This is a substantial oversimplification.

Hydroxychloroquine looked promising for a while, and was being studied in an RCT, but that RCT was ended before completion because of the release of an observational study that suggested strongly negative outcomes from Hydroxychloroquine.

Later, that observational study was retracted because it was discovered to be based on falsified data.

So, unfortunately, we don't know very much about Hydroxychloroquine's effectiveness and risks in use -- but we do know that it has become a massive political hot potato, as your comment indicates.

> but we do know that it has become a massive political hot potato

I posted this a few days ago and got an interesting swing of votes on that comment since then:

> So, unfortunately, we don't know very much about Hydroxychloroquine's effectiveness and risks in use

For effectiveness it looks like we actually do, thanks to an experiment Switzerland accidentally ran on their population over May and June [0]. On May 27, they banned use of hydroxychloroquine, and 13 days later the death rate among resolved cases nearly tripled. Just after the spike began, on June 11 they reversed the ban and 13 days later the death rate drops back down. The graph from that article [1] is very stark, comparing the rate from late March through early July.

[0] http://www.francesoir.fr/societe-sante/covid-19-hydroxychlor...

[1] http://www.francesoir.fr/sites/francesoir/files/20200713_fs_...

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.

Ahhh I meant to say "pretty unlikely that a change in therapy is the cause" above.

This was presumably done a while ago, when hospitals were trying all sorts of weird things.

The "HCQ doesn't work"-narrative is entirely political, not scientific.

The original "quack treatment", which is prophylactic zinc+HCQ (as an ionophore) for at-risk groups, has never been shown to be ineffective in an RCT. It is still being studied.

Of course the studies that show it to be effective aren't of the best quality, but they aren't entirely meaningless either.

Logically, the treatment ought to be for exposed people who aren't yet showing symptoms. Prime candidates would be family members of infected people.