Site-Specific PCI Public Reporting for Operators May Be Misleading

For operators who work at multiple centers, public reporting of PCI outcomes may be less informative when 30-day mortality is reported separately for their cases at each site, data from New York State show.

Instead of this site-specific risk-adjusted mortality rate (RAMR), it would be better to calculate mean RAMR, Nancy Yang, BA (University of Pennsylvania School of Arts and Sciences, Philadelphia), and colleagues say. Their analysis was published online recently as a research letter in JAMA Cardiology.

New York currently releases death rates at both the hospital and the physician level for PCI and cardiac surgery, including TAVR. Although the aim is transparency, many have cited public reporting’s unintended consequences like the potential for risk avoidance as well as the cost and time required to collect data. Still, being tagged as an outlier does appear to encourage better care.

Ashwin S. Nathan, MD (Hospital of the University of Pennsylvania), the new report’s senior author, said that they were curious about how mortality estimates held up when moving from the big picture to home in on an individual physician’s performance at an individual hospital.

“As you get from a large entity to smaller ones, there can be a lot of instability—meaning, a single bad outcome can really affect what the perceived quality is of the physician,” he noted.

This is especially important for operators whose cases are divvied up across sites, with fewer PCIs occurring at each hospital. “If you do 10 interventions and you have, say, one chance bad outcome, it can really look like you have a 10% rate of that bad outcome, which is pretty unrealistic if you look at all major registries of complications including mortality,” he told TCTMD.

This “may be misleading to patients and other physicians who are considering who to refer to,” said Nathan.

As you get from a large entity to smaller ones, there can be a lot of instability—meaning, a single bad outcome can really affect what the perceived quality is of the physician. Ashwin S. Nathan

Based on data from the New York State Department of Health’s website, there were 142,853 PCIs performed by 373 physicians at 61 hospitals from 2014 to 2016. More than half of the doctors (55.5%) practiced at multiple centers, where they performed 57.5% of the state’s PCIs during this 3-year period. There was no difference in physicians’ PCI volume based on whether they practiced at one or more hospitals.

Multisite operators did procedures at a median of two sites, with a median of 56 PCIs done at each hospital. Their mean RAMR was 1.11%, ranging from 0 to 5.33%, while their median site-specific RAMR was 0.52%, ranging from 0 to as high as 47.69%. Using the site-specific math, 15 physicians were classified as “outliers” when it came to mortality; this number dropped to four physicians when calculations were based on the broader metric of mean RAMR.

“Because physicians’ PCI volumes at individual centers may be small, a single mortality event at a hospital where a physician performs a few cases could result in an extremely high reported RAMR owing to the stochastic nature of mortality following PCI, even with perfect risk adjustment,” Yang et al explain, adding that this knowledge could inform quality-improvement initiatives in New York and elsewhere.

Asked how this might occur, Nathan suggested it may be best to stick with the aggregate reporting of data, for entire hospitals or for physicians’ performance as a whole, rather than dividing one clinician’s numbers center by center. “New York in the past has been pretty receptive to changing things in order to make reporting more equitable and make sure the right things were happening [for] patients,” he said. “They’ve done a really good job with that at evolving these reports.”

Overall, when it comes to public reporting, there needs to be a balance between its benefits and its unintended consequences, he advised.