A Statistician's Role in Interim Analysis
Posted on June 13, 2013
Sonia Davis (bio) describes the process for evaluating whether a study should be stopped early.
An exciting part of the job as a statistician — and a very important role — is evaluation of an interim analysis, whether a study should be stopped early. And those are always done in very controlled and regulated and planned ways, whether or not you're in the pharmaceutical industry or whether you are doing an academic research study.
And they need to be planned in advance and have interim looks written into a protocol. And pretty much every study would have — if it's a clinical trial, it has a DSMB, Data and Safety Monitoring Board; if it's an observational study, which generally should have fewer risks and less likely need to stop, it has an OSMB, Observational Study Monitoring Board.
And these boards are always outside of the people that are running the trial. They tend to be academic experts in the clinical area being studied. Statisticians sometimes and ethicists sometimes, are representative of an interest group that is being impacted by the disease or the treatment that's being studied. Those are the ones that should always have the decision about whether or not a study or a treatment gets stopped.
We can't just look on the fly and look at the data, and decide that we think that there's enough information or enough signal for us to stop, or a safety concern for us to stop. It definitely needs to be taken away from those who are running the trial and have a very vested interest in the trial continuing and sent to the people who are on the DSMB.
And for efficacy, there's usually two reasons why you'd want to stop early. One time is that there's a lot of evidence that says, "This drug is working. There's enough evidence. You can stop." The second situation is you could say, "This drug is not working. There is never going to be enough evidence in this trial. You're going to not be able to conclude something. Therefore, you should stop and save your resources." But both of those are very important, big decisions to make along the way.
And so, the job of the statistician who's on the project, usually in the academic setting, the statistician will be delivering some reports to the DSMB or the OSMB, and the suggestions that come back from the committee are then ones that are discussed amongst your group; oftentimes, amongst with NIH [National Institutes of Health] as well.
And it sounds interesting. You're going to be exposed to this other information, have some input into seeing how the pharmaceutical develops. But you have to be prepared to have clear thoughts and clear understanding of the implications of when you should stop a study and when you shouldn't and what would happen to the treatment that's being considered when that study is stopped.
Excerpted from interview with researcher in 2013 at the North Carolina Translational & Clinical Sciences Institute.
The North Carolina Translational and Clinical Sciences (NC TraCS) Institute is the integrated home of the Clinical and Translational Science Awards (CTSA) program at the University of North Carolina at Chapel Hill. It is supported through the NIH's National Center for Advancing Translational Sciences (NCATS) grant ULTR000083.
Please note that the feedback is viewed only by 4researchers staff and is not intended for communication with individual contributors.
Use the form below to submit feedback about this article. If you would like a response, please be sure to include your e-mail address.