Leveraging Datathons to Teach AI in Undergraduate Medical Education: Case Study

Area of Science:

• Medical Education
• Health Informatics
• Data Science

Background:

• Artificial intelligence (AI) and machine learning (ML) are increasingly vital in clinical practice.
• Future physicians require understanding of AI/ML's impact on patient care delivery.

Purpose of the Study:

• Describe trainee-led datathons as an effective method for teaching data science and ML skills.
• Provide insights into datathon implementation and lessons learned for future initiatives.

Main Methods:

• Detail two multi-institutional datathons organized by MDplus.
• Assessed efficacy via opt-in, deidentified post-datathon surveys.
• Analyzed survey responses to evaluate participant experience and identify areas for improvement.

Main Results:

• Approximately 200 medical trainees participated in digital datathons (2023-2024).
• Participants showed improved Python skills for medical data analysis.
• Survey respondents reported enjoyment and enhanced ability to derive clinical insights from data.

Conclusions:

• Datathons are an effective, cost-effective educational tool for medical trainees.
• Datathons improve trainees' capacity to generate clinically meaningful insights from data.
• Datathons enhance skills in data science and AI for improved patient care.