Ron Marks, PhD, serves as CSO and director of biostatistics and is a Clinipace Worldwide cofounder.
It was a pleasure to host the webinar “Navigating Regulatory Biostatistical Requirements Throughout the Clinical Trial Lifecycle” with my colleague, Scott Miller, PhD, a biostatistician at Clinipace. We reviewed common statistical issues (during trial planning, conduct, analysis, and submission phases) and discussed potential approaches to address these issues.
We’ve addressed many excellent questions from attendees regarding the trial planning phase (including protocol deviations, adaptive designs, and alpha spending functions in interim analyses) and the trial conduct phase.
In our last posts, we address questions related to trial analysis and submission. If you have any additional thoughts or related questions to ask our experts, please share your comments below.
Q: What is the best way to present post hoc analyses to the FDA?
A: This is a very interesting question because you can get into really some good discussions about whether or not post hoc analyses are useful. When you write your protocol and especially the SAP, you’re defining all the important planned analyses that are going to take place, which you completely identify before you ever get the database lock and look at your data. That’s clearly what the FDA is going to focus on.
But I think you have an opportunity to really use post hoc analyses, especially in Phase II studies more so than in the Phase III studies, to learn more about your product. We have three different stages of analysis: You have the primary analysis, the secondary analysis on your secondary endpoints, and then either exploratory or observational endpoints and analyses.
While exploratory results are interesting results you want to look at, they can’t be used for registration purposes or labeling. Your primary and secondary analyses are what you hang your hat on to go to the FDA. However, exploratory analyses can be very useful to the client and also informative to the FDA, especially in Phase II studies, to help plan your next study or set of studies that may be coming.
As a result, you don’t have to be nearly as specific with your exploratory analyses as with the primary and secondary analyses. In the analyses of primary and secondary endpoints, you list the specific statistical methodologies you’re going to use in the protocol and SAP and stick to that. However, you may not even know enough about some of the exploratory endpoints to know exactly the analysis to do. This is where you just want to ask “What are the different exploratory endpoints and exploratory analyses I might want to look at?” You might not even do all of them, but if you identify them upfront, you alert the FDA or the regulatory authority, “Here are the other areas of my data I’m interested in learning something about that might help me plan for future studies.”
I believe this type of approach can get a lot of leeway from the FDA. There’s an understanding that you’re not going to draw any firm statistically relevant conclusions from the exploratory analyses because there are analyses that were not planned before you looked at your data—only after you looked at the data. I personally find a lot of opportunity in exploratory or post hoc analyses as long as they’re properly explained in your response to the FDA.
If you’re interested in learning more about Navigating Regulatory Biostatistical Requirements through the Trial Lifecycle, be sure to listen to our webcast in its entirety. Check back soon for the related eBook!