Monday, May 18, 2020

Medicare Stumps Me x 3: #3, One MolDx MACs Deletes LCDs from Other MolDx MACs

Smmary - The MolDx policy system, with uniform LCDs, Z-codes, and pricing, spans almost 30 states and pays circa 80% of Medicare MoPath services.   For the first time since 2011, one of the MACs has deleted MolDx LCDs that are in force elsewhere.  

"Rationales" are provided for this action, but to me, they don't make sense.  I explain why it is... Puzzling!

The MolDx program was founded in 2011.  According to the Noridian website as of 5/17/2020, MolDx "establishes coverage and reimbursement for diagnostic tests."   Here.  The webpage just quoted was revised 3/24/2020, although there's no change-release record of what was revised on March 24.

According to a table on the Noridian MolDx webpage, the MolDx Contractor makes "determinations regarding coverage."  Then, according to the webpage, Noridian "implement[s] MolDx guidelines" and "process[es] claims using MolDx guidelines."   

The MolDx program is very large, spanning about 30 states.  So far, the MolDx program is active in the Palmetto jurisdictions (2), the Noridian jurisdictions (2), the WPS jurisdictions (2), and the CGS jurisdiction (1).  Based on 2018 data I reviewed previously, MolDx states or MACs process about 80% of molecular pathology dollar volume for Medicare Part B, and MolDx MACs have nearly 100% of the Medicare use of unlisted code 81479.

The week of May 11, 2020, marks the first break I am aware of where a MolDx-participating MAC declined to finalize MolDx national system LCDs.  In this case, there is little impact on the tests and patients involved, since Noridian elected to immediately replace the two LCDs with web articles conveying the same coverage content.  However, it's still a first-ever event for those who are closely watching the MolDx system.

What Happened?

In spring 2019, all the MolDx MACs posted drafted LCDs for the Natera Prospera test (detecting donor graft origin DNA and identifying renal transplant rejection) and for the Guardant G360 oncogene liquid biopsy test.  The draft LCDs reflected new coverage for Prospera, and expanded coverage for G360 (from lung cancer, to all solid tumors).  During summer 2019, the various MACs collected public comments on the LCDs.   By winter 2019/2020, the MolDx MACs had formulated a common response to all collated comments, which began appearing along with final versions of the Prospera and G360 LCDs.

By May 10, 2020, the two LCDs were still in "draft" form on the Noridian website, as they had been for nearly a year.   

Then, in rapid succession, Noridian put up 3 articles on its website.  

The first was on May 11, and was titled simply, "Guardant360 and Prospera."  This article stated coverage criteria (exactly as had been proposed in May 2019) for Guardant G360 and then Prospera.  The article also states that "In order to provide a more generic, non-proprietary coverage for plasma based genomic profiling and cell free DNA rejection technology, Noridian is retiring" the corresponding draft policies..."until such time a more comprehensive non-proprietary local coverage determination can be developed."

On May 15, Noridian put up a separate posting that G360 LCD  (DL38127) was being retired with the rational, "In order to provide a more generic, non-proprietary coverage for plasma based genomic profiling, Noridian is retiring this proposed LCD."   A separate but very similar article was also posted on May 15 for Prospera draft LCD DL38131. 

I've collated the three quoted articles in one cloud web document here.

Why It's Puzzling!

Noridian writes on the May 11 article, "Noridian will work with the CMS MolDX contractor to facilitate the future comprehensive LCD."   But Palmetto and other MACs, had already produced a "more generic, non-proprietary coverage" for plasma based cancer gene profiling, which did cover G360 and other tests on a rolling basis with similar performance.   It's LCD L38043 and represents the natural finalization of LCD DL38127 in other MACs.  In short, I don't understand the rationale as stated.  Why note finalize the future comprehensive oncogene liquid biopsy LCD that 's already written, and why not do it now?  I'd be happy if anyone has an idea.

MolDx has already proposed a future, non-proprietary or "umbrella" LCD for donor DNA tumor rejection tests.   This was first proposed a few weeks ago -  Here.  However, it could take about a year to circulate and be finaled in all the MolDx MACs.   See DL38568, liquid biopsies in solid organ transplant.   (Associated article DL58019 lists coverage for Prospera and KSORT).  

 I can't see why Noridian couldn't have finaled the DL38131 LCD at this time, and finaled the future LCD DL38568 in a year (or not, depending on the final decision and public comment and evolving science.)   MACs can always retire an LCD when an NCD or a different future LCD supercedes it.  In short -again - I don't understand the rationale as stated.  I'd be happy if anyone has an idea.

Update - On May 28, Noridian released for comment draft LCD DL38629, DA58168 (here) which is equivalent to Palmetto earlier draft LCDs DL38568.  Article A5818 includes coverage for KSORT and PROSPERA as an attachment to the umbrella LCD DL38568.  

LCDs Have Hard-Won and Statutory Process Controls

Based on CMS rules and recent lawmaking in the 21st Century Cures Act, there are a number of controls on the LCD process.  There is a process for requesting new LCDs, and a process for requesting revised LCDs.  These are logged and documented by MACs, following CMS rules.  There must be posted LCDs, with rationales and citations, there must be a public comment meeting, there must be a final Q&A document responding to public comment.  College of American Pathologist (CAP) made securing these legislative protections a major policy goal several years ago.

None of these rules apply to web articles posted by MACs.   

In addition, based on older lawmaking, there is an LCD appeals process to the administrative judge system, which is not available to articles.  (Although in a case or two, judges have ruled that an article performed the same function "in locus LCD" and therefore the stakeholder had the same legal rights as if the article were in fact an LCD pro forma.)

Stakeholders may also wonder if they should follow the published MolDx rules and processes for MolDx meetings, MolDx dossier submissions, etc, if the LCD resulting may or may not be finalized in one or another participating MAC.  

What Went Different Here?

MolDx medical directors, at various public conferences, have presented the MolDx 11-step process as shown below.  In fact, I've personally seen companies try to jump this process and work directly with Noridian in past years, and they were deflected and pointed to MolDx.  
  1. MolDx provides a central point of contact for new tests.
  2. MolDx provides non-binding advice during preliminary meetings.
  3. Stakeholder submits Formal LCD Request -- to MolDx.
  4. MolDx then coordinates technology assessment and review.  If appropriate, MolDx develops draft LCD.
  5. MolDx circulates the draft LCD to participating MACs and MAC medical directors.
  6. With consensus, the final draft LCD is released in uniform format across MACs.
  7. Release of draft LCD means Medicare has proposed and taken the coverage position stated, at the time of release.   
  8. Public comment follows at each MAC.
  9. Public comments are collated and a consensus Q&A document is produced.
  10. MolDx leads any required LCD revisions and then circulates with participating MACs for buy-in.
  11. Buy-in on the final LCD having been achieved, the final LCDs are rolled out, uniformly, at each MAC.
The final step #11 may drag on for up to six months, which is puzzling if a consensus final LCD was already agreed on at steps #9, #10, and therefore only required the formal act of posting in step #11.