ll the
ea lot closer to mindless empiricism than it ought to be.
is one. It's often not based on evidence because the evidence is so often t oo bent to be useful, and doctors are well aware that a research result doe s not mean reality. So it often falls back on what they found worked before and what other doctors do. Unfortunately a good bit of it is much worse th an that too. The level and degree of reality of evidence varies a lot by to pic.
The Cochrane Collaboration exists to make it less true, but it was only set up in 1993.
Published medical statistics are bent by the usual process - if a trial sho ws that something doesn't work it doesn't get published, but if you do enou gh trials, statistical fluctuation will eventually show up a significant be nefit, and that is the trial that will get published.
Most trials are done on young white Anglo-Saxon males (because there are pl enty of them around the places that create new drugs to test, and quite a f ew of them are willing to risk their health for money). The population tha t takes the drug is usually older and half female, more ethnically mixed, a nd a lot more prudent.
That doesn't "bend" the statistics, but it does make them less useful than they might have been.