Bayesian Hierarchical Modeling of Longitudinal Glaucomatous Visual Fields using a Two-Stage Approach

TL;DR

This paper presents a two-stage Bayesian hierarchical model for analyzing longitudinal visual field data in glaucoma patients, enabling efficient and flexible modeling of complex data structures to detect disease progression early.

Contribution

It extends a two-stage Bayesian approach to handle complex hierarchical glaucoma data, including censoring and measurement variability, improving computational efficiency over traditional methods.

Findings

Successfully modeled glaucoma progression with large hierarchical data

Extended the two-stage approach with model evaluation tools

Demonstrated improved computational feasibility for complex models

Abstract

The Bayesian approach has become increasingly popular because it allows to model quite complex models via Markov chain Monte Carlo (MCMC) sampling. However, it is also recognized nowadays that MCMC sampling can become computationally prohibitive when a complex model needs to be fit to a large data set. To overcome this problem, we applied and extended a recently proposed two-stage approach to model a complex hierarchical data structure of glaucoma patients who participate in an ongoing Dutch study. Glaucoma is one of the leading causes of blindness in the world. In order to detect deterioration at an early stage, a model for predicting visual fields (VF) in time is needed. Hence, the true underlying VF progression can be determined, and treatment strategies can then be optimized to prevent further VF loss. Since we were unable to fit these data with the classical one-stage approach upon which the current popular Bayesian software is based, we made use of the two-stage Bayesian approach. The considered hierarchical longitudinal model involves estimating a large number of random effects and deals with censoring and high measurement variability. In addition, we extended the approach with tools for model evaluation