








Both models use repeated nested internal cross-validation in the Stanford public cohort (n=114; 31 Long COVID). No external validation is claimed.
Green and red thresholds are selected automatically from OOF predictions using a reject-option policy; yellow is the uncertainty zone.
Early Risk is evaluated as prospective risk stratification. PASC Check Lite additionally uses Decision Curve Analysis.
This is a research-use decision-support tool and not a standalone diagnosis.