Colorectal cancer screening attitudes and behavior: a population-based study

Abstract

Background

Even though colorectal cancer (CRC) screening tests for persons 50 years of age or over are recommended to reduce colorectal cancer mortality, screening rates remain disturbingly low.

Methods

Using random digit dialing, 355 telephone interviews were conducted with black and white men and women, 50–79 years of age, who resided in Genesee County, Michigan. The Health Belief Model provided the framework to assess attitudes and practices regarding CRC screening.

Results

For both endoscopic procedures, significantly higher percentages of whites than blacks were aware of the screening procedure (P < 0.05). Overall, fewer than 30% of respondents were adherent to current CRC screening guidelines. Adherence was lowest for black females: 21% for fecal occult blood test, 20% for flexible sigmoidoscopy, and 12% for colonoscopy. Black males compared to black females were about 2.8 times more likely to have had either flexible sigmoidoscopy or colonoscopy (P < 0.05). Physician recommendation was a powerful motivator to screening. Two consistent barriers to screening were the belief that: (a) the test is not needed; and (b) the test is embarrassing.

Conclusions

Interventions directed at physicians and patients are essential to enhance CRC screening rates. CRC survival rates may be improved by physician-guided promotion of screening that focuses on identified barriers.

Introduction

Among cancer death rates for persons 40 to 79 years of age in the United States, colorectal cancer (CRC) ranks second to lung cancer in men and third in women, behind lung and breast cancer, respectively. In 1998, the age-standardized colorectal cancer mortality rate per 100,000 in black men (26.2) exceeded that in white men (19.2); similarly, the rate in black women (19.4) exceeded that in white women (13.2). Colorectal cancer mortality in the United States has been declining since 1985 [1]. During the period 1990–1998, the annual percentage decrease in colorectal cancer mortality was 2.2% in white males, 1.8% in white females, and 0.8–0.9% in black males and females. The United States Preventive Services Task Force in 1996 [2] and again in 2002 [3] strongly recommended CRC screening tests for all men and women at average risk, 50 years of age and older, to reduce CRC mortality.

The goals of CRC screening are to reduce colorectal cancer mortality through early detection and curative intervention and to reduce the colorectal cancer incidence by detecting and removing adenomatous polyps. Recommended colorectal cancer screening examinations include fecal occult blood testing (FOBT), flexible sigmoidoscopy, colonoscopy, and double-contrast barium enema (DCBE). In a meta-analysis of four randomized clinical trials in population samples at average risk using annual or biennial FOBT screening [4], [5], [6], [7], [8] the estimated reduction in colorectal cancer mortality was 16% (OR = 0.84; 95% CI, 0.77–0.92) [5]. The 60-cm flexible sigmoidoscope currently recommended at 5-year intervals enables detection of 50–65% of colorectal cancers and, based on case–control studies, has been estimated to be associated with a 60% reduction in mortality due to cancers in the sigmoid colon and rectum [9], [10]. Although not demonstrated in a randomized clinical trial, the efficacy of colonoscopy has been supported by demonstrating its integral role in the evaluation of participants in the FOBT trials or of patients with distal colorectal adenomas detected by screening sigmoidoscopy. In addition, colonoscopy appears to be more sensitive than DCBE in the detection of true adenomas, and after polypectomy, the incidence of colorectal cancer may be reduced by 76% [11], [12], [13].

Screening recommendations from the American Cancer Society, American College of Gastroenterology, and the United States Agency for Healthcare Research and Quality for the 70–80% of the population at average risk of colorectal cancer include annual FOBT in combination with flexible sigmoidoscopy every 5 years [14]. Although the benefits of alternative screening tests have not been demonstrated, these organizations have suggested that colonoscopy every 10 years or DCBE imaging every 5–10 years may be satisfactory options based on the availability and quality of screening and diagnostic resources.

In a 1999 random telephone survey conducted in the United States, 20.6% of respondents reported FOBT in the previous year, and 33.6% reported having undergone flexible sigmoidoscopy or colonoscopy within the previous 5 years [15]. In their analysis of the year 2000 National Health Interview Survey, Swan et al. reported that underutilization of FOBT or colorectal endoscopy was correlated with lack of health insurance and limited access to primary medical care [16]. The Michigan Cancer Consortium and Michigan Department of Community Health have assigned one of the highest priorities to enhancing utilization of available, recommended CRC screening examinations by Michigan residents at average risk, 50 years of age and older.

Several studies published to date have used a theoretical framework to better understand individuals' attitudes and beliefs concerning colorectal cancer screening [17], [18], [19]. One of the major organizing frameworks that has been used to predict cancer screening behavior is the Health Belief Model [20]. Despite variations in study design and measurement of screening attitudes and behavior, considerable support for the HBM has been documented [21], [22], [23], [24], [25].

While previous studies have addressed one or more of the characteristics below, the present study collectively:

• gathered a population-based sample that includes adequate numbers of black and white men and women to allow for comparisons between gender and race subgroups;

• examined CRC screening attitudes, beliefs, and practices related to individual screening tests (separately and in combination) as opposed to focusing on a single test;

• used a theoretical framework to explore attitudes, beliefs, and practices about colorectal cancer in general and each screening test in particular;

• broadened the scope of the research to examine whether subjects had ever heard about, ever had, and/or were current with recommended screening guidelines for the various CRC screening tests.

A comprehensive telephone interview protocol was developed for use in this research effort. The University of Michigan Institutional Review Board approved the study design and questionnaire. The protocol was designed to operationalize constructs of the Health Belief Model [20]. A significant number of items were included in the questionnaire to tap key constructs: perceived susceptibility, perceived severity, perceived benefits, and perceived barriers. Additional factors thought to contribute to the model's utility, including cues to action and perceived efficacy, were also included.

The interview protocol began with a series of questions related to general health perceptions and practices (including cigarette smoking and date of last physician visit). Because of our particular interest in exploring respondents' attitudes, beliefs, and practices related to the major screening procedures used in the detection of colorectal cancer (digital rectal exam, home-based fecal occult blood test, flexible sigmoidoscopy, and colonoscopy), parallel series of questions were written for each procedure. Each section began by describing each screening test in lay terms, soliciting information about the respondent's history of having the procedure, followed by test-specific benefits and barriers questions. The lay descriptions were intended to highlight specific aspects of the individual tests in order to distinguish one from another (e.g., the preparation for flexible sigmoidoscopy suggested that a person would be asked to take a laxative or use an enema to completely cleanse the colon, while colonoscopy described the need to restrict one's diet as well as drink a special liquid prior to the test). The description of colonoscopy also highlighted the need for someone to drive the individual home from the test.

We drew heavily on an instrument created by Rawl et al. to measure benefits and barriers of the three major colorectal cancer screening tests [24]. This instrument, with well-established psychometric properties, consists of several sections: (a) benefit items for all screening tests (n = 5 items); (b) barrier items for all screening tests (n = 5); (c) barrier items specific to fecal occult blood testing (n = 3); barrier items for flexible sigmoidoscopy and colonoscopy (n = 3); and (d) additional barrier items for colonoscopy alone (n = 2).

Questions related to perceived severity and perceived susceptibility, as well as a dimension referred to as “salience and coherence” (i.e., the extent to which a behavior is judged to make sense in everyday life), were modeled after the work done by Myers et al. [26]. Five of the 10 items in the salience and coherence measure developed by Myers and colleagues were included to assess whether engaging in a preventive health action (e.g., colorectal cancer screening test) is a useful course of action as judged by its technical effectiveness, practical convenience, and personal benefit [26].

In addition, items regarding the frequency of obtaining prostate-specific antigen (PSA) test, Pap test, and mammography were included to determine the relationship of CRC screening behavior with other common screening tests. The protocol included knowledge questions about colorectal cancer and its warning signs, items related to sources of cancer screening information, and additional questions about social support for colorectal cancer screening. Brief comorbidities, family history, and sociodemographics sections concluded the telephone interview protocol.

The pool of eligible subjects consisted of black and white men and women, 50–79 years of age, who resided in Genesee County, Michigan, and had a household telephone. Individuals with a prior history of colorectal cancer, colorectal surgery to remove a polyp, inflammatory bowel disease, or familial adenomatous polyposis were not eligible in this study of “average risk” subjects. A sampling target was established that consisted of approximately 100 individuals within each gender (male, female) and race (white, black) subgroup. In addition, the sample was to be stratified by age (50–64 and 65–79 years) with 50 individuals in each age–race–gender stratum. A random sample of telephone numbers was purchased from a market survey company which also supplied mailing addresses. An informational letter about the study was sent to potential participants prior to contacting them by telephone. Trained interviewing staff from the Survey Research Center at the University of Michigan's Institute for Social Research conducted all telephone interviews. The telephone contact began by first establishing eligibility based on location of residence, age, race, and gender. The screening questions then focused on a history of cancer, polyps, or inflammatory bowel disease. Participants received a check for $10.00 as a token of appreciation for completing the interview.

Demographic variables and variables measuring respondents' familiarity with or use of screening methods were analyzed using two-way contingency tables with accompanying χ² statistics which test the null hypothesis of no difference among race–gender groups. Logistic regression models were used to test whether dimensions represented in the Health Belief Model were significantly related to CRC screening status while controlling for relevant sociodemographic and related factors. Although not reported, the reliability of scales associated with the HBM was assessed using Cronbach's α statistic. These analyses indicated that these measures were each internally reproducible and reliable in this population.