Hacker News Reader

Not nearly as much as you might think. Check out "Bad Pharma" by Ben Goldacre. It's a treatment of the statistical and legal tricks used by drug companies to overstate the efficacy of new drugs, without breaking any actual FDA regulations.

Even after using all the known tricks for cherry-picking results, p-hacking, and the like, there exist much-abused FDA 'fast track' options for getting a 'needed' drug out before the science is confirmed, and quite often the science never gets around to being confirmed. Even big-name drugs like Tamiflu have been shown later to be no better than generics at best and completely useless at worst.

It's also not that hard---perform your own research, under conditions you specify, using analysis methods you dictate, with the ability to retry over and over and over until a favorable result arises by chance, THEN the FDA scrutinizes it.

The same FDA that is staffed by ex-pharmaceutical executives, and the same pharmaceutical companies that are staffed by ex-FDA staff? The FDA is really not something you can reasonably, or should, trust.