Colorectal Cancer Screening Attitudes and Practices in the General Population: A Risk-adjusted Survey

Colorectal Cancer Screening Attitudes and Practices in the General Population: A Risk-adjusted Survey Walter L. Straus, Edward C. Mansley, Karen F. Gold, Qin Wang, Prabashni Reddy, and Chris L. Pashos Objectives: To characterize self-reported colorectal cancer (CRC) screening behavior, and to identify characteristics of CRC screening practices, stratified by risk. Methods: Using random-digit-dial methodology, we conducted telephone surveys in US adults 50 years of age and older. Respondents provided data on utilization of CRC screening tests; demographic characteristics; and awareness, concerns, attitudes and beliefs about the tests, CRC, and health care. On the basis of available guidelines, three definitions of adequate screening were considered. Results: Among persons reporting having ever had a CRC screening exam, the exam was more likely to have been a fecal occult blood test than a radiologic or endoscopic exam (p <.0001). Subjects at increased CRC risk were more likely to have met the screening criteria (p <.001) compared with average-risk subjects. Receipt of information or advice about cancer screening tests, male gender, and concern about managed care were positively associated with adequate screening. Smoking, low health self-monitoring, and an average risk for CRC reduced the probability of CRC screening. Conclusions: Lack of awareness about screening remains common, regardless of CRC risk. Providing information and advice about cancer screening may be the single most important tool available to improve screening rates. KEY WORDS: cancer prevention, cancer screening, colorectal cancer Colorectal cancer (CRC) is the second most common cause of death from cancer in the United States, accounting for 10 percent of all cancer mortality. Early detection of CRC can dramatically increase an individual s survival rate. Data from the National Cancer Institute show that 90 percent of patients who are diagnosed early will survive beyond 5 years. In contrast, 8 percent of patients J Public Health Management Practice, 2005, 11(3), 244 251 C 2005 Lippincott Williams & Wilkins, Inc. diagnosed after their disease has reached an advanced state will survive beyond 5 years. 1 Currently, CRC prevention has a larger number of distinct screening modalities than any other cancer. These methods include the fecal occult blood test (FOBT); endoscopic tests, including flexible sigmoidoscopy (SIG) and colonoscopy (COL); and doublecontrast barium enema (DCBE). 2 Furthermore, the arsenal of screening tools is expanding with recent advances in virtual colonoscopy and stool-based DNA assays. 3 Despite the proven value and diversity of screening procedures, screening rates remain low. Recent data from the US Centers for Disease Control and Prevention (CDC) show that only 40.3 percent of US residents more than age 50 have ever received a home-administered FOBT and only 43.8 percent have ever received an SIG or a COL. Furthermore, only 20.6 percent of those aged more than 50 had a home-administered FOBT within the past year and only 33.6 percent had an SIG or a COL within the past 5 years. 4 In contrast, screening for cervical cancer and breast cancer, other cancers for which there is also evidence that screening reduces mortality, approaches 70 80 percent. 5 The public health community has sought to understand the reasons for these low screening rates. In this This research was sponsored by Merck & Co, Inc. Corresponding author: Walter L. Straus, MD, MPH, Merck Research Laboratories, PO Box 4, BL 1-7, West Point, PA 19486 (e-mail: walter straus@merck.com). Walter L. Straus, MD, MPH, is Senior Director, Epidemiologic Research, Merck Research Laboratories, West Point, Pa. Edward C. Mansley, PhD, is Health Economist, Outcomes Research & Management, Merck & Co, Inc, West Point, Pa. Karen F. Gold, PhD, was formerly Director of Biostatistics and Outcomes Research, Abt Associates, Health Economic Research & Quality of Life Evaluation Services (HERQuLES), Bethesda, Md. Qin Wang, MA, is Lead Programmer Analyst, Abt Associates, HERQuLES, Bethesda, Md. Prabashni Reddy, PharmD, is Associate Director of Clinical Outcomes Research, Abt Associates, HERQuLES, Lexington, Mass. Chris L. Pashos, PhD, Executive Director and Vice President, Abt Associates, HERQuLES, Lexington, Mass. 244

A Risk-adjusted Survey 245 regard, studies involving surveys and focus group interviews have been conducted to uncover facilitators and barriers to CRC screening related to the knowledge, attitudes, and practices of physicians and health care providers, 6 11 as well as those related to the knowledge, attitudes, and behaviors of the general population. 12 18 Recent population surveys, however, have been limited typically to small groups usually restricted to certain geographic, age, and ethnic profiles. Although large, national surveys of colorectal cancer screening have been performed, the effects of only a limited number of health behavior factors on colorectal screening have been assessed. The most commonly used CRC screening guidelines provide different recommendations based on the patient s underlying CRC risk. However, to our knowledge, none of these surveys has compared screening experiences for those at average versus those at elevated risk for CRC. 5,19 The primary objectives of the study presented in this paper are to characterize self-reported CRC screening practice in the United States under different definitions of adequate screening ( screening criteria ), and to better understand the population-reported barriers and facilitators associated with adequate screening. Methods We conducted a national random-digit-dial telephone survey of US households in 2000. Responding households with no adults 50 years and older were not surveyed. Trained interviewers conducted this computerassisted telephone survey. An empirically derived sample of 500 patients was targeted for the study. Up to 12 attempts were made to reach individuals during a random cross-section of the day (eg, afternoon and evening) and over the course of a week. From each respondent, we collected information on demographic characteristics, CRC risk factors, and attitudes toward and personal screening experience with FOBT, SIG, COL, and DCBE. The survey consisted of 60 questions and the interview lasted approximately 30 minutes. Subjects reported the year in which they received the screening tests (if screened). We defined the year in which these tests were conducted as the most recent year in which they received the test. Because our survey was conducted in the middle of the year, a half-year extension was added to the criteria screening intervals (eg, 1 1 / 2 years vs 1 year for FOBT). On the basis of the respondents answers to questions about personal and family medical history, we classified them as at average or increased risk for CRC. 20,21 Subjects were classified as being at increased risk if they reported having a personal history of polyps, colitis, Crohn s Disease or CRC, or having a blood relative with CRC. We also collected and assessed self-reported data on the subjects awareness, concerns, attitudes, and beliefs about CRC, CRC tests, health, and the health care system that might represent facilitators of, or barriers to, the screening tests. Categories of variables considered to be potential barriers or facilitators included (1) demographics; (2) health-related concerns (eg, concern about quality of care, concern about the disease); (3) receipt of information about cancer screening; (4) specific CRC cancer screening beliefs (eg, CRC is preventable, early detection improves survival, CRC is a public or personal health threat); (5) healthy and risky behaviors (eg, alcohol intake, smoking); (6) insurance coverage; (7) employment status; and (8) lack of knowledge of or negative feelings toward CRC and cancer screening tests (eg, discomfort, unpleasantness), and access to care. This instrument was adapted from others used in cancer screening research. 22 24 After a review of recommended CRC screening practices, we identified six sets of screening practice criteria that corresponded to existing research or public health based recommendations. Three sets pertained to subjects at average risk and three to those at increased risk. We focused on criteria broadly representative of the range of existing recommendations. Specifically, we selected the following guidelines for CRC screening: Criterion I, relatively lax requirements; Criterion II, intermediate requirements; and Criterion III, relatively stringent requirements. The guidelines on which these criteria are based are summarized in Table 1. Respondents were evaluated as to whether their screening practices were adequate, on the basis of their conformance with these criteria. To evaluate conformance with Criterion I, subjects were classified by their screening practices into one of four categories (A, B, C, and D, shown in Table 2), adapted from previous work. 2 Screening practices B, C, and D, involving combinations of FOBT, SIG, DCBE, or COL, are considered adequate for subjects at average risk (Table 2). The only difference in the screening behavior criteria based on risk level was that the COL requirement involved a COL every 10 years for subjects at average risk rather than every 5 years for those at increased risk. 2,20 21 Criterion II reflects a stricter definition of adequate screening, involving a COL every 10 years for subjects at average risk or an alternative annual FOBT in conjunction with either an SIG or a DCBE every 5 years. As with Criteria I, the behavior requested of individuals at increased risk is comparable except that the COL should have occurred within the past 5 years. 20,21 Criterion III, involving an even stricter definition of adequate screening, was based on the American College of Gastroenterology s (ACG) preferred CRC

246 Journal of Public Health Management and Practice TABLE 1 Summary of CRC screening guidelines used to develop Criteria I, II, and III American Gastroenterological Association, 1997 (endorsed by multiple organizations) 2 American College of Gastroenterology, 2000 20,21 Average risk ( 50 years and no other risk factors for CRC) FOBT Annually SIG Every 5 years SIG + FOBT Every 5 years + annually Every 5 years + annually (alternative strategy) DCBE Every 5 10 years Every 5 years + annual FOBT (alternative strategy) COL Every 10 years Every 10 years Increased risk Close relative with CRC or adenomatous polyp As for average risk but begin screening at COL every 3 10 years beginning at age 40 or 10 years 40 years younger than age of youngest affected relative Family history of adenomatous polyposis SIG annually beginning at puberty SIG every 1 2 years beginning at 10 12 years Family history of hereditary nonpolyposis CRC COL every 1 2 years age 20 30 years and COL every 2 years beginning at age 20 25; annually after age 40 annually after age 40 History of adenomatous polyps COL every 3 years COL every 3 5 years COL = colonoscopy; CRC = colorectal cancer; DCBE = double-contrast barium enema; FOBT = fecal occult blood test; SIG = flexible sigmoidoscopy. screening strategy. It includes a COL every 10 years for subjects at average risk and every 5 years for subjects at increased risk. This criterion for adequate screening behavior does not allow the alternative use of tests permitted in Criterion II. 20,21 Analyses were conducted using SAS version 6.12. Descriptive statistics and simple tabular counts were performed to describe the sample, classify subjects, evaluate different screening criteria, compare screening behavior classification systems, and compare responses of subjects at average risk with those at increased risk on key variables. Chi-square and t tests were implemented to assess statistical significance. Three logistic regressions were constructed to assess the association between adequate screening behavior (as defined above) and the potential barriers and facilitating factors. Within each of the three sets of criteria, adherence to CRC screening was based on CRC risk (ie, average risk and high risk). 20,21 The data contained very few missing values. Missing values were imputed using mean or median values. Twenty-two subjects could not be classified as being at either average risk or increased risk because of missing data and were included in the average risk group. No imputation was done for screening behavior. Small differences in sample sizes presented in descriptive tables and logistic regressions are attributable to the exclusion of subjects not reporting relevant information. Results Of 4,546 households contacted, 1,897 refused to participate in the 30-minute survey. We then excluded 2,147 TABLE 2 Criteria classification for screening behavior of subjects at average risk Is the screening Criteria behavior adequate? Criterion I, includes FOBT only, FOBT plus other test, and non-fobt only categories A Not B, C, or D B FOBT in the past 1 1 / 2 years, but neither an SIG or a DCBE in the past 5 1 / 2 years nor a COL in the past 10 1 / 2 years C NoFOBT in the past 1 1 / 2 years, but either an SIG or a DCBE in the past 5 1 / 2 years or a COL in the past 10 1 / 2 years D FOBT in the past 1 1 / 2 years, and an SIG or a DCBE in the past 5 1 / 2 years or a COL in the past 10 1 / 2 years Criterion II, includes colonoscopy or FOBT plus other test A Not BorC B COL in the past 10 1 / 2 years C FOBT in the past 1 1 / 2 years AND either an SIG or a DCBE in the past 5 1 / 2 years Criterion III, includes colonoscopy only A Not B B COL in the past 10 1 / 2 years COL = colonoscopy; CRC = colorectal cancer; DCBE = double-contrast barium enema; FOBT = fecal occult blood test; SIG = flexible sigmoidoscopy. Note: Colonoscopy interval for subjects at increased risk was 5 1 / 2 years. No No No

A Risk-adjusted Survey 247 TABLE 3 Demographic characteristics of sample by risk category Average Increased Overall, risk, n = 306 risk, n = 176 N = 482 (63.5%) (36.5%) (100%) Age 51 64 183 (59.8%) 83 (47.2%) 266 (55.2%) 65 79 91 (29.7%) 66 (37.5%) 157 (32.6%) 80+ 32 (10.5%) 27 (15.3%) 59 (12.2%) Primary ethnicity African American 25 (8.2%) 15 (8.5%) 40 (8.3%) Hispanic or Latino 10 (3.3%) 5 (2.8%) 15 (3.1%) White 251 (82.0%) 146 (83.0%) 397 (82.4%) American Native/Alaskan 13 (4.3%) 5 (2.8%) 18 (3.7%) Asian/Pacific Islander 2 (0.7%) 2 (1.1%) 4 (0.8%) Other 5 (1.6%) 3 (1.7%) 8 (1.7%) Gender Male 90 (29.4%) 46 (26.1%) 136 (28.2%) Female 216 (70.6%) 130 (73.9%) 346 (71.8%) p =.024 vs average risk. (47.2%) as there was no resident 50 years or older. Of the households contacted with an individual 50 years or older, 21 percent elected to participate in the survey (N = 502), of whom 482 were 51 years or older. Subjects 50 years old and those not reporting age were excluded because conformance with age-specific screening recommendations could not be assessed with certainty for those individuals. Therefore, responses of the 482 participants who were 51 years of age or older were analyzed. Overall, 36.5 percent (n = 176) of the subjects fulfilled one or more of the criteria that identified them as being at increased risk for CRC. The most common risk factor, personal history of polyps, was present in 21.6 percent of the overall sample and in 59.1 percent of subjects at increased risk. Nearly half (49.4%) of the population at increased risk reported having a blood relative with CRC, while 11.9 percent of them reported history of colitis or Crohn s disease, and 5.1 percent of them reported personal history of CRC. Demographic characteristics, stratified by CRC risk category, are shown in Table 3. Similar distributions for ethnicity and gender are present across the two risk groups. Compared with subjects at average risk, significantly fewer of those at increased risk are in the youngest age group (p =.024). The proportion of men is not significantly different between the two risk groups (p = 0.44). Of the 502 patients who completed the survey, 20 percent were from the Northeast, 20 percent from the West, 35 percent from the South, and 25 percent were from the Midwest (US 2000 Census data: 19% Northeast, 22% West, 36% South, and 23% Midwest). For both risk groups, the FOBT was the test most frequently received, regardless of CRC risk (p <.0001) TABLE 4 Rates of ever received screening by risk category and screening test Average-risk Increased-risk Colorectal cancer subjects subjects screening test (n = 306) (n = 176) p-value Fecal occult blood test 67.1% 80.7%.003 Flexible sigmoidoscopy 43.0% 69.2%.0001 Colonoscopy 22.5% 73.6%.0001 Double-contrast barium 28.6% 58.6%.0001 enema p <.0001 fecal occult blood test vs other colorectal cancer screening tests. (Table 4). Persons at increased CRC risk were significantly more likely than those at average risk to have ever had each of the four types of screening test. For each of the three screening categories, respondents at increased risk were more likely to have received adequate screening (p <.001) than those at average risk for CRC (Table 5). Moreover, the data also suggest that the screening strategy usually recommended for individuals at increased risk (a COL every 5 years) was not followed by a large percentage of those individuals, as 43.7 percent reported not having a colonoscopy within the past 5 1 / 2 years. After adjustment for covariates, six variables were found to be significantly associated with adequate screening (p <.05) in at least one of the three regressions; five variables were significant in two to three of the three regressions, suggesting robustness of results to variations in the definition of adequate screening (Table 6). Persons who were less concerned about their health (as reflected by low health self-monitoring), who smoked, or who were at average risk for CRC were less likely to have satisfied CRC screening criteria. Conversely, male respondents, those who expressed concerns about managed care, or who had ever received information/advice about cancer screening were more likely to have satisfied the screening criteria. Discussion To our knowledge, this is the first study to assess adherence to CRC screening using different sets of screening TABLE 5 Adequate screening behavior by risk level and screening criteria Screening Average-risk Increased-risk criteria subjects (n = 306) subjects (n = 176) p-value Criteria I 48.7% 77.8%.001 Criteria II 23.5% 64.2%.001 Criteria III 15.7% 56.3%.001

248 Journal of Public Health Management and Practice TABLE 6 Association of personal factors with adequate screening behavior for subjects at pooled risk (multivariate logistic regressions) N (%) of subjects conforming with criteria for adequate screening Criterion I Criterion II Criterion III 283 (60%) 183 (39%) 145 (31%) Demographics Male 1.90 (1.11, 3.27) 1.36 (0.79, 2.33) 1.04 (0.59, 1.84) African American/Hispanic 0.69 (0.32, 1.51) 0.98 (0.43, 2.26) 1.02 (0.43, 2.40) Age 1.01 (0.98, 1.05) 1.00 (0.97, 1.04) 1.01 (0.97, 1.04) Health concerns Concern about quality of care 0.96 (0.78, 1.18) 1.03 (0.83, 1.26) 0.96 (0.77, 1.20) Concern about cost of medication 1.18 (0.94, 1.49) 1.14 (0.91, 1.42) 0.95 (0.74, 1.21) Concern about managed care 1.09 (0.91, 1.30) 1.32 (1.10, 1.59) 1.21 (1.01, 1.47) Low health self-monitoring 0.80 (0.70, 0.92) 0.84 (0.72, 0.98) 0.93 (0.80, 1.08) Disease concern 0.95 (0.90, 1.01) 0.95 (0.89, 1.01) 0.96 (0.90, 1.02) Information Ever received about cancer screening tests 2.10 (1.33, 3.31) 2.40 (1.49, 3.88) 1.61 (0.98, 2.65) Cancer screening beliefs CRC preventable via screening 0.98 (0.77, 1.25) 0.88 (0.68, 1.14) 0.85 (0.65, 1.12) Early detection improves CRC survival 0.90 (0.56, 1.45) 1.04 (0.62, 1.74) 1.14 (0.68, 1.92) CRC serious public health threat 0.99 (0.85, 1.16) 1.00 (0.86, 1.18) 1.10 (0.94, 1.30) CRC serious personal threat 0.88 (0.74, 1.04) 0.92 (0.76, 1.11) 0.95 (0.79, 1.15) Health behaviors Smokes 0.48 (0.26, 0.88) 0.60 (0.31, 1.18) 0.47 (0.22, 0.99) Moderate alcohol intake (reference = no alcohol) 1.20 (0.69, 2.07) 1.24 (0.72, 2.14) 0.88 (0.50, 1.56) Heavy alcohol intake (reference = no alcohol) 0.67 (0.37, 1.22) 0.85 (0.45, 1.58) 0.67 (0.35, 1.30) Insurance status Medicare/Veterans Affairs 0.97 (0.52, 1.81) 1.38 (0.73, 2.62) 1.25 (0.65, 2.41) Medicaid 0.80 (0.27, 2.33) 0.60 (0.18, 1.98) 0.54 (0.16, 1.81) Managed care/private 0.82 (0.42, 1.61) 0.88 (0.43, 1.80) 0.84 (0.40, 1.75) Unknown 0.38 (0.10, 1.46) 0.46 (0.11, 2.04) 0.61 (0.14, 2.59) Employment status (reference = full-time) Part time 1.25 (0.55, 2.82) 1.71 (0.75, 3.93) 1.24 (0.51, 3.03) Homemaker 0.75 (0.33, 1.72) 1.32 (0.55, 3.15) 1.75 (0.72, 4.28) Retired 1.26 (0.63, 2.50) 1.02 (0.50, 2.10) 1.18 (0.55, 2.53) Other employment categories 0.57 (0.25, 1.26) 0.86 (0.37, 2.01) 1.54 (0.64, 3.71) Dimensions of knowledge, negative feelings, and access SIG, COL, DCBE Knowledge 0.94 (0.86, 1.03) 0.97 (0.88, 1.06) 0.97 (0.88, 1.07) Fear, discomfort or bothered 1.00 (0.94, 1.06) 1.04 (0.98, 1.11) 1.01 (0.95, 1.08) Cost, insurance, and scheduling adequate 1.00 (0.93, 1.07) 1.02 (0.95, 1.10) 1.03 (0.95, 1.12) FOBT Knowledge 1.00 (0.92, 1.10) 1.03 (0.94, 1.13) 1.04 (0.94, 1.14) Feels fear, discomfort or bothered 0.95 (0.90, 1.01) 0.95 (0.90, 1.01) 0.97 (0.91, 1.03) Cost, insurance, and scheduling adequate 1.05 (0.97, 1.13) 1.00 (0.93, 1.08) 0.99 (0.91, 1.07) Doctor recommended any CRC screening test 0.9 (0.78, 1.04) 0.88 (0.75, 1.04) 0.89 (0.75, 1.06) Average risk 0.29 (0.18, 0.47) 0.18 (0.11, 0.28) 0.17 (0.11, 0.27) COL = Colonoscopy; CRC = colorectal cancer; DCBE = double-contrast barium enema; FOBT = fecal occult blood test; SIG = flexible sigmoidoscopy. Figures in bold indicate variables that are statistically significant. Odds ratio for CRC screening for these variables reflects a 1-point change along a 5-point scale.

A Risk-adjusted Survey 249 criteria. It also appears to be the first survey to compare CRC screening rates in patients at average and increased risk, and to assess potential barriers and facilitators, controlling for the risk for CRC. Our first objective, to characterize self-reported CRC screening behaviors under several criteria, required us to develop a tailored methodology. In the past, there has been a profusion of recommendations for CRC screening, which are typically based on underlying CRC risk, but vary in the recommended screening modalities and frequency of testing. The approach consisted of assessing adequate screening behavior by first creating a set (or sets) of screening behaviors and then assigning criteria evaluations as to whether or not each of the screening behaviors in the set met our standards as an adequate screening behavior. Complicating the identification of barriers responsible for low screening rates is the variety of screening behavior recommendations offered by the cancer research and public health communities at the time this survey was conducted, and the competing evidence justifying alternative recommendations. 25 For example, recommendations 2 endorsed by several medical organizations suggested that an annual FOBT is adequate screening behavior for people at average risk while the ACG had proposed a more aggressive approach. 20,21 Given the controversy as to what qualifies as an adequate screening behavior, the methodology we adopted permitted an examination of how different classifications of screening behaviors empirically relate to one another and how changes in the criteria of adequate screening affect classification of subjects. Furthermore, this methodology provided us a mechanism to assess the robustness of claims regarding barriers to, and facilitators of, CRC screening, by assessing the impact of those factors in a variety of adequate screening behaviors. Ultimately, the methodology permits researchers with different notions of what constitutes adequate screening to incorporate these multiple perspectives. Recently published recommendations for individuals at average risk include a broad range of options, similar to Criterion I. 26 28 The latest guidelines for individuals at increased risk for CRC are more varied; the USPSTF suggests screening at an earlier age while the American Gastroenterological Association (AGA) recommends a more individualized approach dependent on specific familial risk factors. 27,28 It is difficult to determine whether the proportion of subjects classified as being at increased risk in this study (36.5%) is excessive. The prevalence of polyps alone may be as high as 60 percent in industrialized nations. 29 This suggests that while a large proportion of subjects were found to be at increased risk, the proportion may not be unusual. However, the possibility also exists that the proportion at increased risk was unusually high because of selection bias those at increased risk were more likely to have participated in the survey. The self-reported screening rates we found for subjects at average risk are similar to those recently reported by the CDC. 4 In particular, 1999 data from the CDC s Behavioral Risk Factor Surveillance System found that 44.0 percent of respondents aged 50 or older (n = 63,555) had received a home-administered FOBT in the past year and/or an SIG or a COL in the past 5 years compared with 41 percent in 1997. The proportion of those respondents who were at increased risk is unknown. Nevertheless, we applied similar criteria to respondents in our survey and came up with a screening rate of 48.7 percent for patients at average risk (77.8% for those at increased risk). There are some reasons why one would expect the higher rate we observed. Our survey occurred 1 year later, dealt with respondents 51 years or older, and allowed for slightly more relaxed time constraints (1.5 years for the FOBT and 5.5 years for the SIG and COL). We also allowed for a DCBE. Nevertheless, the reader should be aware that selection bias may have produced a sample whose characteristics are not readily generalizable to the overall public s (ie, agreater proportion of respondents at increased risk). This may not have been as much of an issue for the CDC survey. Importantly, our sample was representative of the general population with respect to age, gender, and ethnicity. (US population: 55% female; 81% white; 55% 50 64 years, 35% 65 79 years, 11% 80+ years). Also, self-reports of screening practices may be subject to error due to recall bias, social desirability (preference to appear to conform), and low motivation for determining the exact month and date of screening. However, a recent study in an HMO population found that patient self-report for colorectal cancer screening was reliable, with sensitivity ranging from 89 to 96 percent and the specificity varying from 86 to 97 percent, depending on which test was used. 30 Reassuringly, screening rates for those at increased risk were higher than those for people at average risk. Since many persons defined as being at increased risk became so defined as a result of an abnormal initial screening test (ie, at a time when they were at average risk), this is not surprising. Nevertheless, substantial room for improvement exists for both groups. According to Criterion I, which is similar to the most recent recommendations, more than 51 percent of the respondents at average risk reported that they have not engaged in any adequate screening behavior. Furthermore, the screening strategy generally recommended for individuals at increased risk (a COL every 5 years) at the time the survey was conducted, was not followed by a large percentage of those individuals, as 43.7 percent reported not having a colonoscopy within the past 5 1 / 2 years. On the basis of this study s findings, it is

250 Journal of Public Health Management and Practice suggested that educational programs may be a useful approach to enhancing screening in this group. The significance of gender in this study that men appear more likely to conform to screening criteria than women appears to be counter to the conventional wisdom with respect to health-promotion activities. This may be due to a strong self-selection effect, where only men interested in cancer screening participated in the survey. However, our findings are consistent with a large, national survey that found men were more likely to undergo CRC screening than women. 5 Furthermore, the equal distribution of gender across risk groups suggests that the differences between average risk and increased risk responses is not due to gender bias in the sample. Hence, the gender differences we found may be real, rather than the result of a self-selection problem. Almost 2,000 of the some 2,400 individuals contacted refused to participate in the survey. Given the number of questions and the length of the survey ( 30 min), it was a challenge to recruit willing participants. The study confirms that personal and family history of CRC or related symptoms, and receipt of information/advice related to them, are strongly associated with higher levels of CRC screening. The study reinforces and updates findings of studies done on smaller cohorts restricted by geography, age, or ethnicity as well as a similar national-level study conducted nearly 20 years ago on a comparably sized population. 12 18,31 Previous studies have concluded that information campaigns addressed to such groups as minority ethnic groups, women, and the elderly would be associated with higher screening rates. 12 18,31 Our study broadens this message to the general population. That is, our findings suggest that the screening rates and conformance with screening criteria of men as well as women, younger as well as older subjects, Whites as well as African Americans, Asian Americans, and Hispanics all seem to be higher when they have received information/advice related to cancer screening. We identified relatively few barriers and facilitators, a finding previously reported in studies of other populations. 15 17 Interestingly, we did not find fear and discomfort of screening tests to be a barrier to CRC screening. In a survey of patients and physicians attitudes toward CRC screening tests, 15 percent of patients stated that discomfort was the most important feature in selecting a CRC screening test, in contrast to 64 percent of physicians perceptions of patients values. This suggests that clinicians may not be able to correctly perceive which factors are important to patients. 31 Finally, the negative association between smoking and adequate screening, in conjunction with the positive association between personal health concern and screening, reinforces the differences between key target populations. Effective alternative strategies should be available both for those who are more concerned about health and those who are less concerned. Colorectal cancer remains one of the few neoplasms for which screening has been shown to lead to significant decreases in morbidity and mortality. In addition, several diagnostic screening modalities are available, and recommended screening strategies have been formulated that permit a range of clinical environments. Yet screening rates remain low. Through its focus on lay attitudes toward colorectal cancer screening, this study identifies opportunities to engage patients so they can be properly screened against a leading malignancy that is largely preventable. 26 28 REFERENCES 1. National Cancer Institute. SEER cancer statistics review, 1973 1998. Available at: http://seer.cancer.gov/ Publications/CSR1973 1998. 2001. Accessed April 4, 2002. 2. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology. 1997;112(2):594 642. 3. American Gastroenterological Association. Clinical Symposium: Genetic Testing for Colon Cancer Susceptibility. Program and abstracts of Digestive Week 2001; May 20 23; Atlanta, Georgia. 4. Centers for Disease Control and Prevention. Trends in screening for colorectal cancer United States, 1997 and 1999. MMWR Morbid Mortal Wkly Rep. 2001;50:162 166. 5. Breen N, Wagener DK, Brown ML, Davis WW, Ballard- Barbash R. Progress in cancer screening over a decade: results of cancer screening from the 1987, 1992, and 1998 National Health Interview Surveys. J Natl Cancer Inst. 2001;93:1704 1713. 6. Schroy PC 3 rd, Geller AC, Crosier Wood M, et al. Utilization of colorectal cancer screening tests: a 1997 survey of Massachusetts internists. Prev Med. November 2001;33:381 391. 7. Zack DL, DiBaise JK, Quigley EM, Roy HK. Colorectal cancer screening compliance by medicine residents: perceived and actual. Am J Gastroenterol. October 2001;96:3004 3008. 8. Hawley ST, Levin B, Vernon SW. Colorectal cancer screening by primary care physicians in two medical care organizations. Cancer Detect Prev. 2001;25:309 318. 9. Puschel K, Thompson B, Coronado GD, Lopez LC, Kimball AM. Factors related to cancer screening in Hispanics: a comparison of the perception of Hispanic community members, health care providers, and representatives of organizations that serve Hispanics. Health Educ Behav. October 2001;28:573 590. 10. Lee MM, Lee F, Stewart S, McPhee S. Cancer screening practices among primary care physicians serving Chinese Americans in San Francisco. West J Med. March 1999;170:148 155. 11. Cooper GS, Fortinsky RH, Hapke R, Landefeld CS. Factors associated with the use of flexible sigmoidoscopy as a screening test for the detection of colorectal carcinoma by primary care physicians. Cancer. 15 April 1998;82:1476 1481. 12. Holmes-Rovner M, Williams GA, Hoppough S, Quillan L, Butler R, Given CW. Colorectal cancer screening barriers in

A Risk-adjusted Survey 251 persons with low income. Cancer Pract. September October 2002;10:240 247. 13. James AS, Campbell MK, Hudson MA. Perceived barriers and benefits to colon cancer screening among African Americans in North Carolina: how does perception relate to screening behavior? Cancer Epidemiol Biomarkers Prev. June 2002;11:529 534. 14. Weitzman ER, Zapka J, Estabrook B, Goins KV. Risk and reluctance: understanding impediments to colorectal cancer screening. Prev Med. June 2001;32:502 513. 15. Tang TS, Solomon LJ, McCracken LM. Barriers to fecal occult blood testing and sigmoidoscopy among older Chinese- American women. Cancer Prev. November Drecember 2001;9:277 282. 16. Wolf RL, Zybert P, Brouse CH, et al. Knowledge, beliefs, and barriers relevant to colorectal cancer screening in an urban population: a pilot study. Fam Community Health. October 2001;24:34 47. 17. Mandelson MT, Curry SJ, Anderson LA, et al. Colorectal cancer screening participation by older women. Am J Prev Med. October 2000;19:149 154. 18. Dominitz JA, Provenzale D. Patient preferences and quality of life associated with colorectal cancer screening. Am J Gastroenterol. December 1997;92:2171 2178. 19. Shapiro JA, Seeff LC, Nadel MR. Colorectal cancer-screening tests and associated health behaviors. Am J Prev Med. 2001;21(2):132 137. 20. Rex DK, Johnson DA, Lieberman DA, et al. Colorectal Cancer Prevention 2000: screening recommendations of the American College of Gastroenterology. Am J Gastroenterol. 2000;95(4):868 877. 21. Bond JH, for the Practice Parameters Committee of the American College of Gastroenterology. Polyp guideline: diagnosis, treatment, and surveillance for patients with colorectal polyps. Am J Gastroenterol. 2000;95(11):3053 3063. 22. Weller DP, Owen N, Hiller JE, Wilson K, Wilson D. Colorectal cancer and its prevention: prevalence of beliefs, attitudes, intentions and behavior. Aust J Public Health. 1995;19(1):19 23. 23. Rawl SM, Menon U, Champion VL, Foster JL, Skinner CS. Colorectal cancer screening beliefs: focus groups with first degree relatives. Cancer Pract. 2000;8(1):32 37. 24. Blalock SJ, DeVellis BM, Afifi RA, Sandler RS. Risk perceptions and participation in colorectal cancer screening. Health Psychol. 1999;9:6. 25. Frazier AL, Colditz GA, Fuchs CS, Kuntz KM. Costeffectiveness of screening for colorectal cancer in the general population. JAMA. 2000;284(15):1954 1961. 26. Walsh JME, Terdiman JP. Colorectal cancer screening. JAMA. 2003;289(10):1288 1296. 27. US Preventive Services Task Force. Screening for colorectal cancer: recommendation and rationale. Ann Intern Med. 2002;137:129 131. 28. Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale update based on new evidence. Gastroenterology. 2003;124:544 560. 29. Rickett RR, Auerbach O, Garfinkel L, Hammond EC, Frasca JM. Adenomatous lesions of the large bowel. An autopsy survey. Cancer. 1979;43:1847. 30. Baier M, Calogne N, Cutter G, et al. Validity of selfreported colorectal cancer screening behavior. Cancer Epidemiol Biomarkers Prev. 2000;9(2):229 232. 31. Macrae FA, Hill DJ, St. John DJ, Ambikapathy A, Garner JF. Predicting colon cancer screening behavior from health beliefs. Prev Med. January 1984;13:115 126. 32. Ling BS, Moskowitz MA, Wachs D, Pearson B, Schroy PC. Attitutes toward colorectal cancer screening tests. J Gen Intern Med. 2002;16:822 830.