Laurie
L. Fajardo, MD #{149} Michelle Saint-Germain, Rose, MD #{149} Bruce J. Hillman, MD2
Cathryn
Factors Screening
of well-being,
satisfaction
with
health care, and knowledge about breast cancer. Women who underwent mammography were more likely to have a regular physician, to practice breast self-examination, to be less
concerned
over
the
cost
of mam-
mography, to be generally satisfied with their health care, to believe they have a greater sense of control over their health, and to be more knowledgeable about mammography and breast cancer. Age, health behavior, sense
and
of well-being,
breast
locus
of control,
cancer
experience differentiate between women had undergone mammography those who had not. Index terms: #{149} Cancer ologists Radiology
1
From
1992;
the
helped who and
of Radiology
(L.L.F.,
T.J.M., CR., B.J.H.) and the Southwest Institute for Research on Women (M.S.G.), University of Arizona and University Medical Center, 1501 N Campbell Ave. Tucson, AZ, 85724. Received September 10, 1991; revision requested November 11; revision received January 13, 1992; accepted February 21. Address reprint requests to Current address: University of Virginia Charlottesville. 0 RSNA, 1992
cancer is the most common malignancy and the second heading cause of cancer death among American women (1). Reduction in breast cancer mortality depends on early detection and treatment; undergoing regular screening mammography is one of the most important factors promoting the early detection of breast cancer (2-6). However, both women and their physicians have been show to adopt screening mammography for preventive health maintenance. Early studies focusing on the reluctance of physicians to recommend screening mammography suggested that physician education would result in increased patient referrals for mammography (7,8). Still, while the percentage of women practicing preventive breast care has increased, less than 50% of age-eligible women follow the recommendations of the American Cancer Society for breast cancer screening (8). Recent campaigns have targeted increasing women’s knowledge of breast cancer and its detection and have encouraged women to refer themselves to breast cancer screening clinics. However, even when offered the opportunity to undergo mammography in a setting where the cost is removed,
do not avail cer screening stand
why
many
women
still
themselves of breast (9). To better underwomen
factors,
Department of Radiology, Health Sciences Center,
attitudes,
behaviors screening
undergo
and
can-
AND
Development
health-related
with
METHODS
and
Study
Population We developed questions about
a survey demographic
health general
behavior, health concerns, well-being, satisfaction
health
care, and knowledge
containing information,
sense with
about
of
breast
cancer. The conceptual framework for the survey was adapted from previously described health rating instruments for measuring general health perceptions on the basis of the health belief model and health
locus
of control
(10-14).
Where
appropri-
ate, survey variables also focused specifically on breast-related health. Items initially considered for inclusion in the survey were grouped into scales representing major aspects of behavior that they were hypothesized to measure. To verify that each survey item had the highest relationship to the particular scale to which it was assigned, we administered a pilot survey to three groups of 25
women from the three separate health facilities from which we would draw the study sample (described below). On the basis
of the
responses
questionnaire,
were were
to the
correlations
independently
analysis,
and
discarded.
pilot
examined
invalid
study
among
items
with
or redundant
Criteria
factor items
of discriminant
validity were applied in evaluating itemscale correlations (15). To retain an item in a particular scale, each item-scale correlation was required to be higher for the scale the item was hypothesized to measure than for all other scales.
On the basis of the pilot study results, we developed the 44-question survey used in the actual study, which encompassed six scales of health perceptions. These scales and the variables of which they are comprised are shown in Table 1; the survey instrument is shown in Table 2.
screening
have on compliance mammography.
MATERIALS
III, MD
to Undergo
REAST
Survey
L.L.F. 2
J. Meakem
mammography and to further develop information to guide the promotion of breast cancer screening, we investigated what impact personal
184:59-63
Department
B
barrier
Breast radiography, utilization, screening #{149} Radiology and radi-
00.11
#{149} Thomas
Influencing Women Mammography’
Despite the fact that mammography is a valuable tool for early detection of breast cancer, the majority of ageeligible American women do not avail themselves of screening mammography. To better understand why women do or do not undergo mammography and to further develop guidelines for promoting breast cancer screening, the authors investigated what impact personal factors, attitudes, and health-related behaviors have on compliance with screening mammography by surveying 521 women for information related to health behavior, health concerns, sense
PhD
Of the 521
557 responded
separate
women
health
women from
approached 1989 to 1990
facilities.
at a general
initially, at three
We surveyed
medicine
108
clinic at the
University of Arizona, Tucson, and 115 at a health maintenance organization (Cigna Health Plan of Arizona, Tucson) to sample a cross-section of women who were seeing health care providers for reasons other than breast health. To ensure that we
would have a sufficient number of women who had undergone mammography, we also surveyed 298 women waiting to undergo mammography at a local breast imaging center (Tucson Breast Center). Women aged 35 years or older were surveyed sequentially, and the response rates
59
were similar ( > 91 %) at all three sites. Respondents ranged in age from 35 to 81 years (mean, 54 years). Most women were white (88%); 7% were Hispanic, and 5% were of other racial backgrounds. Sixtynine percent were married or living with a partner; 24% were separated, divorced, or widowed; and 7% were never married. Educational attainment ranged from 8th grade or less to graduate degrees. The mean was some education past the high school level. The mean annual household income averaged $25,000-$30,000 per year but varied from $5,000 or less to more than $60,000. Only responses from women with no symptoms breasts
(ie,
or complaints women who
related to their are potential can-
didates for screening) were data analysis (7’ = 498).
used
in the
Analysis The responses to the items on the questionnaire were coded as a series of Likerttype scales ranging from 1 to 5, with I being lowest and 5 being highest (10,11). A score for each of the six health perception scales was computed for each respondent by using the simple algebraic sum of scores for items that satisfied factor analytic and discriminant validity criteria. Scores for the knowledge scale consisted of the total number of questions answered correctly. We performed general descriptive analyses and tests of association (x2 test). Bivariate
analysis
for
the
survey
variables
was performed on a scale-by-scale basis using the Wilcoxon rank-sum test. We used the results of this analysis to test multivanate models examining combined effects
of the
factors
in the
six
scales.
by
We
used discriminant analysis to determine how well individual study variables helped differentiate between women who had undergone mammography from those who had not. Discriminant analysis is a procedure that enables classification of observations or subjects into discrete groups by using one or more quantitatively measured variables (15,16). Variables
with
2
values
commensurate
with P < .05 or better for showing a difference between women who had undergone mammography and those who had not were used in the final analysis. In this analysis, we developed a predictive model of women likely to seek mammography. That is, by knowing a woman’s score on a series of questions, we can predict whether she has undergone mammography and whether she will be likely to undergo mammography in the future.
pared years than
equal
proportions
races
had
Of the surveyed, undergone once. Of
older,
women
395 (75.8%) had mammography 322 women aged
274 (85%)
mammography 60 #{149} Radiology
had
previously at least 50 years
undergone
at least
once,
corn-
we
and
With
a multivariate
personal
health locus
actions
did
and
previously were
undermore
previously phy.
can
affect
those
undergo
health),
who
mammography,
(45%)
were
it was 311 had
mammogra-
women
did
not
only
correctly
the model. This prior probability
sense (that
experience, differentiate (94%) who
undergone
For
model
behavior, of control
and breast cancer possible to correctly of the 331 women
likely to see their physician regularly, to practice regular breast self-examination, to believe that their physician treats them with respect, and to be satisfied with the way their breasts were examined by health professionals. They were also more likely to exercise regularly and to control their dietary intake of salt, fats, and sugar. Women who underwent mammography also were substantially more knowledgeable about breast cancer
41
classified
analysis of 78:22
with
assumes a for having
undergone mammography, the actual proportion in the sample. With a more nationally representative assumption
of equal
probability
(50:50),
the same model enables correct classification of 72% of the women who already had undergone mammography and improves cation of women
gone
the who
correct classifihad not under-
mammography
to 81%.
mammography. Table
4 lists
often
the
influenced
to undergo
factors
that
a woman’s
most
DISCUSSION
decision
mammography.
The
To maximize
most
the
important factor was a physician’s recommendation. For women who had not undergone mammography, the three most important reasons for
phy
by American
gain
insight
not undergoing the examination were (in decreasing order) physician said she did not need one, inconvenience, and expense. This group of women more often believed that there was little they could do to reduce their chances of dying of breast cancer. A total of 422 respondents provided
information
information
on the
variables
in the discriminant analysis. 422 subjects, 91 (21.6%) had
undergone
or
other
one
women who mammography.
Women who had gone mammography
used these
521 age-eligible
and
at least
pendix.
based on age, of well-being,
examination. individual variables
differed for not undergo
complete RESULTS
of whites
undergone
mammographic Table 3 lists that did
40-49
with 76 of 115 (66%) aged and 36 of 84 (43%) younger age 40 years. Approximately
mammography
(78.4%) had. The five were most successful ating between women
and
Of not
331
variables that in differentiwho had un-
dergone mammography who had not are detailed
and those in the Ap-
tudes sions
use
of mammogra-
women,
into
the
we
factors
must
and
that influence women’s to undergo mammography.
could
then
attideciThis
be used
by
physicians, local breast cancer screening programs, and national campaigns to target specific populations of women who have thus far been resistent to breast cancer screening.
Our results corroborate and extend previous findings that physicians’ recommendations are crucial in influencing patients’ beliefs about and behavior toward mammography (5,7-9). As these investigations also found underutilization of mammography by physicians, it seems evident that an
important tients
educate
first to undergo
their
step
in motivating
pa-
mammography
physicians
about
is to
the July
1992
Table
2
Survey
Instrument
Scale
Survey
Variable
Us. cancer rates
Knowledge
In American
women,
Item
(a) is one of the cancers that causes death but not (c) rarely is a cause of
cancer
breast
most often causes death, (b)
commonly of other cancers,
as often as a number death Breast cancer most often affects women (a) younger than age 35 years, (b) ages 35-50 years, (c) older than age 50 years If your mother or sister had breast cancer (a) you are much more lilcely to get breast cancer yourself, (b) you are only a little more likely to get breast cancer yourself, (c) you are no more likely to get breast cancer than anyone else Breast cancer (a) can be cured in most women if it is found early, (b) is usually incurable, (c) is always fatal The method that is best able to find small breast cancers is (a) self-examination of the breasts, (b) a physician’s examinalion of the breasts, (c) mammography (x-ray of the breast) A woman should undergo mammography (a) when she thinks something is wrong with herbreasts, (b) when her doctor recommends it, (c) on a regular basis Do you have a physician that you see regularly? How many times during the last 3 months have you practiced breast self-examination? Do you currently smoke cigarettes? In an average week, how often do you exercise vigorously for 20 minutes or more? To what extent do you try to controlyour intake of salt, fats, and
Age of incidence Risk
Fatality of breast cancer
Detection
Recommended
Health
of mammography
frequency
Physician Breast self-examination
behavior
Smoking
habib
Exercise
Diet
sugar?
Health
concerns
Breast
concern
How many
times during
personnel General
the past year did you consult
for an illness or concern
medical
about your breasts?
the past month, how much has your health worried or you? My medical insurance coverage should pay for more expenses During
concern
concerned
Insurance
than
Health worry Breast cancer worry Cost of mammography Breast
well-being
General
cancer
General
health
Current
health
it does
I worry about my health a great deal I never worry about getting breast cancer I would undergo mammography more often if my insurance paid for it How likely do you think it is that you will get breast cancer in the future? In general, would you say your health is excellent, good, fair, or
risk
poor? Satisfaction
medical
cancer
ofbreast
If I eat properly and exercise frequently, I will feel healthier and live longer There is little I can do to reduce my chance of dying of breast
Satisfaction
with
care
Medical
examinations
Respect Physician
communication associated
Unpleasantness
Locus
I am not as healthy now as I used to be I am very satisfied with the medical care I receive The medical personnel who have examined my breasts have respected my feelings My doctor always treats me with respect My doctor causes me to worry alot because he/she does not explain medical problems to me The unpleasant way mammography is performed outweighs the benefits of having it done If I take care of myself, I can reduce my chances of getting breast
of control
Reduce
chances of getting
Taking
care of myself
Reduce
chances
with
mammography
cancer of dying
Personal
relationships
cancer Having breast
Personal
history
Have you ever had breast cancer?
Demographics
Family Cancer
history
Has
death
Has a relative
Note.-The complete survey instrument and scoring algOrithm Likert-type scale, with I beinglowest and 5 being highest.
advantages
of the procedure, thus them to encourage screening to patients. Communication between radiologists and referring physicians, as well as radiologist involvement in community breast can-
motivating
cer awareness
programs,
can
physicians with information merits of mammography Volume
184 #{149} Number
I
cancer
provide
about the and the im-
portance decisions Beyond vides
new
are available
from
your
cancer
ruin
or sister had
mother
or dose
friend
the authors on request. Responses
of their influence on women’s to undergo mammography. this, our research also proinformation
would
concerning
the
importance of more global aspects of the patient-physician relationship in promoting the likelihood that patients will seek breast cancer screening. Our study found that women who under-
my relationship
with men
breast cancer? died of breast cancer?
to some questions
were recorded
on a
went mammography had a high level of satisfaction with their general medical care. A related finding is that women who underwent mammography
portrayed
more
positive
health-
related behaviors. In contrast, women who did not undergo mammography were less likely to exercise, watch their diets, or adopt other preventive Radiology
#{149} 61
health
measures
and
believed
that
there was little they could do to reduce their chances of dying of breast cancer. It is possible that undergoing mammography makes women believe they are more in control of their lives. This points to the concept of locus of control as an important avenue for exploration in research on promotion of breast cancer screening. A number
of variables
cited
in pre-
Table
3 between Health
Differentiating Mammography
Health
When
variables such as education and income and the inconvenience and p05sible discomfort associated with mammography. In addition, concern or worry about getting breast cancer was not
an
important
factor
in determin-
ing whether a woman had previously undergone mammography. This suggests that campaigns to encourage breast cancer screening that target women’s fear of breast cancer may be less
effective
making
than
women
in control
those
believe
of their
aimed
they
at
(% responding yes) doctor breast self-examination regularly
81 82 71
diet
72
17 20 58 61