See my 39. Ignore the first sentence.
If there seems to be reason to believe it may have some efficacy, it will likely start the process of using control groups among those who might not normally actually be put under treatment...need a safer population (or such a disaster that it is acknowledged that so many will die that control groups become a "Hail Mary" potential stopgap) to gather enough discerning data..
But using places you mentioned as a way to do it does seem logical - if they ain't gonna get it there anyway, might as well use those stats.