Retractions often seem to take on lives of their own, and they can be complicated lives to manage, even for the most seasoned journal editors. This session focused on milestones in the life of a retraction: its birth as an allegation, the resulting institutional investigation, the journal editor’s decision to retract the paper, and the retracted paper’s afterlife in the literature.

Charlotte Haug, vice chair of the Committee on Publication Ethics (COPE) and editor-in-chief of the Journal of the Norwegian Medical Association, opened the session with a brief background of COPE. She also introduced COPE’s retraction guidelines, which can be found at publicationethics.org/files/retraction%20guidelines.pdf.

Haug’s presentation outlined reasons for retraction of a paper and best practices for retraction. Although various stakeholders can ask for a retraction, the final responsibility lies with journal editors. Haug outlined four situations that call for a retraction: the findings prove unreliable, as a result of either misconduct or honest error; the findings have previously been published elsewhere without proper referencing, permission, or justification (duplicate publication); the text is plagiarized; or the paper reports unethical research. Editors should not retract a paper to punish authors when only a change in authorship is required or when a small part of an article reports flawed data. In those situations, a correction may be most appropriate.

Haug’s presentation ended with a case study that showed the importance of journal leadership in correcting the literature by following the COPE flowchart appropriate to a situation and persevering until the correction was made. Additional examples of cases can be found on COPE’s Web site at publicationethics.org/category/keywords/retractions.

Abraham Fuks, research integrity officer (RIO) at McGill University, spoke about investigating allegations of research misconduct and used the process of McGill University as an example. After a goodfaith allegation is made, the RIO determines whether there is sufficient evidence to warrant an investigation. If so, the allegation is forwarded to the Committee on Research Misconduct (CORM), which determines the relevant facts and validity of the allegations and recommends an appropriate disposition of the case.

The provost considers the CORM’s report and then communicates his or her decision to the various stakeholders. In addition, the provost determines whether any external bodies should be notified of the outcome of the investigation.

Journal editors need to know the results of investigations so that appropriate errata and retractions can be published. But there are still many unanswered questions. For example, Who notifies whom? Who decides when and what to retract? Who writes the retraction? What should it say, and when should it be published? Is there a statute of limitations? How should online versions of articles be handled?

Diane Scott-Lichter, vice president of publishing for the American College of Physicians, provided an editor’s and publisher’s perspective of the retraction process. The premise of her presentation was that “journal editors and publishers are responsible for integrity of the literature.” Scott-Lichter reviewed four types of publication notices: errata, expressions of concern, partial retractions, and retractions in full. The four express, in ascending order of degree, questions or problems concerning the integrity of the work.

The appropriate handling of retractions is of the utmost importance. The list of stakeholders associated with retractions is long, and the inappropriate handling of retractions can adversely affect people, institutions, fields of study, and science as a whole.

The first step is to determine whether an allegation has merit. If there appears to be merit or if there is doubt, the authors should be asked about it. The authors’ responses and the journal’s preliminary investigation determine the course of action, which may be taking no further action, asking the authors to correct their work, or turning the matter over to the institution for a more thorough investigation.

Better communication and collaboration between journals on the one hand and authors, institutions, and oversight bodies on the other can lead to more timely and accurate retractions and thus advance the journal’s objective of protecting its integrity and the integrity of the scientific literature as a whole.

Ivan Oransky, cofounder of Retraction Watch and executive editor at Reuters Health, brought us to the end of the retraction life cycle by exploring the increasing rate of retractions and the effects of retractions on the afterlife of retracted articles. There has been a steep increase in the number of retractions since 1977 to about 400 each year. However, a retraction does not restore full integrity to the scientific literature. First, there is a lapse of time between original publication and retraction, which allows the suspect material to circulate for a longer time unmarked. Even after the retraction has been made, an article can continue to affect the literature greatly. Retracted papers continue to be cited at an alarming rate, and only a small percentage of citations mention that the papers were retracted. For those reasons, the time lapse between publication and retraction noted above is of particular concern.

What can be done? First, journals must do a better job of communicating retractions to their readership and to the scientific community at large. Supporting the reproducibility initiative may lessen the need for retractions. Finally, using CrossMark is a way for an individual reader to know whether the version of a paper that he or she is consulting is the most current one.

In conclusion, the panel agreed that transparency and education are key to implementing best practices for retractions and other corrections in scientific literature. However, the panel also recognized the deeper cultural issue at play here, which also warrants examination: the culture of advancement, tenure, and promotion in academic institutions, which can reward undiscovered misconduct. Examining the behaviors that this culture is motivating may be another step toward meaningful change.