That is quite a spike! Unfortunately the sharpness of the edge makes it pretty likely that a change in therapy is the cause -- time to death from hospital admis

That is quite a spike! Unfortunately the sharpness of the edge makes it pretty likely that a change in therapy is the cause -- time to death from hospital admission has a median of 10-12 days, but this is not a "wall", it's a median. The actual distribution is probably somewhat normal -- you'd expect a gradual increase and decrease around the intervention. We don't know how or when the Hydroxychloroquine was administered, but the mechanism of action seems to be prevention of infection early in the course of infection, generally before hospitalization. To me, this seems more likely coincidental or a reporting change.