METADATA last updated: 2026-03-20 RT file_name: _context-commentary_regulatory-fda-policy.md category: regulatory subcategory: fda-policy gfile_url: https://docs.google.com/document/d/1ej8e1HjKyq1As-FsYFA1jcz2lV-dFv9ya9vmQ68Y2a0 words: 2468 tokens: 3647 CONTENT ## Context This `regulatory/fda-policy` subcategory contains 47 documents spanning 2020–2023 that trace the evolution of FDA's regulatory policy for COVID-19 diagnostic testing. The collection includes seven versions of FDA's overarching COVID test policy guidance, EUA letters and amendment letters, review templates (molecular, antigen, home-use, pooling), fact sheet templates, FDA press announcements, and transition planning documents. Together they form a detailed record of how FDA managed the regulatory environment for COVID-19 tests from the earliest emergency through the end of the Public Health Emergency. This collection is not complete and consists mostly of documents that we downloaded, many of which we reviewed. There are almost certainly other important documents that are not included here. ### AI Summary of FDA Policy Below is an AI (ChatGPT 5.2 Pro and Claude Opus 4.6) generated summary, which may contain errors. The FDA COVID-19 testing policy evolved through three broad phases: - **Emergency expansion (early 2020):** FDA used enforcement discretion to allow CLIA high-complexity labs to develop, validate, and begin using molecular tests before EUA issuance, provided labs notified FDA and submitted an EUA within a set timeframe. Guidance versions 1–4 (Feb–May 2020) progressively broadened these pathways to include state authorization, commercial manufacturers, and serology. - **Standardization and new use cases (mid-2020–2021):** FDA tightened quality oversight (notably revoking the serology umbrella EUA pathway), introduced standardized templates, and expanded authorized intended uses from "suspected COVID-19" into explicit asymptomatic screening and pooled testing. Serial testing emerged as a formal authorization strategy, and variant monitoring became a standard EUA condition. - **Narrowing and wind-down (2022–2023):** FDA reduced the scope of new EUA reviews, encouraged developers toward traditional marketing pathways (510(k)/De Novo), standardized repeat-testing labeling for consumer antigen tests, and published transition plans for returning to normal device regulation. The COVID-19 PHE expired May 11, 2023, but FDA clarified that existing EUAs remain in effect as long as the underlying EUA declaration persists. Key points of variation across the seven guidance versions: | Version | Date | Key Change | | --- | --- | --- | | v1 | Feb 29, 2020 | Initial accelerated policy for CLIA high-complexity lab molecular LDTs | | v2 | Mar 16, 2020 | Expanded pathways: state authorization, commercial distribution prior to EUA, serology policy | | v3 | May 4, 2020 | Additional templates, clearer timelines, tightened oversight | | v4 | May 11, 2020 | Further template and oversight refinements | | v5 | Nov 15, 2021 | Major reset: umbrella EUAs for serial testing, reissued/narrowed molecular LDT umbrella EUA, FDA signals that newly offered tests should generally have EUA or traditional authorization before clinical use | | v6 | Sep 27, 2022 | Narrowed EUA review priorities; encouraged traditional marketing pathways | | v7 | Jan 12, 2023 | Final revision of the overarching COVID test policy guidance | ### AI Compilation of EUA Staffing and Review Timing `_AI_FDA COVID-19 Diagnostic Test EUA Staffing and Review Timing.md` compiles publicly available data on FDA staffing levels and review timelines for COVID-19 diagnostic test (molecular and antigen) EUAs, covering CDRH OHT7/OIR workforce growth from 60 to 180 staff, review-time trends from 14-day to 99-day medians, submission volume and outcomes by test type and developer category, reasons for non-authorization, and the key federal sources for each metric. Three policy innovations are specifically documented in this subcategory: #### Serial screening (Mar 2021 onward) FDA authorized screening claims based on strong symptomatic performance combined with a repeat-testing regimen, rather than requiring standalone evidence in asymptomatic populations. This treated the testing algorithm and frequency as part of the risk-control package. By Nov 2022, serial testing instructions were standardized into enforceable antigen test labeling (two tests over three days if symptomatic; three tests over five days if asymptomatic, at least 48 hours apart). Key archive files include: - `2021-03-16_FDA Website - FDA takes steps to streamline path for COVID-19 screening tools` - `2021-03-16_FDA Template - Supplemental Template for Developers of Molecular and Antigen Diagnostic COVID-19 Tests for Screening with Serial Testing` - `2021-03-13_FDA Fact Sheet - Screening for COVID-19 Deciding Which Test to Use When Establishing Testing Programs` - `2021-10-25_FDA Template - Supplemental Template for Developers of Molecular and Antigen Diagnostic COVID-19 Tests for Screening with Serial Testing` (updated version) - `2022-11-01_FDA Letter - Repeat Testing Revision Letter` - `2022-11-01_FDA Website - Antigen EUA Revisions for Serial Repeat Testing` - `2022-11-17_FDA Website - At Home COVID-19 Antigen Tests-Take Steps` #### Asymptomatic/screening testing (Jul 2020 onward) FDA distinguished between testing asymptomatic individuals under clinical "suspicion" (provider judgment) and broad population screening (no symptoms, no known exposure). The Jul 24, 2020 LabCorp reissuance was the first EUA explicitly authorizing screening of people without known or suspected infection, establishing screening as a distinct intended use with its own evidence and labeling requirements. Key archive files include: - `2020-08-24_FDA Policy IVD - Pooled Sample Testing and Screening Testing for COVID-19` - `2021-03-13_FDA Fact Sheet - Screening for COVID-19 Deciding Which Test to Use When Establishing Testing Programs` - `2020-10-26_FDA Template - Antigen Template for Test Developers` - `2021-10-06_FDA Template - For Developers of Molecular Diagnostic Tests` #### Pooled testing (Jul 2020 onward) FDA authorized sample pooling to conserve reagents and increase throughput, but constrained it through validation requirements (replicate detection thresholds, Ct shift limits, invalid-rate caps), restriction to CLIA high-complexity labs, and behavioral controls (reflex individual testing for positive/invalid pools, specific fact-sheet language). The Apr 20, 2021 amendment letter tied pooled screening specifically to serial testing programs operating at least weekly, treating serial frequency as a mitigation for the sensitivity loss inherent in pooling. FDA distinguished swab pooling from media pooling and provided different validation pathways for each. Key archive files include: - `2020-08-24_FDA Policy IVD - Pooled Sample Testing and Screening Testing for COVID-19` - `2021-04-20_FDA Letter - Amendment Letter` - `2021-04-20_FDA Website - Pooling and Serial Testing Amendment` - `2021-04-20_FDA Fact Sheet - Sample Updated Fact Sheet for Health Care Providers` - `2021-04-20_FDA Fact Sheet - Sample Updated Fact Sheet for Patients` An important distinction throughout the archive is between FDA's two regulatory levers: **EUA issuance** (test-specific, enforceable conditions, required fact sheets) and **enforcement discretion** (FDA choosing not to enforce certain requirements for a period, with conditions). The early guidances relied heavily on enforcement discretion to expand access; later guidances progressively moved back toward requiring EUA or traditional authorization before clinical use. Related subcategories include `regulatory/open-euas` (the open EUA submissions FloodLAMP prepared), `regulatory/fda-townhalls` (a RAG demo over 100 FDA townhall transcripts for COVID-19 test developers), and `regulatory/ldts` (laboratory-developed test policy context). ## Commentary For the primary commentary on FDA policy and the Open EUAs, see `regulatory/open-euas/_context-commentary_regulatory-open-euas.md`. There remains a need for significant progress in FDA policy as it relates to pandemic preparedness and response. The complexity of the regulatory environment documented in this subcategory, with seven guidance versions, multiple overlapping enforcement mechanisms, and evolving intended-use distinctions, illustrates both the scale of the challenge and the difficulty of navigating it in real time as a small developer. ### Deprioritization The most consequential example of this difficulty is FDA's use of "deprioritization" — a discretionary triage mechanism through which CDRH declined to review hundreds of EUA submissions for COVID-19 diagnostic tests (see `regulatory/fl-fda-correspondence/_AI_FDA Deprioritization of COVID-19 Diagnostic EUAs.md` for a detailed AI-generated analysis). By September 2021, GAO reported that FDA had declined to review 558 EUA requests, including 230 from laboratories developing LDTs. These submissions represented enormous effort by startups, established companies, academic groups, and clinical laboratories — organizations and individuals that stepped up during a pandemic to contribute in what is arguably the single most critical area for slowing early-stage spread: testing. A small number of FDA reviewers and agency leadership, exercising broad and largely unpublicized discretion, effectively bottlenecked the efforts of thousands of scientists and businesses working to expand national testing capacity. What followed was the Omicron surge of winter 2021–2022, with severe nationwide testing shortages, hours-long lines, front-page headlines, and real consequences in lost lives, closed schools, economic destruction, and massive societal upheaval. As our own pilot programs documented (see `pilots`), there was tremendous unmet need for accessible testing even among first responders and essential workers. Despite the volume of policy documents, templates, and guidance versions cataloged in this subcategory, the underlying process was highly discretionary, capricious, and unpredictable — and the deprioritization mechanism itself was barely publicized, remaining largely unknown, perhaps even among researchers studying the pandemic response. ### Regulatory Reform No systemic reform of the FDA diagnostic review process has occurred since the pandemic. There were ample mechanisms available — provisional authorization with post-market data collection against established comparators, for example — that could have preserved access to emerging tests while managing quality concerns, but these were not meaningfully pursued. Specific and actionable reform proposals exist, such as the one documented in `regulatory/reg-articles-misc/2021-01-18_Phillips and Dinakar - A Proposal for Increasing the Speed of Validating SARS-CoV-2 Diagnostic Tests`, but to the best of our knowledge none have been implemented or placed on any announced reform agenda. What the FDA did pursue was the 2024 LDT rule, an attempt to expand its control by imposing device-style premarket review requirements on laboratory-developed tests — effectively eliminating the ability of clinical labs to develop and offer their own diagnostics. That rule was vacated by the courts in March 2025, and FDA chose not to appeal. For a fuller account, see the AI-generated report at `regulatory/ldts/_AI_FDA 2024 LDT Rule - Status and Legal History.md`. Discussion of FDA reform is understandably usually dominated by pharmaceutical policy, drug pricing, and drug approval. Drugs are a much larger industry and a much larger share of overall healthcare spending compared to diagnostics. But with respect to pandemic preparedness and response, diagnostic and screening testing remains underappreciated and underaddressed as a domain for regulatory reform. Without reform, there is every reason to expect the same bottlenecks, the same discretionary purges, and the same capacity constraints if another pandemic arrives. This is a significant and tractable area for progress. ### AI Tools for Regulatory Navigation AI tools offer substantial potential for making this kind of dense regulatory material more accessible and navigable. As a related effort within this archive project, we built a demo QRAG (Question Retrieval Augmented Generation) tool over a corpus of 100 FDA virtual town hall transcripts for COVID-19 diagnostic test developers (see `regulatory/fda-townhalls`). That tool takes natural language queries and returns the closest matching quoted FDA authority responses along with an AI-generated summary, demonstrating one approach to making regulatory guidance more searchable and usable. Feel free to try it at [FDA COVID-19 Diagnostics Townhalls - QRAG Demo](https://www.focusonfoundations.org/fda-town-halls-qrag-demo). ### FDA SARS-CoV-2 Reference Panel In 2020, the FDA created a SARS-CoV-2 Reference Panel — a standardized set of samples containing heat-inactivated virus — to independently benchmark the analytical sensitivity (limit of detection) of EUA-authorized molecular diagnostic tests. When the results were published, they revealed large and systematic discrepancies between the LoDs measured with the panel and the self-reported LoDs on which EUA authorizations had been based. The peer-reviewed study by Blommel et al. (reference below) found a statistically significant difference, with a mean FDA panel LoD of 43,750 copies/mL versus a mean EUA-reported LoD of 9,417 copies/mL. A compiled dataset showed that approximately 86% of tests performed worse on the FDA panel, with a median discrepancy of roughly 9x and nearly half of tests showing discrepancies of 10x or greater. A formal complaint was filed under the Information Quality Act in December 2020; the FDA responded nearly three years later, defending the panel and noting the data had been removed as outdated. Key archive files include: - `_AI_FDA SARS-CoV-2 Reference Panel Report` (AI-generated comprehensive report analyzing the episode) - `FDA SARS-CoV-2 Reference Panel Comparative Data - Complied by Matt McFarlane` (compiled LoD comparison data) - `2022-12-15_Paper - Blommel - Authorized SARS-CoV-2 molecular methods show wide variability in the limit of detection` (peer-reviewed study directly analyzing the discrepancy) - `2020-12-22_Complaint Letter to FDA regarding Reference Panel - IQA-Request-Hyman-Phelps-McNamara` (Information Quality Act complaint) - `2023-09-23_FDA Response Letter - To Complaint regarding FDA Reference Panel from Dec 2020` (FDA response, issued nearly three years later) The reference panel episode exposes a fundamental problem in how the FDA evaluates diagnostic test performance for EUA authorization. The entire EUA system relies on self-reported performance data from developers, validated using whatever materials they choose — clinical remnant samples, contrived specimens from commercial providers like Zeptometrix, gamma-inactivated virus from BEI Resources, synthetic RNA from Twist Bioscience, or other sources — each prepared under different protocols, in different matrices, and reported in inconsistent units. When the FDA's own standardized benchmark revealed that most tests performed substantially worse than their authorization data indicated, the agency neither resolved the discrepancy nor reformed the process that produced it. It archived the results, took nearly three years to respond to a formal complaint, and continued operating under the same, now even more clearly flawed framework of self-reported data and rigid performance cutoffs. What the FDA should have done, and still to the best of our knowledge has not done, is use the knowledge it gained from the reference panel to build accessible infrastructure for generating credible, comparable performance data. That means establishing clear, standardized procedures for creating contrived positive specimens for validation work, making reference materials broadly available rather than restricting them to existing EUA holders, and creating a streamlined, permissionless pathway for developers to generate real-world comparative data against established comparator tests at partner clinical laboratories. Instead, the pathway to generating clinical performance data remained gated behind expensive IRB approvals, six-figure clinical study costs, and weeks-long BEI Resources registration delays just to obtain cross-reactivity reagents (see `regulatory/irb` and `regulatory/fl-fda-submissions`). There should have been a mechanism by which a developer whose test demonstrated performance in the right ballpark, even based on contrived specimens, could proceed directly to supervised comparative testing against an established comparator, generate real clinical data, and submit those results as part of a provisional or follow-up authorization, without the full standalone clinical study apparatus. See the `regulatory/irb` subcategory, specifically the archive files `_context-commentary_regulatory-irb` and `_AI_digestion_irb_new-clinical-study-design`. FloodLAMP's experience illustrates the consequence of this gap. The FloodLAMP colorimetric LAMP test was not as sensitive as purified or even direct PCR, but it was consistently positive earlier than FDA-authorized rapid antigen tests in real-world surveillance comparisons, particularly in detecting early and asymptomatic infections — precisely the use case where accessible testing has the greatest public health value (see `pilots/pilot-data/_context-commentary_pilots-pilot-data.md` for case studies and comparisons). Yet it was unable to obtain authorization or even meaningful review (see `regulatory/fl-fda-submissions` and `regulatory/fl-fda-correspondence` for the full record). Meanwhile, tests with reference-panel results showing order-of-magnitude sensitivity shortfalls from their EUA claims remained authorized and on the market. The reference panel discrepancy and the FDA's failure to act on it point to a need for fundamental and substantial reform in how the agency measures, validates, and adjudicates diagnostic test performance.