|
Activity Number:
|
163
- SPEED: Longitudinal/Correlated Data
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2018 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
|
Abstract #329975
|
Presentation
|
|
Title:
|
Modeling a Longitudinal Covariate as Continuous Time Markov Chain in a Survival Framework
|
|
Author(s):
|
Ting-Yu Chen* and Wenyaw Chan and Qiuling Shi and Xin Shelley Wang and Charles Cleeland
|
|
Companies:
|
The University of Texas Health Science Center at Houston and University of Texas Health Science Center at Houston and The University of Texas MD Anderson Cancer Center and The University of Texas MD Anderson Cancer Center and The University of Texas MD Anderson Cancer Center
|
|
Keywords:
|
Continuous Time Markov Chain;
Survival model;
Transition rate;
Patient reported outcomes
|
|
Abstract:
|
Longitudinal variables and survival outcomes have been modeled jointly using the trend of a longitudinal variable as a covariate in the survival model. Very few statistical modeling approaches manage the situation when the longitudinal covariate is a continuous time Markov Chain (CTMC). In this study, we estimate the prognostic value of longitudinal patient-reported symptom scores for overall survival by incorporating the transition rates of a CTMC as covariates. CTMC is obtained by dichotomizing these longitudinal symptom scores, measured via MD Anderson Symptom Inventory (MDASI), as none/mild or moderate/severe level. Results show that faster change of fatigue severity from none/mild to moderate/severe level is associated with a higher risk of death for patients with stage III/IV non-small cell lung cancer undergoing chemotherapy. The method can better identify the effects of symptom level changes on the survival outcome when the individual symptom level is dynamic.
|
Authors who are presenting talks have a * after their name.
