Rectangular Statistical Regions with Applications in Laboratory Medicine

Thomas Mathew, Department of Mathematics and Statistics
Michael Daniel Lucagbo, Department of Mathematics and Statistics

Reference regions are invaluable in the interpretation of results of biochemical and physiological tests of patients. Moreover, when there are multiple biochemical analytes measured on each subject, a multivariate reference region is called for. Our work is on the development and computation of rectangular reference regions both in a multivariate normal setup and in a nonparametric setup. We consider both prediction and tolerance regions to be used as reference regions. The methodologies employed in this work involve the bootstrap (both parametric and non-parametric), Box-Cox transformations and kernel density estimation.