|| Checking for direct PDF access through Ovid
Background: Stroke remains among the leading causes of death and disability worldwide. The recent advances in acute stroke therapy further validate the need for an efficient ‘stroke system,’ of which the first step is rapid and accurate identification of stroke. Currently, there is no agreed upon standard prehospital stroke scale.Methods: In conjunction with the Cochrane Stroke Group we developed an electronic search strategy to identify relevant studies in MEDLINE. We then adapted it to seven other databases. Scales had to have been used in the emergency or prehospital setting, have a discharge diagnosis of stroke by a neurologist and needed to demonstrate at a minimum, the data needed to construct a two by two table. Our database search yielded 8479 references with 16 full texts articles and 4 abstracts (7 scales) meeting our inclusion criteria. We applied the Quadas-2 tool to eleiminate bias.Results: The ROSIER scale demonstrated the highest median sensitivity at 90% (95%CI, range 85%-97%) as well as the smallest confidence intervals. The LAPSS demonstrated the best pooled specificity, 91% (95%CI, range 84%-95%).Conclusion: In the acute stroke setting a highly sensitive scale for stroke identification is of great value. This Cochrane Systematic review demonstrated that of all currently validated scales, the ROSIER scale has the greatest sensitivity for detecting stroke in the the prehospital or emergency setting.