Hacker News new | ask | show | jobs
by estearum 30 days ago
No, you really don't need to "blindly trust" the entities doing the studies.

I work in clinical trials. I know how they work. You do not.

1 comments

>I work in clinical trials. I know how they work. You do not.

Ah, that makes perfect sense.

It is hard for a man to understand something when their salary depends on them not understanding it.

It's even harder when he's completely uncurious about how the world actually works ;)
Being curious does not help much if you are easily satisfied by bullshit answers, like "It works because I said so!"
Imagine thinking that placebo-controlled, double-blinded randomized clinical trials say "it works because I said so."

Surprised you don't fall out of your own shoes in your upside-down world.

I am not distrusting the method. I am distrusting the person that claims to follow it, behind closed doors and comes out with only a result.
Again: This reveals an ignorance about how trials actually work. You should be more curious.

Clinical trials don't happen "behind closed doors and come out with only a result."

Clinical trials are run independently across hundreds if not thousands of research sites. The people who run these studies do not work for the pharma companies. They do not work with each other. They do not know whether the drugs they're giving each patient is the real drug or the placebo. They therefore have no way to coordinate fraud with each other and actually execute it at a scale required to both change the results and be undetectable (i.e. massive, systematic deviations on top of randomized data that produces something other than just "more noise"). Each of the thousands of people who would need to cooperate in this conspiracy would be risking years in prison, the destruction of their businesses, and the cancelation of their medical licenses... for... literally no benefit? The people actually running these studies (which are not the pharma companies themselves) don't get paid more or less money based on the results!

Every step from initial data entry through to data analysis is auditable and inspected by the FDA. The statistical analysis plan is written prior to any data being collected, and any ad-hoc deviation from the pre-established analysis plan must be defended to the FDA. No "data corrections" can happen once analysis has started without very specific audit and approval by FDA for each individual data correction.

Again: this is a really remarkable system that is a lot of fun to learn about if you have some curiosity. It's not perfect, but it's especially strong at the massive scale of the vaccine trials because there would be so many data points that need to be manipulated by so many independent parties that it's just a ridiculous notion.