This article is more than 5 years old. Information may no longer be current.
2014 Update on PCI Without On-Site Surgery: A Consensus Made in Good Conscience
The writing committee of the “SCAI/ACC/AHA Expert Consensus Document: 2014 Update on Percutaneous Coronary Intervention Without On-Site Surgical Backup” is to be commended for its efforts to formulate professional society recommendations for the performance of PCI in facilities without cardiothoracic surgical backup on-site (no-SOS) with the intent of assuring quality patient care. Although the recommendations are sound and evidence based, they stand in stark contrast to the reality of current clinical practice. The writing committee reaffirmed prior recommendations regarding institutional and operator annual PCI volumes, but fell short on guidance regarding definition(s) of “satisfactory outcomes” following PCI, particularly at lower-volume centers.
A few salient features of the consensus document may be summarized as follows.
PCI Volume and Clinical Outcomes
First, the association between a facility’s annual PCI procedural volumes and meaningful clinical outcomes, particularly mortality following PCI, was reaffirmed. The writing committee has resisted downward pressure from hospital associations and some society constituents regarding annual institutional PCI volume credentialing numbers and has adhered to the recommendation of 200 total PCIs/year. This is consistent with the 2013 ACCF/AHA/SCAI update on clinical competence in coronary artery interventional procedures, which noted “a signal exists suggesting that an institutional volume threshold <200 PCIs/year appears to be consistently associated with worse outcomes across various studies.” Similarly, the recommended annual institutional volume of >36 primary PCIs/year was reaffirmed and is supported by recent data that demonstrate a 20% relative increase in hospital mortality following primary PCI done at institutions performing <36 primary PCIs/year (lowest tertile) when compared with those performing >60 primary PCIs/year (highest tertile, HR=1.2; 95% CI, 1.08-1.33; P=.001).
Dean J.
Kereiakes
Timothy D.
Henry
Although consistent with the 2013 clinical competency statement, evidence supporting the recommended individual operator minimum annual PCI volumes for credentialing (≥50 total PCIs and ≥11 primary PCIs/year) as shown in Table 4 of the new document is less clear. For example, the New York State Database suggests that the relationship between operator primary PCI annual volume and hospital mortality is both inverse and statistically significant. Furthermore, among 345,526 PCIs reported to ACC/NCDR by 3,649 physicians from 543 cath PCI participating hospitals, in-hospital mortality was increased for operators who performed <75 PCIs/year for all PCI (OR=1.14; 95% CI, 1.05-1.24; P<.01), primary PCI for STEMI/shock (OR=1.1; 95% CI, 1-1.21; P=.06) and particularly for elective PCI (no STEMI/shock, OR=1.27; 95% CI, 1.11-1.45; P<.001).
Considering the declining number of PCI procedures performed annually in the United States, the reduction in disease prevalence for both CAD and acute MI and the need for cost reduction while maintaining or increasing quality, a more reasonable approach might be to promote regionalization with fewer institutions and encourage operators to maintain higher annual volume thresholds rather than lowering established operator volume thresholds somewhat arbitrarily. This is important as 26% of all centers reporting to ACC/NCDR performed <200 total PCIs/year and 38% perform <36 primary PCIs/year. Of note, approximately 83% of those centers performing <200 PCIs/year have no-SOS. Finally, recent data suggest that the performance of inappropriate PCI using current appropriate use criteria may be increased in lower-volume centers.
Institutional Thresholds
Second, the writing committee reaffirms the statement that it “recommends that an institution without on-site surgery with a volume fewer than 200 PCIs annually, unless in a region underserved because of geography, should strongly consider whether or not it should continue to offer this service.” Importantly, the new document further develops the definition of geographic isolation based on a transport time of > vs. ≤30 minutes and states “the development of PCI facilities within a 30-minute emergency transfer time to an established facility is strongly discouraged” as “multiple low-volume and partial-service PCI centers within a geographic area diffuse PCI expertise, increase costs for the overall health system and have not been shown to improve access.”
Although the writing committee recommends that “any (low volume <200 PCIs/year) laboratory that cannot maintain satisfactory outcomes be closed,” it falls short on recommendations for the definition of “satisfactory outcomes.” Instead, it states, “the writing committee recommends that these (definitions) be defined by each PCI center, including those with on-site surgery, as part of their quality review process, using national benchmark data. Programs failing to meet established criteria for satisfactory performance for two consecutive quarters must undertake efforts to improve their performance, engaging outside experts if necessary. Failure to improve quality metrics should lead to program closure regardless of location.” Notwithstanding the limitations of site self-reported data to ACC/NCDR by persons with widely variable training, understanding and motivation without systematic monitoring, adjudication or audit, the concept of “self-prescribed” definition(s) for satisfactory outcomes to be established by the individual reporting centers is curious.
Furthermore, unless a composite of clinical outcomes is used as the quality metric to be monitored, risk-adjusted mortality alone becomes increasingly problematic as PCI volumes decrease and is inadequate to provide efficient differentiation. Indeed, as the new document cites, “With facility PCI volumes decreasing, it becomes increasingly difficult to determine whether there are significant differences in the data reports from year to year” — let alone over a “two consecutive quarter” time frame. Indeed, for institutions performing 200 PCIs/year, it will take almost 4 years, and for institutions performing 150 PCI/year, it will take 5 years of continuous monitoring to ascertain that a twofold increase in risk-adjusted mortality (from 1% to 2%) is statistically significant and not attributable to chance. Frankly, we need better and more specific guidance. Low-volume centers, by virtue of the wide CIs surrounding the low-incidence occurrence of adverse outcomes such as in-hospital death, should not be given the capacity to hurt unwitting patients under the guise of statistical probability. Hopefully, in the near future, more specific recommendations will be forthcoming as the current situation in practice is concerning.
The Reality of Clinical Practice
In the state of Ohio, almost 90% of the adult population lives within 30 minutes ground transportation of a hospital that performs PCI. Nevertheless, in 2012, six of the 10 Ohio C-PORT (no-SOS) facilities performed <150 PCIs/year and five of 10 were within 15 minutes (or less) ground transportation from a cardiac surgical facility that performed PCI. At these annual PCI volume levels, it would take more than 5 years of continuous monitoring of risk-adjusted mortality to determine if a twofold increase in mortality by an underperforming low-volume institution was actually statistically significant and not attributable to chance. Ohio, like many states in the United States, has considerable geographic reduplication of PCI services. Furthermore, the state has no mechanism to enforce volume or geographic credentialing recommendations. In the absence of established, benchmarked composite clinical outcomes, quality monitoring of these low-volume facilities becomes virtually futile. Although the 2014 consensus document represents a step forward, more concrete recommendations are required to assure quality care and patient safety. Finally, guideline recommendations without a process or hope for implementation are difficult to make in good conscience.