Statistical Methods in Epidemiology--General

Environmental Health Sciences

Investigators

Linked publications & trials

Abstract

Methodologic work progressed in two areas: (1) When data are clustered, e.g. based on litter in a toxicology experiment, or sibship in a human study, conditional logistic regression is often used to allow for dependencies where inherent susceptibility to the endpoint under study varies across clusters. If there are unmeasured factors that vary across clusters and that influence susceptibility to effects of the exposure under study, then there may be residual dependency, which will invalidate such analyses. These factors are called "effect modifiers" in epidemiology. We developed a statistical test for unmeasured effect modification and carried out simulations to confirm that the approach has reasonable power to detect such model violations. When such violations are detected, good alternatives exist in place of conditional logistic regression models, including the within-cluster paired resampling approach that was developed at NIEHS. (2) A number of studies have provided evidence that human fertility has declined over recent decades, raising alarm that a widespread reproductive toxicant exists in the environment. We carried out simulations to assess the bias due to recent trends in the availability of effective methods of birth control, including induced abortion, and showed that contradictory reports in the literature can potentially be explained by these demography-induced biases, and that the study of trends in fertility is particularly problematic.