“

That means it cannot determine cause and effect, the meta-analyses that the authors cite are also mainly based on observational data. On the other hand, large observational studies sometimes are able to detect effects in some groups or for more rare outcomes that trials are not empowered to see, so these studies are still valuable.