EP12 Dong Yao, MD and Shivam Vedak, MD Stanford (Part 1)
Engineering the Future — AI Triage, Prompt Design, and Clinical Decision Support
Guests: Dr. Dong Yao and Dr. Shivam Vedak, Emergency Physicians and AI Researchers, Stanford University
Release Date: 4/29/25
Episode Summary:
In this episode of the STAT AI Podcast, Drs. Ethan Abbott and DJ Apakama, explore two critical frontiers at the intersection of emergency medicine and artificial intelligence: smarter triage and smarter AI design.
First, DJ and Ethan break down a landmark study published in NEJM AI examining how AI-based triage decision support impacted patient flow, clinical decision-making, and emergency department operations across three hospitals. They discuss how machine learning models embedded into triage workflows helped nurses better identify critical patients, reduced waiting times, and improved overall throughput — while preserving human clinical judgment.
Then, they’re joined by Stanford’s Dr. Dong Yao and Dr. Shivam Vedak, two leading voices in the emerging field of clinical prompt engineering. Dong and Shivam share practical lessons from the frontlines of AI tool development, emphasizing why prompt design — the art of crafting precise, reliable model inputs — is just as important as the model itself. They offer insights into how healthcare-specific prompting can bridge gaps between cutting-edge technology and real-world clinical needs.
In This Episode, We Cover:
🧠 The challenges and limitations of traditional triage tools like ESI — and how AI is reshaping risk stratification.
📈 How AI decision support reduced median ED wait times and improved triage precision.
🛠️ Why prompt engineering is emerging as a critical discipline for clinical AI tools.
🧩 Techniques for crafting effective prompts that align with clinical reasoning, not just technical performance.
⚖️ Balancing human judgment and AI recommendations at the bedside.
🚀 The future of AI in emergency medicine: from smarter triage to better diagnostic support.
Referenced Paper:
Taylor RA, et al. The Impact of Artificial Intelligence-Based Triage Decision Support on Emergency Department Care. NEJM AI, 2024.
Takeaway Quote:
“In clinical AI, the model matters — but how you talk to the model matters just as much.” — Shivam Vedak, MD
Connect with the Guests:
Follow STAT AI:
🌐 Website: stataipodcast.com
🐦 Twitter: @STATaipodcast
🎧 Listen: Apple Podcasts | Spotify | Prime Music
