Quality of Colonoscopy in Routine Clinical Practice

Abstract

Purpose: The aim of this study was to perform a population-based assessment of colonoscopy quality and to examine factors influencing quality. Methods: We used a nationally representative 20% sample of Medicare Carrier claims from 2003. Physician specialty was determined by linking to the AMA Physician Masterfile. We examined the indications for colonoscopy, completion rates, rates of polyp/tumor detection and biopsy/polypectomy, incidence of ER visits or hospitalizations within 30 days, and incidence of follow-up colon exams within 1 year. Results: 389,276 outpatient colonoscopy exams were identified in 2003. Race/ethnicity and age were the patient characteristics with the greatest influence on colonoscopy quality, while colonoscopy volume and specialty were the most influential provider characteristics. Hispanics and American Indians were significantly less likely to have a screening or surveillance exam indication than other racial/ethnic groups. Polypectomy rates were highest in Asians, and lowest in Hispanics. Reporting of incomplete exams was highest in African-Americans and Hispanics. Incomplete colonoscopy reporting was similar between provider specialties, but was inversely associated with annual colonoscopy volume. Gastroenterologists had higher polyp detection and polypectomy rates compared to other specialties (44.8% vs. 34.9–42.7% for polyp detection, P < 0.001; 27.1% vs. 17.7–23.1% for polypectomy, P < 0.001). Five percent of patients had an emergency room visit and 6% of patients were hospitalized within 30 days of colonoscopy; the frequency of both events increased with patient age. Risk of hospitalization within 30 days was lower for exams performed by gastroenterologists (5.5%) or colorectal surgeons (5.6%) compared to exams by family medicine physicians (7.2%), general surgeons (7.9%), or internal medicine specialists (6.0%) (P < 0.001). Colon perforation occurred in 0.1%, and lower gastrointestinal bleeding in 1.5%. Gastroenterologists had slightly lower rates of repeat colonoscopy within 1 year vs. primary care providers or general surgeons (4.6% vs. 5.3–6.2%, P < 0.001). Conclusion: Patient age is associated with significantly higher rates of incomplete colonoscopy and of subsequent emergency room visits or hospitalizations, suggesting that the risks and benefits of colonoscopy need to be considered carefully in older patients. Provider specialty is significantly associated with polyp detection and polypectomy rates, risk of hospitalization within 30 days, and use of follow-up colonoscopy. These results have strong implications for appropriate training and credentialing of practitioners providing colonoscopy services.