Presented By: Michigan Lifestage Environmental Exposures and Disease Center
Bayesian models for mercury effects on multiple outcomes in the Seychelles Child Development Study
2021 Environmental Statistics Week with Sally Thurston (Univ of Rochester Med Ctr)
Sally W. Thurston, PhD is an Associate Professor of Biostatistics and Environmental Medicine, University of Rochester, Dept of Biostatistics and Computational Biology.
ABSTRACT: The Seychelles Child Development Study (SCDS) Main Cohort was recruited to investigate the association between prenatal mercury (Hg) exposure from maternal fish consumption and the children’s subsequent development. Multiple neurodevelopmental outcomes were measured at several discrete ages. These outcomes can be grouped into different classes or "domains", such as cognition, motor, or memory. Associations of Hg and covariates may differ across outcomes and domains. Motivated by the SCDS, I will first discuss a model for multiple outcomes at a single age under the assumption that each outcome belongs to one domain and domain memberships are known. I will then briefly describe two extensions: a model in which outcomes may have partial domain membership in more than one domain, and a longitudinal multiple outcomes model for cognition when the tests of cognition change with the age of the child.
The goal of Environmental Statistics Week is to disseminate knowledge of advanced statistical methods most relevant to environmental health research with expert-led discussions on statistical concepts.
ABSTRACT: The Seychelles Child Development Study (SCDS) Main Cohort was recruited to investigate the association between prenatal mercury (Hg) exposure from maternal fish consumption and the children’s subsequent development. Multiple neurodevelopmental outcomes were measured at several discrete ages. These outcomes can be grouped into different classes or "domains", such as cognition, motor, or memory. Associations of Hg and covariates may differ across outcomes and domains. Motivated by the SCDS, I will first discuss a model for multiple outcomes at a single age under the assumption that each outcome belongs to one domain and domain memberships are known. I will then briefly describe two extensions: a model in which outcomes may have partial domain membership in more than one domain, and a longitudinal multiple outcomes model for cognition when the tests of cognition change with the age of the child.
The goal of Environmental Statistics Week is to disseminate knowledge of advanced statistical methods most relevant to environmental health research with expert-led discussions on statistical concepts.
Related Links
Livestream Information
ZoomNovember 2, 2021 (Tuesday) 12:00pm
Meeting ID: 93347765458
Explore Similar Events
-
Loading Similar Events...
