I think it's important that studies are not complete and infallible statements of fact. They're observations that can have underlying flaws in methodology, biases, and just incorrectly interpreting data.

"I smoked weed everyday for years and now I'm schizophrenic" for years is a true sentence for a lot of people, but it's almost impossible to say that there's causation where that state means the same as "Because I smoked weed everyday for years, I'm now schizophrenic"

Self-medicating behavior has always existed.

"I take tylenol every day and now I have a brain tumor" doesn't mean Tylenol caused a brain tumor. Most people would say "That Tylenol usage was because of the headaches being caused by an existing, growing brain tumor"