How is the NIH seeking to encourage investment in diagnostic testing for COVID-19?

By Rachel Sachs, Jacob S. Sherkow, Lisa Larrimore Ouellette, and Nicholson Price

In this blog post series, we have written about the importance of ensuring an adequate supply of diagnostic testing for COVID-19, and about the challenges the United States has faced so far on this front. Many months after COVID-19 began spreading in the United States, there is still great concern that we lack sufficient testing to ensure that we can detect and respond to the virus quickly. In this post, we explain a new initiative from the National Institutes of Health (NIH)—the Rapid Acceleration of Diagnostics (RADx) Initiative—that aims to speed the development and marketing of new COVID-19 tests, how RADx is progressing so far, and what policymakers can learn from the initiative about the design of different types of innovation incentives. 

What is the NIH’s RADx Initiative and how is it progressing so far?

NIH’s RADx Initiative is a $1.5 billion award program “for scientists and organizations to bring their innovative ideas for new COVID-19 testing approaches and strategies.” The Initiative’s goal is to “to make millions of accurate and easy-to-use tests per week available to all Americans by the end of summer 2020, and even more in time for the flu season”—in particular, for at-home and point-of-care testing. For purposes of both assessment and funding distribution, RADx is also designed to expand the Point-of-Care Technologies Research Network (POCTRN), a consortium of five research centers working on the “development of inexpensive and easy-to-use medical devices and information sharing tools that provide timely health status information at the point of care.”

RADx serves as both an accelerator, of sorts, and a regulatory liaison for diagnostic innovators whether they hail “from the basement [or] the boardroom.” In recognizing that the pandemic implicates matters of both public health and national security, RADx is an interagency affair, with coordination among the FDA, CDC, BARDA, and NIH. This, perhaps, is noteworthy itself; a significant portion of criticism of the government response to the pandemic has been a lack of substantive interagency coordination (which, to be sure, is also a problem outside the pandemic). RADx appears to be an attempt to change that. The FDA, which oversees laboratory diagnostics, has also committed to expedited review of successful awardees.

RADx is particularly notable for its funding structure. The program aims to fund companies focusing on the development of a diagnostic product, as they’re developed—rather than general, exploratory research, as is classically the case with grants, or upon demonstrated completion of a particular goal, as with traditional prizes. RADx projects are assessed—and funded—in three phases: Phase 0, a basic concept and, possibly, a working prototype; Phase I, a validation study and a risk review assessment; and Phase II, clinical tests, regulatory approval, and a scale-up of manufacturing. These phases are meant to be completed quickly: NIH’s schema shows that an awardee may be able to complete all phases by late summer or fall of this year—only weeks away. The NIH has eschewed calling RADx a series of “prizes”—even though the agency has used such language in the past—and instead referred to Phase 0 of the system as a “Shark Tank”-like competition. 

The speed and commitment to fund all comers seems real. The NIH last week announced that seven companies have received $250 million in awards through RADx. Some winners are quite small, such as Talis Biomedical, a small biotech company from Menlo Park, California. Others are much larger, like the San Diego biotech behemoth, Quidel (market cap: $11 billion). But further RADx funding isn’t limited to these seven firms; others are still eligible for funding, and yet others can jump ahead in the phase queue if they can demonstrate advanced success. Most importantly, “[t]he seven technologies use different methods and formats and can be performed in a variety of settings to meet diverse needs,” including using CRISPR-based technology. Whatever terminology can be used to describe such a funding mechanism, the public health and research arms of government are paying milestone payments to a wide variety of companies to design a product in order to solve a specific problem without being particular to the methodology employed.

How does the RADx Initiative compare to other innovation policy levers, including grants and prizes?

As we have explained throughout this series, COVID-19 calls for heavy use of government-set policies like grants and prizes from the toolkit of innovation policy choices. The large positive externalities of effective interventions mean that market-based rewards alone will be insufficient relative to social value, so government needs to step in. And government-set rewards can be effective because policymakers know what technological goals need to be met—having effective interventions to prevent, diagnose, and treat COVID-19 infections—and the enormous economic and human costs of the pandemic provide some estimate of the social value of bringing it under control.

When choosing among government-set policies, lawmakers still face a choice in award timing: awards can be provided ex ante to the most promising projects before the results of the research are known (like grants), or ex post to only the projects that turn out to be successful (like prizes). Neither choice is uniformly superior. Ex post awards provide a stronger incentive for success and can reduce informational constraints for government funders—defining a technological goal is often, though not always, easier than choosing which teams are most likely to meet that goal. But ex ante awards can be more effective when innovators (1) are risk averse, (2) face constraints raising capital to cover research costs before results become known, particularly in capital-intensive fields such as health care innovations, or (3) are more present-minded than society as a whole.

Intriguingly, RADx has both ex ante and ex post aspects—and that’s probably a good thing. As one of us has explained in work with Professor Daniel Hemel, the reward-timing dimension of innovation policy “is a continuous spectrum rather than a binary choice … transfers can be scheduled at any time during the R&D process, based on varying demonstrations of success,” and mixing “ex ante and ex post rewards represents an effort to achieve the best of both worlds.” Like a grant, RADx funding comes before the stage of R&D for which the funding is targeted—manufacturing and scale-up of diagnostic testing—and is not contingent on the success of that work. But like a prize, RADx rewards firms that have already achieved early-stage success: the seven firms funded so far have already received Emergency Use Authorization (EUA) status from the FDA or have applications being processed.

Government funding for innovation usually isn’t targeted at this intermediate space of the product development timeline. Rather, the bulk of public R&D funding is focused on grants for earlier-stage “basic” research, leaving more “applied” work for the private sector. As Professor Bhaven Sampat has recently explained, this division emerged from a debate on science and technology governance at the end of WWII, when Vannevar Bush’s vision triumphed over that of Senator Harley Kilgore. Sampat argues that Kilgore’s vision, which “involved the government stepping in to do the applied research that profit-oriented firms would not,” is worth more attention today. Indeed, public funding may be valuable not only for traversing the “valley of death” between identifying a promising intervention and demonstrating its commercial viability, but also for even more downstream challenges such as efficient manufacturing and scale-up. 

RADx is a step in this direction. The NIH has run programs similar to RADx before, such as the Antimicrobial Resistance Diagnostic Challenge, although that program’s $20 million in total funding pales in comparison with the $1.5 billion available for RADx. In addition, RADx involves non-monetary support—including regulatory assistance—that is different in kind from both typical grants and typical prizes. The program thus seems like a novel addition to the innovation policy toolkit, which might serve as a blueprint for future policies.

What should policymakers keep in mind as they design initiatives like RADx for COVID-19?

Designing effective innovation incentives, whether grant-like or prize-like, is hard, and policymakers must take care to get it right. When designing prizes, policymakers need to pre-specify milestones with enough clarity and specificity to ensure that the initiative’s goals are met—here, the development and deployment of millions of effective diagnostic tests for COVID-19—but with enough flexibility that innovators aren’t foreclosed from taking creative approaches to answering the challenge. (The inability to foresee the right solution in the first place is, after all, the key reason for initiatives like RADx rather than the government simply creating a solution in the first place.) To take a counterfactual, if RADx had focused solely on point-of-care tests administered by health-care providers, either explicitly as a selection criterion or implicitly by using prize-like reimbursement criteria rather than direct provision of funds, developers would understandably but problematically focus their efforts on provider-administered tests rather than at-home tests. Such errors are easy to make. Professor Zorina Khan has cataloged many historical examples that have failed on these criteria.

It is also crucial to couple intermediate incentives like RADx with other tools from the menu of push/pull incentives. RADx, as something like a milestone prize, provides a useful tool to help advance products from early conception into production phases, especially for capital-intensive scaling and manufacturing processes. But purely ex ante funding, most straightforwardly grants, may still be needed to drive that early conception and proof-of-principle work, especially for small firms that might still face early-stage capital constraints (though it’s not clear that such constraints are particularly problematic in the COVID-19 diagnostic testing space). And after the role of initiatives like RADx, other innovation policy levers may be necessary to ensure that the resulting products are in fact broadly accessible.

Finally, while RADx and other initiatives like it certainly seem like a good idea, they represent a significant departure from the NIH’s typical approach to research funding, and it will be hard to know how well they work without rigorously measuring outcomes. It is perfectly plausible that RADx will be a crucial initiative to drive scale-up of COVID-19 diagnostics, and that the companies selected will bring very large volumes of excellent tests to the market. But it is also plausible that the selected companies would have reached the same outcome without RADx’s help: substantial federal investments are already being made into the development and reimbursement of COVID-19 diagnostics, and the selected firms might already have sufficient capital and access to regulatory expertise. Because there is no comparison group for the funded entities, future researchers will have difficulty understanding this program’s impact. The NIH should carefully observe and record the process and the results for funded companies (as CARB-X, a similar initiative in the antibiotic space, does). A more rigorous way to distinguish such outcomes is to consider structuring future RADx-like initiatives as policy experiments. Policymakers might consider, for example, randomly varying the funding received by each recipient and observing differences in outcome. Just as we need to know what health interventions and diagnostic tests work for COVID-19 and other diseases, we need to know what policy interventions work to improve the overall innovation ecosystem.

This post is part of a series on COVID-19 innovation law and policy. Author order is rotated with each post.