View Public Comments for CED Public Solicitation

January 20, 2012

Louis Jacques, M.D.
Director, Coverage and Analysis Group
Centers for Medicare & Medicaid Services
Office of Clinical Standards and Quality
7500 Security Blvd.
Mail stop: C1-09-06
Baltimore, MD 21244

RE: Request for Comments on CMS' Coverage with Evidence Development Policy

Dear Dr. Jacques:

On behalf of Medtronic, Inc., I am pleased to respond to the Centers for Medicare & Medicaid Services' (CMS') request for public comment on its coverage with evidence development (CED) policy.¹ Medtronic is one of the world's leading medical technology companies, specializing in implantable and interventional therapies that alleviate pain, restore health, and extend life. We are committed to the continual research and development necessary to produce high-quality products and to support innovative therapies that improve patient and health care system outcomes.

We appreciate the opportunity to comment on the CED policy, building on our past work with CMS on the national coverage process and on coverage determinations for many Medtronic therapies, including CED for implantable cardioverter defibrillators (ICDs). We have appended our 2005 comment letter to the agency's initial solicitation for public input on factors to consider when making a determination of CED.² Medtronic appreciates the value of evidence-based approaches to generate real-world evidence and create data sources to track real-world outcomes of interest to Medicare and its beneficiaries. We have and continue to engage with CMS and the FDA to generate these data and continue to make significant investments on our own to refine the use of our products, better define appropriate indications, and expand the evidence base for our therapies.

Medtronic supports CMS' overarching goal to ensure that Medicare beneficiaries receive appropriate, high-quality health care, including access to life-saving and life-enhancing medical advancements. Medtronic also supports CMS' goal to improve physician and patient decisionmaking. In that regard, it is up to all of us as the stewards of limited clinical research dollars to ensure that the data generated through the CED approach would create meaningful, actionable clinical evidence and would not duplicate existing public and private efforts outside the agency. It is also important that CED be conducted in a manner that is efficient and considers the financial and administrative burdens placed on all parties involved.

To that end, we put forward the following recommendations for the agency's consideration:

1. CMS should publicly issue an evaluation of the successes and challenges of its existing CED policies and host a public meeting to engage stakeholders in ways to improve and mature CED.

2. CMS should limit the use of CED as an outcome of the NCD process. More specifically, CED should be applied rarely in NCDs and only in circumstances where the alternative is non-coverage. In these instances, CED should be used to generate specific evidence that will address, in a timely fashion, the open questions related to whether the technology is reasonable and necessary. These questions should be identified prior to the implementation of CED.

3. In light of past challenges with CED, Medtronic recommends that CMS consider the following refinements to the existing NCD process to ensure appropriate implementation of CED:

   1. CMS' intentions of considering CED should be noted as early as the initiation of the national coverage analysis (NCA).
   2. A value of information (VOI) analysis or at minimum an informal analysis of the return on investment from CED should be conducted before the draft decision; the results of this analysis should be included in the draft guidance.
   3. The draft NCD should propose the membership of a CED implementation steering committee with representation from all appropriate stakeholders and state an interim coverage policy to ensure beneficiary access to the therapy before the data collection mechanism is operational.
   4. The final NCD should name the final steering committee, interim coverage policy, and provide a clear timeline for next steps for appropriate CED implementation.
   5. While the billing and coding instructions are being written, the steering committee would work with CMS to ensure all questions are clearly answered for appropriate implementation of CED.

4. CED should not be issued by local Medicare contractors.

Further detail on these recommendations is provided below. We appreciate your consideration of these comments.

Medtronic appreciates the agency's efforts of considering internal lessons learned to better align CED with the rapidly evolving changes in our healthcare system. We believe that a formal evaluation of the lessons learned from all prior applications of CED is an important step to refine the overall effort and prove the value of additional data collection. We recommend that the agency publicly post its lessons learned and incorporate these lessons in a revised policy framework for CED. Moreover, we recommend that CMS establish a permanent mechanism to evaluate the effectiveness of CED in providing additional evidence to bolster physician and patient decision-making.

For example, as recommended by stakeholders in response to the 2005 draft CED guidance, each year in CMS' report to Congress on the NCD process, the agency's Office of Research, Development, and Information could conduct an internal evaluation of CED initiatives. Alternatively, CMS could contract with an independent entity to review CED policies on an ongoing basis to assess their implementation. The agency should monitor and evaluate beneficiary access to care after the implementation of all NCDs that issued CED to verify that these decisions have expanded access to these therapies and that the data collection efforts have minimized the burden and have resulted in useful and actionable data. CMS should publicly post these reports evaluating the CED initiatives at least every three years.

In considering lessons learned, Medtronic believes it is helpful to draw from three illustrative specific cases of CED to improve future efforts: ICD, left ventricular assist device (LVAD), and cochlear implantation.

In 2005, CMS implemented an NCD on implantable automatic defibrillators, which expanded coverage for ICDs by requiring the systematic collection of implant data via an ICD registry.³ While this registry has served to aggregate data on ICD implantations, several challenges have been observed in its implementation. Particularly, even after the establishment of the registry, there was no formal agreement between CMS and the National ICD Registry Working Group on a protocol with well-defined research questions or a timeline for the duration of data collection. As a result, the study population was not targeted and the collected data was not necessarily relevant to the CED policy.⁴

At the start of the ICD CED registry, it was acknowledged that certain data elements were absent and there were no specific plans on how to obtain the necessary follow-up data for ICD firing, complications that occur after discharge, such as infection and lead dislodgement, and long-term survival data. This is a problematic approach to robust clinical research. Mayo Clinic clinician Dr. Stephen Hammill noted that the ability of the original ICD Registry to address specific areas identified as research gaps in the coverage decision remained uncertain: "some were skeptical whether data from this registry would ever be useful in refining coverage decisions; with little/no history of using observational data to affect coverage and with no control group, some view using such data to change coverage as unlikely, especially after nearly a half-decade of established use."⁵ In April 2010, version 2.0 of the registry was launched to meet many of these limitations.

Registries provide a useful vehicle for real world utilization and safety surveillance observations. But their role and ability to efficiently and effectively answer CED questions have not yet been demonstrated in the example of ICDs. There have been no publications or advancement of the CED discussion based on the ICD CED registry data. There are no specific terms or dates that provide process transparency to the use of the registry to answer definitively the CED questions. We are now seven years in to answering the CED question but there is no clear path forward in terms of process and decision making.

We note that even when a specific CED requirement is applied, alternative studies may be conducted that answer questions intended to be addressed by the CED requirement. This has been the case with ICDs, where Medtronic initiated the OMNI study to prove further the outcome benefit of ICD in relation to the CED questions. The results of OMNI have been shared with CMS and publications are pending. We believe CMS should carefully consider alternative public and private study resources such as OMNI when addressing CED questions, and should take into consideration results of such studies in determining whether and when a CED requirement should conclude.

In 2007, CMS released a reconsideration of its final decision on ventricular assist devices (VAD) as destination therapy, which established INTERMACS as the registry that satisfies CMS' reporting requirements for coverage.⁶ The INTERMACS case demonstrates that establishing a sustainable and fairly distributed funding program is a difficult endeavor that needs to be carefully considered within a clearly defined process. Established in 2005, the registry was initially funded by the NIH and the U.S. Department of Health and Human Services (DHHS). In this case, the burden of third parties (e.g., medical technology companies and hospitals) to provide financial support has steadily increased. Starting in 2010, the NIH contract renewal requires INTERMACS to become mostly self-funded by 2015. As of April 2011, the registry is now funded 38% by NIH, 38% by hospitals, and 22% by industry.⁷ As a result, each hospital participating in INTERMACS will pay $10,000 per year and industry will pay an increasing rate, starting at $500 per patient enrolled to cover the costs of the registry.⁸ Due to these increasing costs, hospitals may choose to stop participating in the registry, which would ultimately impede Medicare beneficiaries' access to the procedure.

In order to ensure there is a sustainable funding mechanism that does not unduly burden providers and other relevant stakeholders, Medtronic suggests that CMS should identify appropriate stakeholders to jointly establish a framework for funding the collection and analysis of the data. If CMS envisions a cost-sharing model that includes the private sector, there should be a process to establish who will convene the various parties when there are multiple stakeholders and product sponsors as well as how decisions will be made concerning who will pay what share and for how long. These funding consultations should also include consideration of reimbursement for providers who will bear an additional data collection and reporting burden. Medtronic encourages CMS to provide reimbursement mechanisms for physicians and other healthcare professionals participating in CED data collection, such as implementing special coding for reimbursement directly to providers for the additional efforts required to collect meaningful and relevant data.

In 2005, CMS published an NCD for cochlear implantations, which specified that cochlear implantations may be covered only in the context of a clinical trial approved by CMS, among other criteria.⁹ However, according to the Medicare Evidence Development and Coverage Advisory Committee's (MEDCAC's) review of this topic in 2011, there have not been any clinical trials established to satisfy the terms of the coverage decision, since the final NCD was published in 2005.¹⁰ As a result, cochlear implantations have remained non-covered for the specified patient population, thus preventing beneficiary access to this treatment.¹¹ This case illustrates the importance of a clear framework outlining the appropriate timeline, data elements, and responsibilities for data collection before the implementation of CED. Moreover, it also shows the need to determine interim coverage of the item or service until the data collection vehicle is operational. The need is particularly acute with respect to items or services whose pre- CED coverage status is non-coverage.

In its 2010 Report to Congress, the Medicare Payment Advisory Commission (MedPAC) cites operational challenges to implementing CED that should be addressed as the agency updates its guidance. Identified challenges include the need to develop: (1) a mechanism to identify potential CED candidates, (2) a timeframe for the consideration and re-consideration of a particular service under CED, and (3) a clear financial and management plan for data collection efforts.¹² Medtronic agrees with MedPAC that CMS needs to develop a clear and predictable decision-making process for applying CED, implementing CED, and reconsidering the evidence generated by CED. In addition, CMS should also put forth a process for concluding CED requirements after the research questions have been sufficiently resolved. CMS should consider the development of such processes in draft form before they are finalized, giving stakeholders an opportunity to comment on the processes before implementation.

Medtronic believes that the next iteration of the CED guidance document should clearly articulate a framework for the appropriate application of CED, developed and vetted in a public forum. We recommend that this guidance be issued in draft form before it is finalized, giving stakeholders an opportunity to comment on the new policy before it takes effect. The previously issued CED guidance was too broad and provided little predictability in understanding how CED would be applied in the evaluation of valuable health technologies.

Moreover, while the agency has publicly stated that CED will be applied rarely, the agency has applied its CED policy at an increasing rate. Before the finalization of the CED guidance, CED policy was applied in fewer than 10% of NCDs from 2000 to 2007.13 In comparison, since the guideline was finalized in 2006, CMS issued 38 NCDs, 21% of which have resulted in CED.14 Therefore, it is imperative for the agency to clearly articulate a defined framework for the appropriate application of CED.

To provide further clarity and predictability to the CED process, Medtronic believes the situations in which CED may be applied should be clearly outlined and identified by CMS. We believe the following circumstances are most appropriate:

• Since the primary goal of CED is to increase access to medical advancements for Medicare beneficiaries, as previously stated by the agency, Medtronic believes that CED should only be applied when the alternative to the policy is national non-coverage or significant noncoverage at the local level because of limited evidence of clinical benefit.

• CED should only be used to answer essential questions of "reasonable and necessary" affecting the health of Medicare beneficiaries. In these instances, CED should be used to generate specific evidence that will address the open questions regarding coverage of the item or service under review that are within the jurisdiction of CMS.

• CED should not be used to generate evidence to address research questions that have already been confirmed by RCTs.

• Moreover, CED should not be applied in situations where the outstanding questions are related to safety, as the mechanism to answer these questions is the responsibility of the FDA and generally should not be addressed through CED. The FDA routinely requires manufacturers to conduct post-approval studies for the purposes of detecting adverse events, estimating adverse event frequencies, evaluating product performance, and identifying groups at risk. The manufacturer is already legally obligated to fulfill these requirements and, therefore, CED is not necessary.

• Lastly, CED should also not be implemented in a way that creates geographic disparities in Medicare beneficiary access or unnecessarily disadvantages providers who may not be financially capable of meeting the data collection requirements or reduces coverage available under the existing local process.

Medtronic believes the implementation of CED would benefit from a more structured approach within the national coverage process. Several of the concerns raised above about previous CEDs resulted from the lack of a systematic and transparent approach to conceptualizing and developing the specific details of the proposed CED. We believe this has occurred because the NCD timeframes and comment periods ¡V which otherwise are quite useful and appropriate ¡V do not provide sufficient time or flexibility to address the myriad questions and issues that must be taken into consideration in developing a suitable approach for individual CEDs. We therefore recommend a series of steps that would clarify and improve the process for establishing CEDs when the need for CED is determined to be appropriate.

Medtronic believes that the agency should announce its consideration of CED at the initiation of the NCA to promptly begin a transparent process of ensuring appropriate application and implementation of CED. During this process, CMS should work with all appropriate stakeholders in the development of the CED requirement and framework for data collection.

Following the agency's initial announcement of considering CED at the initiation of the NCA, Medtronic believes that a thorough VOI analysis should be commissioned.¹⁵ Such an analysis would assist CMS in determining whether the additional information collected as part of a CED requirement represents a worthwhile investment relative to the burden it places on stakeholders and restrictions imposed on beneficiary access. This analysis should be completed before the draft decision and the results should be included in the draft guidance.

In order to properly conduct this VOI, CMS should work with the stakeholder community to clearly identify the key questions that, if addressed with sufficient evidence, would allow the agency to make a determination that the item or service is "reasonable and necessary." Moreover, we urge CMS to acknowledge initiatives, such as ongoing private sector and FDAmandated post-market studies, when considering the need for an additional governmentmandated data collection as a condition for Medicare coverage and conduct a gap analysis of ongoing or planned public and private evidence generation activities to ensure that the proposed CED is not duplicative.

We believe the draft decision memorandum should address the following:

1. Specific research question(s) that CMS believes CED will address;
2. Gap analysis of public and private evidence generation activities; and
3. Impact of additional burden of data collection (i.e., time, financial, administrative) on key stakeholders (e.g., geographic inequality in beneficiary access to treatment)

Before the draft decision is published, Medtronic recommends that the agency publicly discuss and vet the VOI analysis findings in an open forum, such as a MEDCAC. Such discussions will help ensure that CED is appropriately applied and that future data collection efforts through CED will not be duplicative or burdensome.

In order to ensure that following an NCD, CED will advance the common goals of all stakeholders, Medtronic recommends establishing a steering committee with broad stakeholder representation, including all relevant stakeholders, such as patient representatives, hospitals, professional societies, and medical technology companies. The draft decision memo should propose key members for the steering committee designed to develop a clear framework for data collection, addressing outstanding factors, such as timeline, data elements, access and ownership to data, and funding of data collection. The public should have the opportunity to nominate members of this committee via public comments after NCA initiation and comment on the proposed members for this committee following the release of the draft decision.

Additionally, it will be important for the agency to include a draft interim coverage policy as a safeguard for beneficiaries between the final NCD and CED implementation.

Medtronic believes that the final NCD should clearly outline the steps required for appropriate and effective implementation of CED. The final NCD should include a final interim coverage policy, a final list of members for the steering committee, and a clear timeline for the committee to develop a framework for data collection. Clearly outlined next steps will ensure appropriate data collection through CED, while granting Medicare beneficiaries access to promising medical advancements.

Following the final NCD, the steering committee would work with the Coverage and Analysis Group to develop a clear framework for data collection under CED. As such, the committee would aim to answer the following questions:

• What is the appropriate study design to specifically answer the CED questions?
• What is the expected duration that patients will be followed in the study?
• What data elements will be collected and how do they specifically contribute to the CED research questions?
• How should the evidence be collected to ensure analyses can be done on specific subpopulations?
• Who will have access to the data?
• Who will have ownership of the data?
• Who will fund the data collection and analysis of the data?
• Who will conduct the analysis of the data?
• What will be the time intervals of this analysis and review by CMS?

The answer to each of these questions has important implications for the success and scientific character of data collection through CED and should be considered in an open and transparent manner, led by the steering committee. The agency should not move forward without further clarification on these questions. Moreover, the answers to these questions should be posted on the tracking sheet of the decision. Failure to articulate these issues in any NCD with a CED requirement will put an undue burden on beneficiaries, providers, and other stakeholders, and ultimately, decrease the potential for any meaningful data collection to lead to improved decision-making.

CMS should continue to implement CED within the NCD process. Medtronic does not believe that CED should be implemented by local Medicare contractors due to challenges with executing the policy at this level, such as limited sample size to adequately power studies and the potential for duplicative studies with potentially conflicting outcomes across contractors. Moreover, since a CED decision impacts entire populations of Medicare beneficiaries in need of specific classes of products, it is also important that such decisions are made at the national level to ensure that all relevant stakeholders have the opportunity to engage in the decision-making process. The NCD process, with its statutorily defined public comment periods and process requirements, provides stakeholders with the opportunity to monitor and weigh in on CED decisions that impact them.

Medtronic appreciates this opportunity to provide comments to CMS on the CED policy. We appreciate your consideration and are happy to provide further information or assist with any additional questions. Please feel free to contact me at (202)442-3633 or jeff.a.farkas@medtronic.com if you have any questions or wish to discuss our comments in further detail.

Sincerely

Jeff Farkas
Vice President, Health Policy and Payment
Medtronic, Inc.