What should a journal do if an old (more than 15 years old) published paper is flagged on PubPeer for image concerns, but the case cannot be resolved due to the time lapsed? For example, if only low quality images are available online that cannot be analysed conclusively; some of the key authors may no longer be contactable; the raw data is no longer available; an institutional investigation is unlikely.
Questions for the Forum
- After how many years is it acceptable to leave a case unresolved because it would be ineffectual to pursue an investigation? Is there a cut-off point?
- If there is a cut-off point then should the journal respond to the comment/concern, for example, by posting a note on PubPeer to say that due to the age of the paper and without raw data it is not possible to verify the issues?
Journals are receiving increasing numbers of requests like this from readers asking for action to be taken on old papers. The difficulty is that the data are often unretrievable (for example, if the author has died), or have been digitised from an analogue format and are very poor quality. In such cases it may not be possible to establish definitively if falsification or fabrication has taken place or not, and therefore whether to correct or retract the article. Editors may also feel uncomfortable applying current expectations and practices to articles produced some time ago. While there is no expiry period for an investigation or retraction, therefore, journals should make a decision on their actions on a case by case basis, and be honest about the limits of what they can do. If necessary they should feel able to inform the person raising the concern that they have done all they are able to do.
If the article or its findings are felt to be of sufficiently high impact, then the journal should make an effort to investigate the concerns, both for the integrity of the published record, and to avoid a distorting impact on systematic reviews and citations. Some image issues can be fairly easily verified even in low-quality images. Editors might also publish the data that are available, or produce an editor’s note or an expression of concern depending on the severity of the issues. If there is sufficient evidence to bring the research conclusions into doubt, then a retraction may be warranted. An approach to the institution could also be useful, depending on the time elapsed and the institution (or funder)’s policies on data retention. However, it was also recognised that journals have limited resources, and that editors may need to make responsible choices about which concerns are worth investigating.
On the specific case being considered it was not possible to verify the allegations due to low resolution of the figures and no available raw data. The case could not be formally resolved as such and is now considered closed. It was decided not to publish an expression of concern.