Based on that reasoning, we shouldn't give any pharmaceuticals to coronavirus patients. Even simple fever relievers, like acetaminophen, or oral rehydration fluids should be prohibited because we just don't have enough data to know how they'll interact with COVID-19 specifically.

I think this is the fairest counter-argument I've heard so far about safety of HCQ in Covid-19. I still think the background of "you're in a house burning down, some tests tentatively indicate this may be effective..." weigh more against the risk of side-effects or worse progression due to CQ/HCQ. We ought to be able to devise some experiments to add certainty, using mainly retrospective data now available.

Most SLE/Lupus patients take HCQ, and reportedly none in the cohort of the doctor I know have sufficiently severe symptoms to qualify for pCR testing to even establish if they have Covid19. This is Spain with (currently) 708 cases per MM. Of course, caveats like absence of evidence is not (necessarily sufficient) evidence of absence. It may be that SLE or other immuno-modulation treatments have some effect - someone should do analysis with matched non-SLE HCQ and non-HCQ prophylaxis cohorts. But back to my main point with the house burning metaphor for exponential infection - the risk landscape demands we take more risk, give HCQ widely instead of waiting for the conclusion of prospective studies. We can withdraw it if it is shown to be no better than placebo or worsens outcomes when enough data is available to conclude.