P16 Acceptability of key performance indicators (KPI) in the national endoscopy database (NED), a delphi process

Abstract

Introduction Automated Performance Reports to Improve Quality Outcomes Trial (APRIQOT) uses NED to provide endoscopists feedback on colonic detection KPI. Traditional adenoma detection rate is dependent on unavailable histological data. Our aim was to gain expert consensus on which available KPI are acceptable to endoscopists. Method A Delphi panel of UK expert endoscopists was recruited online, purposively for professional background. Panellists interacted using an online form. In round one we provided a summary and acceptability statement for each KPI, participants rated agreement with a 5 point Likert scale and free-text comments. Responses were analysed anonymously. In subsequent rounds participants reviewed all graded consensus statements and comments. Statements were accepted with ≥80% consensus or redrafted. Rounds ran January to April 2019. Results We recruited 21 UK expert endoscopists. 12 were female, 48% gastroenterology background, 29% nursing, 14% surgical and 9% trainees. All statements reached consensus by round 3 (Table 1). The panel agreed KPI adjusted for age, sex and indication were ‘more acceptable’. Polyp measure had risks of ‘gaming’ and distal hyperplastic polyp over reporting, but encompass significant non-adenomatous polyps. Mean number of polyps (MNP) reached consensus after discussing reduction of the ‘one and done’ phenomenon and using a cap of 5 polyps/colon to mitigate skew from polyposis. Proximal polypectomy rate (PPR) was accepted as a secondary ‘tool to improve right sided detection’ and reduce ‘gaming’, despite concerns around contraindications to polypectomy. Conclusion All adjusted KPI were accepted, MNP was selected for trial with robust data to model case-mix.