Objective: To explore the accuracy of EMS neurological examination and documentation of stroke patients delivered to our non-academic metropolitan comprehensive stroke center. Background: With the recent advances in acute stroke therapy, emphasis on prehospital care is increasing. EMS identifying stroke and prenotification has been shown to enhance the care at the receiving hospital. Design/Methods: We used institutional Get-With-The-Guidelines to identify stroke patients brought by EMS. EMS run sheets and ED notes were reviewed and exam findings were extracted and correlated. We excluded patients with TIA, hemorrhagic stroke, and those who were transferred from an outside facility. Weighted Kappa statistic was calculated for individual neurological examination items, and was categorized as follows: poor if <0.40, moderate if 0.4–0.75, and excellent if >0.75. Results: Of 477 ischemic stroke patients brought by EMS, 372 had available run sheets for review. The frequency of reporting neurologic findings was as follows: level of alertness (n=266), focal weakness (n=190), awareness/orientation/attention (n=151), dysarthria (109), non-specific mental status change (105), facial weakness (n=103). Frequency of all other neurologic findings were < 50. All neurologic exam findings by EMS showed poor agreement except two, decreased level of alertness and focal weakness which showed moderate agreement (kappa 0.415 and 0.443 respectively). Conclusions: The documented exam findings in the EMS run sheets is inconsistent. Whether this is due to limitation in examination or documentation is unclear. Of the documented exam findings, only level of alertness and focal weakness correlated with the neurologic examination documented in the ED.