Molecular Pathologist Group Proposes Legislative Solution to Modernize CLIA, Including CMS/Third-Party Premarket Review, as Substitute for FDA’s LDT Framework

August 14, 2015

By Jamie K. Wolszon

As we have previously blogged on numerous occasions, FDA has released a framework for regulation of laboratory-developed tests (LDTs) that has sparked considerable controversy.  The Association of Molecular Pathology (AMP), which represents molecular pathologists and has been an active participant in the LDT debate, announced in a press release that it has met with the Senate Health, Education, Labor and Pensions Committee to propose a legislative solution to modernize the Clinical Laboratory Improvement Amendments (CLIA), including premarket review by the Centers for Medicare & Medicaid (CMS) or its designated third parties, as a substitute.  Under the AMP proposal, FDA would only review laboratory-developed testing protocols (LDPs) – a new term coined by AMP to stress that these are healthcare professional services –  under limited circumstances.  At the same time, it will increase CMS regulation of LDPs compared to today.

The House Energy & Commerce Committee has released a discussion draft which proposes a separate alternative to the Framework.  The 21st Century Cures legislation adopted by the House did not address LDTs.  In the meantime, as we have previously reported, a group of labs and in vitro diagnostic kit companies called the Diagnostic Test Working Group have tried to craft their own legislative solution.  We expect that additional legislative solutions may be proposed in the near future.  Whether any legislation will actually be enacted is anyone’s guess.

As for FDA, we understand that FDA is working on revising the two draft guidances that outlined the framework (the “Framework” guidance and the “Notification” guidance).  As of today, according to, FDA has received 236 comments on the draft “Framework” guidance; however, it has only posted 170 of those comments.  We do not understand why more than six months after the comment period closed (in early February 2015) so many of the comments have yet to be posted.  The website offers no explanation for the missing 66 comments other than the following boilerplate language: “[S]ome agencies may choose to redact, or withhold, certain submissions (or portions thereof) such as those containing private or proprietary information, inappropriate language, or duplicate/near duplicate examples of a mass-mail campaign.”

Discussing the rationale for its own legislative proposal, AMP explains that it is duplicative to have two separate agencies regulate LDPs, and CMS, as the agency that implements CLIA, is the appropriate agency for the task. 

At the heart of the proposal is a risk-based scheme that will determine whether a LDP must undergo premarket review.  To be high-risk the LDP must involve a proprietary computational method or algorithm AND be used “to diagnose a disease, predict risk of disease, or risk of progression of a disease, that is associated with significant morbidity or mortality, or threatens the public health.”  Therefore, an LDP that FDA would now consider a companion diagnostic is high risk under the AMP proposal only if it uses an algorithm or computational method.  The approach of focusing on transparency is reminiscent of FDA’s IVDMIA proposal, in which it proposed to require premarket review for a subset of LDTs, involving an algorithm.  The theory behind the IVDMIA proposal was that these tests were non-transparent black boxes. 

Under the AMP proposal, like high-risk tests, moderate-risk tests would also be those used “to diagnose a disease, predict risk of disease, or risk of progression of a disease, that is associated with significant morbidity or mortality, or threatens the public health.”  However, moderate-risk LDPs would use a methodology that “lends itself to inter-laboratory comparisons or proficiency testing.” 

Low-risk tests, by contrast, would either be adjunctive in nature or protocols for which “the consequence of an incorrect result or interpretation is unlikely to lead to serious morbidity or mortality, either for the patient or the public health.”  The AMP proposal would classify LDPs used for rare diseases, for public health emergencies, and for infectious agents that are not serious threats to the public health as low risk.  The proposal would modify the definition rare disease to match the drug definition of a disease or disorder with an incidence of fewer than 200,000 newly diagnosed individuals per year in the United States.  FDA’s “Framework” draft guidance proposed an exemption from premarket review for rare disorders, but it included a limitation of 4,000 tests per year.  During FDA’s public meeting in January 2015, multiple stakeholders criticized that limited number.

Under the AMP program, FDA would only review the LDP submission if the laboratory voluntarily chooses to go through FDA PMA/510(k) process or if a protocol is high-risk and the laboratory does not want to give the proprietary information to the CMS or third-party reviewer.  In all other instances where premarket review would be required, CMS or a qualified third party acting as a surrogate of CMS would perform the review.  CMS would establish a transparent process to designate non-federal government organizations as third parties.  Current accrediting bodies under the CLIA process, such as the College of American Pathologists (CAP), could be a third party, as could a state for LDPs offered to residents of that state. 

The proposal would direct CMS to stipulate a minimum level of standards for LDP analytical and clinical validity.  The proposal defines clinical validity as “the association of a biomarker or analyte with the presence, absence, predisposition to, or risk of a specific clinical condition.”  The proposal notes that clinical validity is distinct from clinical utility.  Clinical validity can be established by many things including peer reviewed literature; clinical practice guidelines; subject matter expert opinion; bench studies, including use of archived specimens; past experience with similar products; case studies; clinical data; consensus standards; reference standards; data registries, e.g., ClinGen, ClinVar, CancerLinQ, or other curated relevant database; post-market data; or clinical trials, including those conducted outside of the United States.

The need for clinical studies would be substantially curbed.  A third-party reviewer could require a laboratory to conduct a clinical trial only for a high-risk protocol and only if the CMS or the third-party reviewer determines that no other approach can provide the necessary information to support the laboratory’s claims, and provides written justification for that decision.

The LDP review information must be sent to CMS within three days after completion of review of the LDP.  CMS can reclassify the risk of an LDP, which implies that CMS reviews the information sent to them by either the CMS or third-party reviewer.  How often CMS will exercise this right to verify the work of the third-party reviewer within the deadlines outlined in this proposal is an open question.  This has been an issue under FDA’s third-party 510(k) review program as there have been claims that FDA’s secondary review of the reviewer’s work means that any potential efficiency from a third-party review is lost.

If a LDP is high-risk, the CMS or third-party reviewer has 90 days to review the submission.  If the CMS or third-party reviewer does not complete the review within the deadline the approval is automatically granted.  There is no grandfathering for high-risk LDPs.  This timeframe of 90 days is the same as the 510(k) process, and shorter than the PMA review timeframe of 180 days; there is no mention of stopping the clock as there is under the PMA and 510(k) processes.    

If a LDP is moderate risk, the CMS or third-party reviewer has 30 days to review the submission.  If the CMS or third-party reviewer does not complete the review within the deadline, the submission is automatically granted.  There would be grandfathering for moderate-risk tests on the market prior to enactment of the law. 

Low-risk LDPs would not undergo premarket review under the proposal.  Rather, low-risk LDPs would be subject to inspection in the normal course of the laboratory inspection process. 

Laboratories with demonstrated success with approved LDPs in the same or higher-risk classification, will be conditionally approved to begin testing with LDPs that use similar technologies or methodologies while review of the LDP submission is pending. 

In addition to pre-market review, the AMP proposal would require laboratories to prepare summaries about the test that could be posted by CMS.  This summary would be similar to the ones that a 510(k) submitter must provide, and that FDA publishes on its publicly available searchable electronic database.  Low-risk LDPs would not have to provide these summaries.  Moderate-risk tests on the market as of April 24, 2003, would not have to provide these summaries.  (It is not clear why this date was selected).  Once there have been three tests of the same kind the laboratory seeking to market a new LDP does not need to provide clinical validity either in the submission or prepare a summary of the clinical validity for the test but can instead reference the database.

If a laboratory modifies an LDP after the protocol goes through the review process, that modified LDP would have to undergo new review if the change elevates the test to a higher risk classification or if the modification significantly changes the performance characteristics.  Moreover, if a laboratory modifies an FDA-approved or cleared test in a way that significantly changes the performance characteristics, and the modified test is high risk or moderate risk, the modified test would need to undergo premarket review as described above.  If the modification to the approved or cleared device does not change the performance characteristics, the laboratory would need to provide summary information that could be published as described above. 

The proposal also would modify CLIA requirements related to proficiency testing, inspections, recordkeeping, and reporting of laboratory errors.  It also would authorize CMS to collect an annual user fee, limited to cost recovery, from laboratories determined by the number of LDPs the laboratory offers.  

The proposal would call on CMS to issue final updated CLIA regulations within two years after the legislation is enacted. That is a very short time period for issuance of any final rule.  The requirements will be effective two years after the regulations are finalized.

The fate of the proposal is uncertain.  However, it does show that there are alternative approaches to regulating LDPs – or LDTs – than what FDA proposed.