Abstract WP203: Code or Confounder? A Review of Stroke Team Activations and Mimic Diagnoses

Background: Between 2.2% and 17% of all strokes in the United States occur during hospitalization, often in non-acute care settings and among patients admitted for cardiovascular and/or surgical procedures. While emergency departments are supported by well-established stroke systems, the identification and management of in-hospital strokes outside these environments remain under-characterized. These settings often lack standardized training, protocols, or stroke infrastructure, contributing to delays in diagnosis and treatment. This represents a critical gap in stroke care delivery. Methods: We conducted a retrospective review of all acute stroke team activations at the University of Southern California Keck Hospital—a tertiary non-acute care facility without an emergency department—from October 2021 through December 2024. Data were collected on presenting symptoms, imaging, final diagnoses, and use of acute interventions. Primary outcomes included the frequency of confirmed ischemic and hemorrhagic strokes, transient ischemic attacks (TIAs), and use of reperfusion therapies such as intravenous alteplase (IV-tPA) or mechanical thrombectomy. Results: A total of 346 code strokes were activated. Of these, 20.8% were confirmed ischemic strokes, 7.2% hemorrhagic strokes, 4.6% TIAs, and 2.2% stroke recrudescence. The remaining 64.7% were adjudicated as stroke mimics. Reperfusion therapy was infrequent: 6 patients (1.7%) received IV thrombolysis, and 10 (2.9%) underwent thrombectomy. Among ischemic strokes, 37.5% were peri-procedural, related to recent cardiovascular surgeries. Mimics were most often due to encephalopathy or delirium (30.1%), with seizures (6.1%) and headaches (2.0%) less common. Altered mental status alone triggered 13 activations, none due to stroke, though all had neurological causes. Conclusion: Over this 4-year period, only 35.3% of code stroke activations were confirmed cerebrovascular events. More than one-third of ischemic strokes were peri-procedural, and most activations were for mimics—especially encephalopathy and delirium. These findings highlight the high mimic rate and low reperfusion therapy use in a non-acute care setting. We propose targeted training and tailored protocols to improve diagnostic accuracy, reduce unnecessary activations, and optimize care. Stroke nurse evaluations or expedited neurology consults may improve resource use.