Misiu I have already answered that point.
Again, and again, defining subgroups for a medical protocol is an ENTIRELY DIFFERENT THING from defining subgroups as part of a STATISTICAL PROCEDURE, eg the endpoints. The medical protocol needed this severe vs critical distinction, to filter out patients that could not participate to the trial. Fine. Now read the endpoints. None of the endpoints was split according to severe vs critical. None of them. So, different things.
I can repeat it ad infinitum…