• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jgimedspringer.comThis journalToc AlertsSubmit OnlineOpen Choice
J Gen Intern Med. Jul 1999; 14(7): 432–437.
PMCID: PMC1496606

Patient Preferences for Colon Cancer Screening

Michael Pignone, MD, MPH,1 Dawn Bucholtz, MA, MPH,2 and Russell Harris, MD, MPH1

Abstract

OBJECTIVE

To measure patient preferences for four different screening strategies: annual fecal occult blood testing (FOBT) alone; flexible sigmoidoscopy (FSIG) every 5 years alone; both annual FOBT and FSIG every 5 years; or no screening.

DESIGN

Survey.

SETTING

University internal medicine clinic.

PATIENTS

Convenience sample of 146 adults (aged 50–75 years) with no previous history of colon cancer.

INTERVENTION

Three-part educational program on colon cancer screening administered verbally by trained research assistants.

MEASUREMENTS AND MAIN RESULTS

Patient preferences for screening were measured at three points: after descriptive information about colon cancer and screening options (testing procedure information); after information about test performance but with no out-of-pocket costs (test performance information); and finally with hypothetical out-of-pocket costs (cost information). After only descriptive test information, the most popular strategies were FOBT alone (45%) or both tests (38%). Fewer patients preferred FSIG alone (13%). After information about test performance, more subjects preferred both tests (47%), and fewer subjects preferred FOBT alone (36%) (p =.12). With hypothetical out-of-pocket costs, the proportion preferring FOBT alone increased to 53%, while those preferring both tests decreased to 31% (p < .001). Less than 5% of patients preferred no screening.

CONCLUSIONS

Patient preferences for colon cancer screening were modestly sensitive to information about test performance and strongly sensitive to out-of-pocket costs. The heterogeneity of patients' preferences for how to be screened supports informed shared decision making as a possible means of improving colon cancer screening.

Keywords: colon cancer, screening, patient preferences, shared decision making

Colon cancer screening for adults over 50 years of age is now recommended by major policy-making bodies on the basis of evidence that the use of either fecal occult blood testing (FOBT) or flexible sigmoidoscopy (FSIG) decreases mortality from colon cancer.16 There is no direct comparative evidence to indicate which approach is better, or whether a strategy of using both tests is superior to either test alone.

Despite evidence supporting screening, actual performance of screening for colon cancer among eligible adults is low: the National Health Interview Survey found that 17% of adults over 50 years of age had received testing with FOBT within 1 year and 9% had received FSIG within 5 years.7 Levels among patients in primary care practices are also low. In the North Carolina Prescribe for Health study, a survey of North Carolina ambulatory practices, medical record reviews conducted in 1994 indicated that only 32% of adults over 50 years of age had an FOBT in the previous year, and only 11% had undergone FSIG or colonoscopy within the previous 3 years.8

Communication between providers and patients is important for improving the level of colon cancer screening. Patients may not be aware of colon cancer screening unless their provider discusses it with them; providers may believe most patients are uninterested in screening and thus not discuss it unless patients ask about it. When colon cancer screening is considered, there may little discussion about the options for screening, the consequences of positive and negative tests, or how well the tests perform.

Shared decision making has been increasingly advocated as a means of increasing physician-patient communication and improving decision making.9,10 In shared decision making, the provider gives the patient information about the risk of disease, the benefits and risks of screening, and the options for how to be screened. The patient is able to ask questions about the information and express his or her values and preferences. Provider and patient then work together to choose the best option mutually.

Much of the previous work in shared decision making has focused on its role in decisions about therapy. Shared decision making has been shown, for example, to improve adherence and outcomes in the treatment of chronic diseases such as diabetes and hypertension.11 More recently, shared decision making has also been applied to decisions about screening and preventive care. Providing patients with information about the efficacy and consequences of prostate-specific antigen testing, for example, has been shown to influence patient preferences for whether to be screened.12

A major assumption of shared decision making is that informed patients will vary in their decisions when presented with different options. If virtually all informed patients choose the same approach, providers may justifiably feel that they can simply order that one approach; if informed patients vary in their decisions, however, shared decision making may be preferable. Thus, it is important to elicit informed patients' preferences, a process sometimes called values clarification, when attempting to understand if shared decision making is appropriate.

Patient preferences for colon cancer screening have not been studied extensively, so the degree of variation in how patients would prefer to be screened is not well known. One published study of patient preferences13 was conducted among outpatients with a high previous level of screening and high level of educational attainment and thus may not be broadly generalizable. Further, this study asked about preferences for individual screening tests rather than about specific screening strategies, such as the use of two tests (FOBT and FSIG) in combination.

We sought to characterize the variation in patient preferences for colon cancer screening among subjects from a university-based internal medicine clinic that serves a population with a broad range of education, income, and previous experience with colon cancer screening. Using a structured, verbal educational program, we informed patients about colon cancer and screening methods, and measured their preferences for four colon cancer screening strategies: yearly FOBT screening, FSIG every 5 years, the strategy of both yearly FOBT and FSIG every 5 years, or no screening.

METHODS

Institutional Review Board approval was secured in September 1997, and the study was performed from October through December 1997. A nonconsecutive convenience sample of 146 subjects was drawn from patients attending an internal medicine ambulatory care clinic affiliated with a university hospital.

Eligible patients were between the ages of 50 and 75 years; were able to hear, understand English, and answer the questions independently; and felt sufficiently well to participate in the study. Patients who had ever had colorectal cancer were excluded. We did not exclude patients who had recently received colon cancer screening tests; rather, these subjects were asked to answer questions based on decisions they would make when they were eligible to be rescreened.

Trained research assistants used daily appointment logs to identify patients over 50 years of age. Potential subjects were approached while they were waiting to see their providers and asked to participate in a study about preventive health care. Those who consented were assessed for eligibility. Specific reasons for exclusion were not recorded, but the total number of ineligible patients or refusals was estimated to be less than 5% of those approached. The major reason for refusal was a lack of time before the patient's appointment.

The research assistants verbally administered the interview in private examination rooms while patients waited for their appointments. They used a structured educational intervention that was developed from previous research with a pilot set of patient interviews. The interview took 15 to 20 minutes. Subjects were permitted to ask clarifying questions during the presentation. The educational presentation consisted of three segments, each lasting 3 to 4 minutes: (1) risk of colorectal cancer and benefit of asymptomatic screening; (2) description of the screening tests (FOBT and FSIG); and (3) efficacy of the tests, including information about sensitivity, specificity, and positive predictive value. Verbal information was reinforced visually with simple graphics displayed on a flip chart that illustrated each point (see Appendix A).

Information was provided in stages to allow measurement of each segment's effect on preferences. Previous screening experience was assessed at baseline. Patients then received information about the risk of colon cancer and benefit of screening. Next, subjects were given basic descriptive information about the screening tests (testing procedure information). Preferences for the different screening strategies were then assessed for the first time: subjects were asked if they preferred annual FOBT alone, FSIG every 5 years alone, both tests together, or neither test. Colonoscopy and barium enema were not offered as choices because of a lack of evidence supporting their use as screening tests for average-risk patients, but they were recorded if patients mentioned them. Information was next provided about the tests' sensitivity, specificity, and predictive value (test efficacy information) and preferences were again elicited under the assumption of no out-of-pocket costs for testing. Finally, preferences were elicited a third time with the assumption that screening would require out-of-pocket payments of $10 per year for FOBT and $150 for FSIG every 5 years. These charges were based on data from a separate study of community providers in Central North Carolina (R. Harris, unpublished data, 1998)

We also used open-ended questions to better understand the reasons why subjects preferred particular strategies. Research assistants coded patient responses using several predetermined themes derived from pilot testing. Responses not captured by these themes were directly recorded and assigned to a new category by one of the investigators (DB). Demographic information, including ethnicity, highest grade completed in school, and whether patients had health insurance, was collected at the end of the interview. Gender and age were recorded from appointment logs.

Statistical analyses were performed using STATA 5.0 (Stata Corp., College Station, Tex, 1997). Changes in the proportion of subjects choosing to be screened with both tests, as opposed to either test alone, were analyzed with McNemar's χ2test. Potential predictors of screening preferences, including previous experience with screening, attitudes and beliefs about the likelihood of contracting colon cancer, and the impact of the disease, were examined using the χ2statistic. Results were reported as odds ratios (ORs) for choosing the more aggressive strategy of screening with both tests as compared with choosing either test alone.

RESULTS

Demographics and Previous Testing Experience

The mean age of our 146 subjects was 60 years (range, 50 –75 years). More than half (58%) were female. Nearly all patients self-identified as either white (52%) or African American (43%). The overall level of education was low: 53% had not completed high school. Most of the patients (86%) reported having some type of health insurance. Slightly more than half of subjects (53%) reported that they had ever had a test to look for colon cancer, 14% were unsure, and 33% believed they had never had such a test. When asked about specific colon cancer tests, 59% reported ever having a stool test for blood, 22% ever had sigmoidoscopy, and 13% ever had colonoscopy. Only 28% of subjects reported having had any previous conversation (defined as any discussion beyond simply ordering the test) with their provider about colon cancer screening.

Test Preferences

After subjects had received information about colon cancer and descriptive test information on testing procedure, almost half (45%) of the subjects preferred FOBT alone. Another 38% preferred to be screened with both tests. Flexible sigmoidoscopy alone was preferred by 13%, 1% preferred colonoscopy and 3% stated that they did not want any colon cancer screening test.

After receiving information on test performance, preferences were elicited again. Under the assumption that both the stool test for blood and FSIG were covered by insurance, there was a statistically nonsignificant 9% absolute increase in those preferring to be screened with both tests: 47% now chose to be screened with both tests, while 9% fewer subjects (36%) preferred the stool test alone (p = .12 by McNemar's χ2). The proportion preferring FSIG alone did not change appreciably (13% to 12%).

Preferences were then elicited for a third time, in this case assuming out-of-pocket costs of $10 per year for the stool test for blood and $150 for the FSIG every 5 years. With these costs included, preferences changed dramatically: FOBT alone was now preferred by 54% (an 18% absolute increase), only 31% now wanted both tests, and 8% chose FSIG alone. The difference was statistically significant when compared with the preferences elicited under the assumption of insurance coverage (p < .001 by McNemar's χ2) (Table 1).

Table 1
Subjects' Preferences for Colon Cancer Screening (n = 146)*

How Individuals' Preferences Change

Of the 65 subjects who preferred the strategy of FOBT alone after initial descriptive test information, most (74%) continued to prefer FOBT alone after information about test performance and under the assumption of no out-of-pocket costs. Ten subjects (15%) changed their preference to the strategy of both tests, and four changed to FSIG only. Under the assumption of out-of-pocket costs, 57 of the original 65 preferring FOBT again expressed a preference for FOBT alone.

Among the 55 subjects initially preferring both tests, 49 (89%) continued to prefer both tests after test performance information and under the assumption of no out-of-pocket costs. When faced with out-of-pocket costs, however, 15 of the original 55 changed to FOBT alone, and 1 changed to FSIG alone.

Of the 19 subjects initially preferring FSIG alone, 7 changed to both tests after test performance information. Of the 7 who changed, only 2 maintained this preference for both tests when out-of-pocket costs were assumed.

Reasons for Test Preferences

We also asked subjects to explain why they preferred a particular test strategy. The responses given in the final phase of the study (under the assumption of out-of-pocket costs) are presented in Table 2. The subjects preferring both tests did so because they believed that both tests were more effective in detecting cancer. Of the 79 patients who preferred FOBT only, 37 (47%) did so because it was easier to perform and could be done alone, while 28 (35%) reported doing so because of cost. Those preferring FSIG alone had mixed reasons: some felt it was easier, while others chose it because it was a better or more effective test. Subjects' earlier responses in the descriptive testing procedure information phase were similar with the exception that cost was not reported as a reason by those preferring FOBT alone.

Table 2
Reasons for Test Preferences After All Information and Assuming Out-of-Pocket Costs (n = 136)*

Predictors of Screening Preferences

We examined how several potentially important predictor variables affected the screening preferences. The results presented in Table 3 are based on preferences expressed in the final phase of the study and assumed out-of-pocket costs to the patient. Results at the earlier phases were not substantially different.

Table 3
Factors Affecting Patient Preferences for Colon Cancer Screening Strategies After All Information andAssuming Out-of-Pocket Costs*

Recalling a previous discussion about colon cancer screening was modestly associated with a subject's preference to have both tests as opposed to either test alone. Having performed any colon cancer screening test in the past was also associated with a preference for both tests. Having previously undergone FSIG was strongly associated with preferring both tests.

Differences in the perceived likelihood of getting colon cancer, knowing someone closely who had had colon cancer, or the perception that getting colon cancer is especially bad did not affect test preferences. Subjects who reported a willingness to have a screening test when they did not have symptoms were more likely to prefer both tests than those subjects who did not endorse asymptomatic screening.

Men were somewhat more likely to prefer both tests than women (46% vs 25%; OR 2.6; 95% confidence interval 1.2, 5.3). Education, age, race, and insurance status did not predict test preferences (results not shown).

DISCUSSION

In our study of patient preferences for different colon cancer screening strategies, we found that the most popular strategies were annual FOBT alone or the combination of annual FOBT and FSIG every 5 years. Fewer patients expressed a preference for FSIG every 5 years alone. Information about test performance appeared to modestly increase preferences for having both tests, though the result was not statistically significant. Potential out-of-pocket costs increased the proportion preferring to have FOBT alone. The differences among informed patients with regard to how they would prefer to be screened supports shared decision making as a potentially useful approach to improving colon cancer screening.

The majority of patients preferred to have some form of screening, suggesting that when presented with educational information about colon cancer and the different options for screening, patients' willingness to be screened may be greater than currently reported levels of screening performance would suggest. Alternatively, the high proportion preferring some form of testing could be due to subjects' perception that the investigators wanted them to have some form of screening when, in fact, they did not plan to actually be tested.

There has been little previous research on patient preferences for different colon cancer screening strategies. Leard et al. examined preferences among outpatients from Southern California.13 They provided subjects with a “balance sheet” that gave information about the different colon cancer tests (FOBT, FSIG, colonoscopy, and barium enema) and asked subjects which test they preferred. Their patient population was primarily white (87%), well educated, and had substantial experience with colon cancer screening (93% had been tested before). Ninety-six percent of their subjects wished to be screened by some method. They also found substantial variety: 38% preferred colonoscopy, 31% FOBT, 14% barium enema, and 13% FSIG. However, the most popular test in their study, colonoscopy, is not covered as a screening test for average-risk patients by Medicare, which limits the relevance of their findings, and they did not examine the strategy of combined screening with FOBT and FSIG.

Our study of patient preferences was, to our knowledge, the first work in this area to be conducted among a population with a wide range of educational attainment. We offered screening options that would be available to most patients to increase the usefulness and relevance of our results. We examined preferences for the three screening strategies that are currently covered by Medicare for usual-risk patients and recommended by the U.S. Preventive Services Task Force.1

Our study, however, has several limitations. Most important, it was not designed to measure actual testing behavior. It is not currently known how well preferences patients express correlate with actual screening. We are currently conducting a study that examines whether the provision of information affects the proportion of subjects who go on to complete screening tests.

Second, we used a nonconsecutive convenience sample of clinic patients. We believe that the low level of refusal (less than 5%) allowed a good representation of the actual clinic population, but our results should not be considered to be generalizable to the general population of adults over 50 years of age. Also, we cannot exclude unconscious bias among the research assistants in the selection of subjects. We relied on patient self-report of previous screening experience, and we did not assess the level of recent or regular colon cancer screening among our sample or the clinic population as a whole, a factor that may act as a confounder of patient interest or test preference. The reasons why subjects preferred particular screening strategies were categorized by a single coder, so we cannot assess their reproducibility. Finally, we did not vary the order of presentation of the material, so we cannot exclude that some of the changes in preferences could be explained simply by the sequence of questioning or as a result of repeated questioning.

Despite these limitations, we believe this study has two important conclusions. First, patient preferences for colon cancer screening are not monolithic, and efforts to improve the performance of colon cancer screening in primary care settings should encourage discussion and negotiation between patients and providers. Whether screening tests ordered on the basis of a shared decision are more likely to be performed than those ordered unilaterally by the provider has not been studied, and further evaluation is warranted to determine if informed shared decision making improves levels of colon cancer screening and satisfaction with results.

Second, patient preferences are sensitive to out-of-pocket costs. Many patients appear to be unwilling to pay $150 to have FSIG in addition to annual FOBT. The effect of different levels of out-of-pocket costs has not been studied. We plan to examine this area more closely in the future, using the willingness-to-pay approach. Our data suggest, however, that policy makers or insurers should minimize out-of-pocket costs if they want to facilitate more aggressive screening.

Appendix A: Verbal Information Provided to Subjects

Introduction

We are doing a study about colon cancer, trying to find out what patients know and believe about this cancer. We want to find out how to best teach patients about colon cancer. I would like you to answer some questions, which will only take 10 to 15 minutes. Your answers will be kept confidential and will not be shared with your health care provider. If you wish to discuss these issues with your doctor, bring it up at your next visit.

Disease Information

Colon cancer is a common type of cancer: about 140,000 people get colon cancer in the United States each year, and it is the second leading cause of death due to cancer. Both men and women get colon cancer.

Colon cancer is much more curable if it is found in the early stages rather than in the later stages. Colon cancer starts as little growths in the intestine, called polyps. Polyps sometimes develop into cancer. But if the polyps are found before they become cancer, they can be removed and cancer can be prevented. And if the polyps have already developed into cancer, the cancer is much more likely to be treatable if it is found before it spreads.

Colon cancer becomes more likely to occur as you get older. Experts recommend regular testing for colon cancer beginning at age 50.

Descriptive Testing Procedure Information

Now I am going to tell you about the two tests that are used to find colon cancer. These tests are important because most people who get colon cancer don't notice any symptoms until the cancer is in the later stages. These tests can find cancer before you would notice any symptoms.

One test is the stool test for blood. This is a test that is done once a year. For this test, you collect samples of your stool on a card and mail them back to your doctor. You collect three stool samples from three different bowel movements. The test is safe and not difficult to do, but some people consider it inconvenient to take the stool samples. Also, you may be asked to follow a diet that is free of red meat and certain vegetables for 3 days before taking the samples. Not eating these foods improves the test's accuracy.

A second test is the flexible sigmoidoscopy. FSIG is done every 5 years. With this test, the doctor inserts a small flexible tube (about the width of a little finger) with a light on the end into the rectum to look for polyps. A sigmoidoscopy is safe, it does not require anesthesia, and takes only about 10 to 15 minutes. You may have to have an enema before the test. Some people report mild to moderate discomfort, but most say it was not as uncomfortable as they thought it would be.

Test Performance Information

Finally, I would like to give you some information about how well the tests work. No test is perfect. You need to understand how well the tests work before you decide to get them.

First, let's discuss the stool test for blood. Of every 1,000 people who do a stool test for blood, 950 will have a negative test and 50 will have a positive test. Each of these 50 people who tested positive would need to have a full colon examination (called colonoscopy) to find out for sure whether they have cancer, or polyps, or some other cause of bleeding.

Some who tested positive will be found to have no cancer. A little more than half will not actually have cancer. Some will have polyps that their doctor will want to know more about. A few will actually have cancer, but they had to have this test to find it.

For this test to work best, it should be done every year, even if the results are negative. This is because polyps don't always bleed, so cancer can sometimes be missed. For example, it is likely that 1 of the 950 people who had a negative stool test would actually have cancer and would have been missed.

Now let's talk about flexible sigmoidoscopy. It is important to know that FSIG only looks at the lower part of the colon. It cannot find cancer or polyps higher up. But most polyps or cancer are in the lower part of the colon that the sigmoidoscope can see.

Of every 1,000 people who have a flexible sigmoidoscopy, 5 will have cancer detected and about 100 will have large polyps. Their doctor will want to know more about them. Everyone with cancer or large polyps will have a full colon examination to see if they have cancer in the upper part of the colon.

So which test is the better test? Right now, doctors cannot say whether one test is better than the other for finding colon cancer.

Wrap-up

Thank you for taking the time to participate in our study. I would like to remind you that you will not be tested for colon cancer as part of this study, so if you do want to be tested, please talk to your health care provider at your next visit.

REFERENCES

1. US Preventive Health Services Task Force. Guide to Clinical Preventive Services. 2nd ed. Alexandria, Va: International Medical Publishing; 1996.
2. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology. 1997;112:594–642. [PubMed]
3. Mandel J, Bond J, Church T, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med. 1993;328:1365–71. [PubMed]
4. Kronborg O, Fenger C, Olsen J, Jorgensen D, Sondergaard O. Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet. 1996;348:1467–71. [PubMed]
5. Hardcastle J, Chamberlain J, Robinson M, et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet. 1996;348:1472–7. [PubMed]
6. Selby J, Friedman GD, Quesenberry CP, Weiss NS. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med. 1992;326:653–7. [PubMed]
7. Centers for Disease Control and Prevention. Screening for colon cancer. MMWR. 1996;45:107–10. [PubMed]
8. Pignone M, Harris R, Kinsinger L. Changing cancer screening practice: physician or patient activation? J Gen Intern Med. 1997;14(Suppl 1):110.
9. Charles C, Gafni A, Whelan T. Shared decision-making in the medical encounter: what does it mean? Soc Sci Med. 1997;44:681–92. [PubMed]
10. Coulter A, Entwhistle V, Gilbert D. Sharing decisions with patients: is the information good enough? BMJ. 1999;318:318–22. [PMC free article] [PubMed]
11. Greenfield S, Kaplan S, Ware J, Yano E, Frank H. Patients' participation in medical care: effects on blood sugar control and quality of life in diabetes. J Gen Intern Med. 1988;3:448–57. [PubMed]
12. Flood A, Wennberg JE, Nease RF, Fowler FJ, Ding J, Hynes LM. The importance of patient preference in the decision to screen for prostate cancer. J Gen Intern Med. 1996;11:342–9. [PubMed]
13. Leard L, Savides T, Ganiats T. Patient preferences for colorectal cancer screening. J Fam Pract. 1997;45:211–8. [PubMed]

Articles from Journal of General Internal Medicine are provided here courtesy of Society of General Internal Medicine
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

  • Cited in Books
    Cited in Books
    PubMed Central articles cited in books
  • MedGen
    MedGen
    Related information in MedGen
  • PubMed
    PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...