A few medical schools have integrated a module on biostatistics and epidemiology into their curriculum, but it's not common (or at least it wasn't ~20 years ago, if that's changing I'm happy)

It is changing, I had bio stats and epidemiology. How much I paid attention to it is up for debate. I can say that epidemiology from my experience is easy to have a grasp on what's going on, but to understand the nuances of data you have to go through rigorous training.

such as specificity and sensitivity, much less positive predictive value (also very important to know when interpreting test results like doctors do every day)

This is the only thing I think physicians are getting good at. although the older ones may not have the palate that younger physicians have for this kind of thing. Compared to epidemiology sensitivity and specificity have daily implications on patient outcomes so it becomes imperative whereas epidemiology isn't as important for day to day tasks. I like your part on how lay people, such as myself, have little idea of what a computer scientist does and how that can cause frustration. I'm finding as I get older the less I know about what other people do and the more I should listen instead of speculate.