What is learning in med ed?

I’ve just enrolled in the University of Dundee distance PGCert in MedEd.

It’s a little daunting at first. Moodle is a visual cacophony.

The first activity is to ask ourselves what is learning. This gave me a mini-existential crisis. Let me explain.

My position for the last few years has been medical education is the key to improving patient outcomes in the longer term. As medical care has become more complicated and there is more we can do per patient/per problem, our skill set needs to focus on:

a) what is the minimum amount of raw clinical info we need in our working memory to allow clinical decision making heuristics to flow nicely

b) how to access and apply the evidence to this patient in front of me

I think this now may be wrong. AI can be thought of as code that handles uncertainty. That is what a doctor does more of compared to most other health care professionals. Specifically looking at machine learning, I can see that eventually machine learning that has access to huge amounts of patient training data will ultimately beat any doctor. How can a urologist who sees 20,000 patients in his career ever compete with ever improving code that has dealt with 10 million patients?

So what is our role in med ed? I think we have to realise that the role of the doctor will change. Instead of being primarily diagnostic decision makers, we will be those professionals who talk though the options with the patient, what is actually happening to them and basically putting a human coating on the raw machine inputs/outputs. Our role in physical examinations and some initial acute management will probably persist for a while.

So what is learning? If teaching is what goes in, learning is what comes out. It is the improvement in knowledge, thought processes and physical processes that make you better at what you are trying to do.

And if improving patient outcomes is what we are trying to do, maybe this means re-examining how we really make clinical decisions, and recognising the cold truth that perhaps a decent chunk of our decision making could probably be done more safely by AI.