Evaluation of rosier and ABCD2 scores, Whiteley Index, and Bradford Somatic Symptom Inventory results in patients presenting to the emergency department with suspected transient ischemic attack

Objective: This study aimed to evaluate the relationship between clinical outcomes and various scoring systems applied to patients presenting to the emergency department with symptoms that can be seen in stroke and Transient Ischemic Attack (TIA), in addition to diagnostic evaluations. The goal was to use the data obtained to help differentiate between patients with TIA or stroke and those with stroke mimics. Method: This prospective study was conducted at the Emergency Medicine Department of Düzce University between 01/03/2024 and 01/02/2025. Patients who presented within the last 24 hours with syncope, convulsion, or focal neurological symptoms were evaluated diagnostically. In addition, ABCD2, ROSIER, WI-7, and BBBE-44 scores were calculated, and the relationships between these scores and clinical outcomes were statistically analyzed. SPSS 25 was used for data analysis. Patients with head trauma within the last 24 hours, a diagnosis of cancer or epilepsy, or severe aphasia were excluded from the study. Results: Of the 115 patients followed with suspected stroke, 28.7% (n=33) were diagnosed with a neurovascular disease, while 71.3% (n=82) were evaluated as stroke mimics. The most common symptom among the presentations was dizziness (33.9%). Among patients diagnosed with neurovascular disease, the proportion experiencing dizziness (15.2%) was significantly lower than in those not diagnosed with neurovascular disease (41.5%) (p=0.007). When comparing patients diagnosed with neurovascular disease to stroke mimics, the presence of aphasia (p<0.001), paresthesia in the upper extremity (p=0.009), plegia in the lower extremity (p=0.001), ABCD2 score (p<0.001), and ROSIER score (p<0.001) were found to be significantly higher. Although the proportion of patients with a moderate-to-high WI-7 score was lower in the diagnosed group (30.3%) compared to the undiagnosed group (34.1%), this difference was not statistically significant (p=0.692). Logistic regression analysis of patient symptoms and scoring systems in terms of stroke diagnosis showed that the model had a sensitivity of 66.7%, specificity of 95.1%, and an overall accuracy of 87%. Conclusion: A sensitivity of 66.7% indicates that the model correctly identifies about two-thirds of patients diagnosed with neurovascular disease, while a specificity of 95.1% shows that it correctly excludes the vast majority of those without neurovascular disease. The 87% accuracy supports that the model demonstrates overall strong performance. Our regression analysis results suggest that combining scoring systems that screen for stroke mimics with those that predict stroke may provide more reliable diagnostic performance. The model's particularly high specificity and low false-positive rate show promise for distinguishing stroke mimics in clinical practice. Keywords: stroke, ABCD2, ROSIER, Bradford, health anxiety, Whiteley index