sectional study was to increase the understanding of the relationship between 6), psychological variables (depres sion, negative affect, illness related distress), patient caregiver relationship characteristics in the context of caregiving (relationship quality, attachment style), caregiver health behaviors (sleep quality, diet, physical activity) and quality of life during this multifaceted and stressful experience of caring for a loved one with multiple myeloma. Participant Characteristics In the following section, many comparisons are made between the current sample not formal statistical comparisons; rather, they are informal appraisals of relative levels. Caregiver Health Characteristics. Regarding chronic illness, 65% of participants in this study were managing some type of c hronic health condit ion (e.g. hypertension, cancer, hyperthyroidism ) and 26% were managing two or more chronic illnesses In comparison, the CDC estimated that 49% of persons (18 65) are affected by a chronic disease or condition (Ward, Schiller, & Goodman 2014). The higher rates in our sample are consistent with studies that suggest caregivers are at increased risk for medical comorbidities although the cross sectional nature of our study makes it impossible to assess if these comorbidities developed over the course of caregiving (Schultz & Beach, 1999;

PAGE 85

74 Vitaliano, Zhang & Scanlan, 2003) Interestingly, regarding physical activity behavior, 67% of our sample reported meeting the CDC requirement of 150 minutes of moderate physical activity or exercise per w eek, which is a higher percentage than the estimated 49.6% of adults who met the physical activity guidelines in 2012 ( http://www.cdc.gov/nchs/fastats/exercise.htm ). One possible reason for this difference is that Colorado typically has higher than the average national level of physical activity among older adults (Cen ter for Disease Control, 2015). Regarding other health related Interestingly, despite higher than average level of physical activity and positive perception of health and diet quality, 30% of our sample met World Health Organization criteria for obe sity (World Health Organization, 2014). Our sample was lower than the estimated national prevalence in adults older than 60 (35.4%) and higher compared to Colorado, which has a 20 25% obesity rate (Center for Disease Control, 2015). Regarding sleep quality about half of our sample met criteria on the PS QI for sleep difficulty, which was lower compared to another study of caregivers of stem cell transplant (SCT) patients in the acute phase of an allogeneic transplant, who found that all of the caregivers me t the criteria for sleep d isruption (Simoneau et al., 2013 ). Caregiving and Illness Characteristics Consistent with the literature on family caregivers of MM patients, caregivers in o ur sample reported a high amount of daily time dedicated to caregiving for their partners with MM ( Molassiotis et al., 2011 ). Specifically, our sample reported a median of 3 hours dedicated to physical caregiving activities, 2 hours dedicated to practical activities, and 1 hour dedicated to emotional activities daily. Many al so reported a decrease in income following their partn

PAGE 86

75 diagnosis, which has been reported in other studies of SCT recipients (Meehan et al., 2006; Chow & Coyle, 2011). Regarding staging for multiple myeloma, there is a wide range of heterogeneity found in the reported diagnosis/staging of our study sample. This may be due to the existence of various systems that help predict prognosis that are consistently being refined and updated as MM patients have longer rates of survival (Batail le, Annnweiler, & Beauchet, 2013). The two most common system s used are called the Durie Salmon System and the International Staging System for Multiple Myeloma, which are based on different diagnostic factors. Lack of accuracy and specificity among caregi ver report of their system and/or caregi ver specific issues. Also multiple myeloma tends to be a chronic, long standing illness for which pat i ents receive different types of treatment s over the course of their life. T racking specifics about stage and diagnosis can be difficult given that it may fluctuate over time. In addition, the caregiver receives a lar ge amount of complex medical information over a long period of time an d may potentially not have the cognitive resources to remember everything. This phenomenon is likely exacerbated by caregiver distress level. Caregiver Psychological Variables. The prevalence of clinical levels of d epression was 33% in our sample, which is consistent with other similar caregiver studies (Mosher et al., 2013; Langer et al., 2003; Lambert et al., 2013), lower than levels of depression compared to end of life cancer caregivers (Given et al., 2004) and caregiver s of SCT recipients (Laudenslag er et al., 2015; Simoneau et al., 2013 ) and higher levels compared to normative data as measured in community dwelling older

PAGE 87

76 adults (Lewinsohn, Seeley, Roberts, & Allen, 1997). Within the multiple myeloma literature, our sample reported higher levels of depression (31%) than the 13.6% cited in another sample of MM caregivers ( Molassiotis et al., 2011) Regarding positive and negative affect, our sample was observed to have lower levels of negative aff ect compared to caregivers of Multiple Sclerosis patie nts and similar levels of negative affect compared to a group of health care professionals (Bassi et al., 2014). Our sample was also observed to have lower negative affect levels on average compared to a study of parents of pediatric cancer patients (Hexem Miller, Carroll, Faerber, & Feudtner, 2013). T he complexity of emotions that accompanies a child receiving cancer treatment is likely more distressing than having a significant other in his/her early sixties diagnosed with and treated for cancer. Specifi cally, when couples age together there is an expectation that health problems will arise later in the relationship and later in life. On the contrary, it is rarer, and therefore arguably more distressing to have a child experiencing a life threatening illn ess as a parent. One study to date was located that included negative affect scores (PANAS) among caregivers of stem cell transplant recipients Comparatively our sample had slightly lower scores o n the negative affect subscale However there was only a t hree point difference in the average scores among participants in the two studies, which we would expect given the similarity in the two samples. Means and brief description of comparative samples can are summarized in Table 21. Regarding illness related d istress, as measured by the IES R our study had lower levels of illness related distress compared to a study with caregivers of allog eneic SCT recipients (Laudensla ger et al., 2015). Also parents of pediatric oncology patients,

PAGE 88

77 specifically those who are about to receive a SCT, had higher levels of illness related distress compared to our sample (Virtue et al., 2014). However, our sample was found to have higher levels of illness related distress compared to a sample of undergraduate students who had witne ssed at least one traumatic event (defined as the DSM IV criteria) in their lifetime (Adkins, Weathers, McDevitt Murphy & Daniels, 2008). One hypothesis for the observed difference is that our sample is likely re experiencing illness related traumatic even ts with each challenging side effect and medical treatment, potentially placing them at a higher risk for distress. Again means and sample demographic information is presented in Table 21. Regarding relationship quality, our sample reported lower relations hip quality compared to husbands of breast cancer patients ( Boeding, Pukay Martin, Porter, Kirby, Gremore, et al., 2014 ) and similar levels of relationship quality compared to a sample of underinsured primary care patients who are not caregivers (Woods & D enton, 2014). Compared to breast cancer, MM is characterized as a chronic and uncertain illness trajectory. The chronicity of challenges faced by MM caregivers may negatively impact marital quality contributing to the observed difference in marital quality compared to couples facing breast cancer. Finally, the group of primary care patients were thought to is likely that they were facing economic challenges that may have negatively impacted marital quality. As mentioned in the literature review, no studies were identified that included marital satisfac tion in partnered caregivers of MM patients. Regarding attachment to significant other, scores on the MAQ suggest that our sample had higher levels of secure attachment on average compared to an other sample

PAGE 89

78 of cancer caregivers (Kim et al., 2008) and healthy undergraduates (Carver, 1997). Table 21 displays comparative data among studies. Finally regarding quality of life, ou r sample reported higher quality of life compared to spousal caregivers of lung, breast, and prostate cancer patients (Weitzner, McMillan, & Jacobsen, 1999; Weitzner et al., 1999). While the samples included in these studies are demographically similar (e. g. race, level of education, length of illness time gender, etc.), the biggest difference is that about 40% of the caregivers were caring for a partner at the end of life, indicating end of life challenges that may negatively impact quality of life. Avera ge scores and comparative data can be located in Table 21. No studies were found that used the CQOLC in a MM sample, however, within the MM population, a study that used a different measure of Q o L, t he European Organization for Re search and Treatment of Ca ncer Quality of Life Scale, found that caregivers experienced moderately low quality life (Molassiotis et al., 2011). Our sample also reported moderate quality of life, indicating that caregivers of MM patients consistently report lower QoL compared to no n caregiving samples ( Vitaliano, Zhang, & Scanlan, 2003 ). Table 21. Comparative data for psychosocial measures across various studies Measure Sample Score Demographic data Reference CES D (depression) End of life cancer caregivers 14.87 (9.1) 75% 5 5 and up 87% Female 88% Spouses Given et al., 2004 Community dwelling older adults 8.33 (6.84) 14% above 16 63.9 (7.9) yrs 58% Female 77% married Lewinsohn et al., 1997

PAGE 90

79 Measure Sample Score Demographic data Reference SCT caregivers 15.2 (9.6) 4 5% above 16 52.2 (11.3) yrs 77% female 75% married Simoneau et al., 2013 MM caregivers 13.2 (9.6) 31% above 16 62.9 (7.85) yrs 77% female 100% married Present study PANAS (negative affect) S C T caregivers 21.5 (7.8 ) 56.5 (11.8 ) yrs 62% female 78 % spouses Kessler et al, 2014 MS family caregivers 21.2 (7.2) 46.4 (11.7) yrs 59% female 80% married Bassi et al., 2014 Health professionals 18.3 (5.4) 40.9 (8.8) 73% female 69% married Bassi et al., 2014 MM caregivers 18.6 (6.7) 62.9 (7.85) yrs 77% f emale 100% married Present study IES R (impact related distress) Parents of children receiving SCT 31.3 (22.5) 37.4 (8.1) yrs 88% mothers 69% married Virtue et al., 2014 Trauma exposed undergraduates STC caregivers 16.5 (16.8) 30.7 19.4 (1.6) 53% f emale 53.5 yrs 76% female 70% spouses Adkins et al., 2008 MM caregivers 19.4 (12.7) 62.9 (7.85) yrs 77% female 100% married Present study QMI (relationship quality) Husbands of breast cancer patients 38.7 (7.1) No age provided 100% male 100% marrie d Boeding et al., 2014 Underinsured primary care clinic patients 32.0 (10) 46 (12.2) yrs 81% female 100% married Woods & Denton, 2014 MM caregivers 31.2 (7.8) 62.9 (7.85) yrs 77% female 100% married Present study

PAGE 91

80 Measure Sample Score Demographic data MAQ (s ecure attachment) Cancer caregivers 3.69 (0.5) 56.5 (10.6) 50.5% female 100% married Kim et al., 2008 Undergraduates 3.57 (.49) 56% female Age and marital status not reported Carver, 1997 MM caregivers 10.5 (2.0) 62.9 (7.85) yrs 77% female 100% married Present study CQOLC (quality of life) Family caregivers of various cancer patients 96.2 60 yrs 60% female 80% spouses Weitzner et al., 1999 Family caregivers of cancer patients 93.2 59 yrs 67% female 70% spouses Weitzner et al., 1999 MM caregivers 52.3 (21.6) 62.9 (7.85) yrs 77% female 100% married Present study *Note: Comparative sample for IES R difficult to find as many of the extant caregiving studies used the original IES. Caregiver Neuroendocrine Variables. Levels of pro i nflammatory cytokines were similar in this study compared to other studies involving caregivers of cancer patients ( Futterman et al., 1996, Rohleder et al., 2009) and lower than in other stressed caregiver populations, such as parents of children with Auti sm or Attention Deficit Hyperactivity Disorder (Lovell, Moss & Wetherell, 2012). Regarding TNF inducible protein that is typically elevated in individuals with chronic and acute health conditions (Gruenewald et al., 2006; Papanicolaou et al., 1998; Pradhan et al., 2001). Furthermore, basal levels are typically below detection rate for commercially available ELISA kits (10pg/ml). Surprisingly, in our sample, despite many individuals actively managing chronic illnesses the TNF

PAGE 92

81 inducible protein that is generated during an immune response Lack of detection could signify a problem with the preparation of the samples or in the laboratory procedures however it was determined that the ELISA kit was valid as signified by the presence of an accurate standard curve. Discussion of Results Relat ed to Study Aims Many of the study hypotheses were supported and are consistent with the literature. First, within Aim 2, this study found that number of hours participants engaged in physical activity each week was significantly negatively related to lev els of plasma CRP and IL 6 which are indicators of systemic inflammation. This relationship is consistent with the literature that has noted that people who are physically active have lower CRP compared to less active counterparts (Kasapis & Thompson, 2005 ; Kiecolt Glaser et al., 2010). Furthermore, this important finding supports the critical role that regular physical activity may have on physical health, including lowering risk for negative health outcomes such as cardiovascular disease, cancer, type II diabetes, functional decline and arthritis, all of which are associated with inflammation (Kiecolt Glaser, McGuire, Robles, & Glaser, 2002). Specifically, chronic inflammation influences tumor promotion damage s healthy cells and inhibit s angiogenesis (Ki ecolt Glaser et al., 2010). In addition, our study also found that CRP and BMI were positively associated, suggesting that those who are heavier have more systemic inflammation. This association is also consistent with the literature (Kiecolt Glaser et al. 2010). In addition, as hypothesized, CRP was significant ly predicted by caregiver reported quality of diet, sleep quality, age, length of illness time hours of physical activity, and BMI together. Within that model physical activity was a significant pr edictor and was inversely re l ated to CRP.

PAGE 93

82 With regard to Aim 4, the study results indicated that that sleep quality was significantly inversely associated with levels of depression, negative affect, and illness related distress both in bivariate correlati onal analyses and in linear regression models that included length of illness time and perception of disease severity as control variables. These findings are consistent with previous literature suggesting that depression is negatively related to sleep qua lity (Antoni et al., 2006; Van Moffaert, 1994). In addition, diet quality was inversely related to depression and illness related distress in the correlational analyses. Finally, as hypothesized, positive health behavior indicators (sleep quality, diet qua lity, and level of physical activity) significantly predicted caregiver distress while controlling for age, length of illness time, and caregiver perception of disease severity. With regard to Aim 5, quality of life was inversely related to negative affe ct, illness related distress, and depression. Furthermore, a regression model including caregiver level of distress (depression, illness related distress, and negative affect), controlling for perception of illness severity and length of illness time also significantly predicted levels of caregiver quality of life. Depression was a significant individual predictor above and beyond all of the other measures of distress, length of illness time and caregiver perception of illness severity. In addition, longe r length of illness time was also a significantly associated with decreased quality of life, in the regression model. The studies examining the nature of the relationship between length of illness and quality of life have identified fear of recurrence, soc ial isolation, and marital stress as some factors that contribute to lower Q o L in middle to long term survivorship (defined as 3.5 years after diagnosis) (Lewis & Deal, 1995; Mellon, Norhouse, & Weiss, 2006). In addition, depression has been linked

PAGE 94

83 to lowe r quality of life in breast cancer patients undergoing autologous stem cell transplants (Gaston Johansson, Lachica, Fall Dickson, & Kennedy, 2004) and in cardiothoracic t ransplant patients (Myaskovsky et al. 2012) As patients diagnosed with multiple myel oma live longer, these findings indicate that it will be especially necessary to monitor and/or screen caregivers for factors that may negatively impact quality of life and provide interventions accordingly. With regard to Aim 6, this study found that wit hin two regression models, depression is negatively associated with relationship quality and quality of life, while controlling for perception of illness severity and length of illness time However, the full mediation model, which hypothesized that relati onship quality is the mechanism by which depression is related to quality of life, was not found to be significant in our sample. Rather, the data sugge st that depression has a direct relationship with quality of life that is not mediated by relationship q uality The significant relationship between variables measuring similar constructs supported the construct validity of some of the measures. All measures of caregiver distress (depression, negative affect, illness related distress) were positively assoc iated confirming that those measures were measuring a related underlying construct as we conceptualized. This was also the case for the inflammatory markers (CRP and IL 6) which were measuring ation). Also caregiver self reported quality of diet and sleep quality were positively associated. Again, these Not all of the hypotheses of the study were supported. Fo r example, contrary to what we hypothesized, quality of life was not significantly associated with the level of pro

PAGE 95

84 inflammatory markers. One reason for this finding may be that quality of life is a diffuse and multifaceted construct that cannot be accura tely represented by a biological correlate. In addition, we did not find a significant relationship between inflammatory cytokines and distress (negative affect, depression, and illness related distress) despite the fact that a link between IL 6 levels and depression has been supported extensively in the literature, specifically in older people living in the community (Dentino et al., 1999) and adults with untreated Major Depressive Disorder compared to healthy controls (Rawdin et al., 2013; Wright et al., 2005). The lack of significant findings in our sample could also be explained by an underpowered analysis which increases the likelihood that a Type II error occurred, meaning that a significant relationship was not detected in our sample although one may exist. Clinical Implications Research suggests that there are often unmet mental health care needs among MM caregivers (Molassiotis et al., 2011). For example, in our sample, despite the prevalence of clinical levels of depression, only 11% reported re ceiving a diagnosis of major depression, and only 9% (n = 4) reported being treated pharmacologically, suggesting a disparity in obtaining care when caring for others. While this study did not assess whether our sample was receiving psychotherapy, many of the caregivers in our sample had met an intake assessment. Furthermore the treatment of patients at CBCI includes the option of individual and family sessions, as we ll as caregiver support groups run by mental health care professionals (e.g. licensed clinical psychologists, post doctoral fellows, and social workers). Despite this service, caregiv ers have been observed to under utilize

PAGE 96

85 supportive services (T. Simoneau, personal communication, July 6, 2015) likely due to limited time. Caregivers and patients were also often referred to community mental health providers if a need was determined. While the model of including psychological support during treatment is likely helpful in addressing acute mental health concerns, it is unclear at CBCI whether caregivers maintain continuous mental health care after their significant other has completed treatment. Generally, a disparity in accessing mental health care for caregive rs has been found in other studies ( Santin, Treanor, Mills, & Donnelly, 2014) and has been linked to barriers such as inadequate knowledge of mental health services, inadequate time, stigma and avoidance of mental health care, a desire to manage emotional concerns independently and inadequate finances (Mosher, Given, & Ostroff, 2015). Despite limitations to access, there have been recently published frameworks of linking psychosocial care to MM patients and their families, that recognize the unique challeng es faced by MM patient and their families (Kurtin, Lilleby, & Spong, 2013; Zabora et al., 2015). These frameworks recommended regular distress screening and early intervention (e.g. around time of diagnosis) with a focus on psycho education, Cognitive Beha vioral Therapy, caregiver self care, and patient caregiver communication. While our cross sectional study does not allow us to infer causal relationships, we would hypothesize that based on the results, the use of multi modal interventions that focus on he althy behaviors and mood may improve psychological adjustment and potentially increase long term resilience in caregiving populations. Specifically, the results suggested that sleep hygiene, sleep quality, and physical activity would be ideal priority targ ets in interventions aimed at improving the health and mood of caregivers. In addition,

PAGE 97

86 interventions aimed at decreasing level of caregiver depression may simultaneously improve caregiver quality of life. Finally these findings speak to the importance of providing psychosocial screening and treatment to caregivers to mitigate the negative impact that caregiver distress has on their quality of life even in the context of a close, supportive relationship with a partner diagnosed with MM. Limitations There were a number of limitations to this study. These include the low number of participants resulting in underpowered statistical analyses, variable disease progression trajectories leading to heterogeneous caregiver experiences, the selectivity of the care giver sample, the methodological challenges associated with the preparation and analysis of the biological markers, and the limitations of a cross sectional design. As mentioned throughout this document, the low number of subjects impacted the ability to detect significant findings within our sample. Although the original goal of the study was to recruit 65 participants, which would have allowed us to detect a medium effect, we were unable to obtain that number due to slower than anticipated recruitment. As a result, the study may have been unable to detect significant relationships between some of the variables. While statistical techniques such as effect sizes and tentative language about trends towards significance were utilized, weaker relationships among variables may not have been detected due to lack of power. Future research could aim to replicate these findings in a larger sample. A second limitation was the potential for selection bias related to participants who responded to the recruitment e fforts. For example, the majority of the sample was White (88%); African Americans and Latino/as were underrepresented in our sample, based on

PAGE 98

87 the diversity at PSLMC and in the state of Colorado. In addition, the participants who replied to the mail out or agreed to participate during the in person recruitment were a self selected sample with sociodemographic variables including high income/education level and high level of physical activity, and thus may not be completely representative of all caregivers served at CBCI. Thus, these findings may not generalize to all MM caregivers of all ethnic/racial backgrounds. A third limitation is that caregivers were caretaking for loved ones at various points of disease trajectory and for variable amounts of time ( e.g. length of caregiving). The potentially impact a number of psychosocial and health outcomes. On the one hand, those who have served as caregivers for an extended per iod of time may report higher levels of stress. On the other hand, some individuals who have experienced the illness longer may have an attenuated stress response, as they have been exposed to illness related stressors for longer. It is also possible that participants who are new caregivers may report higher levels of distress given their close proximity to diagnosis and issues associated with their adjustment to caregiver activities. In addition, patients experienced a wide range of treatments and side e ffects, some more severe than others, also impacting caregiver adjustment. While some of this variability was accounted for by including caregiver perception of illness severity and time since diagnosis as control variables when assessing certain relations hips, it is impossible to capture all of the caregiving variables that may have impacted stress and psychosocial adjustment of the participants. A fourth limitation involved the collection and analysis of the biological data. For the majority of the parti cipants we employed standardized collection procedures (e.g.

PAGE 99

88 standard sitting time before blood draw, standard collection time (4 hour time window in morning), asking participants to cease nicotine, caffeine, food, and alcohol 3 hours prior to blood draw) and using data from the pre blood draw screen to assist us in controlling for variables or excluding data that may cause variation in inflammation levels. Six individuals were outside of the 9am 1pm time window, due to scheduling challenges, however thes e participants were not observed to be outliers. Also, of those who opted to return the survey via mail, the majority of caregivers returned the survey in less than a week. This difference in timing was not observed to impact the biological data. In addi tion, based on the nature of collecting data in older populations, many of our participants had chronic illnesses and were on medications. While medications such as antidepressants, systemic and respiratory steroids, nonsteroidal anti inflammatories, estro gens, immunomodulators, immunosuppressants, antirheumatic medications, chemotherapy and other anti cancer medications, diabetes medications, anti hypertensives have been shown to impact inflammation, psychoneuroimmunology researchers recognize that an olde r adult who is not on one or multiple medications is rare (Kiecolt Glaser et al., 2010). Thus it is suggested that statistical tests be utilized to determine whether significant differences exist between participants who are on medications or who are suffe ring from chronic health conditions and variables of interest. This was conducted on our sample and no significant differences were found with regard to inflammatory cytokine levels and presence of chronic medical diagnoses, BMI, and medication. While the downside is that the relationships between inflammation and psychosocial factors may be confounded by medications and chronic illnesses, the level of chronic illness in our sample is representative of the caregivers and does make these findings more genera lizable to the

PAGE 100

89 typical caregiver with regard to presence of health conditions. Also, as mentioned in the methods section, the SNP assay procedure used to identify the OXTR genotype yielded no results. There are a few methods to remedy this situation includ ing: repurifying the DNA as well as adding a larger amount of the current DNA sample into the SNP assay. This could be the focus of future research, which is discussed more below. Regarding the TNF methodological issues or not using a sensitive enough assay to detect the TNF One way to addre ss this question is to use a higher sensitivity assay, known as an AlphaLISA. However, in general w e would expect that TNF and IL 6 did to the psychological processes and health behavior indicators. Finally, the cross sectional design is problematic in that it does not allow one to infer causal relationships between the variables. However this design was the most feasible given resources, time restraints and reducing the burden on the participants In addition, there was not a control group of non caregivers with which to compare the descriptive data. Thus, when possible data from similar studies of caregivers which included controls were used to yield c omparisons ( see Table 21 ) Ideally, a control gro up consisting of individuals matched for sociodemographic factors such as age, gender, relationship status and SES as well as health variables such as level of physical activity and BMI would enable us to look at differences between psychological distress, immune functioning, and quality of life in our samples of caregivers. Future Research Directions As mentioned in the introduction, there is a strong link between marital/relational quality and health/well being. Future research should evaluate the relat ionship between

PAGE 101

90 relationship quality and health behaviors such as sleep, diet, and physical activity. Also, it would be interesting to look at how relationship quality and attachment style impacts level of inflammatory markers and caregiver reported psych ological variables such as depression, illness related distress and positive affect. In addition, future research should to understand the relationships between these patient characteristics and caregiver psychological response. In addition, regarding biological data that has been collected and analyzed, it would be interesting to evaluate if a specific quantity of physical activity exists that yields a clinically signi ficant decrease in inflammation that is associated with improved health and greater psychosocial adjustment. It would be interesting to pursue obtaining the OXTR phenotype data to evaluate are related to amount care. Additionally it would be fascinating to determine the relationship between caregiver oxytocin genotype, psychosocial variables, careg iver relationship characteristics in the context of caring for a loved one with MM, and demographic variables (e.g. gender, age). If we were successful at obtaining the OXTR genotype data, another future research goal would be to look at the relationship b etween OXTR genetic expression and psychological variables reported by these caregivers. One might hypothesize that variability in the psychological response and caregiver attachment to their loved one may have an epigenetic effect on the level of express ion of the OXTR.

PAGE 102

91 Finally, the literature suggests that gender differences exist among caregivers regarding level o f distress (Drabe et al., 2015), thus it would be interesting to look at whether gender differences exist among psychological, relationship, caregiving and health behavior variables of our sample.

PAGE 103

92 CCTSI EDUCATIONAL AND TRAINING EXPERIENCES The educational and practical learning experiences that I was afforded during my CCTSI training were tantamount to the implem entation of my dissertation project and my development as an academic researcher. I was first accepted into the CCTSI program in the Fall of 2012. At that time I already had developed a strong interest in psychoneuroimmunology. This interest was based on e arly graduate school experiences analyzing cortisol levels in informal caregivers of hospice patients, who participated in a Cognitive Behavioral Therapy (CBT) supportive intervention. This project coupled with my background in biology sparked my interest in psychoneuroimmunology, specifically incorporating biological markers to elucidate the connection between psychological stress and health. My acceptance into the NIH funded C olorado C linical and T ranslational S ciences I nstitute TL1 Pre doctoral Fellowshi p P rogram allowed me to foster this research passion through seminars, additional coursework, and additional research experiences. These experiences have built upon each other in a cumulative fashion and the sum of these experiences enabled me to include a nd analyze biological markers into my dissertation project. Coursework Through additional coursework, such as Tissue Biology and Disease Mechanism, I learned key concepts about neuropeptides such as oxytocin that are influenced by human behavior and have a molecular link to disease progression. Then through the course Practical Application of Molecular and Cellular Biology Techniques for the Clinician, I learned various basic science labora tory procedures and techniques as well as

PAGE 104

93 the principle functional m echanisms behind such techniques. Examples of procedures include: the Western Blot (used to detect specific proteins of interest in a tissue sample), the Northern Blot (used to detect the presence of mRNA or RNA and thus assess gene expression), Gel Electrophoresis (used to separate macro and micro molecules based on charge and size), Polymerase Chain Reaction (PCR) and Reverse Transcriptase PCR (used to magnify and analyze DNA and RNA composition respectively) and Enzyme Linked Immunosorbent Assay s (used to analyze the presence and concentration of a given antibody of interest). This technique will be discussed in more detail as this was one of the main techniques I utilized for my dissertation project. Research and Laboratory Training Experience s Enzyme Linked Immunosorbent Assays (ELISA) In the Fall of 2012, I sought out the additional research opportunity to conduct and refine my ability to utilize Enzyme Linked Immunosorbent Assays (ELISA). I assisted with a project that was assessing levels of inflammatory cytokines (IL 1, IL 6, and CRP) in cerebral spinal fluid, plasma, and saliva samples of individuals with various psychiatric diagnoses in the Biobehavioral Laboratory of Mary Coussons Read, Professor of Psychology. For this experience, I al so completed a Laboratory Safety training which familiarized me with laboratory and biological material safety precautions, regulations, and ethical matters related to working with biohazardous materials and human samples. Principle and Procedure The ELIS A uses the immunology concept of an antigen binding to an antibody. Specifically, the antigen of interest, which may include a pr otein, peptide, or hormone, binds to a specific antibody. In my study the antigens of interest were CRP, IL 6, and TNF

PAGE 105

94 pre coated onto a microplate. Standards and samples are pipetted into each well and incubated for a period of time, giving binding the opportunity to take place. Then the plates are washed to remove any unbound substances. Following this wash an enzyme linked polyclonal antibody is added to the wells and then incubated again. This technique is known as a sandwich enzyme (the antigen of interest is bound between two antibodies). After incubation the wells are washed again. Finally a chromogenic substrate for the enzyme is added yielding a color change based on the concentration of the enzyme present in each well. The wavelength of the color change is quantified by a microplate reader and the concentrations are determined based on comparisons to a standard curve. The entire process takes between 4.5 6 hours. Radioimmunoassay. In the Spring of 2013, I began training in the lab of Neuroscience Professor, Dr. Celia Sladek, learning how to conduct radioimmunoassays (RIA) to measure plasma oxytocin. For this experience, I utilized volunteer plasma samples to begin learning and refinin g the RIA technique. In addition to learning the RIA technique, I took an intensive Radiation Training course at Anschutz Medical Campus in order to work with radioactive materials. However, for me this technique was difficult to yield consistent results. In my experience, the oxytocin levels were often below detection. This was one of the concerns with measuring plasma oxytocin for my study because oxytocin degrades quickly in the blood and thus can be challenging to quantify. In the Spring of 2014, prior to me having all of my biological samples collected, Dr. Sladek retired and thus I needed to find another lab where I could continue to refine my skills and also run my oxytocin samples.

PAGE 106

95 Principles and Procedures The radioimmunoassay utilizes a similar te chnique to that of the ELISA in that the antigen antibody binding is used to determine the concentration of the antigen of interest. However, the RIA technique labels the antigen of interest with gamma radioactive isotopes of iodine. First, a known amou nt of radioactive antigen is mixed with the first antibody. Then unknown amounts of unmarked (non unknown antigen increase, an increased amount of radioactive an tigen is displaced from the antibody molecules. A second antibody is added forming a precipitate of bound antigen. Finally a Gamma Counter measures the radioactivity of the supranatant (free antigen) and the precipitate (bound antigen) to determine the c oncentration of the unknown antigen. The RIA is highly sensitive and specific and the process takes place over three days. DNA Isolation and SNP Assay Soon after learning that Dr. Sladek was planning to retire, I connected with a colleague who had recentl y begun looking at oxytocin receptor genotypes for a nother study. Upon researching the SNP assay method in depth, reviewing OXTR literature, and talking to others who have included oxytocin in their studies, I decided that this could be a more definitive a nd very interesting way to still include oxytocin in my study. So in the Spring of 2014, I began training in the Laboratory of Dr. Chris Phiel, Associate Professor of Integrative Biology, learning how to conduct DNA Isolation and SNP assays. The goal was t o assess for differences of genetic

PAGE 107

96 polymorphisms in the oxytocin receptor gene (OXTR) and then relate those differences to attachment and mood in my sample of caregivers. Principles and Procedures In order to conduct the OXTR genotyping, DNA must first be isolated. DNA is located inside the cells, and i s extracted out of the cells by adding 500 microliters of lysis buffer to the sample to create an environment in which the amount of water outside the cell is greater than the amount inside the cell. Via a process called osmosis, water moves out of the cell, eventually causing the cell to burst and the DNA to be exposed. In addition to the exposed DNA, proteins, nucleases, and other organic material are present. In order to remove this unwanted material, a solution called proteinase K (PK solution) i s added to the sample. By targeting molecular structures that contain amines (proteins and nucleases) and ignoring molecules that contain phosphates eaving DNA intact. The PK solution is attracted to the nitrogen on the proteins and nucleases and it dissolves them upon bonding. Twenty mi croliters of this PK solution is added and the tube i s vortexed briefly in order to accomplish this task. After placi ng the tube in a 60 C water bath for an hour (the manufacturer s recommended incubation period for the PK solution) the solution is removed from the original tube and placed in a 1.5 milliliter centrifuge tube. To this solution a capture buffer or CT solu tion, i s added to the centrifuge tube containing the sample. This capture buffer controls the pH of the solution and enables the DNA to be separated from the rest of the material. After the 4 00 microliters of CT solution is added, the centrifuge tube i s vo rtexed briefly then placed in a microcentrifuge and spun at 13,000 r pm for five minutes. This causes the heavier DNA molecules to go to the bottom of the centrifuge tube. The DNA that ended up on the bottom is called the

PAGE 108

97 pellet The rest of the material i s removed by pipetting leaving the isolated DNA. Finally, a re hydration buffer is added to the isolated DNA. This buffer help s to stabilize the DNA and prevents it from degrading. After 150 microliters of this solution is added, the sample i s incubated o ver night at room temperature and i s then ready for genotyping. Once the DNA i s ready for genotyping, two microliters con taining 10 nanograms of DNA are removed from the isolated sample and placed in a separat e polymerase chain reaction (PCR) tube. PCR is a biochemical technology in molecular biology used to amplify a single or few copies of a piece of DNA across several orders of magnitude, generating thousands to millions of copies of a particular DNA sequence. In this case, we amplified the OXTR gene DN A sequence using a STEPONE qPCR machine. In order to prepare the DNA sample for qPCR, a master mix containing DNA primers, TAQ polymerase, probes and water i s added to the PCR tube containing the DNA sample. The tube was placed into the machine where it i helix DNA strand leaving single DNA strands in solution. This process is called denaturation. The primers in the master mix isolate the portion of DNA containing the OXTR gene. In the case of the OXTR gen e, 348 of the 3 billion bas e pairs that make up DNA that a re isolated by the primers. TAQ pol ymerase then replicates these isolated portions of DNA in a sense amplifying the OXTR region. The binding of the probes in the master mix enables us to tell which alleles are present in each sample. The alleles present will determine the OXTR genotype (either G/G, G/A, or A/A ) for each participant. There a re 2 probes present. One only binds allele. Located on the probe s a re molecules that fluoresce, or light up, when they bind to speci fic parts of DNA. In their unbound state they do not fluoresce.

PAGE 109

98 Once they bind wavelength of light that is measured by t he S TEPONE qPCR machine. This i s done to determine the genotypes of our participants for the OXTR gene. Unfortunately, the SNP assay procedure used to identify the OXTR genotype yielded no results. Upon consulting with Dr. Phiel and others who have expertise i n this area, it appears that the DNA concentrations from the samples were too low for detection in the SNP assay procedure. There are methods to trouble shoot this challenge, such as including higher amounts of the DNA in the SNP assay or re isolating the DNA, which I hope to pursue during my post doctoral research fellowship. The practical application of all of these experiences required me to adapt into three difference basic science labs, in three difference departments over the course of three years, wh ich was a great learning experience! In addition, the collection of my samples required coordination and communication with multiple parties at the CBCI Clinic including the phlebotomists, schedulers, and laboratory manager. I also shared a space at CBCI w here I would do the initial blood processing which consisted of centrifuging, aliquoting, and freezing samples. Each of these experiences facilitated my professional growth and confidence. Finally, the CCTSI program, grant, and related experiences have bee n the topic of numerous conversations on interviews for internship and post doctoral positions. Those conversations have led to even more opportunity, ideas, collaborations, and continue to positively shape my professional development as an academic resea rcher. In sum, I am very fortunate to have been involved with the CCTSI program.

PAGE 110

99 REFERENCES Adkins, J.W., Weathers, F., McDevitt Murphy, M., & Daniels, J.B. (2008). Psychometric properties of seven self report measures of posttraumatic stress disorder in college students with mixed civilian trauma. Journal of Anxiety Disorders, 22 (8), 1393 1402. The American Cancer Society. (2015). How is Multiple Myeloma Staged? Retrieved on June 22, 2015 from http://www.cancer.org/cancer/multiplemyeloma/detailedguide/multiple myeloma staging The American Cancer Society. (2015). Cancer Facts & Figures. Retrieved on August 15, 2015 from http://www.cancer.org/acs/groups/c ontent/@editorial/documents/document/acspc 044552.pdf. Amico, J. A., Cai, H. & Vollmer, R. R. (2008). Corticosterone release in oxytocin gene deletion mice following exposure to psychogenic versus non psychogenic stress. Neuroscience L etters, 442 (3), 262 266. doi: 10.1016/j.neulet.2008.07.004 Antoni, M. H., Lutgendorf, S. K., Cole, S. W., Dhabhar, F. S., Sephton, S. E., McDonald, P. G., et al. (2006). Opinion The influence of bio behavioural factors on tumour biology: pathways and mec hanisms. Nature Reviews Cancer, 6 (3), 240 248. doi: 10.1038/nrc1820 Baron, R. M., & Kenny, D. A. (1986). Moderator Mediator Variables Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations. Journal of Personali ty and Social Psychology, 51 (6).

PAGE 111

100 Bartholomew, K. & Horowitz, L.M. (1991). Attachment styles among young adults: A test of a four category model. Journal of Personality and Social Psychology, 61 226 244. Bassi, M., Falautano, M., Cilia, S., Goretti, B., Grobberio, M., Pattini, M., et al., (2014). The coexistence of well and ill being in persons with multiple sclerosis, their caregiver and health professionals. Journal of the Neurological Sciences, 337, 67 73. Bakermans Kranenburg, M.J. & Van IJzendoorn, M .H. (2008). Oxytocin receptor (OXTR) and serotonin transporter (5 HTT) genes associated with observed parenting. Social Cognitive and Affective Neuroscience, 3, 128 134. doi: 10.1093/scan/nsn004 Bataille, R., Annweiler, C., Beauchet, O. (2013). Multiple M yeloma International Clinical Lymphoma Myeloma & Leukemia, 13 (6), 635 637. Berkman, L. F. (1995). Role of Social Relations in Health Promotion Psychosomatic Medicine, 57 (3), 245 254. Besser, A., & Prie l, B. (2005). The apple does not fall far from the tree: Attachment styles and personality vulnerabilities to depression in three generations of women. Personality and Social Psychology Bulletin, 31 (8), 1052 1073. doi: 10.1177/0146167204274082 Boeding, S., Pukay Martin, N., Porter, L.S., Kirby, J.S., Gremore, T.M., Keefe, F.

PAGE 112

101 satisfaction predicting support perception. Journal of Family Psychology 28 (5), 675 683. Bowlby, J. (1973). Attachm ent and loss: Vol. 2. Separation, anxiety, and anger. New York: Basic Books. Buchheim, A., Heinrichs, M., George, C., Pokorny, D., Koops, E., Henningsen, P., et al. (2009). Oxytocin enhances the experience of attachment security. Psychoneuroendocrinology, 34 (9), 1417 1422. doi: 10.1016/j.psyneuen.2009.04.002 Campo, R. A., Uchino, B. N., Vaughn, A., Reblin, M., Smith, T. W., & Holt Lunstad, J. (2009). The assessment of positivity and negativity in social networks: the reliability and validity of the social r elationships index. Journal of Community Psychology, 37 (4), 471 486. doi: 10.1002/jcop.20308 Carnelley, K. B., Pietromonaco, P. R. & Jaffe, K. (1994). Depression, Working Models of Other and Relationship Functioning Journal of Personality and Social Psyc hology, 66 (1), 127 140. doi: 10.1037//0022 3514.66.1.127 Carter, C. S. (1998). Neuroendocrine perspectives on social attachment and love. Psychoneuroendocrinology, 23 (8), 779 818. Carver, C. S. (1997). Adult attachment and personality: Converging evidence and a new measure. Personality and Social Psychology Bulletin, 23 (8), 865 883. doi: 10.1177/0146167297238007 Centers for Disease Control and Prevention. (2015). Exercise or Physical Activity. NCHS FastStats Web site. http://www.cdc.gov/nchs/fastats/exercise.htm

PAGE 113

102 Retrieved on May 2, 2015 from http://www.cdc.gov/nchs/data/series/sr_10/sr10_257.pdf Centers for Disease Control and Prevention. (2015). Adult Obesity Facts and Prevalence. Retrieved on A ugust 10 from http://www.cdc.gov/obesity/data/prevalence maps.html. Chen, F. S., Barth, M.E., Johnson, S.L., Gotlib, I.H., & Johnson, S.C. (2011). Oxytocin Receptor (OXTR) Polymorphisms and attachment in human infants. Frontiers in Psychology, 2, 200. doi :10.3389/fpsyg.2011.00200 Child, J.A., Morgan, G.J., Davies, F.E., Owen, R.G., Bell, S.E., Hawkins, K. (May 2003). High dose chemotherapy with hematopoietic stem cell rescue for multiple myeloma. N. Engl. J. Med. 348 (19): 1875 83. Chow, K., & Coyle, N. (2011). Providing Palliative Care to Family Caregivers Throughout the Bone Marrow Transplantation Trajectory Research and Practice: Partners in Care. Journal of Hospice & Palliative Nursing, 13 (1), 7 13. doi: 10.1097/NJH.0b013e3181fce813 Coan, J. A., Schae fer, H. S., & Davidson, R. J. (2006). Lending a hand: Social regulation of the neural response to threat. Psychological Science, 17 (12), 1032 1039. doi: 10.1111/j.1467 9280.2006.01832.x Coon, S. K., McBride Wilson, J., & Coleman, E. A. (2007). Back to basi cs: assessment, communication, caring, and follow up: a lesson from a couple's journey with multiple myeloma. Clinical journal of oncology nursing, 11 (6), 825 829. doi: 10.1188/07.cjon.825 829

PAGE 114

103 Crawford, J. R., & Henry, J. D. (2004). The positive and negati ve affect schedule (PANAS): Construct validity, measurement properties and normative data in a large non clinical sample. British Journal of Clinical Psychology, 43 245 265. doi: 10.1348/0144665031752934 Crispi, E. L., Schiaffino, K., & Berman, W. H. (199 7). The contribution of attachment to burden in adult children of institutionalized parents with dementia. Gerontologist, 37 (1), 52 60. Dahan, J. F., & Auerbach, C. F. (2006). A qualitative study of the trauma and posttraumatic growth of multiple myeloma patients treated with peripheral blood stem cell transplant. Palliative & supportive care, 4 (4), 365 387. Dantzer, R., O'Connor, J. C., Freund, G. G., Johnson, R. W., & Kelley, K. W. (2008). From inflammation to sickness and depression: when the immune sy stem subjugates the brain. Nature Reviews Neuroscience, 9 (1), 46 57. doi: 10.1038/nrn2297 Dentino, A.N., Pieper, C.F., Rao, M.K. (1999). Association of interleukin 6 and other biologic variables with depression in older people living in the community. J A m Geriatr Soc, 47, 6 11. Ditzen, B., Schaer, M., Gabriel, B., Bodenmann, G., Ehlert, U., & Heinrichs, M. (2009). Intranasal Oxytocin Increases Positive Communication and Reduces Cortisol Levels During Couple Conflict. Biological Psychiatry, 65 (9), 728 731. doi: 10.1016/j.biopsych.2008.10.011 Drabe, N., Klaghofer, R., Weidt, S., Zwahlen, d., Buchi, S., J. (2015). Mutual

PAGE 115

104 Respect to Relationship Quality, Physical Complaints, and Sense of Coherence in Couples Coping with Cancer. Psycho Oncology, 24 442 450. Erdfelder, E., Faul, F., & Buchner, A. (1996). GPOWER: A general power analysis program. Behavior Research Methods Instruments & Computers, 28 (1), 1 11. doi: 10.3758/bf03203630 Ershler W.B. & Keller, E.T. (2000). Age associated increase Interleukin 6 gene expression, late life diseases, and frailty. Annual Review of Medicine, 51 245 270. Feeney, B. C., & Collins, N. L. (2001). Predictors of caregiving in adult intimate relatio nships: An attachment theoretical perspective. Journal of Personality and Social Psychology, 80 (6), 972 994. doi: 10.1037/0022 3514.80.6.972 Fekete, E. M., Antoni, M. H., Lopez, C., Mendez, A. J., Szeto, A., Fletcher, M. A., et al. (2011). Stress buffering effects of oxytocin on HIV status in low income ethnic minority women. Psychoneuroendocrinology, 36 (6), 881 890. doi: 10.1016/j.psyneuen.2010.12.003 Ferrell, B.R ., Dow, K. H ., Leigh, S. Ly, J. Gulasekaram, P (1995). Quality of life in long term cance r survivors Oncol Nurs Forum 22 915 922 Field A. (2005). Discovering Statistics using SPSS. 2 nd ed. Sage Publications. London, UK.

PAGE 116

105 Foxall, M. J., & Gaston Johansson, F. (1996). Burden and health outcomes of family caregivers of hospitalized bone marr ow transplant patients. Journal of Advanced Nursing, 24 (5), 915 923. doi: 10.1111/j.1365 2648.1996.tb02926.x Futterman, A. D., Wellisch, D. K., Zighelboim, J., LunaRaines, M., & Weiner, H. (1996). Psychological and immunological reactions of family members to patients undergoing bone marrow transplantation. Psychosomatic Medicine, 58 (5), 472 480. Gaston Johansson, F. Lachica, E.M., Fall Dickson, J.M. & Kennedy, M.J. (2004). Psychological Distress, Fatigue, Burden, and Quality of Life in Primary Caregivers of Patients with Breast Cancer Undergoing Autologous Bone Marrow Transplantation. Oncology Nursing Forum, 31,6 1161 1171. Gimpl, G., & Fahrenholz, F. (2001). The Oxytocin Receptor System: Structure, function, and regulation. Physiological Reviews, 81 (2), 629 683. Given, B., Wyatt, G., Given, C., Gift, A., Sherwood, P., DeVoss, D, et al. (2004). Burden and depression among caregivers of patients with cancer at the end of life. Oncology Nursing Forum, 31(6) 1105 1117. Gonzaga, G. C., Turner, R. A., Keltner D., Campos, B., & Altemus, M. (2006). Romantic love and sexual desire in close relationships. Emotion, 6 (2), 163 179. doi: 10.1037/1528 3542.6.2.163 Gouin, J. P., Glaser, R., Loving, T. J., Malarkey, W. B., Stowell, J., Houts, C., et al. (2009). Attachme nt avoidance predicts inflammatory responses to marital

PAGE 117

106 conflict. Brain Behavior and Immunity, 23 (7), 898 904. doi: 10.1016/j.bbi.2008.09.016 Gouin, J. P., Carter, C. S., Pournajafi Nazarloo, H., Glaser, R., Malarkey, W. B., Loving, T. J., et al. (2010). M arital behavior, oxytocin, vasopressin, and wound healing. Psychoneuroendocrinology, 35 (7), 1082 1090. doi: 10.1016/j.psyneuen.2010.01.009 Graham, J. E., Glaser, R., Loving, T. J., Malarkey, W. B., Stowell, J. R., & Kiecolt Glaser, J. K. (2009). Cognitive Word Use During Marital Conflict and Increases in Pro inflammatory Cytokines. Health Psychology, 28 (5), 621 630. doi: 10.1037/a0015208 Greipp, P. R., San Miguel, J., Durie, B. G. M., Crowley, J. J., Barlogie, B., Blade, J., et al. (2005). International sta ging system for multiple myeloma. Journal of Clinical Oncology, 23 (15), 3412 3420. doi: 10.1200/jco.2005.04.242 Grewen, K. M., Girdler, S. S., Amico, J., & Light, K. C. (2005). Effects of partner support on resting oxytocin, cortisol, norepinephrine, and b lood pressure before and after warm partner contact. Psychosomatic Medicine, 67 (4), 531 538. doi: 10.1097/01.psy.0000170341.88395.47 Gruenewald, T, Seeman, T, Ryff, C, Karlamangla, A & Singer, B. (2006). Combinations of biomarkers predictive of later life mortality. Proceedings of the National Academy of Sciences, 103 (38), 14158 14163. Gutkowska, J., Jankowski, M., Lambert, C., MukaddamDaher, S., Zingg, H. H., & McCann, S. M. (1997). Oxytocin releases atrial natriuretic peptide by combining with oxytocin re ceptors in the heart. Proceedings of the National Academy of

PAGE 118

107 Sciences of the United States of America, 94 (21), 11704 11709. doi: 10.1073/pnas.94.21.11704 Hazan, C. & Shaver, P.R. (1987). Romatic love conceptualized as an attachment process. Journal of Pers onality and Social Psychology 59 511 524. Heinrichs, M., Baumgartner, T., Kirschbaum, C., & Ehlert, U. (2003). Social support and oxytocin interact to suppress cortisol and subjective responses to psychosocial stress. Biological Psychiatry, 54 (12), 1389 1 398. doi: 10.1016/s0006 3223(03)00465 7 Hertzog, C., Van Alstine, J., Usala, P. D., Hultsch, D. F., & Dixon, R. (1990). Measurement properties of the Center for Epidemiological Studies Depression Scale (CES D) in older populations. Psychological Assessment : A Journal of Consulting and Clinical Psychology, 2 (1). doi: 10.1037/1040 3590.2.1.64 Hexem, K.R., Miller, V., Carroll, W.K., Faerber, A. J., & Feudtner, C. (2013). Putting on a happy face: Emotional expression in parents of children with serious illness. Journal of Pain and Symptom management, 45 (3), 542 551. Horowitz, M., Wilner, N., & Alvarez, W. (1979). Impact of event scale measure of subjective stress. Psychosomatic Medicine, 41 (3), 209 218. Juhasz, K., Buzas, K. & Duda, E. (2013). Importance of r everse signaling of the TNF superfamily in immune regulation. Expert Review of Clinical Immunology, 9 (4) 335 348. Kasapis, C. & Thompson, P.D. (2005). The effects of physical activity on serum C reactive protein and inflammatory markers: a systematic revi ew. J. Am. Coll. Cardiol., 45 1563 1569.

PAGE 119

108 Kessler, E.R., Moss, A., Eckhardt, S.G., Laudenslager, M.L., Kilbourn, K., Mauss, I.B., Bowles, D.W., et al. (2014). Distress among caregivers of phase I participants: a cross sectional study. Supportive Care in Ca ncer, Kiecolt Glaser, J. K., Gouin, J. P., & Hantsoo, L. (2010). Close relationships, inflammation, and health. [Review]. Neuroscience and Biobehavioral Reviews, 35 (1), 33 38. doi: 10.1016/j.neubiorev.2009.09.003 Kiecolt Glaser, J. K., Loving, T. J., Stow ell, J. R., Malarkey, W. B., Lemeshow, S., Dickinson, S. L., et al. (2005). Hostile marital interactions, pro inflammatory cytokine production, and wound healing. Archives of General Psychiatry, 62 (12), 1377 1384. doi: 10.1001/archpsyc.62.12.1377 Kiecolt G laser, J.K., McGuire, L., Robles, T.R., & Glaser, R. (2002). Emotions, morbidity, and mortality: new perspectives from psychoneuroimmunology. Annu. Rev. Psychol., 53 83 107. Kiecolt Glaser, J. K., & Newton, T. L. (2001). Marriage and health: His and her s. Psychological Bulletin, 127, 4 72 503 Kiecolt Glaser, J. K., Preacher, K. J., MacCallum, R. C., Atkinson, C., Malarkey, W. B., & Glaser, R. (2003). Chronic stress and age related increases in the pro inflammatory cytokine IL 6. Proceedings of the Natio nal Academy of Sciences of the United States of America, 100 (15), 9090 9095. doi: 10.1073/pnas.1531903100

PAGE 120

10 9 Kim, Y., & Carver, C. S. (2007). Frequency and difficulty in caregiving among spouses of individuals with cancer: Effects of adult attachment and gend er. Psycho Oncology, 16 (8), 714 723. doi: 10.1002/pon.1110 Kim, Y., Carver, C.S., Rocha Lima, C. & Schaffer, K.M. ( 2011). Depressive symptoms among caregivers of colorectal cancer patients during the first year since diagnosis: A longitudinal study. Psy cho Oncology, 22 (2) 362 367. Kim, Y., Deci, E. L., Carver, C. S., & Kasser, T. (2008). Adult attachment and psychological well being in cancer caregivers: The mediational role of spouses' motives for caregiving. Health Psychology, 27 (2), S144 S154. doi: 10 .1037/0278 6133.27,2(Suppl.).S144 Kim, H.S., Sherman, D.K., Mojaverian, T., Sasaki, J.Y., Park, J., et al. (2011). Gene Culture interaction: Oxytocin receptor polymorphism (OXTR) and emotion regulation. Social Psychology and Personality Science, 2, 665 672 Kosfeld, M., Heinrichs, M Zak, P.J Fischbacher, U & Fehr E. (2005). Oxytocin increases trust in humans. Nature, 435, 673 676. doi: 10.1038/nature03701 Kulik, J. A., & Mahler, H. I. M. (2006). Marital quality predicts hospital stay following coronary artery bypass surgery for women but not men. Social Science & Medicine, 63 (8), 2031 2040. doi: 10.1016/j.socscimed.2006.05.022 Kurtin S., Lilleby, K., & Spong, J. ( 2013 ). Caregivers of multiple myeloma survivors. Clinical Journal of Oncology Nursing, 17 (6), 25 32. Kumsta, R., Hummel, H., Chen, F.S., Heinrichs, M.(2013). Epigenetic regulation of the oxytocin receptor gene: implications for behavioral neueroscience. Frontiers in Neuroscience, 7, 83.

PAGE 121

110 Lambert, S.D., Girgis, A., Lecathelinais, C. & Stacey, F. (2013) Walking a mile in their shoes: anxiety and depression among partners and caregivers of cancer survivors at 6 and 12 months post diagnosis. Supportive Care in Cancer, 21 75 85. Langer, S., Abrams, J., & Syrjala, K. (2003). Caregiver and patient marital satisfaction and affect following hematopoietic stem cell transplantation: A prospective, longitudinal investigation. Psycho Oncology, 12 (3), 239 253. doi: 10 .1002/pon.633 Laudenslager, M., Simoneau, T.L., Kilbourn, K., Natvig, C., Phillips, S., Spradley, J. et al. (2015). A randomized control trail of a psychosocial intervention for caregivers of allogeneic hematopoietic stem cell transplan patients: effects on distress. Bone Marrow Transplantation, 50 1110 1118. Lesko, L. M. (1994). Bone Marrow Transplantation Support of the patient and his/her family Supportive Care in Cancer, 2 (1), 35 49. doi: 10.1007/bf00355238 Lewinsohn, P.M., Seeley, J. R., Roberts, R., Allen, N. (1997). Center for Epidemiologic Studies Depression Scale (CES D) As a screening tool for depression among community residing older adults. Psychology and Aging, 12 (2) 277 287. Lewis, F.M. & Deal, L.W. (1995). Balancing out lives: a study of experience with breast cancer recurrence. Oncol Nurs Forum, 22 943 953. Light, K. C., Grewen, K. M., & Amico, J. A. (2005). More frequent partner hugs and higher oxytocin levels are linked to lower blood pressure and heart rate in p remenopausal women. Biological Psychology, 69 (1), 5 21. doi: 10.1016/j.biopsycho.2004.11.002

PAGE 122

111 Lovell, B., Moss, M., Wetherell, M. (2012). The psychosocial, endocrine, and immune consequences of caring for a child with Autism or ADHD. Psychoneuroimmunology, 37, 4, 534 542. Lucht, M. J., Barnow, S., Sonnenfeld, C., Rosenberger, A., Grabe, H. J., Schroeder, W., Rosskopf, D. ( 2009). Associations between the oxytocin receptor gene (OXTR) and affe ct, loneliness and intelligence in normal subjects. Progress in Neu ropsychopharmacology and Biological Psychiatry, 33 860 866. MacDonald, K., & MacDonald, T. M. (2010). The Peptide That Binds: A Systematic Review of Oxytocin and its Prosocial Effects in Humans. Harvard Review of Psychiatry, 18 (1), 1 21. doi: 10.3109/1067 3220903523615 Maier, S. F. (2003). Bi directional immune brain communication: Implications for understanding stress, pain, and cognition. Brain Behavior and Immunity, 17 (2), 69 85. doi: 10.1016/s0889 1591(03)00032 1 Marks, N., Lambert, J.D. & Choi, H. (200 2). Transitions to caregiving, gender, and psychological well being: A prospective U.S. national study. Journal of Marriage and Family, 64 657 667. Meehan, K. R., Fitzmaurice, T., Root, L., Kimtis, E., Patchett, L., & Hill, J. (2006). The financial requir ements and time commitments of caregivers for autologous stem cell transplant recipients. The journal of supportive oncology, 4 (4), 187 190. Mellon, S. Northouse, L.L. Weiss, L.K (2006). A population based study of the quality of life of cancer survivor s and their family caregivers Cancer Nurs 29 120 131

PAGE 123

112 Miles, J. N. V., & Shelvin, M. E. (2001). Applying regression and correlation: A guide for students and researchers. London: Sage Publications. Molassiotis, A., Wilson, B., Blair, S., Howe, T., & Cav et, J. (2011). Living with multiple myeloma: experiences of patients and their informal caregivers. Supportive Care in Cancer, 19 (1), 101 111. doi: 10.1007/s00520 009 0793 1 Molassiotis, A., Wilson, B., Blair, S., Howe, T., & Cavet, J. (2011). Unmet suppor tive care needs, psychological well being and quality of life in patients living with multiple myeloma and their partners. Psycho Oncology, 20 (1), 88 97. doi: 10.1002/pon.1710 Morse, J. Q., Shaffer, D. R., Williamson, G. M., Dooley, W. K., & Schulz, R. (20 12). Models of Self and Others and Their Relation to Positive and Negative Caregiving Responses. Psychology and Aging, 27 (1), 211 218. doi: 10.1037/a0023960 Mosher, C.E., Champion V.L., Hanna, N., Jalal, S.I., Fakiris, A.J., Birdas, T.J., Okereke, I.C., Ke sler, K.A., Einhorn, L.H., Given, B.A., Monahan, P.O. & Ostroff, J.S. (2013). Support service use and interest in support services among distressed family caregivers of lung cancer patients. Psycho Oncology, 22, 1549 1556. Mosher, C.E., Given, B.A., Ostrof f, J.S. (2015). Barriers to mental health care service among distress family caregivers of lung cancer patients. European Journal of Cancer Care, 24 (1), 50 59.

PAGE 124

113 Myaskovsky, L., Posluszny, C.M., Shultz, R., DiMartini, A.F., Swizer, G.E., et al. (2012). Pred ictors and outcomes of healtg related quality of life in caregivers of ca rdiothorasic transplant patients. Am J Transplant, 12, 3387 97. Neumann, I. D. (2002). Involvement of the brain oxytocin system in stress coping: interactions with the hypothalamo pit uitary adrenal axis. In D. O. S. T. D. Poulain (Ed.), Vasopressin and Oxytocin: From Genes to Clinical Applications (Vol. 139, pp. 147 162). Norton, R. (1983). Measuring marital quality: a critical look at the dependent variable. Journal of Marriage and t he Family, 45, 141 151. Orth Gomer, K., Wamala, S. P., Horsten, M., Schenck Gustafsson, K., Schneiderman, N., & Mittleman, M. A. (2000). Marital stress worsens prognosis in women with coronary heart disease The Stockholm Female Coronary Risk Study. Jama Journal of the American Medical Association, 284 (23), 3008 3014. doi: 10.1001/jama.284.23.3008 Palumbo, A., Rajkumar, V.S., San Miguel, J.F., Larocca, A., Nievisky, R., Morgan, G. et al. (2014) International Myeloma Working Group consensus statement for th e management, treatment and supportive care of patients with myeloma not eligible for standard autologous stem cell transplantation. Journal of Clinical Oncology, 32, 587 604. Papanicolaou, D. A., Wilder, R. L., Manolagas, S. C., & Chrousos, G. P. (1998). The pathophysiologic roles of Interleukin 6 in human disease. Annals of Internal Medicine, 128 (2), 127 137.

PAGE 125

114 Petersson, M. (2002). Cardiovascular effects of oxytocin. In D. O. S. T. D. Poulain (Ed.), Vasopressin and Oxytocin: From Genes to Clinical Applica tions (Vol. 139, pp. 281 288). Pinquart, M., & Sorensen, S. (2003). Associations of stressors and uplifts of caregiving with caregiver burden and depressive mood: A meta analysis. Journals of Gerontology Series B Psychological Sciences and Social Sciences, 58 (2), P112 P128. Pradhan, A. D., Manson, J. E., Rifai, N., Buring, J. E., & Ridker, P. M. (2001) C reactive protein, Interleukin 6, and risk of developing Type 2 Diabetes Mellitus. JAMA, 286 (3), 327 334. Blood Cancer Institute. (June 2011). Retrieved from http://www.bloodcancerinstitute.com/why choose us/psychology postdoctoral fellowship training.htm) Quirin, M., Kuhl, J., & Duesing, R. (2011). Oxytocin buffers cortisol responses to stress in individuals with impaired emotion regulation abilities. Psychoneuroendocrinology, 36 (6), 898 904. doi: 10.1016/j.psyneuen.2010.12.005 Radloff, L. S. (1977). The CES D Scale: A self report depression scale for research in the general population. Applied Psychological M easurement, 1 (3). Rawdin, B. J., Mellon, S.H., Dhabhar, F.S., Epel, E.S., Puterman, E, Su, Y., et al. ( 2013) Dysregulated relationship of inflammation and oxidative stress in major depression Brain, Behavior, and Immunity 31 143 152.

PAGE 126

115 Reversi, A., Casso ni, P., & Chini, B. (2005). Oxytocin receptor signaling in myoepithelial and cancer cells. Journal of Mammary Gland Biology and Neoplasia, 10 (3), 221 229. doi: 10.1007/s10911 005 9583 7 Riem, M. M. E., Pieper, S., Out, D., Bakermans Kranenburg, M. J., & va n IJzendoorn, M. H. (in press). Oxytocin receptor gene and depressive symptoms associated with physiological reactivity to infant crying. Social Cognitive and Affective Neuroscience. Advance online publi cation. doi:10.1093/scan/nsq035 Rodrigues, S. M., Sas low, L. R., Garcia, N., John, O. P., & Keltner, D. (2009). Oxytocin receptor genetic variation relates to empathy and stress reactivity in humans. Proceedings of the National Academy of Sciences USA, 106, 21437 21441. Robles, T. F., & Kiecolt Glaser, J. K. (2003). The physiology of marriage: pathways to health. Physiology & Behavior, 79 (3), 409 416. doi: 10.1016/s0031 9384(03)00160 4 Rohleder, N., Marin, T.J., Ma, R., & Miller, G. (2009). Biologic cost of caring for a cancer patient: Dysregulation of pro a nd anti inflammatory signaling pathways. Journal of Clinical Oncology, 27 (18), 2909 2915. Santin, O., Treanor, C., Mi lls, M., & Donnelly, M. (2014). The health status and health service needs of primary caregivers of cancer survivors: a mixed methods appro ach. European Journal of Cancer Care, 23 333 339. doi: 10.1111/ecc.12157

PAGE 127

116 Sbarra, D. A. (2009). Marriage Protects Men from Clinically Meaningful Elevations in C Reactive Protein: Results from the National Social Life, Health, and Aging Project (NSHAP). Psy chosomatic Medicine, 71 (8), 828 835. doi: 10.1097/PSY.0b013e3181b4c4f2 Scantamburlo, G., Hansenne, M., Fuchs, S., Pitchot, W., Marechal, P., Pequeux, C., et al. (2007). Plasma oxytocin levels and anxiety in patients with major depression. Psychoneuroendocr inology, 32 (4), 407 410. doi: 10.1016/j.psyneuen.2007.01.009 Schulz R., Beach, S.R., Ives, D.G., Martire, L,M., Ariyo, A.A., & Kop, W.J. (1999). Association between depression and mortality in older adults: the Cardiovascular Health Study. Arch Intern Med i cine, 160 1761 1768. Simoneau, T.L., Mikulich Gilbertson, S.K., Natvig, C., Kilbourn, K., Spradley, J., Grzywa Cobb, R., Phillips, S. et al. (2013). Elevated peri transplant distress in caregivers of allogeneic blood or marrow transplant patients. Psycho oncology, 22 (9), 2064 2070. Soothill, K., Morris, S.M., Thomas, C., Harman, J.C., Francis, B., McIllmurray, M.B. (2003). The universal, situational, and personal needs of cancer patients and their main carers. European Journal of Oncology Nursing, 7 (1), 5 13. Tabak, B. A., McCullough, M. E., Szeto, A., Mendez, A. J., & McCabe, P. M. (2011). Oxytocin indexes relational distress following interpersonal harms in women. Psychoneuroendocrinology, 36 (1), 115 122. doi: 10.1016/j.psyneuen.2010.07.004

PAGE 128

117 Taylor, S. E., Gonzaga, G. C., Klein, L. C., Hu, P. F., Greendale, G. A., & Seeman, T. E. (2006). Relation of oxytocin to psychological stress responses and hypothalamic pituitary adrenocorticaI axis activity in older women. Psychosomatic Medicine, 68 (2), 238 245. doi: 10.1097/01.psy.0000203242.95990.74 Tom, N., & Assinder, S. J. (2010). Oxytocin in health and disease. International Journal of Biochemistry & Cell Biology, 42 (2), 202 205. doi: 10.1016/j.biocel.2009.10.008 Tost, H., Kolachana,B., Hakimi, S., Lemai tre, H., Verchinski,B. A., Meyer Lindenberg, A. (2010). A common allele in the oxytocin receptor gene (OXTR) impacts prosocial temperament and human hypothalamic limbic structure and function. Proceedings of the National Academy of Sciences, 107, 13936 13941. Trox el, W. M. (2010). It's More than Sex: Exploring the Dyadic Nature of Sleep and Implications for Health. Psychosomatic Medicine, 72 (6), 578 586. doi: 10.1097/PSY.0b013e3181de7ff8 Van Moffaert, M. M. M. P. (1994). and Journal of Psychosomatic Research, 38 9 13. Virtue, S. M., Manne, S. L., Mee, L., Bartell, A., Sands, S., Gajda, T.M., et al. (2014). Psychological distress and psychiatric diagnoses among primary caregivers of children undergoing he mapoietic stem cell transplant: An examination of prevalence, correlates, and racial/ethnic differences. General Hospital Psychiatry, 36 (6), 620 626.

PAGE 129

118 physical health? A met a analysis. Psychological Bulletin, 129 (9), 946 972. von Kanel, R., Dimsdale, J. E., Mills, P. J., Ancoli Israel, S., Patterson, T. L., Mausbach, B. T., et al. (2006). Effect of Alzheimer caregiving stress and age on frailty markers interleukin 6, C reacti ve protein, and D dimer. Journals of Gerontology Series a Biological Sciences and Medical Sciences, 61 (9), 963 969. von Kanel, R., Kudielka, B. M., Preckel, D., Hanebuth, D., & Fischer, J. E. (2006). Delayed response and lack of habituation in plasma inte rleukin 6 to acute mental stress in men. Brain Behavior and Immunity, 20 (1), 40 48. doi: 10.1016/j.bbi.2005.03.013 Ward, B.W., Schiller, J.S., Goodman, R.A. (2012). Multiple Chronic Conditions Among US Adults: A 2012 Update. Preventing Chronic Disease DOI : http://dx.doi.org/10.5888/pcd11.130389 Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect The PANAS scales. Journal of Pe rsonality and Social Psychology, 54 (6), 1063 1070. doi: 10.1037/0022 3514.54.6.1063 Weiss, D., & Marmar, C. (1997). The Impact of Event Scale Revised. In J. Wilson & T. Keane (Eds), Assessing psychological trauma and PTSD. New York: Guildford. Weitzner, M.A., Jacobsen, P.B., Wagner, H ., Friedland J.& Cox., C. (1999). The Caregiver Quality of Life Index Cancer (CQOLC) scale: development and

PAGE 130

119 validation of an instrument to measure quality of life of the family caregiver of patients with cancer. Qual Life Res, 1 2, 55 63. Weitzner, M.A., McMillan S.C., & Jacobsen, P.B. (1999).Family caregiver quality of life: Differences between curative and palliative cancer treatment settings. Journal of Pain and Symptom Management, 17 (6), 418 428. Wright, C. E., Stri ke, P. C., Brydon, L., & Steptoe, A. (2005). Acute inflammation and negative mood: Mediation by cytokine activation. Brain Behavior and Immunity, 19 (4), 345 350. doi: 10.1016/j.bbi.2004.10.003 Woods, S. & Denton, W.H. (2014). The biobehavioral family model as a framework for examining the connections between family relationships, mental, and physical health for adult primary care patients. Families, Systems & Health, 32 (2), 235 240. World Health Organization (2014, May). BMI classification. Retrieved on Mar ch 2015, from http://apps.who.int/bmi/i ndex.jsp?introPage=intro_3.html Zabora, J., Buzaglo, J., Kennedy, V., Richards, T., Schapmire, T., Zebrack, B. et al. (2015). Clinical perspective: Linking psychosocial care to the disease continuum in patients with multiple myeloma. Palliative and Supportive Care, 13 829 838.

PAGE 131

120 APPENDICES A ppendix A: Consent Form Study Title: Caregiver patient relationship and stress response in multiple myeloma Principal Investigator: Kristin Kilbourn, Ph.D., M.P.H. COMI RB No: 12 0430 You are being asked to be in a research study. This form provides you with information about the study. A member of the research team will describe this study to you and answer all of your questions. Please read the information below and or not to take part. Why is this study being done? Spousal caregivers in general are a distressed group and there is limited research examining the physical and mental health of cou ples, one of whom is diagnosed with Multiple Myeloma. This study plans to learn more about the relationships between caregiver distress, patient caregiver relationship characteristics and stress hormone levels in spousal caregivers as their loved one unde rgoes bone marrow transplant. Furthermore, this study aims to understand whether patient caregiver relationship quality impacts general health status. We will collect blood defe nses against illness as well as about your stress response system. These are experimental tests and have no clinical value at this time. We may save left over cells or fluid after the tests are performed for any new test that might be relevant to stress and health. You are being asked to be in this research study because you are 18 years of age or older and living with and/or married to a Multiple Myeloma patient. Other people in this study Up to 80 local caregivers as well as the person they will be tak ing care of will be contacted or enrolled in this research study. What happens if I join this study? If you join the study, you will be asked to do the following: 1) You will be asked to provide about 20 mL of blood only one time. We will get blood by putti ng a needle into one of your veins and letting the blood flow into a glass tube. You may feel some pain when the needle goes into your vein.

PAGE 132

121 A day or two later, you may have a small bruise where the needle went under the skin. The samples will be used to m easure chemicals that sometimes change with stress. 2) You will be asked to fill out a pre blood draw screen which will ask about medical and mental health diagnoses, current medications, and other health behaviors (e.g. caffeine and nicotine use). This will take about 5 10 minutes to complete. 3) You will be asked to fill out a survey about symptoms of stress you may be experiencing, how being a caregiver has affected your life, your general physical and mental well being, and your relationship with the patient. This survey will take about 30 45 minutes to complete. What are the possible discomforts or risks? The study may include risks that are unknown at this time. While in the study you may find out about a psychiatric condition that you did not know abou t before starting the study. Discomforts you may experience while in this study include emotional distress or embarrassment when asked to think about your feelings related to your experience as a spousal caregiver of a Multiple Myeloma patient. If your spouse is still a patient at PSL, you may request to see one of the staff psychologists or psychology fellows. Otherwise, you will be provided with references for community providers if you experience emotional distress or embarrassment if you request or i t is believed to be helpful to you by our staff. If you choose to utilize these services you would be responsible for any associated costs. Some participants may feel burdened by filling out surveys or providing blood samples. You are encouraged to part icipate only if you feel that filling out the surveys and providing blood samples will not be a burden. In this study a trained phlebotomist will insert a needle, connected to a plastic tube, into a vein in your arm. We will use the tube to take a blood sample. This is the standard method used to obtain blood for tests. You will feel some pain when we first insert the tube into your vein. You may have some redness, swelling, or bruising where the tube goes under your skin. In some cases, this type of tube can cause an infection where it goes under the skin. In rare cases, it can cause a blood clot in the vein. You will have this tube inserted for about 30 seconds. A total of approximately 20 mL of blood will be collected from you only once in the co urse of the study.

PAGE 133

122 What are the possible benefits of the study? This study is designed for the researcher to learn more about the patient caregiver relationship, and the physical and mental health of spousal caregivers who are caring for a loved one wit h Multiple Myeloma. This study is not designed to treat any illness or to improve your health. Also there are risks as mentioned in the Discomforts and Risk Section. Who is paying for this study? This research is being funded by the Colorado Clinical an d Translational Sciences Institute. Will I be paid for being in the study? You will not be paid to be in the study. Will I have to pay for anything? It will not cost you anything to be in the study. Is my participation voluntary? Taking part in this st udy is voluntary. You have the right to choose not to take part in this study. If you choose to take part, you have the right to stop at any time. If you refuse or decide to withdraw later, you will not lose any benefits or rights to which you are entit led. If you leave this study, the person you will be taking care of will still receive their normal medical care. Can I be removed from this study? The study doctor may decide to stop your participation without your permission if the study doctor thi nks that being in the study may cause you harm, or for any other reason.

PAGE 134

123 What happens if I am injured or hurt during the study? If you have an injury while you are in this study, you should call Kristin Kilbourn, Ph.D., M.P.H. immediately. Her phone numbe r is 303 556 2687. We will arrange to get you medical care in the unlikely event that you incur an injury that is caused by this research. However, you or your insurance company will have to pay for that care. Things That Must be Reported to The Authoriti es We respect your right to privacy. But there are some things we cannot keep private. If you give us information about child neglect or child abuse, we have to report that to Social Services. If you give us information about someone hurting someone els e, we have to report that to the police. If a court orders us to hand over your study records, we have to hand them over to the court. Who do I call if I have questions? The researchers carrying out this study are Kristin Kilbourn, Ph.D., M.P.H. and Teri any questions you have now. If you have questions later, you may call Kristin Kilbourn, Ph.D., M.P.H. at 303 556 2687. You will be given a copy of this form to keep. You may have que stions about your rights as someone in this study. You can call Kristin Kilbourn, Ph.D., M.P.H. with questions. You can also call the Colorado Multiple Institutional Review Board (COMIRB). You can call them at 303 724 1055. Who will see my research infor mation? We will do everything we can to keep your records private, however it cannot be guaranteed. Both the records that identify you and the consent form signed by you may be looked at by others. They are: People at the Colorado Multiple Institutional Review Board (COMIRB) The study doctor and his/her team of researchers Colorado Clinical and Translational Sciences Institute, who is providing funding for this research study

PAGE 135

124 making sure that we follow all of the rules for research We might talk about this research study at meetings. We might also print the results of this research study in relevant journals. We will always keep the names of the research subjects, like you, private We will ask you to sign a different form that talks about who can see your research records. That form is called a HIPAA form. It will give the names of companies and universities who may see your research records. This authorization does not expire. However, you may withdraw this authorization for use and disclosure of your personal health information by providing a written request to the Investigator. If you withdraw this authorization, the Institution, the Investigator, the research staff, and th e research sponsor will no longer be able to use or disclose your personal health information from this study, except so far as that they have already relied on this information to conduct the study. Agreement to be in this study I have read this paper abo ut the study or it was read to me. I understand the possible risks and benefits of this study. I know that being in this study is voluntary. I choose to be in this study: I will get a copy of this consent form. Signature: Date: Print Name: Consent form explained by: Date: _____ Print Name: Investigator: Date:

PAGE 136

125 Additional Consent for Blood for Research Kristin Kilbourn, Ph.D., M.P.H. would like to keep some of the blood that is taken during the study but is not used for other tests. If you agree, the blood samples will be kept and may be used in future research to learn more about stress disorders. The research that is done with your blood samples is not designed to specifically help you. It might help pe ople who have stress disorders and other diseases in the future. Reports about research done with your samples will not be given to you or your doctor. These reports will not be put in your health records. The research using your blood samples will not affect your care. The choice to let Kristin Kilbourn, Ph.D., M.P.H keep the blood samples for future research is up to you. No matter what you decide to do, it will not affect the care that you will receive as part of the study. If you decide now that you r blood samples can be kept for research, you can change your mind at any time and contact your study doctor to let him or her know that you do not want Kristin Kilbourn, Ph.D., M.P.H. to use your blood samples any longer, and they will no longer be used f or research. Otherwise, they may be kept until they are used up, or until Kristin Kilbourn, Ph.D., M.P.H. decides to destroy them. In the future, people who do research with your blood samples may need to know more about your health. While Kristin Kilbour n, Ph.D., M.P.H. may give them reports about your health, they will not be given your name, address, phone number or any other information that will let the researcher know who you are. Sometimes blood samples are used for genetic research (about diseases that are passed on in families). Even if your blood samples are used for this kind of research, the results will not be told to you and will not be put in your health records. Your blood samples will only be used for research and will not be sold. The res earch done with your samples may help to develop new products in the future, but there is no plan for you to be paid. The possible benefits of research from your blood include learning more about what causes stress disorders and other diseases, how to pre vent them and how to treat them. The greatest risk to you is the release of information from your health records. Kristin Kilbourn, Ph.D., M.P.H. will protect your records so that your name, address and phone number will be kept private. The chance that th is information will be given to someone else is very small. There will be no cost to you for any blood collected and stored by Kristin Kilbourn, Ph.D., M.P.H. The University of Colorado Denver and the hospital(s) it works with have rules to protect inform ation about you. Federal and state laws including the Health Insurance Portability and

PAGE 137

126 Accountability Act (HIPAA) also protects your privacy. This part of the consent form tells you what information about you may be collected in this study and who might se e or use it. We cannot do this study without your permission to see, use and give out your information. You do not have to give us this permission. If you do not, then you may not be able to join this study. We will see, use and disclose your information only as described in this form and in our Notice of Privacy Practices; however, people outside the University of Colorado Denver and its affiliate hospitals may not be covered by this promise. We will do everything we can to keep your records a secret. It cannot be guaranteed. The use and disclosure of your information has no time limit. You can cancel your permission to use and disclose your information at any time by writing to the ou do cancel your permission to use and disclose your information, your part in this study will end and no further information about you will be collected. Your cancellation would not affect information already collected in this study. Kristin Kilbourn, P h.D., M.P.H. UC Denver Department of Psychology Campus Box 173, PO Box 173364 Denver, CO 80217 3364 Both the research records that identify you and the consent form signed by you may be looked at by others who have a legal right to see that information. Collected data may be discussed or presented at research meetings. Results of research may be printed in journals but your name will always be kept private. Please read each sentence below and think about your choice. After reading each sentence, circle doctor or nurse. Remember, no matter what you decide to do about the storage and future use of your blood, you may still take part in the study. I give my permission for my blood to be stored in a central tissue bank at Kristin Kilbourn, Ph.D., M.P.H. for future use by the study investigators: 1. I give my permissions for my blood samples to be kept by Kristin Kilbourn, Ph.D. for use in future research to learn more about how to prevent, detect, or treat stress disorders. Yes No _________Initials

PAGE 138

127 2. I give my permissions for my blood samples to be used for research about other health problems such as stress disorders like heart disease. Yes No _________Initials 3. I give my permiss ion for my study doctor (or someone he or she chooses) to contact me in the future to ask me to take part in more research. Yes No _________Initials I agree to take part in the study having to do with research on blood as indicated above. Sign ature____________________________ Date___________

PAGE 139

128 A ppendix B: Initial Recruitment Letter Patient Hello, Thank you for taking your time to read this letter! I am a doctoral student in the Clinical Psychology program at Un iversity of Colorado Denver. I am presently working on a research study for my doctoral dissertation in partnership with Teri Simoneau, Ph.D., a psychologist, at the Colorado Blood Cancer Institute. You are receiving this letter because you are a patient who has been diagnosed with multiple myeloma and have received treatment at CBCI. At times during your treatment, your significant other has served in a caregiver role. Caregivers often experience high levels of stress. The purpose of this study is to eva luate levels of stress (both self reported stress and stress hormone levels) and quality of life when caring for a loved one with a experiences in order to continue to provide better care to family members/caregivers. Enclosed in this envelope you will find another letter for your significant other. We would benefit from hearing about their experiences as a caregiver. If you are willing, please provide them the encl osed letter requesting their participation in our study. I would be happy to talk with you more about the study. If you are interested in participating in this study or if you have any questions please contact me at the number listed below. Sincerely, S hannon Madore, M.A. Kristin Kilbourn, Ph.D., M.P.H. Clinical Psychology Doctoral Student Assistant Professor Office: 303 556 8567 University of Colorado Denver Cell: 970 389 5604 Department of Psychology Shannon.madore@ucdenver.edu Teri Simoneau, Ph.D. Director of Psychosocial Oncology Colorado Blood Cancer Institute

PAGE 140

129 Follow Up Recruitment Letter Patient Hello, A couple of weeks ago you received a letter regarding a research study for significant others of multiple myeloma patients. I am sending this letter in case the first letter did not reach you or in case your significant other is interested in participating, but has not yet responded. I am a doctoral student in the Clinical Psych ology program at University of Colorado Denver. I am presently working on a research study for my doctoral dissertation in partnership with Teri Simoneau, Ph.D., a psychologist, at the Colorado Blood Cancer Institute. You are receiving this letter becaus e you are a patient who has been diagnosed with multiple myeloma and have received treatment at CBCI. At times during your treatment, your significant other has served in a caregiver role. Caregivers often experience high levels of stress. The purpose of t his study is to evaluate levels of stress (both self reported stress and stress hormone levels) and quality of life when caring for a loved one with a experiences in order to continue to provide better care to family members/caregivers. Enclosed in this envelope you will find another letter for your significant other. We would benefit from hearing about their experiences as a caregiver. If you are willing, please pr ovide them the enclosed letter requesting their participation in our study. I would be happy to talk with you more about the study. If you are interested in participating in this study or if you have any questions please contact me at the number listed be low. Sincerely, Shannon Madore, M.A. Kristin Kilbourn, Ph.D., M.P.H. Clinical Psychology Doctoral Student Assistant Professor Office: 303 556 8567 University of Colorado Denver Cell: 970 389 5604 Department of Psychology Shannon.madore@ucdenver.edu Teri Simoneau, Ph.D. Director of Psychosocial Oncology Colorado Blood Cancer Institute

PAGE 141

130 A ppendix C: Initial Recruitment Letter Caregiver Hello, Thank you for taking your time to read this letter! I am a doctoral student in the Clinical Psychology program at University of Colorado Denver. I am presently working on a research study for my doctoral diss ertation in partnership with a psychologist at the Colorado Blood Cancer Institute, Teri Si moneau, Ph.D. You are receiving this letter because you are serving in a caregiving role for your significance other. Caregivers often experience high levels of stress. The purpose of this study is to evaluate levels of stress (both self reported stress a nd stress hormone levels) and quality of life when caring for a loved one with a chronic condition. We will use this provide better care to family members/caregivers Participation in this study entails a onetime visit to the CBCI clinic. During this visit you will 1) review and sign a consent form to participate in the study (10 20 minutes), 2) have your blood drawn by an appropriately trained CBCI staff member (15 m inutes), and 3) complete questionnaires about your caregiving experiences (25 30 minutes). This appointment will be scheduled at your convenience. I would be happy to talk with you more about the study. If you are interested in participating in this study or if you have any questions please contact me at the number listed below. I sincerely hope you consider being a part of this study. Sincerely, Shannon Madore, M.A. Kristin Kilbourn, Ph.D., M.P.H. Clinical Psychology Doctoral Student Assistant Professor Office: 303 556 8567 University of Colorado Denver Cell: 970 389 5604 Department of Psychology Shannon.madore@ucdenver.edu Teri Simoneau, Ph.D. Director of Psychosocial Oncology Colorado Blood Ca ncer Institute

PAGE 142

131 Follow Up Recruitment Letter Caregiver Hello, A couple of weeks ago you received a letter regarding a research study for significant others of multiple myeloma patients. I am sending this letter in case the first letter did not reach you or in case your significant other is interested in participating, but has not yet responded. I am a doctoral student in the Clinical Psychology program at University of Colorado Denver. I am presently working on a research study for my doctoral disser tation in partnership with Teri Simoneau, Ph.D., a psychologist, at the Colorado Blood Cancer Institute. You are receiving this letter because you are serving in a caregiving role for your significance other. Caregivers often experience high levels of stre ss. The purpose of this study is to evaluate levels of stress (both self reported stress and stress hormone levels) and quality of life when caring for a loved one with a chronic condition. We will use this important information to better understand caregi provide better care to family members/caregivers. Participation in this study entails a onetime visit to the CBCI clinic. During this visit you will 1) review and sign a consent form to participate in the study (10 20 minutes), 2) have your blood drawn by an appropriately trained CBCI staff member (15 minutes), and 3) complete questionnaires about your caregiving experiences (25 30 minutes). This appointment will be scheduled at your convenience. I would be happy to talk with you more about the study. If you are interested in participating in this study or if you have any questions please contact me at the number listed below. I sincerely hope you consider being a part of this study. Sincerely, Shannon Madore, M.A. Kristin Kilbourn, Ph.D., M.P.H. Clinical Psychology Doctoral Student Assistant Professor Office: 303 556 8567 University of Colorado Denver Cell: 970 389 5604 Department of Psychology Shannon.mado re@ucdenver.edu Teri Simoneau, Ph.D. Director of Psychosocial Oncology Colorado Blood Cancer Institute

PAGE 143

132 A ppendix D : Letter to Psychosocial Team to Request Recruitment Assistance Hello, For those of you who have not yet met me, my name is Shanno n Madore and I am a doctoral student in the Clinical Psychology program at University of Colorado Denver. I am presently working on a research study for my doctoral dissertation in partnership with Teri Simoneau, Ph.D. and Athena Baca Chieza, Psy.D. at CB CI. The purpose of this study is to evaluate levels of stress (both self reported stress and stress hormone levels) and quality of life in individuals who are in a romantic relationship with and caregiving for a multiple myeloma patient. As part of their participation, caregivers will be asked to do the following: 1) Review and sign a consent form to participate in the study (10 20 minutes) 2) Have 20ml of blood drawn by an appropriately trained CBCI staff member (15 minutes) study participation is not contingent on blood draws* 3) Complete a questionnaire about caregiving experiences (30 45 minutes) 4) Caregivers may also participate in an 8 week follow up survey (25 30 minutes) As part of this process I am asking for your assistance to recruit caregi vers of multiple myeloma patients when they are participating in the psychosocial intake. The in person recruitment process will entail you briefly describing the purpose of this study and providing participants an informational sheet with my contact infor mation. If individuals express interest, I will follow up with scheduling and data collection. I am planning on presenting more information about the specifics of the study at the Psychosocial Oncology Seminar on May 14. Please let me know if you have an y questions. I can be reached at 970 389 5604 or Shannon.madore@ucdenver.edu Sincerely, Shannon Madore, M.A.

PAGE 144

133 Appendix E : Caregiver Informational Flyer Caregiver patient relationship and stress response in multiple myeloma Caregiver Information Sheet Why is this study being done? Spousal caregivers in general are a distressed group and there is limited research examining the physical and mental health of couples, one of whom is diagn osed with multiple myeloma. This study plans to learn more about the relationships between caregiver distress, patient caregiver relationship characteristics and stress hormone levels in spousal caregivers as their loved one undergoes bone marrow transpla nt. Furthermore, this study aims to understand whether patient caregiver relationship quality impacts general health status. What will I be asked to do if I join this study? You will be asked to provide about 20 mL of blood only one time before your patien We will get blood by putting a needle into one of your veins and letting the blood flow into a glass tube. You may feel some pain when the needle goes into your vein. A day or two later, you may have a small bruise where the needle went under the skin. The samples will be used to measure chemicals that sometimes change with stress. You will be asked to fill out a survey about symptoms of stress you may be experiencing, how being a caregiver has affected your life, your general physical and mental well being, and your relationship with the patient. It will take about 30 45 minutes to complete. You may withdraw at any time during the study. If you would like more information or are interested in participating in this study, please cont act Shannon Madore at 970 389 5604, 303 556 8567 or Shannon.madore@UCdenver.edu

PAGE 145

134 A ppendix F : Pre Blood Draw Screen Name____________________________________ Date___________________ This cov er sheet will be removed to protect your privacy. Thank you for taking the time to complete these questions!

PAGE 146

135 1. Have you taken any medications this morning? If so please list ALL medications taken including over the coun ter or vitamins in the space provided below: __________________________________________________________________ 2. What is your height? __________ 3. What is your weight? __________ 4. If you are currently experiencing any type of acute illness, please circle belo w: Acne or skin infection Flu Allergies Food Poisoning Appendicitis Sinus infection Bronchitis Severe burn Cold Strep Throat Ear infection Venomous animal bite Fever Other: ___________________________ 5. If you are diagnosed with any type of chronic health condition, please circle below: Asthma Epilepsy Bronchiectasis Glaucoma Cancer Hemophilia Cardiac Disease Hepatitis Cardiomyopathy Hyperlipidemia Chronic obstructive pulmonary disorder Hypertensio n Chronic renal disease Hypothyroidism Coronary artery disease Multiple sclerosis Crohn's disease Parkinson's disease Diabetes mellitus types 1 & 2 Rheumatoid arthritis Dysrhythmias Systemic lupus erythematosus Sleep Apnea Other:______________ __________ 6. If you are diagnosed with any type of mental health conditions, please circle below: Bipolar disorder Post traumatic stress disorder Generalized anxiety disorder Major depressive disorder Panic disorder Substance abuse (alcohol nicoti ne, etc.)

PAGE 147

136 Other:_________________________ Schizophrenia 7. If you smoke cigarettes, did you have one this morning? Yes No 8. What was the approximate time since your last meal ? ________

PAGE 148

137 A ppendix G : Caregiver Psychosocial Surve y Name_____________________ _______________ Date___________________ This cover sheet will be removed to protect your privacy. Thank you for taking the time to complete these questions!

PAGE 149

138 Caregiver Assess ment Packet General Instructions Please take your time and answer all questions as best you can. If the responses included in this survey do not fit your answer exactly, please select the best answer for you. For some questions, you may have to write an answer. Please print your answer carefully so that we can read it. Most of the questions in this survey ask you to select only one answer. However, some questions allow you to select more than one answer. Please watch for these questions, which will ind Other questions in this survey will have two sections. Please be sure to answer both sections. Also, please remember that there are no right or wrong answers to the questions in this survey. The best answ ers you can give are the ones that are most correct for you. Thank you for the time and attention you are giving to this survey. When you complete this survey, please place in the stamped envelope provided for you and mail it back to us. I. Demographics and Health Behaviors Remember there are no right or wrong answers, only your best answer. 1. What is your gender? 1. female 2. male 2. What is your age ? ____ 3. What is your racial/ethnic background? 1. White (not of Hispanic origin) 2. Hisp anic 3. African American/Black (not of Hispanic origin) 4. Asian/Pacific Islander 5. Native American 6. Multi Ethnic (please specify) ___________________ 7. Other (please specify ) ___________________ 4. What is your current relationship status ? 1. Married 2. Legally separated 3. Committed relationship (partner opposite sex) 4. Committed relationship (partner same sex) 5. Other ______________ 5. How long have you been with your intimate partner? ___months and/or ____ years 6. I f you are married, for how long have you been married ? ____ __________ 7. Do you have children?

PAGE 150

139 1. Y es If yes how many? ______ What are their ages ? _________ 2. N o Are they living with you? 1. Yes 2. No 8. How many years of education have you completed? 1. Less than high school 2. Some high school 3. High school graduate 4. Some college 5. Associate degree 6. College degree 7. Post graduate degree 9. What is your current employment status? 1. Employed full time (including self e mployed) 2. Employed part time (including self employed ) 3. Full time homemaker 4. Full time or Part time volunteer 5. Full time student 6. On temporary medical leave/disability 7. Retired 8. Unemployed 9. Permanently unable to work 10. W hat is your occupation? _________________________ 11. Approximately what is your annual household income currently? 1. 0 $25,000 2. $25,001 $50,000 3. $50,001 $75,000 4. $75,001 $100,000 5. $100,000+ 12. What was your approximate annual house hold income prior to your partner being diagnosed with cancer? 1. 0 $25,000 2. $25,001 $50,000 3. $50,001 $75,000 4. $75,001 $100,000 5. $100,000+ 13. When was your partner diagnosed with multiple myeloma? Please provide the date as accurately as possible. _________________________________________________________

PAGE 151

140 ______________________________________________________________________ ______________________________________________________________________ ____ __________________________________________________________________ 15. What, if any, symptoms did your partner experience prior to being diagnosed? ______________________________________________________________________ ____________________________________ __________________________________ 16. What was his/her exact diagnosis ( e.g. stage and type of multiple myeloma)? ______________________________________________________________________ 17. Since his/her diagnosis, what treatments has your partner underg one? ______________________________________________________________________ ______________________________________________________________________ 18. Did your partner have experience any side effects from these treatments? If so, please describe. ________ ______________________________________________________________ ______________________________________________________________________ _________________________________________________ _____________________ ______________________________________________________________________ 20. Is your partner currently prescribed corticosteroids ? 1. Yes 2. No 21. If yes, for how long has he/she been taking this medication? ________years __________ months 22. What is your perception of the severity of his/her diagnosis? 1. Very severe 2. Somewhat severe 3. Mildly severe 4. Not at all severe 23. Does your partner experience significant physical pain? 1. All the time 2. Most of the time 3. Some of the time 4. None of the time 24. Caregiving can be defined as the act of caring for someone who is physically ill. Oftentimes caregivers have many different roles including helping or supporting loved ones with physical, emotional and practical needs How long have you been a caregiver? ________years___________months

PAGE 152

141 25. Sometimes caregiving involves providing physical care including but not limited to helping a loved one eat, bathe, dress, stand up, etc. Are these activities you help your partner wi th? 1. Yes 2. No 26. If so, what is the average amount of time you spend each day doing these types of activites?__________hrs____________minutes 27. Approximately how long have you been helping your partner with these activites? ________years___________ months listening to and talking with your partner about his/her multiple myeloma experiences, etc. Are these activities you help your partner with? 1. Yes 2. No 29. If so, what is the average amount of time you spend each day doing these types of activites?__________hrs____________minutes 30. Approximately how long have you been helping your partner with these activites? ________years___________months 31. Some times caregiving involves practical care including but not limited to managing medication, providing transportation, doing household chores, providing financial support. Are these activities you help your partner with? 1. Yes 2. No 32. If so, what is the average amount of time you spend each day doing these types of activites?__________hrs____________minutes 33. Approximately how long have you been helping your partner with these activites? ________years___________months 34. Do you feel you are able to p rovide an adequate level of care for your partner? 1. All of the time 2. Most of the time 3. Some of the time 4. None of the time

PAGE 153

142 Now we would like to ask you some general questions about your health. 35. How would you currently describe y our health? 1. Excellent 2. Good 3. Fair 4. Poor 36. In general, how would you describe your diet? 1. Excellent 2. Good 3. Fair 4. Poor 37. How many meals do you eat each day? ________ 38. On a daily basis how many servings do you ha ve of the following food items? Leafy vegetables _____________ Vegetables _______________ Fruit _________________ Milk/Dairy Products ____________ Eggs _________________ Meat ________________ Whole Grains _______________ 39. Do you engage in any recreation activities? 1. yes 2. no 40. If yes, what are they? ______________________________________________________________________ 41. How many days per week do you exercise or perform moderate activity (brisk walking, bicycling, vacuuming, gar dening, causing some increase in breathing or heart rate) for more than 10 minutes? __________ 42. On average, what is the total time per week that you exercising or engage in moderate activity? ______ hours ________ minutes 43. Do you drink caffei nated beverages? List type and amount per day (eg. 1 double espresso, 2 cokes) ______________________________________________________________________ ______ 44. Do you drink alcohol? 1. yes 2. no

PAGE 154

143 45. If yes, how many drinks per day _____ or week_ _________? 46. Do you use tobacco? 1. yes 2. no 47. If yes, how? 1. Cigarettes 2. Cigar 3. Pipe 4. Chewing tobacco 5. Other __________________ 48. How much do you use tobacco per day __________ or week_________? For how many y ears?_____

PAGE 155

144

PAGE 156

145

PAGE 157

146

PAGE 158

147

PAGE 159

148 II. Psychosocial Measures

PAGE 160

149

PAGE 161

150 Positive and Negative Affect Scale This scale consists of a number of words that describe different feelings and emotions. Read each item and then mark the appro priate answer in the space next to that word. Indicate to what extent you feel this way right now, that is, at the present moment. Use the followi ng scale to record your answers : 1 2 3 4 5 Very slightly or not at all A little Moderately Quite a bit Extr emely _____ Interested _____ Irritable _____ Distressed _____ Alert _____ Excited _____ Ashamed _____ Upset _____ Inspired _____ Strong _____ Nervous _____ Guilty _____ Determined _____ Scared _____ Attentive _____ H ostile _____ Jittery _____ Enthusiastic _____ Active _____ Proud _____ Afraid

PAGE 162

151 Center for Epidemiological Studies Depression Scale Circle the number of each statement which best describes how often you felt o r behaved this way DURING THE PAST WEEK Rarely or none of the time (less than 1 day) Some or a little of the time (1 2 days) Occasionally or a moderate amount of the time (3 4 days) Most or all of the time (5 7 days) During the past week: 0 1 2 3 1) I was bothered by things 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 a s 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 19) I felt that people disliked me 0 1 2 3 0 1 2 3

PAGE 163

152 Impact of Events Scale Below is a list of difficulties people sometimes have during stressful life events. Please read each item, and then indicate how distressing each difficulty has been for you DURING THE PAST SEVEN DAYS how much were you distressed or bothered by these difficulties? Not at all A little bit Moderately Quite a bit Extremely Any reminder bri ngs back illness 0 1 2 3 4 I have trouble staying asleep 0 1 2 3 4 Other things keep making illness 0 1 2 3 4 I feel irritable and angry 0 1 2 3 4 I avoid letting myself get upset when I think abo ut reminded of it 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 I stay away from reminders 0 1 2 3 4 illness pop into my mind 0 1 2 3 4 I am jumpy and easily startled 0 1 2 3 4 I try not to think about my 0 1 2 3 4 I am aware that I have a lot of feelings about my deal with them 0 1 2 3 4

PAGE 164

153 My feelings about my numb 0 1 2 3 4 I find myself acting or feeling as though I was back at the time of diagnosis 0 1 2 3 4 I have trouble falling asleep 0 1 2 3 4 I have waves of strong feeli illness 0 1 2 3 4 I try to remove it from my memory 0 1 2 3 4 I have trouble concentrating 0 1 2 3 4 illness cause me to have physical reactions, such as sweating, trouble breathing, nausea, or a pounding heart 0 1 2 3 4 I have dreams about my 0 1 2 3 4 I feel watchful or on guard 0 1 2 3 4 I try not to talk about my 0 1 2 3 4

PAGE 165

154 Social Relationship Inventory Our relationships with other people may or may not have both positive and upsetting aspects. For the following questionnaire, please answer them in regards to your thoughts and feelings about your partner Remember that your responses are confidential and that there are no right or wrong answers. I AM RATING MY (circle one): Husband Wife Partner HOW LONG HAVE YOU BEEN IN YOUR RELATIONSHIP? ___ Year(s) ___ Month(s) HOW IMPORTANT IS YOUR PARTNER TO YOU (circle one number)? 1 2 3 4 5 6 Not at all important A l ittle important Somewhat important Moderately important Very important Extremely important For the next section, when asked to rate the extent YOUR PARTNER is HELPFUL or POSITIVE, you should ignore any upsetting aspects of the relationship. When asked to rate the extent your PARTNER is UPSETTING, you should ignore any helpful or positive aspects of your relationship. Please circle the ONE best answer. WHEN YOU NEED SUPPORT, SUCH AS ADVICE, UNDERSTANDING, OR A FAVOR... NOT AT ALL A LITTLE SOMEWHAT MOD ERATELY VERY EXTRE MELY HOW POSITIVE IS YOUR PARTNER? 1 2 3 4 5 6 HOW UPSETTING IS YOUR PARTNER? 1 2 3 4 5 6 HOW MIXED OR CONFLICTED ARE YOUR THOUGHTS AND FEELINGS TOWARDS YOUR PARTNER? 1 2 3 4 5 6 HOW UNPREDICTABLE IS YOUR PARTNER? 1 2 3 4 5 6

PAGE 166

155 Qualtity of Marriage Index Please read the following statements and answer them according to your feelings about your significant relationship. 1) We have a good relationship Strongly agree Agree Agree somewhat Disagree somewhat Disagree Strongly disag ree 2) My relationship with my partner is very stable Strongly agree Agree Agree somewhat Disagree somewhat Disagree Strongly disagree 3) Our relationship is strong Strongly agree Agree Agree somewhat Disagree somewhat Disagree Strongly disagree 4) My relation ship with my partner makes me happy Strongly agree Agree Agree somewhat Disagree somewhat Disagree Strongly disagree 5) I really feel like part of a team with my partner Strongly agree Agree Agree somewhat Disagree somewhat Disagree Strongly disagree 6) Overa ll, everything considered, how happy are you in your relationship Very Happy Quite Happy Somewhat Happy Neutral Somewhat Unhappy Quite Unhappy Very Unhappy

PAGE 167

156 Measurment of Attachment Quality Respond to each of the following statements by expressing how much you agree with it (if you do generally agree) or how much you disagree with it (if you generally disagree). Make all your responses on the answer sheet only. Do not leave any items blank. Please be as accurate as you can be throughout, and try esp ecially hard not to let your answer to any one item influence your answer to any other item. Treat each one as though it is completely unrelated to the others. There are no right or wrong answers, you are simply to express your own personal feelings and opinions. Choose from these response options: 1 = I DISagree with the statement a lot 2 = I DISagree with the statement a little 3 = I agree with the statement a little 4 = I agree with the statement a lot 1. When I'm close to my partner, it gives me a sense of comfort about life in general. ___ 2. I often worry that my partner doesn't really love me. ___ 3. I have trouble getting my partner to be as close as I want them to be. ___ 4. I find it easy to be close to my partner. ___ 5. I often w orry my partner will not want to stay with me. ___ 6. My partner wants me to be more intimate than I feel comfortable being. ___ 7. It feels relaxing and good to be close to my partner. ___ 8. I am very comfortable being close to my partner. ___ 9. ___ 10. My desire to merge sometimes scares my partner away. ___ 11. I prefer not to be too close to my partner. ___ 12. I find my partner is reluctant to get as close as I would like. ___ 13. I get un comfortable when my partner wants to be very close. ___ 14. Being close to my partner gives me a source of strength for other activities. ___ You have reached the end of the survey. Thank you for your time!