A Random-Effects Approach to Generalized Linear Mixed Model Analysis of Incomplete Longitudinal Data

TL;DR

This paper introduces a random-effects method for handling missing data in generalized linear mixed models, enabling standard analysis tools to be used effectively for incomplete longitudinal data.

Contribution

It presents a novel approach that transforms GLMMs with missing covariates into complete models, supported by theoretical justification and empirical evaluation.

Findings

Method performs well compared to multiple imputation.

Theoretical basis explains observed simulation patterns.

Effective in healthcare longitudinal data analysis.

Abstract

We propose a random-effects approach to missing values for generalized linear mixed model (GLMM) analysis. The method converts a GLMM with missing covariates to another GLMM without missing covariates. The standard GLMM analysis tools for longitudinal data then apply. The method applies, in particular, to the cases of linear mixed models and logistic regression. Performance of the method is evaluated empirically, and compared with alternative approaches, including the popular MICE procedure of multiple imputation. Theoretical justification of the method is given, and explained, for the patterns observed in the simulation studies. Two real-data examples from healthcare studies are discussed.