The long-term goals of this grant are to develop study designs and methods of statistical analysis for epidemiologic research of cancer and other chronic diseases. In this proposal, we focus on case-control studies in three broad aims: The first is to develop case-control design and analysis solutions to a number of important problems we have identified based on our applied work on """"""""standard""""""""' and counter matched case-control studies. These include appropriate analysis of case-control studies with random digit dial controls, Mantel-Haenszel and polytomous logistic regression estimators for studies with complex control selection, development of new population based case-control study designs that use inexpensive correlate exposure information to improve cost-efficiency, and analysis of case-control studies with missing covariate data. The second is to develop efficiency bound methods appropriate to case-control study designs as well as investigate estimation methods that make use of cohort or external information to improve case-control study efficiency. These results will clarify when auxiliary at-risk and covariate information can be used to improve efficiency in analysis of nested case-control studies and provide methods that exploit this information.
The third aim i s outreach to investigators interested in the methods including maintaining software and data examples on our website and limited consultation. The impact of these projects is to expand the design and analysis """"""""tools""""""""' available to epidemiologic research in a way that will improve validity and cost-efficiency. ? ?
Showing the most recent 10 out of 20 publications