Pooling data when analyzing biomarkers subject to a limit of detection

Abstract

The use of biomarkers to assess exposure and investigate biomedical questions is common in epidemiology. The usefulness of biomarker research, however, is contingent upon the ability to achieve a complete understanding of the role they play within a population. In estimating distributional parameters for a particular biomarker, such as oxidative stress or antioxidant markers, scientists face two main challenges: overcoming the cost of performing a large number of assays and dealing with data subject to a limit of detection. While approaches have been suggested to deal with each of these issues individually, pooling is a strategy that can address both problems. © 2009 Humana Press.