Biostatistics Core Overview

Investigators: Lance Waller, Ph.D. (Co-PI, Emory), W. Dana Flanders, M.D., Sc.D. (Co-PI, Emory), Mitchel Klein, Ph.D. (QA Advisor, Emory)

The Biostatistics Core provides design support, analytic guidance, data management, and collaborative analysis to Center investigators. As outlined in each project description and highlighted below, many of the proposed studies involve a fair amount of statistical agility in analysis, often branching beyond a fixed set of methods and requiring development of appropriate novel methodologies to address specific substantive questions.

The primary objective of the Biostatistics Core is to provide statistical support to the Center and to the associated research projects.  While all projects involve some sort of quantitative data analysis, the most direct associations with the Biostatistics Core are by Projects 2-4 (the health projects).  Project 1 interacts predominantly with the Air Quality Core in order to measure and estimate individual pollutants and mixtures of pollutants, and their temporal and spatial variation.  These measures and estimates of air pollutants and mixtures are used by the Biostatistics Core as exposure measures by the health projects to assess associations with health outcomes as described in each project, adjusted for analytic complications such as measurement error, spatio-temporal correlations, etc. 

The primary functions of the Biostatistics Core are: 
  1. Provide guidance and support to all projects for design issues. 
  2. Provide guidance and support to all projects for epidemiologic modeling, including identification and characterization of mixtures and their health effects.
  3. Methodological development, including:
    • Developmethods for identification of model misspecification
    • Adapt methods for analyses of time series based on LASSO and CART
    • Adapt principle component methods for identifying mixtures that associate with health outcomes
  4. Develop and provide support and guidance for addressing the impact of measurement error.
  5. Archive, document and assure security of analytic data files.