Oops: Published But Not Registered or Randomized

Experience is the name everyone gives to their mistakes, said Oscar Wilde. So our latest “experience”  comes from the Obesity journal, where an unusual correction appeared yesterday. An article published in the journal precisely three months ago posted results for treating obesity that seemed too good to be true. But the paper said this study was randomized and had a pre-registered protocol on ClinicalTrials.gov. It went through peer review.  In short, it sounded “impressive,” which we wrote at the time.

Unfortunately, none of that was true. As the correction makes plain, this study had no randomization and no protocol registration. So calling these results “impressive” was a mistake. This is an experience we will remember.

Retractions Are Hard

We have no idea why the journal and the author did not retract this study.

For one thing, the manuscript stated that “312 patients were randomly assigned” to the test and control groups. But this is false. The correction makes it clear that the assignment of patients to treatment groups was not random. The difference between a study with random treatment assignments and “an observational comparison of two convenience samples” is large.

Second, the study protocol had no pre-registration. Medical journals, including Obesity, require that any clinical study have a pre-registered protocol. The ICMJE is clear that journals should never publish an article without a protocol that the researchers registered before the study began. Again, though, the correction makes it clear that the researchers did not register their protocol. “The trial was not registered in ClinicalTrials.gov,” it says.

Yes, retractions are difficult. They take up a lot of time and cause a lot of angst. But they are also very important for scientific integrity.

Publication Should Not Have Been Possible

Coming right down to it, publication of this trial without a valid protocol registration should not have been possible under guidelines for publication ethics. So we are left wondering why this publication was not retracted when it was clear there was no such registration for the study.

Click here for the original paper and here for the correction. For our prior posts on this, click here and here. For further perspective, click here. To read more about why retractions are both important and difficult, click here.

Experience, photograph © theilr / flickr

Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.


June 23, 2021