Confusion as JNCI yanks press release on embargoed breast density-cancer study after authors see error
Please disregard this MTM, which was sent out on Friday June 29, 2012. The data in the study changed so we are no longer putting out a press release for the study.
The original release was headlined:
Percent Density May Be As Strong a Risk Factor as Variation in Breast Density for Breast Cancer
MTM stands for “memo to the media.” What wasn’t clear was whether the new notice meant the release was being pulled back, or the study itself, but the JNCI press office said it was just the release.
But the Mayo Clinic’s Celine Vachon told Embargo Watch:
Everything in the original press release remains correct.
Our take home message remains the same, that is the automated V density measure (or variation) is a risk factor for breast cancer in three studies.
It performs as well as the percent density measure in these studies. And, the V might be preferable to percent density in the clinical setting, since it is automated.
Further, the V has been seen a risk factor on digital mammography, also. So, the translation of this measure is promising.
The only difference is that the V is not superior to percent density in 2 of the 3 studies, as we initially thought.
But that superior language doesn’t appear in the release. So why yank it? Vachon:
Yes, I’m cautious by nature. The release is indeed correct as written but the paper tables and text need to be updated. I’d like the journal to make these decisions as we’ve never been in a situation like this. Thus, better not to highlight at this point.
So there wasn’t actually anything wrong in the release. The authors just didn’t want to call attention to the study, since given the late hour, it was apparently impossible to change the paper before it was released. The press office tells us:
The authors of the study contacted us with a change in data that due to our production schedule did not allow enough time to make changes to the study. The study will still be released tomorrow, however, there will be no press release to accompany it. An erratum may be published but we don’t know when.
So the inaccurate parts of the paper remain, but since the release could be changed — or withdrawn, as happened — the journal did that.
As far as correcting the paper, Vachon tells me:
We are working with the editorial staff to determine the best way to proceed.
This is an interesting case, in which everyone’s good intentions are clear. It has some similarities with this story about a yanked study, but also important differences. The authors realized they’d made an error, and told the journal. The journal followed the authors’ lead. Perhaps the assumption is that reporters won’t cover anything that isn’t press-released. That may be true at some outlets, but not at most of the ones we know. And judging by the number of readers who forwarded us the JNCI’s alert withdrawing the release, this has garnered some attention among journalists.
Cross-posted on Retraction Watch