This study was presented in part at Digestive Disease Week 2008, San Diego, California.
This study was funded by an unrestricted grant from Procter and Gamble.
Patients with extensive ulcerative (UC) or Crohn's (CD) colitis have an increased risk of colon cancer and require colonoscopic surveillance. This study explores patient attitudes and behavior regarding inflammatory bowel disease (IBD), colonoscopies, and colon cancer risk.
In all, 514 patients with UC or CD colitis for at least 7 years and at least one-third of the colon involved participated in this cross-sectional questionnaire study performed at three tertiary referral IBD clinics.
In all, 288 patients were female, 262 had UC, and 252 had CD. The mean age was 48 (range, 20–88) and mean number of years with symptoms was 20 (range, 7–51); 70.8% reported “my doctor” as an extensive information source. The mean perceived lifetime risk of developing colon cancer without having routine colonoscopies was 56% (SD 24.193). We developed and validated a scale of 10 important messages that IBD patients remember doctors discussing with them (“Doctor Told Scale”). Higher scores correlated with better quality of life (P < 0.001) and positive descriptors of colonoscopies and IBD (P < 0.001). Patients with higher scores perceived a higher chance of getting colon cancer without having surveillance colonoscopies (P < 0.001) and were more likely to report that the correct surveillance interval is every 2 years (P < 0.01).
Patients who remember their doctor's messages are more likely to have a positive outlook about colonoscopies and IBD, have a better quality of life, undergo surveillance colonoscopies at the correct interval, and perceive cancer risk more realistically. (Inflamm Bowel Dis 2012)
It is well established that patients with chronic extensive ulcerative (UC) or Crohn's (CD) colitis have an increased risk of colorectal cancer. From one large meta-analysis, the risk of cancer in patients with UC is estimated at 2% after 10 years, 8% after 20 years, and 18% after 30 years of disease.1 Data from a UK 30-year surveillance program calculated the risk of cancer to be 2.5% at 20 years, 7.6% at 30 years, and 10.8% at 40 years of disease.2 Studies in extensive Crohn's colitis have quoted a similarly increased risk.3, 4
Individuals with chronic extensive UC and Crohn's colitis need periodic colonoscopic surveillance due to their increased risk of colon cancer. The Crohn's and Colitis Foundation of America (CCFA) and the American Gastroenterological Association recommend having a colonoscopy every 1–3 years after 8–10 years of disease.5–7 Although physicians have known about this increased risk of colon cancer for decades, it is not clear whether patients perceive their increased risk. In order to better understand and educate patients, we developed and validated a questionnaire exploring adherence to surveillance colonoscopies, patient risk perception and worry, quality of life, attitudes toward colonoscopies and inflammatory bowel disease (IBD), and sources of information about IBD. This questionnaire was complemented by medical chart review of all individuals participating in this study. In this article we detail the significant findings regarding patient worry and risk perception, sources of information about IBD, attitudes regarding IBD and colonoscopies, self-reported adherence, and recollection of doctor message.
PATIENTS AND METHODS
All patients with UC or Crohn's colitis for at least 7 years and at least one-third of the colon involved were asked to participate in this cross-sectional questionnaire study. The 20-page questionnaire was self-administered by each patient and took ≈40 minutes to complete. A chart review was also performed for each patient. The study was conducted from October 2007 to May 2009. The Inflammatory Bowel Disease Centers of three major Boston hospitals participated in this study: Brigham and Women's Hospital (BWH), Beth Israel Deaconess Medical Center (BIDMC), and Boston Medical Center (BMC).
The questionnaire was first validated by using focus groups of healthcare workers (physicians, physician assistants, nurses, administrative staff) and then piloted in three different focus groups of 20 patients each. After each focus group the questionnaire was revised to make it more understandable to patients and more reflective of common patient concerns regarding colonoscopy and IBD. Consent was obtained from each patient prior to administering the questionnaire and performing chart review. Charts were abstracted for demographic data, medication use, colonoscopy and pathology results, and disease type, location, and severity.
A bank of items on the questionnaire was designed to be a measure of what patients remember their doctor having told them (Fig. 1). For each of the 10 items in the question bank, patients were scored 0 if they marked “My doctor did not tell me” or “My doctor might have told me” and 1 if they marked “My doctor definitely told me.” The sum of each patient's answers ranged from 0 to 10. Patients who were included in the analysis responded to all 10 items on the “Doctor Told Scale.”
The scale is reliable with a Cronbach's alpha of 0.722. The content validity of the scale was verified by close examination of the items that constitute the scale. We confirmed that they accurately reflect important information regarding IBD, colonoscopies, and colon cancer by testing the relevant items in this survey question bank (along with the entire questionnaire) on focus groups of healthcare workers and pilot groups of patients. Of six physicians on a panel reviewing this scale's content, 100% concurred that none of the scale's 10 items should be excluded from the questionnaire. No additional items not already included on the survey instrument were endorsed by 100% of the panel.
To confirm the scale's construct validity, we utilized a multiple-group factor analysis approach.8–10 Multiple groups of variables were selected from the data collected for this study, including multiple traits assessed (e.g., patients' memory of doctor messages as well as patients' perceptions of cancer risk decrease if the messages were heeded) and two methods (survey data collection as well as chart extraction). A principal components analysis without rotation confirmed the multiple groups of variables that would be used for the multiple-group factor analysis, as well as which variables belonged in which variable group. A second set of principal components analyses calculated each variable group's separate single-factor score using the regression method. Both convergent and divergent validity were demonstrated by a correlation matrix of the factor scores (Table 1). Measures that correlate with the Doctor Told Scale, such as using one's doctor as an extensive information source, support convergent validity, while measures assessing distinct phenomena, such as patient perceived impact of doctor recommendations on cancer odds or frequency of recent doctor visits, do not correlate, and thus support divergent validity. We similarly confirmed the scale's concurrent validity (criterion-related validity) through multiple correlations of the Doctor Told Scale with survey-collected items, which are reported in detail in the Results section. The Doctor Told Scale correlated with items that assessed knowledge of the respondents' disease, knowledge of additional doctor messages not queried by the scale, and self-reported adherence to colonoscopy surveillance guidelines.
Table 1. Validation of Doctor Told Scale
To demonstrate its construct validity, we correlate the Doctor Told Scale below with two variables and three factor scores (Spearman's rho)
Multiple-Group Factor Analysis Factors:
Correlation Coefficient With Doctor Told Scale
The Doctor Told Scale (measuring patients' memory of doctor messages) is included in the first factor; correlations of the other two variables in this initial factor with the Doctor Told Scale are statistically significant and included as convergent validity confirmation. Correlations of the Doctor Told Scale with the three factor scores are not significant (P > 0.05) and consequently serve as divergent validity confirmation.
Doctor Told Scale With Doctor Information Source Items Factor*
Doctor cited as a main information source, alone or in combination with other sources
Information from: My Doctor
Information Sources (Factor Score)
Includes items describing all non-doctor information sources, such as nurses/hospital staff, family, friends, and other patients.
Patient-Perceived Impact of Doctor Recommendations on Cancer Odds (Factor Score)
Includes items for assessing patients' perceived impact on their odds of cancer if they have: 1) doctor-recommended colonoscopies; 2) disease worsening; 3) compliance with prescriptions; and 4) blood-relatives with IBD.
Chart Visits (Factor Score)
Includes 2-3 year running average of GI clinic visits and GI surgery visits.
All statistical analyses were performed using SPSS for Macintosh software (v. 16.0, Chicago, IL). Descriptive statistics were calculated to report the items' central tendencies and distributions. When data did not distribute normally, nonparametric statistical analyses were performed.
A Spearman's correlation coefficient was calculated to ascertain any statistical relationship between two ordinal fields or one ordinal and one continuous field. A Mann–Whitney U-test was performed when dichotomous variables indicated group membership and nonnormally distributed fields were the test variables. Kruskal–Wallis H-tests were performed on nonnormally distributed fields when the group variables were nominal and indicated three or more groups.
The Institutional Review Boards at all three hospitals approved of the study.
Demographics (n = 514)
Of the medical records queried at all three hospitals, there were 514 patients with either CD or UC who met the inclusion criteria and consented to participate in this study. Of 411 eligible patients from the BWH group, 296 patients consented for the study and completed the questionnaire (72%). Of 222 eligible patients from the BIDMC group, 180 consented and completed the questionnaire (81%). Of 53 eligible patients from the BMC group, 38 consented and completed the questionnaire (72%). Not every patient answered every question, and the total number of patients who answered each block of questions is noted in each respective section. In all, 288 patients were female, 262 had UC, and 252 had CD. The mean age was 48 (range, 20–88; SD 14.27) and the mean number of years with symptoms was 20 (range, 7–51; SD 9.885). The number of years since the onset of symptoms was derived from patients' medical charts.
Pertinent demographic data including smoking history, race/ethnicity, family history of colon cancer and IBD, relationship status, employment status, income, and health insurance are depicted in Table 2. The variable “family history includes colon cancer” was created as a dichotomous field indicating that the respondent marked at least one hereditary relationship that included a family member diagnosed with colon cancer.
Table 2. Demographics
Stopped prior to IBD diagnosis
Stopped after IBD diagnosis
Use and plan to quit
Use but do not plan to quit
Family History of IBD
Living with someone, not married
Race/Ethnicity (All that applied were marked)
Employment Status (All that applied were marked)
Unemployed, currently looking
Unemployed, not looking
Family History of Colon Cancer
Some graduate school
Less than a high school diploma
Tech, trade, or vocational school
150K or more
No health insurance
Disease Activity and Quality of Life (n = 514)
We included the validated Short Form Inflammatory Bowel Disease Questionnaire11 (SIBDQ) in our patient questionnaire. The SIBDQ is a simple, 10-item questionnaire that can be easily scored and interpreted by clinicians. The SIBDQ measurement of health-related quality of life in patients with IBD gives results parallel to those obtained with the full 32-item IBDQ.12
There were 10 individual questions with scores ranging from 1 (worst) to 7 (best). The sum is divided by 10 to calculate a scale score in the 1–7 range with 1 being poor health-related quality of life and 7 being optimum health-related quality of life. The mean score for CD patients was 5.31 (range, 1.3–6.9; SD 1.21). The mean score for UC patients was 5.39 (range, 1.9–7.0; SD 1.20). These correspond with scores seen in inactive CD and UC, respectively.
Opinions About Colonoscopies (n = 513) and IBD (n = 504)
To better understand patients' opinions regarding colonoscopies and IBD, we asked them to mark a series of descriptors, both positive and negative, regarding each topic. These descriptors were taken from patient interviews during the validation process of the survey instrument. The responses are depicted in Figure 2a–d. These were later correlated with our “Doctor Told Scale.”
Self-reported Adherence to Colonoscopies (n = 513)
Part of our questionnaire asked patients to report how frequently they have “actually had a colonoscopy” after their IBD was diagnosed. In all, 3.4% of patients reported that they had a colonoscopy every 6 months or less, 35.4% every year, 33.4% every 2 years, 23.5% every 3 years or longer, 3.2% reported that they did not know, and 1% reported that their doctor had not told them that they needed a colonoscopy. When asked, “How frequently should you have a colonoscopy, according to your doctor,” 3.7% of patients said their doctor told them they needed one every 6 months, 42.3% every year, 33.1% every 2 years, 10.3% every 3 years, 9.2% did not know, and 1.4% said their doctor did not tell them.
Risk Perception and Worry (n = 505)
Patients' mean perceived percentage risk of developing colon cancer during their lifetime without having routine colonoscopies was 56% (SD 24.193) (Fig. 3). Patients with longer disease duration were more likely to report a lower perceived risk of colon cancer (P < 0.05). In comparison, the mean perceived percentage risk of developing colon cancer with routine colonoscopies was 38.5% (SD 25.019) (Fig. 4). For this item, there was no relationship to disease duration. Patients who were Jewish indicated that they had a lower risk of colon cancer if they did not have routine colonoscopies (P < 0.05). Patients who had a family history of colon cancer indicated that they had a higher risk of colon cancer without routine colonoscopies (P < 0.001), as well as with routine colonoscopies (P < 0.05). Patients who had fewer years of illness indicated that they had a higher risk of colon cancer without routine colonoscopies (P < 0.05).
On a 1–10 scale, with 1 being Not Worried At All and 10 being Extremely Worried, the average worry of IBD patients for developing colon cancer was 5.42. When asked about their level of worry about getting colon cancer when compared to the average person without IBD, on a 1–5 scale with 1 being having no worry and 5 being worrying the most, IBD patients had a mean worry score of 3.45. On the 10-point scale, the mean worry of patients was statistically indistinguishable from the midpoint calculated on the scale (5.5), given the patients' mean worry having a standard error of 0.11. The worry level of IBD patients on the 5-point scale, however, exceeded the midpoint of that scale, which is 3, by more than 10 times the standard error of the mean (SE = 0.041). Both values can be interpreted as an elevated level of worry.
When asked how much risk would cause them to never miss a colonoscopy (in terms of the number of times they would be more likely to develop colon cancer compared to someone without IBD), IBD patients responded that they would have to be on average 11.62 times more likely to get colon cancer. Self-reported adherence to colonoscopy was not related to such a risk threshold or to worry level.
In order to determine whether patients feel they can alter their risk of colon cancer, we asked a number of questions regarding healthy and unhealthy actions and genetics. Patients were given a 1–5 scale with 1 being the lowest chance, 5 being the highest chance, and 3 the same chance, and they were asked to compare their chances of getting colon cancer in their lifetime to the average person without IBD. Roughly two-thirds of patients on each separate item indicated that their odds of colon cancer would be the same as or lower than the average person without IBD if the healthy actions detailed in Figure 5 were relevant to them and they took them (n = 229–476 depending on the individual healthy action and whether it was relevant to each patient; Fig. 5). In all, 58.8% of patients said that if they have colonoscopies as often as their doctor recommends, their chance of colon cancer would be the same as or lower than a person without IBD (n = 493); 86.7% felt that their chance of colon cancer was higher or highest if they had a blood relative who had colon cancer (n = 444), and 83.3% felt their chance of colon cancer was higher or highest if they had a blood relative who had CD or UC (n = 436).
Sources of Information (n = 514)
To understand where our patients were obtaining information about how to manage their IBD, we included a section in our questionnaire on sources of information. With regard to a specific source of information, patients were asked to mark, “I have not used this source,” “I have used this source sometimes,” or “I have used this source extensively.” In all, 70.8% of patients marked “my doctor” as an extensive source of information. Other extensive sources of information are detailed in Figure 6. Of patients citing that they had a main source of information (n = 467), 77.3% listed their doctor as their main source of information, and 80.1% had used their doctor as a main source, alone or in combination with other sources. In all, 10.9% cited CCFA materials as a main course, alone or in combination with others.
Education level was positively correlated to using “my doctor” (P < 0.01) and CCFA materials (P < 0.001) as extensive sources of information. The higher the level of education, the more likely patients were to use these sources of information extensively. Income was not related to source of information.
Doctor Message (n = 509)
Since 70.8% of patients had used their doctor as an extensive source of information, we investigated whether these patients could recall more about what their doctor had told them. As described in Patients and Methods, we developed and validated a scale of 10 important messages that IBD doctors should discuss with their patients, hereafter referred to as the “Doctor Told Scale.” We used the scale as a measure of what patients thought their doctor told them, among a list of doctor messages detailed in Figure 1. Of 509 patients who successfully completed the Doctor Told Scale for each item, the mean score was 5.62, median score 6.0, SD 2.282, and range 0–10.
If patients used their doctor as an extensive source of information, they were much more likely to remember their doctor's messages. The mean score on the Doctor Told Scale was 5.83 (SD 2.23) for those who used their doctor as an extensive source of information. The mean score was 5.13 (SD 2.34) for those who only used their doctor sometimes or who did not use their doctor as a source of information. This group difference is statistically significant (P < 0.01).
Patients who recalled their doctor conveying more messages also reported multiple benefits. Higher scores on the Doctor Told Scale correlated with better quality of life (P < 0.001). Patients with higher scores selected more positive descriptors of the IBD experience (P < 0.001) and chose fewer descriptors indicating emotional and physical discomfort related to colonoscopy (P < 0.05). Lower scores on the Doctor Told Scale correlated with selecting more negative descriptors and emotions about the IBD experience (P < 0.05) and more negative descriptors and emotions about colonoscopies (P < 0.05).
Theoretically, recall of doctor message should relate to, even predict, patients knowing other facts about their disease and its treatment apart from the specific messages included in the scale. Three patient knowledge measures on the survey, independent of the Doctor Told Scale, reinforce the scale's criterion-related validity: Patients who remembered what their doctor told them were more likely to report that the correct surveillance interval is every 2 years (P < 0.01) or 2 or fewer years (P < 0.001), and patients with higher Doctor Told scores perceived they had a higher chance of getting colon cancer without having surveillance colonoscopies (P < 0.001). Furthermore, we related the Doctor Told Scale to self-reports of how frequently patients actually had colonoscopies, and found a positive correlation (P < 0.001). Recalling a doctor's directive is evidenced by adherence to that directive.
In addition, older patients and patients who listed their status as “retired” scored higher on the Doctor Told Scale; i.e., they were more likely to remember their doctor's messages (P < 0.001 for age and P < 0.01 for retired). There was no correlation, either positive or negative, of the Doctor Told Scale with level of worry or with the perception of the impact of healthy actions on the chance of getting colon cancer, e.g., exercise, prayer, reducing stress, keeping sense of humor, and others detailed in Figure 5.
In this study we attempted to determine the factors that most influenced patients' attitudes and actions regarding IBD and colon cancer. It is reassuring that over three-fourths of patients use their doctor as their main source of information and that the doctor message is getting through to patients. The Doctor Told Scale is a measure of what patients remember their doctor having told them. Multiple analyses provide compelling evidence of the scale's construct, content, and criterion-related validity. The scale clinically relevant because it demonstrates that patients who recall what their doctors have said experience better outcomes. This should inspire physicians to better educate their patients and also patients to listen more diligently to their doctors' messages.
This result is similar to the results of smoking cessation studies. A large meta-analysis of physician advice for smoking cessation showed that even a brief advice intervention can increase quitting by 1%–3%. There is also a small benefit of follow-up visits and more intensive interventions.13 Similar to smoking cessation, recalling physician advice in our study was associated with patients self-reporting surveillance at the correct 2-year interval as well as patients having more realistic risk perceptions of colon cancer. It is reassuring that remembering doctor message does improve patient outcomes, and thus physicians must provide accurate information as clearly as possible. The CCFA can certainly help, through physician and patient educational resources and programs.
Although patients use their doctor as an important source of information, it seems that patients are not taking similar advantage of the extensive CCFA materials both printed and online. At the time of our study in 2007–2009, only 9.7% of patients extensively used CCFA mailings and 9.1% of patients extensively used the CCFA website. The fact that patients overestimate their risk of colon cancer by nearly 10-fold means that patients are likely to have significant educational benefit from using CCFA resources. Despite doctor message and various educational resources, there is still a significant impediment to getting surveillance colonoscopies done. Paradoxically, even though IBD patients have a high degree of worry about getting colon cancer and greatly overestimate their cancer risk in this study, they would have to perceive themselves as almost 12 times more likely than the average person to get colon cancer in order to avoid missing a colonoscopy. Self-reported adherence to colonoscopy was not related to worry level or to each individual's risk threshold to ensure adherence.
Other published studies, both in Barrett's esophagus and IBD, complement this overestimation of cancer risk by patients. In a study of Barrett's esophagus by Shaheen et al,14 more than two-thirds of Barrett's patients grossly overestimated their risk of esophageal cancer, with a mean estimated 1-year cancer risk of 13.6%. In a study by Siegel et al15 IBD patients quantified their risk of colon cancer as 23% in the “next 10 years.” Despite this high estimated risk, patients in the Siegel et al study would only agree to a colectomy if their risk of having colon cancer “right now” were at least 73%. In a study by Baars et al16 IBD patients estimated their 10-year colorectal cancer risk at 25%. Patients would agree to a colectomy if their current colorectal carcinoma risk were at least 53%. Our study reflects this overestimation of risk and emphasizes the reluctance of patients to undergo procedures in order to reduce their personal risk.
Some hospital differences should be noted. One difference among the hospitals is patients' worry about getting colon cancer. Patients seen at the BIDMC and BMC are more worried (5.71 out of 10 and 6.39 out of 10, respectively) than BWH (5.12 out of 10) (P < 0.01); however, BIDMC and BMC cannot be ascertained as statistically different from each other. For self-reported adherence, BIDMC patients claim that they come in more often than anyone else and that their doctors tell them to do so (P < 0.001). BWH patients claim they come in less often and that this is also at the advice of their doctors (P < 0.001). No other significant differences among hospitals were noted.
Our measurement of self-reported adherence should be further discussed. In our questionnaire we asked patients to report how frequently they had “scheduled” a colonoscopy and how frequently they had “actually had” a colonoscopy after their IBD was diagnosed. We did not specifically ask if the colonoscopy was for increased symptoms or strictly for surveillance for colon cancer, as we were doubtful patients could recall the difference; furthermore, a colonoscopy for increased symptoms could have fallen within the window of time appropriate for a surveillance colonoscopy, and we chose not to measure patients' ability to ascribe an accurate clinical motivation for their history of individual procedures. We also did not ask them how many colonoscopies they had had after 7 years of disease (a reasonable start of colonoscopic surveillance). While this may be a weakness of the study, realistically we did not think patients could recall the numbers accurately. Our goal in the design of these items was to assess strictly whether patients were indeed showing up for and completing their colonoscopies.
As one of the first studies in IBD to look at the effect of doctor message on patient beliefs and behavior, the strengths of this study are the high number of patients enrolled, the high response rates to the questionnaire, and the extensive analysis of all the interrelated questions and responses. To participate, the patients all visited university-based practices, which suggests a study weakness in that responses found in private practice may not be similar. Additionally, the sample was predominantly white, was one-third Jewish (higher than average for IBD patients), and had attained higher levels of education and higher levels of household incomes on average. Another factor that might affect adherence is that all patients in Massachusetts are mandated to have health insurance; thus, the patients in this study might have had greater medical access than others in different states, consequently increasing their adherence.
In summary, we demonstrate that the majority of patients use their doctor as an extensive source of information about their IBD and colonoscopies. Patients in our study have an elevated degree of worry about getting colon cancer and greatly overestimate their lifetime risk of colon cancer, but express reluctance to undergo colonoscopic surveillance. In addition, many of our patients harbor negative opinions regarding IBD and colonoscopies. It is very helpful that something as simple as remembering the doctor message can have a positive impact on our patients' lives by fostering more realistic beliefs and expectations. Patients who remember what their doctors tell them are more likely to have a more positive outlook about their IBD and colonoscopies, have a better quality of life, report that they have undergone colonoscopies at the correct 2-year interval, and have more realistic risk perceptions about colon cancer.