Can AI Help Paramedics Make Better Decisions?

Can artificial intelligence tools like ChatGPT help paramedics with differential diagnosis and transport decisions? A groundbreaking new study explores AI's accuracy in prehospital care, offering a glimpse into a future where technology acts as a clinical safety net. Host Bram Duffee, PhD, EMT-P, sits down with lead author Eric D. Miller to discuss his team's fascinating findings. This first-of-its-kind study analyzed 104 real prehospital care records. ChatGPT was tasked with generating the most likely diagnosis, which was then compared to the diagnosis made by the treating paramedic. When the two differed, two emergency physicians reviewed the case to determine the more accurate conclusion.

Key Findings:
• High Accuracy: ChatGPT correctly predicted the diagnosis in approximately 75% of cases.
• Conservative Bias: In 46% of disagreements, the AI identified more critical conditions, suggesting it may help reduce under-triage.
• Strong Safety Profile: In only one case out of 104 did the AI suggest a potentially inappropriate transport decision.
Watch the full interview with Eric D. Miller to understand the nuances of this research. Subscribe for more expert discussions on the future of EMS.

Disclaimer: This content is for educational and informational purposes only and does not constitute medical advice. Always follow your local protocols and medical direction.

Book announcement: Bram Duffee is a co-author of the latest book about how hypnosis has the potential to revolutionize medical care for those struggling with emergency / acute distress. Hypnotic Communication in Emergency Medical Settings: For Life-Saving and Therapeutic Outcomes demonstrates how hypnotic communication has the potential to improve patient outcomes in emergency care, integrating insights on the connection between mind and body for paramedics and other first responders. https://routledge.pub/Hypnotic-Commun...

Stay up to date: Review the most recent research in EMS through the monthly article reviews put out by the International PreHospital Medicine Institute https://iphmi.com/ems-articles-of-the...

Connections: If you want to connect with the EMS Research Podcast staff check out our website at www.ProfessorBram.com If you enjoyed this episode please consider subscribing to EMS Research on YouTube: / @4emsresearch

Reference: Miller, E. D., Franc, J. M., Hertelendy, A. J., Issa, F., Hart, A., Woodward, C. A., Newbury, B., Newbury, K., Mathew, D., Whitten-Chung, K., Bauer, E., Voskanyan, A., & Ciottone, G. R. (2025). Accuracy of commercial large language model (ChatGPT) to predict the diagnosis for prehospital patients suitable for ambulance transport decisions: Diagnostic accuracy study. Prehospital Emergency Care, 29(3), 238–242.