REDEFINING LABORATORY REFERENCE RANGES FOR FEMALE REPRODUCTIVE AND THYROID HORMONES

Abstract

The diagnoses and clinical management of many reproductive health conditions rely on categorically interpreting hormone measurements as either within or outside a ‘normal’ reference range. Historically these ranges represent the 95% confidence intervals of a small population without regard to the distributional asymmetry expected by compositional data. We establish a more robust statistical framework that considers continuous quantiles, derived from a large dataset, in order to better define normal reference ranges of reproductive and thyroid hormone levels.