# The FDA Bayesian Guidance

Page: https://stenobird.com/podcast/in-the-interim-7211889/the-fda-bayesian-guidance
Text version: https://stenobird.com/podcast/in-the-interim-7211889/the-fda-bayesian-guidance.md
Podcast: [In the Interim...](https://stenobird.com/podcast/in-the-interim-7211889)
Published: 2026-01-26T12:00:00+00:00
Episode link: https://share.transistor.fm/s/d4747470
Audio file: https://media.transistor.fm/d4747470/f48e6bc2.mp3
Processing state: not_requested
JSON: https://stenobird.com/v1/public/podcasts/in-the-interim-7211889/episodes/the-fda-bayesian-guidance
Duration seconds: 2600

## Resource

In this episode of "In the Interim…", Dr. Scott Berry and Dr. Kert Viele deliver a quick reaction to the FDA’s draft guidance on Bayesian statistics for clinical trials of drugs and biologics. Their assessment addresses the structure, content, and impact of the document, emphasizing evidence-based requirements and guidance scope. The episode breaks down regulatory language, technical expectations, and workflow implications for clinical trial sponsors and statisticians. Key Highlights Clear distinction between trials justified by type 1 error control and trials justified by agreement on Bayesian priors and decision rule. Explanation of how informative priors can be created based on external or historical data. Technical explanation of dynamic discounting/borrowing, especially in Bayesian hierarchical models for rare populations, pediatric-adult extrapolation, related disease subgroups, and platform and basket trials (e.g., ROAR). In-depth look at the necessity of sensitivity and robustness checks for different priors, and the FDA’s design prior and analysis prior terminology. FDA’s requirements for accepting external data sources: data provenance, patient-level comparability, recency, and appropriate covariate adjustments. Comparison with ICH E20 on adaptive designs, providing context for ongoing regulatory harmonization and possible influence on international regulatory directions. Direct warning against attempts to misuse Bayesian methodology as a substitute for scientific rigor; legitimate uses must meet FDA standards and not simply serve to lower evidentiary bars. Resource: FDA News Release: https://www.fda.gov/news-events/press-announcements/fda-issues-guidance-modernizing-statistical-methods-clinical-trials For more, visit us at https://www.berryconsultants.com/

## Actions

- request_transcript: `POST https://stenobird.com/v1/public/podcasts/in-the-interim-7211889/episodes/the-fda-bayesian-guidance/transcription-requests` — Idempotently request low-priority transcript generation for this episode.
- read_markdown: `GET https://stenobird.com/podcast/in-the-interim-7211889/the-fda-bayesian-guidance.md` — Read the agent-friendly Markdown representation of this episode resource.

A page view does not enqueue transcription. Agents should invoke `request_transcript` explicitly when they need this episode processed.

## Transcript

Full transcripts are not published on public pages unless there is a clear rights basis.