What I found noteworthy were Dr Dhody's remarks as
Post# of 148184
he made it very clear and plausible that endpoints like "discharge time from hospital" our GILD friends used are subjective and effectively bogus, and that mortality at 28 days is the best and most objective endpoint for severe/critical cases.
Also regarding change of endpoints during a study he pointed out that it would be unethical.
That's what we need to keep in mind: unlike GILD who couldn't prove a stat sig effect on mortality in three studies, we are aiming for the hardest and most meaningful endpoint.