You are here:
ANALYSES OF NEUROBEHAVIORAL SCREENING DATA: DOSE-TIME-RESPONSE MODELING OF CONTINUOUS OUTCOMES.
Citation:
Zhu, Y., M. R. Wessel, T. Liu, AND V C. Moser. ANALYSES OF NEUROBEHAVIORAL SCREENING DATA: DOSE-TIME-RESPONSE MODELING OF CONTINUOUS OUTCOMES. REGULATORY TOXICOLOGY AND PHARMACOLOGY. Elsevier Online, New York, NY, 41:240-255, (2005).
Description:
Analysis of neurotoxicological screening data such as those of the functional observational battery (FOB) traditionally relies on analysis of variance (ANOVA) with repeated measurements. Although ANOVA is effective in detecting dose-induced neurotoxic effects, it does not directly model the underlying dose-response relationship, nor utilize the shape of the dose-response with respect to time for risk assessment. Such models are a prerequisite for deploying some recent advances in risk assessment methodology such as the benchmark dose method. This paper discusses dose-response modeling of longitudinal neurotoxicity data and illustrates the modeling process through several continuous, FOB outcomes from an EPA study involving acute exposure to triethyltin (TET). Several mathematical functions are presented as candidate dose-time-response models. The use of random effects are discussed for the purpose of characterizing inter-subject variation in dose-response and for improving statistical effectiveness in analysis. The results indicate that simple mathematical functions are useful to describe empirical dose-response relationships observed in longitudinal neurotoxicological data such as those of a FOB, which are not necessarily designed or generated for dose-response modeling purposes.