A Theory-based Educational Pamphlet With Low-residue Diet Improves Colonoscopy Attendance and Bowel Preparation Quality

Abstract

Goals/Background:

Patients who “no-show” for colonoscopy or present with poor bowel preparation waste endoscopic resources and do not receive adequate examinations for colorectal cancer (CRC) screening. Using the Health Belief Model, we modified an existing patient education pamphlet and evaluated its effect on nonattendance rates and bowel preparation quality.

Study:

We implemented a color patient education pamphlet to target individual perceptions about CRC and changed bowel preparation instructions to include a low-residue diet instead of the previous clear liquid diet. We compared the nonattendance rate over a 2-month period before and after the introduction of the pamphlet, allowing for a washout period during which pamphlet use was inconsistent. We compared the Boston Bowel Preparation Scale (BBPS) in 100 consecutive patients who underwent colonoscopy during each of the 2 periods.

Results:

Baseline characteristics between the 2 groups were similar, although patients who received the pamphlet were younger (P=0.03). The nonattendance rate was significantly lower in patients who received the pamphlet (13% vs. 21%, P=0.01). The percentage of patients with adequate bowel preparation increased from 82% to 86% after introduction of the pamphlet, although this was not statistically significant (P=0.44). The proportion of patients with a BBPS score of 9 was significantly higher in the pamphlet group (41% vs. 27%, P=0.03). There was no difference in adenoma and sessile serrated adenoma detection rates before and after pamphlet implementation.

Conclusions:

After implementing a theory-based patient education intervention with a low-residue diet, our absolute rate for colonoscopy nonattendance decreased by 8% and the proportion of patients with a BBPS score of 9 increased by 14%. The Health Belief Model appears to be a useful construct for CRC screening interventions.

In the United States, colorectal cancer (CRC) is the third most common cancer in both women and men and the second leading cause of cancer mortality.¹ The incidence of CRC has been declining since the mid-1980s due to a combination of population-level reductions in modifiable risk factors and an increase in screening.² Although the fecal occult blood test is widely used for CRC screening world-wide, colonoscopy is the predominant screening modality in the United States.³ In 2015, 96% of individuals who received CRC screening completed lower endoscopy, compared with 11% for stool-based testing.⁴ Colonoscopy has been shown to reduce CRC incidence and mortality,⁵ but deriving maximal benefit requires high-quality screening examinations after adequate bowel preparation.⁶ Nonattendance or “no-show” at colonoscopy appointments and inadequate bowel preparation are significant barriers to the concerted national campaign to increase CRC screening rates above 80% by 2018.

Colonoscopy remains a limited resource that is further depleted by nonattendance and poor bowel preparation. Although a recent modeling study in the United States suggested that resources are available to screen 80% of the age-eligible population by colonoscopy, it is unclear how geographic variation and missed or incomplete procedures would impact these results.⁷ A number of studies have assessed reasons for colonoscopy nonattendance, including difficulty in completing bowel preparation and preexisting medical conditions, and have also provided potential solutions.^8–13 A large body of work has examined predictors of inadequate bowel preparation, including medication use and socioeconomic factors, and their corresponding interventions.^13–17 However, despite the fact that both of these factors must be addressed to ensure a complete colonoscopic examination, to our knowledge, no study has assessed whether a single intervention can improve both of these outcomes.

We previously reported on the effectiveness of a patient education pamphlet, which was created using the principles of the Health Belief Model, on bowel preparation quality in a safety-net hospital.¹⁸ In the present study, we investigated the effect of this modified version of the pamphlet and a low-residue diet on both colonoscopy attendance and bowel preparation in an urban Department of Veteran Affairs (VA) hospital.

MATERIALS AND METHODS

Pamphlet Design and Implementation

The original version of the patient education pamphlet used was created by Spiegel et al¹⁹ using principles of the Health Belief Model and based on interviews with patients and providers. Subsequent work by our group demonstrated the effectiveness of a revised pamphlet in a safety-net hospital with a large proportion of patients with low English proficiency.¹⁸ Consistent with standard clinical practice at the time, the previous pamphlets both used a clear liquid diet for bowel preparation. To address the issues of colonoscopy attendance and bowel preparation quality at the VA New York Harbor Health Care System Manhattan Medical Center, we held a series of meetings with an interdisciplinary group of gastroenterology attendings, fellows, and nurses. We adapted the Health Belief Model for screening/surveillance colonoscopy—the leading indications for an outpatient colonoscopy—and used this framework to identify modifiable factors that impact the likelihood of undergoing successful colonoscopy (Fig. 1). We identified cues to action and patient resistance to the clear liquid diet as factors in the model that could be targeted for intervention using an educational pamphlet. We then adapted the English version of the pamphlet we had previously used at the safety-net hospital and added the website address for a video that explains the bowel preparation process using our standard purgative regimen (2L PEG-3350, MoviPrep; Salix Pharmaceuticals). In our predominantly elderly male veteran population, we hypothesized that patients could better remember preparation directions and would be more motivated to complete the procedure by reading the color educational pamphlet compared with our previous set of black and white handouts, which contained little information about colonoscopy or CRC. Patients who are visual learners and have internet access could use the website as a supplementary guide. In addition to targeting patient education with our pamphlet, we also modified the prescribed bowel preparation. Instead of the previously recommended clear liquid diet, we added a low-residue diet menu to the pamphlet to address the common feedback that patients experienced intolerable hunger on a clear liquid diet. The menu was adapted from a randomized controlled trial that showed that a low-residue diet had equal efficacy and better patient tolerance compared with clear liquid diet in both veterans and non-veterans.²⁰ We created both a standard 1-day bowel preparation and an extended 2-day bowel preparation version of the pamphlet, both of which used the low-residue diet (Fig. 2).

FIGURE 1.

Health Belief Model for screening/surveillance colonoscopy. CRC indicates colorectal cancer.

FIGURE 2.

Select sections of patient education pamphlet. Panels show simple facts, cues to action with graphics to explain the bowel preparation process, and elements of the low-residue diet.

The pamphlet was distributed to patients starting in November 2016 as a quality improvement project. Per our routine colonoscopy workflow, all patients were assessed in the clinic before the procedure by either a nurse practitioner or gastroenterology fellow. After the colonoscopy date was selected, a nurse provided one-on-one instruction for the bowel preparation using the pamphlet. By default, patients received the standard 1-day bowel preparation pamphlet. Patients with risk factors for inadequate bowel preparation (eg, previous inadequate bowel preparation, chronic constipation, opioid use) were instructed to undergo a 2-day bowel preparation at the discretion of the clinician. All patients were instructed to undergo split-dose bowel preparation with 2L PEG-3350 (4L for 2-day bowel preparation).

Study Design

After the implementation of the educational pamphlet, we retrospectively evaluated the effectiveness of the pamphlet on colonoscopy attendance rates and bowel preparation quality. The study was approved by the VA New York Harbor Health Care System Institutional Review Board (IRB #01655). Patients over 18 years of age who underwent an outpatient colonoscopy performed for any indication (screening, surveillance, or diagnostic) between September 2016 and April 2017 were eligible for inclusion. We excluded inpatient procedures because the pamphlet was not used in the inpatient setting, procedures that were incomplete due to reasons unrelated to bowel preparation (eg, technical difficulty or patient intolerance), and those that were not scored using the validated Boston Bowel Preparation Scale (BBPS).²¹ Patients who underwent a colonoscopy both before and after the pamphlet were only included in the pre-pamphlet group. To allow for a washout period during which the pamphlet was used inconsistently and previously scheduled colonoscopies were performed, we chose 2 non-contiguous periods before and after the introduction of the pamphlet in November 2016: September to October 2016 for the pre-pamphlet group and February to March 2017 for the post-pamphlet group. Our 2 primary outcomes of interest were the colonoscopy nonattendance or “no-show” rate and rate of adequate bowel preparation quality. For the “no-show” outcome, we analyzed nursing data that tracked attendance for endoscopy patients over the corresponding 2-month periods before and after the pamphlet introduction. A patient was defined as “no-show” if he/she did not present to the endoscopy unit on the appointment date and also did not cancel or reschedule at any point leading up to the appointment. Thus, patients who canceled on the day of the procedure were not considered “no-shows” under this definition. For the bowel preparation quality outcome, we included the first 100 consecutive patients who underwent colonoscopy during each of the 2 periods (Fig. 3). An adequate bowel preparation was defined as a BBPS score of at least 2 in each of the 3 colonic segments. We extracted data on colonoscopy indication, cecal intubation, BBPS score by segment, fellow participation, and adenoma detection from the electronic health record. We assessed mean BBPS score, percentage of procedures with BBPS score of 9, adenoma detection rate, and sessile serrated adenoma detection rate as secondary outcomes.

FIGURE 3.

Flow chart of patients included in study. BBPS indicates Boston Bowel Preparation Scale.

Statistical Analyses

We used the χ² test to compare proportions and the Student t test to compare continuous variables. A 2-sided P-value of <0.05 was considered statistically significant. The analysis was performed using R Statistical Software (Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Patient and Procedure Characteristics

Baseline characteristics of the pre- and post-pamphlet groups are summarized in Table 1. The reasons for exclusions are summarized in Figure 3, with only 3 patients excluded for not having a documented BBPS score. The vast majority of patients were male (95%). Most procedures were performed for screening or surveillance (85%), and the remaining 15% were diagnostic examinations. Patients in the post-pamphlet group were slightly younger (61 vs. 64 y, P=0.03). Cecal intubation rate and fellow involvement were similar in the 2 groups.

TABLE 1.

Patient and Procedure Characteristics

Characteristic	Pre-pamphlet (N=100)	Post-pamphlet (N=100)	P
Age [mean (SD)]	64 (9)	61 (11)	0.03
Gender, male (%)	94	95	0.76
Indication (%)			0.73
Screening	40	38
Surveillance	44	49
Diagnostic	16	13
Cecal intubation (%)	98	94	0.15
Fellow involvement (%)	62	56	0.39

Boldface indicate statistical significance P<0.05.

Colonoscopy Nonattendance Rate

During the 2-month period before the intervention, 55 of 259 (21%) scheduled patients were “no-shows”; in the 2-month post-intervention period, 38 of 295 (13%) scheduled patients were “no-shows” [odds ratio (OR), 0.55; 95% confidence interval (CI), 0.35–0.86; P=0.01]. This corresponded to a 38% relative reduction and 8% absolute reduction in “no-shows,” which meant there was one fewer “no-show” for every 12.5 patients who received the pamphlet.

Bowel Preparation Quality

We compared the BBPS score for the first 100 consecutive patients in each time period who met inclusion criteria (Table 2). The average total BBPS score was similar in the pre- and post-pamphlet groups (7.1 vs. 7.4, P=0.22). The rate of adequate bowel preparation was slightly higher in the post-pamphlet group, although this was not statistically significant (86% vs. 82%, OR, 1.3; 95% CI, 0.63–2.94; P=0.44). The percentage of procedures with a maximum BBPS score of 9 was higher in the post-pamphlet group (41% vs. 27%; OR, 1.9; 95% CI, 1.03–3.43; P=0.03). This corresponded to a 52% relative increase and a 14% absolute increase. Figure 4 shows the BBPS score distribution by group.

TABLE 2.

Pre- and Post-intervention Outcomes

Outcome	Pre-pamphlet	Post-pamphlet	Odds Ratio (95% CI)	P
Nonattendance (%)	21	13	0.55 (0.35–0.86)	0.01
Adequate preparation (%)	82	86	1.3 (0.63–2.94)	0.44
BBPS total (mean)	7.1	7.4		0.22
BBPS=9 (%)	27	41	1.9 (1.03–3.43)	0.03
Adenoma detection (%)	56	55	1.0 (0.58–1.8)	0.60
SSA detection (%)	1	1	1.0 (0.03–39.4)	1.00

Boldface indicates statistical significance P <0.05.

BBPS indicates Boston Bowel Preparation Scale; CI, confidence interval; SSA, sessile serrated adenoma.

FIGURE 4.

Boston Bowel Preparation Scale outcomes before and after patient education pamphlet. Adequate is defined as Boston Bowel Preparation Scale > 1 in all segments.

Adenoma Detection Rate

There was no statistically significant difference between the pre- and post-groups for either the adenoma detection rate (56% vs. 55%; OR, 1.0; 95% CI, 0.58–1.8; P=0.60) or the sessile serrated adenoma detection rate (1% vs. 1%; OR, 1.0; 95% CI, 0.03–39.4; P=1.00).

DISCUSSION

We showed that use of a patient education pamphlet that was created and modified using principles of the Health Belief Model and which incorporated a low-residue diet led to an 8% absolute reduction in colonoscopy nonattendance rate and an adequate bowel preparation rate above 85%. One fewer “no-show” was observed for every 12.5 patients who received the pamphlet. These findings extend the work of Spiegel et al¹⁹ and our previous study¹⁸ by demonstrating that a simple, low-cost intervention can improve bowel preparation quality and also increase colonoscopy attendance rates. To our knowledge, this is also one of the first studies showing the effectiveness of a low-residue diet for bowel preparation in a veteran population.

The Health Belief Model is a widely used framework to conceptualize and implement health behavior change.²² Our theory-based assessment of barriers to colonoscopy identified stronger cues to action and resistance to the clear liquid diet as easy targets for intervention using a simple pamphlet. Since important aspects of the model—such as individual perception and modifying factors—may be different in other clinical settings and patient populations, we recommend adopting a structured theory-based approach when addressing health behavioral issues such as endoscopy attendance and bowel preparation quality.

Colonoscopy nonattendance is important for health care institutions because of the delay in clinical diagnoses as well as financial implications of wasted limited endoscopic resources.²³ Even though the VA health care system has achieved national CRC screening rates above 80% for many years, nonattendance is still a highly prevalent problem. In a large cohort of veterans who had scheduled appointments following a positive fecal occult blood test, 38% missed, canceled, or rescheduled them.⁸ In the current study, we used a strict definition of nonattendance that excluded lastminute cancellations and found a 21% baseline “no-show” rate. It should be noted that due to the difficulty in filling colonoscopy slots on short notice, cancellation and rescheduling in the few days leading up to a procedure almost always results in unused endoscopy time; thus, the percentage of wasted endoscopic capacity was likely substantially higher than our measured nonattendance rates. On the basis of a discrete event simulation model for colonoscopy nonattendance in a non-VA setting, the reduction in “no-show” rate from 21% to 13% that we observed would have resulted in ~$500 net gain in revenue per day.²⁴ Moreover, as most “no-show” patients will likely be rescheduled into a future colonoscopy appointment, access to colonoscopy for other veterans will be compromised. Therefore, a lower nonattendance rate can simultaneously increase cost efficiency for the endoscopy unit and improve patient access to colonoscopy.

Interventions previously shown to improve endoscopy attendance include opt-in scheduling, patient reminders,^8–10,25 educational efforts such as pre-procedure clinics and nursing follow-up contact,^11,26 as well as predictive overbooking.²⁷ We have already incorporated opt-in scheduling, nursing telephone reminders, and pre-procedure clinic visits in our routine practice, although the latter requires an additional appointment for patients and may not be an option in an open-access endoscopy model. Predictive overbooking requires knowledge of predictors of nonattendance and willingness to tolerate scheduling uncertainties on the part of physicians and staff, which limit its widespread adoption. The optimal intervention would be inexpensive, not influence endoscopy scheduling, serve as a procedure reminder, and provide on-demand instructions for bowel preparation, as an inadequately cleaned colon also results in wasted capacity. The patient education pamphlet meets all of these criteria, and the inclusion of a low-residue diet in our version of the pamphlet likely played a major role in increasing attendance by improving compliance with the bowel preparation regimen.

The conventional bowel preparation regimen calls for a clear liquid diet on the day before the procedure, but the difficulty of adhering to this regimen has led to extensive research on the low-residue diet. Although studies of the low-residue diet vary in how extensively it is used during the day—for example, breakfast only versus full day—most have demonstrated that it increases patient satisfaction and tolerability of bowel preparation and does not sacrifice efficacy of colon cleansing.^28–32 Given these findings, the 2015 American Society for Gastrointestinal Endoscopy guidelines recommended a low-residue diet over a clear liquid diet for bowel preparation.³³ However, we only found a single abstract and no full publications that have investigated the efficacy of the low-residue diet in veterans.²⁰ As the VA population, in general, is older and sicker than the civilian population,³⁴ a less restrictive bowel preparation was appealing. Our findings provide evidence that a low-residue diet may improve bowel preparation quality. We believe that the combination of a liberalized diet, answers to frequently asked questions, directions to a website explaining bowel preparation, and a colorful visual guide all led to the improved outcomes noted in our study.

The mean age of our post-pamphlet group was 3 years younger than the pre-pamphlet group. As younger age has been associated with colonoscopy “no-shows,” this demographic difference may have slightly underestimated the effect of the intervention in reducing nonattendance.^10,11 The rates of adequate bowel preparation and maximum BBPS score of 9 both increased in the post-pamphlet group, although only the latter was statistically significant. Nevertheless, the results suggest the intervention had a net positive impact on bowel preparation. A greater proportion of patients with BBPS score of 9 may also improve procedural efficiency by reducing the time spent on irrigation, although we did not measure this in the study. Studies have shown that adenoma detection rate is correlated with bowel preparation quality and that interventions designed to increase adenoma detection are less effective when baseline rates are already high.³⁵ Given the similar rates of adequate bowel preparation before and after the intervention and our high baseline adenoma detection rate, the lack of improvement in adenoma detection rate after the intervention is not surprising.

Our study has several limitations. First, since this was a quality improvement project, we did not randomize individuals to the intervention and instead used historical controls. However, there were no other changes in our colonoscopy-related practices during this period besides the intervention, and the only demographic difference between the groups would likely have biased our results toward the null. We minimized the effect of seasonal weather-related “no-shows” by excluding winter months from the analysis, which was the initial period during which pamphlets were inconsistently used. Second, we did not collect data on other demographic and medical variables that may have influenced nonattendance and adequate bowel preparation rates, such as race/ethnicity, travel distance, comorbid conditions, history of prior bowel preparation, and history of prior “no-shows.” However, there is no reason to believe that these factors would be different in the 2 groups. We only included individuals who had procedures in both periods in the pre-pamphlet group because the results of their first procedure could have influenced the outcomes of their second procedure. Third, because our intervention combined both the educational components of the pamphlet as well as a low-residue diet, we cannot isolate the effect of each component of the intervention on the outcomes. Finally, because this study population consisted predominantly of older, male veterans, these results may not be generalizable to other populations.

We showed that a simple theory-based patient education pamphlet can improve colonoscopy attendance and bowel preparation among veterans. Furthermore, a low-residue diet appears to be effective in this patient population. In future studies, we plan to incorporate patient feedback to further refine the pamphlet and examine additional demographic and medical factors that may influence nonattendance and bowel preparation quality.