RESUMO
OBJECTIVES: Stroke-like symptoms may be difficult to appreciate due to the high incidence of stroke mimics (e.g., delirium) in the inpatient population. Many centers have adopted inpatient-specific stroke protocols with the aim of improving time to diagnosis and treatment. We aimed to assess one of these instruments, the "2CAN" score, in our patient population. MATERIALS AND METHODS: A retrospective chart review was conducted for all inpatients for whom our Brain Attack Team (BAT) was called between January 2015 and June 2019. Patients were excluded if they had stroke prior to current admission, were in the emergency department at the time of BAT call, or had incomplete documentation. The 2CAN score was calculated for each patient. RESULTS: The BAT was activated 201 times, and 110 patients met inclusion criteria. Twenty percent of patients had a history of atrial fibrillation, 72% hypertension, and 36% diabetes. Median NIHSS was 14.5 (IQR 5-24). Only 18% of stroke calls occurred within 24 h of hospital admission. The mean 2CAN score was 2.8. Ninety-seven (88%) patients received a final diagnosis of ischemic stroke and 13 (12%) of stroke mimics. There was no difference between 2CAN scores in the stroke and mimic groups (P = 0.91). A 2CAN score of ≥ 2 had sensitivity 83.5%, specificity 23.1%, PPV 89.0%, and NPV 15.8% for stroke. CONCLUSIONS: The 2CAN score was derived and validated in a single academic center as a tool to recognize inpatient stroke. The 2CAN score had good sensitivity and positive predictive value for stroke in our cohort, but poor specificity.