Accuracy of the field triage protocol in selecting severely injured patients after high energy trauma

Background

For optimal treatment of trauma patients it is of great importance to identify patients who are at risk for severe injuries. The Dutch field triage protocol for trauma patients, the LPA (National Protocol of Ambulance Services), is designed to get the right patient, in the right time, to the right hospital. Purpose of this study was to determine diagnostic accuracy and compliance of this triage protocol.

Study design

Triage criteria were categorised into physiological condition (P), mechanism of trauma (M) and injury type (I). A retrospective analysis of prospectively collected data of all high-energy trauma patients from 2008 to 2011 in the region Central Netherlands is performed. Diagnostic parameters (sensitivity, specificity, negative predictive value, positive predictive value) of the field triage protocol for selecting severely injured patients were calculated including rates of under- and overtriage. Undertriage was defined as the proportion of severely injured patients (Injury Severity Score (ISS) ≥ 16) who were transported to a level two or three trauma care centre. Overtriage was defined as the proportion of non-severely injured patients (ISS < 16) who were transported to a level one trauma care centre.

Results

Overall sensitivity and specificity of the field triage protocol was 89.1% (95% confidence interval (CI) 84.4–92.6) and 60.5% (95% CI 57.9–63.1), respectively. The overall rate of undertriage was 10.9% (95%CI 7.4–15.7) and the overall rate of overtriage was 39.5% (95%CI 36.9–42.1). These rates were 16.5% and 37.7%, respectively for patients with M+I−P−. Compliance to the triage protocol for patients with M+I−P− was 78.7%. Furthermore, compliance in patients with either a positive I+ or positive P+ was 91.2%.

Conclusion

The overall rate of undertriage (10.8%) was mainly influenced by a high rate of undertriage in the group of patients with only a positive mechanism criterion, therefore showing low diagnostic accuracy in selecting severely injured patients. As a consequence these patients with severe injury are undetected using the current triage protocol. As it has been shown that severely injured patients have better outcome in level one trauma care centres further optimisation of this protocol aiming at lowering undertriage is therefore essential, preferably without incrementing overtriage too much.