Thursday, September 19, 2019

Policy Discovery!! Statute Prevents CMS From Issuing Blanket NonCoverage NCDs

By far the most impactful NCD for the diagnostics industry and cancer patients is the March 2018 NCD 90.2, "Next Generation Sequencing (NGS) for Medicare Beneficiaries with Advanced Cancer" - here.

In its original March 2018 form, the NCD provides coverage for FDA-approved diagnostics if they are (1) based on NGS technology, (2) used only once, and (3) used in advanced cancer patients (defined as stage 3,4, metastatic, recurrent, relapsed). 

The original analysis focused only on companion diagnostic tests (e.g. oncogenes like KRAS and ALK), and contained a statement that the scope applied NOT to all NGS tests, but ONLY to tests for prescription of targeted chemotherapies. There was also a statement that "the scope of the review" was limited only to "advanced cancer."  (In total, the phrase "outside the scope" is used 9 times!  Many things were outside scope of review.)

However, and fairly notoriously, in November 2018, CMS issued a transmittal about the NCD that added a section (inserted as ^ "C") that made all services that are not covered by the NCD, nationally non-covered.  See CR10878 (the most recent version of the November 2018 document) at link.

Statute Prevents This Switch

Note that the NCD body (which describes literature search, data review, and scope), repeatedly and explicitly limits the scope to "advanced cancer."  This matches the literature reviewed.  CMS also reviewed only single-use test scenarios.  CMS made germline testing topics entirely out of scope as its review, and the bibliography corroborates this, and the keyword search, and the absence of germline guidelines in the guideline review subsection.

You can't do that and then switch in the decisional section to voiding coverage of all topics never reviewed and not discussed.   See the statute in Section 1862, which is titled "EXCLUSIONS FROM COVERAGE".  (As in the legal rules for "how to" make exclusions from coverage.)  This section defines how and when CMS can (and can't) make exclusions from coverage through NCDs.

In particular, Congress instructs CMS (text immediately prior to 1862(b)):
"In making an NCD, [CMS] shall ensure that...[CMS] has considered applicable information with respect to the subject matter of the determination AND, in the determination, provides a CLEAR STATEMENT of the basis of the determination, AND make available to the public the data considered in the determination."
What this means is really simple.

The CMS determination of an exclusion to coverage MUST be made based on applicable information AND the data used must be public (and part of public comment). 

Since CMS didn't consider germline data in the review (it deliberately excluded it; read the body of the analysis), OR any multiple-use tests, OR any tests in less-advanced cancer (e.g. even neoadjuvant therapy with Herceptin based on Her2-neu positivity, an on-label use of the FMI NGS test, was excluded from review), the "determination" that rests atop the CMS "analysis" ... can't deal with those subjects.  Graphically as follows:

click to enlarge

The rules CMS must follow when issuing an NCD exclusion from coverage.


An analogy would be an NCD analysis that studied aortic valve repair, and reviewed the literature on aortic valve repair.  Thereupon, the determination following this analysis - to be legally compliant with statute - could cover, or non-cover, aortic valve repair technology.   But it could not non-cover all other cardiac services (for example, CABG or stents), which weren't part of the review or offered to the public as data and CMS rationale.  It's not only that this makes sense; but it's also directly written into Medicare law as shown in this article.

Similarly, for the NGS NCD, the scope of the determination must rest upon the scope of the data analysis and the clearly stated scope of the review.  The statute requires it, in black and white.

It's my understanding that CMS has said, what we say or remark in the analysis doesn't matter, the decisional part of the NCD is only the "determination" (if the two contradict in scope).  But no.  The determination MUST rest on the scope and content of the analysis, as statute requires.


See statutory history footnote here.