Hacker News Reader

Comment by toolz | Hacker News Reader

the final study list he compiled shows 80-85% confidence for <30% mortality reduction and the author confirms that in the comments. This is 80-85% confidence that hundreds and hundreds of thousands of lives could've been saved with ivermectin. You didn't read closely enough.

Notably though, at no point does the author even suggest it doesn't work - just that the statistical significance isn't high

View on HN

I'm not sure where you're reaching 'hundreds and hundreds of thousands', perhaps because there's a big range between zero and thirty percent.

In medical trials, statistical significance is the milestone to reach before broadly claiming something 'works'. Dozens of other treatments were put through trials. Some had promising initial results. Most didn't work.

No one is claiming it works - I'm refuting the point you've made which is that it doesn't work - which is contrary to what the data shows. The data shows that it's not reaching the medical milestone you refer to, but it does show that it's more likely helping.

You've claimed it doesn't work - nothing says that. Nothing even suggests that.

Quite the mental pretzel there. If no one is claiming it works, and you claim that I claim that it doesn't work, maybe that's reason at work? Unfortunately, I never claimed it didn't work.

Edit: I see you might have confused me for a parent post by lalaland1125 who did claim this.

Reply on news.ycombinator.com