Epidemiology is useful in identifying hypotheses for further research.

For me yes. Epidemiology studies resulting in significant correlations, such as Harris-Benedict and Framingham, showed me simple behavior modifications which improved my health. They also showed the effects of the things I can't change, such as age, height and sex.

You can point out plenty of flaws epidemiological studies, but RCT's are just as limited. All sorts of studies have their strengths and weaknesses. Fact of the matter is the answers you arrive at are as much a function of the questions you ask as anything else.