Title: Time-varying proportional odds model for mega-analysis of clustered event times with functional covariates
Authors: Tanya Garcia - Texas A\&M University (United States) [presenting]
Karen Marder - Columbia University (United States)
Yuanjia Wang - Department of Biostatistics (United States)
Abstract: Mega-analysis, or the meta-analysis of individual data, enables pooling and comparing multiple studies to enhance estimation and power. A challenge in mega-analysis is estimating the distribution for clustered, potentially censored event times where the dependency structure can introduce bias if ignored and functional covariates introduce a secondary complexity. We propose a new proportional odds model with unknown, time-varying coefficients and functional covariates. The model directly captures event dependencies, handles censoring using pseudo-values, and corrects for the impact of functional genetic information. We demonstrate that a simple transformation of the model leads to an easily estimable additive logistic mixed effect model. Our method consistently estimates the distribution for clustered event times even under covariate-dependent censoring and complex functional covariate structures. Applied to three observational studies of Huntington's disease, our method provides, for the first time in the literature, evidence that varying lengths of CAG repeats in the huntingtin gene (modeled as a functional covariate) has very different impacts on motor and cognitive impairments.