The Importance of High Quality Data on Event Rates

A lot of clinical trials estimate these event rates and it's just like they're pulling it out of their butt, they don't really know and it ends up being an underestimate. And then if the trial comes out non-significant, they can just say it's underpowered. So I would push back a little bit against that, which is like, for this thing to be worth a damn, it needs to at least be this good, right? And so if it's not this good, then it's not worth a damn. Like we can't power a clinical trial to show that something reduces the absolute risk of something by a 10th of a percent - nobody would care about that