Related Experiment Video
Updated: May 23, 2025
03:14
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
474
Virtual Patient Simulations Using Social Robotics Combined With Large Language Models for Clinical Reasoning Training
Alexander Borg1, Carina Georg2, Benjamin Jobs1
1Division of Rheumatology, Department of Medicine Solna, Karolinska Institutet, Karolinska University Hospital, and Center for Molecular Medicine (CMM), Stockholm, Sweden.
Journal of Medical Internet Research
|March 7, 2025
Summary
A social robotic virtual patient combined with a large language model offers a more authentic clinical reasoning training experience for medical students than traditional computer-based simulations.
Area of Science:
- Medical Education Technology
- Artificial Intelligence in Healthcare
- Clinical Reasoning Training
Background:
- Virtual patients (VPs) are computer simulations used in health professions education to develop clinical reasoning (CR).
- Inadequacies in CR can negatively impact patient safety.
- Large language models (LLMs) and social robots present opportunities to enhance VP realism and interactivity.
Purpose of the Study:
- To explore the added value of a social robotic VP platform integrated with an LLM for CR training.
- To compare this novel platform against conventional computer-based VPs for medical students' CR development.
Main Methods:
- An explorative proof-of-concept study conducted in Sweden (May-July 2023).
- 15 medical students used both a social robotic VP and a computer-based VP.
- Quantitative assessment of self-perceived experience and qualitative in-depth interviews were performed.
Main Results:
- The social robotic platform was rated as more authentic (P=.04) and provided a greater overall learning effect (P=.01).
- Students found the social robot superior for training CR, communication, and emotional skills.
- Identified limitations included technical issues and suggestions for improved facial expressions and complex case simulations.
Conclusions:
- LLM-enhanced social robotic VPs offer an authentic and engaging learning experience for medical students' CR training.
- This technology shows promise for achieving educational outcomes in medical curricula.
- Further improvements can enhance the platform's effectiveness and user experience.
