It's a great question, "how do we properly assess students and enhance patient care", and one I'm not sure we'll ever have a satisfactory answer for.
I'm of the opinion that we're at a level of knowledge and technology that we're about to - or already have - crossed the threshold of where rote memorization (so called "walking around knowledge") isn't the best/most efficient way to train doctors.
This was illustrated by the following Tweet:
View attachment 295432
Now, you can interpret that photo many different ways, but I interpret that as "core medical knowledge has at least doubled over the last two decades".
I've said this before about RadOnc's board certification process, chiefly around the oral boards, but I think moving to practical exams where mastery and application of resources is what's most important to be a "good" physician in 2020 and beyond.
Medical schools should heavily emphasize how to read and critique the literature, discern what makes good and bad studies/books/practices etc. At the medical student level, I haven't really envisioned how that process would work. At the RadOnc level, I think we should be plunked down at our normal workstation with all of our normal resources available. I think we should be thrown a mix of "zebra" and "bread-and-butter" cases and judged based on how we utilize all available resources to arrive at a treatment plan. It's a part of Neurosurgery's oral boards - they don't do theirs until 3 years after residency, and you submit all of your cases you've operated on in that time. The board examiners then look at how you've practiced medicine and grill you on your outcomes and what you could have/should have done differently etc.
Medical knowledge is only going to continue to explode and become more complicated. We literally walk around with computers in our pockets able to access the sum total of human knowledge. Me being able to memorize radiation trials from the 1970s is nowhere near as important as my ability to critically evaluate some new trial with a jacked up control arm concluding that we should utilize some $40,000,000 carbon ion treatment for Gleason 7 prostate cancer.