Citation
Women's decision-making prior to enrollment in the STAR trial for breast cancer chemoprevention

Material Information

Title:
Women's decision-making prior to enrollment in the STAR trial for breast cancer chemoprevention
Creator:
Chasco, Emily E
Place of Publication:
Denver, Colo.
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
x, 100 leaves : ; 28 cm.

Thesis/Dissertation Information

Degree:
Master's ( Master of Arts)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Anthropology, CU Denver
Degree Disciplines:
Anthropology
Committee Chair:
Brett, John
Committee Members:
Scandlyn, Jean
Horton, Sarah

Subjects

Subjects / Keywords:
Decision making ( lcsh )
Clinical trials ( lcsh )
Breast -- Cancer -- Chemoprevention ( lcsh )
Women ( lcsh )
Breast -- Cancer -- Patients -- Attitudes ( lcsh )
Breast -- Cancer -- Chemoprevention ( fast )
Breast -- Cancer -- Patients -- Attitudes ( fast )
Clinical trials ( fast )
Decision making ( fast )
Women ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (M.A.)--University of Colorado Denver, 2009.
Bibliography:
Includes bibliographical references (leaves 90-100).
Statement of Responsibility:
by Emily E. Chasco.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
519503237 ( OCLC )
ocn519503237

Downloads

This item has the following downloads:


Full Text
WOMENS DECISION-MAKING PRIOR TO ENROLLMENT IN THE STAR
TRIAL FOR BREAST CANCER CHEMOPREVENTION
By
Emily E. Chasco
B.A., University of Michigan, 2004
A thesis submitted to the
University of Colorado Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts
Anthropology
2009


This Thesis for the Master of Arts
degree by
Emily E. Chasco
has been approved
by
Sarah Horton
ll/l


Chasco, Emily E. (M.A.Anthropology)
Women's Decision-Making Prior to Enrollment in the STAR Trial for Breast Cancer
Chemoprevention.
Thesis directed by Associate Professor John Brett
ABSTRACT
Despite advances in screening and treatment, breast cancer remains an
important public health issue in the United States. Until recently, early detection
provided the best chance for survival. However, chemoprevention, the use of
selective estrogen receptor moderators (SERMs)now offers a new avenue for
prevention of the disease in women at high-risk. Clinical trials are essential to
determine the safety and efficacy of chemoprevention agents. Yet factors that
influence women to participate in these trials are poorly understood. The purpose of
this study was to explore factors that influence the decision-making process of
women at high risk for breast cancer when considering enrollment in a
chemoprevention trial. Eligible participants were between the ages of 50-80, resided
in the Denver metropolitan area, were high-risk for breast cancer, and enrolled in the
Study of Tamoxifen and Raloxifene (STAR). STAR was a non-placebo arm
prevention trial conducted by the National Surgical Adjuvant Breast and Bowel
Project to test tamoxifen, the first agent approved by the FDA for breast cancer
prevention, against raloxifene, a second generation SERM. Taking an exploratory
qualitative approach, data was gathered through a short Demographic Questionnaire
and in-depth interviews conducted with seventeen women between September


2007 and September 2008. The results are discussed within a socio-ecological
framework that examines individual, interpersonal, institutional/organizational,
communityand structural levels of influence. Womens experiences while enrolled
in STAR are also discussed. Results indicate that physician recommendation,
altruistic motivations, having female offspring, and knowing a drug would be given
rather than a placebo were all important factors in participants1 decision-making.
While physician recommendation was important, its influence varied between
participants. Prior experience with breast cancer among family and friends was an
influential factor in shaping participants beliefs around breast cancer. Many women
mentioned concern regarding potential side effects, though it was not a deterrent to
participation. Social support at the interpersonal and community levels played a role
prior to enrollment in STAR and during the study as well. The most common
motivation for trial enrollment was the desire to help other women by furthering
scientific knowledge of breast cancer prevention.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication. *
Signed
John Brett


ACKNOWLEDGEMENT
I would like to extend my gratitude to my advisor, John Brett, for his continued
support and patience throughout this process. I would also like to thank those at the
Colorado Cancer Research Program here in Denver, Colorado, and in particular, Jane
Hajovsky, for their invaluable help- this project would not have been possible without
them. To my fellow students from the University of Colorado Denver and most
importantly to my family: thank you for your advice and encouragement over the past
two years.


TABLE OF CONTENTS
Figures...................................................... ix
Tables....................................................... x
CHAPTER
1.INTRODUCTION............................................... 1
2. BACKGROUND................................................ 4
Breast Cancer and Chemoprevention....................... 4
Study of Tamoxifen and Raloxifene....................... 9
Determining Risk....................................... 10
Womens Decision-Making When Considering
Enrollment in Chemoprevention Trials................... 13
3. THEORETICAL FRAMEWORK.....................................17
Socio-Ecological Approaches............................ 17
Socio-Ecological Approaches, Health Behavior, and Breast
Cancer Chemoprevention Trials.......................... 22
Social Support and Social Network Theories............. 26
4. METHODS & ANALYSIS....................................... 30
Colorado Cancer Research Program....................... 30
Research Design and Qualitative Methods................ 31
Data Collection.................................... 34
vi


Data Analysis.............................................. 37
5. MOTIVATIONS FOR ENROLLMENT....................................... 40
Demographic Characteristics of Participants.................... 40
Individual Factors.......................................... 42
Perception of Breast Cancer Risk........................... 42
Perception of Medical Research............................. 45
Perceived Benefits of and Barriers to Participation........ 47
Interpersonal Factors.......................................... 50
The Role of the Physician.................................. 50
Prior Experience with Breast Cancer........................ 54
Having Female Offspring/ Sisters........................... 56
Social Support............................................. 58
Institutional/ Organizational Factors.......................... 62
Placebo vs. Non-Placebo Trials............................. 62
Community Factors.............................................. 66
Altruistic Motivations..................................... 66
Social Structure, Policy, and Systems Factors.................. 69
Media...................................................... 69
Other Factors.............................................. 70


6. PARTICIPANT EXPERIENCES WHILE ENROLLED IN
THE STUDY OF TAMOXIFEN AND RALOXIFENE.....
74
Individual Factors................................. 74
Interpersonal and Community Factors................ 75
Nurses.......................................... 75
Social Events................................... 76
Physician....................................... 78
Organizational Factors............................. 78
7. CONCLUSION............................................ 80
Directions for Future Research..................... 84
APPENDIX
A. ACRONYMS............................................. 86
B. KEY CONCEPTS AND DEFINITIONS OF THE HEALTH
BELIEF MODEL AND THE THEORY OF REASONED
ACTION................................................ 87
C. DEMOGRAPHIC QUESTIONNAIRE............................ 88
D. INTERVIEW QUESTION GUIDE............................. 89
BIBLIOGRAPHY.................................................. 90
viii


LIST OF FIGURES
Figure
2.1 Five-Year Relative Survival Rates by Stage at Diagnosis...... 5
3.1 Socio-Ecological Levels of Influence......................... 21
5.1 Perceived Risks/Barriers and Benefits of Study
Participation................................................... 50
5.2 Types of Social Support Mentioned by Women
in this Study................................................... 59
5.3 Socio-Ecological Framework- Factors that
Influenced Womens Decision-Making While
Considering Enrollment in STAR.................................. 73
IX


LIST OF TABLES
Table
2.1 FDA-approved uses of SERMs........................................ 9
2.2 Risk factors and their impact on breast cancer................... 12
3.1 Descriptive characteristics of social networks................... 27
3.2 Types of social support.......................................... 28
4.1 Final code list.................................................. 38
5.1 Age range of participants........................................ 40
5.2 Number of female children and sisters
of participants................................................ 41
5.3 Drug assignment during STAR...................................... 64
6.1 Symptoms experienced by women during STAR........................ 75
B.l The Health Belief Model.......................................... 87
B.2 The Theory of Reasoned Action and the Theory of Planned
Behavior....................................................... 87
x


CHAPTER 1
INTRODUCTION
Breast Cancer is an important public health issue in the United States. It is the
second most commonly diagnosed cancer among American women and prognosis
depends significantly upon early detection and stage at diagnosis. Until relatively
recently, prevention was limited to changes in lifestyle factors, but in 1998 the first
chemoprevention drug for breast cancer was approved by the Food and Drug
Administration (FDA 2005). This drug was tamoxifen. The Breast Cancer
Prevention Trial (BCPT) demonstrated that tamoxifen use was shown to reduce the
incidence of breast cancer among women by 49% (Mulley and Sepucha 2002). Since
then, further clinical trials have been conducted on other potential chemoprevention
agents, such as raloxifene. Although these clinical trials are essential to determine the
efficacy of chemoprevention, only a small proportion of those who are eligible to
enroll do so. Current research indicates that a variety of factors may influence
womens decision-making when it comes to considering enrollment in breast cancer
chemoprevention trials, such as physician recommendation, perceived risk of breast
cancer, and concern over side effects, to name a few (Bober et aL 2004; Lovegrove et
al 2000; Yeomans Kinney et al.1998a).
Understanding womens motivations for enrolling in chemoprevention trials is
important for several reasons. Chemoprevention trials are essential for determining
the efficacy and toxicity of the drugs involved, unfortunately they often involve long
time commitments, require the recruitment of many high-risk but currently healthy
participants, and have complicated protocols (Greenwald 1995; Jordan 1995;
Yeomans Kinney et al.1998b). According to one organization, only 3/ to 5% of
adult cancer patients actually participate in cancer clinical trials while approximately
1


20% may be eligibleCCRP 2009c). In at least one chemoprevention trial,
researchers have noted a concern regarding the lower than expected response rate
for enrollment (Lovegrove et al 2000). Clinical trials not only further medical
knowledge around a given health issue, they also provide access to cutting edge
treatment and prevention options for those with a cancer diagnosis, or those at-risk.
By illuminating why some choose to participate and some do not, physicians and
organizations who work in cancer treatment and prevention can better tailor their
approach for recruiting women to clinical trials.
The method through which a high-risk label is assigned to women is an
important point of consideration. Rose notes that, the determinants of incidence are
not necessarily the same as the causes of cases1985:34). Furthermoreit may be
difficult to identify determinants of incidence within a population if exposure is
widespread (Rose 1985). Breast cancer is an excellent example of this phenomenon.
The most significant risk factor for the disease is advanced age, which all women
(barring early death) will be exposed to. However, for the purposes of breast
cancer chemoprevention, old age by itself does not make one high-risk. Rather a
model that examines multiple individual risk factors relating to family history,
reproductive history, and history of other types of breast disease is used to create a
personalized risk score (Constantino et al. 1999; Gail et al. 1999). Chemoprevention
is therefore an intervention that targets individuals at risk for breast cancer.
The primary research aim in this study is to explore the decision-making
process women at high risk for breast cancer undergo when considering enrollment in
a breast cancer chemoprevention trial.A second is to identify factors that influence
the decision-making process related to social relationships and social support. While
this project utilizes a primarily inductive approach, social relationships and social
support were of particular interest heading into the study due to the fact that they have
2


largely been overlooked in the literature in favor of more individual level factors.
Exploratory, qualitative methodology was used involving in-depth interviews with
women who participated in the Study of Tamoxifen and Raloxifene, a clinical trial for
breast cancer chemopreveation. The results are explored using a socio-ecological
framework that analyzes multiple levels of influence on decision-making,
emphasizing the individual, interpersonal and community levels of influence. The
experiences of women during trial participation also emerged during the interview
process, and these experiences are discussed in relation to the socio-ecological
approach that provides the overall framework for this project. Finally, implications
for future research are discussed.
This study was facilitated by the Colorado Cancer Research Program (CCRP),
a non-profit organization based in Denver, Colorado whose mission is to provide the
people of Colorado the opportunity to participate in and benefit from medical
research through cancer clinical trialsCCRP 2009a). CCRP is a Community
Clinical Oncology Program (CCOP), part of a network established by the National
Cancer Institute (NCI) in 1983 to connect eligible patients with NCI-sponsored
clinical trials for the prevention and treatment of cancer (NCI: CCOP electronic
document). The purpose of the CCOPs was to bring cancer trials to local
communities by creating a nationwide network of local health centers and local
providers to conduct clinical trials. The CCOPs, who partner with NCI Research
Bases, conduct patient recruitment and data collection, as well as oversight of the
community health care providers (NCI 2005). CCRP provided not only information
and resources but also conducted the recruitment portion of this project.
3


CHAPTER 2
BACKGROUND
Breast Cancer and Chemoprevention
Despite the many medical advances in breast cancer treatment over the last
century, current incidence and mortality statistics are still startling. One in eight
women will develop breast cancer at some point in their lifetime, according to the
National Cancer Institute (NCI 2008). It is estimated that 182,460 women were
diagnosed with new cases of invasive breast cancer in 2008, a number that does not
include the more than 60,000 cases of in situ1 breast cancer (ACS 2008a). In terms of
mortality, nearly 40,480 women died of the disease that same year (NCI 2008). It is
evident from these numbers that cancer of the breast presents an important public
health problem.
Both incidence and mortality rates for breast cancer remain a cause for
concern. Age-adjusted incidence rates for breast cancer continually increased from
1980 to 2001 for U.S. women of all races and while incidence has decreased since
2001, breast cancer remains the second most common cancer diagnosis in the U.S.
among women after skin cancers (ACS 2008a; Kapp et al. 2008; NCI 2008). The
mortality rate for breast cancer has shown a minor downward trend since 1990 but
this trend has not been equal in all sub-populations of American women (Cyrus-
David and Strom 2001;NCI 2008). African-American women, for example, are
more likely to die of breast cancer than their White counterparts while the highest
mortality rates from breast cancer belong to Native Hawaiians (Kapp et ai 2008;
Newman et al. 2001;Shavers and Brown 2002).
1 In situ refers to carcinoma that is confined to the layer of cells in which it originated.
It has not spread to other organs or tissues (ACS 2009).
4


There may be biological or behavioral differences between sub-populations
that contribute to mortality, nonetheless, it is known that differential access to
screening and treatment has an impact as well. Stage at diagnosis2 3 plays an important
role in breast cancer outcomes and it has been argued that the recent decrease in
mortality may be due to improved screening practices (Key et al. 2001). As Figure
2.1 below demonstrates, while the overall five-year survival rate for breast cancer is
relatively high, it varies quite noticeably by stage at diagnosis.
100
All Local Regional Distant
Stage at Diagnosis
Figure 2.1. Five-Year Relative Survival Rates by Stage at Diagnosis3
(From ACS 2008a:11)
2 At diagnosis, cancer is classified by stagewhich refers to specific characteristics
of the tumor, including size, location, and spread of the disease outside the immediate
area of the tumor. A number typically identifies stage, progressing from early or in
situ carcinoma identified as stage I, to stage IV that denotes invasive cancer in an
advanced stage (ACS 2008a).
3 The terms local, regional and distant refer to the spread of the disease from the
original location of the tumor and roughly correspond to numbered stages as follows:
Local (Stage I), Regional (Stages II and III), and Distant (Stage IV) (ACS 2008a).
5


Differences in access or utilization of breast cancer screening may be due to
multiple factors. Economic circumstances, such as having low-income or being
uninsured, limit access to available screening or prevention measures. Meanwhile,
programs such as the National Breast and Cervical Cancer Early Detection Program
(NBCCEDP) which are in place to assist women in these positions, may be under-
utilized (Cyrus-David and Strom 2001; DeSantis et al 2008). Certain ethnic or racial
groups are known to feel distrustful of the medical profession due to a shared history
of feeling targeted or marginalized by that profession- the most well known example
being the legacy of Tuskegee among African Americans (Armstrong et al. 2005;
Cyrus-David and Strom 2001). Breast cancer is sometimes believed to pose less of a
threat to minority women and this may lead to a lack of knowledge not only of the
disease, but also of the importance of prevention in these communities. When
information does spread to communities, it is not always appropriate to the particular
cultural background of that population (Armstrong et al. 2005; Nguyen et al. 2005).
Geography is also an important consideration in the United States, as states
differ dramatically in everything from the incidence of breast cancer to the types of
screening that are available and the frequency of follow-up care that is provided after
the screening. In part, this is due to the diverse socioeconomic make-up of each state,
including race/ethnicity, income, insurance status, and percentage living in rural
versus urban areas, though DeSantis et al (2008) point out that state policy on cancer
control can play a crucial role in influencing the disparities we see in mortality from
breast cancer.
Figure 2.1 illustrates the importance of early detection. With the advent of
chemoprevention, women who had a high-risk of breast cancer were given the option
of taking tamoxifen to prevent breast cancer rather than focusing solely on screening.
Chemoprevention is defined as the "treatment with either naturally occurring or
6


synthetic chemical agents to prevent, reverse, or arrest the progression of
preneoplastic lesions to invasive cancers (Cyrus-David and Strom 2001:522).
Currently used chemoprevention agents are classified as selective estrogen receptor
moderators or SERMs (Bober et al. 2004). Estrogen plays a key role in promoting
the growth of cells in the breast- a process which when unchecked can lead to breast
cancer (Key et al. 2001;NCI 2006). By affecting the estrogen receptors in the body,
SERMs interfere in the process of carcinogenesis. The use of the term selective in
the name is an important one. SERMs may inhibit estrogen receptors in one part of
the body, while having the opposite effect in other regions. Therefore, a SERM may
simultaneously prevent breast cancer while increasing the risk of another type of
cancer such as endometrial cancer (Cyrus-David and Strom; 2001;NCI 2006; NCI
2008). This is an important consideration when prescribing a course of a SERM to a
patient.
For over 30 years, tamoxifen has been used as an effective treatment for
primary breast cancer in women diagnosed with the disease (Vogel et al. 2006). Still,
it wasnt until the early 1990s that the Breast Cancer Prevention Trial (BCPT) began
recruiting women at high-risk to participate in a study examining the effectiveness of
tamoxifen in preventing breast cancer. The results indicated that tamoxifen was
indeed able to reduce both invasive and non-invasive breast cancer incidence among
women assigned to the drug by 49% (NCI 2008). Tamoxifen therefore became the
first chemoprevention agent approved by the Food and Drug Administration (FDA)
for breast cancer prevention in 1998 (Bober et al. 2004; FDA 2005). A side benefit
of tamoxifen is its positive effect on bone density. However, it also includes risk of
several fairly serious side effects. In addition to potentially exacerbating menopausal
symptoms, tamoxifen increases one's risk of endometrial cancer, thromboembolic
events and cataracts (Bober et aL 2004; NCI 2006).
7


More recently, the use of raloxifene, a second generation SERM prescribed in
the treatment and prevention of osteoporosis, for breast cancer chemoprevention has
been explored (Table 2.1). Research has included the Multiple Outcomes Raloxifene
Evaluation (MORE) and the Continuing Outcomes Relevant to Evista (CORE)
studies; both included a placebo arm. MORE was primarily intended to test the
efficacy of raloxifene in preventing fractures in postmenopausal women with
osteoporosis, though results also indicated a reduced risk of invasive breast cancer
among women enrolled in the raloxifene arm of the study. CORE then picked up
where MORE left off. Using the same study population, women who chose to
continue to take raloxifene for 4 more years following the completion of MORE had a
reduced risk of invasive breast cancer of 69% (Vogel et al 2006). While raloxifene
has been shown to have a decreased risk of thromboembolic disease and cataracts
when compared to tamoxifen, the risk of noninvasive breast cancers such as lobular
carcinoma in situ (LCIS) and ductal carcinoma in situ (DCIS)4, is increased.
However, this increased risk was not found to be statistically significant (NCI 2006;
Vogel et al. 2006).
4 Lobular carcinoma in situ and ductal carcinoma in situ both refer to types of non-
invasive breast cancers, which have not spread beyond the boundaries of the initial
cell layer where they originated. In the case of LCIS, the carcinoma is located in the
lobules, while in DCIS it is in the ducts (ACS 2009).
8


Table 2.1, FDA-aDproved uses of SERMs
SERM Uses
Tamoxifen Treatment: Metastatic breast cancer Adjuvant treatment Invasive breast cancer Prevention: Contralateral breast cancer
Raloxifene Prevention: Osteoporosis
(Adapted from Brown and Lippman 2000:6)
Study of Tamoxifen and Raloxifene
The women interviewed for this study were all participants in the Study of
Tamoxifen and Raloxifene (STAR). STAR was a non-placebo arm prevention trial
conducted by the National Surgical Adjuvant Breast and Bowel Project (NSABP).
The largest breast cancer prevention trial ever conducted at the time (enrollment
began in 1999 and concluded in 2004), STAR compared tamoxifen to raloxifene in
post-menopausal women at high risk for breast cancer (Altshuler and Somkin 2000;
NCI 2008; Vogel et al 2006). The goal of the study was not only to compare the
effectiveness of the two drugs in preventing invasive breast cancer but also to
compare their side effects relative to one another (Vogel et al. 2006).
A population of 19,747 postmenopausal women was randomized into two
groups and given a five-year course of 60mg of raloxifene or 20mg of tamoxifen. In
addition to being identified as high-risk for breast cancer, women had to be 35 years
9


old or more and be postmenopausal (NCI 2006). Risk was determined using the Gail
model and personal health history was also taken into account (Altschuler and
Somkin 2000; Vogel et al. 2006).
Study results indicated that tamoxifen and raloxifene were equally effective in
preventing invasive breast cancer, though raloxifene did not appear to be as effective
at preventing non-invasive breast cancer (Vogel et al. 2006). While the risk of certain
side effects appeared to be similar in the two treatment groups, those taking
raloxifene had lower risk of developing deep vein thromboses, pulmonary embolisms,
cataracts and uterine cancer, though the latter was not statistically significant (Vogel
et al. 2006). As a result of these findings, the study was un-blinded early and women
in the tamoxifen group were given the choice of completing their five years of
treatment with raloxifene instead (NCI 2006)
Determining Risk
Due to the potentially serious side effects related to the use of
chemoprevention agents, treatment is recommended only to women with an elevated
risk of breast cancer. The Gail model is one way of determining risk.
Gail et al. used data from the Breast Cancer Detection Demonstration Project
(BCDDP) to develop a model for estimating the risk of breast cancer for
women in a program of annual mammographic screening who have had no
previous breast cancer and who have no evidence of breast cancer at the time
of their initial screening mammogram. The model estimates the absolute risk
(probability) that a woman in a program of annual screening will develop
invasive or in situ (ductal carcinoma in situ [DCIS]) or lobular carcinoma in
situ [LCIS]) breast cancer over a defined age interval. (Constantino et al.
1999:1541).
To determine risk, this model examines current age, age at both first
menstruation and first live birth, family history of breast cancer in first-degree
10


relatives, number of previous breast biopsies and whether biopsy reveals the presence
of atypical hyperplasia (Constantino et aL 1999; Gail et aL 1999). The most
significant risk factor for breast cancer in women is increased age. This risk increases
most significantly prior to age 50; the rate of increase is smaller during the post-
reproductive period (Kelsey and Berkowitz 1988; Key et al. 2001). Additional risk
factors for breast cancer are summarized in Table 2.2.
11


Table 2.2. Risk factors and their impact on breast cancer
Factor ____________________
Age
Childbearing
Menarche
Breastfeeding
Menopause
Hormones
Oral contraceptives
Benign breast disease
Diet
Alcohol/ Smoking
Anthropometry
Exercise
Ionising radiation
Environmental oestrogens
Family history
High-risk mutations
Imact on Risk ______________
Risk increases with age. This relationship is more pronounced
during the reproductive years.
Nulliparity is associated with increased risk for the disease. Breast
cancer risk is reduced with each pregnancy, while younger age at
first pregnancy also provides more protection.
Later age of menarche lowers risk.
Believed to provide protection against breast cancer, although this is
still controversial.
Younger age at menopause is protective against the disease. Nor
does it matter if it is a natural menopause or one induced by bilateral
oophorectomy.
High levels of endogenous hormones may be associated with
increased risk in post-menopausal women. HRT during menopause
is associated with increased risk as well, though risk begins to
decrease following treatment cessation.
Breast cancer risk increases by 25% while oral contraceptives are in
use, however, begins to return to normal levels following cessation.
Proliferative lesions are associated with an increase in risk, as are
atypical hyperplasias. However non-proliferative lesions are not.
Although a connection between high-fat diets and an increase in
breast cancer risk has been suspected in the past, currently this is
unconfirmed, as is a decreased risk for high soy diets. A diet high in
vegetables may protect against the disease.
Alcohol causes some increase in risk. Currently, smoking and
increased breast cancer risk has not been shown to be associated.
Those who had high birth weight are at increased risk for breast
cancer. Obesity increases risk by 50% in post-menopausal women.
Physical activity decreases risk for breast cancer, particularly in
premenopausal women.
Exposure to radiation is known to increase risk for breast cancer,
though how large an effect diagnostic radiotherapy poses is still
uncertain.
The few studies done so far have not shown as association between
the compounds and risk of disease.
Risk doubles if a woman has a first-degree relative, less so if the
relative is second-degree. Mmilar environments and diet may also
play a role in incidence of breast cancer among family members.
Mutations in the BRCA1, BRCA2, P53, PTEN and ATM genes
increase risk of breast cancer. BRCA1 and BRCA2 are particularly
high risk.
(Adapted from Key et al 2001)
12


Family history in particular plays an important but varied role in determining
risk, indicating the importance not only of biological factors, but that of the
environment as well. The presence of a first-degree relative with breast cancer
doubles ones risk of developing the disease. An increase in risk is also associated
with second-degree relatives, though it is not as pronounced (Kelsey and Berkowitz
1988; Key ^ a/. 2001). Similar environmental factors and lifestyle choices account
for some of the clustering of breast cancer cases seen within families; a small
percentage of these clusters can be traced to specific genetic mutations. Key et al
write that, high-risk alleles probably account for most of the families with four or
more breast-cancer cases, for around 20-25% of the familial breast-cancer risk
overall, and for about 50/ of all breast cancers2001:138). The significance of these
mutations however, is not in the frequency of their occurrence but in the fact that their
presence may drastically increase an individual1 s risk of developing breast cancer.
Five mutations have been tied to increased risk for breast cancer, the most well
known being the BRCA 1 and BRCA 2 genes (Key et al 2001). It has been
estimated that the lifetime risk of breast cancer for an individual with the BRCA 1/2
mutations may be as high as 50-80% (Armstrong et al. 2005).
Womens Decision-Making When Considering
Enrollment in Chemoprevention Trials
Chemoprevention trials, such as STAR, are essential in determining the
effectiveness and safety of new chemoprevention drugs. Still, further research is
needed into the motivations and decision-making processes of women who choose to
enroll in these trials due to difficulty with recruitment (Altschuler and Somkin 2005;
Bober et al. 2004; Yeomans Kinney et al. 1998b). According to the Colorado Cancer
Research Program, when it comes to clinical trials for cancer research, less than one
13


quarter of adult cancer patients who are eligible to participate actually do so5 (CCRP
2009c). In the area of cancer prevention, two studies that surveyed an eligible
population for enrollment in a chemoprevention trial for breast cancer (either IBIS,
the International Breast Cancer Intervention Study, or BCPT), found that only half or
less than half of the women who were eligible actually choose to enroll (Lovegrove et
ai 2000; Yeomans Kinney et al 1998b). There are negatives to trial enrollment.
These trials involve taking drugs that may have potentially uncomfortable or serious
side effects and participation may be time-consuming. In addition, while women
must be at high-risk for breast cancer in order to enroll, they are currently healthy and
this risk does not guarantee the eventual development of breast cancer. Yet some
women do choose to enroll. What factors influence one potential participant to enroll
in a chemoprevention trial for breast cancer while others choose not to?
How women interpret their risk for breast cancer has been shown to be
associated with womens decision-making for chemoprevention. Women who
believed themselves to be at high-risk for breast cancer were more likely to choose
enrollment in chemoprevention trials (Bober et al 2004; Lovegrove et al. 2000;
Yeomans Kinney et al 1998b). While absolute risk as determined by acknowledged
risk factors such as a family history or nulliparity is important, how these risk factors
are interpreted by women within the context of their own lives is also key to
understanding their decision-making. For example, among women who had a first-
degree relative with breast cancer, their perception of their own health compared to
that of the relative was an important predictor of the likelihood they would choose to
enroll in chemoprevention trials (Altschuler and Somkin 2005). Those who chose not
to enroll in a chemoprevention trial were more aware on average of the impact of
5 As cited in Chapter 1,"only 3% to 5% of adult cancer patients actually participate in
cancer clinical trials while approximately 20% may be eligibleCCRP 2009c).
14


lifestyle factors on breast cancer risk and how they could be mitigated than those who
chose to enroll (Altschuler and Somkin 2005; Lovegrove et al. 2000). However,
Cyrus-David and Strom (2001), in a study evaluating general knowledge of
chemoprevention among a sample of high-risk women, found that those with a close
friend or family member who had reduced their risk for breast cancer through
behavioral measures but still developed the disease despite this were less likely to be
open to the idea of chemoprevention.
Psychological factors, such as intrusive thinking and worry about breast
cancer may be associated with choosing trial enrollment as well (Bober et al. 2004).
Still, further research is needed in this area due to contradictory findings. Women
who participated in the STAR trial actually had less anxiety about breast cancer than
those who opted not to, according to a study by Altschuler and Somkin (2005). This
may be due to the fact that, as the same study suggests, women viewed trial
participation as a way to gain some measure of control over their lives. Similarly,
participants were more likely to report that they would gain peace of mind from trial
enrollment than those who chose not to enroll (Yeomans Kinney et al 199^),
Aside from the fear of breast cancer, concerns about the actual clinical trials
and the drugs involved were also a consideration. Trial enrollment was associated
with fewer concerns over the side effects of chemoprevention drugs and more sources
of information regarding tamoxifen use (Bober et aL 2004; Yeomans Kinney et al
1995). Others were concerned with the randomization process used to assign women
to different arms of the trials, the possibility of being assigned a placebo, or the time
commitment involved (Altschuler and Somkin 2005; Yeomans Kinney et al. 1995).
Women who were currently taking estrogen replacement therapy but would have to
stop to enroll in the trial were also less likely to choose enrollment (Yeomans Kinney
et aL 1995).
15


The current literature emphasizes the importance of physician
recommendation either for or against trial participation in womens decision-making.
Physician recommendation for participation was found to be strongly associated with
enrollment in chemoprevention trials, which is consistent with its similarly strong
influence in breast cancer screening through mammography (Yeomans Kinney et al
1995). In fact, those women whose physicians had recommended trial participation
were thirteen times more likely to do so than those whose physician had not made the
recommendation (Bober et aL 2004; Yeomans Kinney et al. 1998a).
The simple act of recommendation is not the only pertinent factor regarding
the role of the health care provider. Physician communication style may also be
important in the decision-making process, particularly the understandability of the
information given to their patients. This is supported by research showing that the
framing of a message by physicians is influential in the decision-making process
(Bober et al, 2004; Lovegrove et al. 2000). In addition, those patients who felt that
their physicians involved them in the decision-making process tended to be more
satisfied with their decisions in the end (Bober et al. 2004). On the other hand, one
qualitative study indicated that poor physician communication resulted in women
being less willing to undergo chemoprevention, though this was not conducted within
the context of decision-making for enrollment in a clinical trial (Cyrus-David and
Strom 2001)
16


CHAPTER 3
THEORETICAL FRAMEWORK
Socio-Ecological Approaches
Socio-ecological approaches to medical anthropology research emphasize the
multiple levels of interaction between an individual, their physician, and the social
environment, and the way in which this interaction influences individual behavior and
beliefs. Although the terms may vary depending on the specific model used, these
levels of influence typically proceed in a fashion from the individual, to interpersonal
relationships, to the community or neighborhood, and finally to the larger social and
cultural environment (Elder et al 2006; Fisher et al, 2005; Sallis and Owen 2002).
While biological processes in the individual may be the most proximate levels of
health, these processes are perceived and interpreted through the lens of an
individuars larger environment. Yet this relationship is reciprocal rather than uni-
directional. The processes taking place at a more distal level of influence also shape
and impact an individuars health experiences while at the same time providing the
larger context in which the individual understands those experiences. Ecology, as
used in this type of approach, refers to the space outside of the person meaning
their physical and social environments, as well as their internal environment (Sallis
and Owen 2002:462).
In moving health promotion from interventions that target individuals to
interventions that also target environments, it is important not only to examine the
different levels of influence on health but also how these levels interact with one
another (Stokols 1992). Stokols identifies four assumptions regarding socio-
ecological approaches:
17


The first is that health is influenced by multiple facets of the physical
and social environments. The role of personal attributes is also
acknowledged. The second assumption is that environments
themselves are multidimensional. Environments can be described as
social or physical, actual or perceived; as discrete attributes (such as
temperature or spatial arrangements); or as constructs (such as
behavior setting or social climate). The third assumption recognizes
that human- environment interactions can be described at varying
levels of aggregation: individuals, families, work and cultural
organizations, communities, or whole populations. Stokols fourth
assumption is that there is feedback across different levels of
environments and aggregates of persons (Sallis and Owens 2002:465-
466).
Individuals may possess the skills and attitudes to make certain choices when
it comes to their health, however, these must be viewed through their interactions
with the social and physical environments. The environments that either constrain or
augment their ability to seek care for a medical condition, access screening programs,
or implement lifestyle changes. The first assumption notes the influence of individual
characteristics on health, while also recognizing that the wider ecological context also
plays a role in health behavior. In a study on self-management of diabetes, Fisher et
al (2005) argue that self-management approaches which rely solely on the abilities of
the patient ignore the relationships between differential access to resources and social
support, and the larger social, physical and policy environments which interact to
influence behavior. Possessing the knowledge of how one manages diabetes is
useless, for example, if the individual is unable to follow through on that knowledge.
Healthy eating and exercise may not be possible due to the built environment in
which the individual lives, while social support for behavior change may transition
throughout the life cycle (Fisher et al. 2005).
18


According to the second assumption, ecology does not simply refer to the
natural or physical environment but rather may include the social and cultural
environments in which individuals live. That environments can be actual or
perceived is illuminating for health behavior research in which individuals beliefs
regarding their environment may dictate how they view barriers to health care or self-
efficacy when it comes to behavior change. Interventions can address these
perceptions. For example, reduced physical activity among girls was the subject of
the trial of activity for adolescent girls, which sought to address the situation through
school-based interventions (Elder et al. 2006). While increasing the availability and
awareness of physical activity opportunities for girls in both schools and the local
community were objectives of the interventions, another was to address "real and
perceived barriers to being physically active'>, among girls (Elder et ai 2006:161,
emphasis added).
The third assumption is critical to socio-ecology, which is a multilevel
approach. Analyses of human-environment interactions have progressed beyond the
infectious disease model of host-agent-enviromnent to include a more complex
classification of what constitutes environmental factors and at what level groups are
studied (Sallis and Owen 2002). The influence of the environment can be studied not
only at the individual level, but also at the level of the family, the social network, the
community, the institution, et cetera. For example, teen smoking patterns have been
researched by examining the influence of the parent-child relationship, the school
environment, neighborhood characteristics, and state policy (Sampson et ai 2002;
Wen et al 2009; Wilcox 2003).
Finally, ecological levels are not discrete entities, bounded by neat lines. The
fourth assumption, that of feedback across the different levels of the socio-ecological
context, illustrates the difficulty faced by researchers attempting to determine where
19


to place influential factors for health behavior, and which direction the influence
travels. In some cases, the feedback relationship is reciprocal, while in others it may
be difficult to actually differentiate the effects of multiple levels from one another due
to the fact that causal forces for behavior change in different environments may be
correlated to one another (Cook 2003). In the case of adolescent smoking, for
example, schools may affect smoking through the peer groups that the schools
themselves make possibleCook 2003:153).
Socio-ecological models vary depending on their applications to research. An
early model was that of Bronfenbrenner, whose microsystem, mesosystem, and
exosystem influences on behavior contributed to later frameworks, including the one
utilized for this study (McLeroy et al.1988; Sallis and Owens 2002). These system
levels proceeded from the most local to the most structural, with microsystem
corresponding to interpersonal influences, mesosystem to the community,
organizational or institutional levels, and exosystem to social, cultural, political and
economic forces. The specific socio-ecological framework applied to the current
study is that used by McLeroy et al. (1988) and Gregson et al (2001), which identifies
five levels of influence. These levels are defined as individual, interpersonal,
institutional/ organizational, community, and social structure, policy, and
systems15 (Gregson et al 2001 :S5).
This particular model was chosen due to the fact that it included not only
community levels of influence, but also organizational, which is particularly relevant
for this study given the impact of both womens beliefs and values, which were
shaped by the community of which they were a part, but also the role of CCRP in
recruiting women to the STAR trial and then providing support to them during their
enrollment. Definitions for each of the levels of influence can be found in Figure 3.1.
20


Social Structure,
Policy, and Systems
Local, state, federal
policies and laws that
regulate or support
healthy actions
Community
Institutional/
Organizational
Social networks,
norms, or standards
(e.g. public agenda,
media agenda)
Rules, regulations,
policies, and informal
structures (worksites,
schools, religious
groups)
Interpersonal
Interpersonal
processes and primary
groups (family, peers,
social networks,
associations) that
provide social identity
and role definition
Individual
Individual
characteristics that
influence behavior
such as knowledge,
attitudes, beliefs, and
personality traits
Figure 3.1. Socio-Ecological Levels of Influence
(Adapted from Gregson et al 2001:55).
21


Socio-Ecological Approaches, Health Behavior,
and Breast Cancer Chemoprevention Trials
To date, research in the area of participation in chemoprevention trials for
breast cancer has largely focused on the individual level, drawing primarily on one of
two theoretical frameworks: the Health Belief Model and the Theory of Reasoned
Action. Concepts drawn from the Health Belief Model and the Theory of Reasoned
Action that have been applied to research on chemoprevention decision-making for
breast cancer include perceived susceptibility (often termed perceived risk), perceived
barriers and benefits, perceived behavioral control {locus of control), self-efficacy,
and normative beliefs (Altschuler and Somkin 2005; Cyrus-David and Strom 2001;
Lovegrove et ai 2000; Yeomans Kinney et ai 1995; Yeomans Kinney et ai 1998a;
Yeomans Kinney et ai 1998b). Further concepts and definitions important to these
two frameworks can be found in Appendix B.
The perception in biomedicine that individuals have a responsibility for their
own health is particularly salient given the focus on decision-making and perceived
risk. One of the key trends in the process of biomedicalization since the mid-1980s
has been the transfer of responsibility for health from the sphere of the physician to
that of the individualwho now has a moral obligation to remain healthy (Clarke e/
a/. 2003:171). Included in this responsibility are a heightened awareness of ones own
personal risk factors and what the individual can do to mitigate this risk, a trend that
Rockhill terms risk privatization2001:365). In order to determine risk, there has
been a large emphasis on what has been called risk factor epidemiology and the
development and application of ever more complex risk assessment tools, which can
be seen clearly in the field of breast cancer (Clarke et ai 2003).
One key example of this is the Gail model. By quantifying the multitude of
scenarios that may make women susceptible to breast cancer, the vague notion of
22


''risk" is transformed into a number, a value that can be easily compared and
stratified. Nor is this value totally subject to fate. While women cannot currently do
anything to change their genes, or the age at which they begin menarche or
menopause, the remainder of the list of breast cancer risk factors could be said to be
subject to alteration. Oral contraceptive use or hormone replacement therapy (HRT);
diet and exercise; alcohol or tobacco use; nulliparity or breastfeeding; all rely to one
degree or another, on choices women make. Therefore, according to this trend
towards risk privatization, women can and should mitigate their own risk by changing
their behavior. Chemoprevention becomes one more route through which women can
lower their risk of breast cancer.
Although this prioritizing of individual risk in decision-making has been a
focus of the research on chemoprevention trial enrollment, it should be recognized
that it is part of larger society and policy level phenomena that is impacting health.
With the advent of genetic testing, individual risk can now be defined on the
molecular level to result in what Lock ef a/, calls the individual genetically at risk
(2007:257) Despite the fact that for most diseases the presence of a particular allele
or mutation rarely gives a simple yes or no answer to whether or not they will actually
develop the condition, risk is calculated using these tests. Though how individuals
themselves interpret their results in a meaningful way for their own lives may be
more complex (Lock et al 2007; Rapp 2000). A similar shift to an individual
responsibility model can be seen in the field of chronic diseases- those with diabetes
for example are encouraged to self-manage their disease (nsher et al. 2005). Even
the language around health at the organizational or policy level has actually
undergone a shift:
Terms such as health maintenance, health promotion, and healthy living
highlight the mandate for work and attention toward attaining and maintaining
health. There has been a steady increase in mandates for self-regulation until,
23


with biomedicalization, there is a shift in the general cultural expectations of
whole populations. In this constant, self-disciplining and other/public-
disciplining, there is no rest for the weary (Clarke et ai 2003:170).
In chemoprevention research, the primary exception to the focus on the
individual level of influence has been the effect of physician recommendation on the
potential participant, which reflects a larger interest in the patient-physician
relationship in general (Kreuter ei aL 2000; Lovegrove et al 2000; Quill and Brody
1996). Physician recommendation falls under two levels of analysis, the individual
and the interpersonal. From the standpoint of the Theory of Reasoned Action,
normative beliefs refer to the attitudes of important referent individuals towards a
particular health behavior (Montano et al 1997:87). A physician could be considered
such an influential person. In contrast, at the interpersonal level, it is not only the
physician^ attitude that matters, but the relationship and level of communication
between the physician and the patient as well (Bober et al. 2004; Lovegrove et al.
2000).
A more holistic view of breast cancer chemoprevention trial enrollment can be
conceptualized in terms of a socio-ecological model that allows for a multilevel
examination of factors that impact womens decision-making. An example would be
to consider not only how the individual perceives risk based on her own known
biological risk factors, but also how those risk factors in that individuals
environment shape her perception of risk. The individuals understanding of the
severity of breast cancer may depend on not only personal experience, but also
relationships with others who have had the disease, knowledge of well-known public
figures with breast cancer, or the messages seen in the media. The view that women
hold of clinical trials may be shaped by their personal history, the history of the sub-
population to which they belong, their education level, or their physician^ opinion.
24


A few factors that may reflect the influence of interpersonal or community
levels have been briefly touched on in the literature. However, they have not been the
primary focus of research but rather were generally included as part of inquiries
conducted using close-ended questionnaires that have not allowed women to expand
on their thoughts. By choosing which factors to include in the questionnaire, the
researchers may incorporate assumptions about how women think about risk or which
factors they perceive to be important without truly examining those assumptions. In
addition, survey research has a history of assuming that each individual is an
independent entity or of treating events as independent of one another, when in fact
individuals are embedded within a larger social context (Coleman 1958). This social
context, including relationships between individuals and shared community values or
beliefsmay shape womens answers and should be acknowledged by researchers.
Interpersonal and community factors that have been identified include the
influence of physician recommendation, the attitude of ones spouse, having had a
friend or family member with the disease, and wanting to help other women (Bober et
ai 2004; Lovegrove et ai 2000; Yeomans Kinney et ai 1995). Unfortunately, there
has been little unifying theory applied to these levels of analysis. While the socio-
ecological model allows for a more holistic understanding of a given health behavior,
lack of specificity can be an issue. In order to provide this specificity, additional
theories or models can be applied to each level of analysis, from individual to society
(Elder et ai 2006; Sallis and Owens 1997).
Of particular interest for this study are social support and social network
theories, which may provide a deeper understanding of the interactions between
participants and others in their environment. Examining individual constructs is
useful, but rarely are the beliefs or actions they describe unconnected. The
patient/physician dyad is interesting not only because of the impact of physician
25


recommendation on the individual, but because that recommendation is taking place
within a relationship- a relationship that may also include other forms of social
support. The role of intrusive thinking and worry has been studied within the context
of trial enrollment. It may be that support given by ones physician mitigates those
concerns. The physicians communication style presumably would depend on the
relationship they have with their patient, whether more authoritative or more
egalitarian. Social networks in particular could be important in illustrating whom
participants spoke to regarding their enrollment or explaining their previous
experiences with breast cancer. The socio-ecological model provides a framework to
organize different levels of influence on health. Social support and social network
theories may explain these influences.
Social Support and Social Network Theories
Current research on decision-making in the area of chemoprevention trials has
focused on the individual, ignoring the fact that individuals exist within a social
environment made up of groups of individuals and relationships between those
individuals. To address this gap, social support and social network theories can be
integrated into the socio-ecological model at the interpersonal and community levels.
The link between health and social support can be traced back to one of the fathers of
modem sociology, Emile Durkheim and his landmark work, Suicide: A Study in
Sociology (1951), originally published in 1897. Durkheim was the first to examine
suicide not as an individual phenomenon, but as a social phenomenon. His proposal,
that social integration was an important protective force against suicide, has been
very influential in the literature on social support and social networks (Kawachi and
Berkman 2001).
26


A social network is the web of social relationships that surround individuals
(Heaney and Israel 2002:185). The descriptive characteristics of this web are
identified in Table 3.1.These characteristics are influential in determining what types
of coping mechanisms individuals may have available when a health issue arises, as
well as how resources, influence and risk travel through social groups (Friedman and
Aral 2001; Heaney and Israel 2002; Roberts et ai 1994). Different social
relationships may provide different types of social support (Table 3.2). An analysis
of the interpersonal aspect of social relationships that influence decision-making
around trial enrollment must move beyond the physician-patient relationship and
what could be called instrumental and/or informational support to further
contextualize the interpersonal environment. In addition to their physician, other
members of an individual5s social networks may also be contributing social support,
such as a partner, child or friend.
Table 3.1. Descriptive characteristics of social networks
Concepts Definitions
Reciprocity Extent to which resources and support are both given and received in a relationship
Intensity Extent to which social relationships offer emotional closeness
Complexity Extent to which social relationships serve many functions
Density Extent to which network members know and interact with each other
Homogeneity Extent to which network members are demographically similar
Geographic dispersion Extent to which network members live in close proximity to focal person
(Adapted from b eaney and Israel 2002:187)
27


Table 3.2. Types of socia support
Concepts Definitions
Emotional Support Expressions of empathy, love, trust and caring
Instrumental Support Tangible aid and service
Informational Support Advice, suggestions and information
Appraisal Support Information that is useful for self- evaluation
(Adapted from Heaney and Israel 2002:187)
The ameliorating influence of strong social support on stress and anxiety
following a breast cancer diagnosis has been studied (Roberts et al.1994). However,
the same has not been true of those who are identified as high-risk for breast cancer.
At the moment these women are healthy but with the emphasis placed on prevention,
these women are choosing t&to 'treat1 the risk of cancer,5 (Clarke et al. 2003). How
women access support through their social networks when making the decision to
treat their risk by enrolling in chemoprevention trials is therefore of interest here.
Another area of research related to social networks that could be further
explored is that of altruistic motivations. The literature suggests that having a relative
or friend who had breast cancer or died of the disease does exert some sort of
influence on decision-making (Altschuler and Somkin 2005). The desire to help
family members at future risk for the disease has also been associated with study
enrollment (Lovegrove et ai 2000). The subject of altruism and clinical trials has
been addressed elsewhere in the clinical trial literature, though not extensively within
the context of breast cancer chemoprevention. Altruism is a motivating factor in
enrollment for placebo-controlled trials of antihypertensive drugs, specifically the
ideas of helping other patients and contributing to scientific knowledgeHalpem
et al 2003:987). These two concepts illustrate the argument of Simon et ai, that
28


altruism in the context of clinical trials is not a monothematic notion but rather
reflects a more diverse set of concepts and concerns (2006:46). In other words,
altruism refers not only to the idea of helping one particular group of people, but may
encompass family, friends, other patients, and the population in general. However, in
the specific study by Simon et al. (2006), altruism was not found to be a major
motivator for trial participation.
Altruism has also been studied in the context of social networks. In a study of
volunteering and its influence on health, an association was found between
volunteering outside one^ direct social network and positive health gains (Brown et
aL 2005). Drawing on these examples from other areas of health research, and the
preliminary findings on quantitative instruments used in breast cancer
chemoprevention research given above, altruistic motivations towards family
members or other members of ones social network appear an area worth
incorporating into the socio-ecological framework used to analyze decision-making
around chemoprevention trials.
29


CHAPTER 4
METHODS & ANALYSIS
Colorado Cancer Research Program
The Colorado Cancer Research Program was an invaluable partner in this
project, providing information and support to the investigator. CCRP is a non-profit
organization located in Denver, Colorado and describes itself as a:
A nonprofit community-based cancer program established to provide
community hospitals and physicians access to a wide range of cancer
research trials in order to provide their patients with greater options for
the treatment, control, and prevention of cancer. [CCRP 2009a]
Colorado is one of 34 states in which Cancer Clinical Oncology Programs (CCOPs)
such as CCRP are located (CCRP 2009). Their role is to bring clinical cancer trials,
which previously were confined to areas surrounding research centers, to
communities across the country allowing individuals to participate through their
physician or local health center.
CCRP provides oversight of the clinical trials, providing data support to the
local network of physicians and hospitals, and quality oversight for study procedures
and data management. In Colorado this is accomplished though a network which
includes seventeen Colorado hospitals, over 100 medical and radiation oncologists,
surgeons and other specialists, and a high trained CCRP oncology program staff1
(CCRP 2009a). It also includes research nurses, provided by CCRP, who serve as
participants primary contact among study personnel, remind them of appointments to
fulfill study requirements, and provide support and information for participants when
they have questions or concerns.
30


Research Design and Qualitative Methods
This study utilized a qualitative exploratory design in order to gain in-depth
understanding of the decision-making process for women at high-risk for breast
cancer considering enrollment in a chemoprevention trial. The strength of qualitative
research does not lie in its generalizeability, as in quantitative research, but rather in
that it provides detailed and nuanced information about a given phenomenon.
Validity rests not on a large sample size, but rather "has to do with description and
explanation, and whether or not a given explanation fits a given description
(Janesick 1994:216). Drawing on a grounded theory approach, an inductive and
iterative analysis was used by which the data were examined for patterns or themes,
which were then linked to existing theory. The theory is therefore grounded in the
data,11 as opposed to deductive approaches which assess the data using previously
defined hypotheses or theoretical constructs (Neuman 2003; Patton 1990).
Qualitative methodologies emphasize the importance of contextualization,
which has been lacking in the literature on decision-making for breast cancer
chemoprevention trials. These studies have typically relied on the use of close-ended
(or a combination of close- and open-ended) surveys and questionnaires, which are
then subjected to statistical analysis. In addition, they tend to make assumptions
regarding the independence of individuals and events, which may be problematic
when these individuals live in social contexts that may be very influential on their
beliefs and behaviors (Coleman 1958). Such an approach may identify important
factors in the decision-making process, yet a more nuanced in-depth understanding of
how and why these factors play a role requires a qualitative approach (Patton 1990).
While qualitative research into breast cancer chemoprevention trials has been rare,
one important study has influenced the research reported here.
31


Altschuler and Somkin (2005) conducted qualitative in-depth, semi-structured
interviews with a population of 51 women eligible for the STAR trial. The results of
their study indicate that age was an important factor in that women who chose to
participate were slightly older on average than those who chose not to participate. A
possibly related finding was that STAR participants had a higher perceived risk of
breast cancer. While most of the factors brought up were similar to results found
using quantitative methods in other articles, two important themes were revealed.
Women who enrolled in the STAR trial were more likely to see participation as a
form of activism and were also more likely to have a close friend who had had breast
cancer, indicating that activism and altruism are areas that should be included in
future studies.
When close-ended surveys are used, participants are forced to choose between
pre-selected answers decided on by the researcher. There is no room for women to
explain why they chose the answer they did; this is crucial as different participants
may arrive at the same action but have alternative motivations for doing so. The
reverse is true as well; two women with the same concerns may see opposite paths to
resolution. This type of quantitative instrument may also force women to choose a
response even if none are truly appropriate for their own circumstances. An example
of this can be seen in the study by Altschuler and Somkin (2005), in which those who
chose to participate in the STAR trial and those who chose not to both considered
their decision to be based on a desire to be in control of their health. However, the
actions that constituted taking control were completely different for each group.
Qualitative approaches, on the other hand, seek to allow participants to tell
their own story, to present their motivations and experiences using their own words,
without imposing preexisting expectations on the phenomenon or setting under
studyPatton 1990:44). A survey which simply asked whether or not women felt
32


themselves to be at increased risk for breast cancer would not necessarily lead to
insights into how that feeling influenced womens behavior. Yet by conducting a
qualitative study, Altschuler and Somkin (2005) were able to differentiate between
women who acknowledged risk intellectually and rather dispassionately, and women
for whom that risk that was a source of concern and anxiety. Not coincidentally,
these groups of women made different decisions in regards to trial enrollment.
Consider these two quotes from this study:
Intellectually, I do (feel at risk), but emotionally, I want to believe that I can
have good health and that I can prevent bad health. I dont dwell on the fact
that I might be getting cancer.
My gynecologist had the feeling from me that Fm an anxious person, and she
thought (joining the trial would be a good idea). So I thought, Shes right,
you know? If I dont participate, I will always think, Well, geemaybe I
should have. And if I do, maybe it will lesson the anxiety (Altschuler and
Somkin 2005:88).
Each of these women would respond that they felt at risk for breast cancer on a close-
ended survey. But by allowing the women to express their thoughts regarding this
risk, we see that they have very different responses to managing these feelings.
The qualitative methodology employed in this study takes the form of semi-
structured in-depth interviews. This choice of methods allows the researcher to
obtain data that allows for an understanding of how the participants organize and
contextualize their own experiences (Gelo et al. 2008). In keeping with this goal,
womens own words are often presented in the discussion of the data, drawing on a
polyphonic approach which allows differing beliefs and opinions to come through
(Marcus and Fischer 1986).
33


Data Collection
Purposive sampling was the most appropriate choice for this study given that
its purpose is ''selecting information-rich cases for study in depth" (Patton 1990:169).
The literature is primarily composed of quantitative studies that have utilized
questionnaires or surveys to gain a broad but not very nuanced view of womens
motivations when it comes to trial enrollment. The goal of the present study was just
the opposite. By selecting women for in-depth interviews who had made the decision
to enroll in STAR, a closer examination of influential factors and how they are related
was made possible. Eligibility for the study was restricted to women between 50 and
80 years of age, living in the Front Range region of Colorado, who had participated in
the STAR trial. In addition, interviews were conducted between June 2007 and
September 2008 and availability during this period was a necessary requirement.
Due to confidentiality concerns, the principal investigator could not access
CCRP^ STAR participant database for recruitment purposes. Instead personnel at
CCRP first identified eligible STAR participants and then handled initial contact.
Introductory letters outlining the purpose and procedure for the study were mailed out
by CCRP in waves of 20-30 letters to 140 eligible women, beginning with those who
enrolled in the STAR trial first. Included with the introductory letters were response
letters. Using an opt-in approach, women who were interested in being contacted
with further information were asked to return the response letters to the offices of
CCRP. The initial response letter asked for a name and phone number, however, a
space to include email address was eventually added as it became apparent that some
women preferred this method of communication. Only once participants gave
permission through these response letters was their contact information passed on to
the principal investigator.
34


Women who were interested in further information were then contacted by the
PI to request participation in a 30-60 minute in-person interview. Interviews were
held in the women1 s homes, or, if preferred, at the offices of CCRP or another public
location. Alternative locations included a library, a place of business, restaurants, a
city park, and a coffee shop. Twenty-five response letters were received during the
study period. Of these, seventeen women completed interviews. The remaining eight
either railed to respond to repeat attempts to contact or were out of town during the
study period.
While a goal of this study was to gain a better understanding of particular
type- women who choose enrollment, it is important to note that the women who sent
back the response letters and then agreed to interviews are a group that has been
doubly self-selected. First enrolling in STAR, then the present study, they twice
made the decision to participate in scientific studies. A potential bias that may result
is that this population may be more comfortable with scientific research than a
randomized sample would be and this may be reflected in the results. Those who
chose not to enroll in STAR or who chose not to speak with those conducting this
study might express very different ideas regarding trial enrollment or scientific
research in general.
Prior to each interview, informed consent was obtained from participants who
were also given the opportunity to request the results of the study once they were
ready for dissemination. Participants were then asked to complete a short
Demographic Questionnaire (Appendix C) that was used to inform the interview.
The form asked for basic information, including age, marital status, ethnicity,
education level and occupation. In addition, the form included several questions that
targeted family make-up and history. These questions included number of female
children, number of sisters, and whether or not the participant had a family history of
35


breast cancer. If the participant responded in the affirmative to the family history
question they were asked to give a short explanation of this history. These questions
were in included on the Demographic Questionnaire in order to provide the researcher
with this information for the interview process. The interviewer was able to use the
participants answers to focus questions regarding the influence of family history and
make-up on the participants1 decision-making process.
With the permission of participants, sixteen out of seventeen interviews were
tape-recorded. The seventeenth was not recorded due to a recorder malfunction;
however, notes were taken during the interview. Field notes were also taken
immediately following every interview in order to summarize the information gained
during the experience and to capture any immediate reactions or thoughts of the
principal investigator. The PI conducted all interviews.
While the purpose of this study was exploratory and inductive, a semi-
structured Interview Question Guide (Appendix D) was used to direct the interview
process. Questions were meant to progress from the general to the specific, allowing
participants to generate their own discourse around their motivations and the
decision-making process prior to addressing specific factors that had arisen in the
literature. Depending on the flow of the interview, questions were on occasion
addressed out of order. Participants were also given the freedom to address related
topics that came up during the interviews. In particular, any factors they felt were
important to their decision that had not been addressed by the question guide were
noted; these factors were then examined by the investigator and incorporated into the
36


question guide for later interviews (Neuman 2003). Interviews lasted from 20
minutes to 2 hours in length6.
Data Analysis
Following the interviews, all field notes were typed and all taped interviews
were transcribed using a word processing program. Names of friends, family
members or physicians given during the interviews were changed during this process
to maintain confidentiality. Important topics or themes that arose in early interviews
were used to inform the question guide for later interviews, in an iterative process.
Transcripts were initially hand-coded using an open-coding process whereby codes
are produced inductively from the text. These codes were then compiled and edited
to reduce repetitive or unnecessary codes and compared to the existing literature. In
cases where inductive codes clearly mirrored a construct already in use in the
literature, the name of the code was changed to reflect that of the construct. This
included perceived risk, perceived benefit and female offspringfor example.
Each code was clearly defined to ensure proper classification and the coding system
was completed (Neuman 2003).
Transcripts were entered into Atlas-ti, a qualitative data analysis program, and
re-coded using the final coding system. The final list of codes can be found in Table
4.1. Patton writes, the first task in qualitative analysis is description1990:374).
Once coding was completed, primary memos were written to summarize the data for
each code and to identify important concepts or themes. In order to write these
memos, all of the data under a specific code were compiled and read by the PI, who
6 The procedures outlined here were submitted to and approved by the Colorado
Community Institutional Review Board, as well as the Human Subjects Research
Committee at the University of Colorado Denver as HSRC Protocol 2007-131.
37


then created a summary of the data. This was repeated for each of the codes in the
final coding system. Through this examination, points that illustrated a consensus
among participants, or recurring regularities, were noted, as were any experiences
or comments that seemed to stand out (Patton 1990:403). Given the qualitative,
inductive nature of this project, it was particularly important to note those things that
participants viewed as significant or essential, not just the researcher (Patton 1990).
These points were then marked as possible concepts or themes to return to at the next
stage of analysis.
Table 4.1. Final code list
Discussion with Physician Decision-Making
Drugs: Tamoxifen Experience with Breast Cancer
Drugs: Raloxifene Importance of Breasts
Placebo Female Offspring
Having a Sister Friend/ Acquaintance with Breast Cancer
Perceived Risk: Study Participation Motivating Factors
Perceived Benefits: Study Participation Health History of Participant
Perceived Risk: Breast Cancer CCRP: Organization
Perceived Risk: Side Effects CCRP: Nurses
Enrollment Process CCRP: Study Events
Overall Study Experience Other Breast Cancer Activities
Perception of Medical Research Previous Trial Participation
Knowledge of Breast Cancer State of Mind
Denial Role of: Physician
Reconstructive Surgery Role of: Friend
Side Effects Experienced Role of: Spouse
Family History: Breast Cancer Role of: Family Member
Family History: Other Cancers Proactive Behavior
Upon completion of the primary memos, these concepts were drawn from the
initial memos and used to create clusters. Clustering is a qualitative method in which
similar concepts, actions, processes, etc. are grouped together into categories or
38


dusters which are then given a name based on the members of the group. Were
trying to understand a phenomenon better by grouping and then conceptualizing
objects that have similar patterns or characteristicsMiles and Huberman 1994:249).
Resulting clusters included The Role of CCRP Social Support, Perceived Risks
of Study Participation, Perception of Medical Studies, among others. Secondary
memos were then written which summarized the connections within and between the
clusters, and interpreted the results.
39


CHAPTER 5
MOTIVATIONS FOR ENROLLMENT
Demographic Characteristics of Participants
Seventeen women were interviewed between June 2007 and September 2008.
Participants ranged from 55 years to 73 years of age, with the majority of participants
under the age of 70 (Table 5.1). Fifteen of the seventeen women selr-identified as
either White, Caucasian, Anglo or European American. Two declined to
answer the question. The majority of participants were currently married, with
thirteen respondents answering as such on the Demographic Questionnaire. Two
were divorced and one participant identified herself as a widow. Only one participant
identified as single. As a group, the participants in this study were well educated with
all having had advanced beyond the high school level. Two had attended some
college six had completed bachelors degreesand four had completed graduate
degrees. A total of five participants had progressed to the post-graduate level.
Table 5.1. Age range of participants
Number of Participants Age Range Percentage
5 55-59 29.4
4 60-64 23.5
5 65-69 29.4
3 70-74 17.7
Family history of breast cancer is an important risk factor for the disease,
particularly having had a first degree relative with breast cancer. Twelve of the
seventeen participants responded affirmatively to the family history question. Of
these twelve, four listed multiple family members diagnosed with breast cancer, two
did not elaborate on the question, and the rest only listed one family member. Those
40


relatives listed included mothers, sisters, maternal aunts, one maternal grandmother
and one daughter.
A previous study posited the idea that having female children might influence
the decision-making process for women considering participation in chemoprevention
trials (Lovegrove et al 2000). Though it was not found to be statistically significant
in that study, it was a subject worthy of further inquiry. In order to further explore
the issue of how female family members might influence decision-making on
enrollment during the interview, participants were asked about the number of female
children and the number of sisters they had on the Demographic Questionnaire. The
results of these questions can be seen below in Table 5.2.
Table 5.2. Number of female children and sisters of participants
Number of Female Children # Participant Responses Number of Sisters # Participant Responses
0 6 1 7 2 3 3 1 0 5 1 8 2 2 3 2
In the case of two participants with female children, their daughters were
adopted. A participant with one adopted daughter did not have any family history of
breast cancer. The second, who had two adopted daughters, replied yes to the
family history question though she did not elaborate. Their adoptive status means
that the daughters are not at increased risk due to their adopted mothers family
history. This was therefore a potentially interesting point for further exploration in
the interviews.
The results of the in-depth interviews are anchored within a discussion of the
analytic levels of the socio-ecological framework, with primary emphasis being given
to those levels that emerged as the most important to women during the interviews:
41


the individual, interpersonal and community levels. The institutional/organizational
and social structure, policy, and systems levels will be addressed more fully in
Chapter 6. It is important to keep in mind Stokols5 four assumptions, however, when
examining the factors that influence womens decision-making around enrollment in
breast cancer chemoprevention trials (Sallis and Owen 2002). The lines between
levels of influence may blur and feedback often occurs that may make classification
of factors to simply one level alone difficult.
Individual Factors
Perception of Breast Cancer Risk
In order to be eligible for the STAR trial, women had to be identified as high-
risk for developing breast cancer. Both this medically defined risk, as well as prior
experience contributed to women's perception of their personal risk for the disease.
An association between having a high perceived risk of developing breast cancer and
the decision to take chemoprevention drugs or enroll in a chemoprevention trial is
supported by the literature (Bober et al. 2004; Lovegrove et al 2000). That perceived
risk is influenced by more than simply a Gail score is supported by the fact that
women tend to overestimate the risk of breast cancer. When asked to estimate the
risk of breast cancer for women with a family history at the ages of 30, 45, and 60,
participants in one study came back with numbers that were too high by a factor of 24
for the 30 year old and by a factor of 4 for the latter two ages (Lovegrove et al. 2000).
Also, while women are likely to overestimate their susceptibility to breast cancer, at
least one study found no association between interest in chemoprevention treatment
and Gail score (Bastian et al 2001). It is therefore not solely a medically defined risk
that is important to women; other factors are influencing how they perceive their own
risk.
42


The women in the current study generally discussed their own risk in relation
to either a family history of the disease or a personal history of other breast disease.
Twelve women had a family history of breast cancer, though this fact did not have a
similar effect in all women. Rapp, in her study on the social impact of amniocentesis,
illustrates that individuals understanding of risk is dependent on their own personal
and social history (2000). In the present study, participants differed in how closely
linked they felt to their family history and in its relevance to their own life. For some,
breast cancer felt like an ever-present part of their lives as in the case of a woman
who had lost both of her sisters and her mother to the disease, saying, 'Tm always
100king over my shoulder. A few women, on the other hand, acknowledged their
history, but felt that it was important not to dwell on it, while still another felt that her
similarity to her fathers side of the family made her unlikely to fall victim to her
mother^ family history. Family history was clearly important in determining risk
status for the trial. Still, the emotional connection women felt to that history and how
it influenced the perception of their own risk different between participants.
Key et al. observe that umost women with the disease do not have a family
history of it, and most women with affected relatives never develop breast cancer
(2001:138). However, family history of breast cancer was so identified with being at
high risk for the disease that at least two of the women without a history did not even
consider breast cancer a threat until one event or another brought it to their attention.
For the firstit was her daughters breast cancer. Despite the fact that all of the
women in her immediate family had cystic mastitis( and had been told that they were
at increased risk for breast cancer, the participant did not truly believe this risk was a
real threat due to a lack of family history; that is until her daughter was diagnosed 7
7 Cystic mastitis is also referred to as fibrocystic disease or fibrocystic changes (ACS
2008b).
43


with breast cancer at the age of 27. Another woman had no family history before her
own experience with LCIS and said that right up to the point when they told her she
would need a lumpectomy, she believed the diagnosis was a mistake. Although
women often overestimate their risk of breast cancer, it appears that in these two
cases at least, a lack of family history caused them to underestimate risk (Colditz
1997). In noting that personal experience can often shape how likely an individual
may perceive a given event, Rapp writes, this homegrown sense of statistics can be
quite powerful.55
Personal health history also contributed to perceived risk for breast cancer.
Some women had previous experience with precancerous tissue in their breasts, dense
breast tissue, or other breast conditions that led to medical oversight. Four women
had been affected by LCIS, while a fifth had atypical ductal hyperplasia, or ADH8.
Four women mentioned having been on HRT prior to their participation in the study,
which is known to increase the risk of breast cancer (Colditz 1997). Nor were
precancerous breast tissue or carcinoma in situ of the breast the only experience
participants had with cancer. Two women spoke of their previous experience with
skin cancer, while cervical and anal cancers had affected another participant each.
This previous experience with cancer was a significant motivating factor, according
to one participant:
But I am a cancer survivor. Um- about twenty, well,itM be twenty-
four years now. Twenty-four years ago I was diagnosed with cervical
cancer and I completely understand that there has yet to date been no
genetic correlation found between cervical cancer and breast cancer.
Typically it5s, you know, breast, ovarian, prostate. But given the
circumstances, I just couldn't help but feel that somehow my body had
8 Atypical ductal hyperplasia, or ADH, refers to a condition in which abnormal
growth occurs in the ductal cells of the breast. Atypical refers to the fact that these
cells are slightly distorted in how they are arrangedACS 2008b).
44


a history of making really poor decisions at the cellular level, and so
that was very concerning to me.
As this quote illustrates, risk was not defined based only on risk factors listed
in the Gail model.A personal health history that includes other health issues may
increase a woman's perceived risk of breast cancer, even if, as in the case of the
woman with cervical cancer, it has not medically been shown that the two conditions
are associated.
Perception of Medical Research
A positive perception of medical research was also a common thread in the
interviews conducted during this study. The value of medical research to womens
health, one participant felt, had already been shown by studies that had revealed the
risks of hormone replacement therapy, while another pointed out the benefits that
have been gleaned from clinical trials for pediatric cancers. Others expressed similar
sentiments in a more general manner, calling these types of studies an important
source of knowledge.
The STAR trial in particular was seen to be important and informative, with
one participant calling it "one of the best studies really of all time, I would say.v In
particular, she appreciated that once the preliminary results were in indicating that
raloxifene was effective, the study was un-blinded and women were notified of which
drug they were on. Women demonstrated a good understanding of the trial in
general, with one feeling it had particularly good oversight. When asked whether
they or others they knew had concerns about enrolling in the trial, most women said
no, with several citing the education level of themselves, their spouses, or their
friends.
45


The relatively high education levels in this group certainly appeared to play a
role here. All of the women had attended at least some college, while all but two held
college degrees. That most felt they understood the value and importance of clinical
trials and how they worked was clear- a few had even participated in previous studies,
while two had considered doing so. While none of the other studies involved taking
an experimental drug, at least two had involved dietary restrictions, and urine or
blood samples. The topics of these other studies included breast cancer, diabetes,
cervical cancer, and cancer prevention through nutrition. These values- high
education, positive views of clinical trials, the importance of scientific knowledge-
while reflected here in individual beliefs, may in fact reveal more about the
community level of influence, which will be discussed further in a later section.
While only about half of the women considered themselves to be fairly
knowledgeable or more knowledgeable than the average person when it came to
breast cancer, other skills or experiences they possessed were of importance here.
One had worked for an oncologist in the past; another was a laboratory scientist. Two
had experience working in health advocacy while another was a history teacher who
subscribed to Harvards Womens Health newsletter. Several had read Dr. Susan
Loves Breast Book (Love 1995) or saved articles on the subject since the diagnosis
of a family member. And while the participants may not have considered themselves
particularly knowledgeable, this may in fact reflect a biased view of what constitutes
average, as seen in one womans statement: I had heard of tamoxifen. I think
everybody has. In reality, research with a more diverse group of women has shown
that knowledge about breast cancer in general, and tamoxifen specifically, is not
widespread even among women at high risk for breast cancer (Cyrus-David and
Strom 2001).
46


Other topics that came up also reflected a fairly good awareness of issues
related to breast cancer, such as the available treatments, well-known publications on
the subject, the specific drugs being tested in the study and their side effects, the
dangers of hormone replacement therapy, and the importance of early detection in
prognosis. Two also brought up the genetic mutations that indicate a hereditary risk
for breast cancer. That this was a proactive group of women was obvious in the fact
that many of them, when faced with a friend or family member's breast cancer or
when making the decision to enroll in the trial, chose to do their own research on the
subject rather than rely solely on information from a physician.
Perceived Benefits of and Barriers to Participation
Concerns were given about certain aspects of study participation, such as the
time commitment involved, their ability to follow through with study activities, and
the side effects associated with the study drugs. However, only a couple of women
expressed these concerns. Most participants viewed study participation as a no-lose
situation or did not view the potential risks as constituting a strong deterrent.
The risk of side effects specifically did not appear to be a major concern for
the participants when considering enrollment in the trial. However, different reasons
were cited for this lack of concern. Two women were not at risk for many of the side
effects due to previous hysterectomies, while others felt that since they were already
having similar symptoms due to menopause, they were not worried about the effects
of the trial drugs (which may mimic these symptoms).
You know, right from the first time they talk to you, they pretty much
assure you, you know, that so far these drugs haventyou know,
provid-, proved to have any side effects, except maybe hot flashes.
Well, we're all used to that anyway...
47


This particular participant also expressed a sentiment echoed by others in that
they felt they were made aware of potential side effects by the study personnel prior
to enrollment and were reassured about the likelihood of them occurring. The
voluntary nature of participation and the possibility of dropping out were also brought
up by a woman who was very familiar with clinical trial protocol in general.
The participants in fact mentioned only two specific potential side effects that
had concerned them prior to enrollment. The first were hot flashes, which were not
considered by any to be a deterrent (though in some cases did prove to be difficult
during the actual STAR experience). However, one woman also mentioned
endometrial cancer as the potential side effect that concerned her most. Interestingly,
she was also the only participant to mention that a friend or family member expressed
concern over side effects. Although she couldnt remember exactly what he said, her
older brother was not supportive of her enrollment, which she believed was due to the
risk of side effects in general and endometrial cancer in particular.
Studies by Bober et ai (2004) and Yeomans Kinney et al.(1995) found that
the decision against tamoxifen use or trial enrollment was associated with concern
over side effects. The results found here may reflect the fact that those who do have
strong concerns were underrepresented in this particular population given that all of
these women made the decision to enroll. Two participants expressed another
sentiment that perhaps accounts for the general dismissal of the side effects associated
with tamoxifen and raloxifene in other participants. These participants saw the threat
of breast cancer as much greater and more severe than any risks associated with
tamoxifen or raloxifene. A woman who had already dealt with cervical cancer and
LCIS in the past had this to say:
I was significantly more worried about the side effects of breast
cancer. And I was never diagnosed with having breast cancer. I did
not have breast cancer. I was in a pre-cancerous state. So my
48


concerns about the side effects of breast cancer seriously, seriously
outweighed any of my concern about the drugs- although I tried to be
knowledgeable about them.
Benefits that might be gained through study participation were more
numerous according to the results of this study (Figure 5.1). Five out of the
seventeen participants specifically stated that they saw no problem or downside to
participation with one woman saying .1 figured that I didnt have too much to lose
either way. STAR was evaluating the effectiveness of raloxifene for the prevention
of breast cancer, compared to the current standard of tamoxifen. However, despite
the fact that it was not yet approved for the purpose, participants were fairly certain
that they would experience benefits from raloxifene if assigned to that arm of the
trial. Time commitment was seen as a potential problem by one participant, however,
others viewed the screening and oversight provided by the study as a benefit to those
who were at risk for developing breast cancer, a view which is supported by the
literature (Lovegrove et al 2000). Other benefits of study participation will be
discussed in later sections.
49


STUDY PARTICIPATION
Figure 5.1. Perceived Risks/Barriers and Benefits of Study Participation
Interpersonal Factors
The Role of the Physician
The role of the physician in recommending study participation has been
heavily emphasized in the literature and in fact has been strongly associated with the
decision to participate in chemoprevention trials (Bober et al. 2004; Yeomans-Kinney
et aL 1998a). While the influence of physicians on womens decision-making could
be seen in the current study, the recommendation of a physician was not a universal
influence among participants. Eight women first heard about the trial from one of
50


their health care providers. Some of the women specifically remember their
physicians broaching the subject with them. Others could not remember the exact
occasion, but believed that was where they first heard about it. The following two
passages illustrate each of these scenarios:
But my doctor is pretty on top of things and uh, just around the first of
the year he notified, he called me and told me hed been looking into it
to decide. And he said that he could either put me on a regimen, a five
year regimen of tamoxifen himself or hed been talking to some of his
doctor friends and there was this STAR study and he thought I was a
real good candidate for it, if I wanted to go that route. But either way
he thought I should do something. Just because I was too high-risk.
So....I said Id do the study.
I believe it was mentioned. Dr. Jones was my doctor at the time and I
believe she did mention it. And so I went ahead and signed up!
In the first case, the woman's high risk for breast cancer made her a candidate
for tamoxifen, a regimen her physician wanted her to follow anyway. The only
decision to make, as she saw it, was whether or not to enroll in the study, not whether
or not to take chemoprevention drugs. Given that all of the women in the STAR trial
were at high-risk for breast cancer, other participants may have viewed the decision
from this vantage point as well; hers was not the only physician to frame the
discussion in this fashion:
Well yeah, Dr. Smith is my primary doctor, and he thought it was fine.
But then he said, Wait a minuteI want you on tamoxifen, I dont
want you on a placebo.,5 But then I explained to him that the other
drug wasnt a placebo. So I went on the study.
Physicians were not the initial source of information about STAR for all of the
women interviewed. In two cases, the women themselves heard about the study
51


through other sources such as the media or someone else already involved and
brought it to their providers attention, while two others couldnt remember who
brought it up first. Regardless of who broached the subjectthe majority (n13) of
women stated that they discussed the study with their physician prior to making the
decision to enroll in the trial, who then provided largely informational support in the
form of both recommendations and reservations during these conversations. Nearly
all physicians were generally supportive of the decision to enroll. The primary
reservation expressed by physicians according to the women interviewed involved the
two cases mentioned above in wmch physicians wanted their patients on a
chemoprevention drug due to their high risk for breast cancer, as opposed to taking a
placebo or doing nothing. One physician mentioned possible side effects, but
according to the participant, he did not view the side effects as a large risk:
His reaction was with, you know, anything like this there are gonna be
side effects. There are side effects if you walk across the street and get
hit by a truck.
While physician recommendation played a role in several womens decisions
to enroll in the trial, the strength of that influence seemed to vary between
participants depending on their relationship with the physician. Some participants
simply mentioned that their physician thought they should enroll, or asked if they
would be willing to participate. Another mentioned seeking out a physician she had a
clinical relationship with for their opinion when another physician told her about the
study. Two physicians knew their patients' personalities well enough to take a more
personalized tack in addressing study enrollment. For example, one participant with a
long history of working for social justice causes and with underprivileged women
said that her physician appealed to her passion for this work.
52


...and I talked to my oncologist. And the upshot of his
recommendation was, he said, Its for the greater good. And that
cinched me; that was it. That did it.
Another appealed to the proactive nature of his patient with whom he had a long and
close relationship due to other health issues she had had in the past.
And when I had the breast tumor and had the diagnosis and my doctor
explained the STAR trial, he even said right up front that he thought
this would be ideally suited for me because he knew, it, it would have
just been impossible for me to just sit and say Wow, Ive just gone
from very, very low risk of breast cancer to very, very high risk.. ..and
Im not going to do anything about it? II just, that just would not
have been good for me mentally... Well, again, my gynecologist/
obstetrician at the time recommended it to me. And he knew me well.
I mean, my, my son was a vaginal delivery after cesarean which at the
time virtually nobody, there was only one hospital and like, one doctor
that would even let you do that, so this was the doctor who had seen
me through that. And then when I was diagnosed with cancer a year
after that, um, he made my cancer diagnosis. He donated blood for me
while I was in the hospital...and so I so respected his opinion and had
such trust in him. The fact that he said he thought this would be really,
really good for me almost made it a done deal.
The importance of physician recommendation has already been established
by the quantitative literature on decision-making for chemoprevention trials. The
results of this study provide context for these findings and suggest that while not all
women consult their physicians prior to making the decision to enroll, for those who
do this is an important conversation. However, the data gained here also indicate that
rather than a one-sided relationship existing between physician and patient in which
the physician possesses all of the information, some women brought the study to their
physicians attention or arranged for their physician to speak to CCRP for more
53


information, perhaps reflecting a more shared treatment decision-making model
(Charles e/a/. 1999).
Prior Experience with Breast Cancer
Prior experience with breast cancer among others in their social network was a
frequent topic of conversation during the interviews conducted for this study. Only
two participants said that they had known those with breast cancer, but as one put it,
no one I was emotionally close to. Twelve had a family history of the disease,
while one without any family history had had a daughter diagnosed with breast
cancer. As one woman put it, I dont think you could go out into this library and
find anybody who doesnt know someone that has gone through this. This was an
important factor because womens previous experiences with breast cancer likely
shaped how they viewed their own risk for the disease as well as its severity.
Therefore, though perceived risk for breast cancer and perceived severity of the
disease are individual-level beliefs that have been examined in the literature, these
beliefs may be influenced by experiences at the interpersonal level as discussed here
(Bober et al. 2004; Lovegrove et al. 2000; Yeomans Kinney et ai 1998b).
The prognosis, treatment, and outcomes of family members or friends with
breast cancer were influential in how women conceptualized the disease. In a few of
the cases of women who had had a family member with the disease, this had spurred
them to take further action to monitor their own risk by seeing an oncologist or
attending a high-risk clinic. While some had only known one or two who had gone
on to recover, others had lost multiple people to either breast cancer or cancer in
general, with some experiences falling between the two extremes.
my mom and my aunt both had breast cancer, they obviously had
the kind that is very slow growing and it was found early in a
mammogram and treated, and so thats more than ten years ago.
54


And people, you know, we knew about it and thought, in fact, she
thought she was in remission, and thought she was gonna be okay.
And then she had a setback and died.
The issue of being proactive about one's health or, on the other side of the
coin, being in denial about one's health, emerged through these discussions. Two felt
that enrolling in the STAR trial was something they could do to take a proactive
approach to their own health. While only three women brought up the issue of denial,
it seemed to have had a powerful impact on them. They had friends, acquaintances or
family members who had waited too long to seek screening or treatment and as a
result had a worse prognosis when they eventually did go in. In all three cases this
person had had suspicious symptoms or felt ill for some time before seeking care.
One woman was not as close to the breast cancer patient she knew and seemed
to primarily feel empathy or pity for her. However, the other two participants were
referring to sisters and expressed a sense of frustration that their siblings ignored what
was going on. Interestingly, both felt that their sisters should have known better and
that the behavior was out of keeping with their character:
And the irony of the thing is that Carrie had an abnormality on her
breast in February and did not do anything about it until August. And
thats one question I wished I had asked her... Why in the world did
you not? You know, I mean, you were salad dressing on the side
and oh no, you shouldnt eat that and all this stuff that she did for
years.. .and yet she had invasive breast cancer, which showed up as a,
like the classic orange peel kind of thing. And she did notice it- she
was on a trip to Seattle or some place and noticed it. And then never
did anything about it until August. And by then it was Stage IV. And
you know, it, it was always, thafs just a mystery that I will never have
an answer to.
55


Having Female Offspring/ Sisters
Interpersonal relationships were also influential in how women perceived the
benefits of enrolling in STAR. Benefits were seen to accrue not only to the
participant herself, but also family members as well. Having female children was an
important consideration in participants'1 decision to enroll in the trial due to the
potential knowledge that could be gained for future cancer treatment. For seven of
the women, the threat breast cancer posed to their daughters and the desire to further
research that might someday protect them were important factors to consider. Just
one participant had a daughter who had already been diagnosed with breast cancer.
She had undergone a mastectomy and chemotherapy, and was well past the five-year
survival mark. However, the fact that her daughter was diagnosed with breast cancer
at the young age of 27 had been a frightening surprise. She had not realized that such
a young woman could develop breast cancer, particularly without a family history,
and called it, the scariest thing that ever happened.
Only three of the ten women who had daughters did not cite this as an
influential factor in their decision. One of these participants said that her daughter
was out of town at the time she was considering enrollment and therefore she did not
discuss it with her, though she did not say either positively or negatively whether the
fact that she simply had a daughter influenced her at all. Another who had worked in
the field of cancer screening and prevention felt she would have participated with or
without having had a daughter. The third had an adopted daughter and no family
history; she simply said that having a daughter did not influence her decision 4tat
Having female siblings was also important in two ways. While twelve women
overall had at least one sister, six of these had a sister who had been diagnosed with
breast cancer. One woman said, tLWhen it happened to Carrie, it was just like it
happened to me,illustrating how a sisters breast cancer may have shaped the way
56


women viewed breast cancer. Others felt having a sister was influential in a more
indirect way; by participating in research that could increase our knowledge around
breast cancer prevention they were potentially helping their sisters who could be at
risk for the disease in the future. A sister with breast cancer is also a first-degree
relative. This therefore may have made women more aware of their own risk, such as
one participant who sought out an oncologist for herself after her sister was
diagnosed.
Lovegrove et al. (2000) did look for an association between number of female
offspring and the decision to enroll in a chemoprevention trial but did not find the
association to be statistically significant, despite the fact that women who chose to
enroll did in fact have more female offspring than their counterparts who made the
opposite decision. Other studies have not examined this variable in depth, but the
connection between having female offspring or siblings and choosing to enroll in
chemoprevention trials is an important one to explore further for several reasons.
First of all, one of the key risk factors for eligibility in the STAR trial is having a
first-degree relative with breast cancer which means that some sort of family history
of the disease and its associated risk is common among the subject pool whether or
not that history is due to actual heritable genetic mutations, such as the BRCA 1/2
mutations. The results of this study seem to indicate that this is an important
contextual factor for many of the women who chose to enroll in STAR. Second, in a
socio-ecological framework, having a sister or daughter diagnosed with breast cancer
may likely be the primary experience they have had with the disease and therefore
shapes a womans perceptions of how severe breast cancer is and how large a risk it
constitutes. It seems reasonable to argue that a family history that includes a sister or
daughter with the disease may make the risk of breast cancer feel more intense than
having had a relative that was farther removed, such as a grandparent or aunt.
57


Social Support
The discussion around social support on an interpersonal level was complex.
It involved support given both from others to the participant, but also from the
participant to others in their lives. Nor was social support always discussed within
the context of the decision-making process for enrollment in this particular trial.
Women often brought up support they had given to others with breast cancer either in
the past, or during their enrollment in the trial, while support was received by
participants before the decision-making period, during, and after they had decided to
enroll in the trial. Figure 5.2 illustrates some of the examples of different types of
social support brought up during the interviews given by and to participants. Some of
these acts actually took place during the period of enrollment and will be discussed in
Chapter 6. However, most took place either during the decision-making process, or
were influential in the participants past. The specific role of important people in
ones social network, including a spouse, family member, or friends will be
discussed.
58


SOCIAL SUPPORT
Given By Given By
Others to Participant
Participant to Others
Emotional:
Family & friends being
supportive of enrollment
decision
Family & friends offering
concerns about enrollment
Camaraderie offered by other
participants
Nurses sympathizing with
participants who are
experiencing side effects
Instrumental:
Spouse enabling trial
participation
Physician cooperating with
trial personnel
Nurses providing reminders
during the trial
Informational:
Physician recommending trial
enrollment
Other breast cancer patients
or survivors offering advice
Study nurse telling potential
participants about study
Appraisal:
Physician discussing risk
factors with participant
Emotional:
Offering support to others who
had breast cancer
Attending appointments with
those who had breast cancer
Joining in a breast cancer walk
Instrumental:
Taking on extra duties for a
work colleague with breast
cancer
Enrolling in the trial as a way
to help family members
Informational:
Researching treatments for
another with breast cancer
Appraisal:
o Telling a daughter about her
risk for disease
Key:
Occurred during decision-
making process
Occurred before decision-
making process or after trial
enrollment
Figure 5.2. Types of Social Support Mentioned by Women in this Study
59


Spouse
Thirteen of the seventeen participants were married at the time of the
interviews. None mentioned a partner other than a spouse or indicated the presence
of one on the Demographic Questionnaire. When participants were asked if they
discussed the study with their spouse or partner prior to enrolling in the study, a few
seemed affronted by the question and emphatically stated that it was their decision, or
that they told their spouse about it but did not ask for permission. Eleven of the
thirteen, however, felt that their husbands were supportive of the decision overall,
though some were more enthusiastic than others. Two husbands had cancer
themselves in the past, which partly contributed to their support.
Primarily this support came in the form of emotional or informational support.
However, one participant in particular felt that her husbands instrumental support
was also essential to her enrollment and subsequent participation in trial activities.
She and her spouse spent a significant portion of each year out of the country at their
other home, which was in a remote location and was only accessible by boat. While
she was able to work out a situation with the study personnel here in the states where
she could fulfill her study requirements outside of the country, she did not feel
comfortable dnving herself alone by boat to her appointments. Although he had to go
to more trouble than most of the husbands of the women I interviewed, this woman's
husband was still very supportive.
Other Family Members
A few other family members were also mentioned in response to this line of
questioning, such as siblings or children. However, while some participants shared
their plans with these family members, most did not feel that they played a large or
significant role in their decision. One participant did have a brother who expressed
60


reservations about her decision to enroll due to concern about side effects, while the
others felt supported in their decision.
Friends
Only four participants could remember discussing their decision at length with
a friend prior to enrolling in the trial. The response two of them received was an
expression of concern regarding potential side effects or the time commitment.
Another had a friend with breast cancer who was taking tamoxifen. While she said
that in the end she would have participated regardless of the womans reaction, she
also delayed saying yes to the study until she could discuss this womans experience
on tamoxifen with her. One participant had a particular friend who had been part of
her health journey since the participants diagnosis of LCIS. This friend had been
with her when she received the diagnosis and was supportive of her decision to enroll
in the trial, which the participant partially attributed to them both being educated
people.
The types of social support provided by friends and family members primarily
took the form of informational and emotional support. Overall, however, women
seemed to feel that the decision to enroll had been theirs alone, and while they may
have discussed it with others, they were not necessarily swayed by advice from
outside. This perhaps reflects the larger cultural tendency to believe that health is the
responsibility of the individual discussed in Chapter 3 (Clarke et al 2003; Fisher et al
2005). Of the advice they did receive, it appears that the recommendation of their
physician was the most significant. It must be remembered however, that these
women all chose to enroll in the trial.A different picture might be gained from a
sample of women who chose not to enroll in the STAR trial. What did appear to be
influential, however, was past experience with breast cancer through the cases of
61


friends and family members. Many of the women who enrolled in STAR had been
sources of support themselves for others who were undergoing treatment for the
disease, not only emotionally, but providing informational and instrumental support
as well which informed their own perceptions of the disease.
Institutional/ Organizational Factors
Placebo vs, Non-Placebo Trials
The study protocol of STAR could be considered part of an institutional level
of influence in that womens decisions to enroll in a chemoprevention trial may
depend on its design. One of the most important themes that arose during the course
of the interviews conducted during this study was the significance given to the fact
that the STAR trial was a non-placebo arm trial. This is a topic that has only very
briefly been touched upon in the literature but should be included in future studies
given the results found here (Yeomans Kinney et al.1995).
Six women said that the nature of the STAR protocol was an influential factor
in their decision due to the fact that they knew they would be getting one drug or
another, rather than a placebo. Knowing the benefits of tamoxifen in preventing
breast cancer and of raloxifene for the prevention and treatment of osteoporosis, and
having good reason to think that raloxifene would likely have a similar effect on
breast cancer, the STAR trial was typically seen as testing two drugs each with their
own benefits against one another, rather than as testing a proven drug against an
unproven one. Womens words therefore reflected this view:
So it seemed like a no-lose situation.
And um, you know, I think it was helpflil to think that this drug would
possibly have the potential to help me. And the one that was
62


approved, theyM already determined that it was helpful. The one that
was being tested, the information I read about it indicated that it
probably was even, was going to be better than the one that had been
approved. So I figured that I didnt have too much to lose either way.
The participants were not alone in feeling this way. In addition to the physicians who
expressed concern in the above section, several of the participants9 spouses expressed
similar feelings about the possibility of their wives enrolling in a study that could
potentially not benefit them. These individuals were usually reassured when told
there was no chance of participants being assigned a placebo.
In discussing their reluctance to enroll in a placebo-arm trial, several women
elaborated on just why they felt it would have altered their decision, citing wasted
time and effort, or a sense of being cheated.
One of the reasons I like this trial was that there was no placebo. I
would be getting something, that it was a test between two drugs. And
that very much affected my decision to join because it seems to me
after youve been in a study, and lets just say that a drug was very
beneficial to the people taking it and you found out you were on the
placebo. It seems to me there would be a dis- a sense of
disappointment; of almost of being cheated that you didnt get the
drug. Although, it seems to me also when you agree to be in a study
thats the chance you take. But, so I liked the fact that 1 was going to
get something.
Its a lot of, theres a lot of, uh, effort on your part to be in the study
and if youre thinking youre probably taking a placebo, why bother?
Not all women felt it would have changed their decision had they been
approached to be involved in a placebo-arm trial. One woman talked about the
benefits her own mother received from tamoxifen when she had breast cancer.
Another woman cited the need for dummy trials as well as non-placebo arm trials.
Generally though, women discussed the lack of a placebo as a factor influencing them
63


to participate in this particular trial. Yeomans-Kinney et aL (1995) also noted an
association between nonparticipation and concern about being assigned to the placebo
arm of the trial. However, in that case, the trial in question was the Breast Cancer
Prevention Trial, which was in fact testing tamoxifen against a placebo. Further
quantitative research is needed to explore whether women who list concerns about
placebos as a deterrent to participation are in fact more likely to participate in non-
placebo trials. The results here seem to indicate that this may be the case.
Tamoxifen Versus Raloxifene
The majority of the participants reported an assignment to the tamoxifen arm
during the trial (Table 5.3); as a result, tamoxifen was discussed during the interviews
with greater frequency than raloxifene. Participants also knew more about tamoxifen
as the FDA had already approved it and many knew others who had taken the drug in
the past for treatment, often with positive results. One, for example, had had a friend
diagnosed with breast cancer who had undergone a mastectomy and been treated with
tamoxifen. She had reached the five-year mark and was still in remission. Another
participants mother had a similar positive experience.
Table 5.3. Dru^ assignment during STAR
Drug Number of Participants
Tamoxifen 12*
Raloxifene 5
Total 17
* One of these twelve switched to raloxifene during the study
However, not all knew friends or family members who had had such good
results from tamoxifen; other stories were told as well.A friend had been treated
with tamoxifen but relapsed. A sister died during the study, despite being treated
64


with tamoxifen. In the case of the second, the participant changed her study drug to
raloxifene mid-trial due to concerns about tamoxifen^ ineffectiveness resulting from
her sisters experience.
There was some speculation during the trial by participants regarding which
drug they were on. The intensity of hot flashes experienced by one woman led her
physician to suspect she was on tamoxifen. Only two women mentioned the fact that
they were given the opportunity to change the drug they were taking, once the trial
had been un-blinded. Of these two, only one chose to do so- the woman discussed
above whose sister had died.
One participant went so far as to compare her drugs with a friend's tamoxifen
and a family members raloxifene. The drugs she was given looked similar to her
sisters raloxifene and it chalked, a trait she associated with raloxifene. Therefore
she was surprised to find out in the end that she had in fact been on tamoxifen. Why
did she go to such trouble? She perceived the potential side effects of tamoxifen to
be more worrisome than those of raloxifene: I didnt think I had to worry about the
side effects of tamoxifen, until five years later, or six years later, when I found out I
was on it!
She was not the only one to consider raloxifene the good one, as one
participant put it. Several of the women assigned to the raloxifene arm thought of it
as the preferable drug, either due to research they had done prior to enrollment or due
to the results of the study. One noticed an improvement in her bone density during
the trial, which she ascribed to being on raloxifene.
65


Community Factors
Altruistic Motivations
One of the most important themes that emerged from the interviews
conducted during this study was that women who enrolled in the trial did so in order
to help others. Fifteen of the seventeen women explicitly stated that they were
motivated to enroll by altruistic sensibilities while making their decision. This was
framed in multiple ways, for example as wanting to help other women to promote
knowledge, protecting other people, and being usefliL A couple saw it as an
obligation or duty; as one said, Somebody has to do this stuff.
A strong sense of the importance of finding better treatments or perhaps even
a cure someday was revealed in these interviews; of women doing what they could to
help other women, whether stranger or relative. Nor did women believe they were
the only ones who felt this way:
Well,I mean, I think that as a general rule, we, as women, want to
make sure that if weve gone through this, someone else doesnt need
to go through it. And if theres a way thatyou know, studies can
show that it not only helps that but helps other things like your bones
and different things that its certainly a benefit to everybody.
Interviewer: And you said, when I called you, something on that, I
wrote it down cause you said, I think youre talking to a group of
pretty altruistic people.
Participant: Yeah.
I: And you think that is true of all the other women in the study too?
P: Oh yeah. Yeah. Because you're not really doing it for yourself. I
said because it was already proven to be a good drug. But its really
for our sisters and our mothers and whatever.
A couple of the women had backgrounds in womens health advocacy or
social justice, which motivated their enrollment. One had worked for an organization
66


active in preventing breast and cervical cancer and had also been active in trying to
link under-served women with preventive care programs providing Pap smears and
mammograms. She not only understood the importance of prevention and screening
for these diseases, but also the benefit to herself of a study that provided these things
for free. This particular woman saw her previous work with vulnerable women as the
most important motivating factor to her participation. A second participant viewed
enrollment in the STAR trial as a logical decision in a lifetime of fighting for social
causes:
Fm kind of a weirdo um, when it comes to civil rights and things like
that. I was at Martin Luther Kings thing in 1963 in Washington. And
Tve been working for an AIDS organization for twenty years now. So
I started out when it was all gay men. And something about social
welfare is very important to me.
The theme of altruism ran very deeply throughout these interviews. This
echoes collective cultural themes in the United States that emphasize personal
responsibility, yet also a shared duty to help others. Cultural themes belonging to a
society are typically widely shared and provide a common point from which to draw
meaning and understanding of life events (Becker 1994). Siminoff and Arnold
(1999) point to similar ideas of altruism reflected in decision-making for organ
donation also. That this value of altruism, of wanting to help others, was not only
espoused by the majority of these women but was also assumed by the women
themselves to be a shared value illustrates just how much it permeates their social
environment. The women quoted above don't just ascribe their desire to help others
67


to their own distinct personality or to their own unique background, but rather
identify it as a trait common of woman as a group, or at least of subgroups of women.
Moreover, that trial participation was viewed as a method through which a
desire to help others could be fulfilled may reflect other shared cultural values.
Becker, in her study on the disruption infertility can cause in the lives of men and
women, points to the importance of the idea of continuity in the United States
(1994:401). Americansshe argues, value order in the universe, productivity and
progress and struggle to make sense of disruptions to this order. Illness or other
unforeseen events are just such disruptions. Becker writes,
The emphasis on will, that hope and determination may will a change in the
course of an illness, reflects efforts to control and order illness through
rational determinism and creates a specific cultural response to illness: to fight
the illness and overcome it (1994:396).
The women in this study do not have breast cancer, yet it is a very real presence in
many of their lives just the same. Enrolling in a chemoprevention trial could, on both
a personal and a social level, provide a tangible way for participants to fight the
disease and to change its potential course in both their own bodies and those of other
women.
Other Breast Cancer Activities
Although study participation was tied to a desire to help others, this was not
necessarily associated with participation in other breast cancer activities in the minds
68


of the women I interviewed. One woman was quite emphatic about this and viewed
the study social events as being in a similar vein to the various fundraising walks.
I get their information about the stuff and I just throw it away. I guess
partlyldontwantljustdontwanttojointheclub.lreallydont-
and I dont go to the March for the Cure or anything. UmI have no
desire to do that". I just dont want to deal with it at that level.
While other women had donated to friends or acquaintances participating in
fundraising in the past, only one was currently involved in fundraising efforts. Five
had participated in various breast cancer walks, but several felt that they had become
too large now and the crowds were a deterrent. Four out of the five who had
participated in walks had a family member diagnosed with the disease, while the fifth
had LCIS in the past.
Social Structure, Policy, and Systems Factors
Media
While it was not a focus of the interviews, multiple participants mentioned the
role of the media as a source of information about the trial or about breast cancer.
Two of the participants had initially heard about STAR through a print article or
television adwhile at least three others couldnt remember for certain but believed
that was where they might have seen it first. These women were then proactive about
finding out further information, again reflecting a sense of individual responsibility
for personal health. When asked how knowledgeable they were about breast cancer
many of the participants cited newspapers, magazines, newsletters, books, and the
Internet as sources of information. One woman noted the increasing amount of
information on the subject, saying, Theres been more to readfor one thing. The
books of well-known personalities within the breast cancer movement, such as Nancy
69


G. Brinker of the Susan G. Komen for the Cure Foundation and Dr. Susan M. Love
were brought up as sources of information that were tapped.
Other Factors
The discussions of motivating factors that took place during this study
primarily focused on more proximal levels of influence that that of social structure,
policy, and systems factors. The one specific factor that was just briefly touched on
was the media. Yet while the women in this study did not discuss it, the very fact that
they were able to participate in STAR related to changes at the policy level. Prior to
the initiation of the CCOP network by NCI, clinical trials were typically located at
large research centers or teaching hospitals. Enrollees in these trials consisted of
those who lived in the vicinity or had the means and ability to travel to these centers
(CCRP 2009b). However, the system of CCOPs changed that by taking clinical trials
to the local level and allowing for increased opportunities for enrollment.
A second important factor that cuts across the community, institutional, and
social structure level is the impact of the history of clinical trials in the United States.
Women in this study, who nearly all identified themselves as White9, in general held
very positive views of medical research, seeing it as an important source of
knowledge. Furthermore, they felt that medical research was typically conducted
with benign intentions- no participant questioned the motivations of those conducting
the STAR trial or of CCRP. This may not be the case with all groups. Certain sub-
populations may be suspicious of the medical profession and the research it conducts.
9 White is the term currently used by the U.S. Census Bureau and so is the term
used here. However, on the Demographic Questionnaire, participants were asked to
write in their race or ethnicity on a blank line. Participants responded with multiple
70


An often-cited example of this would be the legacy of Tuskegee within the African-
American population. A long history of racial discrimination in the United States, of
which Tuskegee is just one example, has contributed to a distrust of the medical
profession among many African-Americans, a fact which has inhibited screening and
treatment for multiple conditions, breast cancer included (Armstrong et ai 2005;
Thomas and Quinn 1991). Other sub-populations may struggle with language
barriers, or understanding study protocol or the intentions of medical research, such
as in a study by Nguyen et ai (2005) with Asian-American women.
Socioeconomic status should also be considered here. In examining
knowledge of reproductive technology among a diverse sample of women, Rapp
noted that "virtually all middle-class women knew about the test from a dense and
overlapping nexus of friends, medical professionals, and books2000:114). A
similar phenomenon was found in the present study. All of the women in this study
were fairly well educated, with several having advanced degrees. More than one
woman mentioned her education contributing to her understanding of either the
STAR trial itself, or medical research in general. Whether or not their knowledge of
breast cancer or chemoprevention was high, most of the women possessed the skills
to research the information they needed prior to enrolling. Furthermore, they Several
worked in occupational fields that exposed them to information about cancer.
Furthermore, all lived in what could be classified as middle-class neighborhoods, and
cost was not cited as a barrier to participation. The results of this study may have
been very different if conducted among a low socio-economic status population.
The criticism that breast cancer discourse in the United States is largely
focused on the interests of white, middle-classprofessional women has been noted
termswhich included White, Caucasian, Anglo, or European American.
Two participants left the question blank.
71


in the past (King 2006:x). The population of this study mirrors these characteristics.
While this criticism has generally been aimed at a movement that has focused on
finding a cure for breast cancer, that it has relevance for chemoprevention is clear.
Cancer clinical trials have had a difficult time recruiting minority participants and
further research is needed to identify important factors for their decision-making
(Advani et al 2003; Nguyen et ai 2005). While there may be cultural beliefs
associated with trial non-participation, Advani et ai (2003) identified income and
education as important factors as well. In that study, African-American patients had
lower education and income levels than their white counterparts; both were shown to
be factors associated with patients being less willing to participate in a clinical trial.
Among Asian-American women, economic and language barriers were also found to
be important (Nguyen et al 2005). These issues must be addressed if
chemoprevention trial populations are to be more diverse than they currently are.
Given the framework of a socio-ecological model then, it is possible to see
how each of these levels of influence fit together to impact womens decision-making
for enrollment in the STAR trial. Figure 5.3 illustrates those factors at each level that
resulted from the interviews conducted here, as well as those identified at the social
structure and policy level that were pertinent to the population in this study. The
institutional/organization level was more influential in examining women's
experiences while enrolled, which is when womens primary interactions with CCRP
took place.
72


Social Structure. Policy & Systems:
Media
Transition of Clinical Trials Out of Research Centers into Communities
History of Medical Research in tlic United States
Socioeconomic Status
ComnuinitY
Altruistic Motivations
Values & Characteristics Shared By Members of a Social Network
: L
_
-.,i * 1!>
__

Interpersonal:
Relationsnip wnh Physician and Physician
Recommendation
Prior Experience with Breast Cancer Among Social
Network
Having Sisters or Female Offspring
Spousal Support
Individual:
Perceived Risk of Breast Cancer
Perception of Medical Research
Perceived Risks & Benefits of Study
Participation
Education Level
!^.


figure 5.3. Socio-Ecological Framework- Factors that Influenced Women1
)ecision-Making While Considering Enrollment in STAR
73


CHAPTER 6
PARTICIPANT EXPERIENCES WHILE ENROLLED IN THE STUDY
OF TAMOXIFEN AND RALOXIFENE
During the interview sessionsconversations naturally turned to womens
general experiences while they were enrolled in the trial. These generally centered
around three levels of influence: individual, interpersonal, and institutional. While
these experiences do not address the specific aims of this study, the information given
may be useful for understanding those factors that make continued study participation
easier or more difficult for women, as well as what aspects of study protocol were
particularly valued by women.
Individual Factors
Despite participants generally reporting a lack of concern regarding the
potential side effects of tamoxifen and raloxifene prior to enrolling in the study, ten
did in fact experience symptoms during the study period. Seven of these ten however
were not positive whether these symptoms either qualified as side effects or if they
were actually related to age, menopause or other health issues, rather than the drugs
themselves.
Symptoms brought up in interviews and the number of women who
mentioned them can be seen in Table 6.1.While most women only experienced one
symptom, three experienced between two and four different symptoms during the
study period. By far the most commonly reported issue was hot flashes, which
ranged from warm flashes, to significant, to horrific depending on the woman
being interviewed. One woman experienced such severe hot flashes she nearly had to
drop out of the study.
74


Nor were side effects restricted to one drug or the other, with women on both
tamoxifen and raloxifene reporting them. While it is true that participants on
raloxifene only reported two of the symptoms (hot flashes and vaginal dryness), it is
difficult to draw any conclusions from this given that over twice as many participants
in this study were on tamoxifen as were on raloxifene. Again, these are symptoms
reported by women during the study period, and which women believed might be
associated with either tamoxifen or raloxifene use. However, in some cases there was
no verification of that fact.
Table 6 Symptoms experienced by women during STAR
Symptom Number of Number of Symptoms Number of
Experienced Women Affected Experienced Women Affected
Hot Flashes 7 Any Symptoms 10
Night Sweats 1 Only 1 Symptom 6
Vaginal Discharge 1 2 Symptoms 1
Vaginal Dryness 1 3 Symptoms 1
Hair Thinning 1 4 Symptoms 1
Charley Horses 1
Weight Gain 1 * One participant was not specmc enough
Nails went bad 1 to categorize
Interpersonal and Community Factors
Nurses
Participants spoke with fondness and appreciation of the nurses they worked
with during the study period. Others have noted the importance of social support
during times of stress associated with health, particularly instrumental and emotional
support (Bloom et al. 2001; Kawachi and Berkman 2001; Thoits 1995). The nurses
provided not only information and answers to questions but also offered emotional
75


support when women were struggling with intense hot flashes or other symptoms.
Participation in the STAR trial required keeping periodic appointments and nurses
provided reminders with phone calls or letters. Several women reported developing
friendships with their original nurses and were disappointed when those particular
nurses left the study, which may be illustrative of the fact that women tend to
maintain more emotionally intimate relationships than menICawachi and Berkman
2001:461-2). One participant in fact, said that she did not have as personal a
relationship with her second nurse as she did with her first, which was one reason for
her dwindling participation in necessary study appointments. The only complaint was
made by a participant who felt that she was given vague answers when she asked her
nurse if physical symptoms she was having could be due to the drug she was taking.
However, she also acknowledged that this was not necessarily the nurse?s fault:
Sometimes I wish, um, they could tell me if something was a side effect or
not, but Pm not sure they knew. I mean, how would they know cause they
didnt have all the study results?
Social Events
The social and informational events held throughout the study period were
another source of support for participants. Many women mentioned the annual
luncheons they attended, and the benefits they gained from these occasions; some
even brought friends or family members with them. The informative nature of the
speakers and study updates these occasions provided were appreciated. In addition,
participants, in particular those who had enrolled at the beginning of the trial when
the group was smaller, felt that these meetings encouraged a sense of camaraderie
among the women, who were able to speak with others going through the same things
they were. One woman spoke of having done the Komen Foundation fundraiser walk
with some of the other STAR participants early in the trial, which she remembered as
76


being organized by CCRP. This is in keeping with previous research into social
networks, which has illustrated that being a part of a larger web of social relationships
increases the type and breadth of support an individual may access when in times of
stress (Bloom et ai 2001; Kawachi and Berkman 2001). Although the social events
were not held often, for some of the women in the study, they offered unique
opportunities to be around like-minded individuals who were taking part in the same
experience that they were.
One of the key motivations women gave for enrolling in the STAR trial was
an altruistic desire to help others which is consistent with studies on clinical trials in
other areas of medical research (Simon et ai 2006). As a follow-up question,
therefore, participants who attended these social events were asked how they felt their
reasons for joining the study compared to other women they spoke to. Two women
noticed a similar desire among other participants to try to help other women, while
another three felt there was a clear connection between family history and
participation. One woman in particular spoke expressively on the subject:
Well, one big difference was that I had no history in my family of
breast cancer. And at some of those very earlier meetings, when we
would go around and introduce ourselves it was not at all uncommon
to hear a woman say, Hi, Im so and so. Ive lost a mother, two
sisters, and a daughter to breast cancer. And we would go around the
room and almost everybody in the room would have a history like that.
I mean, I was practically sobbing. And I remember the very first time
that I introduced myself and I said, ul really feel like an anomaly
because I have no history of breast cancer in my family, although of
course Im terrified for my daughter right now."but I was definitely
an anomaly. Almost all of the other participants had this daunting
family history of breast cancer.
The feeling that lack of family history made one an anomaly in the study was
not unique to this participant. Another participant did not attend the luncheons
77


because she mistakenly thought that they were geared towards those who had had
breast cancer themselves. Others gave reasons such as lack of time, work
commitments or, for those who lived outside Denver, the driving distance for their
lack of attendance.
Physician
Most of the women did discuss the trial with their physician prior to enrolling,
as discussed in a previous section. While most were supportive, it was not universal.
Not all women based their decision on the attitude of their physician; however the
attitude of the physician did make a difference in one womans ability to follow
through with the study requirements. This participant was very well educated and
understood the study protocol; however, her physician would often make it difficult
for her to obtain the required tests by refusing to order them or hinder her ability to
make appointments with the required people. She felt he did not share her concerns
about breast cancer and was primarily concerned about keeping costs down.
Interestingly, she also felt he did not take her concerns seriously in other areas of her
health as well. In contrast, another participant felt study participation was fairly easy
because her physician was very cooperative with the study requirements.
Organizational Factors
Given their role in the study, it was not surprising that the Colorado Cancer
Research Program came up in interviews. Nearly two thirds of the women
interviewed expressed similar sentiments regarding their interaction with study
personnel, viewing them as supportive and efficient, but not overly intrusive, and as
making the participation process easier for enrollees.
78


...I would just say I think that the whole, you know, this organization,
and the whole program is run, was run, very efficiently and effectively
and with a great deal of heart.
And I was always made to feel that this was a very good thing I was
doing, and they were very appreciative.
I think theyre very good people and they, they really like what they
do, I think, and that makes a difference.
Specifically, women appreciated reminders about appointments, the follow-up
process, and the excellent rapport CCRP established with their clients. One in
particular, appreciated the time they took to keep her personal physician informed
about the trial. Another, who spent a significant period of time out of the country
during the study period, felt that they worked very well with her, despite this fact.
In fact, the only concern mentioned came from a few participants who
experienced a change in the study nurse they were assigned to. One said that she
thought she was on her fourth nurse, while another felt that the lack of face-to-face
contact she had with the second nurse she was assigned made her less apt to follow
through on required study activities. The issue of turnover, however, seemed more of
a problem for women who had felt a closer bond to their first nurse than the
subsequent one. Another participant had two nurses but did not see this as an issue:
I just loved them. I had two different nurses and uh I just liked them
both so much. IM1, Til miss her! You know, cause we saw each other
twice a year and I liked that a lot. Nice, nice gal...cant say enough
about them.
In general women appreciated the overall helpfulness and support, which they
experienced in working with CCRP.
79


CHAPTER 7
CONCLUSION
The primary aim of this study was to explore factors that influence womens
decision-making when considering enrollment in a chemoprevention trial for breast
cancer, such as the STAR trial. Given the focus in the literature on individual level
factors, a secondary aim was to identify any factors in the decision-making process
related to social networks or social support. An approach to studying health
behaviors that only considers individual level factors ignores the larger environmental
context in which these individuals live out their lives. This study demonstrates that
while individual beliefs and attributes are important in enrollment decision-making,
factors at the interpersonal, institutional/ organization, community level, and social
structure, policy and systems levels cannot be ignored if researchers hope to increase
enrollment in breast cancer chemoprevention trials.
During the decision-making process, the influence of all five levels in the
socio-ecological model could be seen, with the individual, interpersonal and
community levels the most pronounced. Individual factors included perceived risk of
breast cancer, perceived risks/barriers and benefits to study participation, perception
of medical research and physician recommendation. Perceived risk of breast cancer
was influenced not only by objective measures of risk such as the Gail model but was
also shaped by participants previous health history, family history, and prior
experiences with breast cancer among others in their social networks. The benefits of
study participation outweighed the potential risks or barriers, such as problems with
side effects, managing the time commitment, and the ability to follow through with
study activities. Potential benefits identified included the medical oversight of
participant health, preventing breast cancer and osteoporosis, the knowledge that one
80


was getting a drug rather than a placebo, and the potential to help others and further
medical knowledge.
Interpersonal influences on health included prior experiences with breast
cancer among friends or family members, physician recommendation, having female
offspring or sisters and, to a lesser degree, the support of a spouse or friends. The
importance of physician recommendation to women in this study is supported by the
literature (Bober et al 2004; Yeomans Kinney et al 1995; Yeomans Kinney et al
1998a), though it was not universal nor was it a one-way relationship. Some women
did not consult their physicians, while others were the ones to bring the study to their
physician's attention rather than the reverse. Regardless of who initiated the
conversation, most women did discuss the study with their health care providers prior
to enrolling.
Women with female offspring overwhelmingly considered this to be
influential in their decision-making process. Several spoke with their daughters prior
to enrollment, while others just felt that participation was something they could do to
potentially protect their daughters in the future. A similar sentiment was expressed
regarding sisters in that the results of the STAR trial could inform prevention and
treatment for their sisters in the future if they were ever at risk or diagnosed with
breast cancer. For those women whose sisters had breast cancer, this experience
often informed their view of breast cancer as a disease and many were actively
involved as support providers in these relationships.
Social support occurred in a reciprocal fashion, with women both giving and
receiving support to and from others, which is consistent with the literature on gender
difference and social support (Kawachi and Berkman 2001). Support from a spouse
during the decision-making process appeared to be the most relevant for the
participants, when compared to that of family or friends but most women felt that
81


they had made the decision to enroll on their own. The majority of participants
reported positive emotional support from their spouses. Friends also gave emotional
and informational support during the process, though again, their influence in the
decision-making process appeared to be small. Women themselves reported that their
own role in giving support to others with breast cancer was influential in their
decision, however. As women tend to invest to a greater degree than men in their
social relationships, it is perhaps to be expected that participants in this study
discussed social support either given or received not only within the context of close
members of their social network, such as family members or friends, but also within
their relationships with those farther removed, such as their physician, the study
nurses, and other participants (Thoits 1995).
Study protocol was an important factor at the institutional/ organizational
level, with nearly a third of women saying that the non-placebo design of the study
was influential in their decision. Women generally had good understanding of the
study drugs and were reassured about side effects. Raloxifene was believed by most
to present no more risk than tamoxifen, and the study was seen as comparing two
drugs that likely worked, rather than testing a new drug. The majority of study
participants were assigned to the tamoxifen arm of STAR.
At the community level, several factors overlapped with the influence of the
interpersonal level. Altruistic motivations that included not only a desire to help
family members and friends at risk for the disease, but also women in general were
cited along with a goal of furthering medical knowledge. In addition, members of
social networks often share similar values, norms, or backgrounds; the participants in
this study reinforced that view (Kawachi and Berkman 2001). Few women had
experienced negative feelings towards clinical trial enrollment from friends or family
members, while none exhibited these feelings themselves. More than one participant
82


indicated that the educated nature of her family or friends was in part why they were
supportive of trial enrollment. Although not all women considered themselves
knowledgeable about breast cancer, the interviews revealed a good understanding of
risk, prevention and treatment. Participants also were familiar with sources they
could access for further information.
Participant responses indicate a stronger influence of proximate rather than
distal factors in their decision-making. In fact, these distal factors were more
apparent in their absence from the interviews rather than their presence. Most were
familiar with the media discourse on breast cancer and found it easy to access
information on the disease through media sources or the Internet. Other social
structure factors may help to understand the relatively homogenous demographic
make-up of the population that participated in this study, such as socioeconomic
status, cultural or linguistic background, education levels, and the history of clinical
trials in the United States. Policy wise, the organization by NCI of a network of
CCOPs to bring cancer clinical trials to local communities and health centers has
allowed a larger proportion of the population to participate in these trials and
broadened the pool of potential participants.
Social network and social support theories were useful tools to examine
several layers of the socio-ecological framework between the individual and the
larger social structure, policy and systems levels. In particular, social networks
helped to understand how these participants could share values and characteristics at
the community level, such as citing altruistic motivations as an essential part of their
decision to enroll in STAR. Having had a member of their social network experience
breast cancer, such as a friend, family member, or colleague, was also influential. At
the interpersonal level and organizational level, different types of social support came
into play. Spouses were a source of emotional support in many cases during both the
83


decision-making and participation processes, and in at least one, a crucial source of
instrumental support without which study participation would not have been possible,
Interestingly, in some cases, it was the help or resources women had given to others
with breast cancer in the past that influenced their decision to participate; acts which
were often tied to their altruistic motivations. Participants cited the informational,
instrumental, and emotional support they received from their study nurses as an
important aspect of their study participation. This may be useful for future clinical
trials in chemoprevention to consider in addressing study design.
Directions for Future Research
The results of this study are most helpful in illuminating factors that impact
decision-making on the individual, interpersonal, and community levels. While
institutional/organizational factors such as study oversight and the study design of
STAR were important, data from this study indicate that this level may be more
influential in shaping women's experiences while enrolled in the trial, rather than
their decision-making process prior to enrollment. However, further research is
needed to confirm these results. An analysis that also accounted for large scale social
and policy influences would increase our understanding of which segments of the
population hear about clinical trials and perceive trial participation as a possibility
given transportation, and economic, linguistic and cultural circumstances. One
limitation of this study is its focus on women who chose to enroll in a
chemoprevention trial for breast cancer. Because women who opted out of
enrollment were not interviewed, comparisons to this population are limited. For
example, none of the women in this study viewed risk of side effects as a barrier to
trial enrollment. Most reported positive support for their participation from their
spouses. Women who choose not to participate in a clinical trial may give different
84


answers. Futhermore, this group was even more self-selected in that those who
participated took a second step in agreeing to be interviewed for the study reported
here. A second limitation is the relatively homogenous nature of the population
studied. As a group, they were highly educated, middle-class and primarily White.
Future qualitative research is needed to identify contextual differences in the
decision-making process between participants and non-participants, and to identify
factors that are influential for a more diverse group of women.
85


APPENDIX A
ACRONYMS
ACS American Cancer Society
ADH Atypical Ductal Hyperplasia
BCDDP Breast Cancer Detection Demonstration Project
BCPT Breast Cancer Prevention Trial
CCOP Community Clinical Oncology Program
CCRP Colorado Cancer Research Program
CORE Continuing Outcomes Relevant to Evista
DCIS Ductal carcinoma in situ
FDA Food and Drug Administration
HRT Hormone Replacement Therapy
IBIS International Breast Cancer Intervention Study
LCIS Lobular carcinoma in situ
MORE Multiple Outcomes Raloxifene Evaluation
NBCCEDP National Breast and Cervical Cancer Early Detection Program
NCI National Cancer Institute
NSABP National Surgical Adjuvant Breast and Bowel Project
SERM Selective Estrogen Receptor Moderators
STAR Study of Raloxifene and Tamoxifen
86


APPENDIX B
KEY CONCEPTS AND DEFINITIONS OF THE HEALTH BELIEF MODEL
AND THE THEORY OF REASONED ACTION
Table B.l. The Health Belief Model.
Concept Definition
Perceived susceptibility Ones opinion of chances of getting a condition.
Perceived severity Ones opinion of how serious a condition and its sequelae are.
Perceived benefits Ones opinion of the efficacy of the advised action to reduce risk of seriousness of impact.
Perceived barriers Ones opinion of the tangible and psychological costs of the advised action.
Cues to action Strategies to activate ones readiness.
Self-efficacy Ones confidence in ones ability to take action.
(Adapted from Strecher and Rosenstock 1997:45)
Table B.2 The Theory of Reasoned Action and the Theory of Planned Behavior.
Concept Definition
Behavioral intention Perceived likelihood of performing the behavior
Attitude Behavioral belief Belief that behavioral performance is associated with certain attributes or outcomes
Evaluation Value attached to a behavioral outcome or attribute
Subjective norm Normative belief Belief about whether each referent approves or disapproves of the behavior
Motivation to comply Motivation to do what each referent thinks
Perceived behavioral control Control belief Perceived likelihood of occurrence of each facilitating or constraining condition
Perceived power Perceived effect of each condition in making behavioral performance difficult or easy
(Adapted from Montano et al 1997:91)
87


APPENDIX C
DEMOGRAPHIC QUESTIONNAIRE
Subject ID # _________________
Age________________
Marital Status (circle one):
Separated
Domestic Partner
Single Married
Divorced Widowed
Ethnicity:________________________________
Education Level: Some High School or below
High School Diploma
Some College
Associates Degree
________Bachelors Degree
Graduate Degree
Post-graduate Degree
Occupation: ___
Family History of Breast Cancer? YES NO
If YES briefly explain:
Number of Female Children:
Number of Sisters:
88


APPENDIX D
INTEVIEW QUESTION GUIDE
What motivated you to consider participating in the STAR trial?
Tell me about your experiences in the STAR trial.
How do you think your reasons for participating in the STAR trial compare to
those of other women?
What reservations did you have regarding participation in the STAR trial?
Once you were identified as high risk for breast cancer, with whom did you
discuss your status?
o With whom in your family did you discuss these issues with?
o Why did you choose those particular people?
o What about friends? Who did you turn to outside of your family and
why?
How knowledgeable did you consider yourself about breast cancer before
entering the trial?
o How about chemoprevention?
What steps did you undertake to educate yourself or find out more about
breast cancer after being identified as high risk for the disease?
o How about breast cancer prevention?
Do you have a family history of breast cancer? Who in your family had the
disease?
o How did this family history play a role in your decision to enroll in a
chemoprevention trial?
How did having daughters influence your decision to enroll in the STAR trial?
How did having sisters influence your decision to enroll in the STAR trial?
Have you ever had a close friend who was diagnosed with breast cancer?
o What happened in her case?
o Was she diagnosed before or after you were identified as high risk?
o How did this influence your decision to enroll in the STAR trial?
How did your spouse/partner react when you told him/her you were
considering enrolling in the STAR trial?
o How did this influence your decision?
What were your friends1 thoughts and reaction to your decision to enroll?
Who made up your social network while you were taking part in the STAR
trial?
89


BIBLIOGRAPHY
Advani, Anjali S., Benjamin Atkeson, Carrie L. Brown, Bercedis L. Peterson, Laura
Fish, Jeffrey L. Johnson, John P. Gockerman, Marc Gautier.
2003 Barriers to the Participation of African-American Patients with Cancer in
Clinical Trials. Cancer 97(6):1499-1506.
Altschuler, Andrea and Carol P. Somkin.
2005 Womens Decision Making About Whether or Not to Use Breast Cancer
Chemoprevention. Journal of Women & Health 41(2):81-95.
American Cancer Society.
2005 Breast Cancer Facts & Figures 2005-2006. Atlanta: American Cancer
Society, Inc.
American Cancer Society.
2008a Cancer Facts and Figures 2008. Electronic document,
http://www.cancer.org/downloads/STT/2008CAFFfmalsecured.pdf, accessed
May 19,2009.
2008b Non-Cancerous Breast Conditions. Electronic document,
http://www.cancer.org/docroot/CRI/content/CRI 2 6X Non Cancerous Breas
t_Conditions59.aspaccessed May 19, 2009.
American Cancer Society.
2009 What is Breast Cancer? Electronic document,
http://www.cancer.org/docroot/CRI/content/CRI2 4 IXWhatis_breast can
cer 5.asp?mav=cri, accessed May 19, 2009.
Armstrong, Katrina, Ellyn Micco, Amy Carney, Jill Stopfer, and Mary Putt.
2005 Racial Differences in the Use of BRCA 1/2 Testing Among Women With a
Family History of Breast or Ovarian Cancer. Journal of the American Medical
Association 293(14):1729-1736.
Bastian, Lori A., Isaac M. Lipkus, Maggie N. Kuchibhatla, Haoling Holly Weng,
Susan Halabi, Paula D. Ryan, Celette Sugg Skinner, and Barbara K. Rimer.
2001 Womens Interest in Chemoprevention for Breast Cancer. Archives of
Internal Medicine 161:1639-1644.
90


Full Text

PAGE 1

WOMEN'S DECISION-MAKING PRIOR TO ENROLLMENT IN THE STAR TRIAL FOR BREAST CANCER CHEMOPREVENTION By Emily E. Chasco B.A., University of Michigan, 2004 A thesis submitted to the University of Colorado Denver in partial fulfillment of the requirements for the degree of Master of Arts Anthropology 2009

PAGE 2

This Thesis for the Master of Arts degree by Emily E. Chasco has been approved by Sarah Horton Date

PAGE 3

Chasco, Emily E. (M.A., Anthropology) Women's Decision-Making Prior to Enrollment in the STAR Trial for Breast Cancer Chemoprevention. Thesis directed by Associate Professor J olm Brett ABSTRACT Despite advances in screening and treatment, breast cancer remains an important public health issue in the United States. Until recently, early detection provided the best chance for survival. However, chemoprevention, the use of selective estrogen receptor moderators (SERMs), now offers a new avenue for prevention of the disease in women at high-risk. Clinical trials are essential to determine the safety and efficacy of chemoprevention agents. Yet factors that influence women to participate in these trials are poorly understood. The purpose of this study was to explore factors that influence the decision-making process of women at high risk for breast cancer when considering enrollment in a chemoprevention trial. Eligible participants were between the ages of 50-80, resided in the Denver metropolitan area, were high-risk for breast cancer, and enrolled in the Study ofTamoxifen and Raloxifene (STAR). STAR was a non-placebo arm prevention trial conducted by the National Surgical Adjuvant Breast and Bowel Project to test tamoxifen, the first agent approved by the FDA for breast cancer prevention, against raloxifene, a second generation SERM. Taking an exploratory qualitative approach, data was gathered through a short Demographic Questionnaire and in-depth interviews conducted with seventeen women between September

PAGE 4

2007 and September 2008. The results are discussed within a socio-ecological framework that examines individual, interpersonal, institutional/organizational, community, and structural levels of influence. Women's experiences while enrolled in STAR are also discussed. Results indicate that physician recommendation, altruistic motivations, having female offspring, and knowing a drug would be given rather than a placebo were all important factors in participants' decision-making. While physician recommendation was important, its influence varied between participants. Prior experience with breast cancer among family and friends was an influential factor in shaping participants' beliefs around breast cancer. Many women mentioned concern regarding potential side effects, though it was not a deterrent to participation. Social support at the interpersonal and community levels played a role prior to enrollment in STAR and during the study as well. The most common motivation for trial enrollment was the desire to help other women by furthering scientific knowledge of breast cancer prevention. This abstract accurately represents the content of the candidate's thesis. I recommend its publication.

PAGE 5

ACKNOWLEDGEMENT I would like to extend my gratitude to my advisor, John Brett, for his continued support and patience throughout this process. I would also like to thank those at the Colorado Cancer Research Program here in Denver, Colorado, and in particular, Jane Hajovsky, for their invaluable helpthis project would not have been possible without them. To my fellow students from the University of Colorado Denver and most importantly to my family: thank you for your advice and encouragement over the past two years.

PAGE 6

TABLE OF CONTENTS Figures............................................................................. 1x Tables.............................................................................. x CHAPTER 1. INTRODUCTION ........................................................... 2. BACKGROUND.............................................................. 4 Breast Cancer and Chemoprevention.... ...... .. .. .. .. .. .. .. .. .. .. .. ... 4 Study ofTamoxifen and Raloxifene................................ ... 9 Determining Risk......................................................... 10 Women's Decision-Making When Considering Enrollment in Chemoprevention Trials.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 13 3. THEORETICAL FRAMEWORK........................................... 17 Socio-Ecological Approaches........................................... 17 Socio-Ecological Approaches, Health Behavior, and Breast Cancer Chemoprevention Trials... . . . . . . . . . 22 Social Support and Social Network Theories......................... 26 4. METHODS & ANALYSIS................................................... 30 Colorado Cancer Research Program.................................... 30 Research Design and Qualitative Methods............................ 31 Data Collection....................................................... 34 Vl

PAGE 7

Data Analysis........................................................ 37 5. MOTIVATIONS FOR ENROLLMENT............................................. 40 Demographic Characteristics of Participants......................... 40 Individual Factors.......................................................... 42 Perception of Breast Cancer Risk.................................. 42 Perception of Medical Research................................... 45 Perceived Benefits of and Barriers to Participation............ 47 Interpersonal Factors...................................................... 50 The Role of the Physician.......................................... 50 Prior Experience with Breast Cancer............................. 54 Having Female Offspring/ Sisters.................................. 56 Social Support........................................................ 58 Institutional/ Organizational Factors................................... 62 Placebo vs. Non-Placebo Trials................................... 62 Community Factors....................................................... 66 Altruistic Motivations............................................... 66 Social Structure, Policy, and Systems Factors........................ 69 Media.................................................................. 69 Other Factors......................................................... 70 Vll

PAGE 8

6. PARTICIPANT EXPERIENCES WHILE ENROLLED IN THE STUDY OF TAMOXIFEN AND RALOXIFENE.................. .. 74 Individual Factors..................................................................... 74 Interpersonal and Community Factors.................................. 75 Nurses................................................................. 75 Social Events.......................................................... 76 Physician.............................................................. 78 Organizational Factors.................................................... 78 7. CONCLUSION................................................................. 80 Directions for Future Research.......................................... 84 APPENDIX A. ACRONYMS.................................................................. 86 B. KEY CONCEPTS AND DEFINITIONS OF THE HEALTH BELIEF MODEL AND THE THEORY OF REASONED ACTION................................................................................... 87 C. DEMOGRAPHIC QUESTIONNAIRE.................................... 88 D. INTERVIEW QUESTION GUIDE......................................... 89 BIBLIOGRAPHY......................................................................... 90 Vlll

PAGE 9

LIST OF FIGURES Figure 2.1 FiveYear Relative Survival Rates by Stage at Diagnosis............. 5 3.1 Socio-Ecological Levels of Influence............................................ 21 5.1 Perceived Risks/Barriers and Benefits of Study Participation................................................................................... 50 5.2 Types of Social Support Mentioned by Women in this Study.................................................................................... 59 5.3 Socio-Ecological FrameworkFactors that Influenced Women's Decision-Making While Considering Enrollment in STAR.................................................. 73 IX

PAGE 10

LIST OF TABLES Table 2.1 FDA-approved uses of SERMs........................................................ 9 2.2 Risk factors and their impact on breast cancer................................ 12 3.1 Descriptive characteristics of social networks................................. 27 3 .2 Types of social support.................................................................... 28 4.1 Final code Jist................................................................................... 38 5.1 Age range of participants................................................................. 40 5.2 Number of female children and sisters ofparticipants.................................................................................. 41 5.3 Drug assignment during STAR........................................................ 64 6.1 Symptoms experienced by women during STAR............................ 75 B.1 The Health Belief Model................................................................. 87 B.2 The Theory of Reasoned Action and the Theory of Planned Behavior........................................................................................... 87 X

PAGE 11

CHAPTER I INTRODUCTION Breast Cancer is an important public health issue in the United States. It is the second most commonly diagnosed cancer among American women and prognosis depends significantly upon early detection and stage at diagnosis. Until relatively recently, prevention was limited to changes in lifestyle factors, but in 1998 the first chemoprevention drug for breast cancer was approved by the Food and Drug Administration (FDA 2005). This drug was tamoxifen. The Breast Cancer Prevention Trial (BCPT) demonstrated that tamoxifen use was shown to reduce the incidence of breast cancer among women by 49% (Mulley and Sepucha 2002). Since then, further clinical trials have been conducted on other potential chemoprevention agents, such as raloxifene. Although these clinical trials are essential to determine the efficacy of chemoprevention, only a small proportion of those who are eligible to enroll do so. Current research indicates that a variety of factors may influence women's decision-making when it comes to considering enrollment in breast cancer chemoprevention trials, such as physician recommendation, perceived risk ofbreast cancer, and concern over side effects, to name a few (Bober et al. 2004; Lovegrove et al 2000; Yeomans Kinney et al. 1998a). Understanding women's motivations for enrolling in chemoprevention trials is important for several reasons. Chemoprevention trials are essential for determining the efficacy and toxicity of the drugs involved, unfortunately they often involve long time commitments, require the recruitment of many high-risk but currently healthy participants, and have complicated protocols (Greenwald 1995; Jordan 1995; Yeomans Kinney et al. 1998b). According to one organization, "only 3% to 5% of adult cancer patients actually participate in cancer clinical trials while approximately

PAGE 12

20% may be eligible" (CCRP 2009c). In at least one chemoprevention trial, researchers have noted a concern regarding "the lower than expected response rate" for enrollment (Lovegrove eta!. 2000). Clinical trials not only further medical knowledge around a given health issue, they also provide access to cutting edge treatment and prevention options for those with a cancer diagnosis, or those at-risk. By illuminating why some choose to participate and some do not, physicians and organizations who work in cancer treatment and prevention can better tailor their approach for recruiting women to clinical trials. The method through which a high-risk label is assigned to women is an important point of consideration. Rose notes that, "the determinants of incidence are not necessarily the same as the causes of cases" ( 1985 :34). Furthermore, it may be difficult to identify determinants of incidence within a population if exposure is widespread (Rose 1985). Breast cancer is an excellent example of this phenomenon. The most significant risk factor for the disease is advanced age, which all women (barring early death) will be "exposed" to. However, for the purposes of breast cancer chemoprevention, old age by itself does not make one high-risk. Rather a model that examines multiple individual risk factors relating to family history, reproductive history, and history of other types of breast disease is used to create a personalized risk score (Constantino et al. 1999; Gail et al. 1999). Chemoprevention is therefore an intervention that targets individuals at risk for breast cancer. The primary research aim in this study is to explore the decision-making process women at high risk for breast cancer undergo when considering enrollment in a breast cancer chemoprevention trial. A second is to identify factors that influence the decision-making process related to social relationships and social support. While this project utilizes a primarily inductive approach, social relationships and social support were of particular interest heading into the study due to the fact that they have 2

PAGE 13

largely been overlooked in the literature in favor of more individual level factors. Exploratory, qualitative methodology was used involving in-depth interviews with women who participated in the Study ofTamoxifen and Raloxifene, a clinical trial for breast cancer chemoprevention. The results are explored using a socio-ecological framework that analyzes multiple levels of influence on decision-making, emphasizing the individual, interpersonal and community levels of influence. The experiences ofwomen during trial participation also emerged during the interview process, and these experiences are discussed in relation to the socio-ecological approach that provides the overall framework for this project. Finally, implications for future research are discussed. This study was facilitated by the Colorado Cancer Research Program (CCRP), a non-profit organization based in Denver, Colorado whose mission is "to provide the people of Colorado the opportunity to participate in and benefit from medical research through cancer clinical trials" (CCRP 2009a). CCRP is a Community Clinical Oncology Program (CCOP), part of a network established by the National Cancer Institute (NCI) in 1983 to connect eligible patients with NCI-sponsored clinical trials for the prevention and treatment of cancer (NCI: CCOP electronic document). The purpose of the CCOPs was to bring cancer trials to local communities by creating a nationwide network of local health centers and local providers to conduct clinical trials. The CCOPs, who partner with NCI Research Bases, conduct patient recruitment and data collection, as well as oversight of the community health care providers (NCI 2005). CCRP provided not only information and resources but also conducted the recruitment portion of this project. 3

PAGE 14

CHAPTER 2 BACKGROUND Breast Cancer and Chemoprevention Despite the many medical advances in breast cancer treatment over the last century, current incidence and mortality statistics are still startling. One in eight women will develop breast cancer at some point in their lifetime, according to the National Cancer Institute (NCJ 2008). It is estimated that 182,460 women were diagnosed with new cases of invasive breast cancer in 2008, a number that does not include the more than 60,000 cases of in situ 1 breast cancer (ACS 2008a). In terms of mortality, nearly 40,480 women died of the disease that same year (NCI 2008). It is evident from these numbers that cancer of the breast presents an important public health problem. Both incidence and mortality rates for breast cancer remain a cause for concern. Age-adjusted incidence rates for breast cancer continually increased from 1980 to 2001 for U.S. women of all races and while incidence has decreased since 2001, breast cancer remains the second most common cancer diagnosis in the U.S. among women after skin cancers (ACS 2008a; Kapp eta!. 2008; NCI 2008). The mortality rate for breast cancer has shown a minor downward trend since 1990 but this trend has not been equal in all sub-populations of American women (Cyrus David and Strom 2001; NCI 2008). African-American women, for example, are more likely to die of breast cancer than their White counterparts while the highest mortality rates from breast cancer belong to Native Hawaiians (Kapp eta!. 2008; Newman et a!. 200 I; Shavers and Brown 2002 ). 1 In situ refers to carcinoma that is confined to the layer of cells in which it originated. It has not spread to other organs or tissues (ACS 2009). 4

PAGE 15

There may be biological or behavioral differences between sub-populations that contribute to mortality, nonetheless, it is known that differential access to screening and treatment has an impact as well. Stage at diagnosis2 plays an important role in breast cancer outcomes and it has been argued that the recent decrease in mortality may be due to improved screening practices (Key eta/. 2001). As Figure 2.1 below demonstrates, while the overall five-year survival rate for breast cancer is relatively high, it varies quite noticeably by stage at diagnosis. 100 80 60 % 40 20 0 +----JL...;..._ All Local Regional Distant Stage at Diagnosis Figure 2.1. FiveYear Relative Survival Rates by Stage at Diagnosis3 (From ACS 2008a: 11) 2 At diagnosis, cancer is classified by "stage", which refers to specific characteristics ofthe tumor, including size, location, and spread ofthe disease outside the immediate area of the tumor. A number typically identifies stage, progressing from early or in situ carcinoma identified as stage I, to stage IV that denotes invasive cancer in an advanced stage (ACS 2008a). 3 The terms "local, regional and distant" refer to the spread of the disease from the original location of the tumor and roughly correspond to numbered stages as follows: Local (Stage I), Regional (Stages II and III), and Distant (Stage IV) (ACS 2008a). 5

PAGE 16

Differences in access or utilization of breast cancer screening may be due to multiple factors. Economic circumstances, such as having low-income or being uninsured, limit access to available screening or prevention measures. Meanwhile, programs such as the National Breast and Cervical Cancer Early Detection Program (NBCCEDP) which are in place to assist women in these positions, may be under utilized (Cyrus-David and Strom 200 I; DeSantis eta/. 2008). Certain ethnic or racial groups are known to feel distrustful of the medical profession due to a shared history of feeling targeted or marginalized by that professionthe most well known example being the legacy of Tuskegee among African Americans (Armstrong et al. 2005; Cyrus-David and Strom 2001). Breast cancer is sometimes believed to pose less of a threat to minority women and this may lead to a lack of knowledge not only of the disease, but also ofthe importance of prevention in these communities. When information does spread to communities, it is not always appropriate to the particular cultural background of that population (Armstrong eta/. 2005; Nguyen et al. 2005). Geography is also an important consideration in the United States, as states differ dramatically in everything from the incidence of breast cancer to the types of screening that are available and the frequency of follow-up care that is provided after the screening. In part, this is due to the diverse socioeconomic make-up of each state, including race/ethnicity, income, insurance status, and percentage living in rural versus urban areas, though DeSantis eta/. (2008) point out that state policy on cancer control can play a crucial role in influencing the disparities we see in mortality from breast cancer. Figure 2.1 illustrates the importance of early detection. With the advent of chemoprevention, women who had a high-risk of breast cancer were given the option of taking tamoxifen to prevent breast cancer rather than focusing solely on screening. Chemoprevention is defined as the "treatment with either naturally occurring or 6

PAGE 17

synthetic chemical agents to prevent, reverse, or arrest the progression of preneoplastic lesions to invasive cancers" (Cyrus-David and Strom 2001 :522). Currently used chemoprevention agents are classified as selective estrogen receptor moderators or SERMs (Bober eta!. 2004). Estrogen plays a key role in promoting the growth of cells in the breasta process which when unchecked can lead to breast cancer (Key eta!. 2001; NCI 2006). By affecting the estrogen receptors in the body, SERMs interfere in the process of carcinogenesis. The use of the term "selective" in the name is an important one. SERMs may inhibit estrogen receptors in one part of the body, while having the opposite effect in other regions. Therefore, a SERM may simultaneously prevent breast cancer while increasing the risk of another type of cancer such as endometrial cancer (Cyrus-David and Strom; 2001; NCI 2006; NCI 2008). This is an important consideration when prescribing a course of a SERM to a patient. For over 30 years, tamoxifen has been used as an effective treatment for primary breast cancer in women diagnosed with the disease (Vogel eta/. 2006). Still, it wasn't until the early 1990s that the Breast Cancer Prevention Trial (BCPT) began recruiting women at high-risk to participate in a study examining the effectiveness of tamoxifen in preventing breast cancer. The results indicated that tamoxifen was indeed able to reduce both invasive and non-invasive breast cancer incidence among women assigned to the drug by 49% (NCI 2008). Tamoxifen therefore became the first chemoprevention agent approved by the Food and Drug Administration (FDA) for breast cancer prevention in 1998 (Bober eta/. 2004; FDA 2005). A side benefit oftamoxifen is its positive effect on bone density. However, it also includes risk of several fairly serious side effects. In addition to potentially exacerbating menopausal symptoms, tamoxifen increases one's risk of endometrial cancer, thromboembolic events and cataracts (Bober eta!. 2004; NCI 2006). 7

PAGE 18

More recently, the use ofraloxifene, a second generation SERM prescribed in the treatment and prevention of osteoporosis, for breast cancer chemoprevention has been explored (Table 2.1 ). Research has included the Multiple Outcomes Raloxifene Evaluation (MORE) and the Continuing Outcomes Relevant to Evista (CORE) studies; both included a placebo arm. MORE was primarily intended to test the efficacy of raloxifene in preventing fractures in postmenopausal women with osteoporosis, though results also indicated a reduced risk of invasive breast cancer among women enrolled in the raloxifene arm of the study. CORE then picked up where MORE left off. Using the same study population, women who chose to continue to take raloxifene for 4 more years following the completion of MORE had a reduced risk of invasive breast cancer of 69% (Vogel eta!. 2006). While raloxifene has been shown to have a decreased risk ofthromboembolic disease and cataracts when compared to tamoxifen, the risk of noninvasive breast cancers such as lobular carcinoma in situ (LCIS) and ductal carcinoma in situ (DCIS)4 is increased. However, this increased risk was not found to be statistically significant (NCI 2006; Vogel et al. 2006). 4 Lobular carcinoma in situ and ductal carcinoma in situ both refer to types of non invasive breast cancers, which have not spread beyond the boundaries of the initial cell layer where they originated. In the case of LCIS, the carcinoma is located in the lobules, while in DCIS it is in the ducts (ACS 2009). 8

PAGE 19

T bl 2 1 FDA a e -approve d uses o fSERM s SERM Uses Tamoxifen Treatment: Metastatic breast cancer Adjuvant treatment Invasive breast cancer Prevention: Contralateral breast cancer Raloxifene Prevention: Osteoporosis (Adapted from Brown and Lippman 2000:6) Study ofTamoxifen and Raloxifene The women interviewed for this study were all participants in the Study of Tamoxifen and Raloxifene (STAR). STAR was a non-placebo arm prevention trial conducted by the National Surgical Adjuvant Breast and Bowel Project (NSABP). The largest breast cancer prevention trial ever conducted at the time (enrollment began in 1999 and concluded in 2004), STAR compared tamoxifen to raloxifene in post-menopausal women at high risk for breast cancer (Altshuler and Somkin 2000; NCI 2008; Vogel eta/. 2006). The goal of the study was not only to compare the effectiveness of the two drugs in preventing invasive breast cancer but also to compare their side effects relative to one another (Vogel et al. 2006). A population of 19,747 postmenopausal women was randomized into two groups and given a five-year course of 60mg of raloxifene or 20mg of tamoxifen. In addition to being identified as high-risk for breast cancer, women had to be 35 years 9

PAGE 20

old or more and be postmenopausal (NCI 2006). Risk was determined using the Gail model and personal health history was also taken into account (Altschuler and Somkin 2000; Vogel et al. 2006). Study results indicated that tamoxifen and raloxifene were equally effective in preventing invasive breast cancer, though raloxifene did not appear to be as effective at preventing non-invasive breast cancer (Vogel eta/. 2006). While the risk of certain side effects appeared to be similar in the two treatment groups, those taking raloxifene had lower risk of developing deep vein thromboses, pulmonary embolisms, cataracts and uterine cancer, though the latter was not statistically significant (Vogel eta/. 2006). As a result of these findings, the study was un-blinded early and women in the tamoxifen group were given the choice of completing their five years of treatment with raloxifene instead (NCI 2006). Determining Risk Due to the potentially serious side effects related to the use of chemoprevention agents, treatment is recommended only to women with an elevated risk of breast cancer. The Gail model is one way of determining risk. Gail et al. used data from the Breast Cancer Detection Demonstration Project (BCDDP) to develop a model for estimating the risk of breast cancer for women in a program of annual mammographic screening who have had no previous breast cancer and who have no evidence of breast cancer at the time of their initial screening mammogram. The model estimates the absolute risk (probability) that a woman in a program of annual screening will develop invasive or in situ (ductal carcinoma in situ [DCIS]) or lobular carcinoma in situ [LCIS]) breast cancer over a defined age interval. (Constantino eta/. 1999:1541). To determine risk, this model examines current age, age at both first menstruation and first live birth, family history of breast cancer in first-degree 10

PAGE 21

relatives, number of previous breast biopsies and whether biopsy reveals the presence of atypical hyperplasia (Constantino eta!. 1999; Gail et al. 1999). The most significant risk factor for breast cancer in women is increased age. This risk increases most significantly prior to age 50; the rate of increase is smaller during the post reproductive period (Kelsey and Berkowitz 1988; Key et al. 2001 ). Additional risk factors for breast cancer are summarized in Table 2.2. 11

PAGE 22

Table 2.2. Risk factors and their impact on breast cancer Factor Impact on Risk 0 Age 0 Risk increases with age. This relationship is more pronounced during the reproductive years. 0 Childbearing 0 Nulliparity is associated with increased risk for the disease. Breast cancer risk is reduced with each pregnancy, while younger age at first pregnancy also provides more protection. 0 Menarche 0 Later age of menarche lowers risk. 0 Breastfeeding 0 Believed to provide protection against breast cancer, although this is still controversial. 0 Menopause 0 Younger age at menopause is protective against the disease. Nor does it matter if it is a natural menopause or one induced by bilateral oophorectomy. 0 Hormones 0 High levels of endogenous hormones may be associated with increased risk in post-menopausal women. HRT during menopause is associated with increased risk as well, though risk begins to decrease following treatment cessation. 0 Oral contraceptives 0 Breast cancer risk increases by 25% while oral contraceptives are in use, however, begins to return to normal levels following cessation. 0 Benign breast disease 0 Proliferative lesions are associated with an increase in risk, as are atypical hyperplasias. However non-proliferative lesions are not. 0 Diet 0 Although a connection between high-fat diets and an increase in breast cancer risk has been suspected in the past, currently this is unconfirmed, as is a decreased risk for high soy diets. A diet high in vegetables may protect against the disease. 0 Alcohol/ Smoking 0 Alcohol causes some increase in risk. Currently, smoking and increased breast cancer risk has not been shown to be associated. 0 Anthropometry 0 Those who had high birth weight are at increased risk for breast cancer. Obesity increases risk by 50% in post-menopausal women. 0 Exercise 0 Physical activity decreases risk for breast cancer, particularly in premenopausal women. 0 Ionising radiation 0 Exposure to radiation is known to increase risk for breast cancer, though how large an effect diagnostic radiotherapy poses is still uncertain. 0 Environmental oestrogens 0 The few studies done so far have not shown as association between the compounds and risk of disease. 0 Family history 0 Risk doubles if a woman has a first-degree relative, less so if the relative is second-degree. Similar environments and diet may also play a role in incidence of breast cancer among family members. 0 High-risk mutations 0 Mutations in the BRCAl, BRCA2, P53, PTEN and ATM genes increase risk of breast cancer. BRCA 1 and BRCA2 are particularly high risk. (Adapted from Key eta!. 2001) 12

PAGE 23

Family history in particular plays an important but varied role in determining risk, indicating the importance not only of biological factors, but that of the environment as well. The presence of a first-degree relative with breast cancer doubles one's risk of developing the disease. An increase in risk is also associated with second-degree relatives, though it is not as pronounced (Kelsey and Berkowitz 1988; Key eta!. 2001 ). Similar environmental factors and lifestyle choices account for some of the clustering ofbreast cancer cases seen within families; a small percentage of these clusters can be traced to specific genetic mutations. Key et al. write that, "high-risk alleles probably account for most of the families with four or more breast-cancer cases, for around 20-25% of the familial breast-cancer risk overall, and for about 5% of all breast cancers" (2001: 138). The significance of these mutations however, is not in the frequency of their occurrence but in the fact that their presence may drastically increase an individual's risk of developing breast cancer. Five mutations have been tied to increased risk for breast cancer, the most well known being the BRCA 1 and BRCA 2 genes (Key et al. 2001). It has been estimated that the lifetime risk ofbreast cancer for an individual with the BRCA 112 mutations may be as high as 50-80% (Armstrong et al. 2005). Women's Decision-Making When Considering Enrollment in Chemoprevention Trials Chemoprevention trials, such as STAR, are essential in determining the effectiveness and safety of new chemoprevention drugs. Still, further research is needed into the motivations and decision-making processes of women who choose to enroll in these trials due to difficulty with recruitment (Altschuler and Somkin 2005; Bober et a!. 2004; Yeomans Kinney eta!. 1998b ). According to the Colorado Cancer Research Program, when it comes to clinical trials for cancer research, less than one 13

PAGE 24

quarter of adult cancer patients who are eligible to participate actually do so5 (CCRP 2009c). In the area of cancer prevention, two studies that surveyed an eligible population for enrollment in a chemoprevention trial for breast cancer (either IBIS, the International Breast Cancer Intervention Study, or BCPT), found that only half or less than half of the women who were eligible actually choose to enroll (Lovegrove et a!. 2000; Yeomans Kinney eta!. 1998b ). There are negatives to trial enrollment. These trials involve taking drugs that may have potentially uncomfortable or serious side effects and participation may be time-consuming. In addition, while women must be at high-risk for breast cancer in order to enroll, they are currently healthy and this risk does not guarantee the eventual development ofbreast cancer. Yet some women do choose to enroll. What factors influence one potential participant to enroll in a chemoprevention trial for breast cancer while others choose not to? How women interpret their risk for breast cancer has been shown to be associated with women's decision-making for chemoprevention. Women who believed themselves to be at high-risk for breast cancer were more likely to choose enrollment in chemoprevention trials (Bober et a!. 2004; Lovegrove eta!. 2000; Yeomans Kinney eta!. 1998b ). While absolute risk as determined by acknowledged risk factors such as a family history or nulliparity is important, how these risk factors are interpreted by women within the context of their own lives is also key to understanding their decision-making. For example, among women who had a first degree relative with breast cancer, their perception of their own health compared to that of the relative was an important predictor of the likelihood they would choose to enroll in chemoprevention trials (Altschuler and Somkin 2005). Those who chose not to enroll in a chemoprevention trial were more aware on average of the impact of 5 As cited in Chapter I, "only 3% to 5% of adult cancer patients actually participate in cancer clinical trials while approximately 20% may be eligible" (CCRP 2009c). 14

PAGE 25

lifestyle factors on breast cancer risk and how they could be mitigated than those who chose to enroll (Altschuler and Somkin 2005; Lovegrove et al. 2000). However, Cyrus-David and Strom (200 1 ), in a study evaluating general knowledge of chemoprevention among a sample ofhigh-risk women, found that those with a close friend or family member who had reduced their risk for breast cancer through behavioral measures but still developed the disease despite this were less likely to be open to the idea of chemoprevention. Psychological factors, such as intrusive thinking and worry about breast cancer may be associated with choosing trial enrollment as well (Bober et al. 2004). Still, further research is needed in this area due to contradictory findings. Women who participated in the STAR trial actually had less anxiety about breast cancer than those who opted not to, according to a study by Altschuler and Somkin (2005). This may be due to the fact that, as the same study suggests, women viewed trial participation as a way to gain some measure of control over their lives. Similarly, participants were more likely to report that they would gain peace of mind from trial enrollment than those who chose not to enroll (Yeomans Kinney et al. 1995). Aside from the fear of breast cancer, concerns about the actual clinical trials and the drugs involved were also a consideration. Trial enrollment was associated with fewer concerns over the side effects of chemoprevention drugs and more sources of information regarding tamoxifen use (Bober et a/. 2004; Yeomans Kinney eta/. 1995). Others were concerned with the randomization process used to assign women to different arms of the trials, the possibility ofbeing assigned a placebo, or the time commitment involved (Altschuler and Somkin 2005; Yeomans Kinney eta/. 1995). Women who were currently taking estrogen replacement therapy but would have to stop to enroll in the trial were also less likely to choose enrollment (Yeomans Kinney et al. 1995). 15

PAGE 26

The current literature emphasizes the importance of physician recommendation either for or against trial participation in women's decision-making. Physician recommendation/or participation was found to be strongly associated with enrollment in chemoprevention trials, which is consistent with its similarly strong influence in breast cancer screening through mammography (Yeomans Kinney eta!. 1995). In fact, those women whose physicians had recommended trial participation were thirteen times more likely to do so than those whose physician had not made the recommendation (Bober eta!. 2004; Yeomans Killl1ey et al. 1998a). The simple act of recommendation is not the only pertinent factor regarding the role of the health care provider. Physician communication style may also be important in the decision-making process, particularly the understandability of the information given to their patients. This is supported by research showing that the framing of a message by physicians is influential in the decision-making process (Bober eta!. 2004; Lovegrove et al. 2000). In addition, those patients who felt that their physicians involved them in the decision-making process tended to be more satisfied with their decisions in the end (Bober eta!. 2004). On the other hand, one qualitative study indicated that poor physician communication resulted in women being less willing to undergo chemoprevention, though this was not conducted within the context of decision-making for enrollment in a clinical trial (Cyrus-David and Strom 2001). 16

PAGE 27

CHAPTER 3 THEORETICAL FRAMEWORK Socio-Ecological Approaches Socio-ecological approaches to medical anthropology research emphasize the multiple levels of interaction between an individual, their physician, and the social environment, and the way in which this interaction influences individual behavior and beliefs. Although the terms may vary depending on the specific model used, these levels of influence typically proceed in a fashion from the individual, to interpersonal relationships, to the community or neighborhood, and finally to the larger social and cultural environment (Elder eta!. 2006; Fisher et a!. 2005; Sallis and Owen 2002). While biological processes in the individual may be the most proximate levels of health, these processes are perceived and interpreted through the lens of an individual's larger environment. Yet this relationship is reciprocal rather than uni directional. The processes taking place at a more distal level of influence also shape and impact an individual's health experiences while at the same time providing the larger context in which the individual understands those experiences. Ecology, as used in this type of approach, refers to "the space outside of the person," meaning their physical and social environments, as well as their internal environment (Sallis and Owen 2002:462). In moving health promotion from interventions that target individuals to interventions that also target environments, it is important not only to examine the different levels of influence on health but also how these levels interact with one another (Stokols 1992). Stokols identifies four assumptions regarding socio ecological approaches: 17

PAGE 28

The first is that health is influenced by multiple facets of the physical and social environments. The role of personal attributes is also acknowledged. The second assumption is that environments themselves are multidimensional. Environments can be described as social or physical, actual or perceived; as discrete attributes (such as temperature or spatial arrangements); or as constructs (such as behavior setting or social climate). The third assumption recognizes that humanenvironment interactions can be described at varying levels of aggregation: individuals, families, work and cultural organizations, communities, or whole populations. Stokols' fourth assumption is that there is feedback across different levels of environments and aggregates ofpersons (Sallis and Owens 2002:465466). Individuals may possess the skills and attitudes to make certain choices when it comes to their health, however, these must be viewed through their interactions with the social and physical environments. The environments that either constrain or augment their ability to seek care for a medical condition, access screening programs, or implement lifestyle changes. The first assumption notes the influence of individual characteristics on health, while also recognizing that the wider ecological context also plays a role in health behavior. In a study on self-management of diabetes, Fisher et a!. (2005) argue that self-management approaches which rely solely on the abilities of the patient ignore the relationships between differential access to resources and social support, and the larger social, physical and policy environments which interact to influence behavior. Possessing the knowledge of how one manages diabetes is useless, for example, if the individual is unable to follow through on that knowledge. Healthy eating and exercise may not be possible due to the built environment in which the individual lives, while social support for behavior change may transition throughout the life cycle (Fisher et a!. 2005). 18

PAGE 29

According to the second assumption, ecology does not simply refer to the natural or physical environment but rather may include the social and cultural environments in which individuals live. That environments can be actual or perceived is illuminating for health behavior research in which individuals' beliefs regarding their environment may dictate how they view barriers to health care or self efficacy when it comes to behavior change. Interventions can address these perceptions. For example, reduced physical activity among girls was the subject of the trial of activity for adolescent girls, which sought to address the situation through school-based interventions (Elder et al. 2006). While increasing the availability and awareness of physical activity opportunities for girls in both schools and the local community were objectives of the interventions, another was to address "real and perceived barriers to being physically active" among girls (Elder et al. 2006:161, emphasis added). The third assumption is critical to socio-ecology, which is a multilevel approach. Analyses ofhuman-environment interactions have progressed beyond the infectious disease model of host-agent-environment to include a more complex classification of what constitutes environmental factors and at what level groups are studied (Sallis and Owen 2002). The influence of the environment can be studied not only at the individual level, but also at the level of the family, the social network, the community, the institution, etcetera. For example, teen smoking patterns have been researched by examining the influence of the parent-child relationship, the school environment, neighborhood characteristics, and state policy (Sampson et al. 2002; Wen eta!. 2009; Wilcox 2003). Finally, ecological levels are not discrete entities, bounded by neat lines. The fourth assumption, that of feedback across the different levels of the socio-ecological context, illustrates the difficulty faced by researchers attempting to determine where 19

PAGE 30

to place influential factors for health behavior, and which direction the influence travels. In some cases, the feedback relationship is reciprocal, while in others it may be difficult to actually differentiate the effects of multiple levels from one another due to the fact that causal forces for behavior change in different environments may be correlated to one another (Cook 2003). In the case of adolescent smoking, for example, "schools may affect smoking through the peer groups that the schools themselves make possible" (Cook 2003:153 ). Socio-ecological models vary depending on their applications to research. An early model was that of Bronfenbrenner, whose microsystem, mesosystem, and exosystem influences on behavior contributed to later frameworks, including the one utilized for this study (McLeroy eta!. 1988; Sallis and Owens 2002). These system levels proceeded from the most local to the most structural, with microsystem corresponding to interpersonal influences, mesosystem to the community, organizational or institutional levels, and exosystem to social, cultural, political and economic forces. The specific socio-ecological framework applied to the current study is that used by McLeroy eta/. (1988) and Gregson eta! (200 1 ), which identifies five levels of influence. These levels are defined as "individual," "interpersonal," "institutional/ organizational," "community," and "social structure, policy, and systems" (Gregson eta!. 2001 :S5). This particular model was chosen due to the fact that it included not only community levels of influence, but also organizational, which is particularly relevant for this study given the impact ofboth women's beliefs and values, which were shaped by the community of which they were a part, but also the role of CCRP in recruiting women to the STAR trial and then providing support to them during their enrollment. Definitions for each of the levels of influence can be found in Figure 3 .1. 20

PAGE 31

Social Structure, Policy, and Systems Community ----- Institutional/ Organizational Interpersonal Individual Local, state, federal policies and laws that regulate or support healthy actions Social networks, norms, or standards (e.g. public agenda, media agenda) Rules, regulations, policies, and informal structures (worksites, schools, religious groups) Interpersonal processes and primary groups (family, peers, social networks, associations) that provide social identity and role definition Individual characteristics that influence behavior such as knowledge, attitudes, beliefs, and personality traits Figure 3.1. Socio-Ecological Levels of Influence (Adapted from Gregson eta!. 2001 :55). 21

PAGE 32

Socio-Ecological Approaches, Health Behavior, and Breast Cancer Chemoprevention Trials To date, research in the area of participation in chemoprevention trials for breast cancer has largely focused on the individual level, drawing primarily on one of two theoretical frameworks: the Health Belief Model and the Theory of Reasoned Action. Concepts drawn from the Health Belief Model and the Theory of Reasoned Action that have been applied to research on chemoprevention decision-making for breast cancer include perceived susceptibility (often termed perceived risk), perceived barriers and benefits, perceived behavioral control (locus of control), self-efficacy, and normative beliefs (Altschuler and Somkin 2005; Cyrus-David and Strom 2001; Lovegrove et al. 2000; Yeomans Kinney eta!. 1995; Yeomans Kinney et al. 1998a; Yeo mans Kinney et a!. 1998b ). Further concepts and definitions important to these two frameworks can be found in Appendix B. The perception in biomedicine that individuals have a responsibility for their own health is particularly salient given the focus on decision-making and perceived risk. One of the key trends in the process of biomedicalization since the mid-1980s has been the transfer of responsibility for health from the sphere of the physician to that of the individual, who now has a "moral obligation" to remain healthy (Clarke et al. 2003: 171). Included in this responsibility are a heightened awareness of one's own personal risk factors and what the individual can do to mitigate this risk, a trend that Rockhill terms "risk privatization" (2001 :365). In order to determine risk, there has been a large emphasis on what has been called risk factor epidemiology and the development and application of ever more complex risk assessment tools, which can be seen clearly in the field of breast cancer (Clarke eta!. 2003). One key example of this is the Gail model. By quantifying the multitude of scenarios that may make women susceptible to breast cancer, the vague notion of 22

PAGE 33

"risk" is transformed into a number, a value that can be easily compared and stratified. Nor is this value totally subject to fate. While women cannot currently do anything to change their genes, or the age at which they begin menarche or menopause, the remainder of the list of breast cancer risk factors could be said to be subject to alteration. Oral contraceptive use or hormone replacement therapy (HRT); diet and exercise; alcohol or tobacco use; nulliparity or breastfeeding; all rely to one degree or another, on choices women make. Therefore, according to this trend towards risk privatization, women can and should mitigate their own risk by changing their behavior. Chemoprevention becomes one more route through which women can lower their risk of breast cancer. Although this prioritizing of individual risk in decision-making has been a focus of the research on chemoprevention trial enrollment, it should be recognized that it is part of larger society and policy level phenomena that is impacting health. With the advent of genetic testing, individual risk can now be defined on the molecular level to result in what Locket a/. calls "the individual 'genetically at risk"' (2007:257). Despite the fact that for most diseases the presence of a particular allele or mutation rarely gives a simple yes or no answer to whether or not they will actually develop the condition, risk is calculated using these tests. Though how individuals themselves interpret their results in a meaningful way for their own lives may be more complex (Locket al. 2007; Rapp 2000). A similar shift to an individual responsibility model can be seen in the field of chronic diseasesthose with diabetes for example are encouraged to self-manage their disease (Fisher et al. 2005). Even the language around health at the organizational or policy level has actually undergone a shift: Terms such as "health maintenance," "health promotion," and "healthy living" highlight the mandate for work and attention toward attaining and maintaining health. There has been a steady increase in mandates for self-regulation until, 23

PAGE 34

with biomedicalization, there is a shift in the general cultural expectations of whole populations. In this constant, self-disciplining and other/public disciplining, there is no rest for the weary (Clarke et al. 2003: 170). In chemoprevention research, the primary exception to the focus on the individual level of influence has been the effect of physician recommendation on the potential participant, which reflects a larger interest in the patient-physician relationship in general (Kreuter et al. 2000; Lovegrove et al. 2000; Quill and Brody 1996). Physician recommendation falls under two levels of analysis, the individual and the interpersonal. From the standpoint of the Theory of Reasoned Action, normative beliefs refer to the attitudes of "important referent individuals" towards a particular health behavior (Montano et al. 1997:87). A physician could be considered such an influential person. In contrast, at the interpersonal level, it is not only the physician's attitude that matters, but the relationship and level of communication between the physician and the patient as well (Bober et al. 2004; Lovegrove et al. 2000). A more holistic view of breast cancer chemoprevention trial enrollment can be conceptualized in terms of a socio-ecological model that allows for a multilevel examination of factors that impact women's decision-making. An example would be to consider not only how the individual perceives risk based on her own known biological risk factors, but also how those risk factors in that individual's environment shape her perception of risk. The individual's understanding of the severity of breast cancer may depend on not only personal experience, but also relationships with others who have had the disease, knowledge of well-known public figures with breast cancer, or the messages seen in the media. The view that women hold of clinical trials may be shaped by their personal history, the history of the sub population to which they belong, their education level, or their physician's opinion. 24

PAGE 35

A few factors that may reflect the influence of interpersonal or community levels have been briefly touched on in the literature. However, they have not been the primary focus of research but rather were generally included as part of inquiries conducted using close-ended questionnaires that have not allowed women to expand on their thoughts. By choosing which factors to include in the questionnaire, the researchers may incorporate assumptions about how women think about risk or which factors they perceive to be important without truly examining those assumptions. In addition, survey research has a history of assuming that each individual is an independent entity or of treating events as independent of one another, when in fact individuals are embedded within a larger social context (Coleman 1958). This social context, including relationships between individuals and shared community values or beliefs, may shape women's answers and should be acknowledged by researchers. Interpersonal and community factors that have been identified include the influence ofphysician recommendation, the attitude of one's spouse, having had a friend or family member with the disease, and wanting to help other women (Bober et al. 2004; Lovegrove eta!. 2000; Yeomans Kinney et al. 1995). Unfortunately, there has been little unifying theory applied to these levels of analysis. While the socio ecological model allows for a more holistic understanding of a given health behavior, lack of specificity can be an issue. In order to provide this specificity, additional theories or models can be applied to each level of analysis, from individual to society (Elder et al. 2006; Sallis and Owens 1997). Of particular interest for this study are social support and social network theories, which may provide a deeper understanding of the interactions between participants and others in their environment. Examining individual constructs is useful, but rarely are the beliefs or actions they describe unconnected. The patient/physician dyad is interesting not only because of the impact of physician 25

PAGE 36

recommendation on the individual, but because that recommendation is taking place within a relationshipa relationship that may also include other forms of social support. The role of intrusive thinking and worry has been studied within the context of trial enrollment. It may be that support given by one's physician mitigates those concerns. The physician's communication style presumably would depend on the relationship they have with their patient, whether more authoritative or more egalitarian. Social networks in particular could be important in illustrating whom participants spoke to regarding their enrollment or explaining their previous experiences with breast cancer. The socio-ecological model provides a framework to organize different levels of influence on health. Social support and social network theories may explain these influences. Social Support and Social Network Theories Current research on decision-making in the area of chemoprevention trials has focused on the individual, ignoring the fact that individuals exist within a social environment made up of groups of individuals and relationships between those individuals. To address this gap, social support and social network theories can be integrated into the socio-ecological model at the interpersonal and community levels. The link between health and social support can be traced back to one ofthe fathers of modem sociology, Emile Durkheim and his landmark work, Suicide: A Study in Sociology (1951), originally published in 1897. Durkheim was the first to examine suicide not as an individual phenomenon, but as a social phenomenon. His proposal, that social integration was an important protective force against suicide, has been very influential in the literature on social support and social networks (Kawachi and Berkman 2001 ). 26

PAGE 37

A social network is "the web of social relationships that surround individuals" (Heaney and Israel 2002: 185). The descriptive characteristics of this web are identified in Table 3.1. These characteristics are influential in determining what types of coping mechanisms individuals may have available when a health issue arises, as well as how resources, influence and risk travel through social groups (Friedman and Aral 2001; Heaney and Israel 2002; Roberts et al. 1994). Different social relationships may provide different types of social support (Table 3.2). An analysis of the interpersonal aspect of social relationships that influence decision-making around trial enrollment must move beyond the physician-patient relationship and what could be called instrumental and/or informational support to further contextualize the interpersonal environment. In addition to their physician, other members of an individual's social networks may also be contributing social support, such as a partner, child or friend. T bl 3 1 D a e .. h f I k escnphve c aractenstlcs o soc1a networ s Concepts Definitions Reciprocity Extent to which resources and support are both given and received in a relationship Intensity Extent to which social relationships offer emotional closeness Complexity Extent to which social relationships serve many functions Density Extent to which network members know and interact with each other Homogeneity Extent to which network members are demographically similar Geographic Extent to which network members live in close dispersion proximity to focal person (Adapted from Heaney and Israel 2002: 187) 27

PAGE 38

T bl 3 2 T a e ypes o f 1 SOCia support Concepts Definitions Emotional Support Expressions of empathy, love, trust and caring Instrumental Support Tangible aid and service Informational Support Advice, suggestions and information Appraisal Support Information that is useful for selfevaluation (Adapted from Heaney and Israel 2002: 187) The ameliorating influence of strong social support on stress and anxiety following a breast cancer diagnosis has been studied (Roberts et a/. 1994 ). However, the same has not been true of those who are identified as high-risk for breast cancer. At the moment these women are healthy but with the emphasis placed on prevention, these women are choosing "to 'treat' the risk of cancer" (Clarke et a/. 2003 ). How women access support through their social networks when making the decision to treat their risk by enrolling in chemoprevention trials is therefore of interest here. Another area of research related to social networks that could be further explored is that of altruistic motivations. The literature suggests that having a relative or friend who had breast cancer or died of the disease does exert some sort of influence on decision-making (Altschuler and Somkin 2005). The desire to help family members at future risk for the disease has also been associated with study enrollment (Lovegrove eta/. 2000). The subject of altruism and clinical trials has been addressed elsewhere in the clinical trial literature, though not extensively within the context of breast cancer chemoprevention. Altruism is a motivating factor in enrollment for placebo-controlled trials of antihypertensive drugs, specifically the ideas of"helping other patients" and "contributing to scientific knowledge" (Halpern et al. 2003 :987). These two concepts illustrate the argument of Simonet a/., that 28

PAGE 39

altruism in the context of clinical trials is not a "monothematic notion" but rather reflects a more diverse set of concepts and concerns (2006:46). In other words, altruism refers not only to the idea ofhelping one particular group of people, but may encompass family, friends, other patients, and the population in general. However, in the specific study by Simonet al. (2006), altruism was not found to be a major motivator for trial participation. Altruism has also been studied in the context of social networks. In a study of volunteering and its influence on health, an association was found between volunteering outside one's direct social network and positive health gains (Brown et al. 2005). Drawing on these examples from other areas of health research, and the preliminary findings on quantitative instruments used in breast cancer chemoprevention research given above, altruistic motivations towards family members or other members of one's social network appear an area worth incorporating into the socio-ecological framework used to analyze decision-making around chemoprevention trials. 29

PAGE 40

CHAPTER4 METHODS & ANALYSIS Colorado Cancer Research Program The Colorado Cancer Research Program was an invaluable partner in this project, providing information and support to the investigator. CCRP is a non-profit organization located in Denver, Colorado and describes itself as a: A nonprofit community-based cancer program established to provide community hospitals and physicians access to a wide range of cancer research trials in order to provide their patients with greater options for the treatment, control, and prevention of cancer. [CCRP 2009a] Colorado is one of 34 states in which Cancer Clinical Oncology Programs (CCOPs) such as CCRP are located (CCRP 2009). Their role is to bring clinical cancer trials, which previously were confined to areas surrounding research centers, to communities across the country allowing individuals to participate through their physician or local health center. CCRP provides oversight of the clinical trials, providing data support to the local network of physicians and hospitals, and quality oversight for study procedures and data management. In Colorado this is accomplished though a network which includes "seventeen Colorado hospitals, over I 00 medical and radiation oncologists, surgeons and other specialists, and a high trained CCRP oncology program staff' (CCRP 2009a). It also includes research nurses, provided by CCRP, who serve as participants' primary contact among study personnel, remind them of appointments to fulfill study requirements, and provide support and information for participants when they have questions or concerns. 30

PAGE 41

Research Design and Qualitative Methods This study utilized a qualitative exploratory design in order to gain in-depth understanding of the decision-making process for women at high-risk for breast cancer considering enrollment in a chemoprevention trial. The strength of qualitative research does not lie in its generalizeability, as in quantitative research, but rather in that it provides detailed and nuanced information about a given phenomenon. Validity rests not on a large sample size, but rather "has to do with description and explanation, and whether or not a given explanation fits a given description" (Janesick 1994:216). Drawing on a grounded theory approach, an inductive and iterative analysis was used by which the data were examined for patterns or themes, which were then linked to existing theory. The theory is therefore "grounded in the data," as opposed to deductive approaches which assess the data using previously defined hypotheses or theoretical constructs (Neuman 2003; Patton 1990). Qualitative methodologies emphasize the importance of contextualization, which has been lacking in the literature on decision-making for breast cancer chemoprevention trials. These studies have typically relied on the use of close-ended (or a combination of closeand open-ended) surveys and questionnaires, which are then subjected to statistical analysis. In addition, they tend to make assumptions regarding the independence of individuals and events, which may be problematic when these individuals live in social contexts that may be very influential on their beliefs and behaviors (Coleman 1958). Such an approach may identify important factors in the decision-making process, yet a more nuanced in-depth understanding of how and why these factors play a role requires a qualitative approach (Patton 1990). While qualitative research into breast cancer chemoprevention trials has been rare, one important study has influenced the research reported here. 31

PAGE 42

Altschuler and Somkin (2005) conducted qualitative in-depth, semi-structured interviews with a population of 51 women eligible for the STAR trial. The results of their study indicate that age was an important factor in that women who chose to participate were slightly older on average than those who chose not to participate. A possibly related finding was that STAR participants had a higher perceived risk of breast cancer. While most of the factors brought up were similar to results found using quantitative methods in other articles, two important themes were revealed. Women who enrolled in the STAR trial were more likely to see participation as a form of activism and were also more likely to have a close friend who had had breast cancer, indicating that activism and altruism are areas that should be included in future studies. When close-ended surveys are used, participants are forced to choose between pre-selected answers decided on by the researcher. There is no room for women to explain why they chose the answer they did; this is crucial as different participants may arrive at the same action but have alternative motivations for doing so. The reverse is true as well; two women with the same concerns may see opposite paths to resolution. This type of quantitative instrument may also force women to choose a response even if none are truly appropriate for their own circumstances. An example of this can be seen in the study by Altschuler and Somkin (2005), in which those who chose to participate in the STAR trial and those who chose not to both considered their decision to be based on a desire to be in control of their health. However, the actions that constituted taking control were completely different for each group. Qualitative approaches, on the other hand, seek to allow participants to tell their own story, to present their motivations and experiences using their own words, "without imposing preexisting expectations on the phenomenon or setting under study" (Patton 1990:44). A survey which simply asked whether or not women felt 32

PAGE 43

themselves to be at increased risk for breast cancer would not necessarily lead to insights into how that feeling influenced women's behavior. Yet by conducting a qualitative study, Altschuler and Somkin (2005) were able to differentiate between women who acknowledged risk intellectually and rather dispassionately, and women for whom that risk that was a source of concern and anxiety. Not coincidentally, these groups of women made different decisions in regards to trial enrollment. Consider these two quotes from this study: Intellectually, I do (feel at risk), but emotionally, I want to believe that I can have good health and that I can prevent bad health. I don't dwell on the fact that I might be getting cancer. My gynecologist had the feeling from me that I'm an anxious person, and she thought (joining the trial would be a good idea). So I thought, "She's right, you know?" If I don't participate, I will always think, "Well, gee, maybe I should have." And if I do, maybe it will lesson the anxiety (Altschuler and Somkin 2005:88). Each of these women would respond that they felt at risk for breast cancer on a close ended survey. But by allowing the women to express their thoughts regarding this risk, we see that they have very different responses to managing these feelings. The qualitative methodology employed in this study takes the form of semi structured in-depth interviews. This choice of methods allows the researcher to obtain data that allows for an understanding of how the participants organize and contextualize their own experiences (Gelo et al. 2008). In keeping with this goal, women's own words are often presented in the discussion of the data, drawing on a polyphonic approach which allows differing beliefs and opinions to come through (Marcus and Fischer 1986). 33

PAGE 44

Data Collection Purposive sampling was the most appropriate choice for this study given that its purpose is "selecting information-rich cases for study in depth" (Patton 1990: 169). The literature is primarily composed of quantitative studies that have utilized questionnaires or surveys to gain a broad but not very nuanced view of women's motivations when it comes to trial enrollment. The goal of the present study was just the opposite. By selecting women for in-depth interviews who had made the decision to enroll in STAR, a closer examination of influential factors and how they are related was made possible. Eligibility for the study was restricted to women between 50 and 80 years of age, living in the Front Range region of Colorado, who had participated in the STAR trial. In addition, interviews were conducted between June 2007 and September 2008 and availability during this period was a necessary requirement. Due to confidentiality concerns, the principal investigator could not access CCRP's STAR participant database for recruitment purposes. Instead personnel at CCRP first identified eligible STAR participants and then handled initial contact. Introductory letters outlining the purpose and procedure for the study were mailed out by CCRP in waves of 20-30 letters to 140 eligible women, beginning with those who enrolled in the STAR trial first. Included with the introductory letters were response letters. Using an opt-in approach, women who were interested in being contacted with further information were asked to return the response letters to the offices of CCRP. The initial response letter asked for a name and phone number, however, a space to include email address was eventually added as it became apparent that some women preferred this method of communication. Only once participants gave permission through these response letters was their contact information passed on to the principal investigator. 34

PAGE 45

Women who were interested in further information were then contacted by the PI to request participation in a 30-60 minute in-person interview. Interviews were held in the women's homes, or, if preferred, at the offices ofCCRP or another public location. Alternative locations included a library, a place of business, restaurants, a city park, and a coffee shop. Twenty-five response letters were received during the study period. Of these, seventeen women completed interviews. The remaining eight either failed to respond to repeat attempts to contact or were out of town during the study period. While a goal of this study was to gain a better understanding of particular typewomen who choose enrollment, it is important to note that the women who sent back the response letters and then agreed to interviews are a group that has been doubly self-selected. First enrolling in STAR, then the present study, they twice made the decision to participate in scientific studies. A potential bias that may result is that this population may be more comfortable with scientific research than a randomized sample would be and this may be reflected in the results. Those who chose not to enroll in STAR or who chose not to speak with those conducting this study might express very different ideas regarding trial enrollment or scientific research in general. Prior to each interview, informed consent was obtained from participants who were also given the opportunity to request the results of the study once they were ready for dissemination. Participants were then asked to complete a short Demographic Questionnaire (Appendix C) that was used to inform the interview. The form asked for basic information, including age, marital status, ethnicity, education level and occupation. In addition, the form included several questions that targeted family make-up and history. These questions included number of female children, number of sisters, and whether or not the participant had a family history of 35

PAGE 46

breast cancer. If the participant responded in the affirmative to the family history question they were asked to give a short explanation of this history. These questions were in included on the Demographic Questionnaire in order to provide the researcher with this information for the interview process. The interviewer was able to use the participants' answers to focus questions regarding the influence of family history and make-up on the participants' decision-making process. With the permission of participants, sixteen out of seventeen interviews were tape-recorded. The seventeenth was not recorded due to a recorder malfunction; however, notes were taken during the interview. Field notes were also taken immediately following every interview in order to summarize the information gained during the experience and to capture any immediate reactions or thoughts of the principal investigator. The PI conducted all interviews. While the purpose of this study was exploratory and inductive, a semi structured Interview Question Guide (Appendix D) was used to direct the interview process. Questions were meant to progress from the general to the specific, allowing participants to generate their own discourse around their motivations and the decision-making process prior to addressing specific factors that had arisen in the literature. Depending on the flow of the interview, questions were on occasion addressed out of order. Participants were also given the freedom to address related topics that came up during the interviews. In particular, any factors they felt were important to their decision that had not been addressed by the question guide were noted; these factors were then examined by the investigator and incorporated into the 36

PAGE 47

question guide for later interviews (Neuman 2003). Interviews lasted from 20 minutes to 2 hours in length6 Data Analysis Following the interviews, all field notes were typed and all taped interviews were transcribed using a word processing program. Names of friends, family members or physicians given during the interviews were changed during this process to maintain confidentiality. Important topics or themes that arose in early interviews were used to inform the question guide for later interviews, in an iterative process. Transcripts were initially hand-coded using an open-coding process whereby codes are produced inductively from the text. These codes were then compiled and edited to reduce repetitive or unnecessary codes and compared to the existing literature. In cases where inductive codes clearly mirrored a construct already in use in the literature, the name of the code was changed to reflect that of the construct. This included "perceived risk," "perceived benefit" and "female offspring" for example. Each code was clearly defined to ensure proper classification and the coding system was completed (Neuman 2003). Transcripts were entered into Atlas-ti, a qualitative data analysis program, and re-coded using the final coding system. The final list of codes can be found in Table 4.1. Patton writes, "the first task in qualitative analysis is description" (1990:374). Once coding was completed, primary memos were written to summarize the data for each code and to identify important concepts or themes. In order to write these memos, all of the data under a specific code were compiled and read by the PI, who 6 The procedures outlined here were submitted to and approved by the Colorado Community Institutional Review Board, as well as the Human Subjects Research Committee at the University of Colorado Denver as HSRC Protocol 2007-131. 37

PAGE 48

then created a summary of the data. This was repeated for each of the codes in the final coding system. Through this examination, points that illustrated a consensus among participants, or "recurring regularities," were noted, as were any experiences or comments that seemed to stand out (Patton 1990:403). Given the qualitative, inductive nature of this project, it was particularly important to note those things that participants viewed as significant or essential, not just the researcher (Patton 1990). These points were then marked as possible concepts or themes to return to at the next stage of analysis. Table 4.1. Final code list Discussion with Physician Decision-Making Drugs: Tamoxifen Experience with Breast Cancer Drugs: Raloxifene Importance of Breasts Placebo Female Offspring Having a Sister Friend/ Acquaintance with Breast Cancer Perceived Risk: Study Participation Motivating Factors Perceived Benefits: Study Participation Health History of Participant Perceived Risk: Breast Cancer CCRP: Organization Perceived Risk: Side Effects CCRP: Nurses Enrollment Process CCRP: Study Events Overall Study Experience Other Breast Cancer Activities Perception of Medical Research Previous Trial Participation Knowledge of Breast Cancer State of Mind Denial Role of: Physician Reconstructive Surgery Role of: Friend Side Effects Experienced Role of: Spouse Family History: Breast Cancer Role of: Family Member Family History: Other Cancers Proactive Behavior Upon completion of the primary memos, these concepts were drawn from the initial memos and used to create clusters. Clustering is a qualitative method in which similar concepts, actions, processes, etc. are grouped together into categories or 38

PAGE 49

"clusters" which are then given a name based on the members of the group. "We're trying to understand a phenomenon better by grouping and then conceptualizing objects that have similar patterns or characteristics" (Miles and Huberman 1994:249). Resulting clusters included "The Role of CCRP," "Social Support," "Perceived Risks of Study Participation," "Perception of Medical Studies," among others. Secondary memos were then written which summarized the connections within and between the clusters, and interpreted the results. 39

PAGE 50

CHAPTER 5 MOTIVATIONS FOR ENROLLMENT Demographic Characteristics of Participants Seventeen women were interviewed between June 2007 and September 2008. Participants ranged from 55 years to 73 years of age, with the majority of participants under the age of 70 (Table 5.1 ). Fifteen of the seventeen women self-identified as either "White," "Caucasian," "Anglo," or "European American." Two declined to answer the question. The majority of participants were currently married, with thirteen respondents answering as such on the Demographic Questionnaire. Two were divorced and one participant identified herself as a widow. Only one participant identified as single. As a group, the participants in this study were well educated with all having had advanced beyond the high school level. Two had attended "some college," six had completed bachelor's degrees, and four had completed graduate degrees. A total of five participants had progressed to the post-graduate level. T bl 51 A f a e .. ,ge range o participants Number of Age Percentage Participants Ran_ge 5 55-59 29.4 4 60-64 23.5 5 65-69 29.4 3 70-74 17.7 Family history of breast cancer is an important risk factor for the disease, particularly having had a first degree relative with breast cancer. Twelve of the seventeen participants responded affirmatively to the family history question. Of these twelve, four listed multiple family members diagnosed with breast cancer, two did not elaborate on the question, and the rest only listed one family member. Those 40

PAGE 51

relatives listed included mothers, sisters, maternal aunts, one maternal grandmother and one daughter. A previous study posited the idea that having female children might influence the decision-making process for women considering participation in chemoprevention trials (Lovegrove eta!. 2000). Though it was not found to be statistically significant in that study, it was a subject worthy of further inquiry. In order to further explore the issue of how female family members might influence decision-making on enrollment during the interview, participants were asked about the number of female children and the number of sisters they had on the Demographic Questionnaire. The results of these questions can be seen below in Table 5.2. T bl 52 N b f:fl h"ld a e .. urn er o ema e c I ren an d Sisters o f participants Number of # Participant Number of # Participant Female Children Responses Sisters Responses 0 6 0 5 1 7 1 8 2 3 2 2 3 1 3 2 In the case of two participants with female children, their daughters were adopted. A participant with one adopted daughter did not have any family history of breast cancer. The second, who had two adopted daughters, replied "yes" to the family history question though she did not elaborate. Their adoptive status means that the daughters are not at increased risk due to their adopted mother's family history. This was therefore a potentially interesting point for further exploration in the interviews. The results of the in-depth interviews are anchored within a discussion of the analytic levels of the socio-ecological framework, with primary emphasis being given to those levels that emerged as the most important to women during the interviews: 41

PAGE 52

the individual, interpersonal and community levels. The institutional/organizational and social structure, policy, and systems levels will be addressed more fully in Chapter 6. It is important to keep in mind Stokols' four assumptions, however, when examining the factors that influence women's decision-making around enrollment in breast cancer chemoprevention trials (Sallis and Owen 2002). The lines between levels of influence may blur and feedback often occurs that may make classification of factors to simply one level alone difficult. Individual Factors Perception of Breast Cancer Risk In order to be eligible for the STAR trial, women had to be identified as high risk for developing breast cancer. Both this medically defined risk, as well as prior experience contributed to women's perception of their personal risk for the disease. An association between having a high perceived risk of developing breast cancer and the decision to take chemoprevention drugs or enroll in a chemoprevention trial is supported by the literature (Bober eta!. 2004; Lovegrove eta!. 2000). That perceived risk is influenced by more than simply a Gail score is supported by the fact that women tend to overestimate the risk of breast cancer. When asked to estimate the risk of breast cancer for women with a family history at the ages of 30, 45, and 60, participants in one study came back with numbers that were too high by a factor of 24 for the 30 year old and by a factor of 4 for the latter two ages (Lovegrove et al. 2000). Also, while women are likely to overestimate their susceptibility to breast cancer, at least one study found no association between interest in chemoprevention treatment and Gail score (Bastian eta!. 2001 ). It is therefore not solely a medically defined risk that is important to women; other factors are influencing how they perceive their own risk. 42

PAGE 53

The women in the current study generally discussed their own risk in relation to either a family history of the disease or a personal history of other breast disease. Twelve women had a family history of breast cancer, though this fact did not have a similar effect in all women. Rapp, in her study on the social impact of amniocentesis, illustrates that individuals' understanding of risk is dependent on their own personal and social history (2000). In the present study, participants differed in how closely linked they felt to their family history and in its relevance to their own life. For some, breast cancer felt like an ever-present part of their lives as in the case of a woman who had lost both of her sisters and her mother to the disease, saying, "I'm always looking over my shoulder." A few women, on the other hand, acknowledged their history, but felt that it was important not to dwell on it, while still another felt that her similarity to her father's side of the family made her unlikely to fall victim to her mother's family history. Family history was clearly important in determining risk status for the trial. Still, the emotional connection women felt to that history and how it influenced the perception of their own risk different between participants. Key eta/. observe that "most women with the disease do not have a family history of it, and most women with affected relatives never develop breast cancer" (2001: 138). However, family history of breast cancer was so identified with being at high risk for the disease that at least two of the women without a history did not even consider breast cancer a threat until one event or another brought it to their attention. For the first, it was her daughter's breast cancer. Despite the fact that all of the women in her immediate family had cystic mastitis 7 and had been told that they were at increased risk for breast cancer, the participant did not truly believe this risk was a real threat due to a lack of family history; that is until her daughter was diagnosed 7 Cystic mastitis is also referred to as fibrocystic disease or fibrocystic changes (ACS 2008b). 43

PAGE 54

with breast cancer at the age of 27. Another woman had no family history before her own experience with LCIS and said that right up to the point when they told her she would need a lumpectomy, she believed the diagnosis was a mistake. Although women often overestimate their risk of breast cancer, it appears that in these two cases at least, a lack of family history caused them to underestimate risk (Colditz 1997). In noting that personal experience can often shape how likely an individual may perceive a given event, Rapp writes, "this homegrown sense of statistics can be quite powerful." Personal health history also contributed to perceived risk for breast cancer. Some women had previous experience with precancerous tissue in their breasts, dense breast tissue, or other breast conditions that led to medical oversight. Four women had been affected by LCIS, while a fifth had atypical ductal hyperplasia, or ADH8 Four women mentioned having been on HRT prior to their participation in the study, which is known to increase the risk of breast cancer (Colditz 1997). Nor were precancerous breast tissue or carcinoma in situ of the breast the only experience participants had with cancer. Two women spoke of their previous experience with skin cancer, while cervical and anal cancers had affected another participant each. This previous experience with cancer was a significant motivating factor, according to one participant: But I am a cancer survivor. Urnabout twenty, well, it'd be twenty four years now. Twenty-four years ago I was diagnosed with cervical cancer and I completely understand that there has yet to date been no genetic correlation found between cervical cancer and breast cancer. Typically it's, you know, breast, ovarian, prostate. But given the circumstances, I just couldn't help but feel that somehow my body had 8 Atypical ductal hyperplasia, or ADH, refers to a condition in which abnormal growth occurs in the ductal cells of the breast. Atypical refers to the fact that these cells "are slightly distorted in how they are arranged" (ACS 2008b ). 44

PAGE 55

a history of making really poor decisions at the cellular level, and so that was very concerning to me. As this quote illustrates, risk was not defined based only on risk factors listed in the Gail model. A personal health history that includes other health issues may increase a woman's perceived risk of breast cancer, even if, as in the case of the woman with cervical cancer, it has not medically been shown that the two conditions are associated. Perception of Medical Research A positive perception of medical research was also a common thread in the interviews conducted during this study. The value of medical research to women's health, one participant felt, had already been shown by studies that had revealed the risks of hormone replacement therapy, while another pointed out the benefits that have been gleaned from clinical trials for pediatric cancers. Others expressed similar sentiments in a more general manner, calling these types of studies an important source of knowledge. The STAR trial in particular was seen to be important and informative, with one participant calling it "one of the best studies really of all time, I would say." In particular, she appreciated that once the preliminary results were in indicating that raloxifene was effective, the study was un-blinded and women were notified of which drug they were on. Women demonstrated a good understanding of the trial in general, with one feeling it had particularly good oversight. When asked whether they or others they knew had concerns about enrolling in the trial, most women said no, with several citing the education level of themselves, their spouses, or their friends. 45

PAGE 56

The relatively high education levels in this group certainly appeared to play a role here. All of the women had attended at least some college, while all but two held college degrees. That most felt they understood the value and importance of clinical trials and how they worked was cleara few had even participated in previous studies, while two had considered doing so. While none of the other studies involved taking an experimental drug, at least two had involved dietary restrictions, and urine or blood samples. The topics of these other studies included breast cancer, diabetes, cervical cancer, and cancer prevention through nutrition. These valueshigh education, positive views of clinical trials, the importance of scientific knowledge while reflected here in individual beliefs, may in fact reveal more about the community level of influence, which will be discussed further in a later section. While only about half of the women considered themselves to be fairly knowledgeable or more knowledgeable than the average person when it came to breast cancer, other skills or experiences they possessed were of importance here. One had worked for an oncologist in the past; another was a laboratory scientist. Two had experience working in health advocacy while another was a history teacher who subscribed to Harvard's Women's Health newsletter. Several had read Dr. Susan Love's Breast Book (Love 1995) or saved articles on the subject since the diagnosis of a family member. And while the participants may not have considered themselves particularly knowledgeable, this may in fact reflect a biased view of what constitutes "average," as seen in one woman's statement: "I had heard oftamoxifen. I think everybody has." In reality, research with a more diverse group of women has shown that knowledge about breast cancer in general, and tamoxifen specifically, is not widespread even among women at high risk for breast cancer (Cyrus-David and Strom 2001). 46

PAGE 57

Other topics that came up also reflected a fairly good awareness of issues related to breast cancer, such as the available treatments, well-known publications on the subject, the specific drugs being tested in the study and their side effects, the dangers of hormone replacement therapy, and the importance of early detection in prognosis. Two also brought up the genetic mutations that indicate a hereditary risk for breast cancer. That this was a proactive group of women was obvious in the fact that many of them, when faced with a friend or family member's breast cancer or when making the decision to enroll in the trial, chose to do their own research on the subject rather than rely solely on information from a physician. Perceived Benefits of and Barriers to Participation Concerns were given about certain aspects of study participation, such as the time commitment involved, their ability to follow through with study activities, and the side effects associated with the study drugs. However, only a couple of women expressed these concerns. Most participants viewed study participation as a no-lose situation or did not view the potential risks as constituting a strong deterrent. The risk of side effects specifically did not appear to be a major concern for the participants when considering enrollment in the trial. However, different reasons were cited for this lack of concern. Two women were not at risk for many of the side effects due to previous hysterectomies, while others felt that since they were already having similar symptoms due to menopause, they were not worried about the effects of the trial drugs (which may mimic these symptoms). You know, right from the first time they talk to you, they pretty much assure you, you know, that so far these drugs haven't, you know, provid-, proved to have any side effects, except maybe hot flashes. Well, we're all used to that anyway ... 47

PAGE 58

This particular participant also expressed a sentiment echoed by others in that they felt they were made aware of potential side effects by the study personnel prior to enrollment and were reassured about the likelihood of them occurring. The voluntary nature of participation and the possibility of dropping out were also brought up by a woman who was very familiar with clinical trial protocol in general. The participants in fact mentioned only two specific potential side effects that had concerned them prior to enrollment. The first were hot flashes, which were not considered by any to be a deterrent (though in some cases did prove to be difficult during the actual STAR experience). However, one woman also mentioned endometrial cancer as the potential side effect that concerned her most. Interestingly, she was also the only participant to mention that a friend or family member expressed concern over side effects. Although she couldn't remember exactly what he said, her older brother was not supportive of her enrollment, which she believed was due to the risk of side effects in general and endometrial cancer in particular. Studies by Bober eta!. (2004) and Yeomans Kinney et al. (1995) found that the decision against tamoxifen use or trial enrollment was associated with concern over side effects. The results found here may reflect the fact that those who do have strong concerns were underrepresented in this particular population given that all of these women made the decision to enroll. Two participants expressed another sentiment that perhaps accounts for the general dismissal of the side effects associated with tamoxifen and raloxifene in other participants. These participants saw the threat of breast cancer as much greater and more severe than any risks associated with tamoxifen or raloxifene. A woman who had already dealt with cervical cancer and LCIS in the past had this to say: I was significantly more worried about the side effects of breast cancer. And I was never diagnosed with having breast cancer. I did not have breast cancer. I was in a pre-cancerous state. So my 48

PAGE 59

concerns about the side effects ofbreast cancer seriously, seriously outweighed any of my concern about the drugsalthough I tried to be knowledgeable about them. Benefits that might be gained through study participation were more numerous according to the results of this study (Figure 5.1 ). Five out of the seventeen participants specifically stated that they saw no problem or downside to participation with one woman saying" ... I figured that I didn't have too much to lose, either way." STAR was evaluating the effectiveness of raloxifene for the prevention ofbreast cancer, compared to the current standard oftamoxifen. However, despite the fact that it was not yet approved for the purpose, participants were fairly certain that they would experience benefits from raloxifene if assigned to that arm of the trial. Time commitment was seen as a potential problem by one participant, however, others viewed the screening and oversight provided by the study as a benefit to those who were at risk for developing breast cancer, a view which is supported by the literature (Lovegrove et al. 2000). Other benefits of study participation will be discussed in later sections. 49

PAGE 60

I STUDY PARTICIPATION I l I I PERCEIVED RISKS PERCEIVED BENEFITS Side Effects Prevent Breast Cancer Time Commitment No Placebo Ability to Follow Prevent/Treat through Osteoporosis Screening/Oversight of Health Help Others Further Medical Knowledge Figure 5.1. Perceived Risks/Barriers and Benefits of Study Participation Interpersonal Factors The Role of the Physician The role of the physician in recommending study participation has been heavily emphasized in the literature and in fact has been strongly associated with the decision to participate in chemoprevention trials (Bober et al. 2004; Yeomans-Kinney eta!. 1998a). While the influence of physicians on women's decision-making could be seen in the current study, the recommendation of a physician was not a universal influence among participants. Eight women first heard about the trial from one of 50

PAGE 61

their health care providers. Some of the women specifically remember their physicians broaching the subject with them. Others could not remember the exact occasion, but believed that was where they first heard about it. The following two passages illustrate each of these scenarios: But my doctor is pretty on top of things and uh, just around the first of the year he notified, he called me and told me he'd been looking into it to decide. And he said that he could either put me on a regimen, a five year regimen oftamoxifen himself or he'd been talking to some of his doctor friends and there was this STAR study and he thought I was a real good candidate for it, if I wanted to go that route. But either way he thought I should do something. Just because I was too high-risk. So .... I said I'd do the study. I believe it was mentioned. Dr. Jones was my doctor at the time and I believe she did mention it. And so I went ahead and signed up! In the first case, the woman's high risk for breast cancer made her a candidate for tamoxifen, a regimen her physician wanted her to follow anyway. The only decision to make, as she saw it, was whether or not to enroll in the study, not whether or not to take chemoprevention drugs. Given that all of the women in the STAR trial were at high-risk for breast cancer, other participants may have viewed the decision from this vantage point as well; hers was not the only physician to frame the discussion in this fashion: Well yeah, Dr. Smith is my primary doctor, and he thought it was fine. But then he said, "Wait a minute, I want you on tamoxifen, I don't want you on a placebo." But then I explained to him that the other drug wasn't a placebo. So I went on the study. Physicians were not the initial source of information about STAR for all of the women interviewed. In two cases, the women themselves heard about the study 51

PAGE 62

through other sources such as the media or someone else already involved and brought it to their provider's attention, while two others couldn't remember who brought it up first. Regardless of who broached the subject, the majority (n=l3) of women stated that they discussed the study with their physician prior to making the decision to enroll in the trial, who then provided largely informational support in the form of both recommendations and reservations during these conversations. Nearly all physicians were generally supportive of the decision to enroll. The primary reservation expressed by physicians according to the women interviewed involved the two cases mentioned above in which physicians wanted their patients on a chemoprevention drug due to their high risk for breast cancer, as opposed to taking a placebo or doing nothing. One physician mentioned possible side effects, but according to the participant, he did not view the side effects as a large risk: His reaction was with, you know, anything like this there are gonna be side effects. There are side effects if you walk across the street and get hit by a truck. While physician recommendation played a role in several women's decisions to enroll in the trial, the strength of that influence seemed to vary between participants depending on their relationship with the physician. Some participants simply mentioned that their physician thought they should enroll, or asked if they would be willing to participate. Another mentioned seeking out a physician she had a clinical relationship with for their opinion when another physician told her about the study. Two physicians knew their patients' personalities well enough to take a more personalized tack in addressing study enrollment. For example, one participant with a long history of working for social justice causes and with underprivileged women said that her physician appealed to her passion for this work. 52

PAGE 63

... and I talked to my oncologist. And the upshot of his recommendation was, he said, "It's for the greater good." And that cinched me; that was it. That did it. Another appealed to the proactive nature of his patient with whom he had a long and close relationship due to other health issues she had had in the past. And when I had the breast tumor and had the diagnosis and my doctor explained the STAR trial, he even said right up front that he thought this would be ideally suited for me because he knew, it, it would have just been impossible for me to just sit and say "Wow, I've just gone from very, very low risk of breast cancer to very, very high risk .... and I'm not going to do anything about it?" I, I just, that just would not have been good for me mentally ... Well, again, my gynecologist/ obstetrician at the time recommended it to me. And he knew me well. I mean, my, my son was a vaginal delivery after cesarean which at the time virtually nobody, there was only one hospital and like, one doctor that would even let you do that, so this was the doctor who had seen me through that. And then when I was diagnosed with cancer a year after that, urn, he made my cancer diagnosis. He donated blood for me while I was in the hospital... and so I so respected his opinion and had such trust in him. The fact that he said he thought this would be really, really good for me almost made it a done deal. The importance of physician recommendation has already been established by the quantitative literature on decision-making for chemoprevention trials. The results of this study provide context for these findings and suggest that while not all women consult their physicians prior to making the decision to enroll, for those who do this is an important conversation. However, the data gained here also indicate that rather than a one-sided relationship existing between physician and patient in which the physician possesses all of the information, some women brought the study to their physician's attention or arranged for their physician to speak to CCRP for more 53

PAGE 64

information, perhaps reflecting a more shared treatment decision-making model (Charles et al. 1999). Prior Experience with Breast Cancer Prior experience with breast cancer among others in their social network was a frequent topic of conversation during the interviews conducted for this study. Only two participants said that they had known those with breast cancer, but as one put it, no one "I was emotionally close to." Twelve had a family history ofthe disease, while one without any family history had had a daughter diagnosed with breast cancer. As one woman put it, "I don't think you could go out into this library and find anybody who doesn't know someone that has gone through this." This was an important factor because women's previous experiences with breast cancer likely shaped how they viewed their own risk for the disease as well as its severity. Therefore, though perceived risk for breast cancer and perceived severity of the disease are individual-level beliefs that have been examined in the literature, these beliefs may be influenced by experiences at the interpersonal level as discussed here (Bober et al. 2004; Lovegrove eta/. 2000; Yeomans Kinney et a!. 1998b ). The prognosis, treatment, and outcomes of family members or friends with breast cancer were influential in how women conceptualized the disease. In a few of the cases of women who had had a family member with the disease, this had spurred them to take further action to monitor their own risk by seeing an oncologist or attending a high-risk clinic. While some had only known one or two who had gone on to recover, others had lost multiple people to either breast cancer or cancer in general, with some experiences falling between the two extremes . my mom and my aunt both had breast cancer, they obviously had the kind that is very slow growing and it was found early in a mammogram and treated, and so that's more than ten years ago. 54

PAGE 65

And people, you know, we knew about it and thought, in fact, she thought she was in remission, and thought she was gonna be okay. And then she had a setback and died. The issue of being proactive about one's health or, on the other side of the coin, being in denial about one's health, emerged through these discussions. Two felt that enrolling in the STAR trial was something they could do to take a proactive approach to their own health. While only three women brought up the issue of denial, it seemed to have had a powerful impact on them. They had friends, acquaintances or family members who had waited too long to seek screening or treatment and as a result had a worse prognosis when they eventually did go in. In all three cases this person had had suspicious symptoms or felt ill for some time before seeking care. One woman was not as close to the breast cancer patient she knew and seemed to primarily feel empathy or pity for her. However, the other two participants were referring to sisters and expressed a sense of frustration that their siblings ignored what was going on. Interestingly, both felt that their sisters should have known better and that the behavior was out of keeping with their character: And the irony of the thing is that Carrie had an abnormality on her breast in February and did not do anything about it until August. And that's one question I wished I had asked her. .. "Why in the world did you not?" You know, I mean, you were "salad dressing on the side" and "oh no, you shouldn't eat that" and all this stuffthat she did for years ... and yet she had invasive breast cancer, which showed up as a, like the classic orange peel kind ofthing. And she did notice itshe was on a trip to Seattle or some place and noticed it. And then never did anything about it until August. And by then it was Stage IV. And you know, it, it was always, that's just a mystery that I will never have an answer to. 55

PAGE 66

Having Female Offspring/ Sisters Interpersonal relationships were also influential in how women perceived the benefits of enrolling in STAR. Benefits were seen to accrue not only to the participant herself, but also family members as well. Having female children was an important consideration in participants' decision to enroll in the trial due to the potential knowledge that could be gained for future cancer treatment. For seven of the women, the threat breast cancer posed to their daughters and the desire to further research that might someday protect them were important factors to consider. Just one participant had a daughter who had already been diagnosed with breast cancer. She had undergone a mastectomy and chemotherapy, and was well past the five-year survival mark. However, the fact that her daughter was diagnosed with breast cancer at the young age of 27 had been a frightening surprise. She had not realized that such a young woman could develop breast cancer, particularly without a family history, and called it, "the scariest thing that ever happened." Only three of the ten women who had daughters did not cite this as an influential factor in their decision. One of these participants said that her daughter was out of town at the time she was considering enrollment and therefore she did not discuss it with her, though she did not say either positively or negatively whether the fact that she simply had a daughter influenced her at all. Another who had worked in the field of cancer screening and prevention felt she would have participated with or without having had a daughter. The third had an adopted daughter and no family history; she simply said that having a daughter did not influence her decision "at all." Having female siblings was also important in two ways. While twelve women overall had at least one sister, six of these had a sister who had been diagnosed with breast cancer. One woman said, "When it happened to Carrie, it was just like it happened to me," illustrating how a sister's breast cancer may have shaped the way 56

PAGE 67

women viewed breast cancer. Others felt having a sister was influential in a more indirect way; by participating in research that could increase our knowledge around breast cancer prevention they were potentially helping their sisters who could be at risk for the disease in the future. A sister with breast cancer is also a first-degree relative. This therefore may have made women more aware of their own risk, such as one participant who sought out an oncologist for herself after her sister was diagnosed. Lovegrove et a/. (2000) did look for an association between number of female offspring and the decision to enroll in a chemoprevention trial but did not find the association to be statistically significant, despite the fact that women who chose to enroll did in fact have more female offspring than their counterparts who made the opposite decision. Other studies have not examined this variable in depth, but the connection between having female offspring or siblings and choosing to enroll in chemoprevention trials is an important one to explore further for several reasons. First of all, one of the key risk factors for eligibility in the STAR trial is having a first-degree relative with breast cancer which means that some sort of family history of the disease and its associated risk is common among the subject pool whether or not that history is due to actual heritable genetic mutations, such as the BRCA 112 mutations. The results of this study seem to indicate that this is an important contextual factor for many of the women who chose to enroll in STAR. Second, in a socio-ecological framework, having a sister or daughter diagnosed with breast cancer may likely be the primary experience they have had with the disease and therefore shapes a woman's perceptions of how severe breast cancer is and how large a risk it constitutes. It seems reasonable to argue that a family history that includes a sister or daughter with the disease may make the risk of breast cancer feel more intense than having had a relative that was farther removed, such as a grandparent or aunt. 57

PAGE 68

Social Support The discussion around social support on an interpersonal level was complex. It involved support given both from others to the participant, but also from the participant to others in their lives. Nor was social support always discussed within the context of the decision-making process for enrollment in this particular trial. Women often brought up support they had given to others with breast cancer either in the past, or during their enrollment in the trial, while support was received by participants before the decision-making period, during, and after they had decided to enroll in the trial. Figure 5.2 illustrates some of the examples of different types of social support brought up during the interviews given by and to participants. Some of these acts actually took place during the period of enrollment and will be discussed in Chapter 6. However, most took place either during the decision-making process, or were influential in the participant's past. The specific role of important people in one's social network, including a spouse, family member, or friends will be discussed. 58

PAGE 69

I l 1.--------1, SOCIAL SUPPORT ,l--------,1 Emotional: Given By Others to Participant I Family &friends being supportive of enrollment decision Family &friends offering concerns about enrollment Camaraderie offered by other participants Nurses sympathizing with participants who are experiencing side effects Instrumental: Spouse enabling trial participation Physician cooperating with trial personnel Nurses providing reminders during the trial Informational: Physician recommending trial enrollment Other breast cancer patients or survivors offering advice Study nurse telling potential participants about study Appraisal: Physician discussing risk factors with participant Emotional: Given By Participant to Others 1 Offering support to others who had breast cancer Attending appointments with those who had breast cancer Joining in a breast cancer walk Instrumental: Taking on extra duties for a work colleague with breast cancer Enrolling in the trial as a way to help family members Informational: Researching treatments for another with breast cancer Appraisal: o Telling a daughter about her risk for disease Key: Occurred during decision making process Occurred before decision making process or after trial enrollment Figure 5.2. Types of Social Support Mentioned by Women in this Study 59

PAGE 70

Spouse Thirteen ofthe seventeen participants were married at the time of the interviews. None mentioned a partner other than a spouse or indicated the presence of one on the Demographic Questionnaire. When participants were asked if they discussed the study with their spouse or partner prior to enrolling in the study, a few seemed affronted by the question and emphatically stated that it was their decision, or that they told their spouse about it but did not ask for permission. Eleven of the thirteen, however, felt that their husbands were supportive of the decision overall, though some were more enthusiastic than others. Two husbands had cancer themselves in the past, which partly contributed to their support. Primarily this support came in the form of emotional or informational support. However, one participant in particular felt that her husband's instrumental support was also essential to her enrollment and subsequent participation in trial activities. She and her spouse spent a significant portion of each year out of the country at their other home, which was in a remote location and was only accessible by boat. While she was able to work out a situation with the study personnel here in the states where she could fulfill her study requirements outside of the country, she did not feel comfortable driving herself alone by boat to her appointments. Although he had to go to more trouble than most of the husbands of the women I interviewed, this woman's husband was still very supportive. Other Family Members A few other family members were also mentioned in response to this line of questioning, such as siblings or children. However, while some participants shared their plans with these family members, most did not feel that they played a large or significant role in their decision. One participant did have a brother who expressed 60

PAGE 71

reservations about her decision to enroll due to concern about side effects, while the others felt supported in their decision. Friends Only four participants could remember discussing their decision at length with a friend prior to enrolling in the trial. The response two of them received was an expression of concern regarding potential side effects or the time commitment. Another had a friend with breast cancer who was taking tamoxifen. While she said that in the end she would have participated regardless of the woman's reaction, she also delayed saying yes to the study until she could discuss this woman's experience on tamoxifen with her. One participant had a particular friend who had been part of her health journey since the participant's diagnosis of LCIS. This friend had been with her when she received the diagnosis and was supportive of her decision to enroll in the trial, which the participant partially attributed to them both being educated people. The types of social support provided by friends and family members primarily took the form of informational and emotional support. Overall, however, women seemed to feel that the decision to enroll had been theirs alone, and while they may have discussed it with others, they were not necessarily swayed by advice from outside. This perhaps reflects the larger cultural tendency to believe that health is the responsibility of the individual discussed in Chapter 3 (Clarke et al. 2003; Fisher et al 2005). Ofthe advice they did receive, it appears that the recommendation of their physician was the most significant. It must be remembered however, that these women all chose to enroll in the trial. A different picture might be gained from a sample of women who chose not to enroll in the STAR trial. What did appear to be influential, however, was past experience with breast cancer through the cases of 61

PAGE 72

friends and family members. Many of the women who enrolled in STAR had been sources of support themselves for others who were undergoing treatment for the disease, not only emotionally, but providing informational and instrumental support as well which informed their own perceptions of the disease. Institutional/ Organizational Factors Placebo vs. Non-Placebo Trials The study protocol of STAR could be considered part of an institutional level of influence in that women's decisions to enroll in a chemoprevention trial may depend on its design. One of the most important themes that arose during the course of the interviews conducted during this study was the significance given to the fact that the STAR trial was a non-placebo arm trial. This is a topic that has only very briefly been touched upon in the literature but should be included in future studies given the results found here (Yeomans Kinney et al. 1995). Six women said that the nature of the STAR protocol was an influential factor in their decision due to the fact that they knew they would be getting one drug or another, rather than a placebo. Knowing the benefits oftamoxifen in preventing breast cancer and of raloxifene for the prevention and treatment of osteoporosis, and having good reason to think that raloxifene would likely have a similar effect on breast cancer, the STAR trial was typically seen as testing two drugs each with their own benefits against one another, rather than as testing a proven drug against an unproven one. Women's words therefore reflected this view: So it seemed like a no-lose situation. And urn, you know, I think it was helpful to think that this drug would possibly have the potential to help me. And the one that was 62

PAGE 73

approved, they'd already determined that it was helpful. The one that was being tested, the information I read about it indicated that it probably was even, was going to be better than the one that had been approved. So I figured that I didn't have too much to lose either way. The participants were not alone in feeling this way. In addition to the physicians who expressed concern in the above section, several of the participants' spouses expressed similar feelings about the possibility of their wives enrolling in a study that could potentially not benefit them. These individuals were usually reassured when told there was no chance of participants being assigned a placebo. In discussing their reluctance to enroll in a placebo-arm trial, several women elaborated on just why they felt it would have altered their decision, citing wasted time and effort, or a sense ofbeing cheated. One of the reasons I like this trial was that there was no placebo. would be getting something, that it was a test between two drugs. And that very much affected my decision to join because it seems to me after you've been in a study, and let's just say that a drug was very beneficial to the people taking it and you found out you were on the placebo. It seems to me there would be a disa sense of disappointment; of almost of being cheated that you didn't get the drug. Although, it seems to me also when you agree to be in a study that's the chance you take. But, so I liked the fact that I was going to get something. It's a lot of, there's a lot of, uh, effort on your part to be in the study and if you're thinking you're probably taking a placebo, why bother? Not all women felt it would have changed their decision had they been approached to be involved in a placebo-arm trial. One woman talked about the benefits her own mother received from tamoxifen when she had breast cancer. Another woman cited the need for "dummy trials" as well as non-placebo arm trials. Generally though, women discussed the lack of a placebo as a factor influencing them 63

PAGE 74

to participate in this particular trial. Y eomans-Kinney et a!. ( 1995) also noted an association between nonparticipation and concern about being assigned to the placebo arm of the trial. However, in that case, the trial in question was the Breast Cancer Prevention Trial, which was in fact testing tamoxifen against a placebo. Further quantitative research is needed to explore whether women who list concerns about placebos as a deterrent to participation are in fact more likely to participate in non placebo trials. The results here seem to indicate that this may be the case. Tamoxifen Versus Raloxifene The majority of the participants reported an assignment to the tamoxifen arm during the trial (Table 5.3); as a result, tamoxifen was discussed during the interviews with greater frequency than raloxifene. Participants also knew more about tamoxifen as the FDA had already approved it and many knew others who had taken the drug in the past for treatment, often with positive results. One, for example, had had a friend diagnosed with breast cancer who had undergone a mastectomy and been treated with tamoxifen. She had reached the five-year mark and was still in remission. Another participant's mother had a similar positive experience. T bl 53 D a e .. rug assignment d STAR unng Drug Number of Participants Tamoxifen 12* Raloxifene 5 Total 17 One of these twelve switched to raloxifene during the study However, not all knew friends or family members who had had such good results from tamoxifen; other stories were told as well. A friend had been treated with tamoxifen but relapsed. A sister died during the study, despite being treated 64

PAGE 75

with tamoxifen. In the case of the second, the participant changed her study drug to raloxifene mid-trial due to concerns about tamoxifen's ineffectiveness resulting from her sister's experience. There was some speculation during the trial by participants regarding which drug they were on. The intensity of hot flashes experienced by one woman led her physician to suspect she was on tamoxifen. Only two women mentioned the fact that they were given the opportunity to change the drug they were taking, once the trial had been un-blinded. Of these two, only one chose to do sothe woman discussed above whose sister had died. One participant went so far as to compare her drugs with a friend's tamoxifen and a family member's raloxifene. The drugs she was given looked similar to her sister's raloxifene and "it chalked," a trait she associated with raloxifene. Therefore she was surprised to find out in the end that she had in fact been on tamoxifen. Why did she go to such trouble? She perceived the potential side effects oftamoxifen to be more worrisome than those ofraloxifene: "I didn't think I had to worry about the side effects oftamoxifen, until five years later, or six years later, when I found out I was on it!" She was not the only one to consider raloxifene "the good one," as one participant put it. Several of the women assigned to the raloxifene arm thought of it as the preferable drug, either due to research they had done prior to enrollment or due to the results of the study. One noticed an improvement in her bone density during the trial, which she ascribed to being on raloxifene. 65

PAGE 76

Community Factors Altruistic Motivations One of the most important themes that emerged from the interviews conducted during this study was that women who enrolled in the trial did so in order to help others. Fifteen of the seventeen women explicitly stated that they were motivated to enroll by altruistic sensibilities while making their decision. This was framed in multiple ways, for example as wanting to "help other women," to "promote knowledge," "protecting other people," and being "useful." A couple saw it as an obligation or duty; as one said, "Somebody has to do this stuff." A strong sense of the importance of finding better treatments or perhaps even a cure someday was revealed in these interviews; of women doing what they could to help other women, whether stranger or relative. Nor did women believe they were the only ones who felt this way: Well, I mean, I think that as a general rule, we, as women, want to make sure that if we've gone through this, someone else doesn't need to go through it. And ifthere's a way that, you know, studies can show that it not only helps that but helps other things like your bones and different things that it's certainly a benefit to everybody. Interviewer: And you said, when I called you, something on that, I wrote it down 'cause you said, "I think you're talking to a group of pretty altruistic people." Participant: Yeah. 1: And you think that is true of all the other women in the study too? P: Oh yeah. Yeah. Because you're not really doing it for yourself. I said because it was already proven to be a good drug. But it's really for our sisters and our mothers and whatever. A couple ofthe women had backgrounds in women's health advocacy or social justice, which motivated their enrollment. One had worked for an organization 66

PAGE 77

active in preventing breast and cervical cancer and had also been active in trying to link under-served women with preventive care programs providing Pap smears and mammograms. She not only understood the importance of prevention and screening for these diseases, but also the benefit to herself of a study that provided these things for free. This particular woman saw her previous work with vulnerable women as the most important motivating factor to her participation. A second participant viewed enrollment in the STAR trial as a logical decision in a lifetime of fighting for social causes: I'm kind of a weirdo urn, when it comes to civil rights and things like that. I was at Martin Luther King's thing in 1963 in Washington. And I've been working for an AIDS organization for twenty years now. So I started out when it was all gay men. And something about social welfare is very important to me. The theme of altruism ran very deeply throughout these interviews. This echoes collective cultural themes in the United States that emphasize personal responsibility, yet also a shared duty to help others. Cultural themes belonging to a society are typically widely shared and provide a common point from which to draw meaning and understanding of life events (Becker 1994 ). Siminoff and Arnold ( 1999) point to similar ideas of altruism reflected in decision-making for organ donation also. That this value of altruism, of wanting to help others, was not only espoused by the majority of these women but was also assumed by the women themselves to be a shared value illustrates just how much it permeates their social environment. The women quoted above don't just ascribe their desire to help others 67

PAGE 78

to their own distinct personality or to their own unique background, but rather identify it as a trait common of woman as a group, or at least of subgroups of women. Moreover, that trial participation was viewed as a method through which a desire to help others could be fulfilled may reflect other shared cultural values. Becker, in her study on the disruption infertility can cause in the lives of men and women, points to the importance of the idea of continuity in the United States (1994:40 I). Americans, she argues, value "order in the universe," "productivity and progress" and struggle to make sense of disruptions to this order. Illness or other unforeseen events are just such disruptions. Becker writes, The emphasis on will, that hope and determination may will a change in the course of an illness, reflects efforts to control and order illness through rational determinism and creates a specific cultural response to illness: to fight the illness and overcome it (1994:396). The women in this study do not have breast cancer, yet it is a very real presence in many of their lives just the same. Enrolling in a chemoprevention trial could, on both a personal and a social level, provide a tangible way for participants to fight the disease and to change its potential course in both their own bodies and those of other women. Other Breast Cancer Activities Although study participation was tied to a desire to help others, this was not necessarily associated with participation in other breast cancer activities in the minds 68

PAGE 79

of the women I interviewed. One woman was quite emphatic about this and viewed the study social events as being in a similar vein to the various fundraising walks. I get their information about the stuff and I just throw it away. I guess partly I don't want, I just don't want to join the club. I really don't and I don't go to the March for the Cure or anything. Urn, I have no desire to do that...Ijust don't want to deal with it at that level. While other women had donated to friends or acquaintances participating in fundraising in the past, only one was currently involved in fundraising efforts. Five had participated in various breast cancer walks, but several felt that they had become too large now and the crowds were a deterrent. Four out of the five who had participated in walks had a family member diagnosed with the disease, while the fifth had LCIS in the past. Social Structure, Policy, and Systems Factors Media While it was not a focus of the interviews, multiple participants mentioned the role of the media as a source of information about the trial or about breast cancer. Two of the participants had initially heard about STAR through a print article or television ad, while at least three others couldn't remember for certain but believed that was where they might have seen it first. These women were then proactive about finding out further information, again reflecting a sense of individual responsibility for personal health. When asked how knowledgeable they were about breast cancer many of the participants cited newspapers, magazines, newsletters, books, and the Internet as sources of information. One woman noted the increasing amount of information on the subject, saying, "There's been more to read, for one thing." The books of well-known personalities within the breast cancer movement, such as Nancy 69

PAGE 80

G. Brinker of the Susan G. Komen for the Cure Foundation and Dr. Susan M. Love were brought up as sources of information that were tapped. Other Factors The discussions of motivating factors that took place during this study primarily focused on more proximal levels of influence that that of social structure, policy, and systems factors. The one specific factor that was just briefly touched on was the media. Yet while the women in this study did not discuss it, the very fact that they were able to participate in STAR related to changes at the policy level. Prior to the initiation of the CCOP network by NCI, clinical trials were typically located at large research centers or teaching hospitals. Enrollees in these trials consisted of those who lived in the vicinity or had the means and ability to travel to these centers (CCRP 2009b ). However, the system of CCOPs changed that by taking clinical trials to the local level and allowing for increased opportunities for enrollment. A second important factor that cuts across the community, institutional, and social structure level is the impact of the history of clinical trials in the United States. Women in this study, who nearly all identified themselves as White9 in general held very positive views of medical research, seeing it as an important source of knowledge. Furthermore, they felt that medical research was typically conducted with benign intentionsno participant questioned the motivations of those conducting the STAR trial or of CCRP. This may not be the case with all groups. Certain sub populations may be suspicious of the medical profession and the research it conducts. 9 "White" is the term currently used by the U.S. Census Bureau and so is the term used here. However, on the Demographic Questionnaire, participants were asked to write in their race or ethnicity on a blank line. Participants responded with multiple 70

PAGE 81

An often-cited example ofthis would be the legacy of Tuskegee within the African American population. A long history of racial discrimination in the United States, of which Tuskegee is just one example, has contributed to a distrust of the medical profession among many African-Americans, a fact which has inhibited screening and treatment for multiple conditions, breast cancer included (Armstrong eta/. 2005; Thomas and Quinn 1991 ). Other sub-populations may struggle with language barriers, or understanding study protocol or the intentions of medical research, such as in a study by Nguyen eta/. (2005) with Asian-American women. Socioeconomic status should also be considered here. In examining knowledge of reproductive technology among a diverse sample of women, Rapp noted that "virtually all middle-class women knew about the test from a dense and overlapping nexus of friends, medical professionals, and books" (2000: 114 ). A similar phenomenon was found in the present study. All of the women in this study were fairly well educated, with several having advanced degrees. More than one woman mentioned her education contributing to her understanding of either the STAR trial itself, or medical research in general. Whether or not their knowledge of breast cancer or chemoprevention was high, most of the women possessed the skills to research the information they needed prior to enrolling. Furthermore, they Several worked in occupational fields that exposed them to information about cancer. Furthermore, all lived in what could be classified as middle-class neighborhoods, and cost was not cited as a barrier to participation. The results of this study may have been very different if conducted among a low socio-economic status population. The criticism that breast cancer discourse in the United States is largely focused on "the interests ofwhite, middle-class, professional women" has been noted terms, which included "White," "Caucasian," "Anglo," or "European American." Two participants left the question blank. 71

PAGE 82

in the past (King 2006:x). The population of this study mirrors these characteristics. While this criticism has generally been aimed at a movement that has focused on finding a cure for breast cancer, that it has relevance for chemoprevention is clear. Cancer clinical trials have had a difficult time recruiting minority participants and further research is needed to identify important factors for their decision-making (Advani eta!. 2003; Nguyen et al. 2005). While there may be cultural beliefs associated with trial non-participation, Advani et al. (2003) identified income and education as important factors as well. In that study, African-American patients had lower education and income levels than their white counterparts; both were shown to be factors associated with patients being less willing to participate in a clinical trial. Among Asian-American women, economic and language barriers were also found to be important (Nguyen et al. 2005). These issues must be addressed if chemoprevention trial populations are to be more diverse than they currently are. Given the framework of a socio-ecological model then, it is possible to see how each ofthese levels of influence fit together to impact women's decision-making for enrollment in the STAR trial. Figure 5.3 illustrates those factors at each level that resulted from the interviews conducted here, as well as those identified at the social structure and policy level that were pertinent to the population in this study. The institutional/organization level was more influential in examining women's experiences while enrolled, which is when women's primary interactions with CCRP took place. 72

PAGE 83

Interpersonal: Relationship with Physician and Physician Recommendation Prior Experience with Breast Cancer Among Social Network Having Sisters or Female Offspring Spousal Support Individual: Perceived Risk of Breast Cancer Perception of Medical Research Perceived Risks & Benefits of Study Participation Education Level 5.3. Socio-Ecological FrameworkFactors that Influenced Women': )ecision-Making While Considering Enrollment in STAR 73

PAGE 84

CHAPTER6 PARTICIPANT EXPERIENCES WHILE ENROLLED IN THE STUDY OF T AMOXIFEN AND RALOXIFENE During the interview sessions, conversations naturally turned to women's general experiences while they were enrolled in the trial. These generally centered around three levels of influence: individual, interpersonal, and institutional. While these experiences do not address the specific aims of this study, the information given may be useful for understanding those factors that make continued study participation easier or more difficult for women, as well as what aspects of study protocol were particularly valued by women. Individual Factors Despite participants generally reporting a lack of concern regarding the potential side effects oftamoxifen and raloxifene prior to enrolling in the study, ten did in fact experience symptoms during the study period. Seven of these ten however were not positive whether these symptoms either qualified as side effects or if they were actually related to age, menopause or other health issues, rather than the drugs themselves. Symptoms brought up in interviews and the number of women who mentioned them can be seen in Table 6.1 While most women only experienced one symptom, three experienced between two and four different symptoms during the study period. By far the most commonly reported issue was hot flashes, which ranged from "warm flashes," to "significant," to "horrific" depending on the woman being interviewed. One woman experienced such severe hot flashes she nearly had to drop out ofthe study. 74

PAGE 85

Nor were side effects restricted to one drug or the other, with women on both tamoxifen and raloxifene reporting them. While it is true that participants on raloxifene only reported two of the symptoms (hot flashes and vaginal dryness), it is difficult to draw any conclusions from this given that over twice as many participants in this study were on tamoxifen as were on raloxifene. Again, these are symptoms reported by women during the study period, and which women believed might be associated with either tamoxifen or raloxifene use. However, in some cases there was no verification ofthat fact. T bl 61 S a e .. ,ymptoms expenence db ywomen d STAR unng Symptom Number of Number of Symptoms Number of Ex_perienced Women Affected EX]!_erienced Women Affected Hot Flashes 7 Any Symptoms 10 Night Sweats 1 Only 1 Symptom 6 Vaginal Discharge I 2 Symptoms I Vaginal Dryness I 3 Symptoms I Hair Thinning I 4 Symptoms I Charley Horses 1 Weight Gain I One participant was not specific enough "Nails went bad" I to categorize Interpersonal and Community Factors Nurses Participants spoke with fondness and appreciation of the nurses they worked with during the study period. Others have noted the importance of social support during times of stress associated with health, particularly instrumental and emotional support (Bloom et al. 200 I; Kawachi and Berkman 2001; Tho its 1995). The nurses provided not only information and answers to questions but also offered emotional 75

PAGE 86

support when women were struggling with intense hot flashes or other symptoms. Participation in the STAR trial required keeping periodic appointments and nurses provided reminders with phone calls or letters. Several women reported developing fiiendships with their original nurses and were disappointed when those particular nurses left the study, which may be illustrative of the fact that women tend to "maintain more emotionally intimate relationships than men" (Kawachi and Berkman 2001 :461-2). One participant in fact, said that she did not have as personal a relationship with her second nurse as she did with her first, which was one reason for her dwindling participation in necessary study appointments. The only complaint was made by a participant who felt that she was given vague answers when she asked her nurse if physical symptoms she was having could be due to the drug she was taking. However, she also acknowledged that this was not necessarily the nurse's fault: Sometimes I wish, urn, they could tell me if something was a side effect or not, but I'm not sure they knew. I mean, how would they know cause they didn't have all the study results? Social Events The social and informational events held throughout the study period were another source of support for participants. Many women mentioned the annual luncheons they attended, and the benefits they gained from these occasions; some even brought friends or family members with them. The informative nature of the speakers and study updates these occasions provided were appreciated. In addition, participants, in particular those who had enrolled at the beginning of the trial when the group was smaller, felt that these meetings encouraged a sense of camaraderie among the women, who were able to speak with others going through the same things they were. One woman spoke of having done the Komen Foundation fundraiser walk with some of the other STAR participants early in the trial, which she remembered as 76

PAGE 87

being organized by CCRP. This is in keeping with previous research into social networks, which has illustrated that being a part of a larger web of social relationships increases the type and breadth of support an individual may access when in times of stress (Bloom et al. 2001; Kawachi and Berkman 2001 ). Although the social events were not held often, for some of the women in the study, they offered unique opportunities to be around like-minded individuals who were taking part in the same experience that they were. One ofthe key motivations women gave for enrolling in the STAR trial was an altruistic desire to help others which is consistent with studies on clinical trials in other areas of medical research (Simon et al. 2006). As a follow-up question, therefore, participants who attended these social events were asked how they felt their reasons for joining the study compared to other women they spoke to. Two women noticed a similar desire among other participants to "try to help other women," while another three felt there was a clear connection between family history and participation. One woman in particular spoke expressively on the subject: Well, one big difference was that I had no history in my family of breast cancer. And at some of those very earlier meetings, when we would go around and introduce ourselves it was not at all uncommon to hear a woman say, "Hi, I'm so and so. I've lost a mother, two sisters, and a daughter to breast cancer." And we would go around the room and almost everybody in the room would have a history like that. I mean, I was practically sobbing. And I remember the very first time that I introduced myself and I said, "I really feel like an anomaly because I have no history of breast cancer in my family, although of course I'm terrified for my daughter right now" ... but I was definitely an anomaly. Almost all of the other participants had this daunting family history of breast cancer. The feeling that lack of family history made one an anomaly in the study was not unique to this participant. Another participant did not attend the luncheons 77

PAGE 88

because she mistakenly thought that they were geared towards those who had had breast cancer themselves. Others gave reasons such as lack of time, work commitments or, for those who lived outside Denver, the driving distance for their lack of attendance. Physician Most of the women did discuss the trial with their physician prior to enrolling, as discussed in a previous section. While most were supportive, it was not universal. Not all women based their decision on the attitude of their physician; however the attitude of the physician did make a difference in one woman's ability to follow through with the study requirements. This participant was very well educated and understood the study protocol; however, her physician would often make it difficult for her to obtain the required tests by refusing to order them or hinder her ability to make appointments with the required people. She felt he did not share her concerns about breast cancer and was primarily concerned about keeping costs down. Interestingly, she also felt he did not take her concerns seriously in other areas of her health as well. In contrast, another participant felt study participation was fairly easy because her physician was very cooperative with the study requirements. Organizational Factors Given their role in the study, it was not surprising that the Colorado Cancer Research Program came up in interviews. Nearly two thirds of the women interviewed expressed similar sentiments regarding their interaction with study personnel, viewing them as supportive and efficient, but not overly intrusive, and as making the participation process easier for enrollees. 78

PAGE 89

... I would just say I think that the whole, you know, this organization, and the whole program is run, was run, very efficiently and effectively and with a great deal of heart. And I was always made to feel that this was a very good thing I was doing, and they were very appreciative. I think they're very good people and they, they really like what they do, I think, and that makes a difference. Specifically, women appreciated reminders about appointments, the follow-up process, and the excellent rapport CCRP established with their clients. One in particular, appreciated the time they took to keep her personal physician informed about the trial. Another, who spent a significant period of time out of the country during the study period, felt that they worked very well with her, despite this fact. In fact, the only concern mentioned came from a few participants who experienced a change in the study nurse they were assigned to. One said that she thought she was on her fourth nurse, while another felt that the lack of face-to-face contact she had with the second nurse she was assigned made her less apt to follow through on required study activities. The issue of turnover, however, seemed more of a problem for women who had felt a closer bond to their first nurse than the subsequent one. Another participant had two nurses but did not see this as an issue: I just loved them. I had two different nurses and uh I just liked them both so much. I'll, I'll miss her! You know, cause we saw each other twice a year and I liked that a lot. Nice, nice gal.. .can't say enough about them. In general women appreciated the overall helpfulness and support, which they experienced in working with CCRP. 79

PAGE 90

CHAPTER 7 CONCLUSION The primary aim ofthis study was to explore factors that influence women's decision-making when considering enrollment in a chemoprevention trial for breast cancer, such as the STAR trial. Given the focus in the literature on individual level factors, a secondary aim was to identify any factors in the decision-making process related to social networks or social support. An approach to studying health behaviors that only considers individual level factors ignores the larger environmental context in which these individuals live out their lives. This study demonstrates that while individual beliefs and attributes are important in enrollment decision-making, factors at the interpersonal, institutional/ organization, community level, and social structure, policy and systems levels cannot be ignored if researchers hope to increase enrollment in breast cancer chemoprevention trials. During the decision-making process, the influence of all five levels in the socio-ecological model could be seen, with the individual, interpersonal and community levels the most pronounced. Individual factors included perceived risk of breast cancer, perceived risks/barriers and benefits to study participation, perception of medical research and physician recommendation. Perceived risk ofbreast cancer was influenced not only by objective measures of risk such as the Gail model but was also shaped by participant's previous health history, family history, and prior experiences with breast cancer among others in their social networks. The benefits of study participation outweighed the potential risks or barriers, such as problems with side effects, managing the time commitment, and the ability to follow through with study activities. Potential benefits identified included the medical oversight of participant health, preventing breast cancer and osteoporosis, the knowledge that one 80

PAGE 91

was getting a drug rather than a placebo, and the potential to help others and further medical knowledge. Interpersonal influences on health included prior experiences with breast cancer among friends or family members, physician recommendation, having female offspring or sisters and, to a lesser degree, the support of a spouse or friends. The importance of physician recommendation to women in this study is supported by the literature (Bober et al. 2004; Yeomans Kinney et al. 1995; Yeomans Kinney eta!. 1998a), though it was not universal nor was it a one-way relationship. Some women did not consult their physicians, while others were the ones to bring the study to their physician's attention rather than the reverse. Regardless of who initiated the conversation, most women did discuss the study with their health care providers prior to enrolling. Women with female offspring overwhelmingly considered this to be influential in their decision-making process. Several spoke with their daughters prior to enrollment, while others just felt that participation was something they could do to potentially protect their daughters in the future. A similar sentiment was expressed regarding sisters in that the results of the STAR trial could inform prevention and treatment for their sisters in the future if they were ever at risk or diagnosed with breast cancer. For those women whose sisters had breast cancer, this experience often informed their view of breast cancer as a disease and many were actively involved as support providers in these relationships. Social support occurred in a reciprocal fashion, with women both giving and receiving support to and from others, which is consistent with the literature on gender difference and social support (Kawachi and Berkman 2001). Support from a spouse during the decision-making process appeared to be the most relevant for the participants, when compared to that of family or friends but most women felt that 81

PAGE 92

they had made the decision to enroll on their own. The majority of participants reported positive emotional support from their spouses. Friends also gave emotional and informational support during the process, though again, their influence in the decision-making process appeared to be small. Women themselves reported that their own role in giving support to others with breast cancer was influential in their decision, however. As women tend to invest to a greater degree than men in their social relationships, it is perhaps to be expected that participants in this study discussed social support either given or received not only within the context of close members of their social network, such as family members or friends, but also within their relationships with those farther removed, such as their physician, the study nurses, and other participants (Thoits 1995). Study protocol was an important factor at the institutional/ organizational level, with nearly a third of women saying that the non-placebo design of the study was influential in their decision. Women generally had good understanding of the study drugs and were reassured about side effects. Raloxifene was believed by most to present no more risk than tamoxifen, and the study was seen as comparing two drugs that likely worked, rather than testing a new drug. The majority of study participants were assigned to the tamoxifen arm of STAR. At the community level, several factors overlapped with the influence ofthe interpersonal level. Altruistic motivations that included not only a desire to help family members and friends at risk for the disease, but also women in general were cited along with a goal of furthering medical knowledge. In addition, members of social networks often share similar values, norms, or backgrounds; the participants in this study reinforced that view (Kawachi and Berkman 2001). Few women had experienced negative feelings towards clinical trial enrollment from friends or family members, while none exhibited these feelings themselves. More than one participant 82

PAGE 93

indicated that the educated nature of her family or friends was in part why they were supportive of trial enrollment. Although not all women considered themselves knowledgeable about breast cancer, the interviews revealed a good understanding of risk, prevention and treatment. Participants also were familiar with sources they could access for further information. Participant responses indicate a stronger influence of proximate rather than distal factors in their decision-making. In fact, these distal factors were more apparent in their absence from the interviews rather than their presence. Most were familiar with the media discourse on breast cancer and found it easy to access information on the disease through media sources or the Internet. Other social structure factors may help to understand the relatively homogenous demographic make-up of the population that participated in this study, such as socioeconomic status, cultural or linguistic background, education levels, and the history of clinical trials in the United States. Policy wise, the organization by NCI of a network of CCOPs to bring cancer clinical trials to local communities and health centers has allowed a larger proportion of the population to participate in these trials and broadened the pool of potential participants. Social network and social support theories were useful tools to examine several layers ofthe socio-ecological framework between the individual and the larger social structure, policy and systems levels. In particular, social networks helped to understand how these participants could share values and characteristics at the community level, such as citing altruistic motivations as an essential part of their decision to enroll in STAR. Having had a member of their social network experience breast cancer, such as a friend, family member, or colleague, was also influential. At the interpersonal level and organizational level, different types of social support came into play. Spouses were a source of emotional support in many cases during both the 83

PAGE 94

decision-making and participation processes, and in at least one, a crucial source of instrumental support without which study participation would not have been possible. Interestingly, in some cases, it was the help or resources women had given to others with breast cancer in the past that influenced their decision to participate; acts which were often tied to their altruistic motivations. Participants cited the informational, instrumental, and emotional support they received from their study nurses as an important aspect of their study participation. This may be useful for future clinical trials in chemoprevention to consider in addressing study design. Directions for Future Research The results of this study are most helpful in illuminating factors that impact decision-making on the individual, interpersonal, and community levels. While institutional/organizational factors such as study oversight and the study design of STAR were important, data from this study indicate that this level may be more influential in shaping women's experiences while enrolled in the trial, rather than their decision-making process prior to enrollment. However, further research is needed to confirm these results. An analysis that also accounted for large scale social and policy influences would increase our understanding of which segments of the population hear about clinical trials and perceive trial participation as a possibility given transportation, and economic, linguistic and cultural circumstances. One limitation of this study is its focus on women who chose to enroll in a chemoprevention trial for breast cancer. Because women who opted out of enrollment were not interviewed, comparisons to this population are limited. For example, none of the women in this study viewed risk of side effects as a barrier to trial enrollment. Most reported positive support for their participation from their spouses. Women who choose not to participate in a clinical trial may give different 84

PAGE 95

answers. Futhermore, this group was even more self-selected in that those who participated took a second step in agreeing to be interviewed for the study reported here. A second limitation is the relatively homogenous nature of the population studied. As a group, they were highly educated, middle-class and primarily White. Future qualitative research is needed to identify contextual differences in the decision-making process between participants and non-participants, and to identify factors that are influential for a more diverse group of women. 85

PAGE 96

ACS ADH BCDDP BCPT CCOP CCRP CORE DCIS FDA HRT IBIS LCIS MORE NBCCEDP NCI NSABP SERM STAR APPENDIX A ACRONYMS American Cancer Society Atypical Ductal Hyperplasia Breast Cancer Detection Demonstration Project Breast Cancer Prevention Trial Community Clinical Oncology Program Colorado Cancer Research Program Continuing Outcomes Relevant to Evista Ductal carcinoma in situ Food and Drug Administration Hormone Replacement Therapy International Breast Cancer Intervention Study Lobular carcinoma in situ Multiple Outcomes Raloxifene Evaluation National Breast and Cervical Cancer Early Detection Program National Cancer Institute National Surgical Adjuvant Breast and Bowel Project Selective Estrogen Receptor Moderators Study of Raloxifene and Tamoxifen 86

PAGE 97

APPENDIXB KEY CONCEPTS AND DEFINITIONS OF THE HEALTH BELIEF MODEL AND THE THEORY OF REASONED ACTION Table B.l. The Health Belief Model. Concept Definition Perceived susceptibility_ One's opinion of chances of getting a condition. Perceived severity One's opinion of how serious a condition and its sequelae are. Perceived benefits One's opinion of the efficacy of the advised action to reduce risk of seriousness of impact. Perceived barriers One's opinion of the tangible and psychological costs of the advised action. Cues to action Strategies to activate one's "readiness." Self-efficacy One's confidence in one's ability to take action. (Adapted from Strecher and Rosenstock 1997:45) T bl 82Th Th a e e eory o fR easone dA f c 1on an d th Th e eory o fPl anne dB h e avwr. Conce_pt Definition Behavioral intention Perceived likelihood of performing the behavior Attitude Behavioral belief Belief that behavioral performance is associated with certain attributes or outcomes Evaluation J Value attached to a behavioral outcome or attribute Subjective norm Normative belief Belief about whether each referent approves or disapproves of the behavior Motivation to comply Motivation to do what each referent thinks Perceived behavioral control Control belief Perceived likelihood of occurrence of each facilitating or constraining condition Perceived power Perceived effect of each condition in making behavioral performance difficult or easy (Adapted from Montano et a!. 1997:91) 87

PAGE 98

APPENDIXC DEMOGRAPHIC QUESTIONNAIRE Subject ID # ______ Age _____ Marital Status (circle one): Single Married Separated Divorced Widowed Domestic Partner Ethnicity: ___________ Education Level: ___ Some High School or below High School Diploma ---___ Some College ---Associate's Degree ___ Bachelor's Degree ---Graduate Degree ___ Post-graduate Degree Occupation: ___________________ Family History of Breast Cancer? YES NO If"YES," briefly explain: Number of Female Children: ------Number of Sisters: ------88

PAGE 99

APPENDIXD INTEVIEW QUESTION GUIDE What motivated you to consider participating in the STAR trial? Tell me about your experiences in the STAR trial. How do you think your reasons for participating in the STAR trial compare to those of other women? What reservations did you have regarding participation in the STAR trial? Once you were identified as high risk for breast cancer, with whom did you discuss your status? o With whom in your family did you discuss these issues with? o Why did you choose those particular people? o What about friends? Who did you tum to outside of your family and why? How knowledgeable did you consider yourself about breast cancer before entering the trial? o How about chemoprevention? What steps did you undertake to educate yourself or find out more about breast cancer after being identified as high risk for the disease? o How about breast cancer prevention? Do you have a family history of breast cancer? Who in your family had the disease? o How did this family history play a role in your decision to enroll in a chemoprevention trial? How did having daughters influence your decision to enroll in the STAR trial? How did having sisters influence your decision to enroll in the STAR trial? Have you ever had a close friend who was diagnosed with breast cancer? o What happened in her case? o Was she diagnosed before or after you were identified as high risk? o How did this influence your decision to enroll in the STAR trial? How did your spouse/partner react when you told him/her you were considering enrolling in the STAR trial? o How did this influence your decision? What were your friends' thoughts and reaction to your decision to enroll? Who made up your social network while you were taking part in the STAR trial? 89

PAGE 100

BIBLIOGRAPHY Advani, Anjali S., Benjamin Atkeson, Carrie L. Brown, Bercedis L. Peterson, Laura Fish, Jeffrey L. Johnson, John P. Gockerman, Marc Gautier. 2003 Barriers to the Participation of African-American Patients with Cancer in Clinical Trials. Cancer 97( 6): 1499-1506. Altschuler, Andrea and Carol P. Somkin. 2005 Women's Decision Making About Whether or Not to Use Breast Cancer Chemoprevention. Journal ofWomen & Health 41(2):81-95. American Cancer Society. 2005 Breast Cancer Facts & Figures 2005-2006. Atlanta: American Cancer Society, Inc. American Cancer Society. 2008a Cancer Facts and Figures 2008. Electronic document, http://www.cancer.org/downloads/STT/2008CAFFfinalsecured.pdf, accessed May 19,2009. 2008b Non-Cancerous Breast Conditions. Electronic document, http:/ /www.cancer.org/docroot/CRVcontent/CRI 2 6X _Non_ Cancerous_ Breas t_ Conditions_59.asp, accessed May 19, 2009. American Cancer Society. 2009 What is Breast Cancer? Electronic document, http://www .cancer.orgidocroot/CRVcontent/CRI 2 4 _I X_ What_ is_ breast_ can cer_5.asp?rnav=cri, accessed May 19, 2009. Armstrong, Katrina, Ellyn Micco, Amy Carney, Jill Stopfer, and Mary Putt. 2005 Racial Differences in the Use of BRCA 112 Testing Among Women With a Family History of Breast or Ovarian Cancer. Journal of the American Medical Association 293(14):1729-1736. Bastian, Lori A., Isaac M. Lipkus, Maggie N. Kuchibhatla, Haoling Holly Weng, Susan Halabi, Paula D. Ryan, Celette Sugg Skinner, and Barbara K. Rimer. 2001 Women's Interest in Chemoprevention for Breast Cancer. Archives of Internal Medicine 161:1639-1644. 90

PAGE 101

Becker, Gay. 1994 Metaphors in Disrupted Lives: Infertility and Cultural Constructions of Continuity. Medical Anthropology Quarterly 8(4):383-410. Bloom, Joan R., Susan L. Stewart, Monica Johnston, Priscilla Banks, and Patricia Fob air. 2001 Sources of Support and the Physical and Mental Well-being ofYoung Women with Breast Cancer. Social Science & Medicine 53(11):1513-1524. Bober, Sharon L., Lizbeth A. Hoke, Rosemary B. Duda, Meredith M. Regan, and Nadine M. Tung. 2004 Decision-Making About Tamoxifen in Women at High Risk for Breast Cancer: Clinical and Psychological Factors. Journal of Clinical Oncology 22(24):4951-4957. Bourdieu, Pierre. 1980 The Logic of Practice. Stanford, California: Stanford University Press. Brown, Powel H. and Scott M. Lippman. 2000 Chemoprevention ofbreast cancer. Breast Cancer Research and Treatment 62:1-17. Brown, William Michael, Nathan S. Consedine, and Carol Magai. 2005 Altruism Relates to Health in an Ethnically Diverse Sample of Older Women. Journal of Gerontology 60B(3):P143-P152. Charles, Cathy, Amiram Gafni, and Tim Whelan. 1999 Decision-making in the physician-patient encounter: revising the shared treatment decision-making model. Social Science & Medicine 49:651-661. Clarke, Adele E., Laura Mama, Jennifer R. Fishman, Janet K. Shim, and Jennifer Ruth Fosket. 2003 Biomedicalization: Technoscientific Transformations of Health, Illness, and U.S. Biomedicine. American Sociological Review 68:161-194. 91

PAGE 102

Colditz, Graham. 1997 Hormone Replacement Therapy Increases the Risk of Breast Cancer. In Cancer: Genetics and the Environment, vol. 833: Annals ofthe New York Academy of Sciences. H. Leon Bradlow, Jack Fishman, and Michael P. Osborne, eds. New York: New York Academy of Sciences. Coleman, James S. 1958 Relational Analysis: The Study of Social Organizations with Survey Methods. Human Organization 17(4):28-36. Colorado Cancer Research Program. 2009a About Us. Electronic Document, http://www.co cancerresearch.org/about/about.htm, accessed May 19, 2009. 2009b Clinical Trials. Electronic Document, http://www.co cancerresearch.org/clinical/index.htm, accessed May 19, 2009. 2009c Trial Participation. Electronic Document, http://www.co cancerresearch.org/participate/index.htm, accessed May 19, 2009. Constantino, Joseph P., Mitchell H. Gail, David Pee, Stewart Anderson, Carol K. Redmond, Jacques Benichou, and H. Samuel Wieand. 1999 Validating Studies for Models Projecting the Risk of Invasive and Total Breast Cancer Incidience. Journal of the National Cancer Institute 91(19):1541-1548. Cook, Thomas D. 2003 The case for studying multiple contexts simultaneously. Addiction 98(Suppl 1):151-155. Cyrus-David, M.S. and S.S. Strom. 2001 Chemoprevention ofbreast cancer with selective estrogen receptor modulators: Views from broadly diverse focus groups of women with elevated risk for breast cancer. Psycho-Oncology 10:521-533. DeSantis, Carol, Ahmedin Jemal, Elizabeth Ward and Michael J. Thun. 2008 Temporal trends in breast cancer mortality by state and race. Cancer Causes and Control 19(5):537-545. Durkheim, Emile. 1951 Suicide: A Study in Sociology. New York: The Free Press. 92

PAGE 103

Elder, John P., Leslie Lytle, James F. Sallis, Deborah Rohm Young, Allan Steckler, Denise Simons-Morton, Elain Stone, Jared B. lobe, June Stevens, Tim Lohman, Larry Webber, Russell Pate, Brit I. Saksvig, and Kurt Ribisi. 2006 A description of the social-ecological framework used in the trial of activity for adolescent girls (TAAG). Health Education Research 22(2): 155-165. Fisher, Edwin B., Carol A. Brownson, Mary L. O'Toole, Gowri Shetty, Victoria V. Arwuri, and Russell E. Glasgow. 2005 Ecological Approaches to Self-Management: The Case of Diabetes. American Journal ofPublic Health 95(9):1523-1535. Food and Drug Administration. 2005 Tamoxifen Information. Electronic Document, http://www.fda.gov/cder/news/tamoxifen, accessed May 19, 2009. Friedman, Samuel R. and Sevgi Aral. 2001 Social Networks, Risk-Potential Networks, Health, and Disease. Journal of Urban Health: Bulletin ofthe New York Academy ofMedicine 78(3):411-418. Gail, Mitchell H., Joseph P. Constantino, John Bryant, Robert Croyle, Laurence Freedman, Kathy Helzlsouer, and Victor Vogel. 1999 Weighing the Risks and Benefits ofTamoxifen Treatment for Preventing Breast Cancer. Journal of the National Cancer Institute 91 (21 ): 1829-1846. Gelo, Omar, Diana Braakmann, and Gerhard Benetka. 2008 Quantitative and Qualitative Research: Beyond the Debate. Integrative Pscyhological and Behavioral Science 42(3):266-290. Greenwald, Peter. 1995 Preventive Clinical Trials. In Cancer Prevention, vol. 768: Annals of the New York Academy of Sciences. H. Leon Bradlow, Michael P. Osborne, and Umberto Veronesi, eds. New York: New York Academy of Sciences. 93

PAGE 104

Gregson, Jennifer, Susan B. Foerster, Robin Orr, Larry Jones, Jamie Benedict, Bobbi Clarke, James Hersey, Jan Lewis, and Karen Zotz. 2001 System, Environmental, and Policy Changes: Using the Social-Ecological Model as a Framework for Evaluation Nutrition Education and Social Marketing Programs with Low-Income Audiences. Journal of Nutrition Education 22(S I ):S4-S 15. Halpern, Scott D., Jason H.T. Karlawish, David Casarett, Jesse A. Berlin, Raymound R. Townsend, and David A. Asch. 2003 Hypertensive patients' willingness to participate in placebo-controlled trials: Implications for recruitment efficiency. American Heart Journal 146(6):985992. Heaney, Catherine A. and Barbara A. Israel. 2002 Social Networks and Social Support. In Health Behavior and Health Education. 3rd edition. Karen Glanz, Barbara K. Rimer, and Frances Marcus Lewis, eds. Pp. 185-209. San Francisco: Jossey-Bass. J anesick, V .J. 1994 The Dance of Qualitative Research Design. In Handbook ofQualitative Design. N.K. Denzin and Y.S. Lincoln, eds. Thousand Oaks, CA: Sage Publications. Jordan, V. Craig. 1995 An Overview of Considerations for the Testing of Tamoxifen as a Preventive for Breast Cancer. vol. 768: Annals of the New York Academy of Sciences. H. Leon Bradlow, Michael P. Osborne, and Umberto Veronesi, eds. New York: New York Academy of Sciences. Kapp, Julie M., A. Blythe Ryerson, Steven S. Coughlin, and Trevor D. Thompson. 2008 Racial and etlmic differences in mammography use among U.S. women younger than age 40. Breast Cancer Research and Treatment 113(2): 327-337. Kawachi, Ichiro and Lisa F. Berkman. 2001 Social Ties and Mental Health. Journal ofUrban Health: Bulletin of the New York Academy of Medicine 78(3):458-467. Kelsey, Jennifer L. and Gertrud S. Berkowitz. 1988 Breast Cancer Epidemiology. Cancer Research 48:5615-5623. 94

PAGE 105

Key, Timothy J., Pia K. Verkasalo, and Emily Banks. 2001 Epidemiology ofbreast cancer. The Lancet Oncology 2:133-140. King, Samantha. 2006 Pink Ribbons, Inc. Minneapolis: University of Minnesota Press. Kreuter, Matthew W., Shobhina G. Chheda, and Fiona C. Bull. 2000 How Does Physician Advice Infleunce Patient Behavior? Archives of Family Medcine 9(5):426-433. Lantz, Paula M. and Karen M. Booth. 1998 The Social Construction of the Breast Cancer Epidemic. Social Science and Medicine 46(7):907-918. Lock, Margaret, Julia Freeman, Gillian Chilibeck, Briony Beveridge, and Miriam Padolsky. 2007 Susceptibility Genes and the Question of Embodied Identity. Medical Anthropology Quarterly 21 (3 ):256-276. Love, Susan. 1995 Dr. Susan Love's Breast Book. 2nd edition. Reading, Massachusettes: Addison-Wesley. Lovegrove, E., N. Rumsey, D. Harcourt, and S.J. Cawthorn. 2000 Factors Implicated in the Decision Whether or Not to Join the Tamoxifen Trial in Women at High Familial Risk ofBreast Cancer. Psycho-Oncology 9:193-202. Marcus, George E. and Michael M. J. Fischer. 1986 Anthropology as Cultural Critique. Chicago: The University of Chicago Press. McLeroy, Kelllleth R., Daniel Bibeau, Allan Steckler, and Karen Glanz. 1988 An ecological perspective on health promotion programs. Health Education Quarterly 15:351-377. Miles, Matthew B., A. Michael Huberman. 1994 Qualitative Data Analysis. 2nd edition. Thousand Oaks: Sage Publications. 95

PAGE 106

Montano, Daniel E. and Danuta Kasprzyk. 2002 The Theory of Reasoned Action and the Theory of Planned Behavior. In Health Behavior and Health Education. 3rd edition. Karen Glanz, Barbara K. Rimer, and Frances Marcus Lewis, eds. Pp. 67-98. San Francisco: Jossey Bass. Mulley, Albert G. and Karen Sepucha. 2002 Making Good Decisions about Breast Cancer Chemoprevention. Annals of Internal Medicine 13 7( I ):52-54. National Cancer Institute. Community Clinical Oncology Programs. Electronic document, http://prevention.cancer.gov/programs-resources/programs/ccop, accessed May 19, 2009. National Cancer Institute. 2005 For More Than 20 Years, CCOPs Define Commitment, Success. NCI Cancer Bulletin 2(35): 1-2. National Cancer Institute. 2006 Results ofthe Study ofTamoxifen and Raloxifene (STAR) Released: Osteoporosis Drug Ra1oxifene Shown to be as Effective as Tamoxifen in Preventing Invasive Breast Cancer. Electronic document, http://www.cancer.gov/newscenter/pressre1eases/ST ARresultsAprl72006, accessed May 19, 2009. National Cancer Institute. 2008 Breast Cancer Prevention Studies. Electronic document, http://www. cancer. gov I cancertopi cs/ factsheet/Preventionlbreast -cancer, accessed May 19, 2009. Newman, Lisa A., Kathryn Carotin, Michael Simon, Mary Kosir, William Hymiuk, Ray Demers, Ann Grossbert Schwartz, Daniel Visscher, William Peters, and David Bouwman. 2001 Impact of Breast Carcinoma on African-American Women. Cancer 91(9):1834-1843. Neuman, W. Lawrence. 2003 Social Research Methods. 51h edition. Boston: Pearson Education, Inc. 96

PAGE 107

Nguyen, Tung T., Carol P. Somkin, and Yifei Ma. 2005 Participation of Asian-American Women in Cancer Chemoprevention Research. Cancer Supplement I 04(12):3006-3014. Patton, Michael Quinn. 1990 Qualitative Evaluation and Research Methods. 2nd edition. Newbury Park: Sage Publications. Quill, Timothy and Howard Brody. 1996 Physician Recommendations and Patient Autonomy: Finding a Balance between Physician Power and Patient Choice. Annals of Internal Medicine 125(9):763-769. Rapp, Rayna. 2000 Testing Women, Testing the Fetus. New York: Routledge Roberts, Cleora S., Charles E. Cox, Vicki J. Shannon, and Nancy L. Wells. 1994 A Closer Look at Social Support as a Moderator of Stress in Breast Cancer. Health & Social Work 19(3 ): 157-164. Rockhill, Beverly. 2001 The Privatization of Risk. American Journal of Public Health 91(3):365-368. Rose, Geoffrey. 1985 Sick Individuals and Sick Populations. International Journal of Epidemiology 14(1):32-38. Sallis, James F. and Neville Owen. 2002 Ecological Models of Health Behavior. In Health Behavior and Health Education. 3rd edition. Karen Glanz, Barbara K. Rimer, and Frances Marcus Lewis, eds. Pp. 462-484. San Francisco: Jossey-Bass. Sampson, Robert J., Jeffrey D. Morenoff, and Thomas Garuwn-Rowley. 2002 Assessing "Neighborhood Effects": Social Processes and New Directions in Research. Annual Review of Sociology 28:443-78. 97

PAGE 108

Shavers, Vickie L. and Martin L. Brown 2002 Racial and Ethnic Disparities in the Receipt of Cancer Treatment. Journal of the National Cancer Institute 94(5):334-357. Siminoff, Laura A. and Robert Arnold. 1999 Increasing Organ Donation in the African-American Community: Altruism in the Face of an Untrustworthy System. Annals of Internal Medicine 130(7):607 -609. Simon, Christian, Michelle Eder, Eric Kodish, and Laura Siminoff. 2006 Altruistic Discourse in the Informed Consent Process for Childhood Cancer Clinical Trials. The American Journal ofBioethics 6(5):40-47. Strecher, Victor J. and Irwin M. Rosenstock. 1997 The Health Belief Model. In Health Behavior and Health Education. 2nd edition. Karen Glanz, Frances Marcus Lewis, and Barbara K. Rimer, eds. Pp. 41-59. San Francisco: Jossey-Bass. Stokols, Daniel. 1992 Establishing and Maintaining Health Environments. American Psychologist 47(1 ):6-22. Thoits, Peggy A. 1995 Stress, Coping, and Social Support Processes: Where Are We? What Next? Journal ofHealth and Social Behavior 35:53-79. Thomas, Stephen B. and Sandra Crouse Quinn. 1991 The Tuskegee Syphilis Study, 1932 to 1972: Implications for HIV Education and AIDS Risk Education Programs in the Black Community. The American Journal ofPublic Health 81(11):1498-1505. Vogel, Victor G., Joseph P. Constantino, D. Lawrence Wickerham, Walter M. Cronin, Reena S. Cecchini, James N. Atkins, Therese B. Bevers, Louis Fehrenbacher, Eduardo R. Pajon, James L. Wade III, Andre Robidoux, Richard G. Margolese, Joan James, Scott M. Lippman, Carolyn D. Runowicz, Patricia A. Ganz, Steven E. Reis, Worta McCaskill-Stevens, Leslie G. Ford, V. Craig Jordan, and Norman Wolmark. 2006 Effects of Tamoxifen vs. Raloxi fene on the Risk of Developing Invasive Breast Cancer and Other Disease Outcomes. Journal ofthe American Medical Association 295(23):2727-2926. 98

PAGE 109

Wen, Ming, Heather Van Duker, Lenora M. Olson. 2009 Social contexts of regular smoking in adolescence: Towards a multidimensional ecological model. Journal of Adolescence 32(2009):671692. Wilcox, Pamela. 2003 An ecological approach to understanding youth smoking trajectories: problems and prospects. Addiction 98(Suppl 1):57-77. Yeomans-Kinney, Anita, Sally W. Vernon, Ralph F. Frankowski, Diane M. Weber, Joann M. Bitsura, and Victor G. Vogel. 1995 Factors Related to Enrollment in the Breast Cancer Prevention Trial at a Comprehensive Cancer Center during the First Year of Recruitment. Cancer 76(1 ):46-56. Yeomans Kinney, Anita, Chesley Richards, Sally W. Vernon, and Victor G. Vogel. 1998a The effect of physician recommendation on enrollment in the breast cancer chemoprevention trial. Preventive Medicine 27:713-719. 99

PAGE 110

Yeomans Kinney, Anita, Sally W. Vernon, Wen Shui, Diane V. Weber, Michael Schell, and Victor G. Vogel. 1998b Validation of a Model Predicting Enrollment Status in a Chemprevention Trial for Breast Cancer. Cancer Epidemiology, Biomarkers & Prevention 7:591-595. 100