Tuesday, June 22, 2021

New York Times Article on ADUHELM Costs is Leaky

At the New Yorker, the diligence of its fact checking has been legendary for decades (article here).

That sort of fact-checking is dead as a doornail at New York Times nowadays (a literary reference that can itself be fact-checked to page 1 of Dickens' "A Christmas Carol.")

In a New York Times headline on June 22, "New Drug Could Cost the Government as Much as It Spends on NASA," there are a number of leaky places.  It's not a toss-off piece, the authors including one of the NYT main health journalists, Margot Sanger-Katz.    But the citation trail doesn't hold up very well.

History of Ridiculous Medicare Projections

New drugs and technologies have often triggered hysterical cost projections for the Medicare program.

  • When the prostate cancer drug Provenge (sipuleucil) was approved by FDA in 2010, there was an outcry that its vast expenses would topple the Medicare program, given the prevalence of well over 1M men with prostate cancer.  
    • In fact,  Provenge sales were never surprisingly and the developer (not the Medicare program) went bankrupt.  Provenge sales ran around $200M per year, 2011-2013.
  • When heart transplants began to become practical in the 1980s, there was a groundswell of federal concern that costs would skyrocket and rapidly overwhelm the Medicare program.  (I wrote about the history in 2015 here).  
    • Never happened.  Didn't even get close.[*]
Paper Trail for NYT Figures

Today, NYT proposed $6-29B in new costs per year for Aduhelm.   The key reference is here: "Analysts expect that Medicare and its enrollees, who pay a share of their prescription drug costs, will spend $5.8 billion to $29 billion on the drug in a single year."  There are multiple references to "analysts" and the high estimate is hyperlinked to a Kaiser paper here.   

The Kaiser webpage is supported by the Arnold Foundation, which is notoriously anti-drug-cost.   Although this is given as the source of the headline high-estimate for the NYT, the Kaiser essay doesn't claim to have any actual data or serious data- and modeling-based projections of costs.  Kaiser simply remarks that "if 1Medicare beneficiaries receive Aduhelm, spending on Aduhelm would exceed $57B in a year."  

Well, yes.  If 1M people get a drug at $56K per year, that is $56B.   That's not a projection, it's basic arithmetic.  Kaiser then hyperlinks the 1M patient count as being "on the low end of Biogen's expectations."  Well, they say "it MAY be on the low end of Biogen's expectations."  That's a red flag; the 1M number is either on the low end of Biogen's expectation, or it isn't.   

Following the Kaiser hyperlink takes you to an online Biogen deck deck, specifically, dropping yo at page 26.  All that Biogen says on this page is that 1-2M patients have mild dementia, and there may be gradual uptake of patients (prescriptions), and modest revenue in 2021 may ramping up to "multi billion dollar opportunity" over several years.   

Did you see $56B?  I didn't.  Did you see that "$56B is at the low end?"  No.   The next page in the Biogen deck doesn't have any dollar expectations either, stating generally that "site infrastructure continues to scale" and "revenue [ramps] beginning in 2022."  Another page, page 30, says that Biogen expects "modest revenue" from Aduhelm in 2022, ramping thereafter.  Nowhere does Biogen say that $56 is "on the low end of its expections."  It doesn't even get near such a statement.


Net net, NYT attributes its high numbers to Kaiser, and Kaiser simply spitballs that 1M patients would be $56B costs, and then Kaiser attributes and hyperlinks the projections to a Biogen deck that says nothing of the sort at all.


[*] Back of envelope, we have about 3000 heart transplants per year, and even at $2M each, that would be $6B or 0.15% of a $3.8T health budget.