Correlation is the key word here. Studies like this can find associations, but they're usually not strong enough to point to something (in this case, coffee drinking) as a cause of something else. They also have to try to factor out other things like smoking (coffee drinkers tend to smoke more often, so that sort of association has to be weeded out).
So can drinking lots of coffee help you live longer? It's possible, I guess, but it isn't going to make much of a difference. Tell me, what do you think about studies like this?