Download Post-treatment adjustment and behavior change among women with

University of Iowa
Iowa Research Online
Theses and Dissertations
2006
Post-treatment adjustment and behavior change
among women with breast cancer
Erin Susan Costanzo
University of Iowa
Copyright 2006 Erin Susan Costanzo
This dissertation is available at Iowa Research Online: http://ir.uiowa.edu/etd/56
Recommended Citation
Costanzo, Erin Susan. "Post-treatment adjustment and behavior change among women with breast cancer." PhD (Doctor of
Philosophy) thesis, University of Iowa, 2006.
http://ir.uiowa.edu/etd/56.
Follow this and additional works at: http://ir.uiowa.edu/etd
Part of the Psychology Commons
POST-TREATMENT ADJUSTMENT AND BEHAVIOR CHANGE AMONG
WOMEN WITH BREAST CANCER
by
Erin Susan Costanzo
An Abstract
Of a thesis submitted in partial fulfillment
of the requirements for the Doctor of
Philosophy degree in Psychology
in the Graduate College of
The University of Iowa
July 2006
Thesis Supervisor: Professor Susan K. Lutgendorf
1
ABSTRACT
Anecdotal and qualitative evidence suggests that women may experience
disrupted adjustment during the months following the end of adjuvant breast cancer
treatment, in part due to the loss of a “safety net” associated with regular treatment
coupled with uncertainty regarding cancer status. The present study examined distress
and quality of life, as well as behavioral and cognitive predictors of adjustment, during
the three months following adjuvant treatment for breast cancer. Participants were 89
women with breast cancer who completed measures of distress, quality of life, health
behavior, behavior changes, and common-sense beliefs about cancer at three time points:
toward the end of adjuvant treatment, 3 weeks following the end of treatment, and 3
months post-treatment. Findings indicated that breast cancer survivors were remarkably
well-adjusted following treatment: participants reported low levels of anxiety and
depression and good health-related quality of life. Nonetheless, women acknowledged
significant concerns about ongoing physical symptoms, potential recurrence, and the
process of returning to or building a “new normal.” Results further suggested that
behavior changes were quite common after the end of treatment, particularly positive
changes in health practices. Although good health practices were associated with better
adjustment, making positive changes in the same behaviors often predicted greater
distress. Women’s common-sense beliefs about breast cancer provided insight into
whether women decided to make behavior changes and what behaviors they decided to
change. Women who perceived greater control over their cancer, saw their cancer as an
acute rather than chronic condition, and attributed cancer to controllable causes or
believed that behavioral or psychological factors could prevent recurrence were more
2
likely to make behavior changes and engage in positive health practices. Although it was
predicted that beliefs and behavior changes would interact to predict distress, no
consistent pattern of interactions was found. In sum, breast cancer patients actively
attempt to create a “new normal” following treatment, and changes in health practices
appear to be an important part of this process. Assessing women’s beliefs about their
cancer and providing psychoeducational interventions addressing post-treatment behavior
changes may assist in promoting breast cancer survivors’ psychological and physical
well-being.
Abstract Approved:
Thesis Supervisor
Title and Department
Date
POST-TREATMENT ADJUSTMENT AND BEHAVIOR CHANGE AMONG
WOMEN WITH BREAST CANCER
by
Erin Susan Costanzo
A thesis submitted in partial fulfillment
of the requirements for the Doctor of
Philosophy degree in Psychology
in the Graduate College of
The University of Iowa
July 2006
Thesis Supervisor: Professor Susan K. Lutgendorf
i
Graduate College
The University of Iowa
Iowa City, Iowa
CERTIFICATE OF APPROVAL
_______________________
PH.D. THESIS
_______________
This is to certify that the Ph.D. thesis of
Erin Susan Costanzo
has been approved by the Examining Committee
for the thesis requirement for the Doctor of Philosophy
degree in Psychology at the July 2006 graduation.
Thesis Committee: ___________________________________
Susan K. Lutgendorf, Thesis Supervisor
___________________________________
Alan Christensen
___________________________________
Michael O’Hara
___________________________________
Susan Roman
___________________________________
Jerry Suls
ACKNOWLEDGMENTS
I would like to acknowledge my mentor and dissertation committee chair, Dr.
Susan Lutgendorf, for valuable advice and guidance through all stages of the project. I
am also grateful to current and past members of my dissertation committee, including
Drs. Alan Christensen, Rene Martin, Michael O’Hara, Susan Roman, Jerry Suls, and
Shruti Trehan for helpful feedback on the study design, implementation, and
interpretation of results. In addition, I would like to acknowledge Mary Mattes, Lisa
Bergman, Heena Maiseri, Hetal Pandya, and Sara Sheerer for assistance with enrollment
of participants, data collection, and data entry, Matt Bauer for technical assistance and
programming, and Drs. Justine Richie and Scott Baldwin for statistical advice. I would
also like to express my gratitude to physicians and health care staff who provided
assistance with participant identification and enrollment, including Carolene Robinson,
Mary Fitzgerald, Anne Fleagle, Judy Gilliam, Jeanne Hein, Dr. Mark Karwal, Dr. Lisa
Simmons, and Dr. Ferial Tewfik. Finally, I am grateful to all of the women who
participated in the study.
This work was funded by the Barbara Rosenblum Scholarship for the Study of
Women and Cancer from the Sociologists for Women in Society, a graduate research
scholarship from the American Psychological Foundation/Council of Graduate
Departments of Psychology, and a research grant from the University of Iowa Student
Government.
ii
TABLE OF CONTENTS
LIST OF TABLES.......................................................................................................... v
LIST OF FIGURES ...................................................................................................... vii
CHAPTER I. INTRODUCTION .................................................................................... 1
Breast Cancer ............................................................................................... 1
Breast Cancer Survivorship Issues................................................................ 2
The Re-Entry Period..................................................................................... 6
Strategies for “Reconstructing the Safety Net”............................................ 14
Common-Sense Representations of Illness.................................................. 22
Common-Sense Representations of Cancer................................................. 24
Summary and Proposed Study .................................................................... 33
CHAPTER II. OBJECTIVES AND HYPOTHESES..................................................... 37
Objective 1................................................................................................. 37
Objective 2................................................................................................. 37
Objective 3................................................................................................. 38
Objective 4................................................................................................. 38
CHAPTER III. METHOD............................................................................................. 40
Participants................................................................................................. 40
Procedure ................................................................................................... 40
Measures .................................................................................................... 42
Statistical Analyses..................................................................................... 52
Power Considerations................................................................................. 60
CHAPTER IV. RESULTS ............................................................................................ 62
Participants................................................................................................. 62
Objective 1: Distress and Health-Related Quality of Life............................ 64
Objective 2: Behavioral Coping Strategies.................................................. 68
Objective 3: Common-Sense Models of Breast Cancer and Coping
Behavior..................................................................................................... 72
Objective 4: Interactions Between Common-Sense Beliefs and
Behavior and Distress Outcomes ................................................................ 80
CHAPTER V. DISCUSSION ....................................................................................... 87
Post-Treatment Distress and Quality of Life ............................................... 87
Creating a “New Normal”: Post-Treatment Behavior Changes.................... 94
Post-Treatment Behavior Changes as Coping Strategies ............................. 97
Common-Sense Models of Breast Cancer ..................................................100
Common-Sense Beliefs and Behavior Changes..........................................104
Interactions Between Common-Sense Beliefs and Behavior and
Distress Outcomes.....................................................................................110
Limitations ................................................................................................115
Summary and Implications for Theory and Practice...................................117
iii
REFERENCES ............................................................................................................121
APPENDIX A. TABLES .............................................................................................136
APPENDIX B. FIGURES............................................................................................168
APPENDIX C. SELECTED MEASURES ...................................................................188
iv
LIST OF TABLES
Table A1.
ACS Recommendations for Nutrition and Physical Activity for Cancer
Prevention ...............................................................................................136
Table A2.
AICR Diet and Health Guidelines for Cancer Prevention.........................137
Table A3.
ACS Advice For Reducing the Risk of Breast Cancer ..............................138
Table A4.
Measurement Timeline ............................................................................139
Table A5.
Sample Demographics .............................................................................140
Table A6.
Sample Disease and Treatment Characteristics ........................................141
Table A7.
Mean Distress and Quality of Life Scores Over Time ..............................142
Table A8.
Mean Ratings of Sources of Stress Over Time .........................................143
Table A9.
Post-Treatment Behavior Changes ...........................................................144
Table A10. Means and Percentages of Health and Psychosocial Behaviors ................145
Table A11. Physical Activity 3 Weeks Post-Treatment as a Predictor of CARS
Recurrence Worry 3 Months Post-Treatment ...........................................146
Table A12. Mean Scores on Measures of Common-Sense Models of Breast
Cancer .....................................................................................................147
Table A13. Causal Attributions ..................................................................................148
Table A14. Recurrence Prevention Beliefs .................................................................149
Table A15. Factor Loadings for Behavioral/Psychological Attributions Factors .........150
Table A16. Chronicity Beliefs 3 Weeks Post-Treatment Predict Fat Consumption
3 Months Post-Treatment.........................................................................151
Table A17. Attribution-Behavior Relationships Tested ..............................................152
Table A18. Behavioral and Psychological Attribution at Baseline Predicts Alcohol
Use 3 Weeks Post-Treatment ...................................................................153
Table A19. Attributing Cancer to Substance Use at Baseline Predicts Alcohol Use
3 Weeks Post-Treatment ..........................................................................154
Table A20. Personal Control and Fat Consumption Interact to Predict IES
Intrusion at 3 Months Post-Treatment .....................................................155
Table A21. Chronicity Belief and Change in Avoidance of Stress Interact to
Predict IES Intrusion at 3 Months Post-Treatment....................................156
v
Table A22. Chronicity Belief and Physical Activity Interact to Predict PRIMEMD Anxiety at 3 Months Post-Treatment ................................................157
Table A23. Chronicity Belief and Physical Activity at 3 Weeks Post-Treatment
Interact to Predict PRIME-MD Anxiety at 3 Months Post-Treatment.......158
Table A24. Causal Attributions and Change in Avoidance of Stress Interact to
Predict CES-D Depression at 3 Weeks Post-Treatment ............................159
Table A25. Causal Attributions and Frequency of Breast Self-Exam Interact to
Predict PRIME-MD Anxiety at 3 Weeks Post-Treatment.........................160
Table A26. Causal Attributions and Change in Avoidance of Stressful Situations
Interact to Predict IES Intrusion at 3 Months Post-Treatment...................161
Table A27. Causal Attributions and Alcohol Use Interact to Predict CES-D
Depression at 3 Months Post-Treatment...................................................162
Table A28. Causal Attributions and Change in Fruit and Vegetable Consumption
at 3 Weeks Post-Treatment Interact to Predict IES Intrusion at 3
Months Post-Treatment............................................................................163
Table A29. Causal Attributions and Change in Avoidance of Stressful Situations
at 3 Weeks Post-Treatment Interact to Predict IES Intrusion at 3
Months Post-Treatment............................................................................164
Table A30. Causal Attributions and Fat Consumption at 3 Weeks Post-Treatment
Interact to Predict IES Intrusion at 3 Months Post-Treatment...................165
Table A31. Causal Attributions and Physical Activity at 3 Weeks Post-Treatment
Interact to Predict PRIME-MD Anxiety at 3 Months Post-Treatment.......166
Table A32. Beliefs about Recurrence Prevention and Physical Activity Interact to
Predict PRIME-MD Anxiety at 3 Weeks Post-Treatment.........................167
vi
LIST OF FIGURES
Figure B1. Model of adjustment during the 3 months following the end of breast
cancer treatment ........................................................................................168
Figure B2. Marginally significant relationships between personal control and
behavioral outcomes over time..................................................................169
Figure B3. Significant and marginally significant relationships between beliefs
about the chronicity of one’s cancer and behavioral outcomes over time ...170
Figure B4. Significant and marginally significant relationships between attributing
cancer to behavioral and psychological causes and behavioral outcomes
over time...................................................................................................171
Figure B5. Significant and marginally significant relationships between attributing
cancer to diet and behavioral outcomes over time......................................172
Figure B6. Significant and marginally significant relationships between attributing
cancer to lack of exercise and behavioral outcomes over time ...................173
Figure B7. Significant and marginally significant relationships between believing
that behavioral and psychological factors can prevent a recurrence and
behavioral outcomes over time..................................................................174
Figure B8. Significant and marginally significant relationships between believing
that decreasing or quitting alcohol or tobacco use can prevent a
recurrence and behavioral outcomes over time ..........................................175
Figure B9. Significant and marginally significant relationships between believing
that decreasing stress in one’s life can prevent a recurrence and
behavioral outcomes over time..................................................................176
Figure B10. Interaction between personal control and fat consumption in predicting
intrusion at 3 months post-treatment, β = .26, p = .03 ................................177
Figure B11. Interaction between chronicity belief and change in avoidance of
stressful situations in predicting intrusion at 3 weeks post-treatment, β
= .40, p = .006...........................................................................................178
Figure B12. Interaction between chronicity belief and physical activity at 3 weeks
post-treatment in predicting anxiety at 3 months post-treatment, β =
.23, p = .04................................................................................................179
Figure B13. Interaction between attributing cancer to behavioral and psychological
causes and change in avoidance of stressful situations in predicting
depression at 3 weeks post-treatment, β = .29, p = .01 ...............................180
Figure B14. Interaction between attributing cancer to behavioral and psychological
causes and breast self-exam in predicting anxiety at 3 weeks posttreatment, β = -.28, p = .02 ........................................................................181
vii
Figure B15. Interaction between attributing cancer to behavioral and psychological
causes and change in avoidance of stressful situations in predicting
intrusion at 3 months post-treatment, β = -.50, p = .002.............................182
Figure B16. Interaction between attributing cancer to behavioral and psychological
causes and alcohol use in predicting depression at 3 months posttreatment, β = -.26, p = .039 ......................................................................183
Figure B17. Interaction between attributing cancer to behavioral and psychological
causes and change in fruit/vegetable consumption at 3 weeks posttreatment in predicting intrusion at 3 months post-treatment, β = -.20, p
= .045........................................................................................................184
Figure B18. Interaction between attributing cancer to behavioral and psychological
causes and fat consumption at 3 weeks post-treatment in predicting
intrusion at 3 months post-treatment, β = .22, p = .019 ..............................185
Figure B19. Interaction between attributing cancer to behavioral and psychological
causes and physical activity at 3 weeks post-treatment in predicting
anxiety at 3 months post-treatment, β = -.23, p = .046 ...............................186
Figure B20. Interaction between believing that behavioral and psychological factors
can prevent recurrence and physical activity in predicting anxiety at 3
weeks post-treatment, β = .32, p = .006 .....................................................187
viii
1
CHAPTER I
INTRODUCTION
Breast Cancer
Breast cancer is a highly prevalent medical problem, with current data suggesting
that more than 1 in 8 women will develop breast cancer within their lifetimes (American
Cancer Society, 2006a). It is the most frequently diagnosed cancer among women,
accounting for nearly one-third of cancer cases. Each year, 211,240 women develop
invasive breast cancer and 58,490 develop in situ (localized) breast cancer. Although
98% of women with localized cancer are expected to survive 5 years or more, survival
rates decrease for those with regional (81%) and metastatic breast cancer (26%)
(American Cancer Society, 2006a, 2006b). Despite the relatively high survival rates,
breast cancer is the second leading cause of cancer death among women after lung
cancer, with 40,410 deaths from the disease each year. Breast cancer rates have increased
over the past 20 years but mortality has decreased, in part due to screening and early
detection (American Cancer Society, 2006a).
Treatment for breast cancer is selected based on stage, biological characteristics
of the cancer, and patient preferences (American Cancer Society, 2006a; Beenken &
Bland, 2002; Hortobagyi, 1998; National Cancer Institute, 2006; National
Comprehensive Cancer Network, 2005). Most women with breast cancer undergo either
lumpectomy followed by radiation therapy or simple or total mastectomy. Noninvasive
ductal carcinoma in situ (stage 0) may be treated with lumpectomy plus radiation or
mastectomy. A less frequent form, lobular carcinoma in situ, is simply observed over
time. Localized or regional invasive breast cancer (stages I-IIIa) that has not spread
beyond the breast or axillary lymph nodes is generally treated with lumpectomy plus
radiation therapy or mastectomy. Depending on the extent and biological characteristics
of the cancer, patients may also be treated with chemotherapy. Chemotherapy is generally
2
used to treat both regionally advanced (spread beyond the breast and axillary nodes) and
metastatic breast cancer. Some women receive radiation therapy or chemotherapy prior to
surgery, or neoadjuvant treatment, in order to shrink a bulky tumor and allow for less
extensive surgery. Finally, women who have cancers with estrogen receptors (called “ER
positive”) may take Tamoxifen or other anti-estrogen agents following initial treatment,
which reduces the likelihood of recurrence by 26% and death by 14% (Early Breast
Cancer Trialists' Collaborative Group, 1998).
Breast cancer treatment is generally accompanied by unpleasant side effects.
Physical side effects of surgery include swelling in the arm due to fluid retention
(Bumpers, Best, Norman, & Weaver, 2002; Manusell, Brisson, & Deschenes, 1993) and
pain. Radiation side effects include fatigue, skin changes, breast changes, breast soreness,
difficulty swallowing, swelling due to fluid retention, and shoulder stiffness (National
Cancer Institute, 2000). Chemotherapy side effects include fatigue, nausea and vomiting,
hair loss, infections, anemia, pain, gastrointestinal problems, cognitive dysfunction,
peripheral neuropathy, menopausal symptoms, and infertility (National Cancer Institute,
1999). Psychological effects of treatment for breast cancer have also been wellcharacterized and include depression and anxiety, poor body image, problems with sexual
functioning, and disruption of social relationships (Moyer & Salovey, 1996; Schou,
Ekeberg, Sandvik, Hjermstad, & Ruland, 2005).
Breast Cancer Survivorship Issues
Due to the efficacy of breast cancer treatment as well as early detection and
screening methods, the majority of women today (88%) are surviving at least 5 years
after diagnosis (American Cancer Society, 2006a). Important quality of life issues
(referring to physical, functional, psychological, and social well-being) for breast cancer
survivors have been identified in the literature.
3
There is some evidence that decrements in quality of life and physical functioning
associated with treatment may extend well beyond the period of initial treatment (Gotay
& Muraoka, 1998). A large, epidemiological study found that cancer survivors had lower
scores on the SF-36, a measure of health-related quality of life, in both physical and
mental health domains as compared to individuals who had never had cancer (Baker,
Haffer, & Denniston, 2003; Dorval, Maunsell, Deschenes, Brisson, & Masse, 1998).
Another study found similar results, with breast cancer survivors who had received
adjuvant chemotherapy scoring lower on most SF-36 subscales as compared to women
who had never had cancer (Broeckel, Jacobsen, Balducci, Horton, & Lyman, 2000).
There is also evidence that breast cancer survivors report more somatic concerns as
compared to survivors of other types of cancer (Kurtz, Wyatt, & Kurtz, 1995). In
contrast, however, other work has found that breast cancer survivors were similar to the
general population or healthy controls with respect to health-related quality of life (Fan et
al., 2005; Ganz, Rowland, Desmond, Meyerowitz, & Wyatt, 1998; Helgeson & Tomich,
2005).
Specific physical concerns of breast cancer survivors include early menopause,
loss of fertility, hair loss, nail problems, osteoporosis, pain, fatigue, insomnia, weight
gain, and loss of one’s breast (Beisecker et al., 1997; Ferrell, Grant, Funk, Otis-Green, &
Garcia, 1997). Lymphedema (swelling of the arm due to fluid retention) may also persist
and can be disabling (Bumpers, Best, Norman, & Weaver, 2002; Voogd et al., 2003). In
many cancer survivors, fatigue does not improve significantly over time, with one-third
of individuals who have survived more than 5 years from diagnosis continuing to report
substantial fatigue (Bower et al., 2006). Overall, 37% of cancer survivors reported at least
two weeks of fatigue during the previous month (Cella, Davis, Breitbart, & Curt, 2001).
Difficulties associated with sexuality and intimate relationships, including poor
body image, lack of sexual interest, sexual dysfunction, and dating difficulties, are some
of the most frequently-reported problems among breast cancer survivors (Ganz et al.,
4
1996). Overall, the evidence regarding the extent of sexual dysfunction among breast
cancer survivors is inconsistent. Some studies have found diminution of sexual
satisfaction or functioning among breast cancer survivors as compared to the general
population (Dorval, Maunsell, Deschenes, Brisson, & Masse, 1998), while others have
found no differences (Ganz, Rowland, Desmond, Meyerowitz, & Wyatt, 1998;
Meyerowitz, Desmond, Rowland, Wyatt, & Ganz, 1999). Overall, it appears that at least
a subset of breast cancer survivors experience notable problems with sexual functioning.
For example, although one study found no overall differences in sexual functioning
between women who did and did not have breast cancer, one-third of the breast cancer
survivors reported that breast cancer had negatively impacted their sexual functioning
(Meyerowitz, Desmond, Rowland, Wyatt, & Ganz, 1999). In addition, chemotherapy and
hormonal treatments such as Tamoxifen can induce menopause and associated symptoms
(Goodwin, Ennis, Pritchard, Trudeau, & Hood, 1999), and women who receive
chemotherapy treatment or become menopausal after treatment appear to be at the
greatest risk for sexual dysfunction (Ganz, Desmond, Belin, Meyerowitz, & Rowland,
1999; Ganz, Rowland, Desmond, Meyerowitz, & Wyatt, 1998; Meyerowitz, Desmond,
Rowland, Wyatt, & Ganz, 1999).
Mild cognitive impairment is another common problem among breast cancer
survivors. Chemotherapy may be neurotoxic, causing decrements in cognitive abilities
including memory, concentration, and psychomotor functioning that persist beyond the
end of treatment. Cognitive deficits have been found at the time of chemotherapy
(Brezden, Phillips, Abdolell, Bunston, & Tannock, 2000), 2 years after treatment
(Schagen et al., 1999; van Dam et al., 1998), and 5 years after treatment (Ahles et al.,
2002). Twenty to 30% of breast cancer survivors who received standard chemotherapy
show overall cognitive impairment (V. Jenkins et al., 2006; Schagen et al., 1999; van
Dam et al., 1998).
5
Breast cancer survivors may also confront economic problems. In one study, 11%
of cancer survivors surveyed reported being denied health or life insurance and 18%
experienced work-related problems (M. Hewitt, Breen, & Devesa, 1999). Despite reports
of work-related problems, the majority of cancer survivors who were employed prior to
their diagnosis were also employed 5 to 7 years post-diagnosis, and only one-fourth of
the patients who had stopped working did so because of health-related problems or
disability (Bradley & Bednarek, 2002).
With respect to psychological and social functioning, breast cancer survivors
appear to be remarkably well-adjusted (e.g., Kurtz, Wyatt, & Kurtz, 1995), often
reporting comparable psychological adjustment to individuals who never had cancer. A
1996 review of the literature on adjustment in breast cancer concluded that while older
studies showed that breast cancer survivors had high levels of anxiety and depression,
more recent studies indicate lower levels of distress that are similar to those of the
general population (Moyer & Salovey, 1996). Modern treatment procedures and a more
open attitude toward cancer may be responsible for this shift. Other recent studies have
supported this trend. For example, a study of 864 breast cancer survivors found that
survivors were no worse off than the general population with respect to depression,
marital functioning, sexual functioning, and health-related quality of life (Ganz,
Rowland, Desmond, Meyerowitz, & Wyatt, 1998). Another study found small, nonsignificant quality-of-life differences between long-term breast cancer survivors and
women who had not had cancer, but differences disappeared when women who had
recurrences were removed from the sample (Dorval, Maunsell, Deschenes, Brisson, &
Masse, 1998). Breast cancer survivors also appear to be functioning particularly well as
compared to patients with other chronic health problems (Ganz et al., 1996).
Despite evidence that breast cancer survivors are generally doing well
psychosocially, there may be some domains that remain troublesome for survivors. While
mood and role functioning improved in mastectomy patients from just after surgery to 18
6
months post-surgery, symptom-related distress did not improve (Northouse, 1989). In
addition to symptom-related distress, cancer-related worries including concerns about
death, fear of recurrence, and worry about impact of cancer on the family have been
reported by breast cancer patients (Ferrell, Grant, Funk, Otis-Green, & Garcia, 1997;
Lampic et al., 1994). Intrusive thoughts also appear to be common among breast cancer
survivors (Walker, Nail, Larsen, Magill, & Schwartz, 1996). Clinical levels of PTSD are
less common, but in one study female cancer survivors were somewhat more likely to
meet lifetime criteria for PTSD than were women had never had cancer (6 of 27 survivors
versus no controls). However, by 3 years post-diagnosis, PTSD rates were minimal (Alter
et al., 1996). Similar results were found in other studies of breast cancer survivors, with
PTSD rates around 6% (M. A. Andrykowski & Cordova, 1998). It also appears that
anxiety and cancer-related worry may diminish to almost normal levels 2 or 3 years after
diagnosis. Lampic and colleagues (1994) found that only one-fifth of cancer survivors
reported moderate or severe anxiety, and anxiety levels were particularly low among
patients who had completed treatment more than two years previously. Another recent
study found low levels of cancer-related worry among breast cancer survivors, with
women who were more than 5 years post-diagnosis reporting the lowest levels of worry
(Rothrock, 2002).
The Re-Entry Period
Despite the relatively good psychosocial adjustment of breast cancer survivors
overall, there may be periods of disrupted adjustment and elevated distress. The period
immediately following the end of adjuvant treatment may be one such time. Case studies
and other anecdotal evidence in the literature suggest that the months immediately
following the end of adjuvant treatment, labeled the “re-entry period” by Stanton and
colleagues (2005), is a time of disruption, transition, and increased distress (Lethborg,
7
Kissane, Burns, & Snyder, 2000; Mullan, 1985; Rowland & Holland, 1990; Schnipper,
2001; Ward, Viergutz, Tormey, deMuth, & Paulen, 1992).
Evidence for disrupted adjustment during the re-entry
period
Common sense might suggest that the end of cancer treatment should be a time of
joy and celebration. However, Stanton and colleagues (2005) debunk this notion as one
of several myths associated with the completion of treatment. Although many women do
report a sense of relief and other positive emotions at the end of treatment (e.g., Beisecker
et al., 1997; Lethborg, Kissane, Burns, & Snyder, 2000), many experience significant
distress as well (e.g., Beisecker et al., 1997; Lampic et al., 1994). A physician described
her emotions at her last treatment (McKinley, 2000):
After my very last radiation treatment for breast cancer, I
lay on a cold steel table hairless, half-dressed, and astonished by
the tears streaming down my face. I thought I would feel happy
about finally reaching the end of treatment, but instead, I was
sobbing. . . . Ironically, I also cried because I would not be coming
back to that familiar table where I had been comforted and
encouraged. Instead of joyous, I felt lonely, abandoned, and
terrified. This was the rocky beginning of cancer survivorship for
me. (p. 479)
There is conflicting evidence regarding the level of distress during the re-entry
period. Among early-stage breast cancer patients, mood disturbance decreased from
diagnosis to 3 months post-surgery to 12 months post-surgery (Stanton, Danoff-Burg, &
Huggins, 2002). Similarly, Carver and colleagues (1993) found that distress decreased
from diagnosis to a week post-surgery and remained low and relatively stable through 3,
6, and 12 months post-surgery. Other studies have shown different trends. In one study,
breast cancer patients reported positive changes in health and physical functioning
between 4 and 10 months post-diagnosis, but mental health and well-being did not
improve. In fact, anger and irritation increased, and perceived internal control decreased
(Vinokur, Threatt, Vinokur-Kaplan, & Satariano, 1990). Ell and colleagues (1989)
8
assessed cancer patients’ adjustment 3 through 6 and 9 through 12 months after diagnosis
and found that mental health deteriorated during this period; the same pattern was found
among the breast cancer patients in the sample.
One potential reason for the discrepancy in findings is that different measures
were used; both studies showing declining distress used a measure of recent mood,
whereas both studies showing stable or increasing distress used general indices of mental
health. In addition, the studies showing improvement in distress are more recent and may
reflect treatment advances and greater openness about cancer. However, it is impossible
to tell from the studies described above whether changes in distress were related to the
end of treatment, because all used diagnosis as a point of reference for follow-up
assessments. This poses a problem for examining distress during the re-entry period
because, for example, at 3 months post-diagnosis some women may be undergoing
chemotherapy, others may have just completed radiation treatment, while still others have
completed surgery months before and received no adjuvant treatment. The trajectory of
distress with respect to treatment cessation cannot be determined.
Stronger evidence regarding distress and quality of life during the re-entry period
can be obtained from studies that use the end of adjuvant treatment as a reference point.
Women with gynecologic cancer showed less mood disturbance 4 months after the end of
treatment than at the time of diagnosis, such that they reported comparable levels of
disturbed mood to a sample of healthy women (Andersen, Anderson, & deProsse, 1989).
Among breast cancer patients treated with chemotherapy, depressive symptoms
decreased from the first chemotherapy treatment to one week after the end of
chemotherapy treatment, and decreased even more 6 months after chemotherapy had
ended (Ward, Viergutz, Tormey, deMuth, & Paulen, 1992). A third study found that
many early-stage breast cancer patients exhibited some symptoms of PTSD 4 to 12
months after treatment completion, but only 3% met criteria for PTSD (Gotay &
Muraoka, 1998).
9
Taken together, these studies suggest that distress during the re-entry period is
lower than distress around the time of diagnosis or the beginning of adjuvant treatment
and that 4 months or more after the completion of treatment, cancer patients have
comparable PTSD rates and mood disturbance to the general population. However,
adjustment at diagnosis may not be a good comparison as the time around diagnosis and
beginning treatment is also a period of distress and disrupted adjustment (Frost et al.,
2000; Osowiecki & Compas, 1999; Stanton & Snider, 1993). In addition, although
survivors’ adjustment appears to be comparable to that of the general population 4 or
more months following treatment, it is possible that the initial months following the
completion of treatment may be a time of disrupted adjustment. It may be that studies
with follow-up assessments 4 or 6 months after treatment are missing a distressing
period.
Much of the evidence regarding disrupted adjustment during the months
following treatment completion is anecdotal. Two physicians who are cancer survivors,
including the physician quoted earlier, have written about their own distress following the
end of treatment (McKinley, 2000; Mullan, 1985). Health care providers have described
clinical observations of increased distress during this time (Rowland & Holland, 1990;
Schnipper, 2001). The empirical literature is somewhat scarce and often lacking in
methodological sophistication. Two studies report on the percentage of women that are
distressed at the end of treatment. In response to structured interviews, 30% of breast
cancer patients reported distress associated with the end of treatment in one study (Ward,
Viergutz, Tormey, deMuth, & Paulen, 1992), and 35% reported feeling fear or anxiety
following the end of chemotherapy treatment in another (Beisecker et al., 1997). Overall,
71% of the patients in the former study reported emotional or psychological problems
after treatment (Ward, Viergutz, Tormey, deMuth, & Paulen, 1992).
There is some empirical evidence from prospective studies suggesting that women
are more distressed following treatment than they are later on. Breast cancer patients in
10
Ward and colleagues’ study reported greater depressive symptomatology at one week
after treatment than they did 6 months later (1992). In another example, cancer survivors
who had completed treatment recently had higher levels of anxiety surrounding a followup visit than did patients who had completed treatment more than two years before the
study (Lampic et al., 1994). In a study of cancer survivors during the year following a
bone marrow transplant, the number of individuals reporting significant depression
increased slightly from baseline (24%) to discharge (28%) to 100 days post-transplant
(31%) before decreasing to 20% 1 year post-transplant (McQuellon et al., 1998).
Overall, data regarding adjustment during the re-entry period, specifically the first
few weeks and months following the end of treatment, are scarce. Data on distress during
the year following diagnosis are equivocal, although more recent studies with better
methodology suggest that distress decreased over this year (Carver et al., 1993; Stanton,
Danoff-Burg, & Huggins, 2002). When distress is examined with respect to the end of
treatment, results suggest that cancer survivors are within normal limits 4 or more months
following the completion of treatment (Andersen, Anderson, & deProsse, 1989; Green,
Krupnick, Stockton, & Spertus, 1998; Ward, Viergutz, Tormey, deMuth, & Paulen,
1992). However, anecdotal evidence and data from structured interviews indicates that a
significant proportion of cancer survivors likely experience distress associated with the
completion of treatment (Beisecker et al., 1997; McKinley, 2000; Mullan, 1985; Ward,
Viergutz, Tormey, deMuth, & Paulen, 1992). In sum, there is limited but suggestive
evidence that adjustment may be disrupted during the weeks and months following the
end of treatment.
Re-entry issues related to distress
Why might the months following treatment present adjustment problems? First,
women may still be dealing with the physical effects of treatment. One-fourth of breast
cancer survivors participating in a post-treatment intervention study reported that
11
physical problems were their primary concern (Stanton, Ganz, Rowland et al., 2005).
Beisecker and colleagues (1997) interviewed 21 breast cancer patients after
chemotherapy ended and again 6 months later. At the end of treatment, women had
problems with fatigue, nausea, infections, and hair loss. Six months later, fatigue and
hair-related problems persisted, and women also complained of weight gain, menopausal
symptoms, and nail problems. Other distressing physical problems during re-entry
include lymphedema, decreased libido, and early menopause (Schnipper, 2001).
Many women become distressed because they had not anticipated ongoing
treatment-related problems but had expected to return to “normal” shortly after the end of
treatment (e.g., Beisecker, 1997). In a pilot study for the proposed project, the majority of
the women we interviewed expressed surprise at how long treatment-related effects had
lasted, noting that they had expected their physical functioning to return to normal soon
after treatment ended. Regarding recovery, one cancer survivor stated, “Because the
doctors and nurses never told me the range of what to expect, I had expectations of
wellness that were absolutely unrealistic.” (National Cancer Institute, 2002, p. i). The
idea that patients should recover soon after treatment ends is another common “myth”
about the completion of treatment, according to Stanton and colleagues (2005).
Despite the fact that many cancer and treatment-related physical problems persist
during the post-treatment period, physical status does improve (e.g., Deshields et al.,
2005; Vinokur, Threatt, Vinokur-Kaplan, & Satariano, 1990), and cancer survivors no
longer need to focus intensely on medical treatment. Some authors have suggested having
fewer medical treatment and physical problems to focus on leaves more room for a
psychological struggle (e.g., McQuellon et al., 1998; Schnipper, 2001). Consistent with
this idea, McQuellon and colleagues found that while quality of life improved during the
year following bone marrow transplantation, patients’ concerns (including concerns about
work, family, friends, finances, health, and the future, among others) increased
throughout the year. Interviews with breast cancer survivors 2 to 4 weeks after the end of
12
treatment revealed that women were just beginning to reflect on and process fears and
existential issues at this time (Lethborg, Kissane, Burns, & Snyder, 2000). One of the
themes in the interviews was the reality of diagnosis setting in after treatment ended. In a
post-treatment intervention study of breast cancer survivors, the central concern for 51%
of women was emotions as opposed to physical (26%) or interpersonal (11%) concerns
(Stanton, Ganz, Rowland et al., 2005).
Another common myth described by Stanton and colleagues (2005) is that once
treatment is over, cancer survivors no longer need support. Overall, it appears that social
support remains relatively stable following treatment (e.g., Vikour et al, 1990). However,
McQuellon and colleagues reported minor declines in social support 100 days and 1 year
following bone marrow transplantation (1998). Breast cancer survivors whom we
interviewed reported that although they felt that social support was available, many
people in their lives did not seem to realize that they continued to struggle with cancerrelated physical and psychological issues. Others have reported similar observations (e.g.,
Schnipper, 2001). The need for emotional support following treatment is illustrated by a
social worker’s observations that more women seek out a support group after treatment
has ended than during treatment (Schnipper, 2001).
Interpersonal difficulties, including dating problems and sexual dysfunction, may
exacerbate perceived loss of support (Schnipper, 2001). The loss of regular treatment and
medical appointments also results in feelings of decreased support, described by one
author as being “cast adrift” to deal with the fear of recurrence on one’s own (Lethborg,
Kissane, Burns, & Snyder, 2000). The diminishing role of medical staff was also
described by Mullan (1985) as a “void that leaves many cancer patients and their families
fending awkwardly for themselves in the ‘healthy’ world” (p. 272). A prostate cancer
survivor stated, “I went to [radiation treatment] every day, and they treated me, and we
were like…family. Now there’s this instant separation” (National Cancer Institute, 2002,
13
p. 59). This loss of support from health care providers may actually increase the need for
support from friends and family.
Perhaps the most significant source of distress following the end of treatment is
the feeling of losing a safety net described by many cancer survivors (e.g., Ward, 1992).
Regularly receiving cancer treatments (chemotherapy or radiation sessions) and attending
medical appointments provides an active means for destroying existing cancer and
preventing cancer growth and recurrence. Once treatment is over, this important active
coping strategy is no longer available. Fear of a cancer recurrence, the main concern for
39% of women enrolled in an post-treatment intervention study (Stanton, Ganz, Rowland
et al., 2005), coupled with the loss of one’s primary means for managing cancer, is
particularly distressing for cancer survivors (Lethborg, Kissane, Burns, & Snyder, 2000;
Mullan, 1985). One breast cancer survivor stated, “When I got off chemotherapy, it
terrified me…I was being cut loose. And I thought, ‘Oh my God, what’s going to happen
now?’” (Beisecker, 1997, p. 90). A breast and thyroid cancer survivor stated, “As long as
I was in treatment I was killing the cancer. [After treatment] I was waiting for the other
shoe to fall” (National Cancer Institute, 2002, p. 48). Breast cancer survivors interviewed
2 to 4 weeks after treatment had ended frequently mentioned the fear associated with not
having regular medical checkups (Lethborg, Kissane, Burns, & Snyder, 2000). In another
study, 29% of breast cancer patients who had received adjuvant chemotherapy reported
feeling as if a safety net had been lost once treatment was over (Ward, Viergutz, Tormey,
deMuth, & Paulen, 1992). McKinley (2000) described this well:
I have now finished surgery, chemotherapy, radiation, and
reconstruction; I’m done, according to the medical profession. But
I don’t really feel done. . . . We just move from the quantifiable,
treatable disease to the immeasurable uncertainty of survivorship. .
. . Being in the midst of active treatment means being seen
regularly by a nurse or a physician—being cared for. (p. 479)
14
Strategies for “Reconstructing the Safety Net”
Coping theory and research
A crucial task after treatment ends may be finding means for “reconstructing the
safety net,” in other words, developing new active strategies for dealing with the
uncertainty of one’s cancer status and preventing a cancer recurrence. Coping theory
provides a perspective for conceptualizing this task. Lazarus describes coping as
“ongoing cognitive and behavioral efforts to manage specific external and/or internal
demands that are appraised as taxing or exceeding the resources of the person” (Lazarus,
1993). Attending treatment sessions and regular medical checkups during treatment may
be thought of as an active, behavioral strategy for coping with cancer. Finding new
active, behavioral strategies for coping with the uncertainty of one’s health may decrease
distress and promote adjustment during the re-entry period.
Previous research suggests that active coping is associated with better adjustment
in the face of chronic stressors. A meta-analysis of the coping literature demonstrated that
while avoidant coping strategies, including constructs such as denial and techniques such
as distraction, worked well in the short term, attentional strategies, including constructs
such as active coping and monitoring and techniques such as focusing on sensory
information, were more effective over longer time periods (Suls & Fletcher, 1985). Based
on their results, the authors proposed that avoidant strategies should work best for coping
with less serious, relatively brief stressful life events, whereas attentional strategies
should work best for coping with serious, long-term stressful life events requiring
readjustment, such as breast cancer.
Active strategies appear to “work better” for breast cancer patients and survivors
with respect to managing cancer-related stressors, especially in contrast to avoidance
strategies. For example, in a recent study of breast cancer survivors, seeking information
after treatment had ended predicted better quality of life 6 months later (Ransom,
15
Jacobsen, Schmidt, & Andrykowski, 2005). In another example, Carver and colleagues
(1993) found that breast cancer patients who used coping strategies of acceptance, humor,
and positive reframing over the course of the year following diagnosis had lower
concurrent distress and that acceptance and humor prospectively predicted lower distress.
In contrast, denial and disengagement predicted higher levels of distress both
concurrently and prospectively. In a third prospective study, breast cancer patients who
coped by using cognitive avoidance prior to their diagnostic biopsy had increased distress
and less vigor both after diagnosis and following mastectomy or lumpectomy, whereas
those who sought social support pre-biopsy had increased levels of post-biopsy vigor
(Stanton & Snider, 1993). Similarly, we found that avoidant coping strategies, including
disengagement and cognitive avoidance, were strongly associated with poorer well-being
and more distressed mood among gynecologic cancer patients (Costanzo, Lutgendorf,
Rothrock, & Anderson, 2006).
Most of these studies examined coping with respect to cancer-related stressors
during times when women were receiving active medical treatment. Stanton and her
colleagues have demonstrated that, compared to avoidance, active coping endeavors are
associated with better adjustment during the re-entry period as well. This group assessed
women 20 weeks after the completion of adjuvant treatment and then again 3 months
later. They found that coping by expressing emotions at 20 weeks post-treatment was
associated with less distress, fewer medical visits, and better self-reported health and
vigor 3 months later. In contrast, avoidance was associated with subsequent distress
(Stanton et al., 2000). In another study, active acceptance at diagnosis predicted better
adjustment one year post-surgery among breast cancer patients. Although avoidance was
associated with less distress 3 months post-surgery, it was associated with greater fear of
recurrence one year post-surgery, when most women would have completed adjuvant
treatment (Stanton, Danoff-Burg, & Huggins, 2002). This group also found that writing
about one’s thoughts and feelings about cancer was an effective active coping tool for
16
breast cancer survivors who had completed treatment 20 weeks previously. Women who
were assigned to write about their deepest thoughts and feelings had fewer physical
symptoms and medical appointments for cancer-related problems 3 months later (Stanton
et al., 2002).
The evidence suggests that acceptance rather than avoidance is adaptive during
the active treatment and re-entry phases of cancer. These represent more cognitive means
of coping with cancer and its stressors. Lazarus suggests that behavioral efforts are
important to consider as well (1993), and behavioral efforts may be particularly critical
for replacing the loss of the primary behavioral strategy of attending cancer treatment
sessions. Stanton and her colleagues’ work indicates that the behavioral strategy of
expressing emotions is adaptive during all phases of cancer (Stanton et al., 2000; Stanton
& Snider, 1993) and that writing about cancer may be particularly useful during the reentry period (Stanton et al., 2002). Although this represents an important finding, breast
cancer survivors likely use and may benefit from a broader array of behavioral strategies
during the re-entry period than those traditionally examined in the coping literature.
Post-treatment behavioral interventions and guidelines
Breast cancer survivors whom we interviewed stated that they would like more
guidance from their health care providers regarding what they can do after treatment
ends; for example, whether they should make health behavior or lifestyle changes.
McKinley, the physician and breast cancer survivor quoted previously, also desired more
guidance on strategies for “reconstructing the safety net” after the end of treatment
(2000):
. . . because the struggle with uncertainty after treatment is
completed is usually a silent battle waged outside of the
physician’s office, most physicians don’t think or talk about it. . . .
Oncologists need to focus on preparing breast cancer patients for
survivorship. That is, they must address the loss experienced by
survivors when active treatment is over and they are sent away
from a very intense environment. They must help survivors
17
understand the impact of fear and uncertainty on their lives and
what might help reduce these stressors. (p. 480)
Researchers and health care providers are beginning to recognize the need to help
survivors develop active, behavioral strategies for coping with uncertainty after treatment
ends, as evidenced by the recent development of behavioral interventions and written
guidelines. For example, a recent article described a “coping skills training” telephone
counseling intervention for breast cancer survivors during the year after treatment
completion. Active strategies for dealing with uncertainty, physical change, self-change,
and changes in relationships, among other issues, are addressed (Marcus et al., 1998).
Another intervention, “Taking CHARGE,” aims to ease the transition following breast
cancer treatment using psychoeducational techniques to assist participants in setting and
working toward behavioral goals (Cimprich et al., 2005). Although participants reported
“Taking CHARGE” to be beneficial, no efficacy data with respect to distress or quality of
life are available for either intervention.
Cancer treatment centers may also provide special services for cancer survivors
during the re-entry period. A recent survey of clinics deemed “comprehensive cancer
centers” by the National Cancer Institute (NCI) found that these centers provided services
related to lymphedema management (70% of centers), support groups (49%), long-term
follow-up clinics for medical care (38%), school re-entry programs (19%), nutrition
counseling (14%), reproductive/sexuality counseling (14%), and fatigue management
(3%) (Tesauro, Rowland, & Lustig, 2002). Based on these data, while it appears that the
physical and medical sequelae of cancer treatment are addressed to some extent by many
comprehensive cancer centers, psychosocial, behavioral, and practical issues may not be
adequately addressed at the majority of centers.
The NCI has also published an educational booklet, “Facing Forward,” targeting
cancer patients who have finished treatment, which is available on the web or may be
ordered from NCI (National Cancer Institute, 2002). The introduction states, “Your new
‘normal’ may involve making changes in the way you eat, the activities you so, and your
18
sources of support” (p. 7). The 125-page book guides cancer survivors through active
strategies for managing physical, psychological, and interpersonal difficulties.
Suggestions include healthful behaviors such as attending follow-up appointments,
getting regular cancer screenings, and developing a wellness plan which includes
healthful eating, physical activity, and cutting back on alcohol and smoking. Strategies
for dealing with ongoing side effects include physical therapy, yoga, relaxation exercises,
acupuncture, adjusting activity levels, attending pain clinics, exercise, and medication.
Suggestions for managing menopausal symptoms and sexual dysfunction are also
included. A variety of ideas for coping with distress and other psychosocial problems,
including seeking out counseling with clergy or professionals, joining support groups,
and enrolling in stress management programs are also discussed. Strategies for dealing
with relationships are included, as are ideas for managing practical problems related to
work or insurance.
Annette Stanton, Patricia Ganz, and their colleagues have designed an
intervention, “Moving Beyond Cancer,” for breast cancer patients who recently
completed adjuvant treatment based on the NCI’s Facing Forward publication (P. Ganz,
personal communication, November 2003; A. Stanton, personal communication, April
2003; Stanton, Ganz, Kwan et al., 2005). Similar to the “Taking CHARGE” program, the
focus of the intervention was to help survivors use active, approach-oriented coping
strategies. This group conducted a randomized intervention trial including 558 breast
cancer survivors who had recently completed treatment and were assigned to three
different intervention conditions. All three arms received NCI’s Facing Forward book
and were asked to work though it. One group also watched a corresponding video, and
another group worked with a cancer educator to address their most significant posttreatment problem and received a follow-up phone call two weeks later. A control group
received no additional treatment. The inervention was being conducted at the time our
study began, and the authors have now released their results. A recently published paper
19
reported that participants in the video group reported improvements in energy as
compared to the control group 6 months post-intervention, and these findings were
strongest among women who felt less prepared for “re-entry.” Those in the cancer
educator group showed reduced distress as compared to the control group when
comparing women who felt most prepared for re-entry (Stanton, Ganz, Kwan et al.,
2005).
Recommendations for health behavior following cancer treatment have also been
established by both the NCI and the American Cancer Society (ACS). In their Facing
Forward publication, the NCI recommends that cancer survivors quit smoking, cut back
on alcohol, eat a healthful diet (eat five or more servings of fruits and vegetables, choose
whole grains rather than processed/refined sugars, and limit red meat or other high-fat
meat), maintain a healthy weight, and exercise moderately for approximately 30 minutes
most days (National Cancer Institute, 2002). Because there is little evidence that dietary
or other health habits may prevent recurrence or secondary cancers, the ACS
recommends that cancer survivors follow the guidelines they have established for cancer
prevention (see Table 1 for the ACS prevention guidelines; ACS Nutrition and Physical
Activity Guidelines Advisory Committee, 2001; Brown et al., 2001). The American
Institute for Cancer Research (AICR) has also established health behavior guidelines for
cancer prevention (see Table A2), and the ACS has specific guidelines for breast cancer
prevention (see Table A3) that may be applicable to cancer survivors as well (ACS
Nutrition and Physical Activity Guidelines Advisory Committee, 2001; American
Institute for Cancer Research, 1997, 2002).
Despite the general lack of evidence regarding factors related to prevention of
cancer recurrence, data are strongest for breast cancer recurrence prevention. In
particular, there is evidence that obesity may be associated with both poorer prognosis
and cancer recurrence and that eating a low-fat diet may be associated with a lower
recurrence rate (Brown et al., 2001). The authors of this paper review evidence for a
20
variety of health behaviors in preventing breast cancer recurrence. They conclude that
food safety practices, weight loss after treatment (if overweight), decreased dietary fat
consumption, and increased physical activity have a probable, but unproven benefit.
Increased consumption of fruits and vegetables, decreased alcohol consumption, juice
therapies, and vegetarian diets have a possible, but unproven benefit. Health behavior
changes may also impact quality of life among breast cancer survivors. Recent work
suggests that exercise may decrease fatigue and improve physical dimensions of quality
of life among breast cancer survivors (J. A. Hewitt, Mokbel, van Someran, Jewell, &
Garrod, 2005; Kendall, Mahue-Giangreco, Carpenter, Ganz, & Bernstein, 2005; Turner,
Hayes, & Reul-Hirche, 2004).
In sum, NCI comprehensive cancer clinics provide few services for re-entry
patients and cancer survivors in general beyond follow-up medical care and management
of treatment side effects. However, there are multiple intervention trials underway to help
cancer survivors “reconstruct the safety net” through the development of active,
behavioral coping strategies, and recent results from one of these interventions is
promising. There are also several publications available to survivors that suggest
behavioral strategies for promoting health and preventing cancer recurrence. NCI’s
Facing Forward publication is the most comprehensive to-date and includes a plethora of
information and suggestions regarding strategies not only for promoting physical health
and preventing recurrence but also for managing stress and interpersonal and practical
problems. It is unclear to what extent cancer survivors access and use these publications.
Behavior changes among cancer survivors
To what extent do cancer survivors actually make behavior changes suggested by
NCI, ACS, and AICR? The data are scarce and are largely limited to information
regarding dietary change, physical activity, and complementary and alternative medicine
(CAM) use. A recent study found that 30.2% of breast cancer survivors began a new
21
physical activity sometime after diagnosis (Patterson et al., 2003). With respect to dietary
change, data are inconsistent. Estimates range from 4.4% to 59.5% of cancer survivors
making dietary changes (Begbie, Kerestes, & Bell, 1996; Burstein, Gelber, Guadagnoli,
& Weeks, 1999; Lee, Lin, Wrensch, Adler, & Eisenberg, 2000; Maskarinec, Murphy,
Shumay, & Kakai, 2001; Patterson et al., 2003). Estimates of CAM use range from 9.4%
to 66% (Henderson & Donatelle, 2003; Maskarinec, Gotay, Tatsumura, Shumay, &
Kakai, 2001; Maskarinec, Shumay, Kakai, & Gotay, 2000; Rothrock, 2002).
One group has recently examined both dietary changes and CAM use among
Hawaiian cancer survivors 2 to 3 years after diagnosis. Unfortunately, CAM users were
deliberately oversampled which limits the generalizability of the study. Sixty-nine of 143
survivors made changes in diet. While some of those changes matched recommendations
(e.g., decreasing fat consumption, increasing fruit and vegetable consumption), others
made changes not necessarily recommended such as using herbal or vitamin supplements
(Maskarinec, Murphy, Shumay, & Kakai, 2001). Participants reported that reasons for
diet changes included preventing recurrence, maintaining health, and promoting wellbeing. Another study of the same sample found that participants commonly used religious
coping methods including prayer/meditation, bible study, and attending religious services
(Tatsumura, Maskarinec, Shumay, & Kakai, 2003). Among this sample of CAM users,
predictors of more extensive use of complementary and alternative treatments included
female sex, Caucasian ethnicity, and more education. Breast cancer patients and those
who experienced more nausea and vomiting were also more likely to use CAM (Shumay,
Maskarinec, Gotay, Heiby, & Kakai, 2002).
Taylor and colleagues (1984) examined a variety of behavior changes among
breast cancer survivors 1 to 60 months post-surgery. Behavior change was common in
their sample; 78% of women made behavioral changes, with 23% making three or more
changes. Common behavior changes included changes in diet (49%) and increasing
exercise (26%). The adaptiveness of behavior changes with respect to adjustment was
22
also examined. Dietary changes were not related to adjustment, but increasing exercise
and making more time for leisure activities were associated with better adjustment.
Despite recent attention to educating cancer survivors about effective behavioral
changes and active coping strategies, there are scarce data on the frequency of posttreatment behavioral changes. There is some evidence that emotional expression though
writing, increasing exercise, and making adjustments in lifestyle to allow for more leisure
activities are associated with better adjustment among cancer survivors. The limited
evidence suggests that making active behavioral changes thought to improve health and
well-being and prevent recurrence may promote adjustment following cancer treatment.
Common-Sense Representations of Illness
Several theorists have been interested in individuals’ cognitive representations of
illness or disease and have posited that these cognitive representations or models may
explain and predict important behavioral responses to illness including adherence to
medical regimens, health behavior changes, and other coping or behavior changes. Work
in this area was initiated by Howard Leventhal, who explored individuals’ common-sense
cognitive models of illness as an important part of his self-regulation theory, a model of
adjustment to illness. The theory incorporates both cognitive representations of illness
and emotional responses to illness, considered to be two distinct components. An
important postulate of the model is that coping responses to illness and treatment are
directed by cognitive models of illness (Leventhal et al., 1997; Leventhal, Meyer, &
Nerenz, 1980).
Based on open-ended interviews with hypertensive patients, Leventhal and his
colleagues proposed that cognitive representations of illness could be characterized along
four dimensions: identity (label and symptoms), cause, consequence (short and long-term
effects of illness), and course (timeline of illness) (Leventhal, Meyer, & Nerenz, 1980;
Leventhal & Nerenz, 1985; Leventhal, Nerenz, & Steele, 1984). Lau and Hartman (1983)
23
validated Leventhal’s dimensions in college students’ descriptions of recent acute
illnesses such as colds and influenza. They also proposed a fifth dimention, cure, which
incorporates perceived control over the illness and beliefs about how one effects recovery
from the illness (Lau & Hartman, 1983). This dimension may not have been uncovered in
Leventhal’s initial work because hypertension generally does not go away. The cure
dimension has since been widely accepted and the five dimensions have subsequently
been validated for chronic illness (Meyer, Leventhal, & Gutmann, 1985), acute minor
illness (Lau, Bernard, & Hartman, 1989; Lau & Hartman, 1983), and hypothetical illness
(Bishop, Briede, Cavazos, Grotzinger, & McMahon, 1987). The Illness Perception
questionnaire (IPQ; Weinman, Petrie, Moss-Morris, & Horne, 1996) was developed to
assess common-sense models of illness based on these dimensions and has recently been
revised (Moss-Morris et al, 2002). Initial studies of common-sense illness representations
relied on time-consuming interview techniques, and the IPQ has therefore greatly
simplified assessment. The revised IPQ and its five dimensions have been validated in a
variety of patient populations including HIV, rheumatoid arthritis, diabetes, asthma,
chronic and acute pain, multiple sclerosis, and myocardial infarction (Moss-Morris et al.,
2002).
Common-sense ideas about disease have proven useful in predicting behavioral
responses to illness as hypothesized by self-regulation theory. Individuals’ cognitive
models of their illness have predicted decisions to seek health care (Grunfeld, Hunter,
Ramirez, & Richards, 2003; Martin & Lemos, 2002; Matthews, Siegal, Kuller,
Thompson, & Varat, 1983; Stillman, 1977) and compliance with medical interventions
(Leventhal, Meyer, & Nerenz, 1980). For example, patients beginning treatment for
hypertension were more likely to discontine treatment if they reported experiencing
physical symptoms they believed to be associated with hypertension (there are no known
symptoms of hypertension, but most patients believe there are) and if they believed that
their disease was acute. In other words, hypertensive individuals who perceived that
24
“hypertension” consisted of a collection of symptoms that would be alleviated with an
acute treatment regimen understandably did not adhere to a long-term treatment protocol
(Meyer, Leventhal, & Gutmann, 1985). Similarly, Christensen and colleagues (1999)
found that endorsing “irrational” beliefs about illness was associated with both poorer
health practices (exercise, diet, relaxation, safety, substance use) among undergraduates
and poorer adherence to one’s diabetic regimen among Type I diabetics (Christensen,
Moran, & Wiebe, 1999).
Illness representations may also predict functional outcomes and psychological
adjustment. A recent study examined the benefits of an intervention designed to help
individuals who had recently had a myocardial infarction (MI) develop adaptive beliefs
about their MI. Individuals randomized to the intervention reported feeling more prepared
when leaving the hospital, returned to work more quickly, and experienced fewer angina
symptoms as compared to a control group (Petrie, Cameron, Ellis, Buick, & Weinman,
2002). Common-sense beliefs have been shown to be associated with psychological
adjustment in diverse medical populations including patients with AIDS (Moulton,
Sweet, Temoshok, & Mandel, 1987), end-stage renal disease (Rich, Smith, &
Christensen, 1999), and cancer (Berckman & Austin, 1993; Faller, Schilling, & Lang,
1995; Kohli & Dalal, 1998; Lowery, Jacobsen, & DuCette, 1993).
Common-Sense Representations of Cancer
Leventhal’s five domains of illness representations applied
to cancer
Leventhal’s dimensions of illness representations have been applied to cancer
patients’ common-sense models of their disease and its physiological sequelae. For
example, patients’ descriptions of cancer-related fatigue could be classified along the
dimensions. Of statements made by cancer patients in focus groups, 41% involved fatigue
consequences, 29% were related to identity, 17% described fatigue pattern, and 13%
25
involved cause of fatigue (Barsevick, Whitmer, & Walker, 2002). Common-sense models
of breast cancer more generally can also be classified using Leventhal’s model. In a
chapter related to this topic, Buick (1997) presents unpublished data suggesting that
breast cancer patients’ representations of their disease fell into two clusters. Women in
the first cluster perceived low control and ability to effect a cure along with greater
consequences, longer duration, more self-blame, and more symptoms associated with
cancer. Women in the second cluster endorsed the reverse on all dimensions. Women in
the first cluster were more likely to report that psychosocial factors were responsible for
their cancer, they perceived treatment as more severe, and they had poorer psychological
adjustment and less coping flexibility (Buick, 1997). Other work has compared breast
cancer patient’s illness representations with those of healthy women (Anagnostopoulos &
Spanea, 2005). Taken together, these studies suggest that cancer patients’ common-sense
ideas about their disease can be classified using the traditional model put forth by
Leventhal and his colleagues.
Few studies have included both a comprehensive assessment of illness
representations and their behavioral and psychological correlates. In one such study, the
ability of the traditional five domains of illness representations to predict intentions to
seek care for possible breast cancer symptoms was examined in a community sample of
546 women (Grunfeld, Hunter, Ramirez, & Richards, 2003). Inability to identify breast
cancer symptoms (Leventhal’s “identity” domain) was associated with delay among all
women. Among younger women, low behavioral control was also associated with delay.
Among older women, perceptions of negative consequences predicted delay in seeking
care. In a second study that included a comprehensive assessment of cancer
representations, trait anxiety and side effects of Tamoxifen treatment activated illness
representations. Illness representations in turn triggered somatic sensitization and worry
as well as increased breast self-exam (Cameron, 1997; Cameron, Leventhal, & Love,
26
1998). These studies suggest that common-sense models of cancer may be important
determinants of both health behavior and distress.
Perceived control over cancer
Although there is a paucity of studies that have incorporated a comprehensive
assessment of cancer patients’ illness representations, several studies have examined one
or two domains of cancer patients’ common-sense ideas about their disease. Perceived
control over various aspects of the cancer experience and the disease itself has been a
historically popular area of inquiry. The literature suggests that higher levels of perceived
control are associated with both more proactive behavior and better adjustment.
Specifically, greater perceived control is associated with CAM use (Henderson &
Donatelle, 2003) and utilization of active coping strategies such as problem-solving,
seeking support, and self-control (Hilton, 1989). Relationships among perceived control
and psychological adjustment have been more thoroughly described in the literature.
Greater perceived control is associated with fewer depressive symptoms (Newsom,
Knapp, & Schulz, 1996) and better adjustment (R. A. Jenkins & Pargament, 1988;
Taylor, Lichtman, & Wood, 1984; Thompson, Sobolew, Galbraith, Schwankovsky, &
Cruzen, 1993) while feeling a “loss of control” is associated with poorer adjustment and
greater distress (Lowery, Jacobsen, & DuCette, 1993).
Perceived control over certain aspects of the cancer experience may be more
important with respect to adjustment. In one study, perceived control over emotions and
physical symptoms was most strongly related to better adjustment and control over the
disease itself was least strongly related to adjustment, with control over medical care and
relationships falling in the middle (Thompson, Sobolew, Galbraith, Schwankovsky, &
Cruzen, 1993). Among breast cancer survivors, believing one has control over cancer
course in the present was more important in promoting adjustment than were attributions
about what may have caused cancer in the past (Taylor, Lichtman, & Wood, 1984).
27
Results of one of the few prospective studies in this literature indicated that for patients
with recurrent cancer, perceived control over life events, illness course, and illness
symptoms were associated with fewer concurrent depressive symptoms, while perceived
control over cancer onset was associated with greater depressive symptoms eight months
later (Newsom, Knapp, & Schulz, 1996).
Perceived control may also moderate the relationship between severity of disease
or physical symptoms and psychological adjustment. For example, Andrykowski &
Brady (1994) found that health locus of control interacted with physical decline since
diagnosis to predict distress among leukemia patients. The relationship between physical
decline and distress was strongest for individuals who perceived high internal control
over health (M. A. Andrykowski & Brady, 1994). In another study of women who did not
have breast cancer (either healthy women or women with a benign breast problem),
perceived control over a potential breast cancer prognosis, as assessed by beliefs that
mammograms can find breast cancer early and breast cancer can be cured if found early,
moderated the relationship between breast symptoms and breast cancer worry. Among
women who perceived low control over a breast cancer diagnosis, more breast symptoms
were related to more worry, whereas among women who perceived high control, breast
symptoms were unrelated to worry (Cunningham et al., 1998). The interaction pattern is
contradictory in these two studies, probably due to the different study populations.
Disease severity may in turn moderate the relationship between perceived control and
distress. In one study, greater perceived control was related to less depression among
individuals who perceived their cancer to be highly severe but not among individuals
who perceived their cancer to by low or moderate in severity (Marks, Richardson,
Graham, & Levine, 1986). This finding appears to be more consistent with Cunningham
and colleagues’ study of women who were not yet diagnosed with breast cancer than with
Andrykowski’s and Brady’s study of leukemia patients.
28
In sum, research on perceived control in cancer patients indicates that greater
perceived control is generally beneficial with respect to psychological adjustment and
health behavior. It may be more beneficial for patients to believe they have control over
their symptoms or treatment side effects and to believe they have some control over their
disease in the present than to believe they had control over the onset or cause of their
cancer. The latter belief may even be detrimental to psychological adjustment. Finally, it
appears that disease severity may be an important contextual factor to examine when
relationships between perceived control and adjustment or behavior are studied, but the
role of disease severity is presently unclear.
Conceptualization of cancer timeline
A recent study examined the influence of breast cancer patients’ beliefs about
their disease timeline on post-treatment distress (Rabin, Leventhal, & Goodin, 2004).
Slightly more than half of the sample (54-55% at different time points) perceived their
cancer to be acute in nature, and fewer perceived their cancer to be a chronic or cyclic
condition (36-45%). Results indicated that women who perceived their cancer to be
chronic or cyclic in nature reported more anxiety, depression, and recurrence worry than
did women who perceived their cancer to be an acute condition. This is the only study we
know of to have examined implications of cancer patients’ beliefs about disease timeline.
Causal attributions for cancer
Another primary area of research with respect to common-sense ideas of cancer
involves characterizing cancer patients’ causal attributions for cancer. Causal attributions
appear to differ by type of cancer, as one might expect. Lung cancer patients, for
example, frequently cite smoking, followed by toxins or pollution, as causes of their
cancer (Berckman & Austin, 1993; Faller, Schilling, & Lang, 1995; Mumma &
McCorkle, 1983). Gynecologic cancer patients’ attributions differed by stage, with earlystage patients attributing cancer to their past behavior and advanced-stage patients
29
attributing cancer to God’s will, and a high proportion in both groups attributing cancer
to chance (Gotay, 1985). We have recently reported that long-term cervical and
endometrial cancer survivors rated genetics/heredity as the most important cancer cause,
followed by stress, God's will, hormones, and environmental factors (Costanzo,
Lutgendorf, Bradley, Rose, & Anderson, 2005). Breast cancer survivors had a similar
pattern to the gynecologic cancer survivors we studied, with stress, heredity,
environmental factors, and diet frequently cited as cancer causes. In one study, survivors
attributed breast cancer to stress, a carcinogen, heredity, and/or diet (Taylor, Lichtman, &
Wood, 1984). In a more recent study, breast cancer survivors were most likely to attribute
their cancer to stress, followed by genetics, the environment, and hormones (Stewart et
al., 2001). Israeli breast cancer patients had a slightly different pattern of attributions;
they were most likely to attribute cancer to fate, followed by psychological factors and
family history (Kulik & Kronfield, 2005).
Along similar lines, other studies have described cancer patients’ ideas about how
to control their cancer or prevent recurrence. In general, patients believe that healthy
behavior, stress reduction, and having a positive attitude are important. Lung cancer
patients most frequently identified behavioral strategies such as diet, exercise, relaxation,
and quitting smoking, but they also cited cognitive strategies such as prayer, positive
thinking, and acceptance is important in controlling their disease (Berckman & Austin,
1993). Gynecologic cancer survivors rated medical screening as most important in
preventing recurrence, followed by positive attitude and prayer (Costanzo, Lutgendorf,
Bradley, Rose, & Anderson, 2005). Breast cancer survivors were most likely to cite
having a positive attitude as important in preventing recurrence, followed by health
behaviors including diet, exercise, and stress reduction (Stewart et al., 2001).
As suggested by Leventhal’s self-regulation theory, causal attributions have been
associated with behavioral responses among cancer patients. In general, controllable
causal attributions have been associated with better health practices. For example, breast
30
cancer patients who perceived cancer causes to be controllable were more likely to
modify their health behavior (Buick, 1997). Similarly, gynecologic cancer survivors who
cited potentially controllable causes were more likely to practice healthy behaviors
(Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005). Christensen and colleagues
found that head and neck cancer patients who attributed their cancer to substance abuse
and perceived high control over their cancer were less likely to smoke, while those who
perceived low control were more likely to smoke (Christensen et al., 1999).
Other work has documented relationships between specific causal attributions and
behavioral outcomes. Breast cancer survivors who believed that stress caused their cancer
were more likely to use CAM and antidepressants and were less likely to smoke (Stewart
et al., 2001). Compared to cancer survivors who did not use CAM, CAM users were
more likely to believe that environmental factors, stress and other psychosocial factors,
and infections caused their cancer (Maskarinec, Gotay, Tatsumura, Shumay, & Kakai,
2001). Nonusers were less likely to name a cause or believed that chance was responsible
for their cancer. Qualitative interview data suggested that many patients decided on
particular CAM treatments because of perceived cancer causes. In sum, attributing cancer
to controllable causes or stress is generally associated with proactive behavior including
positive health behavior and CAM use. As would be expected, cancer survivors who
believed that health behaviors were important in preventing recurrence were more likely
to practice them; breast cancer survivors who believed diet or CAM use could prevent
recurrence were more likely to take vitamins and dietary supplements and use CAM
(Stewart et al., 2001).
Several studies have also examined relationships between causal attributions and
psychological adjustment. While one study found that causal attributions were unrelated
to adjustment (Gotay, 1985), results generally show that stronger or more elaborated
causal attributions are associated with poorer adjustment and greater distress (Berckman
& Austin, 1993; Lowery, Jacobsen, & DuCette, 1993). Psychosocial attributions in
31
particular are related to poorer adjustment, including greater distress, depression,
hopelessness, and anxiety (Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005;
Faller, Schilling, & Lang, 1995; Kohli & Dalal, 1998; Kulik & Kronfield, 2005).
Attributing cancer to another person is also associated with poorer psychological
adjustment (Kulik & Kronfield, 2005; Taylor, Lichtman, & Wood, 1984).
These results are somewhat paradoxical given the literature that has been
reviewed thus far. There appears to be a relationship between making causal attributions
for cancer and engaging in healthy behavior or other proactive behavioral responses. In
addition, we have discussed the literature indicating that active, behavioral coping
responses are associated with positive psychological adjustment among cancer patients.
However, making causal attributions about cancer has been associated with poorer
adjustment among cancer patients. A potential solution to this paradox involves
considering the relationship between causal attributions and distress in the context of
behavior. Specifically, individuals who continue to engage in behavior or have persistent
psychosocial conditions believed to be the cause of cancer may remain distressed.
Conversely, individuals who make changes in behavior and psychosocial conditions to
which they attribute their cancer may demonstrate positive adjustment. We recently
tested this hypothesis in a sample of long-term gynecologic cancer survivors (Costanzo,
Lutgendorf, Bradley, Rose, & Anderson, 2005). The hypothesis was confirmed;
attributing cause and recurrence prevention to health behaviors or lifestyle interacted with
health practices in models predicting distress. For example, among women who had not
made positive dietary changes since cancer treatment, rating lifestyle as important in
preventing recurrence was associated with greater distress. In contrast, this belief was
associated with less distress among women who had made a positive change in diet. In
sum, common-sense attributions for cancer appear to be associated with greater distress,
but engaging in behavior believed to be important in preventing cancer or recurrence may
ameliorate this distress.
32
A major weakness in the literature exploring cancer patients’ common-sense
representations of their disease is that nearly all of the studies examining behavioral and
adjustment outcomes are cross-sectional. There are a few exceptions. As previously
mentioned, perceived control over cancer onset was related to more depressive symptoms
8 months later among patients being treated for recurrent cancer (Newsom, Knapp, &
Schulz, 1996). Another study found that among newly diagnosed breast cancer patients,
characterological self-blame for cancer predicted greater distress 6 months and 1 year
after diagnosis, whereas behavioral self-blame was more likely to be concurrently
associated with distress (Glinder & Compas, 1999). In a similar study, characterological
and behavioral self-blame at cancer diagnosis predicted distress 4 months later
(Malcarne, Compas, Epping-Jordan, & Howell, 1995). Taken together, these prospective
studies indicate that blaming oneself for the development of cancer predicts future
distress. However, it is clear that cross-sectional findings regarding common-sense ideas
about cancer with respect to adjustment and behavior outcomes will need to be validated
in a prospective study. Without prospective evidence, it is impossible to determine the
direction of causality or the dynamic nature of these relationships.
Another gap in the literature, also due to the paucity of prospective studies, is the
lack of information regarding whether common-sense ideas about cancer change over
time or are differentially related to outcomes at different points in the cancer experience.
Breast cancer survivors in one study indicated that thinking about causal attributions was
more important to them during the recovery period than earlier in the cancer experience
(Taylor, Lichtman, & Wood, 1984). It may be that patient’s common-sense models of
cancer play a salient role in helping patients to “reconstruct the safety net” following the
end of treatment. Leventhal’s model would predict that representations of cancer direct
behavioral coping efforts. As previously discussed, patients have a relatively uniform
active means for coping with cancer: attending treatment. Following the cessation of
treatment, a variety of possibilities for active, behavioral coping efforts exist and
33
individual differences in common-sense models of cancer may therefore play a more
significant role in cancer survivors’ behavior.
Summary and Proposed Study
Breast cancer survivors appear to be remarkably well-adjusted as compared to
both community samples (Dorval, Maunsell, Deschenes, Brisson, & Masse, 1998; Ganz,
Rowland, Desmond, Meyerowitz, & Wyatt, 1998; Moyer & Salovey, 1996) and to other
individuals with chronic illness (Ganz et al., 1996) despite reports of ongoing physical
problems such as persistent treatment side effects, lymphedema, cognitive problems, and
fatigue (Ahles et al., 2002; Beisecker et al., 1997; Bumpers, Best, Norman, & Weaver,
2002; Cella, Davis, Breitbart, & Curt, 2001; Ferrell, Grant, Funk, Otis-Green, & Garcia,
1997; Kurtz, Wyatt, & Kurtz, 1995; Schagen et al., 1999). However, there may be
periods of disrupted adjustment during the survivorship experience. The completion of
adjuvant treatment (chemotherapy and/or radiation therapy) may be one such period.
Anecdotal evidence (McKinley, 2000; Mullan, 1985; Schnipper, 2001) and information
collected using structured interviews (Beisecker et al., 1997; Ward, Viergutz, Tormey,
deMuth, & Paulen, 1992) suggest that the “re-entry period” is marked by distress,
anxiety, transition, and disruption. Although Stanton and her colleagues define the reentry period as the year following treatment completion, evidence is strongest for
disrupted adjustment during the first 1 to 3 months following cancer treatment (Deshields
et al., 2005; McQuellon et al., 1998; Ward, Viergutz, Tormey, deMuth, & Paulen, 1992)
with survivors showing nearly normal levels of adjustment 4 to 6 months after treatment
ends (Andersen, Anderson, & deProsse, 1989; Ward, Viergutz, Tormey, deMuth, &
Paulen, 1992).
There are several reasons that the post-treatment period may be a particularly
distressing time for cancer survivors, including unrealistic expectations regarding
whether and how quickly they will return to “normal” after the completion of treatment
34
(Beisecker et al., 1997; Stanton, Ganz, Rowland et al., 2005) and the loss of support from
health care providers as well as family and friends who believe the patient has recovered
(Lethborg, Kissane, Burns, & Snyder, 2000; Stanton, Ganz, Rowland et al., 2005).
However, what is most frequently mentioned in the literature is the experience of losing a
safety net (Lethborg, Kissane, Burns, & Snyder, 2000; McKinley, 2000; Ward, Viergutz,
Tormey, deMuth, & Paulen, 1992). The loss of one’s primary means for managing cancer
coupled with uncertainty regarding cancer status appears to be particularly distressing for
cancer survivors. In addition, because women no longer need to focus on their medical
treatment, they may have more time to think about and process thoughts and emotions
related to the cancer experience, which may exacerbate the distress associated with the
loss of the safety net of treatment.
The present study empirically examined whether and to what extent adjustment
was disrupted during the 3 months following the completion of treatment for breast
cancer. Participants’ distress and quality of life were assessed toward the end of adjuvant
chemotherapy and/or radiation treatment, 3 weeks after the end of treatment, and 3
months after the end of treatment. This allowed for an examination of changes in distress
from the “status quo” of treatment to just after treatment ends to 3 months post-treatment.
Previous studies have not compared post-treatment adjustment during this time period
with adjustment during adjuvant treatment.
An important task during the re-entry period is the reconstruction of one’s safety
net. There is ample evidence that active coping strategies are related to better adjustment
among cancer patients and survivors (Carver et al., 1993; Stanton, Danoff-Burg, &
Huggins, 2002; Stanton & Snider, 1993), and it follows that the use of active, behavioral
coping strategies may be an adaptive replacement following the loss of treatment as a
means for managing cancer. There are some guidelines in existence with respect to
recommended health practices for cancer survivors, particularly in the areas of diet,
physical activity, and substance use (ACS Nutrition and Physical Activity Guidelines
35
Advisory Committee, 2001; Brown et al., 2001; National Cancer Institute, 2002). The
NCI’s Facing Forward guide for cancer survivors also provides an array of potentially
adaptive behaviors including many suggestions for health behavior change, stress
management, and services to help cancer survivors cope with physical and psychosocial
problems (National Cancer Institute, 2002). Importantly, there is also evidence that
reducing dietary fat and increasing physical activity levels may decrease the likelihood of
a breast cancer recurrence (Brown et al., 2001). It is presently unclear to what extent
cancer survivors follow such guidelines or behave differently than they did prior to their
cancer diagnosis. An older study found that the majority of breast cancer survivors made
at least one behavioral change, and many made more than that, with changes in diet and
exercise cited most frequently (Taylor, 1984). These changes were generally associated
with concurrent positive adjustment, but there are no prospective data regarding behavior
change and adjustment.
The present study assessed breast cancer survivors’ health behavior, including
diet, exercise, and substance use, as well as other coping behavior, including CAM use,
spiritual or religious activities, and stress management, during treatment and 3 weeks and
3 months after treatment completion. In addition, participants were asked to compare
their current behavior with their behavior before their cancer diagnosis and note the
direction of any changes. This longitudinal examination of behavior changes during the
post-treatment period allowed for an examination of changes with respect to time since
treatment completion. Finally, the relationship between a variety of behavior changes or
behavioral coping strategies and adjustment outcomes was examined.
Whether cancer survivors decide to make behavioral changes and how they
decide which behaviors to change is another important question. Leventhal’s selfregulation theory posits that individuals’ common-sense representations of their illness
guide behavioral responses. Subsequent research indicates that common-sense ideas
about illness determine health-seeking behavior (Grunfeld, Hunter, Ramirez, & Richards,
36
2003; Matthews, Siegal, Kuller, Thompson, & Varat, 1983; Stillman, 1977) and
adherence to medical regimens (Meyer, Leventhal, & Gutmann, 1985). Among cancer
patients, common-sense illness representations are associated with health practices
(Buick, 1997; Christensen et al., 1999; Costanzo, Lutgendorf, Bradley, Rose, &
Anderson, 2005) and information-seeking (Lavery & Clarke, 1996). In addition, there is
evidence that common-sense models are related to psychological adjustment. However,
the nature of these relationships is unclear, and data from our lab suggest that what is
most important with respect to adjustment is that behavior matches common-sense
representations of one’s illness. For example, individuals who believed that poor health
practices played a role in causing their cancer and who were now eating a healthy diet
were less distressed than individuals with the same belief who were not eating a healthy
diet (Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005). Unfortunately, much of
the work on common-sense models in cancer has been cross-sectional which seriously
limits the conclusions that can be drawn.
In the current study, participants’ common-sense models of their cancer were
assessed at all time points and the role of these representations in predicting behavioral
coping efforts was examined. The present study represents the first attempt to examine
these relationships prospectively in cancer patients, as well as one of only a few studies to
conduct a comprehensive assessment of common-sense representations of cancer. In
addition, the interaction between common-sense models of cancer and behavior was
examined as a predictor of post-treatment adjustment. Specifically, we tested whether a
match between common-sense beliefs and behavior was beneficial with respect to
distress and quality of life.
37
CHAPTER II
OBJECTIVES AND HYPOTHESES
See Figure B1 for a visual model integrating the current study’s objectives and
hypotheses.
Objective 1
Examine distress and health-related quality of life during the 3 months following
treatment for breast cancer.
Hypothesis 1
Anxiety and depressive symptoms, as well as cancer-related worry, will increase
from mid-treatment to 3 weeks post-treatment, and will decrease from 3 weeks posttreatment to 3 months post-treatment. Health-related quality of life will improve over
time.
Objective 2
Examine behavioral coping strategies, including health behavior, use of CAM,
spiritual or religious activities, and stress reduction, utilized during the 3 months
following treatment as well as changes in these behaviors since cancer diagnosis.
Determine whether use of behavioral coping strategies concurrently and prospectively
predicts distress during the post-treatment period. Explore the relative efficacy of
different behaviors with respect to distress.
Hypothesis 2
Positive behavior change and greater use of adaptive behavioral coping strategies
will predict less distress, both concurrently and prospectively.
38
Objective 3
Examine common-sense models of breast cancer, including beliefs about personal
control, consequences, illness chronicity, causal attributions, and factors believed to
prevent recurrence, at each time point. Determine whether common-sense representations
of breast cancer concurrently and prospectively predict use of adaptive behavioral coping
strategies, including positive behavior changes, at post-treatment time points.
Hypothesis 3a
Women who perceive greater control over their disease and its sequelae and/or
perceive cancer to be a chronic or recurrent condition rather than an acute condition will
be more likely to utilize behavioral coping strategies.
Hypothesis 3b
Women will choose behavioral coping strategies and make behavior changes that
match their common-sense ideas regarding what caused their cancer and what may
prevent cancer recurrence. For example, women who believe that stress was responsible
for their cancer or that reducing stress may prevent a recurrence will be more likely to
engage in stress-management strategies. Similarly, women who believe that poor diet was
responsible for their cancer or that good diet may prevent a cancer recurrence will be
more likely to eat a healthful diet.
Objective 4
Examine whether common-sense representations of breast cancer and behavioral
coping strategies interact to predict distress concurrently and prospectively during the 3
months following treatment completion.
Hypothesis 4a
Common-sense representations of cancer will interact with behavior to predict
adjustment both concurrently and prospectively. Specifically, women who perceive
39
greater control over their disease and its sequelae and/or perceive cancer to be a chronic
or recurrent condition rather than an acute condition and engage in adaptive behavioral
coping will have less distress as compared to women with the same beliefs who do not
engage in adaptive behavioral coping strategies.
Hypothesis 4b
A match between women’s common-sense ideas regarding cancer cause and
factors that may prevent recurrence and their behavior will predict better psychological
adjustment, while a mismatch will predict poorer adjustment. For example, women who
believe that stress was responsible for their cancer or that reducing stress may prevent a
cancer recurrence and engage in stress-management strategies will have less distress as
compared to women with the same beliefs who do not attempt to reduce life stress.
Similarly, women who believe that poor diet was responsible for their cancer or that a
good diet may prevent a cancer recurrence and eat a healthful diet will have less distress
than women with the same belief who do not make positive dietary changes.
40
CHAPTER III
METHOD
Participants
Inclusion criteria
Participants were women at least 18 years of age who were being treated with
adjuvant chemotherapy, radiation therapy, or both for breast cancer. Women receiving
both types of treatment were included because both groups must adjust to the loss of
treatment as a means for managing their cancer, and the literature indicates that both
groups are distressed following treatment (e.g., Beisecker et al., 1997; Ward et al., 1992).
Women with recurrent or metastatic cancer were excluded. Given these criteria,
participants included in the present study had stages 0 though III breast cancer.
Patients were enrolled from medical and radiation oncology clinics at the
University of Iowa Hospitals and Clinics and University of Iowa outreach clinics (n =
74), Cancer Care of Iowa City (medical oncology) and Iowa City Cancer Treatment
Center (radiation oncology) at Mercy Hospital in Iowa City (n = 14), medical and
radiation oncology clinics at Trinity Cancer Center in Moline, Illinois (n = 14), and
medical and radiation oncology clinics at Aultman Cancer Center in Canton, Ohio (n = 2)
during a 16-month period from March, 2004 through June, 2005. Follow-up data
collection continued through February, 2006.
Procedure
Eligible breast cancer patients were identified by clinic physicians and nurses.
Participants from the University of Iowa and Mercy cancer clinics in Iowa City were
enrolled by study personnel at outpatient radiation or chemotherapy treatment sessions.
Participants from University of Iowa outreach clinics, Trinity Cancer Center in Moline,
IL and Aultman Cancer Center in Canton, OH learned about the study from a clinic
41
nurse, and if they indicated interest by filling out a form giving permission for study
personal to contact them, were contacted by phone and completed consent documents via
mail. At the time of enrollment, informed consent was obtained and participants were
asked to complete the baseline questionnaire packet and return it in a postage-paid
envelope.
Participants were recruited during the middle to later half of their treatment. For
example, women receiving chemotherapy or both chemotherapy and radiation therapy
were generally recruited at cycle 3 of their chemotherapy if they were receiving 4 cycles
or at cycle 6 if they were receiving 8 cycles. Women receiving radiation therapy only
were generally recruited during week 4 or 5 of treatment (women receive an average of 6
or 7 weeks of treatment). This schedule varied on occasion due to changes in treatment
plans (e.g., increasing number of chemotherapy cycles), patients not being identified as
eligible until later in their treatment, or difficulty meeting with potential participants at
the planned treatment session. However, all participants had completed at least two
chemotherapy treatments or two weeks of radiation at the time of enrollment and all were
enrolled before their final treatment. Overall, 83% of participants were enrolled during
the later half of treatment as planned.
The later half of treatment was selected as the baseline assessment time point
because it allowed for maximal adjustment to treatment. There is evidence that distress is
quite high around the time of diagnosis and at the beginning of treatment but decreases
over the course of adjuvant treatment (Frost et al., 2000; Osowiecki & Compas, 1999;
Ward, Viergutz, Tormey, deMuth, & Paulen, 1992). Other evidence suggests that quality
of life improves throughout chemotherapy treatment, but improves more slowly for
women receiving lengthier treatment (Frost et al., 2000; Hurny et al., 1996). Thus, it is
appropriate that the baseline mid-treatment assessment point be timed relative to the
treatment timeline (i.e., approximately three-quarters of the way through treatment) rather
than relative to treatment start (e.g., 5 weeks into treatment for everyone). We also chose
42
not to assess women at the end of treatment because, given evidence that completion of
treatment is distressing (e.g., Beisecker et al., 1997; Lethborg, Kissane, Burns, & Snyder,
2000; McKinley, 2000; Ward, Viergutz, Tormey, deMuth, & Paulen, 1992), distress
levels may begin to rise again as women anticipate the end of treatment and their
transition back into pre-cancer roles. A primary aim of the proposed study was to
compare distress levels after treatment with those during the routine of treatment and thus
a mid-treatment assessment point was thought to be most amenable to this goal.
Following the baseline assessment, participants were asked to complete a packet
of questionnaires 3 weeks following their last treatment session (either chemotherapy or
radiation treatment) and 3 months following treatment completion. Questionnaires for
these two post-treatment time points were mailed to women so that they arrived a few
days before they were to be completed. Participants were asked to return the
questionnaires in a postage-paid envelope. Women were contacted by phone near the date
they were to complete the questionnaires as a reminder and to answer any questions they
had regarding the study. Women were contacted again if the questionnaire was not
received approximately 7-10 days after the date they were to complete the questionnaire.
Questionnaires that were not completed within one month of the end of treatment for the
3-week follow-up or within 3.75 months for the 3-month follow-up were not included in
data analyses.
Measures
An overview of the measures used in the current study and when they were
administered is presented in Table A4.
Demographics
Demographic information including age, completed education, ethnic
background, current employment status, household income, and relationship status were
43
collected at the baseline assessment. Employment and relationship status were reassessed at both post-treatment time points.
Medical information
Participants were asked to report on cancer treatment received, comorbid medical
conditions, and current medications (including Tamoxifen or other anti-estrogen
medication). They were also asked to indicate whether they wore a breast prosthesis or
had undergone reconstructive surgery. Information regarding cancer stage, nodal status,
estrogen receptor status, menopausal status, type and length of cancer treatments,
comorbid medical conditions, and current medication was abstracted from women’s
medical records by the primary author (E.S.C.) at Iowa City clinics or by clinic nurses at
clinics outside of Iowa City.
Depressive symptoms
The Center for Epidemiological Studies Depression Scale (CES-D; Radloff, 1977)
consists of 20 items designed to assess depressive symptomatology such as I had crying
spells and I could not get “going.” Participants rated how often they experienced each
symptom over the past week on a 4-point scale from rarely or none of the time (less than
1 day) to most or all of the time (5-7 days). A cut-off score of 16 has been established as
indicative of probable clinical depression (Ensel, 1986; Radloff, 1977). The scale has
demonstrated good construct validity; the measure discriminated between psychiatric
inpatients and the general population, showed high correlations with negative life events
and other measures of depression, and demonstrated discriminant validity with social
desirability measures (Ensel, 1986; Radloff, 1977). It has also been found to be reliable
and valid among women undergoing treatment for breast cancer and among newly
diagnosed cancer patients (Beeber, Shea, & McCorkle, 1998; D. Hann, Winter, &
Jacobsen, 1999). The CES-D was found to have excellent internal consistency in the
44
present study (Cronbach’s α ranged from .87 to .92). Participants completed the CES-D
at baseline and both post-treatment assessments.
Anxiety
The Primary Care Evaluation of Mental Disorders Patient Health Questionnaire
(PRIME-MD PHQ; Spitzer, Kroenke, Williams, & The Patient Health Questionnaire
Primary Care Study Group, 1999) anxiety scale was used as a measure of general anxiety.
The scale consists of 7 items designed to assess anxiety symptomatology such as feeling
nervous, anxious, on edge, or worrying a lot about different things and feeling restless so
that it is hard to sit still. Participants rated how often they experienced each symptom
over the past week on a 3-point scale from not at all to more than half the days. Items
correspond to DSM-IV diagnostic criteria for generalized anxiety disorder, but were not
used diagnostically in the present study because generalized anxiety disorder requires
symptoms to be present for a minimum of 6 months, and we only examined
symptomatology over the previous week. The PRIME-MD PHQ anxiety scale has been
shown to be a valid diagnostic tool in primary care settings with good agreement between
measure scores and diagnoses by mental health professionals, and it has been correlated
with greater functional impairment, disability days, and health care use (Spitzer,
Kroenke, Williams, & The Patient Health Questionnaire Primary Care Study Group,
1999). In addition, a physician-administered version of the PRIME-MD has been shown
to be a valid measure of mood disorders among cancer patients undergoing radiation
therapy (Leopold et al., 1998). The scale had good reliability in the present sample
(Cronbach’s α ranged from .78 to .80). Participants completed the PRIME-MD PHQ
anxiety scale at baseline and both follow-up time points.
Cancer-related distress
The Impact of Events Scale (IES; Horowitz et al., 1979) is a 15-item scale that
assesses two dimensions derived by factor analysis: intrusive thoughts or rumination and
45
attempts to consciously avoid such thoughts. Participants were asked to rate intrusive
thoughts about cancer and avoidance of these thoughts over the past week on a 4-point
likert-type scale from not at all to often. Examples of items from the intrusion subscale
are I thought about it when I didn’t have to and Other things kept making me think about
it. Examples of avoidance subscale items include I stayed away from reminders about it
and I tried not to talk about it. The IES has been correlated with measures of
posttraumatic stress disorder (Cordova et al., 1995), and has successfully been used to
assess cancer-related anxiety among breast cancer survivors (Bleiker, Pouwer, van der
Ploeg, Leer, & Adèr, 2000; Rothrock, 2002; Tjemsland, Søreide, & Malt, 1996). The
scale demonstrated adequate reliability in the present sample (Cronbach’s α ranged from
.84 to .89 for the intrusion subscale and .72 to .78 for the avoidance subscale).
Participants were asked to complete the IES at baseline and both follow-up time points.
The Concerns About Recurrence Scale (CARS; Vickberg, 2003) was used to
assess worry about cancer recurrence. The CARS consists of two sections; the first
section measures overall fear of recurrence and the second measures content of fears.
Only the four-item overall fear of recurrence section was used in the current study.
Women were asked to rate the frequency, potential for upset, consistency, and intensity
of their worry about cancer recurrence on a 6-point scale. The CARS has been shown to
be correlated with measures of intrusion, avoidance, distress, and well-being (Vickberg,
2003). This scale demonstrated good internal consistency in the current sample
(Cronbach’s α ranged from .91 to .93). Participants completed this scale at baseline and
at both post-treatment assessments.
Participants were also asked to rate to what extent a variety of factors had been a
source of stress for them over the past week on a 5-point scale ranging from not at all to
very much. Fifteen items were chosen based on pilot interviews with breast cancer
survivors and the literature reviewed previously regarding sources of post-treatment
stress among cancer patients, including physical effects of treatment, trying to return to
46
“normal,” fear of recurrence, loss of a “safety net,” and relationship problems. All items
are listed in Appendix C. Items were examined individually rather than treated as a scale.
Participants completed this measure at baseline and at both post-treatment assessments,
although only a subset of the items were administered at baseline because some items
were not relevant until treatment had ended.
Cancer-related symptoms
The Memorial Symptom Assessment Scale (MSAS; Portenoy et al., 1994) is a 32item questionnaire that measures symptoms associated with cancer and its treatment such
as pain, nausea, lack of energy, and hair loss. Participants were asked to rate whether they
had experienced various symptoms in the past week. For each symptom experienced,
participants rated its frequency on a 4-point scale from rarely to almost constantly, its
severity on a 4-point scale from slight to very severe, and to what extent the symptom
bothered or distressed them on a 5-point scale from not at all to very much. Each
symptom endorsed received a score calculated by averaging the frequency, severity, and
distress ratings. The total score is the average of the symptom scores for all 32 symptoms.
A global distress index was also calculated based on the frequency ratings for distressrelated symptoms such as feeling irritable or nervous and the distress ratings of key
physical symptoms such as constipation and dry mouth. MSAS scores have correlated
with quality of life and clinical functioning measures in cancer patients, including women
with breast cancer (Portenoy et al., 1994). The scale has further discriminated between
inpatients and outpatients and between individuals with early versus advanced disease,
and it has also been found to predict survival (Chang et al., 1998). Among breast cancer
patients, MSAS scores are associated with quality of life (D. M. Hann et al., 1997), and
scores have distinguished between responders and nonresponders to chemotherapy
treatment (Modi et al., 2002). The scale demonstrated adequate reliability in the present
47
study (Cronbach’s α ranged from .83 to .90 for the overall symptom score). Participants
were asked to complete the MSAS at baseline and at both post-treatment assessments.
Health-related quality of life
The Medical Outcomes Study Short Form 36 Version 2.0 (SF-36v2) is a 36-item
scale which measures health-related quality of life (Ware, Kosinski, & Dewey, 2000;
Ware, Snow, Kosinski, & Gandek, 1993). The acute version (1-week recall) was used. In
the current study, we were most interested in assessing functional ability and role
functioning. Therefore, six of the eight SF-36v2 scales were used including physical
functioning, role-physical (role limitations due to physical problems), bodily pain,
vitality, social functioning, and role-emotional (role limitations due to emotional
problems) scales. The mental health and general health scales were not included as more
detailed assessments of mental and physical health were covered by other measures. The
SF-36 has been utilized extensively in a variety of medical populations including breast
cancer patients (Turner-Bowker, Bartley, & Ware, 2002). The scale has distinguished
between breast cancer survivors and healthy women (Broeckel, Jacobsen, Balducci,
Horton, & Lyman, 2000), and the physical functioning subscale has discriminated
between individuals receiving no adjuvant treatment versus chemotherapy (Ganz,
Rowland, Meyerowitz, & Desmond, 1998) as well as among individuals in different
phases of breast cancer (Frost et al., 2000). The subscales demonstrated good reliability
in the present study with reliability coefficients ranging from .81 to .93. The SF-36 was
administered at baseline and at both post-treatment assessments.
Common-sense models of cancer
The Illness Perception Questionnaire-Revised (IPQ-R) consists of 38 items
assessing participants’ views of their illness (Moss-Morris et al., 2002). Participants were
asked to rate their agreement with each illness perception with respect to their breast
cancer on a 5-point scale from strongly disagree to strongly agree. The factor-
48
analytically derived scale contains seven subscales measuring acute versus chronic
timeline, cyclical timeline, personal control, treatment control, consequences, illness
coherence, and emotional representations. There are two additional sections of the IPQ-R
that are not part of the scale’s main body. One additional section measures symptoms
believed to be associated with one’s illness, and this section was not used in the current
study. A second additional section assesses beliefs about factors that may have caused
one’s illness. This section was modified as described under causal attributions below.
During the course of the study, we learned that the cyclical timeline subscale was
confusing to some participants. A closer examination of the items revealed that they did
not apply as easily to cancer as did items in other IPQ-R scales due to the focus on illness
“symptoms” (e.g., my symptoms come and go in cycles), and they did not appear to be an
appropriate operationalization of perceiving cancer to be a recurrent condition as we had
intended. Therefore, only the acute versus chronic timeline subscale was used to assess
participants’ perception of cancer timeline.
The IPQ-R scale evidences good reliability and validity. Discriminant validity has
been demonstrated with the Positive and Negative Affect Scale (PANAS) with small to
moderate correlations (Moss-Morris et al., 2002; Weinman, Petrie, Moss-Morris, &
Horne, 1996). The earlier version of the scale has been used successfully in breast cancer
patients to predict adjustment outcomes and health behavior (Buick, 1997). In the present
sample, reliability was generally very good. Acute versus chronic timeline, cyclical
timeline, personal control, illness coherence, and emotional representations scales
demonstrated good internal consistency with Cronbach’s α ranging from .79 to .93. The
treatment control subscale had passable reliability with α ranging from .64 to .88, as did
the consequences subscale with α ranging from .70 to .76. Participants were asked to
complete the IPQ-R at baseline and at both post-treatment time points.
49
Causal attributions
Participants were asked to rate the importance of a variety of factors in causing
their cancer on a 5-point scale from not at all important to very important. Participants
also rated the importance of various factors in preventing cancer recurrence on the same
5-point scale. Items were chosen from previous studies that examined breast cancer
survivors’ beliefs about a variety of potential causal factors and factors that may prevent
recurrence (Stewart et al., 2001; Taylor, Lichtman, & Wood, 1984) and from the casual
factors section on the IPQ-R. An earlier version of the scale was used successfully in a
sample of 150 long-term gynecologic cancer survivors, with individual items predicting
both adjustment and health behavior (Costanzo, Lutgendorf, Bradley, Rose, & Anderson,
2005). Items that were rarely endorsed in this version (e.g., ethnicity) were deleted and
additional items were added that were specific to breast cancer (e.g. use of medication to
prevent recurrence). In addition, as based on the format of the IPQ-R causal attributions
section, participants were asked to rank order the three factors they believe to be most
important in both causing their cancer and preventing its recurrence. These scales can be
found in Appendix C. Participants were asked to complete the section on causal factors at
baseline and at both post-treatment assessments, and they were asked to complete the
section on factors that may prevent recurrence at both post-treatment assessments.
Perceived control
Perceived control was assessed with the IPQ-R personal control and cure control
subscales. The personal control scale contains 6 items assessing control over one’s
illness, and the cure control subscale contains 5 items assessing the belief that treatment
can control one’s illness. Both scales demonstrated adequate reliability and validity as
previously noted (Moss-Morris et al., 2002). An additional 6-item measure assessed
perceived control over different domains of cancer including control over the cause and
long-term course of cancer, side effects and success of treatment, getting needed
50
information, and making medical decisions (see Appendix C). Similar items have been
utilized successfully in head and neck cancer patients (Christensen et al., 1999) as well as
in gynecologic cancer patients (Costanzo, Lutgendorf, Bradley, Rose, & Anderson,
2005). In the present sample, these items could be summed into a single scale that
demonstrated adequate reliability (Cronbach’s α ranged from .67 to .77). Participants
were asked to complete perceived control items at baseline and at both post-treatment
assessments.
Perceived risk of recurrence
Three items were used to assess perceived risk of breast cancer recurrence.
Participants were asked to rate the chance that their breast cancer would recur on a 6point scale from no chance to high chance, how susceptible they felt they were to a
recurrence of breast cancer rated on a 6-point scale from not susceptible at all to highly
susceptible, and the chance that their cancer would recur as compared to other survivors
of breast cancer on a 6-point scale from greatly below average to greatly above average.
The first two items have been used previously (Aiken, Fenaughty, West, Johnson, &
Luckett, 1995; Vickberg & Revenson, 1999) and have been found to be reliable in a
sample of breast cancer survivors (Rothrock, 2002). The third item was used in another
study of breast cancer survivors (Stewart et al., 2001). The three items together
demonstrated good internal consistency in the present sample (Cronbach’s α ranged from
.92 to .93), and were therefore summed as a scale. Participants were asked to respond to
these items at the two post-treatment assessment time points.
Current behavior
Participants were asked to indicate whether they were attending a support group,
receiving psychotherapy or counseling, or using complementary/alternative therapies.
They were also asked to rate how frequently they read material related to cancer on a 5-
51
point scale from never to frequently and how often they performed breast self-exams with
choices ranging from never to daily.
Health practices over the past two weeks were also assessed. Participants were
asked to indicate number of packs of cigarettes smoked per week, number of alcoholic
drinks consumed per week, and average number of hours of sleep per night. These
questions have been previously used successfully in health psychology research and have
demonstrated six-month test-retest reliabilities above .75 (S. Cohen, Doyle, Skoner,
Rabin, & Gwaltney, 1997; S. Cohen, Tyrell, Russell, Jarvis, & Smith, 1993; G. E. Miller,
Cohen, & Herbert, 1999).
Questions on current physical activity were modified from a section of a physical
activity measure used by Kohl and colleagues (1988). Participants were asked to indicate
the number of times per week they engaged in vigorous and moderate physical activity
and estimate the average duration. Examples of vigorous and moderate activities cited in
the questions were modified to make them more relevant to the study population. A
physical activity index was computed by multiplying the frequency and duration of
activity, with vigorous activity weighted more heavily. This index has been correlated
with physical fitness as based on treadmill performance when age is controlled (Kohl,
Blair, Paffenbarger, Macera, & Kronnenfeld, 1988).
Diet was assessed with Rapid Food Screeners for fat and fruit/vegetable intake
(Block, Gillespie, Rosenbaum, & Jensen, 2000). Participants were asked to rate how
frequently they consumed of a variety of foods containing fats using options ranging
from once a month or less to five or more times a week. They were also asked to rate how
frequently they consume different types of fruits and vegetables from less than once a
week to two times or more a day. The screeners were designed to assess the top sources
of fat and fruit/vegetable intake among adults in the United States. Spearman rank-order
correlations between the screeners and a well-validated 100-item food frequency
questionnaire were high (r = .69 for total fat and .71 for fruit/vegetable servings) in a
52
large sample (Block, Gillespie, Rosenbaum, & Jensen, 2000). A review of brief
fruit/vegetable measures found the fruit/vegetable screener to have greater validity than
other similar fruit/vegetable measures (Kim & Holowaty, 2003).
Behavior change
Behavior changes in a variety of domains including health practices (diet,
exercise, alcohol use, smoking, sleep, breast self-exam), CAM use, religious or spiritual
practices, work in and out of the home, leisure activities, stress management, food safety
practices, and writing/journaling were assessed. Women were asked to compare their
current behavior with their behavior prior to diagnosis and rate the direction of change as
less or have now stopped, about the same, or more or have now started. The 24 items,
illustrated in Appendix C, were chosen based on open-ended interviews with breast
cancer survivors who had completed treatment within the past 3 months as well as input
from a clinical psychologist who works extensively with cancer patients (Dr. Nan
Rothrock). Items were also selected from a previous structured interview designed to
assess behavior changes in breast cancer patients in a variety of domains (Taylor,
Lichtman, & Wood, 1984). The items were also chosen on to be consistent with behavior
changes reported by cancer survivors in the literature (Maskarinec, Murphy, Shumay, &
Kakai, 2001; Tatsumura, Maskarinec, Shumay, & Kakai, 2003; Taylor, Lichtman, &
Wood, 1984) as well as suggestions for cancer survivors in NCI’s Facing Forward
publication (National Cancer Institute, 2002) and health practice guidelines for breast
cancer survivors (Brown et al., 2001). Participants were asked to complete this measure
at baseline and at both post-treatment assessments.
Statistical Analyses
Statistical Package for the Social Sciences (SPSS) version 11.0 for Mac OS X
(Chicago) was used to analyze data. Demographic, disease, and treatment characteristics
were described using means, standard deviations, and frequencies. In addition,
53
descriptive statistics were used to characterize outcome variables (CES-D, IES, PRIMEMD anxiety, CARS, MSAS, and SF-36 scale scores). Participants who did not complete
the study were compared to those who completed all time points on demographic,
disease, and treatment variables as well as on all outcome variables to determine whether
attrition was systematic. Univariate ANOVAs were used to examine differences between
individuals who completed versus dropped out of the study.
Age and length of treatment were entered as covariates in all analyses. Length of
treatment was measured in months from the date of either initial surgery or first
neoadjuvant chemotherapy or radiation treatment until the last date of adjuvant
chemotherapy or radiation therapy treatment. Relationships between other potential
covariates and outcome variables were analyzed to determine whether they significantly
predicted outcomes above and beyond age and length of treatment. Specifically,
relationships between disease stage and outcome variables were examined using partial
correlations controlling for age and length of treatment. Relationships between
menopausal status and specific type of treatment (radiation therapy, chemotherapy, or
both) were examined using ANCOVAs with menopausal status or treatment type as the
between-subjects variable, age and length of treatment as covariates, and each outcome
variable as the dependent variable. Covariates demonstrating a significant relationship
with specific outcome variables were included in all analyses involving the outcome
variable with which it was related.
Data reduction
A principal factor analysis (also known as principal axis factor analysis) was
conducted on items measuring causal attributions for cancer to determine whether items
could be collapsed into broader categories. All items were entered into an exploratory
factor analysis and orthogonal (varimax) rotation was used. Factors that emerged were
selected based on both the Scree test and psychometric properties including coefficient
54
α values and interitem correlations. If reliable factors emerged that were theoretically
coherent, category scores rather than individual item scores could be used in analyses.
This procedure was repeated for items measuring factors believed to prevent cancer
recurrence. Also, in the interest of reducing the number of analyses performed, four to six
behavioral changes were selected on the basis of frequency of endorsement and
theoretical interest. For other behavior items (e.g., exercise, diet, attending support
groups, CAM use), four to six behaviors were selected for analyses on the basis of
sufficient variability and theoretical interest. However, all behavior and behavior change
items were characterized in the descriptive analyses.
Objective 1
It was hypothesized that anxiety and depressive symptoms, as well as cancerrelated worry, would increase from mid-treatment to 3 weeks post-treatment and would
decrease from 3 weeks post-treatment to 3 months post-treatment, while health-related
quality of life would increase over time. Mixed models analyses using an unstructured
covariance matrix with time as the repeated measure and time and covariates as fixed
effects were used to examine changes in distress across the three time points. When the
overall test was significant, post-hoc comparisons were performed to compare changes in
distress or quality of life between each combination of time points. An exploratory
analysis examined the effect of treatment type (chemotherapy, radiation therapy, or both)
on distress over time. Mixed models analyses using an unstructured covariance matrix
with time as a repeated measure, and time, treatment type, the interaction between time
and treatment type, and covariates as fixed effects were used to predict distress and
quality of life outcomes. Unlike most other analyses, length of treatment was not
controlled because it is highly related to treatment type and may therefore have masked
the effects of this variable of interest. Analyses were repeated using CES-D, IES
intrusion and avoidance, PRIME-MD anxiety, CARS, MSAS (symptoms and global
55
distress index), and SF36 (physical functioning, role-physical, vitality, social functioning,
and role-emotional scales) scores as outcomes.
Objective 2
Behavioral coping strategies, including health behavior, use of CAM, spiritual or
religious activities, and stress reduction, utilized during the 3 months following treatment
as well as changes in these behaviors since cancer diagnosis were characterized. Means
and standard deviations for behaviors at both post-treatment time points were calculated.
Proportions of individuals increasing, decreasing, or making no change for each behavior
change item were calculated for each post-treatment time point.
In addition to characterizing behavior and behavior changes, the relative efficacy
of different behavioral coping strategies with respect to distress was examined. It was
hypothesized that greater use of adaptive behavioral coping strategies, including positive
behavior change, would predict less distress both concurrently and prospectively. Four
to six behaviors and four to six behavior changes were selected on the basis of frequency
of endorsement and theoretical interest. Adjustment outcomes that were examined in all
analyses described below include depression (CES-D), intrusion (IES), anxiety (PRIMEMD), and cancer-related worry (CARS).
First, concurrent relationships between selected behaviors and distress outcomes
at both post-treatment time points were examined using partial correlations adjusting for
age and length of treatment. ANCOVAs with behavior change (dichotomized as making
a positive behavioral change versus not doing so) as the between-subjects variable and
distress score as the dependent variable were used to test the concurrent relationships
between behavior change domains and distress outcomes at both post-treatment time
points. This analysis was repeated for all selected behavior changes and distress
outcomes.
56
Second, prospective relationships between selected behaviors/behavior changes at
3 weeks post-treatment and distress at 3 months post-treatment were examined.
Hierarchical multiple regression analyses were used to test the prospective relationship
between behavior and distress outcomes. Covariates (age, length of treatment) were
entered in the first step, distress scores at 3 weeks were entered in the second step, and
behavior was entered in the final step. This procedure was repeated for all selected
behaviors and distress outcomes. ANCOVAs with behavior change as the betweensubjects variable, distress score at 3 weeks post-treatment as an additional covariate, and
distress score at 3 months post-treatment as the dependent variable were performed.
Behavior change, the between-subjects variable, was dichotomized as making a positive
behavior change (e.g., decreased fat intake) versus not doing so (e.g., no change or
increased fat intake). This procedure was repeated for all selected behavior changes and
distress outcomes.
Objective 3
Common-sense models of breast cancer were described. In addition to
characterizing common-sense models of breast cancer, concurrent and prospective
relationships between domains of common-sense cancer representations and behavioral
coping strategies were examined. It was hypothesized that women who perceived greater
control over their disease would be more likely utilize adaptive behavioral coping
strategies, including making positive behavior changes. Perceived control was assessed
by the IPQ-R personal control scale.
Means and standard deviations of all IPQ-R scales (acute versus chronic timeline,
cyclical timeline, personal control, treatment control, consequences, illness coherence,
and emotional representations), perceived control items, perceived risk for recurrence
items, causal attributions, and items on factors believed to prevent recurrence were
calculated for each assessment time point.
57
Next, concurrent relationships between personal control and selected behaviors at
both post-treatment time points were examined using partial correlations adjusting for
age and length of treatment. Logistic regression models covarying for age and length of
treatment were used to test concurrent relationships personal control and behavior
changes at both post-treatment time points. Behavior change, the outcome variable, was
dichotomized as described under Objective 2. Second, prospective relationships between
personal control at baseline and behavior 3 weeks post-treatment were examined using
hierarchical multiple regression models. Covariates were entered in the first step,
behavior at baseline was entered in the second step, and personal control was entered in
the final step. Logistic regression models adjusting for covariates and baseline behavior
changes were used to examine relationships between personal control at baseline and
behavior changes 3 weeks post-treatment. Analyses were repeated to assess the
relationship between personal control at baseline and behavior/behavior changes 3
months post-treatment, as well as between personal control 3 weeks post-treatment and
behavior/behavior changes 3 months post-treatment.
It was also hypothesized that women who perceived cancer as a chronic or
recurrent condition rather than an acute condition would be more likely to utilize
adaptive behavioral coping strategies. Therefore, the concurrent and prospective
analyses described above were repeated using the IPQ-R acute versus chronic timeline in
place of personal control. As noted previously, the IPQ-R cyclical timeline scale likely
was not a valid assessment of women’s perceptions of cancer as a “recurrent condition,”
and therefore only the acute versus chronic timeline scale was used in these analyses.
Finally, it was hypothesized that women would choose behavioral coping
strategies and make behavior changes that matched their common-sense ideas regarding
what caused their cancer and what may prevent cancer recurrence. Therefore, the
concurrent and prospective analyses described above were also repeated substituting
attributions regarding causal factors and factors that may prevent recurrence for
58
perceived control. Rather than testing the relationship between all possible attributions
and all possible behaviors, several specific combinations of attributions and behaviors
were identified for testing a priori. For example, the relationships between causal
attribution of diet/eating habits and fat intake, fruit/vegetable intake, and change in
fruit/vegetable intake were examined. Similarly the relationships between the causal
attribution of stress or worry and changes in avoidance of stressful situations and relaxing
activities were tested. Because reliable and theoretically coherent factors emerged in
factor analyses, these factor were also examined in analyses. It is important to note that
beliefs about recurrence prevention were not assessed at the mid-treatment assessment
point, and therefore only the prospective relationships between factors believed to
prevent recurrence 3 weeks post-treatment and behavior 3 months post-treatment were
tested.
Objective 4
It was predicted that common-sense representations of cancer would interact with
behavior to predict distress both concurrently and prospectively. This moderation
hypothesis was tested using interaction terms in multiple regression models (Baron &
Kenny, 1986; Holmbeck, 1997) as described below. In order to contain the number of
analyses to be performed, distress outcomes examined were limited to the CES-D, IES
intrusion subscale, and PRIME-MD anxiety. This allowed for an examination of general
depression and anxiety symptoms as well as cancer-related intrusive thoughts.
The first specific hypothesis was that women who perceived greater control over
their disease and its sequelae and engaged in adaptive behavioral coping strategies
would report less distress as compared to women with the same beliefs who did not
engage in behavioral coping strategies. As noted under Objective 3, perceived control
was assessed by the IPQ-R personal control scale. First, hierarchical multiple regression
analyses were used to determine whether the interaction between perceived control and
59
selected behaviors/behavior changes significantly predicted concurrent distress at both
post-treatment time points. Covariates were entered in the first step, behavior/behavior
change and perceived control were entered in the second step, and the cross-product of
perceived control and behavior/behavior change was entered in the final step. Perceived
control and behavior were centered and behavior change was dichotomized as described
previously under Objective 2. When the interaction term was significant, combinations of
values one standard deviation above and below the mean of perceived control and one
standard deviation above and below the mean of the behavioral variable were substituted
into the regression equation. The resulting four distress values were plotted to examine
the nature of the interaction, and the simple slopes of lines representing scores one
standard deviation above and below the mean on the behavioral variable were tested to
determine whether slopes differed from zero (Aiken & West, 1991). For behavior change
variables, distress values at one standard deviation above and below the perceived control
mean at each level of behavior were similarly plotted and tested to examine the nature of
the interaction. This procedure was repeated for each selected behavior/behavior change.
Second, hierarchical multiple regression analyses were used to test whether the
interaction of perceived control and behaviors/behavior changes at 3 weeks posttreatment predicted distress at 3 months post-treatment. Covariates were entered in the
first step, distress at 3 weeks post-treatment was entered in the second step,
behavior/behavior change and perceived control at 3 weeks post-treatment were entered
in the third step, and the cross-product of perceived control and behavior/behavior change
at 3 weeks post-treatment was entered in the final step. Examination of significant
interactions terms was completed as described above.
It was also hypothesized that women who perceived breast cancer to be a chronic
or recurrent condition rather than an acute condition and engaged in adaptive
behavioral coping strategies would have less distress as compared to women with the
same beliefs who did not engage in behavioral coping strategies. The analyses described
60
above were therefore repeated to test whether the interaction between beliefs about
cancer timeline and selected behaviors/behavior changes significantly predicted
adjustment concurrently and prospectively. The IPQ-R acute versus chronic timeline
scale was used in place of perceived control in the analyses described above.
Finally, it was hypothesized that a match between women’s common-sense ideas
regarding cancer causes and factors that may prevent recurrence and their behavior
would predict less distress, while a mismatch would predict greater distress. Therefore,
the concurrent and prospective analyses described above were also repeated substituting
attributions regarding causal factors and factors that may prevent recurrence for
perceived control. Because reliable and theoretically coherent factors for causal
attributions and recurrence prevention beliefs scales emerged, these factors were used in
place of perceived control.
Power Considerations
Effect sizes (ƒ2) of relationships among behavior, distress, and common-sense
models of cancer in the literature were calculated based on Cohen’s procedure (1992)
when sufficient data were provided. Cohen classifies ƒ2 values of .02 as “small,” .10 as
“medium,” and .35 as “large” (J. Cohen, 1992). Effect sizes for concurrent and
prospective correlations between perceived control and CES-D scores ranged from ƒ2 =
.13 to .59 in Newsom and colleagues’ study of recurrent cancer patients (1996), with
most effect sizes exceeding ƒ2 =.40. Similarly, the relationship between perceived control
and adjustment published in another study of cancer patients had an effect size of ƒ2 = .27
(Thompson, Sobolew, Galbraith, Schwankovsky, & Cruzen, 1993). Relationships
between causal attributions and distress had effect sizes of around ƒ2 =.10 in our study of
gynecologic cancer survivors (Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005).
Relationships between breast cancer survivors’ behavior changes and adjustment had
effect sizes ranging from .06 to .07 (Taylor, Lichtman, & Wood, 1984). Therefore, most
61
relationships identified would be considered to fall in at least the medium effect size
range, with others more appropriately characterized as large effects.
We aimed for a sample of 100 breast cancer patients. While 104 participants were
enrolled in the present study, 88 participants completed some or all of the study
measures. According to Cohen (1992), a sample size of 88 would have a power of .80 to
detect a medium effect size at the α = .05 level using multiple regression models with
three to five independent variables or covariates, as were conducted in the present study.
Due to participant attrition over time, regression models containing four to five
independent variables or covariates at follow-up time points, while more than sufficiently
powered to detect large effects, would have power slightly shy of .80 for detecting
medium effect sizes.
With respect to examining changes in distress over time, effect sizes of
differences between means can be calculated from means and standard deviations, if
available, using the d statistic (J. Cohen, 1992). Cohen classifies a d value of .20 as
“small,” .50 as “medium,” and .80 as “large.” Few studies provided enough information
to calculate effect sizes for changes in distress over time. In one study that did provide
sufficient data for a calculation, the effect size for the change in adjustment from pretreatment to six months post-treatment was d = .53 (Ward, Viergutz, Tormey, deMuth, &
Paulen, 1992), a medium effect. According to Cohen (1992), a sample size of 64 provides
power of .80 to detect medium effects at the α = .05 level in tests of differences between
means. Therefore, the present sample size is sufficient for detecting medium effects or
effects slightly less than d = .50 but will not provide sufficient power for detecting small
effects.
62
CHAPTER IV
RESULTS
Participants
Across all study sites, 104 women consented to participate and were subsequently
enrolled in the study. Of these women, 2 were later deemed ineligible due to the presence
of recurrent breast cancer or to treatment progressing beyond the period of data
collection. A total of 88 participants (86% of those enrolled) completed the baseline
questionnaire packet, 79 (77%) completed the 3-week follow-up packet, and 71 (70%)
completed the 3-month follow-up packet with 71 (70%) completing all three time points.
Chi-square analyses were performed to examine whether women who dropped out
of the study differed on demographic, disease, or outcome variables. Women who were
enrolled in, but did not complete the entire study, did not differ from those who
completed all time points on disease variables including cancer stage, type of treatment,
or length of treatment. They were no more likely to have been prescribed a psychotropic
medication. They also did not differ on most demographic variables including age,
ethnicity, income, or employment status. Women who did not complete the study did
differ on partner status, however, χ2 = 10.47, p = .02. Specifically, non-completers were
more likely to be divorced or separated and less likely to be married or single. ANOVAs
were used to compare women who completed at least one questionnaire packet but later
dropped out to those who completed all time points on study outcome variables at
baseline. After removing an outlier on two distress measures (described below) who
dropped out at the third time point, there were no differences for any study outcome
variables (all p values exceeded .10).
Participants in the current study ranged in age from 32 to 89 years of age with a
mean age of 55.0 years. The majority of participants were married or living with a partner
(74%). The sample was well-educated, with 68% having completed some post-secondary
63
education. At the time of the study, most participants worked full or part-time (63%).
Full demographic data is presented in Table A5. According to participants’ self-report,
27% had been diagnosed with depression and 19% had been diagnosed with an anxiety
disorder in the past. At the time of study entry, 28% of participants were currently
prescribed psychotropic medication, according to medical records.
Participants had stage 0 (6%), I (33%), II (47%), or III (14%) breast cancer. More
details regarding participants’ disease characteristics are provided in Table A6. All
participants had surgery, with the majority undergoing a lumpectomy (80%) and fewer
having a mastectomy (27%; some participants had an initial lumpectomy followed by a
later mastectomy). All patients received chemotherapy (72%) or radiation therapy (86%),
with 58% of participants receiving both types of treatment. Typical treatment regimens
comprised 4 or 8 cycles of chemotherapy (although some participants received up to 12
cycles) and 5 to 7 weeks of radiation therapy. A small subset of participants (11%)
received some or all of their chemotherapy prior to surgery. Beginning with surgery or
first date of neoadjuvant chemotherapy, participants received 2 to 12 months of treatment
with a mean of 6.0 months of treatment. The date of 3-week and 3-month follow-ups
were calculated based on the last adjuvant chemotherapy or radiation therapy treatment.
During the follow-up period, 77% of participants were prescribed hormonal therapy;
either tamoxifen (46%) or arimidex (54%). In addition, 3 participants began herceptin
therapy on a monthly basis at some point during the follow-up period after results of a
study were released demonstrating the benefit of this treatment for some early-stage
patients (Romond et al., 2005).
All variables were examined for outliers. One individual scored more than 4
standard deviations above the mean on the CES-D and more than 3 standard deviations
above the mean on the IES intrusion scale. Although the scores were accurate, no other
participants scored close to this individual. To avoid skewing results, this participant was
64
removed from analyses investigating relationships between other variables and the CESD or IES intrusion scale.
Age and length of treatment were entered as covariates in all analyses. Other
medical variables, including cancer stage, treatment type, and menopausal status were
examined as potential covariates of outcome measures after adjusting for age and length
of treatment. Partial correlations demonstrated a significant relationship between disease
stage and MSAS total score, pr = .22, p = .04 as well as SF-36 physical functioning, pr =
-.24, p = .03 at baseline, and therefore stage was entered as a covariate in models
examining these baseline variables. There were no associations between stage and other
outcomes or between stage and these variables at follow-up time points, all ps > .10.
ANCOVA models demonstrated a significant association between treatment type and SF36 vitality at 3 months post-treatment, F(2, 64) = 4.09, p = .021, with participants who
received chemotherapy only reporting greater vitality than those who received either
radiation only or both radiation and chemotherapy. Therefore, treatment type was entered
as an additional covariate in analyses examining vitality at 3 months post-treatment.
There were no other associations between treatment type and outcomes, all ps > .10. In
addition, there were no associations between menopausal status and outcomes after
adjusting for age and treatment length, all ps > .10.
Objective 1: Distress and Health-Related Quality of Life
Distress
Means and standard deviations of distress measures are provided in Table A7.
Overall, participants in the present study were not very distressed at any study time point.
CES-D mean scores ranged from 9.1 to 11.1, well under the cut off score of 16 indicative
of clinically significant depression. In fact, only 15.9% of participants exceeded this cutoff at baseline, and similarly small proportions of the sample exceeded the cut-off score
at 3 weeks post-treatment (16.9%) and 3 months post-treatment (15.9%). Scores on
65
measures of anxiety, including the IES and PRIME-MD were also low, with women
reporting on average that they experienced anxiety symptoms and intrusive thoughts
rarely or occasionally. The majority of women did not feel as if they had lost a safety net
following treatment, with mean responses falling between “not at all” and “somewhat” at
both follow-up time points. Nonetheless, a significant subset did express this concern,
with 34.3% of women at 3 weeks post-treatment and 36.7% of women at 3 months posttreatment reporting that they felt they had “somewhat” to “very much” lost a safety net.
Worry about cancer recurrence, as assessed by the CARS, was more prominent in the
present sample, with women reporting moderate levels of worry and distress related to
the possibility of a recurrence.
With respect to potential sources of stress, at baseline, participants rated side
effects or physical problems related to cancer and treatment as most stressful, followed
by impact of cancer on the family, and fear of a recurrence, all of which were perceived
as moderately stressful on average. Other sources of stress included concerns about
ability to fulfill responsibilities at work or at home as well as emotions or emotional wellbeing. On average, women experienced little stress related to relationships or lack of
support. At follow-up, side effects or physical problems and fear of recurrence were the
greatest sources of stress, with both rated to be moderately stressful on average. Getting
back to normal or trying to create a “new normal” were also notable sources of stress, as
were impact of cancer on one’s family, fear that the cancer is not gone, and uncertainty
about what to do for one’s health or to prevent a recurrence now that treatment has ended.
Consistent with women’s responses to the “safety net” question, feeling like one has lost
a safety net and not seeing one’s oncologist or other health care staff regularly were rated
as “not at all” to “a little bit” stressful on average. As was the case at baseline,
relationship problems and lack of support were not significant sources of stress. Mean
ratings of stressors at all time points are presented in Table A8.
66
Changes in distress over time
Mixed models analyses covarying for age and treatment length were used to test
the hypothesis that anxiety and depressive symptoms, as well as cancer-related worry,
would increase from mid-treatment to 3 weeks post-treatment, and would decrease from
3 weeks post-treatment to 3 months post-treatment (Hypothesis 1). Results of these
analyses indicated that there were no significant changes in distress over the three study
time points. IES intrusion was the only distress variable that approached significance,
F(2, 74.96) = 2.64, p = .08. All other p values exceeded .10. An examination of distress
measure means indicated that distress decreased slightly across time, with no increase 3
weeks post-treatment as hypothesized (see Table A7). Recurrence worry was the only
exception; mean CARS scores increased slightly from baseline (11.8) to 3 weeks posttreatment (12.4) and then decreased from 3 weeks to 3 months post-treatment (11.5).
Again, however, these changes were not statistically significant. An analysis of the effect
of treatment on distress over time was also performed using a mixed models strategy
covarying for age. There was no main effect of treatment type (chemotherapy only versus
radiation therapy only versus both) on distress or on change in distress over time for any
of the distress measures (all ps > .10 for both main effects and effect on slopes).
Quality of life
Means and standard deviations of health-related quality of life measures are
provided in Table A7. Scores on SF-36 domains were compared with population norms
of U.S. women using information provided on the measure’s official website (SF-36.org,
n.d.-a, n.d.-b). Women in the current study reported health-related quality of life that was
generally on par with population norms. At baseline, participants’ scores on physical
functioning and role-physical (impairment in daily activities due to physical health)
subscales were slightly lower than population norms (higher scores indicate better
functioning), while vitality, social functioning, and role-emotional (impairment in daily
67
activities due to emotional problems) scales were similar to population norms. Notably,
participants in the present study exceeded population norms on the bodily pain scale at
baseline, indicating that the current sample reported less pain and pain-related
impairment than the population at large. At follow-up time points, participants’ scores on
physical functioning and role-physical subscales increased to levels commensurate with
population norms, while vitality, social functioning, and role-emotional scales increased
to the extent that they exceeded population norms. Bodily pain scores also continued to
exceed population norms at follow-up time points.
Changes in quality of life over time
Mean scores on quality of life measures indicated that health-related quality of
life generally increased over time, as hypothesized (see Table A7). Mixed models
analyses adjusting for age, treatment length, and stage were used to test the hypothesis
that quality of life would improve over time (Hypothesis 1). These analyses revealed
significant changes in MSAS symptom scores, F(2, 72.52) = 10.88, p < .001. Follow-up
contrasts indicated that symptoms decreased from baseline to 3 weeks post-treatment and
then remained steady. No significant changes in symptom-related distress as measured by
the global distress index of the MSAS were found, however, p > .10. This may have been
due in part to the greater variability of scores on this scale or to the fact that this scale is
similar to other distress measures, which also did not change over time. With respect to
SF-36 quality of life domains, mixed models adjusting for age and treatment length (as
well as treatment type for the vitality subscale) demonstrated significant changes over
time for physical functioning, F(2, 76.02) = 4.53, p = .01; role-physical, F(2, 75.79) =
13.34, p < .001; vitality, F(2, 77.53) = 3.94, p = .024; and social functioning, F(2, 74.37)
= 6.12, p = .003. In all cases, quality of life improved over time, with contrasts revealing
better quality of life at follow-up time points as compared to baseline (see Table A7).
There were no significant changes for bodily pain or role-emotional scales, both ps > .10.
68
A mixed models analysis of the effect of treatment on quality of life was also performed.
There were no main effects of treatment type on quality of life or on change in quality of
life over time for any of the quality of life measures after adjusting for age, although a
main effect of treatment type on bodily pain approached significance, F(2, 77.80) = 2.93,
p = .059. All other p values exceeded .10 for both main effects and effect on slopes.
Objective 2: Behavioral Coping Strategies
Post-treatment behavior changes
Behavior changes were very common following treatment, particularly
improvement in health behaviors (see Table A9). At 3 weeks post-treatment, participants
often reported that, as compared to before their cancer diagnosis, they had increased the
frequency of breast self-examination (40%), increased fruit and vegetable consumption
(31%), began using or increased use of dietary supplements (30%), and decreased alcohol
use (30%). Among smokers, 58% reported decreasing or quitting smoking. Women were
less likely to report doing more moderate or vigorous physical activity (15% and 4%,
respectively), more often getting adequate sleep (8%), and more frequently choosing noor low-fat foods (5%). Similar percentages of women reported health behavior changes 3
months post-treatment, with a few exceptions. More women reported increasing moderate
physical activity (29%) and choosing no- or low-fat foods more often (19%), and more
smokers had decreased or quit smoking (90%).
Common changes in non-health behaviors included increased prayer or
meditation (25%) and more frequent avoidance of stressful situations (25%). Smaller but
substantial proportions reported journaling or writing more about their experiences (12%)
and engaging in more food safety practices (e.g., washing produce; 12%). Women also
made changes in the way they spent their time. Three weeks after treatment, participants
reported working less both inside and outside of the home (14% and 13%, respectively)
but spending more time with family and friends (22% and 21%, respectively) or engaging
69
in hobbies and relaxing activities (13% and 12%, respectively). Similar percentages of
women reported these behavior changes 3 months after treatment, with the exception of
avoidance of stressful situations: fewer women reporting more frequent avoidance of
stress (16%). Table A9 illustrates behavior changes at both post-treatment time points.
Post-treatment health and psychosocial behaviors
The frequency of key health behaviors, including fat and fruit/vegetable
consumption, physical activity, alcohol use, smoking, sleep, breast self-exam, and
attending scheduled follow-up appointments was also assessed (see Table A10 for means
and percentages). Women’s diets were quite close to dietary recommendations; scores on
fat and fruit/vegetable screeners indicated participants consumed, on average,
approximately 32% of calories from fat (30% or fewer calories from fat are
recommended) and 4 to 5 servings of fruits and vegetables daily (5 or more servings are
recommended). Participants also reported consuming very little alcohol (average of one
drink per week), getting an average of 7 hours of sleep per night, and performing breast
self-exams on slightly more than a monthly basis. Few women had missed or cancelled
follow-up appointments (3% at 3 weeks and 9% at 3 months post-treatment).
Few women reported participating in traditional supportive treatments including
support groups (6% at 3 weeks and 9% at 3 months post-treatment) or psychotherapy (4%
at 3 weeks and 1% at 3 months post-treatment). Women were somewhat more likely to
report that they used complementary or alternative therapies (15% at 3 weeks and 17% at
3 months post-treatment); commonly cited therapies included herbs, massage, yoga, or
healing touch. Women were also asked whether they had started any new activities since
their diagnosis (e.g., joining a choir, taking an art class). Three weeks after treatment had
ended, only 13% had done so, but 28% had started a new activity three months after
treatment ended.
70
Concurrent behavior-distress relationships
Relationships between behavioral coping strategies and distress outcomes at both
post-treatment time points were examined. ANCOVA models covarying for age and
length of treatment were used to test the hypothesis that positive behavior change would
predict less concurrent distress (Hypothesis 2). Five key behavior changes were chosen
on the basis of frequency of endorsement and theoretical interest, including changes in
breast self-exam, fruit/vegetable consumption, alcohol use, physical activity, and
avoidance of stressful situations. Behavior change was dichotomized as making a positive
change in the behavior versus not doing so. Partial correlations adjusting for age and
treatment length were used to test the hypothesis that positive health behaviors would
predict less concurrent distress (Hypothesis 2). Five key behaviors were chosen on the
basis of sufficient variability and theoretical interest, including breast self-exam, fat
consumption, fruit/vegetable consumption, physical activity, and alcohol use. Four
distress outcomes, including CES-D depression, IES intrusion, PRIME-MD anxiety, and
CARS recurrence worry were examined.
At 3 weeks post-treatment, there were no relationships between behavior changes
examined and distress measures, all ps > .10. The majority of behaviors examined were
also unrelated to distress, with a few exceptions. Alcohol use was associated with greater
intrusion, pr = .31, p = .010, and recurrence worry, pr = .32, p = .007. There was also a
marginal relationship between greater fruit/vegetable consumption and less recurrence
worry, pr = -.21, p = .072. There were no relationships between fat consumption,
physical activity, or self breast-exam and any distress measures, all ps > .10.
At 3 month post-treatment, more behavior-distress relationships were found, with
most findings indicating that making a positive behavior change was associated with
greater distress. With respect to behavior changes, those who had increased their
fruit/vegetable consumption reported greater depression, F(1, 65) = 15.15, p < .001;
intrusion, F(1, 65) = 5.84, p = .018; and anxiety, F(1, 65) = 7.98, p = .006; as well as
71
marginally greater recurrence worry, F(1, 65) = 3.75, p = .057. Women who had
decreased alcohol use reported greater recurrence worry than those who did not, F(1, 40)
= 5.11, p = .029. Participants who reported increasing the frequency of self breastexamination also had greater anxiety, F(1, 64) = 5.43, p = .023, and marginally greater
recurrence worry F(1, 66) = 3.90, p = .052. Finally, women who more frequently avoided
stressful situations reported greater intrusion, F(1, 64) = 8.01, p = .006. There were no
significant relationships between changes in physical activity and distress outcomes, all
ps > .10. With respect to the frequency of key behaviors, greater fat consumption was
associated with greater anxiety, pr = .36, p = .003, and there were marginal associations
between greater alcohol use and greater intrusion, pr = .22, p = .076, as well as between
more frequent self breast-exam and less recurrence worry, pr = -.24, p = .050.
Fruit/vegetable consumption and physical activity were not associated with distress
measures, all ps > .086.
Prospective behavior-distress relationships
ANCOVA models adjusting for covariates and distress outcome measures at 3
weeks post-treatment were performed to test the hypothesis that positive behavior
changes 3 weeks post-treatment would predict less distress 3 months post-treatment
(Hypothesis 2). Consistent with concurrent associations, women who had increased
fruit/vegetable consumption 3 weeks post-treatment reported greater depression, F(1, 64)
= 12.42, p = .001, intrusion, F(1, 64) = 6.28, p = .015, and anxiety, F(1, 63) = 7.90, p =
.007, three months post-treatment. Women who had increased the frequency with which
they avoided stressful situations 3 weeks post-treatment also reported greater depression,
F(1, 62) = 6.74, p = .012, and anxiety, F(1, 62) = 5.25, p = .025, and marginally greater
intrusion, F(1, 63) = 3.57, p = .064, 3 months after treatment had ended. Other behavior
changes showed marginal associations with distress outcomes. Those who had increased
physical activity 3 weeks post-treatment reported marginally less recurrence worry 3
72
months post-treatment, F(1, 63) = 3.26, p = .076; decreased alcohol use was associated
with marginally greater depression, F(1, 36) = 3.29, p = .078; and increased breast selfexam was associated with marginally greater recurrence worry, F(1, 61) = 3.40, p = .070.
Hierarchical multiple regression models were used to test the hypothesis that positive
health behaviors 3 weeks post-treatment would predict less distress 3 months posttreatment (Hypothesis 2). In these models, covariates (age and length of treatment) were
entered in the first step, distress 3 weeks post-treatment was entered in the second step,
and behavior at 3 weeks post-treatment was entered in the final step. Behavior at 3 weeks
post-treatment did not account for additional variance in 3-month post-treatment distress
measures in any of the models examined. A model examining whether physical activity
predicted recurrence worry, the only model in which the p value associated with R2
change in the final step was less than .10, is illustrated in Table A11. As illustrated in this
table, in prospective regression models examined, distress scores at 3 weeks posttreatment accounted for a significant proportion of the variance in distress scores at 3
months post-treatment and explained more variance than did any of the other predictors.
Objective 3: Common-Sense Models of Breast Cancer and
Coping Behavior
Domains of cancer representations
The Illness Perception Questionnaire assessed several domains of participants’
common-sense beliefs about breast cancer including beliefs about timeline,
consequences, illness coherence, emotional representations, and personal and treatment
control. Means and standard deviations of all subscales at each time point are presented
in Table A12. On average, women perceived their cancer to be more acute than chronic,
although not strongly so (acute versus chronic timeline), and they did not perceive their
cancer to be a cyclical or condition (although, as previously noted, the cyclical timeline
scale may not have been valid in this sample). They perceived their cancer to have
73
moderate consequences for their lives (consequences) but reported little emotional
distress related to their cancer (emotional representations). Overall, women reported that
they understood their cancer or it made sense to them (illness coherence), they believed
that they had moderate control over their cancer (personal control), and they saw their
treatment as quite effective in controlling and curing their cancer (treatment control).
A measure examining perceived control over various aspects of cancer and
treatment was also administered, and means and standard deviations at each time point
are presented in Table A12. Consistent with results on the IPQ personal control measure,
women reported moderate levels of control over their cancer overall. They perceived
almost complete control over getting needed information about cancer and very high
levels of control over decisions regarding treatment and medical care. Women believed
they had moderate control over the success of their treatment, and they perceived a
modest level of control over treatment side effects and the long-term course of their
cancer. They believed they had little control over the cause of their cancer. Women’s
perceived risk of recurrence was also assessed, and women perceived themselves to be
moderately susceptible to a cancer recurrence (see Table A12). As Table A12 illustrates,
participants’ common-sense beliefs about cancer were consistent over time.
Finally, women rated the extent to which they believed a variety of factors were
important in causing their cancer and in preventing a recurrence. Tables A13 and A14
illustrate mean ratings as well as the percentage of participants who believed each factor
was at least somewhat important in causing cancer or preventing recurrence. At baseline,
women rated hormones, followed by environmental toxins or hazards, genetics or
heredity, and diet as important in causing their cancer (see Table A13). The majority of
participants saw these factors as somewhat to very important in causing their cancer.
Stress or worry and aging were also perceived as important causal factors. In contrast,
few women perceived an injury, a germ or virus, or past poor medical care as important
in causing their cancer. At 3 weeks post-treatment, women believed that medical
74
checkups or screenings were most important in preventing a recurrence, followed by
having a positive attitude, eating a healthy diet, taking medication, and exercising (see
Table A14). Overall, more than 89% of women perceived these behaviors to be at least
somewhat important in preventing a recurrence. They were less likely to believe that
God’s will, chance or luck, and use of complementary or alternative therapies played an
important role in preventing recurrence. As was the case with other dimensions of
participants’ beliefs about their breast cancer, causal attributions and recurrence
prevention beliefs were consistent over time, and the relative ordering of factors also
remained nearly the same across time points (see Tables A13 and A14).
A principal factor analysis using orthogonal (varimax) rotation was conducted on
causal attributions and recurrence prevention items at all time points to determine
whether items could be collapsed into broader categories. For causal attributions items,
one reliable factor emerged at all three time points that consisted of the following
attributions: poor diet, lack of exercise, stress or worry, and mental attitude (e.g., thinking
negatively). This factor was reliable, with Cronbach’s α ranging from .78 to .84 across
time points. The factor appeared to measure the extent to which women believed that
behavioral or psychological factors which they had control over were important in
causing their breast cancer. No consistent or reliable second factor emerged. With respect
to recurrence prevention beliefs, a reliable factor consisting of parallel behavioral and
psychosocial items emerged: eating a healthy diet, exercise, reducing stress, and having a
positive attitude comprised this factor. Cronbach’s α values were .78 and .80. No reliable
second factor emerged. This “behavioral and psychological” grouping of attributions was
remarkably robust as demonstrated by this factor emerging across all time points and for
both causal attributions and recurrence prevention beliefs. Factor loadings are presented
in Table A15.
75
Personal control and coping behavior
Logistic regression models and partial correlations with age and length of
treatment entered as covariates were used to test the hypothesis that women who
perceived greater control over their disease and its sequelae would be more likely to
utilize behavioral coping strategies (Hypothesis 3a). In these models, concurrent
relationships between IPQ personal control and the 5 selected behaviors and 5 selected
behavior changes were examined. Only relationships of marginal significance were found
between personal control and behavioral outcomes, most indicating that personal control
was associated with adaptive behavioral outcomes. For example, at three weeks posttreatment, women who perceived greater personal control over cancer were marginally
more likely to increase avoidance of stressful situations, Wald = 3.18, p = .075, and
greater personal control was also marginally associated with greater fruit and vegetable
consumption, pr = .23, p = .051. At three months post-treatment, perceived control was
marginally associated with less fat consumption, pr = -.24, p = .062.
Logistic and hierarchical regression models were used to examine prospective
relationships between personal control at baseline and post-treatment behavioral
outcomes. Covariates were entered in the first step, the outcome measure at baseline was
entered in the second step, and personal control at baseline was entered in the final step.
This procedure was also repeated to examine relationships between personal control 3
weeks post-treatment and behavior 3 months post-treatment. Personal control at baseline
did not predict behavioral outcomes 3 weeks or 3 months post-treatment above and
beyond the effects of covariates and baseline behavior, all ps > .10. There were also no
relationships between personal control at 3 weeks post-treatment and behavior or
behavior changes 3 months after treatment had ended, all ps > .09. Figure B2 illustrates
marginally significant relationships found between personal control and behavioral
outcomes across time points.
76
Cancer chronicity beliefs and behavior
The hypothesis that women who perceived their cancer to be a chronic rather than
an acute condition would be more likely to utilize behavioral coping strategies
(Hypothesis 3a) was tested using the same statistical methods described above for
perceived control. Concurrent and prospective relationships between the IPQ acute versus
chronic timeline measure and the 5 selected behaviors and 5 selected behavior changes
were examined. In most cases, analyses indicated that perceiving cancer to be more
chronic in nature was associated with maladaptive behavioral outcomes.
With respect to concurrent relationships, at three weeks post-treatment, women
who perceived their cancer to be more chronic in nature were less likely to increase
frequency of breast self-exam, Wald = 4.05, p = .044. Perceiving cancer to be more
chronic was also associated with greater alcohol use, pr = .24, p = .050. At three months
post-treatment, there were no significant relationships between perception of chronicity
and any behavioral outcomes. However, perceiving cancer to be a chronic condition
continued to be marginally associated with greater alcohol use, pr = .22, p = .076.
With respect to prospective relationships, women who perceived cancer to be a
chronic condition at baseline were more likely to decrease alcohol intake 3 weeks posttreatment, Wald = 4.27, p = .039, but were less likely to increase frequency of breast selfexam 3 weeks after treatment had ended, Wald = 5.02, p = .025. Chronicity beliefs at
baseline did not predict any behavioral outcomes at 3 months after treatment had ended,
all ps > .086. Perceiving cancer to be more chronic in nature at 3 weeks post-treatment
was marginally associated with more frequent breast self-exam 3 months after treatment
had ended, pr = .23, p = .075 and was associated with greater fat consumption 3 months
post-treatment, pr = .26, p = .041. In the regression model predicting fat consumption at 3
months post-treatment, chronicity beliefs at 3 weeks post-treatment accounted for
significant additional variance in fat consumption after covariates and fat consumption at
3 weeks post-treatment had been entered into the model (see Table A16). Figure B3
77
illustrates all significant and marginally significant concurrent and prospective
relationships found between chronicity beliefs and behavioral outcomes across time
points.
Causal attributions and behavior
The hypothesis that women would choose behavioral coping strategies and make
behavior changes that match their common-sense ideas regarding what caused their
cancer (Hypothesis 3b) was tested using the same methods described above.
Relationships between the behavioral/psychological scale that emerged in a factor
analysis (see Table A15) and the 5 selected behavior changes and 5 selected behaviors
were examined, as were relationships between pre-specified combinations of causal
attributions and behaviors, which are listed in Table A17. With a few exceptions,
attributing cancer to behavioral and psychological causes was generally associated with
adaptive behavioral outcomes. At 3 weeks post-treatment, women who attributed their
cancer to behavioral and psychological causes were marginally more likely to have
increased the frequency with which they avoided stressful situations, Wald = 3.79, p =
.052. Attributing cancer to behavioral and psychological causes was also associated with
less physical activity, pr = -.24, p = .048, and marginally less alcohol use, pr = -.21, p =
.075. With respect to specific attribution-behavior combinations tested, women who
attributed their cancer to stress were also marginally more likely to have increased the
frequency with which they avoided stressful situations, Wald = 3.12, p = .077.
Attributing cancer to lack of exercise was associated with less physical activity, pr = -.27,
p = .023, while attributing cancer to substance use was marginally associated with more
alcohol use, pr = .23, p = .059.
At three months post-treatment, women who attributed cancer to behavioral and
psychological causes were more likely to have increased fruit/vegetable consumption,
Wald = 5.16, p = .023, and to have decreased alcohol use, Wald = 5.48, p = .019. This
78
attribution was also marginally associated with less alcohol use, pr = -.22, p = .077.
Women who attributed cancer to diet were also more likely to have increased
fruit/vegetable consumption, Wald = 5.10, p = .024. Finally, attributing cancer to lack of
exercise was marginally associated with less physical activity, pr = -.24, p = .061.
With respect to prospective relationships, women who attributed cancer to
behavioral and psychological factors at baseline were more likely to have increased
fruit/vegetable consumption 3 weeks post-treatment, Wald = 6.55, p = .011 and 3 months
post-treatment, Wald = 5.08, p = .024. This baseline attribution also predicted less
alcohol consumption 3 weeks post-treatment, pr = -.24, p = .043 (see Table A18).
Attributing cancer to diet at baseline was also associated with increased fruit/vegetable
consumption 3 weeks post-treatment, Wald = 13.40, p < .001 and 3 months posttreatment, Wald = 6.87, p = .009, and attributing cancer to alcohol or tobacco use at
baseline predicted less alcohol consumption 3 weeks post-treatment, pr = -.30, p = .015
(see Table A19).
Women who attributed cancer to behavioral and psychological factors at 3 weeks
post-treatment were also more likely to have increased fruit/vegetable consumption 3
months post-treatment, Wald = 4.45, p = .035. Those who attributed cancer to lack of
exercise 3 weeks post-treatment were more likely to have increased physical activity 3
months post-treatment, Wald = 3.90, p = .048. Finally, women who attributed cancer to
stress or worry 3 weeks after treatment had ended were marginally more likely to have
increased the frequency with which they avoided stressful situations 3 months posttreatment, Wald = 3.12, p = .077. Figure B4 illustrates all significant and marginally
significant concurrent and prospective relationships found between attributing cancer to
behavioral and psychological causes and behavioral outcomes across time points. Figures
B5 and B6 illustrate the same for diet and exercise attributions.
79
Recurrence prevention beliefs and behavior
The statistical methods described previously were also used to test the hypothesis
that women would choose behavioral coping strategies and make behavior changes that
matched their beliefs regarding what may prevent their cancer from recurring
(Hypothesis 3b). Concurrent and prospective relationships between recurrence prevention
beliefs and selected behavioral outcomes were examined. In general, behaviors matched
beliefs about recurrence prevention. At 3 weeks post-treatment, women who believed that
behavioral and psychological factors could prevent a cancer recurrence were more likely
have increased fruit/vegetable consumption, Wald = 4.41, p = .036, and to have increased
the frequency with which they avoided stressful situations, Wald = 4.72, p = .030, and
they were marginally more likely to have decreased alcohol use, Wald = 3.52, p = .061.
This attribution was also associated with less fat consumption, pr = -.28, p = .017.
Women who believed that decreasing or quitting use of alcohol or tobacco could prevent
a recurrence were more likely to have decreased their alcohol use, Wald = 4.70, p = .030,
and those who believed that decreasing stress could prevent a recurrence were more
likely to have increased the frequency with which they avoided stressful situations, Wald
= 9.36, p = .002, and engaged in relaxing activities, Wald = 6.63, p = .010. The belief that
a healthy diet could prevent recurrence was associated with marginally less fat
consumption, pr = -.21, p = .078.
At three months post-treatment, women who believed that behavioral and
psychological factors could prevent a recurrence were more likely to have increased their
fruit and vegetable consumption, Wald = 8.46, p = .004. Those who believed that a
healthy diet could prevent recurrence were also more likely to have increased fruit and
vegetable consumption, Wald = 6.20, p = .013; those who believed that exercise could
prevent a recurrence were more likely to have increased physical activity, Wald = 4.24, p
= .039; those who believed that reducing or quitting use of alcohol or tobacco could
prevent recurrence were more likely to have decreased alcohol use, Wald = 5.23, p =
80
.022; and those who believed that reducing stress could prevent recurrence were more
likely to have increased the frequency with which they avoided stressful situations, Wald
= 6.23, p = .013, and engaged in relaxing activities, Wald = 3.75, p = .053.
As we did not assess recurrence prevention beliefs at baseline, only the
prospective relationships between beliefs 3 weeks post-treatment and behavior 3 months
after treatment were examined. Only one significant relationship was found: women who
believed that behavioral or psychological factors could prevent recurrence at 3 weeks
post-treatment were more likely to have increased fruit/vegetable consumption at 3
months post-treatment, Wald = 4.74, p = .029. Figure B7 illustrates all significant and
marginally significant concurrent and prospective relationships found between the belief
that behavioral and psychological factors can prevent recurrence and behavioral
outcomes across time points. Figures B8 and B9 illustrate the same for recurrence
prevention beliefs regarding substance use and stress.
Objective 4: Interactions Between Common-Sense Beliefs
and Behavior and Distress Outcomes
Personal control-behavior interactions and distress
Multiple regression models were used to test the hypothesis that women who
perceived greater control over their cancer and its sequelae and engaged in adaptive
behavioral coping would have less distress as compared to women with the same beliefs
who did not engage in adaptive behavioral coping strategies (Hypothesis 4a).
Specifically, we examined whether interactions between personal control and behavior
predicted distress outcomes both concurrently and prospectively above and beyond the
effects of covariates and either predictor alone. In concurrent models, covariates (age and
length of treatment) were entered in the first step, personal control and the behavior or
behavior change in question were entered in the second step, and the cross-product of
personal control and behavior was entered in the final step. Distress outcomes examined
81
included CES-D depression, IES intrusion, and PRIME- MD anxiety. Interactions that
were significant at p < .05 were plotted and the slopes analyzed as described in the
method section to examine the nature of the interaction.
At 3 weeks post-treatment, there were no significant interactions between
personal control and behavior in models predicting distress outcomes (all p values for
interaction terms in the models exceeded .072). At 3 months post-treatment, the
interaction between personal control and fat intake was significant in the model
predicting intrusion, β = .26, p = .033 (see Table A20). An analysis of the simple slopes
of the plotted interaction (see Figure B10) indicated that greater personal control was
strongly associated with less intrusion among women who had lower-fat diets (t = -2.43,
p = .018), while among women with higher-fat diets, greater personal control was not
associated with intrusion (p > .10). There were no other significant interactions between
personal control and behaviors or behavior changes at 3 months post-treatment (all ps >
.077).
Prospective analyses examining whether interactions between personal control
and behavior at 3 weeks predicted distress outcomes at 3 months were also performed
using multiple regression models. In these models, covariates (age and length of
treatment) were entered in the first step, the distress outcome at 3 weeks post-treatment
was entered in the second step, personal control and the behavior or behavior change in
question were entered in the third step, and the cross-product of personal control and
behavior was entered in the final step. No significant interactions between personal
control and behavior emerged in prospective analyses (all ps > .60).
Chronicity beliefs-behavior interactions and distress
Multiple regression methods described above were also used to test the hypothesis
that women who perceived their cancer to be a chronic rather than acute condition and
engaged in adaptive behavioral coping would have less distress as compared to women
82
with the same beliefs who did not engage in adaptive behavioral coping strategies
(Hypothesis 3a). In these models, interactions between chronicity beliefs and behavior
were examined to determine whether they predicted distress above and beyond the effects
of covariates and the individual predictors alone.
At 3 weeks post-treatment, there were no significant interactions between
chronicity beliefs and behavior or behavior changes in models predicting distress (all ps >
.064). At 3 months post-treatment, chronicity beliefs interacted with avoidance of
stressful situations in the model predicting intrusion, β = .40, p = .006 (see Table A21).
An analysis of the simple slopes of the plotted interaction (see Figure B11) indicated that
perceiving cancer to be a chronic condition was strongly related to greater intrusion
among women who had increased the frequency with which they avoided stressful
situations (t = 3.89, p < .001), while there was only a marginal association between
chronicity beliefs and greater intrusion among women who had not made this change (t =
1.87, p = .067). Chronicity beliefs also interacted with physical activity in the model
predicting anxiety, β = .23, p = .044 (see Table A22). The same interaction at 3 weeks
also prospectively predicted anxiety at 3 months, after adjusting for the effects of
covariates, anxiety at 3 weeks, and the individual predictors, β = .23, p = .041 (see Table
A23). Analysis of the plots of both interactions (see Figure B12 for the prospective
interaction model) indicated that perceiving cancer to be a chronic condition was
associated with greater anxiety among women who were exercising more (concurrent
model: t = 3.25, p = .002; prospective model: t = 2.32, p = .030), while there was no
relationship between chronicity beliefs and anxiety among women who were exercising
less (both ps > .10). There were no other significant interactions between chronicity
beliefs and behavior at 3 weeks in prospective models predicting distress at 3 months.
83
Causal attributions-behavior interactions and distress
Multiple regression methods described above were also used to examine whether
interactions between causal attributions and behavior concurrently and prospectively
predicted distress outcomes. The behavioral and psychological attribution factor, a
theoretically meaningful as well as robust and reliable factor as described previously, was
used in these analyses. As such, the initial hypothesis that a match between women’s
common-sense ideas regarding cancer causes and their behavior would predict less
distress while a mismatch would predict greater distress (Hypothesis 4b) was amended to
predict that women who attributed their cancer to controllable behavioral and
psychological causes and engaged in proactive behaviors would have less distress than
women with the sample beliefs who did not do so.
At 3 weeks post-treatment, attributing cancer to controllable behavioral or
psychological causes interacted with making a change in avoiding stressful situations in
the models predicting all three distress outcomes: depression, β = .29, p = .011; intrusion,
β = .25, p = .037; and anxiety, β = .28, p = .021 (see Table A24 for depression outcome).
Analysis of the simple slopes of the plotted interactions (see Figure B13 for depression
outcome) indicated that this attribution was related to greater distress among women who
were more frequently avoiding stressful situations (depression: t = 4.09, p < .001;
intrusion: t = 2.27, p = .027; anxiety: t = 2.93, p = .005) but not among women who had
not made this change (all ps > 10). Attributing cancer to behavioral and psychological
causes also interacted with frequency of breast self-exam in the models predicting
depression, β = -.24, p = .037; intrusion, β = -.25, p = .033; and anxiety, β = -.28, p =
.021 (see Table A25 for anxiety outcome). Analysis of the plots (see Figure B14 for
anxiety outcome) indicated that among women who were performing breast self-exams
less frequently, attributing cancer to behavioral and psychological causes was strongly
associated with greater distress (depression: t = 3.24, p = .002; intrusion: t = 2.16, p =
.035; and t = 2.77, p = .007). Among women who were performing exams more often, the
84
relationship differed depending on outcome: attributing cancer to behavioral and
psychosocial causes was marginally associated with less intrusion (t = -1.87, p = .066)
but was unrelated to anxiety and depression (both ps > .09).
At three months post-treatment, attributing cancer to behavioral and psychological
causes interacted with making a change in the frequency of avoiding stressful situations
in the model predicting intrusion, β = -.50, p = .002 (see Table A26). An analysis of the
simple slopes of the plotted interaction (See Figure B15) indicated that among women
who had increased the frequency with which they avoided stressful situations, attributing
cancer to behavioral and psychological causes was strongly associated with less intrusion
(t = -2.72, p = .008), while this attribution was marginally associated with more intrusion
among women who had not made this change (t = 1.82, p = .074). Overall, women who
more frequently avoided stress but did not attribute cancer to behavioral and
psychological causes were the most distressed. The attribution also interacted with
alcohol use in the models predicting depression, β = -.26, p = .039 (see Table A27), and
anxiety, β = -.24, p = .055 (although this second interaction did not quite reach
significance). An analysis of the plot for depression (see Figure B16) indicated that this
attribution was associated with greater distress among women who drank little or no
alcohol (t = 2.89, p = .005), but among women who drank larger quantities of alcohol,
this attribution was not related to distress (p > .10).
In prospective analyses, attributing cancer to behavioral and psychological causes
interacted with making a change in fruit/vegetable consumption at 3 weeks posttreatment in the model predicting intrusion at 3 months post-treatment, β = -.20, p = .045
(see Table A28). This attribution also interacted with making a change in avoidance of
stress at 3 weeks post-treatment in the model predicting intrusion at 3 months posttreatment, β = -.30, p = .007 (see Table A29). Analysis of the simple slopes of the plotted
interactions (see Figure B17 which illustrates fruit/vegetable consumption) indicated that
attributing cancer to behavioral and psychological causes predicted less intrusion among
85
women who had increased their fruit and vegetable consumption or had increased the
frequency with which they avoided stressful situations (t = -2.44, p = .018 and t = -3.12, p
= .003, respectively), but this attribution was unrelated to intrusion among women who
had not made these change (both ps > .10). Overall, women who had increased their fruit
and vegetable consumption or the frequency with which they avoided stressful situations
but did not attribute their cancer to behavioral and psychological causes were the most
distressed at 3 months post-treatment. This attribution also interacted with fat
consumption at 3 weeks post-treatment in the model predicting intrusion at 3 months
post-treatment, β = .22, p = .019 (see Table A30). An analysis of the plot (see Figure
B18) indicated that among women eating a lower-fat diet, attributing cancer to behavioral
and psychological causes predicted less intrusion (t = -2.37, p = .021), while there was no
relationship between this attribution and intrusion among women eating a higher-fat diet
(p > .10). Finally, attributing cancer to behavioral and psychological causes at 3 weeks
post-treatment interacted with physical activity at 3 weeks post-treatment in the model
predicting anxiety at 3 months post-treatment, β = -.23, p = .046 (see Table A31). An
analysis of the plot (see Figure B19) indicated that this attribution was marginally
associated with less anxiety among women who were exercising more (t = 1.93, p = .059)
but was unrelated to anxiety among women who were exercising less (p > .10).
Recurrence prevention beliefs-behavior interactions and
distress
Multiple regression models described previously were also used to test whether
interactions between participants’ beliefs about recurrence prevention and their behavior
or behavior changes concurrently and prospectively predicted distress outcomes. As was
the case with the analysis of causal attributions, the factor measuring the extent to which
participants believed that health behavior and psychological factors could prevent a
recurrence was used. The behavioral and psychological attribution factor, a theoretically
86
meaningful as well as robust and reliable factor as described previously, was used in
these analyses. Again, the initial hypothesis that a match between women’s commonsense ideas regarding factors that may prevent recurrence and their behavior would
predict less distress while a mismatch would predict greater distress (Hypothesis 4b) was
amended to predict that women who believed that controllable behavioral and
psychological factors could now prevent recurrence and engaged in proactive behaviors
would have less distress than women with the sample beliefs who did not do so.
At 3 weeks post-treatment, only one significant interaction emerged: the belief
that behavioral and psychological factors could prevent a cancer recurrence interacted
with physical activity in the model predicting anxiety, β = .32, p = .006 (see Table A32).
An analysis of the simple slopes of the plotted interaction (see Figure B20) indicated that
this belief was associated with greater anxiety among women who were exercising more
(t = -2.50, p = .015); among women who were not getting as much exercise, the belief
was not significantly associated with anxiety (p = .10). At 3 months post-treatment, there
were no significant interactions between the belief that behavioral and psychological
factors could prevent a recurrence and behavior changes in multiple regression models
predicting distress (all ps > .057). Multiple regression models were also used to examine
whether interactions between the belief that behavioral or psychological factors could
prevent recurrence and behaviors or behavior changes at 3 weeks post-treatment
prospectively predicted distress outcomes 3 months after treatment had ended. No
significant interactions were found; all p values for interaction terms in the models
exceeded .10.
87
CHAPTER V
DISCUSSION
Post-Treatment Distress and Quality of Life
Breast cancer patients in the present study were remarkably well-adjusted during
the months following treatment as assessed by a variety of distress and health-related
quality of life measures. Our findings expand previous work indicating that longer-term
breast cancer survivors are well-adjusted both with respect to distress and quality of life
(Dorval, Maunsell, Deschenes, Brisson, & Masse, 1998; Ganz et al., 1996; Ganz,
Rowland, Desmond, Meyerowitz, & Wyatt, 1998; Helgeson & Tomich, 2005; Moyer &
Salovey, 1996). Significant concerns following treatment remained, however, in spite of
this otherwise optimal adjustment. Specifically, breast cancer survivors reported mild to
moderate distress related to the possibility of a cancer recurrence, ongoing physical
symptoms associated with cancer and treatment, returning to “normal” or attempting to
create a “new normal,” and the impact of cancer on their families.
Contrary to our hypothesis, the post-treatment period was not, by most measures,
a period of disrupted adjustment. There were no significant changes in any of the indices
of distress assessed from mid-treatment to 3 weeks post-treatment to 3 months posttreatment, including depression, general anxiety, intrusion, avoidance, or recurrence
worry. In fact, mean scores for depression, intrusion, and avoidance decreased over time.
General anxiety and recurrence worry increased slightly from mid-treatment to 3 weeks
post-treatment, before decreasing again from 3 weeks post-treatment to 3 months posttreatment; although consistent with our hypothesis, these changes were not statistically
significant. The expected post-treatment increase in distress, described frequently by
anecdotal and clinical reports, was not evident on self-report measures of distress.
Moreover, there was no effect of type of treatment on distress or changes in distress over
time.
88
Consistent with study hypotheses, health-related quality of life generally
improved over time. Cancer and treatment-related symptoms decreased significantly from
mid-treatment to 3 weeks post-treatment and then remained steady. Quality of life
improved significantly from mid-treatment to post-treatment on 4 of 6 scales, including
improvements in physical functioning and role impairment related to physical
functioning, vitality, and social functioning. There were no significant changes in
emotional functioning or bodily pain, but both scales were already at or better than
population norms at the mid-treatment time point. Type of treatment did not impact
quality of life or changes in quality of life over time.
Not only was there little change in distress over time, but distress levels were
uniformly low across all study time points. Mean depression scores were well below the
cut-off score indicative of clinically significant depression, with only 16% to 17% of
participants exceeding the cut-off at any time point. This was particularly notable given
the fact that 27% of the sample reported having been diagnosed with depression in the
past. Participants also reported experiencing anxiety symptoms and intrusive thoughts
only rarely or occasionally. Overall, findings indicate that not only are breast cancer
survivors well-adjusted as has been previously documented, but breast cancer patients
who have completed some or most of their treatment and those who have recently
completed treatment also exhibit little distress.
These results are perplexing given the multitude of personal anecdotes
(McKinley, 2000; Mullan, 1985), clinical observations (Rowland & Holland, 1990;
Schnipper, 2001), and information collected via structured interview (Beisecker et al.,
1997; Ward, Viergutz, Tormey, deMuth, & Paulen, 1992) suggesting that the posttreatment period is marked by distress, anxiety, transition, and disruption. However,
following the time data collection ended in the present study, a new empirical study
examining anxiety, depression, and quality of life during the 6 months following
radiation treatment for breast cancer was published (Deshields et al., 2005). Results
89
indicated that breast cancer patients experienced low levels of anxiety at the end of
treatment through a 6-month follow-up and minimally elevated depressive
symptomatology at the end of treatment that decreased to normal levels two weeks later
and remained low through 6 months post-treatment. Although we did not find elevations
in depression at any time point, these findings are largely consistent with findings in the
present study demonstrating minimal distress during the weeks and months following the
end of treatment.
It may be that clinical or anecdotal evidence leads to a bias toward overestimating
the prevalence of post-treatment distress because generalizations are made based on a few
remarkable cases. Clinicians and researchers may be struck by and thus more apt to
remember particularly distressed patients. It is also possible that participants exhibited a
response shift on distress measures over the course of the study. It has been proposed that
response shifts, caused by changes in internal standards, values, or conceptualization of
distress over time, impact longitudinal assessment of mental health and quality of life
among individuals with health problems (Schwartz & Sprangers, 1999; Schwartz,
Sprangers, Carey, & Reed, 2004; Sprangers & Schwartz, 1999). It may be that
participants in the present study recalibrated their standards of distress as a result of their
cancer experience such that the same experience of distress was associated with lower
scores on distress measures following treatment. Finally, it is notable that 28% of
participants were prescribed a psychotropic medication at the time of the study, the
majority of which were antidepressants. It is possible that ongoing pharmacological
treatment may be responsible for the remarkably low levels of general distress found in
the present study.
Another explanation for the discrepancy between our findings and previous nonempirical work is the possibility that measures of depression, intrusion, and general
anxiety do not capture the type of distress patients are reporting and clinicians are noting.
There is some evidence in our results to support this conjecture. Despite the lack of
90
general distress reported, some specific sources of stress or worry were more prevalent.
In particular, worry about a cancer recurrence was common, with women reporting
moderate levels of recurrence worry, on average, in the present sample. Fear of a
recurrence was one of the top-rated sources of distress at all time points, with women
reporting this to be a moderate source of stress. Our results are consistent with
Vickburg’s work in developing the Concerns About Recurrence Scale used in the present
study; she also found moderate fear of recurrence in her breast cancer sample (2000).
Moreover, a recent study found fear of recurrence to be the predominant concern of
breast cancer patients who had recently completed treatment (Stanton, Ganz, Rowland et
al., 2005).
The initial review of the literature suggested that worry about a cancer recurrence,
in combination with the loss of the “safety net” of regularly attending cancer treatments,
may be one of the most significant sources of distress for patients. Although fear of
recurrence was prevalent in the current sample, the loss of a safety net was not a
significant source of distress for most women. On average, women reported the distress
associated with losing a safety net as between “none at all” and “a little bit.” Nonetheless,
it appears there is a subset of women for whom this is a more significant source of
distress: slightly over one-third of participants reported feeling that they had “somewhat”
to “very much” lost a safety net, a proportion consistent with findings by Lethborg and
colleagues in their study of breast cancer patients (2000). Future work might examine
psychosocial and demographic characteristics of this subset of patients as compared to
women who are less concerned with the loss of the safety net of treatment.
During treatment, physical problems related to cancer and treatment were the top
source of distress, and somewhat surprisingly, this was also the top source of distress 3
months after treatment had ended. This finding was unexpected in light of the generally
very good quality of life women reported across all study time points. Although
participants’ scores on the SF-36 during treatment indicated that they were experiencing
91
slightly lower physical functioning and slightly more impairment in daily activities as
compared to the larger population of U.S. women, participants’ scores increased to levels
commensurate with population norms at both follow-up time points, and participants in
the present study actually reported less pain and pain-related disability as well as greater
vitality (energy, vigor) than the larger population of U.S. women (SF-36.org, n.d.-a). As
was hypothesized to account for distress findings, this discrepancy may also be due to a
response shift. In the case of quality of life measures, women may have recalibrated their
standards or conceptualization of quality of life on the basis of their experience with
cancer and treatment such that scores on these measures become less of a reflection of
objective physical or functional limitations following cancer treatment. Such a shift in
self-reported quality of life among individuals with health problems, resulting in quality
of life scores that are similar to those of population norms even among very disabled
individuals, has been well-documented in the literature (Schwartz & Sprangers, 1999;
Schwartz, Sprangers, Carey, & Reed, 2004; Sprangers & Schwartz, 1999).
An alternative explanation is that citing physical problems as a predominant
source of distress following treatment does not actually indicate high levels of physical
disability, but may reflect survivors’ overly optimistic expectations regarding recovery.
Previous work suggests that women often do not anticipate ongoing treatment-related
problems but instead expect to return to “normal” shortly after treatment ends (Beisecker
et al., 1997; Stanton, Ganz, Rowland et al., 2005), and this was a sentiment expressed by
women interviewed when collecting pilot data for the present study. It may be that even
mild impairment or changes from pre-cancer baseline that remain following treatment
cause distress because they serve as a reminder of patients’ cancer and the fact that they
are no longer “normal.”
This conjecture is consistent with the finding that “trying to get back to normal
life” was another significant source of distress following treatment for women in the
present study, as was “creating a ‘new normal’”. Women rated these issues to be almost
92
as distressing as a fear of recurrence, consistent with the idea that returning to “normal”
may not be easy or realistic. These data may suggest that women are not receiving
sufficient education regarding ongoing physical and emotional effects of cancer and
treatment. The popular culture that often exalts cancer survivors as physically and
emotionally tough individuals may contribute to this perception. For example, the
appearance of cancer survivors on reality television performing remarkable physical
feats, Lance Armstrong’s LIVESTRONG campaign and his extraordinary athletic
achievements, and the prevalence of cancer survivors participating in community athletic
events such as the Komen Race for the Cure®, while certainly encouraging and inspiring
to many cancer patients, may also lead to frustration or disappointment among survivors
who are unable to perform to such high standards.
The initial review of the literature also suggested that loss of support from family,
friends, and health care providers may be a significant source of distress following
treatment, a conjecture that was also made in Stanton’s recent article discussing posttreatment adjustment (2005). This was not the case in the present sample. Loss of
instrumental and emotional support from family and friends and loss of support from
physicians and other health care staff were rated lowest on a list of potential stressors,
with women rating loss of support to be “not at all” a source of stress, on average.
Similarly, relationship problems were not a significant source of distress for women
following treatment.
Although not identified as a prominent problem in the literature review, a
significant source of stress for the present sample was worry about the impact of cancer
on one’s family. Although this worry conferred less distress on average than concerns
about physical problems, recurrence, and finding a “new normal,” it still fell near the top
of the list. This concern did not appear to reflect difficulty in fulfilling responsibilities at
home or elsewhere, as this was not identified to be a significant source of post-treatment
distress in the present sample. It is possible that this concern reflected marital or family
93
discord, as previous work suggests that marital satisfaction and family functioning
declines among women with breast cancer and their spouses (Northouse, Templin, Mood,
& Oberst, 1998). In contrast to that conjecture, however, relationship problems were of
little concern to women in the present sample. It may be that women are concerned about
the emotional well-being of their loved ones in the face of their cancer. Research suggests
that women should not be so concerned, particularly with respect to their partners:
husbands of cancer patients show less distress than do wives of cancer patients and their
distress and quality of life levels approximate those of healthy controls (Hagedoorn,
Buunk, Kuijer, Wobbes, & Sanderman, 2000; Northouse, Mood, Templin, Mellon, &
George, 2000). It may be that women continue to maintain the gender role of caregiver
even during a serious illness, making concerns about family a more significant source of
distress than other concerns about personal aspects of the cancer experience.
In summary, there was little empirical evidence to support the common idea that
adjustment is significantly disrupted following the end of adjuvant treatment for breast
cancer. Women reported remarkably high levels of quality of life across a variety of
health-related quality of life domains and measures of breast cancer symptomatology, as
well as low levels of distress on measures of depression, general anxiety, and stressrelated symptomatology. There was no significant increase in distress 3 weeks after
treatment had ended as had been anticipated based on the literature review. Our results
suggest that anecdotal and clinical reports of disrupted adjustment following the end of
adjuvant treatment may be confined to a relatively small proportion of breast cancer
survivors.
However, our results suggest that breast cancer survivors do have significant
concerns following treatment, and although they do not lead to significant depression or
anxiety symptoms, these concerns are associated with mild to moderate distress among
patients. One somewhat unexpected finding in the present study was that, in the context
of overall very good quality of life among patients in the present sample, physical
94
symptoms related to cancer and treatment was the highest-rated source of stress not only
during treatment but also 3 months after treatment had ended. Women were also
moderately distressed by the possibility of a cancer recurrence. Other significant sources
of worry included trying to return to normal or to create a “new normal” and the impact
of cancer on one’s family. Although not a significant source of stress for most women,
just over one-third of our sample reported feeling a loss of the “safety net” of treatment
and regular visits with their health care providers. Women’s lives are certainly changed
by their cancer experience. Difficulty obtaining one’s previous or new “normal,” both
physically and otherwise, in combination with the realization that cancer may return to
disrupt one’s life again, best characterizes the type of distress reported by women in the
present study.
Creating a “New Normal”: Post-Treatment Behavior
Changes
As previously discussed, creating a “new normal” was a significant source of
stress for many women in the present sample. Creating a new normal may involve
making changes in lifestyle, health behaviors, work, relationships, and priorities. Data
from the present study indicate that breast cancer patients are making substantial efforts
to make life changes following the end of adjuvant treatment.
Changes in health behaviors were some of the most common changes reported by
the present sample. In particular, women frequently reported increasing the frequency of
breast self-examination, eating more fruits and vegetables, taking dietary supplements,
increasing moderate physical activity, and cutting back on alcohol and tobacco use, with
at least 29% participants reporting these positive health behavior changes at one or both
post-treatment follow-up time points. The prevalence of health behavior changes may
reflect participants’ efforts to address concerns about recurrence by making changes
believed to decrease the possibility of recurrence. In fact, most of these changes,
95
including changes in diet, physical activity, and alcohol use, match health guidelines for
cancer survivors as well as for breast cancer prevention provided by the American Cancer
Society, the National Cancer Institute, and the American Institute for Cancer Research
(ACS Nutrition and Physical Activity Guidelines Advisory Committee, 2001; American
Institute for Cancer Research, 1997, 2002; National Cancer Institute, 2002). Research
indicates that food safety practices, weight loss (if overweight), decreased fat
consumption, increased consumption of fruit and vegetables, increased physical activity
and decreased alcohol consumption have a probable or possible benefit in preventing
breast cancer recurrence (Brown et al., 2001). In the current sample, smaller proportions
of women reported increasing food safety practices (12%) and choosing no- or low-fat
foods more often (19%) as compared to other changes.
The prevalence of positive health behavior changes was also reflected in the
generally good health practices of the present sample. Responses to the diet screener
suggested that on average, women adhered to dietary recommendations. Participants
consumed an average of 4.5 servings of fruits and vegetables each day (5 servings are
recommended) and consumed just under one-third of daily calories from fat. Only 7% to
8% of the sample smoked cigarettes, and participants reported consuming little or no
alcohol and performing monthly breast self-exams, on average. Moreover, patients
complied with treatment: less than 10% of the sample had ever missed or even cancelled
a cancer-related appointment. A significant number of women also sought out
nontraditional health services, with 15 to 17% of the sample reporting use of
complementary therapies (e.g., yoga, massage, herbals) following treatment.
Results suggest that the post-treatment period may be an opportune time to
promote, encourage, and educate women regarding positive health practices. Study
participants reported mild distress related to uncertainty regarding what they should do
for their health or to prevent a cancer recurrence. Therefore, assisting women in
identifying appropriate health behavior changes to focus on and providing education or
96
intervention may assist not only in promoting good health but also in alleviating distress.
McKinley, a physician who wrote about her experience with breast cancer, recommended
that oncologists provide more guidance and preparation to cancer survivors who are
completing treatment (2000). Consistent with the idea that breast cancer survivors may be
particularly receptive to this type of guidance, a recent study found that 79% of breast
cancer survivors expressed interest in a health promotion program (Demark-Wahnefried,
Peterson, McBride, Lipkus, & Clipp, 2000), and an intervention trial indicated that
women who simply received a recommendation from their oncologist were exercising
more 5 weeks later as compared to women who did not receive a recommendation (Jones,
Courneya, Fairey, & Mackey, 2004).
Women’s “new normals” also included other lifestyle changes, many of which
appeared to be directed toward enhancing emotional and spiritual health. In particular,
about one-fourth of the sample reported more frequently avoiding stressful situations or
praying or meditating more. Smaller proportions of women reported engaging in hobbies
or relaxing activities more, journaling or writing about their experiences, and working
less both inside and outside of the home. Nonetheless, only a handful of participants
reported trying more traditional treatments aimed at addressing emotional concerns, such
as attending psychotherapy or support groups. Spending time with family and friends also
became a higher priority for some women, with about one-fifth of the sample increasing
the amount of time they spend with loved ones. Previous work has documented
prevalence of changes in diet, physical activity, and CAM use (Burstein, Gelber,
Guadagnoli, & Weeks, 1999; Maskarinec, Gotay, Tatsumura, Shumay, & Kakai, 2001;
Maskarinec, Murphy, Shumay, & Kakai, 2001; Taylor, Lichtman, & Wood, 1984), but
this is the first study that we are aware of that has examined prevalence of a
comprehensive array of post-treatment behavior changes.
97
Post-Treatment Behavior Changes as Coping Strategies
Clearly engaging in positive health practices is important for survivors’ health,
but the initial literature review also conceptualized engaging in positive health behavior
practices or making changes in behavior designed to improve emotional, physical or
spiritual health as potentially adaptive coping strategies. Thought of this way, the
prevalence of behavior changes might reflect women’s efforts to reduce recurrence
worry, develop a “new normal,” or rebuild one’s “safety net” when attending treatment is
no longer an option for managing one’s cancer. In any case, these behavior changes were
hypothesized to be associated with attenuated distress both concurrently and
prospectively. Consistent with this hypothesis, we found that positive health practices
were generally associated with less distress and better quality of life. However, our
analysis of behavior changes revealed a perplexing trend: women who had made positive
health behavior changes often reported greater distress than women who had not made
such changes.
This pattern of findings was particularly striking for relationships between diet
and distress. Women eating healthy diets, as assessed by dietary screeners, reported less
distress and better quality of life, as anticipated. Specifically, more fruit and vegetable
consumption was associated with marginally less recurrence worry 3 weeks posttreatment, and less fat consumption was associated with less anxiety 3 months posttreatment. Paradoxically, women who reported on the behavior change questionnaire that
they had increased fruit and vegetable consumption had particularly poor adjustment. As
compared to women who had not made positive dietary changes, those who reported
increasing fruit and vegetable intake at either post-treatment time point exhibited greater
distress across a variety of measures at 3 months post-treatment, including more
depression, anxiety, and intrusion. Other health behaviors showed the same pattern. For
example, alcohol use was associated with greater intrusion both 3 weeks and 3 months
post-treatment, but women who cut back on alcohol use reported greater recurrence
98
worry 3 months post-treatment. Similarly, more frequent breast self-examination was
associated with less recurrence worry 3 months post-treatment, but women who increased
the frequency with which they performed breast self-examination had greater anxiety and
marginally more recurrence worry in both concurrent and prospective analyses.
These results are somewhat perplexing; it would seem that if practicing healthy
behaviors is associated with less distress and better quality of life, making positive
changes in the same health practices would also have beneficial outcomes. Moreover, we
had hypothesized that making such behavior changes may be an adaptive coping strategy
in that it provides women with an instrumental strategy for promoting health and
preventing recurrence. It could be, however, that distress is needed to invoke behavior
change. Specifically, it may be that distress in fact instigates behavior changes rather than
behavior changes leading to decreased distress, as we had predicted. Our findings that
these relationships held when examined prospectively, however, stands in contrast to this
conjecture. In all prospective analyses, we examined whether the behavior change in
question at 3 weeks post-treatment predicted distress outcomes 2 to 3 months later, after
adjusting for distress 3 weeks post-treatment. As previously noted, even these analyses
indicated that behavior changes following treatment were associated with increases in
distress during the subsequent 2 to 3 months. Moreover, the literature suggests that
anxiety or distress is usually detrimental with respect to health behavior changes; distress
has been associated with poorer health behaviors and can impair comprehension and
recall of messages promoting health behaviors (M. Miller, 2005; Strine, Chapman,
Kobau, & Balluz, 2005). Another possibility may be that the process of making
significant lifestyle changes causes disrupted adjustment.
Given that healthful behaviors in general were associated with attenuated distress
and better quality of life, it may be that making healthful or adaptive behavior changes is
associated with distress in the short-term but may lead to better outcomes with respect to
distress and quality of life over a longer time period. For example, women who recently
99
decided to increase fruit and vegetable consumption may have done so because they were
experiencing distress related to their cancer, and attempting to make such a behavior
change may predict disrupted adjustment in the months that follow. However, if followed
one or more years down the road, women who had made and maintained the behavior
change may show superior adjustment. Consistent with this idea, breast cancer patients
who reported making dietary changes during the year following diagnosis were more
distressed at the time of initial treatment but also experienced a greater decline in distress
over the next 12 months as compared to women who had not made dietary changes
(Maunsell, Drolet, Brisson, Robert, & Deschenes, 2002). It will therefore be important
for future work to examine the long-term implications of behavior changes made during
the post-treatment period.
Moreover, although many women reported making behavior changes, it is unclear
to what extent these changes are maintained following the immediate post-treatment
period. Maintenance of health behavior may be motivated by different psychosocial
factors than those that lead to initial behavior changes (e.g., Rothman, 2000). Future work
might examine to what extent changes in diet, exercise, substance use, and breast selfexamination are maintained, what predicts maintenance among cancer survivors, and
whether women are more likely to maintain changes made following treatment as
compared to changes made at other times in their lives.
In summary, our results suggest that good health practices (healthy diet, regular
physical activity, minimal alcohol use, and regular breast self-examination) are associated
with positive psychological adjustment. However, making changes in health practices and
stress avoidance predicted greater distress in both concurrent and prospective analyses, as
opposed to less recurrence worry and better adjustment, as predicted. These results
suggest that post-treatment behavior changes may not be an effective post-treatment
coping strategy as proposed, at least in the short-term. However, the long-term
100
implications of making behavior changes are unknown, and it is possible that such
changes may be associated with better adjustment if examined over a longer time period.
Common-Sense Models of Breast Cancer
Although health behavior changes were not the effective post-treatment coping
strategies they had been predicted to be, we remained interested in what might predict
whether survivors decide to make behavior changes and what changes they decide to
make. We therefore attempted to characterize breast cancer patients’ beliefs about their
cancer using Leventhal’s domains of illness representations (Leventhal et al., 1980;
Leventhal & Nerenz, 1985; Leventhal, Nerenz, & Steele, 1984, Meyer, Leventhal, &
Gutmann, 1985) because these dimensions are thought to have important implications
with respect to coping responses. The domains include participants’ beliefs about disease
identity, cause, consequences, course or timeline, and control/cure (how one might effect
a cure).
With respect to views about cancer timeline, women perceived their cancer to be
more of an acute than chronic condition, although no strongly so, and they did not
perceive their cancer to be cyclical in nature (although, as noted in the method section, it
is likely that the measure of “cyclical timeline” was confusing to participants). These
results are consistent with previous work that found a small majority of breast cancer
patients (54-55%) perceived their illness to be acute in nature (Rabin, Leventhal, &
Goodin, 2004). Women perceived their cancer to have moderate consequences in their
lives. They believed they had moderate control over their cancer and saw their treatments
as quite effective in controlling and curing their cancer. Moreover, women had fairly
strong ideas about what may have caused their cancer and what they could now do to
prevent a recurrence. We found these cancer representations to be consistent across time
from mid-treatment through 3 months post-treatment. Although previous work has
suggested that breast cancer patients’ statements about fatigue can be classified along
101
these dimensions (Barsevick, Whitmer, & Walker, 2002), this is the first study to our
knowledge to measure and describe breast cancer patients’ beliefs in each domain.
We examined the “cure/control” and “cause” domains in greater depth as these
beliefs were expected to be more complex and multifacted. With respect to perceived
control, women reported moderate levels of control over cancer overall. They perceived
almost complete control over getting needed information about cancer and very high
levels of control over decisions regarding treatment and medical care. Women believed
they had moderate control over the success of their treatment, and they perceived a
modest level of control over the long-term course of their cancer and over treatment side
effects. They believed they had little control over the cause of their cancer. Although
perceived control over cancer is often viewed as a unitary construct, the present data
suggest that women’s control beliefs are more nuanced and complex. Nonetheless, a scale
comprising all aspects of control demonstrated adequate internal consistency, indicating
that women who perceive greater control over one aspect of their cancer are apt to
perceive greater control over other aspects as well.
Although women reported that they perceived little control over the cause of their
cancer, they nonetheless had very strong ideas regarding what may have caused their
cancer, and several of the highly endorsed causes involved behavioral or other
controllable factors. Women rated hormones, environmental toxins or hazards, and
genetics or heredity as the most important causes of their cancer, with more than twothirds of women citing these factors as at least somewhat important in the development of
their cancer. Although these causes are not modifiable, diet and stress or worry were
rated as the fourth and fifth most important causes, with 63% to 68% of women
indicating that a poor diet was at least somewhat important in causing their cancer and
52% to 55% believing that stress or worry was at least somewhat important. Other
potentially modifiable factors were also seen as important in the development of cancer
by significant proportions of women: almost half of women believed that lack of exercise
102
played a role in the development of their cancer and more than one-fourth of women
believed that their mental attitude was at least somewhat important in causing their
cancer.
It was notable that the first-order factor that emerged in exploratory factor
analyses at all time points comprised potentially controllable or modifiable causal
attributions: diet, lack of exercise, stress or worry, and mental attitude. No reliable second
factor emerged in these analyses. This suggests that controllability may be an important
dimension along which causal attributions vary. In the present study, these attributions
involved health behaviors (diet and exercise) and psychological factors (stress or worry,
mental attitude).
Findings in the present study are consistent with those of previous studies that
have found that many breast cancer patients attribute their cancer to stress, heredity,
environmental factors, and diet (Stewart et al., 2001; Taylor, Lichtman, & Wood, 1984).
In general, patients’ beliefs were also at least somewhat consistent with current medical
knowledge; genetics and heredity, along with aging, are strong risk factors for breast
cancer with hormonal factors conferring more modest risk. Alcohol use and obesity
(which is influenced by diet and exercise) are also modest risk factors (American Cancer
Society, 2006a). Despite widespread controversy regarding links between environmental
toxins, such as pesticides, and breast cancer, no clear evidence for this causal link has
been found (Calle, Frumkin, Henley, Savitz, & Thun, 2002).
The belief by more than half of participants that stress played an important role in
the development of their cancer was also not necessarily consistent with current
knowledge; the literature regarding links between stress and the development of cancer is
inconclusive, with some positive (Chen et al., 1995; Levav et al., 2000) and some null
findings (Protheroe et al., 1999; Roberts, Newcomb, Trentham-Dietz, & Storer, 1996).
Recent reviews and meta-analyses have concluded that there is no convincing overall link
between stressful life events and cancer risk (Butow et al., 2000; Duijts, Zeegers, &
103
Borne, 2003; Petticrew, Fraser, & Regan, 1999). It may be that stressful life events close
in proximity to a cancer diagnosis are more salient to women than are other risk factors.
Nonetheless, in other studies stress had been the most prevalent causal attribution
(Stewart et al., 2001; Taylor, Lichtman, & Wood, 1984), and in our recent work,
gynecologic cancer survivors rated stress as the second most important causal factor
(Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005).
Women in the present sample had even stronger ideas regarding what they could
now do to prevent a cancer recurrence. Medical checkups and screenings were seen as
most important in preventing a recurrence, followed by having a positive attitude, eating
a healthy diet, taking medication, and exercising. More than 85% of women cited these
factors as at least somewhat important in preventing a recurrence. Notably, these factors
can be considered controllable or modifiable, and less controllable factors such as God’s
will and chance or luck were seen as much less important in preventing a recurrence. This
pattern may be psychologically protective: it may be important for women to believe that
they can do something instrumental to prevent a recurrence, rather than leaving the
possibility of recurrence in the hands of uncontrollable factors.
Remarkably, a parallel set of controllable beliefs to those comprising the
controllable causal attribution factor emerged in an exploratory factor analysis as a
primary, reliable factor at both post-treatment time points. The beliefs that eating a
healthy diet, exercising, reducing stress in one’s life, and having a positive attitude could
prevent recurrence comprised this factor. Again, beliefs involved both potentially
modifiable health behaviors (diet, exercise) and psychological factors (stress, attitude).
This finding highlights the robustness of this factor across both causal attributions and
recurrence prevention beliefs.
Consistent with our previous study of gynecologic cancer survivors’ recurrence
prevention beliefs (Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005), 92 to 96%
of participants believed that a positive attitude was somewhat to very important in
104
preventing a cancer recurrence. The literature is far from conclusive regarding the effects
of a positive attitude on cancer recurrence, however. There is some evidence that
pessimism or hopelessness is associated with recurrence and poorer survival (Molassiotis,
Van den Akker, Milligan, & Goldman, 1997; Schulz, Bookwala, Knapp, Scheier, &
Williamson, 1996; Watson, Haviland, Greer, Davidson, & Bliss, 1999) while having a
“fighting spirit” is associated with reduced risk of recurrence and better survival (Greer,
Morris, Pettingale, & Haybittle, 1990; Tschuschkle et al., 2002), but there are just as
many studies that have not found these links (M.A. Andrykowski, Brady, & HensleeDowney, 1994; Giraldi, Rodani, Cartei, & Grassi, 1997; Murphy, Jenkins, & Whittaker,
1996; Ringdal, 1995). Moreover, some research has shown that putting on a happy face
could actually be detrimental; suppression of negative emotions has been associated with
poorer prognoses among cancer patients, while expression of emotions has been
associated with better outcomes (Jensen, 1987; Reynolds et al., 2000; Temoshok et al.,
1985; Weihs, Enright, Simmens, & Reiss, 2000). Popular media promoting the
significance of mental attitude in overcoming disease or encouragement from family and
friends to maintain a “positive attitude” may explain participants’ beliefs.
Common-Sense Beliefs and Behavior Changes
Personal control and behavior changes
An important postulate of Leventhal’s self-regulation theory is that cognitive
models of illness guide coping responses (Leventhal et al., 1997; Leventhal, Meyer, &
Nerenz, 1980), and there is evidence that both cancer attributions and control beliefs are
associated with coping behavior among cancer patients (Buick, 1997; Christensen et al.,
1999; Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005; Lavery & Clarke, 1996).
Consistent with these findings and our hypothesis, results indicated that women who
believed they had more personal control over their cancer just after treatment had ended
were marginally more apt to make behavior changes aimed at reducing stress or to
105
practice healthful behaviors. Specifically, women who perceived more control over their
cancer were marginally more likely to increase avoidance of stressful situations, and
there were trends toward links between perceived control and healthful eating: women
who perceived greater personal control over cancer 3 weeks post-treatment ate more
fruits and vegetables and those who perceived greater control 3 months after treatment
had ended ate less fat. Avoiding stressful situations and eating healthfully might be
considered active coping strategies for managing distress related to cancer. The belief
that one can influence one’s cancer may encourage women to cope proactively with their
cancer.
Results should be interpreted cautiously, however. First of all, all of the
relationships found were concurrent associations. It appears that control beliefs are more
apt to influence current behavior rather than affecting changes in behavior over time.
Moreover, we cannot rule out that proactive behaviors may influence control beliefs, or
that a third factor, such a socially desirable self-presentation, influenced reporting of both
perceived control and proactive coping behavior. Second, most relationships found were
trends that approached, but did not reach, statistical significance. In sum, while results
suggest that perceiving control over one’s cancer may encourage the utilization of
proactive coping strategies, relationships were weaker and less robust than anticipated.
Chronicity beliefs and behavior changes
Although there is evidence that beliefs about disease timeline are associated with
distress among cancer patients (Rabin, Leventhal, & Goodin, 2004), to our knowledge,
no previous work has investigated relationships between illness chronicity beliefs and
behavioral outcomes among cancer patients. It was hypothesized that women who
perceived cancer to be a chronic condition would be more apt to engage in proactive
behavior and to make positive behavior changes. Instead, the majority of findings were in
106
the opposite direction: perceiving cancer to be chronic in nature generally predicted
maladaptive behavioral outcomes. Several findings were more complex, however.
For example, women who perceived their cancer to be a chronic condition during
treatment or 3 weeks post-treatment were less likely to increase the frequency with which
they performed breast self-examinations than were women who perceived their cancer as
an acute illness. However, this belief, when measured 3 weeks post-treatment, predicted a
trend toward more frequent breast self-examination 3 months after treatment had ended.
It may be that the behavioral correlates of chronicity beliefs change over time. Perceiving
cancer to be an acute condition might be most effective in promoting behavior changes
during treatment or immediately after treatment ends, whereas perceiving cancer to be a
chronic condition may be more adaptive with respect to cancer screening behavior in the
long run. Those who see cancer as a chronic condition may feel less compelled to
examine their breasts more frequently initially but may be more apt to regularly perform
self-exams as time goes on. Another possibility is that women who are already examining
their breasts frequently are less likely to increase the frequency of breast selfexamination, explaining why the same belief may be associated with two seemingly
contrasting outcomes.
A similarly paradoxical relationship was found between chronicity beliefs and
alcohol use. At both post-treatment time points, perceiving cancer to be a chronic
condition was associated with more alcohol use. However, those who perceived cancer to
be chronic in nature during treatment were more likely to have decreased their alcohol
use 3 weeks after treatment had ended. Similar to speculations regarding breast selfexamination, it may be that women who consumed more alcohol were also more likely to
try to cut back on alcohol use. This would account for what at face value appear to be
contradictory findings. In sum, it appears that women who perceive cancer to be more
chronic in nature are those who are consuming more alcohol but who are also curbing
their alcohol use.
107
Another notable finding was that women who perceived cancer to be a chronic
condition 3 weeks after treatment had ended consumed more fattening foods 3 months
after treatment had ended. In sum, while perceiving cancer to be a chronic condition
(endorsing beliefs that one’s cancer is “permanent” or “will last for a long time”) was
generally associated with maladaptive behavioral outcomes, chronicity beliefs showed
contrasting effects on cancer screening versus other recommended cancer prevention
behaviors, including eating a healthy diet and limiting alcohol use. Viewing one’s cancer
as an acute condition (endorsing beliefs that one’s cancer is “temporary” or “will pass
quickly”) appears to be more adaptive with respect to alcohol use and diet, with those
who perceived cancer to be acute drinking less alcohol (although also less apt to cut back
on alcohol use) and eating less fattening foods. In contrast, women who perceived cancer
to be a chronic condition were regularly engaging in breast self-examination 3 months
following treatment but were less likely to increase exam frequency earlier on. It may be
that women who see their cancer as chronic in nature do not perceive that dietary and
alcohol use changes would be efficacious in controlling their cancer. These women may
believe their cancer is likely to recur, and therefore they may be more apt to perform
breast self-exams in an effort to monitor or detect changes in cancer symptoms.
Cancer attributions and behavior changes
Causal attributions were a much stronger and robust predictor of behavioral
outcomes than were control and chronicity beliefs. Women who attributed cancer to
controllable causes (health behaviors and psychological factors) both during and
following treatment were more apt to engage in healthful behaviors and to make adaptive
behavioral changes, and relationships were strongest and most consistent for alcohol use
and diet. In particular, these attributions predicted less alcohol use both concurrently and
prospectively, as well as increased likelihood of cutting back alcohol use by 3 months
post-treatment. Women who perceived cancer causes to be controllable at baseline were
108
more likely to have increased their fruit and vegetable consumption at both posttreatment time points, and those making controllable attributions at either post-treatment
time point were also more likely to have increased fruit and vegetable consumption 3
months after treatment had ended. Only one finding stood in contrast to the general trend
toward controllable causal attributions predicting more proactive behavior: controllable
attributions were associated with less concurrent physical activity 3 weeks after treatment
had ended. However, this relationship did not hold in prospective analyses or at 3 months
post-treatment.
A very similar pattern of findings emerged for women who believed controllable
health practices and psychological factors could prevent a recurrence, with these beliefs
predicting adaptive changes in both diet and alcohol use. Women with this set of beliefs
were more likely to have increased fruit and vegetable consumption at both posttreatment time points, and this relationship held in both concurrent and prospective
analyses. Similarly, believing that controllable factors could prevent a recurrence was
associated with less fat consumption 3 weeks after treatment had ended. Women who
held this set of beliefs were also more likely to limit alcohol use following treatment.
Although not found to be the case with controllable causal attributions, controllable
recurrence prevention beliefs were also associated with increased avoidance of stressful
situations 3 weeks post-treatment.
These findings are consistent with previous work demonstrating that patients who
attributed cancer to controllable causes were more likely to modify their health behavior
(Buick, 1997; Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005). Results also
suggest that the measures consisting of health practices and psychological factors
identified in our exploratory factor analyses are not only stable and reliable, but are also
meaningful: both controllable causal attributions and recurrence prevention beliefs
predicted important behavioral outcomes.
109
The hypothesis that behavior changes would match specific causal attributions
was also supported, and again, we found strong and consistent relationships for diet and
alcohol use outcomes. For example, women who believed that diet played an important
role in the development of their cancer were more likely to have increased fruit and
vegetable consumption at both post-treatment follow-ups, and this relationship was found
in both concurrent and prospective analyses. Those who believed that a healthy diet could
prevent a recurrence were also more likely to have increased fruit and vegetable
consumption 3 months post-treatment. Similarly, women who believed that substance use
played a significant role in the development of their cancer drank less alcohol 3 weeks
after treatment had ended and those who believed that cutting back on substances could
prevent a recurrence were more likely to have decreased their alcohol use at both posttreatment time points.
Other belief-behavior combinations tested also supported our hypothesis. For
example, women who believed that stress or worry played a role in the development of
their cancer or that reducing stress could prevent a recurrence were more likely increase
the frequency with which they avoided stressful situations and to increase the frequency
with which they engaged in relaxing activities at both post-treatment time points.
Exercise attributions and physical activity outcomes did not fit this pattern
perfectly, however. Women who attributed cancer to lack of exercise at either posttreatment time points reported doing less concurrent physical activity. However, women
who made this attribution 3 weeks after treatment had ended were more likely to have
increased physical activity 3 months post-treatment. Similarly, women who believed that
exercise could prevent a cancer recurrence were also more likely to have increased
physical activity 3 months post-treatment. As noted previously, few women reported
having increased physical activity just after treatment had ended, whereas a substantial
proportion of our sample reported having done so 3 months post-treatment. Increasing
physical activity may be difficult during the immediate post-treatment period due to the
110
physical effects of cancer and treatment, and therefore this attribution may have less
influence over behavior at that time. Exercise attributions appear predict positive changes
in physical activity later on, however.
Results highlight the instrumental role of women’s causal attributions and
recurrence prevention beliefs in their health practices and the behavior changes they
choose to make following treatment. Both the controllability of the belief as well as the
content appear to be important. With respect to Leventhal’s self-regulation theory, our
results suggest that for breast cancer patients, causal attributions may be one of the most
important dimensions of illness beliefs as they were superior to other domains in
predicting coping outcomes. Control and chronicity beliefs were weaker and less
consistent predictors of behavioral outcomes. Previous results suggested that the posttreatment period is an opportune time to encourage cancer survivors to adopt healthful
lifestyle practices, and women’s common sense beliefs about their cancer provide insight
into their decisions to make these changes.
Interactions Between Common-Sense Beliefs and Behavior
and Distress Outcomes
Results have demonstrated that cancer patients’ common-sense models of cancer,
particularly cancer attributions, predict health practices and behavior changes.
Unexpectedly however, we also found that behavior changes often were associated with
distress. Previous work from our lab suggested that what may be most important with
respect to adjustment outcomes is that behavior matches common-sense representations
of one’s illness (Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005). It was
hypothesized that the same would be true in the present sample, perhaps accounting for
these otherwise perplexing findings. However, overall we found a weak and inconsistent
pattern of interactions between women’s beliefs about cancer and their behavior in
models predicting distress.
111
First, it was hypothesized that women who perceived greater control over their
cancer and engaged in adaptive behavioral coping strategies (health practices, positive
behavior changes) would report less distress as compared to women with the same beliefs
who did not engage in behavioral coping strategies. In general, this hypothesis was not
supported; only one significant interaction between perceived control and behavior was
found in models predicting distress. This pattern of the interaction was consistent with
our hypothesis, however: at 3 months post-treatment, greater personal control was
associated with less intrusion among women who were eating low-fat diets but somewhat
more intrusion among women who were eating high-fat diets. As hypothesized, women
who perceived greater personal control over their cancer and engaged in this adaptive
health behavior had lower levels of intrusion than did women who perceived high levels
of control but did not. Previous work has found that greater perceived control is
associated with less distress and better adjustment among cancer patients (R. A. Jenkins
& Pargament, 1988; Newsom, Knapp, & Schulz, 1996; Thompson, Sobolew, Galbraith,
Schwankovsky, & Cruzen, 1993). Although the current results should be interpreted
cautiously given the lack of significant findings, there is some suggestion that personal
control beliefs may predict better adjustment only among women who are engaging in
active behavioral coping strategies.
It was also hypothesized that women who perceived their breast cancer to be a
chronic condition and engaged in adaptive behavioral coping strategies would report less
distress as compared to women with the same beliefs who did not engage in behavioral
coping strategies. Although there were few significant interactions between chronicity
beliefs and behavior, in the interactions that were found, the opposite was true. For
example, among women who had increased the frequency with which they avoided
stressful situations, as well as those who exercised more frequently, perceiving cancer to
be a chronic condition was associated with greater concurrent anxiety or intrusion.
However, chronicity beliefs were unrelated to outcomes among women who had not
112
made positive behavior changes or were not currently practicing healthful behaviors.
Contrary to our hypothesis, women who perceived cancer to be chronic in nature and had
made positive behavior changes or were currently practicing healthy behaviors were
more distressed than women with the same belief who did not do so; moreover, these
women were also more distressed than women who perceived cancer to be an acute
condition.
Rather than leading to attenuated distress as we had hypothesized, chronic
timeline beliefs and active coping appeared to be the most “lethal” combination with
respect to distress. It is unclear why such a combination would lead to greater distress.
Previous work has found that women who perceive cancer to be chronic rather than acute
are more distressed (Rabin, Leventhal, & Goodin, 2004), and this relationship is intuitive.
It is less evident why women who perceived cancer to be chronic and were engaging in
proactive or healthful behaviors would be more distressed than women with the same
belief who were not. Perhaps women who perceive their cancer to be a chronic condition
do not perceive health behaviors to be effective in controlling their cancer, and
continuing to engage in coping strategies that are perceived as ineffective may lead to
distress. It is also possible that the combination of chronic timeline beliefs and healthful
behavior characterizes women who are most fearful of a cancer recurrence. This may
account for the high levels of distress associated with this seemingly detrimental
combination.
Lastly, we hypothesized that a match between women’s common-sense ideas
regarding cancer causes or factors that may prevent recurrence and their behavior would
predict less distress, while a mismatch would be associated with greater distress. Because
we were able to extract robust, internally consistent, and meaningful factors for causal
attributions and recurrence prevention beliefs, we used the factors in our analyses rather
than individual items. As such, the initial hypothesis was amended to predict that women
who attributed their cancer to controllable causes including health behaviors and
113
psychological factors or women who believed such factors can now prevent recurrence
and engage in healthful and proactive behaviors would have less distress than women
with the same beliefs who do not do so.
Some results were consistent with our hypothesis. For example, women who
attributed cancer to controllable causes and were engaging in frequent breast selfexamination reported less depression, anxiety, and intrusion at 3 weeks post-treatment
than did women with the same belief who were not performing regular exams. In all
cases, attributing cancer to controllable causes was associated with greater distress among
women who were not frequently practicing breast self-exam. Similarly, in prospective
models, women who attributed cancer to controllable causes and were engaging in
frequent physical activity or eating low-fat diets at 3 weeks post-treatment reported less
anxiety 3 months post-treatment than did women with the same belief who were
engaging in little physical activity or eating high-fat diets. In these models, attributing
cancer to controllable causes was associated with less distress among women who were
engaging in positive health practices but was associated with greater distress among
women who were not.
Other significant interactions did not support hypotheses, however, and there were
no consistent patterns among these findings. In one example, attributing caner to
controllable causes was associated with less distress among women who had increased
their fruit and vegetable consumption, as one might predict. However, in contrast to
hypotheses, among women who attributed cancer to controllable causes, there was no
difference in distress between women who had increased fruit and vegetable consumption
and those who had and had not made this change. The interaction instead was driven by
the fact that women who did not attribute cancer to controllable causes but who had made
positive dietary changes were much more distressed than other women. A completely
different pattern was found for the interaction between attributions and avoidance of
stressful situations. Attributing cancer to controllable causes was associated with greater
114
distress at 3 weeks post-treatment but less distress at 3 months post-treatment among
women who more frequently avoided stressful situations. In all cases, there was no
difference in distress levels between women who attributed cancer to controllable causes
who had and had not changed the frequency with which they avoided stress. Interactions
between attributions and alcohol consumption revealed different patterns that were also
not consistent with study hypotheses.
With respect to recurrence prevention beliefs, the belief that controllable health
behavior and psychological factors could prevent a recurrence generally did not interact
with behaviors in models predicting distress. The only significant finding stood in
contrast to our hypothesis: at 3 weeks post-treatment, women with controllable
recurrence prevention beliefs who were currently engaging in more physical activity
reported more anxiety than women with the same set of beliefs who were not engaging in
regular physical activity.
In summary, there were no consistent interaction patterns between causal
attributions or recurrence prevention beliefs and behavior in models predicting distress.
While some of the interactions supported study hypotheses, many did not. Given these
inconsistencies, it is not possible to draw sound inferences from these data. These
findings stand in contrast to our previous work with long-term gynecologic cancer
survivors. In a similar set of analyses, we found consistent interactions between causal
and recurrence prevention beliefs and behavior indicating that relationships between
attributing cancer to controllable factors and greater distress were present only among
individuals who continue to engage in behaviors they believed may have caused their
cancer. In that sample, distress was lower among women who had made positive changes
in health practices (Costanzo, Lutgendorf, Bradley, Rose, & Anderson, 2005). The
discrepancy in findings may be a result of the different study populations. The previous
study examined long-term gynecologic cancer survivors, whereas the present study
examined breast cancer survivors’ adjustment in the immediate post-treatment period.
115
One might expect that disease attributions and related health practices would have a less
consistent influence on adjustment many years after treatment ended. However, perhaps
the frequent changes in behavior and adjustment during the immediate post-treatment
period lead to an inconsistent pattern of results. Once changes solidify, the patterns may
be clearer and more consistent. Whatever the reason, results suggest that considering
causal attributions or recurrence prevention beliefs within the context of health practices
or behavior changes did not consistently or robustly explain post-treatment psychological
adjustment among women with breast cancer.
In summary, interactions between common-sense beliefs about breast cancer and
health practices or behavior changes were weak or inconsistent predictors of posttreatment distress. While one interaction between personal control and behavior was
consistent with study hypotheses, indicating that personal control beliefs are associated
with better adjustment among women who are engaging in active behavioral coping
strategies, no other interactions were significant predictors of distress. More interactions
between chronicity beliefs and behavior were found, but the interaction patterns were not
consistent with hypotheses. Instead, we found that women who perceived cancer to be
chronic in nature and had made positive behavior changes or were currently practicing
healthy behaviors were more distressed than women with the same belief who did not do
so. As already discussed, significant interactions between attributions and behavior in
models predicting distress were plentiful but no consistent patterns were found.
Limitations
Results of the present study should be considered in the context of some
limitations. First, due to attrition, we had limited power at the follow-up time points to
detect small to medium effects for models containing several covariates. As such, it is
possible that patterns in the data were missed. In order to appreciate possible trends in the
116
data, marginally significant findings were always examined and were often reported in
the results section.
Another potential limitation of the present study is the limited measurement time
points. For example, a longer follow-up period may better elucidate to what extent
behavior changes are maintained over time. As previously noted, the fact that participants
were likely in a state of transition and flux with respect to adjustment and behavior
changes may have accounted for the inconsistent patterns we found in analyses of
relationships between behavior and distress as well as in the interactions between
common-sense beliefs and behavior changes tested in models predicting distress. Second,
it is unclear whether our baseline measurement time point was ideal. Because it was
timed to be relative to participants’ treatment plans, the baseline questionnaires were
completed at varying lengths of time since diagnosis and prior to the end of treatment.
We chose mid- to late treatment as the baseline time point because we reasoned that
distress would be relatively low at this time as compared to at diagnosis or at the very end
of treatment, making a potential post-treatment increase in distress more evident. Ideally,
however, one would compare post-treatment adjustment to adjustment prior to diagnosis.
Given the near impossibility of conducting such a study, it may have been beneficial to
measure distress at additional points such as just after diagnosis but prior to treatment and
on the last day of treatment to provide a fuller context for interpreting of the nature and
extent of distress following treatment.
Related to this idea, we have also discussed the idea of a response shift over time
in participants’ reports of distress and quality of life (e.g., Schwartz, Sprangers, Carey, &
Reed, 2004; Sprangers & Schwartz, 1999). Participants’ responses to questionnaires may
have different anchors following treatment as compared to prior to treatment or diagnosis
in that internal standards and conceptualization of quality of life may have changed based
on experiences with cancer and treatment. It is unknown to what extent a response shift
may have affected participants’ reports in the present study, and measuring distress and
117
quality of life as shortly after diagnosis as possible when this type of response shift would
be minimal may have been beneficial.
A final important limitation of the current study is the homogenous nature of the
present sample; almost all participants were Caucasian, and the vast majority were
married and well-educated. Therefore, generalizations of the data to other populations of
breast cancer survivors may be limited.
Summary and Implications for Theory and Practice
Women with breast cancer as a whole appear to be remarkably well-adjusted both
during adjuvant treatment and during the months following treatment. This was true
across a variety of measures including anxiety, depression, cancer-related worry, breast
cancer symptomatology, and multiple domains of health-related quality of life. While it
has previously been documented that long-term breast cancer survivors report good
adjustment with respect to distress and quality of life (Dorval, Maunsell, Deschenes,
Brisson, & Masse, 1998; Ganz, Rowland, Desmond, Meyerowitz, & Wyatt, 1998; Moyer
& Salovey, 1996), the present study extends these findings to the immediate posttreatment period. Results also highlight the speed with which cancer and treatmentrelated symptoms diminished and health-related quality of life returned to levels
commensurate with the larger population following the end of treatment.
There was little empirical evidence to support the common belief that breast
cancer patients experience a period of disrupted adjustment following treatment. No
significant increase in distress was found after treatment ended on any measures of
distress, including depression, anxiety, cancer-related intrusion, and recurrence worry.
Findings suggest that clinical and qualitative reports of disrupted adjustment following
the end of adjuvant treatment are likely limited to a small proportion of breast cancer
survivors. This does not mean, however, that cancer survivor’s lives return to normal
after treatment ends. To the contrary, women report significant concerns about ongoing
118
physical symptoms related to cancer and treatment, the possibility of a cancer recurrence,
the impact of cancer on their families, and how to go about rebuilding a new “normal”
after treatment ends. These specific indicators of distress may be more relevant than
general distress indices following treatment and should therefore be the subject of
attention by health care providers and researchers. Moreover, breast cancer survivors
would likely benefit from interventions targeting these concerns.
Some concerns might be addressed by providing education to cancer survivors.
Women’s ongoing concerns about physical sequelae in the context of overall low levels
of symptoms and quality of life at population norms may be due to unrealistic
expectations regarding the speed and degree of physical recovery, and therefore breast
cancer patients may benefit from information regarding typical timelines for the
dissipation of cancer and treatment-related physical problems and common problems that
remain following treatment, such as lymphedema, sexual dysfunction, fatigue, and
changes in body image (Bumpers, Best, Norman, & Weaver, 2002; Cella, Davis,
Breitbart, & Curt, 2001; Meyerowitz, Desmond, Rowland, Wyatt, & Ganz, 1999). Cancer
survivors might also benefit from health care providers acknowledging the need to create
a “new normal” when returning to one’s pre-cancer baseline lifestyle or level of
functioning may not be desirable or attainable. Given survivor’s concerns that their
families are impacted by their cancer, including family members in psychosocial or
educational interventions and support services directed toward families may also be of
benefit.
Results of the present study suggest that the months after treatment ends are an
opportune time for encouraging adaptive changes in health practices, stress management,
and life priorities. Large proportions of the present sample reported making changes in
each of these areas. Given that women exhibit mild distress related to creating a “new
normal,” they may benefit from at least some support and guidance during this time.
Providing information regarding recommended health practices and resources such as
119
support groups, dietary counseling, exercise programs, or stress management classes in
the community (e.g., yoga, mindfulness-based stress reduction) may assist women in
improving their health and reducing distress. Recent work suggests that breast cancer
survivors are receptive to health recommendations provided by their oncologist or via
mail (Demark-Wahnefried, Peterson, McBride, Lipkus, & Clipp, 2000; Jones, Courneya,
Fairey, & Mackey, 2004). The National Cancer Institute publication, Facing Forward
(National Cancer Institute, 2002), is also an excellent resource for this purpose. New
psychoeducational interventions designed to facilitate post-treatment adjustment by
promoting active coping and goal-setting also show promising results (Cimprich et al.,
2005; Stanton, Ganz, Kwan et al., 2005).
Although post-treatment distress was minimal for breast cancer patients as a
group, identifying predictors of adjustment nonetheless is important as variability in
distress and quality of life was present and a subset of patients remained distressed. We
initially conceptualized behavior changes and health practices as potentially adaptive
coping strategies with respect to post-treatment adjustment. We suggested that positive
behavior changes and practicing healthful behaviors are active coping strategies, and the
literature indicates that active coping leads to better adjustment in the face of cancerrelated stressors (Carver et al., 1993; Stanton & Snider, 1993). The finding that positive
health practices predicted less distress and better quality of life following treatment
supports this conceptualization. In contrast, however, making adaptive behavior changes
was generally associated with greater post-treatment distress. We speculate that lifestyle
adjustments may evoke this distress. If this is the case, these findings underscore the need
for guidance and support as women attempt to create a “new normal”. We also speculate
that positive behavior changes, if maintained over time, may eventually lead to attenuated
distress.
Results of the present study also provide insight into how women decide to make
behavior changes and what changes they decide to make. Domains of patients’ common-
120
sense beliefs about illness, described by Leventhal and others (Leventhal, Meyer, &
Nerenz, 1980; Leventhal & Nerenz, 1985; Leventhal, Nerenz, & Steele, 1984; Meyer,
Leventhal, & Gutmann, 1985), were very good predictors of behavior change among
breast cancer patients. In particular, women who believed they had more control over
their cancer were somewhat more apt to make positive behavior changes. Perceiving
cancer as an acute condition was adaptive with respect to alcohol use and diet, while
perceiving cancer as chronic in nature was more adaptive with respect to promoting
breast self-examination. The most robust predictor of behavior changes and health
practices, however, were women’s attributions regarding what may have caused their
cancer and what factors they believed could prevent recurrence. Attributing cancer to
controllable factors was associated with healthful behaviors and positive behavior
changes, and behavior changes matched the content of women’s attributions. Results
suggest that Leventhal’s model is effective in predicting coping outcomes among breast
cancer patients. Findings further indicate that for breast cancer patients, cancer
attributions may be one of the most important dimensions of illness beliefs as they were
superior to other domains in predicting coping outcomes.
In conclusion, breast cancer patients appear to be actively creating a “new
normal” for themselves following treatment, and changes in health practices and stress
management strategies are likely an important part of this process. Assessing patients’
attributions and beliefs about personal control over cancer, as well as providing
psychoeducational interventions and guidance addressing post-treatment behavior
changes, may assist cancer survivors in this important journey.
121
REFERENCES
ACS Nutrition and Physical Activity Guidelines Advisory Committee. (2001). ACS
recommendations for nutrition and physical activity for cancer prevention.
Atlanta, GA: American Cancer Society.
Ahles, T. A., Saykin, A. J., Furstenberg, C. T., Cole, B., Mott, L. A., Skalla, K., et al.
(2002). Neuropsychologic impact of standard-dose systemic chemotherapy in
long-term survivors of breast cancer and lymphoma. Journal of Clinical
Oncology, 20, 485-493.
Aiken, L. S., Fenaughty, A. M., West, S. G., Johnson, J. J., & Luckett, T. L. (1995).
Perceived determinants of risk for breast cancer and the relations among objective
risk, perceived risk, and screening behavior over time. Women's Health, 1, 27-50.
Aiken, L. S., & West, S. G. (1991). Multiple Regression: Testing and Interpreting
Interactions. Newbury Park: Sage.
Alter, C. L., Pelcovitz, D., Axelrod, A., Goldenberg, B., Harris, H., Meyers, B., et al.
(1996). Identification of PTSD in cancer survivors. Psychosomatics, 37, 137-143.
American Cancer Society. (2006a). Breast cancer facts & figures 2005-2006. Atlanta,
GA: Author.
American Cancer Society. (2006b). Cancer facts & figures 2006. Atlanta, GA: Author.
American Institute for Cancer Research. (1997). Food, nutrition, and the prevention of
cancer: a global perspective. Washington, D.C.: Author.
American Institute for Cancer Research. (2002). Simple steps to prevent cancer.
Washington, D.C.: Author.
Anagnostopoulos, F., & Spanea, E. (2005). Assessing illness representations of breast
cancer: a comparison of patients with healthy and benign controls. Journal of
Psychosomatic Research, 58, 327-334.
Andersen, B. L., Anderson, B., & deProsse, C. (1989). Controlled prospective
longitudinal study of women with cancer: psychological outcomes. Journal of
Consulting and Clinical Psychology, 57, 692-697.
Andrykowski, M. A., & Brady, M. J. (1994). Health locus of control and psychological
distress in cancer patients: Interactive effects of context. Journal of Behavioral
Medicine, 17, 439-458.
Andrykowski, M. A., Brady, M. J., & Henslee-Downey, P. J. (1994). Psychosocial
factors predictive of survival after allogenic bone marrow transplantation for
leukemia. Psychosomatic Medicine, 56, 432-439.
Andrykowski, M. A., & Cordova, M. J. (1998). Factors associated with PTSD symptoms
following treatment for breast cancer: test of the Andersen model. Journal of
Traumatic Stress, 11, 189-203.
122
Baker, F., Haffer, S. C., & Denniston, M. (2003). Health-related quality of life of cancer
and noncancer patients in Medicare managed care. Cancer, 97, 674-681.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51, 1173-1182.
Barsevick, A. M., Whitmer, K., & Walker, L. (2002). In their own words: using the
common sense model to analyze patient descriptions of cancer-related fatigue.
Oncology Nursing Forum, 28, 1363-1369.
Beeber, L. S., Shea, J., & McCorkle, R. (1998). The Center for Epidemiologic Studies
Depression Scale as a measure of depressive symptoms in newly diagnosed
patients. Journal of Psychosocial Oncology, 16, 1-20.
Beenken, S. W., & Bland, K. I. (2002). Breast-conserving surgery. In M. H. Torosian
(Ed.), Breast Cancer: A Guide to Detection and Multidisciplinary Therapy (pp.
89-97). Totowa, New Jersey: Humana Press.
Begbie, S. D., Kerestes, Z. L., & Bell, D. R. (1996). Patterns of alternative medicine use
by cancer patients. Medical Journal of Australia, 165, 545-548.
Beisecker, A. E., Cook, M. R., Ashworth, J., Hayes, J., Brecheisen, M., Helmig, L., et al.
(1997). Side effects of adjuvant chemotherapy: perceptions of node-negative
breast cancer patients. Psycho-Oncology, 6, 85-93.
Berckman, K. L., & Austin, J. K. (1993). Causal attribution, perceived control, and
adjustment in patients with lung cancer. Oncology Nursing Forum, 20, 23-30.
Bishop, G. D., Briede, C., Cavazos, L., Grotzinger, R., & McMahon, S. (1987).
Processing illness information: the role of disease prototypes. Basic and Applied
Social Psychology, 8, 21-43.
Bleiker, E. M. A., Pouwer, F., van der Ploeg, H. M., Leer, J. W. H., & Adèr, H. J. (2000).
Psychological distress two years after diagnosis of breast cancer: frequency and
prediction. Patient Education and Counseling, 40, 209-217.
Block, G., Gillespie, C., Rosenbaum, E. H., & Jensen, C. (2000). A rapid food screener to
assess fat and fruit and vegetable intake. American Journal of Preventive
Medicine, 18, 284-288.
Bower, J. E., Ganz, P. A., Desmond, K., Bernaards, C., Rowland, J. H., Meyerowitz, B.
E., et al. (2006). Fatigue in long-term breast carcinoma survivors: a longitudinal
investigation. Cancer, 106, 751-758.
Bradley, C. J., & Bednarek, H. L. (2002). Employment patterns of long-term cancer
survivors. Psycho-Oncology, 11, 188-198.
Brezden, C. B., Phillips, K. A., Abdolell, M., Bunston, T., & Tannock, I. F. (2000).
Cognitive function in breast cancer patients receiving adjuvant chemotherapy.
Journal of Clinical Oncology, 18, 2695-2701.
123
Broeckel, J. A., Jacobsen, P. B., Balducci, L., Horton, J., & Lyman, G. H. (2000). Quality
of life after adjuvant chemotherapy for breast cancer. Breast Cancer Research &
Treatment, 62, 141-150.
Brown, J., Byers, T., Thompson, K., Eldridge, B., Doyle, C., & Williams, A. M. (2001).
Nutrition during and after cancer treatment: a guide for informed choices by
cancer survivors. CA A Cancer Journal for Clinicians, 51, 153-187.
Buick, D. L. (1997). Illness representations and breast cancer: coping with radiation and
chemotherapy. In K. J. Petrie & J. A. Weinman (Eds.), Perceptions of Health and
Illness: Current Research and Applications (pp. 379-409). New York: Routledge.
Bumpers, H. L., Best, I. M., Norman, D., & Weaver, W. L. (2002). Debilitating
lymphedema of the upper extremity after treatment of breast cancer. American
Journal of Clinical Oncology, 25, 365-367.
Burstein, H. J., Gelber, S., Guadagnoli, E., & Weeks, J. C. (1999). Use of alternative
medicine by women with early-stage breast cancer. New England Journal of
Medicine, 340, 1733-1739.
Butow, P. N., Hiller, J. E., Price, M. A., Thackway, S. V., Kricker, A., & Tennant, C. C.
(2000). Epidemiological evidence for a relationship between life events, coping
style, and personality factors in the development of breast cancer. Journal of
Psychosomatic Research, 49, 169-181.
Calle, E. E., Frumkin, H., Henley, S. J., Savitz, D. A., & Thun, M. J. (2002).
Organochlorines and breast cancer risk. CA: A Cancer Journal for Clinicians, 52,
301-309.
Cameron, L. D. (1997). Screening for cancer: illness perceptions and illness worry. In K.
J. Petrie & J. A. Weinman (Eds.), Perceptions of Health and Illness: Current
Research and Applications (pp. 291-322). New York: Routledge.
Cameron, L. D., Leventhal, H., & Love, R. R. (1998). Trait anxiety, symptom
perceptions, and illness-related responses among women with breast cancer in
remission during a tamoxifen clinical trial. Health Psychology, 17, 459-469.
Carver, C. S., Pozo, C., Harris, S. D., Noriega, V., Scheier, M. F., Robinson, D. S., et al.
(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.
Cella, D., Davis, K., Breitbart, W., & Curt, G. (2001). Cancer-related fatigue: prevalence
of proposed diagnostic criteria in a United Stated sample of cancer survivors.
Journal of Clinical Oncology, 19, 3385-3391.
Chang, V. T., Thaler, H. T., Polyak, T. A., Kornblith, A. B., Lepore, J. M., & Portenoy,
R. K. (1998). Quality of life and survival: the role of multidimensional symptom
assessment. Cancer, 83, 173-179.
Chen, C. C., David, A. S., Nunnerley, H., Michell, M., Dawson, J. L., Berry, H., et al.
(1995). Adverse life events and breast cancer: case-control study. British Medical
Journal, 311, 1527-1530.
124
Christensen, A. J., Moran, P. J., Ehlers, S. L., Raichle, K., Karnell, L., & Funk, G.
(1999). Smoking and drinking behavior in patients with head and neck cancer:
effects of behavioral self-blame and perceived control. Journal of Behavioral
Medicine, 22, 407-418.
Christensen, A. J., Moran, P. J., & Wiebe, J. S. (1999). Assessment of irrational health
beliefs: relation to health practices and medical regimen adherence. Health
Psychology, 18, 169-176.
Cimprich, B., Janz, N. K., Northouse, L., Wren, P. A., Given, B., & Given, C. W. (2005).
Taking CHARGE: A self-management program for women following breast
cancer treatment. Psycho-Oncology, 14, 704-717.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.
Cohen, S., Doyle, W. J., Skoner, D. P., Rabin, B. S., & Gwaltney, J. M. (1997). Social
ties and susceptibility to the common cold. JAMA, 277, 1940-1944.
Cohen, S., Tyrell, D. A. J., Russell, M. A. H., Jarvis, M. J., & Smith, A. P. (1993).
Smoking, alcohol consumption, and susceptibility to the common cold. American
Journal of Public Health, 83, 1277-1283.
Costanzo, E. S., Lutgendorf, S. K., Bradley, S. L., Rose, S. L., & Anderson, B. (2005).
Cancer attributions, distress, and health behavior among gynecologic cancer
survivors. Psychosomatic Medicine, 67, 972-980.
Costanzo, E. S., Lutgendorf, S. K., Rothrock, N. E., & Anderson, B. (2006). Coping and
quality of life among women extensively treated for gynecologic cancer. PsychoOncology, 15, 132-142.
Cunningham, L. L., Andrykowski, M. A., Wilson, J. F., McGrath, P. C., Sloan, D. A., &
Kenady, D. E. (1998). Physical symptoms, distress, and breast cancer risk
perceptions in women with benign breast problems. Health Psychology, 17, 371375.
Demark-Wahnefried, W., Peterson, B., McBride, C., Lipkus, I., & Clipp, E. (2000).
Current health behaviors and readiness to pursue lifestyle changes among men
and women diagnosed with early stage prostate and breast carcinomas. Cancer,
88, 674-684.
Deshields, T., Tibbs, T., Fan, M. Y., Bayer, L., Taylor, M., & Fisher, E. (2005). Ending
treatment: the course of emotional adjustment and quality of life among breast
cancer survivors immediately following radiation therapy. Support Cancer Care,
13, 1018-1026.
Dorval, M., Maunsell, E., Deschenes, L., Brisson, J., & Masse, B. (1998). Long-term
quality of life after breast cancer: comparison of 8-year survivors with population
controls. Journal of Clinical Oncology, 16, 487-494.
Duijts, S. F. A., Zeegers, M. P. A., & Borne, B. V. (2003). The association between
stressful life events and breast cancer risk: a meta-analysis. International Journal
of Cancer, 107, 1023-1029.
125
Early Breast Cancer Trialists' Collaborative Group. (1998). Tamoxifen for early breast
cancer: an overview of the randomised trials. Lancet, 351, 1451-1465.
Ell, K., Nishimoto, R., Morvay, T., Mantell, J., & Hamovitch, M. (1989). A longitudinal
analysis of psychological adaptation among survivors of cancer. Cancer, 63, 406413.
Ensel, W. M. (1986). Measuring depression: the CES-D scale. In Social Support, Life
Events, and Depression: Academic Press.
Faller, H., Schilling, S., & Lang, H. (1995). Causal attributions and adaptation among
lung cancer patients. Journal of Psychosomatic Research, 39, 619-627.
Fan, H. G., Houede-Tchen, N., Yi, Q. L., Chemerynsky, I., Downie, F. P., Sabate, K., et
al. (2005). Fatigue, menopausal symptoms, and cognitive function in women after
adjuvant chemotherapy for breast cancer: 1- and 2-year follow-up of a prospective
controlled study. Journal of Clinical Oncology, 23, 8025-8032.
Ferrell, B. R., Grant, M. M., Funk, B., Otis-Green, S., & Garcia, N. (1997). Quality of
life in breast cancer survivors as identified by focus groups. Psycho-Oncology, 6,
13-23.
Frost, M. H., Suman, V. J., Rummans, T. A., Dose, A. M., Taylor, M., Novotny, P., et al.
(2000). Physical, psychological and social well-being of women with breast
cancer: the influence of disease phase. Psycho-Oncology, 9, 221-231.
Ganz, P. A., Coscarelli, A., Fred, C., Kahn, B., Polinsky, M. L., & Petersen, L. (1996).
Breast cancer survivors: psychosocial concerns and quality of life. Breast Cancer
Research & Treatment, 38, 183-199.
Ganz, P. A., Desmond, K. A., Belin, T. R., Meyerowitz, B. E., & Rowland, J. H. (1999).
Predictors of sexual health in women after a breast cancer diagnosis. Journal of
Clinical Oncology, 8, 2371-2380.
Ganz, P. A., Rowland, J. H., Desmond, K., Meyerowitz, B. E., & Wyatt, G. E. (1998).
Life after breast cancer: understanding women's health-related quality of life and
sexual functioning. Journal of Clinical Oncology, 16, 501-514.
Ganz, P. A., Rowland, J. H., Meyerowitz, B. E., & Desmond, K. A. (1998). Impact of
different adjuvant therapy strategies on quality of life in breast cancer survivors.
Recent Results in Cancer Research, 152, 396-411.
Giraldi, T., Rodani, M. G., Cartei, G., & Grassi, L. (1997). Psychosocial factors and
breast cancer: a 6-year Italian follow-up study. Psychotherapy and
Psychosomatics, 66, 229-236.
Glinder, J. G., & Compas, B. E. (1999). Self-blame attributions in women with newly
diagnosed breast cancer: a prospective study of psychological adjustment. Health
Psychology, 18, 475-481.
Goodwin, P. J., Ennis, M., Pritchard, K. I., Trudeau, M., & Hood, N. (1999). Risk of
menopause during the first year after breast cancer diagnosis. Journal of Clinical
Oncology, 17, 2365-2370.
126
Gotay, C. C. (1985). Why me? Attributions and adjustment by cancer patients and their
mates at two stages in the disease process. Social Science and Medicine, 20, 825831.
Gotay, C. C., & Muraoka, M. Y. (1998). Quality of life in long-term survivors of adultonset cancers. Journal of the National Cancer Institute, 90, 656-657.
Green, B. L., Krupnick, J. L., Stockton, P., & Spertus, I. L. (1998). Prevalence of
posttraumatic stress disorder in women with breast cancer. Psychosomatics, 39,
102-111.
Greer, S., Morris, T., Pettingale, K. W., & Haybittle, J. L. (1990). Psycholsocial response
to breast cancer and 15-year outcome. Lancet, 335, 49-50.
Grunfeld, E. A., Hunter, M. S., Ramirez, A. J., & Richards, M. A. (2003). Perceptions of
breast cancer across the lifespan. Journal of Psychosomatic Research, 54, 141146.
Hagedoorn, M., Buunk, B. P., Kuijer, R. G., Wobbes, T., & Sanderman, R. (2000).
Couples dealing with cancer: role and gender differences regarding psychological
distress and quality of life. Psycho-Oncology, 9, 232-242.
Hann, D., Winter, K., & Jacobsen, P. (1999). Measurement of depressive symptoms in
cancer patients: evaluation of the center for epidemiological studies depression
scale (CES-D). Journal of Psychosomatic Research, 46, 437-443.
Hann, D. M., Jacobsen, P. B., Martin, S. C., Kronish, L. E., Azzarello, L. M., & Fields,
K. K. (1997). Quality of life following bone marrow transplantation for breast
cancer: a comparative study. Bone Marrow Transplantation, 19, 257-264.
Helgeson, V. S., & Tomich, P. L. (2005). Surviving cancer: a comparison of 5-year
disease-free breast cancer survivors with healthy women. Psycho-Oncology, 14,
307-317.
Henderson, J. W., & Donatelle, R. J. (2003). The relationship between cancer locus of
control and complementary and alternative medicine use by women diagnosed
with breast cancer. Psycho-Oncology, 12, 59-67.
Hewitt, J. A., Mokbel, K., van Someran, K. A., Jewell, A. P., & Garrod, R. (2005).
Exercise for breast cancer survival: the effect on cancer risk and cancer-related
fatigue (CRF). International Journal of Fertility & Women's Medicine, 50, 231239.
Hewitt, M., Breen, N., & Devesa, S. (1999). Cancer prevalence and survivorship issues:
Analyses of the 1992 National Health Interview Survey. Journal of the National
Cancer Institute, 91, 1480-1486.
Hilton, B. A. (1989). The relationship of uncertainty, control, commitment, and threat of
recurrence to coping strategies used by women diagnosed with breast cancer.
Journal of Behavioral Medicine, 12, 39-54.
127
Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical clarity in the
study of mediators and moderators: Examples from the child-clinical and pediatric
psychology literatures. Journal of Consulting & Clinical Psychology, 65, 599610.
Hortobagyi, G. N. (1998). Treatment of breast cancer. New England Journal of Medicine,
347, 1566-1575.
Hurny, C., Bernhard, J., Coates, A. S., Castiglione-Gertsch, M., Peterson, H. F., Gelber,
R. D., et al. (1996). Impact of adjuvant therapy on quality of life in women with
node-positive operable breast cancer. Lancet, 347, 1279-1284.
Jenkins, R. A., & Pargament, K. I. (1988). Cognitive appraisals in cancer patients. Social
Science & Medicine, 26, 625-633.
Jenkins, V., Shilling, V., Deutsch, G., Bloomfield, D., Morris, R., Allan, S., et al. (2006).
A 3-year prospective study of the effects of adjuvant treatments on cognition in
women with early stage breast cancer. British Journal of Cancer, 94, 828-834.
Jensen, M. R. (1987). Psychobiological factors predicting the course of breast cancer.
Journal of Personality, 55, 317-342.
Jones, L. W., Courneya, K. S., Fairey, A. S., & Mackey, J. R. (2004). Effects of an
oncologist's recommendation to exercise on self-reported exercise behavior in
newly diagnosed breast cancer survivors: a single-blind, randomized controlled
trial. Annals of Behavioral Medicine, 28, 105-113.
Kendall, A. R., Mahue-Giangreco, M., Carpenter, C. L., Ganz, P. A., & Bernstein, L.
(2005). Influence of exercise activity on quality of life in long-term breast cancer
survivors. Quality of Life Research, 14, 361-371.
Kim, D. J., & Holowaty, E. J. (2003). Brief, validated survey instruments for the
measurement of fruit and vegetable intakes in adults: a review. Preventive
Medicine, 36, 440-447.
Kohl, H. W., Blair, S. N., Paffenbarger, R. S., Macera, C. A., & Kronnenfeld, J. J. (1988).
A mail survey of physical activity habits as related to measured physical fitness.
American Journal of Epidemiology, 127, 1228-1239.
Kohli, N., & Dalal, A. K. (1998). Culture as a factor in causal understanding of illness: a
study of cancer patients. Psychology and Developing Societies, 10, 115-129.
Kulik, L., & Kronfield, M. (2005). Adjustment to breast cancer: the contribution of
resources and causal attributions regarding the illness. Social Work in Health
Care, 41, 37-57.
Kurtz, M. E., Wyatt, G., & Kurtz, J. C. (1995). Psychological and sexual well-being,
philosophical/spiritual views, and health habits of long-term cancer survivors.
Health Care for Women International, 16, 253-262.
Lampic, C., Wennberg, A., Schill, J. E., Brodin, O., Glimelius, B., & Sjoden, P. O.
(1994). Anxiety and cancer-related worry of cancer patients at routine follow-up
visits. Acta Oncologica, 33, 119-125.
128
Lau, R. R., Bernard, T. M., & Hartman, K. A. (1989). Further explorations of commonsense representations of common illnesses. Health Psychology, 8, 195-219.
Lau, R. R., & Hartman, K. A. (1983). Common sense representations of common
illnesses. Health Psychology, 2, 167-185.
Lavery, J. F., & Clarke, V. A. (1996). Causal attributions, coping strategies, and
adjustment to breast cancer. Cancer Nursing, 19, 20-28.
Lazarus, R. S. (1993). Coping theory and research: Past, present, and future.
Psychosomatic Medicine, 55, 234-247.
Lee, M. M., Lin, S. S., Wrensch, M. R., Adler, S. R., & Eisenberg, D. (2000). Alternative
therapies used by women with breast cancer in four ethnic populations. Journal of
the National Cancer Institute, 92, 42-47.
Leopold, K. A., Ahles, T. A., Walch, S., Amdur, R. J., Mott, L. A., Wiegand-Packard, L.,
et al. (1998). Prevalence of mood disorders and utility of the PRIME-MD in
patients undergoing radiation therapy. International Journal of Radiation
Oncology, Biology, Physics, 42, 1105-1112.
Lethborg, C. E., Kissane, D., Burns, W. I., & Snyder, R. (2000). "Cast Adrift": the
experience of completing treatment among women with early stage breast cancer.
Journal of Psychosocial Oncology, 18, 73-90.
Levav, I., Kohn, R., Iscovich, J., Abramson, J. H., Tsai, W. Y., & Vigdorovich, D.
(2000). Cancer incidence and survival following bereavement. American Journal
of Public Health, 90, 1601-1607.
Leventhal, H., Benyamini, Y., Brownlee, S., Diefenbach, M., Leventhal, E. A., PatrickMiller, L., et al. (1997). Illness representations: theoretical foundations. In K. J.
Petrie & J. A. Weinman (Eds.), Perceptions of Health and Illness: Current
Research and Applications (pp. 19-45). New York: Routledge.
Leventhal, H., Meyer, D., & Nerenz, D. (1980). The common-sense representation of
illness danger. In S. Rachman (Ed.), Medical Psychology (Vol. 2, pp. 7-30). New
York: Pergamon.
Leventhal, H., & Nerenz, D. (1985). The assessment of illness cognition. In P. Karoly
(Ed.), Measurement Strategies in Health Psychology (pp. 517-555). New York:
Wiley.
Leventhal, H., Nerenz, D. R., & Steele, D. S. (1984). Illness representations and coping
with health threats. In A. Baum, S. E. Taylor & J. E. Singer (Eds.), Handbook of
Psychology and Health (Vol. IV, pp. 219-252). Hillsdale, NJ: Erlbaum.
Lowery, B. J., Jacobsen, B. S., & DuCette, J. (1993). Causal attribution, control, and
adjustment to breast cancer. Journal of Psychosocial Oncology, 10, 37-53.
Malcarne, V. L., Compas, B. E., Epping-Jordan, J. E., & Howell, D. C. (1995). Cognitive
factors in adjustment to cancer: attributions of self-blame and perceptions of
control. Journal of Behavioral Medicine, 18, 401-417.
129
Manusell, E., Brisson, J., & Deschenes, L. (1993). Arm problems and psychological
distress after surgery for breast cancer. Canadian Journal of Surgery, 36, 315320.
Marcus, A. C., Garrett, K. M., Cella, D., Wenzel, L. B., Brady, M. J., Crane, L. A., et al.
(1998). Telephone counseling of breast cancer patients after treatment: a
description of a randomized clinical trial. Psycho-Oncology, 7, 470-482.
Marks, G., Richardson, J. L., Graham, J. W., & Levine, A. (1986). Role of health locus of
control beliefs and expectations of treatment efficacy in adjustment to cancer.
Journal of Personality and Social Psychology, 51, 443-450.
Martin, R., & Lemos, K. (2002). From heart attacks to melanoma: do common sense
models of somatization influence symptom interpretation for female victims?
Health Psychology, 21, 25-32.
Maskarinec, G., Gotay, C. C., Tatsumura, Y., Shumay, D. M., & Kakai, H. (2001).
Perceived cancer causes: use of complementary and alternative therapy. Cancer
Practice, 9, 183-190.
Maskarinec, G., Murphy, S., Shumay, D. M., & Kakai, H. (2001). Dietary changes
among cancer survivors. European Journal of Cancer Care, 10, 12-20.
Maskarinec, G., Shumay, D. M., Kakai, H., & Gotay, C. C. (2000). Ethnic differences in
complementary and alternative medicine use among cancer patients. Journal of
Alternative & Complementary Medicine, 6, 531-538.
Matthews, K. A., Siegal, J. M., Kuller, L. H., Thompson, M., & Varat, M. (1983).
Determinants of decisions to seek medical treatment by patients with acute
myocardial infarction symptoms. Journal of Personality and Social Psychology,
44, 1144-1156.
Maunsell, E., Drolet, M., Brisson, J., Robert, J., & Deschenes, L. (2002). Dietary change
after breast cancer: extent, predictors, and relation with psychological distress.
Journal of Clinical Oncology, 20, 1017-1025.
McKinley, E. D. (2000). Under toad days: surviving the uncertainty of cancer recurrence.
Annals of Internal Medicine, 133, 479-480.
McQuellon, R. P., Russell, G. B., Rambo, T. D., Craven, B. L., Radford, J., Perry, J. J., et
al. (1998). Quality of life and psychological distress of bone marrow transplant
recipients: the 'time trajectory' to recovery over the first year. Bone Marrow
Transplantation, 21, 477-486.
Meyer, D., Leventhal, H., & Gutmann, M. (1985). Common-sense models of illness: the
example of hypertension. Health Psychology, 4, 115-135.
Meyerowitz, B. E., Desmond, K. A., Rowland, J. H., Wyatt, G. E., & Ganz, P. A. (1999).
Sexuality following breast cancer. Journal of Sex & Marital Therapy, 25, 237250.
Miller, G. E., Cohen, S., & Herbert, T. B. (1999). Pathways linking major depression and
immunity in ambulatory female patients. Psychosomatic Medicine, 61, 850-860.
130
Miller, M. (2005). The effects of perceived stress on reactions to messages designed to
increase health behaviors. Journal of Behavioral Medicine, 28, 425-432.
Modi, S., Panageas, K. S., Duck, E. T., Bach, A., Weinstock, N., Dougherty, J., et al.
(2002). Prospective exploratory analysis of the association between tumor
response, quality of life, and expenditures among patients receiving paclitaxel
monotherapy for refractory metastatic breast cancer. Journal of Clinical
Oncology, 20, 3665-3673.
Molassiotis, A., Van den Akker, O. B. A., Milligan, D. W., & Goldman, J. M. (1997).
Symptom distress, coping style and biological variables as predictors of survival
after bone marrow transplantation. Journal of Psychosomatic Research, 42, 275285.
Moss-Morris, R., Weinman, J., Petrie, K. J., Horne, R., Cameron, L. D., & Buick, D.
(2002). The Revised Illness Perception Questionnaire (IPQ-R). Psychology and
Health, 17, 1-16.
Moulton, J. M., Sweet, D. M., Temoshok, L., & Mandel, J. S. (1987). Attributions of
blame and responsibility in relation to distress and health behavior change in
people with AIDS and AIDS-related complex. Journal of Applied Social
Psychology, 17, 493-506.
Moyer, A., & Salovey, P. (1996). Psychosocial sequelae of breast cancer and its
treatment. Annals of Behavioral Medicine, 18, 110-125.
Mullan, F. (1985). Seasons of survival: reflections of a physician with cancer. The New
England Journal of Medicine, 313, 270-273.
Mumma, C., & McCorkle, R. (1983). Causal attribution and life-threatening disease.
International Journal of Psychiatry in Medicine, 12, 311-319.
Murphy, K. C., Jenkins, P. L., & Whittaker, J. A. (1996). Psychosocial morbidity and
survival in adult bone marrow transplant recipients--a follow-up study. Bone
Marrow Transplantation, 18, 199-201.
National Cancer Institute. (1999). Chemotherapy and you: a guide to self-help during
cancer treatment. Bethesda, MD: Author.
National Cancer Institute. (2000). Radiation therapy and you: a guide to self-help during
cancer treatment. Bethesda, MD: Author.
National Cancer Institute. (2002). Facing forward: life after cancer treatment. Bethesda,
MD: Author.
National Cancer Institute. (2006). Breast cancer PDQ: treatment (health professional
version). Bethesda, MD: Author.
National Comprehensive Cancer Network. (2005). Breast cancer treatment guidelines for
patients. Atlanta, GA: American Cancer Society.
Newsom, J. T., Knapp, J. E., & Schulz, R. (1996). Longitudinal analysis of specific
domains of internal control and depressive symptoms in patients with recurrent
cancer. Health Psychology, 15, 323-331.
131
Northouse, L. (1989). A longitudinal study of the adjustment of patients and husbands to
breast cancer. Oncology Nursing Forum, 16, 511-516.
Northouse, L., Mood, D., Templin, T., Mellon, S., & George, T. (2000). Couples' patterns
of adjustment to colon cancer. Social Science and Medicine, 50, 271-284.
Northouse, L., Templin, T., Mood, D., & Oberst, M. (1998). Couples' adjustment to
breast cancer and benign breast disease: a longitudinal analysis. PsychoOncology, 7, 37-48.
Osowiecki, D. M., & Compas, B. E. (1999). A prospective study of coping, perceived
control, and psychological adaptation to breast cancer. Cognitive Therapy and
Research, 23, 169-180.
Patterson, R. E., Neuhouser, M. L., Hedderson, M. M., Schwartz, S. M., Standish, L. J.,
& Bowen, D. J. (2003). Changes in diet, physical activity, and supplement use
among adults diagnosed with cancer. Journal of the American Dietetic
Association, 103, 323-328.
Petrie, K. J., Cameron, L. D., Ellis, C. J., Buick, D., & Weinman, J. (2002). Changing
illness perceptions after myocardial infarction: an early intervention randomized
controlled trial. Psychosomatic Medicine, 64, 580-586.
Petticrew, M., Fraser, J. M., & Regan, M. F. (1999). Adverse life-events and risk of
breast cancer: a meta-analysis. British Journal of Health Psychology, 4, 1-17.
Portenoy, R. K., Thaler, H. T., Kornblith, A. B., Lepore, J. M., Friedlander-Klar, H.,
Kiyasu, E., et al. (1994). The Memorial Symptom Assessment Scale: an
instrument for the evaluation of symptom prevalence, characteristics and distress.
European Journal of Cancer, 30A, 1326-1336.
Protheroe, D., Turvey, K., Horgan, K., Benson, E., Bowers, D., & House, A. (1999).
Stressful life events and difficulties and onset of breast cancer: case-control study.
British Medical Journal, 319, 1027-1030.
Rabin, C., Leventhal, H., & Goodin, S. (2004). Conceptualization of disease timeline
predicts posttreatment distress in breast cancer patients. Health Psychology, 23,
407-412.
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.
Ransom, S., Jacobsen, P. B., Schmidt, J. E., & Andrykowski, M. A. (2005). Relationship
of problem-focused coping strategies to changes in quality of life following
treatment for early stage breast cancer. Journal of Pain and Symptom
Management, 30, 243-253.
Reynolds, P., Hurley, S., Torres, M., Jackson, J., Boyd, P. T., Chen, V. W., et al. (2000).
Use of coping strategies and breast cancer survival: results from the Black/White
Cancer Survival Study. American Journal of Epidemiology, 152, 940-949.
Rich, M. R., Smith, T. W., & Christensen, A. J. (1999). Attributions and adjustment in
end-stage renal disease. Cognitive Therapy and Research, 23, 143-158.
132
Ringdal, G. I. (1995). Correlates of hopelessness in cancer patients. Journal of
Psychosocial Oncology, 13, 47-66.
Roberts, F. D., Newcomb, P. A., Trentham-Dietz, A., & Storer, B. E. (1996). Selfreported stress and risk of breast cancer. Cancer, 77, 1089-1093.
Romond, E. H., Perez, E. A., Bryant, J., Suman, V. J., Geyer, C. E., Davidson, N. E., et
al. (2005). Trastuzumab plus adjuvant chemotherapy for operable HER2-positive
breast cancer. New England Journal of Medicine, 353, 1673-1684.
Rothman, A. J. (2000). Toward a theory-based analysis of behavioral maintenance.
Health Psychology, 19(1S), 64-69.
Rothrock, N. E. (2002). Worry of recurrence in breast cancer survivors. Unpublished
doctoral dissertation, University of Iowa.
Rowland, J. H., & Holland, J. C. (1990). Breast cancer. In J. C. Holland & J. H. Rowland
(Eds.), Handbook of Psychooncology: Psychological Care of the Patient with
Cancer (pp. 188-207). New York: Oxford University Press.
Schagen, S. B., van Dam, F. S., Muller, M. J., Boogerd, W., Lindeboom, J., & Bruning,
P. F. (1999). Cognitive deficits after postoperative adjuvant chemotherapy for
breast carcinoma. Cancer, 85, 640-650.
Schnipper, H. H. (2001). Life after breast cancer. Journal of Clinical Oncology, 19, 35813584.
Schou, I., Ekeberg, O., Sandvik, L., Hjermstad, M. J., & Ruland, C. M. (2005). Multiple
predictors of health-related quality of life in early stage breast cancer. Data from a
year follow-up study compared with the general population. Quality of Life
Research, 14, 1813-1823.
Schulz, R., Bookwala, J., Knapp, J. E., Scheier, M., & Williamson, G. M. (1996).
Pessimism, age, and cancer mortality. Psychology and Aging, 11, 304-309.
Schwartz, C. E., & Sprangers, M. A. G. (1999). Methodological approaches for assessing
response shift in longitudinal health-related quality of life research. Social Science
and Medicine, 48, 1531-1548.
Schwartz, C. E., Sprangers, M. A. G., Carey, A., & Reed, G. (2004). Exploring response
shift in longitudinal data. Psychology and Health, 19, 51-69.
SF-36.org. (n.d.-a). 1998 SF-36 Normative Data. Retrieved April 15, 2006, from
http://www.sf-36.org/research/sf98norms.shtml
SF-36.org. (n.d.-b). SF-36® PCS, MCS and NBS Calculator. Retrieved April 15, 2006,
from http://www.sf-36.org/nbscalc/index.shtml
Shumay, D. M., Maskarinec, G., Gotay, C. C., Heiby, E. M., & Kakai, H. (2002).
Determinants of the degree of complementary and alternative medicine use
among patients with cancer. Journal of Alternative & Complementary Medicine,
8, 661-671.
133
Spitzer, R. L., Kroenke, K., Williams, J. B. W., & The Patient Health Questionnaire
Primary Care Study Group. (1999). Validation and utility of a self-report version
of PRIME-MD: The PHQ primary care study. JAMA, 282, 1737-1744.
Sprangers, M. A. G., & Schwartz, C. E. (1999). Integrating response shift into healthrelated quality of life research: a theoretical model. Social Science and Medicine,
48, 1507-1515.
Stanton, A. L., Danoff-Burg, S., Cameron, C. L., Bishop, M., Collins, C. A., Kirk, S. B.,
et al. (2000). Emotionally expressive coping predicts psychological and physical
adjustment to breast cancer. Journal of Consulting & Clinical Psychology, 68,
875-882.
Stanton, A. L., Danoff-Burg, S., & Huggins, M. E. (2002). The first year after breast
cancer diagnosis: hope and coping strategies as predictors of adjustment. PsychoOncology, 11, 93-102.
Stanton, A. L., Danoff-Burg, S., Sworowski, L. A., Collins, C. A., Branstetter, A. D.,
Rodriguez-Hanley, A., et al. (2002). Randomized, controlled trial of written
emotional expression and benefit finding in breast cancer patients. Journal of
Clinical Oncology, 20, 4160-4168.
Stanton, A. L., Ganz, P. A., Kwan, L., Meyerowitz, B. E., Bower, J. E., Krupnick, J. L.,
et al. (2005). Outcomes from the Moving Beyond Cancer psychoeducational,
randomized, controlled trial with breast cancer patients. Jorurnal of Clinical
Oncology, 23, 6009-6018.
Stanton, A. L., Ganz, P. A., Rowland, J. H., Meyerowitz, B. E., Krupnick, J. L., & Sears,
S. R. (2005). Promoting adjustment after treatment for cancer. Cancer, 104, 26082613.
Stanton, A. L., & Snider, P. R. (1993). Coping with a breast cancer diagnosis: A
prospective study. Health Psychology, 12, 16-23.
Stewart, D. E., Cheung, A. M., Duff, S., Wong, F., McQuestion, M., Cheng, T., et al.
(2001). Attributions of cause and recurrence in long-term breast cancer survivors.
Psycho-Oncology, 10, 179-183.
Stillman, M. J. (1977). Women's health beliefs about cancer and breast self-examination.
Nursing Research, 26, 121-127.
Strine, T. W., Chapman, D. P., Kobau, R., & Balluz, L. (2005). Associations of selfreported anxiety symptoms with health-related quality of life and health
behaviors. Social Psychiatry and Psychiatric Epidemiology, 40, 432-438.
Suls, J., & Fletcher, B. (1985). The relative efficacy of avoidant and nonavoidant coping
strategies: A meta-analysis. Health Psychology, 4, 249-288.
Tatsumura, Y., Maskarinec, G., Shumay, D. M., & Kakai, H. (2003). Religious and
spiritual resources, CAM, and conventional treatment in the lives of cancer
patients. Alternative Therapies in Health & Medicine, 9, 64-71.
134
Taylor, S. E., Lichtman, R. R., & Wood, J. V. (1984). Attributions, beliefs about control,
and adjustment to breast cancer. Journal of Personality and Social Psychology,
46, 489-502.
Temoshok, L., Heller, B., Sagebiel, R. W., Blois, M. S., Sweet, D. M., DiClemente, R. J.,
et al. (1985). The relationship of psychosocial factors to prognostic indicators in
cutaneous malignant melanoma. Journal of Psychosomatic Research, 29, 139153.
Tesauro, G. M., Rowland, J. H., & Lustig, C. (2002). Survivorship resources for posttreatment cancer survivors. Cancer Practice, 10, 277-283.
Thompson, S. C., Sobolew, S. A., Galbraith, M. E., Schwankovsky, L., & Cruzen, D.
(1993). Maintaining perceptions of control: finding perceived control in low
control circumstances. Journal of Personality and Social Psychology, 64, 293304.
Tjemsland, L., Søreide, J. A., & Malt, U. F. (1996). Traumatic distress symptoms in early
breast cancer I: Acute response to diagnosis. Psycho-Oncology, 5, 1-8.
Tschuschkle, V., Hertenstein, B., Arnold, R., Bunjes, D., Denzinger, R., & Kaechele, H.
(2002). Associations between coping and survival time of adult leukemia patients
receiving allogenic bone barrow transplantation: results of a prospective study.
Journal of Psychosomatic Research, 50, 277-285.
Turner, J., Hayes, S., & Reul-Hirche, H. (2004). Improving the physical status and
quality of life of women treated for breast cancer: a pilot study of a structured
exercise intervention. Journal of Surgical Oncology, 86, 141-146.
Turner-Bowker, D. M., Bartley, P. J., & Ware, J. E. (2002). SF-36® Health Survey &
“SF” Bibliography: Third Edition (1988-2000). Lincoln, RI: QualityMetric
Incorporated.
van Dam, F. S., Schagen, S. B., Muller, M. J., Boogerd, W., vd Wall, E., Droogleever, F.
M. E., et al. (1998). Impairment of cognitive function in women receiving
adjuvant treatment for high-risk breast cancer: high-dose versus standard-dose
chemotherapy. Journal of the National Cancer Institute, 90, 210-218.
Vickberg, S. M. J. (2003). The Concerns About Recurrence Scale (CARS): a systematic
measure of women's fears about the possibility of breast cancer recurrence.
Annals of Behavioral Medicine, 25, 16-24.
Vickberg, S. M. J., & Revenson, T. A. (1999). Fears of breast cancer recurrence go
beyond the threat of death. Poster presentation at the Society of Behavioral
Medicine, San Diego.
Vinokur, A. D., Threatt, B. A., Vinokur-Kaplan, D., & Satariano, W. A. (1990). The
process of recovery from breast cancer for younger and older patients. Cancer,
65, 1242-1254.
Voogd, A. C., Ververs, J. M., Vingerhoets, A. J., Roumen, R. M., Coebergh, J. W., &
Crommelin, M. A. (2003). Lymphoedema and reduced shoulder function as
indicators of quality of life after axillary lymph node dissection for invasive breast
cancer. British Journal of Surgery, 90, 76-81.
135
Walker, B. L., Nail, L. M., Larsen, L., Magill, J., & Schwartz, A. (1996). Concerns,
affect, and cognitive disruption following completion of radiation treatment for
localized breast or prostate cancer. Oncology Nursing Forum, 23, 1181-1187.
Ward, S. E., Viergutz, G., Tormey, D., deMuth, J., & Paulen, A. (1992). Patients'
reactions to completion of adjuvant breast cancer therapy. Nursing Research, 41,
362-366.
Ware, J. E., Kosinski, M., & Dewey, J. E. (2000). How to Score Version Two of the SF36 Health Survey. Lincoln, RI: QualityMetric, Incorporated.
Ware, J. E., Snow, K. K., Kosinski, M., & Gandek, B. (1993). SF-36® Health Survey
Manual and Interpretation Guide. Boston, MA: New England Medical Center,
The Health Institute.
Watson, M., Haviland, J. S., Greer, S., Davidson, J., & Bliss, J. M. (1999). Influence of
psychological response on survival in breast cancer: a population-based cohort
study. Lancet, 354, 1331-1336.
Weihs, K. L., Enright, T. M., Simmens, S. J., & Reiss, D. (2000). Negative affectivity,
restriction of emotions, and site of metastases predict mortality in recurrent breast
cancer. Journal of Psychosomatic Research, 49, 59-68.
Weinman, J., Petrie, K. J., Moss-Morris, R., & Horne, R. (1996). The illness perception
questionnaire: a new method for assessing the cognitive representation of illness.
Psychology and Health, 11, 431-446.
136
APPENDIX A
TABLES
Table A1:
ACS Recommendations for Nutrition and Physical Activity for Cancer Prevention
1. Eat a variety of healthful foods, with an emphasis on plant sources.
•
Eat five or more servings of a variety of vegetables and fruits each day.
•
Choose whole grains in preference to processed (refined) grains and sugars.
•
Limit consumption of red meats, especially those high in fat and processed.
•
Choose foods that help maintain a healthful weight.
2. Adopt a physically active lifestyle.
•
Engage in at least moderate activity for 30 minutes or more on 5 or more days of
the week.
•
45 minutes or more of moderate to vigorous activity on 5 or more days per week
may further reduce the risk of breast and colon cancer.
3. Maintain a healthful weight throughout life.
•
Balance caloric intake with physical activity.
•
Lose weight if currently overweight or obese.
4. If you drink alcoholic beverages, limit consumption.
137
Table A2:
AICR Diet and Health Guidelines for Cancer Prevention
1. Choose a diet rich in a variety of plant-based foods.
2. Eat plenty of vegetables and fruits.
3. Maintain a healthy weight and be physically active.
4. Drink alcohol only in moderation, if at all.
5. Select foods low in fat and salt.
6. Prepare and store food safely.
7. Do not use tobacco in any form.
138
Table A3:
ACS Advice For Reducing the Risk of Breast Cancer
1. Engage in vigorous physical activity at least 4 hours a week.
2. Avoid or limit your intake of alcohol to no more than one drink per day.
3. Reduce lifetime weight gain through the combination of limiting your calories and
exercising regularly.
139
Table A4:
Measurement Timeline
Time 1:
Mid-treatment
Time 2:
2 weeks posttreatment
Time 3:
3 months posttreatment
Demographics and health
information (self-report)
Demographics and health
information (self-report)
Demographics and health
information (self-report)
Disease and treatment
information (medical
records)
Disease and treatment
information (medical
records)
Disease and treatment
information (medical
records)
Center for
Epidemiological StudiesDepression Scale
(CES-D)
Center for
Epidemiological StudiesDepression Scale
(CES-D)
Center for
Epidemiological StudiesDepression Scale
(CES-D)
PRIME-MD anxiety
scale
PRIME-MD anxiety
scale
PRIME-MD anxiety
scale
Impact of Events Scale
(IES)
Impact of Events Scale
(IES)
Impact of Events Scale
(IES)
Concerns About
Recurrence Scale
(CARS)
Concerns About
Recurrence Scale
(CARS)
Concerns About
Recurrence Scale
(CARS)
Cancer-specific stressors
items
Cancer-specific stressors
items
Cancer-specific stressors
items
Memorial Symptom
Assessment Scale
(MSAS)
Memorial Symptom
Assessment Scale
(MSAS)
Memorial Symptom
Assessment Scale
(MSAS)
Medical Outcomes Study
Short Form 36 version 2
(SF-36v2)
Medical Outcomes Study
Short Form 36 version 2
(SF-36v2)
Medical Outcomes Study
Short Form 36 version 2
(SF-36v2)
Illness Perceptions
Questionnaire (IPQ-R)
Illness Perceptions
Questionnaire (IPQ-R)
Illness Perceptions
Questionnaire (IPQ-R)
Causal attributions
Causal and recurrence
prevention attributions
Causal and recurrence
prevention attributions
Perceived control items
Perceived control items
Perceived control items
Perceived risk of
recurrence items
Perceived risk of
recurrence items
Health practices and
Behavior/ other behavior items
Behavior Diet measure
Health practices and
other behavior items
Health practices and
other behavior items
Diet measure
Diet measure
Change
Behavior change
questionnaire
Behavior change
questionnaire
Domain:
General
Distress
HealthRelated
Quality
of Life
Cancer
Beliefs
Behavior change
questionnaire
140
Table A5:
Sample Demographics
Percent of sample
Education
Less than high school
High school graduate
Some college
College graduate
Post-graduate degree
4.5
27.3
22.7
21.6
23.9
Employment
Work full-time
Work part-time
Student
Homemaker
Disabled
Retired
45.5
17.0
1.1
15.9
2.3
18.2
Income
<$10,000
$10,001-$25,000
$25,001-$40,000
$40,001-$55,000
$55,001-$70,000
>$70,000
5.5
15.1
12.3
13.7
19.2
34.2
Caucasian
African American
Asian
Native American
Other
93.2
2.3
1.1
1.1
2.3
Ethnicity
Relationship Status
Married or living with partner
Single
Divorced or separated
Widowed
73.8
5.7
12.5
7.9
141
Table A6:
Sample Disease and Treatment Characteristics
Percent of sample
Cancer stage
0
I
II
III
5.7
33.3
47.1
13.8
Nodal Status
Positive
Negative
44.8
55.2
Estrogen Receptor Status
Positive
Negative
77.3
22.7
Treatment a
Mastectomy
Lumpectomy
Chemotherapy
Neoadjuvant chemotherapy
Radiation therapy
Hormonal therapy
Herceptin
27.3
79.5
71.6
11.4
86.4
77.0
3.4
a
Treatment categories are not mutually exclusive.
142
Table A7:
Mean Distress and Quality of Life Scores Over Time
Baseline
3 Weeks
3 Months
Contrasts
11.08
(8.66)
9.82
(8.84)
9.14
(7.76)
Intrusion
11.58
(8.14)
9.88
(7.47)
9.00
(7.45)
Avoidance
11.08
(7.08)
9.99
(6.69)
10.06
(7.44)
PRIME-MD Anxiety
11.18
(2.86)
11.38
(3.09)
10.84
(3.35)
CARS
11.78
(5.06)
12.40
(5.05)
11.50
(4.64)
Total
symptoms*
0.69
(0.47)
0.51
(0.34)
0.50
(0.31)
Global distress
index
0.93
(0.60)
0.79
(0.57)
0.79
(0.52)
Physical
Function*
74.37
(23.02)
78.59
(20.78)
81.88
(21.27)
T1 < T2
T1 < T3
Role-Physical*
60.85
(25.88)
71.63
(24.19)
77.45
(23.95)
T1 < T2
T1 < T3
Bodily Pain
82.05
(19.54)
80.77
(18.57)
81.01
(18.64)
Vitality*
53.05
(19.90)
58.09
(20.60)
61.68
(21.14)
T1 < T3
Social
Function*
79.55
(22.28)
87.66
(16.79)
86.05
(20.73)
T1 < T2
Role-Emotional
83.33
(24.35)
87.18
(18.79)
88.77
(18.35)
Distress Measures
CES-D
IES
Quality of Life Measures
MSAS
SF-36
T1 > T2
T1 > T3
Note. Standard deviations are enclosed in parentheses. Contrasts were performed only if
the overall F test was significant at p < .05. T1 = baseline; T2 = 3 weeks posttreatment; T3 = 3 months post-treatment.
*Overall F test was significant at p < .05.
143
Table A8:
Mean Ratings of Sources of Stress Over Time
Sources of Stress
Baseline
3 Weeks
3 Months
Side effects or physical problems related to cancer
and treatment
2.86
(1.10)
2.35
(1.19)
2.54
(1.27)
Fear of a cancer recurrence
2.56
(1.25)
2.58
(1.14)
2.48
(1.16)
Trying to get back to normal life now that treatment
has ended
2.46
(1.34)
2.25
(1.42)
Creating a “new normal” now that treatment has
ended
2.22
(1.25)
2.22
(1.30)
2.15
(1.22)
2.18
(1.97
2.18
(1.20)
2.12
(1.24)
Worry about the impact of cancer on my family
2.75
(1.37)
Fear that my cancer is not gone
Feeling unsure what to do for my health or to
prevent a cancer recurrence
2.09
(1.17)
2.16
(1.24)
1.96
(1.09)
Financial problems
2.09
(1.14)
1.93
(1.15)
1.99
(1.25)
My emotions or emotional well-being
2.14
(1.09)
1.91
(0.96)
1.84
(0.98)
Concerns about ability to fulfill responsibilities
at work or home
2.24
(1.12)
1.79
(1.08)
1.72
(1.03)
Feeling like I have lost a “safety net” now that
treatment has ended
1.59
(0.91)
1.54
(0.88)
Not seeing oncologist and health care staff
regularly now that treatment has ended
1.34
(0.81)
1.39
(0.75)
Relationship problems
1.50
(1.04)
1.42
(0.87)
1.24
(0.46)
Not getting the assistance I would like from
family or friends
1.39
(0.73)
1.38
(0.80)
1.33
(0.63)
Not getting the emotional support I would like
from family or friends
1.37
(0.68)
1.31
(0.71)
1.25
(0.53)
Note. Standard deviations are enclosed in parentheses. Ratings ranged from 1 “not at all”
to 5 “very much.”
144
Table A9:
Post-Treatment Behavior Changes
3 Weeks Post-Treatment
%
3 Months Post-Treatment
%
Perform breast self-exam more
40.0
Perform breast self-exam more
36.2
Eat more fruits and vegetables
30.8
Eat more fruits and vegetables
30.0
Use dietary supplements more
30.3
Do more moderate physical activity
28.6
Drink alcohol less a
28.9
Use dietary supplements more
28.6
a
Pray or meditate more
24.7
Drink alcohol less
27.5
Avoid stressful situations more often
24.7
Spend more time with friends
26.1
Spend more time with family
22.1
Pray or meditate more
21.7
Spend more time with friends
20.5
Spend more time with family
20.3
Do more moderate physical activity
15.4
Choose no- or low-fat foods more
18.6
Do less household work
14.3
Avoid stressful situations more often
15.9
Do less work outside of the home
13.3
Journal/write about experiences more
13.2
Spend more time engaging in hobbies
12.8
Do less household work
12.9
Engage in more food safety practices
11.8
Engage in more food safety practices
11.6
Journal/write about experiences more
11.7
Smoke less b
11.6
Do more relaxing activities
11.5
Spend more time engaging in hobbies
11.4
Smoke less b
9.2
Do more relaxing activities
10.0
Get enough sleep more often
7.9
Get enough sleep more often
10.0
Choose no- or low-fat foods more
5.2
Do more vigorous physical activity
8.7
Do more vigorous physical activity
3.9
Do less work outside of the home
8.7
Attend religious services more
2.6
Attend religious services more
5.8
Note. We assessed both increases and decreases in each behavior, but only report the
positive or adaptive behavior change direction (e.g., increased fruit/vegetable
consumption rather than decreased fruit/vegetable consumption). When the
adaptiveness of the behavior change was not evident (e.g., is it better to work more or
less outside of the home?), the change direction that more women reported was used
(e.g., more women reported working less outside of the home).
a
36% of participants denied any alcohol use; of those who reported drinking any alcohol
prior to diagnosis, 44.8% reported drinking less 3 weeks post-treatment and 43.1%
reported drinking less 3 months post-treatment.
b
Of those who smoked prior to diagnosis, 58.3% reported they had decreased or quit
smoking 3 weeks post-treatment and 88.9% decreased or quit smoking 3 months posttreatment.
145
Table A10:
Means and Percentages of Health and Psychosocial Behaviors
3 Weeks
3 Months
Fat consumption score
a
19.31
(8.32)
M and SD
19.18 a
(7.38)
Fruit and vegetable consumption score
14.21 b
(5.38)
14.62 b
(5.30)
11.48 c
(10.28)
12.25 c
(13.07)
Alcohol use (drinks/week)
1.16
(2.02)
1.18
(1.92)
Sleep (hours/night)
6.76
(1.35)
6.65
(1.31)
Breast self-exam score
2.31 d
(0.90)
2.10 d
(0.81)
Physical activity score
Percent
Smokes cigarettes
8.1
7.1
Attends a support group
6.4
8.6
Attends psychotherapy or counseling
3.9
1.4
Uses complementary therapies
15.4
17.4
Started a new activity
12.8
27.5
Note. Standard deviations are enclosed in parentheses.
a
Scores indicate 32.3% and 32.3% of calories from fat.
b
Scores indicate 4.5 and 4.6 daily servings.
c
Scores are based on a formula that incorporates frequency and amount of time spent in
moderate and vigorous physical activity.
d
A score of 2 is equivalent to monthly self-exam, with scores higher or lower indicating
more or less frequent self-exam.
146
Table A11:
Physical Activity 3 Weeks Post-Treatment as a Predictor of CARS Recurrence
Worry 3 Months Post-Treatment
Variable
Step 1
R
ΔR2
.390
.152
β (final
model)
p
.008
Age
-.018
.834
Length of treatment
.077
.316
Step 2*
.812
.507
CARS score at 3 weeks*
Step 3
Physical activity at 3 weeks
*Significant at p < .05.
.000
.780
.822
.017
.000
.089
.131
.089
147
Table A12:
Mean Scores on Measures of Common-Sense Models of Breast Cancer
Illness
Perception
Questionnaire
Perceived
Control Scale
Perceived Risk
of Recurrence
Acute vs. Chronic
Timeline a
Potential
Score
Range
6-30
3 Weeks
3 Months
14.63
(5.60)
15.19
(4.51)
15.91
(4.97)
Cyclical Timeline
4-20
10.11
(3.28)
9.69
(2.81)
9.64
(2.77)
Consequences
6-30
21.89
(3.87)
21.33
(3.68)
21.03
(3.84)
Emotional
Representations
6-30
17.41
(4.69)
17.05
(4.57)
16.88
(4.83)
Illness Coherence
5-25
19.31
(3.92)
19.16
(3.36)
19.22
(3.62)
Personal Control
6-30
22.25
(3.93)
20.88
(3.91)
21.43
(3.88)
Treatment Control
5-25
19.98
(2.65)
19.97
(2.39)
19.63
(2.96)
Total Score
0-24
13.51
(4.30)
13.10
(3.86)
13.64
(3.88)
Treatment success
0-4
2.47
(1.07)
2.16
(1.11)
2.33
(1.09)
Cancer course
0-4
1.86
(1.14)
1.79
(1.10)
1.86
(1.12)
Getting information 0-4
3.40
(0.84)
3.23
(0.79)
3.33
(0.91)
Side effects
0-4
1.95
(1.20)
1.89
(1.07)
1.89
(1.05)
Cause
0-4
0.81
(1.03)
0.87
(1.04)
0.92
(0.99)
Decisions
0-4
3.01
(1.03)
3.13
(0.83)
3.30
(0.86)
8.89
(3.51)
9.07
(3.42)
3-18
Note. Standard deviations are enclosed in parentheses.
a
Baseline
Higher scores indicate a more chronic timeline.
148
Table A13:
Causal Attributions
Hormones
Baseline
3 weeks
post-treatment
3 months
post-treatment
M and SD %
2.63
78.1%
(1.35)
M and SD %
2.81
84.5%
(1.24)
M and SD %
2.97
88.0%
(1.13)
Environmental toxins
or hazards
2.26
(1.29)
72.3%
2.32
(1.33)
73.1%
2.41
(1.27)
76.4%
Genetics or heredity
2.13
(1.59)
60.9%
2.24
(1.41)
70.5%
2.41
(1.53)
69.2%
Diet or eating habits
1.93
(1.24)
63.2%
1.87
(1.15)
67.9%
2.00
(1.27)
62.7%
Stress or worry
1.67
(1.51)
51.6%
1.62
(1.33)
54.6%
1.76
(1.40)
54.4%
Aging
1.62
(1.14)
57.4%
1.62
(1.17)
53.3%
1.64
(1.24)
57.6%
God’s will
1.50
(1.58)
46.5%
1.43
(1.53)
43.4%
1.32
(1.47)
38.4%
Lack of exercise
1.41
(1.31)
47.1%
1.28
(1.31)
38.4%
1.31
(1.21)
44.9%
Chance or bad luck
1.33
(1.38)
44.2%
1.17
(1.26)
36.8%
1.15
(1.34)
35.8%
Mental attitude (e.g.,
thinking negatively)
1.01
(1.32)
31.4%
0.92
(1.26)
28.6%
0.77
(1.03)
24.6%
Use of alcohol or
tobacco
0.89
(1.32)
29.4%
0.74
(1.27)
20.6%
0.71
(1.17)
20.1%
An injury
0.46
(0.96)
14.9%
0.59
(1.07)
19.3%
0.50
(0.91)
17.7%
A germ or virus
0.45
(0.85)
13.0%
0.61
(1.03)
17.1%
0.55
(1.01)
13.6%
Poor medical care in
my past
0.42
(0.85)
13.1%
0.46
(0.92)
14.4%
0.38
(0.68)
7.7%
Note. All factors were rated from 0 to 4. Standard deviations are enclosed in parentheses.
Percentages represent the percent of women rating each factor as at least somewhat
important in causing their cancer.
149
Table A14:
Recurrence Prevention Beliefs
3 weeks
post-treatment
3 months
post-treatment
M and SD
3.65
(0.87)
%
94.8%
M and SD
3.51
(0.99)
%
92.7%
Having a positive attitude
3.43
(1.02)
92.1%
3.32
(0.98)
95.7%
Eating a healthy diet
3.31
(0.89)
94.8%
3.42
(0.86)
95.9%
Using medication
3.15
(1.21)
89.3%
3.15
(1.35)
86.4%
Exercise
3.15
(0.94)
95.9%
3.05
(1.09)
89.7%
Prayer
2.60
(1.41)
75.4%
2.50
(1.35)
77.3%
Reducing stress in my life
2.39
(1.18)
77.7%
2.37
(1.31)
74.7%
Decrease/quit use of alcohol or
tobacco
2.26
(1.67)
66.8%
1.81
(1.64)
51.6%
God’s will
1.68
(1.58)
52.7%
1.56
(1.62)
45.6%
Chance or luck
1.26
(1.18)
46.0%
1.22
(1.39)
38.2%
Using complementary therapies
1.21
(1.12)
41.3%
1.22
(1.22)
43.3%
Medical checkups and screenings
Note. All factors were rated from 0 to 4. Standard deviations are enclosed in parentheses.
Percentages represent the percent of women rating each factor as at least somewhat
important in preventing a recurrence of their cancer.
150
Table A15:
Factor Loadings for Behavioral/Psychological Attributions Factors
Baseline
3 weeks
3 months
post-treatment post-treatment
Diet
.693
.774
.881
Lack of exercise
.688
.652
.630
Mental attitude
.814
.593
.613
Stress or worry
.579
.622
.597
Exercise
.715
.786
Having a positive attitude
.680
.730
Eating a healthy diet
.665
.717
Reducing stress
.551
.722
Causal attributions:
Recurrence prevention beliefs:
151
Table A16:
Chronicity Beliefs 3 Weeks Post-Treatment Predict Fat Consumption
3 Months Post-Treatment
Variable
Step 1
R
ΔR2
.093
.009
Age*
Length of treatment
Step 2*
.783
.026
-.139
.098
.000
.027
IPQ acute vs. chronic timeline
at 3 weeks post-treatment*
Note. Higher scores indicate more chronic timeline.
*Significant at p < .05.
.193
.850
.800
p
.766
.604
Fat consumption at 3 weeks*
Step 3*
β (final
model)
.000
.041
.178
.041
152
Table A17:
Attribution-Behavior Relationships Tested
Behavioral Outcome(s) Examined
Causal attribution:
Diet
Change in fruit and vegetable consumption
Fruit and vegetable consumption
Fat consumption
Lack of exercise
Change in physical activity
Physical activity
Alcohol or tobacco use
Change in alcohol use
Alcohol use
Stress or worry
Change in avoidance of stressful situations
Change in relaxing activities
Recurrence prevention belief:
Eating a healthy diet
Change in fruit and vegetable consumption
Fruit and vegetable consumption
Fat consumption
Exercise
Change in physical activity
Physical activity
Decrease or quit use of alcohol
or tobacco
Change in alcohol use
Alcohol use
Reducing stress in my life
Change in avoidance of stressful situations
Change in relaxing activities
153
Table A18:
Behavioral and Psychological Attribution at Baseline Predicts Alcohol Use 3
Weeks Post-Treatment
Variable
Step 1
R
ΔR2
.098
.010
β (final
model)
p
.715
Age
-.059
.338
Length of treatment
-.049
.435
Step 2*
.860
.730
Alcohol use at baseline*
Step 3*
Behavioral and
psychological causal
attribution at baseline*
*Significant at p < .05.
.000
.866
.869
.016
.000
.043
-.127
.043
154
Table A19:
Attributing Cancer to Substance Use at Baseline Predicts Alcohol Use
3 Weeks Post-Treatment
Variable
Step 1
R
ΔR2
.080
.006
β (final
model)
p
.805
Age
-.066
.288
Length of treatment
-.064
.310
Step 2*
.859
.731
Alcohol use at baseline*
Step 3*
Alcohol and tobacco causal
attribution at baseline*
*Significant at p < .05.
.000
.918
.872
.023
.000
.015
-.166
.015
155
Table A20:
Personal Control and Fat Consumption Interact to Predict IES Intrusion at
3 Months Post-Treatment
Variable
Step 1*
R
ΔR2
.323
.104
β (final
model)
p
.031
Age*
-.329
.008
Length of treatment
-.071
.556
Step 2
.365
.029
IPQ personal control
Fat consumption
Step 3*
IPQ personal control X Fat
consumption*
*Significant at p < .05.
.444
.369
-.144
.243
.003
.982
.064
.033
.260
.033
156
Table A21:
Chronicity Belief and Change in Avoidance of Stress Interact to Predict IES
Intrusion at 3 Months Post-Treatment
Variable
Step 1*
R
ΔR2
.314
.098
β (final
model)
p
.038
Age*
-.244
.024
Length of treatment
-.139
.191
Step 2*
.549
.204
.000
IPQ acute vs. chronic
timeline*
.611
.000
Change in avoidance of
stress*
.226
.039
Step 3*
IPQ acute vs. chronic
timeline X Change in
avoidance of stress*
*Significant at p < .05.
.619
.082
.006
.396
.006
157
Table A22:
Chronicity Belief and Physical Activity Interact to Predict PRIME-MD Anxiety
at 3 Months Post-Treatment
Variable
Step 1*
R
ΔR2
.334
.111
Age*
Length of treatment
Step 2*
.479
β (final
model)
p
.024
-.226
.047
.155
.193
.118
.013
IPQ acute vs. chronic
timeline*
.309
.009
Physical activity
.150
.194
Step 3*
IPQ acute vs. chronic
timeline X Physical activity*
*Significant at p < .05.
.529
.051
.044
.234
.044
158
Table A23:
Chronicity Belief and Physical Activity at 3 Weeks Post-Treatment Interact to Predict
PRIME-MD Anxiety at 3 Months Post-Treatment
Variable
Step 1*
R
ΔR2
.340
.116
Age
Length of treatment
Step 2*
.565
-.066
.579
.117
.299
.000
.530
.598
p
.030
.203
PRIME-MD anxiety at 3 weeks*
Step 3
β (final
model)
.039
.000
.205
IPQ acute vs. chronic timeline at 3
weeks
.139
.232
Physical activity at 3 weeks
.140
.199
Step 4*
IPQ acute vs. chronic timeline X
Physical activity at 3 weeks*
*Significant at p < .05.
.638
.049
.041
.231
.041
159
Table A24:
Causal Attributions and Change in Avoidance of Stress Interact to Predict CES-D
Depression at 3 Weeks Post-Treatment
Variable
Step 1*
R
ΔR2
.349
.122
Age*
Length of treatment
Step 2*
.496
Change in avoidance of stress
Behavioral/psychological
attribution X Change in avoidance
of stress*
*Significant at p < .05.
.560
p
.009
-.316
.003
.124
.246
.124
Behavioral/psychological
attribution*
Step 3*
β (final
model)
.005
.500
.000
-.179
.106
.068
.011
.291
.011
160
Table A25:
Causal Attributions and Frequency of Breast Self-Exam Interact to Predict PRIME-MD
Anxiety at 3 Weeks Post-Treatment
Variable
Step 1*
R
ΔR2
.373
.139
Age*
Length of treatment
Step 2
.405
Frequency of breast self-exam
Behavioral/psychological
attribution X Frequency of breast
self-exam*
*Significant at p < .05.
.479
p
.006
-.363
.002
.165
.166
.025
Behavioral/psychological
attribution
Step 3*
β (final
model)
.372
.129
.247
-.052
.647
.065
.021
-.278
.021
161
Table A26:
Causal Attributions and Change in Avoidance of Stressful Situations Interact to
Predict IES Intrusion at 3 Months Post-Treatment
Variable
Step 1*
R
ΔR2
.305
.093
β (final
model)
p
.044
Age*
-.274
.016
Length of treatment
-.001
.996
Step 2*
.445
.105
Behavioral/psychological
attribution
Change in avoidance of stress*
Step 3*
Behavioral/psychological
attribution X Change in avoidance
of stress*
*Significant at p < .05.
.564
.022
-.282
.084
.356
.002
.120
.002
-.503
.002
162
Table A27:
Causal Attributions and Alcohol Use Interact to Predict CES-D Depression at 3
Months Post-Treatment
Variable
Step 1*
R
ΔR2
.303
.092
Age
Length of treatment*
Step 2
.379
Alcohol use
Behavioral/psychological
attribution X Alcohol use*
*Signifcant at p < .05.
.449
p
.046
-.186
.113
.295
.019
.052
Behavioral/psychological
attribution
Step 3*
β (final
model)
.164
.193
.133
-.048
.709
.058
.039
-.263
.039
163
Table A28:
Causal Attributions and Change in Fruit and Vegetable Consumption at 3 Weeks PostTreatment Interact to Predict IES Intrusion at 3 Months Post-Treatment
Variable
Step 1*
R
ΔR2
.316
.100
β (final
model)
p
.031
Age
-.054
.568
Length of treatment
-.061
.486
Step 2*
.678
.360
IES Intrusion at 3 weeks*
Step 3*
.698
.722
.061
Behavioral/psychological attribution at 3
weeks*
Change in fruit and vegetable consumption
at 3 weeks*
Step 4*
Behavioral/psychological attribution X
Change in fruit and vegetable consumption
at 3 weeks*
*Significant at p < .05.
.000
.742
.000
.023
-.205
.039
.234
.010
.030
.045
-.201
.045
164
Table A29:
Causal Attributions and Change in Avoidance of Stressful Situations at 3 Weeks PostTreatment Interact to Predict IES Intrusion at 3 Months Post-Treatment
Variable
Step 1*
R
ΔR2
.305
.093
β (final
model)
p
.042
Age
-.032
.740
Length of treatment
-.140
.125
Step 2*
.673
.360
IES Intrusion at 3 weeks*
Step 3
.756
.704
.043
Behavioral/psychological attribution at 3
weeks*
Change in avoidance of stressful situations
at 3 weeks*
Step 4*
Behavioral/psychological attribution X
Change in avoidance of stressful situations
at 3 weeks*
*Significant at p < .05.
.000
.744
.000
.079
-.278
.020
.206
.028
.058
.007
-.296
.007
165
Table A30:
Causal Attributions and Fat Consumption at 3 Weeks Post-Treatment Interact to Predict
IES Intrusion at 3 Months Post-Treatment
Variable
Step 1*
R
ΔR2
.306
.093
β (final
model)
p
.041
Age
-.070
.494
Length of treatment
-.063
.489
Step 2*
.674
.361
IES Intrusion at 3 Weeks*
Step 3
.649
.681
.009
Behavioral/psychological attribution at 3
weeks
Fat consumption at 3 weeks
Step 4*
Behavioral/psychological attribution X Fat
consumption at 3 weeks*
*Significant at p < .05.
.000
.714
.000
.598
-.060
.516
.072
.460
.047
.019
.224
.019
166
Table A31:
Causal Attributions and Physical Activity at 3 Weeks Post-Treatment Interact to Predict
PRIME-MD Anxiety at 3 Months Post-Treatment
Variable
Step 1*
R
ΔR2
.337
.114
Age
Length of treatment
Step 2*
.572
Physical activity at 3 weeks
Behavioral/psychological attribution X
Physical activity at 3 weeks*
*Significant at p < .05.
.685
.086
.432
.000
.031
Behavioral/psychological attribution at 3
weeks
Step 4*
-.047
.571
.598
.635
p
.028
.213
PRIME-MD anxiety at 3 weeks*
Step 3
β (final
model)
.000
.266
-.084
.450
.083
.467
.045
.046
-.231
.046
167
Table A32:
Beliefs about Recurrence Prevention and Physical Activity Interact to Predict PRIMEMD Anxiety at 3 Weeks Post-Treatment
Variable
Step 1*
R
ΔR2
.360
.129
β (final
model)
p
.011
Age*
-.332
.006
Length of treatment
-.042
.713
Step 2
.380
.015
.576
Behavioral/psychological
recurrence prevention belief
-.060
.613
Physical activity
-.091
.419
Step 3*
Behavioral/psychological
recurrence prevention belief X
Physical Activity*
*Significant at p < .05.
.494
.100
.006
.318
.006
168
APPENDIX B
FIGURES
Figure B1. Model of adjustment during the 3 months following the end of breast
cancer treatment
Objective 1 involves characterizing distress and health-related quality of life (Box C) at
all time points. Objective 2 includes characterizing behavioral coping strategies and
behavior changes (Box B) at all post-treatment time points and examining concurrent
and prospective relationships between Boxes B and C. Objective 3 includes
characterizing common-sense representations of breast cancer (Box A) at all time
points and examining concurrent and prospective relationships between Boxes A and
B. Objective 4 will examine whether boxes A and B interact to concurrently and
prospectively predict distress in Box C. Box D represents important biological
variables that may influence Box C and possibly Boxes A and B as well and will
therefore be entered as covariates in analyses.
3 Months Post-Treatment
IPQ Personal Control
3 Weeks Post-Treatment
• Less fat consumption
No Significant Outcomes
Relationships that are significant at p < .05 are marked with an asterisk.
Figure B2. Marginally significant relationships between personal control and behavioral outcomes over time
Baseline
• More likely to increase
avoidance of stress
• More fruit/vegetable
consumption
No Significant Outcomes
No Significant Outcomes
169
3 Months Post-Treatment
IPQ Chronic vs. Acute Timeline
3 Weeks Post-Treatment
•!More alcohol use*
•!Less likely to increase breast
self-exam*
•!More alcohol use*
Relationships that are significant at p < .05 are marked with an asterisk.
Figure B3. Significant and marginally significant relationships between beliefs about the chronicity of one’s cancer and behavioral
outcomes over time
Baseline
•!More fat consumption*
•!More frequent breast selfexam
• Less likely to increase breast
self-exam*
• More likely to decrease
alcohol use*
No significant outcomes
170
3 Weeks Post-Treatment
3 Months Post-Treatment
•!More likely to increase fruit/
vegetable consumption*
• More likely to decrease alcohol
use*
• Less alcohol use
• More likely to increase
avoidance of stress
•!Less physical activity*
•!Less alcohol use
Relationships that are significant at p < .05 are marked with an asterisk.
Figure B4. Significant and marginally significant relationships between attributing cancer to behavioral and psychological causes and
behavioral outcomes over time
Behavioral and Psychological Causal Attribution
Baseline
•!More likely to increase fruit/
vegetable consumption*
•!More likely to increase fruit/
vegetable consumption*
•!Less alcohol use*
•!More likely to increase fruit/
vegetable consumption*
171
3 Months Post-Treatment
Diet Causal Attribution
3 Weeks Post-Treatment
•!More likely to increase fruit/
vegetable consumption*
No significant outcomes
Relationships that are significant at p < .05 are marked with an asterisk.
Figure B5. Significant and marginally significant relationships between attributing cancer to diet and behavioral outcomes over time
Baseline
No significant outcomes
•!More likely to increase fruit/
vegetable consumption*
•!More likely to increase fruit/
vegetable consumption*
172
3 Months Post-Treatment
Exercise Causal Attribution
3 Weeks Post-Treatment
• Less physical activity
• More likely to increase
physical activity*
Relationships that are significant at p < .05 are marked with an asterisk.
Figure B6. Significant and marginally significant relationships between attributing cancer to lack of exercise and behavioral outcomes
over time
Baseline
• Less physical activity*
No significant outcomes
No significant outcomes
173
3 Months Post-Treatment
•!More likely to increase fruit/
vegetable consumption*
Relationships that are significant at p < .05 are marked with an asterisk.
Figure B7. Significant and marginally significant relationships between believing that behavioral and psychological factors can
prevent a recurrence and behavioral outcomes over time
Behavioral and Psychological Recurrence Prevention Belief
3 Weeks Post-Treatment
•!More likely to increase fruit/
vegetable consumption*
• More likely to decrease alcohol
use
• More likely to increase
avoidance of stress*
•!Less fat consumption*
•!More likely to increase fruit/
vegetable consumption*
174
3 Months Post-Treatment
•!More likely to decrease
alcohol use*
Relationships that are significant at p < .05 are marked with an asterisk.
Figure B8. Significant and marginally significant relationships between believing that decreasing or quitting alcohol or tobacco use
can prevent a recurrence and behavioral outcomes over time
Alcohol/Tobacco Recurrence Prevention Belief
3 Weeks Post-Treatment
•!More likely to decrease
alcohol use*
No significant outcomes
175
3 Months Post-Treatment
3 Weeks Post-Treatment
Relationships that are significant at p < .05 are marked with an asterisk.
Figure B9. Significant and marginally significant relationships between believing that decreasing stress in one’s life can prevent a
recurrence and behavioral outcomes over time
Stress Recurrence Prevention Belief
•!More likely to increase
avoidance of stress*
•!More likely to increase
relaxing activities
•!More likely to increase
avoidance of stress*
•!More likely to increase
relaxing activities*
No significant outcomes
176
High and low points on the x-axis represent scores one standard deviation above and below the mean on personal control, and highand low-fat diet lines represent scores one standard deviation above and below the mean on fat consumption.
Figure B10. Interaction between personal control and fat consumption in predicting intrusion at 3 months post-treatment, β =
.26, p = .03
177
High and low points on the x-axis represent scores one standard deviation above and below the mean on acute versus chronic timeline,
with high scores indicating belief in a more chronic time course and low scores indicating belief in a more acute time course of
one’s cancer.
Figure B11. Interaction between chronicity belief and change in avoidance of stressful situations in predicting intrusion at 3 weeks
post-treatment, β = .40, p = .006
178
High and low points on the x-axis represent scores one standard deviation above and below the mean on acute versus chronic timeline,
with high scores indicating belief in a more chronic time course and low scores indicating belief in a more acute time course of
one’s cancer. High and low physical activity lines represent scores one standard deviation above and below the mean on physical
activity.
Figure B12. Interaction between chronicity belief and physical activity at 3 weeks post-treatment in predicting anxiety at 3 months
post-treatment, β = .23, p = .04
179
High and low points on the x-axis represent scores one standard deviation above and below the mean on behavioral/psychological
attribution.
Figure B13. Interaction between attributing cancer to behavioral and psychological causes and change in avoidance of stressful
situations in predicting depression at 3 weeks post-treatment, β = .29, p = .01
180
High and low points on the x-axis represent scores one standard deviation above and below the mean on behavioral/psychological
attribution, and more and less frequent breast self-exam lines represent scores one standard deviation above and below the mean
on frequency of breast self-exam.
Figure B14. Interaction between attributing cancer to behavioral and psychological causes and breast self-exam in predicting anxiety
at 3 weeks post-treatment, β = -.28, p = .02
181
High and low points on the x-axis represent scores one standard deviation above and below the mean on behavioral/psychological
attribution.
Figure B15. Interaction between attributing cancer to behavioral and psychological causes and change in avoidance of stressful
situations in predicting intrusion at 3 months post-treatment, β = -.50, p = .002
182
High and low points on the x-axis represent scores one standard deviation above and below the mean on behavioral/psychological
attribution, and high and low alcohol use lines represent scores one standard deviation above and below the mean on alcohol
consumption.
Figure B16. Interaction between attributing cancer to behavioral and psychological causes and alcohol use in predicting depression at
3 months post-treatment, β = -.26, p = .039
183
High and low points on the x-axis represent scores one standard deviation above and below the mean on behavioral/psychological
attribution.
Figure B17. Interaction between attributing cancer to behavioral and psychological causes and change in fruit/vegetable consumption
at 3 weeks post-treatment in predicting intrusion at 3 months post-treatment, β = -.20, p = .045
184
High and low points on the x-axis represent scores one standard deviation above and below the mean on behavioral/psychological
attribution, and high- and low-fat diet lines represent scores one standard deviation above and below the mean on fat consumption.
Figure B18. Interaction between attributing cancer to behavioral and psychological causes and fat consumption at 3 weeks posttreatment in predicting intrusion at 3 months post-treatment, β = .22, p = .019
185
High and low points on the x-axis represent scores one standard deviation above and below the mean on behavioral/psychological
attribution, and high and low physical activity lines represent scores one standard deviation above and below the mean on physical
activity.
Figure B19. Interaction between attributing cancer to behavioral and psychological causes and physical activity at 3 weeks posttreatment in predicting anxiety at 3 months post-treatment, β = -.23, p = .046
186
High and low points on the x-axis represent scores one standard deviation above and below the mean on behavioral/psychological
recurrence prevention belief, and high and low physical activity lines represent scores one standard deviation above and below the
mean on physical activity.
Figure B20. Interaction between believing that behavioral and psychological factors can prevent recurrence and physical activity in
predicting anxiety at 3 weeks post-treatment, β = .32, p = .006
187
188
APPENDIX C
SELECTED MEASURES
Sources of Stress
Not at all
A little bit
Somewhat
Moderately
Very much
Instructions: Below are some things that other cancer patients have said are a source of
worry or stress for them. Please rate to what extent the following things have been a
source of stress for you over the past week.
1. Side effects or physical problems related to cancer and
treatment……………………………………………………..
1
2
3
4
5
2. My emotions or emotional well-being……………………….
1
2
3
4
5
3. Not getting the assistance I would like from family or
friends………………………………………………………..
1
2
3
4
5
4. Not getting the emotional support I would like from family
or friends…………………………………………………….
1
2
3
4
5
5. Not seeing my oncologist and health care staff regularly now
that treatment has ended……………………………………..
1
2
3
4
5
6. Relationship problems……………………………………….
1
2
3
4
5
7. Feeling like I have lost a “safety net” now that treatment has
ended………………………………………………………...
1
2
3
4
5
8. Fear of a cancer recurrence…………………………………..
1
2
3
4
5
9. Feeling unsure of what I should do for my health or to
prevent a cancer recurrence now that treatment has ended.....
1
2
3
4
5
10. Trying to get back to normal life now that treatment has
ended………………………………………………………..
1
2
3
4
5
11. Creating a “new normal” now that treatment is over………
1
2
3
4
5
12. Fear that my cancer is not gone…………………………….
1
2
3
4
5
13. Financial problems…………………………………………
1
2
3
4
5
14. Concerns about my ability to fulfill my responsibilities at
work or at home…………………………………………….
1
2
3
4
5
15. Worry about the impact of my cancer on my family……….
1
2
3
4
5
16. Other (please specify):_____________________________
1
2
3
4
5
To what extent has this been a source of stress?
189
Behavior Changes
Instructions: Please circle whether you do the following less, about the same, or more as
compared to before you were diagnosed with cancer. If you did not do an activity before
cancer and do not do it now, circle N/A.
I do this
LESS
or have
stopped
1. Breast self-exam…………………………………………..
1
I do this
about
the
same
2
I do this
MORE or
have now
started
3
N/A
2. Smoke……………………………………………………..
1
2
3
N/A
3. Drink alcohol……………………………………………...
1
2
3
N/A
4. Vigorous physical activity (running, swimming, etc.)……
1
2
3
N/A
5. Moderate physical activity (walking, etc.)………………..
1
2
3
N/A
7. Eat no or low-fat foods……………………………………
1
2
3
N/A
8. Eat fruits and vegetables…………………………………..
1
2
3
N/A
9. Take dietary supplements (vitamins, etc.)………………...
1
2
3
N/A
10. Use complimentary/alternative treatments (acupuncture,
massage, herbs, etc.)…………………………………………
1
2
3
N/A
11. Get enough sleep…………………………………………
1
2
3
N/A
12. Pray or meditate………………………………………….
1
2
3
N/A
13. Attend religious services or activities……………………
1
2
3
N/A
14. Work outside of home…………………………………...
1
2
3
N/A
15. Household work (cleaning, yard work, etc.)……………..
1
2
3
N/A
16. Spend time with family…………………………………..
1
2
3
N/A
17. Spend time with friends………………………………….
1
2
3
N/A
18. Engage in enjoyable hobbies or activities……………….
1
2
3
N/A
19. Attend psychotherapy or counseling…………………….
1
2
3
N/A
20. Relaxation exercises (e.g. self-hypnosis, yoga, tai chi)…
1
2
3
N/A
21. Relaxing activities (e.g. bubble baths, listening to music)
1
2
3
N/A
22. Avoid stressful situations………………………………...
1
2
3
N/A
23. Journaling/writing about my experiences………………..
1
2
3
N/A
24. Food safety practices (e.g. washing produce, cooking
meat well)……………………………………………………
1
2
3
N/A
Please list any other changes you consider important and rate them below.
25.______________________________________________
1
2
3
N/A
26.______________________________________________
1
2
3
N/A
190
Causal Attributions
Instructions: Sometimes people have ideas about factors that may have played a role in
the development of their cancer. We are most interested in your own views about factors
that may have caused your cancer rather than what others including doctors or family
may have suggested to you. Please rate how important you believe each of the following
factors were in the cause of your cancer from not at all important to very important.
Not at all
important
Very
important
1. Genetics or heredity…………………….
0
1
2
3
4
2. Diet or eating habits…………………….
0
1
2
3
4
3. Environmental toxins or hazards………..
0
1
2
3
4
4. Hormones……………………………….
0
1
2
3
4
5. Chance or bad luck……………………...
0
1
2
3
4
6. God’s will……………………………….
0
1
2
3
4
7. Stress or worry………………………….
0
1
2
3
4
8. An injury………………………………..
0
1
2
3
4
9. My mental attitude (e.g. thinking
negatively)……………………………...
0
1
2
3
4
10. Lack of exercise……………………….
0
1
2
3
4
11. Use of alcohol or tobacco……………..
0
1
2
3
4
12. Poor medical care in my past………….
0
1
2
3
4
13. Aging…………………………………..
0
1
2
3
4
14. A germ or virus………………………..
0
1
2
3
4
15. Other:__________________________
0
1
2
3
4
Please list in rank order the three most important factors that you now believe caused
your cancer. You may use any of the items from the box above, or you may have
additional ideas of your own.
1.
2.
3.
191
Recurrence Prevention Beliefs
Instructions: Sometimes people have ideas about what prevents the recurrence of their
cancer. We are most interested in your own views about factors that may prevent
recurrence of your cancer rather than what others including doctors or family may have
suggested to you. Please rate how important each of the following factors are in
preventing your cancer from recurring from not at all important to very important.
Not at all
important
Very
important
1. Medical checkups and screenings............
0
1
2
3
4
2. Eating a healthy diet…………………….
0
1
2
3
4
3. Using complementary therapies (e.g.
massage, acupuncture, herbs)……………...
0
1
2
3
4
4. Using medication (e.g. Tamoxifen)…….
0
1
2
3
4
5. Chance or luck………………………….
0
1
2
3
4
6. God’s will……………………………….
0
1
2
3
4
7. Reducing stress in my life………………
0
1
2
3
4
8. Prayer…………………………………...
0
1
2
3
4
9. Having a positive attitude………………
0
1
2
3
4
10. Exercise………………………………..
0
1
2
3
4
11. Decrease or quit use of alcohol or
tobacco………………………………...
0
1
2
3
4
12. Other:__________________________
0
1
2
3
4
Please list in rank order the three most important factors that you believe prevent the
recurrence of your cancer. You may use any of the items from the box above, or you may
have additional ideas of your own.
1.
2.
3.
192
Perceived Control
Instructions: The following questions ask about the amount of control you believe you
had or have over different aspects of your cancer and its treatment. Please circle the
number that best describes your belief.
No control
at all
Complete
control
1. Success of the treatment for your cancer…...
0
1
2
3
4
2. The long-term course of your cancer
(whether or not it will worsen or return in
the future)…………………………………..
0
1
2
3
4
3. Getting the information you need about your
cancer……………………………………….
0
1
2
3
4
4. The side effects of your cancer treatment….
0
1
2
3
4
5. The cause of your cancer…………………...
0
1
2
3
4
6. The decisions that are made regarding your
medical care and the treatment of your
cancer……………………………………….
0
1
2
3
4