|| Checking for direct PDF access through Ovid
Enormous progress has been made in recent decades in our understanding of lead neurotoxicology in children, but an important obstacle to additional progress is the striking variability that is evident in any plot of a lead biomarker versus a health endpoint.In this article, three potential sources of variability are identified: (1) errors or imprecision in characterizing dose (and/or outcome); (2) incomplete characterization of endpoint variance attributable to factors other than lead; and (3) inter-individual differences in susceptibility to lead neurotoxicity.Strategies are suggested for reducing the variability contributed by these sources, including the development of validated PBPK models and biomarkers of early biological effects; the development of more comprehensive models of outcome variance and, specifically, the application of multi-level models that incorporate supra-individual and supra-family risk factors; and the use of study designs that permit assessments of the effect modifying influence of contextual factors on the form and severity of neurotoxicity.Decomposing the variability in the distribution of observed scores around the best-fit lines that describe the dose-effect relationships for lead neurotoxicity in children is a major research need. Am. J. Ind. Med. 50:720-728, 2007. © 2007 Wiley-Liss, Inc.