Ontology-based student testing through clinical guidelines: An AI approach.

Area of Science:

• Medical Education
• Artificial Intelligence
• Clinical Decision Support

Background:

• Leveraging 25 years of experience with the GLARE system.
• Addressing the need for effective training methods for medical students in clinical guideline adherence.
• Bridging the gap between clinical decision support and educational applications using AI.

Purpose of the Study:

• To propose a novel facility for creating and evaluating medical student tests based on computer-interpretable clinical guidelines (CIGs).
• To enable teachers to define tests by selectively hiding parts of CIGs and presenting case studies to students.
• To automatically assess student actions against CIGs and provide explanations for differences.

Main Methods:

• Development of a new educational facility integrated with the GLARE system.
• Utilizing computer-interpretable clinical guidelines (CIGs) as a 'golden standard' for evaluation.
• Employing knowledge representation and reasoning techniques for automated comparison of student proposals with CIG recommendations.
• Leveraging a medical ontology for identifying actions and explaining discrepancies.

Main Results:

• A functional system supporting the creation of CIG-based tests for medical students.
• Automated evaluation of student responses against established clinical guidelines.
• Explanation generation mechanism highlighting differences between student actions and CIG recommendations using a medical ontology.

Conclusions:

• The proposed facility offers a robust method for training and testing medical students' ability to act on clinical guidelines.
• AI and CIGs can be effectively utilized for educational purposes, moving beyond traditional decision support.
• The system facilitates objective assessment and provides valuable feedback to students, enhancing their clinical reasoning skills.