Bridging Health and Humanity: AI in EM

OVERVIEW

In this EMRA*Cast episode of Bridging Health and Humanity, host Natalie Hernandez speaks with Dr. Maya Yiadom, MD, MPH, MSCI, an associate professor and director of Precision Analytics and Data Integration in Emergency Medicine at Stanford, about how artificial intelligence is reshaping emergency care.

They discuss practical AI already in use (predictive analytics and ambient AI scribes), how AI can be designed from clinical workflows to improve detection and timeliness (for example, speeding recognition of STEMI), and the promise of tools that reduce documentation burden and support decision-making.

Dr. Yiadom also grapples with real risks — biased training data, subgroup performance, privacy and cloud constraints — and emphasizes protecting trainee learning while teaching residents how to use AI responsibly. She closes on an optimistic note: AI as a fail-safe that augments clinicians’ judgment rather than replaces the human art of medicine.

Objectives

By the end of this episode, listeners will be able to:

1. Describe current applications of artificial intelligence in emergency medicine, including predictive analytics, clinical decision support tools, and ambient AI scribes.
2. Explain a workflow-centered approach to AI development, emphasizing need-based innovation and the use of AI as assistive, fail-safe technology rather than replacement decision-making.
3. Discuss the potential impact of AI on resident education and clinical skill development, including strategies to balance efficiency with preservation of clinical reasoning and documentation skills.
4. Identify key equity considerations in AI implementation, including data representativeness, subgroup performance, and the risks of algorithmic bias in diverse patient populations.
5. Recognize the ethical, legal, and privacy challenges associated with AI integration into clinical workflows, including data governance and regulatory barriers.
6. Reflect on how AI may influence patient–physician relationships, including navigating patient-generated AI information and maintaining the human art of emergency medicine.

References

• Stanford Emergency Care Health Service Research Data Coordinating Center
• EMRA AI in EM: The Byte-Sized Benefits and Glitches
• ACEP Artificial Intelligence and the future of Emergency Medicine
• ACEPNow AI Scribes Enter the Emergency Department
• ACEP January 2024 - Artificial Intelligence to Improve Performance in Emergency Medicine (AIIPEM) Program
• JACEP Artificial Intelligence in Emergency Medicine: A Primer for the Nonexpert 
• NEJM AI in Medicine 
• The Harvard Gazette Machine Healing
• Am J Emerg Med Accuracy of cath lab activation decisions for STEMI-equivalent and mimic ECGs: Physicians vs. AI (Queen of Hearts by PMcardio)
• MedPage Today Man Hospitalized After Taking ChatGPT's Health Advice