Assessing Pediatric Life Support Skills Using Augmented Reality Medical Simulation With Eye Tracking: A Pilot Study

Abstract

Background: Augmented reality (AR) and eye tracking are promising adjuncts for medical simulation, but they have remained distinct tools. The recently developed Chariot Augmented Reality Medical (CHARM) Simulator combines AR medical simulation with eye tracking. We present a novel approach to applying eye tracking within an AR simulation to assess anesthesiologists during an AR pediatric life support simulation. The primary aim was to explore clinician performance in the simulation. Secondary outcomes explored eye tracking as a measure of shockable rhythm recognition and participant satisfaction.

Methods: Anesthesiology residents, pediatric anesthesiology fellows, and attending pediatric anesthesiologists were recruited. Using CHARM, they participated in a pediatric crisis simulation. Performance was scored using the Anesthesia-centric Pediatric Advanced Life Support (A-PALS) scoring instrument, and eye tracking data were analyzed. The Simulation Design Scale measured participant satisfaction.

Results: Nine each of residents, fellows, and attendings participated for a total of 27. We were able to successfully progress participants through the AR simulation as demonstrated by typical A-PALS performance scores. We observed no differences in performance across training levels. Eye tracking data successfully allowed comparisons of time to rhythm recognition across training levels, revealing no differences. Finally, simulation satisfaction was high across all participants.

Conclusions: While the agreement between A-PALS score and gaze patterns is promising, further research is needed to fully demonstrate the use of AR eye tracking for medical training and assessment. Physicians of multiple training levels were satisfied with the technology.

Keywords: Augmented reality; eye tracking; medical simulation; pediatric advanced life support.

Figure 1.

CHARM simulator holographic view, shown from the vantage point of the simulation controller and narrator.

Figure 2.

Mean and standard deviations of A-PALS scores. The overall A-PALS score, as well as the individual subscores, are included. Mean scores are grouped by training level.

Figure 3.

Box plots of SDS scores. The overall SDS score, as well as the section scores, are included. Median scores are grouped by training level.