Last week was the FDA Public Meeting at the CDISC SEND F2F conference and the agency gave clarification about the requirement for juvenile tox studies, and I feel I need to comment. The agency has always been very clear that such studies which include certain reproductive endpoints or multiple phases are out of scope. However, when they first updated the technical conformance guide, it was quickly noted that studies that analyze data relative to the post-natal day (PND), as opposed to the study day, cause a particular issue for SEND. The PND is never stated in SEND v3.1/3.1.1. The PND could be used in SENDIG-DART v1.1, however, no examples were included. For that reason, SENDIG-DART v1.2 was produced to illustrate how to represent PND using the existing SEND concepts.
However, in last week’s public meeting, the agency described how juvenile studies, analyzed by PND are in scope for SEND 3.1/3.1.1. I think we all clearly understand that CDISC defines the standards, but the FDA sets the scope of the requirement and there’s nothing to say that the agency is bound by or limited by anything stated by CDISC or any other organization, even if something comes in scope that was never the intention of CDISC when the standard was created.
So, what is the big deal with requiring juvenile studies in SEND v3.1/3.1.1? Buckle up, this is going to get a bit technical!
SEND 3.1 and 3.1.1 only have the ability to report data relative to the study day. Nowhere is the PND number represented. Yet, in the study report for most juvenile studies, the study day number will not be shown. Only the PND is reported. Therefore, the day numbers shown in the SEND package will not be the same as the day numbers shown in the study report. To work around this, some organizations deliberately set each subject’s study day 1 to be the same calendar date as day 1 PND. They may even go as far as reporting by study day, but even if they report by PND, the day numbers will be the same as the study day, so it’s easy to compare the SEND package to the study report.
However, from talking to numerous organizations, this approach is rare. The common approach is to always collect and analyze by PND, as it is the age of a subject at the point a finding is observed that is more important than how many days it has been on the study, or how many days of dose it has been exposed to. If our study day numbers are not aligned to the PND numbers, then any group summaries produced from the SEND datasets would not align with the study report. This would obviously be challenging for the individual receiving and trying to use the SEND package. However, I’d suggest it presents an even greater challenge to those creating the SEND package, as part of that process requires the package to be QC checked against the study report to ensure that the SEND data are an accurate representation of the study data. How can we be sure of this when they each show different day numbers?
So, the presentation at the FDA Public Meeting has me scratching my head slightly. In truth, we can produce the datasets and QC them, but at significant effort, and I’m left wondering how useful these really are when they do not represent the data as shown in the study report?
Does your organization conduct these studies and have you considered how this announcement would impact your ability to generate the appropriate SEND datasets?
‘til next time,
Marc
