FDA, Testing, and COVID-19: A “Mid-Mortem”August 25, 2020
From the start of the COVID-19 pandemic, access to accurate and reliable testing to identify patients exposed to SARS-CoV-2 or infected with COVID-19 was identified as a critical element of an effective public health response. Testing is needed to diagnose individuals with active symptoms, enable early identification (through contact tracing) and quarantine of individuals who may have been exposed, and determine whether an individual has developed antibodies to the disease. And as we head into the Fall, testing is becoming integral to schools and universities seeking to safely resume in-person instruction, and to the health and safety of employees as they return to work.
The Food and Drug Administration (FDA) is the government agency that has played the most pivotal role in determining what tests have been available in the United States. Unfortunately, the issue of testing has too often been politicized, which has impeded a nuanced, rational assessment of FDA’s regulatory efforts and the extent to which FDA has succeeded in, on the one hand, facilitating widespread and timely access to accurate and reliable tests, while on the other, preventing poor quality tests from being disseminated.
Although “post-mortems” are typically performed at the end of a project, the COVID-19 pandemic appears to be poised to continue for the foreseeable future. This makes it all the more important to evaluate FDA’s response to date and to recommend any “mid-course corrections” that could improve the agency’s response during the pendency of this pandemic . . . let alone any future ones that may arise. The list below is anything but exhaustive, but we believe it includes the key areas for improvement.
As FDA regulatory practitioners who have worked with multiple companies trying to introduce COVID-19 tests, we have witnessed first-hand both the challenges faced by companies seeking to bring new tests to the market and the pernicious impact of “bad actors.” With the end of summer approaching and states experimenting with “back to work” and “back to school” programs, this is an opportune time to reflect on the past five months and identify key lessons for regulators, policy makers, and the regulated industry, which we hope can inform the process going forward.
Primum non nocere – First, do no harm
The lack of adequate testing at the outset of the pandemic can be directly linked to FDA’s controversial laboratory-developed test (LDT) policy, which several FDA Law Blog contributors have addressed on numerous occasions. See J. Gibbs, M. Cato & S. Schlanger, New FDA Policy Significantly Limits Serological Testing, FDA Law Blog (Apr. 13, 2020); J. Valentine, D. Clissold, & J. Shapiro, FDA Works Around the Clock, Provides More Detailed Guidance on the Conduct of Clinical Trials Amidst the COVID-19 Pandemic, FDA Law Blog (Apr. 1, 2020). In short, FDA asserts jurisdiction to regulate clinical laboratory tests, but in most cases exercises “enforcement discretion,” allowing clinical laboratories to innovate and respond quickly to emerging clinical needs. However, FDA did not elect to exercise general enforcement discretion for COVID-19 LDTs, meaning that clinical laboratories other than CDC were not permitted to offer testing for SARS-CoV-2 without prior authorization from FDA, either through an Emergency Use Authorization (EUA) or pursuant to an applicable FDA COVID-19 enforcement policy. As was widely reported, CDC’s test – notwithstanding FDA authorization – had significant performance issues, but FDA’s prohibition on offering LDTs without FDA authorization meant there were no alternatives available for several crucial weeks at the beginning of the outbreak in the U.S. Laboratories were ready, willing, and able to offer LDTs, but were blocked by FDA’s policy. Michael Shear et al., The Lost Month: How a Failure to Test Blinded the U.S. to Covid-19, N.Y. Times, Mar. 28, 2020.
FDA’s jurisdiction to regulate LDTs under the medical device provisions of the Federal Food, Drug, and Cosmetic Act (FD&C Act) has long been the subject of debate but has not yet been subject to legal challenge. In a recent Yale Law Journal Forum essay, Barbara Evans and Ellen Wright Clayton present an additional jurisdictional argument based on Section 564 of the FD&C Act, which addresses FDA’s EUA authority specifically, and which FDA cited in guidance as the source for its authority to require clinical laboratories to obtain EUAs. Barbara Evans & Ellen Wright Clayton, Deadly Delay: The FDA’s Role in America’s COVID-Testing Debacle, 130 Yale Law J Forum 78 (2020). As Evans and Clayton note, Section 564 allows FDA to grant EUAs only for medical “products,” defined as drugs, devices, and biological products, but “grants no new powers for the FDA to regulate clinical laboratory services.” Id. at 80. Indeed, they note that the only mention of clinical laboratories under Section 564 occurs in the context of FDA’s authority to categorize the complexity of a “laboratory examination or procedure associated with such device.” Id.; FD&C Act § 564(m)(1). By drawing a distinction between “laboratory examinations or procedures” that are not subject to Section 564 and “products,” which are subject to Section 564, the statutory language makes FDA’s assertion of jurisdiction to require EUAs for clinical laboratory tests highly suspect.
In any event, FDA’s decision to prevent LDTs for SARS-CoV-2 testing was not dictated by decisions by the prior administration. FDA did not have to prevent labs from running samples for SARS-CoV-2 using their LDTs. FDA chose to take that approach.
In short, FDA’s decision not to exercise enforcement discretion for use of LDTs for SARS-CoV-2 significantly delayed the availability of accurate and reliable tests. The adverse consequences of FDA regulation of LDTs in the time of a pandemic was foreseeable. This is not a case of hindsight is 20/20 – the need for LDTs during disease outbreaks had been foreseen in comments to FDA criticizing past FDA efforts to limit LDTs. See, e.g., American Clinical Laboratory Association, Citizen Petition, Docket No. FDA-2013-P-0667-0001 (Jun. 17, 2013) at 15-16 (noting that LDTs allow for real-time response to emergent infectious diseases that is critical to the welfare of patients and the public health and that requiring FDA marketing authorization under such circumstances could have “potentially catastrophic consequences”).
Remarkably, on August 19, 2020, the Department of Health and Human Services (HHS) issued a statement announcing that FDA could not regulate any LDTs – including those for SARS-CoV-2 – without first going through rulemaking (see our blog post here). The HHS notice specifically stated that this position meant that laboratories did not need EUAs to offer LDTs. As a consequence of this announcement, FDA’s policy requiring laboratories to obtain EUAs for LDTs for SARS-CoV-2 has been overridden. This move, of course, cannot cure the damage caused by testing delays at the beginning of the pandemic. But moving forward, laboratories will have significantly more flexibility to quickly offer appropriately validated LDTs for SARS-CoV-2 without FDA authorization as a rate-limiting step.
Measure twice, cut once – The law of unintended consequences
FDA’s requirements for serology (antibody) tests have changed multiple times in the space of a few short months. We understand that FDA has needed to make changes to the regulatory requirements for serology tests as it learns more about the quality of the tests on the market and the data supporting their use, but making changes also imposes costs and confusion. The lack of regulatory predictability has complicated matters for reviewers and companies.
Initially, FDA announced that manufacturers of serology tests could begin marketing their tests without an EUA, so long as they had appropriate analytical and clinical test validation data on file and submitted a “notification” to FDA of their intent to distribute. FDA, Guidance for Clinical Laboratories, Commercial Manufacturers, and FDA Staff, Policy for Diagnostic Tests for Coronavirus Disease-2019 during the Public Health Emergency, at 1 (Mar. 2020). Companies rushed to validate their tests according to FDA’s recommendations and put distribution networks in place.
FDA then abruptly shifted its notification policy in May, announcing that all serology test manufacturers would first need to prepare and submit an EUA, but could continue distribution pending FDA review. FDA, Guidance for Clinical Laboratories, Commercial Manufacturers, and FDA Staff, Policy for Coronavirus Disease-2019 Tests During the Public Health Emergency (Revised) (May 2020). FDA distributed a “template” EUA request disclosing the data required to be included in the submission. Again, industry raced to meet the new requirements, since the penalty for not submitting an EUA by the stated deadline was to be placed on the list of companies not permitted to distribute in the U.S. FDA was giving companies as little as 24 hours to respond to agency interactive review questions.
FDA explained the change in policy was due in part to the improper promotional claims that companies made, as well as changes in understanding of the benefits and risks. The document did not mention another factor that undoubtedly played a role: FDA was roundly criticized in the press and by Members of Congress for allowing serology test kits onto the market that it had not reviewed, some with low or inconsistent performance. (While LDTs typically are not reviewed by FDA before being introduced, serology tests for new analytes ordinarily do require FDA review before commercial distribution.) See Laurie McGinley, Dozens of coronavirus antibody tests on the market were never vetted by the FDA, leading to accuracy concerns, Washington Post, Apr. 19, 2020; Zachary Brennan & David Lim, FDA pushed through scores of inaccurate antibody tests without agency review, Politico, Apr. 27, 2020; Thomas Burton, FDA Sets Standards for Coronavirus Antibody Tests in Crackdown on Fraud, Wall Street J, May 4, 2020.
While some reviewers, particularly in in the March-May timeframe, responded rapidly and substantively to serology EUA submissions, the collaborative nature of the interactions deteriorated, and response time and quality of reviews varied widely between submissions. Some EUAs were issued fairly quickly, while others languished for weeks or months. Eventually, FDA seemed to be overwhelmed by the influx of EUA submissions and would frequently reference its large “backlog” in submissions when companies requested status updates on FDA’s review of their EUA request.
FDA also at some point made an internal decision to “deprioritize” serology tests, but this decision was never publicly announced or acknowledged. FDA’s stated public rationale was that since, under the revised May policy, companies were permitted to begin distribution of their serology tests while the EUA submission was under review, FDA action on the EUA really was not that important, and FDA’s resources could be better utilized. This perspective, though, ignored the real-world impact of lack of an EUA: in many cases, the product is not commercially viable without an EUA authorization because some customers (e.g., state governments and hospital systems) preferred to purchase tests that had completed the FDA review process.
By design or inadvertently, FDA created two distinct marketplaces: some tests were being sold pursuant to FDA’s revised policy, which allowed distribution during the pendency of FDA’s review, while others had an EUA. Another disadvantage to distribution of a test without an EUA is that there is a question of whether a test sold while an EUA request is under review, and not yet authorized, enjoys the strong liability protection available under the PREP Act, which is unquestionably available for tests that have an EUA. See HHS, Advisory Opinion on the Public Readiness and Emergency Preparedness Act and the March 10, 2020 Declaration Under the Act, April 17, 2020 as modified on May 19, 2020.
Transparency and Communication are Key
The failure to acknowledge changes in review priorities is only one of many instances in which FDA has not clearly communicated with regulated industry. The examples below are by no means exhaustive:
- “Point-of-Care” serology tests: As noted above, FDA’s initial policy for COVID-19 tests allowed serology tests to be marketed pursuant to a notification only without prior FDA authorization. FDA’s written policy stated that such tests were intended for use by clinical laboratories and health care providers “at the point of care.” On its face, this appeared to allow the tests to be sold to and used in point-of-care settings, such as doctors’ offices and drive-through testing facilities. It made sense for this policy to apply to point-of-care health care settings, as these are the settings most suited to the use of these simple, rapid serological tests. This interpretation would also have been consistent with broader access to tests. However, in order for these tests to be used by non-laboratorians outside of a CLIA-certified clinical laboratory, FDA must classify them as “CLIA waived.” Many readers of FDA’s policy reasonably inferred that FDA must mean these tests were CLIA waived, because that was the only way providers “at the point of care” could actually use them. Nevertheless, FDA later issued an “FAQ” on its website stating that, in fact, the tests were not waived, and that manufacturers would need to conduct additional studies demonstrating that their tests could safely be used at the point of care and receive an EUA in order to market them to anyone other than a CLIA “high complexity” laboratory. See Gibbs, M. Cato & S. Schlanger, New FDA Policy Significantly Limits Serological Testing, FDA Law Blog (Apr. 13, 2020). Following this clarification, companies were forced to quickly cease distribution for point-of-care use and revise labeling and marketing materials to note the tests’ high-complexity status. Companies that did not notice the clarification on FDA’s website continued distribution for point-of-care use. There were several points in the process where FDA could have avoided the significant confusion over this issue. Procedurally, FDA’s delay allowed confusion to continue. Substantively, FDA’s decision meant that tests could not be run at many sites where they would have been most useful. FDA’s decision also directly clashed with a policy from the Centers for Medicare & Medicaid Services issued almost at the same time that permitted testing at pharmacies, which was not at the time possible because FDA has not yet authorized any serological tests for point-of-care use.
- NCI testing pathway: On April 28, 2020, FDA issued an umbrella EUA, which was described as a “voluntary” program under which manufacturers could, at no cost, submit their tests to the National Cancer Institute for validation testing rather than conducting independent clinical validation testing and submitting an individual EUA. If the validation testing conducted by NCI showed that the test met minimum performance standards outlined by FDA, the test was authorized under the umbrella EUA. This process was described by FDA as designed to cut down on FDA’s review burden, because FDA would not have to perform in-depth reviews of individual EUA submissions. The NCI test data (whether favorable or unfavorable) would be made publicly available.
While in concept this plan had merit – creating a central, standardized U.S. testing site in which FDA had confidence and for which manufacturers did not have to incur expense – the execution of the program was anything but smooth. Companies that volunteered to have their tests evaluated by NCI were given priority in the form of continued communication with FDA. NCI quickly (and foreseeably) became inundated with testing requests, and manufacturers experienced long delays for their results.
At the same time, many companies chose not to pursue the “voluntary” NCI pathway (e.g., in favor of using a commercial testing laboratory). There were some limitations to the umbrella EUA; for example, it only allowed for authorization for use in moderate- or high-complexity CLIA laboratories. If companies wanted to pursue a point-of-care authorization, an individual EUA was required. However, even though companies submitted robust clinical validation studies with their individual EUA, some were told by reviewers that they would need to submit test kits to NCI for testing to “confirm” their clinical validation results in order to obtain an EUA, while others found their EUAs inexplicably stalled while reviewers inquired repeatedly whether their tests had been submitted to NCI. The de facto requirement to obtain NCI testing was never publicly announced or acknowledged, and companies were confused by requests to submit test kits to NCI for testing when they had not elected the umbrella pathway.
Gating the review of serology tests through the NCI, which apparently stemmed from a lack of trust in data submitted by device manufacturers, has caused another host of problems for companies. It was the expectation of FDA that the data generated by NCI would be consistent with the other data provided in the submission. However, NCI data do not necessarily reflect performance in the field. NCI uses a curated panel of specimens that, because of small sample size and limited range of titers, is not necessarily consistent with what might be observed in the real-world when evaluating “all-comers” as part of a clinical study. The test report provided by NCI acknowledges these limitations, stating that the “[s]ensitivity and specificity estimates in this report may not be indicative of the real-world performance of the [Company] [Device Name].” The report explains that the NCI results are “based on serum and plasma samples only and may not be indicative of performance with other sample types, such as whole blood, including finger stick blood.” Additionally, the report states that the “number of samples in the panel is a minimally viable sample size that still provides reasonable estimates and confidence intervals for test performance, and the samples used may not be representative of the antibody profile observed in patient populations.” Notwithstanding these acknowledged limitations by NCI, FDA has in practice treated NCI results as determinative and has rejected EUAs based on NCI data that failed to meet FDA’s threshold for sensitivity, even if the NCI test results conflicted with other data submitted by the manufacturer, often using much larger sample sizes compared to NCI testing.
- Revocation of NCI umbrella EUA: Three months after announcing the “umbrella” pathway, on July 21, 2020, FDA announced without warning that the umbrella EUA was being revoked. (FDA explained that nobody had actually used this pathway it had created. This was only partially true; many tests were in the pipeline awaiting authorization under the umbrella policy, but FDA not yet authorized any tests.) FDA couched the decision as solely “administrative” in nature, but the implications of its actions seemed potentially far reaching, leaving applicants who had pursued the NCI testing pathway uncertain over the fate of their submissions and concerned about whether they needed to undertake costly independent clinical validation. It took several days and multiple emails to regulators to clarify that pending submissions would not be adversely affected by the change. This confusion could have been avoided had FDA more clearly communicated what the change meant.
- Lack of industry input: FDA has three main resources for manufacturers looking to validate serology tests: “Policy for Coronavirus Disease-2019 Tests During the Public Health Emergency (Revised) Immediately in Effect Guidance for Clinical Laboratories, Commercial Manufacturers, and Food and Drug Administration Staff” (May 11, 2020); FAQs on Testing for SARS-CoV-2; and the Serology Template for Commercial Manufacturers. None of these resources mention FDA’s data expectations regarding shelf-life claims for EUAs. As it turns out, FDA will not accept the standard practice of providing accelerated aging studies that validate the full claimed shelf-life with real-time testing to be conducted post authorization. Companies need real-time data in their EUAs, or they may find that they are left with a short-term shelf-life claim that presents commercial challenges or renders the product commercially non-viable.
This specific example highlights another problem: the lack of industry input. Under ordinary circumstances, FDA would be issuing draft guidances and receiving feedback from stakeholders. The lack of a mechanism to receive comments has meant that FDA has been deprived of a critical outside perspective. FDA may know what it is trying to communicate in its policies and FAQs, but sometimes it is not clear to the outside world. For example, one of FDA’s early guidance documents referred to “at home” testing. The ambiguous phrasing led companies reasonably to conclude that FDA’s policy applied to tests conducted at the home, not samples obtained from the home. FDA then criticized companies for providing products where samples were collected at home, even though FDA’s phrasing was unclear.
- Delays in providing guidelines on data expectations: While there have been occasions where FDA has been quick to provide feedback and clarifications to industry, such as through its list of FAQs and weekly “virtual town halls,” there have been occasions where FDA has been slow to provide public guidance on expectations for data supporting EUAs. For example, for a period of about two months, FDA promised an EUA template for home testing “very soon” or by the “end of the week.” Some companies delayed submission of EUAs for home tests, fearing that FDA would release a template with new requirements shortly after submission of an EUA application. Similarly, following FDA’s clarification that an EUA was required to distribute a serology test for point-of-care use, companies requested information from FDA on the expectations for usability and flex studies. FDA did not add these data expectations to the serology test EUA template until weeks later, but in the meantime, provided the usability and flex study expectations to companies individually upon request, leaving companies who did not know to request these guidelines at a disadvantage. Given the unprecedented volume of work, we understand that developing these guidance documents and FAQs will take time. It is important, however, for FDA to provide realistic estimates as to when these documents will be released.
Dirty Harry had it right – An agency’s “got to know its limitations”
During the early months of the pandemic, interactions between reviewers and test developers were, on the whole, collaborative and communicative. Sadly, as the pandemic has dragged on to new heights, this situation has changed. Many companies have had their questions go unanswered by FDA for months. We at HPM have never had so many companies so frustrated by the lack of agency feedback. Multiple companies have reported sitting in a queue for months, with no visible progress and no meaningful communications with FDA. That same frustration clearly comes through in some of the calls to the weekly virtual town hall meeting on diagnostics.
The reduction in collaborative discourse can be attributed in large measure to a sharp spike in submission volume for CDRH. We are well aware that the agency has never faced such an onslaught of submissions. We are sympathetic to the challenges FDA faces, not just with diagnostics but also with other pandemic-related products (e.g., vaccines, drugs, personal protective equipment, hand sanitizer).
Even so, this lack of interaction has had a crippling impact on companies pursuing EUAs. The regulatory landscape is still changing rapidly, and these changes are not always followed with appropriate communications by the Agency. And as applications languish, FDA’s standards can change, dooming companies whose submissions met the standards that had been in place at the time of submission. When emails are sent to FDA requesting updates on the status of a review one is very likely to hear nothing, or to receive this boilerplate response: “We appreciate your patience during this time. Unfortunately, we are not able to provide estimates of review timelines.” While the U.S. suffers from a shortage of tests, companies that raced to develop and validate tests are left in regulatory limbo.
Again, this outcome was completely foreseeable. FDA has received thousands of EUA requests since March – the equivalent of several years of work compressed into mere months. Doing the math, it is physically impossible for FDA, with current staffing levels, to review these submissions in anywhere close to a timely manner. Although FDA has not (and will not, we expect) publicly admit its capacity constraints, the reality is that some submissions will not get reviewed, and some products will not receive authorization because there are simply not enough people to do the work.
However, FDA’s failure to acknowledge limitations in its capacity has significant adverse commercial consequences for test developers and public health consequences. It also has created distrust and disillusionment. For a number of companies, this has been their first experience with FDA – and it has not been a favorable one.
Companies whose products have been rejected or criticized by FDA are also frustrated by their FDA interactions. There is a recurring pattern reported by companies caught in these situations: that FDA has made up its mind and will not listen to any other perspective, and that when FDA communicates publicly its conclusions, the agency fails to provide the necessary context. Once again, the result is a recurring perception that FDA has acted unfairly.
Those Who Fail to Learn History Are Doomed to Repeat It
This will almost certainly not be the last major outbreak of an infectious disease. As Evans and Clayton note, “[w]hile nothing will buy back time already lost, the COVID-19 outbreak casts a useful light on flaws in recent laboratory-testing policy. . . . Lessons from the current pandemic might still help us in the next one.” It is critical that a thorough, non-partisan, scientifically-based and objective review of FDA’s handling of COVID-19 testing be conducted to evaluate what went right and what needs to be improved. FDA and all stakeholders need to learn from these experiences so that we are all more prepared.