How the FDA Could Change the Way It Approves Drugs • Hacker News Board

Re: How the FDA Could Change the Way It Approves Drugs

#41

Post by Alex3917 » Fri, Sep 04, 2015, 12:38 PM UTC

danieltillett wrote: post #15

Alex3917 wrote: post #8

Earlier quoted context omitted.

Keep in mind also that as a patient you're allowed to take basically any drug you want, companies just aren't allowed to sell those drugs to other people and claim they treat a specific disease without FDA approval. So this whole article is basically just shilling for even less stringent requirements to market a drug as being effective, even though that hurdle is already trivially easy to overcome in these cases.

It is rather hard to take a drug if the company won’t give it to you. Pharmaceutical companies are very reluctant to even give drugs to desperate patients because the FDA requires them to report all adverse side-effects even though they have no control over the conditions in which the drug is used. You can actually kill a totally fantastic drug by letting people use it outside of a controlled clinical trial.

> It is rather hard to take a drug if the company won’t give it to you.

Not really, you just have to buy it from a compounding pharmacy.

Re: How the FDA Could Change the Way It Approves Drugs

#42

Post by gwern » Fri, Sep 04, 2015, 3:07 PM UTC

ccw9302 wrote: post #38

BurningFrog wrote: post #37

Earlier quoted context omitted.

1. A false positive rate of 27.9% is not the same thing as approving 27.9% of all applications. 2. The full clinical trials still cost hundreds of millions of dollars.

false positive rate 27.9% means: if the drug is not effective (at all), there is 27.9% probability that the drug would be approved by FDA.

> if the drug is not effective (at all), there is 27.9% probability that the drug would be approved by FDA.

Which is only one of two possibilities, and hence, it cannot be 'the probability the drug is approved by the FDA', for the same reason that 'a p-value of 0.05' does not mean '95% probability the drug works'.

Re: How the FDA Could Change the Way It Approves Drugs

#43

Post by danieltillett » Fri, Sep 04, 2015, 9:33 PM UTC

Alex3917 wrote: post #41

danieltillett wrote: post #15

Earlier quoted context omitted.

It is rather hard to take a drug if the company won’t give it to you. Pharmaceutical companies are very reluctant to even give drugs to desperate patients because the FDA requires them to report all adverse side-effects even though they have no control over the conditions in which the drug is used. You can actually kill a totally fantastic drug by letting people use it outside of a controlled clinical trial.

> It is rather hard to take a drug if the company won’t give it to you. Not really, you just have to buy it from a compounding pharmacy.

I think you might have a misunderstanding of what a compounding pharmacy can do. They can't just whip up some chemical on demand. If a new drug is not for sale they can't compound it.

Re: How the FDA Could Change the Way It Approves Drugs

#44

Post by dang » Sat, Sep 05, 2015, 9:40 AM UTC

pinaceae wrote: post #17

danieltillett wrote: post #14

Earlier quoted context omitted.

There aren’t any. The FDA and it’s related regulatory authorities around the world are responsible for killing more people than any other organisation that has ever existed (conservatively in the hundreds of millions). You might wonder how this could be. Basically the cost of getting a drug approved is such that large sections of the human population (poor people especially) have diseases for which we have no treatme…

refuting your BS would take far longer than you vomiting this utter, utter nonsense into HN. just please never take any drug approved by the FDA ever.

> refuting your BS would take far longer than you vomiting this utter, utter nonsense into HN

This comment violates the HN guidelines badly. We ban accounts that do this repeatedly. Please post civilly and substantively from now on, or not at all.

Re: How the FDA Could Change the Way It Approves Drugs

#45

Post by refurb » Sun, Sep 06, 2015, 4:10 PM UTC

danieltillett wrote: post #29

refurb wrote: post #27

Earlier quoted context omitted.

That's not true at all. Pre-clinical only requires that you show a drug is not toxic and has a potential benefit. Most drugs fail because they are either toxic or don't do anything. Why would you loosen those regulations?

If only that was all you had to show in the preclinical stage. The bigger problem is you can’t show either of these with preclinical testing. If you could then there would be no need for human testing. The reality is most of preclinical testing outside of certain drug classes (e.g. antibiotics) is not that helpful for knowing if a drug will cause problems in humans or will even work. It has got so bad that most drugs…

Your argument is that preclinical testing provides no data? I'm not sure I agree with that.