Randomized controlled trials are not a silver bullet. There is a reason why meta-analyses spend a good amount of ink sorting out and excluding RCTs for lack of power or other design mistakes.

https://www.statnews.com/2023/05/02/do-masks-work-rcts-rando...