About these ads

Retraction Watch

Tracking retractions as a window into the scientific process

Columbia University misconduct retraction highlights what’s wrong with the retraction process

with 20 comments

jcacoverThe Journal of Clinical Anesthesia has a retraction of a 2006 paper by a group from Columbia University that, to our minds, is the poster child for how not to handle such things.

The article, “Dexmedetomidine infusion is associated with enhanced renal function after thoracic surgery,” was written by Robert J. Frumento, Helene G. Logginidou, Staffan Wahlander, Gebhard Wagener, Hugh R. Playford and Robert N. Sladen, who now is chief of critical care at the institution. The paper has been cited 30 times, according to Thomson Scientific’s Web of Knowledge.

Why do we bother to name all the authors? Here’s why: According to the retraction notice, one of them — but don’t ask which — is guilty of research misconduct:

This article has been retracted at the request of the Editor-in-Chief, Dr. Robert R. Gaiser. The corresponding author, Dr. Robert N. Sladen, has requested the retraction at the direction of the Columbia University Standing Committee on the Conduct of Research, due to research misconduct by one of the coauthors.

Of course, we know that’s not every author is potentially guilty. Sladen, after all, is still employed at Columbia, and as vague as it is, the language of the statement suggests he’s not involved. But the mealy-mouthed wording does leave room for everyone else to be the culprit. What’s more, the notice is excruciatingly unhelpful about the nature of said misconduct. Was it plagiarism? Data fabrication? An author inflating his credentials?

And therein lies the real question: Who’s to blame for such a lousy retraction notice — a notice, we hasten to add, that conforms in no way to Committee on Publication Ethics (COPE) guidelines for retraction statements. To wit:

Notices of retraction should mention the reasons and basis for the retraction.

And:

If retraction is due to the actions of some, but not all, authors of a publication, the notice of retraction should mention this.

The answer for who’s to blame, it seems, is the journal and its publisher, Elsevier.

Neither Sladen nor Gaiser would respond to our queries about the notice. But we did get a reply from Michael Purdy, executive vice president for research at Columbia, who told us this:

We treat questions raised about the legitimacy of research findings with the utmost seriousness and we deeply regret that the conduct of one of our former employees in this instance caused the retraction of this paper.

We believe that the research community should be provided sufficient information to understand the impact of the misconduct on the science, if any, in a way that is fair to all authors.  The appropriate avenue to communicate such information to the research community is the journal that published the paper, rather than through less formal channels.  We are following up with the journal in this regard.

Then we heard from Doug Levy, a Columbia spokesperson, who added:

[N]either Columbia University nor Dr. Sladen drafted this retraction statement or saw it before it was published.

Raising yet another question: What legitimate interest does the JCA have in drafting a ridiculously ambiguous — and, we’ll say, potentially defamatory — retraction notice?

We put that question to Herb Niemirow, an Elsevier exec who oversees the JCA. His response to our questions suggested they hadn’t considered what we think are fairly obvious questions when writing the notice:

You have raised valid questions and we are looking into the matter.  If we decide to amend the retraction we will let you know.

Perhaps coincidentally, we found a retraction of another paper on which Frumento — but none of the other JCA authors — was a co-author. It, too, was a 2006 article, titled “The association between duration of storage of transfused red blood cells and morbidity and mortality after reoperative cardiac surgery,” published in Anesthesia & Analgesia. According to the 2009 retraction notice:

Rothman et al.1 discovered an error in a table in our manuscript describing an association between duration of red blood cell storage and outcome.2 Unfortunately, we were never able to correct the apparent errors since we could not find the dataset upon which the manuscript’s results were based. We have since noted that there were 58 cases of acute renal dysfunction in both our 392-subject data base reported in an abstract3 and our 321-subject data base reported in the article,2 which suggests that the article’s analysis of data from patients with acute renal dysfunction is erroneous. Although our conclusions have been corroborated by Koch et al.4 because we cannot find our data, and have identified discrepant results between our published report and our previously published abstract, we believe it is most prudent that we retract the report (see Notice of Retraction on page 1953).

About these ads

20 Responses

Subscribe to comments with RSS.

  1. As an attorney representing scientists involved in misconduct proceedings, I strongly disagree that a retraction notice should either (a) state whether misconduct occurred, or (b) identify the culpable party, if any. Retraction notices are issued, in my view, to alert the wider scientific community that research results are not reliable. Thus, a retraction, if appropriate, should be done promptly, regardless of why the research is no longer reliable. The question of whether a scientist has committed misconduct is a different issue and may not be determined for months, if not years, after the retraction. In addition, a scientist whose work is under review by an institution (and whose career may be hanging in the balance) is entitled, by law, to confidentiality while the review is on-going. Research journals and their editors may identify misconduct but they are not equipped to adjudicate it. That is a task best left to a university panel or, in some cases, to ORI. I have represented scientists whose work had to be retracted and who, months or years later, were cleared of any inappropriate conduct. The JCA notice could have been drafted better (in my view, the mention of misconduct was wrong and potentially defamatory) but retraction notices should not be a proxy for a finding (or an assumption) of misconduct.

    Richard Goldstein, Esq.

    November 25, 2013 at 10:27 am

    • This retraction notice was requested _after_ a finding of misconduct.

      Dan Zabetakis

      November 25, 2013 at 1:28 pm

      • To Mr. Goldstein:

        I appreciate your sharing your perspective. It is useful to hear from an attorney who represents investigators accused of misconduct. There are several issues here that I believe require clarification:

        1. The comments of Retraction Watch are limited to how journals should handle findings of misconduct. They are not relevant to how institutions, the government, and others handle misconduct. As publications, the peer-reviewed scientific journals are protected under the First Amendment provided they have undertaken reasonable diligence to check the accuracy of the published facts. I would guess that most of your clients are involved with allegations of misconduct that affect their employment, and that you primarily are defending their rights as employees.

        2. Healthcare providers practice medicine, and patients are treated, based on the findings published in the peer reviewed literature. Journals have a duty to alert readers as rapidly as possible concerns about the validity of published research.

        3. Journals exercise that duty through full disclosure. It does make a difference whether the misconduct was plagiarism or data fabrication. That is why the type of misconduct must be disclosed. If the misconduct was plagiarism, then there is little reason to rethink the conclusions of the manuscript. If the result is data fabrication, then all conclusions drawn from the study are compromised. That includes the review articles, practice guidelines, meta-analyses, book chapters, and lectures that are based on the results of the compromised study.

        4. I was closely involved with the retraction of articles by Joachim Boldt several years ago. Boldt’s fraudulent research supported the safety of hydroxyethylstarch. Once his research was found to be compromised, the safety analysis swung in the opposite direction. The unimpeached medical literature now suggest that hydroxyethylstarch may be associated with increased complications. We are not talking about abstract theory. We are talking about real patient harm that would have followed lack of full and complete disclosure of the nature of the misconduct.

        5. To you and others who believe that misconduct should not be transparently reported, how would you feel if a family member received a treatment that caused injury, because a medical journal failed to report that the research results were based on fabricated data? How would you feel as a taxpayer knowing that you were paying for expensive healthcare choices based on fabricated data? It’s absolutely unacceptable to have anything less than full disclosure.

        5. The other reason for full disclosure is to protect physicians whose work has not been compromised. As pointed out by Retraction Watch, the failure to name the culprit in this cases compromises the careers of many investigators involved in the article who have done nothing wrong, other than trusting one of their colleagues. There is a reasonable possibility that the perpetrator has engaged in serial misconduct, casting doubt on his or her other publications. Should the integrity of Dr. Sladen’s papers now be viewed as compromised because Dr. Sladen responsibly reported misconduct by an unnamed perpetrator?

        6. In general I agree that institutions must make determinations of research fraud. In the case of Joachim Boldt, my “job” was to have the responsible Ethical Review Board address my concerns. I was unwilling to make an independent determination. This is the usual course of action, as it should be. However, there are cases where the data are sufficiently clear that Journals can make that determination. Two years ago authors copied a paper from Anesthesia & Analgesia, and republished it in the Saudi Journal of Anesthesia under their own names. Although that is 100% plagiarism, the is also fraud. The authors reported a research study they did not conduct. The institution was notified of the misconduct, but a determination of fraud from the institution was not necessary. Res ipsa loquitur.

        7. I am regularly threatened with lawsuits by authors who are unhappy that their paper is being retracted for misconduct. I am unaware of even a single example in which an author accused of misconduct has successfully sued a peer-reviewed medical journal. If you are aware of such cases. please post them here. I have also not seen any “retractions of retractions”, although there are a few strange cases on Retraction Watch that might qualify. Again, I would appreciate your posting such examples for my own education.

        8. I believe the liability issues are exactly the opposite of what you represent. Author misconduct damages the reputation of the journal that published the retraction. It also wasted a considerable amount of time for the journal, both during the initial peer review leading to publication, and during the subsequent review leading to retraction. I think Journals would be justified in suing authors to recover damages incurred by academic misconduct. Again, I am not aware of any such efforts.

        9. Everyone is entitled to an opportunity to respond when issues of research integrity arise. I am unaware of even a single case in which a retraction notice has been posted without a good faith effort by the institution, or the Journal, to allow the author to respond to the allegations of misconduct. However, to maintain public trust, scientific journals must be transparent, objective, and engage in full disclosure. Anything less compromises the scientific enterprise, and the public health that the scientific enterprise serves.

        Again, I appreciate your sharing the perspectives of an attorney who represents investigators accused of misconduct.

        Sincerely,

        Steven L. Shafer, MD
        Editor-in-Chief
        Anesthesia & Analgesia

        Steven L. Shafer, MD

        November 26, 2013 at 12:26 pm

        • Marco

          November 26, 2013 at 1:09 pm

          • To Marco:

            Thank you VERY MUCH for pointing these out. I didn’t know of this example. It makes interesting reading!

            Three observations:

            1. This is why journals must rely on institutions to make determinations of misconduct, unless the facts speak for themselves, as in my example above. This particular case shows just how complex determinations of misconduct can be. Journals do not have the resources, authority, or responsibility to investigate misconduct. The responsibility necessarily rests with the institution.

            2. This is also why full disclosure matters. Readers need to know exactly what the problem was with the initial submission, exactly why this was felt to be evidence of misconduct, and exactly why the author felt aggrieved by the retraction. It can be hard to sort out. However, experts in the field should be able to follow the rationale and determine if the data can be trusted or not.

            3. There is no evidence that the journal was sued. The journal responsibly reported the findings of the institution. That is how journals should act.

            Every case is different. I recommend reading “The Editor as Umpire” by David Bogod (Anaesthesia 2006;61:1133-1137). This is a case from Taiwan in which an academic competitor successfully bullied a hospital into requesting a retraction. I spent about 6 months working on this case with David Bogod, as it affected a manuscript in Anesthesia & Analgesia. We ultimately decided that the institution had acted in bad faith.

            Conversely, I have been completely convinced of academic misconduct, only to have the institution investigate my concerns and exonerate the author. In those cases I have always deferred to the institution and written a letter of apology to the author.

            When I state that journals must rely on institutions, this does not suggest that journals are bound by the findings of an institution. Circumstances can arise where journal editors make a different determination than the institution, but this should only occur when there is substantial evidence that the institution has acted in bad faith.

            Thanks again,

            Steven L. Shafer, MD
            Editor-in-Chief
            Anesthesia & Analgesia

            Steven L. Shafer, MD

            November 26, 2013 at 2:46 pm

    • What about the reputation of the co-authors? Don’t they deserve fair treatment?

      mat

      November 25, 2013 at 2:00 pm

      • The reputations of all authors, including co-authors, should be protected until a final determination has been made that respects everyone’s rights. In my view, a journal should retract (or clarify) the article without any statement from which one could infer misconduct about any of the authors. In the criminal law, there is a presumption the accused is innocent until proven guilty. A similar presumption should apply here.

        A journal (or journalist) shouldn’t be mentioning the issue of misconduct (which only fosters the notion that retractions are due to misconduct) and, I daresay, the scientific community shouldn’t be presuming misconduct until there’s been a final public determination and all the facts have come out.

        Richard Goldstein, Esq.

        November 25, 2013 at 2:23 pm

        • ” In my view, a journal should retract (or clarify) the article without any statement from which one could infer misconduct about any of the authors.”

          I really have to disagree. We, the scientific community, need to know in detail what happened and by whom. I don’t think there are any scientists (except those devoted to misconduct) who wish for there to be a whitewash.

          Dan Zabetakis

          November 25, 2013 at 4:17 pm

          • Dan: I understand your view – you want to know what happened and who did it. My views, legalistic though they may be (and I can’t hide the fact I am an advocate) come from having represented scientists who were accused of misconduct in a public forum and faced the possible end of the career before they had a chance to clear their name. It’s not easy to say where, in each case, the line should be drawn but confidentiality is built in to the law of research misconduct to protect reputations. In cases of honest error (and in my experience those are far more common than deliberate misconduct) I think discretion on the part of a journal and the institution is the better course and may be required by the law.

            Richard Goldstein, Esq.

            November 25, 2013 at 5:20 pm

          • The retraction note states unequivocally that there WAS in fact a finding of misconduct and reports that fact.

            This retraction notice can be and should be made more clear. Here are some *very* simple means in general…

            1) Each individual author, if alive, should be asked if they agree to retraction, whether they stand by the data, and their response should be noted. I see no reason why at any time after an article is published an author shouldn’t be asked to stand by his/her work. Anyone who does not stand by their work should give a reason. Period.
            2) If the appropriate oversight board has (as they have in this case per the note) found a party guilty of misconduct, this should be noted. It should also be noted whether the party found guilty contested the ruling and/or disagrees with the ruling. If the party does not disagree, of course they should be named! If appeals are ongoing, or processes (legal or private) remain that could potentially vindicate the individual and that individual indicated that (s)he reserves the right to appeal, then and only then should a retraction notice state that an investigation is proceeding to determine if misconduct occurred, with a promise for updates.
            3) I see no reason why this journal cannot, even in the presence of an ongoing investigation, list authors explicitly who have been cleared of wrong doing. Stating that OJ Simpson has been found not guilty of killing Ron Goldman does not imply that someone else did, in fact, kill him.

            The retraction of Smith, Western & Colt, 2006 could read: “This article has been retracted at the request of the corresponding author, Colt. The data used in this report are currently the subject of an ongoing research misconduct institutional investigation that is presently being contested. Smith & Colt agree to the retraction of this article. Western, after being contacted by our editors, did not agree to the retraction. Smith & Colt have been cleared of misconduct by the Equador University Misconduct Board (see what I did there, ha. ha.).”

            Finally, if a (publicly funded, in part) University and or/ Medicare (I assume at least one author was a resident in a Medicare funded slot, but maybe not)… or any other funding source… [interestingly there is NO conflict of interest disclosure or acknowledgements of funding in the original paper from what I can tell...]… is unable to provide definitive conflict resolution that they will make public and stand by their conclusion (even if that conclusion is that the data is bad, but all parties were found innocent of misconduct)… they should NOT be in the business of supporting research. Saying that an investigation is ongoing, someone is guilty or not guilty, would all be OK. Not saying anything is not.

            QAQ

            November 25, 2013 at 11:54 pm

    • Thanks for your comment. RW doesn’t hear enough from the point of view of the POV of those accused of misconduct. With that said, I also disagree with much in your comments. As @Dan Zabetakis points out, there had already been a finding of misconduct. Whether the finding was final, or even correct, is beside the point. It’s a fact, and anyone should be able to report or rely on it without being threatened with defamation claims. Even in the absence of such a finding, journal editors may decide to investigate and act on their own determinations. After all, their careers and the reputations of their journals are also at stake. Then, of course, there are the scientists and administrators who will allocate limited research funds, and the physicians who are considering the use of (e.g.) dexmedetomidine infusion. They generally need to hear more than “oops … never mind” when a report is retracted. Vague retractions also have a way of leaving entirely the wrong impression. See, for example, the retraction of a case reported by some Turkish doctors reported on RW just a few posts before this one.

      Your concerns are entirely legitimate; but they are almost always outweighed by a lot of other factors. Not the least of these factors is the constant struggle not to drown every aspect of human activity in a tide of due process. But, perhaps the most important general considerations is this. If the journals and institutions are not free to report and act on the facts, as they understand them, at a relatively early stage, then we will have to accept the costs of a system in which serious misconduct can be buried in procedures and official silence, the research literature is less reliable, and reputations can be broken on rumor and innuendo while the more informed actors are restricted to “no comment.”

      Toby White

      November 26, 2013 at 7:57 am

    • Richard Godlstein wrote “In addition, a scientist whose work is under review by an institution (and whose career may be hanging in the balance) is entitled, by law, to confidentiality while the review is on-going.”

      I undertsand such a law does not exist.

      Which law are you referring to?

      Stewart

      November 27, 2013 at 4:22 am

  2. The retraction notice does in fact conform to the COPE guidelines which you mention. The retraction notice mentions the reason and basis for the retraction (research misconduct) and it mentions that the retraction is due to actions of some, but not all, of the authors. It does omit details about the research misconduct and it omits the name of the author whose actions led to the retraction, but the COPE guidelines which you quote don’t ask for that…..

    mathbobby

    November 25, 2013 at 12:10 pm

  3. I wonder if any retraction watchers would like to press for comments on the sister article to this..

    http://www.ncbi.nlm.nih.gov/pubmed/16202898

    Both studies report different aspects of a single clinical trial and 5 of 8 authors (including both first authors and the last two authors) appear on both papers.

    Perhaps followup with the corresponding author and/or Columbia to determine whether this study is trustworthy is warranted.

    Also, as alluded to above… I think it is extremely abhorrent of both J Cardiothorac Vasc Anesth and J Clin Anesth to publish papers (ESPECIALLY CLINICAL TRIALS!) without conflict of interest statements and, unless funds for this study were provided entirely by the authors personally and/or Columbia University, without acknowledging funding.

    QAQ

    November 26, 2013 at 12:05 am

    • Mr. Goldstein, what experience do you have in the laboratory? I am of the opinion that attorneys (or anyone else for that matter) should not judge the ethics of scientists, especially if they have no de facto scientific experience. Isn’t it better for scientists to determine what is best for science, and not lawyers, attorneys, politicians and journalists. Let us determine what is important in a retraction notice. You can then defend the scientists who have been accused of inappropriate behavior. Incidentally, what is your take on the Bohannon “sting”? As an attorney, that is…

      JATdS

      November 26, 2013 at 2:15 am

  4. ” In cases of honest error (and in my experience those are far more common than deliberate misconduct) I think discretion on the part of a journal and the institution is the better course and may be required by the law.”

    I disagree with this vehemently.

    I think in cases that get to the point where a retraction is contemplated honest error is extremely rare. We see those cases here on RetractionWatch, and they stand out. What distinguishes them is that an honest scientist can explain clearly and openly where the mistake was made and how it came about.

    What we see in the bulk of retractions, and in many papers that are not retracted (for various reasons) is that ‘honest mistake’ is claimed to justify and excuse the most grotesque fabrications. We regularly see cases where a scientist stitches together different images. When caught they often claim a mere ‘mistake’. These things cannot be done by accident or by incompetence. The person or persons who put the figures together cannot fail to know they were faking the data at the time they did it.

    The claim that protecting the guilty may be ‘required by the law’ is risible. Justice may not always be done, but I know of no legal or ethical principle in favor of cover ups.

    Dan Zabetakis

    November 26, 2013 at 10:53 am

  5. Reputation? We have perhaps a real problem between the legal use and the idea of a “scientific reputation”. In legal terms Mr Goldstein may be right – it would be good if another legal opinion could provide input here. Scientific reputation is somewhat different, I think. It is the consequence of a scientist’s contribution to one or more fields. Clearly, being a coauthor on a paper retracted due to misconduct damages a scientific reputation. If the co-author is innocent, then the damage accrued by the innocent party is far greater than that accrued by the party that undertook the misconduct. Why? Because the latter’s “scientific reputation” has clearly taken a hit too.

    ferniglab

    November 26, 2013 at 11:31 am

    • You are correct — there is a “taint” to having one’s name on a paper that is retracted, since about two-thirds of retractions are due to research misconduct [my MEDLINE-study in 2005 in ORI, and others in more recent years]. Thus, when a finding of research misconduct has been made, by an institution or by ORI, I have for decades (as a former ORI officer) encouraged institutional officials and editors to name the person responsible for the plagiarism, falsification or fabrication and to “clear” the other coauthors who were not responsible for or knowledgeable about the misconduct. Too often retractions have been made for misconduct without stating that and without naming the responsible person — sometimes because the senior author or editor were unwilling to wait a few weeks or months for the formal finding.

      Alan Price

      November 27, 2013 at 12:52 am

      • And, of course, when there is an admission by the person who committed the research misconduct, accepting responsibility and clearing the coauthors, then the retraction can so state (without waiting for a formal institutional or ORI finding).

        Alan Price

        November 27, 2013 at 1:25 am

  6. Columbia credits him as a Phd in some older research, then the Phd is dropped and he’s an MPH, MS. What’s up with that?

    Cant doit

    April 7, 2014 at 4:07 pm


We welcome comments. Please read our comments policy at http://retractionwatch.wordpress.com/the-retraction-watch-faq/ and leave your comment below.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 33,798 other followers

%d bloggers like this: