Credentialing for ERCP must be improved

Endoscopy procedures dominate the practice of most gastroenterologists in U.S. The common types, upper endoscopy and colonoscopy, are mostly performed in ambulatory care centers and offices. ERCP (Endoscopic Retrograde Cholangio Pancreatography) is different. It is substantially more dangerous, and is performed only in the hospital setting. Hospitals have credentialing committees to ensure that privileges are restricted to competent practitioners. Unfortunately, a recent survey showed that 21 percent of hospitals in U.S. had no written guidelines for initial credentialing for ERCP, and 59 percent had none for renewal of privileges.¹ This must have resulted in serious misadventures and substantial costs, and exposes hospitals to litigation for allowing incompetent physicians to practice.²

It is important to realize that ERCP is not a single specific "procedure." It is an access tool, like a scalpel or a laparoscope, nowadays allowing a range of therapeutic interventions, of varying complexity and hazard. Three levels of complexity are recognized: Standard, Advanced and Tertiary (Table 1). The standard applications are all biliary, and include removal of common bile duct stones, treatment of post-operative biliary leaks and management of benign and malignant sub-hilar biliary strictures.

Table 1. Levels of complexity
Standard
•    Deep cannulation of bile duct
•    Biliary stent removal/exchange
•    Biliary stone extraction <10 mm
•    Treat biliary leaks
•    Treat extrahepatic benign and malignant strictures
•    Place prophylactic pancreatic stents

Advanced
•    Any standard procedure after Billroth 2 gastrectomy
•    Biliary stone extraction >10 mm
•    Minor papilla cannulation in pancreas divisum, and therapy
•    Remove internally migrated biliary stents
•    Intraductal imaging, biopsy, fine needle aspiration
•    Manage acute or recurrent pancreatitis
•    Treat pancreatic strictures
•    Remove pancreatic stones, mobile and <5 mm
•    Treat biliary strictures at hilum and above
•    Manage suspected sphincter of Oddi dysfunction  

Tertiary
•    Remove internally migrated pancreatic stents
•    Intraductal image-guided therapy (eg PDT EHL)
•    Pancreatic stones impacted and/or >5 mm
•    Intrahepatic stones
•    Pseudocyst drainage, necrosectomy
•    Ampullectomy
•    ERCP after Whipple or Roux-en-Y bariatric surgery

The advanced categories are more technically challenging and more dangerous. They should be addressed only by more highly trained and experienced endoscopists and are usually limited to specialized centers.
 
We propose practical templates that applicants can use to collect the data that committees need to make their decisions. We focus here on Standard procedures, since these conditions are common, ERCP is proven to be valuable and the expertise should be available at the community level. Table 2 provides metrics for initial credentialing of physicians immediately out of training. Table 3 for those changing institutions and Table 4 for re-credentialing. They include suggestions about which responses should raise concern, to justify more detailed enquiries, and/or to restrict privileges.

Table 2. Template for initial ERCP credentialing                

For physicians just out of training

Name:                                                   Date:

Table 3. Template for initial ERCP credentialing                

For physicians more than one year out of training, applying to a new hospital

Name:                                                  Date:

Notes

¹ Standard biliary treatments include bile duct stone extraction and treatments for biliary leaks and non-hilar strictures.

² Placing a temporary stent when failing to remove a stone >10mm diameter can count as a success.

Table 4. Template for ERCP re-credentialing                

Name:                                                   Date:

Notes
¹ Standard biliary treatments include bile duct stone extraction and treatments for biliary leaks and non-hilar strictures.
² Placing a temporary stent when failing to remove a stone >10mm diameter can count as a success.

Initial credentialing (Tables 2 and 3). Details of the applicant's training are clearly important. Most will have trained in U.S. and be board certified. This may provide some reassurance about the applicant’s knowledge base, but does not address technical competence. Endoscopists trained outside U.S. may be highly competent, but it may be necessary to obtain further details. Most guidelines state that trainees must be 'hands-on' for more than 200 cases before competence can be assessed. Few centers can provide that experience in the standard three-year fellowship, so that many now provide a fourth year for training in advanced procedures, including ERCP. These numbers are poor surrogates for competence, since there is no definition about how much of the case the trainee should do for it to count, or whether the intervention was successful.

The situation would be greatly facilitated if there were a national certification process for ERCP, but there is not. A common criterion for granting initial privileges is a letter from the training director. The letter should be read carefully. For concern about vicarious legal liability, many training directors will state that the fellow has been "trained in ERCP," without referring specifically to competence.

Table 3 applies to criteria for initial credentialing for established physicians who are applying at a new hospital. Details of prior ERCP practice should be available. One simple criterion is the number of prior cases, overall and in the last year, but there are no scientific data on a reasonable threshold. One study showed that those doing more than 50 cases each year have better outcomes than those with less experience.³ Much will depend on their prior experience. Someone who has done thousands might do well with 25 each year for a while, but someone recently out of marginal training will struggle. Thus, Table 3 looks at average volumes (3b divided by 3a). It is possible to do something frequently, but not very well (like my golf). The most important metric is whether the practitioner is able to complete the intended procedure (Table 2, 4 a and b). Anything less than 85 percent for Standard procedures should not be acceptable to patients, and thus also to their guardians, the credentialing committees.

"Precut sphincterotomy" is a controversial technique. Although useful and reasonably safe in expert hands, it is dangerous when used by too frequently by practitioners with marginal skills. Thus, concern should arise if a low volume endoscopist is using it in more than 10 percent of his or her standard cases.  (Table 3, 4c, and Table 4, 2c).

Safety is perhaps even more important than success. The fact that the common procedures (upper endoscopy and colonoscopy) are so safe can cause patients and credentialers to see ERCP in the same light. But it bears repeating that it is very much more dangerous, and no safer than many common surgical procedures. Pancreatitis is the commonest adverse event, which can involve prolonged hospitalization, and death. It can occur in over 15 percent of cases when a healthy pancreas is manipulated, but should be less than 5 percent when dealing with the standard biliary cases. NSAID prophylaxis and temporary pancreatic stenting can reduce the risk.

Re-credentialing (Table 4). The data needed to support renewal of privileges are similar to those for initial credentialing, but should be easier to collect (and validate) since many of the data are available in hospital records. Practitioners doing ERCP at more than one hospital will need to summate their experience.

Giving privileges restricted to common standard procedures. We recommend, for the first time, that credentialing committees should recognize the spectrum of ERCP complexity, and have the option to grant privileges restricted to the Standard biliary indications for applicants with less experience and/or less impressive outcome data. For initial credentialing (Table 2), this might apply to endoscopists who did not have fourth year training, who could not document an 85 percent success rate for biliary cannulation, and/or whose recommendation letter is less than convincing. The same concerns would apply for initial credentialing at a new hospital (Table 3). In addition, privileges might be restricted (or denied) if the letter from a prior endoscopy administrator is missing or less than enthusiastic, if the candidate cannot provide the requested data, or if privileges had previously been denied or rescinded, or if there has been a lawsuit involving endoscopy. Having done less than 50 cases/year, or an excessive rate of adverse events, should also raise a flag. It would be concerning also if a low volume practitioner was doing many advanced procedures (Table 3d/c = <75 percent). The same points apply at re-credentialing (Table 4). Clearly the concern rises quickly if several of these criteria apply.

Hurdles to implementation. Most endoscopists probably do not currently keep detailed records of their ERCP practice, as I have repeatedly recommended⁴ and thus will not be able to provide all of the data that we consider pertinent, at least initially. The ASGE and ACG jointly developed a voluntary GIQuIC system for endoscopists to report their performance of colonoscopy and upper endoscopy.⁵ Practitioners can review their own data and benchmark with peers, anonymously. When the system adds ERCP, credentialing committees should mandate its use. Committees will need to be flexible in their implementation as these concepts and systems are introduced, allowing for some missing data. However, breaching one or more of the "thresholds of concern" should trigger a detailed review by individuals familiar with ERCP and its risks. Committees may need to enlist the assistance of outside experts, to add objectivity and legitimacy and to reduce the malpractice risk. Defining criteria for credentialing (and a track record of applying them) should also reduce the risk of a suit concerning restraint of trade. If the outcome of the review is marginal (at initial or re-credentialing) the committee will need to have a clear and logical policy on responses. These should include denial and restriction to standard biliary procedures (pending further data), with or without proctoring or full privileges.

Conclusion
ERCP nowadays involves a spectrum of indications, some more technically challenging and hazardous. Lax credentialing makes hospitals vulnerable to corporate negligence and vicarious liability claims. Patients deserve better. We provide some suggestions for improvement.

To learn more insights from Dr. Cotton, please click here.
 
References
1 Cotton PB, Feussner D, Cote GA, Dufault D.
A survey of credentialing for ERCP in USA.  GIE 2017, in press

2 ACG Releases updated legal opinion of Hospital/Payer Liability for credentialing of undertrained endoscopists. ACG June 2005: 14 (3).

3 Coté GA, Imler TD, Xu H, et al. Lower provider volume is associated with higher      failure rates for endoscopic retrograde cholangiopancreatography. Med Care. 2013 Dec;51(12):1040-7.

4 Cotton PB, Cotés GA.ERCP (Ensuring Really Competent Practitioners). Endoscopy. 2014 Nov;46(11):922-4.

5 www.giquic.org. Accessed June 15 2017

Acknowledgements. We are pleased to thank Dr Gregory Cote, James W. Saxton, Esquire and Darlene K. King, Esquire, Saxton & Stump LLC, for their helpful comments and additions.