Several years ago, a third party contacted the journal with concerns about data irregularities in two randomised controlled trials published about 10 years ago. Both of the papers were published before the journal had strict requirements on data upload to a public repository and availability.
The journal sent an initial email to the corresponding author of both papers (the same author for both papers) notifying them of the concerns and requesting information on the study dataset, ethics approval letter, original protocol, original SAP and the randomisation list. The author responded by threatening legal action against the editor-in-chief and did not provide the requested documents.
A year later, the journal contacted the corresponding author again, requesting the same information. The author responded to the second request stating that the original data for the studies had been destroyed/was not accessible because only hard copies were available in patients‘ files and on an old computer. Standards in record keeping require authors to keep data for 15 years.
The journal published an expression of concern for both articles, stating that concerns regarding the validity of the data in both articles had been raised, and that the journal had requested the data from the authors. The third party contacted the editor-in-chief again, suggesting that the papers should be retracted.
The journal sent a third notice to the author requesting the same information. The author responded stating that the situation was currently being investigated by their university. No further correspondence has been received.
Question for the Forum
- Should the journal retract these articles because the data cannot be verified?
The Forum noted that it would be difficult to defend retraction of the article in the absence of evidence of the data irregularities in the published papers. The complainant needs to support their claim that there are data issues in the published articles. Hence retraction might not be appropriate in this situation if there is no evidence to support the claims. The expressions of concern should remain on the papers. The Forum suggested that if the editor cannot get the original data from the authors, they might consider an independent review of the publications and if the same concerns arise, it might be helpful to the editor in making a decision. The Forum asked if there was any subsequent research over the past 10 years that has supported or refuted the data in the trials.
The editor could contact the authors and the institutions informing them that the journal is considering retraction and give them a deadline to respond. If the authors and the institutions are not responding, and the journal cannot verify what has been published, that could be grounds for retraction.
Is there a higher authority, such as an ethics committee, office or research integrity, or vice president of academic concerns at the institution that the editor might contact?
For the future, the Forum suggested that the journal might want to update their author guidelines with explicit instructions around record keeping and data retention policies, and how long they are expected to retain their data.