In the recent report published by Cavender et al entitled “State Mandated Public Reporting Outcomes of Percutaneous Coronary Intervention in the United States,” the investigators should be congratulated on the herculean mission to analyze the effect of public reporting for percutaneous coronary interventions (PCI) on outcomes. This report categorizes Pennsylvania (PA) as a publicly reporting state, but, in reality, the reporting protocols in PA are much more complex and opaque than the other 2 states that were given this designation (Massachusetts [MA] and New York [NY]). As approximately half (n = 78,317) of the subjects in the “publicly reporting states” arm were from PA, we hope to shed light on the differences in reporting procedures as this may influence the readers’ interpretation of the reported results and conclusions.

In PA, cardiac surgical outcomes from coronary artery bypass graft and valve cases are rigorously evaluated through PA Healthcare Cost and Containment Consortium (PHC4). This has been the case for nearly 2 decades. Both individual operator and hospital outcomes are reported publicly in risk-adjusted fashion, and tremendous attention is paid to these metrics by hospital, patient, and surgeons. In contrast, PCI outcomes are not reported in this fashion in PA. There are several methods of PCI assessment and reporting but none meet the previously mentioned rigorous standard (nor are they similar to reporting in MA or NY). Specifically, PHC4 collects data on 15 heath conditions (septicemia, pneumonia, kidney failure, and so forth) from all PA hospitals. Two of these conditions are “heart attack angioplasty/stent” and “heart attack-medically managed”. Although hospital-based data are publically available, data accrual for the PHC4 heart attack metrics does not contain the clinical characteristic that are present in the Massachusetts Data Analysis Center-Cardiac Study and Research–Cardiac Surgery and PCI (MASSDAC) or the NY PCI registry. Data for PHC4 is accrued through claims-based extraction through a system called UB05 (Uniform claims and Billing) that is radically different from the detailed clinical registries in NY and MA. The statistical methods for risk adjustment by PHC4 is unclear and does not appear validated by any peer-reviewed process as in MA or NY and, as such, outcomes derived from this database cannot be interpreted with the same confidence as those reported in MASSDAC or the NY PCI registry.

The second method of reporting in PA is done semivoluntarily to the National Cardiovascular Data Registry–CathPCI Registry (NCDR CATHPCI). Although the general practice for the large majority of hospitals who perform PCI in PA is to submit data to NCDR CATHPCI, only a small subset of hospitals without cardiac surgery are actually required to do so. The state only monitors outcomes of the latter group. None of the NCDR data in PA are publicly reported; it is provided only to the submitting hospitals as an internal quality improvement tool. The state reserves the right to monitor these data in non-CT surgery hospitals but again does not publicly report the data, nor does the state report individual practitioner outcomes.

Finally, 2 academic medical centers in PA submit data to the University Heath Consortium. These data are used for quality comparisons internally but also are not publically reported.

In summary, in contrast to MA and NY, there is no mandated publicly accessible report of outcomes of elective PCI in PA. Additionally, acute coronary syndrome–related reporting is at a hospital-wide level and not at a practitioner level. The voluntary submission of PCI data from each hospital is then mined through claims data, as opposed to clinically rich registries, make this reporting far different from that in NY and MA. In light of these considerations, the current practices in PA are more analogous to the excluded states that use AHRQ data for outcomes assessments and should temper the conclusions of this study.