-- Department of Environmental Medicine --
This Web site is supported in part by a grant from
the National Institute of Environmental Health Sciences (P30 ES01247).
The Biostatistics Facility Core provides statistical and data management
services to EHSC faculty, postdoctoral fellows, and students. Emphasis is
placed on proper study design prior to experimentation. Faculty within this
Facility Core have developed long-standing relationships with researchers
for the continued analysis of data coming from populations under study,
experiments with animal models of disease and human subjects, toxicokinetic
questions in humans, and basic mechanistic questions in molecular toxicology. The experience of Biostatistics faculty in the field of statistical analysis of environmental studies allows the application of innovative statistical methodology to the design and analysis of laboratory experiments and clinical studies.
Services Provided
Consultation with EHSC Investigators on statistical and computing components
of research projects. - For example, advice on appropriate experimental design, sample size for
clinical studies and laboratory experiments, inclusion/exclusion criteria,
definition of exposure and outcome variables, determination of randomization/stratification schemes
and design of data forms.
Development and implementation of data management systems and software for quality control,
update reports and interim statistical analysis. Analysis of completed studies, including statistical analysis, report writing and development
of graphical displays for publication and presentations.
Highlights of Recent Biostatistics Facility Core Activities
-
The Seychelles study was undertaken to define the potential
developmental effects of
prenatal exposure to methylmercury via fish consumption of mothers.
The study is a cooperative
effort between the University of Rochester and the government of the Seychelles
Islands, a large archipelago in the Indian Ocean.
Biostatistics support for this
project is provided by
Li-Shan Huang and
Sally Thurston.
- Thurston et al. (2007)
used regression techniques to analyze relationships between prenatal exposure
to methylmercury and blood pressure. Thurston et al. (2008) developed novel
methods for combining information from multiple outcome variables and applied
them to the Seychelles data.
- Huang et al. (2005) used generalized additive
models to search for nonlinear trends in outcome data with prenatal mercury
exposure. Huang et al. (2007) used regression trees to explore possible
non-homogeneous susceptibility to methylmercury exposure. These last two
papers involved development and application of novel biostatistical methodology
to the Seychelles study data.
-
The Biostatistics Facility Core has also been involved in a series of animal
studies with Dr. Günter Oberdörster.
- This encompassed
the fitting and interpretation
of one and two compartment uptake and elimination models to data from groups of
animals. This included choice of a model, fitting by nonlinear least squares,
and assessment of goodness of fit and statistical significance.
- A second aspect
is the computation of derived quantities such as half-times and their standard
errors, and comparisons of these quantities between groups.
- Other studies have
involved the fitting and comparison of dose-response models, and analysis of
variance of parameters such as cell counts from animal studies of exposure to
inhaled particles. They have also examined the effect of diesel fumes on
physiological function, and comparisons of the respiratory effects of various
dust samples taken from the World Trade Center site following the 9/11 attacks.
|
