Medicare’s Aducanumab Decision Highlights Needed Reforms To FDA And CMS Regulatory Pathways
Much ink has already been spilled to parse the Centers for Medicare and Medicaid Services’ (CMS’s) January 11 draft national coverage determination (NCD) on monoclonal antibody (mAb) therapies for mild cognitive impairment and early Alzheimer’s disease (AD). In this Forefront article, we examine a critical and underappreciated concern that the decision highlights: the Food and Drug Administration (FDA’s) and CMS’s different decision-making authorities and the issues that arise when the two agencies make different judgments about the same evidence. We also explore what the Alzheimer’s decision may mean for future Medicare drug coverage policy, especially for drugs that treat serious diseases for which existing therapies are inadequate.
An NCD Like No Other
CMS’s eagerly awaited NCD is surprising in its restrictiveness: While many observers correctly predicted the use of “coverage with evidence development” (CED), the proposal to cover mAb therapy only in the context of randomized controlled trials was mostly unanticipated. These trials must be approved by CMS and satisfy specific criteria, including those largely consistent with the designs of pivotal anti-amyloid mAbs Phase 3 studies (such as confirmed amyloid pathology) and beyond (such as diversity and national representativeness of the population diagnosed with Alzheimer’s disease and demonstration of clinically meaningful improvement in cognition and function). This draft decision, issued on January 11, 2022, will be followed by 30 days for public comment, with the final ruling expected in April 2022.
This draft NCD is unprecedented because Medicare has historically paid for FDA-approved drugs, including those approved under the accelerated approval pathway, for on-label indications. NCDs are rarely undertaken for drugs in the first place; instead, Medicare usually assigns a billing code and then reimburses at a percentage of the list price set by the drug company.
But the FDA’s approval of the first mAb therapy, aducanumab (Aduhelm), triggered substantial controversy, including federal investigations of the review process, widespread criticisms of the therapy’s high list price, and announcements from numerous health systems that they would not offer the drug (although some clinical experts and patient advocacy organizations have supported the FDA’s decision and encouraged broad Medicare coverage).
CMS’s decision reveals an agency trying to thread a tricky needle, on the one hand addressing evidence uncertainties and potentially huge budget implications, while on the other respecting the FDA’s scientific expertise and judgment. But Medicare officials clearly came down on the side of caution, reflecting the significant limitations of the trials on which the FDA approval was based and the high rate of serious adverse effects associated with this therapy. The draft NCD reveals one branch of the federal government (CMS) in disagreement with another (FDA), even though both are within the same department (Health and Human Services). Medicare’s decision thus highlights differences in authorities with which the FDA and CMS make independent judgments about the evidence supporting drugs and devices—and illuminates potential problems made particularly visible in the context of the FDA’s accelerated approval pathway.
A Clash Of Titans: When FDA And CMS Don’t See Eye To Eye
In its draft decision memo, CMS concludes that “no trial of an anti-amyloid mAb has confidently demonstrated a clinically meaningful improvement in health outcomes (i.e., cognition and function) for AD patients. Thus, there is insufficient evidence to conclude that the use of monoclonal antibodies directed against amyloid is reasonable and necessary for the treatment of Alzheimer’s disease under §1862(a)(1)(A) of the Social Security Act” (italics added). Notably, there are no laws, regulations, or guidance from CMS that explicitly require evidence of “meaningful improvement in health outcomes” to meet Medicare’s statutory authority to decide which therapies are “reasonable and necessary.” In a final rule issued and then repealed in 2021, Medicare did not mention “improved health outcomes” as a criterion to consider in determining which services are “reasonable and necessary.” However, this standard has been consistently applied by CMS in its coverage decisions for two decades, and it is eminently reasonable for Medicare to limit payment for services to situations in which such evidence exists.
In contrast to Medicare, the evidentiary standards the FDA uses for decisions about the safety and effectiveness of medical products are extensively described in law, regulation, and guidance documents. In the case of aducanumab and other drugs approved through the accelerated approval pathway, the FDA requires “…a determination that the product has an effect on a surrogate endpoint that is reasonably likely to predict clinical benefit…” (italics added). A surrogate endpoint is not itself a fully validated measure of clinical benefit in the context of accelerated approval, which is why the FDA requires confirmatory trials to verify that expected clinical benefits are demonstrated after an accelerated approval. The FDA must make a judgement about the strength of evidence for the association between a surrogate marker and a direct measure of clinical benefit. It is this feature of the accelerated approval pathway, explicitly defined in federal law, that allows faster patient access to promising therapies for serious and rare diseases when there are inadequate alternatives, by accepting a higher level of scientific uncertainty about benefits than required for standard approval. That said, the FDA’s criteria about the quality and quantity of evidence necessary to support a “reasonably likely” judgment are highly variable, and these decisions rarely cite the reasoning from past decisions—reducing the transparency, consistency, and predictability of this process. In the case of aducanumab, strong pushback from FDA advisers, multiple outside experts, and some internal FDA staff suggest that available evidence falls short of the usual level expected to support the judgment that beta amyloid reduction is “reasonably likely” to predict cognitive benefits.
In many ways, CMS and the FDA are in agreement about the available evidence for aducanumab and other mAb therapies. Both conclude that there is not yet definitive evidence supporting the conclusion that these therapies improve patient cognition or function, and both note safety concerns. To support its decision, CMS cites a review of all currently available mAb randomized trial data that concludes, “no biomarker has achieved surrogate status in AD drug development with definite evidence that a change in the biomarker predicts a clinical benefit.” It also cites a meta-analysis by authors from the National Institutes of Health that found a statistically significant, but not necessarily clinically meaningful, difference in health outcomes for patients in mAb trials that have been completed to date. These conclusions are consistent with the FDA determination that beta-amyloid is not yet a well-validated surrogate measure for cognition or function in Alzheimer’s disease. However, federal law provides the FDA with authority to grant accelerated approval for these products, while the same evidence falls short of meeting CMS interpretation of its reasonable and necessary authority.
Importantly, by the FDA’s evidentiary standards, all drugs approved through accelerated approval would fail to meet CMS’s evidence requirements as articulated in the draft mAb NCD. What is unusual in the case of aducanumab is the broad negative reaction of the clinical community to this FDA approval, which is not typically seen for other accelerated approval drugs. In addition, aducanumab is associated with frequent serious side effects. CMS noted these factors as considerations in their draft decision; they also noted that many experts outside the FDA do not agree with the judgment that the beta-amyloid reduction is “reasonably likely” to predict cognitive benefits.
Unlike the FDA’s accelerated approval program, CMS has no explicit legal authority to accept a lower level of evidence (or greater uncertainty) for drugs that treat serious or rare diseases or address unmet medical needs when making their coverage determinations. This creates a conundrum, in that the intent of the accelerated approval pathway is to allow for earlier patient access to drugs for serious diseases while confirmatory evidence is being generated. Decisions by CMS to restrict coverage of accelerated approval drugs to randomized controlled trials effectively neutralizes the intent of the authority explicitly assigned to the FDA, and there has not yet been a substantive policy discussion about whether and how these independent agency authorities could be better aligned.
The intent of the Medicare CED approach is more closely aligned with the FDA’s accelerated approval pathway, echoing the goal of allowing earlier access to promising technologies before there is definitive evidence of improved health outcomes, while supporting the generation of additional evidence through Medicare payment. As described in its CED guidance, CMS states that: “we believe that CMS should support evidence development for certain innovative technologies that are likely to show benefit for the Medicare population, but where the available evidence base does not provide a sufficiently persuasive basis for coverage outside the context of a clinical study.” CMS has used a variety of study designs in previous CED policies, ranging from randomized controlled trials (RCTs) (for example, lung volume reduction surgery in 2005) to patient registries (for example, transcatheter aortic valve replacement in 2019). The proposed decision to use CED for Medicare coverage for mAb drugs in RCTs is more likely to produce high-quality evidence. However, it will also limit the number of Medicare patients able to access these drugs by enrolling in trials to a small fraction of the number that the FDA accelerated approval pathway would have allowed, although some note that this may in fact spare these individuals from serious risks in the context of highly uncertain benefits.
Next Steps
CMS’s draft decision may still change when issued in final form later this spring, and time will tell if the aducanumab/mAb case is a harbinger of a new era for Medicare drug coverage policy. Regardless, the tension between accelerated approval and CED will continue to exist and strongly suggests a need for further interagency dialogue and congressional action. A number of steps would help to strengthen the FDA’s accelerated approval program and better coordinate FDA and CMS decision making.
First, as suggested by the former (and likely future) FDA commissioner, Robert Califf, is to improve the research infrastructure and mechanisms for evaluating products once they are on the market. At his confirmation hearing, Califf noted, “I am a fan of accelerated approval,” but he also said that means “we’ve got to have a better system to evaluate these products as they’re used on the market” (emphasis added). Rather than having ill-tempered debates about the FDA not doing its job properly, we need to think more creatively about how to better ensure high-quality, rapid post-approval evidence generation for all accelerated approval products. These efforts would include further construction of a learning health system that leverages health data to facilitate more efficient, streamlined clinical trials and high-quality observational studies, as Califf and senior leaders from multiple federal health agencies have previously outlined.
Second, there is need for more FDA-CMS coordination prior to product approval, especially for drugs and devices expected to have large budgetary and clinical consequences for Medicare, and for those expected to receive accelerated approval. Such coordination could help to identify ways to strengthen regulatory and reimbursement policy incentives to complete confirmatory studies in agreed upon and reasonable time frames.
Third, CMS should continue to improve the CED policy framework for promising drugs and devices to identify approaches that achieve the framework’s intent: to facilitate better access while also generating reliable evidence as quickly as possible.
Fourth, policy makers should explore statutory language that is parallel to the FDA’s accelerated approval authority in section 901 of the Food and Drug Administration Safety and Innovation Act, clarifying how CMS should apply its “reasonable and necessary” authority for promising treatments that address serious illnesses for which existing therapies are inadequate.
Looking Forward
CMS’s proposed decision on aducanumab and other mAb therapies shows a government in disagreement with itself over how best to provide patient access to a novel Alzheimer’s disease therapy. But the decision helps to clarify important structural differences in the FDA and CMS authorities and evidentiary requirements, and how their applications can affect early patient access to promising therapies. From a health policy standpoint, aducanumab’s enduring legacy may be its role as a catalyst for reconsidering the FDA’s accelerated approval program and the relationship between the FDA and CMS, and stimulating further attention on policies and infrastructure that support efficient, high-quality post-approval evidence generation.
Authors’ Note
The authors are members of the Center for the Evaluation of Value and Risk in Health at the Institute for Clinical Research and Health Policy Studies at Tufts Medical Center. The Center receives funding from government, private foundation, and pharmaceutical industry sources. Drs. Chambers, Lin, and Neumann have consulted with pharmaceutical companies including Biogen and others on issues related to health economics and outcomes research. Dr. Neumann served as an external reviewer on the ICER’s recent evaluation of aducanumab. Dr. Tunis has consulted on research and policy strategy for a wide range of health care organizations, including Biogen and UsAgainstAlzheimer’s; he has been employed in the past by both CMS and the FDA, but was not involved in any discussions involving aducanumab while at either agency.