This material is protected by U.S. copywright law. Unauthorized reproduction is prohibited.
To purchase reprints or request permission to reproduce, e-mail reprints@ons.org.
Older Women and Breast Cancer Screening:
Research Synthesis
Suzanne S. Yarbrough, PhD, RN
Key Points . . .
Purpose/Objectives: To identify the most beneficial ways to support
older women as they make screening decisions using a systematic, epi-
demiologic, narrative review of research regarding benefits and burdens
Older women, especially those over age 74, are underrepre-
of breast cancer screening and treatment.
sented in healthcare research.
Data Sources: Medical and nursing research databases emphasizing
Little is known about the effects of aging on breast cancer
women aged 60 and older.
Data Synthesis: Older women can tolerate screening and treatment,
screening and treatment patterns.
yet they are underserved. The most frequently cited reason to explain
Until stronger research evidence supports public health rec-
this phenomenon is declining health status associated with aging. Re-
ommendations for screening in older women, decisions will
search evidence does not support this claim. No evidence clearly de-
have to be based on individual needs.
scribes relationships among health status, aging, and less screening or
less aggressive treatment.
Conclusions: Older women experience varied health problems. How-
ever, indications that they are less able than their younger counterparts
of chronic comorbid diseases and associated disabilities is
to tolerate screening or treatment for breast cancer do not exist.
indeed high in older populations (Kimmick & Muss). How-
Implications for Nursing: Further research in all aspects of breast
cancer care in older women is required to define and describe risks and
ever, aging may not necessarily guarantee disease and disabil-
benefits of screening within a context of aging and changing health.
ity. Life expectancy is increasing, and the prospect of healthy
Nurses should discuss the risks and benefits of screening with older
aging is improving (National Women's Law Center, 2000;
Rowe & Kahn, 1998; Smith et al., 2003). Women aged 65
may live for an additional 2030 years. Although some
women may experience multiple health problems and disabili-
reast cancer is a leading cause of death in older
ties associated with the aging process, others may be very
women, exceeded only by coronary artery disease and
healthy and active well beyond the age of 65 (Cimprich, 1999;
lung cancer. The majority of the 211,300 new cases of
Rowe & Kahn).
breast cancer in 2003 occurred in women over age 65 (Smith
Breast cancer screening, which includes annual or biannual
et al., 2003). The incidence of breast cancer is four times
mammography for patients aged 4074, annual clinical breast
g r e a t e r in older women compared to younger women
examinations (CBEs), and monthly breast self-examination, is
(Kimmick & Muss, 1997; Smith et al.). Death rates are more
considered the only universally applicable cancer prevention or
than doubled for women aged 85 (202 deaths per 100,000)
control option for breast cancer (U.S. Preventive Services Task
compared to women aged 6574 (98.8 per 100,000) (National
Force [USPSTF], 2002). The ultimate benefit of breast cancer
Women's Law Center, 2000; Smith et al.). Older women bear
screening--reduced mortality--has not been examined in
the greatest burden from breast cancer, yet they are under-
women older than age 74 (Olsen & Gotzsche, 2001; Smith et
served and understudied (Burns et al., 1996; Devesa & Hun-
al., 2003; USPSTF). Therefore, public health recommendations
ter, 2000; Kimmick & Muss).
Aging of the baby-boomer generation, resulting in a dra-
matic rise in the proportion of the population that is elderly,
At the time this article was written, Suzanne S. Yarbrough, PhD, RN,
increases the importance of identifying means to promote
was an assistant professor in the Department of Chronic Nursing
Care in the School of Nursing at the University of Texas Health Sci-
healthy aging and reduce the negative impact of diseases such
ence Center in San Antonio. She currently is an associate dean for
as breast cancer (Balducci, 2000; Kimmick & Muss, 1997).
nursing at Excelsior College in Albany, NY. (Submitted June 2002.
Factors that influence the way that services are provided for
Accepted for publication July 1, 2003.)
older women appear to be predicated on assumptions that
aging, disease, and disability are synonymous. The incidence
Digital Object Identifier: 10.1188/04.ONF.E9-E15
regarding breast cancer screening over age 74 are variable and
Studies were excluded if they were not data based or did not
may affect the way that these services are used to promote
address relationships that might influence screening decisions
healthy aging.
(see Table 1).
Kimmick and Muss (1997) argued that the maximum age
Systematic epidemiologic narrative reviews are used to
limit for screening should be increased from age 74 to age 85
analyze the credibility of causal inferences made in the litera-
because, in part, women may live at least 10 years beyond the
ture. The review, based on criteria specified by Hill (1965),
current cutoff age. According to the work of several research-
evaluated the strength of relationships, consistency of the
ers, screening would provide an important health benefit to
evidence, dose response between the causal agent (screening
older women (Ferrini, Mannino, Ramsdell, & Hill, 2000;
or screening through treatment) and the outcome (reduced
Kimmick & Muss; Marwill, Freund, & Barry, 1996). Screen-
mortality), evidence of a temporal relationship between cause
ing decisions should be based on analysis of screening bur-
and effect variables, and strength of the research design used
dens weighed against each individual's physical and emotional
in each study (Weed, 2000).
ability to tolerate screening and/or treatment if diagnosed with
breast cancer and the potential effect on quality of life, not just
mortality (Ferrini et al.; Kimmick & Muss; Marwill et al.;
Although two studies predicted that mortality rates will
Smith et al., 2003). The purpose of this article is to review
decline by 26% (Kerlikowske, Grady, Rubin, Sandrock, &
research regarding the benefits and burdens of breast cancer
Ernster, 1995) or even 30% (Brogdon, 1998) in women aged
screening and treatment that result from screening in older
40 74, no data-based evidence exists predicting the lifesav-
women (i.e., aged 65 and older). The goal of this literature re-
ing benefit of screening in women over age 74 (Olsen &
view is to identify the most beneficial ways to support older
Gotzsche 2001; USPSTF, 2002).
women as they make screening decisions in the absence of a
One group of researchers used decisional analysis model-
unified public health recommendation.
ing to estimate that a moderate fiscal burden would be asso-
ciated with biennial mammography in women aged 6579
Methods and Design
(Kerlikowske, Salzmann, Phillips, Cauley, & Cummings,
1999). The cost of this screening was lower, $66,000 per
This systematic, epidemiologic, narrative review (Weed,
year of life saved, if women who were identified as having
2000) was undertaken to (a) describe current research that
lower risk for breast cancer based on bone mineral density
addresses relationships between screening practices and out-
were not included in biannual screening. However, if all
comes among older women (aged 65 and older), particularly
women aged 6579 were provided biannual mammograms,
those older than age 74; (b) describe factors that have been as-
the cost rose to nearly $118,000 per year of life saved. A
sociated with mortality or quality-of-life outcomes related to
confounding issue identified by Kerlikowske et al. (1999) in
screening in older women; and (c) compare and contrast the
this cost-benefit analysis was that the minimal benefit in
strength of evidence supporting or refuting screening recom-
terms of a small increase in life expectancy would be out-
mendations for older women. These goals should answer the
weighed by the physical burden and cost of treating ductal
following question: Does sufficient research evidence exist
carcinoma in situ.
supporting or refuting causal inference regarding benefits of
screening in older women (aged 65 and older and especially
Screening and Health Status
those older than age 74) for breast cancer that can be trans-
Burdens associated with screening or treatment for breast
lated into just and beneficial public health recommendations?
cancer were the focus of most studies in this review. Five
The review was begun by searching MEDLINE (medical)
studies were found in which older women's declining capa-
and CINAHL (nursing and allied health) research databases
bilities for living associated with aging (health status) and a
for breast cancer research reports of studies involving women
relationship to screening or breast cancer mortality were de-
aged 65 and older published from 1990, which allowed time
scribed (see Table 2). Different measures were used in each
for the effect of Medicare coverage for mammography to be
study, and data were obtained from preexisting databases or
represented in the literature, to 2001. Further literature was
convenience samples of women or data sets. Therefore, con-
identified through analysis of references cited in those stud-
sistently comparing studies, a criterion for making causal in-
ies. Search terms were older women, breast cancer, breast can-
ferences (Weed, 2000), is difficult. The underlying assump-
cer screening, epidemiology, and treatment.
tion of the identified studies appears to be that aging itself
A total of 109 references were identified that addressed
leads to increased incidence of disease and subsequent
breast cancer in older women. Of those, 46 studies either com-
death; therefore, no beneficial improvement in life expect-
pared older and younger women or described older women's
ancy would be derived from screening for cancer among
responses to varied breast cancer treatments or screening. The
older women.
15 reports that provided the data for this review focused on
In most studies, the presence of disease was conceptualized
women older than age 65 and included women older than age
and operationalized as health status. The Charlson Index, a
74 in sampling. Research reports were included in the review
measure that predicts life expectancy, was used to describe
if the study's purpose was to explain or predict relationships
presence and burden of competing diseases in one study
between contextual factors that could mediate or moderate life
(Kiefe, Funkhouser, Fouad, & May, 1998). In another study
expectancy outcomes for older women and women's abilities
(Satariano & Ragland, 1994), a comorbidity index that was
to tolerate breast cancer screening. In some studies, women's
not correlated to life expectancy was constructed based on dis-
ability to tolerate treatment was identified as an intermediate
eases identified when women entered the study. In two stud-
end point for life expectancy; therefore, the inclusion of treat-
ies, women were asked to provide a rating of their current
ment and treatment outcomes in the review was justified.
Table 1. Summary of Reviewed Research
Data Sources
Study Design
Target Population
Face-to-face inter-
Retrospective cohort
Blustein &
Non-HMO Medicare recipi-
Age, general health rating, ADL perfor-
Weiss, 1998
ents, aged 75 or older, with
mance, medical history, and mam-
no history of breast cancer
mography history over past two years
1990 HCFA Part A and
Cohort and descriptive
3 million or
Burns et al.,
Women aged 65 or older
Age, race, income quintile, visits to
B claims from 10
primary care provider, and receipt of
states and 1990 U.S.
census data by zip
Telephone interviews
C r o s s - s e c t i o n a l , de-
Aged 5090 ( X = 65) and
Chang et al.,
Caregiver's perceptions of facilitators
scriptive, and correla-
a n d a convenience
caregiver for an ill family
and barriers, health beliefs and care-
giver burdens, and screening partici-
S t a n d a r d i z e d vali-
Cross-sectional and de-
N e w l y diagnosed with
Physical, cognitive, and affective dis-
dated instruments
stage I or II disease, in pre-
tress after diagnosis but before treat-
a n d a convenience
treatment, and aged 2579
ment stratified by age
Kentucky cancer reg-
Cross-sectional cohort
Aged 67 or older
Fleming et al.,
Effect of age, comorbidity burden,
istry data merged with
and breast cancer stage on probabil-
Medicare claims
ity of death
Meta-analysis: case con-
Aged 4074
Kerlikowske et
Mammography screening and mor-
trol and prospective co-
trolled trials
al., 1995
horts; 10-year follow-up
Retrospective, descrip-
Medical records
Clinic records for women
Kiefe et al.,
Presence of comorbid diseases
tive, and correlational
aged 43 or older (focus
(Charlson Index), cancer screening
ages are 5074)
(CBE, mammogram, and Pap smear)
Mailed questionnaires
Random sample, cross-
Internists, obstetricians or
Marwill et al.,
Screening practices, agreement with
sectional, and descrip-
that included case-
gynecologists, family or
ACS screening guidelines, and mam-
study vignettes
general practice physi-
mography use by patient characteris-
cians, and geriatricians
Personal interviews
D e s c r i p t i v e random
Women aged 60 or older
Barriers to and lack of knowledge
Michielutte et
and valid, reliable in-
sample of clinic attend-
with no history of breast
about screening
al., 1999
Cross-sectional survey
S t a n d a r d i z e d vali-
2 6 7  (35%
Aged 3088
Quality of life and physical and men-
Pusic et al.,
comparing two cohorts
dated quality-of-life
54% response
tal health by three surgical interven-
q u e s t i o n n a i r e and
from two hos-
random sampling
SEER registry data
Cross-sectional and de-
21,972 at diag-
Medicare eligible, aged 65
Treatment patterns by HMO or fee for
Riley et al.,
nosis; 28,608
or older, newly diagnosed
service by age, race, area of resi-
to analyze mo-
with early-stage breast
dence, cancer history, year of diag-
nosis, and education
Survival analysis fol-
Personal interviews
Aged 4084 identified in
P r e s e n c e of comorbid diseases,
Satariano &
two to four months
lowing two cohorts for
one of nine SEER registries
physical functioning, health prac-
three years after diag-
after diagnosis, medi-
t i c e s , social and economic re-
c a l record review,
n o s i s with invasive
sources; breast cancer stage at di-
breast cancer
and physician inter-
agnosis; age; and treatment status
Medical records
Cross-sectional chart
Aged 65 or older with
Eligibility for BCS, mammography
Solin et al.,
newly diagnosed invasive
history, tumor staging, and treatment
breast cancer
Medical record review
Cross-sectional and de-
Aged 8089 with stage T1
Effect of intervention on length of
Susann et al.,
or T2 tumors
follow-up, adjuvant therapy, recur-
rence, disease-free interval, length of
survival, and cause of death
Age stratified, random
Cross-sectional chart
Postmenopausal patients
Yancik et al.,
Comorbidity burden by age, tumor
sample of SEER tu-
with breast cancer (aged
stage, initial surgical treatment, and
mor registry data and
survival 30 months after diagnosis
medical charts
ACS--American Cancer Society; ADL--activities of daily living; BCS--breast-conserving surgery; CBE--clinical breast examination; HCFA--Health Care Financing
Administration; HMO--health maintenance organization; NA--not available; SEER--Surveillance, Epidemiology, and End Results
Table 2. Health Status and Screening Behaviors in Older Women
Data Sources
Participant Ages
Health Indicator
Functional Status
Longitudinal national
47% were aged 75
Excellent (41%), good
Blustein &
Limitations on ac-
ORs for mammogram (p < 0.05): oldest
79, 30% were aged
(29%), or fair or poor
tivities of daily liv-
age (85+) 0.26; poorest health 0.44,
8 0 8 4 , and 23%
(30%) health
ing 50%
crude 0.41 age adjusted; limitations on
w e r e 85 years or
Hypertension (55%)
activities of daily living 0.57 crude, 0.71
Diabetes mellitus (14%)
age adjusted; stroke 0.58 crude, 0.63
age adjusted; hip fracture 0.46 crude,
0.60 age adjusted; Alzheimer disease
0.54 crude, 0.55 age adjusted
Review of HCFA data
55% were aged 65
Primary care visits per year:
Mammograms/visit (p < 0.01): none
Burns et al.,
Not reported
74, 34% were aged
none (32%), one (10%),
N > 3,000,000
(6%), one (15%), two (20%), three or
7 5 8 4 , and 11%
two (9%), three or more
more (22%)
w e r e 85 years or
M e a n age was 65
Providing care to spouse or
Chang et
Small (N = 52) con-
Not reported
Mammograms in last year (81%)
y e a r s , range was
al., 2001
venience sample of
Burden of care to BSE (r = 0.33, p <
5 0 9 0 years, and
0.05) to CBE (r = 0.28, p < 0.05)
56% were younger
Referral positively associated with
than age 65.
M e a n age was 62
Charlson Index: X = 1.24
Retrospective chart
Kiefe et al.,
Mammography and age not significant
years, and the range
review of records in
(range = 08)
Decreased mammography and CBE and
was 43100.
university clinic
Number of clinic visits per
increased Charlson Index; mammog-
year and clinic type
raphy, disease, and age (p < 0.05)
Congestive heart failure and age 75+
OR = 0.39
Angina and age 5074 OR = 0.40
Arthritis and age 75+ OR = 1.90
Gastrointestinal bleeding and age 65
74 OR = 0.72
Ambulatory and mammography OR =
24% were aged 60
Mammogram and age
Cross-sectional in-
Treated for chronic illness
Not reported
64, 53% were aged
6064 OR = 2.76
et al., 1999
terviews and ran-
6574 OR = 1.99
domly sampled clinic
6 5 7 4 , and 23%
Excellent (14%), very good
w e r e 75 years or
(29%), good (33%), or
Fair or poor OR = 1.00
older. (N = 719)
fair or poor (25%) health
Good or excellent OR = 1.81
Annual examination (85%)
Annual examination OR = 2.31
BSE--breast self-examination; CBE--clinical breast examination; HCFA--Health Care Financing Administration; OR--odds ratio
health (Blustein & Weiss, 1998; Michielutte, Dignan, &
and mammography. That relationship was moderated by older
Smith, 1999). Health was assessed indirectly in the remaining
age (OR = 0.41, age 75 or older, p < 0.05). The OR predict-
two studies as either visits to a primary healthcare provider
ing mammography for women with CHF aged 5074 was
(Burns et al., 1996) or the burden of having to care for a
higher (0.84), but when the sample was restricted to younger
chronically ill family member (Chang, Sarna, & Carter, 2001).
women aged 5064, the OR was nearly equivalent to the old-
Functional capacity, the ability to manage activities of daily
est group (OR = 0.50). The other significant health status pre-
living (ADL) (Blustein & Weiss) or ambulate (Kiefe et al.),
dictor of less mammography in that study occurred in younger
was used as another indirect measure of health status associ-
women. The presence of angina predicted less mammography
ated with health ratings.
for women aged 5074, and gastric bleeding predicted less
In their study, Kiefe et al. (1998) attempted to correlate life
screening in women aged 6574.
expectancy with health status and screening and found that
Mammography and CBE both declined as the number or
lower life expectancy was related to declining screening partici-
burden of chronic illnesses increased, which did not necessar-
pation for women of all ages. However, odds ratios (ORs) for
ily coincide with increasing age, indicating that health status
participating in screening activities changed based on specifi-
unrelated to increasing age is associated with less screening
cally identified diseases without a clear relationship to aging. A
(Kiefe et al., 1998). This study employed a cross-sectional ret-
consistent and direct relationship would have to be demonstrated
rospective chart review to make predictions, the least robust
to support an argument that aging is correlated to declining
study design for making causal inferences (Weed, 2000). There-
health and lowered expectations for longevity (Weed, 2000).
fore, a direct relationship between screening behavior and life
Kiefe et al. (1998) identified a statistically significant, yet
expectancy or, more specifically, an interaction among health
moderate, relationship between congestive heart failure (CHF)
status, screening, and life expectancy cannot be predicted.
Blustein and Weiss (1998) concluded that mammography
Women older than age 75 who had osteoarthritis, which
is not being used appropriately in older women based on cor-
may have less potential for affecting life expectancy and is
relations between low screening rates and reported favorable
one of the more common diseases of aging, were almost twice
perceptions about health. The high percentage of women who
as likely to have mammograms (Kiefe et al., 1998). Others
rated their health as good and had low rates of ADL limita-
found that ADL limitations (OR = 0.41), which may be asso-
tions indicates that many women would be good candidates
ciated with osteoarthritis rather than poorest health (OR =
for screening. The evidence suggests that, although breast
0.71), predicted decreased use of mammography.
cancer risk increases with age, screening declines and health
Only 50% of women older than age 75 reported ADL limi-
status or functional capacity solely attributed to aging do not
tations in a study conducted by Blustein and Weiss (1998).
appear to be the only factors influencing declining screening
Specific conditions influencing ADL capacities were stroke,
participation. These reviewed studies do not indicate what the
hip fractures, and Alzheimer disease. No evidence of a corre-
ultimate impact of screening might be for any group of older
lation between ADL and specific illnesses was reported. More
than 70% of the women in this study reported good to excel-
lent health.
Breast Cancer Treatment Effects in Older Women
The reviewed studies indicate that women who are living
with diseases that may potentially influence life expectancy
Kerlikowske et al. (1999) argued that the personal and fi-
are participating less in breast cancer screening, but no evi-
nancial cost of treating ductal carcinoma in situ, a potentially
dence exists to support predictions regarding how those dis-
indolent form of breast cancer, would increase the burden of
eases will affect outcomes if older women are diagnosed with
screening beyond its potential benefit in older populations.
breast cancer. The relationship between illness or disease and
The apparent underlying assumption in this argument is that
perceptions of health, which might affect longevity, is not
older women are less able than their younger counterparts to
clear. Many older women rate their health as good, which
tolerate treatment and should avoid early-detection strategies.
may indicate that they would want to pursue treatment for
The interaction between comorbid diseases and treatment for
early-stage cancer if that would allow them some positive
breast cancer is indeed a consideration for evaluating public
health effect. In one study, 75% of women aged 60 and older
health recommendations for screening, but this should be
reported good to excellent health. In that case, 75% of the
done in relationship to the ultimate outcome, mortality from
sample also reported being treated for a chronic illness. Good
breast cancer (Johnson, 2000; Kimmick, 2000).
to excellent health, which was not correlated with presence
Satariano and Ragland (1994) reported that age and the
or absence of chronic illnesses, predicted almost double par-
existence of comorbid conditions were related to either treat-
ticipation in mammography (OR = 1.81) (Michielutte et al.,
ment trends or mortality outcomes. In this study, older
1999). Therefore, the specific association among age, health
women received less aggressive treatment (both surgical
status, and screening participation was neither clearly de-
treatment and postoperative adjuvant therapy) than their
scribed nor predicted in the studies reviewed. In addition, no
younger counterparts. However, factors that may have con-
evidence was presented that indicated that screening partici-
tributed to this trend were not described. The primary out-
pation would influence life expectancy negatively or posi-
come measure for the study was the three-year mortality rate
tively for women based on presence or absence of preexist-
(i.e., mortality was tracked in the population for three years
ing health problems.
after they entered the study). Comorbidity was linked to
In two studies, increased numbers of visits to family prac-
deaths attributed to causes other than breast cancer. How-
tice or primary care clinics, presumably for management of
ever, adjustment was made for age prior to reporting this
chronic illness, were positively associated with the use of
finding; this indicates that comorbidity was an independent
mammography but not CBE (Burns et al., 1996; Kiefe et al.,
predictor of death rather than a covariant with age. Deaths
1998). This finding implies a negative relationship exists be-
among the oldest study participants (those aged 7584) were
tween screening and poor health status as a result of more fre-
attributed less frequently to breast cancer (50% of cases)
quent contact with healthcare providers among those with
when compared to the youngest cohort (63% aged 4054).
chronic conditions. Kiefe et al. noted that screening partici-
However, Satariano and Ragland made no attempt to vali-
pation was lower in general medicine clinics compared to
date the accuracy of death certificates, a factor that could
family practice clinics. No evidence indicated whether par-
confound any conclusions drawn from this finding when
ticipation was related to physician referral patterns or factors
women in the study may have had multiple competing dis-
inherent in women (Kiefe et al.).
eases. Therefore, this study did not clarify any relationships
Chang et al. (2001) found that 81% of women who were
between aging and the impact that coexisting diseases may
acting as primary caregivers for an ill family member reported
have on breast cancer mortality.
that they had a mammogram in the past year. The women who
Six other studies were found in which breast cancer treat-
reported screening in this small convenience sample (N = 52)
ment was identified as the end point (Fleming, Rastogi,
tended to be better educated, have insurance, and report lower
Dmitrienko, & Johnson, 1999; Pusic et al., 1999; Riley,
burdens associated with caring for the ill family member.
Potosky, Klabunde, Warren, & Ballard-Barbash, 1999; Solin,
Many (85%) received a referral for mammography. Marwill
Schultz, Hanchak, & Kessler, 1999; Susann et al., 1999;
et al. (1996) found that physicians were more likely to order
Yancik et al., 2001). Secondary analysis of national databases,
mammograms for women who were living at home regardless
chart reviews, or survey questionnaires was used for data col-
of their current health status or potential life expectancy.
lection. Each study measured different treatment and outcome
However, these studies do not add to knowledge about the
variables. Therefore, causal inferences regarding relationships
benefits or burdens of screening for older women or the fac-
between screening participation and the end points measured
tors that may affect screening decisions.
in those studies are difficult to make.
A correlation was found between screening and eligibility
den of comorbid conditions did not specifically evaluate the
for less invasive breast-conserving surgery related to early-
impact of treatment. However, as women age, their risk of
stage tumors among a small sample of older women enrolled
death from all causes increases; if they are diagnosed with
in a health maintenance organization (HMO). Of the women
late-stage breast cancer, their potential for death from breast
who had undergone routine mammography, 79% were eli-
cancer is increased regardless of the presence of comorbid
gible for breast-conserving surgery based on the early histo-
health problems (Fleming et al., 1999).
logic stage of the tumor. Only 48% of those in the unscreened
Early diagnosis associated with screening may be benefi-
group were diagnosed with early-stage disease (Solin et al.,
cial. If screened and treated for early-stage breast cancer, older
1999). Similar patterns were found in another study in which
women who already are suffering from other conditions may
less aggressive therapy was associated with increasing age,
be spared the horrors of dying from breast cancer.
whereas presence of comorbid diseases did not significantly
The findings regarding screening practices among older
affect treatment patterns (Yancik et al., 2001).
women and life expectancy or quality-of-life outcomes are
A higher incidence of late-stage diagnosis was found to
equivocal. Very little research has been conducted with older
occur in fee-for-service clinics (11%) compared to HMOs
women. Factors associated with aging that may affect mortal-
(8%) where greater emphasis is placed on health-promotion
ity or quality-of-life outcomes relative to screening have not
programs (Riley et al., 1999). These findings provide more
been defined or described in a way that facilitates develop-
evidence indicating that screening might reduce mortality or
ment of public health recommendations for older women.
treatment burden to older women by decreasing the percent-
Most studies identified for this review used cross-sectional or
age of late-stage diagnosis. However, no direct evidence of a
retrospective designs, which are the weakest epidemiologic
correlation between screening and cancer stage exists.
designs for making causal inference. Furthermore, the predic-
Women older than age 67 and newly diagnosed with breast
tive power of these studies was limited by use of varied inter-
cancer had no (37%) or only one (32%) comorbid condition.
mediate end points rather than breast cancer mortality.
Those with the highest burden associated with comorbid condi-
Further research in all aspects of breast cancer in older
tions were two times more likely to die within one year follow-
women is required to define and describe the risks and ben-
ing diagnosis of breast cancer. As disease burden increased, one-
efits of screening within a context of aging and changing
year survival decreased. However, worsening cancer stage also
health before scientifically based public health recommenda-
affected one-year survival. In addition, patients with early-stage
tions can be made. All epidemiologic studies addressing
breast cancer aged 85 or older demonstrated similar one-year
breast cancer screening must include representative samples
survival to that predicted for younger populations (Fleming et
of women older than age 74.
al., 1999). These results indicate that early-stage disease is sur-
Although no evidence exists that screening may affect over-
vivable in even the oldest women who are relatively healthy.
all life expectancy, evidence confirms that health status is
Little difference was found among groups of older women
heterogeneous in this population and that many older women
(aged 8089) diagnosed with early-stage (T1 or T2) tumors
could tolerate screening and treatment. In view of the natural
comparing effect of minimal and more aggressive treatment.
decline in life expectancy occurring simultaneously with ag-
About 10% of each group died from breast cancer, but most
ing, quality of life should be an end point evaluated in breast
of these octogenarians survived the entire 10-year duration of
cancer studies with older women, most especially when evalu-
the study (Susann et al., 1999). Existence of comorbid condi-
ating the cost and benefits of screening or treatment. The costs
tions that would allow for more specific comparison to other
(fiscal and quality of life) of treatment for late-stage cancer
populations or analysis of the interaction of general health and
also should be factored into these studies (Balducci, 2000).
breast cancer outcomes was not provided.
More robust research designs will have to be employed if sci-
Two studies addressed quality of life; however, the re-
entists are to provide evidence indicating a relationship be-
searchers used age 55 rather than 65 as the cutoff point for
tween screening and any health outcome associated with
defining older age. Women younger than age 55 who under-
breast cancer (Weed, 2000).
went mastectomy without reconstruction experienced greater
Nurses are in a position to interact with older women in mul-
illness intrusiveness and poorer quality of life than women age
tiple settings outside of medical offices. This offers nurses a
55 and older (Pusic et al., 1999). Cimprich (1999) found that
unique opportunity to talk to older women about breast cancer
women aged 55 and older who were awaiting treatment for
and to support them through a process of making independent
newly diagnosed cancer were more resilient than younger
decisions about their personal health and the potential benefits
women. The older women were more able to tolerate diagnos-
or burdens of breast cancer screening. Rowe and Kahn (1998)
tic procedures and treatment decisions. They reported less dis-
predicted that, in general, most older people successfully will
tress, were more able to sleep, and experienced fewer mood
manage the aging process, with fewer physical impairments and
disturbances, fewer problems with fatigue, and less difficulty
greater connections to active life. Many older people are not
concentrating than younger women with comparable physical
frail; they are functionally and cognitively able to tolerate
conditions and similar diagnoses.
screening and treatment for early-stage breast cancer (Rowe &
Kahn). Nothing was found in the literature review to refute this
argument; therefore, no evidence contraindicates strong nursing
Discussion and Conclusions
interventions to support independent screening decisions in a
The data from this review indicate a pattern of less treat-
population of women where the burden from breast cancer is
ment for older women without actual correlation with the
demonstrably higher than in younger populations.
presence of preexisting diseases or disabilities. Furthermore,
Evidence used to make generalizations about a single best
the data indicate that older women tolerate treatment at least
practice approach to care for older women is insufficient. Rec-
as well as younger women. One study that addressed the bur-
ommending universal annual screening for women aged 75
79 or older is premature until further research clearly demon-
treatment should be ensured through shared, informed, au-
strates relationships among all contextual factors, such as the
tonomous decision making.
presence of comorbid diseases, individual ability to tolerate
screening or treatments, and the impact of screening or treat-
Author Contact: Suzanne S. Yarbrough, PhD, RN, can be reached
ment on quality of life and the ultimate end point, mortality.
at syarbrough@excelsior.edu, with copy to editor at rose_mary
Until then, each woman's right to participate in screening or
Balducci, L. (2000). Geriatric oncology: Challenges for the new century.
A national and state-by-state report card. Retrieved March 5, 2001, from
European Journal of Cancer, 36, 17411754.
Blustein, J., & Weiss, L.J. (1998). The use of mammography by women aged
Olsen, O., & Gotzsche, P.C. (2001). Cochrane review on screening for breast
75 and older: Factors related to health, functioning, and age. Journal of the
cancer with mammography. Lancet, 358, 13401342.
American Geriatric Society, 46, 941946.
Pusic, A., Thompson, T.A., Kerrigan, C.L., Sargeant, R., Slezak, S., Chang,
Brogdon, C.F. (1998). Women and cancer. Journal of IV Nursing, 21, 344
B.W., et al. (1999). Surgical options for early-stage breast cancer: Factors
associated with patient choice and postoperative quality of life. Plastic and
Burns, R.B., McCarthy, E.P., Freund, K.M., Marwill, S.L., Shwartz, M., Ash,
Reconstructive Surgery, 104, 13251333.
A., et al. (1996). Variability in mammography use among older women.
Riley, G.F., Potosky, A.L., Klabunde, C.N., Warren, J.L., & Ballard-Barbash,
Journal of the American Geriatric Society, 44, 922926.
R. (1999). Stage at diagnosis and treatment patterns among older women
Chang, B.L., Sarna, L., & Carter, P.A. (2001). Mammography screening be-
with breast cancer: An HMO and fee-for-service comparison. JAMA, 281,
havior in older women caregivers. Geriatric Nursing, 22(1), 3336.
Cimprich, B. (1999). Pretreatment symptom distress in women newly diag-
Rowe, J.W., & Kahn, R.L. (1998). Successful aging. New York: Pantheon
nosed with breast cancer. Cancer Nursing, 22, 185194.
Devesa, S.S., & Hunter, C.P. (2000). The burden of cancer in the elderly. In
Satariano, W.A., & Ragland, D.R. (1994). The effect of comorbidity on 3-year
C.P. Hunter, K.A. Johnson, & H. Muss (Eds.), Cancer in the elderly (pp.
survival of women with primary breast cancer. Annals of Internal Medicine,
124). New York: Marcel Dekker.
120, 104110.
Ferrini, R., Mannino, E., Ramsdell, E., & Hill, L. (2000). Screening mammog-
Smith, R.A., Saslow, D., Sawyer, K.A., Burke, W., Costanza, M.E., Evans,
raphy for breast cancer: American College of Preventive Medicine practice
W.P., et al. (2003). American Cancer Society guidelines for breast can-
policy statement. Retrieved September 18, 2000, from http://www.acpm
cer screening: Update 2003. CA: A Cancer Journal for Clinicians, 53,
Fleming, S.T., Rastogi, A., Dmitrienko, A., & Johnson, K.D. (1999). A com-
Solin, L.J., Schultz, D.J., Hanchak, N.A., & Kessler, H.B. (1999). Patterns of
prehensive prognostic index to predict survival based on multiple co-
treatment for older women with newly diagnosed breast carcinoma. Ameri-
morbidities. Medical Care, 37, 601614.
can Journal of Clinical Oncology, 22, 107113.
Hill, A.B. (1965). The environment and disease: Association or causation?
Susann, P., Susann, P.W., Ackerman, S., Gibbons, R., Hermann, K., & Kuhn,
Proceedings of the Royal Society of Medicine, 58, 295300.
P. (1999). Treatment outcomes for female octogenarians with breast can-
Johnson, K.A. (2000). Cancer prevention in the elderly. In C.P. Hunter, K.A.
cer. American Surgeon, 65, 399401.
Johnson, & H. Muss (Eds.), Cancer in the elderly (pp. 5767). New York:
U.S. Preventive Services Task Force. (2002). Recommendations and rationale:
Marcel Dekker.
Screening for breast cancer. Retrieved February 25, 2002, from http://
Kerlikowske, K., Grady, D., Rubin, S.M., Sandrock, C., & Ernster, V.L.
(1995). Efficacy of screening mammography: A meta-analysis. JAMA, 273,
Weed, D.L. (2000). Epidemiologic evidence and causal inference. Hematol-
ogy and Oncology Clinics of North America, 14, 797807.
Kerlikowske, K., Salzmann, P., Phillips, K.A., Cauley, J.A., & Cummings, S.R.
Yancik, R., Wesley, M.N., Ries, L.A.G., Havlik, R.J., Edwards, B.K., &
(1999). Continuing screening mammography in women aged 7079 years:
Yates, J.W. (2001). Effect of age and comorbidity in postmenopausal breast
Impact on life expectancy and cost-effectiveness. JAMA, 282, 21562163.
cancer patients aged 55 years and older. JAMA, 285, 885892.
Kiefe, C.I., Funkhouser, E., Fouad, M.N., & May, D.S. (1998). Chronic dis-
ease as a barrier to breast and cervical cancer screening. Journal of General
For more information . . .
Internal Medicine, 13, 357365.
Kimmick, G. (2000). Breast cancer. In C.P. Hunter, K.A. Johnson, & H. Muss
(Eds.), Cancer in the elderly (pp. 233257). New York: Marcel Dekker.
Breast Cancer Action
Kimmick, G., & Muss, H.B. (1997). Breast cancer in older women. Clinics in
Geriatric Care, 13, 265282.
National Breast Cancer Awareness Month
Marwill, S.L., Freund, K.M., & Barry, P.P. (1996). Patient factors associated
with breast cancer screening among older women. Journal of the American
Geriatric Society, 44, 12101214.
Breast Cancer Care
Michielutte, R., Dignan, M.B., & Smith, B.L. (1999). Psychosocial factors
associated with the use of breast cancer screening by women age 60 years
Links can be found using www.ons.org.
and over. Health Education and Behavior, 26, 625647.
National Women's Law Center. (2000). Making the grade on women's health: