Friday, September 29, 2023

FDA Releases Proposed Regulation to Control Lab Developed Tests

On September 29, 2023, the FDA released its long awaited regulation to control the lab-developed test industry.   The rule was released in typescript form, with the Federal Register publication on October 3 (88 Fed Reg 68006-68031).  The rule had been imminent, since it was under review at the Office of Management and Budget since August.  In addition, the FDA's leadership has been stating clearly, particularly since last fall, that they would produce regulation if Congress does not pass law, such as the VALID Act.

See FDA press release here:

See the Federal Register rule page here:

Red Flag: You can't full understand the published 26-page rule without its online 127-page financial supplement, which also has some novel facts and novel arguments besides the finances - here.  While the "averaged" cost to industry over 20 years is around $5B, some $40B costs occur suddenly in years 4/5.   Those are tangible costs; tables project hundreds of billions of health QALY intangible benefits (millions per patient) project out decades into the future.

Altogether, the materials include over 150 pages, over 40 tables, some 90 footnotes and 150 citations.

Comment until December 4.


Early news at Genomeweb here. Follow up by Bonislawski, here.  Fierce Biotech here. AdvaMed here. NYT here.   PBS here.  Stat here.  Medpage here.   RAPS (regulatory professionals) here.  Akin Gump law firm here.  Skadden law firm here.  Goodwin law firm here (pt 1).  CAP requests 120 days' comment.  AMP is "disappointed" and states that lab procedures are not "IVDs" boxed and shipped.   >> See my sidebar on how FDA won the press battle.


On Day 1, Quest and Labcorp were off 1%, Natera off 3%.   (These small slippages took several hours to sink in.)

Trump 1, Trump 2

Recall that  shortly after the Trump election, FDA canceled an Obama plan to regulate LDTs. Recall that ;ater during the Trump administration, HHS forbid FDA to regulate LDTs and (deja vu) the Trump administration could be back.  

Sidebars to This Blog

Sidebar 1 - Legal Excerpts ["Can FDA Do This?"].  See Genzen; see Clement and Tribe.

Sidebar 2 - FDA Press Release, Deconstructed

Sidebar 3 - The Crucial Financial Supplement, 127pp

On a humorous note, a fourth sidebar asks ChatGPT to describe the FDA proposal in the voice and talking points of several diverse stakeholders. I also asked Chat GPT to summarize all of my published materials on the rule - here.

With that said...dive in.

A Micro Regulatory Change!

The FDA's press release does not summarize the rule, leaving readers to go through the whole 83-page rule to understand it.  However, the "core" part of the rulemaking is only a few words long, by "making explicit" that a test is an in vitro diagnostic device (IVD) "when its manufacturer is a laboratory."  At its core, the regulation does nothing except add "^when the manufacturer of these products is a laboratory" to the definition of diagnostic test devices at 21 CFR 809.3 (last page of PDF document). 

  1. In vitro diagnostic products are those reagents, instruments, and systems intended for use in the diagnosis of disease or other conditions, including a determination of the state of health, in order to cure, mitigate, treat, or prevent disease or its sequelae.  [INSERT] ^These products are devices as defined in section 201(h)(1) [Act] and may also be biological products subject to section 351 of the Public Health Service Act, including when the manufacturer of these products is a laboratory.

FDA controls both device performance and marketing claims, and FDA gives substantial weight to both issues in this document.    As to problems with marketing claims, to show these may be contentious or disputed, FDA cites lawsuits over marketing claims in areas such as pharmacogenetics.

Comments run to about November 29.  The FDA notes that it expects "interest" and "new parties" in third-party review of LDTs.

One way to think about all this, you can voluntarily go to the FDA  today for 510(k) authorization of a Class II test (Mammaprint) or a Class III test (Foundation F1 CDx).  Start by striking the word "voluntarily."    As proposed, the zillions of 510K or de novo or PMA applications must be submitted in a couple years, although each one will act to 'stop the clock' on enforcement until that document is reviewed.  

FDA Summary (Sic)

The change is styled as a rollback of "enforcement discretion" of LDTs that FDA has allowed, since the Device Amendments of 1976.  However, the upfront "Summary" section of the FDA proposal is not really a summary, but simply restates that FDA is rolling back its prior discretion.  

Costs and benefits are summarized, in terms highly favorable to FDA: Financial benefits "averaged per year over 20 years" will be $2B to $86B, with a best estimate of benefits of $22B.   These includes avoidance of... health damage being caused by LDTs and lawsuits.  

Here's what you don't read in the summary: Most of the benefits are virtual ($26B per year forever) based on millions of dollars of life per patient not harmed by LDTs.  Costs of compliance regulation will be $6B ($2B-$20B), again, "annualized", but industry costs are highly front-loaded, with about $40B in years 4-5.  These are buried in table 35 & 37 of the online docket aka Supplement.

Click to enlarge, Supplement, Table 37.  Up to $83B in costs, then divide by 20.


This section "need for the rule" is extensive.  FDA  sights news coverage that allegedly raise concerns about under-regulation of NIPS testing and "IVDs offered as LDTs are ubiquitous."   Further, "FDA has witnessed an explosion in the volume, complexity, and scope of IVDs offfers as LDTs for use in determining cancer treatments...used in some of the most complex areas of medicine."  FDA asserts that "current information raises serious questions whether patients can rely on IVDs offered as LDTs."  In addtion to citations to marketing-focused lawsuits, FDA writes that "four different studies have documented high variability in performance among [LDTs]."

As to the value of self-regulation, FDA writes, "We are aware of arguments that [requirements under CLIA] already provide adequate oversight...Our review of the evidence indicates otherwise."  They also dismiss the regulatory option of adding more authority to CMS/CLIA ("enhanced CLIA.")  

There is an extensive section regarding the history of this rulemaking and the FDA's legal authorities.  I provide a sidebar blog where I give a detailed Chat GPT summary of the FDA's 5000-word legal justification.

FDA has no problem cherry-picking references that fit its needs best for a particular paragraph.  There's no attempt to portray the citations as systematic or reflecting averages from meta-analyses.

Description of Policy (Section VI, page 68021ff)

Below, I provide an AI-assisted discussion of the actual implementation rules (which are about 7000 words long).

Proposed Phased Approach: The FDA proposes a five-stage phased approach to regulate LDTs:

Stage 1 (1 Year): End the general enforcement discretion approach with respect to Medical Device Reporting (MDR) requirements and correction and removal reporting requirements. Compliance with these requirements would help the FDA monitor adverse events and safety issues associated with LDTs.

Stage 2 (2 Years): End the general enforcement discretion approach for requirements other than MDR, correction and removal reporting, Quality System (QS) requirements, and premarket review requirements. This includes registration and listing, labeling, and investigational use requirements. Compliance with these requirements will provide better oversight of LDTs.

Stage 3 (3 Years): End the general enforcement discretion approach with respect to QS requirements. For LDTs manufactured entirely within a single Clinical Laboratory Improvement Amendments (CLIA)-certified laboratory and not distributed outside that lab, some QS requirements may continue to be enforced. Compliance with QS requirements is crucial for ensuring the quality of LDTs.

Stage 4 (3.5 Years): End the general enforcement discretion approach with respect to premarket review requirements for high-risk IVDs offered as LDTs. This applies to IVDs eligible for classification into class III. The timeline aligns with the start of fiscal year 2028 and the beginning of a new user fee cycle, providing stability to the marketplace.

Stage 5 (4 Years): End the general enforcement discretion approach with respect to premarket review requirements for moderate and low-risk IVDs offered as LDTs, requiring premarket submissions. This includes 510(k) submissions and De Novo requests. The proposed timeline is extended by 6 months to prioritize high-risk IVDs' review.

Use of Third Party Review Program: Laboratories may seek to use FDA's Third Party review program for their submissions, potentially streamlining the review process.

  •  "FDA also anticipates that laboratories may seek to utilize Third Party Review...we anticipate interest [!] well as new Third Party Review organizations." (p 68027)

Effective Date: The proposed rule, if finalized, would become effective 60 days after publication in the Federal Register.

Economic Impact: The rule includes an [extensive] Preliminary Economic Analysis of Impacts, which assesses the potential costs and benefits associated with the proposed regulation.

Public Health Rationale: The FDA's proposed rule aims to ensure the safety and effectiveness of LDTs, addressing concerns about their quality and performance. It is designed to provide better oversight, enhance patient safety, and align LDTs with established FDA regulations for IVDs.

Risk Categories

The proposed FDA rule for Laboratory Developed Tests (LDTs) describes device risk categories for these tests, particularly in relation to premarket review requirements. In the context of this rule, there is a distinction between high-risk (Type 3) and moderate to low-risk (Type 2) LDTs, which corresponds to the regulatory pathways typically associated with them:

High-Risk (Type 3) LDTs: These are LDTs that may be eligible for classification into Class III devices. In the proposed rule, the FDA plans to phase out the general enforcement discretion approach for premarket review requirements for high-risk LDTs offered as LDTs (Type 3) 3.5 years after the finalization of the rule but not before October 1, 2027

Class III devices typically require a Premarket Approval (PMA) application, which is the most stringent regulatory pathway for medical devices. [E.g. current PMA for a few lab tests like FMI F1 CDx].

Moderate and Low-Risk (Type 2) LDTs: These are LDTs that may be eligible for classification into Class II or Class I devices. The proposed rule suggests ending the general enforcement discretion approach for premarket review requirements for moderate and low-risk LDTs (Type 2) that require premarket submissions (e.g., 510(k) submissions and De Novo requests) 4 years after the finalization of the rule but not before April 1, 2028. Class II devices typically require a 510(k) submission, which is a less stringent regulatory pathway compared to PMAs.

By categorizing LDTs into these risk categories (Type 2 and Type 3), the FDA is proposing different timelines for phasing out the enforcement discretion approach, reflecting the level of risk associated with these tests. High-risk LDTs (Type 3) are subject to a longer phase-out period, aligning with PMA requirements, while moderate and low-risk LDTs (Type 2) have a shorter phase-out period, aligning with 510(k) and De Novo requirements. This approach allows the FDA to prioritize high-risk tests while providing additional time for laboratories to comply with regulatory requirements for moderate to low-risk tests.

It's only mentioned once or twice, but critically, applying for a 510k or PMA will "stop the clock" on enforcement against the LDT involved.


Defining the risk category (and thus, whether 2027 or 2028 deadlines apply) includes defining a clinical use statement (that is, FDA intended use), which itself is a complex process and which has no close predicate for LDTs today.  For example, devices introduced after 1976 are Class III by definition unless they have a legal 1976 predicate or a post-1976 cleared or approved predicate.  The rule makes several references to De Novo reclassification. 

Thoughts and Next Steps?

Note that the FDA is not reclassifying LDTs as IVDs.  Rather, it says that LDTs have always been IVDs, but LDTs have been a type of LDT under enforcement discretion, until now.

For "risks of LDTs," FDA cites poor submissions it receives, literature reporting discrepancies, news articles, and lawsuits.  (Submissions included faulty LDTs for COVID).  These include e.g. ref 33, a "memorandum to file" that is filed within an FDA office.

It's not entirely clear where the borders of LDTs are.  For example, paraffin block to H&E slides are not made as a "kit" for paraffin blocks and H&E slides; so are slides an "LDT IVD?"  We tend to think not, but it's opaque from the regulation.  For example, a pathologist has three slides on a case: H&E, Alk Phos, and Her2.  The latter is a PMA test (e.g. PMA980024).  Where do you draw lines and why?

Some HLA and some forensic tests would continue to have FDA enforcement discretion (p. 68022). But FDA is doubtful that e.g. even when test methodology is "well understood," it ought to be exempt from FDA review (p. 68023).  

NIPT/NIPS tests have never reviewed by the NDA, but are used by the hundreds of thousands.  It's unclear to me how FDA would  frame the risk/benefit of a positive result (to the mother? to the fetus??) or if that could become too-hot-to-handle politically.   Plus, the amniocentesis tests that occur as a reflex are surely not "fully manual" tests A to Z, therefore, are also illegal LDTs. 

CAP now has complex policies for  outsourcing parts of genomic tests, where both parties to the test are themselves CLIA labs.  It's unclear what happens to this under the FDA IVD concept (and FDA seems to touch on this on Supplement, page 13-14, new LDTs that are "not manufactured and used in a single laboratory."  One entity has a slide deck about the CAP Distributive Model; while the FDA online financial assessment supporting the LDT proposal (FDA ref 34) states, "FDA is aware of some entities that claim a connection to laboratories and offer IVDs as developers offering software...with LDTs through laboratory partnerships."

FDA has a somewhat challenging discussion of "grandfathering," I think flagging fo readers that FDA is doubtful it is a good idea, but FDA does foresee comments (p. 68023).  They also ask for comments if academic medical centers (AMCs) should be handled in a different way (p. 68023ff).   FDA seeks comments on leveraging the NYSDOH program and whether it could trigger enforcement discretion (p 68024).  

Expect legal challenges, one reason the FDA has "lawyered up" with pages and pages of legal justificaitons in its proposal including interstate commerce under the Constitution, etc.  (See also sidebar blog.)  If finalized, it would cause a frenzied boom of demand for the limited pool of experienced FDA IVD consultants.  (Vs a complementary boom in staff needed by FDA or by its third-party reviewers).   FDA expects $500M in new FDA costs; at $100,000 per person-year, that would be 5,000 FDA person-years of time.

Minimal summary financials are stated in the proposal (e.g. financial benefits might reach $80B...).  For details see a 127 page FDA website summary of the actual financial guesswork (FDA ref 34, here.)  For example, they estimate that 600 to 1200 to 2400 labs will be affected by this rule (p 24).  LDTs to be regulated may be 40,000, 80,000, or 160,000 (p 25; Suppl. Table 2).  New tests per year may be 3600, or 7200, or 14000 (p 25).  

If finalized expect a boom in third-party review.  For those whoo have said the 1976 PMA/510k regulation is a bad fit for LDTs - well, this is exactly what you get if the only change you make is to apply that 1976 law to LDTs (rather than statutory change).  On the Hill, stakeholders might rally around some kind of statutory change (VALID or other) rather than  the crude rolling-over of old 510k/PMA law onto LDTs, which is what we get in today's announcement. 

If finalized...expect delays running several years longer that the FDA's proposals.

The PR Game: Well played, sir.  The FDA released the 80 page rule on a Friday morning.   It came with an FDA press release that didn't really describe the plan, and focused on the bad actors in the lab industry and the urgent need for public safety.   This FDA press release was the only easy entry point for journalists.  Few would be qualified to read and absorb the 80 page rule (or the 127 page supplement), much of it arcane or legalese.  They'd have only the FDA's own press release to work from.   The weekend would soon come, then a new news week starting Monday, badly aging any Monday interviews leading to Tuesday stories (buried on page 6, or the internet equivalent).   Had the FDA wanted to bury the news further, it could have released at 5 pm Friday.  But it didn't do that: It wanted some news on Friday, but little would come out besides the FDA's own press release spin.  (The NYT only got out a first story by 3 pm.)  In fact, nearly all the stories on Friday and Saturday hewed close to the FDA's highly self-congratulatory press release because little else was available.  Mission accomplished, I would say.

Policy Breakdown

1, Executive Summary, 3pp
2. Abbreviations, 1p
3. Background, 25p
4. Legal authority, 1p.  But see Legal topics in III-C and V-B
5. Regulation proposed, 14p
6. Enforcement policy and roll-in, 21p

See also the 127-page online supplement.

The PR Game: Well played, sir.