Bayesian Statistics for Medical Devices: Progress Since 2010.

Area of Science:

• Biostatistics
• Medical Device Regulation
• Regulatory Science

Background:

• Bayesian statistical methods have been used to support medical device regulatory evaluations since the late 1990s.
• The US Food and Drug Administration (FDA) published guidance on Bayesian statistics for medical devices in 2010.

Purpose of the Study:

• To review recent developments in Bayesian methods for medical device evaluation.
• To illustrate the application of these methods in recent regulatory decisions.
• To discuss future opportunities and challenges for Bayesian statistics in this field.

Main Methods:

• Literature review focusing on recent Bayesian statistical developments.
• Illustration of methods using examples of medical device evaluations.
• Discussion of emerging areas such as AI/ML, real-world evidence, and computational challenges.

Main Results:

• Recent Bayesian developments include hierarchical modeling, borrowing strength from prior data, effective sample size, adaptive designs, pediatric extrapolation, benefit-risk analysis, real-world evidence, and diagnostic device evaluation.
• These methods have been applied in recent medical device evaluations supported by the FDA.
• A list of medical devices approved with Bayesian support since 2010 is provided.

Conclusions:

• Bayesian statistics offer a robust framework for medical device evaluation, with ongoing advancements.
• Future directions include integrating artificial intelligence/machine learning (AI/ML), enhancing uncertainty quantification, and addressing computational challenges.
• Continued development and application of Bayesian methods are crucial for regulatory science and medical device innovation.