J Baillie, P‐A Testoni. Are we meeting the standards set for ERCP? (Gut 2007;56:744–6). Both the editors and authors are grateful to the authors of the BSG audit of endoscopic retrograde cholangio‐pancreatography (ERCP) study (EJ Williams, S Taylor, P Fairclough, et al. Are we meeting the standards set for endoscopy? Results of a large‐scale prospective study of endoscopic retrograde cholangio‐pancreatography practice (Gut 2007;56:821–9)) for pointing out that both their article and the accompanying commentary by Baillie and Testoni refer to two sets of data, which the commentary failed to clearly distinguish. The 2003 NCEPOD study (NCEPOD. Scoping our practice. The 2004 Report of the National Confidential Enquiry into Patient Outcome and Death. London: NCEPOD, 2004, http://www.ncepod.org.uk/2004.htm (accessed 28 Mar 2007)) was a small, retrospective study that looked at all endoscopy‐related deaths within 30 days for in‐patients only over a period of several months. The 2004 BSG survey was a much larger prospective, practice‐based audit of all ERCPs in five selected regions of England, which relied on self‐reporting. Despite the different methodologies, there was broad agreement in both studies regarding mortality rates. The BSG audit looked at cannulation and success rates in addition, and found that at least some elements of practice (eg selection, preparation, pre‐investigation, monitoring, antibiotics and consent) were not as bad in reality as the picture that had been painted by the NCEPOD report, which focussed solely on patients who died. It was the judgment of the NCEPOD panel, looking at retrospective data on mortality, that two‐thirds of deaths in the ASA 3–5 group occurred following procedures that were futile. The subsequent BSG survey found that while one‐third of all the deaths were in the ASA 3–5 group, 90% of the the same group survived their procedures. Although 60% of patients in ASA category 5 died, these were only 13/4561 (0.28%) patients. Possibly 100% of these unfortunate individuals would have died without ERCP.
Table 1 of the paper by Williams et al summarises the findings and recommendations of the NCEPOD that the authors consider relevant to ERCP. It refers to all deaths occurring in hospital within 30 days of therapeutic gastrointestinal endoscopy between 1 April 2002 and 31 March 2003. This is the source of concern expressed regarding lack of resuscitation equipment on site (19%), deaths of patients who were confused (16%) but had nonetheless provided written consent for their procedures, lack of any written consent (21%), lack of clotting tests (80%), inadequate or absent monitoring during sedation, and so on. The BSG audit yielded a large amount of data, not all of which could be included in the final paper. However, there is clear evidence from the BSG data that the current practice of ERCP is better than the gloomy picture painted by the NCEPOD study. For example, during the prospective study, prothrombin time was documented to have been performed within one week of ERCP in 86% of patients, which represents very good compliance with practice guidelines.
The authors wish to apologise for this confusion between the two distinctly separate studies referred to within their commentary, which created a falsely gloomy picture of current British ERCP practice, which they trust this notice will correct.