An official website of the United States government

Audio Transcript | Using Bayesian Statistical Approaches to Advance our Ability to Evaluate Drug Products

Dr. Weber: Welcome to the CDER Small Business and Industry Assistance (SBIA) Chronicles Podcast.

Today’s topic: Using Bayesian statistical approaches to advance our ability to evaluate drug products.

My name is Dr. Ellicia Weber and today we are joined by Jennifer Clark, a Lead Mathematical Statistician in the Center for Drug Evaluation and Research at the FDA. Dr. Clark will be discussing recent changes in the use of Bayesian statistics.

Thank you for joining us today, Dr. Clark!

Dr. Clark: My pleasure!

Dr. Weber: Dr. Clark, for those that are unfamiliar with this topic, let’s begin with an overview of Bayesian statistics.

Dr. Clark: Of course. So, Bayesian Statistics is a particular approach of applying probability to statistical problems. This approach starts with a summary of our prior beliefs based on the relevant, available information. When we collect new data, so for example in the course of a clinical trial, the information from this data is combined with our prior beliefs to provide our current beliefs in terms of probabilities.

Dr. Weber: How is this different from traditional statistical approaches?

Dr. Clark: So, the traditional or classical statistical approaches to decision-making are based on only the new data and they don’t incorporate any prior beliefs. Bayesian statistics can be used in practically all situations in which traditional statistical approaches are used and they may have some advantages. So, for example, when experts from various disciplines have determined that there is high-quality, relevant information external to the clinical trial, Bayesian methods may allow studies to be completed more quickly and with fewer participants, and it’s easier to adapt the design of a Bayesian trial based on the accumulated information compared with a traditional trial.

Dr. Weber: It sounds like there have been fairly recent changes in the use of Bayesian statistics.

Dr. Clark: Yes, well the development of Bayesian statistical methods began centuries ago. The recent increases in computing power have made it more feasible to implement these statistical approaches. By the end of the second quarter of the Fiscal Year in 2024, the FDA actually expects to convene a public workshop to discuss aspects of complex, adaptive, Bayesian, and other novel clinical trial designs. By the end of the Fiscal Year in 2025, FDA also anticipates publishing a draft guidance on the use of Bayesian methodology in clinical trials of drugs and biologics.

Dr. Weber: Are there particular areas where you are seeing the use of Bayesian statistics to be helpful?

Dr. Clark: Yeah, so at the FDA, we are seeing increased use of Bayesian statistics in pediatrics. Since many cases in pediatric drug development occur after it’s been demonstrated that the drug is safe and effective in adults, Bayesian statistics can incorporate the information from adults and use it in understanding the effects of a drug in children.

A recent example in asthma products was discussed at an FDA advisory committee on Nov 8, 2022. So, the primary clinical study enrolled both pediatric and adult patients to estimate the rate of severe asthma exacerbations relative to placebo. And traditionally, we would look at the results of the populations separately and in a single pooled analysis. However, in our evaluation of this product, using Bayesian methods allowed us to borrow variable amounts of information from adults and to evaluate the dependencies of these results on the amount of information borrowed. This ultimately allowed us to make more informed decisions.

Dr. Weber: I see. Are there any additional areas of particular interest?

Dr. Clark: So, we are seeing more Bayesian designs in dose finding trials for drugs and biologics, particularly in oncology where the primary goal is to understand the tolerability of the product and to identify a maximum tolerated dose, or MTD. Here, Bayesian designs allow much more flexibility in the design and dosing in the trial. They can also improve the accuracy with which the maximum tolerated dose is estimated and the efficiency of the study by linking the estimation of toxicities across doses.

The ultra-rare diseases, that’s another area where traditionally designed trials are particularly challenging because of the extremely limited patient population sizes. Bayesian methods provide two key advantages here: first, the ability to incorporate prior information and also the ability to adapt the design more easily.

Some Bayesian approaches use hierarchical models. And these are multi-level statistical models describing how data elements relate to one another. They are particularly useful for assessing how well a drug in a particular subgroups of patients works. So, for example, those subgroups could be defined by age or race. Analyses based on these models can provide estimates for drug effects in these subgroups that are generally more accurate than the estimates obtained by analyzing each subgroup in isolation.