|
Activity Number:
|
213
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #304719 |
|
Title:
|
Joint Modeling of Longitudinally Measured Dialysis Reuse and Time-to-Death in the HEMO Study
|
|
Author(s):
|
Abdus Sattar*+ and Lisa A. Weissfeld and Christos Argyropoulos and Mark Unruh
|
|
Companies:
|
University of Pittsburgh and University of Pittsburgh and University of Pittsburgh Medical Center and University of Pittsburgh
|
|
Address:
|
305 S. Fairmount St., Apt 1, Pittsburgh, PA, 16232,
|
|
Keywords:
|
Joint modeling ; Repeated Poisson process ; Cox proportional hazard model ; Regression splines ; Smoothing splines
|
|
Abstract:
|
Our objective is to analyze time-to-death of all causes mortality from HEMO study, a large multicenter study of hemodialysis. Along with time varying covariates, number of dialyzer reuses has been recorded longitudinally. Due to a protocol change to reduce the number of reuses in certain dialyzers and secular changes in policies in the dialysis units, the distribution of the number of dialyzer reuse has two potential parameters. We will analyze time to all cause mortality using a joint modeling approach by taking the number of reuses as repeated Poisson process. We will compare the result of joint model with the result of a number of competitive survival models: time dependent covariates in Cox model, and regression/smoothing splines. A simulation study will be conducted to understand the true variation in the inferences due the modeling approach.
|
