Abstract

Purpose
Stroke is one of the leading causes of global morbidity and mortality, requiring rapid and accurate early detection. Popular screening tools such as FAST often miss posterior strokes, whereas BE-FAST shows better sensitivity; however, its digital implementation remains limited. To develop, validate, and implement BE-ALERT as a community-based early stroke detection application.
Methods
This study used a Research and Development (R&D) design with the following stages: design, development, validation (content, construct, reliability, diagnostics), and limited community implementation. The population consisted of residents aged ≥18 years and suspected stroke patients in emergency department. Sampling techniques included multistage cluster (community) and consecutive sampling (emergency department). Data analysis included Content Validity Index (CVI), reliability, diagnostic accuracy testing (ROC-AUC, sensitivity, specificity), and usability (SUS). This study followed the STARD (Standards for Reporting Diagnostic Accuracy Studies) to ensure transparent and comprehensive reporting of all methodological and diagnostic accuracy aspects.
Results
Validation results showed a CVI of 0.89 and Cronbach's Alpha reliability of 0.82. Implementation among 160 community respondents showed a significant increase in stroke knowledge (82.1), intention to act quickly (4.5), and a SUS usability score of 74 (good). The ROC curve showed an AUC of 0.87, indicating high diagnostic accuracy. BE-ALERT showed a sensitivity of 0.85, specificity of 0.82, NPV of 0.95, and usability score of 74.
Conclusion
BE-ALERT has the potential to be an accurate, practical, and well-received early stroke detection tool in the community. This application has the potential to be an innovation in community-based stroke screening and education efforts.

• Keywords: Stroke; Early diagnosis; Mobile applications; Sensitivity and specificity; Validation study; BE-ALERT