Comparing the Accuracy of Mass Casualty Triage Systems in a Pediatric Population

Introduction: It was previously difficult to compare the accuracy of different mass casualty triage systems to one another. This pilot study is one of the first attempts to operationalize an expert panel's criterion standard definitions of triage categories in a pediatric population in order to compare accuracy between different systems. Objective: To compare the accuracy of 4 different mass casualty triage systems (SALT, JumpSTART, Triage Sieve, and CareFlight) when used for children. Methods: We observed the emergency department triage of patients less than 18 years old presenting to the only pediatric specialty hospital/Level 1 trauma center in Milwaukee County, Wisconsin. A single, certified EMS provider observed each patient's initial triage in the emergency department and recorded all findings that were necessary to categorize the patient using each of the 4 mass casualty triage systems being studied. Hospital medical records were then reviewed for each patient and assigned a criterion standard triage category based on the treatments received and final disposition. Descriptive statistics were used to compare accuracy, over-, and under-triage rates for each of the triage systems. Results: A total of 115 subjects were enrolled. Of those, 51% were male and 57% were transported by ambulance. When compared to the criterion standard definitions, SALT was found to have the highest accuracy rate (59%; 95% CI 50-68) compared to JumpSTART (57%; 95% CI 48-66), CareFlight (56%; 95% CI 47-65), and TriageSieve (56%; 95% CI 46-65). SALT also had the lowest under-triage rate (33%; 95% CI 24-42) compared to JumpSTART (39%; 95% CI 30-48), CareFlight (39%; 95% CI 30-48), and TriageSieve (39%; 95% CI 30-48). SALT had the highest over-triage rate (6%; 95% CI 2-11) compared to JumpSTART (4%; 95% CI 1-8), CareFlight (5%; 95% CI 1-9), and TriageSieve (5%; 95% CI 1-9). However, the confidence intervals for both the accuracy and under-triage rates overlapped between all triage systems. For each triage system, the most common error was designating a patient as "minimal" that, according to the criterion standard, should have been triaged as "delayed." Conclusion: We found that the 4 most popular mass casualty triage systems preformed similarly in an emergency department-based pediatric population. None of the systems were extremely accurate, and each demonstrated an unacceptable amount of under-triage. Better differentiating between patients categorized as "minimal" and "delayed" may improve the accuracy of mass casualty triage systems.