A Unified National Registry of EMS Agencies with Estimated Service Areas: EMS Capability, Insurance Coverage, and Community Health Outcomes

Abstract

Emergency medical services (EMS) in the United States are operated by fire departments, hospitals, government bodies, and private agencies, yet national rosters rely on self-reporting that underrepresents rural providers. We therefore curated the first unified national EMS registry with estimated service areas. Cross-referencing NEMSIS, NPPES, and state databases with AI-assisted extraction (Gemini, Google Places API), we compiled licensed agencies' locations and operational information, validated at F1 = 0.96 for operational information and 0.97 for addresses. Because agencies rarely have authoritative boundaries, each was assigned a polygon by comparison against fire service areas, county boundaries, ZIP/ZCTA areas, and Voronoi partitions. Using eight states publishing official EMS zones (AZ, KY, MN, MO, ND, NE, OK, UT) as ground truth, 1,607 matched agencies scored intersection-over-union 0.43, area ratio 0.98, and 6 km centroid offset; accuracy was not predictable from geography, so each state's regime requires direct testing. Service areas were then linked to Census socioeconomic data, CDC PLACES health outcomes, and RUCA classifications. Regression analysis shows EMS capability significantly moderates the relationship between insurance coverage and mortality: in BLS-served communities, each point of county uninsurance is associated with +2.60 injury deaths per 100,000 (p<0.001), reduced by −1.69 under ALS coverage (p< 0.001), while for heart-disease deaths the uninsurance gradient without a transporting agency (+7.77 per 100,000, p<0.001) is flattened by −2.35 where one exists (p = 0.037). These findings demonstrate the necessity of nationally curated EMS service-area data for understanding how EMS capability and insurance coverage jointly shape health outcomes.