Comparison of Quality Measures for Detection of Neoplasia at Screening Colonoscopy.

BACKGROUND & AIMS:The proportion of colonoscopies with at least one adenoma (adenoma detection rate [ADR]) is inversely associated with colorectal cancer (CRC) risk and death. The aim of this study was to examine whether such associations exist for colonoscopy quality measures other than ADR. METHODS:We used data from the Polish Colorectal Cancer Screening Program collected in 2000-2011. For all endoscopists who performed ≥100 colonoscopies we calculated detection rates of adenomas (ADR), polyps (PDR), and advanced adenomas (≥10 mm/villous component/high-grade dysplasia [AADR]); and number of adenomas per colonoscopy (APC) and per colonoscopy with ≥1 adenoma (APPC). We followed patients until CRC diagnosed before recommended surveillance, death, or December 31, 2019. We estimated hazard ratios (HRs) and 95% confidence intervals (CIs) using Cox proportional-hazard models. We used Harrell's C statistic to compare the predictive power of the quality measures. RESULTS:Data on 173,287 patients (median age, 56 years; 37.8% male) and 262 endoscopists were used. During a median follow-up of 10 years and 1,490,683 person-years, we identified 395 CRCs. All quality measures were significantly associated with CRC risk and death. The relative reductions in CRC risk were as follows: for ADR ≥24.9% (reference <12.1%; HR, 0.41; 95% CI, 0.25-0.66), PDR ≥42.7% (reference <19.9%; HR, 0.35; 95% CI, 0.24-0.51), AADR ≥9.1% (reference <4.1%; HR, 0.69; 95% CI, 0.49-0.96), APC ≥0.37 (reference <0.15; HR, 0.35; 95% CI, 0.21-0.58), and APPC ≥1.54 (reference <1.19; HR, 0.54; 95% CI, 0.35-0.83). AADR was the only quality measure with significantly lower predictive power than ADR (Harrell's C, 59.7 vs 63.4; P = .001). Similar relative reductions were observed for CRC death. CONCLUSIONS:This large observational study confirmed the inverse association between ADR and CRC risk and death. The PDR and APC quality measures appear to be comparable with ADR.