Abstract Objectives Identifying patients at high risk for atrial fibrillation (AF) after cryptogenic stroke remains a challenge, particularly in settings with limited access to long-term cardiac monitoring. The AFibrisk platform, a free digital decision-support tool, integrates 19 validated AF prediction scores to support post-stroke triage. We aimed to assess the concordance of AFibrisk-supported classification decisions with expert electrophysiologist consensus and compare performance across evaluator groups with different levels of clinical experience. Materials and Methods A prospective, cross-sectional concordance study was conducted using 29 standardized clinical vignettes. Evaluators—3 vascular neurologists, 4 cardiology residents, and 11 neurology residents—classified each case as high or low AF risk using AFibrisk outputs. Expert consensus served as the reference standard. Statistical analyses included inter-group comparisons, inter-rater reliability, and regression models adjusting for group size and response clustering. Results Vascular neurologists demonstrated the highest agreement with the reference standard (mean 90.3%), followed by cardiology residents (85.2%) and neurology residents (77.5%). Differences were statistically significant (ANOVA p = .0199; Kruskal–Wallis p = .0259). Neurology residents showed the greatest intra-group consistency (Light’s κ = 0.607), despite lower accuracy. Classification errors differed by experience: residents tended to overestimate risk, while experts showed occasional underestimation. Overall, 30.1% of responses were “not classified,” with the highest uncertainty among vascular neurologists (43.8%). Discussion and Conclusion AFibrisk improved alignment with expert judgment across evaluator groups and helped standardize decision-making. Our free platform may support AF risk stratification in low-resource environments and reinforce evidence-based heuristics among early-career clinicians, and it is available at www.afibrisk.net
Andrade et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: