> This is evidence of what I call magical thinking–an undue focus on the clinical trial design as having incantatory power.

This isn't complicated. Once we have clinical trial evidence, the protocols tested form the standard of care.

With no standards of care (i.e. no tested vaccine, here), then we can only guess about what might be better than nothing. It's fair to make that decision on a balance of probabilities -- try something if it's 51% likely to help (absent negative side-effects, anyway).

But once we have such a standard, the very deliberate conservatism of our medical science means we need compelling evidence before supplanting it. I think it's very likely that a single dose of Pfizer/Moderna is adequately protective in the medium term (weeks to months), but this was not fully tested. It's not appropriate to start experimenting on the general population based on a hunch.

This community is IT-centric, so think of it like experimenting on a production system. Even if you are 80% sure that something's likely to work out, the risk of failure is so catastrophically high that you don't do it when any other option is open to you.