DIALOGUE: A Generative AI-Based Pre-Post Simulation Study to Enhance Diagnostic Communication in Medical Students...

Area of Science:

• Medical Education
• Artificial Intelligence in Healthcare
• Diagnostic Communication

Background:

• Effective diagnostic communication is crucial for patient outcomes.
• Traditional medical training may not fully equip students with necessary communication skills.
• Generative artificial intelligence (GenAI) offers novel training modalities.

Purpose of the Study:

• To evaluate the efficacy of the DIALOGUE (DIagnostic AI Learning through Objective Guided User Experience) program in improving medical students' diagnostic communication skills.
• To assess the impact of GenAI-mediated training on students' ability to disclose a type 2 diabetes mellitus (T2DM) diagnosis.
• To identify factors influencing improvement in diagnostic communication.

Main Methods:

• A single-arm, pre-post study design involving 30 clinical-phase medical students.
• Pre- and post-test assessments using simulated patient encounters scored by blinded raters.
• Ten asynchronous GenAI training scenarios with automated feedback using the DIALOGUE program.
• Analysis of performance gains across eight communication domains and influencing factors.

Main Results:

• Significant improvement in overall diagnostic communication performance (mean increase of 36.7 points, p < 0.001).
• Proportion of high-performing students increased from 0% to 70%.
• Gains were observed across all domains, with notable improvements in encounter opening, closure, and diabetes-specific explanations.

Conclusions:

• GenAI-mediated training, exemplified by the DIALOGUE program, can substantially enhance medical students' diagnostic communication skills.
• The program shows potential as a scalable and individualized adjunct to traditional medical education.
• Learner characteristics such as baseline empathy and digital self-efficacy influenced training outcomes.