Dumb Articles Proliferate on "Expected" Spending for Alzheimer Drugs

Hyperbolic articles continue to appear regarding the potential cost of new Alzheimer drugs.  

I wrote a blog (and submitted a JAMA comment) in January 2022 on inflated "expected costs" (quote unquote) of Aduhelm.   My blog and links, here.   '

Basically, these academic estimates took what everyone in the pharma or the business world would call an "addressable market" and then said this was "expected sales."  Two different things!!!

For example, if I have a new tennis shoe, the addressable market might be 100M people, but my expected sales might be 25,000 pairs.   

This error has happened before.  There were fears that Provenge, approved in 2010 for prostate cancer, would have sales of zillions of dollars and "bankrupt Medicare."  But my informal estimate recently was that Provenge sales never reached higher than a few-percentage-points of its addressable market.

Here we go again.  See coverage at Fierce Healthcare here.   See the underlying letter in JAMA here by Arbanas et al.  

There's nothing wrong with the JAMA article itself; it takes the total number of US Medicare patients and various other estimates to estimate a potential market.   However, like the January 2022 article, the Arbanas article again uses incorrect terms like "expected spending" which is a term that should not be used until you discount the total market by a market penetration factor (as occurred for Provenge).

An example.   Medicare Part B sales of the Cologaurd test are circa 500,000 a year (observed sales circa $250M).  But if you assumed that "expected sales" was the same thing as "total addressable market," then every Medicare patient would get Cologuard every 3 years, and Medicare sales per year would be 10M-15M units (with "expected" (sic) sales north of $5B.)