Latent drop-out transitions in quantile regression

TL;DR

This paper extends longitudinal quantile hidden Markov models to account for informative missing data, specifically monotone missingness, improving inference reliability in analyzing dependent data like HIV CD4 counts.

Contribution

It introduces a novel approach to handle informative missing data in longitudinal quantile regression models with hidden Markov structures.

Findings

Effective handling of monotone missingness improves model accuracy.

Application to HIV CD4 count data demonstrates practical utility.

Simulation studies confirm robustness of the proposed method.

Abstract

Longitudinal data are characterized by the dependence between observations coming from the same individual. In a regression perspective, such a dependence can be usefully ascribed to unobserved features (covariates) specific to each individual. On these grounds, random parameter models with time-constant or time-varying structure are well established in the generalized linear model context. In the quantile regression framework, specifications based on random parameters have only recently known a flowering interest. We start from the recent proposal by Farcomeni (2012) on longitudinal quantile hidden Markov models, and extend it to handle potentially informative missing data mechanism. In particular, we focus on monotone missingness which may lead to selection bias and, therefore, to unreliable inferences on model parameters. We detail the proposed approach by re-analyzing a well known dataset on the dynamics of CD4 cell counts in HIV seroconverters and by means of a simulation study.