To use data science methods to visualize the strengths and weaknesses of two popular prehospital stroke screening tools, the Los Angeles Motor Scale and the BEFAST scale.

Early recognition of stroke in the pre-hospital setting is essential for optimizing transport decisions and improving outcomes. The Los Angeles Motor Scale (LAMS) has long served as a validated screening tool, while BEFAST has recently been adopted by many EMS systems to capture additional neurological deficits such as balance and vision changes. Comparative analyses between these tools remain limited, and existing evaluations lack interactive, subgroup-based visualization to guide clinical insight.

Comparative analyses between these tools remain limited, and existing evaluations lack interactive, subgroup-based visualization to guide clinical insight. This study develops an interactive framework that compares the diagnostic and triage performance of BEFAST and LAMS across diverse patient subgroups, promoting transparency and data-driven improvement in pre-hospital stroke screening. An interactive confusion heatmap is implemented using Plotly Dash and Flask, enabling dynamic filtering by age, sex, onset time, and hospital type with confidence overlays and guided tooltips for interpretability.

~5,000 EMS records containing BEFAST and LAMS assessments, pre-hospital interventions, and hospital outcomes were analyzed. The cohort was 53% female with a median age of 71 years and an IQR of 60-80 years. Both A positive BEFAST or LAMS were both significantly associated with the outcomes of receiving tPA (P=0.0007) or mechanical thrombectomy (P<0.0001). A higher number on either scale was also significantly associated with receiving intervention for stroke. Using visualization tools, subgroup filtering reveals distinct misclassification patterns not visible in static analyses, identifying population-specific variations in false positives and negatives.

This visualization framework transforms stroke screening evaluation into an interactive and clinically interpretable process, providing EMS agencies and researchers with a transparent platform to refine pre-hospital triage tools and improve patient outcomes.