From the media to Congress, we hear about Medicare Advantage denials and roadblocks. CMS now offers improvements to its rules, but some key facts may be out of place.
##
Background - Recent and New Rulemaking
From postings at the AMA (here, here) to the halls of Congress (here), we hear about Medicare Advantage denial rates. Some new proposed rules may be hiding some critical flaws.
Basically, Medicare Advantage plans have to at least match coverage in fee for service Part A and B, mostly as determined by NCDs and LCDs. And for a long time, that's about all CMS wrote to describe the standard. [fn1]
Then, in final rulemaking April 2023 (88 FR 22120), effective 2024, CMS issued pages of policy and paragraphs of regulations in the Federal Register.
See also an important CMS explanatory memo in February 2024.
In a proposed rule published December 10, 2024, CMS revisits the Medicare Advantage coverage topic with several pages of policymaking and a few paragraphs of revised regulations. Find it here, 89 FR 99340. Comment is open til January 25, 2025.
What CMS Says
The particular recent rule is most famous for reasoning that newly, Part D plans must cover weight-loss drugs like Wegovy (99375ff, news here). There's also an important section called "Guardrails for AI" in plan decision-making (99396ff; blog here). But in this article, I'm drawing attention to Coverage Criteria for Medicare Advantage (99455-461, plus regs at 422.101 (99557-8).
CMS states that Medicare Advantage continues to be a source of confusion and misunderstandings, requiring new explanations and rules to ensure beneficiary coverage.
They focus most strongly on language in NCDs and LCDs. CMS states that these always have clear-cut coverage rules, because any gaps or lack in clarity would have been fixed during public comment. (Hope you were sitting down.)
M.A. plans need only fall the "plain language" of the library of LCDs and NCDs. CMS also emphasize that they mean exactly the "LCDs" because coding/billing articles for an LCD "do not contain coverage criteria, that is the role of the LCD."
OMG... I See Problems
This leaves me wondering if authors of the rule know the Part B world at all. For example, the proposed oncology LCD from Novitas, which is suspended without finalization, is a verbose, repetitive, confusing morass - hardly "plain language." (Here). And it would be impossible to guess its implementation without reference to a huge billing and coding article with over 100 sections, arranged willy-nilly.
A different problem arises in the 28 MolDx states, which have the lion's share of all MoPath payments in Medicare. MolDx issues "foundational LCDs" which give general principles for coverage but the LCD never says what conditions, which cancers, or which tests are covered. Rather, this might be found in the current billing article, or, in the proprietary Palmetto MolDx DEX Database (here).
The Palmettogba.com/MolDx webpage even includes a notice that billing and coding articles (in CMS format) will often lack information on tests and services because this can be found over at the separate DEX registry. (Screen shot from here, then here; taken 12-11-2024).
- Topic is the LCD supplements aka Articles, with tables of billing and coding data,
- "These tables are being removed because coverage information on explicit services that have met coverage criteria can be readily found in the DEX™ Registry".
L38779, MolDx, Minimal Residual DiseaseFor example, in one of MolDx's most important policies, on minimal residual disease testing in cancer L38779, although there are 10 coverage rules, many are self-explanatory (the patient has a history of cancer.) Coverage criteria are presented in general terms (the identification of cancer would lead to a change in management; or stating that the test is demonstrated to identify recurrence berfore there is other evidence.) It would be very hard for a Medicare Adantage plan - or better, for ten or twenty different Medicare Advantage plans - to read those general expressions and all tens exactly and identically know whether a particular test (say, the LabCo test when used in liver cancer when used for completeness of curative resection) is or is not currently covered by Medicare fee-for-service.
And unlike coverage codes tied to CMS-based LCD articles, in DEX, it's impossible to search for tests covered or not covered by any given LCD, to know at one view which MRD tests are covered and which not, nor to know dates of coverage (no dates in DEX).
Bringing It Home: Real World versus the Proposed Medicare Advantage Rule
So if I read the CMS rule correctly, it points to the LCD alone as the near-Biblical definition of coverage. And - just like the Bible - the LCDs and NCDs are so clear they can be uniformly interpreted the same way by everybody just by their "plain text" (CMS's term of art).
So this is the status quo of LCD and Article:
OK, that was familiar.
This new guidance for Medicare Advantage decisions is focused on solely the "plain text" of each LCD:
I've put the Red X, because CMS says coverage info needed by Medicare Advantage won't be, shouldn't be, sought in articles, it is only in the LCDs and by quickly reading their "plain text."
That's even further from the coverage rules needed in the MolDx sphere, where the DEX database (plus Z code descriptors) hold some of the key information. As we quoted earlier, "coverage information on explicit [particular] services that have met coverage criteria can be readily found in the DEX™ Registry"
Again, contrast that to the propose rule, we get the Red X over articles or the DEX, because CMS says that Medicare Advantage plans need solely interpret the "plain text" of LCDs.
###
AI Corner
Google Notebook LM reads this article.
##
CMS Proposed Rule for Medicare Advantage Coverage: A Recipe for Confusion?
This briefing document reviews a December 10, 2024 proposed rule from the Centers for Medicare & Medicaid Services (CMS) concerning Medicare Advantage coverage criteria, as analyzed by healthcare policy expert Bruce Quinn, MD, PhD, in his blog post "CMS Edits Medicare Advantage Coverage Rules: A Big Mess?" (December 11, 2024).
Main Theme: The proposed rule simplifies Medicare Advantage coverage determination by emphasizing the "plain language" of National Coverage Determinations (NCDs) and Local Coverage Determinations (LCDs) as the sole source of truth. However, Quinn argues that this approach ignores real-world complexities in Part B coverage, particularly in specialized areas like molecular diagnostics, creating potential for confusion and inconsistent coverage decisions among Medicare Advantage plans.
Key Points and Facts:
- Background: The proposed rule follows previous efforts by CMS to clarify Medicare Advantage coverage requirements, amidst concerns over denials and barriers to care.
- Focus on "Plain Language" of LCDs and NCDs: The proposed rule asserts that NCDs and LCDs provide clear coverage rules, with any ambiguity resolved during public comment periods. Medicare Advantage plans are expected to adhere to these documents as the definitive guide for coverage decisions.
- Quinn's Critique: Quinn challenges this assertion, highlighting the often convoluted and unclear language of LCDs, citing a pending oncology LCD from Novitas as an example. He also argues that relying solely on LCDs ignores the crucial role of:
- Coding/billing articles: Quinn points out that these articles often contain essential coverage details not found in the LCD itself. He specifically cites the Palmetto MolDx program, where billing articles are explicitly skipped (per a website notification), and coverage information for specific tests resides instead in a separate, proprietary database (DEX).
- MolDx policies: Using the example of minimal residual disease (MRD) testing in cancer (L38779), Quinn demonstrates how general coverage statements in the LCD require interpretation based on specific test applications and clinical contexts, information often found in the DEX database or the wording of Z code descriptors.
- Disparity between Proposed Rule and Reality: Quinn argues that the proposed rule's reliance on "plain text" interpretation of LCDs creates a disconnect with the realities of Part B coverage, particularly in complex, evolving areas like molecular diagnostics. This could lead to inconsistent coverage decisions across different Medicare Advantage plans, disadvantaging beneficiaries.
Key Quotes:
- CMS on NCDs and LCDs: "CMS states that these always have clear-cut coverage rules, because any gaps or lack in clarity would have been fixed during public comment."
- Quinn on LCD Complexity: "For example, the proposed oncology LCD from Novitas, which is suspended without finalization, is a verbose, repetitive, confusing morass - hardly 'plain language.'"
- Quinn on the Role of Coding Articles and DEX: "The Palmettogba.com/MolDx webpage even includes a notice that billing and coding articles (in CMS format) will often lack information on tests and services because this can be found over at the separate DEX registry."
- Quinn on Potential for Inconsistent Coverage: "It would be very hard for a Medicare Advantage plan - or better, for ten or twenty different Medicare Advantage plans - to read those general expressions and all tens exactly and identically know whether a particular test [...] is or is not currently covered by Medicare fee-for-service."
Conclusion:
Quinn suggests that the proposed rule's oversimplification of Medicare Advantage coverage determination, while aiming for clarity, risks exacerbating existing confusion and potentially hindering beneficiary access to care. He advocates for a more nuanced approach that recognizes the complexities of Part B coverage and the need for additional resources beyond the "plain text" of LCDs
###
###
I'm highlighting MolDx here because of its elaborate system of LCDs, articles, and DEX database all interacting with one another to give a picture of coverage, but even then, often only partial (e.g. there may be no definitions of frequency for "use case X" when surveillance is involved).
But this requirement for interpretation isn't limited. Take an NCD. CMS NCD 90.2 governs next generation sequencing used in cancer patients, and has some maddening ambiguities that were not fixed during the comment period, or new strangeness was introduced by CMS's new text appearing only in the final NCD. And there is a baroque, huge, Excel spreadsheet with many pages of lengthy instructions about what is covered or not covered, updated quarterly (see CR13596.zip within prior link). To refer to the "plain text" of the NCD assuming all the Medicare Advantage parties would interpret it the same way 8 years later (without the years of supplements and coverage documents) is a false hope.
Another example of vagueness is the newly updated MRI LCD (L34425) for head and deck. The coverage section spends half its time simply defining MRI "radiofrequency signals when exposed to radio waves," and explaining that various metal fragments and clips may be contraindications. Then, it states that MRI of the orbit, face, and neck may be medically necessary to diagnose and characterize pathology of the orbit, face, and neck. (OK, thanks, guys).
###
fn1
These new rules and policymaking complexify old legacy simple statements about coverage parity which are still found in Medicare Managed Care Manual, Chapter 4, section 10.2.
##
Here's another example of very detailed NCD implementation rules that one would be hard-pressed to extract "from the NCD" which is concise.
