Comparing Traditional and Augmented Reality Training for Identifying Critical Symptoms for Diagnosing Airway Obstruction Cases: A Pilot Study
TimeTuesday, April 134:10pm - 4:30pm EDT
LocationEducation and Simulation
DescriptionRecent developments in the portability, flexibility, and affordability of augmented reality (AR) technology make it a tool that can address the limitations of current medical training. As part of a multi-phase study, this pilot is focused on supporting simulation-based medical education through the use of AR technology. We aim to identify how this technology can enhance the detection and identification of subtle visual cues from the real world in the process of making accurate medical diagnoses. We randomly divided five participants into a control group, who received diagnostic training by reading a standard textbook, and an experimental group, who received diagnostic training by a combination of reading a standard textbook (same one as control group) and AR-based training. Participants were then presented with a tension pneumothorax scenario on the AR patient and were asked to list the symptoms they observed before making a final diagnosis. Investigators recorded both the total amount of time participants required to make a final diagnosis as well as the number of correctly identified symptoms, and the results of the control and experimental groups were statistically analyzed using a t-test to compare the groups. Although there was no significant difference between groups in the time needed to make a diagnosis, the participants in the experimental group took more time to diagnose the AR patient than the participants in the control group. This finding is perhaps due to the AR training providing a more accurate and extensive list of symptoms on which the experimental participants were primed to focus. This phenomenon is therefore potentially a case of the speed-accuracy tradeoff, originally proposed by Paul Morris (Zhai, S., Kong, J., & Ren, X., 2004). However, the AR-trained participants correctly identified a significantly greater number of symptoms (average: 12 symptoms) compared to the textbook-trained participants (average: 5 symptoms). These results indicate the potential of AR-based training to enhance detection and identification of cues in medical diagnostics.