Abstract 249: The Effect of Response Time on Out-Of-Hospital Cardiac Arrest Survival Varies by Patient Subpopulation

Itroduction: Previous research has quantified the impact of EMS response time on the probability of survival from OHCA, but the impact on different subpopulations is currently unknown. Aim: To investigate how response time affects OHCA survival for different patient subpopulations. Methods: We conducted a logistic regression analysis on non-EMS witnessed OHCAs of presumed cardiac etiology from the Toronto Regional RescuNet between January 1, 2007 and December 31, 2016. We predicted survival using age, sex, public location, presenting rhythm, bystander witnessed, bystander resuscitation, and response time, defined as the time interval from 911 call to EMS arrival at the patient. We conducted subgroup analyses to quantify the effect of response time on survival for eight different subpopulations: public, private, bystander resuscitation, no bystander resuscitation, patients ≥65, patients <65, witnessed, and unwitnessed OHCA. We also quantified the effect of response time on survival for pairwise intersections of the subpopulations. We compared our results to Valenzuela et al. (1997), which suggests survival odds decrease by 10% for each minute delay in response time. Results: We identified 22,988 OHCAs. Overall, a one-minute delay in EMS response time was associated with a 13.2% reduction in the odds of survival. The reduction varied by subpopulation, ranging from a 7.2% reduction in survival odds for unwitnessed arrests to a 16.4% reduction in survival odds for arrests with bystander resuscitation. Response time had the largest impact on survival for the subpopulation of OHCAs that were both witnessed and received bystander resuscitation (17.4% reduction in survival odds). Conclusion: The effect of a one-minute delay in EMS response on the odds of survival from OHCA can be as low as a 7.2% reduction and as high as a 17.4% reduction. This variability contrasts with the currently accepted 10% rule that is assumed across the entire population.