Evolution of stroke diagnosis in the emergency room--a prospective observational study

Background: Access to acute neurological care is limited. Especially in nonurban areas, and owing to uncertainties in diagnosing stroke, non-neurologists often misinterpret stroke symptoms. We evaluated the profile of patients with suspected stroke and the accuracy of the admission diagnosis 'stroke' in the setting of a specialized neurological emergency department in a nonurban region.

Methods: In this prospective observational study, (1) data from all 4,174 patients with the discharge diagnosis 'stroke' and (2) data from 1,800 consecutive patients (3 cohorts per year over 3 years) with the admission diagnosis 'stroke' were included over a 3-year period.

Results: The positive predictive value of the admission diagnosis 'stroke' was 0.34; the negative predictive value was 0.97. The rate of misdiagnosis significantly correlated with age and time from symptom onset to presentation. During the study period, the proportion of patients with the admission diagnosis 'stroke' admitted early after symptom onset increased from 19.9 to 27.8% within 3 h and from 26.4 to 32.7% within 4.5 h, respectively. Thrombolysis rates increased (from 9.4 to 15.4%).

Conclusion: The uncertainties in interpreting stroke symptoms and the lack of facilities for treating emergency stroke in nonurban areas may be outweighed by offering access to a specialized neurological emergency room, thus rectifying any misinterpretation of stroke symptoms and shortening in-hospital time windows for treatment. Still, the rate of misdiagnosis is high, requiring expensive resources, despite the constant flow of information to the public. Therefore, more prospective data comparing different emergency room settings are needed which focus in particular on patients with the admission diagnosis 'stroke'.