the quality of most are bad.
The studies coming from France are horribly bad. This is the guy evangelically touting them
https://techstartups.com/wp-content/uploads/2020/03/Didier-Raoult.gif
Besides not being randomized, double-blind controlled, their methodology was bad. Here's the layman's conclusion:
"So we compared the people who got it vs who didn't, and after we excluded patients who died, we found that it worked 100% of the time unless it didn't"