A0579
Title: Semiparametric modeling of biomarker trajectory and variability with correlated measurement errors
Authors: Renwen Luo - BNU-HKBU United International College (China)
Chuoxin Ma - Beijing Normal-Hong Kong Baptist University (China) [presenting]
Jianxin Pan - The University of Manchester (United Kingdom)
Abstract: The prognostic significance of biomarker variability in predicting disease risk is well-established. However, existing methods for evaluating the relationship between biomarker variability and time-to-event outcomes often fail to account for within-subject correlation in longitudinal measurement errors. This oversight can lead to biased parameter estimates and incorrect statistical inferences. Furthermore, these methods typically model biomarker trajectories as linear combinations of spline basis functions with normally distributed random effects. While this approach is common, it imposes significant computational burdens due to the need for high-dimensional integration over the random effects. Additionally, the assumption of normality for the random effects limits the flexibility and applicability of these methods. To address these limitations, a novel semiparametric multiplicative random effects model is proposed for biomarker trajectories, coupled with a Cox proportional hazards model that incorporates biomarker variability as a covariate. The application of this methodology is demonstrated by assessing the impact of systolic blood pressure variability on cardiovascular mortality using data from the Atherosclerosis Risk in Communities (ARIC) study.
