Episode 68
Eleven Tips & One Tip of the Hat
Educators in Medicine,
In this newsletter, we continue our journey through the fundamentals of AI, its applications in medicine, and its transformative role in faculty development and education. Let’s dive into learning.
A friend and mentor shared the May/June FPM with me last weekend. I opened expecting the usual handful of practice pearls, and instead landed on a piece I wanted to forward to you all. Short post this week. I want to point you at it, and tell you what I’ve been doing lately that pairs nicely with it.
🧭 1: A Piece Worth Your Inbox Time
Dr. Sumana Reddy’s article in FPM, “11 Tips for More Effective AI Clinical Searches”, is a fine piece of work. She walks family docs through PICO-style prompting, asking the model to show its reasoning, requesting GRADE ratings, surfacing counterfactuals, and watching for hallucinations. It is practical, it is grounded in clinical realities (cost, side effects, access), and it is the kind of guidance our learners are hungry for. I appreciated that she names the bias traps too (automation, confirmation, expert bias, sycophancy). If you teach residents/students share it with them today.
🪞 2: Lately I’ve Been Going Meta
Here is the wrinkle I’d add. Lately I’ve been leveraging a lot of meta prompting, that is, asking the model to help me write the prompt before I ask the real question. Something like, “I want to compare two antihypertensives for a complex outpatient. Before you answer, draft me the ideal prompt that includes the patient context, the citations I should require, and the counterfactual checks I should ask for.”
Recently I had Claude prompt for itself a python tool to de-identify pdfs locally (on my computer) so I don’t upload any PHI for a research project. It did a solid job (after some coaching), but still far better than if I prompted myself.
The output of step one becomes the input of step two, and the answers get noticeably sharper. Pair this trick with Dr. Reddy’s eleven tips.
☕ 3: One Small Habit
Try this today. Pick one clinical question you would normally type in one breath. Ask the model to write you a better version of that prompt first. Then run the improved prompt. Notice the difference. That tiny extra step is, in my experience, the cheapest upgrade you can give your clinical AI use this week.
💌 As always, thanks for reading. Get in touch and let me know your thoughts!
Thank you for joining us on this adventure. Stay tuned for more AI insights, best practices, and more future editions of AI+MedEd.
For education and innovation,
Karim
Share this with someone, have them sign up here.