Abstract 10: Regional Variation in Out-Of-Hospital Cardiac Arrest Survival in the United States

Abstract

Background: Although previous studies have shown marked variation in out-of-hospital cardiac arrest survival across U.S. regions, factors underlying this variation in survival remain unknown. Methods & Results: Using 2005-2013 data from the Cardiac Arrest Registry to Enhance Survival (CARES), we identified 88,305 adult patients (age >18 years) in 107 U.S. counties with out-of-hospital cardiac arrest at home or in a public location, and geo-coded them to a U.S. county using the address where cardiac arrest occurred We constructed a two-level hierarchical regression model (patient & county) and used median odds ratios (MOR) to quantify regional variation in out-of-hospital cardiac arrest survival. Moreover, we examined the proportion of variation in survival that was explained by 1) patient demographics 2) cardiac arrest characteristics 3) county-level rates of bystander cardiopulmonary resuscitation (CPR) and hypothermia treatment and 4) county-level socio-demographic factors. The mean rate of survival to discharge was 10.0%, and varied markedly across counties (range: 1.4%-18.4%, MOR: 1.33; 95% CI: 1.24-1.38, Figure 1). Compared to counties in the lowest quartile of survival, patients in the highest quartile counties were younger (62.5 vs 61.6 years), more likely to be men (60.8% vs 64.4%), have a shockable rhythm (21.1% vs 26.9%), witnessed arrest (50.3% vs 53.0%), receive bystander CPR (23.4% vs 32.6%), and hypothermia (44.4% vs 62.3%, P for trend < 0.01 for all). County-level rates of survival were positively correlated with rates of bystander CPR (ρ = 0.45, P < 0.0001) and hypothermia treatment (ρ = 0.24, P < 0.0001). Sequential adjustment of demographic and cardiac arrest characteristics explained only 4.3% and 12.4% of the county-level variation in survival, respectively. Inclusion of county-level rates of bystander CPR and hypothermia explained a total of 28.5% of the survival variation, and this proportion increased to 36% after adjustment of other county-level factors. Conclusion: There is substantial variation in out-of-hospital cardiac arrest survival across U.S. counties. Although a large proportion of survival variation was unexplained, most of the variation that could be accounted for was due to county-level differences in rates of bystander CPR and hypothermia treatment.