[entee]

It's important to understand why the efficiency is going down, and it's not because the companies themselves suck more every year. It's because of two main drivers:

1.) The easy targets are mostly gone. All the biology we understand very well either has drugs already, or for some key reason can't be addressed with the drug discovery system we have today (see KRAS, though there are glimmers of home there). This means we have to go after harder diseases that aren't as well understood or have complex etiology. Necessarily that's less efficient.

2.) When you drug something, it eventually becomes incredibly cheap to use that drug, and it becomes the benchmark. (AKA the "better than the Beatles problem") As a pharma company you're competing against the former, better (because you were going after easier biology) versions of yourself all the time. This is also true in software, good software often becomes somewhat commoditized. However, in software we've had exponentially increasing compute capacity for some time, which means the next product can be exponentially better than the previous one simply because new things become actually possible to do when they weren't before. This by and large isn't and can't be true in pharma.

[dv_dt]

I do think some aspect of this is true, but I think how research is directed is systemically off today. You may want to read this overview report (it has footnotes for all the research it cites). But there are quotes in there like:

"While critical medical needs remain unmet, a majority of new medicines developed have no added therapeutic value."

https://www.ucl.ac.uk/bartlett/public-purpose/sites/public-p...

[mattkrause]

That's a frustrating article.

Figure 1 shows that recent neuro/psychatric drugs don't have dramatic effects. That's true--but it's not for want of trying. People have been bashing their heads against the brain for decades, but it turns out to be a really hard problem. Biogen lost $18B of market cap on its aducanumab trials early this year. Eli Lilly, AstraZeneca, Roche, Pfizer, Merck, and Johnson & Johnson have all also had big Alzheimer's trials go sideways. I'm not sure that throwing more money at the same amyloid hypothesis will help. Maybe share buybacks—which get taxed and thereby fund the NIH--might not be the worst idea.

The paper also complains about expensive drugs and me-too drugs. The initial HepC drugs were crazy expensive (nearly six figures), but produced quick cures in virtually all of the patients. This was still a massive win over slow, ineffective treatments that still ended with liver cirrhosis/ cancer/transplants for half of the patients. The subsequent "me- too" drugs, which mostly target the same pathway, don't work appreciably better (hard to demonstrate improvement vs a 95%+ cure rate) but have driven the price down by tens of thousands of dollars.