Utilizing generative conversational artificial intelligence to create simulated patient encounters: a pilot study for...

Area of Science:

• Medical Education Technology
• Artificial Intelligence in Healthcare
• Simulation-Based Training

Background:

• Generative conversational artificial intelligence (AI) presents significant opportunities for enhancing medical education.
• Virtual patient avatars can offer realistic training scenarios for medical professionals.
• No-code platforms enable accessible development of customized AI training tools.

Purpose of the Study:

• To evaluate the feasibility of using a no-code generative AI solution for creating interactive 2D and 3D virtual patient avatars for medical training.
• To assess the potential of AI-driven simulations in improving medical education and trainee preparedness.

Main Methods:

• A virtual patient avatar was developed using the 'Convai' platform with a specific patient backstory.
• Trainee anaesthetists used the AI avatar to practice responding to patient questions regarding interscalene nerve blocks.
• Feedback was collected from anaesthetists to evaluate the tool's feasibility, user-friendliness, and accuracy.

Main Results:

• The AI tool received high median scores for intuitiveness (9 [7-10]) and accuracy (8 [7-10]).
• 87% of surveyed anaesthetists reported comfort using the virtual patient model.
• The simulation provided realistic scenarios for practicing patient communication.

Conclusions:

• No-code generative AI platforms can facilitate the creation of customized medical training tools, enabling practice outside of traditional settings.
• AI-powered simulations offer a valuable supplement to traditional training methods, improving accessibility and realism.
• Further research and refinement are necessary to address limitations such as potential inaccuracies before AI can fully replace human interaction in training.