A paper suggested that a cluster of symptoms, signs, and tests could be combined to diagnose pneumonia in general practice. The paper was rejected after being read by two editors, because it was preliminary and had not been validated in an independent population. The authors submitted a new manuscript the following year, describing the same patients and focusing on the accuracy of individual symptoms and test results in differentiating bacterial from viral chest infections. This time the paper was reviewed and considered at a selection committee before being rejected. The authors did not refer to their previous paper in the covering letter. The next year the authors submitted another new study describing the same patients, but this time exploring which symptoms and signs were associated with an antibiotic prescription and whether this was appropriate. Again they did not mention either of their previous manuscripts in the covering letter. The overlapping resubmissions were detected when one of the readers recognised the subject matter and authors’ names, which prompted a search.
- The authors should have pointed out that the paper was a resubmission. They might have thought that the volume of papers received by the journal would have made it difficult to detect multiple submissions. - The critical point is the difference between each version of the paper as it could be argued that if the authors had improved it, then resubmission would be permissible. - Sometimes authors will not take no for an answer, which is an irritant for journals. The onus is on journals to ensure that there are systems in place to detect such resubmissions. - The editors should write to the authors, explaining that multiple submission of a rejected paper is regarded as poor practice.