We started this series by establishing the difference between partial and full unblinding and later, reviewed the trigger points that may result in unintentional unblinding during various trial scenarios. Here we take a different approach and present additional factors, some of which might be considered controversial. related to unintentional and partial unblinding in clinical trials.
Unblinding prevention is a popular topic in the IRT space. It’s well accepted that an accidental unblinding can have ramifications for the study data as well as for individual patients. However, there are specific aspects related to unblinding that aren’t as universally agreed upon.
At Calyx, we discuss the risks of unblinding with our customers often. I would say further that as a provider of IRT services we have probably seen more risks for unblinding than our customers. And what I can tell you from these many discussions is that what unblinding is – and what it is not – is not a closed issue.
Having worked in the IRT space for 25 years, I have a broad understanding and have heard a broad range of perspectives on the topic, particularly as it relates to “partial unblinding”.
“There are absolutes on what unblinding is, but there are certain aspects that are open for debate and interpretation.”
– Craig Mooney, VP Scientific eTech-Enabled Services, Calyx
From participating in debates as both a clinical trial sponsor (with internal customers) and as a solutions provider, I can tell you there is not universal agreement on the following edge issues:
Is partial unblinding a real concern or an overreaction? Is there data to suggest it has a real-world impact or is it just a perception?
Is there a difference between knowing the treatment assignment of an individual subject vs a group of subjects in an open label study? Does aggregate data lead to the same bias in an open label study as it would in a blinded study?
What about random allocation of supplies vs a predetermined dispensing sequence? Can we protect the blind better through this process?
Are randomly generated container / kit numbers an added safety net for unblinding or are they a site burden?
What are the perspectives on not assigning a kit (post-randomization) if not all material types are available? This one is particularly divisive, and I have seen both positions argued aggressively and with conviction of absolute certainty.
There are absolutes on what unblinding is, such as revealing the treatment assignment on a confirmation intended for blinded staff. But there are certain aspects that are a little less clear and open for debate and interpretation.
As our approaches to randomization and trial supply management continue to advance, it’s becoming more important than ever that we begin these conversations to ensure the integrity of blinded clinical trials.