The 33rd annual Iowa Rheumatology Symposium took place at the Iowa City Marriott on Friday, October 13, 2023. Despite the ominous date, rheumatology enthusiasts from across the Midwest braved stormy conditions to enjoy an engaging and informational day covering the expanding scope of rheumatology.
Bharat Kumar, MD, MME, kicked off the event with introductions: industry vendors were available for engagement in the room next door, and raffle baskets filled with Iowa treats would be awarded to two attendees who filled their vendor passport. Additionally, the room next door offered vast views of the medical campus and periodic meals and snacks to punctuate the day’s activities.
Kumar then announced the first speaker of the day, Edward T. and Ellen K. Dryer Early Career Professor of Rheumatology from University of Michigan, Ray Zuo, MD. Kumar used an introduction provided by ChatGPT – a tactic he would repeat for his introductions for the remainder of the symposium. Zuo opened his presentation, Deadliest Catch: a NETs Story from SLE and APS to COVID-19, discussing neutrophils, his “favorite cells,” noting that the human body makes 100 million of them every day, the same number of stars in the Milky Way.
Following lunch, Kumar moderated a Patient Perspectives panel, in which two patients who also work in University of Iowa Hospitals and Clinics spoke freely and poignantly about their own experiences as patients, reminding clinicians to take the opportunity to actively listen to their patients themselves to create and nurture patient-doctor trust.
Interspersed between presentations were challenging cases presented by fellows, collectively called the Thieves Market. These highly interactive sessions included opportunities for attendees to scan a QR code to participate in brainstorming opportunities for additional testing or management.
In line with the theme of the expanding scope of rheumatology, attendees were challenged to speculate on the role of artificial intelligence (AI) in diagnostic reasoning. Second-year fellow, Ayesha Iftekhar, MBBS, and Kumar posed the question to attendees, “Can you beat ChatGPT in uncovering this mystery diagnosis?”
A lively presentation ensued, as Iftekhar layered more clues about the patient in question in between diagnostic reasoning breaks. Attendees worked at a rapid pace to offer their considerations, and minutes later, the same clues were being fed to the AI-powered language model. Several iterations later, Zuo accurately diagnosed the confounding case. Did ChatGPT?
Yes and no. In practicing their presentation the previous day, ChatGPT had presented Kumar and Iftekhar the correct diagnosis in some of the first of several prompts they gave the AI; during the live presentation, it did not. This led the group to conclude that, although ChatGPT is capable of consuming relevant information and seeming to apply sound logic to it, it is not capable of diagnostic reasoning as a trained rheumatologist uniquely is, nor reliable in producing the same results again and again.
Congratulations to the conference organizers for producing an engaging and thought-provoking CME conference, and our Design Center team for producing the accompanying materials!
Danielle Allen, MBA
Ben Davis, MD, PhD
Rhonda Dunn, ARNP
Bharat Kumar, MD, MME