EBM is widely implemented in all decision making concerning health care. EBM is mostly based on metanalysis from studies that qualify for inclusion. Inclusion criteria may wary. However, you have not mention that EBM maybe fundamentally flawed because almost all paper that show not significance or negative date do not get published. The Cochrane library (http://www.cochrane.org) is the most trusted source of recommendations. Heurestics and logical thinking (mechanistic) can often yield much faster, cheaper and safer answer to biological questions than EBM based on metaanalysis imho.
Even Cochrane has some flaws to their analytics of data. An example of that is their approach to proning patients with severe ARDS where they stated that there was no convincing evidence of benefit. Those people have never been at the bedside nor really interpreted the PROSEVA trial. I know empiric data is the worst, but PROSEVA showed a clear benefit. The studies prior to that didn't have the same methods. Either way, I'm going to end my rant stating that I LOVE EBM but there are caveats. Even worse so, there are so many things that just cannot be studied due to ethical issues.