The 130th Annual Meeting of APHA

Abstract #37909

Douglas J. Taylor, Biostatistics, Family Health International, P.O. Box 13950, Research Triangle Park, NC 27709, (919)544-7040, dtaylor@fhi.org, Lawrence L. Kupper, Department of Biostatistics, University of North Carolina, 3101D McGavran-Greenberg Hall, Campus Box #7420, Chapel Hill, NC 27599-7420, and Stephen M. Rappaport, Environmental Sciences & Engineering, University of North Carolina, 4114F McGavran-Greenberg Hall, Campus Box #7400, Chapel Hill, NC 27599-7400.

In conjunction with measurements of external exposure, biomarker measurements have been used in an attempt to provide information about the uptake, bioactivation, and detoxification of toxic chemicals in humans. Such information would be quite valuable since it would reduce reliance solely on animal models for making extrapolations about disease risks in humans. However, valid and precise quantification of exposure-biomarker relationships has been hampered by at least three problems: i) large variability in observed exposure and biomarker levels both within and between subjects; ii) errors in the measurement of both exposure and biomarker levels; iii) exposure and biomarker levels that fall below known detection limits. In this paper, we describe maximum likelihood methods that appropriately adjust for these problems and that are applicable for biomarkers that are either short-term or long-term in nature. These methods allow valid and precise statistical inferences to be made about important regression parameters and variance components in latent variable models. Functions of these parameters can be used to characterize the relationships between true (but unobservable) exposure and biomarker mean levels. To illustrate their utility, these maximum likelihood methods are used to analyze and interpret some exposure-biomarker data.

Learning Objectives: At the conclusion of this presentation, the participants will

• Recognize the importance of the use of biomarkers to assess human health risks due to exposures to toxic substances.
• Recognize that sophisticated statistical methods are needed to analyze and interpret data involving exposure and biomarker concentrations measured on human subjects.

Keywords: Statistics, Environmental Exposures

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.