All Times EDT
With two years since the release of draft ICH E9 Addendum (R1), scientific community and industry have made great progress in further advancing this topic and resolving various issues with implementing the guidance in clinical practice. The COVID-19 pandemic caused a dramatic increase in the number of intercurrent events (ICE), thus being a natural stress test for proper handling of ICEs in study protocols and statistical analysis plans. This short course will cover our learning from the implementation of the addendum during the pandemic. Specially, the outline of the short course is:
• The key elements and concepts of the ICH E9 (R1) Addendum • Defining estimands in presence of ICEs using potential outcomes based on causal inference framework • Handling ICEs using a mix of strategies depending on the type and ICE, emphasizing improved practices for collecting the reasons for treatment discontinuations (a major type of ICE) • Overview of statistical methods for handling missing values and guidance on using appropriate methods tailored to the estimand(s) of interest. • Overview of the principal stratification methods in the context of estimands incorporating ICEs (if time permits)
Yongming Qu, PhD, Sr. Research Fellow at Eli Lilly and Company. Dr. Qu obtained his PhD in statistics from Iowa State University in 2002. He has had rich experience in supporting and leading phase 1-4 clinical development in pharmaceutical industry. He is an ASA Fellow and has been an active researcher in improving statistical methods in drug development with approximately >80 articles published in statistical and medical journals. Recently, he has published several articles regarding estimands and missing data imputation and is a technical leader in driving the implementing the ICH E9 (R1) in clinical studies in Eli Lilly and Company. He has recently given invited presentations in estimands and missing data in ESFPI Webinar (2020), ENAR webinar (2020), BASS invited presentation (2020), China DIA (2020), PSI (2021), DIA webinar (2021), and RISW invited sessions (2019, 2021).
Ilya Lipkovich, PhD, Sr. Research Advisor at Eli Lilly and Company. Dr. Lipkovich received his Ph.D. in Statistics from Virginia Tech in 2002 and has 20 years of statistical consulting experience in pharmaceutical industry. He is an ASA Fellow and published on subgroup identification in clinical data, analysis with missing data, and causal inference in statistical and medical journals including > 50 articles and a book “Analyzing Longitudinal Clinical Trial Data. A Practical Guide.” He frequently taught short courses and webinars on these topics. Recently, he has published several articles connecting estimands with missing data and causal inference and co-authored a book “Estimands, Estimators and Sensitivity Analysis in Clinical Trials.”
Selected publications co-authored by instructors that are relevant to the short course:
Lipkovich, I., Ratitch, B., & Mallinckrodt, C. H. (2020). Causal inference and estimands in clinical trials. Statistics in Biopharmaceutical Research, 12(1), 54-67.
Luo, J., Ruberg, S. J., & Qu, Y. (2021). Estimating the treatment effect for adherers using multiple imputation. arXiv preprint arXiv:2102.03499. Pharmaceutical Statistics. In press.
Mallinckrodt, C.H., Bell, J. Liu, G., Ratitch, B., O'Kelly, M., Lipkovich, I., Singh, P., Xu. L., Molenberghs, G. (2020). Aligning estimators with estimands in clinical trials: Putting the ICH E9(R1) Guidelines into practice. Therapeutic innovation & regulatory science, 54(2),353-364.
Mallinckrodt, C., Molenberghs, G., Lipkovich, I., and Ratitch, B. (2020), Estimands, Estimators and Sensitivity Analysis in Clinical Trials, Chapman & Hall/CRC Biostatistics Series, Boca Raton, FL: Chapman & Hall/CRC Press.
Mallinckrodt, C.H., Lin, Q., Lipkovich, I., Molenberghs, G. (2012). A structured approach to choosing estimands and estimators in longitudinal clinical trials. Pharmaceutical Statistics,11, 456-461.
Qu, Y., & Dai, B. (2021). Return-to-baseline multiple imputation for missing values in clinical trials. arXiv preprint arXiv:2111.09423.
Qu, Y., & Lipkovich, I. (2021). Implementation of ICH E9 (R1): A few points learned during the COVID-19 pandemic. Therapeutic Innovation & Regulatory Science 55, 984–988.
Qu, Y., Luo, J., & Ruberg, S. J. (2021). Implementation of tripartite estimands using adherence causal estimators under the causal inference framework. Pharmaceutical Statistics, 20(1), 55-67.
Qu, Y., Shurzinske, L., & Sethuraman, S. (2021). Defining estimands using a mix of strategies to handle intercurrent events in clinical trials. Pharmaceutical Statistics, 20(2), 314-323.
Ratitch, B., Bell, J., Mallinckrodt, C., Bartlett, J. W., Goel, N., Molenberghs, G., ... & Lipkovich, I. (2020). Choosing estimands in clinical trials: putting the ICH E9 (R1) into practice. Therapeutic innovation & regulatory science, 54(2), 324-341.
Ratitch, B., Goel, N., Mallinckrodt, C., Bell, J., Bartlett, J.W., Molenberghs, G., Singh, P., Lipkovich, I. and O’Kelly, M. (2020). Defining efficacy estimands in clinical trials: examples illustrating ICH E9 (R1) guidelines. Therapeutic innovation & regulatory science, 54(2), 370-384.
Wiens, B.L., & Lipkovich, I. (2020). The impact of major events on ongoing noninferiority trials, with application to COVID-19. Statistics in Biopharmaceutical Research, 12(4), 443–450.
Zhang, Y., Fu, H., Ruberg, S. J., & Qu, Y. (2021). Statistical inference on the estimators of the adherer average causal effect. Statistics in Biopharmaceutical Research, 1-4.
As a part of the isolation to address the complexity, the depth, and the broadness of the drug development, the 21st-century cures act mandates the FDA to explore the use of novel designs, including Bayesian design, in the drug development. To satisfy a mandate of the Cures Act on the use of novel clinical trial designs including the Bayesian design, the FDA published a guidance on adaptive design “FDA. Guidance for Industry: Adaptive Designs for Clinical Trials of Drugs and Biologics. 2019.” and “Interacting with the FDA on Complex Innovative Trial Designs for Drugs and Biological Products” in 2020. To date, Bayesian analyses have been used in phase I and phase II studies. It has also in late pahse trials, e.g. to help to provide substantial evidence and needed confidence for the approval of Belimumab.
In this course, Dr. Yuan and Dr. Travis will overview Bayesian methods that are utilized to leverage external data into the design and analysis of a new study of interest. External data broadly include historical data, natural history data, data from similar populations, data from drug from the same drug class. The Bayesian framework provides a flexible way to integrate external data to improve inference for the study of interest, potentially addressing many practical issues in clinical trials, such as lack of pediatric patients and patients with rare diseases, high cost and lack of efficiency.
In part I, Dr. Yuan will overview static and dynamic borrowing methods including Bayesian hierarchical model (BHM); power prior, commensurate prior, robust meta-analytic predictive prior developed between 2001-2011 and multisource exchangeability model, calibrated BHM, optimal BHM, and elastic prior developed between 2011-2021. The objective of these methods is to encourage information borrowing when historical and trial data are “similar” and refrain from information borrowing when historical and trial data are “different”. The proc and cons, as well as the connections, of the methods will be elucidated.
In part 2, Dr. Travis will discuss potential application in clinical trials from regulatory perspective and case-study example (or mock example) to demonstrate the use of these methods. Mock R code will be provided during the case-study illustration.
During the training, participants will be able to learn (1) general ideas of Bayesian approaches to borrow information from external data and assumptions these approaches make; (2) understand the difference between static and dynamic borrowing, as well as their potential applications; (3) understand mechanism of control of borrowing strength and the links between the methods. (4) Apply these methods to real examples with hand-on training on programming.
Instructors’ background:
Ying Yuan, PhD, is a Bettyann Asche Murray Distinguished Professor and Deputy Chair in the Department of Biostatistics at the University of Texas MD Anderson Cancer Center. Dr. Yuan has published over 100 statistical methodology papers on innovative Bayesian adaptive designs, including early phase trials, seamless trials, biomarker-guided trials, and basket and platform trials. The designs and software developed by Dr. Yuan’s lab have been widely used in medical research institutes and pharmaceutical companies. Dr. Yuan is the Chair of Data Safety Monitoring Board (DSMB) of MD Anderson Cancer Center, was elected as the American Statistical Association Fellow.
James Travis is a lead mathematical statistician supporting the Division of Pediatric and Maternal Health in the Center for Drug Evaluation and Research. He is a member of the Complex and Innovative Trial Designs Program steering committee and the Office of Biostatistics Bayesian and Pediatrics committees. He has provided many training sessions on Bayesian approaches to FDA colleagues. He has research interests in the use of external data in the analysis of pediatric clinical trials. He joined the FDA in 2014 and received his PhD from the University of Maryland Baltimore County.
In May 2021, the U.S. Food and Drug Administration (FDA) released a revised draft guidance for industry on “Adjustment for Covariates in Randomized Clinical Trials for Drugs and Biological Products”. Covariate adjustment is a statistical analysis method for improving precision and power in clinical trials by adjusting for pre-specified, prognostic baseline variables. Here, the term “covariates” refers to baseline variables, that is, variables that are measured before randomization such as age, gender, BMI, comorbidities. The resulting sample size reductions can lead to substantial cost savings, and also can lead to more ethical trials since they avoid exposing more participants than necessary to experimental treatments. Though covariate adjustment is recommended by the U.S. Food and Drug Administration and the European Medicines Agency (EMA), many trials do not exploit the available information in baseline variables or only make use of the baseline measurement of the outcome.
In Part 1 of the workshop, we introduce the concept of covariate adjustment. In particular, we explain what covariate adjustment is, how it works, when it may be useful to apply, and how to implement it (in a preplanned way that is robust to model misspecification) for a variety of scenarios. We demonstrate the impact of covariate adjustment using completed trial data sets in multiple disease areas. We provide step-by-step, clear documentation of how to apply the software in each setting. Participants will have the time to apply the software tools on the different datasets in small groups.
In Part 2 of the workshop, we present a new statistical method that enables us to easily combine covariate adjustment with group sequential designs. The result will be faster, more efficient trials for many disease areas, without sacrificing validity or power. This approach can lead to faster trials even when the experimental treatment is ineffective; this may be more ethical in settings where it is desirable to stop as early as possible to avoid unnecessary exposure to side effects. The new statistical method and software will be demonstrated using the same real datasets as in the first part. We will provide step-by-step, clear documentation of how to apply the software in these different settings. Participants will have the time to apply the new software tools on the different datasets in small groups.
Course outline:
40 minutes: Introduction and overview of covariate adjustment (Michael Rosenblum)
10 minutes: Discussion (Q&A time)
5 minutes: Break
20 minutes: Software tools demonstration on covariate adjustment (Michael Rosenblum)
25 minutes: Small group work applying software tools on covariate adjustment (Michael Rosenblum and Kelly Van Lancker)
10 minutes: Break
30 minutes: Presentation of new statistical method to combine covariate adjustment and group sequential designs (Kelly Van Lancker)
10 minutes: Discussion (Q&A time)
15 minutes: Software tools demonstration (Joshua Betz)
5 minutes: Break
30 minutes: Small group work applying software tool on group sequential designs (Michael Rosenblum, Kelly Van Lancker and Joshua Betz)
10 minutes: Discussion (Q&A time)
Target Audience:
The intended audience consists of clinicians and statisticians with statistical training. Participants should be familiar with the following concepts: Type I error, power, bias and variance.
Course Objectives:
1. Participants will learn about the benefits and limitations of using covariate adjustment to analyze data from randomized trials, and how it can be applied to improve precision and speed up trials.
2. Participants will learn key concepts from the recent (May 2021) draft guidance from the FDA on covariate adjustment
3. Participants will gain experience implementing covariate adjustment on simulated data sets.
Instructors Background:
Michael Rosenblum is a Professor of Biostatistics at Johns Hopkins Bloomberg School of Public Health. His research is in causal inference with a focus on developing new statistical methods and software for the design and analysis of randomized trials, with clinical applications in HIV, Alzheimer’s disease, stroke, and cardiac resynchronization devices. He is funded by the Johns Hopkins Center for Excellence in Regulatory Science and Innovation for the project: “Statistical methods to improve precision and reduce the required sample size in many phase 2 and 3 clinical trials, including COVID-19 trials, by covariate adjustment”.
Dr. Kelly Van Lancker is a postdoctoral researcher in the Biostatistics Department of the Johns Hopkins Bloomberg School of Public Health. She has obtained a PhD in statistics from Ghent University in Belgium. Her primary research interests are the use of causal inference methods in clinical trials and obtaining valid inference when the analysis involves data-adaptive methods, such as variable selection.
Josh Betz is an Assistant Scientist in the Biostatistics department of the Johns Hopkins Bloomberg School of Public Health, and part of the Johns Hopkins Biostatistics Center. His research includes the design, monitoring, and analysis of randomized trials in practice and developing software to assist with randomized trial design.
Relevant Papers:
Wang, B., Susukida, R., Mojtabai, R., Amin-Esmaeili, M., and Rosenblum, M. (2021) Model-Robust Inference for Clinical Trials that Improve Precision by Stratified Randomization and Adjustment for Additional Baseline Variables. Journal of the American Statistical Association, Theory and Methods Section. https://www.tandfonline.com/doi/full/10.1080/01621459.2021.1981338
Benkeser, D., Diaz, I., Luedtke, A., Segal, J., Scharfstein, D., and Rosenblum, M. (2020) Improving Precision and Power in Randomized Trials for COVID-19 Treatments Using Covariate Adjustment, for Binary, Ordinal, or Time to Event Outcomes. Biometrics. This paper was selected to be a discussion paper. https://doi.org/10.1111/biom.13377
Van Lancker, K., Vandebosch, A., & Vansteelandt, S. (2020). Improving interim decisions in randomized trials by exploiting information on short-term endpoints and prognostic baseline covariates. Pharmaceutical statistics, 19(5), 583-601. https://doi.org/10.1002/pst.2014
COVID-19 pandemic has ignited a world-wide broad interest in development of vaccines. Because of biological nature, there are many unique statistical issues and challenges in design and analyze vaccine clinical trials. Some examples include using immunogenicity to evaluate vaccine effects and consistency in manufacture; the rigorous large studies needed to demonstrate efficacy due to low incidence rate of disease; identifying and using biomarkers based on correlates of protection; stringent safety requirement because of broad administration to millions of healthy individuals; and application of innovative designs to speed up the development.
This half-day short course will provide an overview of study designs and analysis methods for vaccine clinical trials. Following some general introduction of vaccine development, the course will cover topics for statistical methods in analysis of immunogenicity, efficacy, and safety. Unique features for vaccine trials such as non-inferiority design, lot consistency, correlate of protection, super superiority study, and handling of low incidence events, etc. will be discussed. Case examples from various vaccine programs will be presented.
Course outlines: 1) Introduction of vaccine development and design considerations; 2) Statistical methods for immunogenicity including non-inferiority comparison, handling of missing data, issues for multi-valent vaccines, and lot consistency; 3) Statistical method for efficacy including conditional exact methods, adaptive dose range and seamless designs, and Bayesian methods; 4) Correlate of protection, modeling efficacy from immunogenicity markers; 5) Evaluation of safety.
Instructors’ background:
Dr. Wenji Pu is Statistics Director at GlaxoSmithKline (GSK) plc and has more than 18 years of experience in designing and analyzing clinical trials. At GSK, Wenji has worked on many vaccine programs including respiratory syncytial virus (RSV) vaccine, herpes zoster vaccine, and flu vaccine, and has been the resource for providing statistical expertise within GSK vaccine statistics group. His research interest includes repeated measurements, categorical data analyses, survival analyses, missing data, and Bayesian adaptive design.
Abstract: The use of open-source R is evolving in drug discovery, research, and development for study design, data analysis, visualization, and report generation in the pharmaceutical industry. In this workshop, we will explore strategies to use R to prepare tables, listings, and figures in a clinical study report and how to prepare the eCTD submission packages for those TLFs and associated source code. The workshop will have three parts.
Part 1, Delivering TLFs in CSR: provides general information with examples to create tables, listings, and figures.
Part 2, Clinical Trial Project: provides general information with examples to manage a clinical trial A&R project.
Part 3, eCTD Submission package: provides general information in preparing a submission package related to clinical study report (CSR) in electronic Common Technical Document (eCTD).
The training is based on an open-source book “R for Clinical Study Reports and Submission” available at https://r4csr.org/ with a demo project https://github.com/elong0527/esubdemo.
Prerequisite: It is an intermediate level training. We assume people are familiar with data manipulation in R. Some good references include Hands-On Programming with R (https://rstudio-education.github.io/hopr/) and R for Data Science (https://r4ds.had.co.nz/).
The International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) issued an ICH Harmonized Guideline: ICH E9 (R1) Addendum. The US Food and Drug Administration adopted this as guidance in May 2021. The ICH E9 (R1) Addendum introduces the estimand framework for clinical trials to obtain clear and interpretable treatment effects, which enables clear assessment of benefits and risks of treatments. The estimand framework is intended to facilitate the dialogue on drug/biologic development among review disciplines, as well as between Sponsor and Regulator. This course introduces the estimand framework to statisticians, provides tools to specify clinical questions of interest precisely and facilitate cross-disciplinary collaboration, and highlights key concepts with illustrative examples. Multiple opportunities for questions are included throughout the course as well as an interactive practice session for the audience.
Target audience: Statisticians
Objectives: At the end of this training the attendees will be able to: • Understand fundamentals of the estimand framework. • Identify relevant discussion topics and methods for successful cross-disciplinary collaboration, including precise specification of clinical questions of interest. • Recognize and apply important estimand considerations.
Prerequisites: None
Outline • Module 1: Overview of the Estimand Framework (Alexei C. Ionan) • Module 2: Key Considerations s (John Scott) • Module 3: Real Example, Including Productive Cross-Disciplinary Interactions (Susan Mayo and Miya Paterniti)
Relevant Materials
ICH E9 (R1) Addendum on Estimands and Sensitivity Analysis in Clinical Trials to the Guideline on Statistical Principles for Clinical Trials, Step 4, 20 November 2019 https://database.ich.org/sites/default/files/E9-R1_Step4_Guideline_2019_1203.pdf
Instructor Background
Alexei C. Ionan is a Mathematical Statistician in the Division of Biometrics IX of the Office of Biostatistics, supporting application review in the Office of Oncologic Diseases at the FDA. He has over 17 years of experience evaluating, developing, and applying statistical methods in oncology. He leads multiple groups at the FDA. His research interests include Bayesian methods, estimands, decision theory, causal inference, predictive biomarkers, early detection of cancer, and optimal design.
John Scott is Director of the Division of Biostatistics in the FDA's Center for Biologics Evaluation and Research, where he has also served as Deputy Director and as a statistical reviewer for blood products and for cellular, tissue and gene therapies. Prior to joining the FDA in 2008, he worked in psychiatric clinical trials at the Western Psychiatric Institute and Clinic of the University of Pittsburgh Medical Center. He was one of FDA's representatives on the ICH E9(R1) expert working group. Dr. Scott is also the CBER lead for 21st Century Cures and PDUFA VI efforts in Complex and Innovative Trial Design and has been heavily involved in a number of FDA's statistical policy and outreach projects, including the 2019 Adaptive Design Guidance for Drugs and Biologics, the 2020 Guidance on Interacting with the FDA on Complex Innovative Trial Design, and the ICH E20 expert working group on adaptive designs. Dr. Scott has taught numerous internal and external short courses on topics including benefit-risk assessment, multiple endpoints, adaptive clinical trial design, and Bayesian analysis. Dr. Scott holds a Ph.D. in Biostatistics from the University of Pittsburgh, an A.M. in Mathematics from Washington University in St. Louis, and a B.A. in Liberal Arts from Sarah Lawrence College. He is a Fellow of the American Statistical Association and is a past Editor of the journal Pharmaceutical Statistics.
Miya Paterniti is a clinical team leader in the Division of Pulmonology, Allergy, and Critical Care within the Office of New Drugs in the FDA’s Center for Drug Evaluation and Research. She received her M.D. at the University of Maryland and completed her internal medicine residency and fellowship in Allergy and Clinical Immunology at The Johns Hopkins School of Medicine. She is also an Assistant Professor at The Johns Hopkins School of Medicine and a practicing allergist. She has been working on estimands for several years and has presented at multiple conferences on estimands.
Susan Mayo is a senior biostatistical reviewer in Division III, Office of Biostatistics within the Office of New Drugs in the FDA’s Center for Drug Evaluation and Research. She received M.S. degrees from Louisiana State in Applied Statistics and Marine Sciences and began her working career in 1986. She has served as a statistician in small to mid-sized biotechnology and drug delivery companies, consulted for several years with small device companies, and served for many years in large pharmaceutical companies. She joined Office of Biostatistics in 2018 and serves with her colleagues in the Division of Pulmonology, Allergy, and Critical Care. Her interests include estimands, safety statistics, statistical graphics, benefit-risk assessment, and the study of change, used in implementation of these important concepts for benefiting drug development and ultimately, public health.
A Bayesian approach provides the formal framework to incorporate external information into the statistical analysis of a clinical trial. There is an intrinsic interest of leveraging all available information for an efficient design and analysis. This allows trials with smaller sample size or with unequal randomization. Examples include early phases drug development, occasionally in phase III trial, and special areas such as medical devices, orphan indications and extrapolation in pediatric studies. Recently, 21st Century Cure Act and PUDUFA VI encourage the use of relevant historical data for efficient design. An appropriate statistical method in this context needs to leverage “borrowing” of information while considering the heterogeneity between historical and current trial. In this short course, we'll cover different statistical frameworks to incorporate trial external evidence with real life example.
We begin with introducing the meta-analytic predictive (MAP) framework for borrowing historical data. The MAP approach is based on Bayesian hierarchical model which combines the evidence from different sources. It provides a prediction for the current study based on the available information while accounting for inherent heterogeneity in the data. This approach can be used widely in different applications of clinical trial. These applications will be demonstrated using the R package RBesT, the R Bayesian evidence synthesis tools, which are freely available from CRAN.
In the second part of the short course, we focus on the propensity score integrated power prior approach. The power prior is a useful class of informative priors for external control data. The power prior discounts the likelihood of the external control data directly using a power parameter. However, choice of the power parameter is tricky in the real-life applications. An integrated propensity score-based method along with prior power can be useful in this context. A two-stage provides a paradigm for conducting a comparative observational, non-randomized study within the premarket regulatory setting. The power parameters are calculated using trial data and external control divided into homogeneous strata using propensity score. Methodological and practical aspects will be discussed to facilitate real life implementation.
Outline I. Introduction: Motivation and general framework
II. Overview of available methods
III. Meta-analytic Predictive (MAP) Prior
a. MAP approach for the analysis of a new trial using historical controls
b. Design Considerations
c. Implementation in real-life using RBesT: Case studies
d. Extension of meta-analytic framework
IV. Propensity score approaches
a. Score-integrated power prior
b. Composite likelihood approach
c. Design and sample size considerations
d. Real life application
V. Regulatory perspective of using trial external information
VI. Concluding Remarks and Discussion
PRESENTER(S):
Dr. Satrajit Roychoudhury is a Senior Director and a member of Statistical Research and Innovation group in Pfizer Inc. Prior to joining, he was a member of Statistical Methodology and Consulting group at Novartis. He started his career as a research statistician in Schering-Plough Research Institute (now Merck Co.). He has 15 years of extensive experience in working in different phases of clinical trials. His primary expertise includes implementation of innovative statistical methodology in clinical trials. He has co-authored several publications/book chapters in this area and provided statistical training at major conferences. His areas of research include the use of survival analysis, model-informed drug development and Bayesian methods in clinical trials.
Dr. Ram C Tiwari is a Senior Director and head of Statistical Methodology at the BMS/Lawrenceville site in New Jersey. In this position, Ram is responsible for promoting the use of novel statistical methods and innovative clinical study designs in the drug development. Prior to joining BMS, he spent 20+ years serving at NIH/NCI and the FDA, and over 20 years in academia as Professor and Chair of the Department of Mathematics at the University of North Carolina at Charlotte. He received his PhD from Florida State University and he is a Fellow of the American Statistical Association. Ram has published over 200 papers on various topics in Statistics, and a book on “Signal Detection for Medical Scientists: Likelihood Ratio Test-based Methodology”, 2021, Francis & Taylor.
The amount of real-world data (RWD) collected from sources other than protocol-driven clinical studies is increasing ultra-rapidly. Such sources include procedure or disease registries, electronic health records, electronic insurance claims databases, and patient-reported outcomes. The clinical evidence that can be derived from analysis of these RWD is considered as real-world evidence (RWE) that can complement the knowledge derived from traditional well-controlled clinical trials. Statistically, leveraging RWE can be viewed as "borrowing" data collected on patients from RWD sources to augment a prospective investigational study and reduce the required number of prospectively enrolled patients. Leveraging RWE can therefore save time and cost of the investigational study, thereby improving the efficiency of regulatory decision-making.
Incorporating RWD in regulatory decision-making demands much more than "mixing" RWD with investigational clinical trial data. The RWD has to undergo appropriate analysis for deriving the right RWE. Moreover, such analysis has to be integrated with the design and analysis of the investigational study for regulatory decision-making. The standard clinical trial toolbox does not offer ready solutions for incorporating RWD. Therefore, there is an unmet need for sound clinical trial design and analysis for leveraging RWE in clinical evaluations.
Course Outline and Main Topics:
In this course, the instructors will cover a series of methods they have developed for leveraging real-world data in clinical trial design and analysis. Noteworthy, these work has been recognized by the FDA and received The FDA CDRH Excellence in Scientific Research Award-EXTERNAL EVIDENCE METHODS RESEARCH (GROUP) and The FDA Scientific Achievement Award- EXCELLENCE IN ANALYTICAL SCIENCE (GROUP) for extraordinary achievements in the timely development and active promotion of novel statistical methods for leveraging real-world evidence to support regulatory decision-making in 2020.
In Part I of the course, we introduce a new method for proposing performance goals—numerical target values pertaining to effectiveness or safety endpoints in single-arm medical product clinical studies—by leveraging RWE. The method applies entropy balancing to address possible patient dissimilarities between the study’s target patient population and existing real-world patients, and can take into account operation differences between clinical studies and real-world clinical practice.
In Part II of the course, we introduce a method that extends the Bayesian power prior approach for a single-arm study to leverage external RWD. The method uses propensity score methodology to pre-select a subset of RWD patients that are similar to those in the current study in terms of covariates, and to stratify the selected patients together with those in the current study into more homogeneous strata. The power prior approach is then applied in each stratum to obtain stratum-specific posterior distributions, which are combined to complete the Bayesian inference for the parameters of interest.
In Part III of the course, we introduce several extensions of the PS-integrated method in Part II. These extensions include 1) a frequentist PS-integrated composite likelihood approach for incorporating RWE in single-arm clinical studies; 2) leveraging multiple RWD sources in single-arm medical product clinical studies; 3) leveraging RWD for the evaluation of diagnostic tests for low prevalence diseases; 4) augmenting both arms of a randomized controlled trial by leveraging RWD; and 5) PS-integrated approach for survival analysis.
In Part IV of the course, we describe an R package, psrwe, that implements a PS-integrated power prior (PSPP) method, a PS-integrated composite likelihood (PSCL) method, and a PS-integrated weighted Kaplan-Meier estimation (PSKM) method for the methods in Parts II and III. Illustrative examples are provided to demonstrate each of the approaches.
In Part V of the course, we introduce a propensity score-based Bayesian non-parametric Dirichlet process mixture model that summarizes subject-level information from randomized and RWD to draw inference on the causal treatment effect in exploratory analysis.
Instructors’ background:
Dr. Chenguang Wang is a Senior Director and the Head of Statistical Innovation at Regeneron. Previously, Dr. Wang was an Associate Professor with Johns Hopkins University and an FDA Mathematical Statistician at CDRH. Dr. Wang has extensive experience in clinical trial design and analysis in the regulatory setting. Dr. Wang also holds B.S. and M.S. degrees in Computer Science and has abundant experience developing user-friendly statistical software.
Dr. Nelson Lu is a team leader in the Division of Biostatistics, Center for Devices and Radiological Health, Food and Drug Administration. Dr. Lu has extensive experience in the design and analysis of clinical trials and studies involving RWD for pre-market regulatory submissions. Prior to joining the FDA, he worked in Wyeth.
Dr. Wei-Chen Chen is a computational and mathematical statistician in FDA CDRH. Dr. Chen reviews clinical trial submissions for therapeutic devices. Dr. Chen is also specialized in statistical computing and high-performance computing using R, C, Fortran, MPI, and ZeroMQ.
References:
• Wang C, Rosner GL. A Bayesian nonparametric causal inference model for synthesizing randomized clinical trial and real-world evidence. Stat Med. 2019;38(14):2573-2588.
• Wang C, Li H, Chen WC, Lu N, Tiwari R, Xu Y, Yue LQ. Propensity score-integrated power prior approach for incorporating real-world evidence in single-arm clinical studies. J Biopharm Stat. 2019;29(5):731-748.
• Wang C, Lu N, Chen WC, Li H, Tiwari R, Xu Y, Yue LQ. Propensity score-integrated composite likelihood approach for incorporating real-world evidence in single-arm clinical studies. J Biopharm Stat. 2020;30(3):495-507.
• Li, H, Chen, WC, Lu N, Wang C, Tiwari R, Xu Y, Yue L. Novel Statistical Approaches and Applications in Leveraging Real-World Data in Regulatory Clinical Studies. Health Services and Outcomes Research Methodology. 2020. doi: 10.1007/s10742-020-00218-4
• Chen WC, Wang C, Li H, Lu N, Tiwari R, Xu Y, Yue LQ. Propensity score-integrated composite likelihood approach for augmenting the control arm of a randomized controlled trial by incorporating real-world data. J Biopharm Stat. 2020;30(3):508-520.
• Li H, Chen WC, Lu N, Song C, Wang C, Tiwari R, Xu Y, Yue L. Mitigating Study Power Loss Caused by Clinical Trial Disruptions Due to the COVID-19 Pandemic: Leveraging External Data via Propensity Score-Integrated Approaches. Statistics in Biopharmaceutical Research. 2020. doi: 10.1080/19466315.2020.1860813
• Lu N, Wang C, Chen WC, Li H, Song C, Tiwari R, Xu Y, Yue LQ. Leverage multiple real-world data sources in single-arm medical device clinical studies. J Biopharm Stat. 2021:1-17. doi: 10.1080/10543406.2021.1897994.
• Chen WC, Li H, Wang C, Lu N, Song C, Tiwari R, Xu Y, Yue LQ. Evaluation of diagnostic tests for low prevalence diseases: a statistical approach for leveraging real-world data to accelerate the study. J Biopharm Stat. 2021;31(3):375-390.
• Li H, Chen WC, Wang C, Lu N, Song C, Tiwari R, Xu Y, Yue LQ. Augmenting Both Arms of a Randomized Controlled Trial Using External Data: An Application of the Propensity Score-Integrated Approaches. Stat Biosci. 2021. doi: 10.1007/s12561-021-09315-5. PMC8214051.
• Wang C, Gary R, Bao T, Lu N, Chen WC, Li H, Tiwari R, Xu Y, Yue LQ. Leveraging Real-World Evidence for Determining Performance Goals for Medical Device Studies, Statistics in Medicine, 2021. Accepted.
• Lu N, Wang C, Chen WC, Li H, Song C, Tiwari R, Xu Y, Yue LQ. Propensity Score-Integrated Power Prior Approach for Augmenting the Control Arm of a Randomized Controlled Trial by Incorporating Multiple External Data Sources. J Biopharm Stat. 2021. Accepted.
ICH E9(R1) has defined estimands for pharmaceutical clinical trials more precisely and more thoroughly than any other previous document in the pharmaceutical industry. Of particular note is the clarity around intercurrent events (IEs), and the impact of IEs on one’s ability to define, infer, and assess an estimand. The manner in which such IEs are handled determines not only WHAT is to be estimated (the estimand), but also HOW inference should be performed (e.g., via an estimator and its corresponding uncertainty assessment), including the definition and treatment of missing data. While there have been many publications, scientific sessions, and other venues for discussing and debating this topic, in pharmaceutical clinical trials much less attention has been paid to one of the strategies described in ICH E9(R1), namely principal stratification.
The learning objectives of this course are: • To explain the importance of considering principal stratification when estimating a treatment effect of a new pharmaceutical agent; • To demonstrate the use of the Tripartite Estimand Approach (TEA), which includes estimation of the treatment effect in the principal stratum of patients who would adhere to their study treatment; (1) • To demonstrate the use of principal stratification based on the early response of a biomarker that is predictive of a longer-term clinical outcome. (2, 3)
Principles, theory, and methods with accompanying examples and R codes for executing specialized principal stratification analyses will be covered.
=====
1 - Akacha M, Bretz F, Ruberg SJ (2017) Estimands in Clinical Trials - Broadening the Perspective. Statistics in Medicine 36, 1: 5-19.
2 - Ridker PM, MacFadyen JG, Everett BM, Libby P, Thuren T, Glynn RJ; CANTOS Trial Group. Relationship of C-reactive protein reduction to cardiovascular event reduction following treatment with canakinumab: a secondary analysis from the CANTOS randomised controlled trial. Lancet. 2018 Jan 27;391(10118):319-328. doi: 10.1016/S0140-6736(17)32814-3.
3 - Bornkamp, B, Rufibach, K, Lin, J, et al. Principal stratum strategy: Potential role in drug development. Pharmaceutical Statistics. 2021; 20: 737– 751. https://doi.org/10.1002/pst.2104.
=====
Course Outline
Introduction (Ruberg – 40 minutes)
Rationale for estimating the direct treatment effect (instead of the treatment policy effect)
The Tripartite Estimand Approach (TEA) – Principal strata defined by adherence
Principal strata defined by early biomarker response
Causal Inference (Sabbaghi – 60 minutes)
The Rubin Causal Model (RCM) and Potential Outcomes
Statistical theory and Adherence Average Causal Estimators (AdACE) related to clinical trials
Break (20 minutes)
Examples with corresponding R code (Sabbaghi – 60 minutes)
Applications of the TEA compared to ITT, MMRM and composite strategies
Diabetes
Alzheimer’s Disease
Applications for early response biomarkers
Cardiovascular disease
Psychiatry
Epilogue (Ruberg – 30 minutes)
Putting principal stratification into the overall ICH E9(R1) context
Other disease states for consideration
Other issues for analysis and interpretation of results from principal stratum
Labeling considerations
Floor Discussion – Q&A (20 minutes)
=====
Instructor Background
Stephen J Ruberg, PhD
Dr. Ruberg was in the pharmaceutical industry for 38 years, where he worked in all phases of drug development and commercialization – from R&D to Business Analytics. He retired from Lilly at the end of 2017. In his last 10 years at Lilly, he formed the Advanced Analytics Hub for which he was the Scientific Leader and ultimately the Distinguished Research Fellow. Since his retirement, he has formed his own consulting company, Analytix Thinking, which is dedicated to teaching good statistical principles and to consulting on analytical strategies for organizations. He is also an Adjunct Professor of Statistics in the Department of Statistics at Purdue University.
He has been a Fellow of the American Statistical Association (ASA) since 1994, was given the Career Achievement Award by Quantitative Scientists in the Pharmaceutical Industry and was elected a Fellow of International Statistics Institute.
Dr. Ruberg is the originator of the Tripartite Estimand Approach, and one of the key authors on its original publication.1 That treatise suggested the use of causal inference as a mechanism for estimating the treatment effect in the principal stratum of patients who would adhere to their randomized study medication. He has spoken on many occasions on this concept at statistical meetings and as part of his consulting practice. Furthermore, he has co-authored several publications that derived estimators for adherers (Adherers Average Causal Effect – AdACE) as well as its application to real clinical trials. He has participated in additional research to improve the original estimators including the use of multiple imputation and the derivation of variance estimators.
Dr. Ruberg is an oft-invited speaker with considerable experience in teaching and communicating statistical concepts in understandable and compelling ways.
/////
Qu, Y., Fu, H., Luo, J., Ruberg, S.J. (2020) A General Framework for Treatment Effect Estimators Considering Patient Adherence. Stat Biopharm Res 12:1, 1-18.
Qu, Y., Luo, J., Ruberg, S.J. (2021) Implementation of Tripartite Estimands Using Adherence Causal Estimators Under the Causal Inference Framework. Pharmaceutical Statistics, 20(1): 55-67. doi.org/10.1002/pst.2054.
Luo, J, Ruberg, S, Qu, Y (2021) Estimating the treatment effect for adherers using multiple Imputation. (Accepted in Pharm Stat – to appear)
Zhang, Y., Fu, H., Ruberg, S. J. & Qu, Y. (2021) Statistical inference on the estimators of the adherer average causal effect, Statistics in Biopharmaceutical Research, DOI: 10.1080/19466315.2021.1891965
/////
Arman Sabbaghi
Dr. Arman Sabbaghi is a Visiting Scholar in the Department of Statistics at the University of California, Berkeley during the 2021 – 2022 academic year, and an Associate Professor in the Department of Statistics and Associate Director of the Statistical Consulting Service at Purdue University. He became an Elected Member of the International Statistical Institute in 2020. He received his PhD in Statistics from Harvard University in 2014. Dr. Sabbaghi's research interests are in Bayesian data analysis, experimental design, and causal inference. Specific major objectives of his current research are (1) the development of new causal inference methods for the analysis of clinical trials and Big Observational Data plagued by nonadherence , (2) the creation of mathematical tools that facilitate the characterization of broad classes of experimental designs for the study and improvement of processes in engineering and the physical sciences, and (3) the development of efficient and interpretable statistical frameworks and machine learning algorithms for modeling and quality control in additive manufacturing systems.
Advances in cell and gene engineering technologies have given rise to an exponential growth of development of cell and gene therapies (CGT) around the world. The potential benefits of CGT have been explored in a broad range of therapeutic areas including oncology, rare diseases, diabetes, cardiovascular and CNS. In 2017, FDA approved Kymriah, a CAR-T therapy in the DLBL patients, which was considered a landmark approval for this novel therapy. What are cell and gene therapies? What are the unique challenges as compared to other conventional mechanisms such as small molecule and monoclonal antibody drugs? How can statisticians respond to the unique challenges and bring value to a clinical discussion? In this short course, a comprehensive review of CGT will be covered and followed by other topics listed below.
Course Outline: 1.Introduction to cell and gene therapy a.Types of CGT b. Concepts and rationale
2. Clinical and pharmacological considerations a. Endpoint selection b. Manufacturing challenges c. PK/PD modeling d. Operational considerations
3. Statistical considerations a. Design options in Phase 1 b. Design options in Phase 2/3
4.Regulatory considerations a. Review of regulatory guidance b. Regulatory concerns
5. Case studies
Background of Instructors: Dr. Weidong Zhang has 20 years’ experience in drug development in multiple therapeutical areas including oncology, inflammation and immunology and gene therapy technology. He has taught numerous short courses in multiple occasions including ASA/FDA industry statistical workshop, DIA, and ICSA, MBSW etc.
Dr. Srinand Nandakumar is a Senior Director of Biostatistics at Nurix Therapeutics. He has worked on several indications including Allogeneic CarT therapies, monoclonal antibodies and small molecule therapies as novel treatment approaches for treating cancer, immune disorders and CNS disorders. His areas of research interest include development of strategic designs and methodology of integrating RWE to clinical research.
Lynn Navale has over 20 years’ experience in clinical development, with 18 years’ experience working in oncology development. With a background in biostatistics and mathematics, she oversees the technical design, analysis, and data acquisition for clinical trials and clinical development strategy.
Ms. Navale has served as Vice President, Biometrics at Allogene since March 2021. Prior to Allogene, Ms. Navale was the Vice President of Biometrics at Kite Pharma, where she developed and led the Biometrics function including biostatistics, statistical programming, and data management and served as the Biometrics team leader for the U.S. and EU regulatory approvals of Yescarta. While at Kite, she led the statistical design of other Yescarta and Tecartus trials, including ZUMA-2, -3, -4, -5, and ZUMA-7, one of the first randomized trials of anti-CD19 CAR T cell therapy. Previously, from 2003 to 2014, she worked at Amgen in roles of increasing responsibility within Clinical Development Biostatistics where she worked on the US filing of Vectibix and led statistical efforts for the phase 1 through phase 3 development of trebananib. She began her career at Baxter BioScience and was the lead statistician for the trial that led to the U.S. regulatory approval of Advate. Ms. Navale has a B.S. in Math from the University of Michigan and an M.S. in Biostatistics from the University of California Los Angeles.