|
Abstract:
|
A joint modeling likelihood-based approach for studies with repeated measures and informative right censoring is presented. The censoring process is defined as a function of longitudinal trajectories and baseline measures of a specific biomarker determined by individual slopes and intercepts and a set of covariates that could be directly affecting the censoring probability. The individual intercepts and slopes are shared between the longitudinal measurements and censoring process and are therefore considered latent random variables in the model. Maximum likelihood estimates for the population slopes and intercepts, their variance-covariance matrix, and their respective censoring parameters, and estimates of the parameters pertaining to the covariates are all generated. This model was applied on two clinical datasets to determine the predictive utility of specific biomarkers over time on progression of kidney disease.
|
