A reader emailed a society, which forwarded the message to the journal office, noting that he can read the name of a patient in a figure in a published letter to the editor. The letter was published online 3 months earlier and had just appeared in print; it was the print version the reader saw. The reader asked if the patient's name could be removed.
The journal’s author instructions already stated that no identifying patient information should be included. At all stages following submission (journal office review of initial submission, publisher production of processed manuscript, copyeditor review, and author review of proof), the patient’s name was not noticed.
Within 2 days of the reader’s email, the publisher had replaced the figure with a version that did not include the patient’s name. It was agreed that an erratum would not be issued so as not to draw attention to the matter.
Both the journal office and publisher have since instituted new procedures in reviewing figures to detect any patient information. Additionally, new text has been added to the author instructions: “Patient’s identity must be removed in all figures (ie, x-rays, MRIs, charts, photographs, etc). Informed written consent is required from any potentially identifiable patient or legal representative, and should be presented in either the methods section or the acknowledgements”.
The publisher and its legal counsel also created a patient consent form (the rural clinic in another country where the patient was treated did not have one) and we able to contact the author who asked her patient to sign the form, post-publication. The patient agreed and signed the form, and the matter is now closed.
Questions for the COPE Forum
(1) Were we right in deciding not to issue an erratum? Is there any other text that you would recommend adding to the author instructions?
The Forum agreed that the editor should not issue an erratum. Correcting the literature is very important, but patient confidentiality is overriding in this case. No useful purpose would be served by issuing an erratum, particularly as written patient consent has now been obtained, and in fact it would draw attention to the name of the patient. A poll of the Forum audience indicated that only a few people would issue a correction—the majority would not. But some argued that in the interests of transparency, perhaps there should be some acknowledgement that a change was made, even if the change is not specifically mentioned.
The Forum discussed having a universal consent form, which has been discussed by COPE in the past. Hence if a patient consented to publication for one journal, if the article was then rejected and submitted to another journal, another consent form would not have to completed. From a poll of the Forum, there was some support for a universal consent form. About a third of the audience currently use a specific consent form for publication that requires the authors to have gained written consent from the patient before any identifying material can be published.
The Forum suggested an editorial note on the importance of preserving patient confidentiality might be useful, as a reminder to authors. The editor might also like to check that their instructions to authors are clear and up to date on this issue.
The editorial team updated their instructions to authors.