(Total Views: 382)
Posted On: 09/17/2020 11:07:45 AM
Post# of 155666
That's not correct. For moderate, they had a 600 patient study plus an NIH study (although open-label), for severe they had more than a thousand.
Results with n=195 will only be stat sig and convincing if the observed effect is drastic (>=50% reduction) and placebo mortality is >=25%. Binomial endpoints are hard to hit with a low number of patients...
To be absolutely sure the trial is stopped for efficacy, we'd better see something like 66% mortality reduction and 30% placebo mortality!
Results with n=195 will only be stat sig and convincing if the observed effect is drastic (>=50% reduction) and placebo mortality is >=25%. Binomial endpoints are hard to hit with a low number of patients...
To be absolutely sure the trial is stopped for efficacy, we'd better see something like 66% mortality reduction and 30% placebo mortality!
(1)
(0)
