Nicole G Jawitz1, Ziad F Gellad1,2, Li Lin3, Richard K Wood1, David A Leiman4,5. 1. Division of Gastroenterology, Duke University School of Medicine, 200 Morris Street, Suite 6524, Durham, NC, 27701, USA. 2. Durham VA Medical Center, Durham, NC, USA. 3. Department of Population Health Sciences, Durham, NC, USA. 4. Division of Gastroenterology, Duke University School of Medicine, 200 Morris Street, Suite 6524, Durham, NC, 27701, USA. david.leiman@duke.edu. 5. Duke Clinical Research Institute, Durham, NC, USA. david.leiman@duke.edu.
Abstract
BACKGROUND: Variability in colon polyp detection impacts patient outcomes. However, the relative influence of physician, patient, and procedure-specific factors on polyp detection is unclear. Therefore, determining how these factors contribute to adenoma and sessile serrated polyp (SSP) detection is important to contextualize measures of colonoscopy quality such as adenoma detection rate and patient outcomes. AIMS: To determine the relative contribution of physician, patient, and procedure-specific factors in total polyp, adenoma, and SSP detection rates. METHODS: We performed a retrospective study of patients undergoing screening colonoscopy and used a two-level generalized linear mixed regression model to identify factors associated with polyp detection. RESULTS: 7799 average risk screening colonoscopies were performed between July 2016 and October 2017. The patient factor most strongly associated with increased risk of adenoma and sessile serrated polyp detection was white race (OR 1.21, 95% CI 1.05-1.39 and OR 3.17, 95% CI 2.34-4.30, respectively). Adenomatous (OR 1.92, 95% CI 1.04-3.57) and sessile serrated polyps (OR 5.56, 95% CI 1.37-20.0) were more likely to be found during procedures performed with anesthesia care as compared to those with moderate sedation. Physician with a luminal gastrointestinal focus had increased odds of adenoma detection (OR 1.61, 95% CI 1.02-2.50). CONCLUSIONS: In a multi-level model accounting for clustering effects, we identified patient, provider and procedural factors independently influence adenoma and sessile serrated polyp detection. Our findings suggest that to compare polyp detection rates between endoscopists, even at the same institution, risk adjustment by characteristics of the patient population and practice is necessary.
BACKGROUND: Variability in colon polyp detection impacts patient outcomes. However, the relative influence of physician, patient, and procedure-specific factors on polyp detection is unclear. Therefore, determining how these factors contribute to adenoma and sessile serrated polyp (SSP) detection is important to contextualize measures of colonoscopy quality such as adenoma detection rate and patient outcomes. AIMS: To determine the relative contribution of physician, patient, and procedure-specific factors in total polyp, adenoma, and SSP detection rates. METHODS: We performed a retrospective study of patients undergoing screening colonoscopy and used a two-level generalized linear mixed regression model to identify factors associated with polyp detection. RESULTS: 7799 average risk screening colonoscopies were performed between July 2016 and October 2017. The patient factor most strongly associated with increased risk of adenoma and sessile serrated polyp detection was white race (OR 1.21, 95% CI 1.05-1.39 and OR 3.17, 95% CI 2.34-4.30, respectively). Adenomatous (OR 1.92, 95% CI 1.04-3.57) and sessile serrated polyps (OR 5.56, 95% CI 1.37-20.0) were more likely to be found during procedures performed with anesthesia care as compared to those with moderate sedation. Physician with a luminal gastrointestinal focus had increased odds of adenoma detection (OR 1.61, 95% CI 1.02-2.50). CONCLUSIONS: In a multi-level model accounting for clustering effects, we identified patient, provider and procedural factors independently influence adenoma and sessile serrated polyp detection. Our findings suggest that to compare polyp detection rates between endoscopists, even at the same institution, risk adjustment by characteristics of the patient population and practice is necessary.
Authors: Jennifer K Maratt; Joseph Dickens; Philip S Schoenfeld; Grace H Elta; Kenya Jackson; Daniel Rizk; Christine Erickson; Stacy B Menees Journal: Dig Dis Sci Date: 2017-10-17 Impact factor: 3.199