No, I wouldn't. I would have to prove, for example, that someone started smoking after getting human papillomavirus and that there was a mechanism plausibly linking the infection with the craving for cigarettes.
I don't readily know how to make the linguistic distinctions I want to make here. Sure, if you want to say "They first have to have a diagnosis of cancer..." okay, I'm dead in the water.
That's not what I'm suggesting. I'm suggesting that we don't fully understand what causes cancer so it's possible that whatever living thing is a factor in causing cancer may also alter behavior such that it makes a diagnosis of cancer more likely.
This is probably not worth discussing further. "You cannot solve a problem using the same mental models that created it" but those "proven" mental models are a handy means to dismiss someone in conversation as getting off the rails.