News for hackers. Stuff that matters.

A comment on Cheap antidepressant shows promise treating early Covid-19. Parent comment

Re:Cheap antidepressant shows promise treating early Covid-19 by darkerside on Thursday October 28, 2021 @01:03AM (#29022563)
> When the DSMC met on Aug 5, 2021, it recommended that the TOGETHER trial stop randomly assigning patients to the fluvoxamine group, as this comparison had met the prespecified superiority criterion for the primary endpoint (prespecified superiority threshold 97ยท6%).

That seems problematic. Did they stop the study because they had achieved statistical significance?

[ Parent | on HN ]
Re:Cheap antidepressant shows promise treating early Covid-19 by scarmig on Thursday October 28, 2021 @01:28AM (#29022700)
Another issue is that they were testing multiple drugs, none of which had a statistically significant effect. Classic p hacking.

https://xkcd.com/882/

[ Parent | on HN ]
Re:Cheap antidepressant shows promise treating early Covid-19 by scotty79 on Thursday October 28, 2021 @05:24AM (#29024019)
Assesing 7 promising substances to determine which one if any does anything and getting 30% effect on one and no effect on the rest is p-hacking?

Is science allowed to research only one substance for specific purpose at any time to sufficiently avoid p-hacking for you?

Also if I read correctly they got >0.60 not >0.05

[ Parent | on HN ]