Validation of the Telestroke Mimic Score in Mayo Clinic population

Objectives: Telestroke consultations enable hospital providers to administer intravenous (IV) alteplase to patients who would otherwise not receive it due to lack of an in-hospital stroke team. However, up to 30% of acute stroke patient evaluations are deemed to be stroke mimics. Mimics present a challenge with the limitations of a virtual neurological exam. The administration of IV alteplase in these patients is not without risk. With the cost and risk associated with IV alteplase, there are both ethical and practical incentives to avoid administering alteplase to a patient manifesting a stroke-mimic. Recently a retrospective analysis validated a TeleStroke Mimic Score (TM-Score) to help detect stroke mimics. We retrospectively applied this tool to Mayo Clinic Stroke Telemedicine for Arizona Rural Residents (STARR) telestroke database to provide external validation in an independent study population. Materials and Methods: We analyzed 339 patients in the STARR database for validation of the TM-Score, which was applied retrospectively to determine whether it predicted stroke-mimic, using data available during each patient's telestroke consult. We assessed the TM-Score's performance with a receiver-operating characteristic (ROC) curve. A scatter plot of the data was assembled to demonstrate the relationship between the TM-Score and the likelihood of having a stroke mimic, and was compared to the nomogram in the original TM-Score study. Results: When the TM-Score was applied to Mayo Clinic STARR validation cohort, the area under the ROC curve was 0.78, larger than that of the derivation cohort in the original study (0.75). Further analysis suggested that a TM-Score > 25 or < 10 provided a greater degree of confidence that the patient had presented with stroke or stroke mimic, respectively. In STARR cohort, additional variables were significantly different between stroke and stroke mimic populations, including a history of sleep apnea and diabetes. Conclusions: We determined that the original TM-Score was valid when applied to Mayo Clinic STARR telestroke population.