Group 2: Homework 10
Matt Conaway, Epid 655 Group 2, Assign. #10
I. ABSTRACT
As a leading cause of death among women aged 40-75, breast
cancer is a significant public health problem and will continue
to increase in importance if preventive measures are not
implemented. Presently, it is unknown what factors lead a woman
to get screened for breast cancer . It is also not known which
screening practices(breast self-exam, clinical exam, and mammography),
if any, are most effective at detecting breast cancer at an earlier stage.
Thus, the goal of this study is to determine the efficacy of different
combinations of screening practice in detecting breast cancer at an
earlier stage, and to identify factors that influence women to practice
breast cancer screening. Using a case-control design, this study will
compare cases of Stage III breast cancer at The University of Michigan
Breast Care Center with controls of Stage II or less breast cancer taken
from the Breast Care Center and The University of Michigan General
Care Clinic. It is hoped that the study will lead to appropriate
recommendations for screening practices which could be implemented
as policy, which could lead to reduced healthcare and other socioeconomic
costs. Further studies into women's screening practices and mortality
trends of breast cancer could be done to assess the effectiveness of future
screening programs.
II. SPECIFIC AIMS
1.
The overall goal is to study the preventative efficacy of breast cancer
screening (breast self-exam, clinical exam, and mammography)
practices as well as factors influencing women's behavior to practice
breast cancer screening.
2.
This study is important because the effectiveness of screening practices
for breast cancer in order to reduce late stage diagnosis needs to be
assessed so that the appropriate combinations of screening practices
can be recommended as public health policy. It is equally as important
to determine what factors influence breast cancer screening practices in
order to increase the frequency of screening which may lead to a reduction
of late stage diagnosis. Since breast cancer is a primary cause of death of
women aged 40-75, early detection might help reduce the morbidity,
mortality and socioeconomic costs which are associated with this disease.
3.
The specific aims of the study are:
a. To compare level of previous breast cancer screening practices in
relation to stage of breast cancer at diagnosis in women aged 40-75.
Null Hypothesis: There is no difference of breast cancer stage at
diagnosis between women who have had previous breast cancer screening and
women who have not had previous breast cancer screening in women aged
40-75.
b. To identify factors which influence breast screening practices in women
aged 40-75.
Null Hypotheses:
1. There is no difference in screening practices between women less than
50 years old and greater than 50 years old.
2. There is no difference in screening practices among women with varying
levels of education.
3. There is no difference in screening practices among women with
insurance or without insurance.
4. There is no difference in screening practices among women having a
first degree relative with breast cancer and those not having a first
degree relative with breast cancer.
5. There is no difference in screening practices among women seeing a
health care provider annually and those not seeing a health care provider
annually.
6. There is no difference in screening practices among women of different
racial and ethnic groups.
7. There is no relationship between screening practices among women with
varying levels of household incomes.
III. BACKGROUND AND SIGNIFICANCE
Breast cancer is the second most common cancer among women in the world
and the most common cancer among women in developed countries. By the
year 2000, it is estimated that breast cancer will account for 500,000 deaths
annually, therefore it is important to reduce breast cancer mortality. Early
detection of breast cancer through appropriate screening practices could help
control costs of medical and surgical treatment. The findings of our study may
lead to changes in policy by providers, insurance companies, government, and
women themselves.
Currently, it is unknown what factors are most likely to influence a woman to
be screened for breast cancer, what screening practices, if any, are most effective
at detecting the possibility of breast cancer, and if screening contributes to a
reduction of mortality from breast cancer by detecting the problem early enough.
It is hoped, from this study, that the findings will lead to appropriate policy to
prevent breast cancer in women aged 40-75. Screening recommendations could
be followed by women and their health care providers in order to detect breast
cancer at the earliest possible time. Early detection could hopefully lead to earlier,
less costly and less traumatic treatment. Earlier treatment could lead to early cure
and, in turn, reduced morbidity from prolonged treatment as well as reduced mortality
from late stage breast cancer. Furthermore, decreased use of expensive and traumatic
procedures could lead to a reduction of health insurance premiums, long-term care
costs, lost employment due to illness, and other economic burdens on women, their
families, and society in general. Further research will be needed to monitor effectiveness
of subsequent screening programs and any changes in screening practices and patterns
of breast cancer mortality among women aged 40-75.
IV. RESEARCH DESIGN AND METHODS
A. Study population selection procedures
The study is a case-control design.
Cases: Women, aged 40-75, who have stage III or IV breast cancer and/or
are likely to die from breast cancer.
Sources: 1. U of M Breast Care Center (BCC)
2. Possible convenience sample from area oncology
practices (if not enough cases found from U of M BCC.
Controls: Women without breast cancer, or who have breast cancer in stage
I or II.
Sources: 1. U of M BCC
2. A U of M out patient clinic, not associated with
cancer, that has a patient population similar to that of
the BCC, and/or has a similar database.
B. Sample size procedures
The public domain software known as SAMPSIZE will be used to
calculate sample size and power for this study.
C. Subject recruitment procedures: U of M BCC will be asked to list all of the women
in their database who have stage III of IV breast cancer, or who have died
from breast cancer, and list separately those women without breast cancer
or survivors of breast cancer. A unique study number will be assigned to
each individual subject. Only those from the BCC will be able to associate
the unique name of any subject with any data collected to protect
confidentiality.
D. Data Collection: As much demographic and clinical data will be collected
from the BCC database as possible.
Questionnaires will be sent via mail to all study subjects to
acquire information not available in the database. The questionnaires
will be identified by the unique study number so that information
collected from the database can be associated with the appropriate
questionnaire responses. In addition to sending a pen with the first mailing
of the questionnaire, a second mailing will be done to increase the response
rate among the study participants.
The procedure will be similar for collecting information from a
similar U of M clinic, although probably not as much useful data will be
available from their database, if one exists. If a convenience sample must
be used for cases, a similar method for mailing the questionnaires will be
used and for assigning the unique study number to each subject.
Preliminary or more detailed clinical data could be acquired through chart
review by study investigators if necessary.
E. Human subject protection procedures
The cover letter sent with the questionnaire will assure the participant
that the study is voluntary. All respondents will be coded by a number taken
from the questionnaire. Only the U-M BCC staff will know the identity of each
respondent.
F. Variables List
Question a:
outcome variable: stage of breast cancer at diagnosis
predictor variable: level of screening (defined as number and years of
previous screenings)
potential confounders: age, age at menopause, age at first pregnancy,
family history, previous breast disease, race, household income,
interpretation of mammogram by radiologist, variation in equipment, oral
contraceptive use, hormone use, body mass index, place of residence.
Question b:
outcome variable: level of screening
predictor: age, family history, previous breast disease, race, household
income, level of education, payment method.
potential confounders: accessibility to health services, experience with
medical services, ethnicity/race, household income level, age at
menopause, age at first pregnancy. (All predictor variables are potential
confounders also).
G. Exposure questions
The exposure is this study is defined as breast cancer screening.
Various factors such as age, education, access to medical care, and family
history of breast cancer will be evaluated. The goal is to determine which
factor(s) lead a woman to be screened for breast cancer.
H. Outcome determination
The outcome will be defined as stage of breast cancer at diagnosis.
Since screening is a preventive measure, the more positive outcome is an
earlier stage of breast cancer at diagnosis. Thus, if the cancer is detected early
enough, mortality and morbidity could be significantly reduced.
I. Budget
Budget for Epid 655 group 2
ITEM COST
Supplies:
Disks, binder $12.83
Envelopes (100@ 7.63) $70.63
BRE ($50/500) $100.00
Letterhead ($23/500) $69.00
Stamper /date stamp/envelope sealer bottles $8.21
Incentive Pens (not yet purchased) 2.14/dozen = $107
Labels ??
Subtotal $260.67
Copies:
Draft Questionnaire (to test in class) 60 copies x 5pgs = $11.70
Final Questionnaire 1000 x 5pgs = $195
Subtotal $206.70
Postage:
Postage $0.55 x 960 = $528
Return Questionnaires $0.43/response (estimate 70% response) = $180.60
Subtotal $708.60
Final Report $10.00
Subtotal $1185.97/2 (shared cost with group 1)
TOTAL $597.99
J. Time line
The study will be executed in sequential steps according to the following time frame:
February 21, 1997 Test questionnaire
February 26, 1997 Questionnaire is mailed
March 13, 1997 Second mailing
March 30, 1997 Begin coding
April 10, 1997 Begin data analysis and report writing
April 18, 1997 Give final report
Nancy Lynn Bischof /group 2 assign10
ABSTRACT: The goal of this study is to determine the efficacy of three types of breast cancer screening procedures, (breast self exam, clinical breast exam, mammography) , and the factors that influence women to practice these procedures.
The importance of this study includes the
need to determine the most effective breast cancer screening procedures in
order to reduce late stage diagnosis of breast cancer, which will lead to
an overall higher survival rate, and the need to determine which factors
influence women to practice breast cancer screening procedures in order to
better educate women about breast cancer surveillance.
The specific aims of this study are to
compare previous breast cancer screening practices between women age 40-75
who have been diagnosed with stage 0-II breast cancer, and women age 40-75
who have been diagnosed with stage III breast cancer. Also the study will
compare breast cancer screening practices between women age 40-75 who have
and have not been diagnosed with breast cancer.
The study design includes 150 women
diagnosed with stage III breast cancer (cases) and 150 women diagnosed
with stage 0-II breast cancer (controls) from the University of Michigan
Breast Care Center. An additional 150 women without any history of cancer
will be chosen as controls from the University of Michigan General Care
Clinic. All women were age 40-75 when seen at the UM Clinics.
The significance of the study involves the
epidemic of breast cancer
( the most common cancer among women in developing countries), and the
need for data on breast cancer screening practices to increase early stage
dignosis of breast cancer and thus survival rates.
The long term objectives of this study fall into three components.
The first is to determine the efficacy of three commonly used breast
cancer screening procedures, breast self exam, clinical breast exam, and
mammography. The second component of the long term objectives is to
determine what factors influence women to practice each of the three
previously mentioned breast cancer screening procedures. Identification
of important barriers that are separating women from participating fully
in their own health care, such as ethnicity, socioeconomic status,
education level, etc. will enable health educators and health care
providers to better serve women, all of whom are at risk for breast
cancer.
The third component of the studys long term objectives is simply a
reduction in the number of late stage breast cancer diagnoses, which in
turn should lead to a higher survival rate for breast cancer patients.
Realistically, this study is just a beginning. It is hoped that our
initial study will lead to new insights into breast cancer screening
practices, and be a stepping stone to improve womens health.
Hypotheses to be tested include:
a) There is no difference in breast cancer stage at diagnosis between
women who have practiced regular breast cancer screening procedures and
women who have not practiced regular breast cancer screening procedures.
b) There is no difference in breast cancer screening practices between
women age 50 and over and women age 40-49.
c) There is no difference in breast cancer screening practices between
women of different education levels.
d) There is no difference in breast cancer screening practices between
women with health insurance, and women without health insurance.
e) There is no difference in breast cancer screening practices between
women having a first degree relative with breast cancer, and those not
having a first degree relative with breast cancer.
f) There is no difference in breast cancer screening practices between
women seeing a health care provider annually and women not seeing a health
care provider annually.
g) There is no difference in breast cancer screening practices between
women of different race/ethnicity.
h) There is no difference in breast cancer screening pracitices between
women with different household incomes.
The most important gap in research knowledge is in the area of breast
self exam. Data is particularly lacking in the literature, and health
education is hampered by this lack of data. Seven out of ten women with
breast cancer found their own cancer. Breast Self Exams Save Lives. is a
much more compelling reason to practice self exam than being told to just
do it.
The most likely path of improved womens health from this study lies
in its ability to provide new insights into how often and who particpates
in breast cancer screening. Follow up research into breast cancer
screening procedures will be necessary to provide the why.
Study subjects will be identified from the University of Michigan
Breast Care Center (150 cases, 150 controls) and the University of
Michigan General Care Clinic (150 controls). All cases and controls will
be women, presently living, residents of Michigan, who were between ages
40 and 75 when seen at the UM Clinics. All cases of stage III breast
cancer that were diagnosed between January 1, 1992 and December 31, 1996
will be studied. Controls will be chosen by simple random sampling from
the women diagnosed with stage 0-II breast cancer (Breast Care Center)
diagnosed between January 1, 1992 and December 31, 1996, and women without
any type of cancer or benign breast disease (General Care Clinic) who were
seen at UM between January 1, 1992 and December 31, 1996.
Due to data base constraints, the number of cases expected is
approximately 150. The data base is a constraint due to both lack of time
and lack of funds. A 1:2 ratio of cases to controls was determined to be
the best available given budget and time constraints. This results in a
total of 450 subjects for the study. Exposure rate in controls of .56
(Michigan BRFS 1994) and in cases of .40 (Moseson) were used with alpha
=.05 and power =.80 to calculate odds ratios. Given the sample size of
450, the study needs to detect an odds ratio of .5 in order to have an 80%
chance of finding a difference in screening behaviours between cases and
controls.
Subjects will be recruited via a mailed questionnaire with a cover
letter from the UM Breast Care Center. A second questionnaire will be
mailed to non-respondents two weeks after the first. Recruitment of
subjects not responding to either questionnaire will be terminated.
Data will be collected from the UM data bank based on a questionnaire
compiled by the UM Breast Care Center, and from an additional
questionnaire mailed to each study subject. To insure voluntary
participation in the study, an informed consent form will accompany every
questionnaire. The identification of study subjects will be confidential.
The data base manager will code each subject with a number. Only the
number, not the name of the study subject, will be available to
investigators.
Variables on which data will be collected include: age, gender,
household income, race/ethnicity, menstrual history, health insurance
status, history of first degree relatives with breast cancer, history of
breast self exams, history of clinical breast exams, mammographic history.
The principle exposure of interest is history of breast self exam.
Questions include:
a) Do you know how to do a breast self exam? Yes No
b) How often do you practice breat self exam? every month every other
month twice a year never
c) At what age did you begin doing breast self exams?
________________________
Exposure status will be catagorized on the basis of below age 40, age
40-49, age 50 and over.
Outcome status will be determined by the diagnosis reported in the UM data
bank by the attending physician.
BUDGET:
Copies 960 4 pages @ $.03 each $115.20
Envelopes 960 @ $56.00 (2 boxes)
Postage 960 @ $.55 $528.00
Pens, UM 600 $108.00
Business Reply Envelopes 960 @ $96.00 (2 boxes)
Postage BRE 450 @ $.43 $193.50
Total $1,096.70
TIMELINE:
2-28-97 Mail first questionnaire.
3-12-97 Finalize codebook.
3-14-97 Mail second questionnaire.
3-31-97 Assemble preliminary data tables.
4-16-97 Finalize data tables and data analysis.
4-18-97 Presentation.
Angela Christine Zanotti / group2hw10
I.
The overall goal of this study is to determine the preventative
efficacy of breast cancer screening, which includes breast self-exams,
clinical exams, and mammography. This study will also examine factors
that influence a womens behavior to practice breast cancer screening. The
significance of this study is to determine the efficacy of screening
practices for breast cancer in order to reduce late stage diagnosis.
Another relevant issue will be to determine what specific factors
influence breast cancer screening practices. With the objective of
increasing the frequency of screening by evaluating womens previous
screening practices in terms of number and years of prior screenings, this
study hopes to reduce the number of late stage diagnosis. The specific
aim of this study is to compare number and years of previous breast cancer
screening practices in relation to stage of breast cancer at diagnosis in
women aged 40-75. The study will be designed as an unmatched case
control. The cases consist of women, aged 40-75, who have stage III
breast cancer. The source of these cases will be from the University of
Michigan Breast Cancer Center. Controls will be chosen form the BCC as
well as the general care unit at the UM medical center. Controls will be
subjects with stage 0, I, or II breast cancer. The information will be
gathered by a mailed questionnaire as well as the UM data bases. By
successful completion of this study our team will provide information that
may lead to earlier detection of breast cancer which is the most common
malignancy in women.
II.
The main long-term objective is to compare the frequency of
screening practices among women with stage III breast cancer and their
screening procedure prior to diagnosis with women who have stage 0,I or II
stage breast cancer and their previous screening practices. The screening
and stage of breast cancer at diagnosis through biopsy by a trained
medical personnel will be measured in terms of number of screenings in
last five years and the time period that has elapsed since the womens last
screening. This study group intends to identify factors which may
influence or contribute in some way to womens behavior to obtain an
initial screening and continue to be screened by a health care provider,
annually for women aged 40-75 and biannually for women aged 40-49. Our
study team is optimistic that the results will determine the significance
and benefits of practicing regular screening and will also determine risk
factors that may potentially predict reasons some women practice
appropriate screening for their age group and why others do not. Specific
hypotheses to be tested are: A. There is no difference of breast cancer
stage at diagnosis between women who have had previous breast cancer
screening and women who have not had previous breast cancer screening in
women aged 40-75. B: There is no difference in screening practices among
women seeing a health care provider annually and those not seeing a health
care provider annually. C. There is no difference in screening practices
among women with varying degrees of education. These are three of our
major hypotheses that will be our initial focus.
III. We are optimistic that this study will provide a clear
understanding of how regular screening will effect the number of stage III
cases that are diagnosed each day. Regular screening may be significant
to the population of women that have an aggressive, deep tumor that has
appeared between mammograms, or annual visits to a health care provider.
However, we want to conclude how many cases of stage III breast cancer
could be reduced by either diagnosing it in stage I or II before it has
spread to stage III, as well as diagnosing the stage III patients sooner,
in order to begin treatment sessions much earlier then if the diagnosis is
made after a women has developed a more serious form of stage III breast
cancer. Our results may be advantageous to health care providers as well
as insurance companies. Health care providers, especially primary care
physicians, can review our results and perhaps gain information on
potential confounders which influence a womens breast care screening
practices. They could then direct their focus on particular at risk
populations and educate those populations appropriately.
The most likely and most welcomed results of our study would be to
show that regular breast cancer screening has a protective effect on late
stage diagnoses. Our study plans are to provide a foundation for future
studies. One positive outcome will be that our results will provide
information to insurance companies. The companies consultants could
review the results and attempt to have their policies include mammography
as a primary preventative action that would be covered by all women ages
40 and older, if this was not part of their previous policy. Insurance
companies would attempt to increase the screening practices in their
clients that were not practicing regular screening. It would be to their
benefit to do this in order to decrease the costs of aggressive forms of
cancer treatments. We clearly think that our study will realistically
prove to be important for future studies regarding insurance policies and
decreasing the cost of breast cancer treatments. Although the cost of
annual screening may increase, their overall costs in will decrease.
IV.
A. Cases and controls will be selected from the University of Michigan
Breast Cancer Center. Furthermore, patients form a general medical clinic
at the University of Michigan will also be used as controls. The cases
will be stage III breast cancer patients. Exclusion of stage IV patients
was decided due to an expected poor response rate and limited time period
of the study. Although stage IV patients could strengthen our study we
chose to focus on Stage III because there were treatment options for these
patients, as opposed to stage IV. Form this group of stage III women we
have no reasons to further exclude them from the study. Information
gathered on these patients will be found through the University of
Michigan data base as well as supplemental mailed surveys. The controls
selected from the BCC will be very similar to the cases in terms of
demographic characteristics. The BCC has a homogenous population and
therefore the cases and controls have many similarities excluding the
stage of breast cancer. The controls selected from the general clinic may
or may not be as similar to the BCC population. For example, they may not
have breast cancer or they may have had it at one time and our cured.
More then likely this population will have similar demographic variables
to the BCC. One population that we chose not to use as controls was the
gynecology unit. This is because this population of women may have
cervical cancers, or other disorders which have influenced them to
practice regular breast cancer screening.
B. We chose a fixed sample size of 450. This sample size incorporates a
1:2 ratio of cases to controls (150 cases:300 controls). This sample size
is consistent with our study objectives. Since we are studying three
types of breast cancer screening practices, (breast self exam, clinical
exam, and mammography) we need to select enough women to indicate that
they practice at least one of the three screening practices. With 450
subjects we are optimistic that we can obtain this variety. Since we have
seven different predictor variable for influencing a women to practice
breast cancer screening it is likely that 450 subjects will give enough
responses, which will further enable us to indicate that at least one of
the many predictor variables influenced their breast cancer screening
practices. We are relying on a 60% response rate we must include at least
600 subjects in the initial selection process. Given the sample size of
450, we need to detect an odds ratio of at least .5 in order to have at
least an 80% chance of detecting a difference in the actual screening
behaviors in our cases and controls. We do not want to lower the power of
our study to below 80%. This is one reason for our 1:2 ratio.
C & D. We will gain information about the subjects through the data
bases compiled at the UM BCC as well as the general clinic data base.
Although the general clinic will not have as detailed of a data base as
the BCC, we will be able to obtain addresses for the subjects this way. A
questionnaire will be mailed to all of the subjects, approximately 600.
Enclosed with the questionnaire will be a University of Michigan pen which
will serve as an incentive to complete the questionnaire and send it.
After two weeks another questionnaire will be mailed to the subjects who
have failed to respond. Enclosed with the questionnaire will be a cover
letter written and signed by Dr. Chang and other BCC physicians supporting
the study. We are optimistic that this process will assure us at least a
60% response rate. Subjects will be terminated from the study if they are
deceased of if their stage III cancer has since become a stage IV and the
study group learns of that. We will have to exclude controls from the
general clinic if we find that they do not at all share similar
characteristics with the BCC population. After our questionnaires are
received from the subjects they will be coded and compiled into a separate
data base. Epi-Info and SAS statistical package will be our methods of
data analysis. To make this feasible the subjects will be given a number
and referred to on a numerical basis.
E. We have no plans of intervention in our study. We will include with
our questionnaire a brief summary about how signing the questionnaire
gives the study group permission to use this data for this specific study.
Although these women have the choice to participate by sending the
application, we want to make the point that this is confidential as clear
as possible. By sending and signing the questionnaire they have accepted
the option of participating. If they do not participate in the first
mailing they will be mailed a second questionnaire. If they fail to
respond to the second mailing they will be terminated from the selection
process and therefore the study. The data will be handled confidentially.
After mailing the questionnaire the subject will be assigned a numerical
identification number. Although the study group has access to the names
and addresses in order to send the mailings, these names will be changed
to numbers and all we will know are identification numbers and not which
subject possesses which characteristics. There are no physical risks to
the subjects in this study. The only perceived risk is that of
confidentiality. The members of the study group will not be able to
discuss any specifics about subjects other then with other study members,
and only if this is necessary. We understand that these subjects have the
right to give information in exchange for maintaining their
confidentiality.
F. Data will be collected on numerous variables. The list includes:
breast self-exams
clinical exams
mammography
date of birth
height
weight
education level
occupation
income
race/ethinicty
menopause
insurance
alcohol consumption
religion
history of breast cancer in family
G. The principal exposure of interest is that of the three previously
listed types of screening. The status to be measured is that of frequency
and duration of breast cancer screening prior to diagnosis with breast
cancer. For all purposes, controls will be considered to have stage 0
breast cancer. We want to discover how often they complete self-breast
exams in terms of every month, every other month, twice a year, or never.
We want a specific time period in between self exams. We want to know
what age the subjects started doing them and after that date, how
frequently have they practices breast self-exams since. The same
information will be gathered for clinical exams. Mammography questions
may be more specific because the subjects may have better recollection of
a mammogram then a self-breat exam routine. We want to know how many
mammograms have they had in the past five years. We would expect 4 to 5
to be of regular screening levels for a woman aged 50 or older. Overall,
our exposure status will be measured in terms of specific time periods
(one year, two months etc.,) as well as relative time periods (how long
after menopause etc.,).
H. For the previously listed exposure variable of screening, the outcome
variable will be stage of breast cancer. Controls will be stages O, I or
II. Cases will be all stage III subjects. Outcome status for the cases
will be predetermined by the physicians at UM BCC. They will have already
recorded their outcome status. The outcome will be determined of the
controls as well. They have early stage breast cancer or no breast
cancer. For the controls from BCC the outcome is determined from the
physician.
I. Supplies
disks binder 12.83
envelopes 70.63
BRE 100.00
Letterhead 69.00
stamper 8.21
incentives 107.00
subtotal 260.67
Copies 206.70
postage 708.60
final report 10.00
subtotal 1185.97/2 (shared cost with group 1)
TOTAL 597.99
J. Draft questionnaire 2-17
Draft human subjects review form 2-19
Draft questionnaire for pilot testing 2-19
Final subjects review form 2-19
Mail Questionnaire 2-21 2-26
Second Questionnaire 3-14
Final compiled results 4-21
Kamala Talese Bellamkonda / Group 2 Assig. 11
I. ABSTRACT: Breast cancer is the second most common cancer among women in
the world and the most common cancer among women in developed countries.
Early detection of breast cancer increases a womans chance of survival.
Therefore, there is a need to determine the effectiveness of previous
breast cancer screening (breast self-exam, clinical exam, and mammography)
practices, as well as the factors that influence the frequency of breast
cancer screening practices which may lead to a reduction in the incidence
of late stage disease. In this case-control study, subjects from the
University of Michigan Breast Care Center (BCC) will be given
self-administered questionnaires. The data collected from the
questionnaire and the data from the existing BCC database will be used to
compare levels of previous breast cancer screening practices in relation
to stage of breast cancer at diagnosis in women aged 40-75 and to identify
factors which influence breast cancer screening practices in these same
women.
II. SPECIFIC AIMS: The overall objective is to gain more information about
the preventive efficacy of breast cancer screening practices and what
factors influence women, at the University of Michigan Breast Cancer
Center, to practice breast cancer screening. The data collected in this
project will provide women with more information about breast cancer
screening and will increase the number and frequency of women practicing
breast cancer screening techniques. The hypotheses tested comparing the
level of previous breast cancer screening practices in relation to stage
of breast cancer at diagnosis in women aged 40-75 is the following: there
is a difference of breast cancer stage at diagnosis between women who have
had previous breast cancer screening and those who have not. The
hypotheses tested identifying factors which influence breast screening
practices in women aged 40-75 are the following: there is no difference in
screening practices between women who are less than 50 years old and
greater than 50 years old, there is no difference in screening practices
among women with varying levels of education, there is no difference in
screening practices among women with health insurance and those without
health insurance, there is no difference in screening practices among
women having a first degree relative with breast cancer and those not
having a first degree relative with breast cancer, there is no difference
in screening practices among women seeing a health care provider annually
and those not seeing a health care provider annually.
III. BACKGROUND AND SIGNIFICANCE:
A. This research will broaden scientific knowledge concerning breast
cancer screening. This project will provide information about which type
of breast cancer screening practices are most effective in detecting early
stage breast cancer and the factors influencing a woman to practice breast
cancer screening.
B. The findings from this study will provide health care providers with
information about the effectiveness of breast cancer screening techniques
and the factors that influence a woman to practice beast cancer screening.
Health care providers can use this information to know which screening
methods are useful to recommend to women. Knowing the factors that
influence a woman to practice breast cancer screening helps health care
providers know which groups of women might be under-represented in
practicing breast cancer screening techniques. Therefore, women that meet
particular criteria might need to be urged more to practice breast cancer
screening and subsequently targeted by their health care provider.
IV. RESEARCH DESIGN AND METHODS:
A. Study population selection procedures:
This project is in response to RFA: ED-97-2 sponsored by the Department of
Epidemiology at the University of Michigan School of Public Health and the
University of Michigan Breast Care Center. Possible study subjects, which
meet the RFA requirements, are women between the ages of 40-75. The RFA
requires a case-control study design with cases defined as deaths due to
breast cancer and/or incident cases with metastatic disease with high
probability of dying of disease (stage III or IV), and controls would
include all living members of the source population, namely subjects who
have never had breast cancer as well as subjects who had survived breast
cancer. Selecting a sample of subjects will take place by using the women
attending the BCC since the BCC is one of the sponsors of the RFA and more
than likely to cooperate with this projects needs. The BCC has a database
which contains demographic and clinical data on the women attending the
BCC. Women who have not attended the BCC and those who are not in the BCC
database will be excluded from this study.
B. Sample size procedure:
The selected sample size is 450. Mr. John Alfaro who is in charge
of the BCC database estimated that there are 150 cases of stage III or IV
breast cancer (which are defined to be the cases) in the BCC database.
Since this project is using a 2:1 ratio of controls to cases the number of
controls would be 300.
Using the program SAMPSIZE, an odds ratio of close to 0.5 is
necessary to have at least an 80% chance of detecting a difference in
screening behaviors in our cases and controls.
C. Subject recruitment procedures:
Stratified random sampling will be used. The study population will be
divided according to the subjects stage of breast cancer. Cases will be
defined as those who have stage III or IV breast cancer and controls are
defined as those women without breast cancer or survivors of breast
cancer. The U of M BCC will be asked to list all the women in their
database who have stage III or IV breast cancer and list separately those
women without breast cancer or survivors of breast cancer.
Questionnaires will be sent via mail to all study subjects to acquire
information not available in the database. A cover letter from the BCC
director will accompany the questionnaire to ensure the validity of this
study. A second mailing of the questionnaire to the non-respondents will
occur 1.5 weeks after the first mailing. If a subject fails to respond to
the second questionnaire then they will be terminated from recruitment.
D. Data collection procedures:
As much demographic and clinical data will be abstracted from the BCC
database. Questionnaires will provide the information not available in
the database.
E. Human subject protection procedures:
In order to insure that participation in the study is voluntary the women
have the choice to respond to the questionnaire. A unique study number
will be assigned to the unique name of any subject with any data collected
to protect confidentiality. The questionnaires will be identified by the
unique study number so that information from the database can be
associated with the appropriate questionnaire response. Only the person
in charge of assigning the study number to the unique name will know the
unique study name corresponding to the study number. Any risk to the
subject will be minimized since the study is confidential.
F. Variable List:
stage of breast cancer at diagnosis
frequency of breast cancer screening practices
previous breast biopsy other than for breast cancer
age
family history of breast cancer
oral contraceptive use
body mass index
place of residence
level of education
health insurance status
race/ethnicity
household income
age at menopause
age at first pregnancy
previous breast disease
G. Exposure questions:
In order to determine previous exposure to breast cancer screening it will
be necessary to use open-ended questions as well as close-ended questions.
Since breast cancer screening is defined as either breast self-exam,
clinical exam, or mammography questions about the subjects exposure to
these practices will be asked. Determining the approximate date of
initial exposure and frequency of exposure thereafter will provide
information about the length of time the subject has been exposed to
screening. Questions will be subdivided according to the three screening
practices so that the type of screening practice is specifically
identified.
H. Outcome determination:
Outcome status will be determined as stage of breast cancer at diagnosis
and level of screening. The stage of breast cancer at diagnosis will be
determined by the subjects clinical data found in the database. Level of
screening (years and frequency of screening) will be determined by the
data abstracted from the subjects questionnaire.
I. Budget:
Item
Cost
Supplies:
Disks, binder $12.83
Envelopes (1000@ 7.63) $70.63
BRE ($50/500) $100.00
Letterhead ($23/500) $69.00
Stamper etc. $8.21
Incentive pens $2.14/dozen $107.00
Labels ???
Subtotal $260.67
Copies:
Draft questionnaire for class (60 copies x 5 pgs) $11.70
Final questionnaire 1000 x 5 pgs $195.00
Subtotal $206.70
Postage:
Postage $ 0.55 x 960 $528.00
Return questionnaire $0.43/response (70% response rate) $180.60
Subtotal $708.60
Final report $10.00
Grand total $1185.97/2 (shared cost with group 1)
Total $597.99
J. Time Line:
Homework Assignments Duration Start Finish
Draft questionnaire 1d 2/17/97
2/17/97
Draft Human Subject Form 1d 2/17
2/17
Draft questionnaire 1d 2/19
2/19
Final Human Subject Form 1d 2/19
2/19
Mail out questionnaire 6d 2/21
2/26
Final questionnaire 1d 2/26
2/26
Reminder postcard 1d 3/7
3/7
Second mailing 1d 3/14 3/14
Grant proposal 1d 2/28 2/28
Draft code book 1d 3/14 3/14
Analysis plan 1d 4/4 4/4
Oral presentation 1d 4/18 4/18
Final report 1d 4/21
4/21
Anjum / assignment 10, group 2
I Abstract: Our study is important because it is intended to help
reduce mortality by breast cancer through evaluating screening efficacy
and factors which motivate women to obtain screening. The overall study
goals of our study are to 1) study the preventative efficacy of breast
cancer screening and 2) to study factors influencing a womans behavior to
practice screening. More specifically we aim to compare the screening
history of women age 40 - 75 diagnosed with stage 3 and 4 cancer, to the
screening history of women, in the same age range, diagnosed with stage 1
and 2 cancer, or not having disease at all. We also hope to uncover the
significant factors which influence women to receive their initial
screening and subsequent screens prior to the development of disease. In
terms of our study design, we are using a classic case-control study in
which women are being selected from the University of Michigan Breast Care
Center (BCC), the cases are those women age 40 -75 who have been diagnosed
with stage 3 and 4 disease, while the controls are those with stage 1 and
2 disease, or those without disease. Our study may prove to be
significant if we are able to determine that screening early and
frequently results in the reduction of mortality. Also if we are able to
determine the factors which influence womens screening behavior we can
begin targeting special groups who have particularly low compliance to
screening.
II Specific Aims: The research in this study is intended to further
understand screening behavior. We are interested in gauging womens
attitudes towards screening, what factors influence their behavior and
which do not. Essentially this will be a way to gauge how effectively
women are being reached by the educational messages about breast cancer.
Other studies have found that risk factors such as familial history,
previous breast disease, and age have been factors which motivate women to
obtain screening. Two specific factors we are interested in evaluating
are the presence or absence of health insurance, and the presence or
absence of a regular primary care physician. These factors, if proven to
be statistically significant, can potentially help to reform policy on
breast cancer screening. On a broader level the factors are important in
that they may be able to provide further evidence in the case for
universal coverage of health care in America. The other factors we are
interested in are those which have been previously evaluated, specifically
age (the difference between women older than 50 and those younger),
education level, family history, racial/ethnic groups, and household
income.
Also we are interested in assessing the efficacy of screening as a
method of reducing mortality, specifically referring to breast self exam
(bse), clinical exam, and mammography. We hope to find that women who
have been complying to the prescribed screening regime, are less likely to
be diagnosed with later stage breast cancer. We will compare these women
to those who have not been receiving screening regularly. Our hypothesis
is that these women will have later stage disease which is more difficult
to treat in all respects. Essentially we hope to make a strong case for
screening as the best method of prevention currently available. In
addition, we hope to gauge the efficacy of the other screening methods,
namely bse and clinical exam. We hope to conclude the necessity of these
methods in conjunction with mammography to help reduce late stage
diagnosis.
III. Background and Significance: A) Much debate has surrounded the
appropriate age at which to begin mammography screening. The general
state of confusion regarding a standardized policy and approach toward
mammography has had a detrimental impact on womens knowledge and awareness
of breast cancer screening. In contrast the other methods of screening
are little mentioned in the literature. In our study we hope to find that
bse and clinical exam are also important tools in detecting suspicious
findings. These low technology, relatively easy methods of screening have
often been relegated to second status in light of the high technology
mammography approach. It seems somewhat unfortunate that this should be
the case. These methods, which are not dependent on age, may have value
which has been gravely overlooked in the past. It is our hope to find
evidence that these methods are the first steps toward regular screening
compliance and reduced mortality. Secondly the other gap in knowledge we
are attempting to fill is the introduction of the variables, insurance
coverage and access to a regular primary care physicians. To my knowledge
existing studies often look at various socioeconomic factors and physician
referral as factors which influence compliance to mammography. It seems
logical that women without proper insurance coverage or primary care
access are not likely to obtain mammography. Although this may seem like
an obvious point, it may be worth reiterating in order to reinforce the
need for health care reform in our nation.
B.) A well developed questionnaire should act as a useful and reliable
data collection tool. Questions regarding the utilization (or
underutilization) of bse and clinical exams may result in information
previously ungarnered. Hopefully this information will lead us to
conclusions regarding the frequency of their usage. A subsequent study
could then go on to evaluate these screening methods further comparing
women who have consistently practiced bse versus those who have not. This
second study may definitively conclude the value of these screening tools.
Thus our current study, is a stepping stone to more conclusive work in the
future. In my opinion this will improve health in many ways. First
empowering women with the tools to care for themselves, reduces the
dependence on western technological advances for solutions to disease. It
also promotes womens independence and self esteem. A womens involvement
in her own treatment and diagnosis is often lacking in the current
structure of medicine in our society. The psychologic benefits she would
obtain are immeasurable. Of course I am not advocating that bse replace
mammography. It is my opinion, that it should become a more active
component in the findings of disease. Of course it seems unrealistic to
expect bse to be effective for everyone in all cases, it may be a method
only useful in detecting a limited number of lumps or other suspicious
findings. Under the ideal circumstance, a detection of an irregularity
would result in a mammography sooner then was originally scheduled. If
the mammography concurred with the original conclusion, a case of disease
could be identified earlier than without the intervention of bse. The
promotion of bse is undoubtedly a cost-effective method worth any womans
time. Even if it were only to account for the early detection of less
than .5% of cases, it would still be worth promoting.
IV. Research Design and Methods:
A)Study Population Selection Procedures: Study subjects will come from
the University of Michigan Breast Care Center. The cases, ideally, will
be stage 3 and 4 patients. We do realize, however, that the number of
women at this stage may be less than sufficient. We are also aware that
the later stage patients may be unwilling to participate in the study due
to emotional reasons. In the event that we have too few stage 3 and 4
cases, we may be forced to use some stage 2 patients as our cases. As for
controls, they will be much easier to obtain. They will be women with
stage 0 and 1 disease. Many women go to the BCC for screening that are
fortunate enough not to have cancer. These women are all still entered
into the database, and thus should be accessible for our study. The women
with early stage disease should also be easy to identify through the
database of patients. In terms of inclusion/exclusion criteria women
should be between the ages of 40-75, have had prior contact with the BCC,
should not have history of other cancers (ovarian, cervical etc), and
should be generally healthy (no history of severe chronic disease). The
procedure we will use for identifying these women is to go through the
database at the BCC, with the aid of or instruction from staff members,
and find women who meet our criteria who have been entered into the
database within the last 3 years. For those individuals receiving care
from the center, (current patients) the 3 year restriction will not hold.
After these individuals are identified and approved (if approval is
necessary), we can proceed to contact them.
B)Sample Size Procedures: The sample size was determined based on cost and
time constraints. Considering our limited budget and time, our ability to
complete the study would be severely hampered if the sample size were too
large. I believe we have decided on a total of 600 subjects. The power
was determined by first finding OR values from the literature which
provide a realistic idea for our study and then using the Sampsize program
to actually do the power calculations. We found that we could achieve
close to 90% power with an OR of .52, at a variety of exposure levels
varying from .50 to .60.
C)Subject Recruitment Procedure: We would ideally like to contact each
subject with a letter from the BCC informing them of the study we are
anticipating. Shortly after they receive the introduction letter, we will
send them the actual questionnaire accompanied by another letter
explaining more about the study. However considering we have cost and
time constraints, sending two mailing may be completely unfeasible.
Therefore we may be forced to send both the introduction letter and
questionnaire simultaneously. Once a participant responds to the
questionnaire, by sending it back in the enclosed envelope, we have
confirmed her participation in our study. If we do not receive anything
from a participant with in two weeks, we will send a second questionnaire,
asking them to respond. A cutoff date will be established, in order to
determine the last date questionnaires will be accepted. We do not
anticipate making any personal calls or face to face visits with any of
the subjects in our study. Also any returned mail will be considered as
failure to recruit the participant.
D) Data collection procedure: Once a questionnaire is returned to us
through the mail, it will be stamped with the date of receipt and placed
in a box for data entry. Information from the questionnaire will be
aggregated with information found in the BCC database to form a complete
picture of each study participant. Each participant will be given a
identification number, in order to protect confidentiality, and will then
be entered into a software program for future analysis.
E) Human Subjects Protection Procedure: In order to protect the
confidentiality of each participant we will assign each woman an
identification number which will be used throughout the data entry and
analysis phases. This id number will be a series of letters and numbers
which will be traceable to the actual patient if necessary. No person
outside the study will have access to the patients questionnaire or
database record. Also only group members will know how to trace a patient
by her code number. Considering the nature of our data collection
process, any individual who does not choose to return a questionnaire is
in no manner forced to do so. Simply be taking the time to respond the
participant has indicated a willingness to participate in the study.
Since all contact with participants will be through the mail, little
opportunity for coercion is possible. We will explicitly state in the
letter accompanying the survey that participation in the study is strictly
voluntary. Lastly, the nature of our study precludes the participant from
risking their health or safety in any fashion. For some participants the
completion of the questionnaire may be somewhat of an emotional
experience, beyond that no other adverse effect will result as a
consequence of our study.
F) Variable List: Data is to be collected on the following variables:
age, sex, household income, religion, education level, occupation, alcohol
consumption, tobacco use, marital status, family history of breast cancer,
age at menopause, and age at menarche, ethnicity/race, existence of a
primary care physician, insurance status. In terms of screening history
we are looking for how many mammographies ever obtained, how often
mammography was obtained, at what intervals were mammographies obtained,
age of first mammogram, knowledge of bse, how often bse performed, and how
often clinical exam performed.
G) Exposure questions: The exposure variable we are interested in is
screening history. More specifically we would like to know the frequency
(how often screening was obtained), duration (at what age screening began)
and consistency (how often screening occurred after first screen received)
with which screening was practiced. These will be measured in a variety
of ways. We can create a measure that ascertains annual screening behavior
by putting number of screens in one year over total number of lifetime
screens. This would tell us at what point in a womens screening history
screening increased. From this information we could attempt to hypothesis
what factors account for the increase in screening in any given year.
Also we can look at our cases and controls and compare the amount of time
elapsed between each screen and the number of total screens received.
This information may be pivotal in determining the efficacy of screening
for late stage versus early stage disease. We can also use our age
variable to create age specific screening rates to find which age group
has the highest and lowest levels. Thus we can conclude which age groups
which are less compliant to suggested screening regimes.
Budget for Epid 655 group 2
ITEM COST
Supplies:
Disks, binder $12.83
Envelopes (100@ 7.63) $70.63
BRE ($50/500) $100.00
Letterhead ($23/500) $69.00
Stamper /date stamp/envelope sealer bottles $8.21
Incentive Pens (not yet purchased) 2.14/dozen = $107
Labels ??
Subtotal $260.67
Copies:
Draft Questionnaire (to test in class) 60 copies x 5pgs = $11.70
Final Questionnaire 1000 x 5pgs = $195
Subtotal $206.70
Postage:
Postage $0.55 x 960 = $528
Return Questionnaires $0.43/response (estimate 70% response) = $180.60
Subtotal $708.60
Final Report $10.00
Subtotal $1185.97/2 (shared cost with group 1)
TOTAL $597.99
J. Time line:
2/7/97 - 2/17/97 - Work on questionnaire, human subjects form, and
informed consent
2/19/97 - Pilot test questionnaire
2/24/97 - Mail out first questionnaire
3/10/97 - Second mailing for nonrespondents
3/10 - 3/14 - Complete code book
3/21/97 - Final date for receipt of all questionnaires to be included in
data analysis
3/21/97 - 4/14/97 Data analysis period
4/14 - 4/20 - Write final report
4/21/97 - Final report completed
Arum Atmawikarta / Group-2, Assignment #10
I. ABSTRACT:
The importances of this study are: to determine the effectiveness of
screening practices for breast cancer in order to reduce incidence late
stage disease, and to determine the factors that influence the frequency
of breast cancer screening practices which may lead to a reduction in the
incidence of late stage disease. The overall goal of the study is to study
the preventive efficacy of breast cancer screening(breast self-exam,
clinical exam, and mammography) practices and to study factors influencing
womens behavior to practice breast cancer screening. The aims of the study
are to compare the level of previous breast cancer screening practice in
relation to stage of breast cancer at diagnosis , and to identify factors
which influence breast screening practice in women aged 40-75. The study
design is case-control study . The total sample of 450 consists of cases
(150) and controls (300) which are selected by stratified random sampling
and simple random sampling from database of the University of Michigan
Breast Cancer Center ( UMBCC) and General Medical Care The University of
Michigan (GMCUM)
II. SPECIFIC AIMS:
The overall goal is to study the preventive efficacy of breast cancer
screening practices and to study factors influencing womens behavior to
practice breast cancer screening among women aged 40-75 years. In this
study screening practices consists of three categories: breast
self-examination(BSE), clinical breast examination(CBE) and mammography.
The specific aims of the study:
A. To compare the level of previous breast cancer screening practices in
relation to stage of breast cancer at diagnosis in women aged 40-75 years.
This specific aim is formulated in order to test the null hypothesis (Ho)
that there is no difference in breast cancer stage at diagnosis between
women who have had previous breast cancer screening and women who have not
had previous breast cancer screening in women aged 40-75 years.
B. To identify factors which influence breast screening practice in women
aged 40-75 years. This specific aim is formulated in order to test seven
null hypothesis as follow : (1) There is no difference in screening
practices between women diagnosed with breast cancer in aged less than 50
years old and greater than 50 years old, (2) There is no difference in
screening practices among women with varying levels of education, (3)
There is no difference in screening practices among women with insurance
or without insurance, (4) There is no difference in screening practices
among women having at least one first degree relative with breast cancer
and those not having a first degree relative breast cancer, (5) There is
no difference in screening practices among women seeing a health care
provider annually and those not seeing a health care provider annually,
(6) There is no difference in screening practices among women of different
racial and ethnic groups, and (7) There is no relationship between
screening practices among women with varying levels of household income.
III. BACKGROUND AND SIGNIFICANCE.
A. BACK GROUND
Breast cancer is the second most common cancer among women in the world
and the most common cancer among women in developed countries. With
570,000 new cases in the world each year, breast cancer remains the common
malignancy in women and comprises 18 percent of all cancers among female.
Furthermore it was estimated that the incidence of breast cancer among
women aged 50 approaches two per 1,000 women per year, and the disease is
the single common cause of death among women aged 40-50, accounting for
about a fifth of all deaths in this age group ( McPherson, K et al, 1994).
In the United States 180,000 new cases are diagnosed annually and there
are 45,000 death due to breast cancer in this country annually(Hartmann C
Lynn, 1996). Moreover breast cancer incidence rates in the United States
rose by 24 percent between 1973 and 1991 ( King E.S, Schottenfeld, D,
1996).
Previous studies show that screening practices of breast cancer can reduce
mortality. However, there was a controversy about urging women younger
than 50 years to have annual mammograms. The Swedish study shows that in
the age group 65-74 there was a significant reduction in breast cancer
mortality in the screened group with a relative mortality of 0.68 ( Chen
HH, 1995). Case -control study in Hamspstad-London shows that there is
potential similar attendance at routine screening by older women if they
are invited in the same way as younger women (Horton Taylor D et al,
1996). Other study in Nijmegen shows that regular mammography screening
of women over age 65 can reduce breast cancer mortality by approximately
45 percent (Van Dijck JA, et al, 1996). Furthermore study in DOM Project
Utrecht, concluded that early diagnosis of breast cancer by mammography
reduces breast cancer mortality in women 50-64 years old ( Collete HJ et
al, 1992). The most recent meta- analysis provides statistically
significant proof that screening can reduce the death rate from breast
cancer by at least 24 percent for women aged 40-49(Kopans DB, 1995).
Furthermore there is evidence that the tumor size, axillary lymph node
status, and stage of screening-detected breast cancer in women aged 40-49
years indicate that screening is at least as beneficial for these women as
it is for those aged 50-64 years for whom mammography screening is widely
accepted (Sickles EA Kopans DB, 1995).
Several factors influence breast cancer screening practices such as
education level, per capita income, availability of health care, race and
ethnics group, previous breast cancer diseases, family history, influence
of health provider and friends, and cost of screening. Study of factors
that associated with stage of breast cancer in Black and White women show
that access to health care, a lack of mammograms, and an increased of body
mass index significantly contributed to stage differences in Blacks, where
as income was marginally associated with stage for Whites(Carrie P.Hunter
et al, 1993).
Other study show that physician recommendation is one of the strongest
predictors of mammography use(Mandelblatt J et al, 1995). Similar study
shows that encouragement by physicians is an important factor in
increasing cancer screening rates ( Shapira DV, 1996). Study in
California show that age as demographic variable have significant effect
on follow-up rates of breast cancer screening ( Pamela Arnsberger Webber
et al, 1996)
B. SIGNIFICANCE
Breast cancer is the second most common cancer among women in the world
and most common cancer among women in developed countries. By the year
2000, it is estimated that breast cancer will account for 500,000 deaths
annually, therefore it is important to reduce breast cancer mortality.
Early detection of breast cancer through appropriate screening practical
practices could help control cost of medical and surgical treatment. The
importance of study is to determine the effectiveness of previous
screening practices for breast cancer in order to reduce incidence of late
stage disease. In addition it is important to determine what factors
influence the frequency of breast cancer screening practices which may
lead to a reduction in the incidence of late stage disease. The findings
of this study may lead to changes in policy by providers, insurance
companies, government and women themselves.
IV. RESEARCH DESIGN AND METHODS
A. STUDY POPULATION AND SELECTION PROCEDURES:
This study will use a case-control study design. Cases are women aged
40-75 who have stage III or IV breast cancer and or are likely to die from
breast cancer. The sources of the cases is from The University of Michigan
Breast Cancer Center (UMBCC). Controls are women aged 40-75 without breast
cancer, or who have breast cancer in stage I and II. The sources of
controls are from the UMBCC and from General Medical Care the University
of Michigan (GMCUM). Acquisition of subjects : UMBCC will be asked to list
all of the women in their database who have stage III and IV breast
cancer, or have died from breast cancer and list separately those women
without breast cancer or has stage I and II of breast cancer.
B. SAMPLE SIZE PROCEDURE.
Two sampling approaches that might be used in this study there are simple
random sampling and stratified random sampling. Simple random sampling
will be used if the cases and control selected from UMBCC(more
homogeneous). Stratified random sampling will be used in case we use the
sample from the GMCUM as a controls(more heterogeneous). This study
chose fixed sample size of 450. This sample size incorporates a 1 : 2
ratio of cases to controls. Thus the sample size are 150 cases and 300
controls.
In this study the smallest odds ratio(OR) or the closest to the null value
is 0.83 (This is based on meta-analysis of the efficacy of screening
mammography by Kerlikowske, 1995). For the mild OR it will use 0.66 which
is the lower limit of the confidence interval(CI) in Kerlikowske study and
the larger OR is 0.52. By using the Sampsize program then calculated the
OR using this two exposure levels (alpha=0.05 and Power=0.80). Here women
without breast cancer is serving as a surrogate for women with stage 0, I,
and II breast cancer, and women with breast cancer is surrogate for women
with stage III and IV. Exposure levels in controls are 0.58 and 0.56 using
the upper and lower levels of CI from the BRFS data, and 0.50 derived from
plugging 0.40 exposure in cases and mid-level OR(0.66) into Sampsize to
determine the exposure level in control.
_________________________________________________
Exposure Power Power Power
in control (OR=0.83) (OR=0.66) (OR=0.52)
0.50 16.59 51.8 89.5
0.56 16.9 51.7 89.3
0.58 17.0 52.0 89.4
__________________________________________________
Given the constrains of the sample size, we need to detect an OR of close
to 0.50 in other to have at least an 80% chance of detecting a difference
in screening behaviors in cases and control.
C. SUBJECT RECRUITMENT PROCEDURES
The subject recruitment procedures in this study is as follows :
1. Contact with the person who is responsible in database at UMBCC to
discuss the availability of data related to the study and what kinds of
computer program they used and how to access to database.
2. Make the list all of the women in their database who have stage III and
stage IV breast cancer or who have died from breast cancer (for cases),
and list all the women with stage I, II and without breast cancer( for
controls). In case there is not enough sample for controls will be
collected from GMCUM.
3. Two sampling approaches that might be used in this study: simple random
sampling and stratified simple random sampling as explained above. The
total sample for cases are 150, and the total sample for control are 300.
4. The unique study number will be assigned to each individual subjects to
protect confidentiality.
5. Questionnaire will be sent via mail to all study subjects to acquire
information related to the study.
6. Second mailing questionnaire will be sent one week after first mail
questionnaire to the subjects of study who do not response in order to
increase the response rate.
The criterias for terminating recruitment are: when the subjects do not
response the questionnaire by the deadline, subjects do not want to
participate in the study(discontinue to participatein study), and more
subjects want to participate in the study.
D. DATA COLLECTION PROCEDURE.
The data collection procedure will be done in this study as follow :
1. The stage of breast cancer for cases and control comes from database
UMBCC. In case there is not enough sample for control data of stage of
breast cancer will be gathered from GMCUM.
2. Factors influencing breast cancer screening practices will be collected
from the sample of cases and control from database of UMBCC and from
GMCUM.
3. Data will be coded and manually entered into the microcomputer program,
EPI INFO by two different subgroup of the study investigators. Two data
sets will be compared to identify errors in data entries. CHECK Program in
EPI INFO will be used to control quality and insure consistency in data
entry.
4. For statistical analysis beside using EPI INFO we will transfer the
data set to SAS and SPSS Statistical Program.
5. For quality control, frequency distribution and cross tabulation of
important variables will be calculated at weekly as data entered.
6. Management duties for quality control will be assigned to one subgroup
of the research team. This subgroup will be responsible for overseeing the
accuracy of collecting, editing, and analyzing the data.
E. HUMAN PROTECTION PROCEDURE.
1. The purpose of study.
The purpose of the study is to study the preventive efficacy of breast
cancer screening practices and to study factors influencing women behavior
to practice breast cancer screening. A case-control study will be
conducted in order to describe the relation between stages of breast
cancer and factors influencing breast cancer screening practices which
consists: age, level of education, insurance systems, history of breast
cancer, race/etthnic group, regularity(annual) of screening, and income of
household. The study will consist women aged 40-75 years, with or without
breast cancer who has experience in breast cancer screening practice. Four
hundred and fifty women who fit the selection criteria will be randomly
selected from UMBCC computer database and from patient who comes to
GMCUM.
2. The risk of the investigation to the individual.
This study involves no physical contact with the individuals. The greatest
risk to the individual would be the loss of confidentiality of his or her
responses. To protect against this possibility, every precaution is being
taken to assure their confidentiality. The survey is anonymous. A unique
study number will be assigned to each individual subject. Only those from
UMBCC will be able to associate the unique name of any subject with any
data collected ton protect confidentially.
3. The protection of the rights and welfare of individual.
A mailed questionnaire will be used to minimize potential embarrassment to
the respondent. An introductory letter will accompany the questionnaires
to inform all potential subjects of the studys purpose and legitimacy, and
to assure them that participation in the study is strictly voluntary and
anonymous. The following procedure will be used to protect the
confidentiality of participants respondent:
a. No name, social security number, or code will be placed on the
questionnaire which allow for later identification of respondent.
b. Questionnaires will be stored in the office of the principal
investigator ( School of Public Health University of Michigan )
c. Data collected will be pooled and presented in aggregate. Therefore, no
potential will exist for recognition of any particular individuals
responses.
4. Informed consent process.
The participants will be mailed the questionnaire with a cover letter
which describes the study and the consent process. The letter will assure
the participant that he or she is under no obligation to complete this
survey or that they may choose no to answer any question they deem
inappropriate. Completion and return of the questionnaire implies consent.
No further contact will be made with the participant after mailing out of
the questionnaires the study group. Participants will be instructed to
contact the supervising professor with any questions regarding the study.
A telephone number for the professor will be provided.
F. VARIABLE LIST.
The variables will be used in this study are: stage of breast cancer at
diagnosis, level of screening, age, age at menopause, age at first
pregnancy, family history, previous benign disease, race, household
income, mammography by radiologist, variation of equipment, oral
contraceptive use, body weight, body height, place of residence, level of
education, payment method, medical service experience, hormone use, and
accessibility to health service.
G. EXPOSURE QUESTIONS.
The procedure to measure the variable of previous exposure to screening
as follow:
1. There are three kinds of exposure to screening: breast
self-examination(BSE-by the participant based on written guideline),
clinical breast examination(CBE- by physician or health provider(nurse),
and mammography( by radiologist)
2. To measure previous exposure to screening:
a. The frequency of BSE per year during last 3 years will be asked
through the questionnaire.
b. The frequency of CBE per year during last 3 years will be identified
from medical record of database of UMBCC.
c. The frequency of mammography per year during last 3 years will be
identified from medical record of database of UMBCC.
3. Three are three categorizes levels of exposure :
a. High if the frequency of screening 5-6 times per three years
b. Medium if the frequency of screening 2-4 times per three years
c. Low if the frequency of screening 0-1 times per three years
H. OUTCOME DETERMINATION.
1. The outcome of stage of breast cancer at diagnosis by using the staging
system of American Joint Committee on Cancer. Stage I ( T1 N0), Stage IIA
(T2 N2, T2 N0) Stage IIB (T2 N1, T3 N0), Stage IIIA(T1 N2, T2 N2, T3 N1
N2), Stage IIIB (T4, Any N) and Stage IV (Any T, Any N M1). This stage it
can be observed from mammography record of UMBCC.
2. The outcome of level of screening is define as number or frequency of
BSC, CBE, and mammography per year during last 3 years.
I. BUDGET
ITEM
COST
1. Supplies:
a. Disk, binder = $ 12.83
b. Envelopes (100 x $7.63 ) = $ 70.63
c. Business Replay Envelopes (500x$0.50) = $ 100.00
d. Letterhead ( 500x$0.23) = $ 69.00
e. Stamper/date stamp/envelope sealer bottles = $ 8.21
2. Incentive ( UM-Pens) (50 dozen x $ 2.14) = $ 107.00
3. Copies:
a. Draft questionnaire ( 60 copies x 5 pages x $ 0.39) = $ 11.70
b. Final questionnaire (1000 copies x 5 pages x $ 0.39 = $ 195.00
4. Postage:
a. Postage (960 x $ 0.55) = $ 528.00
b. Return questionnaire (70%x 600 x $0.43) = $ 180.60
5. Final report = $ 10.00
===========
Total =$ 1292.97
(This total budget for 2 Group (Group1 and Group2), or $ 646.49 per group
J. TIME LINE
TASK WEEK NUMBER
1 2 3 4 5 6 7 8 9 10 11 12 13
1 Research question X
2 Background and significant X
3 Study and sample design X
4 Draft questionnaires X
5 Draft human subjects X
6 Draft code book and operation manual X X
7 Draft analysis plan X
8 Mail introductory letter X
9 Human subject review X
10 Final Questionnaires to printers X
11 Pilot testing X
12 Code book and operation manual X
13 Questionnaire distribution X X
14 Data entry X X
15 Data Analysis X X X
16 Preparation of final report X X
Sutthi Jareinpituk / Assignment10, Group2
1) Abstract:
As part of the attempt to reduce the breast cancer mortality rate among
the women in the USA, this proposed project is established to study
screening practices. Mammography, clinical breast examination, and breast
self-examination are instruments used to early detect the disease, leading
to prolong lives of women with breast cancer, due to early diagnosis and
treatment. The goals of this project are to study the preventive efficacy
of breast cancer screening practices and to study factors influencing
women behavior to practice breast cancer screening. This project aims to
compare level of previous breast cancer screening practices in relation to
stage of breast cancer at diagnosis in women aged 40-75 and to identify
factors which influence breast cancer screening practices in women aged
40-75. The importance of the study is to determine the effectiveness of
previous screening practices in order to reduce incidence of late stage of
breast cancer and to determine what factors influence the frequency of
breast cancer screening practices which may lead to a reduction in the
incidence of late stage disease. The public health relevance and
significance of this study are as follows: 1) Breast cancer is the most
common cancer among women in the world, especially in developed countries.
2) By the year 2000, it is estimated that breast cancer will account for
500,000 deaths annually, and hence it is important to reduce breast cancer
mortality. 3) Early detection of breast cancer through appropriate
screening practices could help control costs of medical and surgical
treatment. Case-control is a study design for this project with the
sample size of 450 women, composed of 150 cases and 300 controls. The
target population is women in the University of Michigan Breast Care
Center in Ann Arbor, Michigan.
2) Specific Aims:
The three-month project proposal requests the funds required to establish
the study of the efficacy of breast cancer screening practices on a
reduction of breast cancer and the factors influence women behavior to
practice breast cancer screening. This project is a small-scale study, a
part of the study of Epidemiology 655 (Field Studies in Epidemiology) at
the University of Michigan School of Public Health. Hence, the
researchers are the students who are studying this course. The project
will aim to do the following:
A. To compare the level of previous breast cancer screening practices in
relation to stage of breast cancer at diagnosis in women aged 40-75.
The research described in this application is intended to evaluate the
effectiveness and efficacy of screening practices that are concerned with
disease prevention and health promotion. The specific hypotheses to be
tested is that there is no difference in breast cancer stage at diagnosis
between women who have had previous breast cancer screening and women who
have not had previous breast cancer screening in women aged 40-75.
B. To identify factors which influence breast screening practices in
women aged 40-75.
The research described in this application is intended to identify the
factors that enhance the women aged 40-75 to utilize the screening
practices. Also, the factors hindering the women from the access to the
screening practices. The specific hypotheses to be tested are as follows:
1. There is no difference in screening practices between women less than
50 years old and those 50 years old and older.
2. There is no difference in screening practices among women with
different levels of socioeconomic status.
3. There is no difference in screening practices among women with
insurance and without insurance.
4. There is no difference in screening practices among women having at
least one degree relative to breast cancer and those not having at least
one degree relative to breast cancer.
5. There is no difference in screening practices among women seeing a
health care provider annually and those not seeing a health care provider
annually.
3). Background and Significance:
Breast cancer has been studying for decades, and it has been indicated
that the lives of women with breast cancer could be prolonged if their
tumors could be early diagnosed and treated. Many studies indicate that
there is significant evidence of efficacy of breast cancer screening
practices, however, it is still unknown whether the number of previous
screening practices during the past five years influences a stage of
breast cancer. Also, the question of which screening practices among
mammography, breast self-examination, and clinical breast examination is
the most effective, is still needed to study. Moreover, there is
insufficient evidence to determine the factors that encourage or
discourage women to involve in screening practices.
After this study are published, the results obtained can be used
to convince the politics to support breast cancer screening program and to
consider the most effective screening practice as a tool to early detect
breast cancer, leading to cost-effectiveness preventive programs.
Politics play an important role in public health policies, and hence the
knowledge about the efficacy of screening practices should be clearly
shown to them. Incentives, monetary and non-monetary, influence the
policy making both in terms of resource allocation sectors and political
sectors. Knowledge is one kind of non-monetary incentives used to
persuade decision makers to decide which program is the most appropriate.
Also, the factors influencing breast cancer screening practices should be
presented to politicians in order to advocate the policies pertinent to
the promotion of the factors enhancing the practices and those relevant to
the reduction of the factors hindering the practices. When politicians
realize the importance of the screening practices in minimizing the breast
cancer mortality, breast cancer screening policies will be issued, and
then they will greatly impact on increasing in the utilization of breast
cancer screening and the reduction in breast cancer mortality rate. Since
this project is a small-scale study due to the limited budget and time, it
still has some room for a large-scale study to improve the quality in
terms of generalizability and public health importance.
4). Research Design and Methods:
A Study population selection procedures:
A case-control study will be utilized to collect the data. A
case-control study is a type of observational analytic study in which
subjects are selected on the basis of whether they have (cases) or do not
have (controls) a particular disease under study. The groups are then
compared with respect to the proportion having a history of an exposure or
characteristic of interest. The study population includes women aged
40-75, which have shown to be at risk for breast cancer. Inclusion
criteria for the cases are presented as follows: 1. Women aged 40-75
with breast cancer stage 3 or 4. 2. Sources of the cases is the
University of Michigan Breast Care Center (BCC). If the number of the
cases, drawn from BCC, is insufficient, the convenient cases, drawn from
oncology practices in Ann Arbor, will be added. On the other hand,
inclusion criteria for the controls are presented as follows: 1. Women
without breast cancer or those who have stage 1 and 2 breast cancer. 2.
Sources of the controls are BCC and the University of Michigan Outpatient
Clinic. The controls, drawn from the University of Michigan Outpatient
Clinic, are not associated with cancer and are similar to the patient
population at BCC or has a similar database.
B. Sample Size Procedures:
The fixed sample size of 450 is chosen with respect to the budget
of $500 in order to incorporate a 1:2 ratio of cases and controls (150
cases and 300 controls). This sample size is estimated to be sufficient
to investigate the efficacy of screening practices and to capture the
variety of the predictor variables influencing breast cancer screening
practices.
The simple random sampling technique will be employed to collect the data
in this study. According to the meta-analysis of the efficacy of
screening mammography by Kerlikowske in 1995, the significant OR, used to
study the efficacy of screening practices, ranges from 0.83 to 0.52 and
below. Since exposure to the screening practices is a protective factor,
the smaller the OR, the more the efficacy of the screening practices. The
weak, middle, and strong OR are 0.83, 0.66, and 0.52 or above,
respectively. Based on Michigan 1994 BRFS data, they showed that at least
56% of women over 50 without breast cancer had a mammogram in the previous
year. For exposure in cases, it is based on a tumor registry audit of
mammography by Moseson in 1994, which showed that approximately 40% of the
women diagnosed with breast cancer had a mammogram in the year prior to
diagnosis. After the appropriate OR and exposure in the controls are
identified as mentioned above, the sample size program then calculated to
detect the power by using the fixed sample size of 450, the level of
significance of 0.05, three different levels of exposure in the controls,
and three different OR. The results obtained are presented as follows:
Exposure in controls Power (OR=0.83) Power (OR=0.66) Power (OR=0.52)
0.50 16.9 51.8 89.5
0.56 16.9 51.7 89.3
0.58 17.0 52.0 89.4
Given the constrains of our sample size, odds ratio of 0.5 or below are
needed to obtain at least an 80% chance of detecting a difference in
screening behavior in the cases and controls.
C. Subject recruitment procedures:
The subject recruitment procedures are presented as follows: 1.
Contact the persons who are responsible for the breast cancer database at
BCC in order to determine the availability of the data pertinent to the
study and the accessibility to the database. 2. List the women aged
40-75 in the database who have stage 3 and 4 breast cancer for the cases
and those who have stage 1 and 2 breast cancer for the controls. 3.
Random sapling the list for 200 cases and 400 controls in order to reach
approximately 150 cases and 300 control responses. Reminder postcard,
second questionnaires and incentives are intensive efforts to contact
individuals. Pens as an incentive will be sent with the first
questionnaires. Then reminder postcard will be sent, two weeks after
sending questionnaires, to the subjects who do not respond the
questionnaires to increase response rate. Finally, second mailing
questionnaires will be sent one week after reminder postcard are sent to
increase the response rate.
The recruitment of the subjects will be terminated when they do not
respond the questionnaires or they want to discontinue the study. In
addition, when more subjects wish to participate than the study can
accommodate, some of them will be terminated from the study.
D. Data collection procedures:
The data will be collected through the questionnaires and the BCC
database. The unique study number will be assigned to each individual
subjects to avoid confusing and protect confidentiality. Questionnaire
will be sent by mail to all study subjects to obtain information needed.
All data obtained will be coded according to a standardized code book and
then manually entered into the microcomputer EPI INFO program two times by
different investigators for a double-check purpose. EPI INFO, SAS, and
SPSS are the statistical packages used to analyze the data.
E. Human subject protection procedures:
This study involves essentially no risk to the participants in
terms of physical and mental risk. The informed consent documents will be
sent to the subjects and the subject's institutional medical record in
order to ask for permissions. The overall purpose of the study and its
nature will be explained in the consent form. This form will explain that
participation is strictly voluntary and that the participants may choose
not to answer any questions. This Confidentiality is assured since this
study does not intend to identify the persons and the finished
questionnaires will be stored in the principle investigator's office.
Moreover, only the researcher team in this study will have the right to
access to the questionnaires, and the results obtained will not be
presented in a manner which would allow identification of any individuals
respondent.
F. Variable List:
Outcome variable for the study of the efficacy of previous breast
cancer screening is a stage of breast cancer at diagnosis. Predictor
variable is level of screening, defined as the frequency and the number of
years involving in previous screenings. Confounders are age, age at
menopause, age at first pregnancy, family history, previous benign breast
disease, race, household income, interpretation of mammogram by
radiologist, variation in equipment, oral contraceptive use, hormone use,
body mass index, and place of residence.
Outcome variable for the study of factors influencing women
behavior to practice breast cancer screening is level of screening,
defined as the frequency and the number of years involving in previous
screenings. Predictor variables are age, family history, previous benign
breast disease, race, household income, level of education, and payment
method. Potential confounders are accessibility to health services,
experience with medical services, ethnicity/race, household income level,
age at menopause, and age at first pregnancy.
G. Exposure measurement:
The principal exposure variable in this study is the number of
previous screenings over a period of time in years. In order to assess
the efficacy of breast cancer screening, an adequate period of time in
years involving in screening must be identified because breast cancer is a
disease with long latent period. This study will inquire the information
regarding with the number of years involving in screening during the past
five years for the controls, and during five years before diagnosis with
breast cancer for the cases. This information includes all three methods
of screening practices in order to evaluate the efficacy among the
screening practices. Then, the number of years will be categorized into
three groups, high, medium, and low. Based on the recommendation that
access to screening practices once a year on a regular basis can reduce
the mortality rate of breast cancer, the number of years for high, medium,
and low categories is defined as 4 to 5 or above, 2 to 3, and 0 to 1,
respectively. In this study, exposure to screening practices is a
protective factor, and hence higher exposure should minimize the
probability of dying from breast cancer.
H. Outcome determination:
Outcome of this study are stages of breast cancer, ranging from
stage 1 to 4. Stage 1 breast cancer has a tumor size less than or equal
to 2 cm and no involved nodes. Stage 2 breast cancer has a tumor size
equal to 2.1 to 5 cm. with no involved nodes or a tumor size less than or
equal to 2 cm. with movable ipsilateral axillary nodes. Stage 3 has a
tumor size equal to 2.1-5 cm. with matted or fixed nodes or a tumor size
greater than 5 cm. with movable ipsilateral axillary nodes. Stage 4 has
tumors of any size with direct extension to chest wall or skin and distant
metastasis present.
I. Budget:
Supplies Cost
Disks, binder $12.83
Envelopes $70.63
Letterhead $69.00
Stamper $08.21
Incentive $107.00
Copies:
Draft Questionnaires $11.70
Final Questionnaires $195.00
Postage:
Sending questionnaires $528.00
Return questionnaires $180.60
Final report $10.00
Subtotal $1192.37
Share cost with group1 (1192/2) $596.18
Total $596.18
J. Time line
February 07, 1997 Proposed study objectives and
associated variables
February 12, 1997 Choose study and sample design for
the research
February 19, 1997 Final Human Subject Form
February 26, 1997 Final Questionnaires and Pilot
testing
February 27, 1997 Sending questionnaires
February 28, 1997 Grant Proposal
March 14, 1997 Code Book Development
March 19, 1997 Dummy tables
March 24, 1997 Data analysis plan
April 02, 1997 Revised dummy tables
April 04, 1997 Data analysis
April 07, 1997 Interpret results
April 18, 1997 Oral presentation
April 21, 1997 Final report
Laura Williams Group 2 Assignment 10
ABSTRACT: Breast cancer is the leading cause of cancer-related death for
women. It is clear that the prognosis for survival in women with breast
cancer is improved by earlier stage at diagnosis. For women over fifty, it
is known that yearly mammography can improve survival. However, this has not
been shown conclusively for women under fifty. Furthermore, little is known
about the efficacy of clinical breast exam or self breast exam in earlier
detection or survival, especially for women under fifty. The aims of this
study are twofold. The first will be to compare the level of previous breast
cancer screening practices (mammography, clinical breast exam, and self
breast exam) with presence of, and stage of breast cancer diagnosis in women
aged 40-75. The second will be to determine the factors that influence breast
cancer screening practices in these women. The design will be a case-control
study using mailed questionnaires and data already collected by the Uof M
Breast Care Center, (BCC). The cases will be women without breast cancer, or
stage I-II cancer, and controls will be women with stage III or IV breast
cancer. Both cases and controls will be selected from the BCC. This study
potentially may lead to new knowledge of what if any types of screening
practices are beneficial to women under fifty, and if clinical breast exam or
self breast exam are beneficial to women over fifty. It may also show what
factors influence screening practices, so that this knowledge can be
incorporated into future public health policy or educational programs to
increase useful screening practices.
SPECIFIC AIMS: 1. The first aim is to compare the level of previous breast
cancer screening practices (mammography, clinical breast exam, and self
breast exam) with presence of and stage of breast cancer diagnosis in women.
The objective here is to determine what, if any types of breast cancer
screening practices are effective in increasing survival from breast cancer
(we will be using earlier stage at diagnosis as a surrogate for survival),
and for what age groups.
We hypothesize that: A. mammography will be effective in women over fifty.
B. mammography will not be effective in women under fifty.
C. self breast exam will be more effective in women under fifty than over
fifty.
D. clinical breast exam will be effective in both groups.
2. The second aim is to determine the factors that influence breast cancer
screening practices. The factors we will be studying include: age, education
level, income, race, presence and type of insurance, family history of breast
cancer, and having regular medical care.
The objective here is to determine what, if any of these factors influence
breast cancer screening practices (BCSP) among women. We will determine if
the factors affect screening practices positively, negatively, alone, or in
combination.
We hypothesize that: A. BCSP will be highest for women over 50 and under 75.
B. higher education level will be associated with higher levels of BCSP.
C. higher income level will be associated with higher levels of BCSP.
D. white race will be associated with higher levels of BCSP.
E. having private insurance will be associated with higher levels of BCSP.
F. having a family history of breast cancer will be associated with higher
levels of BCSP.
G. having regular medical care (seeing a physician at least once a year in a
continuity setting) will be associated with higher levels of BCSP.
III. BACKGROUND AND SIGNIFICANCE: Breast cancer is the leading cause of
cancer-related death for women. It is clear that the prognosis for survival
in women with breast cancer is improved by earlier stage at diagnosis. For
women over fifty, it is known that yearly mammography can improve survival.
However, this has not been shown conclusively for women under fifty.
Furthermore, little is known about the efficacy of clinical breast exam or
self breast exam in earlier detection or survival, especially for women under
fifty. This is the most important gap in knowledge which this study will
fill.
This study could lead to improved health if, for example, it determined that
although mammography was not effective in women under fifty, self breast exam
was. This knowledge could lead to greater focus on self breast exam
specifically aimed at this age group. The more women in this age group who
were educated about and practiced self breast exam, the earlier the overall
detection and subsequent survival from breast cancer.
IV. RESEARCH DESIGN AND METHODS:
A. Study Population Selection Procedures: Cases and Controls will be
recruited from the U of M Breast Care Center. We will choose a sample from
the women in their data base who have received some care at the BCC. Cases
will be selected as those diagnosed with stage III or IV breast cancer.
Controls will be those diagnosed with no breast cancer, or stage I or II
breast cancer. We will include only women who are alive currently (if this
can be accurately determined by the data base information). After all those
women entered in the database in the last five years are identified and
separated into cases and controls using the listing procedure capabilities of
the data base program, we will chooses a random sample for controls, and all
the cases (due to the relatively small number of cases expected in the data
base). These women will then be sent a questionnaire in the mail to fill out
and return via mail to us.
B. Sample Size Procedures: Sample size was determined using the sampsize
program. We first used previous studies reported in the literature to
determine the expected exposure levels of previous breast cancer screening
practices in our cases and controls. We used women without breast cancer as
a surrogate for our controls, and women with breast cancer as a surrogate for
our controls. We determined that the exposure level of previous mammography
in controls was .56 and in cases was .40. We also knew that we realistically
could expect a maximum of 150 cases from the BCC data base, and our budget
and time constraints limited us in getting cases from other sources. We also
knew that the number of available controls would be much greater. Using 150
cases and 300 controls, an alpha level of .05, and an OR of .52 the power of
our study to detect an effect is 89.3.
C. Subject Recruitment Procedures: See A. above. Each questionnaire will
be identified by a subject number. Two weeks after the questionnaires are
mailed out, all those not returning the questionnaire will be sent a second
questionnaire.
D. Data Collection Procedures: the questionnaires will be made using
Epi-Info. As they are returned, members of the study team will use Epi-Info
to input the data on the questionnaires into the computer and create our own
data base.
E. Human Subject Protection Procedures: The study subjects will be
identified by the BCC data base manager and assigned a unique study
identification number. This is the only way that the subjects will be
identified. Because the BCC cannot provide the man-power to address the
envelopes, study team members will address the envelopes. This, and with the
second questionnaire mailing, are the only times that any study team member
will have access to both the unique study number and the names and addresses
of the study subjects. No listing of this type will be available outside the
BCC and will only be kept there in their possession. This listing should be
destroyed after the completion of the study. When subjects are mailed the
questionnaire they will be asked to participate by completing it. There will
be no penalties for not participating. The subjects will be offered an
incentive for their time of a U of M pen. The subjects will be told of their
rights to not participate if they choose, and will be assured of
confidentiality of responses and identity.
F. Variables: stage of breast cancer at diagnosis, frequency of SBE, CBE
and mammography, age at first SBE, CBE and mammography, most recent SBE,CBE,
and mammography, age, gender, education level, occupation, household income,
race, menopausal state, health insurance, family history of breast cancer,
frequency of routine physician visits, presence of a primary physician,
religion.
G. Exposure Questions: Previous breast cancer screening - Questions
regarding the age at first SBE, CBE and Mammography, most recent, and
frequency of each will be asked on the questionnaire. Exposure level will
be categorized in several ways. First, and most simply, by those ever and
never having each of the three screening procedures. Second, by those having
CBE or Mammography in the last year (current recommendation for women over
fifty, and controversially for women over forty), and those who have not.
(There are no current evidenced based recommendations for self breast exam,
but most in the medical professions have been advocating once a month).
Thirdly, by those practicing SBE once a month, once every other month, 2-5
times a year, once a year, or never. By using age a first of each of the
three screening exams, frequency of exams and age, we could also categorize
the subjects by "average frequency" of exams, and also by compliance with
currently recommended guidelines.
H. Outcome Determination: Cases and controls will be selected on the basis
of outcome as described in A and C.
I. Budget: Draft Questionnaire 60 x 5 pages = $12.00
Final Questionnaire 600 x 5 pages = $60.00 (volume discount)
Second Mailing 300 x 5 pages =$30.00
Cover sheet (BCC letterhead) 1000 = $46.00
Envelopes (U of M) 1000 = $54.00
Postage .55 x 900 = $495.00
Bus. Reply Envelopes 1000=$96
Bus. Reply Postage .43 x 400 = $180.60
Labels 900 = $22.00
Incentive Pens 600 = $96.00
Other Supplies $25.00
Subtotal $1116.6 / 2 (shared cost w/ grp. 1)
TOTAL $558.30
J. Timeline:
Jan. 8-15
Formulate Research Question, literature search, back ground reading
Jan. 16-Feb. 5
Study Design, sampling procedure
Feb. 5-19
Questionnaire Design and testing
Feb. 14-21, Feb. 14-March 1
Human Subjects form and approval, Grant prep. And approval
Feb. 21-26
Print and ready questionnaires for mailing
Feb. 27
Mail out first batch of Questionnaires
March 1-10 Spring Break
March 11-14
Field work protocols/ data entry protocols
March 14
Second Questionnaire mailing
March 14-April 4
Data Entry, Data analysis design plan
April 4-11
Data Analysis
April 4-16
Preparation of written document
April 18
Presentation of Final Report
Supaporn /assignment #10, group.2
I. ABSTRACT:
Breast cancer is the most common cancer occurring in women in the
United States. The reduction in mortality due to breast cancer, which can
be achieved through the mammography screening and clinical breast
examination has been confirmed only among women 50 years and older, while
there has been no demonstration of such benefits from performing breast
self-exam at present. The goal of this study is to study the preventive
efficacy of breast cancer screening, which is defined as breast self-exam,
clinical breast examination, and screening mammography practices among
women under and over 50 years; and also, to study factors influencing
womens behavior to practice breast cancer screening. The study was
designed as a case-control study, which samples consist of women age 40-75
years, from the University of Michigan Breast Care Center. Data of the
subjects will be obtained from the U of M BCC database, and also,
self-administered questionnaires. The benefit of this study will be to
reduce the need and costs for complicated treatment of late stage breast
cancer, and also, the findings from the study may lead to changes in
breast cancer screening policy by providers, insurance companies,
government, and women themselves.
II. SPECIFIC AIMS:
The overall goal of this study is to study the preventive efficacy
of breast cancer screening, which is defined as breast self-exam, clinical
breast examination, and screening mammography practices;
and also, to study factors influencing womens behavior to practice breast
cancer screening.
The specific aims of the study are conducted as followings.
a. To compare the level of previous breast cancer screening practices in
relation to stage of breast cancer at diagnosis in women aged 40-75.
The null hypothesis was established as there is no difference in breast
cancer stage at diagnosis between women who have had previous breast
cancer screening and women who have not had previous breast cancer
screening in women aged 40-75.
b. To identify factors which influence breast screening practices in women
aged 40-75. The null hypotheses are established as followings:
b.1. There is no difference in screening practices in women diagnosed with
breast cancer between women less than 50 years old and 50 and older.
b.2. There is no difference in screening practices among women with
varying
levels of education.
b.3. There is no difference in screening practices among women with
insurance or without insurance.
b.4. There is no difference in screening practices among women having at
least one first degree relative with breast cancer and those not having a
first degree relative with breast cancer.
b.5. There is no difference in screening practices among women seeing a
health care provider annually and those not seeing a health care provider
annually.
b.6. There is no difference in screening practices among women of
different
racial and ethnic groups.
b.7. There is no relationship between screening practices among women with
varying levels of household incomes.
III. BACKGROUND AND SIGNIFICANCE:
A. Because breast cancer is the most common cancer occurring in women in
the United States, and there is sufficient evidence from several studies
indicated that mortality due to breast cancer can be reduced by 30 to 40
percent among women age 50 and older through the practice of screening
mammography and clinical breast examination (4). However, the evidence of
mammography screening efficacy among women age 40-49 years is still in
controversies (1,4,6,7,8).
Early detection of breast cancer through screening mammography,
physician clinical examination, and breast self-examination has been
recommended. Guidelines for cancer screening have been developed based on
studies of efficacy of breast screening, evaluating the risks and benefits
to the public health of screening, and evaluating the economic impact of
the intervention. The standard for efficacy to which breast screening
studies have been held is the reduction of deaths for the screened group
compared with the unscreened control group. This standard ignores the
other aspect of benefit from early stage diagnosis of breast cancer, which
permits breast-conserving surgery, decreases the need for adjuvant
therapy, and reduces the late sequelae related to extensive treatment and
tumor recurrence (3).
Physician clinical examination of the breast is recommended at
various intervals depending on patient age. Clinical breast exam every 1-2
years is recommended for women age 50 and older. Also, annual mammography
should be performed for women age 50 and older. Breast self-examination
has been recommended to perform every month although there has been no
randomized prospective study to demonstrate the benefits in reducing
deaths from breast cancer (2,3).
However, while women age 50 and older are at the highest risk of
breast cancer nationwide, these women have the lowest screening rates (5).
Barriers, or factors influencing women to practice breast cancer screening
include age, race, socioeconomic status, educational levels, physician
recommendation, lack of understanding of the benefit of screening in the
asymptomatic state, cost, insurance coverage, and difficulty with access
(3,5).
B. In order to achieve better understanding of its usefulness, it is of
importance to determine the effectiveness of previous breast screening
practices, including screening mammography, clinical breast examination,
and breast self-examination, in reducing late stage breast cancer at
diagnosis. Based on this knowledge, early detection of breast cancer
through appropriate breast screening practices could be performed
effectively. The impact thus would result in improving heath and reducing
the need for extensive and more complicated treatment. Furthermore, the
factors influencing frequency of breast cancer screening practices are
also the important issues. The findings from the study may lead to changes
in breast cancer screening policies and practices by health care
providers, insurance companies, government, and women themselves.
IV. RESEARCH DESIGN AND METHODS:
A. Study population selection procedures:
This study was designed as a case-control study. The study
population would be obtained from the U of M Breast Care Center. Inclusion
criteria would be women age 40-75 in U of M BCC database. Cases include
the women age 40-75, who have stage III or IV breast cancer and/or are
likely to die/ or those who died from breast cancer. Controls include the
women with the same age range, who are free of breast cancer, survivors of
breast cancer, or who have breast cancer in stage I or II. Exclusion
criteria would be the women of these age who have severe, or multiple
underlying systemic disease, which could affect the staging of breast
cancer. Also, since the data would be collected through mailed
questionnaires, those without addresses , and those who are unable to read
would be excluded from the study.
B. Sample size procedures:
The procedures used to determine the size of the samples to be
studied or the power of the study to detect effects are described as
follows.
The smallest OR, or the closest to the null value of 1, is .83.
This is taken from a meta-analysis of the efficacy of screening
mammography by Kerlikowske,1995 (10). This is the upper-limit of the CI
around the OR found in his study for women over 50 years old. For the
mid-level OR .66 which is the lower-limit of the CI around the OR in
Kerlikowskes study will be used. The largest OR, or the farthest from the
null value of 1, is .52, which is based on theoretical grounds using two
literature sources for exposure prevalence. For exposure in controls, the
Michigan 1994 BRFS data showed that at least 56% of women over 50 without
breast cancer had a mammogram in the previous year. For exposure in
cases, a tumor registry audit of mammography by Moseson, 1994 (9) showed
that approximately 40% of the women diagnosed with cancer had a mammogram
in the year prior to diagnosis. The sampsize program then calculated the
OR using these two exposure levels(alpha=.05, power =.80). Here women
without breast cancer is serving as a surrogate for women with stage 0, I
and II breast cancer, and women with breast cancer is a surrogate for
women with stage III and IV breast cancer.
Exposure levels in controls are .58 and .56 using the upper and
lower levels of the CIs from the BRFS data, and .50, derived from plugging
.40 exposure level in cases and the mid-level OR, .66, into sampsize to
determine the exposure level in controls.
Exp. in Controls Power (OR=.83) Power (OR=.66) Power (OR=.52)
.50 16.9 51.8 89.5
.56 16.9 51.7 89.3
.58 17.0 52.0 89.4
Given the constraints of sample size, we need to detect an odds ratio of
close to .5 in order to have at least an 80% chance of detecting a
difference in screening behaviors in our cases and controls.
A fixed sample size of 450 was chosen, which incorporates a 1:2
ratio of cases to controls (150 cases:300 controls). This sample size is
consistent with the objectives of the study. Since we are studying three
types of breast cancer screening practices (breast self exam, clinical
breast exam, and mammography), we need to have enough subjects to indicate
that they practice at least one of the three screening practices, and with
450 subjects we should be able to obtain this variety. Since we have seven
different predictor variables for factors influencing a woman to practice
breast cancer screening, it seems that 450 subjects will give us enough
responses indicating that at least one of the seven predictor variables
influenced their breast cancer screening practices.
Some constraints that limit the sample size to 450 or smaller are
budget, the database using to collect data, time, and demographics. Based
on this amount of budget, we will be allowed to send out approximately
1000 questionnaires. Since we are relying on a 60% response rate we are
hoping for 600 actual subjects. However, the databases we are using
indicate that the number of Type III breast cancers (cases) is relatively
small. Due to the database constraints (the number of Type III breast
cancer) we are estimating that we can obtain 150 cases and since we are
using a 1:2 ratio of cases to controls then the number of controls will be
300. So our total number of possible subjects is estimated to be around
450. The time factor limits how many subjects we can use. If we had more
time we would not have to rely so much on the U of M database for all of
our cases and we could recruit people on our own. Finally, the demographic
factors of the women using the U of M BCC limits the generalizability of
our study. Therefore, it would not make economical sense to have a lot
more subjects than 450 since the women at the BCC are a relatively
homogenous population.
C. Subject recruitment procedures:
Both cases and controls would be recruited from the U of M Breast
Care Center. The information on all of the women who have stage III of IV
breast cancer, or who have died from breast cancer, and also, the list of
those women without breast cancer or survivors of breast cancer, would be
obtained from U of M BCC database. Questionnaires will be sent via mail
along with incentive gifts (U of M pens) to all study subjects to acquire
information not available in the database. An introductory letter will
also accompany the questionnaires to inform all studied subjects, and to
assure them that participation in the study is strictly voluntary and
anonymous. The description of the following procedure will be informed to
assure that the confidentiality of the subjects will be protected:
(1). No name, social security number, or other code, which will
allow for later identification of a respondent, will be placed on the
questionnaires.
(2). Questionnaires will be stored in a metal locker, and secured
with combination lock.
(3). Data collected will be pooled and presented in aggregate.
Therefore, any particular individuals responses will not be recognized.
To obtain response rate of at least 60 percent, the postcards will
be sent to remind the subjects after one week of sending out the
questionnaires. If needed, the second postcards will be sent to the
nonresponses two weeks after sending out the first postcards.
D. Data collection procedures:
As much selected demographic and clinical data will be collected
abstracted from the BCC database. Questionnaires will be sent via mail to
all study subjects to acquire information not available in the database.
The questionnaires will be identified by the unique study number so that
information collected from the database can be associated with the
appropriate questionnaire responses.
E. Human subject protection procedures:
(1). The rights and welfare of the individual or individuals involved:
According to the method used to collect the data through mailed
questionnaires, this study involves no risk to the participants. Physical
risk will be not involved, as the contact will be made through mailed
questionnaire. Confidentiality will be assured since none of any names,
addresses, or any type of unique identifier will be placed on the
questionnaires. The rights and welfare of the individual will be protected
by assigning a number to the questionnaire which will be sent out to the
subject, then data from questionnaires will be collected using code number
for each individual women. In addition, all materials relevant to this
study will be stored in a metal locker and secured by combination lock.
Only the investigators mentioned in this study will have access to them.
All data collected will be pooled; no data will be presented in a manner
which would allow identification of any individual subject.
Psychological risk due to the questions will be minimized through
extensive pilot testing, expert review and constant revision of the
questionnaires.
(2). The appropriateness of the methods used to secure informed consent;
To make sure that the subject of the study clearly understands
what the experiment is, what the potential risks are, and has freely, and
without pressure of any kind, elected to participate, following approval
by the human subjects review committee, the consent will be informed by
sending a letter, or a cover sheet attached to the questionnaire,
describing the study and what participation will consist of, to the
subjects. This form will explain that participation is strictly voluntary
and that the participant may choose not to answer any questions.
Completion of the questionnaire will be assumed as informed participation.
(3). The risks and potential medical benefits of the investigation:
The potential medical benefits of this study can be achieved from
the following findings:
1. Early detection of breast cancer through appropriate screening
practices could reduce the need for extensive and more complicated
treatment, thus could help control costs of treatment.
2. The findings from the study may lead to changes in breast
cancer screening policy by providers, insurance companies, government, and
women themselves.
F. Variable list:
(1) According to the overall goal of this study, which is stated
as to study the preventive efficacy of breast cancer screening, which is
defined as breast self-exam, clinical breast examination, and screening
mammography practices, the outcome variable is TNM stage of breast cancer
at diagnosis; the predictor variable is the frequency of previous breast
screening practice within period of time, including BSE during the past
year before diagnosis or onset of symptoms, clinical breast exam in the
past 2 years before diagnosis or onset of symptoms, and mammogram during
the past 5 years before symptoms or before diagnosis of breast cancer for
women who had asymptomatic disease; potential confounders can be defined
as age, age at menopause, age at first pregnancy, family history, previous
benign breast disease, previous breast biopsy, race, household income,
variation in interpretation of mammogram by radiologist, variation in
equipment, oral contraceptive use, hormone use, body mass index, place of
residence.
(2) According to the goal of the study, which is established as to
study factors influencing womens behavior to practice breast cancer
screening, the outcome variable is the frequency of previous breast
screening practice within period of time, including BSE during the past
year before diagnosis or onset of symptoms, clinical breast exam during
the past 2 years before diagnosis or onset of symptoms, and mammogram
during the past 5 years before symptoms or before diagnosis of breast
cancer for women who had asymptomatic disease; the predictors are age,
family history, previous benign breast disease, race, household income,
level of education, payment method; and the potential confounders can be
accessibility to health services, experience with medical services,
ethnicity/race, household income level, age at menopause, age at first
pregnancy. In addition, it can be seen that all predictor variables can be
potential confounders as well.
G.Exposure questions:
Data on previous breast screening practice will be obtained from
the U of M BCC database, and collected from the mailed questionnaires.
Women will be asked about the number of breast screening practices
including BSE during the past year before diagnosis or onset of symptoms,
clinical breast exam in the 2 years before diagnosis or onset of symptoms,
and mammogram during the past 5 years before symptoms or before diagnosis
of breast cancer for women who had asymptomatic disease. In analysis
procedure, exposure variable, or history of breast screening practices
will be dichotomized as never or 0 versus 1 or more.
The examples of questions can be seen as followings:
1) Do you know how to do a breast self exam ?
2) How often do you practice breast self exam during the past year ?
3) At what age did you begin doing breast-self exams ?
4) Have you ever had a doctor or nurse exam your breasts during the past
two years?
5) Since your 40th birthday, how often has a doctor or nurse examined your
breasts?
6) Has a doctor ever recommended that you have a mammogram?
7) Have you ever had a mammogram during the past five years?
8) How many mammograms have you had in the last 5 years?
9) At what age did you get your first mammogram?
10) Between the ages of 40 - 50, how often did you get a mammogram?
11) After age 50, how often did you get a mammogram?
H. Outcome determination:
The outcome, or TNM stage at diagnosis of breast cancer for the
first question, and the frequency of screening practices during period of
time for the second question will be obtained from chart or record review
if accessible, or through hospital cancer registry, or from questionnaires
then will be determined by using the TNM dichotomy, which will be
classified as early stage of disease (in situ/ stage I, II) versus
late stage of disease (stage III or greater), or never (0) versus ever (1
or more) screened. Analyses will be performed using the Epid Info
statistical package, and also, using SAS or SPSS for the analysis using
logistic regression model.
I. Budget:
ITEM COST
Supplies:
Disks, binder $12.83
Envelopes (100@ 7.63) $70.63
BRE ($50/500) $100.00
Letterhead ($23/500) $69.00
Stamper /date stamp/envelope sealer bottles $8.21
Incentive; Pens 2.14/dozen $107
Subtotal $260.67
Copies:
Draft Questionnaire (to test in class) 60 copies x 5pgs = $11.70 Copies:
Draft Questionnaire (to test in class) 60 copies x 5pgs = $11.70
Final Questionnaire 1000 x 5pgs = $195
Subtotal $206.70
Postage:
Postage $0.55 x 960 $528
Return Questionnaires $0.43/response (estimate 70% response) = $180.60
Subtotal $708.60
Final Report $10.00
Subtotal $1185.97/2 (shared cost with group 1)
TOTAL $597.99
J. Time line:
February 19, 1997 Human subjects review
February 26, 1997 Final questionnaire
February 28, 1997 Grant proposal
March 3-9, 1997 Pilot testing
March 12, 1997 Code book development
March 12-31, 1997 Administer questionnaires
April 1-5, 1997 Data entry, data cleaning
April 6-14, 1997 Data analysis
April 18, 1997 Oral presentation
April 21, 1997 Written final report
V. REFERENCES
1. Baines CJ. Screening mammography in women 40 to 49 years of age: No
Important Adv Oncol 1995; 243-51.
2. Center for Health Promotion. Breast Cancer Screening and Detection in
Michigan: Recommendations to reduce mortality. Michigan: Michigan
Department of Public Health, 1990.
3. Leitch AM. Controversies in breast cancer screening Cancer 1995;76(10
Suppl): 2064-9.
4. Public Health Service. The National Strategic Plan for the Early
Detection and Control of Breast and Cervical Cancers. Washington, D.C.: U.S.
Department of Health and Human Services, 1993.
5. Webber PA, Fox P, Zhang X, Pond M. An examination of differential
follow-up rates in breast cancer screening. J Community Health
1996;21(2):123-32.
6. Kopans DB. Screening mammography in women 40 to 49 years of age:
Yes Important Adv Oncol 1995; 231-41.
7. Baines CJ. The Canadian National Breast Screening Study. Why? What
next? And so what? Cancer 1995; 76 (10 Suppl): 2107-12.
8. Harris R. Breast cancer among women in their forties: toward a
reasonable research agenda J Natl Cancer Inst 1994;86(6): 410-2.
9. Moseson D, Meharg K. Tumor registry audit of mammography in
community practice. Am J Surg 1994;167 (5): 505-8.
10.Kerlikowske K, Grady D, Rubin SM, Sandrock C, Ernster VL. Efficacy
of screening mammography. A meta-analysis JAMA 1995;273(2):149-54.
Woosung Sohn / homework#10/ group2
I. ABSTRACT
Breast cancer is the second leading cause of cancer death among women.
Reduction in mortality from breast cancer largely depend on successful
interventions aimed at early detection and treatment. We propose a
hospital-based case control study whose main goal is to evaluate the
preventive efficacy of breast cancer screening procedure (breast
self-exam, clinical exam, and mammography). The goal of this study also
include identification of the factors influencing women', behavior to
practice breast cancer screening. A sample of 450 will be selected from
the patients who enrolled in the Breast Cancer Center(BCC) of the
University of Michigan Hospital. The case in our study would consist of
breast cancer patients who were diagnosed at the stage III or later, and
the control would consist of women with breast cancer who were diagnosed
at the stage II of earlier and women without breast cancer. The
information for the study subjects such as demographic data, history of
breast cancer screening, income, education, familial history of breast
cancer as well as the stage at diagnosis would be acquired mainly from the
data base of the Breast Cancer Center(BCC) of the University of Michigan
Hospital and supplemented by individual questionnaire. In analysis, the
likelihood of having regular breast cancer screening would be compared
between the women diagnosed in later stage of cancer (case group) and the
women who were either diagnosed in earlier stage or without breast cancer
(control group). The association between women's breast cancer screening
practice and many possible-influencing factors would be evaluated using
logistic regression models.
II. SPECIFIC AIMS
We propose a hospital-based case control study whose main goal is to
evaluate the preventive efficacy of breast cancer screening procedure
(breast self-exam, clinical exam, and mammography). The goal of this study
also include identification of the factors influencing women', behavior to
practice breast cancer screening.
One of the main hypothesis of this study is that the women who have had
regular breast cancer screening would be detected in the earlier stage of
breast cancer. The second main hypothesis is that a women's breast cancer
screening practice would be influenced by many factors such as age,
income, education, health insurance, familial history of breast cancer,
and so on. A set of specific aims and the null hypotheses to be tested
for each specific aim are as follows:
The specific aims are:
a. To compare level of previous breast cancer screening practices in
relation to stage of breast cancer at diagnosis in women aged 40-75.
Null Hypothesis: There is no difference of breast cancer stage at
diagnosis between women who have had previous breast cancer screening and
women who have not had previous breast cancer screening in women aged
40-75.
b. To identify factors which influence breast screening practices in women
aged 40-75.
Null Hypotheses:
1. There is no difference in screening practices between women less than
50 years old and greater than 50 years old.
2. There is no difference in screening practices among women with varying
levels of education.
3. There is no difference in screening practices among women with health
insurance or without health insurance.
4. There is no difference in screening practices among women having a
first degree relative with breast cancer and those not having a first
degree relative with breast cancer.
5. There is no difference in screening practices among women seeing a
health care provider annually and those not seeing a health care provider
annually.
6. There is no relationship between screening practices among women with
varying levels of household incomes.
III. BACKGROUND AND SIGNIFICANCE
Breast cancer is the second leading cause of cancer death among women.
Reduction in mortality from breast cancer largely depend on successful
interventions aimed at early detection and treatment. However, there are
still controversies about the efficacy of the breast cancer screening and
the study results shows inconsistencies, especially for the women under
age of 50. Majority of the past studies compared the mortality as the
outcome variable. It is necessary to determine the effectivity of breast
cancer screening practices in detection of earlier stage of breast cancer,
since there are not enough evidences for the presumption of the
association of screening-early detection of breast cancer.
There are some information about the differences between women who
regularly take screening and women who do not, race and income, for
example. However, we largely don't know what factors lead women to breast
cancer screening and what factors hinder their screening practice.
Especially, there are not much information for the role of health
insurance, the role of routine medical care (recommendation of family
physician, for example). It is important to determine what factors
influence breast cancer screening practices in order to increase the
frequency of screening which may lead to a reduction of late stage
diagnosis.
The results from our study might lead to a proper guidance of breast
cancer screening for the women of various age groups to reduce late stage
diagnosis. Early detection of breast cancer through appropriate screening
practices could help control costs of medical and surgical treatment as
well as reduce the mortality. Our results could also be used as basis for
policies of medical care system and health insurance reimbursement system
in relation to cancer screening.
IV. RESEARCH DESIGN AND METHODS
A. Study population selection procedure
We propose a hospital-based case control study of which the main source of
population is the patients who currently enrolled in the Breast Cancer
Center(BCC) of the University of Michigan Hospital. Some patients in the
general clinic of the University of Michigan Hospital might be included as
controls if it is necessary.
The general eligibility of the study population are the women who are
enrolled in the BCC (or the general clinic) of the University of Michigan
Hospital and whose age is 40-75 when they were diagnosed.
The case and control group would be sampled from the populations as;
Cases: Women, who were diagnosed of the stage III or IV breast cancer
during past 5 years (1992-1996).
Controls: Women without breast cancer (stage 0), or with breast cancer
which were diagnosed in the stage I or II during past 5 years
(1992-1996).
B. Sample size procedure
Under some constraints such as budget, the databases we are using to
collect our data, time, and demographics, we chose a fixed sample size of
450. This sample size incorporates a 1:2 ratio of cases to controls (150
cases:300 controls). Our budget will only allow us to send out
approximately 1000 questionnaires. Since we are relying on a 60% response
rate we are hoping for 600 actual subjects, at best. However, the
databases we are using indicate that the number of stage III breast
cancers (our cases) is relatively small. Due to the database constraints
we are estimating that we can obtain 150 cases and 300 controls.
Based on a meta-analysis of the efficacy of screening mammography by
Kerlikowske (1995), the smallest OR, or the closest to the null value of
1, is .83, and the largest OR is .52. These are the results from his study
for women over 50 years old. In our study, minimum OR of 1.75 (same as OR
of .52 in protective association) would be detected with a power of 80% in
the given sample size.
C. Subject recruitment procedure
The subjects in each of case and control groups would be sampled from the
populations as;
Cases: Women, who were diagnosed of the stage III or IV breast cancer
during past 5 years (1992-1996), and Controls: Women without breast cancer
(stage 0), or with breast cancer which were diagnosed in the stage I or
II during past 5 years (1992-1996).
The questionnaire would be mailed to each subject with a cover letter
from the BCC of U-M which explains the purpose and significance of the
study and asks cooperation. A pen with U-M logo on it would be also mailed
together as a sign of gratitude. The second round of questionnaire would
be mailed to those who didn't return the questionnaire within 2 weeks. A
unique study number will be assigned to each individual subject, and
confidentiality of the information would be notified to each subject.
D. Data collection procedure
We are going to use the information which are already collected in the
database of BCC of U-M hospital. The data are comprehensively covering the
demographic information, familial history of breast cancer, history of
breast cancer screening, as well as the stage of breast cancer at
diagnosis. More specific questions about the history of breast cancer
screening along with other information would be collected through
questionnaire.
Questionnaires will be sent via mail to all study subjects to acquire
information not available in the database. The questionnaires will be
identified by the unique study number so that information collected from
the database can be associated with the appropriate questionnaire
responses. Preliminary or more detailed clinical data could be acquired
through chart review by study investigators if necessary.
E. Human Subjects
This project will involve review of database information of currently
enrolled breast cancer patients and the records of patients currently
under treatment. The information for the study subjects such as
demographic data, history of breast cancer screening, income, education,
familial history of breast cancer as well as the stage at diagnosis would
be acquired mainly from the data base of the Breast Cancer Center(BCC) of
the University of Michigan Hospital and supplemented by individual
questionnaire. The questions in the questionnaire would be focused on the
history of breast cancer screening and no offensive or personal questions
would be included. The only risk involved concerns confidentiality. We
intend to maintain confidentiality of our patient sample by assigning each
patient a unique identification number, organizing and securing a master
table of names and corresponding identification numbers, and then
replacing these names with their respective numbers on the duplicated
records, patient questionnaires. The risk of compromising patient
confidentiality is small relative to the potential public health and
clinical benefits of this study.
F. Variable List
Question a:
Outcome variable: stage of breast cancer at diagnosis
Predictor variable: history of screening (defined as number and years of
previous screenings) Potential confounders: age, race, age at menopause,
familial history of breast cancer, previous breast disease, race,
household income, hormone use, body mass index, and place of residence.
Question b:
Outcome variable: level of screening (screened or un-screened)
Predictor: age, family history, previous breast disease, race, household
income, level of education, payment method.
Potential confounders: accessibility to health services, experience with
medical services, ethnicity/race, household income level. (All predictor
variables are potential confounders or effect modifier for each other).
G. Exposure questions:
The key exposure variable in our study is the history of breast cancer
screening. We are not merely interested in whether the person has ever
taken breast cancer screening (mammogram for example), but the number and
regularity of recent screening before the cancer diagnosis. This
information would be acquired by asking "the number of screening practice
during 5 years period just before the cancer diagnosis". The data would be
categorized in several classes or dichotomized into ever or never during
the 5 year period.
The questionnaire would include the following questions;
Have you ever had a mammogram for screening purpose?
a) yes
b) no
What was the date of your most recent mammogram for screening?
month_______ year_________
How many mammograms have you had for the purpose of screening in the last
5 years before you diagnosed? ____________
By the nature of breast-self exam, it would be very hard to get precise
history of breast-self exam as a screening practice. Besides the numbers
and regularity, the level of knowledge and method should be included in
the questionnaire for breast-self exam.
H. Outcome determinations
The outcome variable as the stage of breast cancer at diagnosis would be
determined using the data from the database of the BCC of U-M. The case
and control group would be stratified according to the stage at
diagnosis as follows;
Cases: Women, who were diagnosed of the stage III or IV breast cancer
during past 5 years (1992-1996).
Controls: Women without breast cancer (stage 0), or with breast cancer
which were diagnosed in the stage I or II during past 5 years
(1992-1996).
I. Budget
ITEM COST
------------------------------------------------------------------
Supplies:
Disks, binder $12.83
Envelopes (100@ 7.63) $70.63
BRE ($50/500) $100.00
Letterhead ($23/500) $69.00
Stamper /date stamp/
envelope sealer bottles $8.21
Incentive Pens (not yet purchased)
2.14/dozen = $107
Labels
Subtotal $260.67
Copies:
Draft Questionnaire (to test in class)
60 copies x 5pgs = $11.70
Final Questionnaire 1000 x 5pgs = $195
Subtotal $206.70
Postage:
Postage $0.55 x 960 = $528
Return Questionnaires
$0.43/response (estimate 70% response)= $180.60
Subtotal $708.60
Final Report $10.00
Subtotal $1185.97/2
--------------------------------------------------------------------
TOTAL (shared cost with group 1) $597.99
J. Time line
Task Time line (due date)
---------------------------------------------------------------------
1. Contact the staffs of BCC,
including the staffs who are in charge
of the database of BCC. Month 1 (2/12)
2. Development of questionnaire and
pilot testing Month 1 (2/21)
3. Define sampling frame and
select samples Month 1 (2/23)
4. Extract appropriate information from
the database Month 1 (2/28)
5. Analysis of the database information Month 2 (3/19)
6. Mail the first round of questionnaire Month 1 (2/26)
7. Mail the second round of questionnaire Month 2 (3/12)
8. Data analysis Month 2 (3/31)
9. Oral presentation of study result Month 3 (4/18)
10. Report the study results Month 3 (4/21)
Ade Nuriasri / Assignment#10/group2
ABSTRACT
Breast cancer, the most commonly diagnosed cancer among women in
the USA continues to be a major cause of cancer deaths. By the year 2000,
it is estimated that breast cancer will account for 500,000 deaths
annually. Studies have found a big effect of breast cancer screening on
reducing late stage of breast cancer and breast cancer mortality. Also
early detection of breast cancer through appropriate screening practices
could help control costs of medical and surgical treatment. This
case-control study is designed to study the preventative efficacy of
breast cancer screening such as, breast self-examination, clinical
examination, and mammography and also to study factors influencing women
behavior to practice breast cancer screening. Cases and controls will be
collected from the University of Michigan Breast Care Center and General
Care Clinic. Specific aims of this study are to compare level of previous
breast cancer screening practices in relation to stage of breast cancer at
diagnosis and to identify factors which influence breast screening
practices in women aged 40 to 75 years. This study is important in
determining the effectivity of screening practices for breast cancer in
order to reduce late stage diagnosis and determining what factors
influence breast cancer screening practices in order to increase the
frequency of screening which may lead to a reduction of late stage
diagnosis.
SPECIFIC AIMS
The specific aims of this study are to compare level of previous
breast cancer screening practices in relation to stage of breast cancer at
diagnosis and to identify factors which influence breast screening
practices in women aged 40 to 75 years.
Given the richness of the data set that will be available to us , we will
test the hypotheses relevant to the following topics:
- The difference of breast cancer stage at diagnosis between women
who have had previous breast cancer screening in women aged
40 to 75 years.
- The difference in screening practices between women less than 50
years old and greater than 50 years old.
- The difference in screening practices among women with varying
levels of education.
- The difference in screening practices among women with insurance
or without insurance.
- The difference in screening practices among women having a first
degree relative with breast cancer and those not having
a first degree relative with breast cancer.
- The difference in screening practices among women seeing a
health care provider annually and those not seeing a health care
provider annually.
- The difference in screening practices among women of different
racial and ethnic groups.
- The relationship between screening practices among women with
varying levels of household incomes.
Considering our long term objectives which are a) to determine the
effectivity of screening practices for breast cancer in order to reduce
late stage of breast cancer at diagnosis and b) to determine what factors
influence breast cancer screening practices in order to increase the
frequency of screening which may lead to a reduction of late stage
diagnosis; we hope that the finding of this study may lead to changes in
policy by providers, insurance company, government, and women themselves.
BACKGROUND AND SIGNIFICANCE
Breast cancer, the most commonly diagnosed cancer among women in
the USA and other Western countries, continues to be a major cause of
cancer deaths. In the USA women, breast cancer deaths are second only to
lung cancer deaths, which constitute 22 percent of all cancer deaths in
women. Breast cancer incidence rates have increased in the USA since the
early 1960s in all age groups. By the year 2000, it is estimated that
breast cancer will account for 500,000 deaths annually. Therefore, it is
important to reduce breast cancer mortality. Moreover, prevention of
breast cancer is difficult because many associated factors are endogenous
and thus difficult to manipulate. Studies have shown that certain
characteristics have been associated with increased risk of developing
breast cancer such as, family history, hormones, diet, and life style.
Age-specific incidence rates rise rapidly with increasing age, but
breast cancer differs from other common cancers as the rate of increase
declines after age 50, around the age of menopause. By 1980, there was
general agreement that women older than 50, in particular 50-69 year of
age, should have an annual or biennial (every two years) breast physical
examination and mammography. Also, there was then and even now a
controversy about urging women younger than 50 years to have annual
mammograms. National Institute of Health consensus conference on January
1997 said not enough evidence exists to support a recommendation that all
women have regular mammograms from 40 to 49 years old.
However early detection of breast cancer through appropriate
screening practices could reduce breast cancer mortality and help control
costs of medical and surgical treatment and the finding of this study may
lead to changes in policy by providers, insurance company, government, and
women themselves.
RESEARCH DESIGN AND METHODS
Study Population Selection
This study is a case-control study. Cases and controls will be
selected from the University of Michigan Breast Care Center (BCC).
Furthermore, patients from the General Care Clinic of the University of
Michigan (GCC) will also be used as controls. Cases are women who have
stage III or IV of breast cancer and/or are likely to die from breast
cancer. Controls are women without breast cancer or who have breast cancer
in stage I or II, or survivors of breast cancer.
Sample participants will be women, aged 40 to 75 years old, US
citizens who entered the BCC and GCC for the last three years, and should
have permanent address. Individuals with multiple chronic diseases which
could affect the stage of breast cancer will not be included in this
study. The University of Michigan will be asked to list all of the women
aged 40 to 75 in their database. Questionnaires will be sent via mail to
all study subjects to acquire information not available in the database.
The questionnaires will be identified by the unique study number so that
information collected from the database can be associated with the
appropriate questionnaire responses.
Sample size
Since this study is restricted to gathering information within a
short time frame and with limited financial resources and also with
limited databases we are using to collect the data, we decided to limit
our sample size to 450. This sample size incorporates a 1:2 ratio of cases
to controls ( 150 cases and 300 controls).
The smallest Odds Ratio (OR), or the closest to the null value of
1 is .83. This is taken from a meta-analysis of the efficacy of screening
mammography by Kerlikowske (1995) and this is the upper limit of the
Confidence Interval (CI) around the OR found in his study for women over
50 years. For the mid-level OR we will use .66 which is the lower-limit of
the CI around the OR in Kerlikowskes study. The largest OR, or the
farthest from the null value of 1, is .52. This is based on theoretical
grounds using two literature sources for exposure prevalence.
For exposure in controls we used Michigan 1994 BRFS data which
showed that at least 56% of women over 50 without breast cancer had a
mammogram in the previous year. For exposure in cases we used a tumor
registry audit of mammography by Moseson (1994) which showed that
approximately 40% of the women diagnosed with cancer had a mammogram in
the year prior to diagnosis. The sampsize program then calculated the OR
using these two exposure levels (alpha = .05, power = .80). Here women
without breast cancer is serving as a surrogate for women with stage 0, I,
and II breast cancer, and women with breast cancer is a surrogate for
women with stage III and IV cancer.
Exposure levels in controls are .58 and .56 using the upper and
lower levels of the CIs from BRFS data, and .50, derived from plugging .40
exposure in cases and our mid-level OR, .66 into sampsize to determine the
exposure level in controls; as showed in table below:
Exposure Power (OR=.83) Power (OR=.66) Power (OR=.52)
in Controls
.50 16.9 51.8 89.5
.50 16.9 51.7 89.3
.50 17.0 52.0 89.4
Given the constraints of our sample size, we need to detect an
odds ratio of close to .50 in order to have at least an 80% chance of
detecting a difference in screening behaviors in our cases and controls.
Subject recruitment
Cases and controls will be selected from the University of
Michigan Breast Care Center (BCC). Furthermore, patients from the General
Care Clinic of the University of Michigan (GCC) will also be used as
controls. Sample participants will be women, aged 40 to 75 years old, US
citizens who entered the BCC and GCC for the last three years.
Two sampling approaches that might be used to meet the objectives
of the study are simple random sampling for cases and stratified simple
random sampling for controls. A unique study number will be assigned to
each individual subject. Only those from the BCC and GCC will be able to
associate the unique name of any subject with any data collected to
protect confidentiality. Questionnaires will be sent via mail to all study
subjects to acquire information not available in the database. The
questionnaires will be identified by the unique study number so that
information collected from the database can be associated with the
appropriate questionnaire responses.
Data Collection
An introductory cover letter will be mailed with self-administered
questionnaire. The letter will introduce the researchers and explain the
purpose of the study. It will also include why the potential participants
were chosen for the study and how their names were obtained.
The letter will emphasize the importance of each individual
response. Respondents will be assured that the researchers will take
precautions necessary to ensure anonymity and confidentiality of
information provided. No names or any identifying characteristics will be
on the returned questionnaires. The telephone number of the person
supervising the study will be provided to participants in case further
clarification is needed at any point or if any concerns arise.
The introductory letter, the questionnaire, a return envelope with
postage, a University of Michigan pen (as a incentive) will be mailed to
the potential participants. Returned questionnaires will be kept in the
office of the supervisor in order to assure only authorized persons will
have access to them. The cover letter gives a clear description of the
contents of the questionnaire and participants are instructed that they
may refuse to answer any questions with which they do not feel
comfortable.
Questionnaires will be sent via mail to all study subjects to
acquire information not available in the database. The questionnaires will
be identified by the unique study number so that information collected
from the data base can be associated with the appropriate questionnaire
responses. In order to validate the questions, a literature review will be
done to obtain published studies dealing with similar topics.
In order to ascertain the clarity and conciseness of questions,
colleague testing will be carried out. Pilot testing will be done on a
small number of students who meet inclusion criteria for this study. This
will aide in developing a better questionnaire.
There will be 40 questions and that should take approximately 15
minutes to complete. They will be encouraged to reply to all questions,
but they will be clearly instructed that they may refuse to answer any
part of questionnaire. Data will be collected and analyzed from the
database and questionnaires.
No study investigator will be able to associate the name of any
subject with any data collected to protect confidentiality. As much
demographic and clinical data will be collected from the BCC data base and
GCC as possible.
Human Subject
1. The risk of the investigation to the individual consists of the
following:
This study involves no physical with the individuals. The greatest
risk to the individual would be the loss of confidentiality of her
responses. To protect against this possibility, every precaution is being
taken to assure their confidentiality. This study is anonymous. Subjects
can only be identified by using unique study number.
Due to the personal questions, subjects may feel discomfort. To
limit this possibility, before the start of the questionnaire the
individual is informed that they may skip all or any questions they feel
uncomfortable answering.
2. The rights and welfare of the individual will be protected and
the guarding of the individual from the risks enumerated above will be
accomplished by:
A mail questionnaire will be used to minimize potential
embarrassment to the respondent. An introductory letter will accompany the
questionnaires to inform all potential subjects of the studys purpose and
legitimacy, and to assure them that participation in the study is strictly
voluntary and anonymous. The following procedure will be used to protect
the confidentiality of the participants responses:
- No name, social security number, or other code will be placed on
the questionnaire which will allow for later identification of a
respondent.
- Questionnaires will be stored in the office of the principal
investigator.
- Data collected will be pooled and presented in aggregate.
Therefore, no potential will exist for recognition of any particular
individuals responses.
3. Informed consent process:
The opinion of the committee is that informed consent is the legal
way of describing a meeting of the minds in a contract. In this situation,
it means that the subject of the experimental procedure clearly
understands what the experiment is, what the potential risks are, and has
freely, and without pressure of any kind, elected to participate. the best
way to ascertain that the consent is informed is to have the subject
explain back fully to the interviewer what exactly he or she is submitting
to and what he or she believes the risk might be. The content of this
proposed discussion shall be recorded in detail below.
The participants will be mailed the questionnaire with a cover
letter which describes the study and the consent process. The letter will
assure the participant that she is
under no obligation to complete this survey or that they may choose not to
answer any question they deem inappropriate. Completion and return of the
questionnaire implies consent. No further contact will be made with the
respondent after mailing out the questionnaires by the study group.
Variable List
Following is the list of variables on which data is to be
collected: level of screening (defined as number and years of previous
screenings), age, family history, previous breast disease, race, household
income, payment method/insurance, and level of education.
Exposure Questions
Data are collected from the database of University of Michigan BCC
and GCC and questionnaires. For the principle questions on breast
screening we use breast self exam monthly, clinical breast exam annually
and mammogram annually as guidances. To determine we will use dichotomous
variables (0 vs 1 or more) on those practices within a certain period of
time, for example: breast self exam within last year, clinical breast exam
within the last 5 years, and mammogram within the last 5 years.Questions
on exposure of interest (previous exposure to screening) are as follows:
1a. Do you know how to perform breast self-exam?
Yes No
1b. If Yes, how often do you perform breast self-examination within the
past year?
2a. Have a nurse or doctor examined your breasts?
Yes No
2b. If Yes, how many clinical breast-examination do you have within the
last 5 years?
3a. Have you ever had a mammogram?
Yes No
3b. If Yes, how many mammogram do you have in the last 5 years?
Outcome Determination
Outcome status will be determined by level of screening (defined
as number and years of previous screenings), and the stage of breast
cancer at diagnosis.
BUDGET
Item Cost
Supplies:
Disks, binder $ 12.83
Envelopes (100@ $7.63) $ 70.63
BRE ($50/500) $100.00
Cover sheet (BCC letterhead ($23/500) $ 69.00
Stamper/date stamp/envelope sealer bottles $ 8.21
Incentive Pens ($2.14/dozen) $107.00
Copies:
Draft Questionnaire (60 copies x 5 pages)
(to be tested in class) $ 11.70
Final questionnaire (100 x 5 pages) $195.00
Postage:
Postage ($0.55 x 960) $528.00
Return Questionnaires ($0.43x response)
(estimated 70% response) $180.60
Final Report $ 10.00
Subtotal
(shared cost with group 1) $1292.97/2
Total $ 646.50
TIME LINE
Start Finish
Draft study objectives 1/22/97 1/30/97
Refined study objectives 2/7/97 2/7/97
Study and sample design 2/12/97 2/12/97
Draft questionnaire 2/17/97 2/17/97
Draft Human Subjects and
Proposal preparation 2/17/97 2/17/97
Draft Questionnaire-
for testing 2/19/97 2/19/97
Final Human Subjects
and Pilot testing 2/19/97 2/19/97
Mail out Questionnaire,
study protocol 2/24/97 2/24/97
Final Questionnaire to
be handed 2/26/97 2/26/97
Final Grand Proposal
(to be handed) 2/26/97 2/28/97
Second mailing
questionnaire 3/10/97 3/10/97
Draft code book 3/14/97 3/14/97
Analysis plan /
dummy tables 3/19/97 3/19/97
Oral Presentations 4/16/97 4/16/97
Final Report 4/21/97 4/21/97
Burcu Bayirli/ Assignment 10
Abstract: Reduction in late stage diagnosis of breast cancer leads to
reduction in mortality rates from breast cancer. Indeed, it is known
effective breast cancer screening enables early diagnosis. Although the
most frequently practiced breast cancer screening methods are breast
self-examination, clinical examination, and mammography, it is not known
which one of these methods is most effective in detection of early stage
breast cancer. The goal of the present investigation is to study the
preventative efficacy of the various breast cancer screening
practices--breast self-examination, clinical examination, mammography--,
and factors that influence the breast cancer screening behavior of women.
Level of breast cancer screening practices will be compared to the stage
of breast cancer at diagnosis in women between ages 40-75. Factors that
can influence the screening behaviors of women aged 40-75 will then be
identified (i.e., age education, family history, etc.). Women, aged 40-75
with stage II and/or IV breast cancer will serve as cases, whereas women
who have stage 0, I and/or II breast cancer will be controls. The study
population will be selected from the University of Michigan Breast Care
Center and a general care clinic in the University of Michigan. The
present study may enable health care workers to recommend women with the
most efficient screening method and specify the appropriate ages various
screening practices should be performed.
Specific Aims: Level of previous breast cancer screening practices will
be compared to the stage of breast cancer at diagnosis in women aged
40-75. The null hypothesis that there is no difference of breast cancer
stage at diagnosis in women aged 40-75 whether they had previous breast
cancer screening or not will be tested. The results will shed light on
effects of the three screening methods in mortality reduction. It will
also be possible to determine at what ages these methods are most
effective, because both cases and controls will be stratified across
different age groups. Another aim of this study is to identify life-style
factors that influence breast cancer screening behaviors of women between
ages 40-75. It will therefore be tested that whether or not there are
significant differences between the screening practices of women in
relation to education, insurance, family history of breast cancer, regular
medical care, race or ethnicity, and household income. Once the
relationship of these various life-style factors to type of screening is
specified, it will be possible to find ways to enhance breast cancer
screening among different groups of women.
Background and Significance:
A. Although it is known that various breast cancer screening practices
reduce mortality through early detection of breast cancer, previous
research has not clarified what type of breast cancer screening--breast
self-examination, clinical examination, mammography--is most effective in
early diagnosis. Thus, the most important gap covered through the present
investigation will be the determination of the most effective breast
cancer screening practice.
B. The present research could identify the most effective screening
practice for women in the age groups 40-49 and 50-75. It will be hard,
however, to detect differences with regard to life-style factors due to
the lack of diversity in the study population. Because lack of diversity
decreases the representativeness of the sample, the importance and
generalizability of the results about the most effective screening
practice in the two different age groups will also be limited. The
results will still add to the present body of knowledge and will be able
to determine a health policy regarding breast cancer screening for women
in the age groups 40-49 and 50-75 in conjunction with the findings of any
other studies that might be carried out in more diverse study populations.
For instance, one of the screening methods might be more effective for the
younger age group, whereas another might be recommended in the older age
group. The application of most effective method at the right age would
lead to early diagnosis, and thus to reduction in mortality rates.
Moreover, if future investigations with more representative samples could
identify the effect of life-style factors on the level and type of
screening practices, health care workers would be able to develop new
breast cancer screening programs taking into account specific issues among
various groups of women.
Research Design and Methods:
A. Study Population Selection Procedures: The University of Michigan
Breast Care Center will be asked to provide a list of all women who have
stage III and/or IV breast cancer as well as women with stage 0, I, and/or
II breast cancer. In case there are not enough controls with early stages
of breast cancer, women without breast cancer will be selected from a
general care clinic in the University of Michigan. Other than the stage
of breast cancer, age and diagnosis of breast cancer within the last five
years will serve as inclusion criteria. Women who are not aged between
40-75, and who have history of other type of cancer will be excluded from
the study. Each participant will be assigned a unique number to protect
confidentiality.
B. Sample Size Procedures: Sample size is determined to be 450 with a
1:2 ratio of cases to controls. The smallest OR is .83 and the mid-level
OR is .66 both of which are incorporated from Kerlikowskes study (1995).
Largest OR is taken to be .52, based on the review of literature.
According to Michigan 1994 BRFS data, exposure in controls was that 56% of
women over 50 without breast cancer had a mammogram in the previous year.
Tumor registry audit of mammography by Moseson (1994) showed 40% of women
diagnosed with breast cancer had a mammogram in the year prior to
diagnosis. Accordingly, sampsize program calculated the OR using these
exposure levels.(alpha=.05, power=.80). It was determined that we need to
detect an odds ratio of .5 to have an 80% chance of detecting a difference
in screening behaviors in this sample from U of M BCC.
C. Subject Recruitment Procedures: After the names of the possible cases
and controls are obtained form the U of M database, these subjects will be
mailed questionnaires including a letter describing the study and a
consent form. The subjects will be assured all the information gathered
will remain strictly confidential and the data will be presented in
aggregate form once the research is completed. A pen will be included
with the first mailing to increase the response rate. After two weeks,
non-respondents will receive a second mailing. After a period of two
weeks following the second mailing, non-respondents will be excluded from
the study.
D. Data Collection Procedures: Demographical and clinical data that is
available on the questionnaires administered in the U of M BCC will be
collected from their database. Any other information that is necessary
for the purposes of this study will be tried to be collected via the
mailed questionnaires. After the data collection is finished, data will
be coded and entered into the computer. The data entry will be repeated
two times to avoid any mistakes.
E. Human Subject Protection Procedures: Subjects will fill out the
questionnaires and mail it back to us only if they are willing to do it.
Indeed, a consent form will be included with the questionnaire. Once the
data is collected, each subject will be assigned a number, and any link
between these numbers and their names will be destroyed. Researchers
handling the data will be signing a form of confidentiality and all the
information will remain confidential. The data will be presented in
aggregate form at the end of the investigation. Furthermore, a human
subjects review form will be completed and submitted to guarantee that
participants will not be harmed in any way.
F. Variable List: Age, occupation, race, family history of breast
cancer, insurance, household income, education, marital status, height,
weight, sex, medical history, breast self-examination, clinical
examination, mammography, age at menopause, age at menstruation, age at
first pregnancy, estrogen hormone therapy, and smoking and drinking habits
are the variables data will be collected on.
G. Exposure Questions: The exposure variable of interest is screening.
Three types of screening practices are being investigated in this study,
and those are breast self-examination, clinical examination, and
mammography. On the questionnaires, questions that directly seek the
answers whether or not the subjects had these exposures are included. U
of M BCC database will be reviewed to assure that the answers are correct.
Precise estimates can be obtained for mammography and clinical
examination, whereas it is harder to measure breast self-examination. For
measuring breast-self examination, we are totally relying on the response
of the subject. If the same answer is in the database as in the mailed
questionnaire, however, the chances of making a correct estimate
increases.
H. Outcome Determination: Outcome status will be determined through the
records of U of M BCC. Outcome is classified as stages III and/or IV
breast cancer and stages 0, I, and/or II breast cancer.
I. Budget:
Miscellaneous supplies $12.83
Draft questionnaire $11.70
Final questionnaire $206.70
Second mailing $180.60
BCC letterhead $69
Envelopes $54
Postage $528
Business Reply $276.60
Envelopes
Labels $21.00
Final report $10.00
Incentive $20.00
Total $1390.43
These costs will be shared with Group I.
J: Timeline:
1/15/97 Review of literature
2/7/97 Definition of study objectives
2/12/97 Study and sample design
2/19/97 Draft questionnaire
Final human subjects form
2/20/97 Final questionnaire
2/21/97 First mailing
3/14/97 Second mailing
3/21/97 Data entry initiation
4/3/97 Completion of data entry
2/28/97 Grant proposal
3/14/97 Preparation of draft codebook and operations manual
4/4/97 Data analysis
4/21/97 Final report
Linda Hipolite /Assignment #10,Group #2
I. ABSTRACT
The broad long term objectives of this proposal are to study the
preventive efficacy of breast cancer screening ( breast cancer self-exam,
clinical exam, and mammography) practices in women. In addition, to
determine the what factors influence womenUs behavior to practice breast
cancer screening in order to increase the frequency of screening which may
lead to a reduction in the incidence of late stage disease. The specific
aims of the study are 1) to compare the level of previous breast cancer
screening practices in relation to stage of breast cancer at diagnosis in
women aged 40-75 ; 2) to identify factors which influence breast screening
practices in women aged 40-75. The health relatedness of the project is
that early detection of breast cancer through appropriate screening
practices could help reduce mortality . In addition, determining which
factors influence women to practice breast screening are important in
community interventions. The research design of this study will be a
case-control. Acquisition of cases and controls will be from the database
at the University of Michigan Breast Cancer Center. The demographic and
clinical data will be abstracted from the BCC database. In addition,
questionnaires will be mailed to all study subjects to acquire information
not available in the database. Data will be analyzed using the SAS
statistical package. Results of the study will be disseminated via
presentations at the School of Public Health.
II SPECIFIC AIMS
The broad long term objectives of this proposal are to study the
preventive efficacy of breast cancer screening ( breast cancer self-exam,
clinical exam, and mammography) practices in women. In addition, to
determine the what factors influence womenUs behavior to practice breast
cancer screening in order to increase the frequency of screening which may
lead to a reduction in the incidence of late stage disease. The specific
aims of the study are 1) to compare the level of previous breast cancer
screening practices in relation to stage of breast cancer at diagnosis in
women aged 40-75 ; 2) to identify factors which influence breast screening
practices in women aged 40-75. It is hoped that the findings of our study
may lead to changes in policy by providers, insurance companies,
government, and women themselves.
It is hypothesized that a relationship exists between level of
previous screening practices and stage of cancer at diagnosis. Also it is
hypothesized that there are identifying factors which influence breast
cancer screening practices among women.
Null H1: There is no difference in breast cancer stage at
diagnosis between women who have had previous breast cancer screening and
women who have not had previous breast cancer screening in women aged
40-75.
Null H2:
1. There is no difference in influencing factors among women of varying
educational levels, income levels, different racial/ethnic groups, age,
having insurance, or family history of breast cancer and frequency of
screening practices.
III. Background and Significance
A. Breast cancer is the leading diagnosis of cancer in females in North
America and the second leading cause of cancer-related deaths. The
leading cause of death among women age 35-54 yeas of age, breast cancer
represents about 30% of all new cancers in women each year Though 5 year
survival rates fall dramatically with regional; ( 73%) and distant
( 18%) disease involvement, National cancer Institute statistics show that
94% of patients whose breast cancer is detected early, when it is
localized, survive 5 years beyond diagnosis( American Cancer Society,
1995). Because randomized trials of cancer screening have show the
screening is more beneficial in controlling breast cancer
( Fletcher,1993) that in controlling other cancers, oncologists believe
that early detection is the most important means of limiting breast cancer
morbidity and mortality. However, barriers to screening exist. Therefore,
at the present time it continues to be worthwhile and necessary to assess
associations in relation to factors influencing screening behaviors.
B. Although the study sample is very homogenous from the Breast Care
Center (BCC) the study results may be used in conjunction with other
studies to contribute to the knowledge base on efficacy of breast cancer
screening and stage of diagnosis. It is hoped that we can add to the
knowledge on the practice of Breast Self Exam and its influence on seeking
mammography. Also if information on factors which influence the frequency
of breast cancer screening can be elicited it may have an impact on
community intervention programs.
IV. RESEARCH DESIGN AND SIGNIFICANCE
A. The intended study population comprises all women who have been
diagnosed with breast cancer at the University of Michigan Breast Cancer
Center in Ann Arbor, MI. The BCC was chosen due to its extensive database
and accessibility of the information on the potential study sample. The
sample selection will be obtained from the database at the Breast Care
Center at the University of Michigan .
Sample Selection.
According to the database records at the BCC, there are 3,400
cases of diagnosed breast cancer. The individuals referred to the center
are primarily white, high SES women with commercial insurance. The study
design will be a case-control . A list of women with stage III and IV
breast cancer will make up the cases in this study. Similarly a
generated list will be made of women with stage 0, I, II breast cancer
and controls will be chosen for the study . We will seek recruitment of
available women who meet the inclusion criteria at the BCC. Criteria for
inclusion in the study are as follows:
1. Women in the BCC database who were diagnosed with breast cancer
stages 0-IV after 1992.
2. Women between 40-75 years of age.
3. Submission of informed written consent.
B. SAMPLE SIZE PROCEDURES According to the database records at the BCC
there are 3,400 cases of breast cancer. This study will employ a fixed
sample size of 450. This sample size incorporates a 1:2 ratio of cases to
controls ( 150 cases: 300 controls). The sample size in this study is
constrained due to budget, time and the database we will be collecting
information from. We need to detect an odds ratio of close to .5 for a
meaningful difference in screening behaviors in our cases and controls.
Given the constraints of our sample size , with an alpha level of 0.05 and
a power of 80% an ability to reject hypotheses 1 and 2 would require a
sample size of at least 450. A larger sample size would increase the cost
of the study and a smaller sample size would jeopardize the ability to
judge minor relationships in the data with confidence.
C. Subject Recruitment
For this study, we will be recruiting subjects from the BCC in Ann Arbor.
A group from the research team will work closely with the BCC staff
during recruitment, the goal is to maximize recruitment. In addition to
the approval of the BCC director, we will seek an approval from the IRB
board at the University of Michigan. A cover letter will be sent with all
questionnaires from the BCC to encourage participation of controls and
cases. Criteria for inclusion will be the return of the informed consent
form. Criteria for terminating the recruitment of a subject will be if
they have not returned the informed consent within two weeks of the first
mailing or if they deny consent.
D. Data Collection
The data to test the hypothesis that there is no difference in
breast cancer stage at
diagnosis between women who have had previous breast cancer screening and
women who have not had previous breast cancer screening in women aged
40-75 will come from the BCC database. As much selected demographic and
clinical data as possible will be abstracted from the BCC database.
Questionnaires will be sent via mail to all study subjects to acquire
information not available in the database.
The data collection instrument will be self designed . Each
participant will receive a questionnaire. The survey instrument will be
accompanied by a cover letter explaining the nature, purpose, and
importance of the project. Confidentiality of response will be
emphasized. To ensure confidentiality all participants in the sample will
be assigned code numbers . The questionnaires will be identified by the
unique study number so that information collected from the database can be
associated with the appropriate
questionnaire responses. All results will be reported in aggregated
form. In addition, the cover letter will state that the participant is
under no obligation to answer any or all of the questions and that
participation has no effect on the level of treatment at the Breast Care
Center.
The survey instrument .The survey instrument will be developed using
information gathered during a literature search and with the advice of the
faculty at the University of Michigan School of Public Health. The
questionnaire will be pretested by students at the school of Public Health
as well as survivors of breast cancer. The questionnaire will be revised
to facilitate more accurate collection of data. The questionnaire will
be composed of questions eliciting information on factors potentially
related to breast cancer screening including contraceptive, reproductive,
and menstrual histories; selected history of lifestyle exposures; history
of mammograms before diagnosis; previous screening procedures BSE,
Clinical Examinations -frequency ; familiarly history of breast cancer;
and general demographic traits.
The data gathered in the proposed investigation will be complied
and analyzed using the University of Michigan computing sites. The
Statistical Analysis System (SAS) is the data analysis statistical package
which is hoped to be used. Another data analysis package may have to be
used in the event that the database from the BCC is not compatible with
SAS.
E. Human subject protection procedures:
No information will be gathered without the consent of the
participant.
Confidentiality will be protected by issuing each subject a unique
identifying number for analysis. Linking between questionnaires and the
database records can only be made by linking the number with the consent
record at the BCC. All records in the BCC database computer system are
subject to strict confidentiality guidelines. The study protocol will be
submitted for approval by the University of Michigan School of Public
Health Institutional Review Board
F. Variable list
Variables of Research Questions
Question a: outcome variable: stage of breast cancer at diagnosis
predictor variable: level of screening (defined as number and years of
previous screenings)
potential confounders: age, age at menopause, age at first pregnancy,
family history, previous breast disease, race, household income,
interpretation of mammogram by radiologist, variation in equipment, oral
contraceptive use, hormone use, body mass index, place of residence.
Question b: outcome variable: level of screening
predictor: age, family history, previous breast disease, race, household
income, level of education, payment method.
potential confounders: accessibility to health services, experience with
medical services, ethnicity/race, household income level, age at
menopause, age at first pregnancy. (All predictor variables are potential
confounders also).
H. OUTCOME DETERMINATION
We will assess the outcome measure which is stage of cancer at diagnosis.
Curable carcinomas are likely to be smaller tumors, detected at an earlier
stage of disease. The effectiveness of a screening program can be measured
by comparing the stage distribution of breast cancer at diagnosis with
similar data from published reports.
The primary exposure variable of interest in this study is
previous exposure to screening. It is classified into three categories. To
measure mammography and clinical exam information will be abstracted from
the database and compared with data collected by questionnaire. It will be
more precise measurement than breast self exam. Breast self exam will be
self reported via questionnaire. Some bias may be present with this type
of measurement.
I. Budget
All costs are based on sending 1000 mailed questionnaires.
ITEM COST
Supplies:
Disks, binder $12.83
Envelopes (100@ 7.63) $70.63
BRE ($50/500) $100.00
Letterhead ($23/500) $69.00
Stamp $8.21
Incentive Pens $107
Labels ??
Subtotal $260.67
Copies:
Draft Questionnaire (to test in class) 60 copies x 5pgs = $11.70
Final Questionnaire 1000 x 5pgs = $195
Subtotal $206.70
Postage:
Postage $0.55 x 960 = $528
Return Questionnaires $0.43/response (estimate 70% response) =
$180.60
Subtotal $708.60
Final Report $10.00
Subtotal $1185.97/2 (shared cost with group 1)
Equipment
No equipment is necessary
Travel
The data will be collected by questionnaire . No travel is necessary.
Patient Care Costs
No patient care costs will be incurred.
Other Expenses
None
Budget Justification
Personnel
The primary investigator is the only personnel. No expenses are
anticipated.
Consultants Costs
No expenses are anticipated at this time. Statistical consultation is
provided to the investigator at no charge.
Supplies
A good portion of the budgeted moneys in this section will go toward the
cost of the mailed questionnaire and written report. Paper, envelops,
printing costs and stamps will be used for this purpose. Computer disks
will be purchased for storage of the data. Below is an itemize list of
supply expenses.
J. Time line The following timeline describes the approximate time of
the activities involved in the proposed investigation. The survey
instrument will be sent out in . One follow up mailing of a letter will
be made to non-respondents at a two week interval in late . The letter
will point out that the respondent has been sent a questionnaire but has
not yet responded. The letter will urge them to respond and offer to send
another questionnaire. The importance of a high response rate will also
be emphasized.
Returned surveys will be coded by the investigator for follow-up
mailings. Data collection will be cut off in .
Study Activity Schedule
__________________________________________________________________
IRB February 2/19/97
Proposal acceptance February 2/28/97
Finalize Questionnaires February 2/19/97
Initial Mailing of Questionnaires February 2/2/26/97
Follow up letter March 3/14/97
Data collection cut off March 3/28/97
Data analysis April 4/4/97
Presentation of Results April 4/18/97