ARE ENDOSCOPISTS GOOD AND CONFIDENT AT ERCP FLUOROSCOPIC IMAGE INTERPRETATION?

Aims The aim of this audit was to determine whether there was discordance between reporting of ERCP fluoroscopic images by endoscopists and radiologists. Objectives included gauging endoscopists confidence and training at image interpretation.

Methods Patients treated with ERCP in a UK tertiary centre by three experienced endoscopists were identified from a prospective database over a 1-year period. Endoscopy reports were compared for concordance to image reports written by consultant gastrointestinal radiologists who had access to the endoscopy reports. Discordance was classified as minor if it was not of immediate clinical significance or major if it resulted in adverse outcomes.

A survey was shared through social media to endoscopists to gather practice, confidence and experience of image reporting and interpretation.

Results 409 ERCP cases were identified. 36 were not reported by radiologists. There was full concordance in 356 cases, minor discordance in 15 cases and major discordance in 2 cases. The majority of minor discordance was secondary to 8 cases of missed stones; Other cases included incorrectly placed stents, missed fistulas, strictures, additional findings and sites of bile duct injury. Cases of major discordance included a missed stone and an incorrectly placed stent. Both patients suffered from cholangitis and had another ERCP.

100 endoscopists completed the survey (55% trainees, 45% trainers). 23% of endoscopist’s images were routinely reported. Mean confidence for interpretation of images was 70%. 27% had received formal training in image interpretation. 88% felt they would benefit from formal training.

Conclusions Image reporting by endoscopists and radiologists has a discordance rate of 5%. Discordance is likely to increase the risk of patient harm; which occurred in 0.5% of patients in this case series. Image reporting provides a safety net and feedback to endoscopists. Formal training in image interpretation is uncommon and desired by endoscopists; this may improve endoscopist’s image interpretation and increase confidence.