I think the wildcard is that yes, we know there we

I think the wildcard is that yes, we know there were 87 deaths, but do we know how many of those deaths were inside or outside of 28 days? I think this is key to understanding why critical results were only included. I just think that if there were deaths in the severe population, it's logical to assume this population would take longer to die, and thus won't count as an event when measuring mortality.