AI Virtual Standardized Patient Training for Perinatal Dental Communication: A Pilot Study

Confidence in dentist-patient communication is essential for building trust, supporting treatment adherence, and improving oral health outcomes. However, many dental residents feel underprepared for sensitive or complex conversations. Traditional communication training methods, such as standardized patients and role-play, are effective but resource-intensive and difficult to scale. AI-powered virtual standardized patients offer a flexible alternative by enabling simulated practice supported by automated, real-time feedback.

This pilot study explored the feasibility and preliminary educational impact of SOPHIE, an AI-powered virtual patient platform, in enhancing dental residents' confidence in clinical communication.

Ten dental residents from a perinatal dental clinic participated. Residents completed a 30-min orientation and a baseline self-efficacy in patient-physician communication (SE-12) survey, followed by engagement with five virtual patient scenarios providing AI-generated communication feedback. Residents completed the SE-12 survey and the System Usability Scale (SUS) after the practice. A semi-structured focus group explored user experience and perceived educational value. Quantitative data were analyzed descriptively, and qualitative data were analyzed thematically.

Residents reported increased confidence in managing challenging patient conversations and valued the opportunity for repeated, low-stakes practice. SUS scores indicated strong usability and positive learner experience. AI-generated metrics suggested improvements in conversational balance and use of open-ended questions. Qualitative findings reinforced these outcomes, highlighting the platform's support for empathetic, patient-centered communication.

SOPHIE was well accepted and demonstrated promising potential to strengthen communication confidence among dental residents, particularly in perinatal care. Future work should enhance conversational realism, emotional responsiveness, and nonverbal feedback.