Abstract
Medicare coverage for transcatheter aortic valve replacement (TAVR) in the United States (US) is governed by the 2012 National Coverage Determination (NCD 20.32), which enshrined minimum numbers of TAVR, surgical aortic valve replacement, and percutaneous coronary intervention that centers must perform to begin or maintain TAVR programs. In July 2018, the Centers for Medicare and Medicaid Services (CMS) convened a meeting of the Medicare Evidence Development & Coverage Advisory Committee (MEDCAC) to review the evidence for setting minimum procedure volume requirements and to evaluate the impact of such requirements on access to care. In this paper, we summarize the MEDCAC panel deliberations, the evidence presented to the panel, and how the panel members voted. CMS is expected to publish a draft decision in March 2019 that may reshape the TAVR landscape in the US for years to come.
Highlights
- •
MEDCAC reviewed evidence for setting minimum volume requirements for TAVR centers.
- •
We summarize deliberations, evidence presented, and how panel members voted.
- •
CMS’s expected decision in 2019 may reshape US TAVR landscape for years to come.
1
Introduction
On July 24, 2018, the Centers for Medicare and Medicaid Services (CMS) convened a meeting of the Medicare Evidence Development & Coverage Advisory Committee (MEDCAC) to review the 2012 National Coverage Determination (NCD) for Transcatheter Aortic Valve Replacement (TAVR) [ ] in response to a formal request for reconsideration by the Providence Saint John’s Health Center, CA. The 2012 NCD established minimum required surgical aortic valve replacement (SAVR) and percutaneous coronary intervention (PCI) procedure volumes for centers to begin a TAVR program, and thereafter minimum required TAVR volume to maintain a program. The basis for the request for reconsideration was essentially twofold: first, to question the principle of a volume-quality relationship for TAVR and second, to contend that minimum volume requirements limit access to care.
2
Does higher volume equate to higher quality?
Ahead of the meeting, a joint AATS/ACC/SCAI/STS society statement was published, setting out new recommendations for minimum procedure volumes for new and current TAVR sites [ ]. The authors recommended increasing the TAVR volume requirement for sites to maintain a TAVR program from 20 cases per year, as established in the 2012 NCD, to 50 cases per year. Dr. Joseph Bavaria, Penn Heart and Vascular Center, Philadelphia, PA, and Dr. Carl Tommaso, NorthShore Medical Group, Chicago, IL, presented on behalf of the Society of Thoracic Surgeons (STS), American College of Cardiology (ACC), Society for Cardiovascular Angiography and Interventions (SCAI), and the American Association for Thoracic Surgery (AATS). Presenting unpublished data from the STS/ACC Transcatheter Valve Therapy Registry (TVT), Drs. Bavaria and Tommaso argued that low-volume centers had higher mean 30-day mortality compared with high-volume centers and that, consequently, minimum procedure volumes should be required to maintain TAVR programs ( Fig. 1 ). They advocated maintaining the NCD mandate for participation in a single national registry (i.e., TVT) to serve as both a research and quality audit tool. In addition, the joint society statement recommended that TAVR operators obtain interventional cardiology board certification, although at present there is no sub-specialty structural interventional cardiology board certification. Dr. Tommaso suggested that new TAVR operators should be exposed to 100 transfemoral TAVR procedures, including 50 as first operator. He argued that there are now a sufficient number of structural fellowship training positions to train the operators of tomorrow. Other speakers highlighted that it is increasingly difficult to be exposed to the complications of TAVR as the procedure becomes safer and that only volume can ensure that new operators are equipped with the skills required to manage complications.
Dr. Martin Leon, Columbia University, New York, NY, presented conflicting data from MEDPAR ( Fig. 2 ) showing that there was no difference in in-hospital mortality between high- and low-volume centers using the latest-generation TAVR devices. He commented on the TVT registry data of the early US experience (2011–2015) that reported a 1% adjusted mortality difference between high- and low-volume sites [ ]. One reason could be that 30% of these cases were transapical, and many with first-generation devices. In contrast, in the transfemoral subgroup, there was no association between site volume and risk-adjusted mortality. Looking at new-generation devices only, no volume-outcome relationship was found in analyses by the two commercial TAVR manufacturers (Edwards Lifesciences and Medtronic). The 2012 NCD and 2018 joint consensus statement also require a minimal volume for PCI and SAVR for centers starting or maintaining a TAVR program. No evidence was presented to suggest that increasing volume of either PCI or SAVR improved TAVR outcomes. A MEDPAR analysis of TAVR cases between 2012 and 2016 failed to show a relationship between prior or current PCI or SAVR volume and TAVR in-hospital mortality. This finding was backed up by industry representatives, who showed their own analysis of the MEDPAR data showing no evidence of a volume-outcome relationship for TAVR. Nonetheless, many of the speakers and panel members expressed personal beliefs that SAVR and PCI skills were applicable and indeed necessary for TAVR.
2
Does higher volume equate to higher quality?
Ahead of the meeting, a joint AATS/ACC/SCAI/STS society statement was published, setting out new recommendations for minimum procedure volumes for new and current TAVR sites [ ]. The authors recommended increasing the TAVR volume requirement for sites to maintain a TAVR program from 20 cases per year, as established in the 2012 NCD, to 50 cases per year. Dr. Joseph Bavaria, Penn Heart and Vascular Center, Philadelphia, PA, and Dr. Carl Tommaso, NorthShore Medical Group, Chicago, IL, presented on behalf of the Society of Thoracic Surgeons (STS), American College of Cardiology (ACC), Society for Cardiovascular Angiography and Interventions (SCAI), and the American Association for Thoracic Surgery (AATS). Presenting unpublished data from the STS/ACC Transcatheter Valve Therapy Registry (TVT), Drs. Bavaria and Tommaso argued that low-volume centers had higher mean 30-day mortality compared with high-volume centers and that, consequently, minimum procedure volumes should be required to maintain TAVR programs ( Fig. 1 ). They advocated maintaining the NCD mandate for participation in a single national registry (i.e., TVT) to serve as both a research and quality audit tool. In addition, the joint society statement recommended that TAVR operators obtain interventional cardiology board certification, although at present there is no sub-specialty structural interventional cardiology board certification. Dr. Tommaso suggested that new TAVR operators should be exposed to 100 transfemoral TAVR procedures, including 50 as first operator. He argued that there are now a sufficient number of structural fellowship training positions to train the operators of tomorrow. Other speakers highlighted that it is increasingly difficult to be exposed to the complications of TAVR as the procedure becomes safer and that only volume can ensure that new operators are equipped with the skills required to manage complications.
Dr. Martin Leon, Columbia University, New York, NY, presented conflicting data from MEDPAR ( Fig. 2 ) showing that there was no difference in in-hospital mortality between high- and low-volume centers using the latest-generation TAVR devices. He commented on the TVT registry data of the early US experience (2011–2015) that reported a 1% adjusted mortality difference between high- and low-volume sites [ ]. One reason could be that 30% of these cases were transapical, and many with first-generation devices. In contrast, in the transfemoral subgroup, there was no association between site volume and risk-adjusted mortality. Looking at new-generation devices only, no volume-outcome relationship was found in analyses by the two commercial TAVR manufacturers (Edwards Lifesciences and Medtronic). The 2012 NCD and 2018 joint consensus statement also require a minimal volume for PCI and SAVR for centers starting or maintaining a TAVR program. No evidence was presented to suggest that increasing volume of either PCI or SAVR improved TAVR outcomes. A MEDPAR analysis of TAVR cases between 2012 and 2016 failed to show a relationship between prior or current PCI or SAVR volume and TAVR in-hospital mortality. This finding was backed up by industry representatives, who showed their own analysis of the MEDPAR data showing no evidence of a volume-outcome relationship for TAVR. Nonetheless, many of the speakers and panel members expressed personal beliefs that SAVR and PCI skills were applicable and indeed necessary for TAVR.
3
Is a minimum volume required to measure quality?
There was general consensus that quality of care is more important than volume of procedures. Data presented showed many low-volume sites with excellent outcomes. These data were challenged on the grounds of high chances of random sampling variability due to the low volume and low predicted event rate of mortality with TAVR of around 3%. In other words, with a small sample size and a low event rate, zero mortality is meaningless.
The authors of the consensus document presented unpublished data from the TVT registry showing higher variability in mortality at centers with an annual TAVR volume of <50 cases ( Fig. 1 ). These data demonstrated flattening of the variability in mortality once centers performing fewer than 100 cases per year were excluded. Non-risk adjusted 30-day mortality demonstrated little difference in the median mortality rate between low- and high-volume centers. However, there was a significant difference between mean and median mortality rates in low-volume centers, and convergence of these values at high-volume centers, illustrating the problems of statistical analyses in low-volume centers.
Dr. David Shahian, Harvard Medical School, Boston, MA, representing the STS Council on Quality Research, highlighted the difficulty in measuring quality with low volumes and that the signal-to-noise ratio was highly dependent on sample size. He conjectured that a procedure like TAVR with an expected mortality of approximately 3% would require a sample size >100 to detect a doubling of the mortality rate with 80% power. Given the challenges in collecting good quality statistical data on performance of sites with low volumes, the authors of the joint consensus statement proposed a minimum of 50 TAVR procedures per year be required to adequately assess outcome. Other options for improving the statistical analysis were suggested, including measuring composite outcomes to increase the “event rate,” to combine outcomes of TAVR and SAVR so patients understand how well their disease is treated at a center, or to measure outcomes over time periods >1 year.
4
Do minimum volume requirements restrict access to care?
An unintentional consequence of minimum volume requirements is that they may prevent patients, particularly from underserved populations, from accessing treatment. It was suggested that elderly TAVR patients are less likely to travel far from home, perhaps partially explaining the poor penetrance of TAVR in geographically or socially isolated populations, particularly among African-Americans, Hispanics, and women. In a survey of patients over the age of 65 years, 75% would prefer local care even when presented with a 1% increase in mortality compared to higher-volume centers farther afield [ ].
In 2016, there were an estimated 250,000–350,000 patients in the US with symptomatic severe aortic stenosis (both diagnosed and undiagnosed) based on the Optum cohort of patients diagnosed between 2014 and 2015. However, aortic stenosis is still undertreated, with a 52% 2-year mortality, which reduces to 17% when treated with aortic valve replacement. The incidence of severe symptomatic aortic stenosis, estimated from the American Community Service 5-year census data, divided by the number of aortic valve replacements (TAVR or SAVR) performed according to the MEDPAR database, showed that every state had a treatment penetrance of <40%. No evidence was presented to show increase in treatment penetrance with increase in TAVR centers.
In 2017 in the US, there were 1872 centers performing PCI, 1103 performing SAVR and 540 performing TAVR. Although the authors of the consensus document denied seeking to close any of the active TAVR centers in the United States, the future of these centers will be uncertain if CMS decides to base reimbursement on these volume recommendations. In 2017, 204 TAVR sites were performing fewer than 50 TAVRs a year. Dr. Leon inferred that 39% of TAVR centers could fail to meet at least one of the proposed volume requirements and that the impact on the TAVR landscape in the US could be substantial ( Fig. 3 ). Furthermore, of the >500 centers that currently perform SAVR but not TAVR, <25% would qualify to start TAVR programs based on the proposed volume requirements. SAVR mortality in SAVR-only centers is currently higher than in SAVR/TAVR centers (6.7% vs. 4.4%, source: MEDPAR), suggesting that SAVR-only centers are treating higher-risk patients surgically, whereas TAVR centers would triage these same patients to TAVR. Dr. Leon argued that the goal should be to reduce the number of SAVR-only centers and that the harm of limiting access to TAVR to only hospitals that meet volume requirements outweighed the benefits of meeting procedural (i.e., SAVR, PCI, and TAVR) volume requirements to begin a TAVR program.
Dr. Aaron Horne Jr., Methodist Health System, Dallas, TX, representing the Association of Black Cardiologists, spoke passionately regarding current disparities in access to TAVR for minorities in the US and the risk of widening disparities in the future if the minimum volume requirements for institutions are not relaxed. Dr. Horne argued that not only are African-Americans diagnosed less frequently with aortic stenosis, but they are also less likely to be referred for treatment. The evidence suggests that the prevalence of aortic stenosis is similar among African-Americans and Caucasians [ ]. African-Americans compose approximately 13.5% of the total US population, yet represent just 3.8% of all patients treated with TAVR in the TVT registry [ ]. Citing TVT data, Dr. Horne argued that despite the good outcomes of African-Americans undergoing TAVR [ ], the rate of growth in TAVR cases had not appreciably increased in African-Americans compared to Caucasians ( Fig. 4 ). Differences in access to care, socioeconomic disadvantage, and a strong reluctance to travel outside their support network were identified as potential reasons. Dr. Horne argued that increasing the number of TAVR centers by decreasing minimum volume requirements would help improve access for African-Americans, but he recognized that increasing awareness and education was also important. He proposed a national campaign to address TAVR disparities. Representatives of the major societies welcomed a partnership with the Association of Black Cardiologists.
