Retraction Watch

Tracking retractions as a window into the scientific process

JAMA tells readers: “Caution advised.” Here’s why.

with 8 comments

Last week, JAMA issued some unusual notices, letting readers know they should use caution when reading an editorial and letters associated with now-retracted articles by a bone researcher in Japan.

The notices — for papers by Yoshihiro Sato, now up to 14 retractions — remind readers not to heed the results of the now-retracted papers, and alert them to read any associated materials (specifically, an editorial in JAMA and letters in JAMA Internal Medicine) with caution.

The text of the notices describes them as “formal correction notices;” we asked Annette Flanagin, executive managing editor at The JAMA Network, why they chose that approach, instead of an expression of concern or retraction:

We do not wish to penalize authors of such Letters or opinion pieces by retracting them.

The three papers were retracted last year, in June. We asked why the publisher waited more than a year to alert readers about the associated writings:

This is the first time we have used this method to alert readers in this way. We have been thinking about how best to do this for some time. These corrections are related the last retractions that we have published. We plan to use this approach going forward in the event of a future retraction.

Flanagin said she hadn’t seen other journals issue similar notices in the past, but “that does not mean it hasn’t happened:”

I reviewed the retraction policies of ICMJE, CSE, and COPE and didn’t see any guidance on how to handle these types of articles related to retracted articles that were published before the retraction notice. Those policies do comment on assessing the validity of previously published work by the author(s) of retracted articles, but that is different.  I expect that we may add guidance on the articles related to retracted articles that were published before the retraction notice to the next edition of the AMA Manual of Style, which is under revision, if we see that this solution works.

Here’s the text of the notice for the materials associated with the JAMA paper:

The Preliminary Communication “Effect of Folate and Mecobalamin on Hip Fractures in Patients With Stroke: A Randomized Controlled Trial,” was retracted on June 3, 2016. An Editorial and Letter were written to comment on the originally published article and were published before the Retraction notice. This formal Correction notice is being published to alert readers and remind them not to rely on the subsequently retracted article.

The JAMA editorial “Homocysteine and fracture prevention” has been cited 11 times since it was published in 2005, according to Clarivate Analytics’ Web of Science; the most recent citation appeared in 2016. The letter “Inaccurate description of collaborating hospitals in a study of the effect of folate and mecobalamin on hip fractures after stroke,” co-authored by Sato, was cited six times since 2006; the last citation occurred in 2012.

And here’s the text for the notice accompanying the articles in JAMA Internal Medicine:

Two Original Investigations, “The Prevention of Hip Fracture With Risedronate and Ergocalciferol Plus Calcium Supplementation in Elderly Women With Alzheimer Disease: A Randomized Controlled Trial,” and “Risedronate Sodium Therapy for Prevention of Hip Fracture in Men 65 Years or Older After Stroke,” were retracted on June 3, 2016. A series of Letters were written to comment on the originally published articles and were published before the Retraction notice. This formal Correction notice is published to alert readers and remind them to not rely on the subsequently retracted articles.

The letter “Prevention of hip fracture in elderly women with Alzheimer disease,” in Archives of Internal Medicine, was cited once in 2012. “Prevention of hip fracture in elderly women with Alzheimer disease—reply,” co-authored by Sato, has not yet been cited. “Risedronate for the prevention of hip fractures: concern about validity of trials” was cited four times since 2007, the last in 2016; Sato’s reply was cited once, in 2011.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

Comments
  • rfg August 11, 2017 at 4:08 pm

    Excellent approach.

    This approach needs to be expanded broadly.

    If a follow-up paper/letter/review references a retracted paper there should be a notification. This could simply be a footnote in the biography all the way to the “caution advised” notice approach of JAMA.

    Of course, there are also occasions – such as more fabrication – where the follow-up should be retracted as well.

    • Just some guy August 12, 2017 at 7:24 am

      If it’s a response/letter/review I’d like a sentence or two placed at the start of the article. It doesn’t have to be much, something along the lines of “Please note that due to retractions (1,2,3…n), parts of this article citing references (2,7,8,31) may no longer be correct.” would be fine. (Depending on how much of the article relies on those references of course.)

  • rfg August 13, 2017 at 5:04 am

    A very good start would be to add a notice to the PubMed page for papers citing a retracted paper. Could be done for corrected papers too.

    This would be a great service to readers.

    Citations are tracked in Web of Science, Scopus and Google Scholar.

    The challenge would be to link this information to PubMed. It seems like it should be doable

  • rfg August 14, 2017 at 4:03 am

    “PubMed already prominently displays retraction status…”

    Yes – well noted and appreciated.

    A separate issue is mitigating the downstream impact of the retracted article on the scientific literature.

    It would be helpful for readers to know that the authors of an article that cited a paper before it was retracted may have been mislead by the erroneous, falsified, fabricated or otherwise non-reliable information.

    A simple note on PubMed – this paper cites XXXX et al. date (doi:), which was subsequently retracted – would caution the readers that they should not rely on whatever conclusions if any were drawn from the citation of the reacted paper.

    JAMA’s editors got this right. It’s not a trivial concern, Many ill effects of a retracted paper can be perpetuated in the scientific literature via citations made during the often protracted period before the issues with the retracted paper result in its retraction.

  • Cheryl Hamill August 14, 2017 at 4:40 am

    Ah it appears my reply to rfg was not on target – dangers of speed reading. PubMed doesn’t index the full text of papers so unless the papers are in PubMed Central or a source such as AHRQ (where NLM does have the full text) I don’t see how this could happen. CrossRef link-outs to a DOI that go to a publisher website can help where the publisher includes the retraction notice – but not till you get there. Given publishers are now providing their citation data to NLM through the PubMed Data Management system it would be a useful enhancement to PubMed for NLM to request publishers to provide details of cited papers along with other citation information. This must already be in xml format for it to be supplied to sources such as GoogleScholar, Web of Science etc. So rfg is spot on in the comments.

  • aceil August 14, 2017 at 8:15 am

    Concerned about retracted papers being primary to meta analyses. A Lego tower awaiting a crash?

  • knd August 14, 2017 at 6:23 pm

    I find this rather illogical:

    The 3 original JAMA/Arch Int Med papers have been retracted because misconduct occurred and the data are unreliable, but the editorial in JAMA that is based on the unreliable paper and the letter by the authors correcting errors in the original unreliable paper are not retracted?

    Likewise in JAMA int med, 2 letters by the authors discussing more analyses of their unreliable data and defending criticisms of their unreliable papers are not retracted?

    If the purpose of retractions are to correct the literature and ensure its integrity, it would seem logical to retract the authors’ letters in response. It seems reasonable that the other correspondence should remain but with indications that the original has been retracted. You can argue for or against retraction of the editorial. Withdrawal and republication of an updated editorial would be a good but impractical option.

  • Post a comment

    Threaded commenting powered by interconnect/it code.