Proceedings of the 8th International Academic Conference on Teaching, Learning and Education
Year: 2024
DOI:
[PDF]
Case Simulation Large Language Model (LLM) Chatbot for Radiology Education: Guiding Learning in CXR Scenarios
Long Yin Huen, Lei Wang, Zongyou Cai, Lun M. Wong, Tiffany Y. So
ABSTRACT:
Clinical case simulations are an integral part of medical education, offering students the opportunity to apply theoretical knowledge to practical, real-world scenarios. However, traditional case simulation exercises in radiology can be resource intensive. Additionally, as existing radiology educational materials typically have limited engagement with students, there is a growing need for innovative, scalable tools that provide the same level of interactive learning. We aimed to develop a case simulation chatbot to guide students through clinical scenarios involving the identification and interpretation of lines and tubes on chest X-rays. We aim to implement this chatbot seamlessly within a radiology education mobile app. The chatbot is powered by GPT-4o, a large language model (LLM) developed by OpenAI. Published guidelines involving lines and tubes (i.e., endotracheal tubes, central venous catheters) and drafted case data for relevant clinical scenarios were pre-learned by the chatbot before the start of the interaction. The student is briefed with background information and prompted to provide a decision or intervention at various stages of the case. The chatbot reviews the student’s decision-making process, providing tailored feedback and guiding the student through the potential consequences of their decisions. Once the case concludes, the chatbot will provide an overview of the student’s overall performance and ask follow-up questions to reinforce the student’s knowledge on topics that need improvement. The use of chatbots incorporating case simulations and interactive questioning allows tailored learning experiences and allows students to engage in self-learning practices at any time and place with real-time evaluation and feedback.
keywords: active learning; artificial intelligence (AI); e-learning; gamification; medical imaging