Abstract 14956: Out-of-hospital Cardiac Arrest and Bystander CPR in Kent County, Michigan: A Geospatial Analysis

Abstract

Introduction: Geographic clustering of OHCA & bystander CPR (bCPR) has been demonstrated in some US cities. We assessed for geospatial distribution patterns within our county. Objectives: 1) Characterize bCPR performance; 2) Test for geographic clustering of OHCA & bCPR; 3) Identify hot spots & cold spots, defined as areas with disproportionately greater and lesser chance of bystander intervention within a concentration of OHCA. Hypothesis: OHCA & bCPR is geographically clustered within our region, resulting in hot and cold spots of bystander intervention. Methods: Retrospective, observational, population-based, cohort study using the CARES dataset for Kent County, Michigan. We included adult, non-traumatic, EMS-treated OHCA from 2010-14, excluding cases with on-site medical care. We abstracted EMS dispatch addresses, bCPR, & patient characteristics, then geocoded street addresses (ArcGIS) to create a population-adjusted density map of OHCA annual incidence. Moran’s I test assessed for spatial autocorrelation of population-weighted cardiac arrest rate by census block. Difference mapping identified relative differences between cases with/without bCPR, & Getis-Ord Gi statistic assessed for spatial clustering of bCPR, identifying hot and cold spots. Results: Of 1,632 eligible subjects, 1,167 met inclusion criteria. bCPR occurred in 436 (37%), & was performed by family members (58%), lay persons (34%), & off-duty medical providers (8%). Specific techniques included compressions/ventilations (28%), compression only (35%), ventilation only (1%), & unknown (36%). Geospatial analysis revealed significant clustering of OHCA ranging from 0 to >60 annual cases per 100,000 population (p<0.001). Relative difference in bCPR ranged from -11.2 to 0.9 across geographic locales, with a paucity of bCPR (-11.2 to -2.0) in urban and suburban areas. We identified central Grand Rapids & southern suburbs as primary cold spots (99% confidence), whereas eastern & northeastern suburbs contained hot spots (99% confidence). Conclusion: The minority of OHCA received bCPR. OHCA & bCPR are geographically clustered, & bCPR is more/less likely to occur in certain areas. Distinguishing hot & cold spots affords opportunity for targeted public health initiatives.