>>79075602A lot of those studies that involve "People did X for 30 years and Y occurred more/less often" can point out a correlation, but don't directly point to a mechanism of action. It might be that there is some compound in coffee (caffeine or otherwise) that provides some long-term benefit. If so, it is possible there are sources of it other than coffee and/or it could be isolated and administered as a supplement.
Or it could be purely correlative!
As an example, if the study began with a narrow age cohort (5-10 years) that were born into a stable economic situation, and people in their region before and after that age cohort received poor childhood nutrition for some reason. Studies are supposed to account for that. Most even do. But then those caveats of "Well, we're not sure about [detail]" don't make headlines.
Kind of like how women in the United States with a hobby of riding horses live longer.
It's nothing to do with the horse. Women who can afford to own a horse can also afford good healthcare. 
