Abstract

Objectives

This article builds on the literature regarding the association between emergency medical service (EMS) response times and patient outcomes (death and severe injury). Three issues are addressed in this article with respect to the empirical estimation of this relationship: the endogeneity of response time (systematically quicker response for higher degrees of urgency), the nonlinearity of this relationship, and the variation between such estimations for different patient outcomes.

Methods

Binomial and multinomial logistic regression models are used to estimate the impact of response time on the probabilities of death and severe injury using data from French Fire and Rescue Services. These models are developed with response time as an explanatory variable and then with road time (dispatch to arrival) hypothesized as representing the exogenous variation within response time. Both models are also applied to data subsets based on response time intervals.

Results

The results show that road time yields a higher estimate for the impact of response time on patient outcomes than (total) response time. The impact of road time on patient outcomes is also shown to be nonlinear. These results are of both statistical significance (model coefficients are significant at the 95% confidence level) and economical significance (when taking into account the number of annual interventions performed).

Conclusions

When using heterogeneous data on EMS interventions where endogeneity is a clear issue, road time is a more reliable indicator to estimate the impact of EMS response time on patient outcomes than (total) response time.