Slope estimation of covariates that influence renal outcome following renal transplant adjusting for informative right censoring

Abstract

A new statistical model is proposed to estimate population and individual slopes that are adjusted for covariates and informative right censoring. Individual slopes are assumed to have a mean that depends on the population slope for the covariates. The number of observations for each individual is modeled as a truncated discrete distribution with mean dependent on the individual subjects’ slopes. Our simulation study results indicated that the associated bias and mean squared errors for the proposed model were comparable to those associated with the model that only adjusts for informative right censoring. The proposed model was illustrated using renal transplant dataset to estimate population slopes for covariates that could impact the outcome of renal function following renal transplantation.

Keywords:

• discrete geometric distribution
• empirical Bayes estimates
• informative right censoring
• longitudinal data
• likelihood function
• slope estimation

Acknowledgement

This work was supported by National Institutes of Health Grants HL077192 and HL087986.