Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Drug Regulation topic
No spam. Unsubscribe anytime.
FDA participant urges broader use of surrogate endpoints and Bayesian approaches in reviews
Summary
A speaker at an agency roundtable outlined rationale for accepting validated surrogate endpoints, discussed accelerated approvals and argued for Bayesian and continuous statistical methods to improve review efficiency while stressing credibility and pre-specification.
At an informal FDA roundtable, a participant advocated for greater use of validated surrogate endpoints and more flexible statistical frameworks, including Bayesian methods and continuous analysis, to speed drug reviews while preserving credibility.
The speaker said accelerated approval is appropriate when a surrogate is "reasonably likely to predict a clinical benefit," and that fully validated surrogates can support traditional approval. They emphasized that credibility depends on factors including effect size, pre-specified endpoints and appropriate comparators, and argued that regulators should evaluate the "totality of evidence" rather than relying solely on counts of trials.
The speaker also discussed statistical innovation: using Bayesian priors to reflect differing pretest probabilities across disease areas (for example, higher skepticism for Alzheimer’s drugs than for well-established clotting-factor therapies), and the possibility of continuous real-time analysis with stopping rules tuned to each drug’s characteristics.
The participant framed these proposals as internal reforms under discussion and said the agency is "socializing these ideas," with potential guidance and clinical frameworks to follow. No formal guidance or regulatory text was released during the session.
