1. Age-Related Vulnerabilities of Older Adults with Colon
Adenomas
Evidence from Project Prevent
Elizabeth C. Clipp, R.N., M.S., Ph.D.
1–4
Elizabeth H. Carver, R.N., M.S.N.
2
Kathryn I. Pollak, Ph.D.
3,5
Elaine Puleo, Ph.D.
6
Karen M. Emmons, Ph.D.
6
Jane Onken, M.D.
2
Francis A. Farraye, M.D., M.Sc.
7,8
Colleen M. McBride, Ph.D.
3,5
1
School of Nursing, Duke University Medical Cen-
ter, Durham, North Carolina.
2
Department of Medicine, Duke University Medical
Center, Durham, North Carolina.
3
Cancer Prevention, Detection, and Control Re-
search Program, Duke Comprehensive Cancer
Center, Duke University Medical Center, Durham,
North Carolina.
4
Geriatric Research Education and Clinical Center,
Veterans Administration Medical Center, Durham,
North Carolina.
5
Department of Community and Family Medicine,
Duke University Medical Center, Durham, North
Carolina.
6
Center for Community-Based Research, Dana-
Farber Cancer Institute, Harvard University School
of Public Health, Boston, Massachusetts.
7
Section of Gastroenterology, Boston Medical Cen-
ter, Boston, Massachusetts.
8
Department of Medicine, Boston University
School of Medicine, Boston, Massachusetts.
See related editorial on pages 881–2, this issue.
Supported by grants from the National Cancer
Institute (NCI RO1 CA74000-02 and NCI
CA81191), the National Institute of Nursing Re-
search (NINR 1P20NR07795-01), and the Na-
tional Institute on Aging (NIA AG11268) and by
National Grid (Westborough, MA) and Liberty
Mutual (Boston, MA).
The authors thank Maureen Lahti, M.B.B.S.,
M.P.H., for her work on data analysis and Jennifer
Chamberlain, M.B.A., M.H.A., for her assistance in
preparing the current article.
Address for reprints: Elizabeth C. Clipp, R.N., M.S.,
Ph.D., School of Nursing and Department of Med-
icine, Box 3322, Duke University Medical Center,
Durham, NC 27710; Fax: (919) 286-6823; E-mail:
clipp003@mc.duke.edu
Received May 14, 2003; revision received August
13, 2003; accepted August 29, 2003.
BACKGROUND. This report addresses the interface between cancer and aging in the
context of colorectal carcinoma (CRC), the second leading cause of cancer death in
the U.S. overall and the first leading cause among individuals age Ն 75 years.
Because polyp risk increases with age, interventions to prevent recurrent polyps
among older adults likely would reduce CRC morbidity and mortality.
METHODS. Data for this study derive from Project Prevent, a multisite, randomized
controlled trial designed to reduce behavioral risk factors for CRC among 1247
adults who underwent the removal of Ն 1 adenomatous colon polyps. Middle-aged
and older patients were compared on key cognitive-behavioral mechanisms asso-
ciated with CRC risk and established age-related factors associated with adverse
health outcomes. Relations between cognitive-behavioral mechanisms and age-
related vulnerability factors identified subgroups of older polyp patients that may
have an enhanced risk for CRC.
RESULTS. Compared with middle-aged patients, older patients were less concerned
about developing CRC, less motivated to reduce their risk, and less confident that
their behavior change efforts would succeed. As expected, they also reported more
age-related physical, social, and economic vulnerabilities, as expected. Evidence
for enhanced CRC risk was found for older patients with multiple comorbid
conditions, low social support for change, and perceptions of income inadequacy.
CONCLUSIONS. The presence of age-related vulnerability factors may enhance the
risk of CRC among older cancer patients by creating barriers to behavioral change.
Efforts to reduce the cancer burden in older populations will require attention
beyond early detection and surveillance to interventions that account for the
unique physical and psychosocial characteristics of older adults. [See editorial on
pages 000–000, this issue.] Cancer 2004;100:1085–94.
© 2004 American Cancer Society.
KEYWORDS: aging, colorectal carcinoma risk, adenomatous colon polyp, colon
adenoma, cancer control.
Advancing age increases the risk of cancer.1
The incidence of can-
cer among individuals age Ն 65 years is expected to double from
1.3 million to 2.6 million between 2000 and 2050.2
This means that
health care providers are beginning to manage what will become
unprecedented numbers of older cancer patients with excess morbid-
1085
© 2004 American Cancer Society
DOI 10.1002/cncr.20082

2. ity, social support needs, and economic disadvantage.
The comprehensive geriatric model3
suggests that the
cancer experience in older patients is affected ad-
versely by age-related physical and psychosocial vul-
nerabilities.4
Therefore, efforts to reduce the cancer
burden in the older population will require consider-
ation of the unique physical and psychosocial charac-
teristics of older adults.
Colorectal carcinoma (CRC) is the second leading
cause of cancer death in the U.S. overall and the first
leading cause among individuals age Ն 75 years.2
CRC
arises from neoplastic adenomatous polyps,5
the prev-
alence of which increases from 20% to 25% at age 50
years and to 50% percent by ages 75–80 years.6
Re-
search demonstrates that, although most CRC can be
prevented by early endoscopic resection of colon ad-
enomas,5
patients who have polyps removed have a
30% likelihood of developing recurrent polyps,7,8
and
many do not undergo additional screening.9
Because
polyp risk increases with age, interventions among
older adults to prevent recurrent polyps likely would
reduce the absolute number of CRC diagnoses among
those at highest risk for the disease.
Evidence-based recommendations to reduce the
risk of adenomatous polyps and CRC include a diet
that is low in red meat and alcohol10–17
and avoidance
of smoking.18
Research also suggests that normal body
weight should be maintained through regular exer-
cise.19
In addition, micronutrients in fruits and vege-
tables may lower risk, and it also has been demon-
strated that the folate in multivitamins also protects
from CRC.20,21
Although further studies are needed to
identify major risk factors of CRC, older patients post-
polypectomy at least should be informed that the Na-
tional Cancer Institute recommends improving diet
quality, increasing physical activity, and avoiding to-
bacco to lower overall cancer risk. Promotion of these
behaviors among elderly populations has received lit-
tle attention.22
Several theories, including the Health Belief
Model23
and the Precaution Adoption Model,24
sug-
gest that heightened vulnerability and personal accep-
tance of risk help promote successful behavioral
change. In the context of CRC prevention, this means
that interventions that build on the heightened risk
perceptions and related concerns of patients with pol-
yps may increase inclinations to adopt risk-reducing
behaviors.25
Research also suggests that interventions
should be tailored to the patient’s level of readiness to
adopt changes.26
Patients’ expectations of benefits
that result from behavior change also are important
considerations when developing interventions.27
However, among older patients, age-related phys-
ical and psychosocial factors may influence responses
to interventions. Multiple comorbid conditions, inad-
equate social support, or low financial resources may
set up barriers to positive behavioral change. They
also may affect motivation to reduce CRC risk and the
perception of benefits potentially gained from behav-
ior change. For example, fatigue or functional limita-
tions secondary to comorbid conditions may prevent
an older polyp patient from increasing levels of phys-
ical activity. Older patients who lack social support
may eat poorly because they frequently are alone and
are not inclined to prepare nutritious meals. Others
may believe they cannot afford to meet dietary rec-
ommendations on fixed incomes. Awareness of vul-
nerable subgroups would inform the development of
CRC prevention initiatives for the burgeoning popu-
lation of older Americans.
In this report, we examine how well established
physical, emotional, social, and financial age-related
vulnerabilities28
affect CRC risk perceptions and be-
haviors among older adults who have had one or more
adenomatous colon polyps removed. Given the evi-
dence that perceptions of risk motivate efforts to re-
duce risk and that age-related vulnerability factors
may affect older patients’ responses to risk-reducing
recommendations, we hypothesize that age will mod-
ulate the perception of CRC risk and, thus, the likeli-
hood of adopting behavior to reduce CRC risk.
Four specific questions were asked. First, among a
broad age range of patients with colon polyps (ages
40–75 years), we asked about the extent to which older
persons (ages 60–75 years) differed from middle-aged
individuals (ages 40–59 years) on cognitive-behavioral
factors, including actual risk, perceived risk, and worry
about colon cancer. Second, we asked about the ex-
tent to which older patients differed from middle-aged
patients in levels of readiness to reduce CRC risk
through behavioral change and in their expectations
that such changes would be beneficial. Third, com-
pared with middle-aged patients, we asked about the
extent to which older patients provided evidence of
age-related vulnerability factors that may pose addi-
tional challenges to the adoption of risk-reducing be-
havior. Finally, among the older patients, we asked
about the extent to which age-related vulnerability
factors related to CRC risk and cognitive-behavioral
mechanisms that underpin successful behavior
change.
Data to address these questions derive from
Project Prevent, a multisite, randomized intervention
trial to reduce behavioral risk factors for CRC among
patients diagnosed with adenomatous polyps (Na-
tional Cancer Institute Project RO1 CA74000-02). In
this trial, eligible patients were randomized to receive
either usual care or a multiple risk factor intervention
1086 CANCER March 1, 2004 / Volume 100 / Number 5

3. consisting of 1) a health care provider recommenda-
tion letter concerning the importance of health behav-
ior change, 2) tailored self-help materials, 3) a moti-
vational and goal-setting telephone session delivered
by a health educator, and 4) four follow-up telephone
counseling calls and progress reports. The interven-
tion aimed to reduce CRC behavioral risk factors re-
lated to dietary intake, multivitamin intake, physical
activity, smoking, and alcohol use.
MATERIALS AND METHODS
Sample
The study sample included 1247 patients, ages 40–75
years, who participated in Project Prevent, an inter-
vention trial intended to 1) increase the use of daily
multivitamins, 2) increase fruit and vegetable intake,
3) reduce red meat consumption, 4) increase physical
activity, 5) decrease alcohol use, and 6) increase smok-
ing cessation. Patients were eligible if they had adeno-
matous colon polyps removed within 4 weeks of study
recruitment, no history of CRC, capacity for informed
consent, the ability to read and speak English, tele-
phone access for the baseline survey, and physician
approval to participate in moderate physical activity.
Participants completed the interviewer-administered
baseline telephone survey and were assigned ran-
domly to either a usual care group or to receive a
multiple risk factor intervention. The intervention in-
volved telephone counseling and tailored self-help
materials that were designed to help participants re-
duce the targeted CRC risk behaviors. The analyses for
this report utilized the baseline sample of 594 middle-
aged patients (ages 40–59 years) and 653 older pa-
tients (ages 60–75 years).
Measures
Demographics.
Standard demographic measures included age, gender,
education, marital status, race/ethnicity, height,
weight, medical history, and characteristics of the
household.
CRC risk factors
Servings of fruits/vegetables and red meat were as-
sessed using an abbreviated form of the Food Fre-
quency Questionnaire.29
A single item assessed aver-
age weekly multivitamin use (9 response options ranged
from never to 7 days per week). Alcohol consumption
was measured using the Quantity Frequency Index.30
A
modified version of the Community Healthy Activities
Model Program for Seniors (CHAMPS) Activities Ques-
tionnaire for Older Adults31,32
indexed physical activity.
Patients’ smoking status was measured using standard-
ized questions regarding lifetime and current smoking,
intensity, quit attempts, and nicotine dependence.33
Based on these measures, “risk” status was conferred for
each risk factor based on Ͻ 5 servings of fruits and
vegetables per day, consumption of Ͼ 3 red meat serv-
ings per week, taking a multivitamin Ͻ 7 days per week,
consuming Ͼ 1 (women) or Ͼ 2 (men) servings of alco-
hol per day, status as a current smoker, or participating
Ͻ 150 minutes per week in moderate exercise (1, pres-
ence of risk factor; 0, absence of risk factor). Individual
risk factor scores were summed to yield a categoric mul-
tiple risk factor score ranging from 0 (no risk factors) to
6 (all risk factors).34
Cognitive-behavioral mechanisms
Perceived risk was measured by asking participants
how likely they were to get CRC in their lifetime (on
a 5-point Likert scale that was reduced to 3 catego-
ries: unlikely, 50:50 or not sure, or likely). Patients
also were asked about their level of worry/concern
about developing CRC in their lifetime (0–10 scale).
Readiness to change was based on an individual’s
readiness to change all of their risk factors in the
coming 6 months. Participants with no risk factors
were classified in the “maintenance stage.” Those
who were unaware that they had risk factors were
not classified as “ready to change,” regardless of
their response to the question. For those who indi-
cated that they had habits to change, outcome ex-
pectancies were indicated by agreement with the
statement “changing my health habits will reduce
my risk of colon cancer.” Finally, self-efficacy was
indexed by asking patients who had identified risk
factors to rate their confidence in changing all prob-
lem behaviors within the next 6 months (on a
5-point Likert scale, from not at all confident to
extremely confident). This variable is treated as a
continuous variable in the tables.
Age-related vulnerability factors
Physical vulnerability factors included higher levels
comorbid illness and lower levels of perceived health.
Comorbid illness was measured with a revised version
of the Older American Resources and Services (OARS)
questionnaire,35
on which respondents indicated the
presence or absence of major chronic medical condi-
tions (on a 0–10 scale; e.g., stroke, cancer, diabetes,
heart disease, or lung disease). Patients with two or
more comorbid illnesses were considered vulnerable.
Self-rated health was indexed on a 4-point scale (an-
chors: excellent, good, fair, poor).36
It has been found
that this measure is a significant predictor of mortal-
ity37
; study participants with ratings of “fair” or ”poor“
were considered vulnerable.
Patient Age and Colon Adenoma/Clipp et al. 1087

4. Emotional vulnerability factors included negative
affect and lower levels of life quality. Negative affect
was measured by patient reports of the frequency of
feeling downhearted and blue during the past month
(4-point Likert scale with anchors ranging from rarely,
to no time, to most/all of the time). Patients were
considered vulnerable with ratings of “some of the
time” or more often. Quality of life was measured by
asking patients to rate the overall quality of their life
on a 4-point scale (with anchors ranging from excel-
lent to poor). Patients were considered vulnerable
with ratings of “fair” or “poor.”
Social vulnerability factors included social isola-
tion and low levels of social support. Social isolation
was determined by asking patients to report whether
they were married/cohabitating, living with others, or
living alone. Those living alone were considered so-
cially vulnerable. Social support was measured by ask-
ing participants to report the number of confidants
(“of all the people you know, how many do you feel
particularly close to?”). Patients who reported one
confidant or none were considered vulnerable. Social
support for change was captured by patients’ reports
of the extent to which friends and family would sup-
port their efforts to change their health habits (5-point
Likert scale ranging from not at all to extremely).
Those considered vulnerable reported that support for
change would be “a little” or “not at all.”
Financial vulnerability factors included low levels
of objective income and patients’ perceptions that
their income was inadequate for their needs. Annual
household income was indexed by total yearly house-
hold income (categories ranging from Ͻ $15,000 to
$45,001). Patients who reported incomes Ͻ $30,000
were considered vulnerable. Perceived income ade-
quacy was measured by asking patients to consider
their overall income and endorse one of the following:
1) have money for special things, 2) have money for
bills but not extras, 3) must cut back to make bills, 4)
have difficulty paying bills. Patients who endorsed 2, 3,
or 4 were considered financially vulnerable.
Statistical Analyses
Analyses focused on age differences in cognitive-be-
havioral factors (perceived risk, worry about colon
carcinoma, readiness and self-efficacy for change, and
expectations that behavior change would be benefi-
cial) and physical, emotional, social, and financial vul-
nerability factors. Among the older patients only, cor-
relations between vulnerability factors and actual CRC
risk and cognitive-behavioral mechanisms associated
with CRC risk were examined. Analyses employed
contingency table analysis with chi-square testing for
discrete variables and Student t tests for continuous
variables. All analyses were performed using SAS sta-
tistical software (release 8.02; SAS Inc., Cary, NC).
RESULTS
As shown in Table 1, the full sample was comprised of
more men than women (58% vs. 42%, respectively).
Participants’ ages ranged from 40 years to 75 years,
with an average age of 60 years (standard deviation,
8.4 years). Seventeen percent of patients were non-
white. Most participants were married or lived with a
partner, were well educated, and had annual house-
hold incomes Ն $45,000. A minority of participants
(9.3%) did not report income. Most participants had
never been diagnosed with polyps.
Not surprisingly, middle-aged group versus older
group comparisons across demographic indicators re-
vealed three significant correlations. Older patients
were less likely than middle-aged patients to have
attended college or graduate school (P Ͻ 0.0001). Age
and marital status also were found to be related (P
Ͻ 0.0001) such that, compared with middle-aged pa-
tients, older patients were more likely to be widowed
and were less likely to be in a current relationship.
Overall, compared with middle-aged patients, older
patients had lower annual household incomes and
were more than twice as likely to report incomes
Ͻ $30,000 per year (P Ͻ 0.0001). They also were more
likely to refuse to report income or to check “do not
know.”
Disease Factors: To What Extent do Actual and Perceived
CRC Risks Differ by Age?
An initial step in examining the impact of aging on
CRC risk among patients with colon adenomas was
to determine whether perceptions of CRC risk dif-
fered for middle-aged patients (40–59 years) and
older patients (60–75 years) (Table 2). There was no
significant difference between middle-aged and
older patients in the number of risk factors. By
contrast, marked age differences emerged on per-
ceived CRC risk and cognitive mediators of behavior
change. First, there was a strong association be-
tween age and perceived lifetime risk of developing
CRC. Compared with middle-aged patients, older
patients were more likely to perceive that develop-
ing CRC in their lifetime was unlikely or very un-
likely (P Ͻ 0.05) and reported significantly less con-
cern about this possibility (P Ͻ 0.0001). Older
patients also were less likely than middle-aged pa-
tients to believe that changing their health habits
would reduce their CRC risk (P Ͻ 0.05). Compared
with middle-aged patients, older patients reported
significantly less motivation for changing all risk
behaviors associated with CRC (P Ͻ 0.01).
1088 CANCER March 1, 2004 / Volume 100 / Number 5

5. Aging Factors: To What Extent are Older Patients More
Vulnerable?
Next, we examined a broad range of general influences
(physical, emotional, social, financial) that were linked
previously to health adversity in later life28
(Table 3).
Within the physical realm and compared with middle-
aged patients, older patients reported significantly
higher levels of comorbid illnesses (P Ͻ 0.0001) and
lower levels of self-rated health (P Ͻ 0.05). By contrast,
and also consistent with past research,38,39
older pa-
tients’ reports of emotional functioning revealed lower
vulnerability. Compared with middle-aged patients,
older patients reported similar levels of quality of life
and significantly lower levels of depressed feelings (P
Ͻ 0.0001). Socially, older polyp patients in this sample
were significantly more likely than middle-aged pa-
tients to live alone (P Ͻ 0.01). They also had fewer
confidants (P ϭ 0.06) and significantly lower levels of
support for making behavior changes (P Ͻ 0.01). Fi-
nally, age group comparisons in the financial realm
revealed that, although older patients reported signif-
icantly lower incomes compared with middle-aged
patients (P Ͻ 0.0001), they were more likely to per-
ceive that their financial resources were just adequate
for meeting their needs (P Ͻ 0.05).
Cancer-Aging Interface: How Is CRC Risk Affected by
Age-Related Vulnerability?
The following analyses were restricted to older pa-
tients (see Table 4). Compared with older patients
with low levels of comorbid illness, physically vulner-
able older patients (i.e., those with two or more
chronic conditions) reported more risk factors for CRC
(P Ͻ 0.05), a greater likelihood that they would de-
velop CRC in their lifetime (P Ͻ 0.05), and greater
concern for developing CRC in the future (P Ͻ 0.01),
yet greater readiness to make changes (P Ͻ 0.05).
Compared with older patients who rated their health
as good or excellent, those with lower health ratings
had significantly more CRC risk factors (P Ͻ 0.01),
more worry about getting CRC in the future (P Ͻ 0.05),
and perceived themselves at great risk for developing
CRC (P Ͻ 0.01). In addition, older patients who re-
ported fair or poor health were significantly less con-
fident that they could take steps to change all of their
health habits in the next 6 months (P Ͻ 0.05).
TABLE 1
Full Sample Demographic Profile with a Comparison of the Middle-Aged and Older Patient Subsamples
Demographics
No. of patients (%)
P valuea
Entire sample
(ages 40–75 yrs)
(n ‫؍‬ 1247)
Middle-aged group
(ages 40–59 yrs)
(n ‫؍‬ 594)
Older group
(ages 60–75 yrs)
(n ‫؍‬ 653)
Female gender 523 (41.9) 249 (41.9) 274 (42.0) NS
Race
White 1030 (82.9) 487 (82.3) 543 (83.5) —
Black 150 (12.1) 71 (12.0) 79 (12.2) —
Other 62 (5.0) 34 (5.7) 28 (4.3) NS
Education level
Յ High school grad 310 (24.9) 117 (19.7) 193 (29.7) —
Ͼ High school 281 (22.6) 126 (21.3) 155 (23.8) —
College grad 260 (20.9) 149 (25.1) 111 (17.1) —
Postgraduate work 393 (31.6) 201 (33.9) 192 (29.5) Ͻ 0.0001
Marital status
Married/cohabitating 938 (75.5) 466 (78.9) 472 (72.4) —
Divorced/separated 141 (11.3) 73 (12.4) 68 (10.4) —
Widowed 92 (7.4) 14 (2.4) 78 (12.0) —
Never married 72 (5.8) 38 (6.4) 34 (5.2) Ͻ 0.0001
Annual household income
Յ $15,000 91 (7.3) 28 (4.7) 63 (9.7) —
$15,001–30,000 161 (12.9) 44 (7.4) 117 (17.9) —
$30,001–45,000 160 (12.8) 61 (10.3) 99 (15.2) —
Ն $45,001 719 (57.7) 428 (72.1) 291 (44.6) —
Don’t know/refused 116 (9.3) 33 (5.6) 83 (12.7) Ͻ 0.0001
NS: nonsignificant; grad: graduate.
a
P values were based on Student t tests for continuous dependent variables and on chi-square tests for categoric dependent variables. All comparisons were between the middle-aged subsample and the older
subsample. Missing data are not shown for race (five patients), education (three patients), or marital status (four patients).
Patient Age and Colon Adenoma/Clipp et al. 1089

6. Older patients’ reports of the number of confi-
dants in their lives were related significantly to their
readiness for change (Table 5). Those with few confi-
dants were significantly less ready to change all of
their risk behaviors (P Ͻ 0.005) and were less confi-
dent that they could make those changes (P Ͻ 0.005).
The extent to which older patients felt supported
by others in efforts to change CRC risk behaviors was
related to their outcome expectancies, motivation for
change, and self-efficacy for change. Specifically, older
patients who reported low levels of support for chang-
ing their behaviors were significantly less likely to
think that changing all of their risk behaviors would
lower CRC risk (P ϭ 0.02). Older patients who felt less
supported also indicated the lowest levels of readiness
to change (P Ͻ 0.0001) and confidence for changing all
of their risk behaviors (P Ͻ 0.0005).
Because relatively few associations were found
between CRC risk and financial vulnerability, these
data are reported but not tabled. Objective house-
hold income related to the number of CRC risk
factors, such that financially vulnerable older pa-
tients (i.e., those with incomes Ͻ $30,000) reported
significantly more CRC risk factors (P Ͻ 0.01). Com-
pared with patients who reported income, those
who did not report income expressed significantly
less concern for developing CRC in their lifetime.
Perceived adequacy of income (i.e., ability to pay
bills, purchase extras) was related to the number of
CRC risk factors and self-efficacy for making behav-
ior changes. Older patients with lower perceptions
of income adequacy reported significantly more
CRC risk factors (P Ͻ 0.05) and significantly less
confidence that they could change their risk behav-
iors in the next 6 months (P Ͻ 0.01).
DISCUSSION
Understanding aging-cancer interactions is particu-
larly important in the context of CRC, because most
CRC cases occur among individuals age Ն 50 years.
Moreover, CRC incidence nearly doubles each decade
until around age 80 years, and 5-year survival rates are
comparable among persons Ͻ 65 years and Ͼ 65
years, making the older population a key CRC screen-
ing target.40
Ideally, older patients with colon adeno-
mas should have follow-up colonoscopy screening
TABLE 2
Age Comparison of Actual and Perceived Risk, Concern, Readiness, and Self-Efficacy for Change
Enhanced risk domains
Middle-aged group
(ages 40–59 yrs)
(n ‫؍‬ 594)
Older group
(ages 60–75 yrs)
(n ‫؍‬ 653) P valuea
Actual risk:
No. of risk factors (0–6): (mean Ϯ SD) 2.5 Ϯ 1.3 2.4 Ϯ 1.2 Ͻ 0.10
Perceived risk
Likelihood of getting CRC in lifetime: no. (%)
Unlikely 218 (36.7) 285 (43.9) —
50:50 chance/unknown 302 (50.8) 294 (45.3) —
Likely 74 (12.5) 70 (10.8) Ͻ 0.03
Concern:
Level of concern for CRC (0–10) in lifetime (mean Ϯ SD) 5.4 Ϯ 3.1 4.4 Ϯ 3.1 Ͻ 0.0001
Outcome expectancies
Patients who indicated they had habits to changeb
No. of patients 555 558 —
Agree (mean Ϯ SD) 435 Ϯ 78.5 401 Ϯ 72.0 —
Do not agree (mean Ϯ SD)c
119 Ϯ 21.5 156 Ϯ 28.0 0.01
Motivation/readiness (stage of change)
No. of patients ready to change all behaviors (% yes)
Precontemplation/contemplation 298 (50.2) 384 (58.8) —
Preparation/maintenance 296 (49.8) 269 (41.2) Ͻ 0.002
Self efficacy
Confidence (1–5) in ability to change all health habits in
the next 6 mos (mean Ϯ SD) 3.7 Ϯ 0.9 3.7 Ϯ 0.9 NS
SD: standard deviation; CRC: colorectal carcinoma; NS: not significant.
a
P values were based on Student t tests for continuous dependent variables and on chi-square tests for categoric dependent variables. All comparisons were between the middle-aged and older subsamples. Missing
data are not shown for outcome expectancy (changing health habits will reduce CRC risk; n ϭ 159 middle-aged patients; n ϭ 252 older patients).
b
There were 39 patients from the from middle-aged group (6.7% of the data), and 95 patients from the older group (14.9% of the data) who were not asked this question because they did not think they had health
habits to change. In addition, 27 middle-aged patients and 39 older patients responded “don’t know.”
c
Includes “disagree” and “neither.”
1090 CANCER March 1, 2004 / Volume 100 / Number 5

7. which, since July 2001, has been covered by Medicare.
However, not all patients undergo surveillance
colonoscopy. Therefore, and regardless of whether or
not surveillance colonoscopy is performed, behavior
modifications are important in decreasing the risk of
recurrent polyps. Results of this study suggest that the
presence of age-related vulnerability factors may en-
hance CRC risk and indicate special considerations in
the design of interventions to control recurrent ade-
nomas and CRC in older patients.
Preliminary evidence for enhanced risk emerged
in three analyses. First, after the removal of one or
more adenomatous colon polyps, older patients re-
port less concern than middle-aged patients about
developing CRC. They also have lower motivation to
change and lower outcome expectancies regarding
the benefits of behavior change. The second analysis
confirmed the suspected existence among the older
patients of age-related physical, social, and financial
vulnerabilities, namely, multiple morbidities, lower
self-rated health, lower social support, and lower
financial resources. The final analysis focused on
the cancer-aging interface and identified three sub-
groups of older polyp patients with a potential for
enhanced CRC risk.
Members of the first subgroup were older polyp
patients with higher levels of comorbidity. These
chronically ill patients had more CRC risk factors
and, presumably, have illness-related functional
limitations that present barriers to exercise. There
also is mounting evidence that unhealthy behaviors
tend to cluster (i.e., those who are sedentary also are
more likely to eat high amounts of animal fat, low
amounts of fruits and vegetables, and vice versa)
both within the general population41,42
and in can-
cer patients.43
Therefore, front-line practitioners
caring for older patients with a history of colon
adenomas should consider targeting general “life-
style” changes in their recommendations rather
than individual CRC risk factors.
A second subgroup with enhanced risk was com-
prised of older patients who had few social ties or little
social support for change. Compared with more so-
cially engaged older patients, those with social sup-
ports needs held lower expectations that behavior
change would reduce CRC risk, less motivation to
change high-risk behaviors, and less confidence that
efforts to reduce CRC would succeed. These older
patients lacked key individuals who could support
them, for example, in efforts to quit smoking, reduce
alcohol intake, or achieve positive dietary changes.
The relation of social factors to older patients’ at-
tempts to reduce their cancer risk needs systematic
study, especially in patients who have inadequate sup-
port for behavior change.
The third subgroup of patients with enhanced
CRC risk included older adults who perceived that
their incomes were inadequate for meeting needs.
Compared with patients who had “money for little
extras,” those with resources that “barely met bills” or
caused a “struggle to meet bills” had more CRC risk
factors and less confidence that they could make the
lifestyle changes needed to reduce risk.
Taken together, these findings suggest that, al-
though behavioral risk essentially is identical among
middle-aged and older individuals, older adults are
more likely to underestimate that risk. Therefore, by
moderating the perception of CRC risk, age also may
moderate the likelihood of adopting healthier life-
styles. The correlations between age-related vulnera-
bilities and CRC risk also suggest that certain older
patients may be less capable than others of following
CRC risk-lowering recommendations. The extent to
TABLE 3
Age Comparison of Physical, Emotional, Social, and Financial
Vulnerability Factors Age-Related Vulnerability Factors, in
Percentages
Age-related vulnerability factors
Age group (%)
P valuea
Middle aged
(40–59 yrs)
(n ‫؍‬ 594)
Older
(60–75 yrs)
(n ‫؍‬ 653)
Physical
Higher comorbidity (Ն 2 comorbid
illnesses) 46.5 64.8 Ͻ 0.0001
Lower self-rated health (rated fair or
poor) 14.8 19.3 Ͻ 0.05
Emotional
Feels blue (sometimes, occasionally,
most or all of the time) 46.4 34.5 Ͻ 0.0001
Lower quality of life (rated fair or
poor) 13.7 11.1 NS
Social
Lives alone (yes) 13.5 19.0 Ͻ 0.01
Fewer confidants (none or 1) 12.0 15.7 0.06
Lower social support for change
(family/friends would help little or
not at all) 8.7 13.5 Ͻ 0.01
Financial
Income Յ $30,000 12.1 27.6 —
Income Ͼ $30,000 82.3 59.7 —
Don’t know/refused 5.6 12.7 Ͻ 0.0001
Lower perceived adequacy of
income (can just meet bills, must
cut back to meet
bills, difficulty paying bills) 31.7 26.2 Ͻ 0.05
NS: not significant.
a
P values were based on Student t tests for continuous dependent variable and on chi-square tests for
categoric dependent variable. All comparisons were between the middle-aged and older subsamples.
Patient Age and Colon Adenoma/Clipp et al. 1091

8. which these findings apply to other chronic conditions
or cancers in which interventions may lower risk is
unknown. For example, after a myocardial infarction,
are older patients with physical or psychosocial vul-
nerabilities less likely than others to take a daily aspi-
rin?
Research is needed to determine whether preven-
tion efforts in older patients with certain physical,
social, or financial characteristics would benefit from
interventions that reach out to functionally impaired,
socially isolated, or low-income seniors. For example,
CRC risk reduction may be realized by identifying
low-income elders for physical activity programs in
community senior centers, by establishing buddy pro-
grams that link isolated elderly with walking compan-
ions, or by partnering with dietary programs, such as
Meals on Wheels, to promote diets high in fruits and
vegetables and low in red meat. Interventions to re-
duce CRC risk in elderly with low social support may
be implemented more successfully within clinic sup-
port group settings or with frequent phone “visits.”
The results of the current study also suggest that
intervention programs targeting older patients with
polyps should make special efforts to include individ-
uals with poor personal health perceptions. These pa-
tients need directives for behavior change and may be
particularly receptive, because they tend to hold fa-
vorable outcome expectancies. However, study results
suggest that providers may need to build older pa-
tients’ levels of confidence that they can make these
changes effectively. This may be achieved more easily
by presenting risk factor reduction as a lifestyle
change rather than as multiple tasks related to multi-
ple behaviors. In light of the fact that the adenoma-
carcinoma process can take a decade or more, an
approach in which lifestyle change is linked to more
immediate gains in overall health particularly may
benefit older polyp patients who are managing multi-
ple comorbid conditions.
The current study was limited by the use of
several single-item indicators to capture aspects of
health, quality of life, and social support. Similar to
many large-scale behavioral intervention trials, and
particularly Project Prevent, with its focus on mul-
TABLE 4
Physical Vulnerability in Older Patients by Actual and Perceived Risk, Concern regarding Colorectal Carcinoma, and Readiness and Self-Efficacy
for Change
Cognitive-behavioral variables
Physical vulnerability factors
No. of comorbid illnesses Self-rated health
> 2 (n ‫؍‬ 423) < 2 (n ‫؍‬ 230) P valuea
Fair/poor
(n ‫؍‬ 125)
Good/excellent
(n ‫؍‬ 523) P valuea
Actual risk
No. of risk factors (0–6) mean Ϯ SD 2.5 Ϯ 1.2 2.2 Ϯ 1.2 0.02 2.6 Ϯ 1.2 2.3 Ϯ 1.2 Ͻ 0.01
Perceived risk
Likelihood of getting CRC in lifetime: no. (%)
Unlikely 168 (40.0) 117 (51.1) — 42 (33.6) 240 (46.2) —
50:50 chance/don’t know 200 (47.6) 94 (41.1) — 63 (50.4) 230 (44.3) —
Likely 52 (12.4) 18 (7.9) 0.02 20 (16.0) 49 (9.4) 0.01
Concern
Level of concern for CRC (0–10) in the future
(mean Ϯ SD) 4.6 Ϯ 3.2 4.0 Ϯ 3.0 Ͻ 0.01 5.0 Ϯ 3.4 4.3 Ϯ 3.0 0.04
Outcome expectancies (changing health habits will
reduce CRC risk)
No. of patients who indicated they had habits to
change (mean Ϯ SD)
Agree 266 Ϯ 78.9 135 Ϯ 74.6 — 84 Ϯ 80.0 315 Ϯ 77.0 —
Disagree/neither 71 Ϯ 21.1 46 Ϯ 25.4 NS 21 Ϯ 20.0 94 Ϯ 23.0 NS
Motivation/readiness (stage of change)
No. of patients ready to change all behaviors (%)
Precontemplation/contemplation 235 (55.6) 149 (64.8) — 78 (62.4) 304 (58.1) —
Preparation/maintenance 188 (44.4) 81 (35.2) 0.02 47 (37.6) 219 (41.9) NS
Self efficacy
Confidence (1–5) in ability ability to change all
health habits in the next 6 mos (mean Ϯ SD) 3.7 Ϯ 0.9 3.8 Ϯ 0.8 NS 3.5 Ϯ 1.0 3.8 Ϯ 0.8 0.04
SD: standard deviation; CRC: colorectal carcinoma; NS: not significant.
a
P values were based on Student t tests for continuous dependent variables and on chi-square tests for categoric dependent variables. All comparisons are within the older patient subsample.
1092 CANCER March 1, 2004 / Volume 100 / Number 5

9. tiple risk factors, the focus was on rigorously eval-
uating the intervention. Therefore, data collection
resources were allocated heavily toward the use of
gold standard health behavior measures (e.g., diet,
physical activity), whereas other measures necessar-
ily were brief. Although the brief measures were
chosen based on their strong performances in pre-
vious trials, future work on the cancer-aging inter-
face should include multiple-item indicators of
older patients’ physical and psychosocial function-
ing. Conversely, valid and reliable, single-item mea-
sures, such as self-rated health, are efficient and are
administered easily in busy clinical settings. Addi-
tional domains that may interfere with cancer con-
trol initiatives and should be considered for mea-
surement in older populations include self-care
ability, pain, fatigue, and needs for assistance with
instrumental tasks and transportation.
The expected future increase in our nation’s older
population calls for an integrated perspective of can-
cer prevention, detection, surveillance, and treatment
within oncology, geriatrics, nursing, and the behav-
ioral sciences. Common issues that cut across these
perspectives include concomitant illness, quality of
life, financial resources, and social support. Stronger
consideration of aging factors in the design of behav-
ior change interventions may reduce CRC risk by ad-
dressing physical and psychosocial vulnerabilities that
undermine adherence to recommendations for life-
style changes among older patients with polyps.
REFERENCES
1. Hunter CP, Johnson KA, Muss HB. Cancer in the elderly.
New York: Marcel Dekker, 2000.
2. Edwards BK, Howe HL, Ries LA, et al. Annual report to the
nation on the status of cancer, 1973–1999, featuring impli-
cations of age and aging on U.S. cancer burden. Cancer.
2002;94:2766–2792.
3. Cohen HJ. Geriatric principles of treatment applied to med-
ical oncology. Semin Oncol. 1995;22(Suppl 1):1–2.
4. Cohen HJ. Cancer in the elderly: an overview. In: Cassel CK,
Cohen HJ, Larson EB, et al., editors. Geriatric medicine, 3rd
edition. New York: Springer-Verlag, 1997:229–230.
TABLE 5
Social Vulnerability Factors in Older Patients (Ages 60–75 yrs) by Actual and Perceived Risk, Concern Regarding Colorectal Carcinoma, and
Readiness and Self-Efficacy for Change
Cognitive-behavioral variables
Physical vulnerability factors
No. of confidants Support for change
0–1 (n ‫؍‬ 102) > 2 (n ‫؍‬ 547) P valuea
Low (n ‫؍‬ 86)
High (n
‫؍‬ 550) P valuea
Actual risk
No. of risk factors (0–6) (mean Ϯ SD) 2.4 Ϯ 1.2 2.4 Ϯ 1.3 NS 2.4 Ϯ 1.2 2.4 Ϯ 1.2 NS
Perceived risk
Likelihood of getting CRC in lifetime: no. (%)
Unlikely 42 (41.6) 241 (44.3) — 43 (50.0) 236 (43.1) —
50:50 chance/DK 45 (44.6) 247 (45.4) — 32 (37.2) 252 (46.1) —
Likely 14 (13.9) 56 (10.3) NS 11 (12.8) 59 (10.8) NS
Concern
Level of concern for CRC (0–10) in the future (mean
Ϯ SD) 4.6 Ϯ 3.4 4.4 Ϯ 3. NS 3.9 Ϯ 3.4 4.5 Ϯ 3.1 0.13
Outcome expectancies (changing health habits will
reduce CRC risk)
No. of patients who indicated they had habits to
change (%)
Agree 52 (71.2) 349 (78.6) 0.16 40 (66.7) 357 (79.7) 0.02
Disagree/neither 21 (28.8) 95 (21.4) — 20 (33.3) 91 (20.3) —
Motivation/readiness (stage of change)
No. of patients ready to change all behaviors (% yes)
Precontemplation/contemplation 74 (72.6) 306 (55.9) — 67 (77.9) 305 (55.5) —
Preparation/maintenance 28 (27.5) 241 (44.1) Ͻ 0.01 19 (22.1) 245 (44.6) Ͻ 0.0001
Self efficacy
Confidence (1–5) in ability to change all health
habits in the next 6 mos (mean Ϯ SD) 3.5 Ϯ 0.9 3.8 Ϯ 0.9 0.11 3.1 Ϯ 1.0 3.8 Ϯ 0.9 Ͻ 0.001
SD: standard deviation; CRC: colorectal carcinoma; DK: don’t know; NS: not significant.
a
P values were based on Student t tests for continuous dependent variable and on chi-square tests for categoric dependent variables. All comparisons are within the older patient subsample.
Patient Age and Colon Adenoma/Clipp et al. 1093

10. 5. Bond JH. Polyp guideline: diagnosis, treatment, and surveil-
lance for patients with colorectal polyps. Am J Gastroenterol.
2000;95:3053–3063.
6. Winawer SJ, Shike M. Prevention and control of colorectal
cancer. In: Greenwald P, Kramer BS, Weed DL, editors.
Cancer prevention and control. New York: Marcel Dekker,
1995:537–560.
7. Winawer SJ, Zauber AG, O’Brien MJ, et al. Randomized com-
parison of surveillance intervals after colonoscopic removal of
newly diagnosed adenomatous polyps. The National Polyp
Study Workgroup. N Engl J Med. 1993;328:901–906.
8. Citarda F, Tomaselli G, Capocaccia R, Barcherini S, Crespi
M, The Italian Multicentre Study Group. Efficacy in standard
clinical practice of colonoscopic polypectomy in reducing
colorectal cancer incidence. Gut. 2001;48:812–815.
9. Yood MU, Oliveria S, Boyer JG, Wells K, Stang P, Johnson
CC. Colon polyp recurrence in a managed care population.
Arch Intern Med. 2003;163:422–426.
10. Sesink A, Termont D, Kleibeuker J, Van der Meer R. Red
meat and colon cancer: dietary ham-induced colonic cyto-
toxicity and epithelial hypoproliferation are inhibited by
calcium. Carcinogenesis. 2001;22:1653–1659.
11. Willett W. Diet and cancer: one view at the start of the millen-
nium. Cancer Epidemiol Biomarkers Prev. 2001;10:3–8.
12. Sesink A, Termont D, Kleibeuker J, Van der Meer R. Red
meat and colon cancer: dietary ham, but not fat, has cyto-
toxic and hyperproliferative effects on rat colonic epithe-
lium. Carcinogenesis. 2000;21:1909–1915.
13. Bingham S. High meat diets and cancer risk. Proc Nutr Soc.
1999;58:243–248.
14. Negri E, Bosetti C, La Vecchia C, Fioretti F, Conti E, France-
schi S. Risk factors for adenocarcinoma of the small intes-
tine. Int J Cancer. 1999;82:171–174.
15. Giovannucci E, Colditz G, Stampfer M, et al. A prospective
study of cigarette smoking and risk of colorectal adenoma
and colorectal cancer in U.S. women. J Natl Cancer Inst.
1994;86:192–199.
16. Martı´nez ME, McPherson RS, Annegers JF, Levin B. Ciga-
rette smoking and alcohol consumption as risk factors for
colorectal adenomatous polyps. J Natl Cancer Inst. 1995;87:
274–279.
17. Boutron M, Faivre J, Dop M, Quipourt V, Senesse P. To-
bacco, alcohol, and colorectal tumors: a multistep process.
Am J Epidemiol. 1995;141:1038–1046.
18. Giovannucci E. An updated review of the epidemiological
evidence that cigarette smoking increases risk of colorec-
tal cancer. Cancer Epidemiol Biomarkers Prev. 2001;10:
725–731.
19. U.S. Department of Health and Human Services. Physical ac-
tivity and health: a report of the Surgeon General. Atlanta:
Centers for Disease Control and Prevention, National Center
for Chronic Disease Prevention and Health Promotion, 1996.
20. Giovannucci E, Stampfer MJ, Colditz G, et al. Multivitamin
use, folate, and colon cancer in women in the Nurses’
Health Study. Ann Intern Med. 1998;129:517–524.
21. Glynn S, Albanes D. Folate and cancer: a review of the
literature. Nutr Cancer. 1994;22:101–119.
22. Rosenstock IM. Historical origins of the health belief
model. In: Becker MH, editor. The health belief model and
personal health behavior. Thorofare, NJ: Charles B. Slack,
1974:1–8.
23. Janz NK, Champion V, Strecher V. The health belief model.
In: Glanz K, Lewis F, Rimer B, editors. Health behavior and
health education: theory, research and practice, 3rd edition.
San Francisco: Jossey-Bass, 2002:45–66.
24. Weinstein ND. The precaution adoption process. Health
Psychol. 1988;7:355–386.
25. Weinstein N. Unrealistic optimism about future life events.
J Pers Soc Psychol. 1980;39:806–820.
26. Prochaska JO, DiClemente CC. Stages and processes of self-
change of smoking: toward an integrative model of change.
J Consult Clin Psychol. 1983;51:390–395.
27. Bandura A. Social foundations of thought and action: a
social cognitive theory. Englewood Cliffs, NJ: Prentice Hall,
1986.
28. Besdine RW. Clinical approach: an overview. In: Cassel CK,
Cohen HJ, Larson EB, et al., editors. Geriatric medicine, 3rd
edition. New York: Springer-Verlag, 1997:155–168.
29. Schatzkin A. Diet and colorectal cancer: still an open ques-
tion. J Natl Cancer Inst. 1995;87:1733–1735.
30. Polich JM, Orvis BR. Composite quantity frequency index.
In: National Institute of Alcohol Abuse and Alcoholism.
NIAAA treatment handbook series 4: assessing alcohol prob-
lems. Bethesda: National Institute of Alcohol Abuse and
Alcoholism, 1998:457–459.
31. Stewart A, Painter P. Issues in measuring physical function-
ing and disability in arthritis patients. Arthritis Care Res.
1997;10:395–405.
32. Stewart A, Verboncoeur C, Mclellan B, et al. Physical activity
outcomes of CHAMPS II: a physical activity promotion pro-
gram for older adults. J Gerontol A Biol Sci Med Sci. 2001;56:
M465–M470.
33. Fagerstrom KO. Measuring degree of physical dependence
to tobacco smoking with reference to introduction of treat-
ment. Addict Behav. 1978;3:235–241.
34. Emmons K, Marcus B, Linnan L, Rossi J, Abrams D. The
relationship between smoking, physical activity, and dietary
fat intake among manufacturing workers. Prev Med. 1994;
23:481–489.
35. Fillenbaum GG. Multidinensional functional assessment of
older adults. Hillsdale, NJ: Lawrence Erlbaum Associates,
1988.
36. Breslow L. Social ecological strategies for promoting healthy
lifestyles. Am J Health Promot. 1996;10:253–257.
37. McLeroy KR, Bibeau D, Steckler A, Glanz K. An ecological
perspective on health promotion programs. Health Educ Q.
1988;15:351–377.
38. Cassileth B, Lusk E, Brown L, Cross P, Walsh W. Factors
associated with psychological distress in chronically ill pa-
tients. Med Pediatr Oncol. 1986;14:251–254.
39. Ganz P, Coscarelli C, Heinrich R. The psychosocial impact of
cancer on the elderly: a comparison with younger patients.
J Am Geriatr Soc. 1985;33:420–435.
40. Prindiville SA. Screening for colorectal cancer in the elderly.
In: Hunter CP, Johnson KA, Muss HB, editors. Cancer in the
elderly. New York: Marcel Dekker, 2000:41–56.
41. Serdula MK, Byers T, Mokdad AH, Simoes E, Mendlein JM,
Coates RJ. The association between fruit and vegetable intake
and chronic disease risk factors. Epidemiology. 1996;7:161–165.
42. French SA, Hennrikus DJ, Jeffrey RW. Smoking status, di-
etary intake, and physical activity in a sample of working
adults. Health Psychol. 1996;15:448–454.
43. Demark-Wahnefried W, Peterson B, McBride C, Lipkus I,
Clipp E. Current health behaviors and readiness to pursue
life-style changes among men and women diagnosed with
early stage prostate and breast carcinomas. Cancer. 2000;88:
674–684.
1094 CANCER March 1, 2004 / Volume 100 / Number 5