FDA issues guidance on use of Bayesian methods to support drug development

The US Food and Drug Administration (FDA) issued a draft guidance on Friday to assist sponsors in using Bayesian methods to support the safety and effectiveness of new drugs in clinical trials. These methods utilize information from previous studies to accelerate the drug development process.

The guidance pertains to clinical trials utilizing methods for investigational new drug applications (INDs), new drug applications (NDAs), biologics licensing applications (BLAs), and supplemental applications.

FDA announced that issuance of the guidance fulfills a commitment made under the sixth reauthorization of the Prescription Drug User Fee Act (PDUFA). This commitment aims to expedite drug development by utilizing complex adaptive, Bayesian, and other innovative clinical trial designs. The Complex Innovative Trial Design (CID) initiative, which began under PDUFA VI, promotes the application of Bayesian analysis in clinical trials.

Traditional statistical methods for decision-making in new drug development depend solely on new data.

FDA announced that “Bayesian methods can be used in various ways in clinical trials. For example, Bayesian calculations can be used to govern the timing and adaptation rules for an interim analysis in an adaptive design, to inform design elements (e.g., dose selection) for subsequent clinical trials, or to support primary inference in a trial.”

The guidance emphasizes several key aspects of Bayesian methods. It addresses situations where these methods are applicable, outlines general principles for selecting prior distributions, and discusses considerations for using estimands and managing missing data in a Bayesian context. Additionally, it covers the use of software for conducting Bayesian analyses and provides best practices for documenting and reporting the results of these analyses.

Bayesian methods can be used to borrow information from previous clinical trials, extrapolate data for pediatric populations, or for dose-finding studies.

The guidance demonstrated how a sponsor utilized information from a previous clinical trial in a phase 3 study to evaluate REBYOTA, a fecal microbiota transplant product designed to prevent the recurrence of Clostridioides difficile infection (CDI) in individuals with recurrent CDI. “C. difficile is a common cause of antibiotic-associated diarrhea and colitis and is a major public health burden,” it states.

“The primary analysis of the randomized, double-blind, placebo-controlled phase 3 study to evaluate the effectiveness of REBYOTA used a Bayesian model to formally incorporate data from a previous phase 2 placebo-controlled study of REBYOTA,” said the guidance. This analysis demonstrated the effectiveness of REBYOTA, which received approval in 2022.

The guidance said that sponsors can leverage the International Council for Harmonisation’s (ICH) E9 and E9(R1) guidelines for more information regarding the use of estimands and missing data.

To effectively document the use of Bayesian methods, the protocol should clearly outline and justify both the study design and the planned statistical analysis methods. Sponsors need to provide supporting information for the proposed prior distribution, as well as any relevant external data concerning borrowing, the likelihood function, success criteria, and the operational characteristics of the trial.

Additional details or supporting information can be included in supplementary documents, such as a statistical analysis plan or simulation report. The guidance said it is important to submit all relevant information to the FDA during the design phase and as early as possible. This will allow sufficient time for the FDA to provide feedback before the trial begins.

Stakeholders have 60 days to comment on the guidance under docket number FDA-2025-D-3217.

Guidance, Announcement

FDA issues guidance on use of Bayesian methods to support drug development

Related topics

Recommended content

Welcome to the new RAPS Digital Experience