Online Program

Piece-Wise Mixed Effect Model for Renal Function Data Analysis in Transplantation Patients (302992)

Keywords: Piece-wise mixed-effect model, longitudinal data, covariance structure, model diagnosis

In a renal transplantation study, renal function change over time is the primary endpoint in test treatment effect. Longitudinal analysis is generally used in analyzing such data change over time. The renal function is often not linear over all the time period, but it is linear for some time periods, hence we are modeling the data with piece-wise mixed effect model and test treatment effect at specified time points and overall. The background of piece-wise mixed effect model will be introduced and application to renal function data from a Novartis renal transplant clinical trial study will be provided. Many other techniques related to longitudinal data modeling will be provided, including fixed effect selection, covariance structure choice, model diagnosis, and testing. The comparison with other mixed effect models (e.g., polynomial, loglinear, exponential, etc.) will be presented also.