A0591
Title: Mediation analysis with ultra-high dimensional confounders for the study on geriatric depression and Alzheimer's disease
Authors: Yuexia Zhang - The University of Texas at San Antonio (United States) [presenting]
Annie Qu - University of California Irvine (United States)
Yubai Yuan - Penn State University (United States)
Qi Xu - University of California Irvine (United States)
Fei Xue - Purdue University (United States)
Kecheng Wei - Fudan University (China)
Abstract: Depression and Alzheimer's disease (AD) are both prevalent diseases in older adults. Using the data sets from the Alzheimer's disease neuroimaging initiative (ADNI) study, whether geriatric depression is explored has a significant average treatment effect on AD and whether the effect is mediated by some important mediators. To estimate these causal effects consistently, ultra-high dimensional potential confounders are controlled for, including DNA methylation levels. A new ball correlation-based screening method is proposed for confounder selection in mediation analysis. A robust mediation analysis framework is utilized to achieve robustness against model misspecification. Simulation studies show that the proposed method has good finite-sample performance in terms of confounder and mediator selection, effect estimation, and inference. In the real data analysis, it is found that geriatric depression has a significantly positive causal effect on AD. New prevention and treatment strategies are also proposed for geriatric depression and AD by changing the selected confounders and mediators.
