Developing Virtual Standardized Patients for Health Sciences Education using Natural Language Dialogue Systems

This presentation will summarize our work on creating conversational virtual standardized patients for medical education. Our Virtual Standardized Patients are avatar representations of human patients which are controlled by artificial intelligence software. Students can converse using natural language, practice their history taking skills and receive immediate feedback on their interaction. Students can interview the Virtual Patients via a web site, on their iPads, or using a Virtual Reality headset in which the students are virtually in the room with their patients to enhance the authenticity and realism of the experience. We have Virtual Standardized Patients with a variety of chief complaints including back pain, abdominal pain, headache, dizziness, shortness of breath, dysuria, and sadness. We are currently working on a patient with Opioid Use Disorder, a patient with Limited English Proficiency, and a Virtual Cancer Patient for health care providers to practice difficult conversations.

Douglas Danforth is a Professor of Obstetrics and Gynecology at The Ohio State University College of Medicine. He has been involved in Virtual Patient and Virtual Reality research for over 15 years. His lab is funded by the Department of Health and Human Services Health Resources and Services Administration, The Ohio Department of Medicaid, The Institute for Innovative Technologies in Medial Education, The Agency for Health Care Research and Quality, and The National Science Foundation.