As a layperson it feels to me like research in the "megadose on vitamins to prevent illness" area is kind of tainted by the legacy of Linus Pauling, who was no doubt a genius in his fields and whose published literature advocated the position that vitamin C megadose would reduce cancer deaths by 75% ( https://www.theatlantic.com/health/archive/2013/07/the-vitam... ). I guess we know how to avoid the trouble -- make h…
That's an exceptionally oversimplified view on the placebo effect.
There's a very obvious issue: If you take something that helps you due to placebo it may still also harm you due to its side effects. Some vitamins have been shown to increase overall mortality.
But there's another issue, and that is that often what people observe isn't the placebo effect, but something called "regression to the mean". Which is basically the effect that you may take something and you get better, though you would've gotten better anyway, because that's the natural progression of the disease you have. This is very likely a bigger reason for all the "I took it myself and got better, so I know it works!"-stories out there than the placebo effect. And the "who cares why it works?" argument doesn't count for regression to the mean.