285 Adherence to Recommending 10-Year Intervals After Normal Screening Colonoscopy in Average-Risk Individuals: A Snapshot of 2017 for Phase 1 of the Michigan CRC Screening Quality Improvement Project

Abstract

INTRODUCTION: The Michigan CRC Screening Quality Improvement Project (CRC-QIP) is a multi-center, 4-phase project to improve adherence with recommending 10-year intervals after normal screening colonoscopies. Per guidelines, target is 90% adherence with recommended intervals. METHODS: Inclusion criteria: In order to minimize possible confounders, patients were limited to: (a) average-risk, 50–75 year olds; (b) colonoscopy performed in 2017; (c) sole indication-CRC screening; (d) no biopsy, polypectomy, or any abnormal findings on procedure report. Study Setting: In order to minimize confounders when stratifying for specialty and type of practice, only medical centers with (a) GI fellowship program; (b) Hospital-based “open” endoscopy unit (i.e., utilized by private practice gastroenterologists (GIs), colorectal surgeons, and general surgeons). Primary Outcome: Adherence to guideline-consistent intervals defined as repeat colonoscopy in 10 years or discontinue CRC screening due to patient's age when bowel preparation is adequate or <1 year if bowel preparation is inadequate. Primary analysis is frequency of adherence based on endoscopist specialty after adjusting for multiple confounders, including procedure-related factors (e.g., withdrawal time), and endoscopist-related factors (e.g., yrs in practice, solo vs group practice, colonoscopies performed per yr). Hierarchical logistic regression model was used with adherence as the dependent variable with random intercepts for endoscopist performing procedure to account for correlation between endoscopist and site as fixed effects. RESULTS: Among 1,694 eligible patients, mean age was 58.7 ± 6.5, 43% male, and 68% African-American. Adherence was better for academic gastroenterologists with or without GI fellows vs private practice GIs or academic general surgeons (P = 0.04). The latter two groups were adherent in <40% of patients (Table 1). Adherence was significantly better with good/excellent bowel preps compared to all other bowel prep categories (P < 0.05), and patients with poor, fair, or no documentation of prep were adherent <50% (Table 2). CONCLUSION: In this project, adherence did not meet guideline-specified target of 90% among private practice GIs and general surgeons. Phase 2 will include additional analysis to identify factors associated with non-adherence and a mixed methods qualitative study to identify appropriate interventions to improve performance.