ASSOCIATION OF ALL INDICATION ADENOMA DETECTION RATE (ADR) WITH POST-COLONOSCOPY INCIDENT COLORECTAL CANCER (PCCRC)

Introduction
ADR is a colonoscopy quality metric based on the percentage of screening colonoscopies at which 1 or more adenomas are detected. Studies have shown that screening ADR is strongly inversely associated with patients’ risk of PCCRC incidence and death. However, the measurement and reporting of screening ADR is resource intensive often requiring manual record review, including for procedure indication; these issues challenge implementation and sustainability of ADR measurement. Recent studies have shown comparable ADRs for all colonoscopy examinations vs only screening examinations, with one reporting “overall indication” ADR may be associated with PCCRC risk. We aimed to determine whether all indication ADR is associated with risk for PCCRC in a national healthcare system.

Methods
We identified colonoscopies for all indications using CPT codes from 2016-2020 and attributed pathology to results posted within 30 days of the exam. We defined normal colonoscopy as procedures with no associated adenoma or CRC diagnosis within linked pathology reports. We linked these procedures to the cancer registry. We defined PCCRC as a cancer diagnosis occurring more than 6 months after a colonoscopy procedure but before a) a subsequent colonoscopy, b) patient date of death, or c) the end of available data (September 30, 2023). We calculated physician ADR at the time of each colonoscopy as the number of colonoscopies with at least one adenoma or adenocarcinoma detected in their previous 100 colonoscopies. We included only colonoscopies within the prior 2 years. We estimated differences in incident cancer by ADR using logistic regression and including physician specialty (gastroenterology vs. surgery), patient age and patient gender as covariates.

Results
We identified 508,601 normal colonoscopies undergone by 469,350 patients and associated with 1,100 physicians. Physician and patient characteristics, by colonoscopy procedures, are shown in the Table. There were 112 cases of PCCRC, for an incident cancer rate of 0.02%. All indication ADR was significantly associated with incident cancer, for each 5% percent increase in provider ADR, regardless of procedure indication, the odds of having a PCCRC decreased by 8.4%, (OR = 0.916, 95% CI 0.850 to 0.987, p = 0.021). The figure shows the predicted probability for PCCRC associated with physician ADR, for a male patient at the mean age of 47 who undergoes colonoscopy by a gastroenterologist.

Conclusion
All Indication ADR is strongly inversely associated with patients’ risk of incident colorectal cancer. Our findings support all indication ADR as a quality metric. Expansion of colonoscopy indications included in calculation of the ADR beyond screening indication could facilitate measurement and reporting of quality metrics, and shift resources to quality improvement efforts.