Implementing Methods to Capture ERCP Intraprocedural Quality Indicators in a Health Care System

Abstract

Introduction: A GI multisociety taskforce published quality indicators for ERCP practice in 2015, which included a subset of 5, high-priority indicators. Three of these consist of intraprocedural measures with recommended performance targets >90% in patients with normal biliary anatomy, and include: (1) rate of native papillae biliary cannulation; (2) success of extracting common duct stones < 1cm, and (3) success of stent placement for distal biliary obstruction. Methods to easily capture these measures for all ERCPs to enable outcomes analyses would benifit endoscopists, endoscopy units, and health systems. Methods: We developed a process to capture procedural quality measures for all ERCPs. Three hospitals share the same endoscopy reporting software (ProVationMD, ProVation Medical, Minneapolis, MN) and perform the majority of ERCPs in our system. We created “hard stops” in the reporting software that require the endoscopist to record an outcome of either success, failure, or not applicable (N/A) for each measure before the report could be finalized. Data was exported to analyze procedural quality measures, and to compare outcomes by: (a) annual hospital volume- low (<100) vs high (>200), and (b) annual endoscopist volume- low (<50) vs high (>200). Results: A total of 1484, 48, and 389 ERCPs were performed at 3 hospitals over 21, 16, and 16 month periods by 12 endoscopists (5 high volume). Documentation of success, failure, or N/A for each quality indicator was captured in all ERCPs (100%). Overall deep cannulation success was 96%, with higher success for high volume endosopists (98% vs 88%; p<0.001), and at high volume centers (97% vs 87%; p=0.006). Overall success for extracting stones <1 cm was 97%, with higher success for high volume endoscopists (98% vs 91%; p=0.002), and at high volume centers (98% vs 84%; p=0.001). Overall success for stenting strictures below the bifurcation was 98%, with higher success with high volume endoscopists (99% vs 81%; p=0.001), and at high volume centers 99% vs 60%; p=0.002). Cannulation and stent placement metrics were below recommended targets for low volume endoscopists. All three, ERCP procedural measures were lower than recommended targets at low volume centers. Conclusion: Endoscopy documentation software can be formatted to prospectively capture procedural quality indicators for ERCP practice. Implementing such methods allows for 100% collection of these measures, and can be used to analyze ERCP quality in a health system.