Joint Analysis of Longitudinal Data and Competing Risks Survival Times in the Presence of Dependent Observational Times

• Activity Number: 85
• Type: Contributed
• Date/Time: Sunday, August 4, 2013 : 4:00 PM to 5:50 PM
• Sponsor: International Chinese Statistical Association
• Author(s): Tai-Fang Chen Lu*+ and Chyong-Mei Chen
• Companies: Providence University and Providence University
• Keywords: Longitudinal data ; Informative observational times ; Terminal event ; Competing risk ; Random effect

Abstract

The work is motivated by a study of patients with chronic kidney disease (CKD) in Taichung Veterans General Hospital. A total of 1096 patients receiving treatment were enrolled and followed between December, 2001 and July, 2012. In the study, physicians are mainly interested in the decline rate of estimated Glomerular Filtration Rate (eGFR). The collected data tends to be longitudinal data possibly with 2 competing terminal events, death or dialysis. Moreover, the observational times may depend on the repeated measurements. Accordingly, a flexible model is proposed using random effects to describe the association among longitudinal data, competing terminal events and observational times. In addition, it also allows different correlations between two terminal events and the repeated measurements, or the observational times. The corresponding likelihood approach is derived for the statistical inference. Extensive simulation studies reveal that the proposed approach performs well for practical situations, and finally, it is applied to estimate the decline rate of GFR for the CKD data.