Semiparametric joint models for longitudinal biomarkers and time to event data The goal of this project is to develop novel statistics methods to jointly model longitudinal biomarker trajectories and time to event data. The proposed methods are motivated and will be applied to three major applications: 1) liver transplant and kidney transplant available through the United Network for Organ Sharing (UNOS), 2) the end stage renal disease (ESRD) data available through the United States Renal Data System (USRDS), and 3) the Vaginal birth after a prior cesarean (VBAC) data collected at the University of Pennsylvania. The main motivation comes from the fact that biomarkers are usually the surrogates of the underlying disease process and need to be treated as surrogate outcomes in modeling the time to event data, and the trajectories of the biomarkers usually require nonparametric models allowing flexible patterns over time, such as smooth curves, shape-registered curves, and branching curves. Another motivation is that in predicting the event such as death, the cumulative effects of the biomarkers may be more appropriate than the concurrent values, and therefore we propose to combine the ideas of functional data analysis and survival analysis. We will first develop the functional accelerated failure time (AFT) models and their join models with functional mixed effects models. We then extend this framework to include non-Gaussian longitudinal biomarkers. The third specific aims will develop a series of nonlinear functional mixed effect models for curve registration and branching curves, and their joint models with time to event data.
Each specific aim i ncludes methods development, theoretical studies, empirical simulations and applications. We will also develop a user-friendly software package that includes all the proposed features and post it to public domain.

Public Health Relevance

The goal of this project is to develop novel statistics methods to jointly model longitudinal biomarker trajectories and time to event data. The proposed methods are motivated and will be applied to three major applications: 1) liver transplant and kidney transplant ,2) the end stage renal disease (ESRD) data, and 3) the Vaginal birth after a prior cesarean (VBAC) data.

Agency
National Institute of Health (NIH)
Institute
National Institute of General Medical Sciences (NIGMS)
Type
Research Project (R01)
Project #
5R01GM104470-02
Application #
8708158
Study Section
Biostatistical Methods and Research Design Study Section (BMRD)
Program Officer
Marcus, Stephen
Project Start
2013-08-01
Project End
2017-07-31
Budget Start
2014-08-01
Budget End
2015-07-31
Support Year
2
Fiscal Year
2014
Total Cost
Indirect Cost
Name
University of Pennsylvania
Department
Biostatistics & Other Math Sci
Type
Schools of Medicine
DUNS #
City
Philadelphia
State
PA
Country
United States
Zip Code
19104
Yang, Wei; Xie, Dawei; Pan, Qiang et al. (2017) Joint Modeling of Repeated Measures and Competing Failure Events In a Study of Chronic Kidney Disease. Stat Biosci 9:504-524
Harhay, Michael O; Ratcliffe, Sarah J; Halpern, Scott D (2017) Measurement Error Due to Patient Flow in Estimates of Intensive Care Unit Length of Stay. Am J Epidemiol 186:1389-1395
Xu, Kelin; Guo, Wensheng; Xiong, Momiao et al. (2016) An estimating equation approach to dimension reduction for longitudinal data. Biometrika 103:189-203
Liu, Chengcheng; Ratcliffe, Sarah J; Guo, Wensheng (2015) A random pattern mixture model for ordinal outcomes with informative dropouts. Stat Med 34:2391-402
Liu, Ziyue; Guo, Wensheng (2015) Modeling diurnal hormone profiles by hierarchical state space models. Stat Med 34:3223-34
Elmi, Angelo; Ratcliffe, Sarah J; Guo, Wensheng (2014) The estimation of branching curves in the presence of subject-specific random effects. Stat Med 33:5166-76
Liu, Ziyue; Cappola, Anne R; Crofford, Leslie J et al. (2014) Modeling Bivariate Longitudinal Hormone Profiles by Hierarchical State Space Models. J Am Stat Assoc 109:108-118
Wierzbicki, Michael R; Guo, Li-Bing; Du, Qing-Tao et al. (2014) Sparse Semiparametric Nonlinear Model with Application to Chromatographic Fingerprints. J Am Stat Assoc 109:1339-1349