- The FDA is supposed to speed innovations
- Emergency Use Authorization doesn’t require clinical trial data
- Rapid tests should meet a reasonable limit of detection minimum to receive Emergency Use Authorization
- Standardized data should be available
- Revocation can provide a backstop
- Applying a public health standard to public health tools
The United States doesn’t have enough rapid tests. The root cause of this shortage is that the Food and Drug Administration (FDA) doesn’t have an appropriate framework for thinking about tradeoffs during an emergency.
In the United Kingdom, there are over 150 different rapid antigen tests available for sale, many for only a few pounds. And individuals don’t even have to buy the tests themselves, as the government made it a priority to ensure that every individual can receive up to seven free tests every day. In the United States, there are now 16 rapid antigen tests available for sale, but during the peak of the delta wave, there were only 3. Not only are they prohibitively expensive for many Americans, they’re in such short supply that even those with means often can’t find them. Americans could desperately use more competition in the market, but, as the difficulty in acquiring legitimate high-quality masks shows, it is important to ensure that the tests actually work.
The FDA currently has a mechanism to ensure vetted products are available quickly in such an emergency via the aptly-named Emergency Use Authorization, but it has yet to be used to its full potential. One of the major lessons of this pandemic is that we need a serious re-evaluation of Emergency Use Authorization authority to be better prepared for next time.
Here’s what the FDA should do:
- Pre-approval efficacy evidence should be relaxed in favor of speeding deployment of public health tools during a public health emergency under an Emergency Use Authorization
- FDA should lead by providing robust, standardized data to consumers rather than forbidding access.
- FDA should feel free to revoke authorization once a product is shown to have less benefit than plentiful, comparable products
- FDA should recognize that different products can fill different niches when judging “comparable products”
- Authorization criteria should emphasize the public health benefit above and beyond the individual benefit
Rapid tests provide a key example of where the FDA has treated clinical trial data as necessary when other data should be sufficient for authorization during a public health emergency. Despite the broad benefit to the public of having rapid tests widely available, their supposed individual inferiority to PCR has kept almost all of them off the market. The FDA should have pursued an alternative route: (1) Authorize tests that, based on preclinical data, are sensitive and specific and then (2) Once there are sufficient tests for consumers, remove the worst performing tests from the market. The FDA should also support and publish standardized limits of detection and clinical trial data to allow Americans to make informed decisions about their health.
According to the Nuclear Threat Initiative and Johns Hopkins’ Global Health Security Index, the United States scored 100 out of 100 on “capacity to test and approve medical countermeasures” in 2019 and 2021. That was, in large part, thanks to the FDA. According to their own mission statement:
“The Food and Drug Administration is responsible for protecting the public health by ensuring the safety, efficacy, and security of human and veterinary drugs, biological products, and medical devices”
That checks out. Many were confident that the COVID-19 vaccines were safe and effective, even when they only had an Emergency Use Authorization. And others chose to wait until full authorization in order to be vaccinated, because that stamp of approval is meaningful. So the FDA seems to be fulfilling the first part of its mission. But the statement continues (emphasis added):
“FDA is responsible for advancing the public health by helping to speed innovations that make medical products more effective, safer, and more affordable and by helping the public get the accurate, science-based information they need to use medical products and foods to maintain and improve their health.”
Wait, what? The FDA’s mission is actually to speed innovation? And provide us with science-based information? This seems at odds with bureaucratic delays in vaccine approval, a demonstrated inability to relabel a safe and effective treatment, and a uniquely American dearth of COVID-19 tests.
To be clear, the FDA isn’t doing nothing to speed access to critical medical countermeasures. Exercising Emergency Use Authorization has allowed some lifesaving measures to come to market sooner than they would have under the regular approval process. Now, over two years into this crisis, the FDA has also finally worked with the NIH and over-the-counter, at-home COVID-19 test manufacturers to coordinate clinical trials that meet FDA specifications.1 That seems like it will help us dig out of this hole, but we probably shouldn’t have made the hole in the first place.
One big issue slowing innovation is the FDA’s insistence on requiring clinical trial-based efficacy data. This demand presents a problem for a manufacturer trying to design a test for a novel pathogen: It can’t do a well-powered clinical trial until there are enough patients in its clinic. This creates a terrible Catch-22: The goal for emergency use of novel pathogen detection mechanisms should be to stop an outbreak, but manufacturers can’t get the clinical data they need until one already happens.
For COVID-19, we no longer have that particular problem, but the cost of waiting has been immense. Although bringing more COVID rapid tests to the market ASAP would likely help with the current pandemic, policy decisions the FDA makes today may prove even more important in establishing a precedent for the next novel pathogen outbreak.
Fortunately, the congressional authorization for Emergency Use Authorization does not require applications to include clinical data. The law states:
“(2) that, based on the totality of scientific evidence available to the Secretary, including data from adequate and well-controlled clinical trials, if available, it is reasonable to believe that–
“(A) The product may be effective in diagnosing, treating, or preventing…
“(B) The known and potential benefits of the product, when used to diagnose, prevent, or treat such disease or condition, outweigh the known and potential risks of the product”
The FDA is required by law to reasonably believe that the product benefits outweigh the risks based on the totality of evidence available. But while totality of the data can include clinical trials, it doesn’t need to include them. For rapid testing in an emergency, the threshold for evidence should not require clinical trials.
Rapid tests should meet a reasonable limit of detection minimum to receive Emergency Use Authorization
The required minimum threshold of detection is a key difference between the United States and the United Kingdom’s antigen test approval process and is largely responsible for the discrepancy in test availability. The United States requires antigen tests to be 80% as sensitive as the gold standard RT-PCR tests, which are only negative if no viral genomic material is detected after it has been doubled 40 times. This is kind of like needing to magnify a sample under a microscope 40x before determining there’s nothing there, except instead of times 40, it’s times 2^40 (or a bit more than a trillion times magnified). In the UK, the threshold for comparison is only 25 doublings (or a bit more than 30 million times), but at a concordance rate of 90%. This means, in theory, an antigen test in the US needs to be over 30,000 times more sensitive.
The FDA has recognized that the previous threshold for approval was too stringent. In November of 2021, the FDA lowered the concordance rate with PCR from 90% to 80%, but they have not changed the actual threshold for sensitivity: The concordance still needs to be comparable to the 2^40 magnification, rather than the more lenient 2^25 required in the UK. Technically, the FDA requires tests to be both more sensitive and less accurate.
The FDA’s current standard of using clinical trials for approval does not actually test what the minimum limit of detection is, it tests whether the individual with a positive PCR result also has a positive rapid test result. That concordance rate will depend on how much virus each patient has as much as how sensitive the test is. This makes the design of the clinical trial as important for approval as the design of the test.
The totality of evidence needed for a test to receive an EUA, especially very early in an outbreak, should be that (1) the device system is able to safely collect and process whatever sample is relevant and (2) the device is able to specifically detect a minimum concentration of pathogen material that is similar to what is found in an average patient sample. In other words, the device should be similar enough to a previously approved device and sensitive enough to stand a chance at detecting the new pathogen.
The number of times the pathogen’s genetic material needs to be doubled in order to be detected by PCR can be used to calculate appropriate minimum limits of detection. The limit of detection is how much the pathogen can be diluted and still be detected – some minimum number of viral particles per milliliter. The limit of detection should be realistic based on PCR tests from the initial outbreak, but not overly restrictive. The exact level will depend on the pathogen and may need to change as more information is gathered and sufficiently sensitive tests enter the market.
More urgently, it requires the company to perform a clinical trial when the speed premium is highest — right as infections of a novel pathogen are beginning to spread. Trials are not only expensive, increasing the barrier to entry for companies, but they require a large number of infected patients, preventing tests from being available early in an outbreak. This alternative approach focused on minimum levels of detection would be much faster because labs could directly test a sample in a tube. It can also be much safer as it can minimize interactions with infectious viruses.
While there are valid concerns about inferior products being allowed on the market during an emergency, there are other solutions to alleviating those worries besides outright banning their sale. Instead, the FDA should provide standardized and verified measurements about each test directly to the public to allow them to choose the best tests.
The limit of detection, not just concordance with PCR, for every available test should be readily available to the public on the FDA’s website and on product packaging. This information does exist for some COVID-19 tests, but it is not formatted for the general public and may not have been measured in a standardized way, meaning the numbers are not necessarily directly comparable. A benchmarked, universal standard should be required to allow for cross-comparison between tests. The United Kingdom’s Health Security Agency and the University of Oxford are measuring the limit of detection for the most promising tests, but the only information provided to the public is if a test has passed evaluation.
Users should be able to know the comparative power of different devices in order to select the best, or at least know how their results with the available tools may differ. This should include clinical data, once available. The FDA should work with the NIH and CDC to perform these clinical trials alongside diagnostic PCR testing, which would allow direct comparison between devices.
Continuous clinical testing will also allow for updated data as new variants arise without a several month lag. Similarly, if new information suggests protocol modifications are needed (e.g., sampling location should be changed from nose to throat), the company can suggest modifications to their protocol that can be immediately tested. Government supported clinical trials would be most important early in an outbreak, when infected individuals may be hard to find and should be directed to government resources. If a company wishes to later hold their own clinical trial using a protocol that is not conducive to the government-supported framework they should be able to do so and submit that data separately. However, this data should be a supplement to, not a replacement for, the standardized trial. Again, clinical-trial data should be featured prominently on the FDA’s website and at the point of sale.
The federal government response to the COVID-19 pandemic has focused on slogans like “protect yourself”. But Americans can’t protect themselves when the tools critical to doing so, including rapid tests, are in short supply due to government restrictions on sale.
The data needed to make informed decisions is also difficult to find and not directly comparable between tests. By publishing standardized limit of detection numbers and continuously updated, standardized clinical trial information, the FDA can help Americans help themselves by buying the most sensitive and accurate tests. Knowing exactly how well a test is performing in clinical trials will allow users to better interpret their own results. The actual numbers should be provided for true comparison and the FDA should help the public with interpretation through simplifications like “sensitive, more sensitive, most sensitive”. Americans will be able to choose the best tests, incentivizing and providing a framework for device manufacturers to continuously improve their tests.
The FDA, of course, should not let just any rapid test stay on the market. The agency should continue to require reasonable evidence that a test will be efficacious based on pre-clinical data and the best available scientific evidence. And, importantly, if any test performs unacceptably worse than other available tests once the clinical trial data comes in, it should lose authorization and be removed from the market.
It doesn’t get as much attention as it’s pre-market authority, but the FDA has broad power to mandate a recall or revoke authorization if tests or devices fail to perform as intended. In fact, during this pandemic, the monoclonal antibody bamlanivimab had its Emergency Use Authorization revoked when there was an increase in resistant variants, which reduced its benefits relative to its risks. The FDA noted that part of this decision was made because there were other, more efficacious monoclonal antibody cocktails and combination treatments available. The same case can be made for any treatment or test: once there are plentiful and better alternatives, authorization of more marginal products can be revoked.
This recall authority provides a vital backstop that should enable the FDA to be initially more lenient with approval during the early stages of a pandemic, with the knowledge that Emergency Use Authorization is not an irreversible decision and can always be revisited.
The FDA should use the Emergency Use Authorization to approve drugs and devices when they would benefit the public in a public health emergency, which may require the agency to act with imperfect data. They also may need to approve products that meet a public health need, rather than a personal medical need — the standard the FDA traditionally uses to judge medical devices, like tests.
In the context of testing, this approach requires recognizing that PCR tests and rapid tests have different uses. PCR tests tell an individual whether they’re infected, but not whether they’re likely to be infectious. That makes PCR tests good for disease diagnosis and personal treatment planning, but much less useful for public health purposes (e.g., should this person isolate or quarantine). And even if we wanted to only use PCR tests, there is currently a shortage and people are waiting as long as two weeks to get results — a response time that is useless for public health purposes.
Rapid tests, in contrast, tell an individual whether they’re likely to be infectious, but are less accurate than PCR tests as to whether the individual currently has any viral material in them at all. But infectiousness is precisely the information that is most valuable to the public during a pandemic, and the FDA’s current personal medicine standard effectively does not recognize this benefit.
Moving forward requires a shift in the way the FDA is using Emergency Use Authorization. First, during a public health emergency, the primary goal should be the health of the entire public. This requires using a different cost-benefit analysis than during non-emergency times.2 For rapid tests, it should involve setting minimum limits of detection rather than requiring stringent concordance with PCR-based clinical trials. Second, the FDA should use its Emergency Use Authorization more liberally.
The FDA, rightfully, wants to protect the legitimacy of their stamp of approval. However, they can do this by better distinguishing Emergency Use Authorization from full authorization: Emergency Use can be used in an emergency and when current data suggests the benefit to the greater public outweighs the cost (and waiting for further information has its own cost) while full authorization can remain the gold standard of “safe and effective”. When information or conditions change, the Emergency Use Authorization can be revoked. But, during the emergency, products that enable a more effective public health response should be allowed to reach the market. That’s exactly what the authority was intended for.
The FDA urgently needs to revisit how they use Emergency Use Authorization and recognize the key role it plays in enabling the agency to make different tradeoffs during an emergency. The FDA needs to make this transition now, both to save lives during this emergency and to be better prepared for the next one.
- However, a requirement to join the program is that the tests are already authorized in a major international market.
A great example from Dr. Carroll’s piece is Indiana rationing rapid antigen tests by providing them to the people most at risk of severe illness, those over 50, when instead they should have been given to those most at risk of spreading the virus, those 19-49.