Abstract
BACKGROUND: Trauma is time sensitive, and minimizing prehospital (PH) time is appealing. However, most studies have not linked increasing PH time with worse outcomes because raw PH times are highly variable. It is unclear whether specific PH time patterns affect outcomes. Our objective was to evaluate the association of PH time interval distribution with mortality. METHODS: Patients transported by emergency medical services in the Pennsylvania trauma registry from 2000 to 2013 with a total PH time (TPT) of 20 minutes or longer were included. TPT was divided into three PH time intervals: response, scene, and transport time. The number of minutes in each PH time interval was divided by TPT to determine the relative proportion each interval contributed to TPT. A prolonged interval was defined as any one PH interval contributing equal to or greater than 50% of TPT. Patients were classified by prolonged PH interval or no prolonged PH interval (all intervals < 50% of TPT). Patients were matched for TPT, and conditional logistic regression determined the association of mortality with PH time pattern, controlling for confounders. PH interventions were explored as potential mediators, and PH triage criteria used identify patients with time-sensitive injuries. RESULTS: There were 164,471 patients included. Patients with prolonged scene time had increased odds of mortality (odds ratio, 1.21; 95% confidence interval, 1.02-1.44; p = 0.03). Prolonged response, transport, and no prolonged interval were not associated with mortality. When adjusting for mediators including extrication and PH intubation, prolonged scene time was no longer associated with mortality (odds ratio, 1.06; 95% confidence interval, 0.90-1.25; p = 0.50). Together, these factors mediated 61% of the effect between prolonged scene time and mortality. Mortality remained associated with prolonged scene time in patients with hypotension, penetrating injury, and flail chest. CONCLUSION: Prolonged scene time is associated with increased mortality. PH interventions partially mediate this association. Further study should evaluate whether these interventions drive increased mortality because they prolong scene time or by another mechanism, as reducing scene time may be a target for intervention. LEVEL OF EVIDENCE: Prognostic/epidemiologic study, level III.