The investigators propose to continue the development of two classes of statistical methods relevant to cancer epidemiology involving extended exposure histories and temporal modifying factors.
One aim concerns design aspects of sampling within a cohort and the other concerns the analysis of exposure-time-response relationships. These two broad aims have remained unchanged since the beginning of the grant. But, as time has progressed, attention has been narrowed to particularly fruitful topics. Design and analysis methods for sampled cohort data. Much of the work in the last grant cycle was devoted to building a theoretical foundation for the analysis of a broad class of nested case-control studies designs. This work has resulted in some important innovations both in the design and analysis of matched and nested case-control studies and the investigators intend to continue along these lines in this grant cycle. They further wish to expand the theoretical developments to methods for case-cohort designs and unmatched case-control studies. Methods of exposure-time-response modeling. Methods have been developed for describing exposure-response relationships for extended time-dependent exposure histories, taking into account the modifying effect of time-related variables such as age at exposure and latency. During this grant cycle the investigators intend to develop descriptive methods for visualizing time-modifying effects, develop methods to test the """"""""additvity of risk"""""""" assumption commonly made in models for latency, and develop estimators of absolute risk for time-dependent exposures for cohort and nested case-control data.
Showing the most recent 10 out of 20 publications
