The present study was designed as a prospective randomized controlled trial of triage test scores before and after a triage lecture for 30 min and triage exercises for 30 min for two study groups. One group received trauma card-based training and the other group received triage training based on direct instruction. Henceforth, they are referred to the trauma card group and the direct instruction group, respectively.
Sixteen firemen shifts located at four rescue stations in two Swedish municipalities were asked through their superiors to participate voluntarily in the study. Eight trauma card sessions and eight direct instruction sessions were set up to reach all 16 shifts. All shifts were randomized to either trauma card-based training or direct instruction using the randomization function in Microsoft Excel 2010 (Microsoft, Redmond, WA, USA v14.0.7153.5000). Before starting, all participants were informed in both verbal and written form about the purpose of the study and their ability to withdraw at any time without explanation. All participants verbally consented to participate in the study on those conditions. Demographic data such as age, educational background, years in service and multi-casualty experience were collected. All sessions took place during October, 2013.
A Microsoft PowerPoint 2010 (Microsoft, Redmond, WA, USA v14.0.7153.5000) lecture (18 slides, 30 min) on SALT triage based on the disaster medicine literature and published research papers [11, 14] was created in collaboration with national trauma instructors and disaster medicine experts. The aim, goals and learning objectives were defined including the history and purpose of triage, the origin of SALT triage and the SALT triage algorithm. Ten trauma victims with different injuries were chosen from the ETS patient bank. According to the SALT algorithm, the victims consisted of three red priority 1 victims with an Injury Severity Score (ISS) >15, three yellow priority 2 victims (ISS 8–12), three green priority 3 victims (ISS <3) and one black victim (dead on scene).
Both study groups attended the 30-min PowerPoint lecture. In the trauma card sessions, the participants were told in groups of 3–5 people, to triage 10 patients. Information was given to them in form of trauma cards (Fig. 1). First, all patients were triaged according to their visible injuries. When consensus was reached in the groups, the patients were triaged again according to their vital signs and the additional findings after the primary survey. When all groups were satisfied with their triage decisions, the presenter discussed the correct triage category for each patient.
In the direct instruction sessions, the same 10 patients were individually presented without the trauma cards by the presenter as a continuation of the lecture. In both groups, the starting condition was a bus crash with three ambulances on site and more on the way. To facilitate learning, a poster of the SALT triage algorithm was available for both groups during the patient scenarios (Fig. 2). The time for discussing patients was restricted to 30 min in both groups making the whole session 60 min. All lectures were held by the same presenter with a passive observer to ensure conformity. To further minimize bias, all participants were asked not to talk about the study with colleagues before all teaching sessions were completed.
To maximize the number of participants, and according to an agreement with the fire departments, a test encompassing 15, for the participants new, casualties used in earlier studies was used to measure triage performance [7, 15]. The same test was distributed to all participants before and after the training sessions. By not informing the participants of this arrangement, the incentive to memorize the pretest was minimized. The score was calculated as the number of correctly assigned triage categories. A time limit of 15 min to complete the test was set.
Six months after the initial training, a follow-up study was performed. All firemen who had participated in the first session were asked to take part. The follow-up did not involve any training. The same test with the 15 casualties was used, again under a time restriction of 15 min.
Exclusion criterion was a call out exceeding 45 min during any test or lecture. No participant was excluded due to this criterion. The data were stored on paper and unprocessed until data collection was complete. Microsoft Excel 2010 was then used to store the data in digital form. All participants and study groups were coded to ensure anonymity and were blinded to the analyser.
Ethics, consent and permissions
The Regional Ethical Review Board in Linköping was consulted regarding ethical approval. The study was not required to undergo a formal approval process. Informed consent was obtained from all study participants. No remuneration was given to any participant.
Calculations were blinded and made using IBM SPSS Statistics for Windows (Ver. 21.0 IBM Corporation, Armonk, NY). The Shapiro-Wilks test of normality was used. Because the tests before and after the training did not show normality, the Wilcoxon signed rank test was used to compare the scores at baseline and after the intervention as well as the post-test scores and the follow-up scores; the Mann–Whitney U test was used to investigate if previous experience of multi-casualty incidents, years in service, level of education or age or educational model was related to improvement or retention. A P value less than 0.05 was considered statistically significant. Due to the exploratory nature of the subgroup analyses, no correction for multiplicity was done. All data are presented as the median [interquartile range (IQR)].