Using Geographic Information Systems to Identify High-Risk Neighborhoods for Out of Hospital Cardiac Arrest (OHCA) in San Diego

Objective: To compare the results of three spatial statistical methods for the detection of high-risk census tracts, which are those areas that have a high incidence of out-of-hospital cardiac arrest (OHCA) and low prevalence of bystander cardiopulmonary resuscitation (CPR), in order to identify possible sites for a targeted community-based intervention.Methods: Secondary analysis of prospectively collected EMS data from the City of San Diego, California (1,251,184 million people). Consecutive adult (≥18 years) OHCAs restricted to those of cardiac etiology and treated by EMS from Jan.1st 2003 to Dec. 31st 2009. Locations of OHCA were geocoded with BatchGeo and ArcGIS. Census tracts with a high incidence of OHCA and low prevalence of bystander CPR were considered high-risk census tracts. We used the following three cluster analysis methods to identify high-risk census tracts: Local Moran's I, Getis-Ord Gi* and Empirical Bayes reliability adjusted rates. The overlapping census tracts between these three methods were then identified. Results: A total of 2304 arrests in 276 census tracts occurred during the study period, with 2,301 arrests included in the final sample. The majority of OHCA patients were Caucasian (n=1389, 60.3%), male (n=1490, 64.7%), had a presenting rhythm of asystole (n=1311, 64.7%), with the OHCA events occurring at home (n=1834, 79.7%). Local Moran's I identified 11 census tracts, Getis-Ord Gi* identified 9 high-risk census tracts, and the Empirical Bayes method identified 3 high-risk census tracts. Two census tracts were identified by all three methods, while an additional 3 census tracts were identified by at least two out of the three methods. Conclusion: This is the first study to apply three different methods for the detection of high-risk areas using OHCA data. The 2 census tracts, identified in all three analyses, appear to be possible sites for targeted community-based interventions to improve CPR training and cardiovascular disease education.