Proposal Description

This session explores the cutting-edge statistical methods for causal inference within complex data landscapes, an area of paramount importance to biomedical research. As large health databases burgeon, offering a rich soil for causal research, the intricacy of these datasets presents unique challenges. These complexities range from multifaced treatment regimens, where initiation may hinge on specific biomarkers, to the undetermined optimal timing for cardiovascular interventions, demanding insights from comprehensive patient data. Additionally, this session delves into the quandaries of drawing reliable causal conclusions from datasets marred by missing information and seeks to unveil how these advanced methodologies can personalize our understanding of disease progression.
Featuring an illustrious and diverse group of speakers, this session promises to illuminate the forefront of methodological innovations. These include joint mixed membership models that intertwine multivariate longitudinal and survival data, flexible Bayesian approaches to addressing time-varying confounding for causal inference, and novel Bayesian nonparametric techniques adept at navigating complex longitudinal treatments obscured by missing data.
The proposed session will be comprised of four speakers. Each speaker will have 25 minutes to present his/her work. The presentation will be followed by a 10-minute open discussion from the floor.

Session Speakers: The proposed speakers are as follows:

Xinyuan Song, Professor, Departement of Statistics, The Chinese University of HongKong
Email: xysong@sta.cuhk.edu.hk
Presentation title: “Joint mixed membership modeling of multivariate longitudinal and survival data for learning the individualized disease progression”

Liangyuan Hu, Associate Professor, Department of Biostatistics and Epidemiology, Rutgers University
Email: liangyuan.hu@rutgers.edu
Presentation title: “A flexible Bayesian approach for causal inference with time-varying confounding and survival outcomes”

Arman Oganisian, Assistant Professor, Department of Biostatistics, Brown University
Email: arman_oganisian@brown.edu
Presentation title: “Bayesian Nonparametric Inference for Optimal Treatment Strategies with Missing Covariates: Applications in Acute Myeloid Leukemia”

Michael J Daniels, Professor, Department of Statistics, University of Florida
Email: daniels@ufl.edu
Presentation title: “A Bayesian Nonparametric Approach to an HIV Assessment Survey with Missing Data and Skip Conditions”

Session Organizer: The proposed session is organized by Liangyuan Hu (liangyuan.hu@rutgers.edu), Associate Professor, Department of Biostatistics and Epidemiology, Rutgers School of Public Health.

Session Chair: The proposed session will be chaired by Qixuan Chen (qc2138@cumc.columbia.edu), Associate Professor, Department of Biostatistics, Columbia University