PDF
Citation
Print
Introduction
Colorectal cancer (CRC) is the second leading cause of cancer-related mortality worldwide and its incidence has increased rapidly in many Asian countries [1,2]. CRC is also a major public health concern in Korea. In 2018, it was the fourth most commonly diagnosed cancer, the third most prevalent cancer, and the third leading cause of cancer-related deaths [3].
Colonoscopy is considered the gold standard for CRC screening based on its ability to visualize the complete colon as well as directly remove neoplastic lesions. Accordingly, colonoscopy is widely used for primary CRC screening in several countries, including the United States, Canada, Germany, and Poland [4]. In 2004, CRC screening programs were introduced as part of the Korean National Cancer Screening Program (KNCSP) [5]. Therein, all Korean males and females over 50 years of age are invited to undergo annual fecal immunochemical testing (FIT). Individuals with a positive FIT result can subsequently undergo further investigation by colonoscopy or double contrast barium enema [5].
Although colonoscopy is widely considered a safe procedure for CRC, its effectiveness, especially in the proximal colon, seems to be limited [6]. Substantial incidence rates of interval CRCs (iCRCs) are observed following colonoscopy [7,8]. As iCRC has a deleterious impact on screening outcomes, the rate of iCRC is now accepted as a key performance indicator of organized CRC screening programs [7-9]. Although previous studies have reported the epidemiology, risk factors, and possible etiologies of iCRC, several key questions remain unsolved. Among these, there is limited data available concerning iCRCs that develop subsequent to diagnostic colonoscopy following a positive primary screening test.
Procedural volume is often regarded as a proxy for proficiency in medical and surgical fields [10]. Higher annual colonoscopy volumes are associated with a lower rate of colonic perforation. Moreover, there is a trend towards decreased overall potential adverse events with higher procedural volumes [11]. The procedural volumes of endoscopists are linked to the occurrence of iCRC following a negative colonoscopy, as evidenced by large population-based studies [10].
Therefore, we aimed to investigate the rate of iCRC as a quality metric subsequent to a negative follow-up colonoscopy following a positive FIT within a nationwide population-based CRC screening program. Further, we examined whether endoscopists with higher annual colonoscopy volumes have lower incidence of iCRC.
Materials and Methods
1. Study population and endoscopist volume
The study population comprised individuals aged 50 years and older who received colonoscopies after positive FITs from January 1, 2019 until December 31, 2020 in the KNCSP. The KNCSP database contains information on participant’ sex, age, insurance type, income, screening methods, and results. In addition, information on colonoscopy examinations include the endoscopist’s ID. During this period, a total of 240,170 individuals received colonoscopies. We excluded colonoscopy examinations performed by endoscopists who had not conducted a colonoscopy in the previous year. Finally, a total of 216,907 individuals were included in the study population (Fig. 1).
For each endoscopist, the annual colonoscopy volume was calculated as the total number of colonoscopy procedures in each year of the study. An endoscopist’s volume was allowed to vary from year to year, based on their actual procedural counts. The endoscopists were then categorized into different volume groups, with high-volume category serving as the reference. The annual endoscopist volume was divided into five groups (V1, 1-9; V2, 10-29; V3, 30-59; V4, 60-119; V5, ≥ 120). Due to the coronavirus disease 2019 pandemic causing a decrease in CRC screening in 2020 compared to 2019 [12], endoscopist volume groups were defined based on the number of procedures instead of percentiles, indicating the relative standing from the lowest to the highest values. The screening results in 2019 and 2020 were linked to the endoscopists’ annual volumes in 2018 and 2019, respectively.
2. Definition of iCRC
There is no universally accepted definition of iCRC. In colonoscopy studies, a cancer is classified as an interval cancer if it develops within a defined time frame after a screening colonoscopy (e.g., 3, 5, or 10 years), depending on the study [7,8]. iCRC is generally defined as CRC that was diagnosed between the time of a screening colonoscopy and the scheduled time of CRC screening [13,14]. In this study, the iCRC was calculated as the number of interval cancers diagnosed within one year (equal to the national CRC screening interval in Korea) divided by the number of negative test results [5,15].
Within the KNCSP for CRC, the results were judged as one of the following pre-defined statuses: “normal”, “colon polyp”, “suspected CRC”, “CRC”, or “other diagnoses”. We defined a negative colonoscopy as “normal”, “colon polyp”, or “other diagnoses” [16]. Following a negative colonoscopy, CRC was identified based on the International Classification of Disease, 10th revision codes C18-20 within one year after screening [15], based on the co-payment reduction program data. The Korean government implemented the co-payment reduction program to control catastrophic medical costs and the occurrence of unmet needs due to high health expenditure for all inpatient and outpatient services for those diagnosed with cancer, cerebrovascular disease, heart disease, rare diseases, severe incurable diseases, and severe burns [17]. Thus, individuals diagnosed with a new cancer in Korea can benefit from reduced out-of-pocket payments for medical services through the National Health Insurance program.
3. Statistical analysis
First, we described the sociodemographic characteristics of the individuals who underwent colonoscopies and had negative colonoscopy results. Second, the number of endoscopists and colonoscopy examinations according to the previous year’s annual volume, the screening year, and the type of medical institution were calculated. In addition, the average rate of iCRC per endoscopist was calculated for each variable and assessed using ANOVA or t tests. The incidence of iCRC was calculated as the proportion of iCRC cases among all negative colonoscopies performed by each endoscopist. The rate of iCRC was expressed as the number of cancers per 1,000 colonoscopy examinations. Third, a robust Poisson regression model was used to confirm the association between the rate of iCRC and endoscopist volumes, adjusted by the type of medical institution and screening year. Robust Poisson models are less susceptible to the influence of outliers when estimating the relative risks (RR) [18]. Finally, subgroup analysis was conducted to further elucidate the association between iCRC rates and endoscopist volumes, according to the type of medical institution and the screening year. We also conducted the sensitive analysis by defining quintiles of endoscopists’ procedural volumes. The RR and their respective 95% confidence intervals (CIs) were calculated. All statistical analyses were performed using SAS ver. 9.4 (SAS Institute Inc., Cary, NC).
Results
1. Sociodemographic characteristics of participants who underwent colonoscopy
Among the 216,907 participants who underwent colonoscopy examinations, a total of 207,480 (95.7%) had negative colonoscopy results (Table 1). Men accounted for 52.5% of the participants, of whom 94.7% had negative test results. In comparison, 96.7% of the women had negative colonoscopy results. Participants aged 70 years or older accounted for 26.9% of the study population, of whom, 93.9% had negative results.
2. Rate of iCRC by endoscopist volume, type of medical institution, and screening year
A total of 10,412 endoscopists performed 216,907 colonoscopies from January 1, 2019, until December 31, 2020, in the KNCSP (Table 2). Endoscopists in the V1 group performed < 10 examinations (46.1% of endoscopists), accounting for 18.6% of the total colonoscopy examinations. Conversely, the V5 group, who performed 120 or more colonoscopies (2.2% of endoscopists), represented 14.8% of the examinations. Among all the endoscopists, 39.5% performed colonoscopy examinations at hospitals, representing 43.5% of the total examinations.
Overall, the average rate of iCRC per endoscopist was 8.46 per 1,000 examinations. The rate of iCRC was 9.24 per 1,000 examinations in the low-volume V1 group, and 9.00‰, 6.56‰, 5.00‰, and 4.27‰ in the V2, V3, V4, and high-volume V5 groups, respectively. We also calculated the iCRC rate for each endoscopist based on the endoscopist’s procedure volume as well as the sociodemographic characteristics of the participants who received colonoscopies from that specific endoscopist (S1 Table).
3. Risk of iCRC according to endoscopist’s procedure volume
Compared to the group of endoscopists with the highest procedure volume (V5 group), the iCRC rate was 2.21 times higher in the V1 group. Similar trends were observed in the other groups (V2: RR, 2.15; 95% CI, 1.57 to 2.94; V3: RR, 1.56; 95% CI, 1.15 to 2.13; V4: RR, 1.18; 95% CI, 0.83 to 1.67) (Table 3). The association between the type of medical institution, the screening year, and the risk of iCRC was not statistically significant.
Upon comparing iCRC rates by endoscopist volume, while considering the type of medical institution (Fig. 2), we confirmed the association between iCRC rates and endoscopist volume. In hospitals, the risk of iCRC was 2.15 times higher in the V1 group compared to that in the V5 group. In clinics, the risk in the V1 group increased by 3.91 times compared to that in the V5 group. In 2020, the risk of iCRC in the V1 group increased by 2.04 times compared to that in the V5 group. Finally, in 2019, this risk in the V1 group increased by 2.42 times compared to that in the V5 group.
The results of the subgroup analysis, considering participant’s sociodemographic characteristics, demonstrated an association between iCRC rates and endoscopist volumes (S2 Fig.).
Discussion
Our study demonstrates some important findings regarding the association between colonoscopy procedural volumes and the rate of iCRC, which serve as key quality indicators for colonoscopy effectiveness. Across all endoscopists in the study, the mean iCRC rate per endoscopist was 8.46 per 1,000 examinations. The mean rate of iCRC was higher in males than in females and in older adults than in younger adults [19,20]. We found a substantial variation in the rate of iCRC across endoscopists. Endoscopists with the highest procedural volumes had the lowest rates of interval cancers following colonoscopy, while lower-volume endoscopists exhibited higher iCRC rates.
When patients underwent colonoscopy between 6 months and 36 months before CRC was diagnosed, they were considered to have a new or missed cancer, with the rate ranging from 0.4% to 7.9% [21]. U.S. population–based analysis to include validation of cancer diagnoses through registry data, we documented an interval cancer frequency of 7.2% [22]. Recent studies reported an increase of up to 8% in the prevalence of interval cancers [22,23].
In Korea, the prevalence of interval cancer was reported to be 6.2% [24]; however, this figure was based on data obtained by telephone calls from a single tertiary-referral center. Another Korean study reported that the prevalence rate of iCRC was 7.9% among all diagnosed cases of CRC, specifically focusing on early iCRC occurrences within 6-36 months after index colonoscopy [16]. Modifying the time cutoff to include iCRC affects the overall proportion [11,13,25]. For our study, we defined colonoscopy iCRCs as those occurring within 1 year (i.e., a complete FIT screening cycle) after index colonoscopy and before the date of the next recommended examination [8,14]. In our study, the mean rate of iCRC was calculated per endoscopist, and unit of analysis was the endoscopist. Therefore, the rates of interval cancer need to be compared with caution.
Although the exact etiology of interval cancer remains unknown, factors implicated in the development of interval cancers include missed lesions at the time of colonoscopy, incomplete resection of previous neoplastic lesions, distinct tumor biology, and the serrated pathway of carcinogenesis. However, recent evidence suggests that the occurrence of interval cancers is related to the training of the endoscopist and quality of the colonoscopy rather than to the tumor biology [24].
Our findings suggest that endoscopists procedural volumes were likely to play an important role in optimizing performance. Decreasing the volume of surgical procedures requiring advanced technical skills can result in poorer results both on an institutional and individual basis [11,26,27]. Various studies have convincingly shown that the risk of post-endoscopy cancer depends on the experience of the endoscopist [4]. Low-volume facilities may not have management supervision [28]. Colonoscopies performed by low-volume endoscopists have been reported to be associated with an increased risk of bleeding, perforation, and cecal intubation rates [11].
As a result, several national and regional gastroenterology guideline committees have recommended minimum annual colonoscopy volume benchmarks [29-31]. Although some data have suggested that low-volume endoscopists have higher rates of colonoscopy-related adverse events [23], it remains unclear whether this association also exists for interval cancer. Although our analysis could not establish a specific lower threshold, the present study suggests that very low-volume endoscopists may be most at risk.
A strong emphasis has been placed on the importance of organized population screening with constant monitoring of colonoscopy performance [4]. In 2008, the Ministry of Health and Welfare launched the National Screening Quality Assessment Program in the KNCSP. This external evaluation assesses colonoscopy accuracy indicators every 3 years, such as sensitivity, specificity, adenoma detection rate, and interval cancer. However, it does not provide feedback. In addition, institutions with fewer than 50 screening cases are excluded from the evaluation of accuracy indicators due to challenges in generating statistically reliable indicators when considering the incidence of colon cancer among a small number of colonoscopies.
However, regular audits and feedback should be considered to improve the overall quality and decrease quality variations at the national level, regardless of the number of examinations performed [13,32]. In addition, it should be considered that high-volume endoscopists may need to decrease their per-procedure time because of denser endoscopy schedules, which could potentially compromise quality.
Our study had some limitations. First, endoscopist characteristics are related to achieving appropriate levels of colonoscopy competence. However, we did not consider these variables due to a lack of data availability. Second, defining colonoscopy performance is difficult, and case volume is not the only factor involved [33]. The performance of colonoscopy is associated with various factors, such as the endoscope system (e.g., high-magnification scopes), bowel preparation quality, withdrawal time, cecal intubation, and endoscopy room quality. Third, we need to comprehensively investigate other performance indicators such as the positivity rate, sensitivity, specificity, positive predictive value, and false-positive rate in future studies. Fourth, follow-up colonoscopy should ideally be performed for all screened participants to evaluate the incidence of iCRC. However, 64.4% of the eligible population participated in the 2020 Korean CRC screening program [3]. Fifth, our study does not reflect colonoscopy in opportunistic cancer screening. Finally, a positive FIT prior to a negative colonoscopy could indicate that a neoplasm had already been present but remained undetected during the colonoscopy [8].
Despite these limitations, this research has several important implications for clinical practice and policy. First, we demonstrated an association between endoscopist volume and interval cancers from a nationwide population-based CRC screening program. Second, the findings emphasize that low-volume endoscopists have higher risks of interval cancer. If the impact of colonoscopy on cancer prevention is to be maximized, this issue should be addressed by monitoring endoscopist volume and variations in performance. Quality metrics should be targeted for all providers, regardless of minimum requirements for endoscopist volume.
XML Download