You don't, nor stats. This can help: hypothetically take away the apparent placebo effect. For RSBQ the drug arm would have had a mean effect of -4.61 and placebo arm is 0. Would the trial have been a stat-sig success?
This is in fact how the placebo effect works; you can't magically assume it's just in (or even just greater) the placebo arm. And this is how ANOVA stats work. And why the market gave thumbs down to the results, and why the FDA said we'd have to do another trial to proceed with them. It's not because they're fools and don't know what you 'know'. And it's why a larger phase 2b/3 wouldn't 'fix' anything. Btw this also doesn't mean another, bigger trial, will just yield the same results. Don't get confused. Drug responses are IID within a randomized population. A bigger trial could yield better - or worse - results. So why do it? Because it will yield more credible results. 2-73 has shown potential efficacy in Rett. But we have to show that credibly. Hence the full phase 3.
News 
Market Data 
Discover 
AI
