FDA Fellowship - Endpoint Strategy Selection and Statistical Analysis Plan for Non-Malignant Hematologic Disease

Research Project

The participant will collaborate with statistical and clinical reviewers to research and develop a practical framework for selecting (primary) endpoint strategies and determining the corresponding statistical analysis plan for non-malignant hematologic diseases in confirmatory trials, starting with the use of historical clinical trials from NDA/BLAs and selected ongoing trials from INDs in myelofibrosis, with the intent to apply the framework to other rare, non-malignant hematologic diseases including PNH, SCD, Anemia of CKD, WAHI, ITP, Hemophilia and others.

• Survey and synthesize endpoint strategies (e.g., co-primary, composite, Hochberg, win ratio, etc.) as well as the statistical analyses from recent submissions/literature with emphasis on clinical interpretability, multiplicity control, estimands, and alignment with common regulatory expectations.
• Analyze operating characteristics under realistic data-generating scenarios (endpoint correlation, effect assumptions, alpha requirements), compare study power and other metrics (e.g., probability that endpoints are individually significant given declared success of the study by a strategy) to understand the pros/cons of different strategies.
• Investigate regulatory flexibility relevant to rare diseases (e.g., alpha relaxation, smaller study power, no requirement for nominal significance for each endpoint, etc.) and balance statistical rigor with clinical feasibility.
• Explore whether the real-world evidence (RWE) approaches as well as Bayesian borrowing can help address sample size and ethical concerns for rare diseases.
• Collaboration between the statistician at DB9 and the clinical team at the Division of non-malignant hematology to map endpoints to disease biology and patient-meaningful benefit; develop annotated case studies highlighting the trial operational characteristics.
• Compare and justify endpoint strategies with respect to FWER control, power, and interpretability in the context of FDA’s regulatory mission;
• Conduct and explain simulation-based operating characteristic assessments, including power, type I error, and other metrics of interest, and communicate their implications for protocol design;
• Translate clinical objectives into estimand definitions and aligned statistical strategies;
• Contribute to regulatory science by documenting case studies and best practices for non-malignant hematologic diseases that generalize to other areas to improve consistency.

Mentor

The mentor for this opportunity is Alison Moliterno (Alison.Moliterno@fda.hhs.gov). If you have questions about the nature of the research, please contact the mentor.

Learning Objectives

• Under the guidance of a mentor, you will learn from the following:
• Methods curriculum: You will be provided guided readings in familiar issues for many non-malignant diseases and also on rare diseases via available guidance and also some statistical guidance (e.g., ICH E9, multiple endpoints, estimands).
• Applied analytics training: You will learn to conduct simulations using R or other programming languages to compare study power and other metrics for different endpoint strategies over different settings and scenarios.
• Participation in weekly internal meetings/groups, receive guidance from mentors, and discuss research. Communicate with mentors as needed during the weekday.
• Pursue the opportunity to co-author a manuscript presenting the finding from this program.
• Obtain a comprehensive understanding of clinical trial design in non-malignant hematological diseases

Qualifications

• The qualified candidate should be currently pursuing or have received a master's or doctoral degree in the one of the relevant fields.
• I am a U.S. citizen, or I have lived in the United States for at least 36 out of the past 60 months. (36 months do not have to be consecutive.)
• I have read the FDA Ethics Requirements.