MedSimAI: Simulation and Formative Feedback Generation to Enhance Deliberate Practice in Medical Education
Project Overview
The document highlights the application of generative AI in education, specifically through the introduction of MedSimAI, an innovative simulation platform aimed at improving medical education via realistic and interactive patient encounters. It tackles significant challenges present in traditional simulation-based learning, including issues of scalability, accessibility, and the quality of feedback provided to students. By leveraging large language models, MedSimAI creates dynamic clinical interactions and offers immediate feedback, thereby fostering self-regulated learning and encouraging deliberate practice among medical students. A pilot study evaluating the platform indicated promising levels of engagement and perceived educational benefits, although it also identified areas for enhancement, particularly regarding students' use of self-regulated learning features. Overall, the findings suggest that generative AI can significantly enrich educational experiences in the medical field, while also underscoring the need for continued refinement to maximize its effectiveness in promoting autonomous learning.
Key Applications
MedSimAI: AI-powered simulation platform for medical training
Context: Medical education for first-year medical students at a northeastern medical school in the United States
Implementation: Co-developed with a multidisciplinary team, tested in a pilot study with 104 students, providing AI-simulated patient encounters and feedback.
Outcomes: Increased accessibility to clinical practice, immediate structured feedback, and positive student engagement with opportunities for deliberate practice.
Challenges: Underutilization of self-regulated learning features, occasional AI oddities in patient interactions, limited diversity in patient representation.
Implementation Barriers
Accessibility
High costs and resource demands of traditional simulation methods limit accessibility to training opportunities.
Proposed Solutions: Utilizing AI to create scalable, cost-effective simulation options.
Quality of Feedback
Challenges in delivering high-quality, timely feedback due to limited instructor training and resources.
Proposed Solutions: Implementing automated assessment frameworks that provide immediate feedback.
Engagement
Low engagement with self-regulated learning features among students.
Proposed Solutions: Embedding MedSimAI encounters into course requirements and providing structured reflection exercises.
Project Team
Yann Hicke
Researcher
Jadon Geathers
Researcher
Niroop Rajashekar
Researcher
Colleen Chan
Researcher
Anyanate Gwendolyne Jack
Researcher
Justin Sewell
Researcher
Mackenzi Preston
Researcher
Susannah Cornes
Researcher
Dennis Shung
Researcher
Rene Kizilcec
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Yann Hicke, Jadon Geathers, Niroop Rajashekar, Colleen Chan, Anyanate Gwendolyne Jack, Justin Sewell, Mackenzi Preston, Susannah Cornes, Dennis Shung, Rene Kizilcec
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai