On June 7, 2021, the US Food & Drug Administration (FDA) Center for Drug Evaluation and Research (CDER) announced the accelerated approval of aducanumab, a monoclonal antibody targeting amyloid, in the treatment of Alzheimer’s Disease. This decision was made in the face of two confirmatory clinical trials that had been terminated for futility, a negative opinion from the CDER Office of Biostatistics, and an overwhelmingly negative recommendation (0 yes, 10 no, 1 abstention) from the Peripheral and Central Nervous System (PCNS) Advisory Committee that reviewed the evidence (Dr. Emerson was a member of that Advisory Committee). 

This module presents a statistical perspective on clinical trial design issues that adversely impacted the available evidence; on problems with data analyses that were presented to the committee; and on the arguments put forth as justification for accelerated approval using a surrogate endpoint that had not been validated. The goal of this module is to examine how different choices in the drug development process might have prevented, or at least mitigated, the controversies in the approval process. In particular, this module will cover:

• the important role that screening pilot studies play in drug discovery and how the results of screening studies can best inform the design of confirmatory RCT (both in terms of powering the study and monitoring safety),
• the appropriate choice and implementation of sequential sampling in clinical trial designs (with particular emphasis on futility boundaries for longitudinal primary endpoints),
• the special problems that arise when time-varying treatment effects are of concern (as might exist due to either treatment mechanism of action or changing RCT conditions),
• the interpretation of discordant RCT results (under both presumptions of a null or an alternative hypothesis),
• the dangers of conditioning on post-randomization variables (with particular focus on nonadherence or changes in concomitant therapies), and
•  the proper validation of surrogate endpoints that might be used in drug approval.