On January 11, 2022, CMS announced a plan to cover the Alzheimer drug aducanumab (Aduhelm) and other Alzheimer amyloid drugs, only in randomized controlled trials, for which Medicare would pay direct drug costs. Entry point here.
On January 14, 2022, JAMA Health Forum published an article, Mafi et al, on "Estimated annual spending on Aducanumab in the US Medicare Program" - here.
I believe the authors make an error. The authors discuss the potential Medicare market that might be available to aducanumab - from 1M to 6M patients. This is likely correct and is the focus of the authors' expertise. The concept of "potential market" will be familiar in business terms as the "total addressable market" or TAM.
However, and this went by the editors of JAMA, the authors pivot directly to expected Medicare expenditures, and state an "upper bound limit" (which is correct) and what they call "a lower bound limit" (sic) which is absolutely not correct. They describe their results as "Medicare spending estimates" which they are not. Among "limitations" they mention a few, such as misclassification of dementia, but do not mention they project dollar values for expected Medicare costs without allowing any variable for market penetration.
They open by stating that aducanumab "could threaten Medicare's solvency." This is unlikely; it is a high multi hundred billion dollar program. New therapies, such as aggregated checkpoint inhibitors, are larger dollar volumes that aducanumab would be. Whether aducanumab was $100M or $1B in sales, it would not singlehandedly threaten the larger program's solvency, any more than Provenge for prostate cancer did (it reached $300M in sales). Even worse, the "minimum lower bound of Medicare cost" for aducanumab is presented as $7B with narrow plus/minus error bars.
I submitted a public comment which I clip below the break.
January 15, 2022
Total Addressable Market (TAM) vs Actual Anticipated Costs
BRUCE QUINN, MD PhD | Physician and policy consultant
The article moves back and forth between the framework of total addressable market (sometimes abbreviated TAM) and best expected costs. These are two quite different things.
While the addressable population being carefully estimated here may vary from 1M to 6M, the uptake and hence the "expected" or "anticipated" costs may be lower. Many recommended programs and drugs and health interventions have much lower uptake than the addressable market.
For example, the total addressable market for sipuleucil (Provenge) was very large: all men in Medicare with metastatic prostate cancer. Yet despite an on-label national coverage determination (NCD), actual uptake was quite limited. The NCD was published in 2011; yearlong sales in 2012 and 2013 were circa $300M each, not multiples of billions of dollars, and roughly 3000 men per year. The prevalence for metastatic prostate cancer, and thus eligible men, had to be multiples of 35,000, the annual death rate from metastatic prostate cancer.
For another example, many of the millions with dementia might qualify for a CSF exam, yet total Medicare CSF exams for all causes are only 28,000 services per year (codes 62270 + 62328).
CONFLICT OF INTEREST: I have consulted for BIOGEN on diagnostic test policy for dementia.
Though unfortunately it's not open access, see an outstanding historical article on how Medicare faced (with hysteria) early heart transplant decisions; Evans 2017, J Heart Lung Trans 36:1294.
No direct relation to my article above, but on the topic of Medicare and accelerated approval costs, see article on "Medicare spending on accelerated approval drugs," Dec 2021, here.