Re: Facility Level Variation in Hospitalization, Mortality, and Costs in the 30 Days Following Percutaneous Coronary Intervention: Insights on Short-Term Healthcare Value from the VA CART Program

Bradley et al.¹ make substantial contributions to our understanding of costs during percutaneous coronary intervention (PCI) episodes of care. The authors conclude, among other things, that the financial savings of reducing PCI readmission is limited, since readmission costs account for only 5.8% of the total cost of PCI care for all patients. By examining readmission costs as a proportion of total PCI costs, however, the cost of readmissions appear relatively small. The denominator used in the analysis includes PCI costs for all patients, including those patients not readmitted. Because only a fraction of all patients are readmitted (10.8%), comparing the costs of readmission to the total costs of the care episode would make virtually any readmission costs look comparatively small. However, many of the index costs, including, for example, the procedural costs associated with PCI in patients with myocardial infarction, are necessary for the delivery of high quality, evidence-based care. Readmissions, on the other hand, are predominantly undesired and a source of preventable healthcare costs.

For example, we have shown that nearly half of PCI readmissions may be preventable² and readmissions after PCI are more costly than those associated with any other procedural episode of care.³ Given the high costs and high preventability of PCI readmissions, we believe that readmission after PCI should remain an important, actionable quality metric.

Furthermore, including readmissions after acute myocardial infarction (AMI) in the Medicare Hospital Readmission Reduction Program is not adequate to capture the substantial potential benefit of reducing PCI readmissions. Since the PCI population is larger than the overlapping AMI population, the cost savings of focusing on PCI readmissions is high. From the Medicare perspective, PCI readmissions are more expensive than AMI readmissions ($359 million versus $136 million yearly³). Many patients with PCI do not have AMI during the index hospitalization.

To illustrate this point, we have queried Medicare fee-for-service data. From 1999-2011, there were 562,356/3,682,904 (15.3%) all-cause readmissions within 30 days of PCI. Only 184,110 of those readmitted PCI patients (32.7%) had AMI during the index hospitalization during which PCI was performed. As such, focusing on AMI only as a quality metric misses more than two-thirds of PCI readmissions. Those two-thirds of readmitted patients represent an actionable and important target for quality improvement.

We agree with Bradley et al. that opportunities to improve PCI value exist by focusing on variation in the total cost of care. Since PCI readmissions are very common, very costly, and highly preventable, we believe that PCI readmission should remain an important quality metric and focus for improving value in cardiology care.