By Rachel Sachs, Jacob S. Sherkow, Lisa Larrimore Ouellette, and Nicholson Price
The antiviral drug remdesivir, developed by California-based biopharmaceutical company Gilead Sciences in collaboration with government scientists, has emerged as the new standard of care for COVID-19 patients. The drug has also been beset by controversy, from Gilead’s withdrawn orphan drug application to conflicting reports of clinical trial results to an opaque and haphazard distribution process. We examine the role of legal institutions in the remdesivir rollout in two posts focused on the two (separable) parts of innovation policy. This week focuses on the innovation incentives for producing information about remdesivir and other COVID-19 treatments, including the FDA’s role in deciding whether and under what conditions a new drug can reach the market. Next week we will turn to pharmaceutical allocation mechanisms, including payment and distribution.
What do we know about how well remdesivir works?
Remdesivir emerged early as a potential COVID-19 treatment because it had already been developed as a potential treatment for hepatitis or Ebola, although it failed clinical tests for efficacy and was never approved by the FDA. Gilead began distributing remdesivir for COVID-19 patients under a “compassionate use” basis in late January, until anecdotes of successful outcomes and promotion by the President led to overwhelming demand by late March.
Many of the clinical trials investigating remdesivir for treating COVID-19 can be tracked at ClinicalTrials.gov. (Trial pre-registration is required for FDA-approved drugs, though compliance is low.) On April 29, The Lancet published results from a randomized controlled trial in Wuhan, which led to disappointment for finding no statistically significant clinical benefits. The trial had been terminated early, however, due to insufficient new cases and only enrolled 237 patients rather than the original plan of 453. On the same date, preliminary results were released from a 1,063-patient trial conducted in the US by the NIH’s National Institute of Allergy and Infectious Diseases (NIAID). The results are thin but somewhat promising: NIAID found that 100mg of intravenous remdesivir decreased recovery time from 15 days to 11 days (p<0.001), and decreased mortality—although, not to standard levels of statistical significance—from 11.6% to 8.0% (p=0.059).
The NIAID study is “the most important and rigorously designed” remdesivir trial yet, but it has still been subject to multiple lines of criticism. The clinical benefit is modest. The primary endpoint of the trial was changed shortly before releasing the results, although this appears to have occurred before the data was unblinded. The study was stopped early: promising results caused patients randomized to the placebo to be offered remdesivir when only 480 of the 1,063 patients had recovered, limiting the ability to collect more data. And NIAID has so far released only a press release rather than the full study results, making it difficult to reconcile these results with the Chinese study’s finding that there was no difference in viral load reduction. In short, substantial questions remain about not just the optimal timing and dosage, but also remdesivir’s overall efficacy. Even under the most optimistic reading of these results, remdesivir is not a silver bullet that will fundamentally shift the course of the pandemic—but it appears to be the best we have for now.
How will remdesivir change COVID-19 treatment in the US?
Primarily on the basis of this NIAID-administered clinical trial, on May 1 the FDA issued an Emergency Use Authorization (EUA) for remdesivir. Most importantly, an EUA is not the same as a typical FDA approval. One principal difference between the two is in the approval requirements: under an EUA, the FDA asks whether the product in question “may be effective,” which is short of the FDA’s typical approval standards. The EUA also lasts only for the duration of a particular public health emergency (unless it is revoked earlier), after which time the product would be pulled from the market if it failed to obtain full approval. There are a wide range of other limitations that accompany EUAs (as detailed in the FDA’s guidance on the topic), including limitations on advertising that are not as easily imposed on products approved through the normal process.
The EUA does allow physicians to prescribe and administer remdesivir for their patients. (Assuming they can obtain the drug, as we will consider in Part II of this post next week.) But those physicians are under greater limitations on the conditions of distribution and administration than they would typically be for newly marketed products. Typically, once the FDA approves a new drug, physicians can prescribe it to patients beyond the scope of the drug’s label. This might represent a small deviation from the drug’s intended use—prescribed to patients with mild disease, rather than moderate-to-severe disease, for instance—or a larger deviation, such as when a drug is used in an entirely different condition. In some cases, such as for opioids and other drugs with particular risks, the FDA does impose more restrictive conditions on prescriptions, but those are not the norm. In the EUA context, such restrictions are standard, and the FDA has restricted the administration of remdesivir to COVID-19 patients with relatively severe disease.
Now that remdesivir has been authorized for prescription, some (including the FDA’s former chief scientist) have argued that the EUA for hydroxychloroquine issued on March 28 ought to be revoked. At the time, the FDA granted the EUA because, in the agency’s view, (1) it was “reasonable to believe” that hydroxychloroquine “may be effective” for COVID-19 and (2) “there is no adequate, approved, and available alternative” to these products for treating COVID-19. More than a month later, now that remdesivir has received an EUA, have these dynamics changed? In some ways yes, but in other ways, no.
The clinical trial evidence supporting a moderate benefit of remdesivir does change the dynamics that led to the EUA for hydroxychloroquine. Essentially, the existence of a new drug with some demonstrated efficacy against COVID-19 changes the political and scientific dynamics that led to the hydroxychloroquine EUA. Particularly in light of new studies (albeit observational ones) continuing to show no benefit of hydroxychloroquine, it may no longer be reasonable to think that the drug has efficacy against COVID-19. As the FDA noted in its EUA for remdesivir, “there is no adequate, approved, and available alternative” to remdesivir for COVID-19.
However, it is not obvious that the “adequate, approved, and available” criterion used in the EUA for hydroxychloroquine has changed. Remdesivir is not approved, but is authorized. And questions of remdesivir’s availability loom large in its distribution process, as we will discuss next week. Further, the fact that hydroxychloroquine is already FDA-approved for other indications means that removing the EUA may have little impact on physicians’ ability to prescribe it for COVID-19, although physicians themselves are already prescribing less of the product, likely due to the lack of evidence for its efficacy. The FDA should take a hard look at the continuing viability of the EUA for hydroxychloroquine over the coming days, but removing the EUA may have little policy impact.
How can we get better evidence of remdesivir’s effectiveness?
Under an EUA or not, it’s important to demonstrate whether remdesivir is actually effective in improving patient outcomes. In an ideal world, the potential importance of remdesivir would command better studies than those that have been conducted so far—a larger sample size and more treatment arms, for example. But the quality of clinical trial design is often tethered to the FDA approval regime. In general, once the approval threshold can be crossed, drug marketers have no further incentives to design more robust trials. Indeed, they have every incentive not to, quailed by the fear that negative information about the drug’s safety or efficacy could threaten a likely approval. A 2012 Institute of Medicine report described just this “oversight” for FDA-required post-approval clinical studies.
Having entities other than the drug’s principal marketer conduct such studies may help. For instance, government agencies with public health mandates have stronger incentives to ensure the best possible information about approved drugs makes its way to clinicians and the public. (This is one of the reasons why Professor Rebecca Eisenberg referred to FDA’s role as not just gatekeeping, but also “the development of credible information about the effects of drugs.”) Such a project, in fact, has recently been launched by the WHO, and included several government agencies around the world, Gilead (the commercial marketer of remdesivir), and AbbVie, a company with an interest in a remdesivir combination therapy. Government and nonprofit researchers may still face political pressure or career-related incentives to focus on positive results, but they are still more likely to report negative results than industry-funded scientists.
In addition, government involvement in coordinating such trials has an international virtue—the ability to coordinate trials across countries. While drug approval is typically a domestic concern, the global nature of the COVID-19 pandemic, and harm possibly wrought by conducting duplicate trials across borders, makes international coordination all that more important. Indeed, this is a principal motivating factor behind the WHO’s study and its attempt to coordinate public health agencies from ten different countries. Government agencies have the power to persuade, purchase, partner, and pilot the best studies needed to generate evidence of remdesivir’s efficacy—or, as it may turn out, lack thereof.
What does this mean for incentivizing other anti-COVID-19 drugs?
Given that remdesivir has at most a modest clinical benefit, having a treatment that would greatly reduce mortality should still be a first-order policy priority. This is perhaps an obvious point, but an important one: we should still be putting a lot of thought, effort, and money into developing treatments for COVID-19. That includes non-pharmaceutical treatments like medical devices (including ventilators) or more effective procedures, but most obviously it means finding drugs that work—ideally substantially better than remdesivir. Last night, the FDA released guidance for developing new COVID-19 drugs, and policymakers should make sure firms have ample incentives to investigate potential candidates.
In pursuing this priority, policymakers have access to a well-stocked toolbox of options. Patents provide the innovation incentive backdrop, letting companies appropriate a substantial portion of the social benefit of their invention. But market-based rewards may under-incentivize treatments for pandemic infections for a variety of reasons. For example, most drug candidates are also existing drugs that are being evaluated to see if they might work against COVID-19, rather than totally new drugs (for which the development timeline is just too slow)—and patents are a relatively weak incentive for this type of development.
Grants, prizes, reimbursement policies, and tax incentives are other big tools in the toolbox for pharmaceutical innovation, all of which can be tailored to correct for failures in market-set rewards, including through conditions on transparency and affordability. Tax incentives reduce development costs (and indeed, one benefit of Gilead’s seeking orphan drug status for remdesivir would have been a hefty tax credit). Insurance reimbursement policies can be shaped to provide rewards for particular innovative treatments, effectively acting as prizes for solving a particular problem and getting the solution to patients. Grants similarly provide up-front funding for projects that reviewers deem promising and important. Government-set incentives such as grants and prizes are least effective when it is hard for the issuer to know what problems are important to solve or how valuable the solutions are, but that’s not the case here; improving outcomes for COVID-19 is a clearly-defined problem and—as long as trials are done right, measuring success should be straightforward.
We think policymakers should keep two key points in mind as they sort through these options. First, substantial public funding is needed, but failures in the political market for allocating this funding can be just as bad as market failures in the patent system. One promising innovation policy for limiting discretion (and political whim) in funding allocations is the Advance Market Commitment (AMC), a form of prize in which the government commits to spending a certain amount per patient successfully treated. The AMC model was successfully used to incentivize development of a pneumococcal vaccine, and it has been championed in the COVID-19 vaccine context: two months ago Daniel Hemel and Lisa Ouellette suggested a commitment of $500 per person vaccinated, and last week Susan Athey, Michael Kremer, Christopher Snyder and Alex Tabarrok suggested $100. But AMCs are not limited to vaccines: we think the model could also be used to incentivize more effective treatments, and it has the potential for bipartisan appeal.
Second, effective coordination is critical. Public funding of research and clinical trials is sometimes criticized as picking winners and losers (though that criticism may be overblown). Here, we have even more robust pushing by the government, including the allocation of patients to clinical trials, funding for development and manufacture, streamlining of approval at FDA (with the possibility of political pressure), and the general prioritization of particular research strategies. This sort of full-press effort seems entirely justified, but it’s important to recognize that we need multiple candidates, and that efforts to prioritize one treatment shouldn’t sap the energy of ongoing research into other, better cures.