Adenoma detection rates during screening colonoscopy performed by interventional gastroenterologists and general gastroenterologists.

Introduction
Colonoscopy with polypectomy is critical for reducing the incidence and mortality from colorectal cancer (CRC) as the primary screening test or after an abnormal fecal immunochemical or stool DNA test result.1,2 There are many proposed colonoscopy quality indicators, with a few having strong data supporting them as key measures to identify a high-quality screening colonoscopy. Since we conducted our study, guidelines were updated to ≥35% adenoma detection rate (ADR) and at least an 8-minute withdrawal time.3 The ADR is the detection of at least 1 adenoma in a number of screening colonoscopies.4 It is considered the most important quality indicator because it is objective and measurable, reproducible, and directly proportional to the risk of developing or dying from interval or subsequent CRC.5
Colonoscopies performed by gastroenterologists compared with other specialties have been associated with improved quality metrics including ADR and reduction of the risk for incident CRC after a colonoscopy, but the results are heterogeneous and difficult to assign to the specialty alone.6,7
Colonoscopy credentialing standards are recommended to ensure delivery of high-quality care and assure patients and institutions that the individual has achieved a minimal level of expertise to perform a procedure effectively.8 The number of procedures needed to achieve colonoscopy proficiency is not clear. The American Society for Gastrointestinal Endoscopy recommends at least 275 completed colonoscopies for credentialing, but competency assessment based on quality metrics may be better at identifying high-quality procedural proficiency.8, 9, 10 The number of procedures to achieve expertise remains unknown, but at least 1 study by Rabeneck et al6 found that ≥300 colonoscopies per year were best to meet high-quality procedure indicators.
The ADR among gastroenterologists who subspecialize in interventional endoscopy is not known. Interventional gastroenterologists dedicate most of their procedural time to their areas of expertise, thus performing fewer screening colonoscopies than general gastroenterologists. In this study, we aimed to compare the ADR between interventional gastroenterologists and general gastroenterologists to determine differences in ADRs and implications for patient outcomes.

Methods
An Institutional Review Board–approved retrospective analysis of screening colonoscopies was conducted at The Ohio State University Wexner Medical Center from January 2020 to January 2023. De-identified information was obtained by querying the electronic procedure database (GI Quality Improvement Consortium, North Bethesda, Md, USA). All endoscopies were performed at the Ohio State University Wexner Medical Center Endoscopy facility. Colonoscopies were performed with either standard or pediatric Olympus 190-series colonoscopes (Olympus America, Center Valley, Pa, USA). Various imaging modalities and cap assistance were not documented and used at the discretion of the endoscopist. All included endoscopists were credentialed to perform independent colonoscopies, were board certified in gastroenterology, had performed >500 colonoscopies, and were at least 2 years after training. Interventional gastroenterologists were defined as endoscopists who perform therapeutic procedures, including endoscopic retrograde cholangiopancreatography and endoscopic ultrasound.
Inclusion criteria were patients aged between 45 and 75 years with an indication for colon neoplasia screening. Colonoscopies for diagnostic indications were excluded. Patients with a Boston Bowel Preparation Scale score <6, inflammatory bowel disease, previous colon resection, a history of colon polyposis syndrome, and undergoing surveillance colonoscopy were excluded.
Patient characteristics and quality measures were summarized and compared for interventional gastroenterologists and general gastroenterologists using t tests and χ2 tests as appropriate. Multivariable regression was used to assess the difference in ADR between interventional gastroenterologists and general gastroenterologists while controlling for patient age, sex, race, previous colonoscopy, family history of CRC, fellow involvement, and sedation type.

Results
A total of 9755 screening colonoscopy procedures were included in the analysis. Of the study participants, 54.8% were female with a mean (standard deviation [SD]) age of 56.2 (7.7) years. There were 407 screening procedures performed by 15 interventional gastroenterologists, and 9348 screening colonoscopies were performed by 41 general gastroenterologists, resulting in a mean (SD) of 27 (53) and 228 (216) colonoscopies per endoscopist per year for interventional gastroenterologists and general gastroenterologists, respectively (P < .001) (Table 1).
Patients who had a screening colonoscopy by an interventional gastroenterologist were more likely to be male than those screened by a general gastroenterologist (P < .001) and were slightly older, with an average age of 57.5 years for interventional gastroenterologist-screened patients versus 56.2 years for general gastroenterologists (P < .001). There were no racial disparities between the groups. A family history of CRC was similar between the groups. More patients in the general gastroenterologist cohort had a previous colonoscopy (8.1% vs 3.4% in the interventional gastroenterologist cohort, P < .001). Most patients received sedation provided by anesthesia, with a smaller number having moderate sedation, with a similar distribution between the groups. In procedures without intervention, withdrawal time was ≥6 minutes in 95.6% of the colonoscopies in the interventional gastroenterologist cohort and 98.2% in the general gastroenterologist group (χ2 = 13.56; P < .001). Although our study was performed under previous standards, we were able to assess withdrawal times ≥8 minutes, which was 87.7% of the colonoscopies in the interventional gastroenterologist cohort and 81.1% in the general gastroenterologist group (P < .001) (Table 2).
On univariate analysis, ADR was higher among interventional gastroenterologists compared with general gastroenterologists (46.9% vs 41.5%, P = .033) (Table 2). The unadjusted odds ratios are detailed in Supplementary Table 1, available online at www.igiejournal.org. Sex-stratified univariate analysis showed ADR among males was 57.0% for interventional gastroenterologists versus 49.0% for general gastroenterologists (P = .024), whereas ADR among females was similar between the groups (34.8% vs 35.4%, P = .932). However, in multivariate logistic regression controlling for patient demographic and clinical characteristics, provider type was not associated with adenoma detection (odds ratio = 0.99; 95% confidence interval, 0.80-1.22; P = .919) (Table 3).

Discussion
Studies indicate that colonoscopies performed by gastroenterologists meet high-quality measures at an improved rate and diminish the risk of postcolonoscopy incident CRC compared with other specialties.6 Variability in quality measures, and especially the ADR between gastroenterologists who perform interventional or advanced endoscopic procedures and those who perform more standard procedures, has not been previously reported, to our knowledge. In this study, we compared colonoscopy quality outcomes from interventional gastroenterologists and general gastroenterologists. Despite the low volume of screening colonoscopies for interventional gastroenterologists, the ADR, arguably the most reliable measure of future CRC incidence and mortality, was >40% and similar between the groups. This is reassuring because it shows that despite a lower volume of screening colonoscopies, interventional gastroenterologists at our center maintained high-quality screening colonoscopy. Interestingly, neither of the 2 groups achieved the mean frequency of 300 colonoscopies per year, which has been reported to be protective of incident CRC.6 It is possible that in gastroenterology training, where more than 1000 colonoscopies per trainee before graduation is common in the United States, and the experience of the gastroenterologists may have allowed the ADR to remain high despite a lower frequency of screening colonoscopies. Colonoscope withdrawal time was higher for general gastroenterologists, but it was optimal, >95% for both groups, supporting that both groups performed high-quality procedures.
There are study limitations that should be considered. The study was retrospective, and the procedures were performed in a single-center academic institution by gastroenterologists. Results may not be applicable in a nonacademic setting, or when performed by endoscopists who are not gastroenterology trained. Longitudinal outcomes (eg, interval CRC incidence) and external validation were also not pursued or determined at this time. Although the study was conducted during a period when the accepted withdrawal-time benchmark was ≥6 minutes, it is important to note that updated guidelines now recommend a minimum of ≥8 minutes3; acknowledging this distinction ensures appropriate historical context for interpreting our results and prevents misinterpretation of withdrawal times as suboptimal by current standards.
A multicenter study on colonoscopy quality measures and outcomes within the subspecialties in gastroenterology would be helpful to identify gaps in knowledge and possible interventions to enhance outcomes and continue to reduce the incidence of CRC.
In conclusion, this study showed similar ADR between interventional gastroenterologists and general gastroenterologists after controlling for patient characteristics. The results of this study are reassuring and support interventional gastroenterologists who may perform a lower volume of screening colonoscopies continuing to be included in CRC screening practices.

PATIENT CONSENT
Patient consent was waived via our Institutional Review Board, as data were collected from a deidentified database.

Disclosure
The following authors disclosed financial relationships: P. P. Stanich: Grant/research support from Freenome, Guardant, Janssen Pharma, and Pfizer Inc. G. Papachristou: Grant/research support from AbbVie; speaker for Nestlé; consultant for Olympus. S. Krishna: Grant/research support from Mauna Kea Technologies, Taewoong Medical USA, and US Biotest. L. Lara: Consultant and speaker for AbbVie Inc. All other authors disclosed no financial relationships.