There are many kinds of studies. Retrospective vs. Prospective, Randomized Controlled Trial, and Double-blind. Of all these, many consider the Randomized Controlled Trial the most important feature - more important even than the double-blind aspect.
Double blind means that neither the doctor nor the patient knows if he has received a placebo. This is important in treatments where the placebo effect might be relevant, such as treatment of headaches or pain in general. It also avoids the potential that if the doctor knows who is receiving the real treatment, he might affect the outcome with that very knowledge.
The problem with double blind studies is they are more difficult and time-consuming to manage - especially in a crisis situation such as we have now. I don’t think the placebo effect is likely to be a major factor in HCQ treatment, and I think doctors in these trying times are all trying their best to give everyone the best treatment they can. I doubt that with the stakes as high as they are, any ethical doctor would bias the outcome of treatment based only on his knowledge of who is receiving the real treatment.
So here is my suggestion: Skip the double-blind and placebo step for now. But do utilize randomized assignment of who gets HCQ and who doesn’t. That would remove what I think is the biggest source of bias in the results - the decision on who to treat with HCQ and who not to to treat. If that one decision is made in a truly random manner (subject to some reasonable and objective guidelines) then we can be fairly sure that any difference in outcome is truly due to the use of HCQ (or HCQ in some combination).
I think they call that RCT (randomized controlled trial).