Clinical Editor's Corner: Kern

What is the Annual Volume Requirement for a PCI Operator?

Morton J. Kern, with contributions from Richard Bach, Barnes Jewish Hospital, St. Louis, Missouri; Steven R. Bailey, University of Texas Health Sciences Center at San Antonio, Texas; Charles Chambers, Harrisburg, Pennsylvania; David J. Cohen, Beth Israel Deaconess Medical Center, Boston, Massachusetts; Larry S. Dean, University of Washington, Seattle, Washington; Greg Dehmer, Temple, Texas; Ted Feldman, Evanston, Illinois; Kirk Garratt, Christiana Hospital, Newark, Delaware; Ajay Kirtane, Columbia University, New York City, New York; Lloyd Klein, Chicago, Illinois; Mitchell W. Krucoff, Duke University Medical Center, Raleigh, North Carolina; Jeffrey J. Popma, Beth Israel Deaconess Medical Center, Boston, Massachusetts; Duane Pinto, Beth Israel Deaconess Medical Center, Boston, Massachusetts; Fred Resnic, Lahey Clinic, Burlington, Massachusetts; Gregg Stone, Columbia University, New York City, New York; Zoltan Turi, Hackensack University Medical Center, Hackensack, New Jersey; George Vetrovec, Medical College of Virginia, Richmond, Virginia.   

Morton J. Kern, with contributions from Richard Bach, Barnes Jewish Hospital, St. Louis, Missouri; Steven R. Bailey, University of Texas Health Sciences Center at San Antonio, Texas; Charles Chambers, Harrisburg, Pennsylvania; David J. Cohen, Beth Israel Deaconess Medical Center, Boston, Massachusetts; Larry S. Dean, University of Washington, Seattle, Washington; Greg Dehmer, Temple, Texas; Ted Feldman, Evanston, Illinois; Kirk Garratt, Christiana Hospital, Newark, Delaware; Ajay Kirtane, Columbia University, New York City, New York; Lloyd Klein, Chicago, Illinois; Mitchell W. Krucoff, Duke University Medical Center, Raleigh, North Carolina; Jeffrey J. Popma, Beth Israel Deaconess Medical Center, Boston, Massachusetts; Duane Pinto, Beth Israel Deaconess Medical Center, Boston, Massachusetts; Fred Resnic, Lahey Clinic, Burlington, Massachusetts; Gregg Stone, Columbia University, New York City, New York; Zoltan Turi, Hackensack University Medical Center, Hackensack, New Jersey; George Vetrovec, Medical College of Virginia, Richmond, Virginia.   

Many of the editor’s pages come from issues raised by practicing interventionalists. Recently, a query came to our colleague, Dr. Zoltan Turi, from a chief of cardiology, who asks, “I was recently told that the American Heart Association (AHA) was lowering the recommended annual volume requirement for each percutaneous coronary intervention (PCI) operator to 50 cases per year, down from 75. I cannot find that in writing and thought you might know if this is true. Also, it is commonly believed that lifetime PCI experience should negate the need to maintain an annual PCI volume, but I cannot find an authoritative answer as to what constitutes adequate lifetime experience. Could you provide a reference or your expert opinion?”

Mort Kern, Long Beach, California: I recall that the Society for Cardiovascular Angiography and Interventions (SCAI) recommended 75 cases/year as the minimal annual volume several years ago and that this made its way into the AHA/American College of Cardiology (ACC) guidelines for PCI operator annual volumes. I also recall that most California operators do less than 50 cases per year.  

Richard Bach, St. Louis, Missouri: The revised operator number of 50 cases per year comes directly from the update to the ACC/AHA/SCAI clinical competence statement in 2013 by Harold et al.1  

George Vetrovec, Richmond, Virginia: The most recent published guidelines relating to operator volumes support 50 cases per year. The bottom line is ideally 50 per year averaged over a 2-year period and performed in an institution that does >200 per year. The “Senior” experienced option is pretty vague (see Fanaroff et al2). 

Lastly, a recent NCDR publication (Fanaroff et al2) shows that 44% of interventions are performed by operators performing <50 per year, 27% by operators performing 50-100 per year, and only 29% of operators perform more than 100 per year. Figure 1 demonstrates a correlation between volume and mortality. An accompanying editorial cautioned against volume conclusions, since the lower volume operators tended to be in more rural areas, potentially performing a “service” for patients in these areas regarding ST-elevation myocardial infarction (STEMI).   

Fred Resnic, Burlington, Massachusetts: I believe the revised guidance came from the ACC/AHA/SCAI 2013 Clinical Competency Statement. The relevant concluding statement on revised institutional and operator volume guidance is below:  

“2.8.3. Conclusions
In the current era, volume-outcome relationships are not as robust as those that were shown when balloon angioplasty was the only treatment modality. More recent data supports a modest volume-outcome relationship for variables other than mortality, but these data have limitations and are not consistent across all studies. An institutional volume threshold <200 PCIs/annually appears to be consistently associated with worse outcomes, but above this level there was no relationship between even higher annual volumes and improved outcomes. 

“Accordingly, the writing committee recommends a minimum institutional volume threshold of 200 PCIs per year. There is less evidence to support a threshold for individual operator volume for both elective and primary PCI. It is the writing committee’s recommendation that interventional cardiologists perform a minimum of 50 PCI procedures per year (averaged over a 2-year period) to maintain competency. The writing committee cautions against focusing on specific volume recommendations, and emphasizes that procedural volume is one of several variables to consider when determining operator competency. Volume is not a surrogate for quality and should not be substituted for risk-adjusted outcomes and other measures of quality. Periodic case review and ascertainment of the appropriateness of procedures should be performed for all operators and at all institutions. Our writing committee strongly encourages the participation in a local or national registry, such as the NCDR® CathPCI Registry, which can help measure performance, assess appropriateness of procedures, and promote continuous quality improvement.”

Greg Dehmer, Temple, Texas: George (Vetrovec) is 100% correct in his assessment of the 2013 Competency document. I had the privilege of being on the writing committee (as well as a few others who are on Mort’s list) and the discussions about operator volume were “vigorous” and, at times, contentious. Ted Bass and John Harold deserve credit for getting the group through these challenging discussions. The concluding paragraph (as noted in Fred Resnic’s comment) reminds us that there is concern about low-volume labs (<200 annually), as well as lingering concerns about low-volume operators. All that being said, if I ever need a PCI and have a choice, I’m finding a high-volume operator who works in a high-volume lab with a good quality assurance program in place.     

Steve Bailey, San Antonio, Texas: George, thanks for adding the references. The other reference is from Badeka et al, Circ 20133 that supports 50 procedures/year with the following conclusions: “We noted that patients undergoing PCI would experience a 0.53% absolute risk reduction in mortality, after adjustment for other clinical variables, in the hands of an operator with annual PCI volume of 50 to 75 procedures/yr as compared with an operator with annual volume of <50 procedures/yr. … We also found that patients undergoing PCI experience a 2.74% absolute risk reduction in secondary outcome with operator PCI volume of 50 to 75 procedures/yr as compared with an operator with annual volume of <50 procedures/yr.” While Badheka et al add information, direct data and outcomes would be helpful, particularly regarding our training program numbers and continued certification.

Jeffrey J. Popma, Boston, Massachusetts: Many of us are doing 200+ transcatheter aortic valve replacements (TAVRs) per year, but less PCIs as a result. Any consideration for this? Otherwise — I am going to have to take more STEMI call…

Lloyd Klein, Chicago, Illinois: I was chair of a SCAI committee which reviewed this data in 2011, and there has never been really great data about volume relationship with competence. Undoubtedly, the more experience any operator has, the more expert they are at that task. The question is whether any specific volume threshold defines competence, and no parameter predicts outcomes.  

Ajay Kirtane, New York City, New York: In the complex PCI space, these volumes-outcomes relationships are emerging and are more than univariate associates (chronic total occlusion [CTO] PCI success in New York State, left main coronary angioplasty [LMCA] outcomes in China). As noted above, Fanaroff et al2 addressed this relationship as well. The issues of prior experience aside, I would have to say that the AHA stance (if true) is disappointing. I surmise that the dropping volume requirements from multiple societies is likely one of the (few) untoward side effects of offering widespread primary PCI at so many U.S. hospitals. I wonder if we didn’t need to justify cath labs on this basis, whether this would be as defensible an issue.

David J. Cohen, Boston, Massachusetts: We discussed exactly the issue that Ajay is raising.3 The desire/need to provide 24/7 primary PCI at virtually all U.S. PCI centers has led to substantial dilution of operator-specific PCI volumes.

Larry S. Dean, Seattle, Washington: Great conversation. My two additions: 1) The State of Washington just used the 2013 statement to decrease the certificate of need (CON) volumes for off-site surgical backup for elective PCI to 200. The (real) reason was that most sites with current CON weren’t meeting the previous 300 target. Obviously, at some point, the State is going to have to deal with the flawed notion that PCI access should use a model that distributes that access to multiple sites, many with marginal and/or declining volumes. 2) The issue of structural heart disease (SHD)/TAVR and its impact on operator volumes is a real one. How do you maintain competence in both PCI and SHD in high-volume SHD centers? What does certification in PCI and PCI volume requirements have to do with competence in TAVR? SHD should be a separate specialty, in my humble opinion. Obviously, there are more questions to ask…and we need to start thinking about how we are going to answer them.

Lloyd Klein, Chicago, Illinois: The odd thing is the data connecting volume and outcomes is better with STEMI than in any other situation. Keep in mind that 60% of U.S. interventionists don’t even do 50 cases/year.4-6 And that outcomes are far more influenced by case selection than volume. Also, I am sorry, Ajay, but volume associated with outcomes is univariate, and goes either almost or totally away when including case risk. The National Cardiovascular Data Registry (NCDR) is not much help here; the most recent article earlier this year does show a minimal advantage of comparing the lowest volume with the highest volume operators, but not the intermediate ones.

George Vetrovec, Richmond, Virginia: There is another aspect of low operator volume that has not been addressed. I fear that low-volume operators either are unaware of the potential limits of PCI, or just don’t want to admit their limits and don’t refer patients to high-volume operators. The consequence is that patients who might benefit from PCI revascularization are never referred and given an opportunity for a PCI benefit. The real issue for many low-volume operators is they may be technically competent, but with a small case volume denominator, they can’t take the chance of high-risk procedures, because of the higher risk of adverse events. Nonetheless, the operators don’t often refer.

Lloyd Klein, Chicago, Illinois: This conversation perpetuates several only partially correct myths, based on cognitive biases, about the volume-quality relationship. The most important point to be made is that volume and quality are not synonymous. Although some (but not all) high-volume operators are high quality, and some (but not all) low-volume operators are low quality, there is no proven relationship that allows one to infer quality from volume and vice versa. There is no threshold value of annual, or lifetime volume that when crosses it, or does not, that defines competence.

Moreover, the likelihood is that volume is more of a function of one’s maturity of practice, strength of one’s referral base and local practice, and connections in the contracted, managed care world than it is of one’s competence. Furthermore, the idea that one would want to select the highest volume operator at the highest volume center to do one’s stent clearly does not fit with the data. I would want to know the outcomes of all of the operators available particularly with the specific challenges of MY case, and ensure they had plenty of experience with that. If that meant a lower volume operator, or an intermediate one, I would gladly choose that person. Also, at the time of night or week of my acute coronary syndrome (ACS) or MI, maybe that person isn’t available, or perhaps he/she is out of network. The quality question isn’t who is the best, but rather who is competent to do my, or anyone’s, stent. We have to recognize that if we limited singing to the few birds with the best voices, the woods would be silent indeed. We have a higher moral responsibility than that.

George Vetrovec, Richmond, Virginia: While I absolutely agree with Lloyd’s points about the complexity of the volume-outcome relationship, the lack of a clear volume threshold, and the fact that there are undoubtedly many high-quality PCI operators who do relatively low volumes of procedures, the one unequivocal truth is that the higher the volume an operator (or an institution) does, the easier it is to evaluate its quality.  

Simply put, when an operator or an institution is low volume, it is completely impossible to distinguish good quality from good luck — at least by evaluating outcomes. This task is much easier (although still not straightforward) for higher volume operators or institutions. So, one of the underappreciated benefits of increased procedural volume is that it becomes easier to evaluate quality.

Mitch Krucoff, Raleigh, North Carolina: One aspect that is remarkably unexplored is a probable difference between the operator doing 50 PCIs/year who has done 5,000 previous cases and the newly graduated fellow doing 50 cases/year. There is likely a difference between an operator with 10,000 cases in their career, but can’t stand up long enough anymore to do more than 50/year now. There is competence with experience, which may decrease with age.  

Ajay Kirtane, New York City, New York: It is clear that volume is not the only arbiter of quality. But, unfortunately, it is one of the only enforceable current metrics of competency besides board certification (and we all know how much that has to do with actual competency). While Lloyd’s points about the highest volume operators may be true in individual cases, the true reason for minimum volume and competency requirements are to ensure that: 1) The operator is facile in a technical sense in the lab (this includes the ability to understand the equipment, troubleshoot, etc.); 2) Staff are comfortable with the operator, his/her decision-making; 3) The operator is able to present an objective option on all technical options for the patient with reasonable experiential knowledge (George’s point); 4) When random complications occur, the operator is equipped to deal with them competently, especially because time is of the essence. 

Of course, case mix and selection will influence the frequency of complications. If we are truly self-reflective, we will all agree that if we haven’t seen one [a complication] in a while, we are more sluggish in dealing with them (especially given the team-based nature of how to manage them effectively). PCI success can be influenced by the different strengths/qualities of the individuals, teams, hospitals, etc. But empirically speaking, shouldn’t we be more aspirational regarding patient care than a 50-PCI case standard? Are we really at the point where we use the difficulties in studying the complexity of this and establishing true competency (e.g., via simulators, observation, etc.) as an excuse to societally endorse operators who are literally doing less than 1 intervention a week?

I mean no disrespect to operators doing less cases. I even realize that some lower volume operators may even be on this [email discussion], and have great prior volume and are thoughtful and experienced (and arguably more so than some of the highest volume operators in this country). But that doesn’t change the fact that those lower volume operators would no doubt be better in a crisis if they did more PCIs. In our lab, we do the more complex cases together with operators who have lower volumes and I can assure you that both operators (lower and higher volume) are enriched by the experience. Some of this takes humility, but if we put the patient first, this isn’t really that difficult. To use Lloyd’s analogy, we want all the birds to be heard, but some birds could use a little help from the stronger voices to follow the tune. I personally don’t profess to know what the magic number is, but less than one a week seems to be really stretching it.

Ted Feldman, Evanston, Illinois: The focus on PCI volume rather than total interventional volume, to Jeff Popma’s prior point about higher volume TAVR operators, does not account for the facility in technique and in managing patients, procedures, judgments, and complications that comes from the whole procedure mix. 

Lloyd W. Klein, Chicago, Illinois: There are problems assessing outcomes in everyone, but the higher the number, the less “noise” and the more accurate the assessment. To Mitch, there is no data to support that lifetime numbers assure that competence. To Ajay, I am sympathetic, but the data does not support your contentions. In such a highly political atmosphere as any cath lab, it helps to know what you can conclude about other people’s work and what you cannot.

Duane Pinto, Boston, Massachusetts: I observe that low-volume operators have a higher prevalence of a phenotype of suboptimal performance in the lab, defined as using more resources (time) and needing to be bailed out, ‘babysat’ by others. True, I have a hard time proving it with outcomes extant in the literature, though. Let’s face it, our definition of what we would all consider doing a low-quality PCI still has nearly no chance of resulting in an in-hospital death. Maybe we shouldn’t care as a group about things that may have an association with an outcome besides death: length of stay (LOS), restenosis, cost of equipment, staff overtime, as long as the patient survives? I may have a cognitive bias, but it’s still 20 years of cognition and lots of “can you come into room 2?” 

Conversely, there are high-volume operators who also s#$%k [read: “under-perform”  MK] for different reasons and these folks are just as hard to describe, especially when inappropriate PCI probably has lower mortality (I initially had spelling error “morality”) than indicated PCI.  

I agree that this is complicated, but it seems to me that volume relationships to in-hospital mortality are not the appropriate outcomes to identify these operators, describe the quality of their PCIs, or to monitor their practice improvement.  

Mort Kern, Long Beach, California:  Duane, who then should be disqualified? 

Gregg Stone, New York City, New York: This whole question falls under the rubric of “you don’t have to randomize parachutes”. Of course outcomes will be generally better by high-volume operators — whether PCI, plumbing, or shooting basketballs. The only way they can’t be is if the operator limits the types of cases s/he does (which is wise practice for that person, but not particularly helpful to patients, as less skilled docs will often not refer tough cases to a more skilled operator), or if they are routinely getting help from others and taking the credit. The fact that the data don’t clearly show what we all know in our hearts to be true means you should question the limitations of the methodology of the way the data is collected and/or analyzed, not the underlying assumption.

Charles Chambers, Harrisburg, Pennsylvania: Mort... Human nature is the challenge here, the ability to critically assess one’s self, as selflessly as possible, remembering the key is the patient. You can’t regulate this one.

Kirk Garratt, Newark, Delaware: The slippery issue of how to measure quality when the things we can measure don’t really help us very much is an old one. Bill Weintraub and I wrote a “Perspectives” piece in Circulation earlier this year that looks at this problem.7 Sadly, there isn’t a good way out of the conundrum, but in a subsequent article, we recommend very much against using low-frequency, hard-endpoint metrics like mortality after PCI as a measure of a physician’s worth.

The Bottom Line

PCI outcomes and quality have a relationship to the number of procedures of both the operator and the lab.  More is better, but a quantitative link is hard, if not impossible, to establish. The published number from the guidelines says that 50 cases/year is acceptable, but a review of performance of both high- and low-volume operators and their outcomes must be performed to ensure the low-volume experienced operator and the low-volume novice have similar good results. Appropriateness for all operators, especially high-volume operators, should also be part of continuous quality improvement in any lab.

I hope this conversation stimulated your interest and self-reflection to continue to provide high-quality care to our PCI patients. 

References

  1. Harold JG, Bass TA, Bashore TM, et al. ACCF/AHA/SCAI 2013 update of the clinical competence statement on coronary artery interventional procedures: a report of the American College of Cardiology Foundation/American Heart Association/American College of Physicians Task Force on Clinical Competence and Training (Writing Committee to Revise the 2007 Clinical Competence Statement on Cardiac Interventional Procedures). J Am Coll Cardiol. 2013; 62: 357-396.
  2. Fanaroff AC, Zakroysky P, Dai D, et al. Outcomes of PCI in relation to procedural characteristics and operator volumes in the United States. J Am Coll Cardiol. 2017 Jun 20; 69(24): 2913-2924.
  3. Badheka AO, Patel NJ, Grover P, et al. Impact of annual operator and institutional volume on percutaneous coronary intervention outcomes: a 5-year United States experience (2005-2009). Circulation. 2014; 130: 1392-1406. 
  4. Klein LW, Harjaj KJ, Resnic F, et al. 2016 Revision of the SCAI position statement on public reporting. Catheter Cardiovasc Interv. 2017 Feb 1; 89(2): 269-279.
  5. Klein LW, Ho KKL, Singh M, et al. Quality assessment and improvement in interventional cardiology. A Position Statement of the Society of Cardiovascular Angiography and Interventions. Part II. Public reporting and risk-adjustment. Catheter Cardiovasc Interv. 2011; 78: 493-502.
  6. Maroney J, Khan S, Powell W, Klein LW. Current operator volumes of invasive coronary procedures in Medicare patients: Implications for future manpower needs in the catheterization laboratory. Catheter Cardiovasc Interv. 2013; 81: 34-39.
  7. Weintraub WS, Garratt KN.  Challenges in risk adjustment for hospital and provider outcomes assessment. Circulation. 2017; 135: 317-319.

Disclosure: Dr. Kern is a consultant for Abiomed, Merit Medical, Abbott Vascular, Philips Volcano, ACIST Medical, Opsens Inc., and Heartflow Inc.