The Future of CAS Reimbursement
Steve Phurrough, MD, MPA, CMS Director of Coverage and Analaysis, explains the process by which new procedures, such as CAS, achieve CMS reimbursement.
There is currently a national noncoverage decision for carotid angioplasty and stenting (CAS). What is CMS’s basis for that stance?
Currently, published clinical evidence does not exist that CAS improves outcomes of patients with carotid artery disease over the outcomes achieved with carotid endarterectomy (CEA). Therefore, we have a policy of noncoverage except when the stent placement and angioplasty occur as part of a Category B investigational device exemption (IDE) trial approved by the FDA.
What is the difference between Category A and Category B IDEs?
In order for a device or procedure to be used in a clinical trial for the purpose of FDA approval, it has to be granted either a Category A or Category B IDE. The Category A distinction indicates that the device is experimental and has not yet received any type of FDA approval, whereas Category B distinction denotes a device with some kind of previous FDA evaluation or approval history. Most devices fall into Category B, except in cases of brand new technology. The current regulation states that CMS provides coverage only for Category B IDE trials.
What will it take for CAS to gain CMS reimbursement beyond IDE-related instances?
When we have a noncoverage decision for a particular procedure, we must first complete another national coverage determination process to remove the noncoverage classification. Any party can request that we do that. When someone does, we want to see what kind of clinical evidence is available that would support this technology and thus warrant opening a national coverage decision. At the present time, there is very little published evidence on the successful use of stenting in carotid obstructed disease. For CMS to open a national coverage decision, we would need the requester to submit some published evidence from completed trials.
Did CMS provide reimbursement for CEA without completed studies published in a peer-reviewed journal?
Coverage for CEA began prior to the national coverage decision process.
Why did CMS choose to increase reimbursement for drug-eluting stents (DESs) prior to their FDA approval, but not CAS procedures?
DESs are a bit different than carotid stents. Stents have been used in coronary arteries for a number of years, and we have covered them for that indication. CMS felt that the addition of a drug to that stent would still fall within the current coverage, so we did not initiate a national coverage decision on DESs. CMS, as it does on a routine basis, looked at the additional costs DES technology would impose on hospitals and decided to increase the reimbursement for that indication when the FDA approved it. The decision was written in such a fashion that if the FDA found problems with the device and did not approve it, the change in payments would not occur. Our goal in this case was to not delay an appropriate increase in payment for new technology. Manufacturers of carotid stents have the opportunity to discuss the same process with CMS as their devices approach FDA approval.
Does CMS generally issue a noncoverage decision for a particular blood vessel that has not previously been approved for a procedure, such as stenting of the renal arteries?
The decision varies. Regarding the carotid artery, there was some information that angioplasty alone had poorer results than CEA, thus we choose to make a noncoverage decision on CAS. Until industry developed devices that prevented blood clots from breaking off and going to the brain, there was some evidence that patients did worse with angioplasty than they did with surgery. We did not address angioplasty in general in other vessels.
Is there any correlation between the FDA approval of a device and CMS decisions regarding reimbursement for that device?
In general, CMS does not provide coverage for anything that does not have some form of FDA approval. In some cases, we might provide coverage for a device that has FDA approval for one indication and not for another, but if it doesn’t have any FDA approval, we would not provide coverage. However, FDA approval does not, in and of itself, result in Medicare coverage. In many cases it does, but not across the board.
Can you name any exceptions to that general rule?
We just had one such exception with implantable defibrillators. The FDA approved a new indication for implantable defibrillators based on the MADIT II trial. Their indication was for a broad population of patients, but CMS covered a narrower population. This illustrates a case in which there are some patients that are approved by the FDA to be treated with a device for which CMS will not provide reimbursement.
What criteria would CMS use to decide that the reimbursement population would be smaller than the FDA approval population?
All of our coverage decisions are based on a concept that is called evidence-based medicine. We define the term in a Federal Register Notice that we published back in 1999, which can be found on our Web site (http://www.cms.gov/coverage/8a1.asp). Essentially, we are looking for published, peer-reviewed literature that reports the results of a comprehensive clinical trial that demonstrates this particular technology benefits the Medicare population and that its benefits outweigh its risks. For someone to ask us to provide coverage, they must show us that type of trial data. Our preference, in almost every case, is that the trial data are published in a peer-reviewed journal, because these publications have physicians who provide an expert review of the articles before they are published.
Are there any particular requirements with respect to that study, such as randomization, that will affect your decision?
When you are looking at medical literature, there are standard conventions for what is considered an adequate trial. We do not exclude any particular type of trial, but we place more weight on those that are more comprehensive. A randomized, controlled trial is considered to be more informative than a nonrandomized trial, and a blinded trial is better than an unblinded trial. Different kinds of biases can enter into a trial, and the fewer biases a trial has, the more you can be comfortable that its results are based on the intervention they were testing and not on chance alone. We do not look solely for randomized trials, but if there are no such trials for us to review, those that are available must be very good.
How does the FDA exemption status affect CMS reimbursement?
The FDA exemption is given to a device that is undergoing a trial based on a protocol that the FDA has approved. When the FDA gives an exemption, it does not mean that the manufacturer can set up its own trial; it has to be an FDA-approved trial. The FDA has approved some carotid stent trials and is in the process of approving several more. Some of these trials involve high-risk patients, whereas others do not. If all of the FDA trials that are either underway or planned are completed, most types of patients will be covered. The only trial that is nearly complete—the SAPPHIRE trial—is only evaluating high-risk patients. If upon completion of that trial the FDA thinks the evidence is sufficient, then its approval will probably only be for procedures involving high-risk patients. Presumably, our coverage decision would also likely only pertain to high-risk patients. We would look to see if the results could be generalized from the studied population to other populations and possibly make a broader coverage decision.
If coverage is restricted to high-risk patients, how does CMS ensure that only such patients are included in the request for reimbursement?
We do not have a mechanism to ensure that hospitals and providers are honest in their billing for every claim that is submitted to us. We have quality-improvement organizations as well as surveys and investigations that are conducted on a retrospective basis. These surveys check to see whether providers and hospitals are following the rules on individual kinds of events. In general, we trust them to be honest. If they are not, and an investigation shows that they are not, then they have committed fraud. Routinely, we assume that hospitals follow the rules we lay out.
If CMS were provided with trial data for one or more completed trials, could CMS expedite the coverage decision if it believes there is significant clinical benefit to the Medicare population?
As previously stated, we prefer data from published clinical trials to have the benefit that peer review offers. We also like to see the actual data to understand results that may not have made it into the published report. Once we have that data, then our National Coverage Decision process takes a minimum of 90 days and perhaps longer if we need to ask assistance from our advisory committee.
How soon after review of such data could a decision be implemented?
Once we decide to make a positive coverage decision, system changes must be made to implement that decision. This process could require 60 to 90 additional days. n
Interview by Craig McChesney, Publisher of Endovascular Today with Steve Phurrough, MD, MPA, Director, Coverage and Analysis Group, Centers for Medicare & Medicaid Services. Dr. Phurrough may be reached at firstname.lastname@example.org.