|
Activity Number:
|
531
- Precision Medicine: Methods, Tools, and Applications
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 11, 2022 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
|
Abstract #319253
|
|
|
Title:
|
Simultaneous Modeling of the Mean and Within-Subject Variability Trajectories of a Longitudinal Biomarker Together with a Competing Risks Time-to-Event Outcome
|
|
Author(s):
|
Gang Li* and Shanpeng Li and Jin Zhou and Hua Zhou
|
|
Companies:
|
University of California, Los Angeles and UCLA and UCLA and UCLA
|
|
Keywords:
|
Competing risks data;
Competing risks data;
Longitudinal data;
Massive sample size;
Nonignorable missing data;
Within-Subject Variability
|
|
Abstract:
|
Recent research reveal the within-subject (WS) variation of some longitudinal biomarkers as strong risk factors for many health outcomes, which have motivated some interesting methodology development to simultaneously model both the mean and within-subject variability of longitudinal biomarker(s). In this research, we propose a joint model of the mean and within-subject variability trajectories of a longitudinal biomarker together with a time-to-event outcome. A customized EM algorithm is derived for semiparametric maximum likelihood estimation. We also develop high performance computational algorithms to make the proposed method scalable to massive biobank data. The developed method is applied to a large diabetes study to evaluate how an intensive treatment targeting HbA1c affects the mean and variability of fasting plasma glucose (FPG) over time, and how the mean and variability trajectories of FPG predicts the risk of cardiovascular disease.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2022 program
|
