Differences between trained and naive participants on triage classification using gaze tracking
OA Version
Citation
Abstract
The Simple Triage And Rapid Treatment (START) program is a well-known triage-training program that helps first responders correctly sort victims into proper triage categories. START recognizes four categories: 1) Green/Minor: aka walking wounded (able to walk about with minor injuries that are not immediate or life threatening); 2) Yellow/Delayed: victim presents with injuries that needs medical attention after the red/immediate victims are treated; 3) Red/Immediate: severely injured, needs immediate help and gets priority; 4) Black/Expectant: the victim is either deceased or there is little that can be done to save them. Using the START triage program as a model, images of injured victims were used to simulate a typical field triage experience. While viewing the images, participant’s gaze patterns were recorded. Two subject groups were used: one group of subjects received triage training (experimental group) while the other group received unrelated, yet comparable, training (control group). The purpose of this experiment was to test two hypotheses. First, it is expected that participants in the experimental training group will demonstrate an increase in accuracy of triaging victim images as compared to the non-trained participants in the control group. Secondly, the
experimental group will demonstrate a change in fixation patterns in regards to salient features versus non-salient features, compared to those without training in the control group. Both experimental and control groups were given images to properly categorize into the triage categories. Experimental and control groups viewed the same images twice (baseline and second view/post-training). Each image ranged in severity, from minor to severe, and contained text boxes with vital sign information necessary to properly triage the victim images; presented within each image was a separate text box for each of the following: airway, breathing rate and circulation (presence of radial pulse). Data was collected using the BeGaze gaze eye tracker during the experiment. This recorded the participants’ eye movements and was used to analyze their gaze patterns during the visual stimulus. Data collected for resulting triage accuracy and gaze patterns were compared using Fischer Exact tests as well as 1-tailed t-tests. The Fischer Exact tests failed to reach significance to support the alternative hypothesis that participants with the triage training would increase in accuracy when triaging the victim images. The Fischer Exact test was negatively impacted due to a small sample size; however, descriptive statistics demonstrated a potential increase in accuracy by the trained group. Looking at fixation patterns, data shows a significant increase within the experimental group in regards to attention to the text boxes (salient features of the images) with respect to baseline entry times to the second viewing (post-training viewing) entry times.
Description
Thesis (M.A.)--Boston University