Cancer Causes & Control

, Volume 19, Issue 4, pp 339–359

Predictors of colorectal cancer screening behaviors among average-risk older adults in the United States

Authors

  • Hind A. Beydoun
    • Department of Epidemiology, College of Public HealthUniversity of Iowa
    • Center for Human Nutrition, Department of International HealthJohns Hopkins Bloomberg School of Public Health
Review Article

DOI: 10.1007/s10552-007-9100-y

Cite this article as:
Beydoun, H.A. & Beydoun, M.A. Cancer Causes Control (2008) 19: 339. doi:10.1007/s10552-007-9100-y

Abstract

Objective

To critically evaluate recent studies that examined determinants of CRC screening behaviors among average-risk older adults (≥50 years) in the United States.

Methods

A PUBMED (1996–2006) search was conducted to identify recent articles that focused on predictors of CRC initiation and adherence to screening guidelines among average-risk older adults in the United States.

Results

Frequently reported predictors of CRC screening behaviors include older age, male gender, marriage, higher education, higher income, White race, non-Hispanic ethnicity, smoking history, presence of chronic diseases, family history of CRC, usual source of care, physician recommendation, utilization of other preventive health services, and health insurance coverage. Psychosocial predictors of CRC screening adherence are mostly constructs from the Health Belief Model, the most prominent of which are perceived barriers to CRC screening.

Conclusions

Evidence suggests that CRC screening is a complex behavior with multiple influences including personal characteristics, health insurance coverage, and physician–patient communication. Health promotion activities should target both patients and physicians, while focusing on increasing awareness of and accessibility to CRC screening tests among average-risk older adults in the United States.

Key words

Colorectal cancerScreeningAging

Introduction

Colorectal Cancer (CRC) is the second leading cause of cancer mortality, accounting for nearly 11% of all cancer deaths, and the fourth most frequently diagnosed cancer in older adults living in the United States [17]. CRC incidence and mortality rates rank second to lung cancer in men and third after lung cancer and breast cancer among women [8]. In the United States, the lifetime risk for CRC is 1 in 18 or close to 6% [9, 10]. In 2005, the American Cancer Society (ACS) estimated a total of 104,950 colon cancers, 40,340 rectal cancers, and 56,290 CRC deaths among men and women in the United States [11]. More than 90% of CRC occurs in individuals above 50 years of age and 75% of CRC results from benign, adenomatous polyps, among average-risk older adults with no significant colorectal symptoms, family history, or genetic predispositions [10, 1214]. On the other hand, CRC survival rates vary by stage at diagnosis, ranging between 94% among patients with localized CRC and 9% among patients with metastatic CRC [10].

Evidence suggests that both primary and secondary prevention strategies are effective methods for achieving a reduction in CRC incidence and mortality rates[4]. Epidemiologic evidence suggests that the incidence of CRC in the United States could be reduced by as much as 50% if all individuals modified identified risk factors for CRC through weight reduction, alcohol and tobacco control, physical activity, reduction in red meat consumption, increased intake of fiber, fruits, and vegetables, regular use of anti-inflammatory medications, multivitamins containing folic acid, and postmenopausal hormones in women [2, 10] . Furthermore, published analyses show that CRC screening is a cost-effective means of reducing mortality from CRC through early detection and treatment as well as through prevention of CRC [7, 1435].

The ACS, the National Cancer Institute (NCI), and the United States Preventative Services Task Force (USPSTF) agree on the importance of CRC screening [1, 3, 10]. However, due to the availability of a wide range of CRC screening options, specific CRC screening recommendations are complex [9, 13, 3638]. Current practice guidelines favor colonoscopy every 10 years as the preferred screening option in asymptomatic men and women aged 50 years and older with no prior CRC, no family history of CRC, and no genetic predisposition to CRC [9, 13, 36, 39, 40]. Alternatively, “annual FOBT” and/or “sigmoidoscopy (or double-contrast barium enema) every 5 years” are recommended for these average-risk older adults [9, 13, 16, 36, 40]. More frequent CRC screening tests are recommended for intermediate and high-risk individuals, especially those genetically susceptible to Hereditary Nonpolyposis Colorectal Carcinoma (HNPCC) [10, 4145]. Each CRC screening modality has a different risk-benefit profile, with colonoscopy having the most risks and benefits [46]. Due to the invasive nature of endoscopy, virtual colonoscopy and DNA stool tests may show promise as CRC screening tools [18].

Despite the availability of specific guidelines for CRC screening, population-based studies, including the Behavioral Risk Factor Surveillance System (BRFSS) and the National Health Interview Survey (NHIS), have consistently shown under-utilization of CRC screening procedures among older adults in the United States [16, 21, 3234, 39, 40, 4751]. In particular, the BRFSS is an ongoing state-based telephone survey of non-institutionalized civilian individuals 18 years or older [6]. One of the main objectives of the BRFSS is to monitor progress toward the Healthy People 2010 targets including higher utilization rates of preventive healthcare services [6].

BRFSS (1991–2004) trends suggest that each year less than 60% of men and women 50 years of age or older have used CRC screening services according to the established guidelines (Fig.  1) [5, 21, 39, 40, 47, 4952]. For instance, the 1999 BRFSS indicated that only 40% of age-eligible older adults residing in 50 states, the District of Columbia, and Puerto Rico ever received a home-administered FOBT and 44% had undergone either sigmoidoscopy or colonoscopy. In addition, 21% received FOBT in the preceding year and 34% had a sigmoidoscopy or colonoscopy within the preceding 5 years [3]. Similarly, the 2004 BRFSS indicated that 57.3% of older adults had an FOBT within one year or a lower endoscopy (sigmoidoscopy or colonoscopy) within ten years preceding the survey [40]. Therefore, innovative behavioral interventions are needed to increase the proportion of age-eligible United States men and women who are likely to initiate CRC screening and adhere to CRC screening guidelines [17, 5358]. A prerequisite to the design, implementation and evaluation of such behavioral interventions is a better understanding of factors that may influence CRC screening behaviors among age-eligible men and women in the United States [38, 5963]. The low prevalence of CRC testing has been compounded by wide racial, ethnic, and socioeconomic disparities in CRC screening behaviors. Previously conducted studies have associated personal, health care access, delivery, and utilization issues as well as psychosocial characteristics with the utilization of CRC screening procedures among age-eligible men and women in the United States [21, 39, 40, 47, 51].
https://static-content.springer.com/image/art%3A10.1007%2Fs10552-007-9100-y/MediaObjects/10552_2007_9100_Fig1_HTML.gif
Fig. 1

Trends in colorectal cancer screening among age-eligible adult men and women in the United States: Behavioral Risk Factor Surveillance System (1991–2004) [3, 5, 21, 39, 40, 47, 4952]

The main objective of this article is to critically evaluate recent studies that examined determinants of CRC screening behaviors among average-risk older adults (≥50 years) in the United States. Accordingly, we have conducted a systematic review of observational studies published between 1996 and 2006 that identified predictors of CRC screening behaviors. Our long-term goal is to identify high-risk groups and modifiable characteristics that could be targeted by future intervention studies.

Materials and methods

A PUBMED search (1996–2006) was conducted to identify English-language human studies which included keywords (“United States” or “American”) and “colorectal cancer screening” in their titles or abstracts. We identified a total of 152 (121 original and 31 review) articles. Of those, 37 observational studies [1, 35, 8, 12, 16, 46, 52, 53, 55, 59, 60, 6487] were included because their primary focus was on identifying predictors of CRC screening behaviors (intention, use, initiation, or adherence) among average-risk older adults in the United States. Studies that focused on predictors of any type of cancer screening, predictors of CRC screening knowledge, attitudes and beliefs, and those without an abstract were not included in this review. We also excluded studies that were restricted to CRC intermediate or high-risk older adults. In most studies, adherence to CRC screening guidelines was defined as “colonoscopy every 10 years,” “annual FOBT,” and/or “sigmoidoscopy every 5 years” [5, 8, 55, 60, 66, 69, 73, 81, 82, 87].

Several of the selected studies presented secondary analyses from the BRFSS [16, 52, 73, 76], the NHIS [55, 74], or other pre-existing survey data [59, 66]. Other studies were conducted using medical charts abstraction [1, 81] or primary data collection through face-to-face interviews [5, 66], telephone [8, 65, 69, 72, 79, 86], mailed [67, 71, 83, 87], and self-administered surveys[3, 12, 60, 68, 80]. Furthermore, some of these studies incorporated focus group discussions [53, 60]. While most studies were population-based, some were conducted in a clinical setting [1, 35, 12, 46, 53, 77, 78, 81, 82]. Most of the studies enrolled racially and ethnically diverse groups, while a few were restricted to Hispanic, African American, or Asian communities. Other studies were restricted to men or women. Sampling strategies ranged from a convenience sample of 15 African-American women [64] to a nationally representative sample of 61,068 older adults [65]. In Table 1, we have presented the characteristics of studies included in the systematic review. Tables 24 present the results of these studies by type of CRC screening predictor.
Table 1

Predictors of CRC screening initiation and adherence among average-risk older adults in the United States—Summary of articles included in systematic review (1996–2006)

Ref #

Data source, sample sizes, and restrictions

Outcome

Predictors examined

Predictors identified

Behavioral Risk Factor Surveillance System

[75]

2002 BRFSS

FOBT past year

Age, Gender, Ethnicity, Education, Income, Health insurance, Usual source of care, State of residence

↑ Non-Hispanic

Hispanic (n = 5,680)

LE past 10 years

↑ Education

Non-Hispanic (n = 104,733)

↑ Income

↑ Health insurance

↑ Usual source of care

[73]

2002 BRFSS

Adherence to CRC screening guidelines

Age, Health insurance, Personal physician, PSA screening, Poor health, Smoking status

Age (U-shaped)

n = 22,304

↑ PSA screening

Men only

↑ Health insurance

↑ Personal physician

↑ Poor health

↑ Non-smoker

[52]

1999 BRFSS

FOBT past year

Urban–rural residence

↑ Urban influence

Men (n = 23,565); Women (n = 37,847)

SC past 5 years

[16]

1999 BRFSS

CRC screening under utilization

Gender, Age, Race/Ethnicity, Marital status, Education, Employment, Income, Health insurance, Alcohol use, Smoking status, Current general health, Healthcare access, Health check-up, Cholesterol, Mammography, Pap Test

↑ Age (50–64 years)

↓ Education

↓ Health insurance

[76]

2001 BRFSS

Up-to-date CRC screening (FOBT, FS, CP)

Age, PSA screening, State of residence

↑ Age (50–69 years)

n = 49,315

↑ Up-to-date PSA screening

Men only

National Health Interview Survey

[55]

2000 NHIS (n = 12,677)

CRC screening non-adherence (multivariate)

Race, Ethnicity, Urban/Rural Residence Other factors

↑ African-American

↑ Hispanic

↑ Urban residence

[74]

1987, 1992, 1998, 2003 NHIS

Recent FOBT and/or LE

Age, Gender, Ethnicity, Marital status, Education, Household income, Health insurance, Usual source of care, Physician consultation, Number of doctor visits, Mammography, PSA screening, Pap test, Smoking status

↑ 65+ years

↑ Male

↑ Non-Hispanic

↑ Education

↑ Usual source of care

↑ Talk to general doctor

↑ 2–5 doctor visits past year

↑ Health insurance

↑ Ex-smoker

Population-based studies

[70]

Japanese-Americans (n = 341)

CRC screening use

Age, Gender, Income, Education, Acculturation, Marital status, Regular access, Provider–Patient communication, Frequency of contact with family, Emotional Family support, Frequency of contact with close friends, Emotional friend support, Subjective norms, Perceived Behavioral control, Perceived benefits

↑ Usual source of care

↑ Provider–patient communication

↑ Emotional family support

↑ Family/Friend subjective norm

[71]

2003 University of Michigan Health Systems Clinic

Mailed survey

Stratified random sample (n = 1,184)

CRC screening attempt

CRC non-use

CRC preference:

FOBT

Invasive test

Age, Gender, Race, Education, Employment, Income, Health insurance, Family history of CRC, “Accuracy of results,” “Forgot,” “Discomfort,” “Prior colonoscopy”

“Accuracy of results”

“Forgot”

“Discomfort”

“Prior colonoscopy”

[72]

2001 Medicare consumers (North and South Carolina)

Ever use of CRC screening

Age, Gender, Race, Education, Marital status, Medicaid eligibility, Geographical region, Usual source of care, Check-up in past year, CRC risk, Chronic conditions, Physician recommendation, Knowledge/Awareness

↑ Physician recommendation

↑ Knowledge/Awareness

Current with CRC screening

 

Telephone survey

Random sample (n = 1,901)

[83]

Mailed survey

and 5-year retrospective claims data

CRC screening use

Age, Gender, Race/Ethnicity, Marital status, Income, Education, Employment status, Charlson Co-morbidity Index, Health status, Perceived risk/Benefit, Decision-making preference, Help scheduling appointment Discussion of results or follow-up Choice among screening modalities

↑ Help scheduling appointment

↑ Discussion of results or follow-up

↓ Choice among screening modalities

Insured primary care patients (n = 4,966)

 

5A’s: “assess,” “advise,” “agree,” “assist,” and “arrange”

[84]

Hispanic population (50–80 years)

FOBT use

Age, Gender, Marital status, Education, Family income, Acculturation (low, moderate, high)

↓ Acculturation

Endoscopy (SG, CP, PS) use

[60]

Member of 16 Appalachian churches

Adherence to CRC screening guidelines

Age, Gender, Psychosocial facilitators and barriers

↑ Age

↑ Male

23 focus groups (n = 205)

Psychosocial facilitators and barriers

Self-administered survey (n = 839)

[59]

2002 Maryland Cancer Study (n = 2,994)

FOBT past year, SG past 5 years, or CP past 10 years

Gender, Race, Ethnicity, Marital status, Education, Income, Health insurance, Place of residence, Health status, Smoking, Alcohol use, Body Mass Index, Level of concern about cancer, Family history of CRC, Knowledge of CRC screening, Physician recommendation for FOBT, Physician recommendation for endoscopy, Usual source of care, Health access

↑ Clinician recommendation

[85]

Hispanics (n = 137); Non-Hispanics (n = 491)

FOBT use

Age, Gender, Ethnicity, Marital status, Education, Income, Acculturation, Smoking status, Health insurance

↓ Hispanic (SG/CP use)

SG/CP use

[86]

226 individuals

CRC screening non-use

Age, Gender, Race, Marital status, Education, Employment status, Household income, Predisposing (Knowledge, Beliefs, Values), Enabling (Skills, Availability/Accessibility of resources) and Reinforcing factors (Social support)

↑ Lack of familiarity with CRC guidelines and tests

(52–80 years)

Telephone survey

↓ Availability and accessibility to screening tests

Precede–Procede framework

[87]

554 patients at multi-specialty healthcare group

Mailed survey

Endoscopic CRC screening compliance

(gender-specific)

Demographics, Health behaviors, Psychosocial factors

↑ Female provider for female patient

↑ Length of patient–provider relationship

↓ Fear and embarrassment

(Women  >  Men)

↑ Perceived benefits of CRC screening

[66]

598 controls from the North Carolina Colon Cancer Study

Face-to-face interview

CRC screening compliance

Age, Race, Income, Regular physician, General medical examination

↑ Age (≥60)

↑ Regular doctor

↑ General medical exam

[67]

Japanese residents of United States metropolitan areas Mailed survey

CRC, (FOBT, SG/CP) screening

Age, Gender, Marital status, Acculturation, Physician recommendation, Health insurance, Perceived benefits, Perceived costs, Perceived susceptibility, Family history of CRC, Social support, Frequency of gastrointestinal, symptoms

↑ Age (SG/CP)

↑ Male (FOBT)

↑ Acculturation (CRC)

↑ Physician recommendation (CRC)

↑ Health insurance (SG/CP)

↓ Perceived psychological costs (CRC)

↑ Married (FOBT)

↑ Susceptibility (SG/CP)

[68]

Chinese-American men and women at 3 senior centers (n = 203)

Self-administered survey

FOBT past year

SG past 5 years

Age, Gender, Marital status, Education, Employment, Household income, Neighborhood, Citizenship, Family history of CRC, Years of residency, Salience and coherence, Self-efficacy, Social influence, Efficacy of screening, Worries or fears, Perceived susceptibility

↓ Years of US residency (FOBT)

↑ Education (SG)

↓ Worries and fear of test results

(FOBT & SG)

↑ Perceived susceptibility

(FOBT & SG)

[69]

Latino, non-Latino White, Vietnamese (n = 775)

Telephone survey

CRC screening adherence

FOBT past year

SG past 5 years

CP past 10 years

Gender, Race, Marital status, Education, Health insurance, Employment status, Regular doctor, Doctor visits, Self-rated heath, Acculturation, language spoken with doctor, Doctor ethnicity

↓ Vietnamese (CRC)

↓ Vietnamese

(SG, not FOBT or CP)

↓ Latino (CRC)

[64]

15 African-American women (45–69 years)

CRC screening

Barriers

Fear; afraid of pain; doctor never recommended any tests; had no symptoms; competing health histories; embarrassment; and flaws in risk perception

Non-use

[65]

Nationally representative sample (n = 61,068)

Telephone survey

Current CRC screening:

- SG/CP in past 5 years

or

- FOBT in past year

Age, Gender, Race/Ethnicity, Education, Employment status, Health insurance, Geographic region, Self-reported general health, Time since last physician visit, Smoking status, Alcohol use, Body Mass Index, Vitamin use, Physical exercise, Low-fat diet

↓ Age (50–54 years) (CRC)

↑ Females (FOBT)

↓ Asian and Pacific Islanders (CRC)

↑ Asian and Pacific Islanders (SG/CP)

↓ Hispanic (CRC)

↑ Hispanic (SG/CP)

↓ Education  <9th grade (CRC)

↓ No routine doctor visit in past year (CRC)

↑ No routine doctor visit in past year (SG/CP)

↓ Everyday smokers

[8]

Black and White men and women in Genesee County, Michigan (age: 50–79 years)

(n = 355)

Telephone survey (RDD)

CRC screening adherence

Age, Gender, Race, Marital status, Education, Employment status, Income, Health insurance, Family history, Other cancer screening, Susceptibility, Knowledge, Physician recommendation, Benefits and barriers, Salience and coherence

↓ Black Females (FOBT, FS, CP)

↓ Black Females (FOBT, FS, CP)

↑ Physician recommendation

“Test is not needed”; “Test is embarrassing”

[79]

Massachusets residents (n = 954)

Telephone survey

(RDD)

CRC screening

Age, Gender, Race/Ethnicity, Marital status, Education, Household income, Health insurance, Family history of CRC, Regular medical check-up, Check-up, Mammography, PSA, Vitamin supplement use, Smoking status

↑ Regular medical check-up

↑ Mammography (women)

↑ PSA (men)

↑ Family history

↑ Vitamin supplement use

[80]

Chinese-American women age 60 years and older (n = 100)

1+ (FOBT)

Acculturation, Physician recommendation, Annual household income, Marital status, Knowledge of someone with CRC, Family history of CRC, Personal history of CRC, Personal risk of CRC, Time since last physical examination, Health insurance coverage, Sigmoidoscopy coverage

↑ Acculturation (FOBT & SG)

↑ Physician recommendation (FOBT)

Recruited from senior centers in two metropolitan areas

1+ (SG)

 

Self-administered survey

[4]

African American adults attending church-based health promotion program (n = 850)

Recent CRC screening

Age, Gender, Health Belief Model

Barriers (CRC)

(FOBT or SG/CP)

Benefits (SG/CP, not FOBT)

[53]

Focus groups (n = 9)

CRC screening

Cultural factors

↓ Discrimination

Mid-Atlantic Latino patients(n = 70)

↓ New immigrant

↑ Language

Primary care providers (n = 27)

Other factors

Primary care site characteristics

Provider characteristics

Cost/Insurance coverage

Knowledge

Attitudes

Ordering of priorities

Screening procedures

[81]

Chinese Americans attending community health clinic in Seattle

FOBT past 12 months, sigmoidoscopy past 5 years and/or colonoscopy past 10 years

Gender, Age, Health insurance Language

Retrospective review

383 medical charts (July, 2003–September, 2004)

Clinic-based surveys and retrospective reviews

[77]

Institutional administrative database (1998–2002)

Non-FOBT CRC screening (FS, BE or CP 12 months after mammography)

Age, Health insurance, Breast Imaging Reporting and Data System classification, Recommendations after screening mammography, Year of mammography

↑ Health insurance with commercial managed care organization or Medicaid

17,790 women who sought mammography

[5]

Cross-sectional, retrospective study at three Midwestern Clinics

CRC screening compliance

Age, Gender, Income, Education, Ethnicity, Marital status, Employment status, Health status, Health insurance

↑ Health insurance

Face-to-face interviewing

Quota sampling (n = 104)

[1]

Community health centers (n = 8)

Medical chart abstraction

1,176 patients

CRC screening

Age, Gender, Race, Use of preventive care services, Charlson co-morbidity index, Family history

↑ Age

↑ Male

↑ African-American

↑ Preventive care services in past year

↑ Chronic illnesses

↑ CRC family history

[78]

Women (50–75 years) who received mammography at a single academic medical center in 1998. Retrospective review:

CRC screening completion after mammography

Age, Health insurance, BI-RADS classification, classification of mammographer recommendation, Routine screening mammography in 12 months

↑ Managed care insurance

Radiology Information System (n = 3,357)

[12]

Convenience sample of adults attending internal medicine and family practice clinics of community teaching hospital (n = 193)

CRC screening intention

Age, Gender, Ethnicity, Income, Health insurance, Language, Religion, Knowledge, Smoking status, Past experience with CRC screening

↑ Age

↓ Catholics (CRC, CP)

↑ Knowledge of someone with colon cancer or colon polyps (CRC)

Self-administered survey

↑ Ex-smokers (CRC)

↑ Previous experience with CRC screening

[82]

women seeking mammography at large urban breast diagnostic facility (n = 280)

CRC screening adherence

Age, Ethnicity, Marital status, Education, Family income, Health insurance, Physician recommendation, Perceived severity, Perceived susceptibility, Perceived benefits, Perceived barriers, Self-efficacy

↑ Physician recommendation

↑ Perceived benefits

↓ Perceived barriers

↑ Self-efficacy

[3]

Outpatients at Mayo clinic (150 never screened; 150 previously screened)

Self-administered survey

Ever CP screening

Age, Gender, Regular physician Screening mammography, Knowledge, Barriers

↑ Regular primary physician

↑ Screening mammography

↓ Knowledge

“Volume of bowel preparation”

“Adequate analgesia”

“No recommendation from primary physician”

“Embarrassment”

[46]

Washington State Medicare beneficiaries (1994, 1995, 1998)

Use of screening and diagnostic colon test

Year, Age, Gender, Race, Place of residence

↓ Age (>80 years) (CRC)

↓ Female (CRC); ↓ Males (FOBT)

↓ non-White (CRC & FOBT)

BRFSS—Behavioral Risk Factor Surveillance System; CRC—Colorectal cancer; FOBT—Fecal occult blood test; FS—Flexible sigmoidoscopy; CP—Colonoscopy; LE—Lower Endoscopy; NHIS—National Health Interview Survey; PS—Proctoscopy; PSA—Prostate-Specific Antigen; SC—Sigmoidoscopy or colonoscopy; SG—Sigmoidoscopy; ↑ Positive association with outcome; ↓ Negative association with outcome

Table 2

Effects of socio-demographic factors on colorectal cancer screening behaviors—Summary of articles included in systematic review (1996–2006)

Ref #

Main result

Gender

[74]

Men reported higher use of endoscopy than women if they had a usual source of health care, had talked to a general doctor, and had two to five visits to the doctor in the past year

[60]

Male gender was a predictor of CRC screening adherence (males: 66.7% vs. females: 58.7%, p = 0.03)

[81]

No significant differences between users and non-users of FOBT, SIG, and COL in terms of age, gender, insurance status, or language

[67]

Gender was related to screening via fecal occult blood testing (OR, 2.5; 95% CI: 1.5–4.1)

[1]

CRC screening was predicted by male gender (male: 45.0% vs. female: 39.0%, p = 0.09)

[87]

Women reported significantly more embarrassment and fear about having FS than men

Women were more willing to consider having a FS if a female endoscopist performed the procedure

[65]

FOBT was more common in women than in men (OR, 1.8; 95% CI: 1.6–2.0)

[8]

Adherence was lowest for black females: 21% for fecal occult blood test, 20% for flexible SIG, and 12% for COL

Black males compared to black females were about 2.8 times more likely to have had either flexible SIG or COL (p < 0.05)

[46]

Women were less likely to be screened in all 3 years (1994, 1995, 1998) (p < 0.001). Men were less likely to receive a fecal occult blood test (p < 0.001)

Age

[74]

Men and women 65+ years had higher rates of recommended CRC test (males, 55.8%; females, 48.5%) than persons 50–64 years (males, 41.0%; females, 31.4%)

[60]

Older age was a predictor of CRC screening adherence (50–64: 51.1%, 65–74: 66.7%, ≥75: 61.5%, p = 0.05)

[81]

No significant differences between users and non-users of FOBT, SIG, and COL in terms of age, gender, insurance status, or language

[73]

In men, age-predicted CRC screening with an inverse-U correlation

[1]

In the multivariate analysis, CRC screening was predicted by older age (aOR, 1.03; 95% CI: 1.01–1.05)

[66]

In the multivariable analysis, older age was a significant predictor of current CRC screening status with an aOR (95% CI) of 2.9 (1.7–4.8) for those 60–69 years compared to respondents 50–59 years and aOR (95% CI) Of 3.2 (1.9–5.5) for those 70 and older compared to respondents 50–59 years

[67]

In the multivariate analysis, age was related to SIG/COL screening (aOR, 1.07; 95% CI: 1.03–1.11)

[12]

Patients preferring no screening were significantly younger than those who expressed a preference for CRC screening (p = 0.008)

[16]

Underutilization of CRC screening tests was highest in persons aged 50–64 years

[65]

The lowest rate of current CRC screening use was reported by those aged 50–54 years

[46]

Individuals older than 80 years were less likely to be screened in all 3 years (1994, 1995, 1998) (p < 0.001)

Race

[55]

In the multivariate model, the odds for being adherent with current CRC screening recommendations were lower for African Americans (aOR, 0.82; 95% CI: 0.71–0.95) than Whites

[1]

In the multivariate analysis, CRC screening was predicted by being African-American (aOR, 1.38; 95% CI: 1.04–1.84) as compared to being White

[69]

CRC screening rates were generally lower in Vietnamese. Vietnamese were less likely than Whites to have had SIG in the past 5 years (OR, 0.26; 95% CI: 0.09–0.72), but ethnicity was not an independent predictor of FOBT or COL. Only 22% of Vietnamese would find endoscopic tests uncomfortable compared with 79% of whites (p < 0.05). Vietnamese were more likely than whites to plan to have SIG in the next 5 years (OR, 2.24; 95% CI: 1.15–4.38), but ethnicity was not associated with planning to have FOBT or COL

[65]

The lowest rate of current CRC screening use was reported by Asian/Pacific Islanders (34.8%)

SIG/COL was more common in Asian/Pacific Islanders (OR, 2.4; 95% CI: 1.5–3.9) relative to Whites

[8]

Adherence was lowest for black females: 21% for fecal occult blood test, 20% for FS, and 12% for COL

Black males compared to black females were about 2.8 times more likely to have had either FS or COL (p < 0.05)

[46]

Non-whites were less likely to be screened in all 3 years (1994, 1995, 1998) (p < 0.001)

Ethnicity

[75]

After adjusting for differences in education, income, insurance, and having a usual source of health care, Hispanic respondents remained less likely than non-Hispanic respondents to report CRC testing (OR for FOBT, aOR, 0.66; 95% CI: 0.56–0.81; OR for LE, 0.87; 95% CI: 0.77–0.99)

Greater disparity in screening rates between Hispanics and non-Hispanics was observed in Colorado, California, and Texas than in other states

[55]

In the multivariate model, the odds for being adherent with current CRC screening recommendations were lower for Hispanics (OR, 0.71; 95% CI: 0.59–0.86) than Whites

[74]

Use of CRC tests was higher among men and women if they were not Hispanic

SIG/COL was more common in Hispanics (OR, 1.4; 95% CI: 1.1–1.7) relative to Whites

[85]

Hispanics were less likely than non-Hispanic whites to have ever received an FOBT (40.6% vs. 55.7%, p = 0.003) or SIG/COL (26.9% vs. 44.4%, p < 0.001). No significant difference across ethnic groups was observed in the prevalence of recent screening using FOBT (29.8% for Hispanics vs. 34.5% for non-Hispanic whites; p = 0.41), but recent use of SIG/COL was lower for Hispanics (24.1% for Hispanics vs. 33.7% for non-Hispanic whites; p = 0.06)

[69]

CRC screening rates were generally lower in Latinos and Vietnamese. Vietnamese were less likely than Whites to have had SIG in the past 5 years (OR, 0.26; 95% CI: 0.09–0.72), but ethnicity was not an independent predictor of FOBT or COL. While 21% of Latinos would find performing an FOBT embarrassing, only 8% of whites and 3% of Vietnamese felt this way (p < 0.05)

[65]

The lowest rate of current CRC screening use was reported by Hispanics (31.2%)

Education, income, marital status

[74]

Use of CRC tests was higher among men and women if they had higher educational attainment

[75]

Rates of CRC screening test use were lower for respondents who reported less education and lower income, regardless of Hispanic ethnicity

[67]

Being married or cohabiting (OR, 2.3; 95% CI: 1.3–3.9) was related to screening via FOBT

[16]

Never-use of CRC screening tests was highest in persons with less than high school education

[85]

Having few years of education was directly associated with failure to ever receive an FOBT or SIG/COL

[68]

Receipt of a flexible SIG within the past 5 years (22.2%) was associated with higher levels of education

[65]

The lowest rate of current CRC screening use was reported by those with education less than the ninth grade (34.4%)

aOR—adjusted odds ratio; CI—confidence interval; CRC—Colorectal cancer; OR—odds ratio; RR—relative risk; FOBT—fecal occult blood test; FS—Flexible sigmoidoscopy; COL—colonoscopy; SIG—sigmoidoscopy

Table 3

Effects of health care access, delivery, and utilization characteristics on colorectal cancer screening behaviors—Summary of articles included in systematic review (1996–2006)

Ref #

Main result

Health insurance

[74]

Use of CRC tests was higher among men and women if they had health insurance

[75]

Rates of CRC screening test use were lower for respondents who reported no health insurance, regardless of Hispanic ethnicity

[81]

No significant differences between users and non-users of FOBT, SIG, and COL in terms of age, gender, insurance status, or language

[73]

In men, a positive correlate of adherence to CRC screening was having health insurance (aOR, 1.39; p < 0.01)

[77]

In multivariate analysis, being insured by a commercial-managed care organization (aOR, 3.36; 95% CI: 1.59–7.10) or by Medicaid (aOR, 3.27 95% CI: 1.36–7.89) was a significant predictor of non-FOBT CRC screening

[5]

Insured compared to uninsured participants were significantly more likely to have ever completed any CRC testing (77% vs. 33%; p < 0.001), and were more likely to have undertaken testing according to current US guidelines (62% versus 17%; p < 0.001)

[85]

Having health insurance coverage was directly associated with FOBT (aOR, 2.93; 95% CI: 1.19–7.19) or SIG/COL (aOR, 2.78; 95% CI: 1.06–7.27) compliance

[67]

Health insurance was related to SIG/COL screening (HMO/commercial plan vs. Medicare/Medicare-HMO: aOR, 4.0; 95% CI: 1.2–13.2)

[78]

Managed care insurance was the only significant predictor of CRC screening completion after screening mammography in eligible women after adjusting for other variables (aOR, 1.73; 95% CI: 1.21–2.47, p < 0.0001)

[16]

Underutilization of CRC screening tests was highest in persons who lacked health insurance

[65]

The lowest rate of current CRC screening use was reported by those with no health care coverage (20.4%) or coverage by Medicaid (29.2%)

One of the most important modifiable predictors of current CRC was health care coverage (OR, 1.7; 95% CI: 1.5–1.9)

Regular physician, regular check-up, and other access characteristics

[70]

While usual source of care was directly associated with CRC screening adherence, better provider–patient communication was directly and indirectly associated with adherence via increased perceived benefits

[83]

Multivariable model results indicated that the likelihood of CRC screening was greater among patients reporting help scheduling an appointment (assist) (aOR, 2.69; 95% CI: 1.95–3.72) and those reporting a discussion of results or follow-up (arrange) (aOR, 1.63; 95% CI: 1.18–2.24), and lower among patients offered a choice among screening modalities (agree) (aOR, 0.57; 95% CI: 0.37–0.86) as well as among those who wanted more screening information (aOR, 0.65; 95% CI: 0.43–0.97)

[74]

Use of CRC tests was higher among men and women if they had a usual source of care or had talked to a general doctor

[75]

Rates of CRC screening test use were lower for respondents who reported no usual source of healthcare, regardless of Hispanic ethnicity

[60]

Better provider communication was a predictor of CRC screening adherence

[73]

In men, a positive correlate of adherence to CRC screening was having a personal physician (aOR = 2.01; p < 0.01)

[87]

The odds of having the endoscopic procedures increased with the length of time the patients were under the care of their primary care providers and how strongly patients believed that one should have an FS even without symptoms

[66]

In the multivariable logistic regression model, having a regular doctor was significantly associated with current CRC screening status (aOR, 3.8; 95% CI: 1.7–8.3)

[65]

The lowest rate of current CRC screening use was reported by those who had no routine doctor’s visit in the last year (20.3%)

One of the most important modifiable predictors of current CRC screening was routine doctor’s visit in the last year (OR, 3.5; 95% CI: 3.2–3.8)

SIG/COL was more common in persons without routine doctor’s visits in the preceding year (OR, 3.3; 95% CI: 2.8–4)

[3]

Never-screened patients were less likely to have a regular primary physician (80% vs. 95%, p = .0003) compared with screened patients

[79]

Logistic regression results indicated that family history of CRC (aOR, 1.98; 95% CI: 1.02–3.86), receiving a regular medical checkup (aOR, 3.07; 95% CI: 2.00–4.71), current screening by mammography in women and PSA in men (aOR, 4.40; 95% CI: 2.94–6.58), and vitamin supplement use (aOR, 1.87; 95% CI: 1.27–2.77) were significant predictors of CRC screening

Physician recommendation

[72]

Lack of knowledge/awareness and the physician not ordering the test were commonly cited reasons for not having CRC tests. Predictors of receiving a physician recommendation included socio-demographic (younger age, white race, more education), health status (increased CRC risk, co-morbidity), and healthcare access (had a routine/preventive care visit in the past 12 months) factors

[59]

Clinician recommendation for a screening test was the best predictor of CRC screening in two age categories (50–64 years and 65+ years)

It was a very strong indicator and consistently improves the odds of use by a factor of at least 8 for any screening test

[67]

Physician recommendation was a predictor of CRC screening

[12]

Patients with previous experience of CRC screening preferred future screening

[82]

Multiple regression indicated that self-efficacy, physician recommendation, perceived benefits of and perceived barriers to screening accounted for 40% of variance in CRC screening adherence. However, there was no evidence for two mediational models with perceived benefits and perceived barriers as the primary mechanisms driving adherence to CRC screening

[8]

Physician recommendation was a powerful motivator to CRC screening

Use of preventive services

[60]

Being current with other cancer screening was a predictor of CRC screening adherence (mammography (yes vs. no): 61.7% vs. 33.9%, p < 0.001; PSA (yes vs. no): 72.4% vs. 41.5%, p < 0.001)

[73]

In men, adherence to PSA screening (aOR, 3.24; p < 0.001) exerted the largest independent effect on CRC screening adherence

[1]

CRC screening was predicted by having engaged in other preventive cancer screenings in the previous year

[66]

In multivariable logistic regression, participation in a general medical exam was significantly associated with current CRC screening status (aOR, 3.7; 95% CI: 2.1–6.7)

[16]

Underutilization of CRC screening tests was highest in persons who lacked preventive services

[76]

Men are more likely to report having ever been screened for prostate cancer than for CRC. 75% of those aged 50 years or older have had a PSA test vs. 63% for any CRC test (RR, 1.20; 95% CI: 1.18–1.21). Up-to-date PSA screening is also more common than CRC screening for men of all ages

[3]

Never-screened patients were less likely to have undergone a prior screening mammography (87% vs. 96% of women, p = 0.02) compared with screened patients

[79]

Logistic regression results indicated that family history of CRC (aOR, 1.98; 95% CI: 1.02–3.86), receiving a regular medical checkup (aOR = 3.07; 95% CI: 2.00–4.71), current screening by mammography in women and PSA in men (aOR, 4.40; 95% CI: 2.94–6.58), and vitamin supplement use (aOR, 1.87; 95% CI: 1.27–2.77) were significant predictors of CRC screening

Residence

[55]

Residents of urban areas had higher odds (OR 1.19; 95% CI: 1.06–1.34) of being up-to-date with CRC screening than rural residents

[75]

Greater disparity in CRC screening rates between Hispanics and non-Hispanics was observed in Colorado, California, and Texas than in other states

[52]

Approximately 16.2% (95% CI: 15.3% –17.2%) of persons aged ≥50 years who resided in rural areas had received an FOBT test in the past year, compared with 22.0% of those living in the larger metropolitan areas (95% CI: 21.4%–22.7%). About 28.2% (95% CI: 27.1%–29.4%) of those who resided in rural areas had received a SIG or COL in the past 5 years, compared with 35.2% of those in the larger metropolitan areas (95% CI: 34.5%–36.0%)

[76]

In state-level analyses of this age group, men were significantly more likely to be up to date on prostate cancer screening compared with colorectal cancer screening in 27 states, while up-to-date CRC screening was more common in only 1 state

aOR—adjusted odds ratio; CI—confidence interval; COL—colonoscopy; HMO—Health Maintenance Organization; OR—odds ratio; RR—relative risk; FOBT—fecal occult blood test; FS—Flexible Sigmoidoscopy; PSA—prostate cancer screening; SIG—sigmoidoscopy

Table 4

Effects of other objective and subjective influences on colorectal cancer screening behaviors—Summary of articles included in systematic review (1996–2006)

Ref #

Main result

Religion, language, acculturation and other cultural factors

[53]

Comments on CRC screening fell into 10 content areas: primary care site or provider characteristics (25% patient/21% provider comments); knowledge (18% patient/12% provider comments); cost/insurance coverage (10%/25%); attitudes (14%/7%); ordering of priorities (10%/11%); language (12%/7%); procedural issues regarding screenings (8%/10%); discrimination (2%/1%); and issues related to being a new immigrant (.2%/6%)

[70]

Adherence to CRC screening was most strongly associated with family/friend subjective norms about CRC screening use. Emotional family support, but not the size of the networks, was indirectly related to adherence via increased family/friend subjective norms, while emotional friend support was directly related to adherence

[84]

CRC screening was underused in Hispanics compared to non-Hispanic Whites. There was a trend that acculturation level was inversely correlated with having an endoscopy in the past 5 years. This trend was also seen with having a FOBT in the past year or an endoscopy in the past 5 years. However, the association disappeared after adjusting for factors pertaining to utilizing other health care services. Additionally, after stratifying by gender, the association between the two variables was diluted

[60]

Having support from others for screening was a predictor of CRC screening adherence (A lot: 62.9% vs. Somewhat/Very little: 52.9%, p = 0.03)

[81]

No significant differences between users and non-users of FOBT, SIG, and COL in terms of age, gender, insurance status, or language

[67]

Acculturation in terms of language proficiency was a predictor of CRC screening (FOBT in past 2 years: OR, 2.9; 95% CI: 1.3–6.4; SIG/COL in past 5 years: OR, 3.7; 95% CI: 1.3–10.1)

[12]

Catholics were most likely to prefer no CRC screening compared with non-Catholics (Catholics: 69.6% vs. Protestants: 94.6%; Other: 89.0%; p = 0.001)

[80]

Logistic regression models found greater acculturation (aOR, 5.54; 95% CI: 1.85–16.60) to be a significant predictor of having had an FOBT at least once, and found both greater acculturation (aOR, 8.70; 95% CI: 2.07–36.55) and lack of physician recommendation (aOR, 0.59; 95% CI: 0.40–0.89) to be significant predictors of having had a SIG at least once. No significant predictors were found for regular adherence to CRC guidelines, which include having undergone an FOBT in the past year and SIG in the past 5 years

Perceived benefits

[53]

Comments on CRC screening fell into 10 content areas: primary care site or provider characteristics (25% patient/21% provider comments); knowledge (18% /12%); cost/insurance coverage (10%/25%); attitudes (14%/7%); ordering of priorities (10%/11%); language (12%/7%); procedural issues regarding screenings (8%/10%); discrimination (2%/1%); and issues related to being a new immigrant (.2%/6%)

[70]

While usual source of care was directly associated with CRC screening adherence, better provider–patient communication was directly and indirectly associated with adherence via increased perceived benefits

[87]

The odds of having the endoscopic procedures increased with the length of time the patients were under the care of their primary care providers and how strongly patients believed that one should have an FS even without symptoms

[82]

Multiple regression indicated that self-efficacy, physician recommendation, perceived benefits of and perceived barriers to screening accounted for 40% of variance in CRC screening adherence. However, there was no evidence for two mediational models with perceived benefits and perceived barriers as the primary mechanisms driving adherence to CRC screening

[4]

Barriers were significantly negatively related to recent FOBT (aOR, 0.91; 95% CI: 0.86–0.97) and recent SIG (aOR, 0.92; 95% CI: 0.87–0.97). Benefits were significantly related to having a recent SIG (aOR, 1.11; 95% CI: 1.01–1.23) and a recent COL (aOR, 1.27; 95% CI: 1.10–1.48) but not to recent FOBT. All models were adjusted for age, gender and education

Perceived barriers

[53]

Comments on CRC screening fell into 10 content areas: primary care site or provider characteristics (25% patient/21% provider comments); knowledge (18% /12%); cost/insurance coverage (10%/25%); attitudes (14%/7%); ordering of priorities (10%/11%); language (12%/7%); procedural issues regarding screenings (8%/10%); discrimination (2%/1%); and issues related to being a new immigrant (.2%/6%)

[60]

Failure of providers to recommend screening; lack of knowledge about the need for screening; belief that screening was not necessary without symptoms

Fear of cancer; lack of knowledge about screening methods other than COL; reliance on physicians for screening information; and need for people to feel at risk for screening to occur

[86]

The most frequently observed barrier was lack of familiarity with CRC screening guidelines and tests. Availability and accessibility to screening tests was a barrier that was difficult to overcome for many participants. The majority of participants faced two or more barriers

[67]

Perceived psychological costs were a predictor of CRC screening (“Strongly disagree” versus “Strongly agree”; FOBT in past 2 years: OR, 5.5; 95% CI: 1.2–26.0; SIG/COL in past 5 years: OR, 14.0; 95% CI: 2.5–77.9)

[68]

Receipt of a FOBT within the prior 12 months (37.9% of sample) was associated with lower level of worries or fears of test results. Receipt of a flexible SIG within the past 5 years (22.2% of sample) was associated with lower levels of worries or fears of test results

[82]

Multiple regression indicated that self-efficacy, physician recommendation, and perceived benefits of and perceived barriers to screening accounted for 40% of variance in CRC screening adherence. However, there was no evidence for two mediational models with perceived benefits and perceived barriers as the primary mechanisms driving adherence to CRC screening

[8]

The test is not needed; the test is embarrassing

[3]

The four most reported deterrents to screening COL were: “volume of bowel preparation"; “adequate analgesia"; “no recommendation from primary physician" and “embarrassment"

[4]

Barriers were significantly negatively related to recent FOBT (aOR, 0.91; 95% CI: 0.86–0.97) and recent SIG (aOR, 0.92; 95% CI: 0.87–0.97). Benefits were significantly related to having a recent SIG (aOR, 1.11; 95% CI: 1.01–1.23) and a recent COL (aOR, 1.27; 95% CI: 1.10–1.48) but not to recent FOBT. All models were adjusted for age, gender, and education

Knowledge

[72]

Lack of knowledge/awareness and the physician not ordering the test were commonly cited reasons for not having CRC tests. Predictors of receiving a physician recommendation included socio-demographic (younger age, white race, more education), health status (increased CRC risk, co-morbidity), and healthcare access (had a routine/preventive care visit in the past 12 months) factors

[60]

Better knowledge about screening guidelines was a predictor of CRC screening adherence (know at least one: 77.5% vs. Know none: 51.9%, p < 0.001)

[53]

Comments on CRC screening fell into 10 content areas: primary care site or provider characteristics (25% patient/21% provider comments); knowledge (18% /12%); cost/insurance coverage (10%/25%); attitudes (14%/7%); ordering of priorities (10%/11%); language (12%/7%); procedural issues regarding screenings (8%/10%); discrimination (2%/1%); and issues related to being a new immigrant (.2%/6%)

[12]

Patients with knowledge of someone with colon cancer or colon polyps reported a significantly higher preference for screening than those without such knowledge

[3]

Never-screened patients had less understanding of the incidence and treatment outcomes of colon cancer

Susceptibility and Family History

[60]

Greater perceived susceptibility was a predictor of CRC screening adherence (More that average: 87.2% vs. Average: 62.7% vs. Less than average: 57.5%, p < 0.001)

Family history was a predictor of CRC screening adherence (Yes: 77.4% vs. No: 59.0%, p < 0.001)

[1]

CRC screening was predicted by having a family history of CRC

[67]

Susceptibility was related to SIG/COL screening

[68]

Receipt of an FOBT within the prior 12 months (37.9% of sample) was associated with higher level of perceived susceptibility to CRC. Receipt of a flexible SIG within the past 5 years (22.2% of sample) was associated with higher level of perceived susceptibility of CRC

[79]

Logistic regression results indicated that family history of CRC (aOR, 1.98; 95% CI: 1.02–3.86), receiving a regular medical checkup (aOR, 3.07; 95% CI: 2.00–4.71), current screening by mammography in women and PSA in men (aOR, 4.40; 95% CI: 2.94–6.58), and vitamin supplement use (aOR, 1.87; 95% CI: 1.27–2.77) were significant predictors of CRC screening

Health and Lifestyle

[60]

Being physically active was a predictor of CRC screening adherence (Yes: 62.4% vs. No: 61.1%, p = 0.001)

[74]

Use of CRC tests was higher among men and women if they were former smokers

[73]

Failure to adhere to CRC screening was associated with self-reported good health (aOR, 0.87; p < 0.01) and being a current smoker (aOR, 0.65; p < 0.01)

[1]

CRC screening was predicted by having a greater number of chronic illnesses

[12]

Ex-smokers (compared with all others) were more likely to want screening for CRC

[65]

The lowest rate of current CRC screening use was reported by every-day smokers (32.1%)

SIG/COL was more common in persons with poor self-reported health (OR, 1.3; 95% CI: 1.2–1.5)

aOR—adjusted odds ratio; CI—confidence interval; COL—colonoscopy; HMO—Health Maintenance Organization; OR—odds ratio; RR—relative risk; FOBT—fecal occult blood test; FS—Flexible Sigmoidoscopy; PSA—prostate cancer screening; SIG—sigmoidoscopy

Results

Observational studies using a wide range of methodologies found similar predictors of CRC screening behaviors among average-risk older adults. Socio-demographic factors associated with CRC screening include male gender [1, 8, 46, 55, 60, 65, 67, 74, 81, 87], older age [1, 60, 66, 74], marriage [55], higher education [16, 68, 75, 85, 88], higher income [73, 75], urban area of residence [52, 55], and White race or non-Hispanic ethnicity [16, 68, 75, 85]. Lifestyle and health factors that could influence CRC screening include smoking history [73, 74, 77], chronic diseases [1], and family history of CRC [55, 60, 79]. Healthcare access, delivery, and utilization factors associated with CRC screening behaviors include presence of a usual source of care [74], physician recommendation[5, 8, 46, 54, 55, 67, 72, 82, 89], utilization of other preventive health services [16, 66, 73, 74, 84], and health insurance coverage [5, 16, 64, 67, 7375, 80, 85, 89]. Finally, a wide range of psychosocial factors [3, 4, 8, 12, 53, 60, 64, 6770, 72, 8284, 90, 91] were shown to influence the decision to undergo CRC screening among age-eligible men and women.

In a study of 64,048 US older adults who participated in the 1999 BRFSS, compliance with CRC screening guidelines was lowest in individuals 50–64 years of age who reported a lower educational attainment, and for those who lacked health insurance coverage or did not use other preventive healthcare services [16]. In another study [75] based on 5,680 Hispanic and 104,733 non-Hispanic United States older adults who participated in the 2002 BRFSS, a smaller percentage of Hispanics (41.9%) reported having had either FOBT test within the past year or a lower endoscopy (sigmoidoscopy or colonoscopy) within 10 years as compared to non-Hispanics (55.2%). In addition, CRC screening rates were associated with education, income, health insurance, and usual source of health care, regardless of Hispanic ethnicity [75].

Using data from the 2000 NHIS which included 12,677 older adults, James and colleagues [55] found that adherence to CRC screening recommendations was lower for Hispanics and African Americans than Whites, and higher for residents of urban versus rural areas. Among individuals who were not up-to-date with CRC screening, the same disparities existed for receiving CRC screening or recommendations from their physician [55]. Meissner and colleagues[74] used the 2003 NHIS to examine correlates of gender-specific CRC screening test use. In 2003, both men and women reported higher rates of adherence to colonoscopy (32.2% and 29.8%, respectively) than FOBT (16.1% and 15.3%, respectively) or sigmoidoscopy (7.6% and 5.9%, respectively). Men reported higher use of endoscopy than women if they had a usual source of healthcare, had talked to a general doctor, and had two to five visits to the doctor in the past year. Men and women 65 years and older had higher rates of any recommended CRC test (55.8% and 48.5%, respectively) than persons 50–64 years (males, 41.0%; females, 31.4%). Finally, adequate use of CRC screening tests also was higher among both genders if they were not Hispanic, had higher educational attainment, were former smokers, had health insurance, or a usual source of care, or if they talked to a general doctor [74].

A telephone survey of CRC screening behavior was conducted among 954 Massachusets residents aged 50 years and older. The main predictors of current CRC screening participation were family history, receiving a regular medical checkup, current screening by mammography in women, and PSA in men [79]. Another population-based study was conducted by Fisher and colleagues in North Carolina to determine if CRC screening rates are different between Blacks and Whites, after controlling for potential confounders. They found that the main predictors of compliance with CRC screening guidelines were older age, presence of a regular doctor, and participation in a general medical exam [66].

A cross-sectional study by Sheikh and colleagues described CRC screening preferences among a convenience sample of 193 adults attending internal medicine and family practice clinics at a community teaching hospital. While 55% of respondents preferred sigmoidoscopy and FOBT, 29% chose colonoscopy and 16% rejected CRC screening. Older age, non-Catholic status, knowledge of someone with colon cancer or colon polyps, ex-smoker status, and previous experience with CRC screening were identified as positive predictors of patient preference for CRC screening [12].

Equitable access to CRC screening tests is a major concern among age-eligible older adults in the United States [5]. Since the year 2001, colonoscopy is covered by Medicare for US men and women aged 65 years and over. However, colonoscopy remains a more expensive alternative for socioeconomically disadvantaged and uninsured individuals below the age of 65 years [3]. Thus, one of the major perceived barriers to CRC screening for individuals between the ages of 50 and 64 years is the lack of health insurance coverage for CRC screening procedures, especially for colonoscopy.

Several studies have focused on the role of health insurance coverage on adherence to CRC screening guidelines. For instance, a cross-sectional study of 104 patients attending three Midwestern medical clinics found that insured patients were significantly more likely than the uninsured to have ever completed any type of CRC testing (77% vs. 33%) and to have undertaken testing according to current guidelines (62% vs. 17%). In addition, insured individuals were more likely to know about CRC screening and to report physician recommendations [5]. Other studies have restricted their analyses to either insured [72] or uninsured [1] patients. A telephone survey of 1,901 Medicare consumers in North and South Carolina showed that 31% had never been tested for CRC and 18% were tested but were not current with CRC screening. Overall, 28% did not receive any physician recommendation for CRC. Younger age, White race, higher education, increased CRC risk, co-morbidity, and having had a routine or preventive care visit in the past 12 months were important predictors of physician recommendation [72]. In a medical record review of 1,176 uninsured patients attending eight community health centers in Florida, the following factors predicted CRC screening according to recommendations: male gender, older age, being African American, having a greater number of chronic illnesses, a family history of CRC, and having engaged in other preventive cancer screenings in the previous year [1].

Many investigators have adopted theoretically grounded models in an attempt to understand subjective influences of CRC screening behavior in age-eligible average-risk individuals in the United States. Theoretical models that were frequently examined in recent studies include the Health Belief Model (HBM), the Transtheoretical Model (TTM), the Patient/Provider/System Theoretical Model, the Precede–Proceed Model, and the Social Support Networks Model. Other studies used components of theoretical models such as perceived barriers, perceived benefits, acculturation, and health literacy skills. In particular, the term “barrier” was loosely defined to reflect conditions that may preclude CRC screening such as lack of health insurance coverage, or conditions that may discourage CRC screening such as lack of physician recommendation as well as personal attitudes and beliefs vis-à-vis CRC testing.

The most commonly adopted theoretical model is the HBM, a well-established theory that targets the intrapersonal level of influence. It has been shown to be a good fit for preventive health behaviors undertaken by people who are well and not experiencing signs or symptoms for the purpose of remaining well, including both medical and non-medical activities. HBM was originally developed to explain why people did not take advantage of screening and immunization programs. The following are HBM constructs that were often used in primary and secondary cancer prevention programs: (1) Perceived susceptibility (belief regarding chance of getting a condition), (2) Perceived severity (belief in seriousness of condition), (3) Perceived threat (collective effect of perceived susceptibility and perceived severity), (4) Perceived benefits (beliefs concerning the effectiveness of taking a particular action in reducing illness threat or producing other desirable outcomes), (5) Perceived barriers (beliefs about the negative aspects of taking a particular action), (6) Cues to action (internal and external stimuli that remind the person of the need to change and that trigger action), and (7) Self-efficacy (one’s beliefs about their ability to successfully accomplish the specified behavior). In addition, individual characteristics such as age, gender, ethnicity, socioeconomic status, and knowledge are important effect modifiers for the influence of constructs described above on CRC screening behaviors [92].

Many studies have attempted to identify factors that can predict CRC screening based on the HBM. In a study of 300 outpatients at the Mayo Clinic, four perceived barriers to colonoscopy were identified: “volume of bowel preparation,” “adequate analgesia,” “no recommendation from primary physician,” and “embarrassment.” Patients who had a colonoscopy also had better knowledge of CRC incidence and treatment outcomes [3]. Similarly, two major perceived barriers to CRC screening (“the test is not needed,” “the test is embarrassing”) were identified in a study of 355 White and Black older adults who participated in a telephone survey in Michigan [8].

In a recent study by Tessaro and colleagues [60], members of 16 Appalachian churches participated in a self-administered survey (n = 839) and 23 focussed on group discussions (n = 205). Adherence to CRC screening guidelines was facilitated by older age, male gender, being current for other cancer screening, physical activity, perceived support from others for screening, better provider communication, knowledge about screening guidelines, perceived susceptibility to CRC, and family history of CRC. Some of the perceived barriers to CRC screening adherence included “failure of providers to recommend screening,” “lack of knowledge about the need for screening,” “belief that screening was not necessary without symptoms,” “fear of cancer”, “lack of knowledge about screening methods other than colonoscopy,” “reliance on physicians for screening information,” and “need for people to feel at risk for screening to occur” [60].

A convenience sample of 15 African-American women between 54 and 69 years of age was identified from a moderately large church in a major urban city. The following perceived barriers to CRC screening were reported: “fear,” “afraid of pain,” “doctor never recommended any tests,” “had no symptoms,” “competing health histories,” “embarrassment,” and “flaws in risk perception” [64]. A survey examining the relationship between HBM-perceived barriers and benefits and self-reported CRC screening behavior was conducted among African American adults in North Carolina. As expected, perceived barriers were negatively associated with recent FOBT or sigmoidoscopy, whereas benefits were significantly related to having a recent sigmoidoscopy or colonoscopy but not a recent FOBT [4] .

The impact of patient–physician CRC screening discussions on screening use was examined through a mailed survey and retrospective claims data among 4,966 insured primary care patients aged 50–70 years. The survey collected information on patient–physician CRC screening discussion content (5A’s: assess, advise, agree, assist, and arrange). The likelihood of CRC screening was increased among patients who reported to help scheduling an appointment, those who reported that their physician was ready to discuss results or follow-up, and lower among patients offered a choice among CRC screening modalities or those who wanted more CRC screening information [83].

The role of social support networks was assessed in a population-based sample of 341 Japanese Americans aged 50 year and over. A structural equation model showed that CRC adherence was most strongly associated with family/friend subjective norms about CRC screening use. Emotional family support was indirectly associated with adherence via increased family/friend subjective norms, while emotional friend support was directly related to CRC adherence [70].

Conclusion

In summary, most of the studies reviewed were theoretically grounded, and many of them (explicitly or implicitly) measured determinants of CRC screening behavior that were borrowed from the HBM. Objective influences of CRC screening non-adherence were female gender, young age ( <65 years), racial or ethnic minority status (African-American, Hispanic, Asian), less years of education, and lack of health insurance coverage. One HBM construct that was consistently shown to influence CRC screening was perceived barriers (e.g., fear, embarrassment, lack of physician recommendation). Other HBM constructs such as perceived susceptibility (e.g., family history), cues to action (e.g., presence of physician recommendation), and self-efficacy (e.g., use of preventive health services) were only implicitly examined in the selected observational studies. Finally, knowledge of CRC incidence, mortality, and screening effectiveness appear to be important psychosocial predictors of CRC screening initiation and compliance.

Physician recommendation appears to be a major influence on initiation of and compliance with CRC screening. The literature suggests that the presence of this health access, delivery, and utilization characteristic is a cue to action, while the absence of it is a barrier. Another important characteristic is absence of health insurance coverage, which is perceived by older adults as a barrier to CRC screening. However, whereas many studies have attempted to explain racial and ethnic disparities in CRC screening, only few studies have attempted to disentangle the effects of health insurance coverage and other socioeconomic indicators such as income.

Most of the CRC behavioral change studies identified in the literature did not distinguish between sub-types or circumstances of CRC screening behaviors, such as initiation of CRC screening versus adherence to CRC screening guidelines. Whereas initiation is a one-time activity, adherence is a combination of a number of activities that either fit or do not fit existing recommendations. It is plausible that different HBM components may be influencing CRC screening initiation and CRC screening adherence. In particular, the nature of the perceived benefits and barriers is likely to differ among individuals who have never received CRC screening versus those who have done so in the past. In addition, it can be argued that self-efficacy plays a different role among those who received CRC screening in the past as compared to those who did not. Specifically, those without past experience with CRC testing will be starting out with low self-efficacy and will require a greater increase in self-efficacy to adhere, although they may not achieve a degree of self-efficacy as high as an individual who received CRC testing in the past. By contrast, those who received a CRC test in the past may not need to experience a further increase in self-efficacy [9294].

Another important issue is using FOBT, flexible sigmoidoscopy, and colonoscopy as an outcome measure for studies of screening behavior, when these tests are also used for diagnosis of colorectal symptoms. In most studies, the reason for getting tested is not known. One potential effect of this problem is to overestimate the impact on screening behavior of predictive factors such as increasing age and male gender, which may also be related to increased likelihood of receiving diagnostic testing. In most cases, completion of a test for diagnostic purposes achieves the same effect as completing it for screening purposes. Therefore, older individuals may require less effort to promote screening simply because they get diagnostic testing frequently.

The behavioral change literature consists predominantly of cross-sectional studies that examine knowledge, attitudes, and beliefs as correlates of CRC screening behaviors. Thus, very few of these studies had a longitudinal component that would establish the temporal relationship of behavioral influences with initiation and adherence to CRC screening guidelines [95]. Furthermore, many of these studies adopted the HBM which is well suited to a one-time activity such as immunization, and does not explicitly take into consideration the fact that certain behaviors are adopted over time as a result of movement from one stage of behavioral change to the next, and that these behaviors require maintenance. This HBM limitation can be addressed through incorporation of constructs from the Transtheoretical Model [93, 94].

In conclusion, evidence suggests that CRC screening is a complex behavior with multiple influences including socio-demographic characteristics, physician–patient communication, health insurance coverage, and psychosocial factors. Future intervention studies need to focus on increasing awareness of CRC screening guidelines and accessibility to CRC screening tests among average-risk older adults in the United States. In addition, behavioral interventions need to be tailored to specific communities of older adults who were less likely to initiate and comply with CRC screening. These include African-American, Asian, and Hispanic communities, individuals who report a lower educational attainment and the uninsured.

Acknowledgments

The authors would like to acknowledge the Hardin Library for Health Sciences at the University of Iowa and the Welch Library at Johns Hopkins Medical Institutions for providing the needed references and sources. Conflict of Interest: None.

Copyright information

© Springer Science+Business Media B.V. 2007