Given how absolutely trivial it is to do a randomized intervention study here (you can't do double-blind, but dose-response is easy enough) there's not a lot of

Given how absolutely trivial it is to do a randomized intervention study here (you can't do double-blind, but dose-response is easy enough) there's not a lot of reason to care about n=1 results except plausible reports of rare seriously negative outcomes. If it works, it would be within the resources of an ambitious high school class to prove it. If it doesn’t, "more research will be needed"