‘Inappropriate’ PCI: New York State Efforts Appear to Make a Dent
Three years after the 2009 AUC were released, NY hospitals began getting feedback on performance—and, it seems, putting it into practice.
Following the initial release of appropriate use criteria (AUC) for coronary revascularization in 2009, hospitals and physicians nationwide have grappled with how the ratings should best be used to inform practice. Now, new data from New York State suggest that an emphasis on appropriateness is having an impact, though researchers say it’s unclear whether the shifts are due to local efforts or nationwide trends.
So far, there have been two-and-a-half versions of the AUC: the original 2009 document, its 2012 update, and as of December 2016, criteria dedicated solely to acute coronary syndromes. The second half of the newest version, which will be geared toward elective cases, is still in the works.
Back in February 2012, the New York State Department of Health began sharing appropriateness data with hospitals. Around the same time, in November 2011, the New York State Medicaid Redesign Team recommended that reimbursement be withheld for stable PCI cases rated as inappropriate.
Lead author of the current study, Edward Hannan, PhD (State University of New York, Albany), who is responsible for analyzing the New York data and compiling the quarterly reports sent to hospitals, told TCTMD it is hard to tease out the effects of the state-level reporting from worries over Medicare reimbursement, not to mention the growing awareness about appropriateness in the United States as a whole, thanks to numbers gathered and published by the American College of Cardiology’s National Cardiovascular Data Registry (NCDR).
Much like the NCDR, “the intent of the health department is also to improve quality of care,” rather than be punitive, Hannan emphasized.
To New York hospitals, he said, “we fed back quite a bit of information on the nature of the type of inappropriateness, . . . so that they could see which scenarios were particularly problematic. We also gave them results by individual cardiologists, so they could see whether or not some cardiologists were doing substantially worse than other ones and [if the hospital’s] rates were attributable only to a subset of cardiologists.”
The hospital reports are confidential, but they occur against the backdrop of public reporting of PCI mortality in New York, Hannan pointed out.
There’s also the concern generated over whether AUC—meant to take a broad look at practice patterns—would in fact be used to determine Medicare payments in individual cases, Ajay Kirtane, MD (NewYork-Presbyterian/Columbia University Medical Center), stressed to TCTMD. Ultimately, this controversial proposal was taken off the table this past spring, and no payments were denied.
“Especially in the Medicaid population, which is potentially vulnerable, I was concerned that this would lead to underutilization—so much scrutiny and so much fear of reimbursement being withheld that even for patients who might clinically meet criteria for revascularization, it wouldn’t be offered to them because hospitals and providers might fear not being paid,” said Kirtane. “That’s just not in the spirit of what we try to do in clinical medicine.”
The New York State findings were published yesterday in the Journal of the American College of Cardiology.
Fewer ‘Inappropriate’ Stable PCI Cases Over Time
For the study, Hannan along with senior author Spencer B. King III, MD (St. Joseph’s Health System, Atlanta, GA), and colleagues used New York’s Percutaneous Coronary Interventions Reporting System to look at all stable PCI patients treated between 2010 and 2014 at 58 nonfederal hospitals in the state. Results were divided into two periods: before versus after 2012, when New York began giving feedback to hospitals.
The percentage of PCIs deemed “inappropriate” decreased between 2010 and 2014, for all patients (from 18.2% to 10.6%), Medicare patients (from 15.3% to 6.8%), and non-Medicare patients (from 18.6% to 11.2%; P < 0.001 for all). Further analyses based on timing of treatment found no overall trend across the entire 5-year period but did show a 40% lower odds of PCI being inappropriate in 2012-2014 versus 2010-2011 (OR 0.60; 95% CI 0.44-0.83).
In 2010-2011, nearly half of the 50 hospitals with an annual volume of at least 20 stable PCI cases had an inappropriateness rate above 20%. By 2012-2014, only 8 of those 50 hospitals exceeded the 20% rate.
Additionally, the total number of PCIs performed in patients without ACS or prior CABG that were rated as inappropriate also dropped by 69% for all patients (from 2,956 cases in 2010 to 911 in 2014) by 75% for Medicaid patients (from 340 cases to 84). “These decreases in inappropriateness volumes were much larger than the decreases in inappropriateness rates because of the large decline in patients with nonacute conditions going to PCI,” the paper notes.
Indeed, the percentage of all PCIs performed in stable patients decreased from 37% in 2010 to 20% in 2014. Such a pattern could potentially be explained to some degree by upcoding, in which providers are justifying PCI as appropriate by recording patients’ disease as more severe than it actually is.
No matter how much auditing is done, there are upcoding dangers with regard to symptoms, because the symptoms are not clear-cut clinical measurements. Edward Hannan
“No matter how much auditing is done, there are upcoding dangers with regard to symptoms, because the symptoms are not clear-cut clinical measurements,” Hannan explained. “It’s always difficult to confirm the accuracy of that, [since the information is] passed on from patient to physician to medical record without any hard measurement that’s done.”
A rise in Canadian Cardiovascular Society (CCS) class over time would provide evidence that upcoding is happening, he said, but because the definition of CCS class changed in 2012, this trajectory is hard to document.
What Explains the Numbers?
In an accompanying editorial, Frederik A. Masoudi, MD (University of Colorado School of Medicine, Aurora), and colleagues report their own analysis of AUC-related trends, based on NCDR data. Their overall findings are consistent with those of Hannan et al; however, they found different patterns that hint the threat of denied Medicare reimbursement had a real effect.
According to Masoudi et al, Medicare patients saw a sharp drop in inappropriate PCI just after the proposal to deny reimbursement, lending support to the idea that worries over payment fueled some practice shifts. In contrast, inappropriate PCI among non-Medicare patients declined more steeply ahead of the announcement, the NCDR data show. Patients who neither lived in New York nor were Medicare beneficiaries saw steady decreases over time.
Hannan, for his part, is unconvinced by Masoudi et al’s NCDR analysis. “They don’t have all hospitals in New York State in the NCDR,” he pointed out. “Our reporting in our paper [done by quarter] did not show that kind of thing. It showed roughly the same kinds of trends, not a huge drop off like that, so I’m more inclined to believe what we reported.”
The editorialists also take on the issue of upcoding, saying it must be considered when examining the interplay between state and national trends. “Certainly, incentives for ‘gaming’ increase proportionally with financial disincentives for performing procedures classified as inappropriate,” they stress. AUC might have a role in providing value-based care, they say, but care must be taken to ensure that there aren’t incentives that could spur underuse.
Balancing Appropriateness With the Risk of Underuse
Kirtane, too, cautioned against the potential for underuse. With the advent of AUC, “we as a field have done a better job of actually scrutinizing the clinical scenarios and try to make sure what we do actually makes sense for patients,” he said, citing fractional flow reserve as an example. But some patients don’t fit neatly into the AUC categories and represent “more nuanced” cases, he added: “We really have to make sure that we’re not undertreating patients for whom these ratings might not fully apply.”
Hannan acknowledged that, in terms of inappropriateness, “The expectation is not to drive this rate down to zero. . . . The main thing is to keep track of it and see how it is going.”
Hospitals are reporting that they keep the AUC “pinned up on boards in their cath labs and that they look at this,” Hannan noted. “The goal is not to make sure that they always adhere to it but that if they don’t adhere to it, they know they have a good reason [for that choice]. The results are showing that that’s already happening. So I think things are going well, and I think it is valuable to keep sending them their own data.”
Hannan EL, Samadashvili Z, Cozzens K, et al. Changes in percutaneous coronary interventions deemed "inappropriate" by appropriate use criteria. J Am Coll Cardiol. 2017;69:1234-1242.
Masoudi FA, Curtis JP, Desai NR. PCI appropriateness in New York: If it makes it there, can it make it everywhere? J Am Coll Cardiol. 2017;69:1243-1246.
- Hannan and Kirtane report no relevant conflicts of interest.
- Masoudi has a contract with the American College of Cardiology for his role as chief science officer of the National Cardiovascular Data Registry.