We had previously identified a large body of preclinical trials (>40) reported by Iwamoto, Sato and colleagues. A striking feature of these trials was the similarity between many of the papers. Between March and October 2017, we undertook a systematic assessment of the body of preclinical trials to determine what had occurred and whether there were any other red flags. We identified 52 articles published by the group over 18 years, systematically assessed them using a draft form of the REAPPRAISED checklist [9], and it was immediately obvious there were numerous red flags [12]: the majority of papers reported data that had already been reported from the same experiment in another paper(s), without citing previously published data. The amount of unacknowledged duplicate reporting was considerable. The median proportion of duplicate data within pairs of publications was 45%. Most of the publications reporting data from the same experiment had methodological discrepancies; many papers had implausible or impossible data; 75% of papers had evidence of authorship misconduct; hardly any papers included a funding statement for expensive laboratory research; and there was a high proportion of text duplication in the introduction and discussion sections.
By the time we began to notify journals of these issues, 16 clinical papers by the group had been retracted. Based on our previous dispiriting experience with the clinical studies and our initial contacts with journals regarding the animal studies, we decided to proceed to notifying the institutions and relevant authorities straight away. In October 2017, we notified two US universities (New York University Winthrop Hospital and Texas Tech University), Keio University, as well as the US Office of Research Integrity, MEXT in Japan, and a relevant Japanese specialist society.
In what had become a familiar experience, the process was very slow, took multiple emails to elicit a response, and eventual institutional reports were brief, failed to address the detailed concerns raised, were not reported publicly, and did not recommend any action. So, we were forced to embark on the same laborious process of notifying the individual journals and publishers. Once again, the same issues arose as previously: a few journals/publishers responded, investigated and took action promptly, whereas others did not respond to emails, did not investigate, or did not address concerns. As with the clinical studies, the involvement of publishers ranged from minimal to directing the process. Actions again were very inconsistent: for example, two journals initially determined that “there are some inconsistencies and perhaps sloppiness in reporting” about a pair of papers from the same experiment. One journal took no action, the other issued an erratum. We continued to present a case for compromised publication integrity and eventually, one and two years later, respectively, the two affected journals issued EOCs. One of the journals finally decided to retract 18 months after issuing the EOC, a decision that was clearly expedited by the convening of an independent panel of experts. None of the concerns had changed during this time, just the Journals’ response to them. An unresolved EOC for the paired paper remains in place, 5 years later.
We contacted COPE (Committee on Publication Ethics) about several journal/publisher responses, but once again, in a familiar theme, little useful was achieved. Typically, COPE responded sluggishly, then merely reviewed and endorsed the process that the journal undertook, without regard to whether the decision and action of the journal adequately addressed the detailed concerns raised. All correspondence with COPE was filtered through an administrator and at no point did one of its committee members contact us. In the case of Journal of Neurological Sciences (discussed earlier in Part 2), COPE obtained a response from the publisher after it ceased responding to us, extracted the information that the suite of papers was for retraction, but, at least initially, was not able to prompt the publisher to follow through on its undertaking. Nor did COPE pursue this matter to ensure the necessary retractions. Only after we re-contacted COPE at a later time, did it apply pressure to the journal, forcing it to take action.
Despite the inability of the institutions to determine whether any actions should take place and the initial reluctance of many journals to act [12], we continued to correspond with affected journals and by February 2021, 23 preclinical publications had been retracted. However, in 2024 many remain unretracted, despite being subject to the same integrity concerns as those which were retracted.
Concerns about publication integrity 