Does Practice Make Perfect?^∗ Free Access

Introduction

We have all heard the phrase “practice makes perfect” from the time we were young and first tried to master a skill. It just makes sense that the more practice and experience one has, the more skillful one will become at any task. The words of Vince Lombardi, legendary coach of the Green Bay Packers, add another perspective to this old adage. Obviously, Coach Lombardi was not referring to the practice of medicine, but as physicians transition from a fee-for-service, volume-driven reimbursement scheme to one based more heavily on the quality of care delivered, perhaps we should be asking ourselves: How do we make our practice (of interventional cardiology) perfect?

An appreciation for the relationship between volume and outcome has existed for many years and is not limited to percutaneous coronary intervention (PCI). Almost 40 years ago, Luft et al. (2) examined mortality rates for 12 surgical procedures of varying complexity in 1,498 hospitals to determine if there was a volume-mortality relationship. They found an inverse relationship between volume and mortality for complex surgical procedures, but not for more common straightforward operations. The volume-outcome relationship for PCI was extensively reviewed in the 2013 American College of Cardiology Foundation/American Heart Association/Society for Cardiovascular Angiography and Interventions (ACCF/AHA/SCAI) Clinical Competence Statement for coronary artery interventional procedures (3). Based on a review of the accumulated literature up to that time, the writers concluded that an institutional volume threshold <200 PCIs annually appeared to be consistently associated with worse outcomes, but above this level there was no relationship between higher annual volumes and improved outcomes. Evidence suggesting a volume-outcome relationship for individual operators was much weaker, but did exist in 2 large studies each with over 100,000 patients and has been shown for complex PCI procedures such as unprotected left main stenosis and chronic total occlusions (4–7).

In this issue of JACC: Cardiovascular Interventions, Inohara et al. (8) revisit the relationship between facility and operator case volume and in-hospital outcomes of PCI. Using the Japanese PCI Registry, 323,322 PCIs performed at 625 hospitals by 4,211 operators were analyzed. It was striking to read that over 80% of the facilities in Japan have an annual PCI volume ≤200 cases. In-hospital mortality occurred in 0.9% and the composite endpoint of in-hospital mortality plus periprocedural complications (tamponade, shock requiring mechanical or inotropic support, definite stent thrombosis, emergency surgery, and bleeding requiring transfusion) occurred in 2.2% of the study cohort. Facilities and operators were categorized into deciles based on the annual volume of PCIs performed and the occurrence of the 2 outcome variables determined for each decile. The authors found that the adjusted risk for in-hospital mortality and the composite endpoint was significantly higher in hospitals within the lowest decile (<150 PCIs/year) compared with all higher deciles. No significant relationship was observed between operator volume and outcomes even at their lowest decile, which was 1 to 23 PCIs/year. The same findings were observed when the analysis was confined to emergency or urgent PCI cases. Therefore, the authors’ findings are consistent with many prior studies confirming that the risk of in-hospital mortality and selected other outcomes is higher at low-volume facilities. Moreover, in agreement with the conclusions made in the 2013 ACCF/AHA/SCAI Clinical Competence Statement, no definite volume-outcome relationship could be proven for individual operators.

As the authors’ findings from this large Japanese cohort confirm the findings of many other studies, can we finally put the question of a volume-outcome relationship for PCI to rest? At least for annual facility volume, we believe the answer is yes—low-volume facilities in both the United States and Japan have less favorable outcomes compared with higher-volume facilities. This enforces the dictum “practice makes perfect” and thus questions why low-volume facilities exist if outcomes are worse. Those that operate low-volume facilities, however, will quickly respond that they exist to provide critical and timely access to PCI services in their local communities especially for patients with ST-segment elevation myocardial infarction. No doubt that critical access is true for some isolated locations, but 2 separate studies have shown that approximately 80% of the U.S. population lives within a 1-h drive time of a PCI center with a median driving time of approximately 11 min (9,10). Moreover, Horwitz et al. (11) showed that hospitals are more likely to introduce new invasive cardiac services when neighboring hospitals already offer such services and confirmed that the increase in the number of hospitals offering invasive cardiac services did not lead to a corresponding increase in geographic access. Although support for a facility volume-outcome relationship is solidified when conclusions from the present study are added to the existing literature it is unlikely this will lead to the closure of low-volume facilities. How to address low volume facilities is a challenge looming for large health care systems as the quality of care receives more focus and economic pressures encourage the consolidation of services and volume to enhance efficiency and quality, respectively.

With this additional support for a facility volume-outcome relationship, why does an operator volume-outcome relationship remain elusive and unproven? Fortunately, the serious complications from PCI are now quite rare. Trying to show a statistically significant difference in outcomes when the number of annual cases per operator is small remains difficult. Contemporary data on the number of annual cases per operator in the United States is hard to find, but older data estimated that over one-half of U.S. operators performed <40 cases annually and it is unlikely that operator volume will increase (12). As reported in this article, the median annual number of PCIs/operator in Japan is even lower at 28 (8). In the authors’ data it is important to recognize that the adjusted odds ratios for operator in-hospital mortality and the composite outcome metric was above 1.0 for many of the deciles over the fourth decile. Not all of these deciles had confidence limits that excluded unity, but it does raise a question about the true relationship between operator volume and outcomes. Is it possible that higher-volume operators are more likely to perform higher risk and challenging procedures, with resultant higher complication rates even after adjustment? Without individual patient-level analysis, this true relationship may be challenging to discern.

What is the answer as we strive for a higher level of quality in interventional cardiology? What did Vince Lombardi mean when he said “Practice does not make perfect. Only perfect practice makes perfect” (1)? How does this apply to the catheterization laboratory? Similar to football, PCI is a team sport. The operator may be the quarterback, but every member of the team has an important role and the team must practice together for the procedure to be a success and the patient experience positive. However, as cautioned by others and as emphasized in the 2013 ACCF/AHA/SCAI Clinical Competence Statement volume alone may not be an appropriate surrogate for quality (3,13). Practice (or case volume) by itself is of little value if proper execution is lacking. Proper execution of a PCI means the case was performed for appropriate indications, the patient was properly informed of the risks and benefits, the procedure was technically sound, and that appropriate evidence-based therapies were used before, during, and after the procedure (14,15). This is a goal that can be achieved in low- and high-volume laboratories and by low- and high-volume operators, but only if we are focused on executing every element of the PCI perfectly. Finally, processes must be in place at hospitals to review quality data regularly and provide feedback to operators, if the benefit of quality-based reporting is to be optimized.

Footnotes