Project Overview

Description:

The Randy Rhodes AI patient simulation is designed to provide first and second-year medical students with a realistic practice tool for patient interactions. The simulated patient, Randy, is a diabetes patient.

Purpose:

To replace the antiquated EPBLM system and offer a more engaging, interactive learning experience for medical students.

Goals

Primary Objectives:

Create a realistic AI patient for medical student training.

Enable students to practice diagnosing patient conditions through conversation.

Improve the learning experience compared to the EPBLM system.

Measurable Targets:

High student engagement and satisfaction with the simulation.

Accurate simulation of patient trust dynamics.

Development and Implementation

Technology Stack:

Unreal Engine: Used for building the simulation environment.

Metahumans: Utilized for creating the physical body of the patient.

Convai: Managed the personality and body language of the patient.

ChatGPT: Created personality documents to inform the Convai AI model.

Process and Stages:

Initial design and planning of the patient simulation.

Development of the physical model using metahumans.

Integration of personality and body language with Convai.

Creation of detailed personality documents with ChatGPT.

Testing and refinement of the simulation.

Key Milestones:

Completion of the physical model.

Successful integration of Convai and ChatGPT.

Positive feedback from initial student interactions.

Achievements

Significant Breakthroughs:

Development of a trust document to dynamically manage patient trust.

Successful integration of various technologies to create a cohesive simulation.

Metrics and Data:

Overwhelmingly positive student feedback.

Increased engagement and enjoyment reported by students.

User Feedback:

Students appreciated the realism and interactive nature of the simulation.

Challenges and Solutions

Major Obstacles:

Integrating multiple platforms to function seamlessly.

Ensuring a realistic and engaging user experience.

Solutions:

Collaborative effort across development teams to ensure compatibility.

Rigorous testing and feedback loops to refine the simulation.

Outcome

Final Results:

The simulation met and exceeded initial goals.

High levels of student satisfaction and improved learning outcomes.

Impact:

Enhanced the educational experience for medical students.

Provided a practical tool for developing patient interaction skills.

Future Steps:

Explore further improvements and expansions of the simulation.

Consider additional simulations for other medical conditions.

Conclusion

Summary:

The Randy Rhodes AI patient simulation project was a success, providing a modern, interactive learning tool for medical students.

Reflections:

The team enjoyed the development process and overcame significant challenges, leading to a highly effective educational tool.

Excited about future opportunities in the education sector and continuing to innovate in AI-driven simulations.