Physician adherence to societal guidelines following colonoscopy with polypectomy

Introduction

Colorectal cancer (CRC) is a leading cause of cancer-related morbidity and mortality worldwide [1]. Most colon cancers develop from benign adenomatous/serrated polyps; however, less than 5% of the polyps become cancer [2]. Transformation of a polyp into an adenocarcinoma is a gradual process that occurs over 5-10 years and involves acquired genetic, epigenetic and molecular changes [3]. This gradual process creates the opportunity for intervention and prevention.

Most CRCs are sporadic and risk factors include family history, age, environmental risk factors, excessive alcohol consumption, smoking, and certain foods [4]. A very important risk factor for CRC is the presence of polyps (especially advanced polyps) during the index colonoscopy [3]. Studies have shown that colonoscopy with polypectomy reduces the relative risk of CRC by 53% [5], reduces the incidence of CRC by 48%, and reduces mortality from CRC by 65% [6,7].

Guidelines of gastroenterological societies recommend surveillance following polypectomy, according to the patient’s risk of developing additional polyps. The degree of risk is determined by the polyp size, its histological characteristics and the level of dysplasia. These risk factors form the basis for surveillance recommendations (Supplementary Table 1). Too stringent recommendations (shorter intervals between colonoscopies) will impose a significant burden on the patient and the healthcare system, while recommendations that are too lenient (long intervals between colonoscopies) can lead to missed pre-cancerous polyps and the development of cancer [8,9].

Materials and methods

Study design and participants

This was a prospective observational and interventional study conducted at 2 academic centers in Israel. The study population were consultant gastroenterologists who perform elective colonoscopies with polypectomy. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki, as reflected in a priori approval by the institution’s human research committee.

In the first phase of the study, we examined current practices. Reports of colonoscopies with polypectomies were prospectively collected over a period of 3 months, and the endoscopist’s recommendations after the colonoscopy were reviewed and compared with the guidelines. We included patients undergoing colonoscopy with polypectomy and excluded patients with a history of CRC, patients with inflammatory bowel disease, and patients with familial adenomatous polyposis, Lynch syndrome, MYH-associated polyposis, or juvenile polyposis.

After completing the first phase of the study, we conducted an intervention that included several components. First, each physician received a personal report regarding his post-polypectomy recommendations and their concordance with societal guidelines. Second, we presented (in an anonymous form) the results from the first phase in a departmental meeting and reaffirmed current guidelines, including a reference chart to help with decision making (Fig. 1). We then initiated a discussion of the steps that could be taken in order to increase the overall number of recommendations and improve adherence to guidelines. We updated the electronic form (Fig. 2) and, finally, we sent an e-mail with an explanation of the new system and the guidelines reference chart to all participating physicians (Fig. 1).

Figure 1 The chart created for the new reporting platform and introduced in the intervention phase, in order to facilitate more accurate recommendations

BBPS, Boston bowel preparation score

Figure 2 Example of the dedicated page for post-polypectomy recommendations in the medical file. The physician can select the appropriate recommendation for each polyp type and can also view a page with the guidelines (Fig. 1). The computer then automatically generates a letter, mailed to the patient along with the pathology report BBPS, Boston bowel preparation score; EMR, endoscopic mucosal resection

In the second phase, following the intervention, we again prospectively collected data over a period of 3 months. The exact time frame for data collection for both phases of the study was not disclosed to the physicians.

Results

A total of 646 patients who underwent colonoscopy with polypectomy were included in the study, 349 cases in the first phase and 297 in the second phase. In this population, 65% were male and the average age was 65.67±10.01 years. Table 1 shows the baseline demographic and clinical characteristics of our cohort.

Table 1 Baseline characteristics of the study population. There was no significant difference between populations (before and after intervention)

In the first phase of the study, 266/349 patients (76%) received a clearly documented written recommendation for surveillance colonoscopy. A recommendation consistent with societal guidelines was given in 174/266 (65%) patients. Compared with the guidelines, the recommended timing of surveillance was too early in 66 patients (25%) and too late in 26 patients (10%) (Table 2).

Table 2 Physicians’ recommendations to guidelines between before and after intervention. Dark grey represents too early recommendations, light gray represent too late recommendations, and medium gray represents consistent recommendations

In the second phase of the study, 254/297 patients (86%) received a recommendation for surveillance colonoscopy. A recommendation consistent with societal guidelines was given in 199/254 (78%) patients. Compared with the guidelines, the recommended timing of surveillance was too early in 48 patients (19%) and too late in 7 patients (3%) (Table 2).

The patients who did not receive any documented recommendation in the first and second phases might not have received any recommendation, or they might have received a recommendation not documented. Some recommendations were given by telephone or in handwriting on the pathological report sent to the patient. These data were unfortunately not available.

After the intervention, patients received significantly more recommendations (86% vs. 76%, P=0.003), and these were much more consistent with societal guidelines (78% vs. 65%, P=0.001), (Table 2, Fig. 3).

Figure 3 Comparison between recommendations and adherence to guidelines prior and after intervention

We compared different variables between adherent and non-adherent groups (Supplementary Table 2). We found that bowel preparation (P=0.001), pathologic size of polyp (P=0.001), dysplasia (P=0.003), and high-risk polyps (P=0.01) had a statistically significant influence on adherence, both before and after intervention.

Discussion

Recommendation for surveillance colonoscopy should balance the need for preventing CRC, while maintaining adequate use of medical resources and minimizing patient discomfort and morbidity. Previous data on compliance with guidelines are conflicting and, to the best of our knowledge, none of the previous studies examined the effect of intervention on adherence to guidelines.

In a large observational study, including 25 centers in the United States, 36% of physicians were non-compliant with guidelines, with a variation of 3-80% between institutions. This study also showed that, in 94.6% of the cases, physicians recommended surveillance colonoscopy earlier than required according to the guidelines. In that study, recommendations were mainly influenced by the histologic nature of the resected polyp and by the quality of bowel preparation [11]. In another study, more than 50% of physicians recommended repeat colonoscopy after 3 years or less for small adenomas, which according to the guidelines should have been after 5 years. In that study, unnecessary short intervals for repeat colonoscopy were also recommended for hyperplastic polyps in 24% of cases [12]. On the other hand, Menees SB et al evaluated physician compliance with guideline recommendations in average-risk patients undergoing colonoscopy with polypectomy of 1-2 small polyps. They found that more than 90% of the recommendations were consistent with the guidelines, and that this was again significantly influenced by the quality of bowel preparation [13]. These studies had several limitations, including their retrospective design, reliance on physicians’ reports, which can promote recall bias, and the small number of cases.

We performed a 2-phase prospective study, which also included an intervention session in 2 academic centers. Similar to previous reports, the results from our first phase of the study showed that 35% of physicians were non-compliant with guidelines; however, following an intervention we were able to decrease non-compliance to 22%. Our intervention also resulted in a significant increase in the number of patients who received a clear, well documented written recommendation.

Previous studies showed that age, bowel preparation quality, and number/types of polyps were associated with adherence to guidelines [11,13]. In our study, inadequate bowel preparation was associated with an accurate recommendation for early repeat colonoscopy (within 1 year). This is probably because an inadequate preparation warrants repeat colonoscopy regardless of other parameters, as stipulated in societal guidelines.

High-risk polyps (high-grade dysplasia, more than 3 polyps, polyps size >1 cm) were associated with non-adherent recommendations. This might be because of bad estimation of polyp size during colonoscopy, or because recommendations were issued prior to the pathological report.

Our revised electronic reporting platform allowed the physicians to fill in the recommended time for interval colonoscopy, based on the guidelines, the endoscopic report, and the final pathologic report. This user-friendly platform facilitated better compliance by our physicians, which resulted in more accurate recommendations.

Our study had several limitations: we checked the electronic form for a short period, and only a few months after using it. Long-term follow up is required in order to better understand the long-term impact. Our platform is suitable for the specific reporting software we use in our hospitals; different electronic forms should be developed for different reporting platforms. Our platform was developed for hospital settings, but a different approach may be needed for community medicine.

In conclusion, we have shown that poor adherence to societal guidelines for post-polypectomy surveillance can be overcome by a simple intervention in the form of guideline reaffirmation and a structured reporting platform.