Objective: Improve prehospital identification of acute ischemic stroke patients with large vessel occlusion (LVO) by using a trauma system-based emergency communication center (ECC) to guide the emergency medical service (EMS). Methods: We trained 24 ECC paramedics in the Emergency Medical Stroke Assessment (EMSA). ECC-guided EMS in performance of the EMSA on patients with suspected stroke. During the second half of the study, we provided focused feedback to ECC after reviewing recorded ECC-EMS interactions. We compared the sensitivity, specificity, and area under the receiver operator characteristics curve (AUC) and 95% confidence interval of ECC-guided EMSA to the NIH Stroke Scale (NIHSS) for predicting a discharge diagnosis of LVO. Results: We enrolled 569 patients from September 2016 through February 2018. Of 463 patients analyzed, 236 (51%) had a discharge diagnosis of stroke and 227 (49%) had a nonstroke diagnosis. There were 45 (19%) stroke patients with LVO. For predicting LVO, there was no significant difference between the EMSA AUC = .68 (.59-.77) and the NIHSS AUC = .73 (.65-.81). An EMSA score greater than or equal to 4 had sensitivity = 75.6 (60.5-87.1) and specificity = 62.4 (57.6-67.1) for LVO. During the first 9 months of the study, the EMSA AUC = .61 (.44-.77) compared to an AUC = .74 (.64-.84) during the second 9 months. Conclusions:ECC-guided prehospital EMSA is feasible, has similar ability to predict LVO compared to the NIHSS, and has sustained performance over time.