Incidence of interval colorectal cancer attributable to an endoscopist in clinical practice

Abstract

Background and aims: Endoscopists who encounter an interval colorectal cancer (I-CRC) may be concerned about the implications because I-CRCs may represent a lapse in colonoscopy quality and a missed opportunity for prevention. We wanted to determine the I-CRC rate per colonoscopy examination and to examine the effect of colonoscopy volume and adenoma detection rate (ADR) on the number of I-CRCs attributable to an endoscopist.

Methods: We determined the rate of I-CRC diagnosis per outpatient colonoscopy examination by measuring the incidence of CRC diagnosis in practice and by assessing, via literature review, the percentage of cancers that are interval. We also estimated the number of attributable I-CRCs as a function of ADR and colonoscopy volume.

Results: Among 93,562 colonoscopies performed in 2013 to 2015 by 120 physicians in 4 diverse U.S. medical centers, 526 CRCs were diagnosed (.6%). Of 149,556 CRCs in the published literature, 7958 were I-CRCs (5.25% ± .94%). With rates of .6% (CRC per colonoscopy) and 5.25% (I-CRC per CRC), the rate of I-CRC is 1 per 3174 colonoscopies (95% confidence interval, 1 per 2710 to 1 per 3875). An endoscopist at the median of outpatient colonoscopy volume (316/year) in the lowest ADR quintile of detection (7%-19%) would have an I-CRC attributed every 8.2 years, or 4.2 I-CRCs in a 35-year career, versus every 16.7 years, or 2.0 I-CRCs in a 35-year career, for an endoscopist in the highest ADR quintile (33%-52%).

Conclusions: An average-volume endoscopist will have 2 to 4 attributable I-CRCs in a 35-year career, but the frequency will vary depending on colonoscopy volume and ADR.