Agent based simulation for training and assessing students in the field of anesthesiology

Abstract

Preoperative evaluation is a critical skill for anesthesiologists. Training and assessing the performance of residents with standardized patients is expensive, time consuming and resource intensive. In certain cases, virtual humans may provide more fidelity than standardized patients. This technology offers a scalable, portable solution to such a problem. We created a virtual human preoperative patient interview simulator (Avatar) as a joint project between the Icahn School of Medicine at Mount Sinai (New York, NY) and LogicJunction, Inc. (Cleveland, Ohio). Users ask free-text questions as well as perform physical examination and order laboratory studies and receive feedback on their performance. We randomized a cohort of first year anesthesiology residents to perform a preoperative assessment on the Avatar or a standardized patient. While average interview time was increased with participants interviewing the Avatar, total number of questions and performance on objective feedback criteria was similar between the two groups.