TITLE: Marginal Models in Analysis of Correlated Binary Data with TimeDependent Covariates in Biomedical Clinical Trials
INSTRUCTORS:Jeffrey R. Wilson, Arizona State University and Din Chen, UNC-Chapel Hill
MODERATOR:Walter R. Young
This tutorial is based on the textbook: “Marginal Models in Analysis of Correlated Binary Data with Time-Dependent Covariates” co-authored by Jeffrey R. Wilson, Elsa VazquezArreola, and (Din) Ding-Geng Chen, published by Springer in 2020, which uses R and SAS to conduct the computations. It provides a thorough presentation of correlated binary data with time-dependent covariates. It gives a detailed step-by-step illustration of their implementation using R and SAS. Longitudinal data contain correlated data due to the repeated measurements on the same subject. The changing values usually consist of time-dependent covariates and their association with the outcomes present different sources of correlation. Most methods used to analyze longitudinal data average the effects of time-dependent covariates on outcomes over time and provide a single regression coefficient per time-dependent covariate. Such approaches deny researchers the opportunity to follow the changing impact of time-dependent covariates on the outcomes. This tutorial addresses such issues through the use of partitioned regression coefficients. Examples of correlated data with time-dependent covariate include Cervical Dystonia Dataset where data are from a multicenter, randomized controlled trial of botulinum toxin type B (BotB) in patients with cervical dystonia from nine U.S. sites measured at baseline (week 0) and weeks 2, 4, 8, 12, 16 after treatment began and Midlife in the U.S. (MIDUS), a survey of adults age 25- 74 in 1994/95, with ongoing longitudinal follow-up since 2002. Data includes psychosocial, behavioral, sociodemographic, and health status characteristics, and other detailed data collection for subsamples including cognitive assessments, neuroscience data, and biomarkers.
Topics to be Covered:
1. Fundamentals of estimation of regression coefficients in cross-sectional data:
a. Ordinary least squares to obtain the regression coefficient estimates
b. Generalized Method of Moments estimates
2. Presenting data matrix for data with time-dependent covariates. Present the partitioned matrix.
3. Present correlated data with time-dependent covariates: Illustrate longitudinal data and the analysis using linear mixed models for continuous endpoints, generalized linear mixed model, and GEE for categorical endpoints.
4. Bayesian analysis in this partitioned data matrix using MCMC is applied
Dr. Jeffrey Wilson is a Professor of Statistics and Biostatistics at Arizona State University. Dr. Wilson’s research experience includes grants as PI and co-PI from the NIH, NSF, USDA, Arizona Department of Health Services, and the Arizona Disease Research Commission. He is presently the Statistics Associate Editor for The Journal of Minimally Invasive Gynecology and a former Chair of the Editorial Board of the American Journal of Public Health. He has published more than 85 articles in leading journals such as Statistics in Medicine, American Journal of Public Health, Journal of Royal Statistics Society, Computational Statistics, and Australian Journal of Statistics, among others. He has consulted with pharmaceutical companies and hospitals while representing them before the FDA and other federal government healthcare agencies. He has taught specialized Biostatistics classes at Mayo Clinic. He has led similar courses for Phoenix Children’s Hospital, Barrow Neurological Center, St. Joseph’s Hospital, and Banner Hospital. He is the former Director of the School of Health Management and Policy He is a former Director and co-Director of the Biostatistics Core in the NIH Center for Alzheimer at Arizona State University.
Dr. Din Chen is now the Wallace H. Kuralt distinguished professor in Biostatistics at the University of North Carolina-Chapel Hill. Before this, Dr. Chen was a professor in biostatistics at the University of Rochester Medical Center, the Karl E. Peace endowed eminent scholar chair and professor in biostatistics from the Jiann-Ping Hsu College of Public Health at the Georgia Southern University. Dr. Chen is an elected fellow of the American Statistical Association (ASA), an elected member of the International Statistics Institute (ISI) and a senior expert consultant for biopharmaceuticals and government agencies with extensive expertise in clinical trial biostatistics. Dr. Chen has more than 200 referred professional publications and co-authored/co-edited 25 books on biostatistics clinical trials, biopharmaceutical statistics, interval-censored survival data analysis, meta-analysis, public health statistics, statistical causal inferences; statistical methods in big-data sciences, and Monte-Carlo simulation-based statistical modeling. Dr. Chen has been invited nationally and internationally to speak and give short courses at various scientific conferences.He is a committee member of the Deming Conference and has been invited to give various tutorials at Deming Conference since 2011.
Both speakers have extensive experience in biostatistical and binary data while consulting to government and industries in SAS and R.