Biostatistics Weekly Seminar


Nonparametric Estimation for Time-varying Missing Covariates in Longitudinal Models

Panpan Zhang, PhD Candidate
Department of Biostatistics, Epidemiology, and Informatics, University of Pennsylvania

Longitudinal models are useful in biomedical research for studying predictors of disease progression. When data are missing, methods for missing data often require parametric assumptions about the distributions of covariates. In this paper we propose a nonparametric method for longitudinal data, in which one or more covariates are missing for a subsample of subjects, but an auxiliary variable is available for everyone. This type of data can arise when a predictor is expensive or difficult to measure and therefore only collected for a fraction of subjects, but all subjects have a more easily accessible variable that is related to the missing covariate. We use empirical and kernel density estimates to obtain nonparametric density estimates instead of relying on parametric assumptions about the conditional distribution of the missing data given the observed. Our proposed method can handle time-independent or time-varying missing covariates and auxiliary variables. In addition, the auxiliary variable can be discrete or continuous. We derive the asymptotic distribution of the estimator and show that it is consistent and asymptotically normally distributed. Through simulations we show that our estimator has good finite sample properties and is more efficient than the complete case estimator. We show that the variance is well estimated using the asymptotic theory and therefore does not require bootstrapping. We apply our method to a Parkinson's disease dementia study.


Zoom (Link to Follow)
20 December 2021
1pm


Speaker Itinerary

Topic revision: r2 - 13 Dec 2021, DandanLiu
 

This site is powered by FoswikiCopyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback