Joint modeling of longitudinal CD4 cell counts and time-to-default from HAART treatment: a comparison of separate and joint models

Abstract

In HAART and other follow-up clinical trials, both longitudinal and survival data are generated. Joint models are used to describe the joint behaviour of such data. This study has discussed Bayesian joint modeling approaches using a five years HAART data obtained from Jimma University Specialized Hospital HIV Outpatient Clinic, Ethiopia. The objective is to compare separate and joint models of longitudinal CD4 cells counts and default time processes of HIV/AIDS patients. A linear mixed effects model, assuming homogenous and heterogenous CD4 variances, is used for modeling the CD4 counts and a Weibull survival model is used for describing the default times. Then, both processes are linked using unobserved random effects through the use of a shared parameter model. The analysis of both the separate and the joint models reveal that the assumption of heterogenous (subject-specific) CD4 variances brings a substantial improvement in the model t. The parameter estimates of both the separate and joint models are consistent. However, the joint model is parsimonious and fits the data better. The final joint model relates the hazard of defaulting to two characteristics of the repeated CD4 counts; patient-specic slopes and CD4 variability.