The U.S. Food and Drug Administration (FDA) has issued a new guidance on Bayesian statistical methods in the design and analysis of medical device clinical trials, that the agency says may result in less costly and more efficient patient studies.

The final guidance is titled "Guidance for the Use of Bayesian Statistics in Medical Device Clinical Trials," and reflects public input gathered in 2006 after the release of the draft guidance. According to the guidance, Bayesian statistics is a method to learn from evidence as it accumulates. By comparison, traditional, or frequentist, statistics use information from previous studies only at the design stage, not as a part of formal analysis at the data analysis stage. The Bayesian method "uses Bayes' Theorem to combine prior information with current information on a quantity of interest." The prior information and trial results are part of a "continual data stream" and inferences are updated with each instance of new data.

"This final guidance on the use of Bayesian statistics is consistent with the FDA's commitment to streamline clinical trials, when possible, in order to get safe and effective products to market faster," said FDA Commissioner Margaret A. Hamburg, M.D., in an FDA press release. "This is a terrific example of regulatory science in practice at FDA."

The FDA's guidance says a correct Bayesian approach could be less burdensome than a frequentist approach. The FDA is mandated to consider the least burdensome and appropriate means of evaluating effectiveness of a device that would have a reasonable likelihood of resulting in approval.

The guide states that the Bayesian approach is also good for surveillance purposes. Information may be updated as provided by a pre-market clinical trial with post-market data "if you can justify exchangeability between pre- and post-market data."

Some of the suggested information to be included in clinical trial protocols to facilitate the statistical issues unique to Bayesian trial designs include: prior information to be used and the assumptions made; a criterion for success of the study; justification of the proposed sample size; tables of the probability of satisfying the study claim; evaluation of the prior probability of the study claim; and simulations of the trial at the planning (or IDE) stage. These and more suggestions are explained in detail within the guide.

The FDA reports that in a June 2009 public meeting, the Medicare Evidence Development & Coverage Advisory Committee encouraged Medicare policymakers to consider Bayesian approaches when reviewing trials or technology assessments for the national coverage analysis process.

Read more:

The guidance may be accessed at: http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm071072.htm#6

Adapted in part from a FDA press release:
Link: http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm199993.htm

FDA Announces Initiative to Facilitate the Development and Availability of Medical Devices
www.dotmed.com/news/story/1871

More Industry Headlines