The problem is exactly what you say in the last paragraph.
For example, despite the fact that CVP and ScvO2 have been proven not to correlate with outcomes, despite the fact that the PAC has been proven repeatedly to cause more problems than it solves, despite the fact that the literature is full of studies suggesting that less fluids and early vasopressors are better, despite the fact that the same literature suggests, logically, that there is no sense giving fluids where they don't increase cardiac output, despite the fact that few measurements correlate well with that increase, I see uninformed (would almost say uneducated) doctors applying EGDT day in and day out, flooding patients with fluids. Even the current Surviving Sepsis Campaign guidelines are not rooted only in evidence, but also in tradition. Speaking about tradition, one has to point out how internists keep giving abnormal saline all around the country, when all evidence points against it. Same goes for not understanding SIRS, the great plague of this century. It doesn't mean sepsis in 50% of the cases, but a lot of people get agitated when they diagnose it. I have been called for "SIRS" in anemic elderly patients who were tachycardic, hypotensive and had cold extremities. And the list goes on. "EBM" and its algorithms probably kill as many patients as it save, by having them undergo unnecessary/wrong treatments and procedures.
Doctors generally suck at math, as a group, so why let them interpret statistical conclusions in the first place? Letting them play EBM, when they have no friggin idea what's statistically correct and what is not, is like letting children play with guns or decide world politics.