A multicenter pragmatic study of an evidence-based intervention to improve adenoma detection: the Quality Improvement...

Low adenoma detection rates (ADRs) at colonoscopy are linked to significantly higher interval cancer rates, and vary between colonoscopists. Studies demonstrate that lesion detection is improved by: withdrawal time of ≥ 6 minutes; use of hyoscine butylbromide; position change; and rectal retroflexion. We evaluated the feasibility of implementing the above "bundle" of interventions into colonoscopy practice, and the effect on ADR.

A longitudinal cohort design was used. Implementation combined central training, local promotion, and feedback. The uptake marker was change in hyoscine butylbromide use. Comparisons were between the 3 months before and the 9 months after the implementation phase, globally, by endoscopy unit and by quartile when colonoscopists were ranked according to baseline ADR. Chi-squared or Fisher's tests were used to evaluate significance.

12 units participated. Global and quartile analyses included data from 118 and 68 colonoscopists and 17 508 and 14 193 procedures respectively. A significant increase in hyoscine butylbromide use was observed globally (54.4 % vs. 15.8 %, P < 0.001), in all endoscopy units (P < 0.001) and quartiles (P < 0.001). A significant increase in ADR was observed globally (18.1 % vs. 16.0 %, P = 0.002) and in the lower two colonoscopist quartiles (P < 0.001), with a nonsignificant increase in the upper middle quartile and a significant fall to 21.5 %. in the upper quartile. The significant variations in ADR among the upper three quartiles disappeared.

In routine clinical practice, introduction of a simple, inexpensive, evidence-based "bundle" of measures is feasible and is associated with higher global ADR, driven by improvements amongst the poorest performing colonoscopists.