ChatGPT as a Tool for Medical Education and Clinical Decision-Making on the Wards: Case Study

Area of Science:

• Artificial Intelligence in Medicine
• Medical Education Technology
• Natural Language Processing

Background:

• Large language models (LLMs) offer advanced AI capabilities for accessing medical knowledge.
• LLM adoption in medical education is growing, but empirical evidence for clinical settings is limited.

Purpose of the Study:

• To qualitatively evaluate potential use cases and limitations of LLM technology in ward-based medical education.
• To explore the integration of LLMs into real-time clinical teaching environments.

Main Methods:

• Exploratory evaluation of ChatGPT-3.5 (OpenAI) in general internal medicine attending rounds.
• Qualitative phenomenological inquiry analyzing conversation logs and clinical notes.
• Utilized web-based interface for conversational and discrete queries.

Main Results:

• Identified use cases: addressing knowledge gaps, differential diagnosis, cognitive support, and complex care management.
• Explored additional uses: patient communication, ethics education, documentation support, and productivity.
• Highlighted limitations: LLM biases, misinformation, ethics, and health equity concerns.

Conclusions:

• ChatGPT serves as a novel tool for ward-based learning via rapid querying and discussion.
• Further research is essential to understand educational contexts, risks, and impacts on trainee development.