Eye Tracking the Visual Attention of Nurses Interpreting Simulated Vital Signs Scenarios: Mining Metrics to Discriminate Between Performance Level

Nurses welcome innovative training and assessment methods to effectively interpret physiological vital signs. The objective is to determine if eye-tracking technology can be used to develop biometrics for automatically predict the performance of nurses whilst they interact with computer-based simulations. A total of 47 nurses were recruited, 36 nursing students (training group) and 11 coronary care nurses (qualified group). Each nurse interpreted five simulated vital signs scenarios whilst “thinking-aloud.” The participant's visual attention (eye-tracking metrics), verbalisation, heart rate, confidence level (1-10, 10 = most confident), and cognitive load (NASA-TLX) were recorded during performance. Scenario performances were scored out of ten. Analysis was used to find patterns between the eye-tracking metrics and performance score. Multiple linear regression was used to predict performance score using eye tracking metrics. The qualified group scored higher than the training group (6.85 ± 1.5 versus 4.59 ± 1.61, p =2= 0.80, p = <;0.0001). This shows that eye tracking alone could predict a nurse's performance and can provide insight to the performance of a nurse when interpreting bedside monitors.