On BART I met a graduate student in epidemiology. “What are the strengths and weaknesses of epidemiology?” I asked. Strengths:
1. It asks important questions. What causes cancer? for example.
2. The results are useful. They can guide public policy. If you learn that smoking causes cancer, you can start an anti-smoking campaign. Epidemiological results can also lead to informative experiments: Epidemiology suggests that X causes cancer, you do an experiment to test that conclusion.
Weaknesses:
1. Health is complicated, controlled by many things. Presumably this is why studies often have conflicting conclusions.
2. There is enough flexibility in data analysis that your original hypothesis may influence the way that you analyze your data.
I use epidemiology all the time — here, for example. It often makes an interesting idea more plausible. My ideas about depression, derived from studying the effects of seeing faces, became more plausible to me because of the epidemiology of depression.