Abstract WP80: Comparison of Dallas ED Screening Tool to Identify Stroke (DESTINY) With Other Stroke Screening Tools

Abstract

Background: Screening is a valuable tool for triaging, protocol activation, and resource allocation in emergency medicine, yet no tool has been identified as superior in the prehospital and hospital setting. DESTINY was developed as a single tool to screen for all stroke subtypes. Methods: This study is a Phase II prospective two part adaptive design validation study of DESTINY in ED during the early triage of suspected stroke patients from 05/01/2022 to 08/08/2022. CSTAT (questions: 2, 4, 5, 10) and VAN (questions: 3, 4, 5, 6, 7, 10) were embedded in the tool. DESTINY was incorporated into EHR and standardized workflow for nursing triage in ED. The NIHSS at presentation (median 2), primary diagnosis, and stroke subtypes were recorded. During the blinded phase, only CSTAT score was provided for clinical decisions. Primary outcomes included predictability of acute symptomatic: stroke, PCI, LVO, and MeVO. Secondary outcomes included comparison of DESTINY performance to other screening tools. Score completion failures were retrospectively scored by vascular neurologist provided chief complaint and initial exam, blinded to final diagnosis and imaging. Results: DESTINY was used to screen 284 consecutive patients; 32% had score completion failures. Acute symptomatic stroke occurred in 39% (n=110) (sTIA, 20; PCI, 31; CRAO, 5; MeVO, 22; LVO, 19; hemorrhagic, 11, other, 13). DESTINY > 1 had better sensitivity to other tools to predict acute symptomatic stroke (sensitivity 0.95 [95% CI 90.3-98.8], specificity 0.44 [95% CI 36.9-51.6]); MeVO (sensitivity 0.91 [95% CI 78.9-1.0], specificity 0.31 [95% CI 25.3-36.5]) and comparable to NIHSS to predict PCI. DESTINY score > 3 or > 1 and at least one positive from questions 3 to 7 (sensitivity 1.0, specificity 0.63 [95% CI 56.8-68.5]) was more sensitive to other scales at typically defined thresholds. Conclusion: In high volume and fast-paced EDs, DESTINY has promise as single tool to identify patients with acute symptomatic stroke and LVO.