Hacker News Reader

Comment by koheripbal | Hacker News Reader

While that's good news - it isn't statistically significant.

You need to do a double-blind test to confirm results.

This is an example of real-world evidence (https://en.wikipedia.org/wiki/Real_world_evidence), gathered outside the context of any kind of drug trial, blinded or not.

While these two outcomes may not achieve statistical significance in the mathematical sense...

- there are compelling mechanistic hypotheses for why the treatments might work

- drug doses are being supplied for free by manufacturers

- infected people don't have the luxury of time to wait for RCT results

Most importantly, a placebo controlled RCT would be unethical for a disease with high mortality rates.

You would run a randomized controlled trial with treatment control (not placebo controlled) when we have an effective treatment. These are about as good as it is assumed placebo effects figure the same in all tested treatment groups.

A double blind study is currently done in China, results should be there in a couple weeks.

A new trial is being run in Nebraska too.

Well that seems unreasonable? Double-blind just gives better comparability of results with lower N. Statistics wouldn't object to, say, giving all Italian patients the treatment and most Chinese patients not. You'd just end up with a more complex model and wider uncertainty, but if the medicine worked on a noticeable scale, you'd pick it up eventually.

Reply on news.ycombinator.com