Hacker News Reader

Comment by truth_sentinell | Hacker News Reader

Well, that's my point. If they aren't replicating the study, then that's wrong. It's like only validating input on the client and not server-side. It would be much more difficult for a company to fake both FDA results and their own.

And public health is what's at stake here, so there should never be enough tests.

View on HN

That's understandable. But the idea here is that the entire process of testing a drug is extremely transparent to the FDA. In a typical scientific paper, the researcher worked on something for months or years, all the while keeping the structure of the study and the data they were collecting secret. After they finish, they then write a paper summarizing the methods and findings. This paper is (ideally) then peer-reviewed, and eventually replicated by other researchers to determine if the methods described produce the expected results.

Whereas with a drug trial, the company is telling the FDA from the very start "Hey, we have a drug we think might treat diabetes. We want to get this approved and on the market, so here is our plan for how we will conduct the trial." They then conduct the research in a much more open manner than typical lab studies in other fields.

Finally, there is the Institutional Review Board that is charged with ensuring all academic studies do not harm their participants in any way. Lying about your methodology in a medical study would get you expelled from your profession so fast your head will spin.

http://www.fda.gov/ScienceResearch/SpecialTopics/RunningClin...

I would argue that the flipside of this being a public health matter is that people are suffering and dying of a condition that could be helped by a drug that we have to test for years and years. If we can conduct a transparent and rigorous double blind study and determine the safety and efficacy of the drug, we should be doing it.

Reply on news.ycombinator.com