Scientific misconduct and correcting the scientific literature.

Journal editors are among those who must face the issue of when and how to correct the scientific literature when an allegation or finding of scientific misconduct occurs. The author describes an instructive incident of tainted data and a subsequent allegation of misconduct that involved a federally-sponsored study where some data had been fabricated. The journals that had published or were considering articles from that study were not told about the problem for almost four years after the initial allegations of misconduct. The author then provides information to throw light on such questions as: Who has the responsibility to ensure that a manuscript that may contain falsified or fabricated data is not published? Who has the responsibility to correct the literature when falsified or fabricated data have been published, and at what point should that correction be made? For example, should it be when the problem of data is suspected or when it is proven? And if proven, proven by whom? How is the larger scientific community to be notified about the problem? Where and when should the correction or retraction appear, and what should it tell readers about the basis for the retraction or correction? She also presents data from 25 cases to show the various lengths of time involved in correcting the literature after allegations of research misconduct had been made. The author concludes that the record shows how disconnected journal editors have been from the scientific misconduct process and that expectations differ regarding the obligations of authors, research institutions, and federal agencies about informing a journal when an allegation of scientific misconduct is made about a publication in its pages. The 25 cases show that substantial delays in notifying the journal and the public about allegations and findings of scientific misconduct are endemic, and that all parties have far to go in appreciating their roles in maintaining the integrity of the biomedical literature.