Just because two factors are correlated does not mean that one causes the other. My favorite example of this is from my grad school days, when my advisor mentioned that the crime rate follows ice cream eating habit. Therefore ice cream eating causes crime, right? Well, no, both happen to be warmer-weather activities.
The Chicago Tribune had an interesting thought piece on this today by Ronald Kotaluk, their science reporter: False alarms and the need to know (free registration)
What are people to make of studies suggesting that antidepressants may increase the risk of suicide in adolescents, that breast-enhancement surgery may also make some women more prone to kill themselves, or that antibiotics may increase the chance of breast cancer?
Well, scientists don't know what to make of them either, and they make no apologies.
The article describes three types of studies, each with greater complexity that scientists (mostly drug research) conduct. In every case, the issue becomes both when and how to present the findings. It's more difficult than we like to admit to describe waht the results mean and how conclusive they are.