This research explores how Large Language Models (LLMs) can create realistic simulations of challenging patient interactions to enhance medical communication training.

• Creates virtual patients with specific communication styles like "accuser" and "rationalizer" personas
• Enables multilingual training scenarios to accommodate diverse cultural contexts
• Provides medical students with exposure to difficult interpersonal dynamics rarely encountered in traditional training
• Offers scalable, accessible practice opportunities for developing crucial communication skills

This innovation addresses a critical gap in medical education by allowing healthcare professionals to practice managing difficult conversations in a safe environment before encountering similar situations with real patients.

Modeling Challenging Patient Interactions: LLMs for Medical Communication Training