Variation in Prehospital Routing Patterns for Patients With Suspected Stroke

Abstract

BACKGROUND: Our objective was to describe emergency medical services (EMS) transport destination practices for patients with suspected stroke using a novel linked data set of EMS patient care data and verified emergency department (ED) stroke-related capabilities. METHODS: This was a cross-sectional evaluation of EMS 9-1-1 encounters for patients with suspected stroke from 2019 to 2021 from the ESO Data Collaborative, linked to verified ED capabilities from the 2021 National ED Inventory-USA survey. Our primary outcome was transport to an ED with any stroke care capabilities. We used mixed effects logistic regression models with a random intercept for agency to estimate the odds of transport to an ED with stroke care capabilities. We used intraclass correlation coefficients and median odds ratios (ORs) to quantify the contribution of covariates and variability between EMS agencies. RESULTS: The analytic sample included 314642 encounters (median age 72years, 53% female) from 1781 EMS agencies, including 719 (40%) agencies primarily serving rural areas. Overall, 32142 (10%) patients were transported to facilities lacking verified stroke care capabilities. There was significant between-agency variation, with median ORs consistently >2.0 (eg, fully adjusted model, median OR, 3.91 [95% CI, 3.71–4.11]). Rural, versus urban, location was associated with lower odds of transport to an ED with stroke care capabilities (OR, 0.15 [95% CI, 0.14–0.17]). CONCLUSIONS: In this analysis of EMS encounters, 1 in 10 patients with suspected stroke was transported to a facility without verified stroke care capabilities. Variation across EMS agencies and lower odds of appropriate routing in rural settings highlight critical gaps in prehospital stroke systems of care.