Authors
S. OUAZZANI (1), A. LEMMERS (1), F. MARTINEZ (2), M. DELHAYE (1), M. ARVANITAKIS (1), O. LE MOINE (1), P. DEMETTER (3), J. DEVIERE (1), P. EISENDRATH (1) / [1] Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium, Department of Gastroenterology, Hepatopancreatology and Digestive Oncology, [2] Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium, IT Department, Medicotechnical Team, [3] Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium, Department of Pathology
Introduction Indicators for colonoscopy quality assessment were developed and promoted during this last decade. However technical and human resources constraints limit local implementation of continuous recording of endoscopic quality indicators (QI). Automatic system of data extraction and presentation could help endoscopy units in their seek for quality improvement.
Aim
We hereby report our local experience in implementing colonoscopy QI record trough an automatic data extraction from two separate databases, and assess the colonoscopy quality at unit and individual levels.
Methods We locally adapted a company reporting system for colonoscopy by adding in a dedicated tab, selected procedure indicators. Endoscopic QI data from reporting system database (DB) and pathological results from another DB were extracted and merged together in a separated DB. On a regular period basis or on request, key QI are calculated and extracted. It includes adenoma detection rate (ADR), polyp detection rate (PDR), caecal intubation rate (CIR), quality of bowel preparation (using the Boston bowel preparation scale) and type of sedation. During a first period of 6 months starting in January 2016, endoscopists were encouraged to fulfill the dedicated tab on a voluntary basis. In a second period, fulfill of QI was turned to be mandatory. The completeness of QI recording was evaluated across both periods. Performance measures of all endoscopists were compared to global results of our department and to published targets.
Results During the first 6 months "free-filling" period, 1935 colonoscopies were performed with a QI tab fully filled in 63.1% of cases. In medical protocols, the CIR for screening colonoscopy was 93.1%, mean Boston bowel preparation score was 7.2±0.66, with 87% of cases with adequate preparation (Boston score >5), 94.6% of colonoscopies having been performed under propofol sedation. Among QI data, automatically extracted QI (bowel preparation quality, type of sedation) were filled in the specified QI tab in 99.9% and 97.5% respectively; whereas manually filled QI (progression, number of polyps resected and indication) were filled in the specific tab in 79.6%, 76.6% and 76.3% respectively. During this period, the ADR was 32% (range: 0%-61.3%). The PDR, an indicator that does not need a link to the pathology DB, was 37.7% with a mean of 0.94 polyp resected by colonoscopy. During the 4 months "mandatory-filling" period (July-October 2016), 1161 colonoscopies were performed with a QI tab fully filled (both for automatically and manually filled QI) in 100% of cases (the difference with the first period was statistically significant; p<0.0001). The global CIR for screening colonoscopy was 97.9%. Mean Boston bowel preparation score was 7.2±0.76 with 88% of cases with adequate preparation (90% among outpatients and 83.2% among inpatients). Colonoscopies were performed under propofol sedation in 94.9%. During this second period, the global ADR was 32.9% (range: 0%-66.7%). The PDR was 45.8% with a mean of 1.17 polyp removed by colonoscopy.
Conclusions This study illustrates that quality indicators for colonoscopy assessment in a Belgian tertiary hospital endoscopy unit could be easily implemented with limited human resources by adapting a company reporting system and link it to the pathology department database. Mandatory filling of QI items is the key for system implementation success. Our results were consistent with goals required by international guidelines. This system allows giving feedback to individual endoscopists for self-performance assessment and might be easily adapted in the future following guidelines updates.
