EXPERTS ASK FDA TO CLARIFY, SIMPLIFY BAYESIAN STATISTICS GUIDANCE

The FDA recently received comments on its "Guidance for the Use of Bayesian Statistics in Medical Device Clinical Trials -- Draft Guidance for Industry and FDA Staff," issued May 23. Experts are asking for specific examples of when to use Bayesian statistics.

The guidance describes the Bayesian approach and how it is used to make device clinical trials more efficient. The guidance suggests that the Bayesian statistical approach could lead to smaller trial samples, but Thomas Louis, a healthcare professional (HCP) with Johns Hopkins University, said, "there is absolutely no guarantee that sample size will be reduced."

"It would be helpful to elucidate more specifically those trial circumstances in which Bayesian methods are truly needed, preferably with worked examples to illustrate the point to nonspecialists," said Stuart Pocock, a UK-based healthcare expert with the Medical Statistics Unit of the London School of Hygiene and Tropical Medicine. Device industry representatives and those they collaborate with during clinical trials are concerned about "getting drawn into Bayesian designs and analyses when they do not really understand what it is all about," Pocock said. (http://www.fdanews.com/ddl/33_28/)