Inappropriate PCIs Slashed by Half Since First AUC Released, But Key Questions Remain

ORLANDO, FL—Nearly 6 years after publication of the first appropriate use criteria (AUC) for coronary revascularizations, the number of PCI procedures classified as inappropriate has dropped by 50%, new numbers show. Over the same period, the number of procedures performed for nonacute indications in the United States has declined substantially, while the number of PCIs performed for ACS has remained more or less stable. 

The data hint that the majority of US hospitals and operators have acted swiftly to integrate clinical trial results and expert guidelines and distance their practices from the stories of “unnecessary stenting” attracting international media attention in the mid-to-late 2000s.

Nihar Desai, MD, MPH, of Yale-New Haven Hospital (New Haven, CT), presented the findings culled from the National Cardiovascular Data Registry (NCDR) CathPCI Registry here at the American Heart Association 2015 Scientific Sessions. The study was published simultaneously in the Journal of the American Medical Association. 

“The overall message is that there are signs of great progress,” Desai told TCTMD. “What this shows is professional responsibility: there was a concern identified, the profession reacted very strongly, it made the investment to capture the data, and made a commitment to address this. And the result is, we’ve seen a huge reduction in inappropriate PCI and signs of important progress, although there is still room for improvements.”

The findings indicate that clinicians are doing a better job of limiting nonacute PCI procedures to those patients most likely to benefit from revascularization, although Desai acknowledged that there were some hospitals whose rates of inappropriate procedures did not shift over the study period. There’s also the possibility that the data reflect some improvements in documentation on the part of physicians but also, potentially, some “intentional upcoding” of subjective data elements, such as symptom severity, he said.

The Path to Appropriateness

For the study, Desai and colleagues reviewed trends in patient selection, changes in PCI appropriateness, and hospital variation in numbers/proportion of inappropriate PCI using data from 766 hospitals participating in the registry.  Procedures were categorized as acute (ACS) cases or “nonmappable,” which captured all procedures for which invasive testing was not done or was not available. Any procedures that did not fall into those 2 categories were deemed nonacute. Nonacute and acute cases were then assigned a rating of appropriate, uncertain, or inappropriate based on the updated 2012 AUC.

In all, 2.7 million PCI procedures were included in the analysis. Acute PCI volumes ranged from 377,540 to 374,543 over the study period. By contrast, nonacute PCIs dropped from 89,704 in 2010 to 59,375 in 2014. At the same time, the proportion of nonacute PCIs classified as inappropriate decreased from 26.2% to 13.3%. In absolute numbers, those figures translate into almost 22,000 patients treated “inappropriately” in 2010 compared with just under 8,000 in 2014.

Among other key findings between 2010 and 2014 in the nonacute procedure patients:

•       Performance of FFR increased from 8.1% to 30.8%
•       Use of noninvasive testing increased from 64.6% to 72.5%
•       Findings of high-risk characteristics on noninvasive testing rose from 22.2% to 33.2%
•       Presence of CCS grade III/IV angina rose from 15.8% to 38.4%
•       Use of at least 2 antianginal medications rose from  22.3% to 35.1%
•       Presence of multivessel CAD rose modestly from 43.7% to 47.5%

Drivers of Appropriateness

A number of factors lie behind these improvements, Desai and colleagues say. Starting in 2011, the CathPCI registry started giving hospitals feedback on their PCI appropriateness as well as that of other hospitals. PCI appropriateness was also included in the American Board of Internal Medicine’s “Choosing Wisely” campaign. Several large, high-profile randomized trials including COURAGE and BARI 2D are also credited with spurring a slow-down in PCI procedures. At the same time, a number of high-volume operators and hospitals made headlines for implanting stents deemed unnecessary, attracting a number of Department of Justice complaints and class action lawsuits.

William Boden, MD, of Albany Medical College (Albany, NY), commenting on the study for TCTMD, agreed that “certainly COURAGE played a role,” along with several other trials. Together, those helped convince interventionalists that the clear benefits of PCI in the acute setting did not extend to more stable patients.

“What we’re seeing here are some of the downstream effects, and more patients being treated with guideline-directed, intensive medical therapy.”

The AUC, Boden continued, “have set the standard” in that physicians and cardiologists “are being much more prudent in selection of patients” and “much more vigilant about making sure we are not overusing or abusing PCI.”

Different Patients or Stubborn Operators?

While most hospitals have reduced their rate of inappropriate procedures—some drastically, from as high as 70% to less than 10%—others have not seen their rates markedly decline. Among the best-performing hospitals, less than 6% of nonacute PCIs were inappropriate, while at the worst-performing hospitals, nearly 1 in 4 nonacute PCIs met the criteria for inappropriate.

Gregory Dehmer, MD, Scott & White Healthcare (Temple, TX), who co-authored the 2009 and 2012 AUC documents, called these numbers “exactly what one would have hoped the AUC would find.” Namely, that some hospitals and operators have adjusted their behavior, while others who have not can be identified and informed. Moreover, there may be factors not covered by the AUC that might explain some higher rates of inappropriate PCIs. Centers with high rates of solid organ transplantation, for example, may have surgeons who “want things fixed” before they’ll consider a patient for a kidney or liver transplant.

“So do the different rates of inappropriate PCIs reflect a genuine difference in patient populations, or is it because we have a few stubborn practitioners, who aren’t really in this ballgame yet and need to be?” Dehmer asked. “This provides that benchmark data to allow facilities to say well, we are at the opposite end of the curve than we’d like to be.”

Manesh Patel, MD, of Duke University Medical Center (Durham, NC), a co-author on the current study as well as the 2009 and 2012 AUC documents, told TCTMD that the lack of improvement at some hospitals isn’t entirely unexpected.

“This shows that there are continued opportunities to study why hospitals have changed and why others haven’t changed as much,” he said. Key questions to be answered include how hospitals first react to and apply the data they are provided from the registry. 

When Money Talks

Asked whether financial motives may explain the lack of decline in inappropriate procedures at some hospitals, Desai hedged his bets. “We would be naïve if we were to say that the financial incentives that contributed to overuse in the first place aren’t still a factor at some institutions, so there may be some overuse that persists because of the financial benefits. That’s in part what’s led to the uptake of AUC by pay-for-performance strategies and payors, something the AUC “were never intended to be used for and, in my opinion, is a step too far.” 

This might also work in the other direction. Dehmer, questioned as to whether the drop in nonacute procedures due to “inappropriateness” is hitting the pocketbooks of interventionalists, noted that 76% of PCIs included in the analysis were in the acute setting, so the decline has only affected the remaining one-quarter of procedures.

“I don’t think that’s going to make an enormous difference, financially, for the individual interventional cardiologist,” he said. “But even if it did, if it leads to better decision-making in patients, it’s still the right thing to do.”

Hard Outcome Data Lacking

Also discussing the study results with TCTMD, Gregg Stone, MD, of Columbia University Medical Center (New York, NY), said the study “convincingly” shows that the country as a whole is moving in the direction of more appropriate use. The problem for Stone is that no prospective study has demonstrated that overall health and outcomes are actually improved when AUC are followed. The criteria are loosely based on guidelines, themselves pulled from the best evidence available at the time. In the case of the AUC, this includes a number of large, randomized trials that no longer reflect current care. COURAGE, for example, used BMS, while BARI 2D and FREEDOM used first-generation stents that are no longer on the market, he noted. 

“I do believe there has been inappropriate PCI being done and hopefully this is decreasing. What concerns me is that there may be underuse of PCI and bypass in appropriate patients, whether or not that’s defined by AUC,” Stone said. “In other words, there may also be an increase in appropriate PCI that is not being done, and that’s not being captured in these data.”

An accompanying editorial by Robert Harrington, MD, of Stanford University (Stanford, CA), also points to the fact that the AUC are based on retrospective data, which limits their applicability. “What is needed is a national system that allows immediate real-time decision support for clinical activities fully integrated with clinical research capabilities that use constantly accumulating data and sophisticated data analytics, including randomization where appropriate,” he writes.

Desai, asked about outcomes data, acknowledged that it was a “really important” issue not broached in the current study. A retrospective analysis of COURAGE, he noted, addressed this question in that trial and supported the utility of AUC.

Sources: 
1. Desai NR, Bradley SM, Parzynski CS, et al. Appropriate use criteria for coronary revascularization and trends in utilization, patient selection, and appropriateness of percutaneous coronary intervention. JAMA. 2015;Epub ahead of print. 
2. Harrington RA. Appropriate use criteria for coronary revascularization and the learning health system: a good start [editorial]. JAMA. 2015;Epub ahead of print. 

Related Stories:

• Washington State Sees Increase in Appropriate PCI After AUC Introduction
• Revascularization May Benefit Stable CAD Patients With Unclassifiable Appropriateness Scores
• New Revascularization Appropriateness Criteria Feature PCI Upgrades