[uj8efdkjfdshf]

Realistically, while I feel that the author is on the right track with this data driven approach to evidence based medicine, the fact remains that the advantages/disadvantages between drugs within the same class of medicine tend to be negligible compared to drug costs. There are then other specific concerns that must be considered eg pregnancy/breastfeeding, liver failure, kidney failure and often what ends up is that one defaults to a standard drug per drug class with alternatives in specific circumstances.

The real benefit IMO would be marketing this to countries with a top down healthcare purchasing system (eg the UK) or as a tool to drug companies/researchers looking to make their own meta analyses (eg abstrackr). It might also be better instead to diversify into correlating symptom clusters with diseases because then the utility to the end user is bigger.

[allcentury]

I had the same thought:

Sell this product to insurance companies so they can recommend cheaper/safer drugs to their customers, or show the more expensive one and if it truly had better outcomes. Everyone benefits from that.

That said, the medical establishment is a tough tough domain to move.

[lotsofpulp]

Insurance companies already have teams of doctors and pharmacists analyzing studies to ensure they are pursuing the most effective and cost effective healthcare. They’re called Pharmacy and Therapeutics (P&T) Committeea.

[wayoutthere]

I worked at a company in the early 00s that did exactly this -- natural language processing of medical records for meta analysis. Even when it worked, we couldn't sell it for much. The budget of each study was too low to really build a business around, and the low number of potential buyers didn't help things.

Eventually we pivoted the tech to a B2B SaaS play because it was too hard to sell.