A published paper has been under legal scrutiny due to fabricated data. The court has concluded that the evidence presented undermined the credibility of the study. We have read the COPE guidelines for retracting an article and have checked the flowchart 'What to do if you expect fabricated data'. From reading the guidelines it seems that the editor has the right to retract the paper and to do this promptly. However, because we have not had a case in which a paper has been discredited through a civil court process before, we would like to seek the advice of COPE before we ask for a retraction.
The case in more detail:
• 2006: journal publishes paper that explores the link between two drugs, A and B and vision loss.
• 2006: the drug company responds in a published letter, pointing out flaws in author’s case control study (mainly well known biases of observational studies).
• 2007: author becomes expert witness in case against drug company, based on his paper.
• 2009, June: journal receives letter from author outlining some changes he wants to make to the paper—corrections, and referring to re-contacting people and re-analysing, reaffirming result of link with vision loss in men with a history of MI who have taken drug A (but only drug A now and not B).
• 2009, August: drug company tells journal that author is expert witness and casts doubt on his letter. They want to release source documents to the journal that have been disclosed to them by the author’s university as part of the court case. The university is resisting because it wants the issues argued in court.
• 2009, October: journal consults COPE, who suggest not publishing the letter until the outcome of the court case is known. COPE clearly discussed it as a conflict of interest issue (letter driven by the fact that the author was the expert witness).
• 2009, August: drug company writes to say author has been excluded as expert witness in the trial because errors in the trial call the study’s reliability into question. Provides transcript of relevant court ruling. This makes it clear that the author became an expert witness in February 2007, after publication of paper but before submission of the letter. He clearly became an expert witness on the basis of the study. Transcript says that author acknowledges inaccuracies—11 of 27 patients who originally reported use of the drug before developing vision loss turned out, from original trial forms, not to have started drug A or B until after diagnosis of vision loss. The ‘history of MI’ in some of the men turned out to be a family history of MI. Author also claimed to have re-contacted patients during the trial but court found no evidence of that. Court concluded that these undermined the credibility of the study. The court also ruled out the letter from the author as being similarly inaccurate.
It is clear that the author as an expert witness was arguing that his study showed causation. It was a small observational study: it could never show causation. We have not yet approached the author for his comments.
So, we would like to take this back to COPE to ensure our proposed course of action below is the correct one. We do not think we have had a case in which a paper has been discredited through a civil court process so we want to make sure that the actions taken are still acceptable.
(a) Ask the author whether in the view of the court opinion he wants to retract the paper?
(b) If author says no, refer to his university, asking it to investigate and consider ordering retraction of the paper.
The advice from the Forum was to retract the article and to quote the reasons for retraction verbatim from the civil court ruling, pending any legal appeal. In the event of a legal appeal by the author, the editor should probably publish an expression of concern in the journal. The editor was advised to contact the author’s institution and inform them of the journal’s proposed course of action and ask them if they had anything further to add regarding the case.
Following COPE's advice, the editor wrote to the author informing him that the journal wished to retract the paper. The retraction notice will be published in the next issue of the journal.