Alexander Waldthaler et al., 2024: Predicting ERCP procedure time - the SWedish Estimation of ERCP Time (SWEET) tool
Alexander Waldthaler 1 2, Anna Warnqvist 3, Josefine Waldthaler 4, Miroslav Vujasinovic 1 2, Poya Ghorbani 1 5, Erik von Seth 1 2, Urban Arnelo 5 6, Mathias Lohr 1 5, Annika Bergquist 1 2
1Department of Upper Abdominal Diseases, Karolinska University Hospital, Stockholm, Sweden.
2Department of Medicine Huddinge (MedH), Karolinska Institutet, Stockholm, Sweden.
3Division of Biostatistics, Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden.
4Department of Clinical Neuroscience (CNS), Karolinska Institutet, Stockholm, Sweden.
5Department of Clinical Science, Intervention, and Technology (CLINTEC), Karolinska Institutet, Stockholm, Sweden.
6Department of Diagnostics and Intervention (DDI), Surgery, Umeå Universitet, Umea, Sweden.
Abstract
Background: The duration of an endoscopic retrograde cholangiopancreatography (ERCP) is influenced by a multitude of factors. The aim of this study was to describe the factors influencing ERCP time and to create a tool for preintervention estimation of ERCP time.
Methods: Data from 74 248 ERCPs performed from 2010 to 2019 were extracted from the Swedish National Quality Registry (GallRiks) to identify variables predictive for ERCP time using linear regression analyses and root mean squared error (RMSE) as a loss function. Ten variables were combined to create an estimation tool for ERCP duration. The tool was externally validated using 9472 ERCPs from 2020 to 2021.
Results: Mean (SD) ERCP time was 36.8 (25.3) minutes. Indications with the strongest influence on ERCP time were primary sclerosing cholangitis and chronic pancreatitis. Hilar and intrahepatic biliary strictures and interventions on the pancreatic duct were the anatomic features that most strongly affected ERCP time. The procedure steps with most influence were intraductal endoscopy, lithotripsy, dilation, and papillectomy. Based on these results, we built and validated the SW: edish E: stimation of E: RCP T: ime (SWEET) tool, which is based on a 10-factor scoring system (e.g. 5 minutes for bile duct cannulation and 15 minutes for pancreatic duct cannulation) and predicted ERCP time with an average difference between actual and predicted duration of 17.5 minutes during external validation.
Conclusions: Based on new insights into the factors affecting ERCP time, we created the SWEET tool, the first specific tool for preintervention estimation of ERCP time, which is easy-to-apply in everyday clinical practice, to guide efficient ERCP scheduling.
Endoscopy. 2024 Aug 7. doi: 10.1055/a-2371-1367. Online ahead of print. PMID: 39111738
Comments 1
It is not on ESWL. However, Lithotripsy as such, and without detail on the technique used, is presented as the procedure with the longest mean procedural time of 113 minutes. It can be anticipated that further publications will follow. This interesting paper shows the advantage of large databases collected over long time periods in a Scandinavian country. In Germany we are only now and slowly approaching implementation of such systems. In Germany, there is a significant but paradoxid aversion of the population to state data collection, although Google and Co have been doing this in the private sector since decades.
Such data can be used for many different purposes and with very different primary and secondary and positive and negative effects on medical care and health care business: Benchmark setting, Quality control, reimbursement calculations, case selection, procedure selection … It can be assumed that these data will be quoted in very different areas, that is the reason why this paper is important.
Peter Alken