Smoothed Simulated Pseudo Maximum Likelihood Estimation for Nonlinear Mixed Effects Models with Censored Responses

• Activity Number: 275 - Joint Models for Complex Data: An Update on Computational Issues, Solutions, and Applications
• Type: Topic-Contributed
• Date/Time: Wednesday, August 11, 2021 : 1:30 PM to 3:20 PM
• Sponsor: International Chinese Statistical Association
• Author(s): Yue Song and Rui Wang*
• Companies: Harvard T. H. Chan School of Public Health and Harvard Pilgrim Health Care Institute
• Keywords: random effects; left-censoring; viral rebound

Abstract

Nonlinear mixed effects (NLME) model has been widely applied to analyzing longitudinal data. Implementation of NLME models can be complicated by left-censored observations, representing measurements from bioassays where exact quantifications below a certain threshold are not possible. In addition, although normality of random effects is commonly assumed, violations of this assumption could lead to biased inferences of the variance components for approaches requiring the correct specification of the likelihood function. In this article, we propose a method for fitting NLME models based on the smoothed simulated pseudo likelihood function. The consistency and asymptotic normality of the resulting estimator are established. One advantage of the proposed method is its flexibility in the specification of random effects distributions. This work is motivated by modeling of the HIV RNA viral load trajectories among patients who underwent an antiretroviral treatment interruption. We demonstrate satisfactory finite-sample performance of the proposed methods through simulation studies and illustrate these using a combined dataset from six treatment interruption studies.