Wednesday, September 6, 2017

The Holes in the CMS/Novartis Announcement: The CMS Pricing of Kymriah ($475,000)

On August 30, 2017, Novartis announced the FDA's approved of the first chimeric antigen receptor T-cell therapy (CAR-T): Kymriah (tisagenlecleucel).  The therapy has a listed price of $475,000 in its first indication, for relapsed acute B-cell leukemia in patients who under age 25.

Numerous press stories cited an "agreement between Novartis and CMS" that CMS would pay only for patients with successful outcomes.   CMS issued a press release that was incredibly vaguely worded, and then, apparently pulled down from the CMS website.

     Update September 13: Congr. Doggett Asks CMS to Clarify "Agreement;" see here.

What happened?

In Novartis' own press release (here), we have this statement:
Groundbreaking Collaboration with the Centers for Medicare and Medicaid Services 
Novartis also announced a novel collaboration with the United States Centers for Medicare and Medicaid Services (CMS) focused on improving efficiencies in current regulatory requirements in order to deliver value-based care and ensure access for this specific patient population. 
This approach is intended to include indication-based pricing for medicines and supports payments for a medicine, such as Kymriah for its initial indication, based on the clinical outcomes achieved, which would eliminate inefficiencies from the healthcare system. Other value-based approaches related to future indications for Kymriah and CAR-T cell therapies are under discussion.  
Furthermore, Novartis is collaborating with CMS to make an outcomes-based approach available to allow for payment only when pediatric and young adult ALL patients respond to Kymriah by the end of the first month. Future potential indications would be reviewed for the most relevant outcomes-based approach. 
"Novartis has been at the forefront of outcomes-based pricing and is very pleased to work with CMS on this first-of-its-kind collaboration with a technology that has the potential to transform cancer care," said Joseph Jimenez, CEO of Novartis. "We look forward to continuing to work with CMS to potentially expand this approach to other products and disease states."
Novartis' remarks were summarized in the secondary sources, such as Endpoints (here), Fortune (here),  and at more length in an article by Agus & Goldman in Fortune (here).  For example, the latter authors, who are experts on value-based pricing, write:
There was another announcement from another government agency, the Centers for Medicare and Medicaid Services (CMS). But it also could have far-reaching impact on how we pay for—and ultimately treat—cancers of all kinds. CMS would cover Kymriah’s announced price of $475,000 for the treatment. That grabbed headlines, but in the long run it does not compare to the other part of the announcement from CMS. For the first time the agency agreed to a payment plan which is dependent on whether the drug actually produces a benefit to the patient. Novartis will get paid $475,000 only if patients respond to the drug by the end of the first month of therapy.
CMS Press Release Is Much More Vague 

I dug up the CMS announcement last week and found it amazingly vague.  It didn't refer to any agreement.  It only stated vaguely that CMS was interesting in better pricing for drugs.   When I went back to the CMS website today, September 6, the press release had disappeared.  However, I contacted CMS and they described the problem as a web glitch.  By September 7, the unchanged press release was back up at CMS (here).
With today’s [approval of Kymriah]...the Centers for Medicare & Medicaid Services (CMS) is continuing to explore the development of payment models and arrangements... 
CMS is working actively with all stakeholders, including state officials, on innovative payment arrangements. These arrangements may, for example, include outcome-based pricing for medicines...CMS will be issuing future guidance to explain how... 
CMS is committed to further exploring the development of innovative pricing systems that reflect the value delivered to patients. Through the authority provided to the Center for Medicare and Medicaid Innovation (CMMI), CMS will aim to identify and alleviate regulatory barriers in Medicare and Medicaid as may be necessary to test payment and service delivery models that involve value-based payment arrangements.
CMS doesn't say they've "agreed to a payment plan" (as stated by Agus & Goldman, but rather that CMS is "exploring" and "committed to further exploring" and "working actively" (is there another kind of working?)

What Could the CMS-Novartis Process Have Been?
  • Drug prices aren't set willy-nilly by CMS, and not at the spur of the moment judgment of thoughtful administrators.  
  • Drug prices are set by law and regulation in both the hospital and freestanding settings (the "average sales price" or ASP lawmaking dating to 2003.)  
    • See Hays v Leavitt, 2009, halting CMS from implementing drug pricing differently than allowed by law just because CMS found a pricing method that made more sense than the law, in the view of the agency.
  • But can't a company voluntarily agree to a different pricing scheme with CMS?   Without due process?  I'm not so sure.  What if CMS had determined that it only pays $50,000 for Kymriah, and only after proof one-year outcomes?  
    • A satisfied pharma would never complain, and the issue would never reach court.  But an unhappy pharma?  Wouldn't a pharma then complain (and in court) that administrators have no such ad hoc powers to assign one-off coverage and prices by fiat?  
    • Put this way:  If CMS can make up any rule or policy it wants overnight, that you like and agree to, it can make any new rule overnight, that you don't like at all
  • For Medicare, coverage decisions, by law, are called National Coverage Determinations or NCDs and follow a regulatory process laid out in law.  I don't think CMS can ignore those laws at the drop of a hat when it seems like the right moment.  But here, CMS announced it was covering Kymriah out of the blue - a de novo agency coverage decision.
  • CMS senior staff don't pay for drugs: The claims are paid by the eight regional contractors, following specific instructions from CMS.  How would you define everything in the Novartis agreement?  Those long explicit instructions to contractors would usually be transmitted publicly.  And they'd depend on existing rules and regulations.  
    • And by the way, pharma companies don't bill CMS: Hospitals do.  
    • The hospital would get paid -- or not paid -- by CMS after it's already bought the drug.   (Maybe Novartis could rebate each hospital upon proof that CMS had denied the hospital payment - but CMS payments or payment denials initially involve only the hospital that submitted the bill.)    
  • And Kymriah doesn't currently involve Medicare at all, where CMS has more control by hiring and instructing its contractors.  Rather, Kymriah is for patients under age 25 who would have to be Medicaid (CHIP) patients to fall under CMS authority.  
    • And there are 50 independent Medicaid programs, plus many variations of managed care Medicaid from state to state.  
    • And there are bodies of law on how Medicaid pays for drugs, including state laws.   
    • CMS rarely - if ever in its fifty year history - has made a single line item product coverage-and-payment decision that was binding from Florida to Alaska on all the Medicaid programs.  I don't think it can do so.   
While the first indication for Kymriah is for young Medicaid patients, later indications will follow quickly and they involve Medicare.  What if it's a drug given for inpatients?  Then, it bumps up across DRG bundles with no extra payment for drugs administered.  Or bumps into the Oncology Care Model bundle.  And for inpatients, even the New Tech DRG add on payment by law only pays 50% of the added cost - not much good for a $475,000 drug.

With this background, try reading the Novartis quotations and CMS quotations again.  I see the goal, but I don't see the process.

CMMI May Be the Missing Link, Because CMMI Can Do Anything

CMMI, which CMS now prefers to call the Innovation Center, was created by the ACA in 2010.  It has a budget of about a billion dollars a year and has established numerous small and large demonstration programs.

By law, CMMI can conduct Demonstration Programs under which it may waive any Medicare law for the purpose of the demonstration program.   This quickly led to concerns among conservatives that CMMI is bad law - that Congress can't give the agency a large legal rulebook, and then, delegate to the same agency the authority to waive any or all of the laws that create the agency.   (See Grace-Marie Turner in Forbes last year, here; see also here).  (And Congress agreed with Marie-Turner, see House Ways & Means, May 2016, here.)

For example, many stakeholders objected in 2016 when CMMI proposed a novel set of policies for Medicare drug payment, overriding existing laws, and national in scope and of unlimited duration.
(Hence my title: CMMI can do anything.  A "demonstration" program could be nationwide and 100 years long, because Congress never said how long it could be or how big it could be.  At least until a court had to make sense of the open-ended statutory authority.  Since CMMI can waive any law in a demo program, and a demo program can be of any scope, it seems CMMI could lower the Medicare age to 60 or raise it to 70 or cap it at 85. And most of CMMI's policy creations are shielded by law from court review.*)
I think that's why, vague as it was, CMS's press release stated that "Through the authority provided to the Center for Medicare and Medicaid Innovation (CMMI), CMS will aim to identify and alleviate regulatory barriers in Medicare and Medicaid..."

So CMMI could produce a program, specific to Kymriah, and that applies to all 50 state Medicaid programs, and that specificies a "Kymriah demonstration program" payment process to be implemented immediately and nationwide.  But it would be a "demonstration program" that was abruptly and unilaterally created by executive branch burocrats, national in scope, exactly what Republicans went crazy over in 2016 when Obama was in charge.

A Less Gentle CMS, For Better and Worse?

Many of us who have worked on CMS policy issues for a decade have seen instances where CMS declined to make a policy change, because CMS wasn't totally, completely sure it had enough authority, perhaps even under notice-and-comment rulemaking; because somebody of nerdy temperament buried in the Office of General Counsel of CMS had cold feet.

With the direction that Novartis' announcement went in, CMS was appearing to act much faster and more dramatically than it had been able to, under its previous understanding of the limits and guardrails of its legal authority.  



(d) Implementation [of CMMI].—

(1) Waiver authority.—The Secretary may waive such requirements of [CMS law:] titles XI and XVIII and of sections 1902(a)(1), 1902(a)(13), and 1903(m)(2)(A)(iii) as may be necessary solely for purposes of carrying out this section with respect to testing models described in subsection (b).
(2) Limitations on review.—There shall be no administrative or judicial review under section 1869, section 1878, or otherwise of—
(A) the selection of models for testing or expansion under this section;
(B) the selection of organizations, sites, or participants to test those models selected;
(C) the elements, parameters, scope, and duration of such models for testing or dissemination;
(D) determinations regarding budget neutrality under subsection (b)(3);
(E) the termination or modification of the design and implementation of a model under subsection (b)(3)(B); and
(F) determinations about expansion of the duration and scope of a model ...


Inside Health Policy recently reported that CMS may be issuing a request for public comments on the operations and future goals of CMMI; here.   Updatge:  It's out; September 20; here.

For an essay on CMMI by Manatt, September 12, 2017, here.

For a law journal article that includes a section on "unprecedented and unauthorized" CMMI overreach, here (PDF page 14ff).