|
|
Volume
2:
No. 2, April 2005
ORIGINAL RESEARCH
Development of a Brief Survey on Colon Cancer Screening
Knowledge and Attitudes Among Veterans
Michael S. Wolf, PhD, MPH, Alfred Rademaker, PhD, Charles L.
Bennett, MD, PhD, M. Rosario Ferreira, MD, Nancy C. Dolan, MD,
Terry C. Davis, PhD, Franklin Medio, PhD, Dachao Liu, MA, June Lee,
Marian Fitzgibbon, PhD
Suggested citation for this article: Wolf MS, Rademaker
A, Bennett CL, Ferreira MR, Dolan NC, Davis TC, et al.
Development of a brief survey on colon cancer screening knowledge
and attitudes among veterans. Prev Chronic Dis [serial online] 2005 Apr [date cited]. Available from: URL: http://www.cdc.gov/pcd/issues/2005/
apr/04_0104.htm.
PEER REVIEWED
Abstract
Introduction
Poor knowledge of and negative attitudes toward available screening tests may account in part for colorectal cancer screening rates being the lowest among 17 quality measures reported for the Department of Veterans Affairs health care system, the largest integrated health system in the United States. The purpose of this study was to develop a brief assessment tool to evaluate knowledge and attitudes among veterans toward colorectal cancer screening options.
Methods
A 44-item questionnaire was developed to assess knowledge, attitudes, and beliefs about colorectal cancer and screening and was then administered as part of an ongoing randomized controlled trial among 388 veterans receiving care in a general medicine clinic. Sixteen candidate items on colorectal cancer knowledge, attitudes, and beliefs were selected for further evaluation using principal components analysis. Two sets of items were then further analyzed.
Results
Because the Cronbach a for beliefs was low (a = 0.06), the
beliefs subscale was deleted from further consideration. The
final scale consisted of seven items: a four-item attitude
subscale (a = 0.73) and a three-item knowledge subscale (a = 0.59).
Twelve-month follow-up data were used to evaluate predictive
validity; improved knowledge and attitudes were significantly
associated with completion of flexible sigmoidoscopy (P =
.004) and completion of either flexible sigmoidoscopy or
colonoscopy (P = .02).
Conclusion
The two-factor scale offers a parsimonious and reliable measure of colorectal cancer screening knowledge and attitudes among veterans. This colorectal Cancer Screening Survey (CSS) may especially be useful as an evaluative tool in developing and testing of interventions designed to improve screening rates within this population.
Back to top
Introduction
Colorectal cancer is the third most common cancer and the third leading cause of cancer death in the United States (1). The U.S. Preventive Services Task Force, the American Cancer Society, and the American Gastroenterological Association have developed guidelines for colorectal cancer screening and recommend that persons aged 50 years or older who are at average risk for the disease be screened periodically (2-4). Despite these recommendations and multiple studies finding that colorectal cancer screening is cost-effective, screening rates are the lowest for any other cancer screening test, with only half of persons aged 50 years and older having received any of the available methods (5).
One potential barrier to effective screening is inadequate knowledge of both the disease and the possible options for undergoing various types of screening tests (6-17). Poor knowledge related to colorectal cancer is associated with compromised perceptions of cancer risk and low rates of screening services use (6,7,12,14,16). Targeted efforts are needed to improve both the overall awareness of colorectal cancer and the availability of often limited resources for invasive screening procedures, such as flexible sigmoidoscopy or colonoscopy. As health education and colorectal cancer screening programs are developed, valid and reliable measures of knowledge and attitude are needed to explicitly assess the efficacy of these efforts.
One prior study reports on a brief instrument that measures beliefs and attitudes toward colorectal cancer screening (17). This assessment was conducted in a mailed survey of primarily white, employed men. However, it has not been evaluated in other settings characterized by higher rates of racial/ethnic minorities or among persons of lower socioeconomic status; these groups have previously been found to be at greater risk for low screening compliance (18-21). Another population at greater risk for low screening compliance is veterans who receive care in the Department of Veterans Affairs (VA) health care system, the largest integrated delivery system in the country. Out of 17 quality measures routinely included in a nationwide VA quality improvement effort, colorectal cancer screening rates are the lowest (22).
As part of a randomized clinical trial effort to improve colorectal cancer screening rates within a VA outpatient general medicine clinic, we recently reported on knowledge and attitudinal barriers to screening participation among veterans (23). Our intervention targeted improvements in both veterans' perceptions about the disease and screening options in addition to their compliance. In this study, we developed and validated a brief measurement tool for evaluating knowledge and attitudes toward colorectal cancer screening among veterans.
Back to top
Methods
Recruitment of participants
We designed a 44-item questionnaire to measure patient knowledge, attitudes, and beliefs associated with colorectal cancer screening and administered the questionnaire to 388 veterans. Male veterans aged 50 years and older who had not received colorectal cancer screening (defined as having a fecal occult blood test [FOBT] within one year or a flexible sigmoidoscopy or colonoscopy within five years) were recruited from general medicine clinics at the VA Chicago Health System between May 1, 2001, and December 31, 2002. Patients were ineligible if they 1) had received a FOBT within one year; 2) received a flexible sigmoidoscopy or colonoscopy within five years; 3) had a personal or family history of colorectal cancer or polyps; or 4) had a personal history of inflammatory bowel disease. In addition, individuals with dementia, impaired vision, hearing problems, or acute illness were deemed ineligible to participate in the study. We excluded patients with impaired vision because the instrument we employed to assess health literacy required the ability to view a list of words. The study protocol was approved by the Northwestern University Institutional Review Board.
Between May 2001 and December 2002, research assistants approached 589 eligible participants as they waited for their scheduled outpatient visit. Of these, 156 (26.4%) refused to be in the study, and 56 (9.6%) did not complete the study questionnaire primarily because their general medicine physician was ready to begin their visit. In all, 388 (65.9%) individuals completed the entire baseline interview, including the questionnaire. No compensation was offered for participation. After the informed consent process, participants took part in a 10- to 15-minute, face-to-face interview that included sociodemographic items, a literacy assessment, and the 44-item questionnaire. The literacy assessment consisted of administering the Rapid Estimate of Adult Literacy in Medicine (REALM), a screening instrument used to determine the ability of patients to read and pronounce common medical terminology and lay terms for body parts and illnesses (24,25). Raw REALM scores are converted to grade ranges: 0-18 = third grade and below, 19-44 = fourth to sixth grade, 45-60 = seventh to eighth grade, and 61-66 = high school. Follow-up interviews were conducted with 227 of these patients six to 12 months after the baseline interview, beginning November 2001 through December 2003. Patients' screening status was obtained through medical record review also during this period.
Development of the colorectal cancer questionnaire
The 44-item questionnaire included items designed to assess knowledge of colorectal cancer and specific screening tests and attitudes and beliefs toward colorectal cancer and available screening options. Knowledge questions were adapted from the 1992 National Health Interview Survey (NHIS) Cancer Control Supplement, with modifications to reflect current terminology (e.g., use of the term flexible sigmoidoscopy or flex sig instead of proctoscopy) (26). Attitudinal and belief items were developed based on findings from focus groups conducted among this same population of veterans (27). Reader comprehension of the questionnaire items was evaluated using five one-hour cognitive interviews among a convenience sample of community-based, screening-eligible adults. All interviews were conducted by one of the research investigators (Ferreira) and followed available guidelines established for properly conducting cognitive interviews in survey development (28). Interview techniques included both "concurrent think-aloud" and specific probes. The interviews were tape-recorded and abstracted for relevant information, which was used to modify the questionnaire. The modified questionnaire was then administered to a pilot group of 15 patients who were approached in the general medicine clinics (29).
During the pilot process, we obtained patient feedback to items in the pilot test and maintained reading levels of instructions, items, and response options appropriate for lower-literate patients; we used a common measure of document readability (Flesch-Kincaid) to gauge reading levels. Principles described by Doak et al were also applied to maximize item comprehension (30). The final version of the questionnaire registered as having a fifth-grade level of reading comprehension. Even though we planned to administer the instrument through an interview, the readability of the document provided us additional assurance that the questionnaire could be appropriately understood by most patients.
Of the 44 items in the questionnaire administered to the 388 veterans, 10 were associated with knowledge, 29 were associated with attitudes, and five were associated with beliefs. After the administration of the questionnaire, five of the knowledge items were selected by the research team as appropriate for analysis; other knowledge items were excluded because they were conditional questions that were not answered by everyone. Of the 29 attitude items, six were selected for analysis; again, other attitude items were excluded because they were conditional questions not answered by everyone. All five of the belief questions were selected for analysis. Thus, a total of 16 items were selected for analysis.
To prepare for data analysis, we scored questions so that low values reflected high knowledge and attitudes consistent with screening, or "correct" beliefs; high values reflected low knowledge and attitudes inconsistent with screening, or "incorrect" beliefs. Scoring for the knowledge scale was dichotomous (1 = yes, 2 = no); a response of yes required follow-up patient confirmation of understanding of the concept in question. For subjects who responded no or who were determined to have inadequate knowledge of the test in question, simple standard descriptions of both FOBT and flexible sigmoidoscopy were provided by the interviewer to ensure a proper frame of reference. The attitude scale was scored from 1 to 3 on level of worry (1 = not very or not at all worried, 2 = somewhat worried, 3 = very or extremely worried). For both subscales and the total scale, the score was determined by the sum of all nonmissing items. Items on the belief scale were scored for an initial analysis, but the belief construct was not included in a final analysis.
Psychometric analyses
Principal components (PC) analysis was used to assess the
construct validity of the 16 items selected for initial analysis.
Cronbach a was used to examine reliability
(internal consistency) of the derived knowledge, attitudes, and
beliefs subscales. The value of Cronbach
a ranges between 0 and 1; if items within a scale are
perfectly correlated, then a = 1; if items are totally
unrelated, then a = 0. An α coefficient of 0.70 or higher
is considered to be acceptably reliable, indicating that items
within the same scale measure the same underlying construct. A
low Cronbach a for the belief scale and low factor
loadings of the belief variables resulted in deletion of this
subscale. Final PC analysis on the remaining seven knowledge and
attitude items was performed to determine whether these items
followed the knowledge and attitude pattern. To confirm
reliability of the knowledge and attitudes subscales,
correlations between the full scale and items within the
subscales were calculated.
An additional question of interest was: Do individuals who improve their knowledge and attitude exhibit different screening behavior than individuals who do not improve knowledge and attitude? To assess the predictive validity of the total score with screening behavior, change in the total score between two time points (initial questionnaire and follow-up questionnaire) was related to screening behavior using Fisher's exact test. It was postulated that an improvement in knowledge and attitudes would be related to an improvement in screening behavior.
Back to top
Results
Respondents had a mean age of 67.3 years (SEM = 0.52); 41.4% were African American; 59.6% had completed high school, and 22% had completed college. Respondents' reading abilities averaged at the seventh- to eighth-grade level (mean REALM score = 57.3, SEM = -0.7), with 36% having reading skills lower than the eighth-grade level. More than two thirds (69.1%) of the men in the study were unemployed or retired, and 38% were married. One third of respondents reported their health as either very good or excellent.
Initial analysis consisted of the evaluation of 16
candidate items on colorectal cancer knowledge, attitudes, and
beliefs using PC analysis. Item factor loadings
for the three-factor solution are shown in
Table 1. Because
factor loadings for beliefs were low, the belief subscale was
deleted from further consideration. Also, decisions were made to remove
additional items based on lower factor
loadings and/or conceptual fit with remaining items. Thus, the items “likely to get a flexible sigmoidoscopy
(or FOBT) if friend recommended,” “know testing
age,” and “heard of colorectal cancer” were
deleted from further consideration.
The plan and procedure of item retention resulted in
provisional compositions that could be mapped to two factors:
knowledge and attitudes. These two sets of items were further
analyzed using PC analysis to assess construct validity and
Cronbach a to evaluate internal consistency (Table
2). All seven items were retained. The final scale consisted of
seven items: a four-item attitude subscale and a three-item
knowledge subscale (Table 3).
Higher correlations were observed between items within
subscales and their corresponding full scale, while low
correlations were expected and subsequently attained between
items within subscales and the noncorresponding full scale
(Table
4).
Twelve-month follow-up data were used to evaluate the
predictive validity of the knowledge and attitudes scale and each
of the two subscales for completion of a colorectal cancer
screening test (Table 5).
Because low values of all items in the
attitudes subscale reflected favorable attitudes consistent with
screening, and low values of all items in the knowledge subscale
represented high knowledge, decrements over time on these
subscales and the overall knowledge and attitude scale
represented an improvement in attitudes consistent with screening
and/or an improvement in knowledge.
We would assume such
improvements in knowledge and attitudes would be associated with
screening completion among eligible individuals noncompliant with
existing screening recommendations. A minimum decrement over time
(i.e., an improvement) of more than four points in the total
knowledge and attitude summary scale was significantly associated
with higher levels of colorectal cancer screening completion. A
decrease of more than four points over time on the full scale was
significantly associated with completion of flexible sigmoidoscopy (P = .004) and completion of either flexible
sigmoidoscopy or colonoscopy (P = .02).
Back to top
Discussion
We have developed a seven-item scale that can be used to
measure knowledge and attitudes toward colorectal cancer
screening among U.S. veterans. This instrument (Appendix),
the Colorectal Cancer Screening Survey (CSS), was designed to be
a brief and simple measure of knowledge and attitudes of veterans
toward colorectal cancer screening tests. The results of this
study suggest that the two-factor solution offers a parsimonious
and reliable measure. It is the first psychometric tool to our
knowledge to measure colorectal cancer screening knowledge and
attitudes among veterans, a population that is predominantly
low-income; nearly half in this study were African
American. The CSS was also developed for all levels of literacy.
Items were determined to be at a fifth-grade reading level and
had simple response options. Moreover, the instrument was
interviewer-administered.
Although adequate knowledge and positive attitudes alone may not be sufficient to ensure completion of colorectal cancer screening tests, both are common barriers that have been previously linked to noncompliance. Several studies have found that the absence of clinical symptoms was the most important factor associated with noncompliance with returning FOBTs or undergoing a flexible sigmoidoscopy procedure (7-16). Other attitudinal barriers include fear and anxiety about cancer and perceptions that colorectal cancer screening tests are uncomfortable, embarrassing, or generally unpleasant. The goal of many patient-directed interventions has been to overcome these barriers; the CSS could serve as a valuable indicator of an intervention's efficacy to improve intermediary outcomes.
Interestingly, the CSS had the highest predictive validity with the completion of a flexible sigmoidoscopy screening test, and was less likely to predict screening use when return of FOBTs was considered. This discrepancy may reflect both the level of difficulty of personal endorsement for colorectal cancer screening participation between the available testing options, as well as resources within the VA health care system. For example, the decision to have an FOBT may depend less on knowledge and attitudes than the decision to agree to a more invasive procedure such as flexible sigmoidoscopy or colonoscopy. It may be easier to agree to complete an FOBT with poorer knowledge and a less positive attitude toward colorectal cancer screening than to agree to complete a flexible sigmoidoscopy and colonoscopy, since an FOBT asks less of a patient. Patients may complete the procedure based on physician recommendation without recognizing it as a colorectal cancer screening test. However, flexible sigmoidoscopy and colonoscopy procedures require repeat visits and extensive preparation and take more time to explain and to engage subjects in decision making. Although the relationship did not reach significance, it is noteworthy that those with improved knowledge and attitude scores on the CSS scale had higher rates of colonoscopy screening, a test that is often exceedingly difficult to receive in a timely manner within the VA healthcare system because of limited trained clinical staff and resources.
Limitations to this study should be noted. First, our study is based on a cohort of male veterans. Additional assessments in other settings that provide care for large numbers of racial/ethnic minorities, both male and female, and/or who are of low socioeconomic status, such as the county medical systems, are needed. Second, the CSS may benefit from further psychometric evaluation that could improve upon the knowledge subscale and also evaluate the reliability of CSS scores over time. Further evaluation might also include test-retest reliability and discriminant validity assessments. Evidence of sensitivity to change will be necessary to eventually determine whether the CSS is an applicable evaluative tool for screening interventions.
In conclusion, the CSS may be a useful tool for testing the effect of interventions designed to improve colorectal cancer screening among veterans through improving patient knowledge and attitudes. Because veterans with low knowledge and negative attitudes toward screening tests may not be quickly or easily identified in clinical settings, the CSS might eventually be considered for use as a screening assessment to identify veterans who are at risk for colorectal cancer screening noncompliance.
Back to top
Acknowledgments
This project was funded by the Department of Veterans Affairs
(PCI 99-158) and the National Cancer Institute (R01 CA86424).
Back to top
Author Information
Corresponding Author: Michael S. Wolf, PhD, MPH, Assistant
Professor of Medicine, Institute for Health Services Research and
Policy Studies, Northwestern University, Feinberg School of
Medicine, 676 N St Clair St, Suite 200, Chicago, IL 60611. Dr.
Wolf is also affiliated with The VA Midwest Center for Health
Services and Policy Research, the VA Chicago Healthcare System,
Chicago, Ill, and the Robert H. Lurie Comprehensive Cancer
Center, Northwestern University, Chicago, Ill. Telephone: 312-695-0459. Email:
mswolf@northwestern.edu.
Author Affiliations: Alfred Rademaker, PhD, Feinberg School of
Medicine; Charles L. Bennett, MD, PhD, M. Rosario Ferreira, MD, The VA Midwest Center for
Health Services and Policy Research, Feinberg School of Medicine,
and Robert H. Lurie Comprehensive Cancer Center; Nancy C. Dolan, MD, Marian
Fitzgibbon, PhD, Feinberg School
of Medicine and Robert H. Lurie Comprehensive Cancer Center;
Terry C. Davis, PhD, Louisiana State University Medical School,
Shreveport, La; Franklin Medio, PhD, Medical University of
South Carolina, Charleston, SC; Dachao Liu, MA, Feinberg School of Medicine;
June Lee, The VA Midwest Center for Health Services and Policy Research.
Back to top
References
- Cancer Facts and Figures. American Cancer Society. Atlanta (GA): American Cancer Society; 2005.
- U.S. Preventive Services Task Force. The guide to clinical
preventive services: report of the United States Preventive
Services Task Force (3rd ed.) McLean (VA): International Medical
Publishing, Inc.; 2002 Jan.
- Pignone M, Rich M, Teutsch SM, Berg AO, Lohr KN.
Screening for colorectal cancer in adults at average risk: a
summary of the evidence for the U.S. Preventive Services Task
Force. Ann Intern Med 2002;137:132-41.
- Byers T, Levin B, Rothenberger D, Dodd GD, Smith RA.
American Cancer Society guidelines for screening and surveillance for early
detection of colorectal polyps and cancer: update 1997. American Cancer Society
Detection and Treatment Advisory Group on Colorectal Cancer. CA
Cancer J Clin 1997;47:154-60.
-
Screening for colorectal cancer – United States,
1997. MMWR Morb Mortal Wkly Rep 1999 Feb 19;48(6):116-21.
- Brown ML, Potosky AL, Thompson GB, Kesseler LG.
The
knowledge and use of screening tests for colorectal and prostate
cancer: data from the 1987 National Health Interview Survey. Prev
Med 1990;19:562-74.
- Farrands PA, Hardcastle JD, Chamberlain J, Moss S.
Factors
affecting compliance with screening for colorectal cancer. Community
Med 1984;6:12-9.
- Price JH.
Perceptions of colorectal cancer in a socioeconomically disadvantaged population. J Community
Health 1993;18:347-62.
- Wong NY, Nenny S, Guy RJ, Seow-Choen F.
Adults in a
high-risk area are unaware of the importance of colorectal cancer
screening: a telephone and mail survey. Dis Colon Rectum 2002;
45(7):946-50.
- Kelly RS, Shank C.
Adherence to screening flexible sigmoidoscopy in asymptomatic patients. Med Care 1992;30:1029-42.
- Myers RE, Balshem AM, Wolf TA, Ross EA, Millner L.
Adherence to continuous screening for colorectal neoplasia. Med
Care 1993 Jun; 31(6):508-19.
- Brenes GA, Paskett ED.
Predictors of stage of adoption for
colorectal cancer screening. Prev Med 2000;31:410–16.
- Ling BS, Moskowitz MA, Wachs D, Pearson B, Schroy PC.
Attitudes toward colorectal cancer screening tests. J Gen Intern
Med 2001; 16(12):822-30.
- Vernon SW, Myers RE, Tilley BC, Li S.
Factors associated
with perceived risk in automotive employees at increased risk of
colorectal cancer. Cancer Epidemiol Biomarkers Prev 2001;
10(1):35-43.
- Wolf RL, Zybert P, Brouse CH, Neugut AI, Shea S, Gibson G,
et al.
Knowledge, beliefs, and barriers
relevant to colorectal cancer screening in an urban population: a
pilot study. Fam Community Health 2001;24(3):34-47.
- Vernon SW.
Participation in
colorectal cancer screening: a
review. J Natl Cancer Inst 1997;89:1406-22.
- Vernon SW, Myers RE, Tilley BC.
Development and validation of
an instrument to measure factors related to colorectal cancer
screening adherence. Cancer Epidemiol Biomarkers Prev 1997;
6:825-32.
- Mandelblatt J, Andrews H, Kao R, Wallace R, Kerner J.
The
late-stage diagnosis of colorectal cancer: demographic and
socioeconomic factors. Am J Public Health 1996 Dec;86(12):1794-7.
- Cooper GS, Yuan Z, Rimm AA.
Racial disparity in the
incidence and case-fatality of colorectal cancer: analysis of 329
United States counties. Cancer Epidemiol Biomarkers Prev
1997;6(4):283–85.
- Ionescu MV, Carey F, Tait IS, Steele RJ.
Socioeconomic
status and stage at presentation of colorectal cancer. Lancet
1998;352: 1439.
- Tavani A, Fioretti F, Franceschi S, Gallus S, Negri E,
Montella M, et al.
Education, socioeconomic
status and risk of cancer of the colon and rectum. Int J
Epidemiol 1999;28(3):380-5.
- Jha AK, Perlin JB, Kizer KW, Dudley RA.
Effect of the
transformation of the Veterans Affairs Health Care System on the
quality of care. N Eng J Med 2003;348:2218-27.
- Dolan NC, Ferreira MR, Davis TC, Fitzgibbon ML, Rademaker A,
Liu D, et al.
Colorectal cancer screening knowledge, attitudes, and beliefs among veterans:
does literacy make a difference? J Clin
Oncol 2004;22(12):2617-22.
- Davis TC, Crouch MA, Long SW, Jackson RH, Bates P, George RB, et al.
Rapid assessment of literacy levels of adult primary care
patients. Fam Med 1991 Aug;23(6):433-5.
- Davis T, Crouch M, Long S. Rapid estimate of adult literacy in medicine (REALM): examiner's instruction sheet. Red Lake (MN): Red Lake Hospital; 1998 Apr.
- Benson V, Marano MA.
Current estimates from the National
Health Interview Survey, 1992. Vital Health Stat 10 1994 Jan;(189):1-269.
- Davis T, Dolan N, Ferreira MR, Tomori C, Green KW, Sipler AM,
et al.
The role of inadequate health literacy skills in
colorectal cancer screening. Cancer Invest 2001;10(1):193-200.
- Willis GB.
Cognitive
interviewing and questionnaire design: a
training manual. Atlanta (GA): Centers for Disease Control and Prevention,
National Center for Health Care Statistics, Office of Research and Methodology; 1994
Mar.
- Bennett CL, Gorby NS, Ferreira MR, Dolan NC, Fitzgibbon ML, Rademaker AW, et al. A focused CQI intervention can increase colorectal cancer screening rates among veterans: results from the Chicago VA project. Bethesda (MD):
Presentation at the 28th Annual Meeting of the American Society of
Preventive Oncology; 2004
Mar 14-16.
- Doak CC, Doak LG, Root JH. Teaching patients with low literacy skills, 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 1996.
Back to top |
|
