Similarity Analysis of Complete Blood Count (CBC) Reference Interval Distributions Across Ethnic and Geographic Populations

Blood reference intervals (RIs) are central to diagnosis and therapeutic monitoring, yet most were derived from Western populations and assumed universal. This risks misclassification in regions with diverse demographic, physiological, or genetic profiles. We examined one of the most used panels, the Complete Blood Count (CBC), by compiling RI data from 28 countries and applying a multi-stage analytical framework. Structural similarity was assessed using multiple clustering strategies combining different linkage rules with Euclidean, correlation, and information-theoretic distances. To benchmark sensitivity, we introduced a Two-Level Cohesion Score quantifying continent-level grouping. UMAP embeddings with feature importance scores identified analytes potentially driving geography-related separation. Using BMI as a cross-country positive control, CBC reference intervals showed no reproducible clustering by geography or population genetics; weak, unstable signals were limited to MCV and HGB, unlike BMI. These findings indicate that CBC physiology is not geographically coordinated but instead reflects laboratory equipment, calibration, or logistical practices. Our results support moving from one-size-fits-all global RIs toward adaptive and personalized reference frameworks that link precision and predictive medicine with diagnostic equity in patient care.