Title: Inferring random change point from longitudinal data subject to left-censoring by segmented mechanistic nonlinear models
Authors: Hongbin Zhang - CUNY (SPH) (United States) [presenting]
Abstract: Random effects change-point models are commonly used to infer individual-specific time of event that induces trend change of longitudinal data. Linear models are often used before and after the change point. However, linear models may fit the completely observed data well, but may be inappropriate when certain portion of data are censored. In applications such as HIV studies, a mechanistic nonlinear model can be derived for the process based on the underlying data-generation mechanisms and such nonlinear model may provide better ``predictions" for the censored values. We propose a random change point model in which we model the longitudinal data by segmented nonlinear mixed effect models and address the left-censoring with the data. We propose a Monte Carol EM based method for the inference. We apply the model on an HIV surveillance data to estimate the time from HIV diagnosis to initiation of antiretroviral therapy (ART) initiation and evaluate the method with simulation to gain insights.