I think MDs are taught to believe data and graphs, but not *your* data and graphs. I think that's the explanation for why doctors in general do not "trust their own lying eyes." They've been taught to believe the statistics generated by RCTs which were run by big drug companies, analyzed by statisticians-for-hire, and rubber stamped by the FDA. Beyond that, they don't care about data and they know their own experience is "anecdotal" so they ignore it.

I've also been told by at least one statistician who worked with MDs that MDs really don't understand statistics. They don't understand confidence intervals, T-tests, non-parametic statistics, bayesian analysis, the limitations of correlation or whatever topic. (And why would they? It's a different rather confusing specialty.)

The problem is that MDs *think* they understand statistics better than they do. And they expect to be heeded because they got straight As in college or something.