A Competency Framework for Medical AI Education: Mixed Methods Study

Area of Science:

• Medical Education
• Artificial Intelligence
• Health Informatics

Background:

• Clinicians face challenges adopting artificial intelligence (AI) in healthcare due to a lack of understanding, trust, and interpretation difficulties.
• Existing AI competency frameworks lack clinical specificity, and evidence for framework-based medical training is limited.

Purpose of the Study:

• To develop a specialized medical artificial intelligence (AI) competency framework.
• To design and pilot an AI training program based on the developed framework for medical professionals.

Main Methods:

• A mixed-methods approach integrating the UNESCO AI framework with the Miller pyramid model to create the competency framework.
• Expert input from 24 stakeholders and deductive content analysis were used for framework refinement.
• A 2-round Delphi process with 9 educators and a pilot workshop with 28 participants evaluated the training program's feasibility.

Main Results:

• A 6D 4-level medical AI competency framework was successfully developed, emphasizing AI foundations and application skills.
• The framework informed a 5-module training program, achieving full consensus among educators via the Delphi process.
• The pilot workshop demonstrated high participant satisfaction and engagement, with moderate confidence, indicating program feasibility.

Conclusions:

• The developed framework offers a structured reference for designing AI training in medical education.
• Preliminary findings support the feasibility of the AI training module, but broader, long-term evaluations are necessary.
• Future research should expand the framework and training program to diverse settings and assess long-term impact.