Comprehensive Reimbursement Solutions
January 25, 2013
Louis Jacques, MD
Director, Coverage and Analysis Group
Centers for Medicare and Medicaid Services
7500 Security Boulevard S3-02-01
Baltimore, MD 21244
RE: Draft Guidance for the Public, Industry, and CMS Staff Coverage with Evidence Development in the context of coverage decisions
Dear Dr. Jacques:
I would like to thank the Coverage and Analysis Group for the time and energy put into the Draft Guidance on Coverage with Evidence Development (CED). As a consultant to the device industry, I am in complete agreement with the related goals of driving innovation and reimbursing for promising new technologies stated in the National Bioeconomy Blueprint. However, based on 25 years of medical device reimbursement and evidence creation experience, including first-hand knowledge of CED, I am concerned that a number of the provisions outlined in the Draft Guidance are not in keeping with those goals.
The Guidance states that Coverage with Appropriateness Determination (CAD) is no longer considered a form of CED. Between it and Coverage with Study Participation (CSP) CAD resulted in broader coverage. While I am not advocating the return of CAD, eliminating it in favor of the more narrowly construed CSP does not support the goal of using CED more widely.
Because CSP results in a fraction of national coverage, its use is a double-edged sword. By responding to a request for national coverage with coverage under a clinical trial, given the cost and logistical barriers to utilization that that entails, it can deprive a large number of beneficiaries’ access to leading-edge medical technologies. It can also stifle rather than drive innovation. The tipping point between the two outcomes is whether the CED requirement truly expands coverage to areas not fully evidenced. If instead, it’s applied to areas where evidence is actually sufficient, it harms beneficiaries and reduces the incentive to innovate.
To be used more effectively CED should continue to be used sparingly. Stepping back to look at the problem we are attempting to solve, how often do manufacturers actually apply for a national coverage determination with no adequate evidence whatsoever? And if so, couldn’t those few be told they’re not ready earlier in the process? In order for the CED construct to have validity, what we should typically see is national coverage requests resulting in some national coverage, with CED offered as an expansion of that coverage around the edges. Reflecting both the adequacy and softness of the evidence would demonstrate the correct application of the concept. In order for this to occur, the involvement of an advisory body like MEDCAC should be required before a national coverage determination, based solely on CSP, can be issued. Using CED sparingly and involving MEDCAC only when exclusively CSP coverage is under consideration addresses CMS’ concern regarding MEDCAC resources.
CED should only be done as part of a NCD, and should not support evidence creep. Coverage should continue to be based on evidence of clinical effectiveness, including adequate representation of the Medicare population. CED should not be used to support comparative effectiveness research which, although important, is a post-coverage research question. It must also follow the hierarchy of evidence. CSP evidence lower on the hierarchy should not be used to restrict coverage based on higher levelresearch. For example, registry data should not subsequently be used to limit coverage determined by a randomized controlled trial.
For CED to be performed, it should be paid for. This already occurs under the 1995 Interagency Agreement, if an IDE is required. However, for the Clinical Trial NCD that is not the case. Here the investigational item or service itself is excluded from payment. This should be changed for studies performed under CED. Being given a CSP requirement must at least be equivalent to an IDE Category B determination that the basic questions with regard to safety have already been addressed. It should be funded as such, even when an IDE is not required.
Thank you for the opportunity to comment.
Bob Thompson, MPA, MS