Introduction:Triage at mass gatherings in Australia is commonly performed by staff members with first aid training. There have been no evaluations of the performance of first aid staff with respect to diagnostic accuracy or identification of presentations requiring ambulance transport to hospital.Hypothesis:It was hypothesized that triage decisions by first aid staff would be considered correct in at least 61% of presentations.Methods:A retrospective audit of 1,048 presentations to a single supplier of event health care services in Australia was conducted. The presentations were assessed based on the first measured set of physiological parameters, and the primary triage decision was classified as “expected” if the primary and secondary triage classifications were the same or “not expected” if they differed. The performance of the two triage systems was compared using area under the receiver operating characteristic curve (AUROC) analysis.Results:The expected decision was made by first aid staff in 674 (71%) of presentations. Under-triage occurred in 131 (14%) presentations and over-triage in 142 (15%) presentations. The primary triage strategy had an AUROC of 0.7644, while the secondary triage strategy had an AUROC of 0.6280, which was significantly different (P = .0199).Conclusion:The results support the continued use of first aid trained staff members in triage roles at Australian mass gatherings. Triage tools should be simple, and the addition of physiological variables to improve the sensitivity of triage tools is not recommended because such an approach does not improve the discriminatory capacity of the tools.

Objective: Using physiological markers to detect patients at risk of deterioration is common. Deaths at music festivals in Australia prompted scrutiny of tools to identify critically unwell patients for transport to hospital. This study evaluated initial physiological parameters to identify patients selected for transport to hospital from a music festival. Methods: A retrospective audit of 2045 presentations at music festivals in Victoria, Australia, was performed. Presentation heart rate, systolic blood pressure, respiratory rate, oxygen saturation, temperature, and Glasgow Coma Scale were assessed using area under the receiver operating characteristic curve (AUROC) analysis, with a prespecified threshold of 0.7. Results: The only measured variable to exceed the prespecified cutpoint was initial systolic blood pressure, with an AUROC of 0.72 and optimal cutpoint of 122 mmHg. Using commonly accepted cutpoints for variables did not improve detection performance to acceptable levels, nor did using combination systems of cutpoints. Conclusions: Initial physiological variables are poor predictors of the decision to transport to hospital from music festivals. Systolic blood pressure was significant, but only at a clinically insignificant value. Decisions on which patients to transport from an event site should incorporate more information than initial physiology. Senior clinicians should lead decision-making about hospital transport from music festivals.