In Chapter 4 the Commission found that the raw material which formed the basis for the main part of the PhD project, is manipulated and fabricated. This contributes to a series of later publications having to be disregarded. The Commission has also found several cases of data manipulation and fabrication in the subsequent scientific career of Jon Sudbø. The Commission found that this was due to scientific dishonesty on the part of Jon Sudbø.
The comments that Sudbø has given to the draft report did not provide the Commission with reason to make substantial changes to the preliminary conclusions reached during the investigation.
The Commission has not found grounds to believe that others have participated in manipulating and fabricating research data or in any other way committed scientific dishonesty, as this is defined in the recently adopted Research Ethics Act section 5 (2) (not yet in force).
In its terms of reference, the Commission was asked to seek an explanation of the facts discovered. The Commission has asked itself how these – in retrospect – obvious and gross acts could take place, in collaboration with a series of well qualified coauthors and collaboration partners, and at a renowned research institution.
In introduction, the Commission will emphasize the obvious, namely that there will always be a possibility for the dishonest person to cheat and defraud others. No system is water tight in this respect, and this also applies to Norwegian research. The question here is first and foremost whether it is possible to identify factors that have contributed to the acts discovered by the investigation.
The Commission has by no means any foundation for drawing definite conclusions about what caused these circumstances. However, the Commission finds reason to point out certain factors which may contribute to illuminate how and why “things turned out as they did”. To the Commission, a part of the explanation is that a series of “unfortunate” factors occurred simultaneously, and these must be seen in context to illuminate the case.
Criticizable facts are summarized in Chapter 7.
The Commission has got the clear impression that Jon Sudbø acted relatively freely and independently, both as a research recruit (research fellow) and researcher. This impression is confirmed by several of Sudbø’s
colleagues. Reith explains this by Jon Sudbø appearing as an exceptionally clever PhD candidate and that he was also relatively experienced and not as young as many of the other research fellows. In his comments to the draft report, Sudbø objected to Reith’s description of Sudbø’s relatively free role, and in that connection pointed out that he had little or no experience as a researcher at the time when he started his PhD project. Sudbø conceives this as an attempt by Reith to disclaim responsibility.
The Commission’s total impression based on the information that has come to light during the investigation, is all the same that Reith seems to have had an active but nevertheless relatively superficial role as a supervisor. Reith seems to have observed what went on and discussed this underway with Sudbø, without having been involved in the analyses and treatment of data as such. Nor can he have carried out random checks or suchlike to ensure quality. It seems as if Reith has replied to questions from Sudbø, but not himself put control questions or the like. However, Reith has provided contact with both the Cancer Registry and Gade and thus been a key initiator and door opener. Then he has allowed Sudbø to work independently with the data analyses themselves. The Commission here refers to the fact that Sudbø as mentioned got access to both the material and the patient data. This made it possible to manipulate and change around data lists, blocks and analysis results without anyone discovering it. In light of the description of the facts in Chapter 4, the Commission finds reason to state that Reith should have been more alert as a supervisor and followed Sudbø’s treatment of the data material more closely and not least checked to a larger extent certain factors that are erroneously described in publications in which he is listed as a senior author.
Another element which may provide a certain explanation of the facts that have been discovered during the investigation, is that no formal approval or review of the project exists at all. No permissions, evaluations from the Norwegian Board of Health (now: The Directorate for Health and Social Affairs), the Data Inspectorate/the Norwegian Social Science Services or the Regional Committee for Medical Research Ethics exist, something which also contributes to elude further investigations and quality assurance of the project. As mentioned above, this flaw is something for which Sudbø himself, Reith and the Radiumhospitalet must assume the responsibility.
However, the Commission will underline that there is no reason to believe that such failing compliance with the formalities has been a conscious strategy in order to be able to carry on a dishonest research activity. The explanation of these circumstances seems mainly to be that sufficiently effective and good enough routines have not been in place as regards the control and supervision of research projects at the Radiumhospitalet. The fact that Sudbø did not have a formal employment relationship to the Radiumhospitalet, is evidently enough no acceptable explanation. The Commission has indeed found (section 4.2.3) that the research took place under the auspices of the Radiumhospitalet and that this institution had the overall responsibility for Sudbø’s research activities. Even if instructions and control bodies existed at the
institution, it does not seem as if these were satisfactorily implemented. The project was not submitted to the Protocol Committee, for example, and no one saw reason to question it. Researchers associated with the department indeed seem to have had a relatively relaxed relationship to the formalities. This applies in relation to the retrieval, delivery and treatment of human biological material and sensitive patient information, recommendations from the Regional Committee for Medical Research Ethics, licenses for data processing and dispensation from the duty of secrecy. The Commission has also found reason to express doubt as to whether the organization of this research project has been defensible. This organization is an institutional responsibility.
Furthermore, it may seem that when someone was worried that quality routines may not have been followed, this was not followed up by the management or colleagues, maybe because one did not have any system or tradition regarding this type of warning. It must be noted here that one person retained all the documentation from the contact with Sudbø, exactly because this person had his/her suspicions. This was at the end of the 1990’s. However, the person concerned did not want to get into a difficult “whistleblower position” and therefore refrained from making further investigations.
Concerning the relationship to coauthors, what happened in connection with the publication of New England Journal of Medicine 2001 is apparently characteristic of Sudbø’s choice of and collaboration with coauthors in subsequent works. The Commission here relies on written and oral feedback from almost 60 coauthors. A possible explanation of why the research cheating was not discovered earlier may partly be found in the fact that the coauthors mainly appeared as sub-suppliers or as senior guarantors (i.e. senior researchers who primarily played a central role on the overall level (idea, planning, compilation and the like) and who by reason of their professional weight have contributed to giving the works a professional legitimacy). Moreover, the authors have mainly related to Sudbø. In other words, there has apparently not been much communication “horizontally” between the various coauthors or research groups. Most of the communication took place via Sudbø. Sudbø has pointed out that he did not prevent communication between the coauthors in any way, but the Commission cannot see that he has contributed to or has called for such a contact between the coauthors. In this way, Sudbø has been in control of which set of data, which information and the like the individual coauthor possessed. It is worth pointing out that this information probably was limited. None of the coauthors seem to have had access to the underlying data material and patient data. They have probably not seen this as a natural or obvious part of the assignment given them either. However, such a distribution of work is not an unusual phenomenon in medical publications – a fact that obviously must be taken into account in the assessment of whether there is a basis for criticizing the coauthors. As an
example of this, reference is made to the collaboration between Jon Sudbø and J. Jack Lee in connection with the article which was published in New England Journal of Medicine 2004:
At one time Sudbø sends a radically changed file to J. Jack lee at MD Anderson41. The file shows “data cleaning along the way” as expressed by Lee in an email to Gunnar Sæter of January 13, 2006 after this case had become known. The Commission has compared two files: a) “sudbo8” which is the file that was used in the article in New England Journal of Medicine 2004 and b) “updated rawdata” which according to Lee is the first file he received from Sudbø for the 2004 article. The comparison was based on the variable patnid (patient number) which is on both files. The Commission has included the following variables: age, tobacco, year leukoplakia and year cancer. An excerpt of the 150 observations is provided in Table 3 in the report. As previously mentioned, it is conspicuous that “year leukoplakia” has changed from the first to the last file without age having changed. This is mathematics which does not tally if age is age at leukoplakia. In an email from Sudbø to Lee of August 8, 2003 Sudbø writes: ”One point: the age of the subjects have been recalculated to what they were at the time of initial diagnosis”. The latter ought to mean that age at leukoplakia is what is concerned, see section 4.2.7.
On a general basis, the Commission finds reason to note that many research fellows and researchers will have the opportunity to manipulate and even fabricate raw material, without either supervisors or coauthors being able to discover it. Research is indeed to a large degree based on trust, as it necessarily has to be. It is important and necessary that researchers have trust in one another, that one may trust that what others supply is genuine. The employer must be able to trust its employees, and journals must be able to trust researchers. Research participants and the public in general must also have trust in researchers. But such trust cannot be without limits. There must be a certain control and sound skepticism, because at one time or another warning lamps should start to flash thereby causing checks and controls to be made. But the fact that very few researchers find it obvious that any of their partners will fabricate research data may explain something of that which in retrospect easily may appear as failing alertness.
On the other hand, there are, as the Commission sees it, certain descriptions in the articles which more people should have reacted to. This may be coauthors, supervisors, superiors, critics, colleagues and others. The Commission would in particular refer to the account of the reclassification of the dysplasias which allegedly was made by four independent pathologists,42 as well as ploidy analyses which allegedly also was made by four independent observers.43 This was neither usual nor the case. The Commission also refers to the allegations relating to the reporting to the Cancer Registry, a follow-up program for patients and the allegation that one was in possession of almost complete smoking data. These are matters which one maybe
in retrospect should be able to say that someone should have discovered, or at least put a question mark at – both inside and outside the professional community at the institution, taking into account that this information was published in the internationally most renowned medical journal, namely the New England Journal of Medicine, which is remarkable for Norwegian researchers. At the same time, it must be admitted in the coauthors’ defense that the supervisor as well as coauthors in conversations with the Commission has had difficulties in and used much time to reconcile themselves with articles, which they themselves had taken part in and felt ownership to, being in fact based on fabricated data. The latter shows how unbelievable and unusual fabrication of research data in Norwegian research communities is perceived to be by researchers.