I mean, people have tried doing studies like that for CQ/HCQ. They are riddled with questionable assumptions (that patient groups are similar or treatment regim

I mean, people have tried doing studies like that for CQ/HCQ. They are riddled with questionable assumptions (that patient groups are similar or treatment regimens are equivalent) that make their conclusions completely unreliable. In the end we will most likely find that all positive anecdotal evidence found was just confirmation bias. That's why RCTs are so important.