Sections

Section Co-Chiefs: Professors J. Jack Lee and Ying Yuan

Goals   

• Foster a collaborative environment for statistical research on adaptive clinical trial designs

• Promote and support the implementation of novel designs in clinical trials

• Provide education and knowledge transfer of adaptive designs through  seminars, webinars and software development

Participation:  Faculty, statistical analysts, programmers, postdoctoral researchers and students affiliated with the Department of Biostatistics may participate in the activity of the section.

Implementation Plan

•  Develop a software portal that integrates tools for novel clinical trial designs, including tools for trial conduct and data analysis

•  Establish a core team to provide support to clinicians who are interested in applying novel designs in clinical trials

•  Organize seminars, webinars and discussion sessions for disseminating information and exchanging new ideas regarding the design and conduct of novel clinical   trials

Section Chief: Professor Sanjay Shete

Major foci of the Section of Behavioral and Social Statistics include developing and implementing:

• Risk prediction models: to accurately identify individuals with greater likelihood of using risky behavior associated with adverse health consequences, including cancer development. An example is a model to predict youth who are likely to engage in smoking experimentation and for whom early intervention is showing or has shown promise.

• Methods to evaluate interventions: Time-varying SMART designs and data analysis methods for evaluating adaptive interventions to decrease or prevent behaviors that adversely affect health.

• Statistical approaches to determine how much of the total risk associated with a particular behavior (such as initiating tobacco use) is attributable to genetic factors and non-genetic factors.

•  Computational and quantitative methods for analyzing family and population-based data to identify causal genetic and non-genetic risk cancer factors: (i) robust methods for performing linkage analysis between a quantitative trait and a genetic marker; (ii) identifying imprinted loci; (iii) tests for genetic-association studies, such as using evidence about deviation from Hardy-Weinberg proportion in case subjects; (iv) efficient methods for analyzing secondary phenotypes; (v) variable selection approaches; and (vi) linkage disequilibrium-based approaches to identify rare variants.

• Statistical approaches for identifying complex relationships among risk factors and diseases: multiple mediator models for case–control data to identify complex relationships among genetic variants, behavior that will likely compromise one’s health, the development of a chronic disease associated with that behavior, and the development of cancer. An example is the relationships among genetic variants, smoking cigarettes or using other tobacco sources, developing chronic obstructive pulmonary disease, and developing lung cancer.

• Mediation models: Latent class dynamic mediation modeling to determine the benefits achieved by health behavior interventions. This relatively new approach can explicitly account for the possibility of different subgroups within the study population and the mediation effect varying over time.

• Software to implement these statistical methods; dissemination of free downloadable software.

This work often involves collaborations with researchers in the Departments of Behavioral Science, Health Disparities Research, and Health Services Research. In addition to Sanjay Shete, Ph.D., Biostatistics faculty researchers contributing to behavioral and social statistics include Yisheng Li, Ph.D., and Jian Wang, Ph.D.   

Section Co-Chiefs: Professors Ying Yuan, Liang Li, and Yu Shen

As a hospital-centered institution, MD Anderson does not have ready access to populations without cancer who are yet at high risk of developing cancer. Therefore, it is essential to establish state-level or national-level statistical and data coordinating centers through participation in research consortia.

Through this initiative to participate in research consortia, the Section of Early Cancer Detection and Biomarkers within the Department of Biostatistics works closely with Samir Hanash, M.D., Ph.D., Director of the Red and Charline McCombs Institute for the Early Detection and Treatment of Cancer and the proteomics platform leader for the Moon Shots Program™.

This section includes the Early Cancer Detection Data & Statistical Coordinating Center, which works synergistically with large multi-center cancer early detection studies to provide support for grant applications, statistical study designs, and protocol and object-oriented multi-user domain development. The center also assists with IRB approvals; provides a database/information management system for a study; provides site training and monitoring, specimen tracking, quality control and quality assessment, and data analysis; interprets study findings with the principal investigators; and provides long-term data and document archiving. MD Anderson’s Executive Director of Research Information Systems & Technology Services participates with this center.

The center currently supports two additional programs/research efforts:

• The Consortium to Study Chronic Pancreatitis, Diabetes, and Pancreatic Cancer  (PI: Anirban Maitra, Professor of Pathology and Translational Molecular Pathology and Scientific Director of the Sheikh Ahmed Pancreatic Cancer Research Center at MD Anderson). This consortium is funded through August 2020 by a U01 mechanism supported by the National Institute of Diabetes and Digestive and Kidney Diseases and the National Cancer Institute.

• The Texas Hepatocellular Carcinoma Consortium (PI: Hashem El-Serag, Professor of Gastroenterology at Baylor College of Medicine in Houston). This consortium is funded through a multi-investigator research award of the Cancer Prevention Research Institute of Texas (CPRIT). There are 5 projects funded by CPRIT for the Texas Hepatocellular Carcinoma Consortium. 

In addition to Ying Yuan, Liang Li, and Yu Shen, faculty in the Department of Biostatistics who contribute to early cancer detection and biomarker research include Kim-Anh Do, Ph.D., and Jing Ning, Ph.D.