A1062
Title: Predictive partly conditional model for longitudinal outcomes in the presence of informative dropout and death
Authors: Dandan Liu - Vanderbilt University Medical Center (United States) [presenting]
Abstract: Assessing time-dependent risk factors in relation to the risk of disease progression is challenging yet important, especially for chronic diseases with slow progression. An important consideration when modelling outcomes related to ageing is the potential for dropout in the study or death prior to the disease occurring. The predictive partly conditional model (PPCM) is extended to characterize disease progression at time t with longitudinal outcomes in the presence of time-dependent covariates at time $s (s < t)$ when informative death and dropout are present. Inverse probability weighting is adopted to separately account for the probability of death and dropout with the generalized estimating equation approach and establish the conditions for consistency. Extensive simulation studies are conducted to assess the properties of the proposed model and demonstrate the implementation of the proposed model using existing statistical software. Application to data from the National Alzheimer's Coordinating Center will be conducted to illustrate this method.
