Related Experiment Video
Updated: Jul 12, 2025
Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
Published on: June 13, 2025
Bariatric Evaluation Through AI: a Survey of Expert Opinions Versus ChatGPT-4 (BETA-SEOV)
Amir Hossein Davarpanah Jazi1, Mohammad Mahjoubi2, Shahab Shahabi3
1Department of Surgery, Minimally Invasive Surgery Research Center, Division of Minimally Invasive and Bariatric Surgery, School of Medicine, Rasool-E Akram Hospital, Iran University of Medical Sciences, Niyaesh Avenue, Sattar Khan Street, Tehran, Iran.
Background:
Recent advancements in artificial intelligence, such as OpenAI's ChatGPT-4, are revolutionizing various sectors, including healthcare. This study investigates the use of ChatGPT-4 in identifying suitable candidates for bariatric surgery and providing surgical recommendations to improve decision-making in obesity treatment amid the global obesity epidemic.
Methods:
We devised ten patient scenarios, thoughtfully encompassing a spectrum that spans from uncomplicated cases to more complex ones. Our objective was to delve into the decision-making process regarding the recommendation of bariatric surgery. From July 29th to August 10th, 2023, we conducted a voluntary online survey involving thirty prominent bariatric surgeons, ensuring that there was no predetermined bias in the selection of a specific type of bariatric surgery. This survey was designed to collect their insights on these scenarios and gain a deeper understanding of their professional experience and background in the field of bariatric surgery. Additionally, we consulted ChatGPT-4 in two separate conversations to evaluate its alignment with expert opinions on bariatric surgery options.
Results:
In 40% of the scenarios, disparities were identified between the two conversations with ChatGPT-4. It matched expert opinions in 30% of cases. Differences were noted in cases like gastrointestinal metaplasia and gastric adenocarcinoma, but there was alignment with conditions like endometriosis and GERD.
Conclusion:
The evaluation of ChatGPT-4's role in determining bariatric surgery suitability uncovered both potential and shortcomings. Its alignment with experts was inconsistent, and it often overlooked key factors, emphasizing human expertise's value. Its current use requires caution, and further refinement is needed for clinical application.
