The central notion of evidence-based medicine is that our understanding of human biology is imperfect. Some of the idea we come up with for treating and preventing disease are effective, but most are not, worse than useless. So we need careful, rigorous statistical studies before implementing those ideas on a wide scale. A good example of doing this the wrong way was when when doctors started recommending having babies sleep prone, which roughly doubled the incidence of sudden infant death syndrome for the next several decades.
It seems to me that our understanding of psychology, sociology, economics, political science, and education is at least as imperfect as our understanding of biomedicine.