March 19-22, 2023

ENAR 2023 Educational Program | TUTORIALS

Tutorials are roughly 2 hours in length and focus on a particular topic or software package. The sessions are more interactive than a standard lecture, often encouraging participants active engagement and hands-on participation.


Monday, March 20 | 8:30 am – 10:15 am
T1 | Penalized regression with Bayesian Shrinkage Priors and Application to Real-World Data

Arinjita Bhattacharyya, Biostatistics and Research Decision Sciences, Merck & Co., Inc.; Department of Bioinformatics & Biostatistics, University of Louisville, KY
Riten Mitra, Department of Bioinformatics & Biostatistics, University of Louisville, KY

Course Description:

Prediction, classification algorithms, and count data models are commonly used in day-to-day problems, including clinical research, that involves selecting the most significantly associated features with binary, multinomial, count, or zero-inflated responses. These features can include several biological and demographic characteristics, such as genomic biomarkers and health history, essential for early disease detection, therapy selection, subgroup stratification, and controlled disease monitoring. Penalized regression with Bayesian shrinkage priors have been popular methods to address the curse of dimensionality in high-dimensional regression. We will present this Bayesian hierarchical framework that implements and compares benchmark methods for generalized linear models with application to real-life data in R. The short course aims to give participants hands-on experience, understand the techniques in a nutshell, and extrapolate self-implementation in the future.

Statistical/Programming Knowledge Required:
Some familiarity with R and Rstudio; and statistical models such as logistic regression, count models, and Bayesian methods are desired.

Dr. Arinjita Bhattacharyya is a Ph.D. graduate in Biostatistics from the University of Louisville, KY. She is currently a Senior Scientist at the Biostatistics and Research Decision Sciences (BARDS) in Merck & Co., Inc. She has completed her bachelor's and master's in Statistics from the University of Calcutta, and the University of Pune, India, respectively. She received the Graduate Dean's Citation Award and Doctoral Dissertation Completion Award and has internship experiences at Janssen Pharmaceuticals and the University of Michigan.

Arinjita currently supports Phase 1 clinical trials and is involved in designing clinical trials and writing study protocols and reports. Her daily life includes researching and implementing statistical methods in discovering and developing new therapeutic options for patients with unmet medical needs. Her primary research interests are clinical trials, Covid-19, HIV, mental health, oncology, shrinkage priors, prediction, methods for omics data, subgroup analysis, and dose-response models (Google Scholar). She loves to make new connections, mentor students, and is actively involved with ASA, JEDI, CAUSE, IISA, and ENAR CENS among others.

Dr. Riten Mitra is an Associate Professor at the Dept. of Bioinformatics & Biostatistics, University of Louisville, KY. He completed his undergraduate and master's from Indian Statistical Institute, Kolkata, and earned his Ph.D. in Biostatistics from the University of North Carolina at Chapel Hill in 2010. He was a postdoctoral fellow in the Biostatistics department at the MD Anderson Cancer Center from 2010-12 and at the University of Texas Austin from 2012-13. His research is primarily centered around Bayesian hierarchical models. Applications include biological networks, high throughput genomics, and high-dimensional clinical data. Other research focus includes Bayesian Graphical Models, NP-Bayes clustering algorithms, and NP-Bayes models for subgroup analysis and clinical trials.

Monday, March 20 | 10:30 am – 12:15 pm
T2 | Cell-type-aware Differential Analysis for Bulk Transcriptome Data

Hao Feng, Department of Population and Quantitative Health Sciences, Case Western Reserve University

Course Description:

The real-world clinical tissue samples are composed of diverse cell types. Recent statistical method advances in signal decomposition have enabled transcriptome studies at cell type resolution. In this tutorial, we will cover common software packages for cell type aware analysis in bulk transcriptome data. First, we will briefly discuss signal deconvolution, to introduce the basics in estimating pure cell type reference profiles and estimating cell type proportions. Next, we will mainly focus on widely adopted tools to conduct cell-type-specific Differential Expression Genes (csDEG) analysis. Novel methods and research progress in this domain will also be covered. This tutorial will have both the lecture component and the hands-on coding practice. We will provide R code implementation for methods introduced in this tutorial.

Statistical/Programming Knowledge Required:
Basic R programming.

Hao “Harry” Feng, is an Assistant Professor in the Department of Population and Quantitative Health Sciences at Case Western Reserve University. His research focuses on the development and the application of statistical bioinformatics methods to better understand high-throughput -omics data, especially in epigenomics. He proposed several novel statistical models in epigenomics data modeling, single-cell data analysis, cell-free DNA methylation and signal deconvolution. He developed a number of open-source software tools that are available on R-CRAN and Bioconductor, with > 24,000 downloads annually. He received his Ph.D. in Biostatistics and Bioinformatics from Emory University

Monday, March 20 | 1:45 pm – 3:30 pm
T3 | Data Monitoring Committee (DMC) Service 101

Heidi Spratt, University of Texas Medical Branch
Manisha Desai, Stanford University
Erinn Hade, NYU Grossman School of Medicine
Kathie Hartmann, Vanderbilt University Medical School
Chris Lindsell, Vanderbilt University Medical School

Course Description:

Clinical trials are considered the benchmark for concluding causality and are influential to clinical practice, but they carry inherent risk to participants. A common strategy to monitor the safety of participants and to ensure the integrity of the trial is to appoint a Data Monitoring Committee (DMC). While critical to the protection of human participants, few DMC members have ever received training.

This tutorial will cover expected duties for service on a DMC from a statistical perspective as well as guidance for how/when to weigh in on key aspects of a trial, from study design through to study completion. Additionally, we will present some information pertinent to what statistical concepts might be discussed and what is to be expected from those that serve on a DMC. This session will end with time for Q & A to allow participants to ask additional questions they might have about DMC service.

Statistical/Programming Knowledge Required:

Dr. Heidi Spratt is a tenured Associate Professor in the new Department of Biostatistics and Data Science at UTMB. She is a collaborative biostatistician with primary interests in bioinformatics, biomarker discovery, and machine learning. Her applied / collaborative research focuses on clinical studies applied to a wide range of disorders, diseases, and populations including those in the fields of infectious diseases (specifically dengue fever and Chagas disease), liver cancer, internal medicine and geriatrics, and pediatric respiratory infections. She has taught various courses at UTMB including an Introductory Biostatistics course aimed at basic science students, Categorical Data Analysis, Introduction of Bioinformatics, and Biomedical Informatics. Nationally, she is the President-Elect of the Association for Clinical and Translational Statisticians. She has served on 5 DSMB’s including 3 for the NIDDK as well as 2 industry sponsored trials.

Dr. Manisha Desai is Professor of Medicine and of Biomedical Data Science, and, by courtesy, of Epidemiology and Population Health at Stanford University. She is founding director of the Quantitative Sciences Unit (QSU), a collaborative group of about 30 faculty and staff who practice data science to answer critical biomedical questions. A major role of the QSU has been to serve as the data coordinating center for large multi-center trials. In this role, the QSU has been involved with coordinating data monitoring activities. Additionally, the QSU serves as the Independent Statistical Group for numerous Data Monitoring Committees (DMCs). Dr. Desai has enjoyed chairing and participating on a number of DMCs for a variety of clinical trials. She is passionate about training others in DMC activities.

Dr. Erinn Hade is an Associate Professor in the Division of Biostatistics, Department of Population Health in the NYU Grossman School of Medicine. Her research interests are centered on the design and inference of observational and randomized studies focused on evaluating the effectiveness of interventions and implementation strategies. This work has been motivate through collaborations in reproductive, perinatal and pediatric medicine, maternal health and traumatic brain injury. She has served as the trial statistician reporting to the Data and Safety Monitoring Board (DSMB) for numerous NIH or PCORI trials, currently serves as the biostatistician on the NYU CTSI DSMB, and has participated as a DSMB member for NIH trials.

Dr. Katherine Hartmann is Vice President for Research Integration, charged with creating collaborative synergy and new initiatives in the research enterprise, and Associate Dean for Clinical Translational Scientist Development at Vanderbilt. She is among the few obstetrician-gynecologists in the US who have a PhD in epidemiology and a research dominant career. She began her research career as a clinical trialist focused on behavioral interventions and has served on DSMBs nearly continuously since 2002. This includes 16 years of service to the NIH Contraceptive Research Network, and 14 years as chair of a multi-site NIH research network that has evolved a strong and effective relationship with the network DSMB leading to mutually reviewed tools to enhance standardization of protocols, facilitate timely and consistent adverse event reporting, ensure consistency of consenting participants, conduct communication with participants intercurrent in trials, and guide IRB modifications among other topics.

Dr. Chris Lindsell is director of the Vanderbilt Institute for Clinical and Translational Research Methods program, co-Director of the Center for Health Data Science, and professor of biostatistics and biomedical informatics at Vanderbilt University Medical Center. His research portfolio is focused on health systems and services, biomarker discovery and validation, and clinical trials in acute care environments. He has published over 300 peer-reviewed papers and has led the data coordinating center for numerous multi-center clinical trials and epidemiologic studies including ACTIV6, the IVY Network and CONNECTS, NHLBI’s network of networks for COVID-19 research. He holds patents for using clinical information, biomarkers and transcriptomics for prognosis and prediction in sepsis and septic shock and is working towards precision trials to improve the care of critically ill patients. He helps co-lead Vanderbilt University Medical Center’s Learning Health System where he supports the design and execution of pragmatic effectiveness trials.

Monday, March 20 | 3:45 pm – 5:30 pm
T4 | Obuchowski-Rockette-Hillis Methodology for Analysis of Radiologic Diagnostic Imaging Studies

Stephen L. Hillis, Departments of Radiology and Biostatistics, University of Iowa

Course Description:

In radiologic diagnostic imaging studies, the goal is typically to compare the performance of readers (usually radiologists) for two or more tests or modalities (e.g., digital versus film mammograms, CT versus MRI). Typical reader-performance measures are functions of the ROC curve, such as the area under the ROC curve. Because the typical reader performance measures are not indexed by case, a conventional linear mixed model with reader and patient treated as random effects cannot be used to account for reader and patient. Presently, the standard analysis approach is the Obuchowski-Rockette-Hillis (ORH) method.

In this tutorial, I will discuss the ORH method in detail, provide illustrative real-data examples and discuss software for implementing it, including the new R package MRMCaov currently being developed at the University of Iowa. Participants should have a basic knowledge of linear mixed models.

Statistical/Programming Knowledge Required:
Participants should have a basic knowledge of linear mixed models.

Stephen L. Hillis is a research professor in the Departments of Radiology and Biostatistics at the University of Iowa. He received his PhD in statistics in 1987 and his MFA degree in music 1978, both from the University of Iowa. He is the author of more than 100 peer-reviewed journal articles and four book chapters. Since 1998, his research has focused on methodology for multi-reader diagnostic radiologic imaging studies, with an emphasis on developing the analysis method first proposed by Obuchowski and Rockette.

Tuesday, March 21 | 1:45 pm – 3:30 pm
T5 | Statisticians and the Media: Find Your Voice, Build Your Network

Natalie Dean, Emory University

Course Description:

In this tutorial, we will discuss strategies for engaging with social and traditional media. Scientists are flocking to sites like Twitter to access the latest research findings and to network with colleagues. In addition to being a place to learn, scientists can share content and develop their individual voice. As one’s following grows, a scientist can reach colleagues in other fields, and even journalists, policy-makers, and the general public. In the first half of the tutorial, we will discuss the potential value of social media, developing your content, and navigating pitfalls. In the second half of the tutorial, we will talk through the logistics of engaging with traditional media outlets, including doing live or pre-recorded interviews, corresponding with journalists, and writing editorials. I will share lessons learned from my own experiences during the COVID-19 pandemic. Come ready to imagine the type of content you want to share with your community!

Statistical/Programming Knowledge Required:

Dr. Natalie Dean is an Assistant Professor in the Department of Biostatistics and Bioinformatics at Emory’s Rollins School of Public Health. She received her PhD in Biostatistics from Harvard University, and previously worked as a consultant for the WHO’s HIV Department and as faculty at the University of Florida. Her primary research area is infectious disease epidemiology, with a focus on innovative study designs for evaluating vaccines during public health emergencies. During the COVID-19 pandemic, she has been active in public engagement, with authored pieces

Tuesday, March 21 | 3:45 pm – 5:30 pm
T6 | A Primer for Meta-analysis with Real-world Applications

Usha Govindarajulu, Icahn School of Medicine at Mount Sinai

Course Description:

Stemming from some of our initial papers in which we explored and discussed smoothing methods in relation to their use in epidemiological studies and in survival analysis, in this tutorial, I will discuss these various smoothing methods and their potential for other applications. I will especially plan to focus on various spline methods like cubic splines (restricted, natural) and penalized splines (with B-spline basis functions) that we used before and I will also cover newer extensions of these methods, like constrained splines as well as splines that can also work in a generalized additive framework. The focus will be to discuss the theory of these methods but also applications of these methods in existing datasets. Various R packages for handling these methods will be utilized during the tutorial so the attendees should have moderate working knowledge of R.

Statistical/Programming Knowledge Required:
Moderate working knowledge of R

Dr. Usha Govindarajulu is currently an Associate Professor in the Center for Biostatistics of the Department of Population Health Science and Policy at the Icahn School of Medicine at Mount Sinai. She obtained her PhD in Biostatistics from Boston University and post-doctoral work in Biostatistics from the Harvard School of Public Health. She teaches longitudinal data analysis and survival analysis and her major statistical interests are in survival analysis, frailty models, causal inference, smoothing, machine learning, and non-parametrics.