Introduction: No existing mass casualty triage system has been scientifically scrutinized or validated. A recent work group sponsored by the Centers for Disease Control and Prevention, using a combination of expert opinion and the extremely limited research data available, created the SALT (sort-assess-lifesaving interventions-treat/transport) triage system to serve as a national model. An airport crash drill was used to pilot test the SALT system.
Objective: To assess the accuracy and speed with which trained paramedics can triage victims using this new system.
Methods: Investigators created 50 patient scenarios with a wide range of injuries and severities, and two additional uninjured victims were added at the time of the drill. Students wearing moulage and coached on how to portray their injuries served as "victims." Assuming proper application of the SALT system, the patient scenarios were designed such that 16 patients would be triaged as T1/red/immediate, 12 as T2/yellow/delayed, 14 as T3/green/minimal, and 10 as T4/black/dead. Paramedics were trained to proficiency in the SALT system one week prior to the drill using a 90-minute didactic/practical session, and were given "flash cards" showing the triage algorithm to be used if needed during the drill. Observers blinded to the study purpose timed and recorded the triage process for each patient during the drill. Simple descriptive statistics were used to analyze the data.
Results: The two paramedics assigned to the role of triage officers applied the SALT algorithm correctly to 41 of the 52 patients (78.8% accuracy). Seven patients intended to be T2 were triaged as T1, and two patients intended to be T3 were triaged as T2, for an overtriage rate of 13.5%. Two patients intended to be T2 were triaged as T3, for an undertriage rate of 3.8%. Triage times were recorded by the observers for 42 of the 52 patients, with a mean of 15 seconds per patient (range 5-57 seconds).
Conclusions: The SALT mass casualty triage system can be applied quickly in the field and appears to be safe, as measured by a low undertriage rate. There was, however, significant overtriage. Further refinement is needed, and effect on patient outcomes needs to be evaluated.