aamc.org does not support this web browser.

    Standardized Patient Script and Feedback Rubric for Discussing Borderline Personality Disorder with an AI-Simulated Patient

    The standardized patient script and feedback rubric for discussing borderline personality disorder (BPD) is a simulation tool designed for medical students to practice clinical interviewing using large language models (LLMs). It includes a patient script (with LLM instructions and parameters) and a feedback rubric based on the Handbook for Good Psychiatric Management for Borderline Personality Disorder and Tasman’s Psychiatry. Students copy and paste the script into an LLM to initiate the simulation. After completing the interaction, they paste the rubric into the LLM to receive standardized feedback. This resource supports structured training in communicating a BPD diagnosis and management. 

    Note: The authors used AI tools, specifically the free version of OpenAI’s ChatGPT including GPT-4o and GPT-3.5, to develop the AI instructions and parameters in the script and feedback rubric through practice simulation and through multiple conversational techniques. The same LLM was used to proofread and refine the script and rubric text to ensure grammatical accuracy, conciseness, and clarity for use by other LLMs. Finally, the LLM was used to proofread this submission text. These tools were used from June - August 2025. 

    Last updated October 2025 

    Authors

    Kamil Falkowski, Medical Student, Geisinger Commonwealth School of Medicine 

    Jeswin Chethalan, Medical Student, Geisinger Commonwealth School of Medicine

    Erich Miller, Medical Student, Geisinger Commonwealth School of Medicine

    John-michael Perez, MD, Attending Physician Psychiatry, Geisinger Behavioral Health

    Jessica Goldhirsh, MD, Attending Physician Psychiatry, Geisinger Behavioral Health