Well, the point (I think?) of researching causation is so that you can make use of it. If I know that doing X makes me healthier, that's valuable for my decisi

Well, the point (I think?) of researching causation is so that you can make use of it. If I know that doing X makes me healthier, that's valuable for my decision of whether to do X; if doing X is merely associated with being healthier, but causation is not established (and in fact it's possible for doing X to harm you while still being associated with better health—consider "any harmful activity that young healthy people like to do and old people don't do"), then that's not useful for me. I dunno, I guess you could be interested in learning to classify people. If they do X, then that increases your estimate of their current healthiness regardless of the underlying causation. If you're an insurance company, an actuary, etc., then maybe that's all you need to know.