To gain a better understanding of the problem of dealing with publications whose integrity is subsequently challenged, experience in a well-documented case of research fraud was reviewed. At the University of California San Diego, a faculty committee evaluated 135 publications of Robert Slutsky, MD, and reported to each of the corresponding 30 journals whether each article was valid, questionable, or fraudulent, requesting publication of the criteria and the conclusions. Journals responded slowly to this request; half required additional letters over a 2-year period to elicit a reply. Of the 13 journals that had only valid articles, 5 printed a statement to that effect. Statements concerning 46 of 60 nonvalid articles were eventually published. Journals' inconsistent identification of published statements made it difficult to retrieve them by electronic searching. Only 7 notices covering 15 articles were found by searching under the Medical Subject Heading "Retraction of Publication"; scanning the entire bibliography retrieved 18 articles with retraction notations. A poll showed that journals rarely have written procedures for responding to allegations of research misconduct; in our experience, journals were reluctant to accept authorized retractions or corrections when coauthors failed to act.
KIE: The problems of dealing with scientific publications whose integrity is subsequently challenged are documented in this review of a well-known case of research fraud. A faculty committee at the University of California at San Diego School of Medicine evaluated 135 publications by faculty member Robert Slutsky for evidence of fraudulent data. Friedman chronicles the difficulties the committee encountered in trying to make public its evaluations of Slutsky's data through the journals that originally had published his articles, which were asked to publish the committee's criteria and conclusions. The committee experienced problems both with locating and identifying the published notices, and with retrieving citations to retractions and validations using the computerized database MEDLINE. Friedman notes that few journals have policies or procedures for responding to requests for retractions, but believes that they have a responsibility to their readers and to science to do so.