Karim Shater

Operationalizing Instability in Rule-Based Complete Blood Count Phenotyping Using Uncertainty-Aware Machine Learning

Karim Shater, Catharina Gerhards, Osman Evliyaoglu, Stefanie Nittka, Andreas Fischer

Background: Complete blood count (CBC) phenotypes are routinely assigned using deterministic rule-based thresholds. While operationally efficient, such rules may lead to unstable phenotype assignments for results close to clinical cutoffs in the presence of analytical variability…Read more