Statistical Methods In Epidemiology--general

点击翻译按钮获取中文摘要

This project develops new statistical methods for epidemiology with broad applications and also methods as needed for ongoing projects in epidemiology, particularly those related to reproductive studies. The work this year involved two main projects. (1) One project concerns pooled assessment of expensive-to-assay biomarkers based on human samples. Earlier work had shown that in a case-control setting one can pool together specimens from sets of cases and sets of controls and carry out a set-based analysis. With a slightly modified logistic model that analysis can estimate the individual-level risk parameters and loses almost no power compared to analysis based on individual assays. This means that if an exposure is based on an expensive assay that uses human samples, one can markedly improve efficiency by pooling specimens prior to assay. In recent work, we have extended these methods to apply using outcome-dependent pooling in the context of time-to-event data, e.g. time to pregnancy studies. Although these methods can produce huge cost savings and spare precious samples, the methods impose restrictions that limit the flexibility of models that can be fit. For example, one cannot explore secondary endpoints and one cannot fit nonlinear dose-logit-response models. We are now doing additional work aimed at developing semi-parametric methods for "prospective pooling," which will overcome some of the limitations caused by the need to use outcome-dependent pooling. We are now developing semi-parametric methods for "prospective pooling" in time-to-event settings. Such methods should allow powerful and inexpensive assessments of the effects of environmental exposures on survival and other failure-time outcomes. (2) A second project is developing methods for analysis of outcomes related to environmental exposures where the exposure biomarker is often at a level below the assay limit of detection. Methods developed for survival analysis can be employed and permit confounder adjustment. We are applying this to data from the National Health and Nutrition Examination Survey data relating toxic analytes in blood to the presence of a biomarker, anti-nuclear antibodies, for autoimmunity. We have a paper describing this method under revision for the American Journal of Epidemiology. In future work we hope to extend this work to mixtures of chemicals, making use of existing toxic equivalency factors to create a censored toxicity-weighted average.