I was watching House M.D. and to a degree I think the show epitomises the way medicine should be done. (Yes, I understand the show is detached from reality, but I'm talking about the attention doctors pay to their patients, notwithstanding House's demeanor.)
Our "science" encompasses many different aspects: big picture or visionary, exploratory, hypothesis-driven, correlative, engineering, correlative, etc. A combination of all this is what we call "science" and it's important to figure out how these different aspects weigh in any project we undertake (which I think we don't do enough of).
For example, sequencing the human genome is more along the former side, and drug discovery and protein design is more towards the latter. I personally would like everything to lead to hypotheses the lend themselves to many ways of falsification, which will lead to an understanding of how (the part in House where the camera is seen going through the body) but it doesn't happen all the time. The reality is that a lot of what we do is engineering (i.e., if it works, it's good enough).
Medicine in general is highly correlative, but it should be more balanced in terms of the other aspects.