Mass Casualty Incident Training in Immersive Virtual Reality: Quasi-Experimental Evaluation of Multimethod Performance Indicators

Abstract

Background: Immersive virtual reality (iVR) has emerged as a training method to prepare medical first responders (MFRs) for mass casualty incidents (MCIs) and disasters in a resource-efficient, flexible, and safe manner. However, systematic evaluations and validations of potential performance indicators for virtual MCI training are still lacking.

Objective: This study aimed to investigate whether different performance indicators based on visual attention, triage performance, and information transmission can be effectively extended to MCI training in iVR by testing if they can discriminate between different levels of expertise. Furthermore, the study examined the extent to which such objective indicators correlate with subjective performance assessments.

Methods: A total of 76 participants (mean age 25.54, SD 6.01 y; 45/76, 59% male) with different medical expertise (MFRs: paramedics and emergency physicians; non-MFRs: medical students, in-hospital nurses, and other physicians) participated in 5 virtual MCI scenarios of varying complexity in a randomized order. Tasks involved assessing the situation, triaging virtual patients, and transmitting relevant information to a control center. Performance indicators included eye-tracking-based visual attention, triage accuracy, triage speed, information transmission efficiency, and self-assessment of performance. Expertise was determined based on the occupational group (39/76, 51% MFRs vs 37/76, 49% non-MFRs) and a knowledge test with patient vignettes.

Results: Triage accuracy (d=0.48), triage speed (d=0.42), and information transmission efficiency (d=1.13) differentiated significantly between MFRs and non-MFRs. In addition, higher triage accuracy was significantly associated with higher triage knowledge test scores (Spearman ρ=0.40). Visual attention was not significantly associated with expertise. Furthermore, subjective performance was not correlated with any other performance indicator.

Conclusions: iVR-based MCI scenarios proved to be a valuable tool for assessing the performance of MFRs. The results suggest that iVR could be integrated into current MCI training curricula to provide frequent, objective, and potentially (partly) automated performance assessments in a controlled environment. In particular, performance indicators, such as triage accuracy, triage speed, and information transmission efficiency, capture multiple aspects of performance and are recommended for integration. While the examined visual attention indicators did not function as valid performance indicators in this study, future research could further explore visual attention in MCI training and examine other indicators, such as holistic gaze patterns. Overall, the results underscore the importance of integrating objective indicators to enhance trainers' feedback and provide trainees with guidance on evaluating and reflecting on their own performance.

Keywords: disaster medicine; emergency medicine; emergency simulation; eye tracking; mass casualty incident; medical education; prehospital decision-making; virtual reality.

Figure 1

Immersive virtual reality (iVR) lab (left) and use of teleportation in iVR (right).

Figure 2

(A) density plots per group for triage accuracy, (B) triage speed, and (C) information transmission efficiency. The dashed lines mark the average per group. MFR: medical first responder.

Figure 3

Spearman correlation analysis of the performance indicators and immersive virtual reality (iVR) experience (above diagonal: ρ values, below diagonal: P values). iVR experience refers to prior iVR experience as assessed in the preliminary questionnaire; prior iVR experience was not winsorized because of the ordinal data structure. DOAF: duration of average fixation; FC: fixation count; VI: vehicle impact.