Analysis of adenoma detection rate of colonoscopy among trainees

Article information

Ann Coloproctol. 2023;.ac.2023.00199.0028

Publication date (electronic) : 2023 September 26

doi : https://doi.org/10.3393/ac.2023.00199.0028

Young Min Song, Kyung Su Han, Byung Chang Kim, Chang Won Hong, Bun Kim, Min Chul Kim, Myeong Jae Jin, Dae Kyung Sohn

Center for Colorectal Cancer, National Cancer Center, Goyang, Korea

Correspondence to: Dae Kyung Sohn, MD, PhD Center for Colorectal Cancer, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang 10408, Korea Email: gsgabal@ncc.re.kr

Received 2023 March 10; Revised 2023 June 9; Accepted 2023 July 20.

Abstract

INTRODUCTION

Colorectal cancer (CRC) is the third most common cancer and the third leading cause of cancer death in Korea [1]. Colonoscopy plays a crucial role in the prevention and early detection of CRC; therefore, it is widely used [2, 3]. However, there are some concerns about whether colonoscopy is being performed properly; therefore, many studies on the quality and monitoring of colonoscopy have been conducted [4, 5]. Based on these research results, the need for a colonoscopy training program was raised, and gastrointestinal endoscopy departments in the United States have developed colonoscopy training programs and evaluation tools.

Indicators commonly included in colonoscopy training are cecal intubation time, cecal intubation rate, and adenoma detection rate (ADR) [6]. Among these, ADR is an important indicator in evaluating the colonoscopy quality. A low ADR is associated with the risk of interval CRC, detected on screening colonoscopy. A research result indicated that if ADR is increased by 1%, CRC can be reduced by 3% [7]. Currently, the recommended ADR level by the American Society for Gastrointestinal Endoscopy (ASGE) is ≥25% (30% in men and 20% in women) in the first colonoscopy for screening purposes.

This study aimed to analyze the quality indicators of colonoscopy, including ADR, to identify the changes and differences among trainees participating in a colonoscopy training program according to the training period.

METHODS

Data collection

The following patient and colonoscopy data were retrospectively collected through the colonoscopy result sheets: date of performing colonoscopy, patients’ age and sex, the total examination time of colonoscopy, history of abdominal surgery, history of colorectal surgery, and the number of resected polyps. The pathologic reports were reviewed to confirm the pathological diagnosis of all resected polyps. Based on these results, several endoscopic parameters were calculated. The definitions of the analyzed indicators are summarized in Table 1.

Table 1.

Definition of terms

RESULTS

Between March and November 2022, 8,075 colonoscopies were performed at the National Cancer Center (Fig. 1). Of these, 3,820 colonoscopies were selected and analyzed. Of which, 2,005 cases (52.5%) were observed by the 3 trainees. Table 2 shows the characteristics of patients who underwent colonoscopy and the outcomes of colonoscopy. Among the 3,821 patients, 1,889 (49.4%) were female, and the median age was 60 years. Furthermore, 437 patients (11.4%) had a history of abdominal surgery, whereas 344 patients (9%) had undergone colorectal surgery for CRC. The median procedure time was 13.0 minutes (IQR, 10–17 minutes) for the trainees and 10.2 minutes (IQR, 8–14 minutes) for the experts, showing a significant difference between the 2 groups (P<0.001). Among the calculated quality indicators of colonoscopy, only PDR showed no significant difference between the trainees and the experts (62.9% vs. 64.4%, P=0.348). There were significant differences in ADR (50.6% vs. 57.6%, P<0.001), ADR for men (57.2% vs. 62.8%, P=0.014), ADR for women (43.2% vs. 52.9%, P<0.001), AADR (9.3% vs. 13.5%, P<0.001), APC (1.08 vs. 1.30, P<0.001), and APP (70.9% vs. 77.7%, P<0.001) between the 2 groups.

Fig. 1.

The flowchart of included colonoscopy cases.

Table 2.

Patients’ characteristics and colonoscopy outcomes in entire patients (n=3,820)

The analysis results in the screening/surveillance group, defined to increase the reliability of ADR, are presented in Table 3. The procedure time was 13 minutes (IQR, 10.0–17.0 minutes) for trainees and 10 minutes (IQR, 8.0–13.5 minutes) for experts, showing a significant difference between the 2 groups (P<0.001). However, unlike in the analysis of entire patients, all calculated quality indicators, including ADR (51.8% vs. 53.4%, P=0.551), showed no significant differences between the trainees and the experts.

Table 3.

Patients’ characteristics and colonoscopy outcomes in the screening/surveillance group (n=1,606)

The changes in PDR, ADR, and APP of trainees in the screening/surveillance group were divided by 2 months (Fig. 2). When the training period was divided into 3 sections, the trainee's ADR improved a little to 49.4%, 52.6%, and 53.6%, respectively; however, there was no significant difference among the 3 sections (P=0.623). Furthermore, there was no significant difference in the trainees' PDR among the 3 stages; however, it did gradually increase to 62.3%, 67.7%, and 68.4%, respectively (P=0.290). APP showed a significant increase from 63.8% in the first 2 months to 73.2% in the second 2 months (P=0.004); however, there was no significant difference in APP between the second and the last 2 months (73.2% vs. 71.6%, P=0.612).

Fig. 2.

Changes in the trainees’ quality indicators in the screening/surveillance group. The experts’ quality indicators were the following: adenoma per polypectomy (APP), 72.5%; polyp detection rate (PDR), 65.0%; and adenoma detection rate (ADR), 53.4%.

The changes in ADR of the 3 trainees according to the 2-month training period are presented in Table 4. In an analysis of the screening/surveillance group, overall, there was a significant difference in ADR among the 3 trainees (P=0.008); however, in the last 2 months, there was no significant difference in ADR among the 3 trainees (P=0.606).

Table 4.

Changes in the ADR of the 3 trainees for the screening/surveillance group

DISCUSSION

This study revealed that the ADR of trainees participating in a colonoscopy training program could be higher than the recommended level, even in the early stages of training. In the present study, 3 trainees learned not only the basic techniques of colonoscopy but also the knowledge and skills of the experts during the colonoscopy training conducted by several colonoscopy experts. Through these training methods, the 3 trainees continuously showed a high ADR level of ≥40 from the beginning of the training. At the beginning, there was a significant variance in ADR among trainees. However, as the training continued, especially in the third section of the training period, the difference in ADR among the trainees decreased and converged to the experts' ADR level. These results suggest that the training education program was working properly.

Colonoscopy is recognized as the most important method for preventing CRC, and ADR is the index of colonoscopy most directly involved in CRC prevention. Therefore, many studies on improving ADR have been conducted [9–11]. So far, water assistance, elongation of withdrawal time, retroflexion in cecum, second look for proximal colon, and dynamic change in patient position have been suggested as ways to increase ADR [5, 12–14]. Improving ADR is an equally important issue for trainees new to colonoscopy. There have been some studies that have focused on ADR in trainees newly learning colonoscopy. Van Doorn et al. [15] compared the ADRs of 7 gastroenterology fellows. They excluded the initial 100 colonoscopies per trainee, and only colonoscopies of patients aged ≥18 years without underlying diseases were analyzed. There was a difference in ADR among the 7 fellows during the training period. Furthermore, only 2 of the 7 fellows had improved ADRs after training. Based on this result, it was argued that ADR should be improved by implementing feedback and benchmarking from the early stage of training. Gianotti et al. [16] analyzed average-risk colonoscopy performed by 10 gastroenterology fellows. They confirmed that 325 and 539 colonoscopy cases for men and women, respectively, were needed to reach the ideal ADR. They also showed that ADR improved through attending supervision. Based on these results, they argued that a large number of colonoscopies under attending supervision is necessary to increase the ADR of trainees. Consistent with previous studies, differences in ADR among trainees were also observed in the present study. As the training continued, the difference decreases, and eventually, the trainee's ADR reaches the expert level, which was somewhat consistent with previous studies.

In the present study, colonoscopies performed by trainees during the first 3 months were excluded. This was because colonoscopy insertion by trainees did not work well during this period; therefore, they focused on increasing the cecal intubation rate. This period was considered the stage of learning the handling skills necessary for withdrawal. Furthermore, as mentioned above, because the colonoscopy procedure was performed only under expert supervision, the ADR during this period was not considered the trainees’ ADR. Moreover, a screening/surveillance group for patients with a colonoscopy interval of ≥3 years was set up in the present study. This was because the number of screening colonoscopies with an interval of >5 years was only 242 cases between June and November 2022, which was a small number to analyze. Currently, a 3-year colonoscopy interval is recommended for surveillance after polypectomy in Korea [17, 18].

As part of colonoscopy training, the 3 trainees in the present study practiced diagnosing polyps macroscopically by checking the pathologic results of resected polyps. An indicator, APP, was introduced to judge the results of this training. In the screening/surveillance group, the APP of the 3 trainees showed a significant increase (P=0.007), and in the second and third 2 months of training, the results were similar to the experts' APP of 72.5. Therefore, the diagnostic ability of polyps on visual appearance in the endoscopic field of view is considered to be improved relatively quickly by training. Furthermore, APP can be used as one of the indicators of colonoscopy training.

The outcomes of the present study suggest some implementations. First, an ADR of ≥25, the recommended level, can be achievable from the early phase of training. All 3 trainees in this study showed an ADR of ≥40 from the first 2 months. This is thought to be due to the fact that they were tested with a withdrawal time of 9 minutes or more under the supervision of experts. Also, the attending supervision and feedback from the expert staff played a big role. The expert should give the trainee enough time to observe and then observe together. In addition, the expert should regularly monitor the trainee’s ADR and give feedback to the trainee and allow the trainee to learn various cases through seminars. Second, Trainee's ADR can be improved through training. Although there was a difference in ADR among trainees in the early phase of training, as the training continued, trainee's ADR improved overall, even to the expert's ADR level and the difference in ADR among trainees decreased.

It could be criticized that trainee 1's ADR decreased in the third 2 months, which is the biggest reason for the decrease in the difference in the ADR between trainees. Trainee 1 had already achieved an ADR above the expert's ADR in the first 2 months, but due to the relatively long examination time, it seems that trainee 1 was trying to maintain a high level of ADR while minimizing patient discomfort during the training period (Supplementary Table 3). The decrease in the difference in ADR between the trainees in the last 2 months can be seen as a result of their efforts to achieve good quality colonoscopy in their own way, as it is important to increase the ADR, but also to minimize patient discomfort for a good colonoscopy.

This study had some limitations. First, it is a retrospective study; therefore, it was difficult to accurately check the data, such as the history of abdominal surgery. Second, because the analysis was divided by periods rather than the number of performed colonoscopies, the changes in quality indicators could not be accurately reflected. This method was used, however, because the number of colonoscopy experiences before independently performing colonoscopies differed for each trainee. Third, because the trainees’ ADR in the present study is much higher than that in other existing studies on trainees, there may be questions about this study’s results. However, ADR may differ from country to country and institution to institution. Furthermore, because the experts’ ADR in this study was presented as a reference and the trainees’ ADR was confirmed to have reached the expert level, this study’s results are believed to be reliable. Moreover, there may be criticism that it is inappropriate to perform colonoscopies independently after just 3 months of training. The lack of data on quality indicators other than ADR, such as the cecal intubation rate, is also a limitation. Future research should include these factors.

However, in many studies on colonoscopy training, the trainees have <300 colonoscopy experience; however, the 3 trainees who participated in the present study performed >800 colonoscopies between March and November 2022. Therefore, because each trainee had accumulated a sufficient number of colonoscopy experiences, this study has an advantage in terms of learning curve analysis. In addition, the reliability of ADR was secured by analyzing both entire patients and the screening/surveillance group. This study also shows comparable reference values by presenting quality indicators of the colonoscopy experts together. Finally, by targeting colorectal surgical trainees, this study is distinguished from many existing studies targeting gastroenterology trainees.

In conclusion, the ADR of colonoscopy trainees can be higher than the recommended level, even in the early stages of training; however, there are variances among trainees. During the training period, ADR can be improved, even reaching the experts’ level, and the difference in ADR among trainees decreases. Therefore, efforts to increase ADR should be made actively from the beginning of training and continued during the training period. These efforts include things that can be done by expert staffs, such as attending supervision and implementing feedback, and things that can be done by the trainees themselves, such as benchmarking the skills of experts, elongating withdrawal time, and dynamic change in patient position.

SUPPLEMENTARY MATERIALS

Article information Continued