Second-Guessing the FDA: CMS’s Expanding Regulatory Role

November 15, 2010
Michael Mccaughan

Oncology, ONCOLOGY Vol 24 No 12, Volume 24, Issue 12

Nothing gets biopharma policy watchers more worked up than the possibility that the Centers for Medicare & Medicaid Services (CMS) will second-guess FDA approval decisions. In reality, though, CMS often has no choice but to apply its own interpretation to issues that also fall under the FDA’s jurisdiction-and implementation of the health care reform is bringing more of those cases to the fore.

Nothing gets biopharma policy watchers more worked up than the possibility that the Centers for Medicare & Medicaid Services (CMS) will second-guess FDA approval decisions. In reality, though, CMS often has no choice but to apply its own interpretation to issues that also fall under the FDA’s jurisdiction-and implementation of the health care reform is bringing more of those cases to the fore.

Michael McCaughan

The idea of FDA/CMS collaboration makes plenty of folks in the biopharmaceutical industry nervous. Their fear is understandable: the specter of a direct link between government reimbursement authorities and regulatory authorities prompts visions of impossibly difficult clinical trial requirements and/or unsustainably restrictive coverage and reimbursement policy.

Sponsors much prefer a bright line between the role of regulatory authorities and that of the payer agency. Better to make separate cases for approval and coverage, the thinking goes, than face the potential for coordinated action to limit your commercial opportunity.

Unfortunately for sponsors, reality doesn’t match up with either the projected collaboration they fear or the neat compartmentalization they long for.

There is no unseen hand operating behind the two regulatory agencies when it comes to decisions about new products. But there is also no use in pretending that the two agencies operate in isolation, with complete independence from each other when it comes to those decisions. Instead, there is a haphazard and at times unpredictable feedback loop between the two-a state of affairs made abundantly clear during the long back-and-forth over the safety issues associated with the erythropoietin stimulating agents.

The Food & Drug Administration and the Centers for Medicare & Medicaid Services have overlapping missions, and their standards and priorities can and sometimes do come into conflict. They are also part of the same cabinet department, so the notion of greater coordination is a perennial subject of attention.

Initiatives to improve inter-agency collaboration have been a priority of every recent Secretary of Health & Human Services (Donna Shalala under Clinton, Tommy Thompson and Michael Leavitt under Bush, Kathleen Sebelius under Obama). Those initiatives can produce operational efficiencies, such as those resulting from centralized purchasing or a unified email system. But the very fact that better coordination is a recurring theme demonstrates how far from true coordination the two agencies remain.

That really is nothing new. But, if it feels like a more pressing issue, that’s because it is. Thanks to the Medicare Modernization Act, CMS is well on its way to becoming the majority payer for prescription drugs in the US.

The explosion in “specialty” pharmaceutical markets-disproportionately paid for by Medicare Part B-is only accelerating that trend.

And all that, before the newly enacted health care reform law.

The landmark legislation of this past spring will not only further expand CMS’s role as market organizer for pharmaceuticals in the US, it also increases the number of areas where CMS will have a separate regulatory role layered on top of the FDA’s existing authority. Examples include:

• Identifying “pediatric only” drugs for purposes of Medicaid rebate calculation.

• Defining “new formulations” that trigger a separate Medicaid rebate provision.

• Defining “brands” and “generics” in the context of the Part D Donut Hole discount.

And, if the current state of affairs is any indication, sponsors may end up deciding that they would rather see more careful coordination between the two agencies, rather than find themselves caught in the cross-currents of conflicting authorities.

Provenge Provides Context
There is no doubt that an unexpected step by CMS can have sweeping implications for the biopharma sector.

Exhibit A: CMS’s decision to open a national coverage analysis for Dendreon Corporation’s prostate cancer immunotherapy Provenge.

In this case, CMS appears to be second-guessing the FDA’s decision to approve the product in the first place, asking whether there is sufficient evidence to justify coverage of the agent. The request for a national coverage analysis (NCA) seems quite explicitly to question whether there is sufficient evidence to support coverage for the labeled patient population.

That, at least, is how the Biotechnology Industry Organization (BIO) is reading the NCA request. BIO, citing statutory terms governing the coverage of off-label use of cancer therapies, urged CMS to drop the process altogether.

Instead, CMS is moving forward quickly, convening its Medicare Evidence Development & Coverage Advisory Committee (MEDCAC) on November 17. And, as if to reinforce the view that it is not obligated to cover FDA-approved indications for cancer therapies, the meeting agenda is headlined “On-Label and Off-Label Use of Autologous Cellular Immunotherapy Treatment of Metastatic Prostate Cancer.”

CMS “has called this meeting to consider the currently available evidence regarding the impact of labeled and unlabeled use of autologous cellular immunotherapy treatment on health outcomes of patients with metastatic prostate cancer,” the meeting announcement states. And, if past committees are any guide, MEDCAC will be asked to judge the quality of evidence supporting the labeled use of the therapy.

That is still several steps away from an action by the agency to deny coverage for the on-label use, but it does suggest that the assumption that CMS must cover FDA-approved indications for a cancer therapy could be open to discussion.

CMS’s decision to open the coverage decision has nothing to do with the health care reform implementation process; yet the timing-so soon after the enactment of the new law, and immediately after the appointment of Administrator Don Berwick to oversee the implementation-gives the issue a symbolic weight as a possible sign of a new era.

The fact that the coverage review was announced soon after the completion of a long-debated agreement to allow data sharing between the FDA and CMS only highlighted the sense that the Provenge issue is a harbinger of things to come. (The FDA/CMS memorandum of understanding is itself an example of the challenges of building a more collaborative process for the two agencies; it began as an initiative under the Bush Administration but took eight full years to reach formal agreement.)

The Provenge issue is all the more remarkable in that the Medicare statute is unusually explicit in what it has to say about cancer therapies, including standards for non–FDA-approved indications under Part B. The fact that there is no explicit and unequivocal statement in the statute requiring coverage of an FDA-approved cancer therapy may therefore seem surprising.

Nothing Simple About Collaboration in Part D
The Medicaid exclusion language became even more complicated for CMS once Medicare Part D was enacted. Under the Medicare Modernization Act, Part D was prohibited from covering drugs included on the Medicaid exclusion list. Congress took what was an optional exclusion for state Medicaid agencies, and turned it into a mandatory exclusion list for the new Part D benefit.

That came as a surprise to some manufacturers and caused a number of headaches for CMS and health care providers alike when the program launched. For instance, the exclusion of barbiturates and benzodiazepines from the Medicaid outpatient drug benefit took on a new significance when Part D became the primary payer in the nursing home setting.

As the part D program matured, CMS began to pay more attention to the fact that Part D plans had, in some cases, been providing coverage for prescription drugs that were never approved by the FDA. The status of so-called “legacy” drugs had been something of a gray area, with sponsors asserting that some prescription drugs were “grandfathered” under the Federal Food, Drug, and Cosmetic Act and did not require approval of a New Drug Application (NDA).

That changed shortly after the launch of Medicare Part D, when the FDA announced its “unapproved drugs initiative,” spelling out unequivocally its position that unapproved drugs are “illegally marketed” and outlining a risk-based enforcement plan.

As the issue gained attention, CMS began working with the FDA to develop a list of National Drug Code numbers that could not be traced back to an approved NDA or Abbreviated New Drug Application (ANDA). Effective January 1, CMS directed Part D plan sponsors to consult the so-called “non-matched” drug list as a first step in determining eligibility for payment of Part D claims.

That may sound like a model of interagency collaboration, but look at it another way: it took four years for CMS to work out a system to help Part D plans determine whether a given drug is or is not FDA-approved.

And the model that is in place does not, apparently, portend ongoing cooperation on similar issues.

Whose Job Is It?
In an August 16th report, the Health and Human Services Inspector General looked at payment under Part D for drugs determined to lack sufficient evidence of efficacy under the Drug Efficacy Study Implementation (DESI) review process. As the Inspector General notes, after the FDA makes a determination of lack of efficacy, sponsors can appeal the determination, meaning that the product can remain on the market in a sort of regulatory limbo.

The statute governing Medicaid and Part D is clear: DESI-ineffective drugs are not covered.

Nevertheless, the Inspector General found, Part D paid for about $43 million worth of such drugs in the first two years of Part D. While that is a small sum in the context of about $115 billion in Part D spending over that period, it appears to be a black-and-white case of waste.

The Inspector General’s proposed solution is based on the “non-matched” list approach: CMS should create “a comprehensive, up-to-date list of less-than-effective [LTE] drugs” and verify “the accuracy of the list with the FDA on a routine basis.” The list should be disseminated to sponsors and used as a basis for rejecting claims.

That sounds simple enough. But in May 20th comments on the draft report, CMS demurred. “The FDA is the agency tasked with making regulatory drug status determinations and making this information publicly available,” CMS acting Administrator Marilyn Tavenner said. “Therefore, CMS believes it is the responsibility of the FDA to produce a comprehensive, up-to-date list of the DESl LTE drugs (and their respective NDCs [National Drug Codes]).”

“We believe the FDA is in the best position to accomplish this and encourage the OIG [Office of the Inspector General] to work with FDA in recommending mechanisms to disseminate comprehensive DESI lists to all stakeholders,” Tavenner explained to the Inspector General. “The lack of a complete and accurate listing of all marketed drug products and their NDCs are of ongoing concern to CMS and Part D sponsors since the inception of the Part D program. The fact that it is difficult to identify Federal Register notices associated with DESI products and that these notices are not easily retrievable further complicates the issue.”

Tavenner then proposed a solution that underscores the existing lack of coordination between the two agencies: “As per the methodology in the draft report, the OIG presented the FDA with a list of LTE drugs compiled using various data sources. CMS would appreciate if you could clarify whether the FDA confirmed that the OIG’s list represents both an accurate and complete list of FDA-determined DESI LTE drugs. Also, did the OIG inquire whether or not the compiled list represented all marketed NDCs available for LTE drugs or drugs identical, related, or similar to DESI LTE drugs? . . . Interpretation of what constitutes an [identical, related or similar] drug is another source of confusion and possible error that could benefit from the existence of a comprehensive FDA list of DESI drugs. Lastly, was the OIG provided with an explanation from the FDA as to why a comprehensive list of DESI drugs has not yet been made publicly available? If so, it would be informative to include this in the report.”

The Inspector General is sympathetic, but only up to a point. “We acknowledge that FDA plays an important role in identifying less-than-effective drugs, and we modified our second recommendation to reflect that role,” the report states. The Inspector General rewrote the recommendation to suggest that CMS “collaborate” with the FDA to create and maintain the list.

“However, as the administrator of the Medicare Part D program, CMS has the primary responsibility to ensure that sponsors are not paid for less-than-effective drugs. Furthermore, CMS has a responsibility to beneficiaries to ensure that the drugs being prescribed through the Part D program are safe and effective. Therefore, we continue to recommend that CMS regularly disseminate a list of less-than-effective drugs to all sponsors to ensure that they are provided with the information necessary to appropriately administer their Part D plans.”

The new donut hole discount program that begins in 2011 appears to have created a similar issue for CMS, since the new law calls for a 50% discount on brand name drugs in the donut hole-but not on generics.

After initially proposing to include “authorized generics” as “brands” for purposes of the donut hole discount, CMS settled on defining brands as drugs marketed under an NDA and generics as those marketed under an ANDA. (Some “authorized generics” are marketed under license of the approved NDA; many are under a separate ANDA, so that distinction is eliminated.) However, several plan representatives have alerted CMS to the potential for confusion about whether a given drug is in fact marketed pursuant to an ANDA or an NDA. That issue was raised during a workshop on the new discount program June 1, with CMS officials asking plans to submit more specific examples.

Such technical issues involved in the implementation of the health care reform law don’t have the headline-making sizzle of a coverage review of Provenge. But they do underscore the evolving nature of FDA/CMS collaboration, and the real world challenges of making the process work without leaving sponsors trapped in the middle.