Upload
others
View
5
Download
1
Embed Size (px)
Citation preview
Circadian Factors in Coping with Chronic Stress
by
Daniela Bellicoso
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy
Department of Psychology University of Toronto
© Copyright by Daniela Bellicoso 2017
ii
Circadian Factors in Coping with Chronic Stress
Daniela Bellicoso
Doctor of Philosophy
Department of Psychology University of Toronto
2017
Abstract
Circadian rhythms, cyclic changes that repeat approximately once every 24 hours, regulate daily
temporal processes of physiology and behaviour in all living organisms. Rhythmicity is adaptive,
providing advantages for survival, growth, and reproduction. For human beings, this includes
rhythmic regulation of mental, physical, and emotional responses to stressors. In this dissertation,
we addressed how human beings facing major stressors such as chronic disease (in this case,
cancer), are influenced by circadian timing, specifically chronotype and sleep quality. A broad
epidemiological study was conducted among individuals in different roles (patients, medical
staff, and familial caregivers) facing a common stressor to assesses their ability to cope with the
situation. Breast cancer patients provided real-time reports of their coping across the day.
Patients and familial caregivers completed retrospective average ratings for their coping across
the day across treatment, along with details on use of coping behaviours. Oncology staff
provided ratings of their burnout, and Professional Quality of Life (ProQoL). These data were all
assessed in relation to chronotype and sleep quality, and to an extent, in relation to personality.
Chronotype and sleep quality influenced coping within each group, but their impacts were not
correlated. Nonetheless, working at chronotypically optimal times improved ProQoL. ‘In the
moment’ coping ratings from patients reflected their chronotype, as did their recalled coping.
‘Openness’ was positively linked with ProQoL among staff, and with an engagement coping
iii
style among patients and caregivers. The data reflect an influence of circadian timing on the
expression of coping responses during chronic stress. A better understanding of changes in
coping ability as they relate to one’s innate rhythms will allow for the development of a
cognitive and emotion-based chronotherapy regime intended to maximize proactive coping
among individuals facing chronic stress, such as providing care or undergoing treatment for
cancer.
iv
Acknowledgments
Without the help of the following people, this PhD would not have been possible – I owe each of
you many thanks for the kindness you showed me.
My Primary Supervisor, Dr. Martin Ralph: Thank you for taking a chance on me and allowing
me to begin conducting research in your lab as an undergraduate student, over 10 years ago.
Your guidance, patience, and kindness through the years have allowed me the freedom to explore
my own ideas and grow as a researcher in my own way. Thank you ever so much.
Dr. Marg Fitch: Thank you for lending your expertise in the field of coping and adaptation to
illness. Your invaluable guidance on how to approach the various topics explored in this research
are greatly appreciated. It was a pleasure to work with you, and get to know you over the years.
Dr. Maureen Trudeau: There are no words to express my gratitude for the guidance, kindness,
and support you’ve shown me. Without you, this research would not have been possible. It was a
pleasure to learn from you. Thank you for the innumerable and invaluable opportunities for
research, learning, and personal growth that you provided.
The kind staff at Sunnybrook Hospital – Odette Cancer Centre & Princess Margaret Hospital:
Thank you for taking the time out of your busy schedules to offer suggestions on topics and
questions for research, and to identify participants for my studies. Your participation in my
studies and taking the time to sharing some of your own struggles relating to facing chronic
stress were invaluable. I hope that my research, even if only in a small way, helps to make your
job slightly more manageable.
The Breast Cancer Patients and their Caregivers: Your willingness to partake in research, without
any personal reward during your most difficult days is a testament to the good nature and kind
spirit you each possess. Thank you for taking the time to chat with me, share your stories, and
open-up to a stranger about some of your own daily struggles. Thank you for your selflessness.
My husband Matthew, and my sister Elisa: You’ve both given me courage in so many ways. The
love and encouragement you both provided were invaluable, and greatly appreciated.
v
My grandparents, who came to this wonderful country over 60 years ago, in search of a better
life for their families: The life lessons of hard work and determination that you each taught me
helped me to achieve this degree. Thank you for the sacrifices you made, so I could have so
many opportunities and such a wonderful education.
And finally, my parents: You each instilled a love of learning, and a strong work ethic from early
on in my life. You have been my biggest cheerleaders, never failing to give encouragement, and
let me know how much you believed in me. The experiences and opportunities you provided
from early on opened my eyes to so many things around me, and piqued my curiosity to always
learn more and understand why things are a certain way. While you’ve always told me how
proud I make you, I’m proud to be the daughter of two such wonderful people who allowed me
to explore, to be me, and to make my own path in the world.
vi
Table of Contents
Acknowledgments ........................................................................................................................ iv
Table of Contents ......................................................................................................................... vi
List of Tables ................................................................................................................................ xi
List of Figures ............................................................................................................................. xiv
Chapter 1 ....................................................................................................................................... 1
General Introduction ............................................................................................................ 1
1.1 Context, Hypotheses, and Rationale .........................................................................................1 1.2 Coping and Survival Strategies .................................................................................................3
1.2.1 Coping .....................................................................................................................................3 1.2.2 Brief-COPE Questionnaire ...................................................................................................3 1.2.3 Circadian Rhythms and Coping Strategies .........................................................................6 1.2.4 Sleep and Coping ....................................................................................................................7 1.2.5 Depression, Stress, and Coping .............................................................................................9 1.2.6 Burnout, Stress, and Coping in Caregivers .......................................................................10 1.2.7 Personality and Coping .......................................................................................................11
1.3 Cancer as a Chronic Stress ......................................................................................................13 1.3.1 Emotional Distress in Patients ............................................................................................14
1.3.1.1 Why Choose Breast Cancer Patients .........................................................................................14 1.3.1.2 Specific Distress Involving Genetic Issues ................................................................................15
1.3.2 Emotional Distress in Caregivers .......................................................................................15 1.3.2.1 Specific Distress Involving Partners ..........................................................................................15 1.3.2.2 Specific Distress Involving Oncology Staff ...............................................................................16
1.4 Rhythmicity ...............................................................................................................................16 1.4.1 Biological Clocks in Nature .................................................................................................16 1.4.2 Biological Clocks in Mammals ............................................................................................17 1.4.3 Biological Clocks in Human Beings ....................................................................................18 1.4.4 Cognition, Emotion, and Circadian Rhythms ...................................................................18 1.4.5 Chronotype ...........................................................................................................................18
1.4.5.1 Genetic Basis of Chronotype ......................................................................................................19 1.4.5.2 Questionnaires ............................................................................................................................19
vii
1.4.5.3 Performance Variances ..............................................................................................................20 1.4.5.4 Emotionality ................................................................................................................................21
1.5 Sleep ...........................................................................................................................................21 1.5.1 Two-Process Model of Sleep ................................................................................................22 1.5.2 The Functions of Sleep .........................................................................................................22 1.5.3 The Impact of Sleep and Sleep Loss ...................................................................................23
1.5.3.1 Sleep, Cognition, and Memory ...................................................................................................23 1.5.3.2 Polysomnography and Actigraphy .............................................................................................25 1.5.3.3 Sleep Quality ...............................................................................................................................25
1.5.3.3.1 Cancer ....................................................................................................................................26 1.5.3.3.2 Breast Cancer Patients ...........................................................................................................27 1.5.3.3.3 Caregivers ..............................................................................................................................28 1.5.3.3.4 Oncology Staff ........................................................................................................................29
1.6 Breast Cancer Background .....................................................................................................29 1.6.1 Cancer Biology .....................................................................................................................29 1.6.2 Breast Cancer Biology .........................................................................................................30 1.6.3 Genetics and Mutations .......................................................................................................31 1.6.4 Incidence ...............................................................................................................................31 1.6.5 Diagnosis and Treatment .....................................................................................................32
1.6.5.1 Staging ........................................................................................................................................32 1.6.5.2 Procedures ..................................................................................................................................34 1.6.5.3 Other Tumour Characteristics ...................................................................................................34 1.6.5.4 Treatment ....................................................................................................................................36
1.6.5.4.1 Localized .................................................................................................................................36 1.6.5.4.2 Systemic ..................................................................................................................................37
1.6.5.5 Prognosis ....................................................................................................................................39
Chapter 2 ..................................................................................................................................... 40
General Methods ................................................................................................................. 40
2.1 Procedures .................................................................................................................................40 2.1.1 Nursing Study .......................................................................................................................40 2.1.2 Hospital Staff ........................................................................................................................40 2.1.3 Patients and Caregivers .......................................................................................................41
2.2 Instruments ...............................................................................................................................41 2.2.1 General Questionnaires .......................................................................................................41 2.2.2 Group Specific Questionnaires ...........................................................................................42
2.2.2.1 Nursing Staff ..............................................................................................................................42
viii
2.2.2.2 Oncology Staff ............................................................................................................................42 2.2.2.3 Patients and Caregivers ..............................................................................................................43
2.2.2.3.1 Questionnaire Package ..........................................................................................................44 2.3 Statistics .....................................................................................................................................45
Chapter 3 ..................................................................................................................................... 46
Burnout Among Oncology Nurses: Influence of Chronotype and Sleep Quality ......... 46
3.1 Abstract .....................................................................................................................................46 3.2 Introduction ..............................................................................................................................47 3.3 Materials and Methods ............................................................................................................49
3.3.1 Participants ...........................................................................................................................49 3.3.2 Procedure ..............................................................................................................................50 3.3.3 Measures ...............................................................................................................................50 3.3.4 Statistical Analysis ...............................................................................................................51
3.4 Results ........................................................................................................................................51 3.5 Discussion ..................................................................................................................................61
3.5.1 Limitations ............................................................................................................................64 3.6 Conclusions ...............................................................................................................................65
Chapter 4 ..................................................................................................................................... 66
Chronobiological Factors for Compassion Satisfaction and Fatigue Among
Ambulatory Oncology Caregivers ............................................................................................. 66
4.1 Abstract .....................................................................................................................................66 4.2 Introduction ..............................................................................................................................67 4.3 Materials and Methods ............................................................................................................69
4.3.1 Participants and Procedures ...............................................................................................69 4.3.2 Measures ...............................................................................................................................70 4.3.3 Statistical Analysis ...............................................................................................................70
4.4 Results ........................................................................................................................................71 4.4.1 Descriptive Group Statistics ................................................................................................71 4.4.2 Correlation Analysis ............................................................................................................73
4.5 Discussion ..................................................................................................................................80 4.5.1 Chronotype and ProQoL .....................................................................................................80 4.5.2 Sleep and ProQoL ................................................................................................................81 4.5.3 Personality and ProQoL ......................................................................................................82
4.5.3.1 Agreeableness .............................................................................................................................82
ix
4.5.3.2 Emotional Stability .....................................................................................................................82 4.5.3.3 Openness .....................................................................................................................................83 4.5.3.4 Conscientiousness and Extraversion .........................................................................................83
4.5.4 Job Satisfaction and ProQoL ..............................................................................................83 4.5.5 Limitations ............................................................................................................................86
4.6 Conclusion .................................................................................................................................86
Chapter 5 ..................................................................................................................................... 91
Primary Circadian Impacts on Patients, Caregivers, and Dyads .................................. 91
5.1 Abstract .....................................................................................................................................91 5.2 Introduction ..............................................................................................................................92 5.3 Materials and Methods ............................................................................................................94
5.3.1 Participants and Procedures ...............................................................................................94 5.3.2 Measures ...............................................................................................................................95 5.3.3 Statistical Analysis ...............................................................................................................95
5.4 Results ........................................................................................................................................96 5.4.1 Descriptive Statistics ............................................................................................................96
5.4.1.1 Overall Demographic Comparisons ..........................................................................................96 5.4.1.2 Morningness-Eveningness Distribution ....................................................................................97 5.4.1.3 Sleep Quality Comparison ..........................................................................................................97 5.4.1.4 UTIME Performance Results ....................................................................................................97
5.4.2 Analysis 1: Total Patient Population Analysis ................................................................105 5.4.2.1 Multifactorial Correlation Analysis .........................................................................................105 5.4.2.2 Global Correlations Among Patients’ MEQ, PSQI and UTIME data ...................................105 5.4.2.3 Total Patient Population Mixed Measures ANOVAs .............................................................107
5.4.3 Analysis 2: Total Caregiver Population Analysis ............................................................111 5.4.3.1 Global Correlations Among Caregivers’ MEQ, PSQI and UTIME data ..............................111 5.4.3.2 Caregiver Population Mixed Measures ANOVAs ...................................................................113
5.4.4 Analysis 3: Matched Patient and Caregiver Population Analysis .................................116 5.4.4.1 Global Associations and Differences Among Matched Patient and Caregiver MEQ, PSQI
and UTIME data ..........................................................................................................................................116 5.4.4.2 Matched Patient and Caregiver Group Mixed Measures ANOVAs .......................................118
5.5 Discussion ................................................................................................................................123 5.5.1 On the Independent Impacts of Sleep and Chronotype .................................................123 5.5.2 Patient Cognitive Function and Emotional Regulation ..................................................126
5.5.2.1 Chronotype-UTIME Correlations ...........................................................................................126 5.5.2.2 Multivariate Analysis of UTIME Correlations in the Patient Population .............................127
x
5.5.3 Caregiver Cognitive Function and Emotional Regulation .............................................131 5.5.4 Matched Patient and Caregiver Cognitive Function and Emotional Regulation ........132 5.5.5 Limitations ..........................................................................................................................135
5.6 Conclusion ...............................................................................................................................136
Chapter 6 ................................................................................................................................... 138
Coping Behaviour in Chronic Disease ............................................................................ 138
6.1 Abstract ...................................................................................................................................138 6.2 Introduction ............................................................................................................................139 6.3 Materials and Methods ..........................................................................................................142
6.3.1 Participants and Procedures .............................................................................................142 6.3.2 Measures .............................................................................................................................143 6.3.3 Statistics ..............................................................................................................................144
6.4 Results ......................................................................................................................................144 6.5 Discussion ................................................................................................................................178
6.5.1 Personal Coping Assessment .............................................................................................178 6.5.2 Coping Behaviours: Changes in Coping Scores ..............................................................180 6.5.3 In the Moment Coping vs. Coping Behaviours ...............................................................189 6.5.4 Limitations ..........................................................................................................................190
6.6 Conclusion ...............................................................................................................................192
Chapter 7 ................................................................................................................................... 194
Conclusion ......................................................................................................................... 194
7.1 Chronotype ..............................................................................................................................195 7.2 Sleep Quality ...........................................................................................................................197 7.3 Personality ...............................................................................................................................199 7.4 Future Directions ....................................................................................................................200
Bibliography .............................................................................................................................. 202
xi
List of Tables
Table 3.1 Demographic characteristics and questionnaire response ratings of participants…………………………………………………………..
52
Table 3.2 Comparison of burnout ratings between respondents with good and bad sleep quality and between MEQ types ………………………….
54
Table 3.3 Analysis of bivariate correlations for participants’ questionnaire response ratings ……………………………………………………..
56
Table 3.4a Summary of hierarchical multivariate regression analysis for variables predicting personal burnout among oncology nurses (N = 64) ………………………….………………………………………..
58
Table 3.4b Summary of hierarchical multivariate regression analysis for variables predicting work related burnout among oncology nurses (N = 64) ………………………….…………………………………..
59
Table 3.4c Summary of hierarchical multivariate regression analysis for variables predicting client related burnout among oncology nurses (N = 64) ………………………….………………………………….
60
Table 4.1 Means and frequencies of participant demographic and questionnaire
data ………………………….………………………………………. 72
Table 4.2 Non-parametric correlations between ProQoL domains (columns) and covariates (rows) ………………………….…………………….
75
Table 4.3 Final models in backwards multiple regression with professional quality of life (CS, BO, and STS) as dependent variables, and significant values* (as tested by univariate regression) for MEQ, PSQI, JSS and TIPI as independent variables ……………………….
77
Table 4.4 Models in backwards multiple regression with professional quality of life (CS, BO, and STS) as dependent variables, and significant values* (as tested by univariate regression) for MEQ, PSQI, TIPI and working on multiple vs. single cancer sites as independent variables ………………………….………………………………….
79
Table S4.1 Spearman correlations (rs) between continuous variables, and point biserial correlations for categorical variables (demographics, chronotype and sleep quality) ……………………………………….
88
Table S4.2 Spearman correlations (rs) between continuous variables, and point biserial correlations for categorical variables related to job satisfaction ………………………….……………………………….
89
Table S4.3 Spearman correlations (rs) between continuous variables, and point biserial correlations for categorical variables related to personality and professional quality of life ……………………………………...
90
xii
Table 5.1 Means and frequencies of demographic and questionnaire data for patients and caregivers ………………………………………………
99
Table 5.2 Means and frequencies for chronotype and sleep quality questionnaires for patients and caregivers …………………………..
100
Table 5.3 Spearman correlations for chronotype versus sleep quality and UTIME among patients ……………………………………………..
106
Table 5.4 Mixed ANOVAs on patient sleep quality and UTIME: main effects and interactions ………………………….…………………………..
110
Table 5.5 Spearman correlations for caregiver chronotype versus sleep quality and UTIME ………………………….………………………………
112
Table 5.6 Mixed ANOVAs on caregiver sleep quality and UTIME: main effects and interactions ……………………………………………...
115
Table 5.7 Median and Mann Whitney U significance values comparing MEQ, PSQI, and UTIME response scores between patients and their caregivers ………………………….………………………………...
117
Table 5.8 Mixed ANOVAs on patient AND caregiver sleep quality and UTIME: main effects and interactions ……………………………...
122
Table 6.1 Descriptive data for patient coping logs ……………………………. 146 Table 6.2 Median and Friedman test data for changes in raw patient coping
log scores across treatment, with Wilcoxon signed-rank tests with Bonferroni correction applied for differences in raw patient coping log scores across treatment ………………………………………….
148
Table 6.3 Mean ± standard deviation and paired samples t-test data for pre- and post-chemotherapy coping rating comparisons ………………...
150
Table 6.4 Pearson correlations between MEQ and PSQI and patients’ coping log UTIME scores …………………………………………………...
152
Table 6.5 Descriptive data for Brief COPE are mean ± standard deviation …... 154 Table 6.6a Median and Friedman test data for changes in Brief COPE scores
across treatment among patients and caregivers ……………………. 157
Table 6.6b Wilcoxon signed-rank tests with Bonferroni correction applied for differences in Brief COPE scores across treatment …………………
158
Table 6.7 Summary of Multiple Regression Analyses for Brief-COPE Scores Across Treatment, assessing the predictive value of participant role, chronotype, and sleep quality ……………………………………….
163
Table 6.8 Descriptive data and independent samples t-test for BFAS between men and women. Descriptive data are mean ± standard deviation, for patients and caregivers …………………………………………..
166
xiii
Table 6.9 Summary of Multiple Regression Analyses for Brief-COPE Scores Across Treatment, assessing the predictive value of participant role, chronotype, sleep quality and personality …………………………..
172
xiv
List of Figures
Figure 5.1a UTIME scores across treatment for the total patient group ………… 101 Figure 5.1b UTIME scores across treatment for patients without a caregiver
involved in the study ………………………….…………………….. 102
Figure 5.1c UTIME scores across treatment for patients with a caregiver involved in the study ………………………….……………………..
103
Figure 5.1d UTIME scores across treatment for caregivers …………………….. 104
1
Chapter 1!
! General Introduction
1.1! Context, Hypotheses, and Rationale
Circadian rhythms, cyclic changes that repeat approximately once every 24 hours, regulate daily
temporal processes of biology and physiology in all living organisms. Rhythmicity is expressed
in processes at all levels of biological organization. At a molecular level, cell division and
replication are known to be rhythmic. At a higher level, cognition, physical functioning, and
emotionality have all been linked to a rhythmic preference for morning (M) versus evening (E)
performance, known as chronotype. Our biology, physiology, and behaviour are governed by
clocks. In this same vein, previous studies have explored the impact of rhythmicity on various
domains of health. Different findings in areas such as cardiology (e.g., Portaluppi et al., 2012),
musco-skeletal function (e.g., Riley & Esser, 2017), gynecology/obstetrics (e.g., Dogru et al.,
2016), and mental health (e.g., Hasler, Allen, Sbarra, Bootzin, & Bernert, 2010) support evidence
for health and rhythms.
Coping strategies, specifically, coping with chronic trauma that accompany health issues, has not
been examined in depth in relation to rhythmicity. Very little is known about how human beings
might respond at different times of day to a major stressor such as chronic disease, along with
other environmental factors that will influence an individual’s behaviour.
This dissertation focuses on how human beings facing major stressors (specifically, chronic
disease) are influenced by circadian timing. Based on existing information on rhythmicity of
emotional responses (e.g., Costa-Martins et al., 2016; Lenaert, Barry, Schruers, Vervliet, &
Hermans, 2016), a person’s ability to cope with not only the chronic disease but the other
stressors in their environment might vary according to the timing of their circadian cycle. Given
that research has suggested that how an individual deals with their disease or stressors can
impact outcome (e.g., Demytteraera et al., 1998; Shehmar & Gupta, 2010) it is important to
understand if and how these abilities might vary through the day. Furthermore, stress can
contribute to perpetuating a disease and disrupting circadian rhythms (McEwen & Karatsoreos,
2015). We should recognize that if this is the case, then not only will patients be affected, but
2
also everyone involved with the patient and going through the highly stressful situation will also
be influenced. It is reasonable to predict that the stress that occurs due to the disease itself,
together with the stresses felt by patients and caregivers will have a mix of influence on coping.
We initiated a large scale epidemiological study of a group of individuals facing the same
stressor (cancer), but each in a different role (patient, medical staff, familial caregiver), looking
at their ability to cope with the situation. We examined a number of important variables believed
to influence coping behaviour and memory for coping, by conducting various cross-sectional and
longitudinal studies. Using surveys and data logs, we collected measures of participants’
chronotype, sleep quality, personality, and coping. To examine chronotype and sleep quality, this
requires an understanding of the rhythmic processes, starting with the generation of rhythms by
biological clocks, and the expression of that timing in the regulation of behaviour. It requires
also an understanding of the individual and the disease itself. It also involves an understanding
of sleep stages and implications for good versus poor quality sleep. We narrowed down the focus
to address a group of breast cancer patients where the disease itself is relatively well defined, as
is treatment for the disease. The prognosis is also relatively well defined.
Given that circadian rhythms influence cognitive function and emotional regulation throughout
the day in the general population, (e.g., Blatter & Cajochen, 2006; Ottoni, Antoniolli & Lara,
2012) it is important to understand how these rhythms fluctuate among both cancer patients and
caregivers faced with a chronic daily stressor such as cancer. While it is understood that
disturbed rhythms can perpetuate poor health, less is known in general about circadian rhythms
and emotionality and emotional responding, including behaviour and memory for behaviour. An
understanding of emotionality and rhythmicity is required. It was hypothesized that rhythmicity
would impact individuals’ responses to chronic stress, but the influence of circadian rhythms
would vary depending on one’s role, and whether coping ratings were retrospective or in the
moment. A better understanding and documentation of changes in coping ability as they relate to
one’s innate rhythms will allow for the development of a cognitive and emotion-based
chronotherapy regime intended to maximize proactive coping among cancer patients and their
caregivers, both in the hospital and the home.
3
1.2! Coping and Survival Strategies
1.2.1! Coping
Coping refers to those “constantly changing cognitive and behavioral efforts to manage specific
external and internal demands that are appraised as taxing or exceeding the resources of the
person” (Lazarus & Folkman, 1984, p. 141). These taxing demands that exceed one’s personal
resources are known as stress. Stress is a real or perceived interruption to the homeostasis of
one’s physical state or mental well-being. Stress can result from a range of positive and negative
demands. For example, the stress of organizing a large event, or the physical stress that comes
from exercise both have the potential to exceed one’s mental or physical resources, respectively,
to deal with the situation. However, the body’s response to these self-sought out tasks can mimic
the stress response seen when responding to a negative stressor (National Research Council,
2008). Individuals are each taxed differently by various demands, and will not necessarily
respond the same way to a particular stressor. The individual efforts people perform to cope with
stress affects their physical, psychological, and social well-being (Folkman & Lazarus, 1980). As
such, it is important to understand how a specific stressor can elicit various coping responses
from different individuals, and the effect these efforts will have on well-being.
1.2.2! Brief-COPE Questionnaire
The cognitive and behavioural efforts an individual may use to cope or deal with the stress of a
situation can range from healthy to negative, yet it can be difficult to clearly distinguish which
specific efforts fall into either category. The Ways of Coping Questionnaire, developed by
Folkman and Lazarus (1980), suggests certain efforts can be categorized into problem or emotion
focused coping styles. Problem focused coping strategies are intended for problem solving or
performing some action to change the source of the stress. Emotion focused coping efforts are
intended to manage or reduce the emotional distress being created by or associated with the
stressor. It is important to note that of the two techniques, neither is necessarily more or less
positive or negative than the other. This distinction between problem and emotion focused
coping is very basic, and it is important to note that not all efforts aimed at reducing a stressor
necessarily fall neatly into one category or the other (Carver, Scheier, & Weintraub, 1989).
4
The COPE is a 60-item, 15-scale measure, and contains many items which can be considered as
either emotion or problem focused coping. The COPE was also found to be correlated in varying
degrees to certain personality traits. Findings demonstrated that functional coping strategies are
generally linked to personality traits from various questionnaires that are seen as beneficial,
while less functional coping strategies showed inverse correlations with desirable personality
traits. The COPE was designed to address three key issues the authors believed existed with
previous coping measures:
1.! Provide a more complete and comprehensive assessment of the various coping efforts
people may engage in to deal with a stressor
2.! Reduce ambiguity and produce questions with a direct focus
3.! Develop a scale that is theoretically rather than empirically based, focusing on specific
theoretical arguments that assess functional properties of coping strategies
(Austenfeld & Stanton, 2004; Carver, Scheier, & Weintraub, 1989)
This research assesses coping strategies using the Brief COPE, a 28-item 14-scale measure
which is a condensed version of the original COPE questionnaire. The Brief COPE was created
to:
•! Minimize time demands on participants
•! Revise the questionnaire to exclude two irrelevant scales
•! Slightly refocus three scales
•! Include a self-blame scale evidence that had since been proved was important
The Brief COPE continues to address the three key goals of the original COPE. Like the original
COPE, the Brief COPE continues to include both adaptive and dysfunctional measures of
coping. The items of the Brief COPE can be presented in three formats:
•! Situational retrospective (e.g., “I’ve been doing things to try and take my mind off the
situation”)
•! Dispositional (e.g., “I do things to try and take my mind off the situation”)
•! Situational concurrent (e.g., “I’m doing things to try and take my mind off the situation”).
5
The 14 scales of the Brief COPE refer to coping in the following ways:
1.! Active Coping: taking measures to attempt to remove or avoid the stress, or ameliorate its
effects
2.! Planning: coming up with strategies to deal with the stressor and best handle the situation
3.! Positive Reframing: managing or reframing distressing emotions resulting from the
stressor or situation in positive terms
4.! Acceptance: accepting the reality of and attempting to deal with the situation or stressor;
opposite of denial
5.! Humour: making fun of the stressor or situation in an attempt to make light of the
situation
6.! Religion: using religion to provide a source of comfort, or to clear or organize one’s
thoughts about a stressor
7.! Using Emotional Support: seeking out moral support, sympathy, and or understanding
from others about having to deal with the stressor or situation; this is emotion-focused
coping
8.! Using Instrumental Support: getting advice, assistance, and or information from others on
how to deal with the stressor or situation; this is strictly problem-focused coping
9.! Self-Distraction: focusing away from the stress; intentionally performing activities to take
one’s mind off the stressor
10.!Denial: attempting to push away or ignore the reality of the situation; opposite of
acceptance
11.!Venting: focusing on the stress
12.!Substance Use: using alcohol or drugs to think less about the stressor
13.!Behavioural Disengagement: lessening or giving up one’s attempts to deal with the
stressor and/or achieve goals the stressor interferes with
14.!Self-Blame: criticizing oneself as being responsible for the stressor or situation
Certain scales of the Brief COPE are clearly dysfunctional or adaptive, or carry a distinct
negative or positive tone. For example, behavioural disengagement is clearly negative as it
involves one giving up any attempt of working with the situation or stressor. Conversely,
acceptance carries a strong positive tone as it requires one come to terms with the stressor, which
creates opportunity to move forward and deal with the stresses one is facing. Yet other categories
6
are less clearly defined on whether they are adaptive or dysfunctional coping strategies. For
example, humour can be used to shed light on the situation and possibly make it easier to face a
stressor, however, this can also become dysfunctional if someone uses humour to not have to
face the severity of a stressor or situation. It is important to keep the nature of each scale in mind
when assessing the various coping strategies used by patients. Furthermore, it is important to
assess if certain coping styles as indicated by particular scales tend to co-occur (Carver, 1997).
1.2.3! Circadian Rhythms and Coping Strategies
While the literature focusing specifically on circadian rhythms and coping is sparse, research
does indicate that the cognitive and emotional processes which regulate one’s use of particular
coping strategies are under circadian control. Functioning of cognitive and emotional processes
vary over the course of the 24 h day. An appropriately timed wake and sleep cycle that reflects
one’s internal biological clock facilitates maximal cognitive and emotional performance.
Conversely, a wake sleep schedule that does not mirror one’s biological clock can reduce an
individual’s cognitive and or emotional regulatory abilities (Wright, Lowry, & LeBourgeois,
2012).
Given that the use or disuse of cognitive, emotional, and or inhibitory control in different
combinations plays a role in each of the coping strategies people commonly use (i.e., such as
those assessed in the Brief COPE), these dimensions provide an important link between circadian
rhythms and the coping strategies an individual may use. Circadian oscillators are known to
regulate cognition based functions, such as maintaining alertness, and learning and memory
formation and recall. Performance of these cognitive functions is significantly reduced when
occurring out of synchrony with one’s innate circadian rhythm, such as at one’s off peak time as
indicated by their chronotype (Krishnan & Lyons, 2015). One’s inhibitory control is also
modulated by chronotype. For example, on a task measuring vigilance, M type individuals
maintained high performance when tested in the morning, whereas their performance decreased
with time on task in the evening. Conversely, E type individuals showed worse inhibitory control
with greater time on task in the morning session and greater performance in the evening testing
session (Lara, Madrid & Correa, 2014). This study indicates that for cognitive measures, the
negative effects of time on task can be mediated by testing an individual at his or her
chronotypically optimal time in accordance with their circadian rhythm. Inhibitory control is
7
particularly important to coping as it may allow for blocking certain coping strategies that may
be negative or dysfunctional, and that an individual may be aware are not in their best interest,
but may be a natural response. At one’s chronotypically optimal time, it may be easier to work to
actively avoid particular unwanted strategies, whereas at one’s off peak times, inhibitory control
may be lacking and subsequently facilitate the use of these otherwise blocked out strategies.
In addition to changes in cognitive regulatory abilities across the day in line with one’s circadian
rhythm, it appears that chronotype is also linked with changes in emotionality and mood. When
measuring affect across the day for an entire week in healthy individuals, M type individuals
showed the quickest rise in positive affect in the morning between 9 a.m. and noon, followed by
a dramatic decrease after 9 p.m. Conversely, N (neither type, i.e., intermediate between M and E
types) and E types did not demonstrate the same rapid rise in positive affect in the morning as
was seen among M types (Clark, Watson & Leeka, 1989). Among healthy individuals,
depressive symptomatology is more common among E type individuals than M types, suggesting
an E chronotype may be a predisposing factor for depression (Hidalgo et al., 2009). These results
have been replicated even among individuals with different physical health levels. For example,
between normal versus overweight females, E typology remains associated with a greater
number of depressive symptoms (Pabst, Negriff, Dorn, Susman & Huang, 2009), suggesting E
types report more depressive symptomatology regardless of physical health. Even among healthy
individuals, one’s innate rhythm may predispose an individual towards greater use of negative or
dysfunctional coping strategies, and overall poorer stress management abilities.
1.2.4! Sleep and Coping
Everyone copes differently. The coping strategies one uses will influence their life in various
ways, including impacting their sleep quality. Cognitive arousal among healthy individuals
dealing with stressful life events has been linked to sleep disruptions and or chronic insomnia
(Friedman, Brooks, Bliwise, Yesavage & Wicks, 1995). It is believed that adequate sleep
duration may act as a biobehavioural regulatory and restorative process that regulates one’s daily
emotional experiences and allostatic loads of emotional stress (Vandekerckhove & Cluydts,
2010). Among individuals with major depression, one’s use of avoidance behaviours as a coping
mechanism, along with the intrusion of unwanted thoughts are known to contribute to poor sleep
(Hall et al., 1997). Among physically healthy persons, emotional arousal caused by anxiety is
8
also known to produce sleep disruptions, due in large part to activation of the corticotropin-
releasing hormone system which is recruited for reacting to emotional stress and is believed to
regulate spontaneous waking (Staner, 2003).
Cancer patients and their caregivers – both in the home and hospital – are under cognitive and
emotional strain. Cancer patients’ sleep is known to generally be poor. Among healthy
individuals, sleep plays an important role in mediating coping as they face regular daily events.
Therefore, it is important to assess sleep as a variable associated with changes in coping across
the day in a cancer patient or caregiver’s ability to face the various stresses involved in their role.
Among early stage breast cancer patients, the use of avoidance coping has been linked with
greater time needed to fall asleep across the treatment trajectory. Similar results have been found
in men with prostate cancer, along with decreased sleep onset time (i.e., sleep latency) both at
baseline and across treatment when approach coping strategies are used (Thomas, Bower, Hoyt
& Sepah, 2010). In a study of a varied sample of men with cancer, use of avoidance coping at
baseline was associated with poorer sleep at follow up testing. The authors suggested that using
avoidance coping towards cancer-related stressors or circumstances is likely due to poorer mood
and reduced sleep (Hoyt, Thomas, Epstein & Dirksen, 2009). In this case, it is possible that
poorer mood creates greater emotional arousal, leading to subsequent sleep disruptions. Familial
cancer patient caregivers report similar results. The use of less functional coping strategies (e.g.,
venting, self-distraction, self-blame) have been associated with increased sleep disruptions
(Aslan, Sanisoglu, Akyol & Yetkin, 2009; Carter & Acton, 2006; Northouse, Williams, Given &
McCorkle, 2012). Associations have also been found in some studies suggesting that the use of
positive, proactive coping strategies among caregivers has been associated with reduced numbers
of sleep disturbances (Zhang, Yao, Yang, & Zhou, 2014). Interestingly, in studies among
patients and caregivers, while research seems to consistently point to a positive correlation
between less functional coping styles and increased reports of poor sleep, not all studies seem to
find this association between functional or proactive coping strategies and better sleep quality.
Research on the association between sleep quality and coping strategies required to face the
stresses of being an oncology staff member (e.g., oncologist, oncology nurse, etc.) is more
scarce. Given the cognitive and emotional burden of caring for cancer patients – both treatable
and terminal – it is important to have well developed, functional coping strategies. In a study of
nurses following a shiftwork schedule, on average, sleep quality was found to be poor, as rated
9
by the PSQI. Among this same sample, high emotional disturbance was correlated with poorer
overall sleep quality and greater sleep disturbances (Lee, Chen, Tseng, Lee & Huang, 2015).
While no actual coping measure was used in this study, the high level of emotional disturbance
suggests a lack of coping strategies being used to mediate the emotional demands of the job.
While ambulatory oncology staff do not all follow shiftwork schedules, there are still several
cognitive and emotional demands to be dealt with on a daily basis that require well developed
functional, proactive coping strategies.
While the current literature indicates that sleep and coping are related, it is important to
understand how sleep may be associated with one’s perceived coping abilities as they change
across the day. A better understanding of the influence of sleep on various coping strategies will
guide the development of strategies to teach necessary healthy coping behaviours to patients and
caregivers to better help them in their role. A better understanding of changes in coping across
the day as they relate to one’s sleep quality will allow for necessary assistance to be provided at
times of day when additional help may be required to cope in a proactive fashion.
1.2.5! Depression, Stress, and Coping
In North America, major depressive disorder is the leading cause of disability, and by 2020 is
projected to become the second leading cause of disability worldwide (Muscatell, Slavich,
Monroe & Gotlib, 2009). Depression, which refers to a range of mental health problems, is
characterized by persistent traits such as low mood state, little or no positive affect, and
functional and social impairment. According to Radloff (1977) who designed the Centre for
Epidemiological Studies Depression Scale (CES-D), key major components of depressive
symptomatology include low mood, feelings of guilt and worthlessness, a sense of helplessness
and hopelessness, little or no appetite, psychomotor retardation or delay, and sleep disturbances.
Stress – either chronic or acute – can play a role in the development of depression (Hammen,
Kim, Eberhart & Brennan, 2009; Muscatell, Slavich, Monroe & Gotlib, 2009). One’s reaction to,
or interpretation of a stressor influences the mental impact it has on the individual and affects the
stressors’ contribution to depression onset. When faced with stressors, individuals who exhibit
depression or a high number of depressive symptoms are more likely to use dysfunctional coping
strategies based on avoidance and denial; healthy controls are more likely to use positive,
adaptive coping strategies aimed at accepting the stressor and making plans to move forward
10
towards one’s goals (Orzechowska, Zajaczkowska, Talarowska & Galecki, 2013). Given the
high stress levels typically experienced by cancer patient and caregiver populations, it is
necessary to understand how one’s use of specific coping strategies can reflect depression onset;
such an understanding would allow treatment to be provided before the depression worsens.
1.2.6! Burnout, Stress, and Coping in Caregivers
Burnout is believed to be a consequence of a broader feeling known as compassion fatigue (CF),
that often develops among caregivers of trauma victims and/or patients with a grave illness,
particularly after providing care for an extended period (Stamm, 2010). CF refers to the negative
outcomes of being a caregiver to such individuals, and includes specific feelings such as
exhaustion, frustration, depression, or even fear resulting from working with this population. In
addition to burnout, the other outcome of CF is secondary traumatic stress (STS), resulting due
to prolonged exposure to traumatized individuals, and manifesting as an ongoing combination of
fear, intrusive imagery, and/or sleep disturbance.
While caring for sick or traumatized individuals over an extended period can be mentally and
emotionally draining, there is also the opportunity for personal reward stemming from knowing
one has contributed to ameliorating the quality of life of another individual in a time of need.
These positive feelings, known as compassion satisfaction (CS) are the opposite of CF, and refer
specifically to the pleasure or fulfillment one feels from helping others, in particular those faced
with illness or trauma, and carrying out this role well (Stamm, 2002). Unlike CF, CS does not
break down into further subcategories. Stamm (2010), suggests that together, CS and CF
represent a worker’s overall professional quality of life. The Professional Quality of Life Scale
(ProQoL) produces a rating of these two components.
The coping strategies a caregiver uses to deal with the stresses associated with caring for
traumatized and/or grievously ill patients may contribute to their level of CS or CF. However,
the literature on specific coping strategies that contribute towards increased or decreased CS
and/or CF is sparse. The coping strategies used may alter one’s perception of the stresses faced
on a daily basis, producing higher or lower levels of CS or CF. Understanding which specific
coping strategies are associated with increased or decreased levels of CS and CF is important.
This information can gauge which strategies increase the satisfaction one obtains from their
caregiving role, and help to assess and provide suggestions to alter one’s coping style when it is
11
known to reduce satisfaction and increase fatigue. Given that previous research has shown that a
caregiver’s mental state impacts the quality of care provided to patients (e.g., Beach et al., 2005),
this research will contribute to the field by providing suggestions for appropriate coping styles to
maximize CS and reduce CF.
1.2.7! Personality and Coping
Personality, which is made up of the individual characteristics that shape a person’s behaviour,
feelings, and thoughts, has been linked with circadian rhythms and sleep quality (e.g., Cavallera,
Gatto & Boari, 2014; Duggan, Friedman, McDevitt & Mednick, 2014; Hintsanen et al., 2014;
Hsu, Gau, Shang, Chiu & Lee, 2012). In addition, personality has been linked with coping, such
that certain coping styles occur in increasing or decreasing frequency with certain personality
traits. Some research suggests that even prior to coping, one’s personality predicts frequency of
exposure to stressors, type of stressors experienced, and subsequent appraisals of the stressors.
For example, neuroticism contributes to one’s exposure to interpersonal stress, a tendency to
classify events as highly threatening, and to feel unequipped with the necessary coping
resources, while scoring higher on conscientiousness is reflective of lower stress exposure, likely
due to advanced planning and impulse control (Carver & Connor-Smith, 2010). The idea that
personality predicts stress exposure may be especially true in average daily life scenarios where
one has some control over the roles he or she takes on, the interactions that may be had or
avoided during the day, or the scenarios where a person may find themselves. However, this
theory may not apply in scenarios where a person has less control over the stressor.
Among cancer patients, lifestyle choices may at times contribute to the disease’s development,
but on the whole, people are equally susceptible to a cancer diagnosis in general, regardless of
personality. Among familial or spousal caregiver, one’s relationship with the cancer patient, and
not personality, determines this role. In both these groups, personality will contribute to one’s
adjustment to the role, and their ability to cope with the stressors being presented. In these cases,
it is important to understand how personality predisposes an individual to cope with stressors
when they are presented. An individual’s decision to work in oncology (either as an oncologist,
oncology nurse, pharmacist, etc.), may be more so determined by personality compared to cancer
patients and familial caregivers. This relates back to the previously mentioned point that
personality may predispose people to find themselves in particular scenarios. However, in regard
12
to selecting a career, personality may contribute to one’s reasons for taking on a paid caregiver
role in the field of oncology, in addition to mediating how they will cope with the various
stressors that will be presented in this career. In each of these cases, personality will affect how
one copes with the stresses of their role. While research on coping strategies and personality
traits among cancer patients, and spousal and oncology staff caregivers is sparse, studies
conducted among the general population may serve as a starting point.
Research has suggested various associations between personality traits and choice of coping
strategies. Bartley and Roesch (2011) found that college students high in conscientiousness were
more likely to engaged in problem-focused coping strategies (e.g., problem-solving, cognitive
decision making) which in turn was associated with greater positive affect when faced with a
daily stressor. In a meta-analysis assessing the relationship between coping and personality traits,
high conscientiousness and extraversion were more common among individuals using problem-
solving and cognitive restructuring, which generally reflects problem-focused coping strategies.
Extraversion was also noted to produce support-seeking coping behaviours. This same analysis
noted that neuroticism was more predictive of emotion-focused coping strategies, but also
showed a link to support-seeking, similar to that found among individuals high in extraversion
(Connor-Smith & Flachsbart, 2007). Among intensive care unit (ICU) nurses, while personality
was not associated with workplace stress, it was linked with various coping styles used by the
participants to approach daily stresses. Among nurses reporting greater conscientiousness and or
agreeableness, there was a greater association with active coping to work towards making the
situation better, and planning by using strategies to resolve a stressor. Nurses high in openness
were also more likely to cope by using positive reframing in order to see stressors in a more
positive light. These personalities showed associations with coping strategies indicative of
problem focused coping to resolve the problem. Conversely, high neuroticism was strongly
related to venting as a coping strategy to verbally express negative feelings, reflective of an
emotion-based coping style aimed at reducing one’s negative feelings about the threat as
opposed to altering the source of stress itself (Burgess, Irvine & Wallymahmed, 2010).
In the research discussed in this dissertation, in addition to assessing personality and coping in
this cancer-related group, we addressed the role of personality among patients and caregivers as
a mediating factor on the association between chronotype and sleep quality on coping, and one’s
ability to handle the stresses associated with their role, both across the day and overall.
13
Understanding the association between chronotype, sleep quality and personality and their
influence on coping will allow for more tailored strategies to foster healthy coping strategies
among patients and caregivers with different chronotypes and sleep patterns, as well as
personality types.
1.3! Cancer as a Chronic Stress
Stress results when the demands or outcomes of a positive or negative situation impinge upon or
threaten one’s behavioural, emotional, or physical state. More specifically, stress is deemed as an
unpleasant negative emotional experience resulting in behavioural, biochemical, and
physiological changes intended to adapt to the stressor, either via its manipulation, alteration, or
accommodation (Baum, 1990). Stressors can be acute or chronic. Acute stressors are those that
are specific episodic events with relatively discreet onset and offset points, and are general
occurrences in everyday life (e.g., an interview, narrowly missing a car accident, etc.). Chronic
stressors refer to ongoing events that continue over a prolonged period of time (e.g., poverty,
long term illness). The body’s stress response system is activated when dealing with acute and
chronic stress to prepare the body for the challenges that must be faced. However, chronic long-
term activation of this system – at both a low or high stress level – is detrimental to one’s mental
and physical health, contributing to a range of health problems varying in severity (e.g., high
blood pressure, obesity, anxiety, and depression, etc.). Interestingly, events and factors that cause
stress for one individual may not produce the same stresses for another person. Different people
can react to stress in a variety of ways, with the distinction typically depending on one’s
perception of the stressor (e.g., can it be overcome, how will it affect life, etc.) (Baum, 1990;
Hammen, Kim, Eberhart & Brennan, 2009).
Cancer can be contributed to by the detrimental effects that stress causes on the body, and itself
acts as a long-term chronic stressor to both the patient and his or her caregivers, including
medical professionals, and friends and family members. For patients, the gravity and uncertainty
of cancer overall causes stress, but various forms of cancer and the according treatments may
cause their own specific stresses such as the loss of a body part, or changes in one’s sexual
functioning. For medical caregivers, being tasked with continuously providing care and
assistance to ill individuals, coupled with being charged with their care and not always seeing
14
favourable outcomes produces a chronic stress. As friends and family members providing care,
chronic stress can result for many reasons including constant worry for a loved one, and changes
to one’s own personal life including taking on a new role. In each case, one’s personal coping
strategies will determine how an individual will cope with and be affected by the stresses
associated with cancer.
1.3.1! Emotional Distress in Patients
It is well understood that in addition to the physical symptoms associated with breast cancer and
its treatment, there are increased risks for developing psychological problems and suffering
emotional distress (Barre, Padmaja, Saxena & Rana, 2015). Bultz and Carlson (2006) reviewed
the literature and found that in North America the incidence of distress among cancer patients
across the trajectory of the illness (diagnosis, treatment, survival, and or palliation) is between
35-45%. Emotional distress during these times can stem from a number of reasons such as worry
about one’s health and survival, changes in family life and or job, financial security, marital
problems, etc. Research indicates that when cancer patients at large experience greater emotional
distress, they are more likely to visit emergency facilities and community health services, thus
placing a larger burden on the healthcare system, while often leaving the emotional distress
unresolved (Carlson & Bultz, 2004). In addition to the economic burden that emotional distress
can cause when patients seek additional emergency or community care, the emotional distress
compounds with the physical symptoms patients may already be experiencing thus adding extra
burden to the situation. While a certain level of emotional distress is normal, and a patient can
still continue to function, emotional distress can often become all-consuming and debilitating.
Emotional distress may stem from poor coping, and may also fuel continued poor coping styles.
For these reasons, it is important to understand solutions to treat or alleviate emotional distress,
and maximize positive, proactive, and healthy coping strategies among patients.
1.3.1.1! Why Choose Breast Cancer Patients
Survival rates for patients diagnosed with stages I to III breast cancer have increased
significantly in recent decades, with five-year survival rates between approximately 72% to
almost 100% (Canadian Cancer Society, 2016b). While survival rates from breast cancer are
increasing, incidence of breast cancer is still relatively high; it was estimated that in 2016 alone,
approximately 25,700 Canadian women would be diagnosed with breast cancer (Canadian
15
Cancer Society, 2016a). Therefore, while breast cancer survival itself has increased significantly
in recent decades, it continues to be a commonly diagnosed cancer. Given the relatively strong
prognosis, early stage breast cancer patients are a unique population who can experience worries
and fears that contribute to emotional distress, but can also see an end to their treatment and the
eventual high likelihood of recovery. Conversely, among metastatic breast cancer patients, while
some treatment might be aimed at tumour reduction and slowed growth, patients typically
understand that their disease or complications from it will lead to eventual death. The eventual
“light at the end of the tunnel” so to speak that applies to early stage breast cancer patients allows
for a shorter period of study between diagnosis, treatment, and remission, and the according
changes in emotional distress that come with being diagnosed, seeing effects of treatment, and
beginning the road to recovery.
1.3.1.2! Specific Distress Involving Genetic Issues
Given the increased risk of developing breast cancer when one is a BRCA1 or BRCA2 carrier, it
is likely that these individuals may experience increased distress. While learning that one carries
a BRCA1 or BRCA2 mutation can potentially promote efforts towards breast cancer prevention,
concern over adverse emotional, cognitive, and behavioural consequences have also been
expressed (Lerman & Croyle, 1996, Lerman & Schwartz, 1993). Baum et al. (1997) suggest that
one’s level of emotional distress following genetic testing indicating that one is a carrier is
mediated by questions regarding disease occurrence, timing, potential severity, treatment course,
and potential for preventability. Other factors that may influence one’s distress relating to
carrying a BRCA1/BRCA2 mutation include quality of genetic counselling and medical
surveillance available, and age and gender. Time since testing also influences one’s emotional
distress related to being a carrier, such that with time, distress appears to decrease among
carriers, and return to pre-testing levels. Patients who receive confirmation of being
BRCA1/BRCA2 carriers with a personal cancer history experience less distress post diagnosis
than those without a personal cancer history (Hamilton, Lobel & Moyer, 2009).
1.3.2! Emotional Distress in Caregivers
1.3.2.1! Specific Distress Involving Partners
Patient’s partners are typically deeply affected by the cancer diagnosis and are often the patient’s
primary support or caregiver outside the hospital. Only recently has the emotional distress of
16
caregivers begun to be studied. Merckaert et al. (2013) reviewed the literature and found that
high levels of distress are experienced by between 10-50% of caregivers, and their distress level
is often similar to that of their partners. Partners have reported receiving less social support
compared to patients, and their distress increases with disease progression into palliation (Davis-
Ali, Chesler, & Chesney, 1993; Kurtz, Given, Kurtz & Given, 1994). In a recent study among
cancer patient caregivers, 85.6% reported unmet needs (i.e., the requirement or desire for the
provision of actions or resources to help achieve an optimal state of well-being), while 69.1%
reported positive for emotional distress, 26.5% for depression, and 34.9% for anxiety
(Sklenarova et al., 2015). Given the frequency of the issue, it is important to understand causes
of and changes in emotional distress among caregivers in order to provide the necessary tools to
alleviate these feelings.
1.3.2.2! Specific Distress Involving Oncology Staff
Hospital staff who must provide care to cancer patients are also known to experience emotional
distress, often in the form of burnout, fatigue, and low job satisfaction. Work stress and the
stressful situations associated that are faced by doctors, nurses, radiation therapists, and
pharmacists working in the field of oncology and dealing directly with patients can combine to
produce feelings of burnout and fatigue. Even among those oncology professionals who report
high levels of personal accomplishment, occupational distress levels are still rated as being high.
It is important to understand the causes of emotional distress and patterns of change in stress
management across the day in order to provide the necessary tools within the hospital setting to
reduce these negative feelings. Reducing these negative feelings will both reduce the number of
employees needing to take long-term leave as a result of work related burnout or fatigue. At the
same time, reduced emotional distress, burnout, and fatigue among home medical staff is known
to increase the overall satisfaction of their patients (Ruggieri, Zeppegno, Gramaglia, Gill,
Deantonio & Krengli, 2014; Vahey, Aiken, Sloane, Clarke, & Vargas, 2004).
1.4! Rhythmicity
1.4.1! Biological Clocks in Nature
Biological clocks that moderate physiology and behavior reflect cyclic changes in the
surrounding environment. The systematic changes can be moderated by the earth’s rotation, and
17
the sun and moon, and can range in duration based on daily (circadian), annual (circannual),
lunar (circalunar), or tidal (circatidal) cycle length. Rhythms provide a significant adaptive
advantage by allowing an organism to anticipate patterns such as food availability, mating
seasons, migratory periods, times to avoid predation by other organisms, etc. Circadian clocks
are the best understood sources of these daily rhythms. These regulate specifically the
endogenous cyclic changes exhibited by organisms from cyanobacteria to humans across the 24-
hour day. A defining characteristic of these clocks is that their action persists when the organism
is held in constant conditions (Chaudhury & Colwell, 2002; McClung, 2011; Reppert & Weaver,
2001).
1.4.2! Biological Clocks in Mammals
In mammals (including humans), the circadian system is a hierarchical organization of clocks
with a master oscillator that coordinates rhythmicity throughout.
The suprachiasmatic nuclei (SCN) of the hypothalamus has the primary responsibility of
synchronizing physiology and behavior with daily changes in the environment (Ralph, Foster,
Davis & Menaker, 1990). Environmental light received by specialized cells in the retina transmit
this information to the SCN, thereby entraining its rhythm according to the surroundings (Morin
& Allen, 2005).
It is believed that the two primary adaptive reasons for organisms to have biological clocks are
for organizational and anticipatory purposes. From an organizational perspective, biological
clocks allow for energy conservation by ensuring metabolic or behavioural processes are
sequenced to reduce energy use and increase overall efficiency. For example, in the evening, the
biological clock is organized to begin lowering overall body temperature to maximize the body’s
energy reserves; simultaneously, metabolism is also slowed as an organizational measure to
enhance energy conservation since the body does not require a constant elevated supply of
glucose to support activity. From an anticipatory perspective, biological clocks allow organisms
to predict changes both within the organism (serving again as an energy conservation measure)
and in the surrounding environment (to allow the organism to use energy efficiently, while
simultaneously protecting itself). For example, research in rats has demonstrated anticipatory
behaviours for meal times across the day. In free feeding conditions, rats with SCN lesions
demonstrated anticipated restricted food access at 24-hour intervals, regardless of the absence of
18
a functioning master circadian clock, suggesting circadian oscillators beyond the SCN are
entrainable by restricted feeding schedules (Stephan, Shwann & Sisk, 1979).
1.4.3! Biological Clocks in Human Beings
Virtually all cells and tissues within the human body express circadian rhythms, which influence
several bodily processes. At the molecular level, phase is determined according to a set of
canonical clock genes comprised of a transcription translation feedback (TTF) oscillation, that
may be set primarily by light-dark cycle, and influenced by a variety of nonphotic zeitgebers. At
the hormonal level, several rhythmic patterns of expression for various bodily hormones have
been studied. Most notably, the human melatonin cycle shows rhythmic oscillations across the
24-hour day with melatonin synthesis confined predominately to the subjective night (Arendt,
2006). A peak in hunger times across the day also appears to be tied to circadian rhythms,
allowing for a specific peak in evening hunger (dinner time) before the overnight fasting period
(Scheer, Morris, & Shea, 2013). Human performance on physical activities is also known to
fluctuate daily in accordance with circadian timing. For example, several aspects of sports
performance including flexibility and muscle strength show time-of-day variances tied to
circadian rhythms (Atkinson & Reilly, 1996).
1.4.4! Cognition, Emotion, and Circadian Rhythms Performance fluctuations across the day on higher level cognitive processes are driven by
circadian rhythms (e.g., Natale, Alzani, & Cocogna, 2003; Schmidt, Collette, Cajochen &
Peigneux, 2007). Cognitive function is an umbrella term referring to the general mental
processes by which information is acquired, and includes various categories including but not
limited to attention, memory, and language. Emotional regulation requires a high level of
cognitive functioning and control.
1.4.5! Chronotype
M versus E preference is often called chronotype, which broadly refers to self-perceived
changing performance capabilities across the day. Chronotype variances support the idea that
people have distinct innate personal preferences for carrying out tasks with varying demands
earlier or later in the day. These personal preferences may apply to several tasks, including those
with a cognitive, physical, or emotional basis. Furthermore, chronotype is known to shift across
19
the lifespan; children are typically M types, with tendency to shift towards E types with the onset
of adolescence followed by a cross-cultural shift back towards M type by approximately age 50.
(Schmidt, Collette, Cajochen & Peigneux, 2007). Chronotype refers to circadian phase and is
linked to the circadian clock. For example, preference for sleep timing has been correlated with
melatonin secretion profiles which are moderated by the circadian clock (e.g., Burgess & Fogg,
2008).
In humans, the most clearly visible and easily measurable display of rhythmicity is the daily
sleep-wake cycle. While individuals exposed to the same solar cycles should all be similarly
entrained to the daytime, human circadian behaviour research has demonstrated a significant
interindividual variability in the temporal relationships that exist with the light/dark cycle
(Roenneberg et al., 2007). Significant interindividual variability exists for the timing of sleep and
wake behaviour. People are commonly characterized as larks (early risers) and owls (late risers),
as a categorization for their M versus E preference, respectively. Kleitman (1963) found that
human sleep-wake patterns demonstrated that the existence of M and E type individuals is
correlated with “early” or “late” peaks in body temperature and performance efficiency curves
throughout the day.
1.4.5.1! Genetic Basis of Chronotype
There has been long standing interest in understanding the genetic basis for circadian
rhythmicity, and its contribution to chronotype – a behavioural phenotype. Chronotype is
normally distributed in the population across the age span, and many common genetic variants
with modest effects have been found to influence these phenotypes (Kalmbach et al., 2017).
Research indicates that a series of molecules function together in a feedback loop to form a core
circadian clock in every cell, and the SCN appears to maintain synchrony among these cellular
clocks (Kalmbach et al., 2017; Lowrey & Takahashi, 2011). Twin and family studies suggest that
genetic factors explain up to 50% of the population variability in chronotype (Kalmbach et al.,
2017).
1.4.5.2! Questionnaires
Several questionnaires have been devised to assess chronotype, however the gold standard in the
literature remains the Horne-Östberg Morningness Eveningness Questionnaire (MEQ) (Adan &
20
Natale, 2002; Horne & Östberg, 1976). The MEQ is a 19-item multiple choice measure that asks
respondents to indicate their optimal time of day for performing various activities between
morning and night, and yields a total score in a range of 16-86 (16 = extremely evening types; 86
– extremely morning type). More recently, the Munich Chronotype Questionnaire (MCTQ) was
devised, and suggests that the timing of midsleep is also an indicator of chronotype on work- and
free-days (Roenneberg, Wirz-Justice & Merrow, 2003). However, while a satisfactory
correlation exists between MEQ and the uncorrected-MCTQ score, the correlation diminishes as
corrections for various scenarios are applied to the MCTQ score (Di Milia, Adan, Natale, &
Randler, 2013).
1.4.5.3! Performance Variances
Following early research demonstrating that chronotype is associated with performance curves
across the day, further studies have examined this concept in greater detail. Basic performance
timing among athletes appears tied in some extent to chronotype, such that some physical
activity performance appears better at one’s optimal time of day (Brown, Neft, & LaJambe,
2008). In terms of the psycho-physical factors behind performing a physical activity, completion
of such tasks at one’s chronotypically optimal time is linked to increased enjoyment and
affective response, and general motivation to exercise. Engaging in physical activity during one’s
chronotypically preferred time of day is likely to predict greater long term adherence to an
exercise regime (Vitale, Colagiuri & Weydahl, 2013).
Changing cognitive function throughout the day is related to chronotype. Morning and evening
type individuals show optimal functioning of executive control abilities when completing
cognitive-based attentional tasks at their chronotypically optimal time of day; conversely,
executive control abilities are diminished at non-optimal times of day (Lara, Madrid & Correa,
2014). Performance on some memory-related tasks tends to decrease across the day in M type
individuals, and to increase across the day in E type individuals (Hasher, Zacks & Rahhal, 1999).
The same results have been found between different age groups, on a range of cognitive and
executive function type tasks (e.g., Hahn et al., 2012; May, Hasher & Stoltzfus, 1993). The
existence of optimal times for performance on lab-based tasks depending on one’s chronotype
and age indicates that this phenomenon exists throughout the lifespan. While these tasks can be
tested at controlled times in the day in a laboratory setting, it is important to consider that
21
optimal times of day exist in regular day-to-day life as well, and as such, performance variances
will exist here as well.
1.4.5.4! Emotionality
Chronotype is also related to emotionality. Studies have indicated that individuals with a greater
propensity towards morningness rate more highly in positive affect than E types, while those
with tendency towards eveningness display greater tendency for higher depression scores (Clark,
Watson & Leeka,1989; Hasler, Allen, Sbarra, Bootzin, & Bernert, 2010). When comparing
younger and older adults, M types from both age groups also rated higher on their overall
positive affect as compared to age matched E types (Biss & Hasher, 2012). Among young
teenagers with a tendency for eveningness, significantly more emotional and behavioural
problems were reported compared to age matched participants with a tendency for morningness
(Gau et al., 2007). Emotional dysregulation has also been linked to chronotype. Among patients
who have attempted suicide, eveningness tendencies were most common among those patients
who had attempted violent suicide, and rated highly on tests of impulsivity; similarly, suicidal
ideation is highest among patients with major depressive disorder and a preference for E (Bahk,
Han & Lee, 2014; Selvi et al., 2011). Aggression and antisocial behaviour also appear more
prevalent in individuals with a preference for E (Schlarb, Sopp, Ambiel & Grünwald, 2014). The
literature suggests E types are more likely to experience mental, physical and behavioural
disorders than M types. It is unclear whether this is purely based on conflicts between optimal
sleep-wake times versus societally imposed timing structures in daily work and life, or whether
an E chronotype itself is specifically associated with the onset of these disorders. It is important
to consider that morningness may provide a protective factor, and it is necessary to understand
how to protect E types against the onset of these issues. These findings show a clear link between
chronotype and emotionality. However, it remains to be seen how changes in one’s emotional
state across the day, particularly when considering coping, relate to chronotype.
1.5! Sleep
In today’s world, the traditional timing of human behaviours and activities linked to the sleep-
wake and light-dark cycles has been altered due to the use of artificial lighting sources. Societal
22
demands including work schedules, a global network and a wide range of modes of
telecommunication, active lifestyles, and daily tasks also contribute to these shifts in timing of
human behaviour. While it is understood that a need for sleep exists, many people, particularly in
the modern world, do not place high importance on the body’s need for sleep. In the 24-hour
world in which we live, night time sleep is often replaced with other activities, resulting in sleep
deprivation (Ferrara & De Gennaro, 2001).
1.5.1! Two-Process Model of Sleep
Adequate and good quality sleep is necessary for maintaining optimal cognitive, mental,
physical, and emotional health. There is debate regarding how much sleep is necessary each day,
with a wide range of hours of necessary sleep being reported by various researchers. While little
consensus exists on the general number of hours that should be necessary for proper daytime
functioning and health maintenance, on the whole, it is understood that sleep deprivation
negatively impacts mental and physical wellbeing (Alhola & Polo-Kantola, 2007; Ferrara & De
Gennaro, 2001). The most common theory of sleep regulation is the two-process model; it
suggests that sleep is dependent on both a homeostatic sleep process (which reflects the
increased need for sleep with sustained wakefulness), and a circadian process (which determines
a window during which sleep is most likely to occur). These two processes are believed to
interact to determine an individual’s sleep/wake cycle, and to influence changes in one’s level of
alertness and attention throughout the day (Achermann, 2004; Alhola & Polo-Kantola, 2007). A
simple result of environments (real world or experimental) that interrupt sleep is sleep
deprivation, which can drive a greater propensity for sleep during wake time. This is a natural
consequence of sleep disruption, not to be confused with insomnia, which can cause sleep
deprivation, but likely has a primary cause within the sleep regulation system itself. Whereas
insomnia refers to the inability to obtain adequate and or good quality sleep despite the
opportunity to sleep, sleep deprivation refers to truncated sleep length due to an environmental
circumstance that restricts the opportunity to sleep.
1.5.2! The Functions of Sleep
A key function of sleep – to regulate and restore brain energy – is well documented (Dworak,
McCarley, Kim, Kalinchuk & Basheer, 2010). In rats, sleep brings about a surge in adenosine
triphosphate (ATP), which acts as the energy currency of brain cells (Dworak et al., 2010). The
23
function of the hypothalamic pituitary adrenal (HPA) axis is also altered by partial sleep
deprivation. In a study of partial sleep deprivation (between 04:00 and 08:00 hours), the normal
quiescence of cortisol secretion showed a significant delay, producing elevated evening cortisol
levels on the following evening (Leproult, Copinschi, Buxton & Van Cauter, 1997). These
findings suggest sleep deprivation can have adverse health effects. In the United States, average
sleep duration decreased from between 8 to 9 hours in the 1960s, to approximately 7 hours in the
1990s, paralleling the increased prevalence in the country’s diabetes and obesity cases. A review
of the literature supports the notion that chronic partial sleep loss results in dysregulation of
glucose pathways and neuroendocrine control of appetite, resulting in overeating and reduced
energy expenditure, subsequently leading to an increased risk for obesity and or diabetes
(Knutson, Spiegel, Penev & Van Cauter, 2007). On the whole, this research indicates that sleep
not only affects brain function directly, but other downstream pathways tied into one’s health
and biological well-being.
1.5.3! The Impact of Sleep and Sleep Loss
In addition to a biological function, sleep also influences one’s performance and mental state.
Sleep quality is associated with the onset of, or prevention against a variety of disorders.
Individuals with poor sleep quality are significantly more likely to experience at least one
common mental disorder compared to individuals who report good quality sleep (Rose et al.,
2015). Problems with cognition have also been reported in the literature among individuals with
poor sleep quality and particularly among those with untreated sleep apnea, which is known to
cause reduced sleep quality (Vaessen, Overeem, & Sitskoorn, 2015). As with chronotype, it is
evident that sleep quality is also a significant contributor to both one’s health and ability to
function throughout the day.
1.5.3.1! Sleep, Cognition, and Memory
In response to demanding work, school and social schedules in daily life today, many people
trade sleep for additional time to devote to these activities, resulting in high occurrences of sleep
deprivation in modern society. Insufficient sleep dramatically impacts alertness and attention,
without which it is extremely difficult to carry out complex cognitive processing (Lim & Dinges,
2008; 2010). In a study of sleep deprivation using a psychomotor vigilance test among otherwise
healthy subjects, a general slowing of response times and average reaction time was observed
24
(Lim & Dinges, 2008). Sleep restriction research has demonstrated that compared to individuals
sleeping 8 hours, those on a restricted sleep cycle of 6 hours for 2 weeks demonstrated cognitive
performance deficits equivalent to 2 nights of sleep deprivation. Furthermore, subjects on a 6-
hour restricted sleep cycle were relatively unaware of the cognitive deficits they were displaying,
which may explain why many members of the general public assume the effects of sleep
restriction are relatively benign (Van Dongen, Maislin, Mullington & Dinges, 2003).
Furthermore, it seems age may influence the extent to which psychomotor vigilance on
attentional tasks is affected by sleep deprivation. A review of the literature indicates that among
younger adults, sleep deprivation results in poorer performance and slowed reaction time on
vigilance tests, along with unintentional sleep episodes during sleep deprivation or restriction
periods as compared to older adults. Conversely, older adults may show slower reaction times on
such measures, yet compared to younger adults, their overall performance is less impaired by a
night or sometimes more of lost sleep (Killgore, 2010). In general, cognitive processing appears
reduced across the lifespan when an individual loses sleep, however the extent of disruption may
vary by age group.
Sleep loss is also known to affect one’s emotional functioning and processing abilities, which
subsequently alters associated cognitive processes such as memory for events, judgment and
decision making skills. Sleep deprivation specifically appears to result in greater intolerant and
unforgiving behaviour, along with more self-focused tendencies, compared to when one is fully
rested (Killgore, 2010). Sleep plays a role in one’s emotional evaluations. Individuals on a sleep
deprivation schedule for one night reported greater negative mood and perceived neutral images
significantly more negatively compared to individuals on an undisrupted sleep schedule. The
negative emotional bias for neutral stimuli did not appear related to the negative mood known to
typically follow sleep loss (Tempesta et al., 2010). Furthermore, research has also demonstrated
that sleep deprivation results in a greater tendency towards negative cognitions and intolerance
for frustrating social scenarios or greater emotional reactivity to threatening visual stimuli. In a
two-night sleep deprivation study, participants were asked to respond to a cartoon demonstrating
two characters experiencing a frustrating situation. Sleep deprived participants were more likely
to direct blame or show hostility towards other people or objects in the environment, while
simultaneously showing less willingness to suggest amends that would be mutually satisfying for
both parties. These findings were indicative of a weakened inhibition of aggression and poor
25
willingness to behave in ways that would increase effective social interaction (Kahn-Greene,
Lipizzi, Conrad, Kamimori & Killgore, 2006). In studies of REM sleep deprivation, emotional
responses to visually threatening stimuli were also enhanced relative to baseline after one night
of REM-deprivation. This work highlights the specific behavioural regulatory role played by
REM sleep in order to modulate emotional responsivity (Rosales-Lagarde et al., 2012). In
general, it appears that sleep deprivation plays a key role in regulating various aspects of mood
and emotional processing, particularly in elevating negative-response styles in reaction to
potentially trying situations.
1.5.3.2! Polysomnography and Actigraphy
Objective sleep measures often require participants to use polysomnography, and sleep in a
laboratory setting, many times for more than one night. Among patients receiving medical care,
particularly those being treated for a chronic illness (e.g., cancer) some researchers may use
alternate measures to assess sleep quality without having individuals change their routine and
sleep away from home. While not as detailed as polysomnography, actigraphy can be used as a
noninvasive measure of one’s daytime versus nighttime activity and sleep across the 24 h day.
Actigraphy devices are typically worn on the non-dominant wrist and record motor activity
across the day. Sleep estimates are made based on the amount of movement versus rest. The data
are translated into a histogram, with the x-axis representing a 24 h time period, and the y-axis
representing activity frequency. However, unlike polysomnography, actigraphy does not detail
the various stages of sleep that one passes through, or the amount of time spent in each state.
Nonetheless, actigraphy is still deemed an effective and accurate tool to measure circadian
activity rhythms, and has been especially noted for use among cancer populations due to its
nonintrusive nature. (Fiorentino & Ancoli-Israel, 2007; Roscoe et al., 2007).
1.5.3.3! Sleep Quality
Subjective sleep measures are not based on physiological measures such as REM sleep, or
changes in hormone levels, but rather on one’s self-reported ratings of their sleep. Various
measures exist, and can be as simple as a sleep log – where an individual tracks their daily sleep
and wake times, or can include retrospective questions requiring an individual to reflect on
various aspects of their sleep (i.e., duration, number of awakenings, body temperature sensation,
etc.) over an extended period. Subjective sleep measures reporting on one’s own sleep appear to
26
reflect objective polysomnographic data (Armitage, Trivedi, Hoffmann & Rush, 1997), however,
subjective sleep quality reports on the sleep of another individual (e.g., by parents on their
child’s sleep), were less verified by objective measures (Choi, Yoon, Kim, Chung & Yoo, 2010).
The Pittsburgh Sleep Quality Index (PSQI) (Buysse, Reynolds III, Monk, Berman, Kupfer, 1988)
is a commonly used subjective sleep measure. It calculates a score based on one’s recall for sleep
quality, latency, duration, efficiency, and disturbances, along with one’s use of medications and
daytime dysfunction over the past month. It is a short questionnaire, and has good reliability and
validity. It is intended for use in research and clinical settings, and is intended for easy patient
use.
1.5.3.3.1! Cancer
Sleep quality refers to both objective and subjective aspects of sleep. Objective characteristics of
sleep quality can include factors such as sleep duration, latency, and number of arousals, while
subjective aspects refer to the degree to which sleep is restful or deep (Buysse, Reynolds III,
Monk, Berman, Kupfer, 1988). Sleep quality is associated with sleep disorders. Sleep disorders
refer to issues such as difficulty falling asleep or maintaining sleep, low sleep efficiency, and
insomnia, and can become chronic. Sleep quality among cancer patients as a collective group is
typically reported as poor. There is a high prevalence of problems with achieving and
maintaining sleep, low sleep efficiency (amount of time spent asleep while in bed), along with
high levels of sleepiness during waking hours (Fiorentino & Ancoli-Israel, 2007). In a telephone
survey study of 150 lung and breast cancer patients reporting on their sleep quality over the past
month, 44% of participants reported sleep problems, yet only 16.6% of individuals reported these
sleep complaints to their health care providers (Engstrom, Strohl, Rose, Lewandowski &
Stefanek, 1999). While the importance of sleep is generally well understood, this research
indicates that the prevalence of sleep problems and general poor sleep quality among cancer
patients is often not well understood and subsequently left unaddressed.
It is important to also distinguish between cancer related fatigue and cancer related sleep
disorders. Cancer related sleep disorders are those that may result due to the effects of the
disease itself or its treatment (e.g., chemotherapy, radiation, surgery, etc.). Cancer related fatigue
however refers to the feelings of tiredness, weakness, and or low energy, that remain unrelieved
by good quality sleep or even rest. These feelings can occur both as a side effect of the disease
27
itself and or its treatment, however the exact cause is unknown. Poor sleep quality and sleep
disturbances or problems are known to affect patients at all stages of treatment, and are often
reported even following treatment, indicating that poor sleep is often a chronic problem within
this population (Fiorentino & Ancoli-Israel, 2007; Hofman, Ryan, Figueroa-Moseley, Jean-
Pierre, & Morrow, 2007; Roscoe et al., 2007).
1.5.3.3.2! Breast Cancer Patients
Much of the research on cancer patient sleep quality has focused on breast cancer patients. In a
sample of nonmetastatic breast cancer patients recruited during routine follow-up exams after
having completed treatment, the average sleep score on the PSQI was 7.0, which is higher than
the recommended cutoff for good quality sleep as determined by the questionnaire. This
indicates that on average, breast cancer patients report significant sleep problems and poor sleep
quality (Carpenter & Andrykowski, 1998). In a study of 72 women with breast cancer recruited
at various time points of the diagnosis, treatment, and recovery trajectory, 61% reported
significant sleep problems on the PSQI. In particular, sleep in this population was characterized
by reduced total sleep time, pain, nocturia, hot flashes, and trouble breathing caused by coughing
or snoring (Fortner, Stepanski, Wang, Kasprowicz, & Heith Durrence, 2002). These data confirm
previous research indicating that sleep quality is often poor among breast cancer patients.
Insomnia is often reported as a common contributor to poor sleep quality among breast cancer
patients. Specifically, it is characterized by trouble initiating or maintaining sleep, non-
restorative sleep, must persist for at least one month, and must result in distress or interfere with
one’s ability to function. Insomnia as a contributor to poor sleep quality may be more common
among women with breast cancer for three key predisposing, precipitating and perpetuating
reasons. Female gender is a known predisposing factor in the general population to cause
insomnia. Among breast cancer patients, female gender may contribute to higher than average
rates of insomnia, together with factors such as increasing age. Depression is also known to be
more common among females and has a strong link to insomnia. Given the high number of
breast cancer patients reporting depression (on average approximately 20 – 30%), this may be a
common concomitant health mental health issue that contributes to insomnia and poor sleep
quality in this population. Insomnia may be precipitated by treatment side effects which may
include increased frequency and severity of hot flashes due to the sudden onset of drug induced
28
menopause. Finally, it may be perpetuated by the disruptions to the typical sleep-wake cycle,
such as frequent napping or desynchrony of one’s circadian clock. Given the number of common
factors that can contribute to the onset of chronic insomnia, and its negative effects on sleep
quality, it is especially important that the skills to prevent insomnia onset be taught to cancer
patients. Both cognitive and behavioural therapies separately and in combination have been used
with success to teach patients the necessary skills to avoid and or combat the onset of insomnia
(Bower, 2008; Fiorentino & Ancoli-Israel, 2007).
1.5.3.3.3! Caregivers
Familial cancer patient caregivers play a key role in providing care to cancer patients in both the
home, and additional assistance even in a hospital or hospice setting. The difficulties faced by
caregivers are believed to contribute to the sleep disturbances they face during the cancer care
process. Research has indicated that key worries, fears and emotional burdens of cancer
caregivers relate to the nature and metastasis of the patient’s disease, helping patients cope with
or face the emotional consequences of the illness, and managing life disruptions caused by
cancer (Aslan, Sanisoglu, Akyol, & Yetkin, 2009; Osse, Vernooij-Dassen, Schadé, & Grol,
2006). These worries, fears, and burdens are believed to contribute to sleep disturbances found to
be common among cancer caregivers such as insomnia, nocturnal sleep disturbances, chronic
sleep loss. Furthermore, sleep problems are found to persist among several caregivers following
the death of a loved one, particularly if the patient’s symptoms were unrelieved during their final
three months of life (Berger et al., 2005). These sleep disturbances among caregivers are known
to negatively impact on the individual’s ability to carry out daily tasks and activities, quality of
life, emotional status, and performance as a caregiver. Conversely, improved sleep quality allows
caregivers to improve the quality of care which they provide to patients (Aslan, Sanisoglu,
Akyol, & Yetkin, 2009). Given the positive versus negative effects that varying degrees of sleep
quality can have both directly on the caregiver and on the patient, it is important to understand
specifically how functionality is affected across the day, particularly as it relates to one’s
circadian rhythm. A better understanding of the variances in caregiver performance as they relate
to circadian rhythms will produce a clearer picture of performance variances across the day, at
both the physical and emotional levels of functionality.
29
1.5.3.3.4! Oncology Staff
Far less research has focused specifically on the role of sleep quality among oncology staff (i.e.,
oncologists, and nurses and pharmacists working in oncology). Among the general population of
nurses, independently of the stressful work situations faced on a daily basis, sleep debt is known
to result in feelings of physical and emotional fatigue, reduced cognitive functioning, and a
general sense of burnout (Bellicoso, Ralph, Trudeau, 2014). In general, physicians also obtain
inadequate sleep, with some even working more than 24 consecutive hours. These interrupted
sleep cycles result in disrupted circadian rhythms. Evidence has demonstrated that this loss of
sleep experienced by many physicians can directly affect their personal health as well as the
health and safety of patients. A review of the literature indicates that lack of sleep among
physicians has been associated with several negative outcomes including misinterpretation of
patient records and scans, increased procedure error rates, and lack of empathy towards and poor
communication with patients (Eddy, 2005). Given the negative side effects sleep loss and poor
sleep quality can have on one’s self and the patients one cares for, sleep quality among oncology
caregiving staff in hospitals should be examined relative to changes in their performance ability.
A better understanding of performance fluctuations, and how they are related to sleep quality can
yield solutions to maximizing oncology staff performance and quality of patient care.
1.6! Breast Cancer Background
1.6.1! Cancer Biology
While several forms of cancer exist, in general terms, cancer is a disease that begins in the cells.
The normal cycle of cellular reproduction involves cellular division and subsequent death of
older cells. When this typical cycle is interrupted, uncontrolled cell growth results, with cells
producing tumours, lumps, or spreading through the bloodstream and lymphatic system to other
areas of the body. Tumours that stay in one controlled section of the body are known as benign,
and are typically non-life threatening. However, malignant (spreading) tumours have the
potential to invade nearby tissues and travel to other areas of the body. Once a malignant tumour
has travelled to other areas, these new tumours are known as metastases. Even though tumours
have the ability to metastasize to other body areas, the particular cancer is named after the
30
location in the body where it originated (e.g., breast cancer begins in the breast but can
metastasize outside the breast tissues).
1.6.2! Breast Cancer Biology
Breast cancer involves the growth of a malignant tumour that originates in any of the cells of
breast tissue. The breast consists of both fatty and connective tissue, and contains lobules (milk
production glands) and ducts (pathway for milk transport between lobules and nipple); breast
cancer can originate in any of these tissues. The nipple and areola are also susceptible to the
development of breast cancer, however development of cancer is more rare in this tissue. While
significantly less common, men can also develop breast cancer in the cells of their limited breast
tissue.
During replication, cells of the breast tissue can sometimes become mutated, which can result in
the irregular proliferation of cells. These irregular cells may create benign breast growths (e.g.,
cysts, fibroademonas, or benign tumours), or sometimes result in malignant breast tumours. Most
commonly, irregular cell growth is observed in the milk ducts, but can also occur in any other
breast tissue. These malignant tumours have the potential to grow in size as a solid mass, to
produce several smaller tumours in the breast, to spread into the lymph nodes in the underarm, or
to break off and travel to distant areas of the body, including other organs, the bones, or the
brain. So long as disease spread is confined to the breast and or nearby lymph nodes of the
underarm, the disease is considered to have the possibility of being cured. However, once the
tumour metastasizes into distant areas of the body, care is considered palliative.
Several types of breast cancer exist; however, some forms are better understood than others.
Certain tumour-specific biological factors play a role in determining tumour cell proliferation
and subsequent treatment. Circulating hormones are a key contributor to many types of breast
cancer. Estrogen is known to play a key role in the promotion of both normal and abnormal
breast cell growth, and an expression of estrogen receptors is believed to account for up to
approximately 70% of breast cancers (Russo & Russo, 2006; Shi, 2015). These breast cancers
are known as ER-positive, and can be treated with hormones to prevent recurrence. Progesterone
receptor (PR) positive tumours can also be treated with hormone suppression therapy, and are
known to have a better long-term prognosis than PR-negative breast cancer tumours (Arpino et
al., 2005). The over expression of human epidermal growth factor receptor 2 (HER2) in breast
31
tumour cells results in rapid cell division and causes faster tumour cell growth and replication
when left untreated. Over expression of the HER2 gene is exhibited in approximately 25% of
breast cancers and can be suppressed by the anti-HER2 antibody Trastuzumab (Hicks &
Kulkarni, 2008). Breast tumours that are ER-negative, PR-negative, and HER2-negative are
known as triple negative breast cancers.
1.6.3! Genetics and Mutations
In addition to the many forms of breast cancer, the mutation of certain genes are also known to
influence one’s risk of developing breast cancer. When functioning normally, BRCA1 and
BRCA2 produce suppressor proteins that function to repair DNA damage. When an individual
has a mutation in either of these genes, the person is at an increased risk of DNA not being
properly repaired, resulting in cells being more likely to develop additional genetic alterations
that increase one’s risk of developing breast or ovarian cancer. These harmful BRCA1 or BRCA2
mutations are heritable, and can be passed on from one’s mother or father. In the United States,
approximately 12% of women in the general population will develop breast cancer at some point
in their lives (Howlader et al., 2014). Instead, by the time they are 70 years old, women who
have a BRCA1 mutation are estimated to have a 55-65% chance of developing breast cancer, and
those with a BRCA2 mutations are believed to have a 45% chance of developing breast cancer
(Antoniou et al., 2003; Chen & Parmigiani, 2007). These are only estimates, and further research
may indicate differences in these numbers. Genetic testing is available to detect these mutations.
1.6.4! Incidence
Incidence refers to the total number of new cases of breast cancer being diagnosed and does not
include the number of deaths caused by the disease. The most recent statistics by the Canadian
Cancer Society (2016a) indicate that in 2016, it was estimated that 25,700 new cases of breast
cancer would be diagnosed, making up approximately 26% of the cancers diagnosed in all
women in Canada. As of 2009, it was believed that Canadian women had a one in nine chance of
developing breast cancer (Canadian Cancer Society, 2016a).
With the increasing population in Canada, the number of breast cancer diagnoses has
increased since the 1980s, however the increased incidence rate has paralleled the rate of
population growth (Canadian Breast Cancer Foundation, 2016). Data for 2016 estimated that
32
approximately 51% of new breast cancers diagnosed in Canadian women would be seen between
those aged 50 to 69 making this the most common age to receive such a diagnosis. It was also
estimated that approximately 32% of newly diagnosed breast cancers would be found among
those over 69 years of age, while roughly 17% of diagnoses would be for individuals less than 50
years old. While the incidence of breast cancer among Canadian women under 50 is relatively
lower than that among older women, the disease tends to be more aggressive among this cohort.
For example, only 12% of cancer related deaths in women over 60 are due to breast cancer,
while 22% of cancer related deaths among women 30-59 are due to breast cancer (Canadian
Breast Cancer Foundation, 2016).
The detection of breast cancer at earlier stages and greater prevention of deaths caused by
breast cancer can be attributed to both a better understanding of the biology of various breast
tumours, along with the development of modern screening technologies and treatment.
1.6.5! Diagnosis and Treatment
1.6.5.1! Staging
Breast cancer staging is significant to one’s prognosis and determines treatment plan.
Classification of breast cancer staging, according to the American Joint Committee on Cancer
(AJCC) is based on a combination of three key components (tumour size, regional node
involvement, and metastases status) (Edge et al., 2010, p. 347-376). Additional details regarding
hormone receptor status of the tumour tissue, human epidermal growth factor receptor 2 (HER2)
status, and the patient’s menopausal state and general health will also affect one’s treatment plan,
but are not related to staging. One’s stage is generally characterized based on TNM
classifications, which include:
Tumour (T) Classification:
T0 No evidence of primary tumour
Tis Carcinoma in situ
T1 Invasive tumour ≤ 2cm
T2 Invasive tumour > 2cm but ≤ 5 cm
T3 Invasive tumour > 5cm
33
T4 Invasive tumour of any size, either attached to or invading surrounding tissues outside the breast, axillary and mammary lymph nodes; includes inflammatory carcinoma
Lymph Node (N) Classification:
N0 No detectable disease spread to regional tissue; node negative
N1 Spread to underarm lymph nodes; not fixed to other nodes/tissue; node positive
N2 Spread to underarm lymph nodes; fixed to other nodes/tissues; node positive
N3 Spread to breast bone or distant underarm lymph nodes; fixed to other nodes/tissues; node positive
Metastasis (M) Classification:
M0 No evidence of disease outside the breast and nearby region (clinically or radiographically)
M1 Detectable disease spread > 0.2mm beyond the breast and nearby underarm region
This TNM information is then reduced into categories known as anatomic stages or prognostic
groups. A general breakdown of these stages includes:
Stages:
Stage 0: Tis N0 M0
Stage I: T1 N0 M0
Stage IIA: T0-2 N0-1 M0
Stage IIB: T2-3 N0-1 M0
Stage IIIA: T0-3 N1-2 M0
Stage IIIB: T4 N0-2 M0
Stage IIIC: Any T, N3 M0
Stage IV: Any T, Any N, Any M1
Additional further breakdowns of stage I includes invasion of micrometastases to nearby lymph
nodes.
34
1.6.5.2! Procedures
To determine TNM staging, various procedures are conducted to assess tumour size, node
involvement, and if necessary, check for distant metastases. Various procedures may be used for
staging upon determining the presence of breast cancer.
Axillary Lymph Node Dissection: An axially lymph node dissection involves the removal of
lymph nodes from the underarm to verify whether cancer cells have spread outside the breast to
this region. This is a more aggressive lymph node dissection procedure, compared to a sentinel
node biopsy, which involves removing fewer nodes.
Sentinel Node Biopsy: The sentinel node is the first lymph node in a collection of nodes which
lymph fluid surrounding a tumour passes through; cancer cells typically spread to the sentinel
node first. A sentinel node biopsy involves the injection of blue radioactive dye into the breast
tissue surrounding the tumour area, which passes to the sentinel nodes. The node/nodes stained
blue are removed and assessed for presence of cancer cells. If no cancer is detected, the
remaining nodes are left in place; if cancer is detected, further lymph node dissection is required.
This is less aggressive lymph node removal method.
Ultrasound, x-ray, CT scan, MRI, and Bone Scan: These procedures assess the presence of
distant disease metastases outside the breast and nearby lymph node region to other areas
including bones and other organs or tissues.
Measurement of Tumour Size: This measurement reflects the longest dimension of the tumour.
1.6.5.3! Other Tumour Characteristics
As briefly discussed in section 1.6.2 Breast Cancer Biology, in addition to TNM staging, certain
biological traits concerning tumour make up must be consider as they will affect one’s treatment
plan and subsequent disease-free survival. Understanding the ER and PR status of a tumour will
indicate the degree to which a tumour will likely respond to hormone suppression therapy. ER-
positive and PR-positive tumours typically have a greater association to longer periods of
disease-free survival when compared to any combination of tumours that are negative for either
one or both of these hormone receptors. The benefit of an ER-positive PR-positive tumour is due
to the ability to take hormone suppression drugs that will prevent either the release of hormones
35
or subsequent uptake by neoplastic breast cells, thereby reducing the chance of tumour cell
growth. Hormones can sometimes be an alternate to chemotherapy and sometimes even radiation
if a tumour is small enough, however, they can also be taken in conjunction with treatment plans
involving chemotherapy or radiation. Over-expression of HER2, while typically associated with
more aggressive, fast growing tumours can now also be suppressed using Trastuzumab
(Herceptin ®), thereby preventing the expression of the tumour cell growth factor and reducing
the risk of recurrence. Trastuzumab is given in conjunction with chemotherapy, as it is most
effective when combined. Tumours that do not display an over-expression of ER, PR, or HER2,
known as triple negative breast cancers, cannot be treated with hormone suppression therapy or
growth factor antibodies. Instead, the only treatment options are surgery, radiation and
chemotherapy. Typically, triple negative tumours are known to have a worse disease-free
survival rate in the one to four years following initial diagnosis due to a high rate of recurrence
when compared to tumours with positive ER/PR status and or HER2 status. However, long-term
disease-free survival after eight years among triple negative disease survivors appears to be very
good (Dent et al., 2007).
When the benefit of chemotherapy is uncertain, in some cases the Oncotype DX ® (a gene
expression profiling test) may be conducted to assess whether such treatment will provide a
significant additional survival benefit. In Canada, the Oncotype DX ® test is given only to
patients with breast cancer that is hormone receptor positive, HER2 negative, and node negative.
Such patients have the potential to be successfully treated using some combination of surgery,
radiation, and or hormone treatment. However, the Oncotype DX ® test can assess whether
chemotherapy may provide an additional survival benefit based on the analysis of a sample of
tumour tissue. The test assesses the biology of each tumour tissue sample by assessing the
activity of 21 genes, and using these results to calculate a recurrence score on a scale of 0 (no
likelihood of recurrence) to 100 (certain chance of recurrence) if chemotherapy is or is not given.
Chemotherapy’s benefit is sometimes debated given the immediate and long-term side effects
that a patient may suffer. As such, it is important to ensure that only patients who will likely
benefit from chemotherapy receive the treatment. The Oncotype DX ® test can help to increase
both the doctor’s and patient’s confidence in the decision of whether or not to use chemotherapy.
36
1.6.5.4! Treatment
Patients receive individual medical treatment plans based on the particular type of cancer they
have, and the characteristics and stage of the tumour. The goals of any method of treatment
include controlling the spread of and potentially curing the current cancer, and preventing its
future recurrence. When a cure is not possible, treatment may be provided to shrink and reduce
the spread of the disease, reduce cancer related symptoms, prolong life, and increase quality of
life. Treatment may involve local or systemic therapies.
1.6.5.4.1! Localized
Local treatment is targeted to the specific area where cancer is present. Surgery and radiation
therapy are considered local treatments. Depending on cancer stage, one or both treatments may
be used.
Surgical: Various surgical procedures exist for treating breast cancer. The end goal of surgery is
to remove the tumour and achieve clean surgical margins with no disease tissue remaining.
Breast conserving surgery (lumpectomy) is often recommended when a tumour is small and
enough healthy breast tissue will remain to achieve a desirable cosmetic effect. A lumpectomy is
generally accompanied by a sentinel node biopsy and or an axillary node dissection.
Alternately, a mastectomy may be conducted to remove the entire breast. Typically, a
mastectomy is carried out for a larger tumour or higher stage, or if someone has a small tumour
but wishes to avoid radiation therapy. The 3 options for mastectomy include:
•! Modified Radical Mastectomy: The entire breast and nipple, some of the lymph nodes,
and the chest muscle lining (fascia) are removed. Typically, the nerves and chest muscles
are left intact and not removed. This is the most common form of mastectomy today for
invasive tumour treatment. Sentinel node biopsies have replaced more extensive removal
of lymph nodes.
•! Radical Mastectomy: This is no longer the typical surgical procedure for a mastectomy
today as it does not provide a significant survival benefit over the modified version and
can greatly reduce one’s range of motion and posture, but may be used if there is tumour
extension into the chest wall muscle. The entire breast and nipple, all the lymph nodes,
37
and chest muscle fascia are removed. Much of the chest muscle and nerves are also
removed.
•! Total Mastectomy: The least invasive mastectomy, it involves removal of the entire
breast, nipple, and chest muscle fascia. The chest muscle and nerves along with lymph
nodes are left intact. The skin of the breast may also be spared if the patient is opting for
a breast reconstruction. This procedure is either conducted prophylactically or if the
cancer is early stage or non-invasive.
Radiation: For treatment of breast cancer, radiation may either be used following surgery and or
chemotherapy to ensure that any residual tumour cells at the original tumour site are destroyed,
or to shrink tumours either in the breast or in areas of the body with distant disease spread in
patients with metastatic stage IV disease. While in general radiation for local control and
prevention acts as a lifesaving measure, at times patients do not undergo radiation because the
benefits they receive will be outweighed by secondary effects that may develop as a result of
radiation including damage to the cardiac muscle or increased risk of a secondary cancer.
Radiation may be avoidable in those who presented with a very small early stage tumour, and/or
those who subsequently had a mastectomy for a very small tumour, and/or those who had a high
dose of chemotherapy for a small tumour with favourable characteristics. The typical course of
radiation therapy for early breast cancer treatment in Canada is external beam therapy that
involves daily exposure to targeted doses (between 16 and 25) of radiation aimed specifically to
the tumour site.
1.6.5.4.2!Systemic
Systemic therapy is aimed to care for the whole body, preventing tumour recurrence at both the
original site and distant sites. Chemotherapy and hormone therapy are systemic treatment options
for breast cancer. Depending on the stage and characteristics of one’s tumour, neither, one, or
both of these treatments may be necessary.
Chemotherapy: Chemotherapy is used to kill cancer cells that are actively dividing in order to
damage the cells and reduce their ability to reproduce and spread. Several different
chemotherapy regimes may be used for breast cancer depending on the tumour staging and
characteristics, but most are given in a regimented cycle over a specific number of weeks or
months meant to interrupt the cancer cell division and replication cycle. Chemotherapy can be
38
used adjuvantly to prevent recurrence locally and distantly following surgery and tumour
removal by destroying any remaining cancer cells, or neoadjuvantly to shrink a tumour prior to
surgery and kill any microscopic distant metastases. Neoadjuvant chemotherapy is often used in
breast cancer cases where the tumour is too large to conduct surgery without causing significant
damage to the surrounding tissue, when the tumour has affected too much tissue to properly
close the surgical wound, or at times when a surgery date cannot be scheduled soon enough.
Neoadjuvant therapy also allows for the measurement of tumour response to the chemotherapy
drugs – if there is no response to the treatment, it can be stopped and a patient can proceed to
alternate modes of treatment such as surgery, radiation, or hormone treatment. Neoadjuvant and
adjuvant chemotherapy are equivalent in terms of survival outcome. Chemotherapy may not
always be necessary depending on the particular tumour characteristics and may be avoided in
cases where its immediate side effects or potential future complications (e.g., reduced cardiac
function, organ damage, increased risk for secondary cancers) may outweigh the benefits it
provides.
Hormone Therapy: Hormone therapy is a long term, non-cyclic treatment available as treatment
to prevent breast cancer recurrence. The typical hormone targeted by such therapy is estrogen.
These drugs may be used to prevent or slow the growth of cancer cells that are otherwise fueled
by the presence of estrogen. Depending whether a breast cancer patient is pre- or post-
menopausal, tamoxifen may be given to block hormone binding to estrogen specific receptor
sites, or an aromatase inhibitor may be given to prevent estrogen production.
Targeted: As a greater understanding of the specific cells associated with particular cancers
develops, it is possible to create therapies that target these particular aberrant cells fueling the
growth of particular tumours. These treatments, known as targeted therapies, are meant to
specifically attack cancer cells while generally sparing the normal cells. Targeted therapy can be
used in the treatment of some breast cancers where there is an over expression of HER2
receptors (HER2 positive) on the cancer cell surface that facilitate rapid cell proliferation.
Trastuzumab can be used to block HER2 receptors in patients with HER2 positive breast cancer,
and thereby slow or prevent the proliferation of cancer cells. Trastuzumab is given intravenously,
and is given in conjunction with chemotherapy drugs.
39
1.6.5.5! Prognosis
Disease prognosis can play a role in the way patients and their caregivers cope with the
diagnosis. For breast cancer, several factors affect their prognosis. Stage is a key prognostic
factor such that a diagnosis between stages I to III, even with increasing severity, still allows for
treatment options intended to be curative. A stage IV breast cancer diagnosis means treatment
will be to prolong life, but will ultimately not cure the disease. In general, across all stages, the
higher the stage at time of initial diagnosis, the worse one’s prognosis, but this does not mean
long-term survival is not possible. This indicates that greater tumour size and or greater nodal
involvement is associated with a greater chance of local or distant disease recurrence.
Additional factors outside TNM staging include grade (rate of tumour cell reproduction) and
type (e.g., ER/PR/HER-status) of tumour. Higher grade, poorly differentiated cell structure, or
inflammatory type of tumour are typically faster growing and have a worse prognosis. Greater
invasion of the blood vessels around the tumour results in a generally poorer prognosis.
Hormone receptor status affects prognosis, such that greater estrogen receptor status is associated
with a better prognosis, while breast tumours with an overall negative hormone receptor status
typically have a greater risk of recurrence. If distant metastasis does occur, those with hormone
receptor positive tumours typically exhibit first metastases to the bone, while those with hormone
negative tumours exhibit first metastases more commonly to the lung or liver. A more recently
understood prognostic factor is HER2 status. Until recently when the Trastuzumab drug became
available, HER2 positive tumours had a worse prognosis because they exhibited over expression
of a growth promoting protein produced on the tumour cell surface. Trastuzumab blocks HER2
positive receptors to slow and reduce the division of the corresponding cells. Finally, age at
initial diagnosis affects prognosis, such that those of ≤ 35 years of age typically have a worse
prognosis including greater risk of recurrence given the often more aggressive, higher-grade
nature of the tumour, than that typically seen among breast cancer patients who are older and
postmenopausal at time of diagnosis.
40
Chapter 2!
! General Methods This research collectively assessed how coping ability among patients and caregivers (medical
and familial) facing a common stressor (cancer) was influenced by chronotype, sleep quality, and
personality. Individuals in different positions but facing a common stressor were hypothesized to
react in ways specific to their roles. Both cross-sectional and longitudinal data collection
schedules were used. Oncology staff completed cross-sectional assessments of their coping
(based on measures of burnout, compassion fatigue (CF), and compassion satisfaction (CS)).
Patients and caregivers partook in longitudinal data collection schedules. Among each group
involved, specific surveys were used to collect data. Using these questionnaires, we developed an
assessment of how individuals in each group coped with cancer in relation to their chronotype,
sleep quality, and personality. The following is a description of the individual groups involved in
this study, and the various questionnaires used to assess participants across the study.
2.1! Procedures
2.1.1! Nursing Study
A literature search was conducted to verify a gap in the research on the influence of chronotype
and sleep quality on burnout among oncology nursing staff. A set of surveys were compiled to
assess demographic information and job satisfaction, chronotype, sleep quality, and burnout.
Ethics approval was received from Sunnybrook Health Sciences Centre (SH) and Princess
Margaret Hospital (PMH) to approach and conduct research among ambulatory oncology nurses.
2.1.2! Hospital Staff
A literature search was conducted to verify a gap in the research on the influence of chronotype
and sleep quality on CS and CF among oncology hospital staff providing cancer patient care. A
set of surveys was compiled to assess demographic information, professional quality of life,
chronotype, sleep quality, personality and job satisfaction. The professional quality of life
measure assesses CS/CF. Ethics approval was received from Sunnybrook Health Sciences Centre
41
(SH) to request voluntary participation from oncologists, oncology nurses, radiation therapists,
pharmacists, and pharmacy technicians working at the Sunnybrook Odette Cancer Centre. Upon
receiving ethics approval and prior to reaching out to participants, the questionnaire package was
created in an online survey format using Survey Monkey.
2.1.3! Patients and Caregivers
An extensive literature search was conducted to assess chronotype, sleep quality, and coping
among cancer patients and their familial caregivers. A gap in the literature was found on the
influence of chronotype and sleep quality as factors influencing coping, with personality as a
potentially mediating factor, among breast cancer patients undergoing chemotherapy and their
spousal caregivers. Questionnaire packages were assembled to assess demographic information,
chronotype, sleep quality, recalled coping across the day on various domains, and use of coping
behaviours. These were completed at baseline, midpoint and endpoint. A daily log was created
so participants could track their perceived changes in coping across the day on a Likert scale of 1
(poor) to 5 (excellent). Ethics approval was received from SH to approach early stage (I, II, and
III) breast cancer patients, and their spousal caregivers for participation. While spousal
caregivers were only granted ethics approval to participate if doing so in conjunction with their
partner, patients could participate with out without a spousal caregiver so long as they were
going to receive chemotherapy and were given a breast cancer diagnosis between stages I to III.
2.2! Instruments
2.2.1! General Questionnaires
Horne-Östberg Morningness Eveningness Questionnaire (MEQ): a 19-item assessment tool that
measures M or E chronotype. Each question is scored in the range of 0-6 points, depending on
question. Individual items are tallied to produce a single total score in the range of 16 and 86.
Scores in the range of 16 to 41 indicate E-tendency; 42-58 indicates a chronotype of neither (N)
morning nor evening, but likely more a tendency towards a preference for midday; 59-86
indicates a preference for M.
Pittsburgh Sleep Quality Index (PSQI): a 19-item assessment tool that provides a global sleep
quality index. Each question is scored on a scale of 0-3 (0 indicates no particular sleep
42
disturbance based on that dimension). Following data transformation, a global sleep quality
rating is provided between 0 and 21. In the general population, a global score of ≤ 5 indicates
good sleep quality, while scores > 5 indicate poor sleep quality. Further studies have suggested
that among cancer patients, a global score ≤ 8 indicates good quality sleep, while > 8 indicates
poor sleep quality (Carpenter & Andrykowski, 1998; Vargas, Wohlgemuth, Antoni, Lechner,
Holley & Carver, 2010).
2.2.2! Group Specific Questionnaires
2.2.2.1! Nursing Staff
Copenhagen Burnout Inventory (CBI): a 3-scale, 19-item measure that assesses personal, work-
related, and client-related burnout. A corresponding burnout score is provided for each category.
Each component variable is rated with a score of 0, 25, 50, 75, or 100. The total score for each
burnout category is based on the mean value of the component questions in that section. On each
scale, a higher score is indicative of a greater degree of burnout for that specific dimension
(Kristensen, Borritz, Villadsen & Christensen, 2005).
2.2.2.2! Oncology Staff
Job Satisfaction Survey (JSS): a 9-subscale, 36-item measure of satisfaction with a range of
conceptually distinct topics relating to one’s job. Developed specifically for use with human
service, public, and nonprofit sector organizations, the scale provides 9-component scores, and
one total composite score. Individual response choices are scored on a scale of 1 (disagree very
much) to 6 (agree very much); negatively worded items require reverse scoring. Individual
component scores are summed to produce the value for each subscale between 4 (dissatisfied)
and 24 (satisfied), with those between 12-16 indicating ambivalence. All subscale values are
totaled to provide a total composite score between 36 (dissatisfied) and 216 (satisfied), with
those between 108-144 indicating overall ambivalence (Spector, 1985).
Ten Item Personality Inventory (TIPI): a 10-item scale that produces five outcome scores:
Extraversion, Agreeableness, Conscientiousness, Emotional stability, and Openness to
experience. Each outcome score is based on the mean score of two individual questions
describing the particular trait, scored on a scale of 1 (disagree strongly) to 7 (agree strongly);
reverse scoring is used where indicated by questionnaire scoring rules. The TIPI is used as a
43
short version measure to assess the Big Five Personality Traits. The five subscales range as
follows:
•! Extraversion: reserved, quiet to extraverted, enthusiastic
•! Agreeableness: critical, quarrelsome to sympathetic, warm
•! Conscientiousness: dependable, self-disciplined to disorganized, careless
•! Emotional Stability: anxious, easily upset to calm, emotionally stable
•! Openness to Experience: conventional, uncreative to open to new experiences, complex
(Gosling, Rentfrow, & Swann Jr., 2003).
Professional Quality of Life (ProQoL), Version 5: a 3-subscale, 30-item measure of CS, burnout,
and STS. The 3 scales are distinct and cannot be combined to produce a total score. Burnout and
STS are the 2 subcomponents of CF, the counterpart to CS. Each question is scored on a scale of
1 (never) to 5 (very often). Reverse scoring is applied to applicable questions as indicated by the
scoring guide. Total scores for each subscale are achieved by summing the individual questions,
to achieve raw scores in the range of 10 to 50. Each subscale follows the same raw cutoff score
scale: 22 or less indicates low CS, burnout, or STS; between 23 and 41 indicates average CS,
burnout, or STS; 42 or more indicates high CS, burnout, or STS (Stamm, 2010).
2.2.2.3! Patients and Caregivers
Daily coping logs: these were devised for the purposes of this research for participants to keep
track of their coping in the moment across the day. Coping logs were intended for the use by
patients undergoing chemotherapy and their caregivers. Coping is assessed on a scale of 1 (poor)
to 5 (excellent). The three time slots for coping assessment were: 7:00 am to 11:00 am, 1:00 pm
to 5:00 pm, and 7:00 pm to 11:00 pm. The timing slots were assigned as such to reflect a
morning, afternoon, and evening coping time, and to allow for a reflection of coping across the
day. Participants were asked to give a rating of their coping “in the moment” at any point in the
time slot, but to skip/leave blank any coping times that had been missed, even if just by a few
minutes. Participants were told not to reflect back on their coping, but rather give a rating of how
he or she felt in the moment. A specific definition of coping was not provided, instead
participants were asked to give a coping rating based on their own personal understanding of
coping and to provide a rating in terms of everything being faced and in light either their own
44
diagnosis or that of their partner. No coping definition was given in order to remove any bias
towards searching for positive or negative coping traits.
2.2.2.3.1!Questionnaire Package
University of Toronto Inventory of Morningness and Eveningness (UTIME) Questionnaire: asks
respondents to report their recalled best, average, and worst performance times for various
activities they could be asked to perform across the day. UTIME data can be used comparatively
against scores of other questionnaires. Against MEQ scores, the UTIME can be used to
understand how self-reported preference for M versus E (based on MEQ score) is associated with
one’s own recalled performance across the day on various tasks (UTIME score). Against daily
coping log data, the UTIME can be used to assess how one’s retrospective memory of emotional
stability on various coping-based tasks compares to in the moment ratings.
Big Five Aspects Scales (BFAS): a 100-item measure that produces 10-outcome scores that
represent the 2 distinct but related aspects within each of the Big Five domains of personality.
Each question is scored on a scale of 1 (strongly disagree) to 5 (strongly agree). Reverse scoring
is applied where indicating by the scoring guide. Completed questions within each of the 10
scales are summed and averaged to produce a score. The Big Five scores are computed by
averaging a score for the two aspects within each domain. Higher scores (both the distinct pairs,
and the overall domain) indicate a greater propensity for displaying that personality trait
(DeYoung, Quilty, & Peterson, 2007).
Brief-COPE: a modified version of the original 14-subscale, 60-item COPE scale. The Brief
COPE is a 14-subscale, 28-item measure of coping that assesses multiple responses known to be
relevant to coping. The scale does not provide a breakdown of whether the coping responses fall
into categories such as problem-focused or emotion-focused, but rather presents stressor
response-types that may be relevant to either effective or ineffective coping. Each question is
scored on a scale of 1 (I haven’t been doing this at all) to 4 (I’ve been doing this a lot). To
achieve a score for each subscale, take the mean of the two corresponding questions for that
particular scale. The score for each subscale can range between 1 and 4, with higher scores
indicating a greater tendency to engage in the particular coping technique described by the
particular subscale. This scale does not provide a total or overall score that represents a dominant
coping style (Carver, 1997).
45
2.3! Statistics
SPSS (Statistical Package for the Social Sciences) version 20.0, 23.0, and 24.0 (Armonk, NY)
for Mac were used to perform various descriptive and analytical tests on the data.
46
Chapter 3!
Bellicoso, D., Ralph, M. R., & Trudeau, M. E. (2014). Burnout among oncology nurses: Influence of chronotype and sleep quality. Journal of Nursing Education and Practice, 4, 80-89. doi: https://doi.org/10.5430/jnep.v4n8p80
! Burnout Among Oncology Nurses: Influence of Chronotype and Sleep Quality
3.1! Abstract
The study sought to clarify the impact of chronotype and sleep quality on feelings of personal,
work-related, and client-related burnout among ambulatory care oncology nurses following
regular dayshift work schedules. Ninty-four participants from two Toronto, ON hospitals took
part. The Horne-Östberg Morningness Eveningness Questionnaire, Pittsburgh Sleep Quality
Index, and the Copenhagen Burnout Inventory were used to assess the impact of chronotype and
sleep quality, together with subjective measures of job and place of employment satisfaction and
work stressfulness on burnout. Findings showed that participants reporting greater tendency for
evening-type or neither-type chronotype, and/or poor sleep quality had significantly higher levels
of personal, work-related, and client-related burnout than individuals with either a morning
tendency and/or good sleep quality. Work stressfulness also contributed to elevated burnout.
Working at one’s optimal time and obtaining good quality sleep contributes to decreased
burnout. Future studies should consider the effect of chronotype and sleep quality on mediating
burnout among shift work oncology nurses. When creating nursing work schedules, employees’
chronotype and associated sleep quality should be considered to achieve decreased burnout,
optimal performance, and potentially increased employee retention and patient care quality and
satisfaction.
Keywords: Burnout, Chronotype, Circadian Rhythms, Nursing, Oncology, Sleep Quality
47
3.2! Introduction
The impacts of chronotype and sleep quality on burnout have yet to be studied and require
greater examination. Chronotype reflects one’s performance capability as it changes throughout
the day and the associated preference for morning (M) vs. evening (E) activities (Horne &
Östberg, 1976; Nielsen, 2010). Sleep quality refers to both quantitative (e.g., sleep duration and
latency, and number of arousals) and subjective (e.g., degree of restfulness) aspects of sleep –
however, inter-individual differences exist in the elements of sleep composition and their relative
importance (Buysse, Reynolds III, Monk, Berman, Kupfer, 1988). Healthcare workers are
particularly prone to burnout, and oncology nurses in general are at great risk, as the uncertainty
associated with patient outcome will elevate stress contributing to poor quality sleep (Chen &
McMurray, 2001; Cubrilo-Turek, Urek & Turek, 2006; Potter et al., 2010). Given the negative
feelings associated with burnout, it is important to understand how chronotype and sleep quality
both contribute to, and can be used to alleviate burnout.
Chronotype is associated with a range of outcomes. For example, research has shown systematic
differences in blood oxygenation levels in the brain across the day that relate with one’s
chronotype (Peres et al., 2011). In terms of measurable behaviour, one’s executive functioning
and alerting reactions have also shown daily fluctuations that correspond to chronotype
(Matchock & Mordkoff, 2009). Chronotype is linked with the time of sleep onset (Horne &
Östberg, 1976; Roenneberg, Wirz-Justice & Merrow, 2003). Significant differences in wake and
sleep onset times exist between M and E type individuals, with E types going to bed
approximately 99 minutes after M types and awaking approximately 114 minutes after them,
without significant differences in each group’s sleep duration (Horne & Östberg, 1976).
Adequate and good quality sleep is a basic human need that is restorative both physically and
cognitively (Karagozoglu & Bingöl, 2008; Stepanski, 2002). Like chronotype, sleep quality is
also known to affect a range of measurable outcomes. In relation to health, poor sleep quality has
been shown to increase many negative physical and mental health problems (Araghi et al., 2013;
Fujiwara, 2013). In terms of executive function, good sleep quality has been shown to increase
academic performance and executive functioning, while poor sleep quality, particularly among
elderly individuals, is associated with reduced cognitive performance (Lemma, Berhane, Worku,
Gelaye & Williams, 2014; Miyata et al., 2013). While proper and adequate sleep is important for
48
everyone, many people do not obtain the necessary amount of slow-wave and REM sleep each
night, with approximately 30% of the general population reporting sleep problems (van
Litsenburg et al., 2011). While the concept of sleep quality is difficult to define objectively, it
does involve certain key quantitative and subjective constructs such as sleep duration and
restfulness, respectfully (Buysse, Reynolds III, Monk, Berman, Kupfer, 1988). Shortened sleep
duration impedes physical and cognitive restoration, and decreases the following day’s
wakefulness, while regularly oversleeping can increase tiredness due to the strenuous and energy
consuming nature of REM sleep (Bonnet & Arand, 1995). Sleep restfulness relies both on the
quality of one’s previous waking period and its ability to generate proper homeostatic sleep
drive, and how one perceived their personal level of tiredness following a sleep period
(Bersagliere et al., 2012). Thus, sleep quality – based both on its quantitative and subjective
aspects, is important in allowing a person to go through the sleep stages necessary for their body
to be properly restored, well rested, and ready for the next day. However, while the importance
of adequate and good quality sleep has been studied abundantly, many conventional societal
practices interfere with achieving proper sleep quality when they conflict with the hours
associated with innate chronotype.
Work schedules in continuous disharmony with one’s circadian rhythm induce stress, which
negatively affects sleep quality (Karagozoglu & Bingöl, 2008). Disturbed sleep and its associated
fatigue make it difficult to carry out daily activities and work tasks. Among nurses, sleep debt
induces stress independently of the high stress work situations they face on a daily basis,
oftentimes leading to feelings of extreme emotional and physical fatigue, decreased cognitive
function, weariness, and exhaustion, which together are known as burnout (Brand et al., 2010;
Vela-Bueno et al., 2008).
Burnout is determined from feelings of emotional and physical fatigue or exhaustion, cognitive
weariness, and chronic energy resource depletion due to continued exposure to stress. However,
there is no concrete definition of the term, nor is there a standardized general procedure from
which to obtain a burnout diagnosis (Brand et al., 2010; Korczak, Huber & Kister, 2010). A
common hypothesis is that stress-induced sleep disturbances over time compound to produce
mental and physical exhaustion, which are generally agreed to be key contributors to burnout
(Söderström, Ekstedt, Åkerstedt, Nilsson & Axelsson, 2004). In line with this hypothesis, the
Copenhagen Burnout Inventory (CBI) used in the present study suggests that at the core of
49
burnout are fatigue and exhaustion (Kristensen, Borritz, Villadsen & Christensen, 2005). The
CBI subscribes to the definition of burnout that suggests it is a state of emotional, physical, and
mental exhaustion resulting from prolonged involvement in emotionally demanding work
situations (Schaufeli & Greenglass, 2001). In healthcare organizations, burnout is a significant
phenomenon due to its demonstrated negative effects on patient satisfaction and safety, and on
nurse retention and turnover (Potter et al., 2010). Among clinical care oncology nurses, higher
burnout rates have been found compared to hospice care oncology nurses, suggesting that
working in a hospital setting with the goal to heal sick patients rather than care for those that are
dying, may increase stress levels, and cause subsequent sleep disturbances which may contribute
to the gap in burnout reports (Ostacoli et al., 2010).
While many studies have examined the triggers and protective factors for burnout, the influence
of chronotype and sleep quality on this phenomenon have not been studied. We sought to
demonstrate the potential for chronotype and sleep quality to mediate feelings of burnout among
ambulatory care oncology nurses working in a hospital setting on a fixed Monday to Friday
daytime work schedule by using a set of standardized questionnaires to assess the influence of
chronotype and sleep quality on burnout. Our objectives were to understand the separate links
between good versus bad sleep and one’s ratings of the various types of burnout, and the link
between chronotype and burnout. It was also important to understand the differential
contribution of various predictors (chronotype, sleep quality, job satisfaction, overall place of
employment satisfaction, and work stressfulness) on the various types of burnout. It was
predicted that M types and respondents with better sleep would exhibit less overall burnout
compared to E types and people reporting poor sleep quality. It was also predicted that in
addition to one’s chronotype and sleep quality, stressfulness and satisfaction ratings relating to
one’s job would differentially influence each type of burnout.
3.3! Materials and Methods
3.3.1! Participants
Registered ambulatory care oncology nurses working at one of two primary care hospitals in
Toronto, Ontario, Canada – University Health Network Princess Margaret Cancer Centre (PMH),
and Sunnybrook Health Sciences Centre Odette Cancer Centre (SB) – were recruited to
participate. A total of 64 nurses completed questionnaires with useable results (PMH, n = 42, all
50
female, age 27 – 63, mean age 45.29 ± 10.18; SH, n = 22, 21 females, 1 male, age 28 – 65, mean
age 50.54 ± 11.70). The mean ages for these groups are not significantly different (see Table
3.1). Inclusion criteria were as follows: participants must be ambulatory care oncology nurse
responsible for the care and support of oncology patients, be currently employed in one of the
hospitals, follow a regular dayshift work schedule Monday to Friday, and not be connected with
the present study. The study was granted approval by the ethics committees of both hospitals.
Informed consent was obtained from all participants prior to giving them the questionnaire
package. Study participation was voluntary.
3.3.2! Procedure
A preliminary meeting was held to present the aims of the study, and an anonymous survey
package was distributed to nurses who agreed to partake in the present study. Time was given in
the meeting to complete the survey package. Nurses provided demographic information and job
descriptions (age, gender, job satisfaction, place of employment satisfaction, and stressfulness of
current job) in a background information survey.
3.3.3! Measures
The Horne-Östberg Morningness Eveningness Questionnaire (MEQ) is a 19-item measure that
was used to assess tendency for M or E chronotype. Each component question is scored on a
scale in the range of 0-6 depending on the question, based on guidelines provided by the
questionnaire’s authors, to obtain a global score between 16 and 86. Based on one’s global score,
a rating of definitely E- (16 to 30), moderately E- (31 to 41), neither- (42 to 58), or moderately
M- (59 to 69) or definitely M-type (70-86) is given. The Pittsburgh Sleep Quality Index
(PSQI) is a 19-item measure that yields a global sleep quality index. Each question is scored on
a frequency scale of 0 to 3 (0 indicates no occurrence of a particular sleep disturbance). A global
sleep quality rating is provided on a scale of 0 to 21 (following data transformations), where a
global score of ≤ 5 indicates good sleep quality, while > 5 indicates poor sleep quality. The
primary outcome variables – personal, work-related, and client-related burnout – were measured
using the Copenhagen Burnout Inventory (CBI) which is a 3-scale, 19-item measure used to
assess these 3 types of burnout that provides a corresponding burnout score for each category.
Each component question is rated with a score of 0, 25, 50, 75 or 100. Total scores for each
burnout category are based on the mean of the component questions pertaining to that section.
51
The higher a scale’s associated score, the greater the degree of burnout that is associated with
that specific dimension.
3.3.4! Statistical Analysis
A priori tests indicated that to detect a moderate effect size, with 80% power and α = .05, sample
size calculations for a one-tailed test required 27 subjects. Statistical analyses were performed
using SPSS (Statistical Package for the Social Sciences) version 20.0 for Mac. Analyses were
run as follows: demographic and test variables of the two hospital groups of nurses were
compared using unpaired t -tests. Each group of nurses was analyzed by their MEQ
classification, and again by the PSQI good or poor quality sleep rating using t -tests to look for
burnout differences. Correlations were run between the test variables to help determine where
significant associations lay between different items. Lastly, multivariate regression analyses were
run to clarify the differential contribution of the predictors of interest to the outcome measures. A
p-value < .05 was accepted as statistically significant.
3.4! Results
Of the 94 nurses who volunteered to participate, 64 (68.09%) provided useable data, resulting in
a strong effect size with 98.96% power (α = .05, one-tailed test). At PMH, 42/60 (70.00%)
returned useable data, while 22/34 (64.71%) with useable data were returned from SH. None of
the demographic characteristics and questionnaire response ratings between the two groups are
statistically different for these categories, indicating that the two groups are well matched (see
Table 3.1).
52
Table 3.1 Demographic characteristics and questionnaire response ratings of participants Nurse group PMH (n = 42) SB (n = 22) Frequency Frequency Gender (M:F) 0:42 1:21 Mean (SD) Mean (SD) t-test* Age in years 45.29 (10.18) 50.54 (11.70) NS Oncology nurse experience (yrs) 15.10 (11.64) 14.19 (10.25) NS Job satisfaction 3.46 (1.09) 3.93 (0.76) NS Overall place of employment satisfaction
3.70 (1.01) 4.05 (0.95) NS
Work stressfulness 3.82 (1.03) 3.68 (1.09) NS MEQ 57.02 (8.47) 59.82 (9.52) NS PSQI 6.60 (3.59) 5.50 (2.20) NS CBI
Personal 50.20 (19.63) 44.32 (16.94) NS Work 51.70 (22.16) 44.48 (16.52) NS
Client Related 21.22 (19.13) 21.02 (14.52) NS Note. NS = not significant. *p < .05, 2-tailed.
53
The three separate CBI categories each conceptualize burnout as a continuous variable on a scale
of 0 to 100. Table 3.2 shows burnout ratings among respondents with good (≤ 5) or poor (> 5)
sleep quality as rated by the PSQI, and burnout ratings among MEQ types. Given the low
number of moderately E types (n = 2) and definitely M types (n = 4), these two categories have
been added to the neither type and moderately M type groups, respectively. This break point also
allows for a nearly equal frequency distribution of participants with 48.44% of respondents
qualifying as moderately E type or neither type, and 51.56% qualifying as M types. It is
important to note that while both of these conceptualizations of burnout in Table 3.2 are broken
down into groups of n = 31 and n = 33, the individuals making up these divisions are not
necessarily the same across both sets of data.
54
Table 3.2 Comparison of burnout ratings between respondents with good and bad sleep quality and between MEQ types
Sleep Quality (PSQI) MEQ Good (≤ 5)
Bad (> 5) t-test Moderately
& Definitely M type
N & Moderately E type
t-test
CBI Personal 39.65 (18.81) 56.19 (15.13) -3.88*** 42.17 (21.16) 54.57 (13.58) 2.78**
Work 42.51 (21.77) 55.52 (17.41) -2.65** 42.86 (22.23) 55.99 (16.38) 2.68** Client Related 15.59 (15.51) 26.52 (17.94) -2.60* 15.15 (16.23) 27.69 (16.82) 3.03**
*p < .05, 2-tailed. **p < .01, 2-tailed. ***p < .001, 2-tailed.
55
As shown in Table 3.3, MEQ and PSQI scores are not significantly correlated. Significant
correlations exist between all CBI scores and the associated MEQ and PSQI scores. Stressfulness
associated with place of employment is significantly correlated to all burnout subscales, while
job satisfaction is significantly correlated only with work related burnout. Overall place of
employment satisfaction was not correlated with chronotype, sleep quality, or any of the burnout
ratings.
56
Table 3.3 Analysis of bivariate correlations for participants’ questionnaire response ratings CBI MEQ PSQI Personal Work Client MEQ 1.00 -.19 -.33** -.26* -.36** PSQI -.19 1.00 .52** .43** .39** Job Satisfaction .05 -.12 -.21 -.38** -.04 Overall place of employment satisfaction
-.12 .11 -.07 -.23 .08
Work Stressfulness .03 .26* .62** .66** .32** *p < .05, 2-tailed. **p < .01, 2-tailed.
57
To clarify the differential contribution of the predictors of interest on the various categories of
burnout, multivariate regression analyses were run. Table 3.4a shows the relationship between
personal burnout and each predictor. Statistical model #2 (Step 2) yields a satisfactory proportion
of variances explained, as indicated by the associated R2 value (.61). Table 3.4b shows the
relationship between work related burnout and each predictor. For this relationship, statistical
model #3 (Step 2) yields a satisfactory proportion of variances explained as indicated by the
associated R2 value (0.62). Table 3.4c shows the relationship between client related burnout and
each predictor. For this relationship, statistical model #2 (Step 2) yields a satisfactory proportion
of variances explained as indicated by the associated R2 value (0.31). Note that the R2 value
explaining the proportion of variance for client related burnout due to these predictors is much
lower than that seen for personal and work related burnout. This suggests that other factors not
being considered in this study may influence client related burnout.
58
Table 3.4a Summary of hierarchical multivariate regression analysis for variables predicting personal burnout among oncology nurses (N = 64) Variables b SE b ßa Step 1
Constant 60.91 14.56 MEQ -0.52 0.23 -.24* PSQI 2.77 0.63 .47***
Step 2 Constant 34.52 11.93
MEQ -0.61 0.18 .29*** PSQI 1.90 0.50 .32***
Work Stressfulness 9.88 1.51 .55*** Note. R2 = .33 for step 1; ΔR2 = .28 for step 2. a standardized coefficient *p < .05, 2-tailed. ***p < .001, 2-tailed.
59
Table 3.4b Summary of hierarchical multivariate regression analysis for variables predicting work related burnout among oncology nurses (N = 64) Variables b SE b ßa Step 1
Constant 58.18 17.39 MEQ -0.43 0.27 -.18 PSQI 2.52 0.74 .39***
Step 2 Constant 26.07 13.70
MEQ -0.54 0.20 .23** PSQI 1.46 0.58 .23*
Work Stressfulness 12.04 1.74 .61*** Step 3
Constant 45.41 14.72 MEQ -0.52 0.19 -.22** PSQI 1.37 0.55 .21*
Work Stressfulness 11.20 1.68 .57*** Job Satisfaction -4.69 1.68 -.23**
Note. R2 = .22 for step 1; ΔR2 = .35 for step 2; ΔR2 = .05 for step 3. a standardized coefficient *p < .05, 2-tailed. **p < .01, 2-tailed. ***p < .001; 2-tailed.
60
Table 3.4c Summary of hierarchical multivariate regression analysis for variables predicting client related burnout among oncology nurses (N = 64) Variables b SE b ßa Step 1
Constant 44.00 14.23 MEQ -0.59 0.23 -.30* PSQI 1.84 0.63 .33**
Step 2 Constant 32.03 14.77
MEQ -0.63 0.22 .32** PSQI 1.44 0.62 .26*
Work Stressfulness 4.49 1.88 .27* Note: R2 = .24 for step 1; ΔR2 = .07 for step 2. a standardized coefficient *p < .05, 2-tailed. **p < .01, 2-tailed.
61
According to the experimental hypothesis, chronotype and sleep quality significantly influenced
all the CBI subscale scores. In addition, each respondent’s perception of work stressfulness
significantly influenced each CBI subscale score. Job satisfaction was a significant factor
influencing the work-related CBI subscale. These findings correspond with the analysis of
bivariate correlations indicated in Table 3.3.
3.5! Discussion
Chronotype and sleep quality are known to influence a wide range of important aspects of daily
life such as physical and mental health, blood oxygenation, and cognitive functioning to name a
few (Peres et al., 2011; Fujiwara et al., 2013; Miyata et al., 2013). However, the influence of
one’s chronotype and sleep quality on the degree of burnout has not been greatly explored. Given
the high degree of burnout associated with several careers, and particularly with oncology
nursing, it was important to consider the impacts of chronotype and sleep quality on burnout.
This information will be informative to the design of schedules and approaches that reduce
burnout among members of this group. The present study investigated the contribution of
chronotype and sleep quality, together with situational factors, on symptoms of burnout among
ambulatory care oncology nurses working in outpatient clinics in one of two hospitals in
Toronto, Ontario, Canada. It was found that innate chronotype and sleep quality strongly
predicted burnout, such that individuals with an M type tendency, and respondents with better
sleep quality, experienced less burnout when following a set dayshift work schedule.
The two groups of nurses from each hospital were well matched for demographic characteristics.
A comparison of the two nurse groups found the absence of a significant difference in terms of
job satisfaction, overall place of employment satisfaction, place of employment stressfulness,
age, and years as a nurse, thus ruling out the possibility that burnout was triggered by place of
employment satisfaction differences. The two groups were also well matched for their MEQ and
PSQI scores.
No significant difference was found between the two nurse groups, regarding the presence of
feelings of personal, work-related, and client-related burnout as rated by the CBI. However,
significant differences in tendency for feelings of burnout were found between individuals
62
reporting good versus poor sleep quality based on PSQI cutoff ratings. Respondents with good
sleep quality reported significantly lower levels of burnout across all three subscales, compared
to participants who described themselves as having poor quality sleep. Regarding chronotype,
individuals with a greater propensity for morningness reported lower levels of burnout across all
three subscales compared to participants with a tendency for eveningness or neither M nor E. In
both cases of burnout analysis – by sleep quality and by chronotype – personal and work related
burnout were reported to be in the midrange level of severity, while client-related burnout
appeared to be much lower. It is possible that client related burnout may have been reported as
being much lower than personal and work related burnout scores if nurses are inclined to express
compassion towards patients as opposed to being frustrated with them. However, no data was
collected on this topic therefore no formal conclusions can be drawn. In the future, it will be
important to look at both compassion fatigue and compassion satisfaction to examine changes in
nurses’ feelings on these topics as they relate to chronotype and sleep quality.
In agreement with the idea by Söderström et al. (2004) that stress and sleep quality are
associated, correlational analysis demonstrated that individuals expressing greater work
stressfulness tend to have poorer sleep quality. Interestingly however, while the data were not
shown, chronotype and sleep quality were not associated in this sample. This may be due in part
to the sample size. Had more extreme chronotypes been obtained, this may have led to an
association. Furthermore, this result may change if sleep quality and chronotype were to be
examined in shiftworkers. It is more likely that in such a group, people would be working at
hours that may significantly conflict with their chronotype, or if they are extreme M or E types,
may even compliment them. Furthermore, shiftworkers likely have fragmented sleep which
would affect their sleep quality ratings, and this together with chronotype, might yield different
results on the relationship between chronotype and sleep quality.
To further investigate the evidence supporting the significant contribution made by chronotype
tendency, sleep quality, job satisfaction, place of employment satisfaction, and overall place of
employment stressfulness, in determining feelings of burnout, multivariate regression analyses
were applied. Chronotype tendency and sleep quality were significant factors influencing all of
the CBI subscale scores, however, here again the effect of these two factors on client related
burnout was much less than on personal and work-related burnout. As such, it appears that both
propensity for morningness and good sleep quality lower the risk of burnout among oncology
63
nurses working regular day shifts Monday to Friday. Overall workplace stressfulness was a
contributor to each burnout subscale score, indicating that reduced workplace stressfulness
protects against each type of burnout examined by this study. Job satisfaction also contributes
significantly to workplace related burnout, suggesting that higher job satisfaction leads to
reduced burnout in this domain. It did not however, influence personal or client related burnout.
Overall place of employment satisfaction was not a significant contributing factor, suggesting
that one’s burnout is more so influenced by the specific role or duties they must carry out, rather
than by the institution as a whole. The lack of correlation between overall place of employment
satisfaction and any of the burnout ratings supports the finding that overall place of employment
satisfaction was not a differential predictor for any of the burnout ratings examined.
According to the experimental hypothesis, this study’s findings support the view that chronotype
and sleep quality are key predictors of personal, work-related, and client-related burnout
symptoms, indicating that oncology nurses following a permanent dayshift schedule with a
propensity for morningness present with relatively low levels of burnout, compared to their
colleagues working at the same time but who have a tendency for eveningness or neither type of
chronotype. Furthermore, ambulatory care oncology nurses obtaining better sleep quality also
experience significantly decreased feelings of burnout across all three domains as compared to
their colleagues who report poor sleep quality. Many people – either through necessity or
voluntarily – work at times that conflict with their chronotype, oftentimes foregoing adequate
sleep in order to do so. This results in poorer sleep quality for a number of reasons including but
not limited to shortened sleep, going to bed at times not reflective of one’s chronotypic needs,
and even sleeping at times of day when one should actually be awake and functioning, to name a
few. While other personal and environmental factors likely also underlie these significant
burnout differences, allowing nurses to begin work at slightly modified times that better
accommodate their chronotypic needs and may help modify or enhance their sleep situation and
quality, might significantly help to reduce levels of personal, work related, and client related
burnout. However, further studies need to be completed in order to understand which of these
burnout subscale categories would be more or less affected by such changes.
This study’s results have important implications in the healthcare profession, not only regarding
nurse related burnout, but also relating to patient care. If burnout brings on feelings of cognitive
weariness, one’s decision-making skills may be affected such that their ability to make sound
64
judgments may be hindered by increased feelings of burnout. Physical and emotional fatigue
may also be heightened, potentially reducing one’s ability to react quickly in dire situations to
help a patient in distress. Emotional control or regulation may also be compromised, which could
elevate a patient’s level of distress should they be forced to deal with a nurse who is emotionally
burnt out. Burnout is known to increase the risk of absence from work (Westermann Kozak,
Harling & Nienhaus, 2014). An increasing number of people absent from work due to burnout
will further compound the workload and related stress felt by many nurses working in the same
department who will need to take over the responsibilities of their coworkers. Based on these
negative effects of burnout, it is in a medical institution’s best interest to consider how key
contributors such as chronotype and sleep quality can be better managed and used to alleviate
symptoms of burnout. Given the large number of nurses following shift work schedules both in
and out of the field of oncology, future studies should address burnout among oncology nurses
engaging in shift work in order to understand how chronotype and sleep quality could be used to
structure schedules that allow for decreased reports of burnout.
3.5.1! Limitations
One limitation of this study is related to sample size. Future studies using larger numbers may
likely be able to achieve a greater and more equal distribution of E and M type participants. This
might allow for reanalysis of the current statistics in this study that suggest there is no correlation
between chronotype and sleep quality among oncology nurses, as well as a greater understanding
of how extreme the symptoms of burnout are among E type participants working earlier in the
day. Secondly, in the future it would be beneficial to study client-related burnout, by including
other objective measures to rate one’s degree of compassion, in order to understand if reduced
ratings in this domain are related to feelings of compassion for patients, or whether some other
factor was involved in causing nurses to report such markedly lower burnout level in this
domain. The role of other factors such as extracurricular activities, medical conditions, and
medications were not investigated, as many nurses did not report this information in the personal
survey, and it would be important to understand the role of these variables in future.
65
3.6! Conclusions
Individual chronotype and sleep quality are important determinants of alertness and cognitive
ability variance throughout the day. The findings of this study help to clarify the contributions of
chronotype and sleep quality to burnout, together with workplace stressfulness and job
satisfaction, and as such fully support initiatives for programs that would modify work schedules
to accommodate the needs of one’s chronotype and that would also make suggestions for
modifications that would ameliorate one’s sleep quality. Such initiatives would result in
decreased burnout among ambulatory oncology nurses, and may subsequently increase quality of
patient care and satisfaction. This research also presents further research opportunities for
understanding burnout among oncology nurses following shift work schedules, by considering
the disruption their work brings to their circadian rhythm and sleep quality.
66
Chapter 4!
Bellicoso, D., Trudeau, M., Fitch, M. I., & Ralph, M. R. (2017). Chronobiological factors for compassion satisfaction and fatigue among ambulatory oncology caregivers. Chronobiology International, Epub Ahead of Print. doi: 10.1080/07420528.2017.1314301
! Chronobiological Factors for Compassion Satisfaction and Fatigue Among Ambulatory Oncology Caregivers
4.1! Abstract
Primary caregivers for victims of chronic illness and or trauma experience both positive
and negative emotional consequences. These are broadly classified as compassion satisfaction
(CS) and compassion fatigue (CF). Because one of the components of CF, burnout, varies with
chronotype and sleep quality, we assessed the influence of chronobiological features on the
broader constructs of CS and CF. Responses from primary ambulatory care oncology staff
working dayshifts were assessed for potential relationships of chronotype and sleep quality with
CS and CF using the professional quality of life scale (ProQoL). These were analyzed further in a
multivariate model that included personality and job satisfaction as cofactors. We found that
sleep quality was a key contributor to CS development, and CF reduction. Morningness was
positively linked to CS, but the univariate association was masked in the multivariate model. Job
satisfaction (contingent rewards, nature of work, and operating procedures) heavily influenced CS
and CF development. Agreeableness and openness showed positive correlations with CS and
negative with burnout, while emotional stability was linked to reduced CF. While job satisfaction
and personality predictably played roles in the development of CS and CF, sleep quality and
chronotype contributed significantly to benefits and negative consequences of providing care.
67
4.2! Introduction
Compassion fatigue (CF) describes a set of negative consequences experienced by caregivers to
one or many people who are dealing with trauma. It includes feelings of exhaustion, frustration,
depression, and even fear driven by working with such individuals. Burnout has been recognized
as one of the major results that often develops among caregivers who experience chronic CF
from working with trauma victims and or individuals with long term illness (Stamm, 2010).
Burnout is preceded by a progressive increase in the sense of hopelessness and or difficulty in
dealing with work (Brand et al., 2010). It may be expressed as feelings that one’s efforts are
useless, or that workloads are unmanageable. Common responses are anger, frustration, sadness,
cognitive weariness, and emotional and physical fatigue (Bellicoso et al., 2014; Stamm, 2010).
Along with burnout, CF can also produce secondary traumatic stress (STS) resulting from
exposure to individuals who have experienced recently, or are currently experiencing, traumatic
events. STS may be expressed as a lingering combination of fear, sleep disturbance, and or
intrusive imagery.
In contrast, compassion satisfaction (CS) is a positive outcome of providing care. CS refers to the
pleasure or fulfillment a caregiver feels from doing a job well and from helping others,
particularly those faced with traumas or chronic illness (Stamm, 2002). Therefore, taken together,
CF and CS reflect a worker’s total professional quality of life. These positive and negative
outcomes of caregiving are known to exist among those working in various positions in the
healthcare field, (e.g., nurses, physicians, therapists, and even audiologists) (Drury et al., 2014;
El-Bar et al., 2013; Severn et al., 2012; Sodeke-Gregson et al., 2013). A quantification of the two
components is provided by the Professional Quality of Life (ProQoL) Scale.
Given the prevalence and impact of CS and CF among healthcare workers, it is important to
understand factors that contribute to their development and management. In a previous study,
sleep quality and chronotype were shown to be significant factors in the propensity and
development of burnout among oncology nurses (Bellicoso et al., 2014). This raised the question
of what roles these two factors play in the broader constructs of CS and CF.
Chronotype refers to the relative preference shown by individuals for performing various
activities during the morning (M) hours (M types) versus evening (E) hours (E types) independent
of societally imposed time constraints (Horne & Östberg, 1976; Neilsen, 2010). Basic biological
68
processes such as cellular functioning, hormonal fluctuation, and fluctuations in body temperature
have been linked to chronotype and circadian rhythmicity (Brown et al., 2008; Horne & Östberg,
1976; Lack et al., 2009). Chronotype is also correlated with overt, measurable behaviours
including executive control, and physical performance (Brown et al., 2008; Lara, Madrid, &
Correa, 2014). On several tasks with a high cognitive demand, a synchrony effect has been
reported, showing that performance is often best at optimal (morning for M types, afternoon or
evening for E types) over other times of day (Hasher, Chung, May & Foong, 2002).
A key measurable expression of chronotype and rhythmicity is the sleep wake pattern (Horne &
Östberg, 1976; Roenneberg, Wirz-Justice, & Merrow, 2003). M and E types consistently show
no difference in sleep duration, but demonstrate significantly earlier sleep wake timing (Horne &
Östberg, 1976; Lara et al., 2014). While average sleep duration is consistent across chronotypes,
M types often report better sleep quality (Merikanto et al., 2012; Nielsen, 2010; Vitale et al.,
2015).
Importantly, chronotype and sleep quality have known associations with other factors that
ultimately affect CS and CF. Personality has been found to be associated with both chronotype
and compassion. Poorer sleep quality has been found in individuals reporting low
conscientiousness together with high neuroticism, while those reporting higher levels of
extraversion, agreeableness, and conscientiousness typically experience better sleep quality
(Duggan, Friedman, McDevitt, & Mednick, 2014; Hintsanen et al., 2014). Morningness has been
positively correlated with high conscientiousness and agreeableness; eveningness has been
correlated with greater neuroticism (e.g., Cavallera, Gatto & Boari, 2014; Duggan, Friedman,
McDevitt, & Mednick, 2014; Hsu, Gau, Shang, Chiu, & Lee, 2012). Studies that have examined
specifically the relationship between CS/CF and personality are sparse. Some research has
suggested a link with optimism, which is a broadly representative trait on the Big Five factors of
personality (Sharpe, Martin & Roth, 2011). In a group of genetic councilors, those reporting low
optimism had increased CF (moderate to high burnout), and low to moderate CS (Injeyan,
Shuman, Shugar, Chitayat, Atenafu, & Kaiser 2011). Among volunteer counselors caring for the
terminally ill, burnout is greater among those individuals reporting high neuroticism (Bakker,
Van der Zee, Lewig, & Dollard, 2006). Comparative roles for personality traits, chronotype, and
sleep quality on the development of CS and CF within the same subject pool have not been
assessed.
69
In addition, CS and CF are likely to be influenced by work conditions and job satisfaction. Low
job satisfaction is linked with greater stress and lower job ratings compared to workers reporting
high job satisfaction. Those reporting high job satisfaction tend to also report better health than
their colleagues (Rahman & Sen, 1987). Assessing the effect of job satisfaction when considering
CS and CF will provide a more complete picture of how specific personal and environmental
factors contribute to one’s overall professional quality of life.
All the factors that have been studied or theorized to contribute to CS and CF are likely to
interact. For example, chronic health issues have been linked to sleep disturbance, both as
outcomes as well as causative factors. Many societal conventions such as responsibilities in the
home, social protocols, and daily work schedules, interfere with achieving proper sleep quality
when they are desynchronized with the hours associated with one’s innate chronotype. Therefore,
to provide a more complete perspective on the factors contributing to CS and CF, in a defined
work environment, we have assessed the complex effects and interactions among sleep timing
and quality, chronotype, personality, job satisfaction, with self-ratings of CS and CF in oncology
staff working in a large hospital.
4.3! Materials and Methods
4.3.1! Participants and Procedures
Primary ambulatory care staff at the Odette Cancer Centre of the Sunnybrook Health Sciences
Centre (Toronto, Ontario, Canada) were sent a general email requesting their voluntary
participation in this study. The job categories included medical/radiation/surgical oncologists,
registered ambulatory care oncology nurses, radiation therapists, and pharmacists/pharmacy
technicians. The email explained the purpose of the study, and provided a link to participate. The
inclusion criteria indicated that the study was only intended for oncologists, oncology nurses,
radiation therapists or pharmacy team members working in the Odette Cancer Centre, providing
direct care to patients, and following a dayshift Monday to Friday work schedule. It was
explained that by proceeding to answer the survey questions, participants were granting their
consent to use the data provided. Participants were informed that their responses would not be an
indication of their work performance, nor would their individual scores be revealed to
70
supervisors or made public. While participation was voluntary, study subjects were given the
option of providing an email address to participate in a draw for one of three iPad minis, which
were awarded based on a random draw once the entire study had been completed.
In total, 140 staff members responded to the questionnaire. Twelve of the submissions were
discarded because the respondents either did not fit the job description criteria, worked an
irregular shift schedule, or the questionnaire was submitted incomplete, leaving a total of 128
completed useable questionnaires. The test group comprised 25 oncologists, 44 nurses, 17
pharmacy team members, and 42 radiation therapists.
4.3.2! Measures
The questionnaire package included six survey items: background information, Horne-Östberg
Morningness Eveningness Questionnaire (MEQ; Horne & Östberg, 1976), Pittsburg Sleep
Quality Index (PSQI; Buysse, Reynolds III, Monk, Berman, Kupfer, 1988), Job Satisfaction
Scale (JSS; Spector, 1985), Ten Item Personality Inventory (TIPI; Gosling, Rentfrow, & Swann,
2003), and the Professional Quality of Life Scale (ProQoL; Stamm, 2010). The short TIPI scale
was used instead of the full BFAS (DeYoung, Quilty, & Peterson, 2007) to assess personality to
reduce the number of questions for participants to complete, and the time commitment, increasing
the number of volunteers. The JSS was included because workers will have some degree of
opinion on tasks they perform, and the environment in which they work, which will affect
various domains of ProQoL and CF/CS beyond the direct impact of caregiving per se (see Renzi,
Tabolli, Ianni, Di Pietro, & Puddu, 2005).
4.3.3! Statistical Analysis
Data from oncologists, nurses, radiation therapists, pharmacists, and pharmacy technicians was
analyzed collectively, as a group. Descriptive data and demographic frequencies were used to
describe the sample. Questionnaires (MEQ, PSQI, ProQol, TIPI and JSS) were scored according to
their scoring manuals or published scoring guidelines in the original article. Descriptive analyses
were conducted on MEQ, PSQI, each of the nine components and the total JSS scores, the TIPI
scales, and the three components of ProQoL to assess whether data were normally distributed.
Natural log transformation was performed on burnout, STS and PSQI to produce normally
distributed variables later used in the regression analyses. Univariate linear regression was used
to assess which continuous and categorical variables were significantly associated with each
71
outcome variable (CS, burnout and STS). Factors significantly associated with study outcome
variables were included in a backwards multivariate linear regression analysis to build predictive
models of each ProQoL score. A p value < .05 was accepted as statistically significant. Statistical
analyses were performed using SPSS (Statistical Package for the Social Sciences) version 23.0
for Mac.
4.4! Results
4.4.1! Descriptive Group Statistics
Descriptive and frequency statistics for continuous and categorical demographic characteristics
and questionnaire data from the 128 participants are presented in Table 4.1.
72
Table 4.1 Means and frequencies of participant demographic and questionnaire data
Mean (SD) Frequencies (%) Basic Demographics Age 44.41 (11.83) yrs Gender
Males 27 (21.1%) Females 101 (78.9%)
No. of years in healthcare 20.48 (12.81) yrs No. of years in oncology 15.73 (11.27) yrs No. of Sites Worked
Single 25 (19.5%) Multiple 103 (80.5%)
Relationship Status Single 42 (32.8%)
Spouse 86 (67.2%) Current Position
Oncologist 25 (19.5%) Outpatient Nurse 44 (34.4%) Pharmacy Team 17 (13.3%)
Radiation Therapist 42 (32.8%) MEQ
Total Group of Scores 56.09 (9.907) Definitely Evening 1 (0.8%)
Moderately Evening 12 (9.4%) Neither 59 (46.1%)
Moderately Morning 47 (36.7%) Definitely Morning 9 (7.0%)
PSQI Total Group of Scores 5.30 (3.05)
Good Quality Sleep (≤ 5) 76 (59.4%) Poor Quality Sleep (> 5) 52 (40.6%)
Job Satisfaction Pay 14.13 (4.58)
Promotion 11.91 (4.37) Supervision 18.09 (4.76)
Benefits 14.00 (4.17) Contingent Rewards 14.70 (4.68)
Operating Procedures 12.68 (3.50) Coworkers 17.18 (3.72)
Nature of Work 20.04 (3.69) Communication 15.39 (3.96)
Total 138.13 (26.63) Personality
Extraversion 4.17 (1.45) Agreeableness 5.54 (1.09)
Conscientiousness 6.17 ( .93) Emotional Stability 5.46 (1.30)
Openness to Experience 5.39 (1.09) ProQoL
CS 40.16 (6.04) BO 22.40 (5.63)
STS 20.63 (5.80)
73
Respondents were predominantly female (78.8%), were married (67.2%), and or worked with
multiple cancer sites (80.5%). The mean MEQ score (56.06) indicates that on average,
participants were classified as neither M nor E type individuals; however, the average score is
only two points below the cutoff for being moderately M type. 43.7% scored as M type and 10.2%
were E type.
The mean sleep quality score (5.30) is slightly above the cutoff for good quality sleep. 40.6% of
respondents met the criteria for varying degrees of poor quality sleep, while 21.0% of the
respondents who were rated as having good quality sleep were right on the borderline of poor
quality sleep.
JSS scores were all comparable to those presented in the original article (Spector, 1985).
Cumulative scores ranged from 36 to 216 with an average of 138.13 (the highest possible JSS
score is 216). Therefore, respondents overall were above 50% for this index of job satisfaction.
Individual components of job satisfaction range from a score of 4 (not satisfied) to 24 (satisfied).
Average scores for individual components of job satisfaction averaged between 11.91 (promotion)
and 20.04 (nature of work).
Personality is scored on a scale of 1 (strongly disagree) to 7 (agree strongly). On average, the
strongest trait among respondents was conscientiousness (6.17), and the weakest was
extraversion (4.17), which was still above the halfway mark of the scale.
The mean score for CS (40.16) indicates that respondents fell in the high end of the range for
average satisfaction derived from their job (ProQoL manual, Stamm, 2010). The mean burnout
score fell just outside the cutoff to qualify as low burnout, suggesting participants were close to
the low end of an average degree of burnout. The mean STS score indicates that participants
typically feel a low level of STS related to their job.
4.4.2! Correlation Analysis
We first conducted a comprehensive univariate analysis using all component variables of the five
scales and demographic data. The correlations are presented in three supplementary tables
highlighting relationships with (a) demographic, chronotype, and sleep quality (Supplementary
Table S4.1); (b) job satisfaction (Supplementary Table S4.2); and (c) personality and
professional quality of life (Supplementary Table S4.3).
74
ProQoL components were strongly but differentially correlated with MEQ, PSQI, JSS, and TIPI
(Table 4.2). The majority of correlations for CS were positive, with a range of significant
correlation coefficients between rs = 0.179 to 0.768. Most correlations with burnout were
negative, with a range of significant correlation coefficients between rs = - 0.216 and - 0.589.
Fewer significant correlations existed for STS, but again, the majority of the significant
correlations were negative, in the range of rs = - 0.034 to - 0.512. As expected, the correlations
between the three ProQoL domains and PSQI were the inverse of the rest of the significant
correlations for the group, with a range of rs = - 0.322 to 0.463.
75
Table 4.2 Non-parametric correlations between ProQoL domains (columns) and covariates (rows) Compassion
Satisfaction Burnout Secondary
Traumatic Stress MEQ .185* -.106 .053 PSQI -.322** .463** .270** Job Satisfaction
Pay .316** -.280** -.020 Promotion .272** -.216* .066
Supervision .386** -.398** -.110 Benefits .203* -.204* -.034*
Contingent Rewards .452** -.478** -.157 Operating Procedures .233** -.414** -.257**
Coworkers .470** -.504** -.235** Nature of Work .768** -.589** -.237** Communication .413** -.504** -.177*
Total .543** -.552** -.171
TIPI Extraversion .179* -.133 -.100
Agreeableness .390** -.399** -.269** Conscientiousness .126 -.128 -.247**
Emotional Stability .384** -.543** -.512** Openness to Experience .249** -.295** -.180*
Spearman’s rho is shown * p<0.05; **p<0.01
76
Following univariate regression analyses of the predictive value of MEQ, PSQI, and each
component of the JSS and TIPI scales on each ProQoL domain, we assessed the extent to which
significant univariate predictors combined as a multiple regression model (backwards multiple
regression; data not shown) to predict CS, burnout, and STS (Table 4.3). MEQ was not
significantly related to the ProQoL domains when considered with the other factors, and is
therefore not included in Table 4.3. However, the univariate regression analysis for MEQ alone
indicated that chronotype can account for 2.7% of the explained variability in CS development
(univariate regression data not included in tables). Based on the correlation data (Table 4.2), it
appears that greater M type tendency is associated with higher CS. The adjusted R2 values were
higher for CS and burnout than STS, suggesting that other factors not explored here may
contribute to the development of STS. Overall, components of job satisfaction and certain
personality factors contributed heavily to all ProQoL domains. Interestingly, according to these
models, PSQI only contributes to burnout, while MEQ does not contribute significantly to any of
the separate domains of ProQoL when considered together with other factors.
77
Table 4.3 Final models in backwards multiple regression with professional quality of life (CS, BO, and STS) as dependent variables, and significant values* (as tested by univariate regression) for MEQ, PSQI, JSS and TIPI as independent variables ProQoL domain Included covariates B (SE) ß p value Adjusted R2 CS .642 Constant 6.878 (2.490) .007 JSS Contingent Rewards 0.176 (0.082) .136 .035 Nature of Work 1.036 (0.109) .632 .000 TIPI Agreeableness 0.965 (0.319) .174 .003 Openness to Experience 0.849 (0.307) .153 .007 Burnout .593 Constant 4.193 (.154) .000 PSQI 0.088 (0.032) .172 .007 JSS Contingent Rewards -0.010 (0.004) -.174 .013 Operating Procedures -0.020 (0.004) -.268 .000 Nature of Work -0.018 (0.005) -.255 .000 TIPI Emotional Stability -0.059 (0.013) -.297 .000 Openness to Experience -0.038 (0.015) -.160 .011 STS .281 Constant 3.750 (0.112) .000 JSS Operating Procedures -0.015 (0.006) -.186 .016 TIPI Emotional Stability -0.105 (0.016) -.484 .000
Note. B (SE) represents the unstandardized coefficient (standard error). ß represents standardized coefficients. *Only covariates with p < 0.10 in the univariate linear regression were included in the analyses
78
In these regressions, the explained variance is higher for CS (adjusted R2 = .642), and burnout
(adjusted R2 = .593) as compared to the variance for STS (adjusted R2 = .281).
Furthermore, CS and burnout, and burnout and STS shared similar covariates, yet CS and STS
did not share covariates.
Job satisfaction significantly predicts the development of each component of ProQoL. Given that
job satisfaction may be masking the contribution of the other covariates being examined,
backwards multiple regression was run again to assess the various domains of ProQoL, this time
without including the influence of job satisfaction (Table 4.4). The explained variance in these
multiple regressions is lower than when the models included JSS scores.
79
Table 4.4 Models in backwards multiple regression with professional quality of life (CS, BO, and STS) as dependent variables, and significant values* (as tested by univariate regression) for MEQ, PSQI, TIPI and working on multiple vs. single cancer sites as independent variables ProQoL domain Included covariates B (SE) ß p value Adjusted R2 CS .253 Constant 39.469 (3.568) .000 PSQI -3.290 (0.940) -.274 .001 TIPI Agreeableness 1.991 (0.429) .358 .000 Work on multiple vs. single cancer sites -2.589 (1.182) -.171 .030 Burnout .395 Constant 3.425 (0.131) .000 PSQI 0.166 (0.037) .326 .000 TIPI Agreeableness -0.049 (0.018) -.206 .009 Emotional Stability -0.067 (0.016) -.338 .000 STS .253 Constant 3.590 (0.093) .000 TIPI Emotional Stability -0.110 (0.017) -.509 .000
Note. B (SE) represents the unstandardized coefficient (standard error). ß represents standardized coefficients. *Only covariates with p < 0.10in the univariate linear regression were included in the analyses; MEQ did not remain a significant predictor for any of the dependent variables
80
Of the three ProQoL domains, the explained variance is highest for burnout (adjusted R2 = .395)
when job satisfaction is not included in the multiple regression; the explained variance for CS
and STS is equal in this model (adjusted R2 = .253). PSQI score is a covariate in determining
both CS and burnout when JSS scores are not included. Working with multiple instead of a single
cancer site is also a significant covariate in determining CS when JSS scores are not included,
however it is not a significant contributor to any other models. As with the original models,
burnout shares a particular set of covariates with CS, and burnout shares a different set of
covariates with STS; no similar set is shared between CS and STS.
4.5! Discussion
The primary objective of this study was to understand how the internal factors of chronotype and
sleep quality influence the various components contributing to ProQoL. This study focused on
healthcare professionals working directly with cancer patients in a hospital setting. Given the
known pervasive influences of both circadian rhythms and sleep on human physical, mental, and
emotional performance, it is not a surprise to find that both contribute in some way to aspects of
professional quality of life measured by the ProQoL Scale. A significant finding from this study
is that chronotype and sleep quality have distinctly different effects on CS and CF.
4.5.1! Chronotype and ProQoL
While the three components of ProQoL (CS, burnout, and STS) are consistently and strongly
predicted by aspects of job satisfaction, personality, and sleep quality, chronotype was only
correlated with CS (Table 4.2). We corroborated this using univariate linear regression analysis
(data not shown), which indicated that MEQ score predicted CS, but neither burnout nor STS.
However, in the multiple regression model (which includes personality, sleep quality, and JSS as
covariates), MEQ score did not predict CS. Therefore, these employees are able to earn
satisfaction from their work regardless of their chronotype. Chronotype appears to be a weak factor
compared with the others in producing CS. Unfortunately, the impact of chronotype could be
underestimated due to the limited range of chronotypes represented in the data set. Moreover, the
impact of chronotype depends on the synchrony of performance demand with the subject’s optimal
81
time (estimated by the MEQ score). Because the subjects were day shift workers only, the optimal
performance times would coincide with performance demands for the majority of participants, thereby
reducing the impact of chronotype on the dependent measures of CS and CF. Neither burnout nor STS
were linked to MEQ chronotype. This suggests that the underlying CF can develop independently
of whether or not one works at their optimal time.
4.5.2! Sleep and ProQoL
PSQI scores were negatively correlated with CS, and positively correlated with burnout and STS
(Table 4.2); therefore, better sleep quality is associated with higher CS and lower CF. As part of a
multivariate regression model, sleep quality negatively predicts burnout, regardless of whether or
not job satisfaction is included in the model (Table 4.3 and 4.4). As sleep score increases
(indicating worse sleep quality), one’s burnout rating also increases. Given that poor sleep quality
has been associated with perseveration (e.g., Banks & Dinges, 2007), a possible explanation is
that perseverative thoughts relating particularly to difficult work experiences result in this
increased burnout. Rzeszutek and Schier (2014) documented evidence supporting this conclusion,
indicating that among a sample of therapists, perseveration proved to be a significant predictor of
burnout symptoms.
When job satisfaction (JSS) is not included, PSQI score is an even greater predictor of burnout
(Table 4.4). This suggests that rest and good quality sleep increase one’s ability to achieve
satisfaction from providing compassionate care for patients. However, sleep quality is no longer a
significant predictor of CS if job satisfaction is included, suggesting that the benefits gained by
job satisfaction outweigh those of good sleep quality in this population.
Sleep quality does not appear to be a significant predictor of STS in this population, regardless of
the inclusion of JSS in the model, which reflects the ProQoL literature indicating that sleep
difficulties are the result of STS (Stamm, 2010) (Table 4.3 and 4.4). This implies one’s level of
STS is a predictor for sleep quality or sleep disturbance onset. Interestingly, the ProQoL survey
literature does not indicate that either CS or burnout contribute to an inability to sleep or
diminished sleep quality, and only lists poor sleep as an outcome of STS. While the lack of
predictive value for PSQI on STS in this study may be due to sample size, it seems likely that in
general, as part of a model, sleep quality is not a predictor of STS.
82
4.5.3! Personality and ProQoL
Personality traits were all positively correlated to CS and negatively correlated with burnout and
STS; although not all correlations were significant (Table 4.2). There were no overlapping
personality traits acting as predictors for both CS and STS in multivariate regression models;
however, both CS and STS each have shared personality traits predictors with burnout (Table 4.3
and Table 4.4). The same trend exists both when considering or excluding components of JSS as
predictors. This pattern of shared versus non-shared personality trait predictors for CS, burnout,
and STS occurs regardless of whether JSS components are included in the multiple regression
model.
4.5.3.1! Agreeableness
Agreeableness significantly predicts CS (higher agreeableness predicts greater CS), regardless of
whether or not JSS is considered as a predictor (Table 4.3 and 4.4). In the TIPI, agreeableness is
described on a scale of critical or quarrelsome (low agreeableness) to sympathetic and warm
(high agreeableness) (Gosling, Rentfrow, & Swann, 2003). Those who are high in agreeableness
are likely to achieve high CS from their job as they may be able to get along with their peers and
show understanding for their patients, leading to an overall positive outcome and sense of
satisfaction from their work.
When JSS components are not being considered, agreeableness also predicts burnout (higher
agreeableness is linked to lower burnout rates). Agreeableness may allow a person to avoid
overthinking certain hospital policies that may be burdening them, or make it easier to get along
with patients or family members who may be acting difficult, subsequently reducing feelings of
burnout and allowing a person to obtain greater satisfaction from the tasks they perform.
4.5.3.2! Emotional Stability
Emotional stability significantly predicts burnout and STS, regardless of whether or not JSS is
considered as a predictor. Poor emotional stability predicts higher levels of burnout and or STS
(Table 4.3 and 4.4). The TIPI describes emotional stability on a scale of anxious or easily upset
(low) to calm and emotionally stable (high) (Gosling, Rentfrow, & Swann, 2003). Those high in
emotional stability are likely better able to remain calm in the face of emotionally taxing
83
situations at work, allowing them to avoid feeling overwhelmed and bogged down by their
experiences, and able to separate their work versus private life. Being able to avoid these feelings
reduces one’s level of burnout and STS.
4.5.3.3! Openness
Openness to experience also predicts both CS and burnout (higher openness is linked to greater
CS and lower burnout), but only when JSS components are considered as predictors in the
multivariate regression model (Table 4.3). The TIPI describes openness to experience on a scale
of conventional and uncreative (low) to open to new experiences and complex (high) (Gosling,
Rentfrow, & Swann, 2003). When considering job satisfaction, openness to experience may help
an individual think outside the box, and be more creative in ways to perform their job, allowing
them to gain greater satisfaction from finding new ways to help others, or being innovative in
ways to avoid feelings of fatigue or work around hospital policies, subsequently reducing
burnout. However, future studies should further examine this relationship to understand if it is
specifically linked to JSS components, or if it is a stable strong predictor of CS and burnout.
4.5.3.4! Conscientiousness and Extraversion
Neither of these traits was significantly correlated with all the ProQoL indicators. High
conscientiousness was linked to higher STS but not to CS, nor to the rate of burnout.
Extraversion was only weakly linked to CS, and not to CF.
While these results do not indicate shared predictors between CS and STS, the existence of
overlapping personality traits that simultaneously influence CS and burnout, and burnout and
STS suggest that burnout has specific aspects that directly relate to one’s level of CS and STS.
4.5.4! Job Satisfaction and ProQoL
Average job satisfaction scores were comparable to those obtained from a larger population of
human service workers. This indicates that the current sample of participants is reflective of the
scores seen in the general population. Spearman correlations showed a significant positive
relationship between all aspects of the JSS and CS, and a significant negative association with
burnout. The positive individual correlations between the components of the JSS and CS indicate
that as one’s job satisfaction increases, so does one’s level of CS, and vice versa. The negative
84
correlations between the JSS and burnout indicate that those with higher burnout, report poorer
job satisfaction, and vice versa. However, not all aspects of the JSS are correlated with STS.
Furthermore, while almost all correlations between the JSS and STS are negative, the correlation
between the component promotion and STS is a positive correlation, but it is not significant.
Further research with a larger sample size should reassess this association to verify if a significant
correlation can be found, and whether it will be a positive or negative association.
Spearman correlations (Table 4.2) indicate that the specific components of job satisfaction that
most positively correlate with CS are nature of the work, coworkers, and contingent rewards.
Communication was also showed a significant positive correlation with CS. Nature of the work
refers to an overall sense of pleasure derived from the work tasks themselves, and indicates that
there is a sense of pride in one’s job along with the belief that one’s work serves a benefit.
Contingent rewards refer to the recognition and appreciation one receives for their job well done.
Overall, these four domains of job satisfaction refer to the benefits one sees their work provides,
the personal satisfaction one gains from the job, and the accolades received for one’s efforts.
When considering JSS as part of a multivariate regression model (Table 4.3) as a predictor of CS,
the only two components that remain significant predictors are contingent rewards and nature of
work. While total JSS score is significantly correlated to CS, it did not predict CS as part of
model, suggesting that one’s type of job, and the rewards the job brings (in the form of accolades
and appreciation from others), when considered together with other factors such as personality
and sleep quality, are the strongest predictors.
The Spearman correlations between the JSS and burnout (Table 4.2) are all significant and
negative, indicating that high job satisfaction ratings are associated with lower burnout. The four
components of JSS most significantly correlated to CS are also the most significantly correlated
to burnout. Understandably, when one receives little personal pleasure from the nature of his or
her job, and or little recognition in the form of praise or accolades from peers or patients, it is
more difficult to see that one’s work is appreciated. Accordingly, if one feels their work is
serving little or no benefit, it becomes difficult to obtain satisfaction when one feels their efforts
are undervalued. Operating procedures, which refer to one’s impression of the manageability of
workload and the extent to which one feels efforts to do a good job are either facilitated or
85
blocked by organizational policy, are also significantly negative correlated with burnout. A sense
of struggling with a large and unmanageable workload, or that one’s efforts to care for patients
are being blocked by hospital policy, would likely increase one’s job-related sense of frustration
and sense of fatigue, subsequently bolstering work-related feelings of burnout. Further research is
necessary to understand which specific operating procedures in the hospital are contributing to
increased feelings of burnout, in order to identify appropriate coping strategies. As part of a
multivariate regression model, the three components that remain significant predictors of burnout
are contingent rewards, operating procedures, and nature of work (Table 4.3).
The majority of correlations between the JSS and STS are negative, yet unlike the correlations
with CS and burnout, not all correlations with STS are significant (Table 4.2). The correlations
between the JSS and STS are also smaller than those with CS or burnout. Compared with CS and
burnout, one’s job satisfaction is less correlated to their level of STS. Of the correlations between
select components of the JSS and STS, only one’s satisfaction with work operating procedures
remains a significant negative predictor of STS when considered as part of a multivariate
regression model (Table 4.3). This model suggests low satisfaction with operating procedures is
predictive of higher STS. It is possible that certain hospital rules make staff more vulnerable to
the negative emotional effects of the job. However, as with the impact of operating procedures on
burnout, it is necessary to conduct further studies among caregivers working in a hospital setting
to understand how and why certain operating procedures are increasing one’s vulnerability to
developing the negative domains of ProQoL.
The total score for JSS, does not appear to have any predictive value as part of a model on the
development of CS, burnout or STS. This is interesting to note as it appears that specific
components of job satisfaction play more prominent roles in contributing to the development of
particular domains of ProQoL than does one’s overall level of job satisfaction. This can prove
beneficial in the workplace as it indicates specific areas that can directly affect employee levels
of CS and CF. Furthermore, as with personality, there were no overlapping predictors of job
satisfaction for CS and STS. However, both CS and burnout, and burnout and STS shared certain
JSS predictors, again supporting the idea that burnout shares associations with one’s levels of CS
and STS.
86
4.5.5! Limitations
While as a group, this sample allows for a significant effect size, it was not large enough to allow
independent assessments of the professional categories (oncologists, outpatient nurses,
pharmacists, pharmacy technicians, radiation therapists). Separate assessments will paint a clearer
picture of whether certain careers are differentially influenced by M versus E preference, sleep
quality, job satisfaction components, or different personality traits in the development of CS,
burnout, and STS. The individuals included in this study are a cross section of oncology
caregivers, and are sampled from a single hospital. Subject samples from different hospitals will
be necessary to determine how factors such as job satisfaction might contribute to CS and CF; it
may be important to give this current population further examination given the elevated level of
professional quality of life which they report.
The group included a limited range of chronotypes, with disproportionate makeup of N and M
type subjects as compared to E types. A broader range approaching that of the larger population,
may reveal a greater influence of M versus E preference on one’s CS and CF. Given that the
group was sampled from dayshift workers, these individuals may have a greater propensity
towards morningness in this population. Still, even without a large number of E type dayshift
workers, a univariate regression effect of chronotype on CS score was still present.
4.6! Conclusion
We have found that job satisfaction, personality, and sleep quality all are significant predictors of
CS, burnout, and STS. On the other hand, MEQ score was a significant predictor of CS only at
the univariate regression level, and did not show significance in the multiple regression model for
either CS, burnout or STS. But while chronotype was only a weak predictor in our study, it
cannot be discounted as a factor, and may be more relevant in other groups. Importantly, the M
chronotype was over-represented in our subject group, possibly due to self-selection of work
schedule. A more normal distribution of MEQ scores across the E to M spectrum may reveal a
more substantial influence of chronotype, with a greater proportion of individuals working at their
less-than-optimal times.
87
Job satisfaction is important in promoting CS while reducing burnout and STS. However, as CS
and STS shared no predictor variables, healthcare workers could be obtaining CS from their job
while simultaneously developing STS from job related stresses and emotions. It is particularly
important to monitor for the development of STS among all workers, even those who seem
satisfied with their job.
Sleep quality is highly predictive of burnout. Although our data do not indicate the causal
relationship (poor sleep could lead to burnout, or burnout due to other factors, could produce poor
sleep), it seems likely that the relationship would be reciprocal. The strong relationship suggests at least
that, regardless of all other factors, promoting good sleep quality among ambulatory healthcare
providers working in hospitals is beneficial in many ways.
88
Supplementary Table S4.1 Spearman correlations (rs) between continuous variables, and point biserial correlations for categorical variables (demographics, chronotype and sleep quality)
Demographics MEQ PSQI Age Genderƒ Yrs in
Oncology Yrs in
Healthcare No. of Sites
Worked‡
Age 1.000 .092 .796** .928** .009 .360** .034 Genderƒ .092 1.000 .082 .149 .206* .027 .071 Yrs in Oncology .796** .082 1.000 .861** .061 .274** .022 Yrs in Healthcare .928** .149 .861** 1.000 .061 .361** .063 No. of Sites Worked‡ .009 .206* .061 .061 1.000 .147 .017 MEQ .360** .027 .274** .361** .147 1.000 -.172 PSQI .034 .071 .022 .063 .017 -.172 1.000 Job Satisfaction
Pay .190* -.186* .112 .134 -.064 .104 .035 Promotion .003 -.173 -.055 -.036 -.124 .091 -.099
Supervision .057 .041 -.119 .004 -.063 -.002 -.037 Benefits -.025 -.201* -.043 -.049 -.033 .138 -.103
Contingent Rewards -.033 -.135 -.139 -.103 -.054 .019 -.177* Operating Procedures -.016 .030 -.074 -.050 .086 .019 -.109
Coworkers .073 -.068 -.031 .034 -.004 .115 -.282** Nature of Work .199* -.115 .031 .126 -.084 .200* -.317** Communication -.062 -.002 -.236** -.111 .172 .045 -.213*
TOTAL .042 -.123 -.104 -.028 -.024 .098 -.187* TIPI
Extraversion .147 .121 .078 .127 -.034 .032 -.125 Agreeableness .033 .470 .012 .060 .029 .100 -.073
Conscientiousness .142 -.028 .074 .129 -.045 .151 -.047 Emotional Stability .176* -.031 .185* .196* .049 .105 -.266**
Openness to Experience .005 .130 .040 .034 .002 .007 -.160 ProQoL
CS .145 -.021 .779 .092 -.057 .178* -.324** Burnout .105 .041 .109 .114 -.051 -.075 .406**
STS -.046 .159 -.101 -.014 -.030 .118 .272** * p < .05, ** p < .01 ƒ Coding: 1=male, 2=female; ‡ Coding: 1=single, 2=multiple
89
Supplementary Table S4.2 Spearman correlations (rs) between continuous variables, and point biserial correlations for categorical variables related to job satisfaction
Job Satisfaction Survey Pay Promotion Supervision Benefits Contingent
Rewards Operating Procedures
Coworkers Nature of Work
Commu-nication
TOTAL
Age .190* .003 .057 -.025 -.033 -.016 .073 .199* -.062 .042 Gender ƒ -.186 -.173 .041 -.201* -.135 .030 -.068 -.115 -.002 -.123 Years in Oncology .112 -.055 -.119 -.043 -.139 -.074 -.031 .031 -.236** -.104 Years in Healthcare .134 -.036 .004 -.049 -.103 -.050 .034 .126 -.111 -.028 No. of Sites Worked‡ -.064 -.124 -.063 -.033 -.054 .086 -.004 -.084 .172 -.024 MEQ .104 .091 -.002 .138 .019 .019 .115 .200* .045 .098 PSQI .035 -.099 -.037 -.103 -.177* -.109 -.282** -.317** -.213* -.187* Job Satisfaction
Pay 1.000 .531** .432** .541** .483** .109 .398** .354** .388** .686** Promotion .531** 1.000 .501** .460** .587** .088 .319** .307** .404** .681**
Supervision .432** .501** 1.000 .321** .620** .333** .513** .486** .600** .771** Benefits .541** .460** .321** 1.000 .492** .142 .342** .269** .422** .656**
Contingent Rewards .483** .587** .620** .492** 1.000 .245** .542** .500** .567** .823** Operating Procedures .109 .088 .333** .142 .245** 1.000 .318** .110 .425** .395**
Coworkers .398** .319** .513** .342** .542** .318** 1.000 .513** .593** .709** Nature of Work .354** .307** .486** .269** .500** .110 .513** 1.000 .490** .602** Communication .388** .404** .600** .422** .567** .425** .593** .490** 1.000 .774**
TOTAL .686** .681** .771** .656** .823** .395** .709** .602** .774** 1.000 TIPI
Extraversion .120 -.017 .046 -.036 -.019 -.063 .018 .099 .051 .046 Agreeableness -.120 -.250** .015 -.137 -.086 .134 .103 .314** .025 -.065
Conscientiousness -.134 -.059 -.032 -.081 -.118 -.090 -.172 .158 -.001 -.101 Emotional Stability .059 -.006 .164 -.043 .130 .083 .269** .324** .192* .154
Openness to Experience
-.061 -.204* -.101 .125 -.111 -.119 -.029 .103 -.029 -.097
ProQoL CS .273** .221* .380** .187* .416** .151 .436** .742** .388** .470**
Burnout -.263** -.167 -.342** -.223* .432** -.373** -.483** -.499** -.474** -.495** STS .011 .126 -.108 -.021 -.122 -.214* -.245** -.248** -.143 -.137
* p < .05, ** p < .01; ƒ Coding: 1=male, 2=female; ‡ Coding: 1=single, 2=multiple
90
Supplementary Table S4.3. Spearman correlations (rs) between continuous variables, and point biserial correlations for categorical variables related to personality and professional quality of life.
Ten Item Personality Inventory ProQoL Extraversion Agreeableness Conscientiousness Emotional
Stability Openness to Experience
CS Burnout STS
Age .147 .033 .142 .176* .005 .145 .105 -.046 Gender ƒ .121 .064 -.028 -.031 .130 -.021 .041 .159 Years in Oncology .078 .012 .074 .185* .040 .025 .109 -.101 Years in Healthcare .127 .060 .129 .196* .034 .092 .114 -.014 No. of Sites Worked‡ -.034 .029 -.045 .049 .002 -.057 -.051 -.030 MEQ .032 .100 .151 .105 .007 .178* -.075 .118 PSQI -.125 -.073 -.047 -.266** -.160 -.324** .406** .272 Job Satisfaction
Pay .120 -.120 -.134 .059 -.061 .273** -.263** .011 Promotion -.017 -.250** -.059 -.006 -.204* .221* -.167 .126
Supervision .046 .015 -.032 .164 -.101 .380** -.342** -.108 Benefits -.036 -.137 -.081 -.043 -.125 .187* -.223* -.021
Contingent Rewards -.019 -.086 -.118 .130 -.111 .416** -.432** -.122 Operating Procedures -.063 .134 -.090 .083 -.119 .151 -.373** -.214*
Coworkers .018 .103 -.172 .269** -.029 .436** .483** -.245** Nature of Work .099 .314** .158 .324** .103 .742** -.499** -.248** Communication .051 .025 -.001 .192* -.029 .388** -.474** -.143
TOTAL .046 -.065 -.101 .154 -.097 .470** -.495** -.137 TIPI
Extraversion 1.000 .065 .130 .141 .421** .171 -.146 -.124 Agreeableness .065 1.000 .240** .484** .292** .363** -.375** -.282**
Conscientiousness .130 .240** 1.000 .290** .136 .181* -.143 -.264** Emotional Stability .141 .484** .290** 1.000 .312** .405** -.528** -.504**
Openness to Experience .421** .292** .136 .312** 1.000 .267** -.279** -.169 ProQoL
CS .171 .363** .181* .405** .267** 1.000 -.651** -.189* Burnout -.146 -.375** -.143 -.528** -.279** -.651** 1.000 .502**
STS -.124 -.282** -.264** -.504** -.169 -.189* .502** 1.000 * p < .05, ** p < .01; ƒ Coding: 1=male, 2=female; ‡ Coding: 1=single, 2=multiple
91
Chapter 5!
! Primary Circadian Impacts on Patients, Caregivers, and Dyads
5.1! Abstract
We assessed the influence of chronotype on one’s memory for their sleep quality and coping,
(including their compassion and prosocial behaviour), among cancer patients and spousal
caregivers across chemotherapy treatment. Mixed ANOVAs were used to assess patient and
caregiver responses for potential relationships between chronotype and one’s recalled sleep
quality or coping (assessed by the University of Toronto Inventory of Morningness Eveningness
(UTIME)). Among patients without a caregiver involved in the study, chronotype and sleep were
correlated; caregivers and patients with a caregiver participating in the study, did not show a
correlation between chronotype and sleep quality. In the total patient group, peaks in recalled
coping, compassion and prosocial behaviour reflected chronotype; morning (M) types typically
indicated earlier peak performance compared to neither (N) and evening (E) types. To an extent,
peak performance times were recalled earlier in the day as treatment progressed. Caregivers only
reported a chronotype dependent performance difference at endpoint, relating to their need for
time and space to themselves. Patient chronotype is significantly related to recalled performance
across the day and to an extent to sleep quality. Results among patients with caregivers involved
in the study suggest the possibility that spending too much time going through the cancer
treatment process together may produce poorer sleep quality. Caregivers may not report
chronotype dependent performance fluctuations for several reasons, including being required to
function at their non-optimal time, or believing they were providing consistent care at all times.
92
5.2! Introduction
Current published data suggest that two in five Canadians will develop cancer in their lifetime. It
was estimated that in 2016 alone, there would be 202,400 new cancer diagnoses (Canadian
Cancer Statistics, 2016a). While research on cancer care is dominated by diagnosis, treatment,
and patient care, the disease also has profound effects on caregivers and others who are
associated with the affected individuals. Caregivers in hospitals or hospices, including doctors,
nurses, radiation therapists, and or pharmacists, who provide both physical and emotional
medical attention as part of their professional and voluntary work. Outside of the hospital, non-
vocational caregivers, including typically family members and/or friends also provide physical
and emotional support. In this chapter, we refer to caregivers as the individuals who each patient
identifies as their primary provider of non-vocational care or support, outside of the hospital. We
specifically sought caregivers that were spouses of patients also involved in the study, with the
intention that the same person would be providing consistent care across treatment, and would
identify themselves as being closely involved in the patient’s care. While other family members
and friends can also be caregivers, we sought caregivers that would be closely involved across
treatment throughout the day in their spouse’s care, with the intention of examining the influence
of such a role on patients, the spousal caregivers themselves, and on the dyad relationship.
While cancer research is dominated by understanding, diagnosing and treating the disease per se,
other factors beyond the biology of disease also affect patient and caregiver wellbeing. The
potential impact of temporal biology, specifically, has been given comparatively little attention
in either the understanding of cancer or its treatment (e.g., Hede, 2009), yet it has long been
recognized that both are subject to circadian modulation (e.g., Innominato, Levi, & Bjarnason,
2010; Levi, 2001; Levi, 2006).
Chronotype and sleep timing are expressions of the temporal biology of human beings. When
quantified, they are a means of recognizing and categorizing human behavior and physiology
based on both physiological changes and self-perceptions of optimal timing for physical,
emotional, and mental performance. Chronotype is a stable trait that persist for extended time
periods across one’s lifespan, and demonstrates large shifts coinciding with major changes
during development and aging (Schmidt, Collette, Cajochen & Peigneux, 2007). As chronotype
93
shifts, it affects changes in sleep timing (Koscec, Radosevic-Vidacek & Bakotic, 2014; Simpkin
et al., 2014).
Chronotype has been shown to strongly influence cognitive and emotional regulation and
reactivity in human beings. Cognitive task performance typically changes across the day with
high and low performance correlated with chronotype (for review, see Schmidt, Collette,
Cajochen & Peigneux, 2007). Emotional regulation is related to cognitive function. For example,
cognitive reappraisal, a form of emotional regulation, requires individuals to reinterpret the
significance of emotion-eliciting stimuli or situations. Fluid cognitive ability is necessary to
adjust and reinterpret one’s assessment of stimuli and adjust emotional responding (Opitz, Lee,
Gross & Urry, 2014). In the general population, performance recollection for activities with
different cognitive, physical and emotional demands varies. We have shown previously that
memory for performance throughout the day on tasks with high cognitive demands is
significantly higher than it is for emotionally based tasks (Bellicoso, 2010). Since emotional
regulation, and coping in particular, depend on various cognitive processes (e.g., problem
solving, reframing), it is important to understand how coping is related to chronotype, as this
may shed light on how people form memories for their coping ability.
Sleep quality influences performance ability on a variety of domains. Cognitive function is
enhanced by good sleep quality, and reduced following a poor night of sleep (e.g., Lim &
Dinges, 2008). Cancer patients and caregivers (including medical staff, family and friends) often
report sleep disruptions and poor sleep quality (Ancoli-Israel, Moore & Jones, 2001; Zhang,
Yao, Yang & Zhou, 2014). Sleep disruptions among patients have been attributed to a variety of
predisposing, precipitating, and perpetuating factors. For example, Savard and Morin (2001)
suggest cancer patients may have a heightened predisposition to poor sleep quality due to factors
such as familial history of insomnia, aging, hyper-arousability, and female gender. In the general
population, these factors contribute to poor sleep. Savard and Morin (2001) suggest precipitating
factors that bring about poor sleep may include the cancer itself, medical treatments, disease
related pain or treatments, and the emotional impact that a diagnosis carries.
Therefore, we have hypothesized that the prognosticated outcome of cancer therapy is
determined not only by the aggressiveness of the disease and the type of treatment but also by
both the inherent rhythmicity of the biology of cancer and the rhythmic constitution of the
94
patient and their environment. For example, chemotherapy negatively impacts cognitive function
and emotional regulation, and has been linked with increased cognitive dysfunction, known as
chemo brain or chemo fog, a well-established clinical syndrome where patients often report
cognitive deficits at some time during and or following treatment (Wang, Walitt, Saligan, Tiwari,
Cheung & Zhang, 2015). Both cancer patients and caregivers also typically report (for a variety
of reasons) that stress negatively impacts cognitive function, reflected in a reduction of prefrontal
cortex activity underlying many high order cognitive processes (Arnsten, 2009). Cognitive
deficits have been reported by caregivers of ill friends or family members with a range of
physical and or mental conditions, who themselves report elevated stress levels (Jonsdottir et al.,
2013; Oken, Fonareva & Wahbeh, 2011).
In this chapter, we focus on the influence of chronotype and sleep quality on coping with cancer.
Specifically, we look at how the patient and caregiver provide different perspectives on how they
react to the medical situation, and how each views the others’ behavior. We sought to understand
how chronotype and sleep quality influence the judgment of the patient and caregiver, by
assessing how these two factors are related to the subjective recall of sleep quality and coping
ability. Because a cancer diagnosis precipitates a sequence of medical and personal responses,
we attempted to determine how these factors change across the course of cancer therapy.
5.3! Materials and Methods
5.3.1! Participants and Procedures
Prospective patients receiving cancer treatment at Sunnybrook Health Sciences Centre, Odette
Cancer Centre (Toronto, Ontario, Canada) and when applicable, their spousal caregivers, were
identified by their medical oncologist or nurse, and briefed on the general idea of the study. If
patients and caregivers reported interest in participating, they were approached by the study
coordinator, who explained the details and time commitment of the study. Inclusion criteria
indicated that patients had to be female and receiving adjuvant or neoadjuvant chemotherapy for
stage 1, 2, or 3 breast cancer; caregivers had to have a spouse undergoing either of these
treatments who was enrolled in the study. Participants must be 19 or older, be able to read, write
and speak English at least at a basic level. Participants continuing to follow a shiftwork work
95
schedule during the treatment period were excluded. One female caregiver partook in the study,
however, neither her nor her partner returned correctly completed, usable data.
Subjects were informed that their participation was voluntary and would have no bearing on their
treatment. They were informed that while no formal remuneration would be provided, their
participation in the study might provide the opportunity to reflect on personal coping ability, and
factors that influenced their coping. Usable data from 109 participants (86 patients, 23
caregivers) was included in this study. Among those who had signed up but withdrew, the
general reason for withdrawal was that people felt the study would be too time consuming.
5.3.2! Measures
Participants completed baseline, midpoint and endpoint surveys. Baseline questionnaire
packages included background demographics, the Horne-Östberg Morningness Eveningness
Questionnaire (MEQ), Pittsburgh Sleep Quality Index (PSQI), and the University of Toronto
Inventory of Morningness Eveningness (UTIME) (for descriptions, refer to section 2.2
Instruments). Midpoint and endpoint packages included the PSQI and UTIME. Baseline
packages were completed prior to or on the first day of chemotherapy treatment. Midpoint
packages were completed at the halfway session of a patient’s treatment (the treatment session
number varied depending on the particular type of chemotherapy being received). Endpoint
packages were typically completed at the patient’s final chemotherapy session, however, given
that some patients’ final chemotherapy sessions were cancelled upon assessing their health that
day, some endpoint packages were completed at what would have been the final chemotherapy
session.
Actigraphy was attempted as an objective measure of participants’ circadian rhythms. However,
voluntary compliance (e.g., wearing the watch, keeping continuous activity and sleep reports) for
seventeen participants (14 patients; 3 caregivers) was highly variable among participants, and
due to the possibility that actigraphy might have a negative effect on the patients’ treatment, this
approach was dropped from the study.
5.3.3! Statistical Analysis
These data were assessed in three separate analyses: (1) comparing data between patients as an
entire group, (2) comparing data between caregivers as an entire group, and (3) comparing data
96
between patients and their matched caregivers. Descriptive and frequency data were calculated
for each separate analysis. The MEQ was scored according to the criteria presented by Horne and
Östberg (1976). The PSQI was calculated according to the manual, and using software
downloaded from the PSQI website (Buysse, Reynolds III, Monk, Berman, Kupfer, 1988).
UTIME questions were scored according to the criteria presented in Bellicoso (2010). Spearman
correlations were calculated as they are more robust to outliers in non-normally distributed data
than Pearson correlations.
The analyses of patient and caregiver populations were conducted separately, followed by a
comparison of patients with their caregivers. Exploratory analyses to assess whether data were
normally distributed were conducted, and data were winsorized when necessary. For the
population analyses, mixed ANOVAs were performed (with one between subject factor –
chronotype, and one within subject factor – PSQI or UTIME score). Natural logarithm
transformations were not used as they did not generally work well to normalize data, at times
further skewing the data. For the comparison of patients with their caregivers who also took part
in the study, the Mann-Whitney U Test (Wilcoxon Rank Sum Test) was used due to the non-
parametric nature of the data and small sample size. Mixed ANOVAs (with two between subject
factors, and one within subject factor) were used to assess changes in PSQI and UTIME scores
across treatment.
5.4! Results
5.4.1! Descriptive Statistics
5.4.1.1! Overall Demographic Comparisons
Descriptive and frequency statistics for continuous and categorical demographic variables, along
with the data from the MEQ, PSQI, and UTIME questionnaires are presented in Table 5.1. The
average age of the 86 breast cancer patients completing the study was 50.4 yr ± 11.32. Among
the patient participants living with a partner (72), 69 (80.2%) listed their partner as their primary
caregiver. All participating caregivers (23) were living with their patient partners during the
study. The average age of patients with participating partners was not different from that of
patients without, and the average age of the partners was not significantly different from
patients’ age.
97
Overall, about half of the patient population (52.3%) felt that alternate or outside factors would
not influence how they would cope with the disease or treatment. This belief was more strongly
felt by patients who were either married or living with a partner (57.1%). Significantly, among
the patients with caregivers involved in the study, the majority (60.9%) believed the opposite -
that other factors would influence their coping. On the other hand, the majority of caregivers
(60.9%) did not share this belief.
5.4.1.2! Morningness-Eveningness Distribution
The average MEQ score for the overall patient population was 56.3 (i.e., indicative of being N
type). The distribution of patient MEQ scores indicated that 47.7% were M types, 46.5% were N
types, and 5.8% were E types. However, the distribution of scores is different for patients
without caregivers in the study compared with those with caregivers who participated. Although
it is a smaller group, the patients with caregivers in the study were strongly M type (16/23 or
69.4%), and the average MEQ for this group is M type (MEQ average score = 59.96); whereas
the majority of those without caregivers in the study were N type (33/63 or 52.4%) with an
average MEQ score of 55. Furthermore, caregivers in the study were also more strongly M types
(12/23 or 52.2%) (Data are shown in Table 5.2).
5.4.1.3! Sleep Quality Comparison
The average sleep quality score indicated by the PQSI (Table 5.2) for the overall patient
population across treatment indicated poor sleep. Sleep quality worsened from baseline to
endpoint (8.36, to 8.65), though not significantly (p = .384). Between all three phases, this was
the largest change in sleep quality (baseline to midpoint: p = .871; midpoint to endpoint: p =
.422). On the average, sleep quality was higher for patients without caregivers participating than
for those with caregivers. Sleep quality remained essentially identical for all participants
throughout the treatment. Sleep quality was higher for caregivers than for patients throughout the
study.
5.4.1.4! UTIME Performance Results
Responses to seven UTIME scenarios were accumulated from patients and caregivers at three
points in the treatment (beginning, midpoint, end). Midpoint was timed to coincide with any
change in treatment regimen which is part of a normal protocol. Average scores for each group
98
and time are presented graphically in Figure 5.1. Group means did not differ significantly, and
did not change significantly over time. However, there were some consistent results and trends in
responses to the different scenarios. The consistently lowest score was for UTIME #5 with a
score always indicative of peak feeling of fatigue in the second half of the day. The second
lowest score was consistently seen for UTIME #7 (when do you most, average, and least need
time and space to yourself. Relatively high scores were provided consistently in response to
UTIME #1 (when are you able to best, average, and worst cope with the stress associated with
your diagnosis?) and UTIME #4 (when are you most, average, and least alert?). The result
indicates that patients felt that they could cope with the diagnosis and treatment earlier in the day
than other issues that they might face, which corresponds well with their reports of increased
alertness earlier in the day.
In addition, the group of patients whose caregivers participated in the study showed a reduced
score on all but one UTIME question at the midpoint of their treatment. This may reflect a
different outlook or self-perception that is related to an extended presence of the caregiver. The
trend may also be related to the lower sleep quality that was found in this group. The reason for
this is unclear, and is addressed in the Discussion.
99
Table 5.1 Means and frequencies of demographic and questionnaire data for patients and caregivers
Total patient population Patients without study involved caregiver
Patients with study involved caregiver
Total caregiver population
Mean (SD) Freq. (%) Mean (SD) Freq. (%) Mean (SD) Freq. (%) Mean (SD) Freq. (%) Total population size 86 63 23 23 Age 50.38
(11.32) 49.79
(10.81) 52.00
(12.73) 54.22 (13.21)
Relationship Status In relationship, not living together
3 (3.5%) 3 (4.8%) - -
Living together 6 (7.0%) 5 (7.9%) 1 (4.3%) 1 (4.3%) Married, living together 66 (76.7%) 44 (69.8%) 22 (95.7%) 22 (95.7%) Divorced 2 (2.3%) 2 (3.2%) - - Separated 1 (1.2%) 1 (1.6%) - - Single 7 (8.1%) 7 (11.1%) - - Widowed 1 (1.2%) 1 (1.6%) - - Patient with partner as caregiver
Yes 69 (80.2%) 46 (73.0%) 23 (100%) n/a No 3 (3.5%) 3 (4.8%) - n/a Not applicable 14 (16.3%) 14 (22.2%) - n/a Alternate coping influences Yes 40 (46.5%) 26 (41.3%) 14 (60.9%) 8 (34.8%) No 45 (52.3%) 36 (57.1%) 9 (39.1%) 14 (60.9%) Not applicable 1 (1.2%) 1 (1.6%) - 1 (4.3%)
100
Table 5.2 Means and frequencies for chronotype and sleep quality questionnaires for patients and caregivers Total patient population Patients without study
involved caregiver Patients with study involved caregiver
Total caregiver population
Mean (SD) Freq. (%) Mean (SD) Freq. (%) Mean (SD) Freq. (%) Mean (SD) Freq. (%) A.! MEQ All scores 56.33 (9.37) 55.00
(9.707) 59.96
(7.425) 57.22
(8.464)
Morning type 41 (47.7%) 25 (39.7%) 16 (69.4%) 12 (52.2%) Neither type 40 (46.5%) 33 (52.4%) 7 (30.4%) 10 (43.5%) Evening type 5 (5.8%) 5 (7.9%) - 1 (4.3%) B.! PSQI Baseline 8.36 (3.63) 7.92 (3.521) 9.57 (3.727) 6.87 (3.152) Midpoint 8.42 (3.24) 8.29 (3.314) 8.78 (3.089) 6.91 (3.147) Endpoint 8.65 (3.37) 8.32 (3.440) 9.57 (3.057) 6.87 (3.622)
101
Figure 5.1a UTIME scores across treatment for the total patient group
102
Figure 5.1b UTIME scores across treatment for patients without a caregiver involved in the
study
103
Figure 5.1c UTIME scores across treatment for patients with a caregiver involved in the study
104
Figure 5.1d UTIME scores across treatment for caregivers
105
5.4.2! Analysis 1: Total Patient Population Analysis
5.4.2.1! Multifactorial Correlation Analysis
The treatment of breast cancer occurs in stages and is a lengthy and highly personal
(individualized) process. Coping, therefore, is affected by several factors. In this section, we
consider the influences of chronotype on sleep quality and coping performance at each phase of
assessment across chemotherapy treatment.
5.4.2.2! Global Correlations Among Patients’ MEQ, PSQI and UTIME data
Spearman correlations were determined first for the data from the three questionnaires (MEQ,
PSQI, and UTIME) (Table 5.3). Among the 86 breast cancer patients, correlations between
chronotype and sleep quality were not significant, although there was an apparent trend toward
significance as the course of treatment progressed (Table 5.3). However, a distinction emerged
between patients with and without caregivers. Spearman correlations between MEQ and PSQI
for the 23 breast cancer patients whose partner partook in the study were not significantly
correlated, and the non-significant baseline correlation became even less significant by endpoint.
Contrary to this, among the 63 breast cancer patients without a caregiver partaking in the study,
the correlations suggested a significant negative association between chronotype and sleep
quality across treatment.
Spearman correlations between the seven UTIME scores and MEQ and for the complete group
of 86 breast cancer patients were significantly correlated for nearly all questions across treatment
(Table 5.3). The MEQ-UTIME #5 correlation (when are you most, average, and least fatigued
during the day?) never reached significance at any assessment phase in the cancer treatment
period. UTIME #7 showed weak correlations with MEQ and reached significance (p = .003) at
the midpoint of treatment, but was again not significant by endpoint.
Sleep quality and UTIME scores were not significantly correlated across treatment (data not
shown), except for UTIME #5 at midpoint (rs = .218, p = .045), and UTIME #3 at endpoint, (rs =
.268, p = .026); UTIME #3 only applied to patients with a partner (when is your partner able to
best, average, and worst help you cope with the stress of being a cancer patient?). However,
neither correlation reached significance at either of the other two assessment phases.
106
Table 5.3 Spearman correlations for chronotype versus sleep quality and UTIME among patients
Assessment Phase MEQ vs. Baseline Midpoint Endpoint rs p rs p rs p
A.! PSQI Total patient population -.101 .353 -.163 .133 -.170 .118 Patients, caregiver involved .388 .067 .377 .076 .312 .147 Patients, caregiver not involved -.312* .013 -.339** .007 -.328** .009
B.! UTIME UTIME #1 .531** .000 (86) .468** .000 (86) .411** .000 (86) UTIME #2 .578** .000 (70) .583** .000 (67) .481** .000 (69) UTIME #3 .441** .000 (69) .344** .004 (67) .373** .002 (69) UTIME #4 .479** .000 (85) .545** .000 (86) .458** .000 (86) UTIME #5 .097 .374 (86) -.023 .834 (85) .153 .159 (86) UTIME #6 .513** .000 (68) .519** .000 (67) .492** .000 (69) UTIME #7 .202 .063 (86) .317** .003 (85) .173 .114 (85)
Note: numbers in brackets beside MEQ vs. UTIME p-values are n values * p < .05 ** p < .01
107
5.4.2.3! Total Patient Population Mixed Measures ANOVAs
Mixed measures ANOVAs were used to assess the influence of chronotype on changes in sleep
quality and UTIME scores at each assessment phase (baseline, midpoint and endpoint) across
chemotherapy treatments among the total group of 86 breast cancer patients (Table 5.3). The
within subject factor was assessment phase (score at baseline, midpoint, and endpoint). The
between subject factor was chronotype (moderately E type, N type, moderately M type, and
definitely M type); the two largest groups of these between subject factors were N types and
moderately M types.
A standard procedure was used to assess changes in sleep quality or UTIME scores at
assessment times across treatment between chronotypes. Data were assessed for outliers based
on inspection of boxplot values greater than 1.5 box lengths from the edge of the box. Outliers
(detected only among UTIME scores and not PSQI) were corrected for by winsorizing the data
in order to use a standard procedure for each analysis. The Shapiro-Wilk’s test (p > .05) was
used to assess normality of distribution for sleep quality or UTIME score at each between and
within subject level. Homogeneity of variance (p > .05) was assessed by Levene’s test of
homogeneity of variance; any violations were left untransformed and noted. Homogeneity of
covariance (p > .05) was assessed by Box’s test of equality of covariance matrices; any
violations were noted and the interaction term was not interpreted but noted for reference.
Mauchly’s test of sphericity was used to assess whether the assumption of sphericity was met (p
> .05); the Greenhouse-Geisser estimate was used when the assumption was violated. Mixed
ANOVA outputs for sleep quality and UTIME are presented in Table 5.4.
There was no significant interaction between chronotype and assessment phase on either sleep
quality or UTIME score among the total group of 86 breast cancer patients, p > .05. Only
significant between and within subject effects as noted in Table 5.4 are discussed below.
There were no significant main effects at the between or within subject level when assessing
changes in sleep quality between chronotypes or within subjects at different time points.
108
When assessing changes in UTIME #1 scores across assessment phases between chronotypes,
between subjects, the main effect of chronotype showed a significant difference in UTIME #1
score at the different assessment phases p < .001. Definitely M types had a mean UTIME #1
score 0.150 (CI 95%, .002 to .298) points higher than moderately E types, p < .05. Moderately M
types had a mean UTIME #1 score 0.076 (CI 95%, .031 to .122) points higher than N types, p <
05, and 0.122 (CI 95%, .026 to .218) points higher than moderately E types, p < .01. This
indicates that for UTIME #1 (best, average, and worst ability to cope with cancer diagnosis
across the day), patients report reaching their peak coping ability in chronotypically sequential
order, beginning with definitely M types reporting reaching their best coping point earliest in the
day compared to N and E types. Within subjects, the main effect of time showed a statistically
significant difference in UTIME #1 score at the different assessment phases, p = .019. UTIME
#1 score at endpoint showed a statistically significant increase of 0.052 (95% CI, -.090 to -0.014)
points, p < .01, over mean score at midpoint.
When assessing changes in UTIME #2 scores across assessment phases between chronotypes,
between subjects, the main effect of chronotype showed a significant difference in UTIME #2
score at the different assessment phases, p < .001. Definitely M types had a mean UTIME #2
score 0.123 (CI 95%, .005 to .241) points higher than N types, p < .05. Moderately M types had
a mean UTIME #2 score 0.077 (CI 95%, .028 to .127) points higher than N types, p < .001.
Within subjects, the main effect of time showed a significant difference in UTIME #2 score at
the various assessment phases, p = .038. There was a significant increase in UTIME #2 score
across treatment of 0.060 (95% CI, -.118 to -.002), p < .05, from baseline to endpoint.
When assessing changes in UTIME #3 scores across assessment phases between chronotypes,
within subjects, the main effect of time suggested there was a significant difference in UTIME
#3 score at the different assessment phases, p = .044. However, when considering pairwise
comparisons in mean UTIME #3 score within subjects, there were no statistically significant
differences with significance values below p < .05. The closest value to significance was at
endpoint, where mean UTIME #3 scores were 0.064 (CI 95%, -.001 to .130) points above
baseline scores, p < .100 (p = .057).
When assessing changes in UTIME #4 scores across assessment phases between chronotypes,
between subjects, the main effect of chronotype showed a significant difference in UTIME #4
109
score at the different assessment phases, p < .001. Definitely M types had a mean UTIME #4
score 0.177 (CI 95%, .025 to.329) points higher than moderately E types, p < .05. Moderately M
types had a mean UTIME #4 score 0.081 (CI 95%, .033 to .128) points higher than N types, p <
.001, and 0.159 (CI 95%, .060 to .258) points higher than E types, p < .001.
When assessing changes in UTIME #5 scores across assessment phases between chronotypes,
between subjects, the main effect of chronotype suggested there was a significant difference in
UTIME #5 score at the different assessment phases, p = .041. However, when considering
pairwise comparisons in mean UTIME #5 score between various chronotypes, there were no
significant differences with significance values, p < .05. The closest value to significance was
found between N types who had a mean UTIME #5 score -0.060 (CI 95%, -.121 to .001) points
lower than moderately E types, p = .059. While this is not considered statistically significant,
examining Tukey HSD in multiple comparisons shows that N types do have a mean UTIME #5
score -.060 (95% CI, -.119 to -.000) points below moderately E types, a significant difference, p
= .047.
When assessing changes in UTIME #6 scores across assessment phases between chronotypes,
between subjects, the main effect of chronotype showed a significant difference in UTIME #6
score at the different assessment phases, p < .01. Definitely M types had a mean UTIME #6
score 0.104 (CI 95%, .006 to.20) points higher than N types, p < .05. Moderately M types had a
mean UTIME #6 score 0.075 (CI 95%, .034 to .117) points higher than N types, p < .001.
When assessing changes in UTIME #7 scores across assessment phases between chronotypes,
between subjects, the main effect of chronotype showed a significant difference in UTIME #7
score at the different assessment phases, p = .045. Moderately M types had a mean UTIME #7
score of 0.041 (95% CI, 0.000 to 0.081) points higher than N types, p < .05.
110
Table 5.4 Mixed ANOVAs on patient sleep quality and UTIME: main effects and interactions Chronotype
(Between Subject) Assessment phase
Baseline, Midpoint, Endpoint (Within Subject)
Assessment phase x Chronotype (Within Subject)
F df p η2 F df p η2 F df p η2 PSQI 1.014 3, 82 .391 .036 0.168 2, 164 .845 .002 0.401 6, 164 .877 .014 UTIME #1 9.593 3, 82 .000** .260 4.223 1.843, 164 .019* .049 1.422 5.528, 164 .214 .049 UTIME #2 7.328 3, 62 .000** .262 3.366 2, 124 .038* .051 1.950 6, 124 .078 .086 UTIME #3 2.740 3, 61 .051 .119 3.204 2, 122 .044* .050 2.170 6, 122 .050 .050 UTIME #4 11.468 3, 81 .000** .298 0.196 1.788, 162 .798 .002 1.590 5.365, 162 .162 .162 UTIME #5 2.888 3, 81 .041* .097 0.861 2, 162 .425 .011 0.556 6, 162 .765 .020 UTIME #6 9.404 3, 61 .000** .316 1.781 2, 122 .173 .028 1.012 6, 122 .421 .047 UTIME #7 2.083 3, 81 .045* .094 1.635 2, 162 .198 .020 0.981 6, 162 .440 .035
* p < .05 ** p < .01 NOTE: PSQI: Shapiro Wilks violated for N types at baseline (p = .017) UTIME #1: Levene’s test violated at endpoint (p = .016); Box’s test violated (p = .015); Mauchly’s test of sphericity violated, χ2 (2) = 7.229, p = .027 UTIME #2: Levene’s test violated at endpoint (p = .008); Box’s test violated (p = .013) UTIME #3: Shapiro Wilk’s violated for N types at endpoint (p = .022); Box’s test violated (p = .000) UTIME #4: Shapiro Wilk’s violated for N types at baseline (p = .012); Levene’s test violated at baseline (p = .001) and midpoint (p = .000); Mauchly’s test of sphericity violated, χ2 (2) = 10.080, p = .006 UTIME #5: Shapiro Wilk’s violated for moderately M types at baseline (p = .016) and midpoint (p = .044), and definitely M types at midpoint (p = .044); Levene’s test violated at baseline (p = .049); Box’s test violated (p = .023)
111
5.4.3! Analysis 2: Total Caregiver Population Analysis
5.4.3.1! Global Correlations Among Caregivers’ MEQ, PSQI and UTIME data
Spearman correlations indicated no significant association between MEQ and PSQI scores for
the 23 caregivers (Table 5.5). The association was positive, and significance values at each
assessment phase varied greatly from one another. No trend suggested the association was
moving towards or further away from significance. Spearman correlations between MEQ and
UTIME score (Table 5.5) were also non-significant for all questions, except for UTIME #7 at
endpoint, p < 001. Baseline data suggested a positive trend in the association between MEQ and
all UTIME questions. However, while the association remained positive for all other questions at
midpoint and endpoint, MEQ and UTIME #5 showed a negative relationship at midpoint and
endpoint. As in Analysis 1, there were only few significant correlations between PSQI and
UTIME scores (not reported here), however they were not consistently significant across
treatment. The correlation between PSQI and UTIME #2 was only significant at baseline (rs =
.444, p = .034), and endpoint (rs = .441, p = .035). UTIME #3 and UTIME #6 only showed a
significant correlation to PSQI at endpoint (rs = .458, p = .028, and rs = .531, p = .009,
respectively). No significant correlations existed between PSQI and UTIME score at midpoint.
112
Table 5.5 Spearman correlations for caregiver chronotype versus sleep quality and UTIME
Assessment Phase MEQ vs. Baseline Midpoint Endpoint rs p rs p rs p A.! PSQI
.091 .681 .156 .477 .054 .805 B.! UTIME UTIME #1 .130 .553 .215 .325 .374 .078 UTIME #2 .102 .645 .167 .448 .205 .348 UTIME #3 .394 .063 .170 .438 .340 .112 UTIME #4 .165 .451 .107 .626 .282 .192 UTIME #5 .194 .375 -.099 .654 -.168 .442 UTIME #6 .320 .137 .079 .720 .059 .790 UTIME #7 .366 .094 .036 .870 .653** .001
** p < .01
113
5.4.3.2! Caregiver Population Mixed Measures ANOVAs
A standard procedure was used to perform mixed measures ANOVAs to assess changes in sleep
quality or UTIME scores at assessment times across treatment between chronotypes among the
total group of 23 spousal caregivers to breast cancer patients. The within subject factor was
assessment time (score at baseline, midpoint, and endpoint). The between subject factor was
chronotype (divided into two categories: moderately E type and N type versus moderately and
definitely M type). There was only one moderately E type, and one definitely M type individual,
therefore, these scores were joined with N type and moderately M type chronotypes,
respectively.
Data were assessed for outliers based on inspection of boxplot values greater than 1.5 box
lengths from the edge of the box. Outliers (detected only among UTIME #3 - #7) were corrected
for by winsorizing the data in order to use a standard procedure for each analysis. The Shapiro-
Wilk’s test (p > .05) was used to assess normality of distribution for sleep quality or UTIME
score at each between and within subject level. Homogeneity of variance (p > .05) was assessed
by Levene’s test of homogeneity of variance; any violations were left untransformed and noted.
Homogeneity of covariance (p > .05) was assessed by Box’s test of equality of covariance
matrices; any violations were noted and the interaction term was not interpreted but noted for
reference. Mauchly’s test of sphericity was used to assess whether the assumption of sphericity
was met (p > .05); the Greenhouse-Geisser estimate was used when the assumption was violated.
Mixed ANOVA outputs for sleep quality and UTIME are presented in Table 5.6.
Between subjects, the main effect of chronotype did not show a significant difference in sleep
quality or UTIME#1 - #7 scores at the different assessment phases, p > .05. Within subjects, the
main effect of time did not show a significant difference in sleep quality or UTIME#1 - #7 score
at the different assessment phases, p > .05. There were no significant interactions between
chronotype and assessment phase across treatment on sleep quality or UTIME #1-#6. However,
there was a significant interaction between chronotype and assessment phase on UTIME #7
score, p = .010. Specifically, there was a highly significant difference in UTIME #7 score
between chronotypes at the endpoint of treatment, F(1, 21) = 14.649, p = .001, partial η2 = .411.
UTIME #7 score was significantly lower at endpoint among N and moderately E types (M =
114
.441, SE = .025) compared to M types (M = .561, SE = .023), suggesting N and moderately E
types reached their peak time for most needing space to themselves later in the day than M types.
115
Table 5.6 Mixed ANOVAs on caregiver sleep quality and UTIME: main effects and interactions Chronotype
(Between Subject) Assessment phase
Baseline, Midpoint, Endpoint (Within Subject)
Assessment phase x Chronotype (Within Subject)
F df p η2 F df p η2 F df p η2 PSQI 0.140 1, 21 .712 .007 0.004 2, 42 .996 .000 0.406 2, 42 .669 .019 UTIME #1 0.778 1, 21 .388 .036 0.082 2, 42 .922 .004 2.284 2, 42 .114 .098 UTIME #2 0.298 1, 21 .591 .014 1.585 2, 42 .217 .070 0.732 2, 42 .487 .034 UTIME #3 1.240 1, 21 .278 .056 0.942 2, 42 .398 .043 1.139 2, 42 .330 .051 UTIME #4 0.100 1, 21 .754 .005 2.097 2, 42 .135 .091 2.861 2, 42 .068 .120 UTIME #5 1.309 1, 21 .265 .059 0.389 2, 42 .680 .018 1.289 2, 42 .286 .058 UTIME #6 0.090 1, 21 .767 .004 1.945 2, 42 .156 .156 1.335 2, 42 .274 .274 UTIME #7 3.329 1, 21 .083 .143 0.893 2, 42 .417 .043 5.147 2, 42 .010* .205
* p < .05 NOTE: PSQI: Levene’s test violated at baseline (p = .025) UTIME #3: Shapiro Wilk’s violated for M types at midpoint (p = .036); Levene’s test violated at midpoint (p = .006) UTIME #4: Shapiro-Wilk’s violated for E and N types at midpoint (p = .013); Levene’s test violated at midpoint (p = .011); Box’s test violated (p = .018)
116
5.4.4! Analysis 3: Matched Patient and Caregiver Population Analysis
5.4.4.1! Global Associations and Differences Among Matched Patient and Caregiver MEQ, PSQI and UTIME data
The Mann Whitney U (non-parametric test) was used to assess differences between matched
patients and caregivers on their MEQ, PSQI, and UTIME scores (Table 5.7). A non-parametric
test was used due to the small sample size, and the non-normal distribution of data. The only
significant results were the differences between patients and caregivers on all three PSQI scores.
Distributions of the patient and caregiver scores for PSQI at baseline, midpoint, and endpoint
were not similar, as assessed by visual inspection. Sleep quality was statistically significantly
worse among patients at baseline, midpoint, and endpoint (Mdn = 10.0, 9.0, and 10.0) than
among caregivers (Mdn = 8.0, 8.0, and 9.0) (baseline: U = 158.500, z = -2.339, p = .019;
midpoint: U = 172.500, z = -2.032, p = .042; endpoint: U = 136.000, z = -2.837, p = .005). In
these cases, the null hypothesis (H0 = the distribution of scores for the two groups are equal) is
rejected, and the alternate hypothesis (HA = the distribution of scores for the two groups are not
equal), is accepted. For age, chronotype, and UTIME questions assessing memory for their
personal and their partner’s coping abilities, patients and caregivers did not differ significantly.
117
Table 5.7 Median and Mann Whitney U significance values comparing MEQ, PSQI, and UTIME response scores between patients and their caregivers Medians Patient Caregiver U z p Age 52.00 52.00 292.000 0.605 .545 MEQ 61.00 59.00 214.500 -1.100 .271 PSQI
Baseline 10.00 6.00 158.500 -2.339 .019* Midpoint 9.00 7.00 172.500 -2.032 .042* Endpoint 10.00 7.00 136.000 -2.837 .005**
Baseline UTIME #1 .5857 .5329 205.000 -1.307 .191 UTIME #2 .5545 .5659 278.500 0.308 .758 UTIME #3 .5417 .5000 213.000 -0.909 .364 UTIME #4 .5865 .5710 229.000 -0.780 .435 UTIME #5 .4308 .4516 230.500 -0.747 .455 UTIME #6 .5392 .5000 196.000 -1.295 .195 UTIME #7 .5256 .5000 256.500 0.080 .937
Midpoint UTIME #1 .5139 .5256 271.000 0.143 .886 UTIME #2 .5000 .5192 284.000 0.429 .668 UTIME #3 .4900 .5000 302.500 0.836 .403 UTIME #4 .5545 .5929 297.000 0.714 .475 UTIME #5 .4304 .4870 292.000 0.604 .546 UTIME #6 .4910 .5256 278.500 0.308 .758 UTIME #7 .5000 .5000 231.500 -0.726 .468
Endpoint UTIME #1 .5705 .5256 231.500 -0.725 .468 UTIME #2 .5256 .5288 258.000 -0.143 .886 UTIME #3 .5000 .5256 320.000 1.221 .222 UTIME #4 .5604 .6046 351.000 1.901 .057 UTIME #5 .4981 .4402 190.000 -1.638 .101 UTIME #6 .5252 .5321 286.500 0.484 .629 UTIME #7 .5481 .5000 271.000 -1.044 .297
* p < .05 ** p < .01
118
5.4.4.2! Matched Patient and Caregiver Group Mixed Measures ANOVAs
Three-way mixed ANOVAs were run to understand the effects of one’s role (patient or
caregiver), chronotype, and assessment phase across treatment (baseline, midpoint, endpoint) on
sleep quality and UTIME performance scores. The within subject factor was time (score at
baseline, midpoint, and endpoint). The between subject factors were chronotype, (divided into
two categories: moderately E and N type, versus moderately and definitely M type, due to the
small group size), and role (patient versus caregiver). ANOVAs are robust to deviations from
normality, and therefore a good choice when analyzing irregularly distributed data.
Data were assessed for outliers based on inspection of boxplot values greater than 1.5 box
lengths from the edge of the box. Outliers (for sleep quality and UTIME #1-#7) were corrected
for by winsorizing the data in order to use a standard procedure for each analysis. The Shapiro-
Wilk’s test (p > .05) was used to assess normality of distribution for sleep quality or UTIME
score at each between and within subject level. Homogeneity of variance (p > .05) was assessed
by Levene’s test of homogeneity of variance; any violations were left untransformed and noted.
Mauchly’s test of sphericity was used to assess whether the assumption of sphericity was met (p
> .05); the Greenhouse-Geisser estimate was used when the assumption was violated. Mixed
ANOVA outputs for sleep quality and UTIME are presented in Table 5.8, including F-values,
degrees of freedom, levels of significance, and partial eta squared (η2).
When assessing changes in sleep quality at each assessment phase across treatment between
chronotypes and roles, the three-way interaction between treatment time, role, and chronotype
was not statistically significant, p = .848. All two-way interactions were not statistically
significant (p > .05).
When assessing changes in UTIME scores at each assessment phase across treatment between
chronotypes and roles, the three-way interaction between assessment phase, role, and chronotype
was only statistically significant for UTIME #4, p = .027, and UTIME #7, p = .002. This
indicates that the interactions of role and chronotype are different at the individual assessment
phases for UTIME #4 and #7 (recalled alertness, and recalled need for time and space to oneself,
respectively).
119
For UTIME #4, significance of a simple two-way interaction was accepted at a Bonferroni-
adjusted alpha level of .017 (i.e., 0.05 ÷ 3 = 0.017, divide by 3 for 3 assessment phases). There
was a significant interaction of role and chronotype at midpoint, F(1, 42) = 21.005, p < .001, but
not at baseline, F(1, 42) = 2.817, p = .101, or endpoint, F(1, 42) = 0.258, p = .614. Significance
of a simple main effect was accepted at a Bonferroni-adjusted alpha level of .025 (i.e., 0.05 ÷ 2 =
0.025, divide by 2 for 2 chronotype groups). There was a statistically significant simple simple
main effect for moderately E and N type chronotypes at midpoint, F(1, 42) = 24.697, p < .001,
but not for moderately and definitely M type, F(1, 42) = 1.215, p = .277. Pairwise comparisons
were performed for significant simple simple main effects. Bonferroni corrections were made
with comparisons within each simple simple main effect considered as a family of comparisons.
Adjusted p-values are reported. Mean UTIME #4 score was higher for moderately E and N type
caregivers (M = 0.601, SD = 0.018) than patients (M = 0.458, SD = 0.023), a mean difference of
.143 [95% CI, 0.085 to 0.201], p < .001. There was also a statistically significant simple simple
main effect for patients at midpoint, F(1, 42) = 19.785, p < .001, but not for caregivers, F(1, 42)
= 3.747, p = .060. All pairwise comparisons were performed for statistically significant simple
simple main effects. Bonferroni corrections were made with comparisons within each simple
simple main effect considered as a family of comparisons. Adjusted p-values are reported. Mean
UTIME #4 scores were higher for patients with moderately and definitely M type chronotype (M
= 0.578, SD = 0.015) than those with moderately E and N type chronotypes (M = 0.458, SD =
0.023), a mean difference of .120 [95% CI, 0.066 to 0.175], p < .001.
For UTIME #7, significance of a simple two-way interaction was accepted at a Bonferroni-
adjusted alpha level of .017 (due to there being three simple two-way interactions, i.e., .05 ÷ 3 =
.017). At p < .017, there was a significant simple two-way interaction of role and chronotype at
endpoint, , F(1, 42) = 6.762, p = .013, but not at baseline, F(1, 41) = 0.366, p = .549, or
midpoint, F(1, 42) = 5.692, p = .022. Significance of a simple simple main effect was accepted at
a Bonferroni-adjusted alpha level of .025 (due to there being two simple simple main effects for
chronotype, i.e., .05 ÷ 2 = .025). There was a significant main effect of chronotype for caregivers
at endpoint, F(1, 42) = 11.414, p = .002. All pairwise comparisons were performed for
statistically significant simple simple main effects. Bonferroni corrections were made with
comparisons within each simple simple main effect considered a family of comparisons.
Adjusted p-values were reported. Among caregivers, UTIME #7 score was higher in moderately
120
and definitely M types (M = 0.561, SD = 0.026) than in moderately E and N types (M = 0.435,
SD = 0.027), a mean difference of 0.126, 95% CI [0.051 to 0.201], p = .002.
At the level of two-way interactions, certain UTIME assessments demonstrated interactions
between chronotype and role, or between assessment phase and role; no significant interactions
were found between assessment phase and chronotype. For UTIME #2, a significant interaction
was found between role and chronotype, p = .034. Significance of a simple main effect was
accepted at a Bonferroni-adjusted alpha level of .025 (due to there being two simple main effects,
i.e., .05/2 = .025). There was a significant main effect of chronotype for patients, F(1, 42) =
11.806, p = .001, η2 = .219, but not for caregivers, F(1, 42) = 0.249, p = .621, η2 =.006. All
pairwise comparisons were performed for significant simple main effects. Bonferroni corrections
were made with comparisons within each simple main effect considered a family of
comparisons. Adjusted p-values are reported. Mean UTIME #2 score was higher in moderately
and definitely M type patients than moderately E and N type patients, a mean difference of .088,
[95% CI, -0.059 to 0.036], p = .001.
It appeared that for UTIME #5 there was a two-way interaction between role and chronotype, p =
.038. However, statistical significance of a simple main effect was accepted at a Bonferroni-
adjusted alpha level of .025; therefore, with this correction there was determined to be no
significant two-way interactions, p > .025.
For UTIME #6, a significant two-way interaction exists between role and chronotype, p = .034.
Significance of a simple main effect was accepted at a Bonferroni-adjusted alpha level of .025.
There was a significant simple main effect of chronotype for patients, F(1, 41) = 9.499, p = .004,
but not for caregivers F(1, 41) = 0.043, p = .837. All pairwise comparisons were performed for
significant simple main effects. Bonferroni corrections were made with comparisons within each
simple main effect considered a family of comparisons. Adjusted p-values are reported. Mean
UTIME #6 score was higher in moderately and definitely M type patients than in moderately E
and N type patients, a mean difference of .081 [95% CI, 0.028 to 0.135].
For UTIME #3, there was a significant two-way interaction between assessment phase and role,
p = .014. Significance of a simple main effect was accepted at a Bonferroni-adjusted alpha level
of .017 (due to there being three simple main effects, i.e., .05/3 = .017). There was a statistically
significant simple main effect of participant role at endpoint, F(1, 41) = 6.275, p = .016, η2 =
121
.133, but not at baseline, F(1, 41) = 0.569, p = .455, η2 = .014, or midpoint, F(1, 41) = 4.114, p =
.049, η2 = .09. All pairwise comparisons were performed for significant simple main effects.
Bonferroni corrections were made with comparisons within each simple main effect considered
as a family of comparisons. Adjusted p-values are reported. Mean UTIME #3 scores were higher
in caregivers, a mean difference of .055, [95% CI, .011 to .099], p = .016.
122
Table 5.8 Mixed ANOVAs on patient AND caregiver sleep quality and UTIME: main effects and interactions
Role x Chronotype (Between Subjects)
Assessment Phase (Within Subjects)
Assessment Phase x Role (Within Subjects)
Assessment Phase x Chronotype (Within Subjects)
Assessment Phase x Role x Chronotype
(Within Subjects) F df p η2 F df p η2 F df p η2 F df p η2 F df p η2 PSQI 0.620 1, 42 .436 .015 1.069 2, 84 .348 .025 1.146 2, 84 .323 .027 0.341 2, 84 .712 .008 0.165 2, 84 .848 .004 UTIME #1 3.115 1, 42 .085 .069 2.302 1.575, 84 .119 .052 1.568 1.575, 84 .219 .036 2.337 1.575, 84 .116 .053 0.384 1.575, 84 .633 .009 UTIME #2 4.793 1, 42 .034* .102 3.349 2, 84 .040 .074 0.189 2, 84 .828 .004 0.371 2, 84 .691 .009 0.461 2, 84 .632 .011 UTIME #3 0.341 1, 41 .563 .008 2.038 2, 82 .137 .047 4.502 2, 82 .014* .099 0.026 2, 82 .974 .001 1.942 2, 82 .150 .045 UTIME #4 6.758 1, 42 .013 .139 0.622 1.683, 84 .513 .015 4.300 1.683, 84 .023 .093 0.593 1.683, 84 .527 .014 4.088 1.683, 84 .027* .089 UTIME #5 4.610 1, 42 .038* .099 0.615 2, 84 .543 .014 2.125 2, 84 .126 .048 0.137 2, 84 .872 .003 1.617 2, 84 .205 .037 UTIME #6 4.784 1, 41 .034* .104 2.078 2, 82 .132 .048 1.280 2, 82 .283 .030 0.023 2, 82 .978 .001 2.192 2, 82 .118 .051 UTIME #7 0.308 1, 41 .582 .007 1.891 1.642, 82 .166 .044 0.220 1.642, 82 .759 .005 0.922 1.642, 82 .386 .022 7.647 1.642, 82 .002** .157
* p < .05 ** p < .01 NOTES: PSQI: Levene’s test violated at baseline (p = .007) UTIME #1: Levene’s test violated at baseline (p = .042); Mauchly’s test of sphericity violated, χ2 (2) = 12.916, p = .002 UTIME #3: Shapiro-Wilk’s violated for E and N type patients at midpoint (p = .016), and M type caregivers at midpoint (p = .036); Levene’s test violated at midpoint (p = .002) UTIME #4: Shapiro-Wilk’s violated for E and N type caregivers at midpoint (p = .013); Levene’s test violated at midpoint (p = .004); Mauchly’s test of sphericity violated, χ2 (2) = 8.547, p = .014 UTIME #5: Levene’s test violated at baseline (p = .006) UTIME #7: Mauchly’s test of sphericity violated, χ2 (2) = 9.838, p = .007
123
5.5! Discussion
Influences of chronotype and sleep quality on coping among cancer patients have been described
previously in only a few separate reports (e.g., Hansen & Lassen, 2012; Ancoli-Israel et al.,
2014). In this study, we sought to build on this background, recognizing that chronotype and
sleep quality are strongly associated with cognitive function and emotional regulation in the
general population (e.g., Clark, Watson & Leeka, 1989; Killgore, 2010; Lara, Madrid & Correa,
2014; Lim & Dinges, 2008). These associations could provide insight into how coping might be
affected by chronobiology either directly or indirectly through temporal regulation of executive
functions. To investigate this, we performed a series of mixed measures ANOVA analyses to
assess correlations among chronotype, sleep quality, and executive functions on the ability to
cope with the disease. Assessments were made at three points in the treatment, as it was possible
that coping strategies might be adjusted as patients dealt with the experience. In addition, the
involvement of caregivers was considered as a potentially strong environmental influence on
coping strategy.
5.5.1! On the Independent Impacts of Sleep and Chronotype
Chronotype has been defined alternatively as either a time of day preference for physical and
mental performance (e.g., Horne & Östberg, 1976) or the timing of sleep (Merrow et al., 2003).
The two metrics, preference and mid-sleep time, are correlated, and both are correlated with
sleep quality in previous work (Keller, Grünewald, Vetter, Roenneberg & Schulte!Körne, 2017;
Rique, Fernandes Filho, Ferreira & de Sousa-Muñoz, 2014). Although our findings did not
produce a significant correlation between chronotype and sleep quality, the results indicated a
trend in the group of 86 patients which became progressively more negative, moving towards
significance over the course of treatment. This may indicate that sleep quality at the outset of
chemotherapy was not associated to chronotype, but that as treatment progressed, patients began
shifting towards following sleep patterns and habits that reflected their own chronotype. While
sleep is poor across treatment in this group, the trend towards a correlation between MEQ and
PSQI may indicate that at the outset of treatment, sleep is poor due to one’s stress stemming
from their diagnosis. As treatment progresses, while sleep quality does not ameliorate, one’s
sleep patterns may be less disorganized due to stress and irregular sleeping and people may be
following sleep schedules that reflect their chronotype.
124
While chronotype does not predict sleep quality as strongly in this total group as it does in the
general population, the negative correlation indicates that better sleep quality was associated
with increasing tendency towards morningness (lower PSQI scores linked to higher MEQ),
which reflects the relationship in the general human population (Soehner, Kennedy & Monk,
2011; Vardar, Vardar, Molla, Kaynak & Ersoz, 2008). Sleep quality among all patients was poor
across treatment, both by PSQI standards (score > 5) and by research standards that have
suggested a PSQI score > 8 among cancer patients indicates poor sleep quality (Carpenter &
Andrykowski, 1998; Vargas, Wohlgemuth, Antoni, Lechner, Holley & Carver, 2010). Several
patients in our study reported that their sleep was of poor quality, and was worsening. Therefore,
the findings suggest there is a common factor shared among patients that affects sleep.
Interestingly, the relationship between chronotype and sleep quality differed among the 23
patients who had caregivers involved in the study and the 63 patients who were either single
(without a spousal caregiver), or had a partner who was not involved in this study. In the former
group, sleep quality was not significantly correlated to chronotype at any assessment phase
across treatment, and the data showed a positive association. Also, the correlations became
increasingly less significant as treatment progressed. For the latter group, there was a significant
negative correlation between chronotype and sleep quality, and the strength of the correlation
increased across treatment. This correlation strongly indicated an association between good sleep
quality and an increased tendency for morningness.
Overall, the change in correlation exhibited by the group without a caregiver involved indicates
that the initial dissociation between chronotype and sleep returns to reflect the correlation found
in the general population over the course of treatment; whereas in patients with a caregiver
involved, sleep quality further separates from chronotype. It appears that an additional factor
affects the typical association between sleep quality and chronotype when a spousal caregiver is
closely involved in the treatment. Future studies should consider whether there are changes in the
types of factors contributing to poor sleep quality (e.g., high worry and anxiety at the outset
which decreases and is maybe replaced by steroid drug side effects such as insomnia with high
alertness as treatment progresses), and what factors may produce a difference in the relationship
between chronotype and sleep quality between patients with and without caregivers closely
involved in their treatment and care.
125
Among the 23 caregivers, MEQ and PSQI showed a positive but non-significant association,
similar to the association in the 23 matched patients. In this group, the positive association
suggests poorer sleep quality (increasing PSQI score) is related to increasing MEQ score
(tendency toward morningness). Like the 23 matched patients, these results do not reflect the
relationship observed in the general population. While it is possible that this caregiver group may
have been too small to yield an accurate statistical result, it is possible that other factors are
affecting caregiver sleep quality independent of chronotype, subsequently masking the effects of
one’s circadian clock. Future studies should examine whether a shared factor negatively
influences sleep and disrupts the sleep-chronotype relationship between patients and their
caregivers closely involved in the treatment and care process. Research should also examine
whether poor patient or caregiver sleep negatively influences sleep quality of the alternate group.
While sleep quality scores across treatment among the total patient and caregiver groups were
poor, the numbers did show slight fluctuations between assessment phases. However, among the
total group of patients and the caregivers, there were no within or between (i.e., chronotype
dependent) subject changes in sleep quality across treatment. This suggests that regardless of
chronotype, patients and caregivers alike experienced poor sleep, and sleep scores were not
significantly different across assessment phases. Future research must consider whether a
particular set of factors consistently influences poor sleep across treatment, and or whether
various chronotypes are more prone to the effects of certain disruptive influences.
While patients and caregivers both have poor sleep quality, all patient PSQI scores (including
those with and without a caregiver) are significantly worse than those of caregivers. This trend
exists from treatment outset, even before chemotherapy drugs have been administered, indicating
that this may not necessarily be a pharmacologically induced difference at all times. Among
caregivers, the poor sleep is not likely pharmacologically induced, unless the caregiver is taking
other unknown medications (drug history was not collected in this study). Between matched
patients and caregivers, patients demonstrate significantly worse sleep across treatment, however
there is no significant chronotype difference. One possible explanation for the poor sleep among
patient caregiver duos may be that patients and caregivers are both cognitively depleted at
similar times as a result of having similar chronotypes, and therefore may ruminate together in
the evening at their off times, subsequently increasing one another’s stress levels. It is important
to assess in future how patients and caregivers affect one another’s sleep quality.
126
5.5.2! Patient Cognitive Function and Emotional Regulation
5.5.2.1! Chronotype-UTIME Correlations
Cognitive based tasks show performance fluctuations across the day. Research has indicated that
M types often show performance declines across the day on executive function and
attention/vigilance tasks, while E types tend to show a later performance peak time in the day on
the same tasks (for review, see Schmidt et al., 2015). One key aspect of this study sought to
assess subjective recall of one’s emotional regulation (an output of cognitive control) in the form
of coping, in relation to one’s chronotype, among cancer patients and their caregivers.
Normalized UTIME scores above .5000 indicate peak performance in the first half of the day,
with increasing scores above .5000 indicating a peak closer to one’s wakeup time, while scores
below .5000 reflect peak performance in the second half of the day. Among patients, all UTIME
scores, except UTIME #5, were significantly positively correlated with chronotype, indicating
that higher MEQ scores (i.e., greater tendency for M) are significantly associated with higher
UTIME scores (i.e., peak performance is earlier in the day). These data support similar findings
in previous studies that chronotype and one’s perception of peak performance time is correlated.
In other words, a positive correlation indicates that higher MEQ scores are associated with
earlier perceived optimal performance times (higher UTIME scores). Among patients, M types
appear to show earlier peaks on the 7-question battery than E types, indicating that participants
typically report reaching their peak emotional regulatory abilities at times reflective of their
chronotype.
Mean UTIME #7 (most need time and space to yourself) scores for baseline and midpoint were
both below .5000, and just slightly above .5000 at endpoint. The baseline and midpoint averages
suggest that while MEQ was correlated with UTIME #7, patients still typically required personal
time or space to themselves later in the day compared to their peak times on other activities. The
positive correlation here indicates that M types, while still showing a peak need for time/space to
themselves in the second half of the day, still require this time earlier than their N and E type
counterparts. Requiring time and space to oneself later in the day may be reflective of the
documented afternoon energy decrease, often known as the “post lunch dip” (Monk, 2005).
UTIME #5, which assessed changes in fatigue across the day, did not show a significant
relationship with MEQ. Research shows that fatigue is one of the most commonly reported
127
symptoms of cancer, and is prevalent during both treatment and recovery. Specifically,
prevalence rates of fatigue have been reported among breast cancer patients at 4% prior to
chemotherapy, and spiking to 91% following surgery and chemotherapy (Carr et al., 2002). The
lack of correlation, and the shift from positive and negative correlation between MEQ and
UTIME #5 scores indicates patients’ chronotype were not associated with their recalled sense of
fatigue across the day. It is important to note that this is the UTIME question patients most
struggled to answer, with many people expressing confusion (note: no quantifiable data available
on how many people had problems with this question). It seems that many patients had difficulty
recalling their fatigue levels across the day. It has been reported that fatigue is difficult to
quantify or interpret in research, as there is no consensus on a standard definition, and varying
criteria are used to define its presence or severity (Institute of Medicine, 2008). Given that
fatigue is a commonly reported symptom, and has no standard against which it can be measured,
patients may have difficulty quantifying shifts in their level of fatigue, particularly if an element
of fatigue persists across the day. This may result in a range of different assessments about
personal fatigue, therefore removing the chance of correlation between MEQ and UTIME #5.
While no correlation exists between MEQ and UTIME #5 scores, mean UTIME #5 scores were
always below .5000, indicating that the average peak fatigue time was in the second half of the
day. This shows a consistent pattern that on average patients recalled to some degree that they
were more fatigued later in the day.
5.5.2.2! Multivariate Analysis of UTIME Correlations in the Patient Population
Mixed ANOVAs were conducted to assess within and between subject changes in cognitive and
emotional regulation across the treatment. Each UTIME item was chosen to address a different
aspect of executive function as it might pertain to the cancer patient’s situation.
UTIME #1: Emotional Regulation, “Coping” – changes in one’s ability to cope with the stress
associated with their cancer diagnosis across the day. Between subjects, chronotype was
significantly related to changes in coping ability across the day, evidenced by definitely and
moderately M types recalling reaching peak coping ability earlier than E types. Moderately M
types also reached peak coping ability significantly earlier than N types. While a significant
difference in coping may not have existed between definitely versus moderately M types, or
between definitely M versus N types, these preliminary data support a chronotype related
128
difference in emotional regulatory abilities. This supports the idea that people have a memory for
their optimal times to face their stressor. Within subjects (i.e., regardless of chronotype), across
treatment, patients collectively showed an earlier peak coping time at endpoint over midpoint,
suggesting people reached their optimal coping ability for the stresses associated with having
cancer earlier in the day. While it is possible that several factors contribute to this shift to earlier
peak emotional regulatory abilities, it seems likely that increasing exhaustion levels as treatment
progressed may impact this trend. There are likely to be multiple factors that contribute to this,
considering especially the novelty of a situation that is outside most peoples’ experience. It
seems reasonable to expect that the initial diagnosis and entry into treatment would be sufficient
to override (mask) chronotypic variation, or to disrupt underlying circadian regulation entirely.
UTIME #2: Prosocial Behaviour – changes in one’s ability to help their partner cope with the
knowledge of the patient’s illness. Prosocial behaviours refer to those actions that benefit or
increase the welfare of others, sometimes at a cost to one’s self (Dunfield, 2014; Gesiarz &
Crockett, 2015), and are believed to be an output of cognitive function, specifically perspective
taking, which in this case is the ability to take the viewpoint of someone in distress (Underwood
& Moore, 1982). This question was assessed among all patients who had a caregiver, including
those with a caregiver partaking in the study, and those whose partner was their main caregiver
but was not involved in the study. N types showed UTIME scores significantly lower than
definitely and moderately M types, indicating that while the mean UTIME #2 score was always
above .5000 and suggestive of best performance in the first half of the day, later chronotypes
(lower MEQ score) showed delayed peak performance times (lower UTIME score) compared to
their earlier chronotype counterparts. This supports the idea that cognitive function is highly
positively correlated to one’s chronotype (Schmidt, Collette, Cajochen & Peigneux, 2007).
UTIME #3: Attitudes on Caregiver Helpfulness – attitude toward caregivers regarding how well
they could help a patient cope. This question was assessed among all patients with a caregiver,
including those with a caregiver partaking in the study, and those whose partner was their main
caregiver but was not involved in the study. No between subject (chronotype) difference was
seen, indicating that patients did not feel they were given assistance by their partner at times
necessarily corresponding to their best or worst coping times in the day. This demonstrates that
one’s partner is not necessarily able to provide assistance depending on the patient’s
chronotypically preferred time of day, and therefore may be providing assistance at their own
129
personal best time of day. A within subject difference (regardless of chronotype) showed that on
average, patients believed their partners were best capable to help them somewhat earlier in the
day at endpoint compared to baseline. This may indicate that overall, caregivers themselves may
have had a tendency towards better ability to cope and provide assistance to others earlier in the
day, suggesting a preference towards morningness.
UTIME #4: Alertness – changes in patients’ degree of alertness across the day. Alertness is
necessary to engage in complex cognitive processing, and carry out cognitive tasks (Killgore,
2010). This includes emotional regulation and coping. As one’s propensity for morning
increased, their corresponding peak period in the day for alertness was also earlier. All
differences between definitely and moderately M types, N types and E types were sequentially
ordered, such that definitely M types had the earliest peak, and E types the latest peak. These
results demonstrate that among individuals faced with a chronic stressor, alertness – an output of
cognitive function – fluctuates simultaneously with chronotype, supporting the overall link
between one’s circadian rhythm and changes in their cognitive abilities. While this question may
have had an increased chance of Type 1 error (a “false positive”), the results relating to UTIME
#4 seem plausible given that they are in line with the chronotype ordering trends seen in other
UTIME questions.
UTIME #5: Fatigue – daily changes in fatigue according to chronotype. E types showed an
earlier peak than all three other groups. Yet the difference in scores between M and N types was
not significant; only E types showed a significantly earlier peak fatigue time compared to N
types. This may be due to either skewed data resulting from a small E type group, or it is
possible that E types are more fatigued earlier before their energy levels increase for the day. No
within subject differences were found for changes in fatigue across treatment. Furthermore, lack
of a clear definition or way to measure fatigue will likely contribute to poor memory for one’s
changes in fatigue across the day. Lack of a clear way to quantify fatigue for one’s own self
likely makes it more difficult to store clear memories for changes in one’s fatigue. In future, a
clearer definition of fatigue along with a more defined benchmark against which patients can
assess fatigue is required to help them quantify their subjective ratings of fatigue.
UTIME #6: Compassion – patients’ level of compassion for their partner as a caregiver.
Compassion is a prosocial behaviour that has been conceptualized in several ways, and can be
130
summarized as the vicarious experience resulting from witnessing the distress of another’s
suffering, that spurs the subsequent desire to help (for review, see Goetz, Keltner & Simon-
Thomas, 2010). Compassion is associated with perspective taking (Gleichgerrcht & Decety,
2013), and therefore requires cognitive reasoning ability. CS and CF were found to be associated
to chronotype among various oncology staff providing patient care in the hospital setting (see
Chapters 3 and 4). This question was assessed among all patients who had a spousal caregiver,
whether he was or was not involved in the study. Similar to UTIME #2 which also assessed
prosocial behaviour, definitely and moderately M types had earlier peak compassion
performance times in the day compared to N types. While E types did not display a significantly
later peak time compared to M or N types, they did exhibit a delayed peak. These data indicate
that when faced with a chronic stressor themselves, individuals still recall showing compassion
to those who work to help them, and report that their compassion is greatest at times
corresponding to their chronotype. No within subject differences were found across treatment,
suggesting that patients remained fairly consistent in the times at which they recalled feeling
compassion towards their caregivers. A significant ordered difference between moderately and
definitely M types, N types, and E types for compassion abilities should be examined in future
studies among larger populations.
UTIME #7: “Private Time” – patients’ need for time and space to themselves. The results
showed a clear statistically significant between subject difference, where earlier chronotypes
recalled an earlier peak time for requiring personal time and space compared to later
chronotypes. Moderately M types reported a significant earlier peak compared to N types across
treatment. Mean normalized UTIME #7 scores were indicative of needing time and space in the
second half of the day at baseline and midpoint, and just slightly before the second half of the
day by endpoint. This suggests that time and space to one’s self is needed later in the day
sometime in the afternoon, possibly when an individual is less cognitively alert. This may
indicate that one’s resources are slightly depleted at this time and therefore an individual requires
some personal space and reduced stimulation. Nonetheless, those with a later chronotype showed
a delay in the time of day they reported personal time and space was required, suggesting that in
line with their later peak performance compared with earlier chronotypes, these individuals also
reach their energy depletion lows later in the day than their counterparts. The somewhat earlier
peak at endpoint may suggest patients reach a point where personal time is required earlier,
131
possibly as a result of frustration with a long treatment process, and or general energy depletion
and cachexia (general wasting syndrome including appetite, energy, muscle and weight loss) that
is often reported by cancer patients (e.g., Aoyagi, Terracina, Raza, Matsubara, Takabe, 2015).
The within subject trend that was visible in some questions looking at changes in cognitive
function peaking earlier in the day by the end of treatment indicates people’s optimal
performance times were shifting. Future research is required to understand what may be causing
this phenomenon. While chronotype was likely a relatively stable and unchanging trait, it is
possible that people’s degree of fatigue was influencing their need to deal with various tasks and
demands earlier in the day before they were too depleted sometime later in the day. Given that
fatigue has been reported in up to 91% of patients undergoing chemotherapy treatment (Carr et
al., 2002), it is possible that patients noticed this change in their ability and adjusted their
performance times accordingly. Future studies with larger patient groups may be able to more
clearly assess this difference.
5.5.3! Caregiver Cognitive Function and Emotional Regulation
Among the 23 caregivers, there were only two statistically significant correlations between MEQ
and UTIME score between baseline, midpoint and endpoint, and none that were consistently
significant. However, as seen with the total group of patients, all associations between MEQ and
UTIME were consistently positive (except for UTIME #5, which again measured fatigue). This
positive association may suggest that a greater tendency for M would result in an earlier peak
performance time across the day, and vice versa for E types, however, due to lack of
significance, no conclusions can be drawn. The change in association direction between MEQ
and UTIME #5 from positive to negative may again be due to the lack of definition and
quantifiable ways to assess fatigue. As seen in the total patient population, caregivers also show
mean UTIME #5 scores below .5000, indicating that they recall their highest fatigue time is in
the second half of the day. It is interesting to note the highly significant positive correlation
between MEQ and UTIME #7 (“Private Time”). The mean normalized UTIME #7 score at
endpoint (.5008) suggests an average of most needing space to one’s self in the midday. This
significant correlation and ANOVA results indicate that M type caregivers recall needing space
earlier than N and E type caregivers. One possible reason this chronotype difference was not
found at baseline or midpoint, aside from small sample size, is that caregivers may have been
132
hesitant earlier in treatment to take time for themselves when they felt it was necessary if they
were already performing other duties related to helping their partner. Given that patients noted
differences in their caregivers’ abilities to provide care, it is possible that caregivers themselves
may have attempted to provide consistent care across the day but were unable to notice changes
in their performance. Alternately, as suggested in Bellicoso (2010), it might be counter to the
heuristic they maintain about their caregiving abilities being consistent across the day to report
fluctuations in their performance. Future studies with a larger caregiver subject group should
further assess these questions to examine whether caregivers are actually not reaching peak and
trough performance times as dictated by their chronotype in order to continue performing tasks
related to helping their spouse, or whether a larger sample size will allow these chronotype based
differences to be observed.
5.5.4! Matched Patient and Caregiver Cognitive Function and Emotional Regulation
Among the 23 pairs of matched patients and caregivers, the two groups were fairly similarly
matched on MEQ and UTIME scores.
Three-way mixed ANOVAs were used to assess differences in UTIME score across treatment
between patients and caregivers split on two chronotype categories (definitely and moderately M
type versus N and E type chronotypes).
UTIME #1: Emotional Regulation, “Coping” – change in ability to cope with the stress
associated with one’s own or their partner’s cancer diagnosis across the day. No significant
within or between subject differences were found, suggesting that among matched patients and
caregivers partaking in the study, participants recalled coping fairly equally across treatment
regardless of their role or chronotype. It is also possible that no significant differences were
found due to the small 23 pair sample size. However, among the total group of patients, a
difference in coping based on chronotype was found to exist, therefore it is possible that the
group size was too small in this analysis to detect a significant difference. Larger studies with
more normal chronotype distributions should reexamine matched patients and caregivers to
assess whether there is in fact no difference in one’s memory for changes in their coping ability,
or whether clear distinctions exist between patient-caregiver and chronotype groups.
133
UTIME #2: Prosocial Behaviour – change in ability to help one’s partner cope with the stress of
their cancer diagnosis/role as a caregiver. As in the previous two groups, this question assessed
changes in prosocial behaviour as an output of shifts in cognitive function across the day. While
caregivers did not show a significant chronotype based difference in prosocial ability across the
day, M type patients showed a significantly earlier peak recalled performance time for prosocial
behaviour compared to N types (there were no E types in the 23 subject patient subset). This
suggests patients recalled aiding their partners at their own personal chronotypically optimal
times, while caregivers did not recall significant differences in their provision of care across the
day. These results among patients reflect those seen in the larger group. It is possible that due to
the small sample size, differences among caregivers’ recalled coping abilities are not uncovered,
however, given that a small sample size for patients yielded a chronotype based difference, it is
possible that caregivers are not forming a memory for differences in their prosocial abilities. As
caregivers, they may feel they are in a position to provide consistent assistance to their partner,
and recalling performance differences may be counter to the heuristic they hold of themselves.
UTIME #3: Attitudes on Helpfulness – assessed differences in perceived helpfulness among
patients or caregivers for when their partner could best help them cope with the stress of being a
patient or caregiver. A significant within subject effect between time and role indicated that at
endpoint, caregivers felt their partners best helped them cope earlier in the day compared to
when patients felt their caregivers assisted them. These data suggest that caregivers recalled
patients reaching an earlier peak performance ability for prosocial behaviours, subsequently
indicating an earlier peak in cognitive abilities. This supports the slightly higher mean
chronotype score found among the 23 patients compared to their 23 caregivers which would
likely lead to differences in peak cognitive performance times. Future studies among larger
groups should reexamine whether this difference exists across treatment.
UTIME #4: Alertness – changes in patients’ and caregivers’ degree of alertness across the day as
a benchmark for changes in their cognitive function. While a significant effect of treatment time
point was seen at midpoint among patients, studies with larger population sizes will be necessary
to further examine this association. Moderately and definitely M type patients reported earlier
alertness peaks compared to N type patients. Larger sample sizes can assess whether this
association exists at isolated time points across treatment, or whether this is normal across
treatment. Nonetheless, this is reflective of the same data seen in the larger total group of patients
134
where chronotype was linked with one’s peak alertness time across the day. Interesting to note is
that among N and E type chronotypes, caregivers recalled a significantly earlier peak in alertness
compared to patients. Whereas caregivers reported peaking on average late in the first half of the
day, patients recalled peaking sometime in the earlier portion of the second half of the day. This
is particularly interesting given that caregivers had a slightly later mean chronotype score
compared to their matched patients. This suggests female patients with a caregiver feel they take
slightly longer to reach peak alertness than their male caregivers who recall reaching their peak
sooner. This needs to be further examined in larger population samples, but these results may
suggest women feel they take longer to reach heightened cognitive function than males. Overall,
the findings relating to MEQ versus UTIME #4 suggest patients and caregivers show differences
in their recalled levels of alertness across treatment, and that earlier chronotypes recall peaking
earlier in the day compared to later chronotypes, continuing to suggest an association between
one’s chronotype and fluctuations in their recalled cognitive ability.
UTIME #5: Fatigue – daily changes in fatigue according to chronotype. Among matched patients
and caregivers, it appeared that a difference existed in one’s memory for fluctuations in their
level of fatigue based on their role or chronotype. However, upon further consideration, there
were no significant differences in this area. The lack of significant difference between recalled
differences in fluctuations in one’s level of fatigue may result from lack of a clear definition of
fatigue, as previously noted. The small sample size may also contribute to not uncovering
significant findings. Finally, among this matched group specifically, it is possible that patients
and caregivers did not focus on their fluctuating fatigue levels across the day, as many may have
felt familial obligations to other members of their household and to one another, and as such,
may not have stopped to reflect on and form a clear memory of their own fatigue.
UTIME #6: Compassion – patients’ and caregivers’ levels of compassion for their partner and
the stresses faced she/he faced as a patient/caregiver. Caregivers did not report recalling a
chronotype dependent difference in their compassion levels towards their partner the patient. As
discussed previously, it is possible that caregivers are accurately recalling being consistent in
their degree of compassion across the day. Alternately, caregivers may not accurately reflect on
and or recall fluctuations in their levels of compassion, as this would have been counter to their
duty to care for and be compassionate towards their partner in their time of sickness. Yet among
patients with caregivers in the study, similar to the pattern seen among the total group of patients,
135
this subset of patients continued to be compassionate to their caregiver in a manner reflective of
their own chronotype; specifically, M types showed an earlier peak compassion time compared
to N type. While caregivers may believe it is their duty to be compassionate towards the patient
at all times, patients may feel that as the sick individual, it is acceptable to limit the degree of
prosocial behaviour they must express towards others, and therefore may be more open to
reflecting on and reporting performance differences in their behaviour. This chronotype based
difference among patients in recalled fluctuations in compassion supports the other observations
relating to chronotype based changes across the day for cognitive domains. These findings on
performance memory differences support the idea that one’s recall for their cognitive
performance ability is closely linked to chronotype, and that patients need to be open to
accepting differences in their performance in order to form a memory.
UTIME #7: “Private Time” – patients’ and caregivers’ need for time and space to themselves.
While no specific difference were found between M versus N type patients for recalled need for
personal time and space, the results found among just caregivers were repeated here. At
endpoint, M types caregivers again recalled an earlier peak time compared to their N and E type
counterparts. While this is a repeated finding, unlike in the total group of patients, this subset of
patients did not report a chronotype related peak timing difference in their memory for requiring
personal time and space. Unlike the other 63 patients, these 23 patients may have recalled a less
differentiated pattern for changes in the need for personal space across the day. It is possible that
these women whose partners were closely involved in the study were accustomed to having
people around and present more often, so they may have learned not to follow a chronotype
dependent schedule for taking personal time and space.
5.5.5! Limitations
At the outset of this study, it was believed that many caregivers would be available for
participation in the study. However, for several reasons (including but not limited to time
constraints, low interest, not being able to attend hospital visits with patients), it was not possible
to recruit the large number of caregivers we had initially hoped for. This reduced the population
size of caregivers, and the number of matched patient-caregiver dyads. Future studies should
include a larger number of patient caregiver dyads in order to better understand how close
simultaneous involvement by the patient and caregiver across chemotherapy affects each
136
individual. Only 14 patients involved in the study were single. Future research may wish to
examine a larger group of single women in comparison to women with a closely involved
caregiver to understand differences in their recall for performing optimally at one time or
another.
Future studies may wish to reassess the use of actigraphy in such a study. While an attempt to
use actigraphy was made in this study, no useable data were obtained. Actigraphy data would
provide an objective measure of patient activity across the day, which may help to better
understand sleep, and perform a more in depth analysis comparing memory recall among those
patients who are more or less active.
5.6! Conclusion
While a significant correlation between sleep quality and chronotype exists among patients who
either have no caregiver, or whose caregiver was not involved in this study, this relationship does
not exist among patients whose caregiver was involved. Nor does it exist among the participating
caregivers. Given the opposite (positive) direction of the correlation between MEQ and PSQI
scores for patients with a caregiver involved in the study as compared to the (negative) direction
of the correlation for patients without a caregiver, and the fact that involved caregivers show the
same positive direction as their spouses, it seems possible that this group of individuals is
influencing sleep quality among one another. Chronotype is a fairly stable trait, while sleep
quality can change, therefore it seems reasonable to suggest sleep quality is being affected
among these individuals to shift the correlation direction. While a comparison statistic was not
calculated, sleep quality scores were consistently worse among patients whose caregiver was
involved in the study, compared to those patients without a caregiver involved. Future studies
should examine whether patients and their caregivers who are closely involved in and go through
the treatment process with little separation from one another are actually causing further
disruption to one another’s sleep. If so, it would be beneficial to suggest taking some time and
space individually to enhance one another’s sleep and possible ability to cope.
Among the total group of patients, while significant clearly ordered differences between
definitely M, moderately M, N and E type chronotypes were not found for all UTIME questions,
137
the general trend showing that earlier chronotypes were consistently recalling reaching peak
performance times earlier than their later counterparts supports the idea that chronotype is
strongly tied to one’s cognitive abilities and their memories for their cognitively based tasks.
Caregivers did not show significant differences in their performance for the various cognitive
and emotional regulatory tasks that were assessed across the day. There are several possible
explanations for this. First, it is possible that the sample size was too small to reveal a significant
difference. Second, as patients and participating caregivers have a different gender
representation, responses based on subjective recall, may be different. For example, as
mentioned, it is possible that caregivers felt they sought to give their best effort to help their
spouse and to remain consistent in their behaviour across the day, and therefore may have
recalled being consistent if they did not stop to actually reflect on changes. Finally, it is possible
that many caregivers were not acting at their own chronotypically optimal time, and instead
aimed to give their best performance for various activities when they believed it was most
necessary by their caregiver.
Overall, the study demonstrated that among cancer patients, and to an extent among caregivers,
chronotype is related to one’s subjective recall of sleep quality and their memory for changes in
cognitive and emotional regulatory performance across the day. Similar patterns have been
reported previously (Bellicoso, 2010), however further examination is needed to understand why
patients with closely involved caregivers and the caregivers themselves show patterns opposite to
the general population. The subject of fatigue which remains poorly defined and without a
standard measure, while still somewhat associated to chronotype, does not have as clear a
relation as found in other tasks. For the emotional regulatory tasks assessed in this study –
personal coping, prosocial behaviour and compassion – patients may have had clearer standards
against which to measure changes in their behaviour compared to the emotional regulatory tasks
used in previous studies that showed little correlation with chronotype (see Bellicoso, 2010).
It is important that cancer patient chronotypes be assessed in order to help them schedule their
day in order to allow them to face trying situations at times when they will be best equipped to
deal with the situation. Caregivers should also be assessed for their chronotype, as a better
understanding of their optimal times in the day may allow them to structure their day accordingly
and take the necessary measures to also help themselves during this difficult time.
138
Chapter 6!
! Coping Behaviour in Chronic Disease
6.1! Abstract
Little is known about the influence of cancer patient and spousal caregiver chronotype and sleep
quality on their self-ratings of coping, or on their specific behaviours used to deal with cancer
related stressors. Given the influence both factors have on cognitive ability, and the lack of
importance society places on sleep, it is important to understand how these factors relate to
changes in coping across treatment. Patient self-rated coping responses, and patient and caregiver
coping behaviours were assessed for potential relationships with chronotype and sleep quality
across chemotherapy treatment. Personality and the difference between patients and caregivers
was also considered when assessing predictive influences on the use of coping behaviours.
Chronotype was related across treatment to in the moment self-rated coping scores, and at
baseline to the use of coping behaviours such as self-distraction, substance use, and into
treatment for the use of religion. Sleep quality was not linked to self-ratings of coping, but was
positively predictive of engagement style coping behaviours by endpoint. Patients and caregivers
typically reported decreased use of coping behaviours as treatment progressed, except for
acceptance which increased from baseline to endpoint. Openness and industriousness had the
most consistent predictive value for various coping behaviours across treatment. When rating
overall coping ability and use of behaviours early on, people rely on chronotype. Chronotype is
not predictive of coping behaviour use as treatment progresses. As in the typical population,
people do not always factor in the influence of sleep quality for changes in the behaviour across
the day.
139
6.2! Introduction
In 2016, breast cancer was predicted to be the most prevalent cancer diagnosed in females in
Canada, comprising 26% of newly diagnosed cancers in women (Canadian Cancer Statistics,
2017). In addition to the necessary medical treatment, many patients depend on the care and
support of family and or friends who act as non-vocational caregivers, a group who similar to
patients, also experience their own stresses related to providing cancer patient care. Historically,
much of the research among breast cancer patients has focused on the role of clinical variables
such as prevention, diagnosis and treatment. However, in recent years, the scope of cancer
related research has broadened to include understanding how patients and caregivers cope with
the stresses associated with having cancer or caring for a cancer patient. Among patients and
caregivers, individuals in both groups show within and between group differences in their coping
habits (Carter & Acton, 2006; Sharma, Chakrabarti & Grover, 2016).
Coping is a complex process, that refers to those “constantly changing cognitive and behavioral
efforts to manage specific external and internal demands that are appraised as taxing or
exceeding the resources of the person” (Lazarus & Folkman, 1984, p. 141). In addition to
appraisal of the situation, one’s assessment of the resources they have available to deal with the
situation will also influence their coping ability. Coping can be looked at in different ways,
including, but not limited to one’s own overall rating of how they feel they are coping, and by
looking at the specific behaviours one uses to deal with a stressor. Separate from their
assessment of the situation and the resources available to deal with the stressor, the number of
variables that might affect one’s ability to cope, are endless. Understanding how even a select
few variables relating specifically to an individual affect coping, including one’s chronotype,
sleep quality, and personality, can provide a better picture of how coping changes across the day,
and give indications for the coping strategies certain people might employ. This will allow for
recognizing signs for those individuals who may have greater difficulty coping at certain times,
or understanding who may be at risk of engaging in harmful coping behaviours.
People facing a major stressor are often asked “how are you doing?” or “how are you coping?”.
Yet, the majority of coping measures that exist do not ask respondents that question, but rather
assess one’s use of coping behaviours (e.g., Brief COPE), strategies (e.g., Cancer Coping
Questionnaire), or perceived ability to cope effectively with life challenges (e.g., Coping Self
140
Efficacy Scale), to name a few (Carver, 1997; Chesney, Neilands, Chambers, Taylor & Folkman,
2006; Moorey, Frampton & Greer, 2003). None of these questionnaires give an indication of
how one’s self-perceived ability to cope changes across the day, or when in the day a person
feels he or she copes best or worst. Furthermore, these questionnaires are retrospective; none of
the measures ask respondents to report how they believe they are coping at a specific moment in
time. Long term memories are influenced by both one’s memory of the event and the emotion
experienced during memory retrieval, therefore it is possible that one’s memory of their coping
may differ from how they actually felt at a specific time or on a specific day. While these
questionnaires provide valuable information, an understanding of how one is coping in the
moment at specific times of the day would give a clearer understanding of one’s ‘ups and downs’
during their waking hours.
Functioning of cognitive and emotional processes vary over the course of the 24 h day.
Appropriately timed wake and sleep cycles that reflect one’s internal biological clock facilitate
maximal cognitive and emotional performance, while wake sleep schedules that do not reflect
one’s biological clock can reduce an individual’s cognitive and or emotional regulatory abilities
(Wright, Lowry, & LeBourgeois, 2012). Chronotype has been linked with changes in
emotionality and mood, such that morning (M) type individuals typically show the quickest rise
in positive affect in the morning between 9 a.m. and noon followed by a dramatic decrease after
9 p.m., while neither (N) and evening (E) types do not demonstrate the same rapid rise in
positive affect in the morning as seen among M types (Clark, Watson & Leeka, 1989). As such,
it is likely that depending on a patient or caregiver’s chronotype, there will be earlier and later
peaks in optimal coping times across they day. A clearer understanding of how one’s chronotype
relates to coping ability and the use of particular coping habits across treatment can help support
workers to understand when in the day patients and caregivers might be more or less able to deal
with information related to the disease, or when in the day additional help might be needed.
Sleep loss negatively alters both cognitive processes such as event memory, judgment and
decision making, and emotional functioning and processing abilities, often resulting in poor
mood and elevating negative response styles in reaction to potentially trying situations (Lim &
Dinges, 2008; 2010; Rosales-Lagarde et al., 2012). Cancer patients and caregivers commonly
report poor sleep quality, suggesting they are more likely to have difficulties with cognitive and
emotional functioning and processing abilities. Avoidance coping has been documented among
141
cancer patients reporting reduced sleep; among caregivers, similar less functional coping
strategies such as venting, self-distraction, and self-blame have been associated with increased
sleep disruptions (Aslan, Sanisoglu, Akyol & Yetkin, 2009; Carter & Acton, 2006; Northouse,
Williams, Given & McCorkle, 2012; Hoyt, Thomas, Epstein & Dirksen, 2009). Conversely, the
use of positive, proactive coping strategies among caregivers has been associated with reduced
numbers of sleep disturbances (Zhang, Yao, Yang, & Zhou, 2014). However, none of these
studies have tracked patients and caregivers across treatment to see whether there is a change in
the relationship between coping and sleep quality, or whether sleep is related to coping at
specific time points. A better understanding of when sleep might be more or less related to one’s
coping can provide suggestions to patients and caregivers on crucial times in treatment when
their sleep is most important in helping them face stressors. Among those reporting poor sleep, it
can be estimated when there may be more or less problems with coping so that additional outside
assistance can be provided.
Personality refers to an individual’s characteristics that influence or determine patterns of
behaviour, feeling, and thought, and has been linked with circadian rhythms and sleep quality
(e.g., Cavallera, Gatto & Boari, 2014; Duggan, Friedman, McDevitt & Mednick, 2014;
Hintsanen et al., 2014; Hsu, Gau, Shang, Chiu & Lee, 2012). Personality has been linked with
coping, such that certain coping styles occur in increasing or decreasing frequency with certain
personality traits. A study among intensive care unit (ICU) nurses found that those reporting
greater conscientiousness, agreeableness, and or openness showed associations with coping
strategies indicative of problem focused coping in order to approach and resolve the workplace
related problem causing stress. Conversely, high neuroticism was strongly related to venting as a
coping strategy to verbally express negative feelings, reflective of an emotion-based coping style
aimed at reducing one’s negative feelings about the threat as opposed to altering the source of
stress itself (Burgess, Irvine & Wallymahmed, 2010). Similarly, when faced with the stress of
their own or their partner’s ill health, cancer patients and their caregivers may exhibit personality
related coping behaviours. Furthermore, as with chronotype and sleep quality, it is important to
understand whether personality has a consistent or varied influence on one’s use of particular
coping habits across treatment. A better understanding of personality’s influence on coping
behaviours will suggest which patients and caregivers may be more likely to use certain coping
142
behaviours, and allow support workers to identify those who may be at risk or require greater
assistance.
As chronotype, sleep quality, and personality each have the potential to influence coping at a
particular instance, it is important to look more closely at how these factors might interact across
time. We hypothesized that one’s chronotype, yet not necessarily their sleep quality, will
influence in the moment coping. Regarding specific coping behaviours, it is believed personality
will have a varied influence across treatment, depending on the various stressors patients are
faced with at different time points across treatment. Chronotype and sleep quality will also be
related to coping behaviours, however they may not exert a simultaneous influence. As predicted
for in the moment coping ratings, given that society often disregards the influence of sleep on
their behaviour therefore operating under a sleep debt, many people may not consider changes in
sleep quality and resulting influence on their abilities when they rate their behaviour.
Conversely, even when tired, people maintain some understanding of their own preference for
morning versus evening.
A greater understanding of how these factors will influence one’s self-perceived in the moment
rating of their own coping and one’s general use of particular coping behaviours over time is
important. For example, it can help support workers provide personalized assistance to patients
and caregivers who might be at risk of using harmful or less positive coping strategies. Not only
can this information be beneficial to cancer patients and their caregivers, it can be tested and
applied in other populations faced with chronic illness.
6.3! Materials and Methods
6.3.1! Participants and Procedures
Prospective patients receiving cancer treatment at Sunnybrook Health Sciences Centre, Odette
Cancer Centre (Toronto, Ontario, Canada) and when applicable, their caregivers, were identified
by their medical oncologist or nurse, and briefed on the general idea of the study. If patients and
caregivers reported interest in participating, they were approached by the study coordinator, who
explained the details and time commitment of the study. Patients receiving adjuvant and
neoadjuvant treatment were included in the study. If patients reported living with a life partner
143
who would be their primary caregiver, the partner was also given the opportunity to partake in
the same study. Inclusion criteria for patients included being 19 or older, receiving chemotherapy
for stage 1, 2, or 3 breast cancer, being able to read, write and speak English at least at a basic
level, and being female. While men can develop breast cancer, only female patient participants
were recruited. Individuals (patients or caregivers) who were going to continue following a
shiftwork work schedule during the treatment period were excluded. Caregiver inclusion criteria
required subjects to be 19 or older, have a spouse undergoing chemotherapy for breast cancer
treatment and enrolled in the study, and able to read, write and speak English at least at a basic
level. One female caregiver partook in the study, however, neither her nor her partner returned
correctly completely, usable data.
Potential participants were informed that their decision to participate or not would have no
bearing on their treatment. They were informed that their decision to take part was voluntary,
and while no formal remuneration would be provided, their participation in the study might
provide the opportunity to reflect on personal coping ability, and factors that influenced their
coping.
A total of 109 participants (86 patients, 23 caregivers) provided usable data that were included in
this study. Among those participants who withdrew, the general reason for withdrawal was that
people felt the study would be too time consuming. Of the 86 patients who took part, 38 patients
provided correctly completed data for their daily coping log. Only 3 caregivers correctly
completed their daily coping log, therefore their data was not assessed due to the small sample
size.
6.3.2! Measures
This section of the study builds on the work from the previous chapter, making use of the Horne-
Östberg Morningness Eveningness Questionnaire (MEQ), Pittsburgh Sleep Quality Index
(PSQI), and the University of Toronto Inventory of Morningness Eveningness (UTIME), while
also incorporating a daily coping log, the Brief COPE as an assessment of coping responses, and
Big Five Aspect Scales (BFAS) as a measure of personality. Follow-up surveys at the treatment
midpoint and endpoint sessions contained the PSQI and UTIME. Descriptions of these
questionnaires can be found in the General Methods section of this dissertation, section 2.2
Instruments. Baseline packages were completed prior to or on the first day of chemotherapy
144
treatment. Midpoint packages were completed at the halfway session of a patient’s treatment (the
treatment session number varied depending on the particular type of chemotherapy being
received). Endpoint packages were generally completed on the day of a patient’s final
chemotherapy session, however, given that some patients’ final chemotherapy sessions were
cancelled upon assessing the patient’s health that day, some final packages were completed at
what would have been the final chemotherapy session. Daily coping logs included 3 Likert scales
per day, one for morning, afternoon, and evening that were completed in the given time period or
left blank if a patient missed responding at the appropriate time. Daily coping logs were kept
across chemotherapy. Of the 86 patients who completed survey data, only 38 correctly
completed coping logs which are used in these data.
6.3.3! Statistics
Statistical analyses were performed using SPSS (Statistical Package for the Social Sciences)
version 23.0 for Mac. Pearson correlations were used to assess the strength and direction of the
association between daily coping log data and one’s MEQ and PSQI scores. Friedman Tests and
Wilcoxon Signed-Rank tests were used to assess whether there were any statistically significant
differences in coping log scores and coping behaviours at baseline, midpoint, and endpoint.
Paired sample t-tests were used to assess differences in pre- and post- chemotherapy treatment
coping scores. Multiple regression analyses were used to determine how well chronotype, sleep
quality, participant role and personality predicted the use of each coping behaviour at baseline,
midpoint, and endpoint.
6.4! Results
Descriptive data in Table 6.1 are mean ± standard deviation. In general, patient’s average daily
coping scores centered around three, which represented a rating of average on a scale of one to
five. From morning to evening as the day progressed, raw coping log scores tended to taper
down slightly, except for on the five days post chemotherapy at baseline, where the evening
coping score (3.19 1 ± .60) was higher than the afternoon coping score (3.14 ± .59), but still
below the morning coping rating (3.20 ± .72). Peak coping times as demonstrated by UTIME
scores occurred earliest in the day on chemotherapy treatment days (baseline, midpoint,
145
endpoint). When comparing an average of the five days prior to chemotherapy, and the five days
after, peak coping time occurred earlier in the day on the five days after treatment. In the five
days prior to treatment, patients showed a slightly later peak coping performance time in the day.
146
Table 6.1 Descriptive data for patient coping logs Mean (SD) Raw Patient Coping Log Scores
7:00 am to 11:00 am Baseline (n = 28) 3.32 (1.09) 1:00 pm to 5:00 pm Baseline (n = 33) 3.08 (.95)
7:00 pm to 11:00 pm Baseline (n = 35) 3.01 (1.13)
7:00 am to 11:00 am Average of 5 Days Post Baseline (n = 38) 3.20 (.72) 1:00 pm to 5:00 pm Average of 5 Days Post Baseline (n = 38) 3.14 (.59)
7:00 pm to 11:00 pm Average of 5 Days Post Baseline (n = 38) 3.19 (.60)
7:00 am to 11:00 am Average of 5 Days Pre Midpoint (n = 37) 3.90 (.82) 1:00 pm to 5:00 pm Average of 5 Days Pre Midpoint (n = 37) 3.86 (.77)
7:00 pm to 11:00 pm Average of 5 Days Pre Midpoint (n = 37) 3.75 (.88)
7:00 am to 11:00 am Midpoint (n = 33) 3.47 (1.20) 1:00 pm to 5:00 pm Midpoint (n = 31) 3.40 (.80)
7:00 pm to 11:00 pm Midpoint (n = 33) 3.10 (.94)
7:00 am to 11:00 am Average of 5 Days Post Midpoint (n = 37) 3.27 (.79) 1:00 pm to 5:00 pm Average of 5 Days Post Midpoint (n = 37) 3.23 (.84)
7:00 pm to 11:00 pm Average of 5 Days Post Midpoint (n = 37) 3.14 (.85)
7:00 am to 11:00 am Average of 5 Days Pre Endpoint (n = 38) 3.93 (.70) 1:00 pm to 5:00 pm Average of 5 Days Pre Endpoint (n = 38) 3.81 (.81)
7:00 pm to 11:00 pm Average of 5 Days Pre Endpoint (n = 38) 3.75 (.82)
7:00 am to 11:00 am Endpoint (n = 33) 3.92 (.84) 1:00 pm to 5:00 pm Endpoint (n = 32) 3.85 (.91)
7:00 pm to 11:00 pm Endpoint (n = 33) 3.61 (.87)
7:00 am to 11:00 am Average of 5 Days Post Endpoint (n = 38) 3.23 (.75) 1:00 pm to 5:00 pm Average of 5 Days Post Endpoint (n = 38) 3.09 (.69)
7:00 pm to 11:00 pm Average of 5 Days Post Endpoint (n = 38) 2.95 (.76)
Peak Coping Time – Patient Coping Log UTIME Scores Baseline (n = 26) .55 (.08)
Average of 5 Days Post Baseline (n = 36) .51 (.06) Average of 5 Days Pre Midpoint (n = 36) .49 (.07)
Midpoint (n = 30) .53 (.06) Average of 5 Days Post Midpoint (n = 33) .51 (.06) Average of 5 Days Pre Endpoint (n = 37) .49 (.07)
Endpoint (n = 34) .54 (.07) Average of 5 Days Post Endpoint (n = 36) .51 (.07)
147
Pairwise comparisons (Friedman tests) were performed to assess changes in raw coping scores
across treatment from baseline to endpoint (Table 6.2). Friedman tests were used instead of one-
way repeated measures ANOVAs due to the non-normal distribution of raw coping scores. Post
hoc analyses with Wilcoxon signed-rank tests were conducted with a Bonferroni correction
applied resulting in a significance level set at p < .017, rounded: p < .02 (i.e., 0.05/3, where three
represents number of comparisons; this allows the 7:00pm to 11:00pm midpoint versus endpoint
comparison to be significant, p = .019) when Friedman test results were significant (Table 6.2).
148
Table 6.2 Median and Friedman test data for changes in raw patient coping log scores across treatment, with Wilcoxon signed-rank tests with Bonferroni correction applied for differences in raw patient coping log scores across treatment
Median Friedman test Wilcoxon signed-rank tests Baseline vs.
Midpoint Baseline vs.
Endpoint Midpoint vs.
Endpoint Baseline Midpoint Endpoint df χ2 p z p z p z p Coping Log Times
7:00 am to 11:00 am 3.00 3.00 4.00 2 6.633 .036* -1.208 .227 -2.624 .009† -2.198 .028 1:00 pm to 5:00 pm 3.00 3.00 4.00 2 12.029 .002* -1.216 .224 -3.216 .001† -2.300 .021
7:00 pm to 11:00 pm 3.00 3.00 4.00 2 10.500 .005* -.436 .663 -2.230 .026 -2.347 .019† * p < .05 † p < .02
149
Pairwise comparisons (paired-samples t-test) were performed to assess whether there was a
statistically significant change in coping log ratings before and after midpoint and endpoint
chemotherapy treatments, in the morning, afternoon, and evening (Table 6.3). Of the six
comparisons, the only two outliers in difference score data were found in a comparison of coping
ratings five days before and after evening coping at chemotherapy midpoint, as assessed by
inspection of a boxplot value greater than 1.5 box-lengths from the edge of the box. The
remaining five pre- and post-chemotherapy treatment coping ratings had no outliers. The outliers
in the evening coping rating data were included after running both paired-sample t-tests with and
without the outliers and yielding similarly statistically significant results. The differences scores
were normally distributed, as assessed by Shapiro-Wilks test for all six comparisons (p > .05).
Across treatment, patients coped better (higher score) in the five days prior to treatment than in
the five days post treatment.
150
Table 6.3 Mean ± standard deviation and paired samples t-test data for pre- and post-chemotherapy coping rating comparisons Pre Chemo
Coping Post Chemo
Coping
Mean (SD) Mean (SD) df t p Midpoint
Morning (7:00am – 11:00am) 3.909 (.826) 3.274 (.803) 35 4.785 .000 Afternoon (1:00pm – 5:00pm) 3.865 (.783) 3.238 (.846) 35 5.352 .000 Evening (7:00pm – 11:00pm) 3.754 (.895) 3.141 (.865) 35 4.985 .000
Endpoint Morning (7:00am – 11:00am) 3.931 (.703) 3.233 (.754) 37 5.477 .000 Afternoon (1:00pm – 5:00pm) 3.809 (.813) 3.090 (.695) 37 5.066 .000 Evening (7:00pm – 11:00pm) 3.753 (.823) 2.951 (.760) 37 5.751 .000
151
Pearson correlations were conducted between patient’s coping log UTIME scores and their MEQ
score, as well as with their PSQI scores across treatment (Table 6.4). Pearson correlations were
used as the majority of data points involved were fairly normally distributed. Correlations
between coping log UTIME scores and MEQ scores were significant on the five days post
baseline and endpoint, and the five days prior to chemotherapy at midpoint and endpoint. While
not statistically significant, correlations between MEQ and coping log UTIME scores were
approaching significant at baseline and endpoint. However, except for one significant correlation
between PSQI and average coping score five days post midpoint treatment, no other associations
are significant. MEQ and PSQI score among this population group also do not show any
significant relationship.
152
Table 6.4 Pearson correlations between MEQ and PSQI and patients’ coping log UTIME scores
MEQ PSQI Baseline Midpoint Endpoint r p r p r p r p Baseline (n = 26) .385 .052† .082 .690 Average of 5 Days Post Baseline (n = 36) .361 .031* .014 .934 Average of 5 Days Pre Midpoint (n = 36) .584 .000** -.205 .230 Midpoint (n = 30) .145 .445 .124 .512 Average of 5 Days Post Midpoint (n = 33) .283 .111 .387 .026* Average of 5 Days Pre Endpoint (n = 37) .569 .000** -.006 .970 Endpoint (n = 34) .332 .055† .140 .431 Average of 5 Days Post Endpoint (n = 36) .376 .024* -.062 .718
† Not far from significance * p < .05 ** p < .01
153
Descriptive data for Brief COPE responses (Table 6.5) indicate that acceptance is the highest
used coping strategy by patients and caregivers at baseline, and continued to be the most highly
engaged in coping strategy at endpoint, at an increased level. Patients started out with slightly
higher levels of acceptance at baseline compared to caregivers, but caregivers had slightly higher
levels of acceptance by endpoint. Among patients, emotional support was the second most used
coping strategy at baseline, and remained the second highest at endpoint, however at a decreased
level. Among caregivers, the second most engaged in coping strategy across treatment was active
coping, however at a decreased level at endpoint compared to baseline. Among patients, active
coping was their third most used coping strategy, yet, was engaged in at a higher level than seen
among caregivers across treatment. The least engaged in coping strategy among patients and
caregivers across treatment was behavioural disengagement, but at a decreased level at endpoint
as compared to baseline. Caregivers consistently engaged in greater substance use than patients
across treatment and showed increased levels across treatment as compared to baseline, whereas
among patients substance use decreased from baseline to endpoint.
154
Table 6.5 Descriptive data for Brief COPE are mean ± standard deviation.
Patients Caregivers Baseline
Mean (SD) Midpoint Mean (SD)
Endpoint Mean (SD)
Baseline Mean (SD)
Midpoint Mean (SD)
Endpoint Mean (SD)
Self-Distraction 2.95 (.87) 2.72 (.88) 2.73 (.82) 2.28 (.74) 2.20 (.75) 2.22 (.81) Active Coping 3.08 (.80) 2.96 (.79) 2.81 (.82) 2.78 (.89) 2.48 (.94) 2.37 (.79)
Denial 1.51 (.72) 1.31 (.60) 1.21 (.42) 1.15 (.35) 1.13 (.46) 1.09 (.25) Substance Use 1.27 (.62) 1.14 (.51) 1.16 (.48) 1.30 (.47) 1.43 (.79) 1.33 (.54)
Emotional Support
3.26 (.81) 3.06 (.81) 2.97 (.81) 2.00 (.67) 1.83 (.86) 1.78 (.75)
Instrumental Support
2.91 (.79) 2.69 (.84) 2.60 (.87) 1.91 (.62) 1.70 (.75) 1.67 (.72)
Behavioural Disengagement
1.19 (.43) 1.11 (.35) 1.06 (.25) 1.13 (.38) 1.00 (.00) 1.04 (.21)
Venting 2.12 (.82) 2.14 (.73) 2.14 (.79) 1.35 (.44) 1.41 (.65) 1.33 (.42) Positive
Reframing 2.80 (.95) 2.89 (.83) 2.88 (.85) 2.39 (.80) 2.57 (.97) 2.11 (.85)
Planning 2.97 (.80) 2.62 (.86) 2.68 (.85) 2.52 (.87) 2.24 (.86) 2.13 (.87) Humour 2.14 (.95) 2.19 (1.04) 2.15 (.89) 1.50 (.75) 1.48 (.86) 1.48 (.68)
Acceptance 3.33 (.63) 3.49 (.54) 3.45 (.57) 3.28 (.64) 3.37 (.68) 3.48 (.79) Religion 2.48 (1.29) 2.42 (1.24) 2.43 (1.19) 1.89 (1.07) 1.76 (1.13) 1.89 (1.11)
Self-Blame 1.64 (.82) 1.41 (.57) 1.40 (.56) 1.17 (.32) 1.26 (.50) 1.26 (.40)
155
Pairwise comparisons (Friedman tests) were performed to assess differences in Brief COPE
scores across treatment from baseline to endpoint (Table 6.6a).
Among patients and caregivers, Friedman tests were used instead of one-way repeated
measures ANOVAs due to the non-normal distribution of Brief COPE scores. Only among
caregivers were some Brief COPE scores normally distributed: self-distraction (across
treatment), and active coping (only normally distributed at midpoint and endpoint). Post hoc
analyses with Wilcoxon signed-rank tests were conducted with a Bonferroni correction applied
resulting in a significance level set at p < .017 (i.e., 0.05/3, where three represents number of
comparisons) when Friedman test results were significant (Table 6.6b). Among patients, there
were significant changes between baseline and midpoint, and baseline and endpoint for various
Brief COPE domains, but no differences between midpoint and endpoint scores. Among
caregivers, only one significant change existed between Brief COPE scores across treatment,
from midpoint to endpoint.
Among patients, there was a significant change across treatment in Brief COPE scores for
Denial, Substance Use, Emotional and Instrumental Support, Behavioural Disengagement,
Planning, Acceptance, and Self-Blame, p < .05. Post hoc analysis were conducted with Wilcoxon
signed-rank tests with Bonferroni correction applied (p < .017). Denial, Planning, and Self-
Blame showed changes between baseline versus midpoint, and midpoint versus endpoint.
Median Denial coping levels for baseline, midpoint, and endpoint were 1.00 (1.00 to 2.00, 1.00
to 1.50, and 1.00 to 1.125 respectively). There were significant decreases in denial coping from
baseline to midpoint (p = .009), and baseline to endpoint (p < .0005). Median Planning scores
for baseline, midpoint, and endpoint were 3.00 (2.50 to 3.50), 2.50 (2.00 to 3.25), and 2.50 (2.00
to 3.50) respectively. Planning scores decreased from baseline to midpoint (p = .001), and from
baseline to endpoint (p = .006). Median Self-Blame scores for baseline, midpoint, and endpoint
were all 1.00. Self-Blame decreased from baseline to midpoint (p = .002), and baseline to
endpoint (p = .002). Substance Use, Emotional and Instrumental Support, Behavioural
Disengagement, and Acceptance only showed changes in scores between two times of treatment.
Median Substance Use coping levels for baseline, midpoint, and endpoint were each 1.00. There
was decreased coping via Substance Use from baseline to midpoint (p = .012). Median
156
Emotional Support coping levels for baseline, midpoint, and endpoint were 3.50 (2.50 to 4.00),
3.00 (2.50 to 4.00), and 3.00 (2.38 to 3.50) respectively. Patients sought less Emotional Support
from baseline to endpoint (p = .002). Median Instrumental Support coping levels for baseline,
midpoint, and endpoint were 3.00 (2.00 to 3.5), 2.50 (2.00 to 3.50), and 2.50 (2.00 to 3.50)
respectively. Use of Instrumental Support to cope decreased from baseline to endpoint (p =
.002). Median Behavioural Disengagement coping levels for baseline, midpoint, and endpoint
were all 1.00. Engaging in Behavioural Disengagement decreased across treatment from baseline
to endpoint (p = .011). Median Acceptance scores for baseline, midpoint, and endpoint were all
3.50 (3.00 to 4.00). Acceptance score changed from baseline to midpoint (p = .015).
Caregivers only exhibited significant changes in Brief COPE Positive Reframing scores across
treatment, p < .05. Post hoc analysis were conducted with Wilcoxon signed-rank tests with
Bonferroni correction applied (p < .017). Median Positive Reframing scores for baseline,
midpoint, and endpoint were 2.00 (2.00 to 3.00), 2.50 (2.00 to 3.50), and 2.00 (1.50 to 3.00),
respectively. Caregiver decreased Positive Reframing from midpoint to endpoint (p = .008).
157
Table 6.6a Median and Friedman test data for changes in Brief COPE scores across treatment among patients and caregivers Baseline Midpoint Endpoint Friedman test Mdn Mdn Mdn df χ2 p Patients
Self-Distraction 3.00 2.50 2.50 2 3.556 .169 Active Coping 3.00 3.00 3.00 2 3.508 .173
Denial 1.00 1.00 1.00 2 26.406 .000* Substance Use 1.00 1.00 1.00 2 12.822 .002*
Emotional Support 3.50 3.00 3.00 2 9.653 .008* Instrumental Support 3.00 2.50 2.50 2 12.531 .002*
Behavioural Disengagement
1.00 1.00 1.00 2 11.955 .003*
Venting 2.00 2.00 2.00 2 0.422 .810 Positive Reframing 3.00 3.00 3.00 2 1.033 .596
Planning 3.00 2.50 2.50 2 14.171 .001* Humour 2.00 2.00 2.00 2 0.575 .750
Acceptance 3.50 3.50 3.50 2 7.557 .023* Religion 2.00 2.00 2.25 2 0.182 .913
Self-Blame 1.00 1.00 1.00 2 9.329 .009* Caregivers
Self-Distraction 2.00 2.50 2.50 2 0.033 .983 Active Coping 3.00 2.50 2.50 2 5.040 .080
Denial 1.00 1.00 1.00 2 2.000 .368 Substance Use 1.00 1.00 1.00 2 1.040 .595
Emotional Support 2.00 2.00 2.00 2 3.085 .214 Instrumental Support 2.00 1.50 1.50 2 2.576 .276
Behavioural Disengagement
1.00 1.00 1.00 2 3.500 .174
Venting 1.00 1.00 1.00 2 0.136 .934 Positive Reframing 2.00 2.50 2.00 2 6.694 .035*
Planning 2.50 2.00 2.00 2 4.946 .084 Humour 1.00 1.00 1.00 2 1.067 .587
Acceptance 3.50 3.50 4.00 2 4.105 .128 Religion 1.50 1.00 1.50 2 3.211 .201
Self-Blame 1.00 1.00 1.00 2 3.379 .185 * p < .05
158
Table 6.6b Wilcoxon signed-rank tests with Bonferroni correction applied for differences in Brief COPE scores across treatment
Baseline vs. Midpoint Baseline vs. Endpoint Midpoint vs. Endpoint z p z p z p Patients
Denial -2.604 .009* -3.630 .000* -1.655 .098 Substance Use -2.503 .012* -1.745 .081 -0.741 .459
Emotional Support -1.720 .085 -3.149 .002* -1.242 .214 Instrumental Support -2.169 .030 -3.062 .002* -0.848 .396
Behavioural Disengagement
-1.761 .078 -2.551 .011* -0.828 .408
Planning -3.184 .001* -2.741 .006* -0.613 .540 Acceptance -2.443 .015* -1.530 .126 -0.875 .382 Self-Blame -3.100 .002* -3.084 .002* -0.367 .714
Caregivers Positive Reframing -1.053 .292 -1.467 .142 -2.632 .008*
* p < .017
159
Multiple regression analyses were run to predict individual Brief COPE styles across the course
of treatment from a participant’s role in the study (patient or caregiver), chronotype (MEQ
score), and sleep quality (PSQI score) at baseline midpoint or endpoint (Table 6.7). Log
transformations (Lg10) were carried out (for data in Table 6.7) for Denial, Substance Use,
Behavioural Disengagement, and Self Blame data at baseline, midpoint, and endpoint. Reverse
log transformations (Reverse_Lg10) were carried out for Acceptance data at baseline, midpoint,
and endpoint. There was linearity as assessed by partial regression plots and a plot of studentized
residuals against the predicted values. The Durbin-Watson statistic was used to assess
independence of residuals. The assumption of homoscedacity was met as assessed by visual
inspection of plots of studentized residuals versus unstandardized predicted values. No evidence
of multicollinearity was found as assessed by tolerance values greater than 0.1. There were a
minimal number of studentized deleted residuals ± 3 standard deviations and leverage values
greater than 0.2, and no values for Cook’s distance above 1. The assumptions of normality were
met as assessed by Q-Q plots. At different time points in the study, participant role, chronotype,
and sleep quality each contributed differently to the use of particular Brief COPE styles.
Regression coefficients and standard errors are presented in Table 6.7.
Note that the multiple regression analysis performed for behavioural disengagement at baseline
may not have independence of errors (residuals), based on assessment by the Durbin-Watson
statistic (.042). According to Field (2009) values less than one or greater than three are cause for
concern as these are on the fringes of the data range for the Durbin-Watson test statistic. A value
of approximately 2 for the Durbin-Watson test statistic indicates that no correlation exists
between residuals. At midpoint and endpoint, the Durbin-Watson test statistics (2.067 and 2.185,
respectively) indicates independence of errors. Caution should be taken when interpreting the
multiple regression analyses for behavioural disengagement at baseline.
Across treatment, multiple regression models consistently significantly predicted the use of Self-
Distraction, Emotional Support, Instrumental Support, Venting, and Humour (Self-Distraction:
baseline: F(3, 105) = 8.055, p < .0005 , adj. R2 = .164, midpoint: F(3, 104) = 3.494, p < .018, adj.
R2 = .065, endpoint: F(3, 105) = 5.466, p < .002, adj. R2 = .110; Emotional Support: baseline:
F(3, 105) = 16.594, p < .0005, adj. R2 = .302, midpoint: F(3, 104) = 13.738, p < .0005, adj. R2 =
.263, endpoint: F(3, 105) = 15.471, p < .0005 , adj. R2 = .287; Instrumental Support: baseline:
160
F(3, 105) = 10.331, p < .0005, adj. R2 = .206, midpoint: F(3, 104) = 8.920, p < .0005, adj. R2 =
.182, endpoint: F(3, 105) = 8.061, p < .0005 , adj. R2 = .164; Venting: baseline: F(3, 105) =
7.367, p < .0005, adj. R2 = .150, midpoint: F(3, 104) = 7.447, p < .0005, adj. R2 = .153, endpoint:
F(3, 105) = 9.291, p < .0005, adj. R2 = .187; Humour: baseline: F(3, 105) = 3.319, p < .050, adj.
R2 = .061; midpoint: F(3, 104) = 3.194, p < .050 , adj. R2 = .058; endpoint: F(3, 105) = 4.211, p <
.010, adj. R2 = .082). In each of these models, participant role was a significant predictor, p < .05,
both by itself, and when combined with other significant factors; patients consistently engaged in
greater Self-Distraction and Venting, sought more Emotional and Instrumental Support, and used
more Humour to cope than caregivers. For the use of Self-Distraction, chronotype was
significantly predictive at baseline (p < .05); those with higher chronotype scores (i.e., tendency
towards being M type) more likely report greater use of self-distraction to cope. While
chronotype was no longer a significant predictor of Self-Distraction at endpoint, sleep quality
significantly contributed to the model, p < .05; higher PSQI scores, (i.e., increasingly poor sleep
quality) predict greater use of Self-Distraction to cope. Similarly, sleep quality was also a
significant predictor for the use of Venting at endpoint (p < .05) and a trend moving towards
significance for this variable had been present since baseline; however unlike Self-Distraction,
chronotype was not a significant contributor to the multiple regression model at baseline. Sleep
quality and chronotype do not significantly predict Humour at any time, however it appears that
the significance levels for chronotype and sleep quality across treatment run counter to one
another, reflecting a similar pattern to that seen in the multiple regression models for Self-
Distraction and Venting.
For Active Coping, the multiple regression model was not significantly predictive at baseline,
F(3, 105) = .792, p = .501, adj. R2 = -.006. At midpoint, the multiple regression model indicated
an overall shift towards significance, but the model did not reach significance, F(3, 104) = 2.428,
p = .070 , adj. R2 = .039. While the midpoint model was not significant, the value of participant
role reached significance, p < .05, suggesting patients engage in more Active Coping than
caregivers. At endpoint, the multiple regression model was significantly predictive of Active
Coping, F(3, 105) = 3.985, p < .010, adj. R2 = .077; sleep quality acted as a significant predictor,
p < .05, such that higher PSQI scores, (i.e., increasingly poor sleep quality) predict greater use of
Active Coping.
161
Across treatment, multiple regression models were not significantly predictive of Denial or
Acceptance (Denial: baseline: F(3, 105) = 2.681, p = .051, adj. R2 = .045, midpoint: F(3, 104) =
.849, p = .470 , adj. R2 = -.004, endpoint: F(3, 105) = 1.446, p = .234 , adj. R2 = .012;
Acceptance: baseline: F(3, 105) = .443, p = .723 , adj. R2 = -.016, midpoint: F(3, 104) = 1.430, p
= .238 , adj. R2 = .012, endpoint: F(3, 105) = 1.800, p = .340 , adj. R2 = .004). Neither coping
behaviour was significantly predicted by multiple regression models, yet certain factors reached
or approached significance: for Denial, participant role was significant at baseline, p < .05; for
Acceptance, chronotype was not a significant predictor, but it predictive value approached
significance as treatment progressed (baseline: p = .290, midpoint: p = .130, endpoint: p = .080).
For Substance Use, the multiple regression model was not significantly predictive at baseline
(F(3, 105) = 1.409, p = .244, adj. R2 = .011) or endpoint (F(3, 105) = 1.066, p = .367, adj. R2 =
.002). At midpoint, the multiple regression model did significantly predicted Substance Use, F(3,
104) = 3.157, p < .05 , adj. R2 = .057; participant role acted as a significant predictor, p < .05,
suggesting caregivers engaged in greater Substance Use than patients at midpoint.
At baseline, midpoint, and endpoint, the multiple regression model was not statistically
significantly predictive of Behavioural Disengagement, F(3, 105) = .156, p = .925 , adj. R2 = -
.024; F(3, 104) = 1.766, p = .158, adj. R2 = .021; F(3, 105) = 2.611, p = .055 , adj. R2 = .041,
respectively. While neither participant role, chronotype, or sleep quality were significant
predictors across treatment (p < .05), it is important to note that chronotype and sleep quality
each approached significance as treatment progressed. For chronotype, the significance value
progressed from p = .869 at baseline, p = .168 at midpoint, and p = .069 at endpoint. For sleep
quality, significance value progressed from p = .886 at baseline, p = .267 at midpoint, and p =
.072 at endpoint. Participant role did not approach significance from baseline to endpoint.
Multiple regression models for Positive Reframing and Planning were only significant at
endpoint (Positive Reframing: baseline: F(3, 105) = 1.451, p = .232, adj. R2 = .012, midpoint:
F(3, 104) = 0.926, p = .431, adj. R2 = -.002, endpoint: F(3, 105) = 6.251, p = .001, adj. R2 =
.127; Planning: baseline: F(3, 105) = 2.010, p = .117 , adj. R2 = .027, midpoint: F(3, 104) =
2.195, p = .093 , adj. R2 = .032, endpoint: F(3, 105) = 5.559, p < .001, adj. R2 = .112). For both
models, at endpoint participant role was a significant predictor; patients engaged in greater use of
162
both strategies to cope. While neither model was significantly predictive at baseline, participant
role appeared as a significant factor in each, p < .05. Sleep quality was a significant predictor of
Planning at endpoint; worse sleep quality (i.e., higher PSQI score) predicted greater use of
Planning to cope. It appears there was a trend moving towards significance across treatment.
For Religion, the multiple regression model was significant at baseline, midpoint and endpoint,
F(3, 105) = 3.530, p < .017, adj. R2 = .066; F(3, 104) = 4.105, p < .008, adj. R2 = .080; F(3, 105)
= 3.684, p < .014 , adj. R2 = .069, respectively. Participant role was a significant predictor at
baseline and midpoint, p < .05, but not at endpoint, p > .05; patients more likely used Religion to
cope than caregivers. Chronotype was a significant predictor across treatment, p < .05; greater
preference for morningness predicted increased use of Religion to cope. Sleep quality was not a
significant predictor at any time point.
Multiple regression models significantly predicted the use of Self-Blame at baseline, F(3, 105) =
4.316, p < .007, adj. R2 = .084, and endpoint, F(3, 105) = 3.008, p < .034, adj. R2 = .053. At
midpoint, the multiple regression model did not significantly predict coping via Self-Blame, F(3,
104) = 1.036, p < .380, adj. R2 = .001. At baseline, participant role was the only significant
predictor, p < .05. At endpoint, participant role was no longer significant, but sleep quality had
become a statistically significant predictor, p < .05. Sleep quality did not show a progressive
trend towards significance across treatment. At baseline, while chronotype did not reach
significance, it was relatively close to a significant p value (p = .058).
163
Table 6.7 Summary of Multiple Regression Analyses for Brief-COPE Scores Across Treatment, assessing the predictive value of participant role, chronotype, and sleep quality
Baseline Midpoint Endpoint Brief COPE Included
Covariates B (SE) ß p Adjusted
R2 B (SE) ß p Adjusted
R2 B (SE) ß p Adjusted
R2 Self-Distraction .164 .065 .110 Constant
Participant role MEQ PSQI
1.269 (.531) -0.687 (.192) 0.029 (.008) 0.007 (.022)
-.319 .300 .027
.019 .001* .001*
.762
1.658 (.586) -0.500 (.203) 0.015 (.009) 0.030 (.026)
-.235 .153 .108
.006 .015*
.109
.262
1.985 (.541) -0.399 (.190) 0.003 (.008) 0.067 (.023)
-.195 .033 .276
.000 .038*
.717 .004**
Active Coping -.006 .039 .077 Constant
Participant role MEQ PSQI
3.340 (.599) -0.296 (.197) 0.000 (.009) 0.002 (.023)
-.147 .005 .008
.000
.136
.962
.937
2.540 (.572) -.446 (.198) .003 (.009) .027 (.026)
-.218 .038 .102
.000 .026**
.694
.297
1.601 (.546) -.373 (.192) .014 (.008) .047 (.023)
-.184 .157 .198
.004
.055
.095 .040*
Denial .045 -.004 .012 Constant
Participant role MEQ PSQI
.075 (.119) -.090 (.039) .002 (.002) .004 (.005)
-.219 .119 .082
.529 .024*
.212
.394
.034 (.101) -.049 (.035) .001 (.002) .002 (.004)
-.138 .037 .047
.734
.166
.706
.640
.135 (.077) -.029 (.027) -.002 (.001) .002 (.003)
-.106 -.131 .068
.083
.279
.179
.489
Substance Use .011 .057 .002 Constant
Participant role MEQ PSQI
-.134 (.105) .022 (.035) .003 (.002) .002 (.004)
.062 .183 .058
.205
.524
.060
.555
-.124 (.091) .082 (.032) .002 (.001) .004 (.004)
.248 .151 .092
.177 .011*
.115
.343
.008 (.086) .052 (.030) .001 (.001) .001 (.004)
.169 .040 .019
.930
.088
.680
.852
Emotional Support
.302 .263 .287
Constant Participant role MEQ PSQI
3.678 (.565) -1.247 (.186)
.013 (.008)
.013 (.021)
-.548 .124 .052
.000 .000*
.126
.530
2.758 (.570) -1.249 (.197)
.006 (.009) -.007 (.025)
-.535 .061
-.023
.000 .000*
.469
.787
1.877 (.539) -1.125 (.189)
.013 (.008)
.042 (.022)
-.494 .127 .157
.001 .000*
.124
.064
Instrumental Support
.206 .182 .164
Constant Participant role MEQ PSQI
3.939 (.554) -1.012 (.182)
.001 (.008) -.001 (.021)
-.483 .014
-.047
.000 .000*
.870
.592
2.282 (.574) -.992 (.198) .006 (.009) .009 (.026)
-.445 .060 .031
.000 .000*
.495
.734
2.103 (.574) -.877 (.202) .004 (.009) .032 (.024)
-.391 .040 .122
.000 .000*
.657
.181
Behavioural Disengagement
-.024 .024 .048
Constant .089 (.083) .289 -.061 (.058) .296 .065 (.048) .178
164
Participant role MEQ PSQI
-.018 (.027) .000 (.001) .000 (.003)
-.065 -.016 -.014
.511
.869
.886
-.030 (.020) .001 (.001) .003 (.003)
-.145 .134 .109
.139
.168
.267
.002 (.017) -.001 (.001) .004 (.002)
.009 -.175 .177
.928
.069
.072 Venting .150 .153 .187 Constant
Participant role MEQ PSQI
2.004 (.497) -.712 (.180) -.003 (.008) .036 (.021)
-.357 -.037 .156
.000 .000*
.680
.088
1.947 (.491) -.671 (.170) -.002 (.008) .038 (.022)
-.358 -.026 .159
.000 .000*
.771
.085
1.639 (.492) -.745 (.173) .002 (.008) .043 (.021)
-.382 .027 .187
.001 .000*
.755 .039*
Positive Reframing
.012 -.002 .127
Constant Participant role MEQ PSQI
2.878 (.608) -.444 (.220) .002 (.010)
-.021 (.025)
-.196 .018
-.081
.000 .046*
.854
.405
2.642 (.605) -.317 (.209) .003 (.009) .010 (.027)
-.149 .031 .038
.000
.132
.751
.705
2.096 (.579) -.705 (.203) .007 (.009) .043 (.024)
-.319 .074 .166
.000 .001*
.414
.077
Planning .027 .032 .112 Constant
Participant role MEQ PSQI
2.596 (.539) -.424 (.195) .004 (.009) .016 (.022)
-.209 .046 .068
.000 .032*
.627
.483
1.840 (.592) -.324 (.204) .007 (.009) .043 (.026)
-.153 .078 .160
.002
.116
.417
.105
1.611 (.568) -.435 (.200) .008 (.009) .069 (.024)
-.202 .088 .271
.005 .031*
.340 .005*
Humour .061 .080 .082 Constant
Participant role MEQ PSQI
2.747 (.606) -.633 (.219) -.011 (.010) -.002 (.025)
-.273 -.102 -.006
.000 .005*
.280
.947
2.600 (.702) -.702 (.243) -.008 (.011) .004 (.031)
-.276 -.069 .013
.018 .002*
.499
.644
1.071 (.586) -.622 (.206) -.003 (.009) .027 (.024)
-.285 -.208 .106
.001 .003*
.760
.266
Acceptance -.016 .012 .006 Constant
Participant role MEQ PSQI
.300 (.107) .017 (.039)
-.002 (.002) -.001 (.004)
.043
-.104 -.018
.006
.664
.290
.868
.253 (.104) .036 (.036)
-.002 (.002) .004 (.005)
.097
-.148 .095
.017
.326
.130
.338
.325 (.111) -.018 (.039) -.003 (.002) .001 (.005)
-.046 -.171 .016
.004
.644
.080
.869
Religion .066 .080 .069 Constant
Participant role MEQ PSQI
.825 (.806) -.637 (.291) .032 (.013)
-.016 (.033)
-.206 .229
-.044
.308 .031* .016*
.644
.613 (.824) -.697 (.285) .032 (.013)
-.002 (.037)
-.231 .240
-.005
.459 .016* .012*
.958
.377 (.785) -.501 (.276) .031 (.012) .037 (.033)
-.173 .237 .108
.632
.072 .013*
.263
Self-Blame .084 .001 .053 Constant
Participant role MEQ PSQI
.317 (.117) -.099 (.042) -.044 (.002) .006 (.005)
-.219 -.177 .122
.008 .021*
.058
.196
.247 (.103) -.047 (.036) -.002 (.002) -.004 (.005)
-.130 -.103 -.082
.019
.190
.293
.413
.176 (.093) -.017 (.033) -.002 (.001) .008 (.004)
-.051 -.149 .203
.062
.597
.117 .038*
Note. *p < .05; B = unstandardized regression coefficient, SEB = standard error of the coefficient; ß = standardized coefficient; Participant role: patient = 0, caregiver = 1
165
Comparing personality facets, descriptive data are mean ± standard deviation (Table 6.8).
There was homogeneity of variances for all personality facets, as assessed by Levene’s test for
equality of variances (p > .05). Among patients and caregivers, volatility was the lowest
expressed trait, however patients show a slightly higher average volatility level than caregivers,
though not significantly. The second lowest expressed trait among both groups is withdrawal,
which is again slightly less in caregivers than among patients, though not significantly. Among
patients, the highest expressed trait is compassion, while among caregivers, politeness was the
highest expressed trait. Patients were significantly more compassionate than caregivers, 0.43
(95% CI, 0.19 to 0.66), t(107) = 3.601, p < .000). There was no significant difference in
politeness between patients and caregivers. Patients were significantly more enthusiastic than
caregivers, 0.35 (95% CI, 0.05 to 0.53), t(107) = 2.293, p = .024. The only two traits expressed
among caregivers at a slightly higher level than among patients were industriousness and
assertiveness, yet there was no significant difference.
166
Table 6.8 Descriptive data and independent samples t-test for BFAS between men and women. Descriptive data are mean ± standard deviation, for patients and caregivers Patients
(n = 86) Caregivers
(n = 23) Independent Samples t-test
Mean (SD) Mean (SD) df t p Withdrawal 2.68 (.71) 2.45 (.60) 107 1.391 .167
Volatility 2.56 (.72) 2.40 (.69) 107 .949 .345 Compassion 4.33 (.51) 3.90 (.50) 107 3.601 .000**
Politeness 4.18 (.44) 4.02 (.49) 107 1.495 .138 Industriousness 3.75 (.67) 3.77 (.54) 107 -.099 .921
Orderliness 3.76 (.65) 3.53 (.62) 107 1.494 .138 Enthusiasm 3.84 (.68) 3.49 (.56) 107 2.293 .024*
Assertiveness 3.50 (.69) 3.55 (.57) 107 -.326 .745 Intellect 3.62 (.67) 3.60 (.62) 107 .122 .903
Openness 3.60 (.63) 3.33 (.65) 107 1.761 .081 * p < .05; ** p < .01
167
Multiple regression analyses were run to predict individual Brief COPE styles across the course
of treatment from a participant’s role in the study (patient or caregiver), chronotype (MEQ
score), sleep quality (PSQI score) at baseline midpoint or endpoint, and personality (BFAS)
(Table 6.9). Log transformations (Lg10) were carried out (for data in Table 6.9) for Denial,
Substance Use, Behavioural Disengagement, and Self Blame data at baseline, midpoint, and
endpoint. Reverse log transformations (Reverse_Lg10) were carried out for Acceptance data at
baseline, midpoint, and endpoint. There was linearity as assessed by partial regression plots and
a plot of studentized residuals against the predicted values. The Durbin-Watson statistic was used
to assess independence of residuals. The assumption of homoscedacity was met as assessed by
visual inspection of plots of studentized residuals versus unstandardized predicted values. No
evidence of multicollinearity was found as assessed by tolerance values greater than 0.1. There
were a minimal number of studentized deleted residuals ± 3 standard deviations and leverage
values greater than 0.2, and no values for Cook’s distance above 1. The assumptions of normality
were met as assessed by Q-Q plots. At different time points in the study, participant role,
chronotype, sleep quality and personality each contributed differently to the use of particular
Brief COPE styles. Regression coefficients and standard errors are presented in Table 6.9.
For Self-Distraction, the multiple regression model was significantly predictive, at baseline,
F(13, 95) = 2.277, p < .012, adj. R2 = .133, and endpoint F(13, 95) = 1.938, p < .035, adj. R2 =
.101; the model was not significant at midpoint, F(13, 94) = 1.369, p = .189, adj. R2 = .043.
Participant role and chronotype were significant predictors at baseline, p < .05, while only sleep
quality was a significant predictor at endpoint, p < .05. This mimics the same trend, shifting from
an influence of participant role and chronotype at outset, to a predictive influence of sleep quality
by endpoint as seen when personality factors were not considered. The models suggested that at
baseline, patients and or those reporting higher preference for morningness, and at endpoint,
those reporting poorer sleep quality (i.e., higher PSQI score) more likely engaged in Self-
Distraction to cope. While no personality factors were significant predictors of Self-Distraction,
politeness appeared to be approaching significance as treatment progressed (baseline: p < .809,
midpoint: p < .176, endpoint: p < .080).
The multiple regression model only significantly predicted Active Coping at baseline, F(13, 95)
= 2.602, p < .004, adj. R2 = .162. Orderliness significantly predicted Active Coping at baseline; a
168
greater tendency towards orderliness predicts increased Active Coping. At midpoint and
endpoint, the multiple regression model was not significantly predictive of Active Coping, F(13,
94) = 1.264, p = .249, adj. R2 = .031, and F(13, 95) = 1.605, p = .097 , adj. R2 = .068,
respectively. While the model was not predictive of Active Coping at midpoint, participant role
did reach significance, p < .05; patients are more likely than caregivers to engage in Active
Coping at midpoint. At endpoint, Active Coping did not reach significance, p = .097.
At baseline and midpoint, the multiple regression model significantly predicted Denial, F(13, 95)
= 2.142, p < .018, adj. R2 = .121, and F(13, 94) = 2.031, p < .026, adj. R2 = .111, respectively.
There were no consistent significant predictors across treatment. At baseline, participant role and
openness were significant predictors; patients and or those reporting lower openness were more
likely to engage in Denial. At midpoint, withdrawal was the only statistically significant
predictor for the multiple regression model; higher withdrawal signals increased use of Denial to
cope. By endpoint, the multiple regression model did not predict Denial, F(13, 95) = 1.184, p =
.303, adj. R2 = .022.
The multiple regression model was significantly predictive of engaging in Substance Use to cope
at baseline F(13, 95) = 1.946, p < .034, adj. R2 = .102 and midpoint, F(13, 94) = 2.542, p < .005,
adj. R2 = .158. Chronotype and industriousness predicted engaging in Substance Use to cope at
baseline and midpoint; greater tendency towards morningness (i.e., higher MEQ score), and
lower propensity for industriousness predicts increased use. At midpoint, participant role also
became a significant predictor; caregivers were more likely to engage in Substance Use to cope.
At endpoint, the multiple regression model was not significantly predictive of engaging in
Substance Use, F(13, 95) = 1.741, p = .065, adj. R2 = .082; while the model narrowly missed
being significant, participant role continued to be a significant predictor.
Across treatment, multiple regression models consistently significantly predicted the use of
Emotional and Instrumental Support, and Venting as means to cope (Emotional Support:
baseline: F(13, 95) = 4.042, p < .0005, adj. R2 = .268, midpoint: F(13, 94) = 4.037, p < .0005,
adj. R2 = .270, endpoint: F(13, 95) = 3.846, p < .0005, adj. R2 = .255; Instrumental Support:
baseline: F(13, 95) = 3.265, p < .0005, adj. R2 = .214, midpoint: F(13, 94) = 2.711, p < .003, adj.
R2 = .172, endpoint: F(13, 95) = 2.565, p < .004, adj. R2 = .158; Venting: baseline: F(13, 95) =
3.546, p < .0005, adj. R2 = .235, midpoint: F(13, 94) = 3.805, p < .0005, adj. R2 = .254, endpoint:
169
F(13, 95) = 4.671, p < .0005, adj. R2 = .306). Participant role was a significant predictor across
treatment for Emotional and Instrumental Support, and Venting models, p < .01, such that
patients reporting engaging in greater use of these methods to cope than caregivers; for
Instrumental Support, it was the sole significant predictor. For Emotional Support, intellect was a
significant predictor at midpoint, p < .05; higher intellect was predictive of less need for
Emotional Support. For Venting, withdrawal was a significant predictor at baseline and midpoint
(p < .05), but not at endpoint; reporting higher withdrawal predicted greater venting at baseline
and midpoint. By endpoint, withdrawal was no longer significantly predicted Venting; it
appeared that as treatment progressed withdrawal became a progressively less significant
predictor. The presence of participant role as a significant predictor in each of these models
across treatment reflects the same trend observe when personality was not included, suggesting
one’s role as a patient or caregiver plays a consistent role in determining their use of these
coping behaviours even when considering a range of different cofactors.
The multiple regression model was significant of Behavioural Disengagement at baseline, F(13,
95) = 4.197, p < .0005, adj. R2 = .278, and endpoint, F(13, 95) =1.857 , p < .045, adj. R2 = .093,
but not midpoint, F(13, 94) = 1.395, p = .177, adj. R2 = .046. At baseline, increased propensity
for withdrawal, and decreased levels of openness predicted greater use of Behavioural
Disengagement; assertiveness and orderliness narrowly missed reaching significance (p = .052
and p = .076, respectively). The model is not significantly predictive of Behavioural
Disengagement at midpoint, however it should be noted that both participant role and chronotype
score narrowly miss reaching statistical significance (p = .053 and p = .054, respectively) as
predictors. At endpoint, while the model was statistically significant, there were no significant
predictors. The two predictors with p-values closest to significance were enthusiasm (p = .076)
and sleep quality (p = .099). There was no prior trend at baseline or midpoint that indicated a
shift towards significance in these two predictors.
Positive Reframing was significantly predicted by the multiple regression model across treatment
(baseline: F(13, 95) = 2.207, p < .015, adj. R2 = .127, midpoint: F(13, 94) = 2.075, p < .023, adj.
R2 = .115, endpoint: F(13, 95) = 2.800, p < .002, adj. R2 = .178). Openness consistently predicted
positive reframing across treatment, p < .05; increased propensity for openness predicted greater
use of Positive Reframing. At baseline, greater tendency for industriousness predicted increased
Positive Reframing. At midpoint, tendency towards compassion predicted increased Positive
170
Reframing. At endpoint, increased ratings of enthusiasm and or being a patient predicted higher
levels of Positive Reframing. While orderliness consistently predicted Positive Reframing, these
other factors should not be ignored.
The multiple regression model was significantly predictive of Planning as a coping strategy at
endpoint, F(13, 95) = 2.390, p < .008, adj. R2 = .143. Higher intellect and poor sleep quality (i.e.,
higher PSQI score) significantly predicted Planning at endpoint. When personality was not
included in the multiple regression model, PSQI score also significantly predicted Planning
behaviours only at endpoint. The multiple regression model was not significantly predictive of
Planning at baseline, F(13, 95) = 1.407, p = .170, adj. R2 = .047, or midpoint, F(13, 94) = .997, p
=. 461, adj. R2 = .000. While the model was not significant at baseline, intellect appeared as a
significant predictor variable, p < .05.
At baseline and midpoint, multiple regression models did not significantly predict the use of
Humour as a coping strategy, F(13, 95) = 1.653, p = .084, adj. R2 = .073, and F(13, 94) = 1.632,
p = .090, adj. R2 = .071, respectively. While the models were not statistically significant at
baseline or midpoint, at both time points, participant role appeared as a significant predictor, p <
.05; patients were more likely than caregivers to use Humour to cope. By endpoint, the multiple
regression model significantly predicted using Humour to cope, F(13, 95) = 2.513, p < .005, adj.
R2 = .154; intellect was the only significant predictor – higher intellect predicts greater use of
Humour. Participant role (p = .078) was the next most significant predictor at endpoint; patients
more likely to used Humour to cope. Intellect did not approach significance at baseline or
midpoint as a predictor.
Across treatment, the multiple regression models were not significant predictors of Acceptance
(baseline: F(13, 95) = 1.710, p = .071, adj. R2 = .079, midpoint: F(13, 94) = .542, p = .892, adj.
R2 = -.059, endpoint: F(13, 95) = 1.017, p = .441, adj. R2 = .002). While the model was not
significant at baseline, withdrawal and volatility were significant predictors, p < .05. This is the
only time in this study where volatility had some predictive value on a coping behaviour. While
enthusiasm was not a significant predictor of Acceptance across treatment, it appeared to be
approaching significance as treatment progressed; this trend suggested that the higher one’s level
of enthusiasm, the less likely they engaged in Acceptance to cope. The lack of a significant
model reflects the pattern observed when personality was not included in the multiple regression,
171
suggesting that neither participant role, chronotype, sleep quality, or personality strongly predict
one’s use of Acceptance to cope with the stress being a cancer patient or caregiver.
Using Religion to cope was significantly predicted across treatment by multiple regression
models (baseline: F(13, 95) = 2.149, p < .018, adj. R2 = .121, midpoint: F(13, 94) = 2.384, p <
.008, adj. R2 = .144, endpoint: F(13, 95) = 2.252, p < .013, adj. R2 = .131). At baseline,
chronotype and assertiveness were significant predictors, p < .05; preference for morningness
and increased assertiveness predicted greater use of religion. Openness narrowly missed being a
significant baseline predictor, p = .052. At midpoint, participant role, industriousness, and
openness were significant predictors, p < .05, while assertiveness (p = .069) and chronotype (p =
.056) were not; being a patient, and reporting higher industriousness and openness increased
propensity for using Religion to cope at midpoint. At endpoint, industriousness and assertiveness
were significant predictors, p < .05, while participant role (p = .064), chronotype (p = .064), and
assertiveness (p = .068) were not significant; high industriousness and assertiveness predicted
greater use of Religion to cope. The lack of consistent significance of chronotype as a predictor
across treatment differs from when personality was not included in the model.
Multiple regression models significantly predicted engaging in Self-Blame at baseline, F(13, 95)
= 3.228, p < .0005, adj. R2 = .211, and endpoint, F(13, 95) = 2.375, p < .008, adj. R2 = .142. At
midpoint, the multiple regression model narrowly missed significance, F(13, 94) = 1.780, p =
.058, adj. R2 = .087. At baseline, withdrawal and industriousness significantly predicted
engaging in Self-Blame; lower propensity for withdrawal and higher industriousness predict
greater use of Self-Blame to cope. While the model was not significantly predictive at midpoint,
it should be noted that withdrawal narrowly missed being a significant predictor, p < .05. By
endpoint, industriousness and openness were significant predictors; lower industriousness and
higher openness predicted greater use of Self-Blame. While at baseline industriousness positively
predicted Self-Blame, by endpoint, industriousness negatively predicted engaging in Self-Blame;
withdrawal and openness, while not significant across treatment, also switched the direction to
which they contributed from baseline to endpoint.
172
Table 6.9 Summary of Multiple Regression Analyses for Brief-COPE Scores Across Treatment, assessing the predictive value of participant role, chronotype, sleep quality and personality
Baseline Midpoint Endpoint Brief COPE Included
Covariates B (SE) ß p Adjusted
R2 B (SE) ß p Adjusted
R2 B (SE) ß p Adjusted
R2 Self-Distraction .133 .043 .101 Constant
Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
.736 (1.503) -.655 (.220) .029 (.009)
-.001 (.024)
.033 (.172)
.020 (.148) -.360 (.241) .057 (.233)
-.016 (.195) .198 (.159) .205 (.153)
-.135 (.183) .215 (.172) .183 (.142)
-.305 .299
-.003
.026
.016 -.217 .029
-.012 .146 .156
-.102 .160 .134
.625 .004* .003*
.978
.848
.893
.139
.809
.934
.215
.186
.463
.214
.201
3.221 (1.581) -.418 (.230) .017 (.010) .038 (.030)
.053 (.186)
-.019 (.156) -.076 (.253) -.331 (.243) -.083 (.207) .093 (.166) .289 (.165)
-.240 (.193) -.032 (.181) .022 (.151)
-.196 .176 .140
.042
-.016 -.046 -.172 -.060 .069 .221
-.183 -.024 .016
.044
.072
.091
.202
.775
.901
.764
.176
.690
.575
.083
.216
.860
.882
1.797 (1.258) -.277 (.213) .006 (.009) .052 (.026)
.236 (.173)
-.019 (.144) .117 (.230)
-.396 (.223) .027 (.195) .177 (.154) .055 (.149)
-.249 (.179) .103 (.168) .150 (.141)
-.135 .060 .214
.196
-.016 .074
-.214 .020 .137 .044
-.198 .081 .115
.221
.195
.549 .048*
.176 .897 .614 .080 .892 .254 .711 .167 .541 .291
Active Coping .162 .031 .068 Constant
Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
.996 (1.383) -.168 (.202) -.009 (.009) .015 (.022)
-.314 (.158) .206 (.136) .090 (.222)
-.008 (.214) .091 (.180) .342 (.146) .023 (.141) .052 (.169) .122 (.158) .068 (.131)
-.083 -.098 .064
-.264 .178 .058
-.004 .070 .270 .019 .042 .097 .053
.473
.408
.305
.506
.050* .133 .686 .970 .615
.021* .870 .758 .440 .606
3.362 (1.530) -.497 (.222) .003 (.010) .037 (.029)
-.305 (.180) .071 (.151)
-.117 (.245) -.026 (.235) -.147 (.200) .123 (.160) 019 (.159)
.033 (.186)
.196 (.175)
.023 (.146)
-.242 .029 .141
-.252 .060
-.073 -.014 -.112 .094 .015 .026 .153 .017
.030 .028*
.783
.201
.094
.641
.634
.911
.463
.447
.904
.860
.266
.877
1.029 (1.471) -.359 (.215) .014 (.009) .019 (.026)
.255 (.175)
-.145 (.145) .052 (.233)
-.210 (.225) .147 (.197)
-.133 (.156) .074 (.150) .228 (.181)
-.112 (.170) .231 (.142)
-.177 .151 .078
.213
-.124 .033
-.114 .113
-.104 .059 .183
-.089 .178
.486
.097
.142
.477
.149
.319
.823
.355
.458
.397
.626
.210
.511
.109
Denial .121 .111 .022 Constant
Participant role -.062 (.289)
-.091 (.042)
-.221 .831
.034* .203 (.252)
-.051 (.037)
-.143 .423 .172
.286 (.206) -.031 (.030)
-.111
.168
.310
173
MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
.003 (.002)
.002 (.005)
.048 (.033)
.020 (.028) -.012 (.046) .026 (.045)
-.029 (.038) .018 (.030) .011 (.029) .054 (.035)
-.034 (.033) -.056 (.027)
.170
.033
.198
.085 -.039 .071
-.109 .069 .044 .215
-.132 -.214
.086
.739
.149
.478
.792
.559
.449
.559
.707
.127
.306 .043*
.000 (.002)
.001 (.005)
.068 (.030) -.038 (.025) -.024 (.040) -.036 (.039) .042 (.033) .007 (.026) .043 (.026)
-.002 (.031) -.035 (.029) -.041 (.024)
.021
.030
.325 -.185 -.087 -.111 .187 .033 .197
-.009 -.160 -.181
.834
.779
.025* .134 .557 .361 .203 .780 .109 .951 .225 .092
-.001 (.001) .022 (.004)
.015 (.024)
-.007 (.020) -.011 (.033) .005 (.032)
-.021 (.027) .018 (.022) .005 (.021) .013 (.025)
-.038 (.024) -.011 (.020)
-.111 .052
.095
-.047 -.053 .019
-.116 .102 .028 .075
-.218 -.061
.291
.641
.528
.713
.729
.881
.457
.414
.825
.613
.116
.586 Substance Use .102 .158 .082 Constant
Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
-.160 (.253) .043 (.037) .005 (.002) .000 (.004)
.046 (.029)
-.017 (.025) .028 (.041)
-.016 (.039) -.078 (.033) -.005 (.027) .012 (.026)
-.005 (.031) .002 (.029) .019 (.024)
.120 .305
-.011
.219 -.083 .102
-.051 -.343 -.024 .053
-.025 .009 .083
.528
.253 .003*
.911
.115
.498
.492
.679 .019*
.844
.655
.859
.948
.431
.020 (.230) .074 (.033) .004 (.001) .005 (.004)
-.003 (.027) .021 (.023)
-.010 (.037) -.007 (.035) -.073 (.030) -.006 (.024) -.012 (.024) .012 (.028)
-.001 (.026) -.003 (.022)
.224 .283 .128
-.016 .108
-.038 -.024 -.345 -.029 -.058 .061
-.007 -.014
.930 .029* .004*
.212
.909
.365
.793
.842 .017*
.804
.626
.662
.959
.894
-.095 (.222) .074 (.032) .002 (.001) .000 (.004)
.009 (.026) .030 (.022) .040 (.035)
-.034 (.034) -.050 (.030) .016 (.023)
-.001 (.023) .005 (.027)
-.022 (.026) .011 (.021)
.240 .117
-.012
.051
.166
.169 -.121 -.254 .082
-.004 .028
-.116 .055
.668 .024*
.250
.909
.725
.179
.255
.326
.093
.500
.975
.849
.385
.612
Emotional Support
.268 .270 .255
Constant Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness
2.035 (1.462) -1.255 (.214)
.012 (.009)
.011 (.023)
.020 (.167)
.076 (.144) -.187 (.235) .168 (.227) .066 (.190)
-.008 (.154)
-.551 .115 .042
.015 .058
-.107 .082 .045
-.005
.167 .000*
.202
.643
.904
.598
.428
.461
.730
.959
3.463 (1.512) -1.137 (.220)
.006 (.009) -.008 (.028)
.143 (.178)
-.042 (.149) .256 (.242)
-.292 (.232) .040 (.198)
-.061 (.159)
-.487 .057
-.026
.104 -.032 .141
-.138 .027
-.041
.024 .000*
.526
.782
.423
.773
.293
.211
.841
.702
2.950 (1.477) -1.057 (.215)
.015 (.009)
.028 (.026)
.275 (.176) -.230 (.146) .203 (.233)
-.301 (.226) -.085 (.197) .036 (.156)
-.464 .151 .104
.205
-.175 .116
-.146 -.058 .025
.049 .000*
.099
.286
.121
.117
.386
.187
.667
.818
174
Enthusiasm Assertiveness
Intellect Openness
.110 (.149)
.160 (.178)
.100 (.167)
.007 (.138)
.079
.115
.071
.005
.465
.371
.549
.961
.088 (.158)
.268 (.184) -.365 (.173) -.120 (.144)
.061
.187 -.250 .081
.579
.149 .038*
.406
.080 (.151)
.092 (.181) -.063 (.171) -.005 (.143)
.058
.066 -.044 -.003
.596
.613
.712
.972 Instrumental Support
.214 .172 .158
Constant Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
3.506 (1.393) -1.094 (.204)
.003 (.009) -.014 (.022)
.056 (.159)
-.102 (.137) -.264 (.224) .056 (.216)
-.151 (.181) .255 (.147) .019 (.142) .149 (.170) .251 (.159)
-.125 (.132)
-.523 .036
-.058
.045 -.085 -.164 .029
-.113 .193 .015 .116 .191
-.093
.014 .000*
.701
.534
.726
.456
.240
.797
.406
.086
.892
.383
.119
.347
4.908 (1.539) -1.028 (.224)
.009 (.010)
.009 (.029)
.200 (.181) -.172 (.152) -.007 (.246) -.213 (.236) -.033 (.201) -.017 (.161) .023 (.160) .113 (.187)
-.304 (.176) -.033 (.146)
-.461 .088 .030
.152
-.134 -.004 -.106 -.023 -.012 .017 .083
-.219 -.023
.002 .000*
.360
.764
.273
.258
.976
.369
.869
.917
.886
.547
.088
.823
3.513 (1.543) -.807 (.225) .008 (.010) .025 (.027)
.197 (.184)
-.182 (.152) -.158 (.244) -.266 (.236) -.217 (.206) .193 (.163) .284 (.158)
-.182 (.190) .071 (.178) .090 (.149)
-.360 .084 .093
.149
-.141 -.092 -.132 -.151 .136 .208
-.132 .050 .063
.025 .001*
.387
.369
.287
.235
.518
.263
.296
.241
.074
.340
.693
.547
Behavioural Disengagement
.278 .046 .093
Constant Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
-.025 (.177) -.017 (.026) .001 (.001)
-.003 (.003)
.057 (.020)
.023 (.017)
.023 (.028)
.023 (.027) -.024 (.023) -.033 (.019) -.009 (.018) .042 (.022)
-.031 (.020) -.039 (.017)
-.062 .110
-.100
.347
.147
.105
.090 -.138 -.191 -.051 .249
-.178 -.222
.886
.509
.219
.265
.006* .180 .430 .412 .289 .076 .635 .052 .130
.021*
.128 (.154) -.044 (.022) .002 (.001) .003 (.003)
.001 (.018) .001 (.015)
-.035 (.025) .017 (.024)
-.023 (.020) .010 (.016)
-.008 (.016) -.008 (.019) .008 (.018)
-.007 (.015)
-.211 .200 .109
.009 .005
-.219 .092
-.175 .077
-.060 -.063 .059
-.052
.409
.053
.054
.319
.953
.967
.154
.464
.249
.532
.633
.666
.666
.640
.075 (.124) -.001 (.018) -.001 (.001) .004 (.002)
.002 (.015) .020 (.012) .004 (.020) .008 (.019) .011 (.017)
-.012 (.013) -.023 (.013) -.022 (.015) -.005 (.014) .004 (.012)
-.007 -.145 .179
.021 .197 .033 .052 .102
-.108 -.214 -.102 -.049 .037
.547
.947
.153
.099
.884
.110
.822
.672
.497
.369
.076
.477
.712
.735
Venting .235 .254 .306 Constant
Participant role MEQ
-.915 (1.311) -.581 (.192) .002 (.008)
-.291 .021
.487 .003*
.819
1.226 (1.229) -.633 (.179) .005 (.008)
-.337 .061
.321 .001*
.503
1.855 (1.219) -.758 (.178) .012 (.008)
-.389 .134
.131 .000*
.131
175
PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
.021 (.021)
.499 (.150)
.100 (.129)
.263 (.210) -.021 (.203) .149 (.170)
-.133 (.138) -.009 (.134) .256 (.160) .058 (.150)
-.124 (.124)
.093
.423
.087
.171 -.012 .117
-.106 -.007 .209 .047
-.097
.314
.001* .441 .214 .918 .383 .337 .949 .113 .697 .321
.030 (.023)
.296 (.145)
.152 (.121) -.065 (.196) -.203 (.189) -.024 (.161) -.070 (.129) .131 (.128) .253 (.150)
-.057 (.141) .037 (.117)
.124
.268
.140 -.045 -.119 -.020 -.059 .114 .219
-.049 .031
.200
.044* .213 .741 .285 .881 .586 .308 .095 .685 .754
.025 (.022)
.279 (.145)
.155 (.120) -.192 (.193) -.272 (.187) -.062 (.163) -.062 (.129) .155 (.125) .076 (.150) .069 (.141) .100 (.118)
.109
.243
.138 -.128 -.155 -.050 -.050 .130 .064 .057 .080
.250
.057
.201
.320
.148
.702
.633
.217
.613
.624
.400 Positive Reframing
.127 .115 .178
Constant Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
1.531 (1.590) -.347 (.233) -.010 (.010) -.018 (.025)
-.124 (.182) .205 (.156)
-.316 (.255) -.094 (.247) .532 (.207) .097 (.168) .311 (.162)
-.174 (.194) -.101 (.182) .364 (.150)
-.153 -.098 -.069
-.092 .157
-.181 -.046 .367 .068 .224
-.125 -.072 .251
.338
.138
.318
.483
.498
.192
.218
.704 .012*
.562
.058
.372
.578 .018*
-.307 (1.515) -.164 (.220) -.004 (.009) -.002 (.028)
-.095 (.179) .316 (.149)
-.147 (.242) .030 (.233) .284 (.198) .030 (.159) .226 (.158)
-.002 (.185) -.016 (.174) .433 (.144)
-.077 -.038 -.007
-.076 .258
-.089 .015 .209 .022 .173
-.001 -.012 .319
.840
.459
.702
.950
.595 .037*
.545
.898
.155
.849
.156
.993
.927 .003*
-.806 (1.505) -.542 (.220) .000 (.009) .014 (.027)
.131 (.179) .153 (.148)
-.132 (.238) .063 (.231) .362 (.201) .030 (.159) .309 (.154)
-.096 (.185) .058 (.174) .322 (.146)
-.245 -.002 .052
.100 .120
-.078 .032 .255 .022 .228
-.071 .042 .228
.594 .015*
.982
.610
.467
.305
.579
.784
.075
.850 .048*
.605
.739 .029*
Planning .047 .000 .143 Constant
Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
.523 (1.482) -.315 (.217) .006 (.009) .016 (.024)
-.031 (.169) .242 (.146)
-.032 (.238) -.034 (.230) -.006 (.193) .071 (.156) .128 (.151)
-.156 .066 .068
-.026 .207
-.020 -.019 -.005 .056 .103
.725
.150
.520
.509
.854
.100
.894
.881
.974
.649
.401
.811 (1.603) -.157 (.233) .009 (.010) .037 (.030)
.096 (.189) .045 (.158) .286 (.256)
-.350 (.246) -.007 (.210) .111 (.168) .024 (.167)
-.074 .090 .138
.077 .037 .174
-.183 -.005 .082 .018
.614
.501
.395
.219
.612
.777
.267
.159
.972
.512
.887
-.017 (1.496) -.299 (.218) .014 (.009) .055 (.026)
.296 (.178)
-.006 (.147) .136 (.236)
-.360 (.229) -.003 (.200) .025 (.158) .168 (.153)
-.139 .143 .216
.233
-.005 .082
-.186 -.002 .019 .127
.991
.173
.147 .041*
.099 .968 .565 .119 .986 .874 .275
176
Assertiveness Intellect
Openness
-.111 (.181) .377 (.169) .090 (.140)
-.090 .298 .070
.540 .028*
.522
-.104 (.195) .093 (.184) .158 (.153)
-.080 .071 .117
.597
.612
.302
-.133 (.184) .377 (.173) .073 (.145)
-.101 .280 .053
.470 .032*
.617 Humour .073 .071 .154 Constant
Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
3.772 (1.673) -.526 (.245) -.007 (.010) .014 (.027)
-.221 (.191) .212 (.164) .431 (.269)
-.406 (.260) -.314 (.217) .146 (.177)
-.006 (.171) .072 (.204) .040 (.191)
-.188 (.158)
-.227 -.070 .051
-.161 .159 .242
-.194 -.212 .100
-.004 .050 .028
-.127
.027 .034*
.490
.611
.251
.201
.112
.121
.152
.410
.971
.726
.835
.238
2.152 (1.859) -.542 (.270) -.002 (.012) .024 (.035)
-.085 (.219) .247 (.183) .415 (.297)
-.297 (.285) -.318 (.243) .055 (.295) .074 (.194) .180 (.226) .062 (.213)
-.177 (.177)
-.213 -.015 .074
-.056 .169 .210
-.129 -.195 .034 .048 .115 .039
-.109
.250 .048*
.881
.494
.700
.180
.166
.301
.194
.776
.702
.429
.770
.320
-.125 (1.508) -.392 (.220) .005 (.009) .029 (.027)
.110 (.179) .186 (.149) .358 (.238)
-.213 (.231) -.330 (.202) .146 (.160) .140 (.154)
-.011 (.185) .385 (.174)
-.136 (.146)
-.180 .048 .113
.085 .148 .213
-.108 -.236 .106 .105
-.008 .282
-.097
.934
.078
.621
.278
.541
.214
.136
.360
.105
.362
.365
.955 .030*
.355
Acceptance .079 -0.059 .002 Constant
Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
.412 (.283) .014 (.041) .000 (.002) .001 (.005)
.086 (.034)
-.071 (.028) .028 (.045)
-.007 (.043) -.052 (.038) -.040 (.030) -.011 (.029) .053 (.035)
-.026 (.033) -.027 (.027)
.036 .013 .014
.370
-.313 .093
-.020 -.206 -.161 -.044 .218
-.106 -.109
.149
.736
.896
.897
.013*
.013* .532 .871 .173 .187 .716 .134 .429 .319
.231 (.288) .029 (.042)
-.002 (.002) .005 (.005)
-.005 (.034) .015 (.028)
-.052 (.046) .029 (.044)
-.017 (.038) .009 (.030) .023 (.030)
-.016 (.035) .023 (.033)
-.011 (.027)
.078
-.125 .104
-.025 .070
-.181 .087
-.072 .040 .101
-.069 .098
-.047
.425
.491
.251
.364
.873
.601
.263
.512
.650
.755
.448
.654
.493
.689
.657 (.297) -.026 (.043) -.003 (.002) .000 (.005)
-.014 (.035) -.003 (.029) -.063 (.047) -.011 (.046) -.018 (.040) -.020 (.031) .060 (.030)
-.014 (.036) -.031 (.034) .033 (.029)
-.065 -.155 .004
-.060 -.014 -.207 -.030 -.072 -.080 .248
-.058 -.123 .131
.029
.554
.144
.975
.694
.914
.182
.813
.649
.525
.050
.700
.376
.250
Religion .121 .144 .131 Constant
Participant role MEQ PSQI BFAS
Withdrawal
-.865 (2.171) -.597 (.317) .027 (.013)
-.027 (.035)
.226 (.248)
-.194 .199
-.076
.124
.691
.063 .045*
.439
.365
-.064 (2.120) -.767 (.308) .025 (.013)
-.022 (.040)
.092 (.250)
-.254 .188
-.058
.052
.976 .015*
.056
.576
.713
-.922 (2.032) -.556 (.296) .024 (.013)
-.022 (.036)
.402 (.242)
-.191 .184
-.065
.234
.651
.064
.064
.539
.100
177
Volatility Compassion
Politeness Industriousness
Orderliness Enthusiasm
Assertiveness Intellect
Openness
.017 (.213) -.091 (.348) -.238 (.337) .393 (.282)
-.105 (.229) .097 (.222) .537 (.264)
-.393 (.248) .404 (.205)
.010 -.038 -.086 .199
-.054 .052 .284
-.204 .205
.937
.793
.480
.167
.648
.661 .045*
.116
.052
.181 (.209) -.114 (.339) -.251 (.325) .587 (.277)
-.079 (.222) -.132 (.221) .475 (.258)
-.459 (.243) .402 (.202)
.104 -.048 -.092 .303
-.041 -.071 .255
-.243 .208
.388
.738
.443 .037*
.723
.552
.069
.062 .049*
-.061 (.200) .019 (.321)
-.369 (.311) .580 (.272)
-.035 (.215) -.154 (.208) .461 (.250)
-.372 (.235) .438 (.197)
-.036 .009
-.141 .312
-.019 -.087 .259
-.205 .236
.762
.953
.238 .035*
.871
.460
.068
.116 .028*
Self-Blame .211 .087 .142 Constant
Participant role MEQ PSQI BFAS
Withdrawal Volatility
Compassion Politeness
Industriousness Orderliness Enthusiasm
Assertiveness Intellect
Openness
.628 (.204) .043 (.030) .000 (.001)
-.001 (.003)
-.052 (.023) -.004 (.020) .002 (.033)
-.019 (.032) .068 (.026)
-.012 (.022) -.021 (.021) -.004 (.025) -.001 (.023) -.005 (.019)
.140
-.018 -.020
-.286 -.022 -.007 -.068 .349
-.064 -.115 -.022 -.003 -.018
.003
.154
.843
.833
.029* .850 .960 .554
.011* .566 .305 .871 .980 .779
.116 (.264) -.028 (.038) .000 (.002)
-.007 (.005)
.062 (.031) -.002 (.026) -.013 (.042) -.006 (.040) -.046 (.034) -.003 (.028) .021 (.027)
-.009 (.032) .006 (.030) .033 (.025)
-.078 .005
-.142
.287 -.010 -.044 -.018 -.199 -.014 .094
-.040 .025 .141
.660
.460
.964
.184
.050
.938
.767
.885
.182
.909
.447
.783
.850
.194
.122 (.238) .004 (.035)
-.001 (.001) .007 (.004)
.027 (.028)
-.018 (.023) .001 (.038)
-.003 (.036) -.080 (.032) .006 (.025) .016 (.024)
-.014 (.029) .012 (.027) .049 (.023)
.011
-.044 .167
.133
-.093 .002
-.010 -.364 .030 .077
-.069 .057 .224
.607
.915
.656
.113
.345
.435
.986
.931 .014*
.797
.510
.623
.657 .036*
Note. * p < .05; B = unstandardized regression coefficient, SEB = standard error of the coefficient; ß = standardized coefficient; Participant role: patient = 0, caregiver = 1
178
Pearson correlations were run between UTIME coping log scores at baseline, midpoint and
endpoint, and the respective Brief COPE scores, to see whether any particular coping behaviours
were associated with how patients’ coping changes across the day. Of the 14 correlations at each
time point, only substance use at endpoint showed a significant correlation to one’s coping
across the day (r = .352, p = .041). None of the other correlations were significant. Substance use
did not show any significant correlation to coping score across the day at baseline (r = -.013, p =
.948), or midpoint (r = .015, p = .938).
6.5! Discussion
Various studies have individually investigated chronotype, sleep quality, coping and personality
among cancer patients (e.g., Fiorentino & Ancoli-Israel, 2007; Papantoniou et al, 2015;
Porcerelli, Bornstein, Porcerelli & Arterbery, 2015; Zaza, Sellick & Hillier, 2005). These topics
have also been studied in numerous combinations among the general population (e.g., Afshar et
al., 2015; Vitale et al., 2015). We assessed coping two ways: based on one’s personal assessment
of how well they were coping, and based on one’s use of various behaviours as mechanisms for
coping with the situation. Chronotype and sleep quality were compared to self-rated coping, and
to coping behaviours, to understand how one’s rhythm and quality of sleep relate to and predict
coping. Personality was also examined in relation to one’s coping behaviours to further identify
people likely to engage in various coping behaviours over the course of chemotherapy treatment.
6.5.1! Personal Coping Assessment
On self-reported coping logs, patients consistently reported decreased coping ability as the day
progressed. However, on treatment days, patients reported increased coping ability at midpoint
and endpoint versus baseline. Increased self-reported coping ability may be the result of various
factors, including learning what to expect (i.e., from the treatment itself, side-effects, personal
recovery time, etc.), and realizing that one is better prepared to handle the situation than
originally thought.
Comparing coping in the days pre- and post- chemotherapy, patients reported poorer coping in
the days following chemotherapy, suggesting people are likely feeling an onset of negative
179
treatment related side effects in this time. Conversely, in the days prior to treatment, most
patients have likely recovered from the negative symptoms resulting from their previous
chemotherapy infusion. Anecdotally, in the days prior to a chemotherapy treatment session,
some patients report being happy that they are one step closer to completing the entire treatment
process, which may also influence more positive coping ratings. This result does not indicate that
patients actually cope more or less pre- or post- treatment, but rather there is a difference in the
extent to which they feel they are coping.
Plotting coping scores graphically indicates that patients reach their peak coping times earlier in
the day on treatment days and the five days post treatment (prior to midday), and later in the day
(early to midafternoon) in the five days prior to treatment. Treatment days and the five days post
chemotherapy are likely more demanding on patients (i.e., spending several hours out of the
house, waiting in the hospital, and facing side effects), compared to the five days pre-
chemotherapy when many patients are able to follow a more regular routine, without large
blocks of time spent in a hospital and or side effects. Prior to chemotherapy, having less external
stressors and being in better health may allow patients to function longer at slightly higher levels
before feeling they are less able to cope, resulting in a later peak coping time.
Significant positive correlations between patients’ peak coping scores across treatment and their
MEQ scores indicates that changes in coping were associated with one’s chronotype, as
predicted at the study’s outset; earlier chronotypes reached their peak coping times before later
(less morning preference) individuals. While significance for correlations between MEQ score
and UTIME coping at baseline and endpoint was not achieved, the correlation was approaching
significance; future studies with larger or more normally distributed groups across the
chronotype spectrum may achieve a significant correlation. Pre-chemotherapy correlations were
significant, suggesting that when patients have typically recovered from most of their previous
treatment’s side effects and are preparing for their next chemotherapy session, their coping
across the day is reflects their chronotype. The lack of significant correlation between MEQ
score and peak coping at midpoint and in the days post midpoint suggests people’s coping is not
reflective of their chronotype at this time. Several factors may influence this outcome. For
example, many patients, reaching their treatment midpoint may lead people to wonder whether
their treatment is working, and or they may feel fatigued and realize there is still another half of
their treatment course ahead which may be daunting. For several patients their chemotherapy
180
drug regiment changed at midpoint, possibly leaving many people worrying about new side
effects, or with many new thoughts and emotions in their mind which may have affected their
coping across the day. These thoughts and emotions may override the effects of chronotype and
cognitive functioning on coping ability.
Lack of many significant correlations between sleep quality and peak daily coping score reflects
the same trend observed in Chapter 5, where recalled coping ability and feelings across the day
did not correlate significantly with sleep quality. The associations between sleep quality and
peak coping were not uniformly positive or negative, unlike the consistently positive association
between MEQ and peak coping scores. Chronotype and sleep quality were not related in this
population, unlike in the general population (Vitale et al., 2015). The sole significant correlation
between sleep quality and peak coping in the five days post chemotherapy at midpoint is likely a
false positive (Type 1 error). The lack of correlation between peak coping and sleep quality
suggests sleep does not influence patients’ ratings of their own ability to cope in this population.
This is important to consider since poor sleep quality is associated with reduced cognitive
function, which has been linked to less positive or adaptive coping (Byun, Gay & Lee, 2016;
Goretti, Portaccio, Zipoli, Razzolini & Amato, 2010). In our society, sleep debt is often
overlooked as many people try to schedule as many activities in their day, and many people
mistakenly believe they can easily recover from sleep debt (Cohen et al., 2010). Therefore,
patients likely continue to disregard the effects of their sleep debt on their coping ability.
6.5.2! Coping Behaviours: Changes in Coping Scores
Patients and caregivers completed the Brief COPE at baseline, midpoint, and endpoint. Among
patients, use of the majority of coping behaviours changed across treatment. Conversely,
caregivers use of various coping strategies across treatment remained relatively consistent. Of
the fourteen Brief COPE behaviours, with the exception of self-distraction, patients and
caregivers reported that their most often used coping strategies were adaptive (i.e., acceptance,
emotional support, active coping, planning, instrumental support, positive reframing, religion,
and humour). The remaining coping strategies were maladaptive or less solution based. While
patients and caregivers typically reported engaging in more proactive coping behaviours, they
did not engage in the same behaviours at the same rate. Except for acceptance, the coping
behaviours that demonstrated significant changes in use across treatment typically displayed
181
decreased use. This reflects previously reported results that found many coping reactions are
more prominent early on when dealing with a crisis compared to later (Carver et al., 1993).
Self-distraction was the only coping behaviour predicted by chronotype and sleep quality,
regardless of personality trait inclusion in multiple regression. This relationship is modeled in
part in other coping behaviours, however there is a clear shift in predictive influence for self-
distraction across treatment. Self-distraction has been classified as engagement coping (focuses
on dealing with the stressor or distressing emotions) and accommodative coping (aims to adapt
to effects of the stressor) (Carver & Connor-Smith, 2010). While literature on self-distraction
and chronotype is sparse, some research suggests evening types (i.e., lower MEQ score) show
less emotionally adaptive coping (Ottoni, Antoniolli & Lara, 2012), suggesting a similar
relationship to that observed between chronotype and self-distraction. While literature on sleep
quality and coping styles is also sparse, research suggests maladaptive behaviours increase with
poor sleep (Lenjavi, Ahuja, Touchette & Sandman, 2010), which indicates an opposite
relationship to that found between sleep quality and self-distraction in the multiple regression
models. One possible explanation might be that poor sleep quality results in greater self-
distraction as individuals utilize greater self-distractive thought and behaviours when unable to
sleep. The shift in predictive value across treatment from chronotype to sleep quality is
important, as it highlights that at the outset of treatment, many people are likely still following a
schedule based on their chronotypic demands. However, as treatment progresses and fatigue (a
commonly reported feeling by cancer patients and their caregivers) increases, people may be
sleeping at odd hours, not necessarily reflective of their chronotype. In which case, their ability
to perform becomes dependent on the sleep they have obtained, instead of an innate rhythm.
Patients and caregivers most reported using acceptance across treatment. Both groups reported
consistent increases in acceptance as treatment progressed. Patients may demonstrate greater
acceptance as treatment progresses, as side effects are experienced, and the diagnosis and
treatment become more of a reality. Interestingly, acceptance was not significantly predicted by
either of the multiple regression models at any time across treatment. While the multiple
regression model xincluding personality was not significant at baseline, the predictive roles of
withdrawal and volatility did reach significance; these are the two opposite components of
neuroticism. Volatility refers to high emotional reactivity, and difficulty controlling emotional
impulses, and withdrawal encompasses traits of negative affect, such as anxiety and depression.
182
It is believed that these are part of a joint system governing sensitivity to threat and punishment
(DeYoung, Quilty & Peterson, 2007). Neuroticism has been negatively correlated to acceptance
in previous research (Afshar et al., 2015). While heightened withdrawal and decreased volatility
contributed to increased acceptance at baseline, at midpoint and endpoint, withdrawal and
volatility negatively contributed to greater acceptance. It is possible that at baseline, while low
volatility may have helped someone accept the diagnosis, increased withdrawal may have
contributed to greater acceptance by reducing the amount of questioning a person did, instead
making them submit to the situation. Future research with larger populations should reexamine
this relationship to assess whether the previously documented inverse relationship between
neuroticism and acceptance exists across treatment with withdrawal and volatility.
While patients’ acceptance levels increased, denial decreased across treatment. The increase in
acceptance appears to plateau after midpoint, yet denial continues to decrease across treatment,
suggesting patients continued coming to terms with their cancer diagnosis. Kreitler (1999) found
that cancer patients reporting denial in the early phases of coping following diagnosis seemed to
report reduced anxiety compared to those with lower denial during this period. Furthermore,
Kreitler (1999) found that patterns of denial tended to be highest earlier in treatment, while
longitudinal studies showed that denial decreased as time passed for a period of up to two years
post treatment. This decreasing pattern is reflected in our own research. Future research in the
same population should consider whether denial is directed towards one’s emotions or the
information relating to the diagnosis given that emotional denial may provide some positive
benefit while information denial may lead to a worse prognosis or delays in treatment. Multiple
regression models for denial were only predictive when personality was included, then only at
baseline and midpoint. While the multiple regression model was not significant across treatment
when personality was not included, participant role appeared as a significant predictor at baseline
only, in models with and without personality, suggesting at this time, patients show greater
denial at baseline than caregivers. Lower openness at baseline contributed to increased denial,
reflecting previously reported results (Carver & Connor-Smith, 2010), while higher withdrawal
at midpoint also predicted increased denial.
Patients sought emotional support as the second most used coping strategy at baseline, and the
second highest at endpoint, but at a significantly reduced level than at baseline. Good emotional
support protects against poor emotional and mental health (Clough-Gorr, Ganz and Silliman,
183
2007). As treatment progressed, patients may have sought less emotional support for many
reasons, including having learned to deal with some of the stressors they faced, or feeling that
after many months, they could no longer turn to others to continuing seeking support. By
endpoint, patients may have required less emotional support as their acceptance for the situation
generally increased. Instrumental support was sought at a lesser extent than emotional support,
and its use decreased across treatment. The multiple regression models predicting emotional and
instrumental support were significant at all time points across treatment, with and without the
inclusion of personality. For both types of support, in multiple regression analyses with and
without personality, being a patient consistently predicted seeking greater emotional support
across treatment. When including personality in the model, only intellect significantly predicted
emotional support at midpoint; lower intellect predicted greater need for emotional support.
Individuals higher in intellect may have greater emotional intelligence (i.e., ability to recognize
and analyze emotional information and conduct general problem solving), and better understand
and process their own emotions instead of seeking emotional support from others; this reflects
previous work which suggests high emotional intelligence is negatively related to perceptions of
acute and chronic stress (Singh & Sharma, 2012). The lack of predictive value of sleep quality
and chronotype, along with personality on emotional and instrumental support suggest other
factors not explored here contribute to the use of these coping behaviours.
Caregivers reported fairly consistent use of various coping behaviours across treatment. Given
the small group of caregivers in this study, changes may not be as easily observed as in a large
population. Alternately, if caregivers are consistent in their coping strategies across treatment, it
may be due to their stressor remaining relatively consistent: dealing consistently with a partner
facing a chronic health condition; conversely, patients deal with many changes in their
wellbeing, their acceptance of the diagnosis, and their ability to handle daily life. Caregivers only
reported a significant decrease in positive reframing, from midpoint to endpoint; patients
reported no changes in positive reframing across treatment. Since positive reframing is typically
aimed at managing distressful emotions rather than dealing with the stressor itself, the overall
decline in caregiver use of this coping strategy may be due to learning to handle the stressful
situation (i.e., partner’s cancer diagnosis and treatment) as time progresses. The slight non-
significant increase in positive reframing from baseline to midpoint, may result from caregivers
feeling unable to help their partner as treatment progresses, and therefore using this behaviour to
184
deal with their emotions and cope with the situation. As chemotherapy treatment nears
completion and the stressor diminishes, caregivers may be less dependent on positive reframing.
Future research should examine this by assessing changes in caregiver ratings of the stressor.
While caregivers reported significant changes in positive reframing, patients displayed higher
levels overall, and by endpoint, multiple regression models, with and without personality,
demonstrated this difference: participant role predicted a significant difference in endpoint use of
positive reframing. Chronotype and sleep quality were not significant contributors. Across
treatment, high openness significantly predicted greater use of positive reframing, likely because
it allowed individuals to consider the possibility of ‘finding the silver lining’ in an otherwise
negative situation. At baseline, industriousness increased positive reframing, and while it was not
significantly predictive at midpoint or endpoint, it hovered near significance. Volatility only
significantly predicted positive reframing at midpoint, suggesting that this association should be
reexamined in future studies to verify its validity. Enthusiasm was only significantly predictive at
endpoint, however it also hovered near significance at baseline. Enthusiasm is a component of
extraversion, which has been positively related to positive reframing (Afshar et al., 2015). Future
studies should reexamine the predictive role of industriousness and enthusiasm, as well as
volatility, to understand their predictive value on positive reframing and identify individuals at
risk for low use of this coping strategy.
Overall, patients reported using more planning behaviours (i.e., understanding and coming up
with strategies on how to deal with the problem) to cope than caregivers. While caregivers use of
planning was fairly consistent across treatment, patients reported significantly reduced midpoint
and endpoint planning compared to baseline. Planning is an active and adaptive coping strategy
(Burker, Evon, Loiselle, Finkel & Mill, 2005; Carver, Scheier & Weintraub, 1989). Planning
behaviours may decrease as time passes once patients establish steps and patterns to face their
illness (e.g., begun treatment, adopted strategies to overcome stress, progression toward recovery
is occurring, etc.). Multiple regression models predicting planning were only significant at
endpoint. With and without personality, endpoint sleep quality significantly predicted planning
such that worse sleep quality (i.e., higher sleep scores) predicted greater planning. It is possible
that planning behaviours occur with sleeplessness, as people are awake, and may ruminate over
possible strategies to overcome their stressor. As treatment progresses, the significance of sleep
quality as a predictor increases. Future studies should assess whether ruminating about planning
185
strategies interferes with sleep quality, or whether poor sleep quality leads to greater planning.
Chronotype did not have a predictive role for planning behaviour across treatment. Intellect
predicted planning behaviour at baseline and endpoint, not midpoint. While intellect appeared to
predict use of planning at baseline, the model overall was not significant, suggesting the model
prevented against Type I error (i.e., incorrect rejection of a null hypothesis), yet the role of
intellect should not be disregarded, especially due to its significant predictive value at endpoint
when the model is significant. Retesting these findings in a larger population may confirm
whether intellect plays a significant predictive role across treatment among cancer patients and
caregivers for their use of planning behaviour. Previous studies have suggested that intellect is a
predictive trait for one’s use of planning as a coping behaviour (e.g., Lawson, Bundy, Belcher &
Harvey, 2013), therefore it is important to reconsider the results of this study in a larger
population experiencing the same stresses. This would clarify whether intellect is a strong
general predictor across the population of planning behaviours when dealing with a stressor, and
if there are specific times in treatment or whether across treatment intellect predicts planning.
Active coping was the second most engaged in behaviour by caregivers, and the third most used
patient coping behaviour, yet patients reported more active coping than caregivers. Neither
patients nor caregivers demonstrated significant changes in active coping across treatment.
Patients may report greater active coping than caregivers because they are facing the actual stress
of the disease itself, whereas caregivers face the stress of their partner being sick. Dealing with
the disease itself and faced with one’s own mortality daily may make an individual more likely
to engage in actions to alleviate the associated stresses. While caregivers report engaging in
active coping, their rate of engagement may be lower if they feel they cannot do anything to treat
the disease itself or the effects of the disease on their partner. When not including personality,
the multiple regression model for active coping was only significant at endpoint, and was
predicted by sleep quality (poorer sleep quality increases active coping). When including
personality, the multiple regression model was only significant at baseline, suggesting that higher
orderliness predicted greater active coping. Midpoint multiple regression models with and
without personality were non-significant, however participant role appeared as the only
significant predictor, suggesting participant role was related to active coping use. The model
again appears to be preventing against Type I error, which may have resulted due to having
multiple non-significant predictors included. However, all predictors were included in the
186
standard multiple regression in an attempt to model how all factors interacted together to create
an influence in real life.
Patients and caregivers remained relatively consistent in their use of humour and religion to help
cope across treatment, with no significant changes in their use. Previous research with cancer
patients identified humour as a valuable therapeutic tool for decreasing anxiety and discomfort,
while laughter specifically has been noted as a healthy method to relieve stress, let patients feel a
sense of control, and aide in body relaxation (e.g., Christie & Moore, 2005). Humour may have
been used less among caregivers than patients because while patients can make light of their own
poor health, other individuals, including caregivers, may see it as inappropriate to make light of
the patient’s ill health. In multiple regression models, when not considering personality,
participant role was the only significant factor predicting one’s use of humour, while chronotype
and sleep quality were not significantly predictive. When personality was included, the only
significant multiple regression model was at endpoint, where only intellect significantly
positively predicted one’s use of humour; participant role was close to but not statistically
significant. However, at baseline and endpoint, when considering personality, while neither
model was significant, participant role contributed to one’s use of humour. Future studies with
larger population groups and a more equal distribution of patients and caregivers must reexamine
this difference to verify whether participant role actually does contribute to a significant
predictive model across treatment.
Female patients also used religion to cope more than male caregivers, reflecting findings in the
general population which suggest that across the age span, women consistently engage in more
religious practices and have more spiritual beliefs than men (Jones et al., 2011). As a coping tool
in other chronic illnesses and conditions, women have reported greater use of religion to cope
than men (e.g., Dunn & Horgas, 2004). It is important to note that the religious affiliation of
people in this study was not surveyed, therefore it is unclear if people were predominantly of one
particular religious background, or held a variety of religious beliefs. Some studies suggest that
in the Christian faith, women are more religious than men, but that among other religions such as
Judaism or Hinduism, women are less involved in the religion (Loewenthal, McLeod &
Cinnirella, 2000). Future research should examine participants’ faith to understand if women do
rely on more religion-based coping than men, or if this is specific to women of a particular faith.
When personality was not included in multiple regression, participant role predicted significant
187
differences in one’s use of religion to cope at baseline and midpoint, while chronotype was
significantly predictive across treatment. When personality was included, chronotype was only
significant at baseline, and approached significance at midpoint and endpoint, while participant
role was only significant at midpoint and approached significance at baseline and endpoint.
Personality factors played significant predictive roles in the use of religion across treatment as
well: industriousness and openness significantly positively predicted use of religious coping at
midpoint and endpoint, while only openness approached significance at baseline; assertiveness
positively predicted religious coping at baseline and neared significance at midpoint and
endpoint. These findings support those of previous studies suggesting a positive relationship
between religious beliefs and openness, industriousness – which is part of conscientiousness, and
assertiveness – which is part of extraversion (Khoynezhad, Rajaei & Sarvarazemy, 2012;
Saroglou, 2002) Future studies with larger populations must reexamine the predictive influence
of personality and participant role on one’s use of religion as a means to cope, particularly to
verify whether the variables that neared significance are actually significant across treatment.
Patients and caregivers demonstrated relatively consistent levels of venting across treatment,
with patients reporting greater venting behaviour than caregivers. Venting can be positive –
allowing one to discuss their feelings or play out their thoughts and devise a solution for a
problem, or negative – allowing one dwell on the negative aspects of the situation and impede
actively coping with the stressor (Carver, Sheier & Weintraub, 1989). Future research on venting
and what causes it should also examine if people are using it as a positive or negative behaviour,
to devise appropriate strategies to address the individual’s concerns. Multiple regression models
for venting were significant across treatment, with and without personality. Participant role was a
consistent predictor of venting, supporting previous research that suggests women engage in
venting more than men (Sullivan, 2002). Without personality in the multiple regression, sleep
quality was a significant predictor at endpoint, such that poorer sleep predicted greater venting;
at baseline and midpoint sleep quality was not significant but was not far off. However, when
personality was included in the model, sleep quality was no longer a significant predictor.
Withdrawal significantly contributed to baseline and midpoint venting, interestingly suggesting
that the more withdrawn an individual is, the greater their venting. Withdrawal became a less
significant predictor as treatment progressed, and was no longer significant at endpoint. While it
might seem that withdrawal would be negatively predictive of venting since a person high in
188
withdrawal would be more apt to keep to themselves, these results support previous research that
found individuals high in neuroticism (of which withdrawal is a component) engage in greater
venting of emotions (Vollrath, Torgersen & Alnaes, 1998). Chronotype was not a significant
predictor in either model of venting across treatment.
Across treatment, patients and caregivers consistently reported low behavioural disengagement
as a coping strategy. Only at baseline did personality traits of withdrawal and openness
significantly predict behavioural disengagement levels. While the model appeared to be
significant at endpoint when considering personality, no predictors were significant at p < .05,
suggesting that some of the almost significant variables (sleep quality, p = .099; enthusiasm, p =
076) allow the model to appear significant. When personality was not included, sleep quality also
neared significance. It is possible that with all variables in the model, the true effect of sleep
quality is not seen; future research should reassess the relationship between sleep quality and
behavioural disengagement.
Substance use was reported to be low among patients and caregivers. Substance use is
considered a maladaptive coping strategy (Carver, Scheier & Weintraub, 1989). Literature on
drug and alcohol use suggests women typically abuse drugs and alcohol at a lesser rate than men,
yet the number of women using these substances is on the rise (Becker & Hu, 2008). Among
caregivers, research indicates that those who experience higher social and emotional burdens as a
result of caregiving are at increased risk for overconsumption of alcohol (Rospenda, Minich,
Milner & Richman, 2010). Individuals displaying risk factors for demonstrating such
maladaptive coping strategies should be monitored. While only participant role was a significant
predictor when not considering personality, once personality was added to the multiple
regression model, chronotype also became a significant predictor of substance use at baseline
and midpoint (i.e., increased tendency for morningness predicted increased substance use, and
vice versa). This is interesting given that substance use is typically more commonly reported
among individuals reporting later/evening chronotypes, even independently of delayed sleep
phase disorder (Reid et al., 2012; Tavernier, Munroe & Willoughby, 2015). It may be that this
difference is a result of engaging in substance use recreationally versus to cope with stress (with
participants engaging in substance use to cope with stress). Industriousness also predicted
substance use (i.e., higher industriousness predicted decreased substance use, and vice versa), as
189
seen in previous research (Turiano et al., 2012). Poor sleep quality is also known to relate to
substance use (e.g., Wong, Roberson & Dyson, 2015), however that was not observed here.
Patients’ self-blame decreased across treatment, suggesting patients came to realize or accept
that their cancer diagnosis was not their fault. Denial scores were significantly positively
correlated to self-blame scores across treatment (data not shown). While it is unknown whether
denial was directed at information or emotions in this study, this correlation suggests that while
patients may have denied their feelings and or the information being received, an element of self-
blame still existed. As denial diminished, it seems patients felt less blame. Interestingly, no
significant correlation between self-blame and acceptance exists, but overall the decrease in
patient self-blame suggests subjects accepted that the diagnosis and or resulting stress was not
their fault. When not considering personality, sleep quality was associated with self-blame at
endpoint. It is possible that self-blame led to poor sleep, or sleeplessness led to rumination and
self-blame. When personality was included in the multiple regression model, this relationship no
longer appeared significant, however the significance of sleep quality over the course of
treatment did increase in value. Industriousness was significantly predictive at baseline and
endpoint, however the relationship changed directions. The negative endpoint relationship
reflects previous findings (e.g., O’Brien & DeLongis, 1996). At baseline, it is possible that
before fully processing the situation, individuals high in industriousness wondered how when
they are typically organized, diligent people, they could have developed cancer, resulting in a
positive relationship with self-blame.
6.5.3! In the Moment Coping vs. Coping Behaviours
Daily coping logs did not correlate with one’s use of coping behaviours. This suggests that while
patients report having a memory for their use of various coping behaviours (indicated by their
completion of the Brief COPE), they did not rely on their memory for use of these behaviours
when rating how they felt they were coping in the moment. Given that chronotype does relate to
one’s rating of their in the moment coping, there is some element of rhythmicity related to one’s
sense of their coping ability across the day. While further research is needed to validate this, it is
possible that one’s in the moment rating is based on how they are feeling depending on their
cognitive and emotional regulatory abilities. Comparing one’s in the moment coping ratings
across the day to their performance on an emotional regulatory task across the day could shed
190
light on whether one’s rating of their coping ability is in fact tied to one’s emotional and
cognitive control capacities across the day. Conversely, one’s recall for use of various coping
behaviours may be based on salient memories for behaviours that an individual recalls using. To
verify whether one’s memory for use of particular coping behaviours is correct, it would be
important to have the patient, and possibly an additional rater (such as a family member who is
frequently with the patient) keep a record for one’s use of various behaviours. One’s memory for
the use of various coping behaviours may vary from a patients’ actual use of particular
behaviours. If this is correct, it is also possible that the gaps in one’s memory for their use of
particular coping behaviours might explain why certain coping behaviours showed varied
relationships with chronotype, sleep quality and personality across treatment, due to poor
memory for the use of particular behaviours.
6.5.4! Limitations
While many patients reported having a partner, many partners did not attend appointments and or
chemotherapy treatments regularly or even at all. Of the partners who did attend, many were not
interested in participating in studies, often remarking that they were “not the sick one”. It proved
to be more difficult than expected to recruit caregivers for study participation. It is possible that
the commonly echoed comment “I’m not the sick one” suggests caregivers did not want to incur
attention and take away from their partner. Alternately, it is possible that caregivers often felt
they were already too overwhelmed with the responsibilities of their caregiving goal to take on
another commitment (i.e., study participation). In future, it will be important to recruit a greater
sample size of caregivers to better examine coping (both one’s personal reflection of their own
coping abilities and one’s use of various behaviours to cope) and how it relates to chronotype
and sleep quality.
Participants were divided based on their role – patient or caregiver. Given the nature of breast
cancer, it is most common to recruit female participants. A single male participant was initially
recruited but unable to participate as his disease had progressed to stage 4. Future studies need to
examine coping in both male and female patients and caregivers. This will provide a more clear
understanding of how chronotype and sleep quality, together with one’s personality, influence
the way different genders cope with their role of patient or caregiver. This will allow for a better
191
understanding of who is at risk of developing maladaptive coping strategies and to look for
solutions to prevent this from happening.
In this study, participants rated their own daily coping, use of various coping strategies, and
personality. We recognized that study participants may be biased insofar as they believe that
their answers are socially desirable to those assessing their responses (social desirability bias). In
future, it would be beneficial to obtain these same ratings from an independent observer (e.g.,
family member, friend). In addition, previous research has suggested individuals are not
particularly accurate at rating their emotional behaviour fluctuations across the day (Bellicoso,
2010); an independent observation in some cases, particularly when fluctuations in one’s
emotional abilities run counter to one’s image or sense or of self, might provide a more realistic
description. These discrepancies between how one views themselves versus their actual outward
actions can also be seen for personality. For example, a person might see themselves in a more
favourable light than others around them might, thereby biasing a personality score, and vice
versa. Alternately, an individual might see themselves as consistently emotionally stable across
the day, whereas their outward responses in reaction to experiences and individuals throughout
the day might be different. It would help to compare participants’ responses to how other’s view
their coping and personality. While one might argue that in having caregiver reports on the
patient and the patient’s self-reports it is difficult to identify the ‘truth’, these two ratings could
be beneficial in several ways. For example, if patient and caregiver both provide very different
coping ratings, it is important to understand what each individual is assessing when providing a
rating; it is possible that each individual is focusing on specific, yet different, aspects of
behaviour. Having self-ratings and independent ratings would allow a more detailed picture of a
participant’s coping and personality. Furthermore, understanding caregivers’ perception of how
their patient is coping may shed light on why caregivers believe they must act in certain ways in
their own role.
In multiple regression models, to avoid masking the influence of certain variables, all variables
were included in standard multiple regression analyses. There, some models appear non-
significant while components of the model still appear as statistically significant predictors.
When several variables are included in a model to attempt to predict an outcome variable, it is
possible that too many predictors are contributing to the outcome, therefore diluting the
importance of individual variables. In this case, the model itself is not significant, which prevents
192
against Type 1 error (which would cause a false positive, i.e., incorrect rejection of a null
hypothesis). It is still important to make note of those variables that are significant predictors, as
in smaller models they may play a significant role. Alternately, for behavioural disengagement at
endpoint when considering personality, the model appears significant, but there are no significant
predictors. This is possible when predictor variables are highly correlated (multicollinearity).
While the tolerance values for this model were all above .100, they were still lower than
tolerance values found for other models in this study. It is possible that while tolerance values
were not below an acceptable level, the fact that they were low combined with having several
predictors in the model diluted the overall significance any one predictor could hold on its own.
It should be noted that tolerance values for all three models predicting behavioural
disengagement across treatment, while not below .100, were all low.
6.6! Conclusion
Chronotype, sleep quality, and participant role each contribute in some degree to the use of
various coping behaviours across treatment. Most interesting is that chronotype and sleep quality
did not simultaneously influence coping behaviours at the three time points across treatment.
While chronotype and sleep quality do not contribute heavily to the use of all coping strategies,
there appears to be a clear pattern that chronotype affects use of particular coping strategies at
baseline, while sleep quality distinctly only significantly contributes to various coping
behaviours at endpoint. Chronotype did not significantly influence very many behaviours, yet its
consistent role predicting self-distraction at baseline, regardless of one’s personality, should be
noted. It also had a relatively consistent influence on the use of religious coping. Self-distraction
is considered an engagement or accommodative type coping strategy; religion is not necessarily
an engagement or disengagement strategy, and instead can be seen in either way, depending if
someone turns to their faith to seek strength to face the problem, or relies on their faith to ignore
the issue. Coping behaviours relating to sleep are engagement type strategies (Carver & Connor-
Smith, 2010). It is possible that either people are thinking how to best deal with the stressor,
which keeps them awake resulting in poor sleep, or alternatively, that people are unable to sleep
and instead spend their time ruminating and partaking in these engagement style coping
behaviours.
193
While various personality factors contributed to using particular coping behaviours, openness
had the most consistent predictive value, positively relating to engagement based behaviours, and
negatively contributing to disengagement style behaviours. Industriousness also significantly
influenced coping, but less consistently than openness. It followed a similar pattern for
predicting engagement and disengagement coping as openness. It is possible that including
personality did not allow the roles of chronotype and sleep quality to be significant. Future
studies should only include significant predictors in the model to reexamine the relationship.
The significant relationship between one’s chronotype and their self-rating of coping ability
across the day suggests people were aware of fluctuations in their ability across the day, and may
have accepted these changes as part of their overall preference for performing tasks at one time
or another. The lack of correlation between sleep quality and coping log scores across treatment
suggests patients did not consider their sleep debt. It is possible that patients may have accepted
that a sleep debt would be experienced during cancer treatment, and therefore did not factor in its
outputs, (e.g., fatigue) when rating their coping. This is particularly interesting considering the
distinction between chronotype contributing to coping behaviours earlier in treatment, and sleep
quality contributing to use of behaviours at endpoint. While people may believe their daily
schedule and functionality is unaffected by sleep debt, it appears that as treatment progresses and
poor sleep quality persists thereby increasing sleep debt, people engage in certain behaviours,
possibly out of habit, instead of actively thinking of using them at one time of day or another.
This distinction suggests it is important to use self-rated coping measures, and behaviour
assessments, to distinguish between what people believe versus what they are actually doing.
These results should be considered to help patients and caregivers follow schedules that optimize
their performance based on chronotype and cognitive function especially for stressful situations,
and to maximize the possibility of having a full night sleep based on their body’s rhythm.
194
Chapter 7!
! Conclusion The main objective of this dissertation was to assess the degree to which coping is influenced by
chronotype and sleep quality in a prevalent model of chronic stress. In breast cancer, we
examined both patients as well as caregivers and oncology staff, as the issue of coping with
chronic disease depends on the attitudes and interactions of the numerous parties that are
involved with the patient. Coping was examined in several ways, both retrospectively and in the
moment. Among patients and spousal caregivers, coping was assessed using one’s recall of their
ability to cope across the day (including feelings of how well they coped, ability to show
compassion to their ill partner or caregiver, and their need for time and space to themselves), and
based on their use of various coping behaviours over time. Patient coping was also assessed in
the moment across the day. Among oncology staff, coping was assessed based on levels of
compassion satisfaction (CS), compassion fatigue (CF), and burnout. CS, CF, and burnout were
used as coping measures based on the idea that one’s ability to manage the specific and
potentially taxing external and internal demands of their job would produce various levels of
each of these domains (with inability for or low coping leading to low CS and high CF and
burnout, and vice versa). In addition to chronotype and sleep quality, personality was also
examined as a secondary factor that would influence one’s ability to cope with chronic stress.
In this population, chronotype and sleep quality influenced various measures of coping when
faced with chronic stress. Chronotype and sleep quality were not significantly correlated except
specifically in patients without a caregiver involved in the study. Sleep timing and chronotype
are correlated in a world-wide study, and it is possible that the size of the sample taken in our
study was too small to detect this. Furthermore, the inclusion of both cancer patients and their
caregivers in one group may have introduced considerably more variance in sleep habits than
would be found in a general population.
This study examined individuals predominantly working or being treated at Sunnybrook Health
Science’s Centre (SH) in the Odette Cancer Centre; a smaller population of nurses from Princess
Margaret Hospital (PMH) were also included. These are both teaching hospitals affiliated with
the University of Toronto. Although SH and PMH provide care to individuals from all over the
195
greater Toronto area, there are other major healthcare centers across the city. Therefore, the
target subject pool as well as the hospital environment and working conditions, may have been
limited. When nurses at SH and PMH were compared, they showed no significant differences on
the domains being examined. Nonetheless, it is possible that differences exist in populations that
are either working or receiving cancer treatment, in different facilities, cities, and or countries.
7.1! Chronotype
Chronotype was related to recalled coping assessments among patients and oncology staff,
however not significantly among caregivers. Patient reports also indicated that one’s chronotype
was linked with their in the moment coping ratings. Among patients, on recalled and in the
moment coping, people reported reaching their optimal coping ability in sequential order, with M
types showing peak times earlier than N types, who were earlier than E types. Among M types,
distinctions were also reported with definitely M types reaching their peak performance before
moderately M types. Recalled fatigue levels did not correlated with Horne-Östberg Morningness
Eveningness Questionnaire (MEQ) scores, however among patients it appeared that E types
reached their peak fatigue level earlier than M and N types (who did not differ significantly in
their peak times), suggesting E types do feel more fatigued earlier in the day, which may
subsequently delay their peak performance time for other coping behaviours. In the moment
coping also showed a chronotype dependent relationship, indicating that patients do realize
changes in their ability across the day. Unlike previous studies (e.g., Bellicoso, 2010) where
memory for emotional regulatory behaviours was poorly correlated to chronotype, it is possible
that continually facing the same chronic stressor on a daily basis has allowed people to be aware
of their behavioural abilities in the moment and retrospectively upon recall.
Among oncology staff, chronotype was also related to coping as measured by burnout (using the
Copenhagen Burnout Inventory, CBI), and Professional Quality of Life (including CS, and CF -
consisting of a different burnout measure, and secondary traumatic stress, STS). Among nurses
only, even when considering other factors such as work stressfulness and job satisfaction,
chronotype was a significant predictor of one’s burnout score; M types following dayshift
schedules showed lower burnout than N and E types on the same work schedule. However, when
burnout and secondary traumatic stress (STS) were examined using the ProQoL among a mixed
196
group of oncology caregiving staff, chronotype did not contribute to the onset of burnout, or
STS. When considered alone, at the univariate level chronotype did relate to coping as measured
by CS, suggesting staff with earlier chronotypes working dayshift schedules more likely coped
better and were able to enjoy the positive outcomes of providing care. However, when other
factors such as sleep quality, job satisfaction, and personality were included in a model with
chronotype as predictors of CS, the predictive role of chronotype was no longer significant. None
the less, the small role of chronotype in helping people obtain positive feelings and benefits from
providing care should not be dismissed. It is possible that in a larger population, or among
specific subgroups of oncology staff, chronotype plays a stronger role in mediating one’s level of
CS.
Among spousal caregivers, chronotype was not strongly linked to one’s recalled coping ability.
Anecdotally many caregivers would report that they were ‘not the sick one’ and ‘were doing
fine’, often saying suggesting they saw how sick their partner was and comparatively speaking
were doing fine. While caregivers did report a chronotype dependent need for time and space to
themselves at endpoint, there were no other chronotype dependent differences in recalled coping
performance across the day. It is possible that male caregivers do not pay strong attention to
performance fluctuations across the day and therefore do not form strong memories. It is also
possible however that caregivers typically felt they had to be “on duty” and perform optimally
around the clock when caring for their partner, leading them to believe their performance did not
fluctuate across the day. While it is possible that their performance actually did not fluctuate, this
is highly unlikely, given previous research indicating that even regardless of one’s awareness,
performance does fluctuate. Letting caregiver’s know that it is acceptable to take a break for
themselves or to have fluctuations in their caregiving abilities across the day may also reduce
stress and ameliorate their poor sleep quality.
When considering recalled coping behaviours, chronotype appeared to be a more prominent
predictor earlier in treatment (baseline and midpoint), but not often at endpoint. While
chronotype showed a fairly consistent relationship with performance fluctuations and one’s
ratings of their in the moment coping among patients, chronotype did not show a strong
relationship with use of specific coping behaviours among patients and caregivers. When asked
about how they were coping at a particular time of day, in the moment or retrospectively,
participants could consider a broad overview of how they were coping, instead of considering a
197
specific coping behaviour. Reflecting on one’s overall coping may not necessarily cause an
individual to think of the specific behaviours they are engaging in – it is possible that there are
too many different behaviours people use to face a chronic stressor, and instead people may store
a summative memory of their coping overall across the day. Future studies should assess what
factors people consider when rating their overall coping, and whether the individual contributors
relate differently to chronotype.
Overall, the role of chronotype in predicting coping should be considered in future studies.
Awareness for the fact that people are not always able to give their best coping performance for
chronic stressors is important in reducing employee burnout and providing additional help when
necessary to patients and caregivers before they may engage in destructive coping behaviours or
develop problems such as depression. Among caregivers, reducing CF and or increasing CS will
enhance the overall employee experience, and likely lead to more productive staff, and lower
turnover. Positive patient coping during cancer leads to better coping during survival (e.g., Jim,
Richardson, Golden-Kreutz & Andersen, 2006). Future studies should further examine the long-
term influence of positive coping with a chronic stressor such as cancer, together with an
understanding and acceptance that coping behaviour does fluctuate as predictors of greater
emotional recovery upon treatment completion.
7.2! Sleep Quality
Sleep quality was related to various recalled coping assessments among patients,
caregivers and oncology staff, but did not correlate strongly with in the moment measures of
coping. It is possible that in the moment coping was not strongly linked with sleep quality
because society often disregards the importance of sleep, believing and reporting that they can
function well on little sleep. Future studies examining sleep wake times as opposed to sleep
quality may discover different results, possibly showing peak coping times at specific distances
from one’s wake up or sleep time. The results of this study suggest patients disregard the
influence of their sleep quality (which was typically reported as poor) when immediately
examining their behaviour. However, long term, people may store a memory of their chronic
poor sleep, and relate that with fluctuations in their performance.
198
Among oncology nurses, and oncology staff as a collective group, sleep quality significantly
predicted burnout, both with and without the inclusion of work stressfulness and or job
satisfaction. Among the collective group of staff, when not considering job satisfaction, sleep
quality also significantly predicted the onset of CS, indicating that sleep quality influences
various aspects of coping (i.e., mediating stress perceptions leading to burnout, and helping to
find a positive effect of providing care in a chronically stressful environment). While poor sleep
quality can reduce CS and or increase burnout, it is important to note that several steps can be
taken to remedy poor sleep quality. Overall, good sleep quality has the potential to serve as a
protective factor influencing positive coping when facing chronic stress.
The intergroup differences between patients with and without a caregiver partaking in the study
are interesting to note and should be examined in future. Patients and caregivers who partook in
the study together did not show a significant correlation between their chronotype score and
sleep quality. Instead, patients without a caregiver involved in the study (i.e., either they had no
life partner, or their partner – who may or may not have been their primary caregiver and or may
or may not have attended treatments with them – did not enroll in the study), showed a highly
significant relationship between sleep quality and chronotype. Their results suggest that later
chronotypes had worse sleep quality, similar to other research findings (e.g., Merikanto,
Paavonen, Saarenpää-Heikkilä, Paunio & Partonen, 2017; Yun, Ahn, Jeong, Joo & Choi, 2015).
Instead, patients and caregivers involved in the study together showed no significant relationship
between chronotype and sleep quality, and suggested M types would likely have worse sleep
quality than E types, which is counter to many published findings, as well as the correlation
direction reported among nurses and oncology staff as a whole in this study. These results
suggest patients and caregivers participating together may be interrupting one another’s sleep,
possibly ruminating together over worries and fears as both are tired and cognitively depleted.
While these results do not suggest that it is necessarily better to not have a caregiver, it is
possible that having some space to oneself is beneficial.
Patient and caregiver sleep quality did not differ significantly by chronotype or across treatment.
However, patients’ and caregivers’ sleep quality did predict use of certain engagement style
coping behaviours, but only at endpoint. While sleep quality was reported as poor among both
groups from treatment outset, they were likely less cognitively depleted and sleep quality did not
influence coping behaviour choice or use earlier on. Chronic poor sleep quality reduces cognitive
199
abilities, particularly attentional and executive control (e.g., Alhola & Polo-Kantola, 2007;
Stavitsky, Neargarder, Bogdanova, McNamara & Cronin-Golomb, 2012). The results from our
study support this previous work, with sleep quality significantly influencing highly cognitive
coping behaviours such as self-distraction, active coping, planning, venting and self-blame at
endpoint. By endpoint, patients and caregivers alike have reported chronic poor sleep quality,
which begins to impact their coping. Future studies should examine sleep quality’s influence on
coping behaviours into the radiation therapy portion of treatment to assess whether the effect
continues. Research should also examine the specific components of one’s sleep that are
contributing to poor sleep quality, and whether particular components are influencing coping.
7.3! Personality
Like chronotype, personality is considered a relatively stable trait, that has the ability to change
over time, but is typically stable for extended periods. While the main factors for assessment in
this study were chronotype and sleep quality, we believed personality should also be examined
as it is present in all individuals and would undoubtedly influence certain coping decisions or
styles. While personality cannot be changed, it can be screened to identify individuals who may
have greater difficulty coping or those who may do very well on their own without additional
help. In our findings, when personality was included in a model it typically diminished the
effects of chronotype and sleep quality. However, those instances where chronotype and sleep
quality persisted in exerting an influence on coping even when considering personality are
important to note as it suggests a strong predictive value for chronotype and sleep quality. In
those cases where chronotype and sleep quality were no longer significant once personality was
included, the roles of chronotype and sleep quality should not be dismissed – it is possible that in
a larger population their roles might become more pronounced.
Among the total group of oncology staff included in this study, personality traits were
consistently predictive of CS, burnout, and STS. Emotional stability consistently predicted CF,
with and without the inclusion of job satisfaction, while agreeableness consistently predicted CS.
Burnout was also consistently predicted by traits shared with each CS and STS, suggesting that
burnout is possibly influenced by one’s positive and negative take on their work experiences and
stresses. When job satisfaction was no longer included as a predictive variable for ProQoL, while
200
the influence of sleep quality increased, not all personality variables increased their contribution
to CS, burnout, and STS. Those variables that were predictive with and without JSS increased in
their predictive value, however without the JSS, some variables lost their predictive significance.
These variables may have been related to job satisfaction and without it in the model were no
longer significantly predictive. Among oncology staff, openness was significantly predictive of
CS and burnout when job satisfaction was included in the model (i.e., high openness led to
increased CS and diminished burnout, and vice versa), yet was replaced by agreeableness as
dually predictive when JSS was not included. Emotional stability was predictive of burnout and
STS regardless of the inclusion of JSS, suggesting one’s level of emotional stability should be
screened when hiring individuals to work in this field in order to prevent CF.
The most commonly predictive personality trait among patients and caregivers was openness,
which unlike chronotype or sleep quality exhibited predictive value at various points across
treatment. When personality was not included in the model, sleep quality was predictive of
greater number of coping behaviours; upon including personality, sleep quality was only
predictive of self-distraction and planning behaviours. Openness was significantly predictive of
engagement style coping behaviours, suggesting a higher degree of openness predisposes an
individual to use more behaviours aimed at diminishing the stress of being a patient or caregiver.
In a larger group of participants, openness may demonstrate a more consistent predictive role
across treatment for the use of particular coping strategies.
7.4! Future Directions
Many of the patients who participated in this study were able to take time off work, often with
continued pay and benefits. This is not always possible for many individuals, and may vary
depending on the types of jobs held by patients. While chronotype, sleep quality, and personality
were each found to play significant roles influencing coping (both recalled and in the moment)
among patients, caregivers, and oncology staff, it is possible that the influence of these factors
may show greater variation depending on the demographics of the population studied. Given
cancer’s ability to affect anyone, regardless of age, race, gender, ethnicity, or socioeconomic
status, individuals from different walks of life will have their own mediating factors that
influence coping. For example, inner city hospitals (e.g., St. Michael’s Hospital) may show a
201
varied patient population in terms of socioeconomic status, where some patients may come from
affluent households, while others may come from low income areas, or even be homeless.
Individuals in each situation will have their own personal stresses that contribute to the overall
stress they feel from the situation. Future studies must examine the influence of chronotype and
sleep quality together with a broader description of one’s living situation and socioeconomic
status in order to provide a complete picture of results that is more generalized to the entire
population. In future, making the study available on an app or on computer with reminders to
participants for when to complete specific components of the study would reduce the time
constraint on the researcher and allow for study implementation at a greater number of hospitals.
Implementing the study at a greater number of hospitals would allow for a broader sample of the
population and how they cope.
Sunnybrook Health Sciences Centre and Princess Margaret Hospital are teaching hospitals,
servicing large populations. While nurses at the two sites did not show significant differences on
the domains examined, future research should examine oncology staff working at a range of
hospitals, including teaching and non-teaching hospitals, in and out of Toronto. It is important to
understand whether all oncology staff are feeling similar levels of stress, and understanding how
chronotype and sleep quality are relating to their coping as measured by ratings of their ProQoL.
Finally, staff, patients, and familial caregivers dealing with the stresses of other chronic illnesses
need to be examined to understand how stress and burnout related to chronotype, sleep quality,
and personality. For example, heart disease is also a chronic illness, with potentially fatal side
effects, as is liver or kidney disease. These chronic illnesses can also require regular hospital
visits, medication, and severe side effects. Working with such a population can be stressful for
staff, however it remains to be understood whether this would be the similar type of stress faced
by oncology staff, and whether their coping is similar. Patient and caregiver coping for such
illnesses – in the moment and recalled – is also susceptible to the influence of chronotype and
sleep quality. Overall, it is important to understand whether the results from the current study can
be generalized to other populations facing chronic health issues. A better understanding of
whether people faced with chronic illness cope similarly or not would help develop necessary
tools to ameliorate coping among staff, patients, and caregivers, and provide outside help when
necessary.
202
Bibliography
Achermann, P. (2004). The two-process model of sleep regulation revisited. Aviation, Space, and Environmental Medicine, 75, A37-A43. Adan, A. & Natale, V. (2002). Gender differences in morninness-eveningnes preference. Chronobiology International, 19, 709-720. Afshar, H., Roohafza, H. R., Keshteli, A. H., Mazaheri, M., Feizi, A., & Peyman, A. (2015). The association of personality traits and coping styles according to stress level. Journal of Research in Medical Sciences, 20, 353-358. Alhola, P. & Polo-Kantola, P. (2007). Sleep deprivation: Impact on cognitive performance. Neuropsychiatric Disease and Treatment, 3, 553-567. Ancoli-Israel, S., Liu, L., Rissling, M., Natarajan, L., Neikrug, A. B., Palmer, B. W., … Maglione, J. (2014). Sleep, fatigue, depression and circadian activity rhythms in women with breast cancer before and after treatment: A 1-year longitudinal study. Supportive Care in Cancer, 22, 2535–2545. doi:10.1007/s00520-014-2204-5 Ancoli-Israel, S., Moore, P. J., & Jones, V. (2001). The relationship between fatigue and sleep in cancer patients: A review. European Journal of Cancer Care, 10, 245-255. Antoniou, A., Pharoah, P. D. P., Narod, D., Risch, H. A., Eyfjord, J. E., Hopper, J. L., … Easton, D. F. (2003). Average risks of breast and ovarian cancer associated BRCA1 or BRCA2 mutations detected in case series unselected for family history: A combined analysis of 22 studies. American Journal of Human Genetics, 72, 1117-1130. doi: 10.1086/375033 Aoyagi, T., Terracina, K. P., Raza, A., Matsubara, H., & Takabe, K. (2015). Cancer cachexia, mechanism and treatment. World Journal of Gastrointestinal Oncology, 7, 17-29. doi: 10.4251/wjgo.v7.i4.17 Araghi, M.H., Jagielski, A., Neira, I., Brown, A., Higgs, S., Thomas, G.N., & Taheri, S. (2013). The complex associations among sleep quality, anxiety-depression, and quality of life in patients with extreme obesity. Sleep, 36, 1859-1865. doi: http://dx.doi.org/10.5665/sleep.3216 Arendt, J. (2006). Melatonin and human rhythms. Chronobiology International, 23, 21-37. Arnsten, A. F. T. (2009). Stress signaling pathways that impair prefrontal cortex structure and function. National Review of Neuroscience, 10, 410-422. doi: 10.1038/nrn2648 Armitage, R., Trivedi, M., Hoffmann, R., & Rush, A., J. (1997). Relationship between objective and subjective sleep measures in depressed patients and healthy controls. Depression and Anxiety, 5, 97-102.
203
Arpino, G., Weiss, H., Lee, A. V., Schiff, R., De Placido, S., Osborne, C. K, & Elledge, R. M. (2005). Estrogen receptor-positive, progesterone receptor-negative breast cancer: Association with growth factor receptor expression and tamoxifen resistance. Journal of the National Cancer Institute, 97 (17), 1254-1261. doi: 10.1093/jnci/dji249 Aslan, O., Sanisoglu, Y., Akyol, M., & Yetkin, S. (2009). Quality of sleep in Turkish family caregivers of cancer patients. Cancer Nursing, 32, 370-377. doi: 10.1097/NCC.0b013e3181a87fb9 Atkinson, G. & Reilly, T. (1996). Circadian variations in sports performance. Sports Medicine, 21, 292-312.
Austenfeld, J. L., & Stanton, A. L. (2004). Coping through emotional approach: A new look at emotion, coping, and health-related outcomes. Journal of Personality, 7, 1335-1363. doi: 10.1111/j.1467-6494.2004.00299.x Bahk, Y.-C., Han, E., & Lee, S.-H. (2014). Biological rhythm differences and suicidal ideation in patients with major depressive disorder. Journal of Affective Disorders, 168, 294- 297. doi: 10.1016/j.jad.2014.07.001 Bakker, A. B., Van der Zee, K. I., Lewig, K. A., & Dollard, M. F. (2006). The relationship between the Big Five personality factors and burnout: A study among volunteer counselors. Journal of Social Psychology. 146, 31–50. doi: 10.3200/SOCP.146.1.31-50 Banks, S., & Dinges, D. F. (2007). Behavioral and physiological consequences of sleep restriction. Journal of Clinical Sleep Medicine, 3, 519–28. Barre, V. P., Padmaja, G., Saxena, R. K., & Rana, S. (2015). Impact of medical intervention on stress and quality of life in patients with cancer. Indian Journal of Palliative Care, 21 (2), 203-208. doi: 10.4103/0973-1075.156503 Bartley, C. E. & Roesch, S. C. (2011). Coping and daily stress: The role of conscientiousness. Personality and Individual Differences, 50, 79–83. doi: 10.1016/j.paid.2010.08.027 Baum, A. (1990). Stress, intrusive imagery, and chronic distress. Health Psychology, 9 (6), 653- 675. Baum, A., Friedman, A. L., & Zakowski, S. G. (1997). Stress and genetic testing for disease risk. Health Psychology, 16, 8-19. Beach, S. R., Schulz, R., Williamson, G. M., Miller, S., Weiner, M. F., & Lance, C. E. (2005). Risk factors for potentially harmful informal caregiver behavior. 53, 255-261. doi: 10.1111/j.1532-5415.2005.53111.x Becker, J. B. & Hu, M. (2008). Sex differences in drug abuse. Frontiers in Neuroendocrinology, 29, 36-47. doi: 10.1016/j.yfrne.2007.07.003
204
Bellicoso, D. (2010). A task’s cognitive demands influence self-reported performance variances throughout the day (Master’s Thesis). Available from University of Toronto Research Repository. https://tspace.library.utoronto.ca/bitstream/1807/25428/1/Bellicoso_Daniela_201011_M A_thesis.pdf Bellicoso, D., Ralph, M. R., Trudeau, M. E. (2014) Burnout among oncology nurses: The influence of chronotype and sleep quality. Journal of Nursing Education and Practice, 4, 80-89. doi: https://doi.org/10.5430/jnep.v4n8p80 Ben-Zur, H. & Michael, K. (2007). Burnout, social support, and coping at work among social workers, psychologists, and nurses: The role of challenge/control appraisals. Social Work in Health Care, 45, 63-82. doi: 10.1300/J010v45n04_04 Berger, A. M., Parker, K. P., Young-McCaughan, S., Mallory, G. A., Barsevick, A., Beck, S. L., … Hall, M. (2005). Sleep/Wake disturbances in people with cancer and their caregivers: State of the science. Oncology Nursing Forum, 32, E98-E126. doi: 10.1188/05.ONF.E98-E126 Bersagliere A, Raduazzo ID, Nardi S, Gatta A, Amodio P, Achermann P, & Montagnese, S. (2012). Induced hyperammonemia may compromise the ability to generate restful sleep in patients with cirrhosis. Hepatology, 55, 869-878. doi: 10.1002/hep.24741 Biss, R. & Hasher, L. (2012). Happy as a lark: Morning-type younger and older adults are higher in positive affect. Emotion, 12, 437-441. doi: 10.1037/a0027071 Blatter, K. & Cajochen, C. (2007). Circadian rhythms in cognitive performance: Methodological constraints, protocols, and theoretical underpinnings. Physiology and Behavior, 90, 196-208. doi: 10.1016/j.physbeh.2006.09.009 Bonnet, M. H. & Arand, D. L. (1995). We are chronically sleep deprived. Sleep, 18, 908-911. Bower, J. E. (2008). Behavioral symptoms in breast cancer patients and survivors: Fatigue, insomnia, depression, and cognitive disturbance. Journal of Clinical Oncology, 26, 768- 777. doi: 10.1200/JCO.2007.14.3248 Brand, S., Beck, J., Hatzinger, M., Harbaugh, A., Ruch, W., Holsboer-Trachsler, E. (2010). Associations between satisfaction with life, burnout-related emotional and physical exhaustion, and sleep complaints. The World Journal of Biological Psychiatry, 11, 744- 754. doi: 10.3109/15622971003624205 Brown, S. A., Kunz, D., Dumas, A., Westermark, P. O., Vanselow, K., Tilmann-Wahnschaffe, … Kramer, A. (2008). Molecular insights into human daily behavior. PNAS, 105, 1602- 1607. doi: 10.1073/pnas.0707772105 Brown, F. M., Neft, E. E., & LaJambe, C. M. (2008). Collegiate rowing crew performance varies by morningness-eveningness. Journal of Strength and Conditioning Research, 22, 1894- 1900. doi: 10.1519/JSC.0b013e318187534c
205
Bultz, B. D. & Carlson, L. E. (2006). Emotional distress: The sixth vital sign – future directions in cancer care. Psycho-Oncology, 15, 93-95. doi: 10.1002/pon.1022 Burgess, H. J. & Fogg, L. F. (2008). Individual differences in the amount and timing of salivary melatonin secretion. PLoS One, 3, e3055. doi: 10.1371/journal.pone.0003055 Burgess, L., Irvine, F., & Wallymahmed, A. (2010). Personality, stress and coping in intensive care nurses: A descriptive exploratory study. Nursing in Critical Care, 15, 129-140. doi: 10.1111/j.1478-5153.2009.00384.x Burker, E. J., Evon, D., Loiselle, M. M., Finkel, J., & Mill, M. (2005). Planning helps, behavioral disengagement does not: coping and depression in the spouses of heart transplant candidates. Clinical Transplant, 19, 653-658. doi: 10.1111/j.1399-0012.2005.00390.x Buysse, D. J., Reynolds III, C. F., Monk, T. H., Berman, S. R. & Kupfer, D. J. (1989). The Pittsburgh sleep quality index: A new instrument for psychiatric practice and research. Psychiatry Research, 28, 193-213. Byun, E., Gay C. L., & Lee, K. A. (2016). Sleep, fatigue and problems with cognitive function in adults living with HIV. The Journal of the Association of Nurses in AIDS Care, 27, 5-16. doi: 10.1016/j.jana.2015.10.002 Canadian Breast Cancer Foundation (2016). Breast Cancer in Canada, 2016. Retrieved at: http://www.cbcf.org/central/AboutBreastCancerMain/FactsStats/Pages/Breast-Cancer- Canada.aspx Canadian Cancer Society’s Advisory Committee on Cancer Statistics. Canadian Cancer Statistics 2016. Toronto, ON: Canadian Cancer Society; 2016a. Retrieved at: http://www.cancer.ca/~/media/cancer.ca/CW/cancer%20information/cancer%20101/Cana dian%20cancer%20statistics/Canadian-Cancer-Statistics-2016-EN.pdf?la=en Canadian Cancer Society (2016b). Survival statistics for breast cancer. Retrieved at: http://www.cancer.ca/en/cancer-information/cancer-type/breast/prognosis-and- survival/survival-statistics/?region=on Carlson, L. E. & Bultz, B. D. (2004). Efficacy and medical cost offset of psychosocial interventions in cancer care: Making the case for economic analyses. Psycho- Oncology, 13, 837-849. doi: 10.1002/pon.832 Carpenter, J. S. & Andrykowski, M. A. (1998). Psychometric evaluation of the Pittsburgh sleep quality index. Journal of Psychosomatic Research, 45, 5-13. Carr, D., Goudas, L., Lawrence, D., Pirl, W., Lau, J., DeVine, D., … & Miller, K. (July 2002). Management of Cancer Symptoms: Pain, Depression, and Fatigue. Evidence Report/Technology Assessment No. 61 (Prepared by the New England Medical Center Evidence-based Practice Center under Contract No 290-97-0019). AHRQ Publication No. 02-E032. Rockville, MD: Agency for Healthcare Research and Quality.
206
Carter, P. A., & Acton, G. J. (2006). Personality and coping: Predictors of depression and sleep problems among caregivers of individuals who have cancer. Journal of Gerontological Nursing, 32, 45-53. Carver, C. S. (1997). You want to measure coping but your protocol’s too long: Consider the Brief-COPE. International Journal of Behavioral Medicine, 41, 92-100. Carver, C. S. (2005). Enhancing adaptation during treatment and the role of individual differences. Cancer, 104 (11 Suppl.), 2602–2607. doi: 10.1002/cncr.21247 Carver, C. S. & Connor-Smith, J. (2010). Personality and coping. Annual Review of Psychology, 61, 679-704. doi: 10.1146/annurev.psych.093008.100352 Carver, C. S., Pozo, C., Harris, S. D., Noriega, V., Scheier, M. F., Robinson, D. S., … Clark, K. C. (1993). How coping mediates the effect of optimism on distress: A study of women with early stage breast cancer. Journal of Personality and Social Psychology, 65, 375- 390. Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping strategies: A theoretically based approach. Journal of Personality and Social Psychology, 56, 267-283. Cavallera, G. M., Gatto, M., & Boari, G. (2014). Personality, cognitive styles, and morningness- eveningness disposition in a sample of Yoga trainees. Medical Science Monitor, 13, 238- 246. doi: 10.12659/MSM.889030 Chaudhury, D. & Colwell, C. S. (2002). Circadian modulation of learning and memory in fear conditioned mice. Behavioural Brain Research, 133, 95-108. Chen, S. & Parmigiani, G. (2007). Meta-analysis of BRCA1 and BRCA2 Penetrance. Journal of Clinical Oncology, 25, 1329-1333. doi: 10.1200/JCO.2006.09.1066 Chen, S.M. & McMurray, A. (2001). “Burnout” in intensive care nurses. Journal of Nursing Research, 9, 152-164. Chesney, M., Neilands, T. B., Chambers, D. B., Taylor, J. M. & Folkman, S. (2006). A validity and reliability study of the coping self-efficacy scale. British Journal of Health Psychology, 11, 421-437. doi: 10.1348/135910705X53155 Choi, J., Yoon, I., Kim, H., Chung, S., & Yoo, H. J. (2010). Differences between objective and subjective sleep measures in children with attention deficit hyperactivity disorder. Journal of Clinical Sleep Medicine, 6, 589-595. Christie, W. & Moore, C. (2005). The impact of humor on patients with cancer. Clinical Journal of Oncology Nursing, 9, 211-218. doi: 10.1188/05.CJON.211-218 Clark, L. A., Watson, D., & Leeka, J. (1989). Diurnal variation in the positive affects. Motivation and Emotion, 13, 205-234.
207
Clough-Gorr, K. M., Ganz, P. A., & Silliman, R. A. (2007). Older breast cancer survivors: Factors associated with change in emotional well-being. Journal of Clinical Oncology, 25, 1334-1340. doi: 10.1200/JCO.2006.09.8665 Cohen, D. A., Wang, W., Wyatt, J. K., Kronauer, R. E., Dijk, D. J., Czeisler, C. A., & Klerman, E. B. (2010). Uncovering residual effects of chronic sleep loss on human performance. Science Translational Medicine, 2, 14ra3. doi: 10.1126/scitranslmed.3000458 Connor-Smith, J. & Flachsbart, C. (2007). Relations between personality and coping: A meta- analysis. Journal of Personality and Social Psychology, 93, 1080-1107. doi: 10.1037/0022-3514.93.6.1080 Costa-Martins, J.M., Moura-Ramos, M., Cascais, M. J., da Silva, C. F., Costa-Martins, H., Pereira, M., Coelho, R., Tavares, J. (2016). Adult attachment style and cortisol responses in women in late pregnancy. BMC Psychology, 4, 1. doi: 10.1186/s40359-016-0105-8 Craig, C. D. & Sprang, G. (2010). Compassion satisfaction, compassion fatigue, and burnout in a national sample of trauma treatment therapists. Anxiety, Stress & Coping, 23 (3), 319- 339. doi: 10.1080/10615800903085818 Cubrilo-Turek, M., Urek, R., & Turek, S. (2006). Burnout syndrome – assessment of a stressful job among intensive care staff. Collegium Antropologicum, 30, 131-135. Davies, K. A., Macfarlane, G. J., Nicholl, B. I., Dickens, C., Morriss, R., Ray, D. & McBeth, J. (2008). Restorative sleep predicts the resolution of chronic widespread pain: Results from the EPIFUND study. Rheumatology, 47, 1809–1813. doi: 10.1093/rheumatology/ken389 Davis-Ali, S., Chesler, M. A., & Chesney, B. K. (1993). Recognizing cancer as a family disease: Worries and support reported by patients and spouses. Social Work in Health Care, 19, 45-65. Demytteraera, K., Bonte, L., Gheldof, M., Vervaeke, M., Meuleman, C., Vanderschuerem, D., & D’Hooghe, T. (1998). Coping style and depression level influence outcome in in vitro fertilization. Fertility and Stress, 69, 1026-1033. doi: https://doi.org/10.1016/S0015- 0282(98)00089-2 Dent, R., Trudeau, M., Pritchard, K. I., Hanna, W. M., Kahn, H. K., Sawka, C. A., … Narod, S. A. (2007). Triple negative breast cancer: Clinical features and patterns of recurrence. Clinical Cancer Research, 13 (15), 4429-4434. doi: 10.1158/1078-0432.CCR-06-3045 DeYoung, C. G., Quilty, L. C., Peterson, J. B. (2007). Between facets and domains: 10 aspects of the Big Five. Journal of Personality and Social Psychology, 93, 880–96. doi: 10.1037/0022-3514.93.5.880 Di Milia, L., Adan, A., Natale, V., & Randler, C. (2013). Reviewing the psychometric properties of contemporary circadian typology measures. Chronobiology International, 30, 1261- 1271. doi: 10.3109/07420528.2013.817415
208
Dogru, S., Dogru, H. Y., Tapar, T., Karaman, S., Arici, S., Ozsoy, A. Z., … Karakis, A. (2016). Do urgent caesarian sections have a circadian rhythm? Turkish Journal of Anesthesiology and Reanimation, 44, 124-7. doi: 10.5152/TJAR.2016.57614 Drury, V., Francis, K., & Aoun, S. (2014). Compassion satisfaction, compassion fatigue, anxiety, depression and stress in registered nurses in Australia: Phase 2 results. Journal of Nursing Management, 22, 519-531. Duggan, K. A., Friedman, H. A., McDevitt, E. A., & Mednick, S.C. (2014). Personality and healthy sleep: The importance of conscientiousness and neuroticism. PLoS ONE, 9, e90628. doi: doi: 10.1371/journal.pone.0090628 Dunfield, K. (2014). A construct divided: Prosocial behavior as helping and comforting subtypes. Frontiers in Psychology, 5, 1-13. doi: 10.3389/fpsyg.2014.00958 Dunn, K. S. & Horgas, A. L. (2004). Religious and nonreligious coping in older adults experiencing chronic pain. Pain Management Nursing, 5, 19-28. doi:10.1016/S1524- 9042(03)00070-5 Dworak, McCarley, Kim, Kalinchuk & Basheer, 2010). Sleep and brain energy levels: ATP changes during sleep. The Journal of Neuroscience, 30, 9007-9016. doi: 10.1523/JNEUROSCI.1423-10.2010 Eddy, R. (2005). Sleep deprivation among physicians. BC Medical Journal, 47, 176-180. Edge, S. B., Byrd, D. R., Compton, C. C., Fritz, A. G., Charlotte, F. I., & Trotti, A. (Eds.) (2010). Breast. In AJCC Cancer Staging Handbook, 7th Ed., pp. 347-376. New York, NY: Spring. El-Bar, N., Levy, A., Wald, H. S., & Biderman, A. (2013). Compassion fatigue, burnout and compassion satisfaction among family physicians in the Negev area – a cross sectional study. Isreal Journal of Health Policy, 15, 1. doi: 10.1186/2045-4015-2-31. Engstron, C. A., Strohl, R. A., Rose., L., Lweandowski, L. & Stefanek, M. E. (1999). Sleep alterations in cancer patients. Cancer Nursing, 22 (2), 143-148. Ferrara, M. & De Gennaro, L. (2001). How much sleep do we need? Sleep Medicine, 5, 155-179. doi: 10.1053/smrv.2000.0138 Fiorentino, L. & Ancoli-Israel, S. (2007). Sleep dysfunction in patients with cancer. Current Treatment Options in Neurology, 9, 337-346. Folkman, S. & Lazarus, R. S. (1980). An analysis of coping in a middle-aged community sample. Journal of Health and Social Behavior, 21, 219-239. Fortner, B. V., Stepanski, E. J., Wang, S. C., Kasprowicz, S., & Heith Durrence, H. (2002). Sleep and quality of life in breast cancer patients. Journal of Pain and Symptom Management, 24, 471-480.
209
Friedman, L., Brooks, J. O., Bliwise, D. L., Yesavage, J. A., & Wicks, D. S. (1995). Perceptions of life stress and chronic insomnia in older adults. Psychology and Aging, 10, 352-357. Fujiwara, Y., Habu, Y., Ashida, K., Kusano, M., Higuchi, K., & Arakawa, T. (2013). Sleep disturbances and refractory gastroesophageal reflux disease symptoms in patients receiving once-daily proton pump inhibitors and efficacy of twice-daily Rabeprazole treatment. Digestion, 88, 145-152. doi: http://dx.doi.org/10.1159/000354071 Gau, A. S.-F., Shang, C.-Y., Merikangas, K. R., Chiu, Y.-N., Soong, W.-T., & Cheng, A. T.- A. (2007). Association between morningness-eveningness and behavioral/emotional problems among adolescents. Journal of Biological Rhythms, 22, 268-274. doi: 10.1177/0748730406298447 Gesiarz, F. & Crockett, M. J. (2015). Goal-directed, habitual and Pavlovian prosocial behavior. Frontiers in Behavioral Neuroscience, 9, 1-18. doi: 10.3389/fnbeh.2015.00135 Gleichgerrcht, E. & Decety, J. (2013) Empathy in clinical practice: How individual dispositions, gender, and experience moderate empathic concern, burnout, and emotional distress in physicians. PLOS One, 8, e61526. doi:10.1371/journal.pone.0061526 Goetz, J. L., Keltner, D., & Simon-Thomas, E. (2010). Compassion: An evolutionary analysis and empirical review. Psychological Bulletin, 136, 351-374. doi:10.1037/a001880 Goretti, B., Portaccio, E., Zipoli, V., Razzolini, L., & Amato, M. P. (2010). Coping strategies, cognitive impairment, psychological variables and their relationship with quality of life in multiple sclerosis. Neurological Sciences, 31, S227-30. doi: 10.1007/s10072-010-0372-8 Gosling, S.D., Rentfrow, P.J., & Swann, W.B. Jr. (2003). A very brief measure of the big five personality domains. Journal of Research in Personality, 37, 504–28. doi: 10.1016/S0092-6566(03)00046-1 Hahn, C., Cowell, J. M., Wiprzycka, U. J., Goldstein, D., Ralph, M., Hasher, L., & Zelazo, P. D. (2012). Circadian rhythms in executive function during the transition to adolescence: The effect of synchrony between chronotype and time of day. Developmental Science, 15, 408-16. doi: 10.1111/j.1467-7687.2012.01137.x Hall, M., Buysse, D. J., Dew, M. A., Prigerson, H. G., Kupfer, D. J., & Reynolds, C. F. (1997). Intrusive thoughts and avoidance behavior are associated with sleep disturbances in bereavement-related depression. Depression and Anxiety, 6, 106-112. Hamilton, J. G., Lobel, M., & Moyer, A. (2009). Emotional distress following genetic testing for hereditary breast and ovarian cancer: A meta-analytic review. Health Psychology, 28 (4) 510-518. doi: 10.1037/a0014778 Hammen, C., Kim, E. Y., Eberhart, N. K. & Brennan, P. A. (2009). Chronic and acute stress and the prediction of major depression in women. Depression and Anxiety, 26, 718-723. doi: 10.1002/da.20571
210
Hansen, J. & Lassen, C. F. (2012). Nested case-control study of night shift work and breast cancer risk among women in the Danish military. Occupational and Environmental Medicine, 69 (8), 551 – 556. doi: 10.1136/oemed-2011-100240. Hasher, L., Chung, C., May, C. P., & Foong, N. (2002). Age, time of testing, and proactive interference. Canadian Journal of Experimental Psychology, 56, 200-207. Hasher, L., Zacks, R. T., & Rahhal, T. A. (1999). Timing, instructions, and inhibitory control: Some missing factors in the age and memory debate. Gerontology, 45, 355-357. Hasler, B. P., Allen, J. J. B., Sbarra, D. A., Bootzin, R. R., & Bernert, R. A. (2010). Morningness-eveningness and depression: Preliminary evidence for the role of the behavioral activation system and positive affect. Psychiatry Research, 176, 166-173. doi: 10.1016/j.psychres.2009.06.006 Hede, K. (2009). Cancer and the circadian clock: Has the time finally come? Journal of the National Cancer Institute, 101, 550-553. doi: 10.1093/jnci/djp085 Heinzelmann, M., Lee, H., Rak, H., Livingston, W., Barr, T., Baxter, T., Scattergood-Keepper, L., Mysliwiec, V. & Gill, J. (2014). Sleep restoration is associated with reduced plasma C reactive protein and depression symptoms in military personnel with sleep disturbance after deployment. Sleep Medicine, 15, 1565-1570. doi: 10.1016/j.sleep.2014.08.004 Hicks, D. G. & Kulkarni, S. (2008). Review of biologic relevance and optimal use of diagnostic tools. American Journal of Clinical Pathology, 129, 263-273. Hidalgo, M. P., Caumo, W., Posser, M., Coccaro, S. B., Camozzato, A. L., & Chaves, M. L. (2009). Relationship between depressive mood and chronotype in healthy subjects. Psychiatry and Clinical Neurosciences, 63, 283-290. doi: 10.1111/j.1440- 1819.2009.01965.x Hintsanen, M., Puttonen, S., Smith, K., Toornroos, M., Jokela, M., Pulkki-Råback, L., Hintsa, T., Merjonen, P., Dwyer, T., Raitakari, O. T., Venn, A. & Keltikangas-Järvinen, L. (2014). Five factor personality traits and sleep: Evidence from two population based cohort studies. Health Psychology, 33, 1214-1223. doi: 10.1037/hea0000105 Hofman, M., Ryan, J. L., Figueroa-Moseley, C. D., Jean-Pierre, P., & Morrow, G. R. (2007). Cancer-related fatigue: The scale of the problem. The Oncologist, 12, 4-10. doi: 10.1634/theoncologist.12-S1-4 Horne, J. A. & Östberg, O. (1976). A self-assessment questionnaire to determine morningness- eveningness in human circadian rhythms. International Journal of Chronobiology, 4, 97- 110. Horowitz, M., Wilner, N., & Alvarez, W. (1979). Impact of events scale: A measure of subjective stress. Psychosomatic Medicine, 41, 209-218.
211
Howlader, N., Noone, A. M., Krapcho, M., Garshell, J., Miller, D. Altekruse, S. F., … Cronin, K. A. (eds). SEER Cancer Statistics Review, 1975-2011, National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/csr/1975_2011/ , based on November 2013 SEER data submission, posted to the SEER web site, April 2014. Retrieved at: https://www.cancer.gov/about-cancer/causes-prevention/genetics/brca-fact-sheet#r6 Hoyt, M. A., Thomas, K. S., Epstein D. R., & Dirksen, S. R. (2009). Coping style and sleep quality in men with cancer. Annals of Behavior Medicine, 37, 88-93. doi: 10.1007/s12160-009-9079-6 Hsu, C., Gau, S. S., Shang, C., Chiu, Y. & Lee, M. (2012) Associations between chronotypes, psychopathology, and personality among incoming college students. Chronobiology International, 29, 491-501. doi: 10.3109/07420528.2012.668995 Injeyan, M. C., Shuman, C., Shugar, A., Chitayat, D., Atenafu, E. G., & Kaiser, A. (2011). Personality traits associated with genetic counselor compassion fatigue: The roles of dispositional optimism and locus of control. Journal of Genetic Counselling, 20 (5), 526- 540. doi: 10.1007/s10897-011-9379-4 Innominato, P. F., Levi, F. A., & Bjarnason, G. A. (2010). Chronotherapy and the molecular clock: Clinical implication in oncology. Advanced Drug Delivery Reviews, 62, 979-1001. doi: 10.1016/j.addr.2010.06.002 Institute of Medicine (IOM). (2008). Cancer care for the whole patient: Meeting psychosocial health needs. Nancy E. Adler and Ann E. K. Page, eds. Washington, DC: The National Academies Press. Jim, Richardson, Golden-Kreutz & Andersen, 2006). Strategies used in coping with a cancer diagnosis predict meaning in life for survivors. Health Psychology, 25, 753-761. doi: 10.1037/0278-6133.25.6.753 Jones, J. M., St. Peter, J. R., Fernandes, S. J., Herrenkohl, T. I., Kosterman, R., & Hawkins, J. D. (2011). Ethnic and gender variation in religious involvement: Patterns of expression in young adulthood. Review of Religious Research, 53, 207–225. doi:10.1007/s13644-011- 0006-5 Jonsdottir, I. H., Nordlund, A., Ellbin, S., Ljung, T., Glise, K., Währborg, P., & Wallin, A. (2013). Cognitive impairment in patients with stress-related exhaustion. Stress, 16, 181- 190. doi: 10.3109/10253890.2012.708950 Kahn-Greene, E., Lipizzi, E. L., Conrad, A. K., Kamimori, G. H. & Killgore, W. D. S. (2006). Sleep deprivation adversely affects interpersonal responses to frustration. Personality and Individual Differences, 41, 1433-1443. doi: 10.1016/j.paid.2006.06.002 Kalmbach, D. A., Schneider, L. D., Cheung, J., Bertrand, S. J., Kariharan, T., Pack, A., & Gehrman, P. R. (2017). Genetic basis of chronotype in humans: Insights from three landmark GWAS. Sleep, 40, 1-10. doi: 10.1093/sleep/zsw048
212
Karagozoglu, S. & Bingöl, N. (2008). Sleep quality and job satisfaction of Turkish nurses. Nursing Outlook, 56, 298-307, 307.e1-.e3. doi: 10.1016/j.outlook.2008.03.009 Keller, L. K., Grünewald, B., Vetter, C., Roenneberg, T., & Schulte!Körne, G. (2017). Not later, but longer: Sleep, chronotype and light exposure in adolescents with remitted depression compared to healthy controls.European Child and Adolescent Psychiatry, Epub ahead of print. doi: 10.1007/s00787-017-0977-z Khoynezhad, G., Rajaei, A. R., & Sarvarazemy, A. (2012). Basic religious beliefs and personality traits. Iranian Journal of Psychiatry, 7, 82-86. Killgore, W. D. S. (2010). Effects of sleep deprivation on cognition. Progress in Brain Research, 185, 105-129. doi: 10.1016/B978-0-444-53702-7.00007-5 Kleitman, N. (1963). Sleep and wakefulness. Chicago, Il: University of Chicago Press. Knutson, Spiegel, Penev & Van Cauter, 2007). The metabolic consequences of sleep deprivation. Sleep Medicine Reviews, 11, 163-178. doi: 10.1016/j.smrv.2007.01.002 Korczak, D., Huber, B., & Kister, C. (2010). Differential diagnostic of the burnout syndrome. GMS Health Technology Assessment, 6, 1-9. doi: 10.3205/hta000087 Koscec, A., Radosevic-Vidacek, B., & Bakotic, M. (2014). Morningness-eveningness and sleep patterns of adolescents attending school in two rotating shifts. Chronobiology International, 31, 52-63. doi: 10.3109/07420528.2013.821128 Kreitler, S. (1999). Denial in cancer patients. Cancer Investigation, 17, 514-534. Krishnan, H. C. & Lyons, L. C. (2015) Synchrony and desynchrony in circadian clocks: Impacts on learning and memory. Leaning & Memory (Cold Springs Harbor, N.Y.), 22 (9), 426- 437. Kristensen, T.S., Borritz M., Villadsen, E., & Christensen, K.B. (2005). The Copenhagen burnout inventory: a new tool for the assessment of burnout. Work Stress, 19, 192-207. doi: http://dx.doi.org/10.1080/02678370500297720 Kurtz, M. E., Given, B., Kurtz J. C., & Given, C. W. (1994). The interaction of age, symptoms, and survival status on physical and mental health of patients with cancer and their families. Cancer, 74(Suppl 7), 2071-2078. Lack, L., Bailey, M., Lovato, M., & Wright, H. (2009). Chronotype differences in circadian rhythms of temperature, melatonin, and sleepiness as measured in a modified constant routine protocol. Nature and Science of Sleep, 1, 1-8. Laerd Statistics (2015). Statistical tutorials and software guides. Retreived from: https://statistics.laerd.com
213
Lara, T., Madrid, J. A., & Correa, A. (2014). The vigilance decrement in executive function is attenuated when individual chronotypes perform at their optimal time of day. PLoS ONE, 9, e88820. doi: 10.1371/journal.pone.0088820 Lawson, V. L., Bundy, C., Belcher, J., & Harvey, J. N. (2013). Changes in coping behavior and the relationship to personality, health threat communication and illness perceptions from the diagnosis of diabetes: a 2-year prospective longitudinal study. Health Psychology Research, 1, 93-99. doi: 10.4082/hpr.2013.e20 Lazarus R. & Folkman S. (1984) Stress, appraisal, and coping. New York: Springer Lemma, S., Berhane, Y., Worku, A., Gelaye, B., Williams, M.A. (2014). Good sleep quality is associated with better academic performance among university students in Ethiopia. Sleep & Breathing, 18, 257-263. doi: 10.1007/s11325-013-0874-8 Lee, C. Y., Chen,H. C., Tseng, M. C., Lee,H. C., & Huang, L.H. (2015). The relationships among sleep quality and chronotype, emotional disturbance, and insomnia vulnerability in shift nurses. The Journal of Nursing Research, 23, 225-235. doi: 10.1097/jnr.0000000000000095 Lenaert, B., Barry, T. J., Schruers, K., Vervliet, B., & Hermans, D. (2016). Emotional attentional control predicts changes in diurnal cortisol secretion following exposure to a prolonged psychosocial stressor. Psychoneuroendocrinology, 63, 291-5. doi: 10.1016/j.psyneuen.2015.10.013 Lenjavi, M. R., Ahuja, M. A., Touchette, P. E., & Sandman, C. A. (2010). Maladaptive behaviors are linked with inefficient sleep in individuals with developmental disabilities. Journal of Neurodevelopmental Disorders, 2, 174-180. doi: 10.1007/s11689-010-9048-1 Leproult, R., Copinschi, G., Buxton, O. & Van Cauter, E. (1997). Sleep loss results in an elevation of cortisol levels the next evening. American Sleep Disorders Association and Sleep Research Society, 20, 865-870. Lerman, C. & Croyle, R. T. (1996). Emotional and behavioral responses to genetic testing for susceptibility to cancer. Oncology, 10, 191-195. Lerman, C. & Schwartz, M. (1993). Adherence and psychological adjustment among women at high risk for breast cancer. Breast Cancer Research and Treatment, 28, 145–155. Levi, F. (2001). Circadian chronotherapy for human cancers. The Lancet Oncology, 2, 307-315. doi: http://dx.doi.org/10.1016/S1470-2045(00)00326-0 Levi, F. (2006). The circadian timing system, a coordinator of life processes. Implications for the rhythmic delivery of cancer therapeutics. Conference Proceedings: Annual International Conference of the IEEE Engineering in Medicine and Biology Society, Suppl: 6736-6739. doi: 10.1109/IEMBS.2006.260934
214
Lim, J. & Dinges, D. F. (2008). Sleep deprivation and vigilant attention. Annals of the New York Academy of Science, 1129, 305-322. doi: 10.1196/annals.1417.002 Lim & Dinges (2010). A meta-analysis of the impact of short-term sleep deprivation on cognitive variables. Psychology Bulletin, 136, 375-389. doi: 10.1037/a0018883 Loewenthal, K. M., McLeod, A. K., & Cinnirella, M. (2000). Are women more religious than men? Gender differences in religious activity among different religious groups in the UK. Personality and Individual Differences, 32, 133-139. Lowrey, P. L. & Takahashi, J. S. (2011). Genetics of circadian rhythms in mammalian model organisms. Advances in Genetics, 74, 175-230. doi: 10.1016/B978-0-12-387690-4.00006- 4 Luo, J. Zhu, G., Zhao, Q., Guo, Q., Meng, H., Hong, Z. & Ding D. (2013). Prevalence and risk factors of poor sleep quality among Chinese elderly in an urban community: Results from the Shanghai aging study. PLoS ONE, 8, e81261. doi: 10.1371/journal.pone.0081261 Matchock, R. L. & Mordkoff, J. T. (2009). Chronotype and time-of-day influences on the alerting, orienting, and executive components of attention. Experimental Brain Research, 192, 189-198. doi: 10.1007/s00221-008-1567-6 May, C. P., Hasher, L., & Stoltzfus, E. R. (1993). Optimal time of day and the magnitude of age differences in memory. Psychological Science, 4, 326-330. McClung, C. R. (2011). Circadian rhythms: Lost in post translation. Current Biology, 21 (10), R400-R402. doi: 10.1016/j.cub.2011.04.024 McEwen, B. S. & Karatsoreos, I. N. (2015). Sleep deprivation and circadian disruption: Stress, allostastis and allostatic load. Sleep Medicine Clinics, 10, 1-10. doi: https://doi.org/10.1016/j.jsmc.2014.11.007 Merckaert, I., Libert, Y., Lieutenant, F., Moucheux, A., Farvacques, C., Slachmuyider, J.-L., & Razavi, D. (2013). Desire for formal psychological support among caregivers of patients with cancer: Prevalence and implications for screening their needs. Psycho-Oncology, 22, 1389-1395. doi: 10.1002/pon.3153 Merikanto, I., Kronholm, E., Peltonen, M., Laatikainen, T., Lahti, T., Partonen, T. (2012). Relation of chronotype to sleep complaints in the general Finnish population. Chronobiology International, 29, 311-317. doi: 10.3109/07420528.2012.655870 Merikanto, I., Paavonen, E. J., Saarenpää-Heikkilä, O., Paunio, T., & Partonen, T. (2017). Eveningness associates with smoking and sleep problems among pregnant women. Chronobiology International, 34, 650-658. doi: 10.1080/07420528.2017.1293085 Miyata, S., Noda, A., Iwamoto, K., Kawano, N., Okuda, M., & Ozaki, N. (2013). Poor sleep quality impairs cognitive performance. Journal of Sleep Research, 22, 535-541. doi: http://dx.doi.org/10.1111/jsr.12054
215
Monk, T. H. (2005). The post-lunch dip in performance. Clinics in Sports Medicine, 24, e15-23, xi-xii. doi: 10.1016/j.csm.2005.12.002 Moorey, S., Frampton, M., & Greer, S. (2003). The cancer coping questionnaire: A self-rating scale for measuring the impact of adjuvant psychological therapy on coping behaviour. Psycho-Oncology, 12, 331-344. doi: 10.1002/pon.646 Morin, L. P. & Allen, C. N. (2005). The circadian visual system, 2005. Brain Research Reviews, 51, 1-60. Muscatell, K. A., Slavich, G. M., Monroe, S. M., & Gotlib, I. H. (2009). Stressful life events, chronic difficulties, and the symptoms of clinical depression. Journal of Nervous Mental Disorders, 197, 154-160. doi:10.1097/NMD.0b013e318199f77b Natale, V., Alzani, A., & Cicogna, P. (2003). Cognitive efficiency and circadian typologies: A diurnal study. Personality and Individual Differences, 35, 1089-1105. National Research Council (US) Committee on Recognition and Alleviation of Distress in Laboratory Animals. Recognition and Alleviation of Distress in Laboratory Animals. Washington (DC): National Academies Press (US); 2008. Committee on Recognition and Alleviation of Distress in Laboratory Animals. Neilsen, T. (2010). Nightmares associated with the eveningness chronotype. Journal of Biological Rhythms, 25, 53-62. doi: 10.1177/0748730409351677 Nordstokke, D. W., Zumbo, B. D., Cairns, S. L., & Saklofske, D. H. (2011). The operating evaluation data. Practical Assessment, Research & Evaluation, 16, 1-8. Available: http://pareonline.net/getvn.asp?v=16&n=5 Northouse, L., Williams, A. L., Given, B., & McCorkle, R. (2012). Psychosocial care for family caregivers of patients with cancer. Journal of Clinical Oncology, 30, 1227-1234. doi: 10.1200/JCO.2011.39.5798 O’Brien, T. B. & DeLongis, A. (1996). The interactional context of problem-, emotion-, and relationship-focused coping: The role of the big five personality factors. Journal of Personality, 64, 775-813. Oken, B. S., Fonareva, I., & Wahbeh, H. (2011). Stress-related cognitive dysfunction in dementia caregivers. Journal of Geriatric Psychiatry and Neurology, 24, 191-198. doi:10.1177/0891988711422524 Opitz, P. C., Lee, I. A., Gross J. J., & Urry, H. L. (2014). Fluid cognitive ability is a resource for successful emotional regulation in older and younger adults. Frontiers in Psychology, 5, 1-13. doi: 10.3389/fpsyg.2014.00609
216
Orzechowska, Zajaczkowska, Talarowska & Galecki, 2013). Depression and ways of coping with stress: A preliminary study. Medical Science Monitor, 19, 1050-1056. doi: 10.12659/MSM.889778 Osse, B. H. P., Vernooij-Dassen, M. J. F. J., Schadé, E., & Grol R. P. T. M. (2006). Problems experienced by the informal caregivers of cancer and their needs for support. Cancer Nursing, 29, 378-388. Ostacoli, L., Cavallo, M., Zuffranieri, M., Negro, M., Sguazzotti, E., Picci, R. L., … Furlan, P. M. (2010). Comparison of experienced burnout symptoms in specialist oncology nurses working in hospital oncology units or in hospices. Palliative & Supportive Care, 8, 427- 432. doi: 10.1017/S1478951510000295 Ottoni, G. L., Antoniolli, E., & Lara, D. R. (2012). Circadian preference is associated with emotional and affective temperaments. Chronobiology International, 29, 786-193. doi: 10.3109/07420528.2012.679329 Pabst, S. R., Negriff, S., Dorn, L. D., Susman, E. J., & Huang, B. (2009). Depression and anxiety in adolescent demales: The impact of sleep preference and body mass index. Journal of Adolescent Health, 44, 554-560. doi: 10.1016/j.jadohealth.2008.11.012 Papantoniou, K., Castaño-Vinyals, G., Espinosa, A., Aragonés, N., Pérez-Gómez, B., Burgos, J., Gómez-Acebo, I., Llorca, J., Peiró, R., Jimenez-Moleón, J. J., Arredondo, F., Tardón, A., Pollan, M., & Kogevinas, M. (2015). Night shift work, chronotype and prostate cancer risk in the MCC-Spain case-control study. International Journal of Cancer, 137, 1147- 1157. doi: 10.1002/ijc.29400 Peres, I., Vetter, C., Blautzik, J., Reiser, M., Pöppel, E., Meindl, T., Roenneberg, T., & Gutyrchik, E. (2011). Chronotype predicts activity patterns in the neural underpinnings of the motor system during the day. Chronobiology International, 28, 883-889. doi: http://dx.doi.org/10.3109/07420528.2011.619084 Portaluppi, F., Riseo, R., Smolensky, M. H., Hermida, R. C., Ayala, D. R., & Fabbian, F. (2012). Circadian rhythms and cardiovascular health. Sleep Medicine Reviews, 16, 151-66. doi: 10.1016/j.smrv.2011.04.003 Potter, P., Deshields, T., Divanbeigi, J., Berger, J., Cipriano, D., Norris, L., & Olsen, S. (2010). Compassion fatigue and burnout: Prevalence among oncology nurses. Clinical Journal of Oncology Nursing, 14 (5), E45-E62. doi: http://dx.doi.org/10.1188/10.CJON.E56-E62 Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385-401. Ralph, M. R., Foster, R.G., Davis, F. C., & Menaker, M. (1990). Transplanted suprachiasmatic nucleus determines circadian period. Science, 247, 975-978. Ray, S. L., Wong, C., White, D. & Heaslip, K. (2013). Compassion satisfaction, compassion fatigue, work life conditions, and burnout among frontline mental health care
217
professionals. Traumatology, 19 (4), 255-267. Reid, K. J., Jaksa, A. A., Eisengart, J. B., Baron, K. G., Lu, B., Kane, P., Kang, J., & Zee, P. C (2012). Systematic evaluation of Axis-I DSM diagnoses in delayed sleep phase disorder and evening-type circadian preference. Sleep Medicine, 13, 1171-1177. doi:10.1016/j.sleep.2012.06.024 Renzi, C., Tabolli, S., Ianni, A., Di Pietro, C., & Puddu, P. (2005). Burnout and job satisfaction comparing healthcare staff of a dermatological hospital and a general hospital. Journal of the European Academy of Dermatology & Venereology, 19, 153–7. doi: 10.1111/j.1468-3083.2005.01029.x Reppert, S. M. & Weaver, D. R. (2001). Molecular analysis of mammalian circadian rhythms. Annual Review of Physiology, 63, 647-676. doi: 10.1146/annurev.physiol.63.1.647 Riley, L. A. & Esser, K. A. (2017). The role of the molecular clock in skeletal muscle and what it is teaching us about muscle-bone crosstalk. Current Osteoporosis Reports, Epub ahead of print. doi: 10.1007/s11914-017-0363-2 Rique, G. L., Fernandes Filho, G. M., Ferreira A. D., & de Sousa-Muñoz, R. L. (2014). Relationship between chronotype and quality of sleep in medical students at the Federal University of Paraiba, Brazil. Sleep Science, 7, 96-102. doi: http://dx.doi.org/10.1016/j.slsci.2014.09.004 Roenneberg, T., Kuehnle, T., Juda, M., Kanermann, T., Allebrandt, K., Gordijn, M., & Merrow, M. (2007). Epidemiology of the human circadian clock. Sleep Medicine Reviews, 11, 429-438. doi: 10.1016/j.smrv.2007.07.005 Roenneberg, T., Wirz-Justice, A., & Merrow, M. (2003). Life between clocks: Daily temporal patterns of human chronotypes. Journal of Biological Rhythms, 18, 80-90. doi: 10.1177/0748730402239679 Rosales-Lagarde, A., Armony, J. L., del Rio-Portilla, Y., Trejo-Martinez, D., Conde R., & Corsi- Cabrera, M. (2012). Enhanced emotional reactivity after selective REM sleep deprivation in humans: An fMRI study. Frontiers in Behavioral Neuroscience, 6, 1-13. doi: 10.3389/fnbeh.2012.00025 Roscoe, J. A., Kaufman, M. E., Matteson-Rusby, S. E., Palesh, O. G., Ryan, J. L., Kohli, S., Perlis, M. L., & Morrow, G. R. (2007). Cancer-related fatigue and sleep disorders. The Oncologist, 12, 35-42. doi: 10.1634/theoncologist.12-S1-35 Rose, D., Gelaye, B., Sanchez, S., Castañeda, B., Sanchez, E., Yanez, D., & Williams, M. A. (2015). Morningness/eveningness chronotype, poor sleep quality, and daytime sleepiness in relation to common mental disorders among Peruvian college students. Psychology of Health Medicine, 20, 345-352. doi: 10.1080/13548506.2014.951367
218
Rospenda, K. M., Minich, L. M., Milner, L. A., & Richman, J. A. (2010). Caregiver burden and alcohol use in a community sample. Journal of Addictive Diseases, 29, 314–324. doi:10.1080/10550887.2010.489450. Russo, J. & Russo, I. H. (2006). The role of estrogen in the initiation of breast cancer. The Journal of Steroid Biochemistry and Molecular Biology, 102 (1-5), 89-96. doi: 10.1016/j.jsbmb.2006.09.004 Ruggieri, V., Zeppegno, P., Gramaglia, C., Gill, S., Deantonio, L., & Krengili, M. (2014). A survey of Italian radiation oncologists: Job satisfaction and burnout. Tumori, 100, 307- 314. doi: 10.1700/1578.17212 Rzeszutek M, Schier K. (2014). Temperament traits, social support, and burnout symptoms in a sample of therapists. Psychotherapy (Chicago, III), 51, 574–9. doi: 10.1037/a0036020 Saita, E., Acquati, C., & Kayser, K. (2015). Coping with early stage breast cancer: Examining the influence of personality traits and interpersonal closeness. Frontiers in Psychology, 6, 1-8. doi: 10.3389/fpsyg.2015.00088 Sardiwalla, N., VandenBerg, H., & Esterhuyse, K. G. F. (2007). The role of stressors and coping strategies in the burnout experienced by hospice workers. Cancer Nursing, 30, 488-497. doi: 10.1097/01.NCC.0000300159.67897.c7 Saroglou, V. (2002). Religion and the five factors of personality: A meta-analytic review. Personality and Individual Differences, 32, 15-25. Savard, J. & Morin, C. M. (2001). Insomnia in the context of cancer: A review of a neglected problem. Journal of Clinical Oncology, 19, 895-908. doi: 10.1200/JCO.2001.19.3.895 Schaufeli, W.B. & Greenglass, E. R. (2001). Introduction to special issue on burnout and health. Psychology & Health, 16, 501-510. doi: 10.1080/08870440108405523 Scheer, F. A. J. L., Morris, C. J., & Shea, S. A. (2013). The internal circadian clock increases hunger and appetite in the evening independent of food intake and other behaviors. Obesity (Silver Spring), 21, 421-423. doi: 10.1002/oby.20351 Shehmar, M. & Gupta, J. K. (2010). The influence of psychological factors on recovery from hysterectomy. Journal of the Royal Society of Medicine, 103, 56-59. doi: 10.1258/jrsm.2009.090198 Schlarb, A. A., Sopp, R., Ambiel, D., & Grünwald, J. (2014). Chronotype-related differences in childhood and adolescent aggression and antisocial behavior – A review of the literature. Chronobiology International, 31, 1-16. doi: 10.3109/07420528.2013.829846 Schmidt, C., Collette, F., Cajochen, C., & Peigneux, P. (2007). A time to think: Circadian rhythms in human cognition. Cognitive Neuropsychology, 24, 755-789. doi: 10.1080/02643290701754158
219
Schmidt, C., Collette, F., Reichert, C. F., Maire, M., Vandewalle, G., Peigneux, P., & Cajochen, C. (2015). Pushing the limits: Chronotype and time of day modulate working memory- dependent cerebral activity. Frontiers in Neurology, 6, 1-9. doi: 10.3389/fneur.2017.00149 Selvi, Y., Aydin, A., Atli, A., Boysan, M., Selvi, F., & Besiroglu, L. (2011). Chronotype differences in suicidal behavior and impulsivity among suicide attempters. Chronobiology International, 28, 170-175. doi: 10.3109/07420528.2010.535938 Severn, M. S., Searchfield, G. D., & Huggard, P. (2012) Occupational stress among audiologists: Compassion satisfaction, compassion fatigue, and burnout. International Journal of Audiology, 51, 3-9. doi: 10.3109/14992027.2011.602366. Sharma, N., Chakrabarti S., & Grover, S. (2016). Gender differences in caregiving among family – caregivers of people with mental illness. World Journal of Psychiatry, 6, 7-17. doi: 10.5498/wjp.v6.i1.7 Sharpe, J. P., Martin, N. R., & Roth, K. A. (2011). Optimism and the big five factors of personality: Beyond neuroticism and extraversion. Personality and Individual Differences, 51, 946–51. doi: https://doi.org/10.1016/j.paid.2011.07.033 Shi, L., Xia, T.-S., Wei, X.-L., Zhou, W., Xue, J., Cheng, L., … Ding, Q. (2015). Estrogen receptor (ER) was regulated by RNPC1 stabilizing mRNA in ER positive breast cancer. Oncotarget, 6, 12264-12278. doi: 10.18632/oncotarget.3654 Shin, H., Park, Y. M., Ying, J. Y., Kim, B., Noh, H., & Lee, S. M. (2014). Relationships between coping strategies and burnout: A meta-analytic approach. Professional Psychology: Research and Practice, 45, 44-56. doi: 10.1037/a0035220 Simpkin, C. T., Jenni, O. G., Carskadon, M. A., Wright Jr., K. P., Akacem, L. D., Garlo, K., & LeBourgeois, M. K. (2014). Chronotype is associated with the timing of the circadian clock and sleep in toddlers. Journal of Sleep Research, 23, 397-405. doi: 10.1111/jsr.12142 Singh, Y. & Sharma, R. (2012). Relationship between general intelligence, emotional intelligence, stress levels and stress reactivity. Annals of Neurosciences, 19, 107-111. doi: 10.5214/ans.0972.7531.190304 Sklenarova, H., Krümpelmann, A., Haun, M. W., Friederich, H.-C., Huber, J., Thomas, M., Winkler, E. C., Herzog, W., & Hartmann, M. (2015). When do we need to care about the caregiver? Suppotive care needs, anxiety, and depression among informal caregivers of patients with cancer and cancer survivors. Cancer, 121, 1513-1519. Sodeke-Gregson, E. A., Holttum, S., & Billings, J. (2013). Compassion satisfaction, burnout, and secondary traumatic stress in UK therapists, who work with adult trauma clients. Journal of Psychotraumatology, 4, eCollection. doi: 10.3402/ejpt.v4i0.21869.
220
Söderström, M., Ekstedt, M., Åkerstedt, T., Nilsson, J., & Axelsson, J. (2004). Sleep and sleepiness in young individuals with high burnout scores. Sleep, 27, 1369-1377. Soehner, A. M., Kennedy, K. S., & Monk, T. H. (2011). Circadian preference and sleep-wake regularity: Associations with self-report sleep parameters in daytime-working adults. Chronobiology International, 28, 802-809. doi:10.3109/07420528.2011.613137 Spahni, S., Morselli, D., Perrig-Chiello, P. & Bennet, K. M. (2015). Patterns of psychological adaptation to spousal bereavement in old age. Gerontology, 61, 456-68. doi: 10.1159/000371444 Spector PE. (1985). Measurement of human service staffsatisfaction: Development of the job satisfaction survey. American Journal of Community Psychology, 13, 693–713 Stamm, B. H. (2002). Measuring compassion satisfaction as well as fatigue: Developmental history of the compassion satisfaction and fatigue test. In C. R. Figley (Ed.). Treating compassion fatigue (pp. 107-119). New York: Brunner-Routledge. Stamm, B. H. (2010). The Concise ProQol Manual, 2nd Ed. Pocatello, ID: ProQOL.org. Staner, L. (2003). Sleep and anxiety disorders. Dialogues in Clinical Neuroscience, 5, 249-258. Stavitsky, K., Neargarder, S., Bogdanova, Y., McNamara, P., & Cronin-Golomb, A. (2012). The impact of sleep quality on cognitive functioning in Parkinson’s disease. Journal of the International Neuropsychological Society, 18, 108–117. doi: 10.1017/S1355617711001482. Stepanski, E. J. (2002). The effect of sleep fragmentation on daytime function. Sleep, 25, 268- 276. Stephan, F. K., Schwann, J. M, & Sisk, C. L. (1979). Anticipation of 24-hour feeding schedules in rats with lesions in the suprachiasmatic nucleus. Behavioral and Neural Biology, 25 (3), 346-363. Sullivan, A. (2002). Gender differences in coping strategies of parents of children with Down syndrome. Down Syndrome Research and Practice, 8, 67-73. Tavernier, R., Munroe, M., & Willoughby, T. (2015). Perceived morningness-eveningness predicts academic adjustment and substance use across university, but social jetlag is not to blame. Chronobiology International, 32, 1233-1245. doi: 10.3109/07420528.2015.1085062 Tempesta, D., Couyoumdjian, A., Curcio, G., Moroni, F., Marzano, C., De Gennaro, L., and Ferrara, M. (2010). Lack of sleep affects the evaluation of emotional stimuli. Brain Research Bulletin, 82, 104-108. doi: 10.1016/j.brainresbull.2010.01.014
221
Thomas, K., Bower, J., Hoyt, M. A., & Sepah, S. (2010). Disrupted sleep in breast and prostate cancer patients undergoing radiation therapy: The role of coping processes. Psychooncology, 19, 767-776. doi: 10.1002/pon.1639 Tsai, F. J., Huang, W. L. & Chan, C. (2009). Occupational stress and burnout of lawyers. Journal of Occupational Health, 51(5), 443-450. Turiano, N. A., Whiteman, S. D., Hampson, S. E., Roberts, B. W., & Mroczek, D. (2012). Personality and substance use in midlife: Conscientiousness as a moderator and the effects of trait change. Journal of Research in Personality, 46, 295-305. doi:10.1016/j.jrp.2012.02.009 Underwood, B. & Moore, B. (1982). Perspective-taking and altruism. Psychological Bulletin, 91, 143-173. Vaessen, T. J. A., Overeem, S., & Sitskoorn, M. M. (2015). Cognitive complaints in obstructive sleep apnea. Sleep Medicine Reviews, 19, 51-58. doi: 10.1016/j.smrv.2014.03.008 Vahey, D. C., Aiken, L. H., Sloane, D. M., Clarke, S. P., & Vargas, D. (2004). Nurse burnout and patient satisfaction. Medical Care, 42 (2 Suppl), II57-66. doi: 10.1097/01.mlr.0000109126.50398.5a Vandekerckhove, M. & Cluydts, R. (2010). The emotional brain and sleep: An intimate relationship. Sleep Medicine Reviews, 14, 219-226. doi: 10.1016/j.smrv.2010.01.002 Vanwesenbeeck, I. (2005). Burnout among female indoor sex workers. Archives of Sexual Behavior, 34 (6), 627-639. doi: 10.1007/s10508-005-7912-y Van Dongen, H. P. A., Maislin, G., Mullington, J. M., & Dinges, F. (2003). The cumulative cost of additional wakefulness: Dose-response effects on neurobehavioral dunctions and sleep physiology from chronic sleep restriction and total sleep deprivation. Sleep, 26, 117-126. van Litsenburg, R.R., Huisman, J., Hoogerbrugge, P. M., Egeler, R. M., Kaspers, G. J., & Gemke, R. J. (2011). Impaired sleep affects quality of life in children during maintenance treatment for acute lymphoblastic lymphoma: an exploratory study. Health and Quality of Life Outcomes, 9, 25-31. doi: 21496357 http://dx.doi.org/10.1186/1477-7525-9-25 Vardar, E.,Vardar, S. A., Molla, T., Kaynak, C., & Ersoz, E. (2008). Psychological symptoms and sleep quality in young subjects with different circadian preferences. Biological Rhythms Research, 39, 493-500. doi: 10.1080/09291010701736884 Vargas, S., Wohlgemuth, W. K., Antoni, M. H., Lechner, S. C., Holley H. A., & Carver, C. S. (2010). Sleep dysfunction and psychosocial adaptation among women undergoing treatment for non-metastatic breast cancer. Psychooncology, 19, 669-673, doi: 10.1002/pon.1603 Vela-Bueno, A., Moreno-Jiménez, B., Rodríguez-Muñoz, A., Olavarrieta-Bernardino, S., Fernández-Mendoza, J., De la Cruz-Troca, J. J. Bixler, E. O., & Vgontzas, A. N. (2008).
222
Insomnia and sleep quality among primary care physicians with low and high burnout levels. Journal of Psychosomatic Research, 64, 435-442. doi: 10.1016/j.jpsychores.2007.10.014 Vitale, J. A., Colagiuri, G., & Weydahl, A. (2013). Influence of chronotype on responses to a standardized, self-paced walking task in the morning vs. afternoon: A pilot study. Perceptual & Motor Skills: Exercise & Sport, 116, 1020-1028. doi: 10.2466/06.19.PMS.116.3.1020-1028 Vitale, J. A., Roveda, E., Montaruli, A., Galasso, L., Weydahl, A., Caumo, A., & Carandente, F. (2015). Chronotype influences activity circadian rhythm and sleep: differences in sleep quality between weekdays and weekend. Chronobiology International, 32 (3), 405-415. Vollrath, M., Torgersen, S., & Alnaes, R. (1998). Neuroticism, coping and change in MCM-II clinical syndromes: Test of a mediator model. Scandinavian Journal of Psychology, 39, 15-24. Wang, X. M., Walitt, B., Saligan, L., Tiwari, A. F. Y., Cheung C. W., & Zhang, Z. J. (2015). Chemobrain: A critical review and causal hypothesis of link between cytokines and epigenetic. Cytokine, 72, 86-96. doi: 10.1016/j.cyto.2014.12.006 Westermann, C., Kozak, A., Harling, M., & Nienhaus, A. (2014). Burnout intervention studies for inpatient elderly care nursing staff: systematic literature review. International Journal of Nursing Studies, 51, 63-71. doi: 10.1016/j.ijnurstu.2012.12.001 Wong, M. W., Roberson, G., & Dyson, R. (2015). Prospective relationship between poor sleep and substance-related problems in a national sample of adolescents. Alcoholism, Clinical and Experimental Research, 39, 355-362. doi: 10.1111/acer.12618 Wright, K. P., Lowry, C. A., & LeBourgeois, M. K. (2012). Circadian and wakefulness-sleep modulation in humans. Frontiers in Molecular Neuroscience, 5, 1-12. doi: 10.3389/fnmol.2012.00050 Yun, J. A., Ahn, Y. S., Jeong, K. S., Joo, E. J., & Choi, K. S. (2015). The relationship between chronotype and sleep quality in Korean firefighters. Clinical Psychopharmacology and Neuroscience, 13, 201-208. doi: http://dx.doi.org/10.9758/cpn.2015.13.2.201 Zaza, C., Sellick, S. M., & Hillier, L. M. (2005). Coping with cancer: What do patients do? Journal of Psychosocial Oncology, 23, 55-73, doi:10.1300/J077v23n01_04 Zhang, Q., Yao, D., Yang, J., & Zhou, Y. (2014). Factors influencing sleep disturbances among caregivers of cancer patients in northeast China. PLOS One, e108614. doi: 10.1371/journal.pone.0108614 Zhang, L., Zhao, J., Xiao, H., Zheng, H., Xiao, Y., Chen, M. & Chen, D. (2014). Mental health and burnout in primary and secondary school teachers in the remote mountain areas of Guangdong Province in the People’s Republic of China. Neuropsychiatric Disease and Treatment, 10, 123-130. doi: 10.2147/NDT.S56020