A randomised controlled trial was submitted, showing that a new treatment, which is a combination of familiar compounds, is highly beneficial in a common but largely untreatable problem. The authors came from several different countries and included people from the company that manufactures the treatment. The editors had great difficulty finding reviewers for the paper as many simply returned it, saying that they could not produce an opinion. The reviewer who did eventually do so said that the results were not credible and that all the signs suggested that the paper might be fraudulent. The statistical adviser was asked for his opinion, and, although he agreed that the results were very unlikely, he was less convinced that there had been any data manipulation. He suggested that the editors request the raw data. The editors were unsure what to do at this point. Their previous experience of asking for raw data was that it involved a highly complex and very expensive exercise. They wondered if instead they should simply alert the authors’ employers—there are six different employers from four different countries—and ask them to investigate. The editors almost certainly did not want to publish the trial.
_ The editor should write to the authors saying that a reviewer has expressed some concern about the data. _ Request the raw data from the authors, specifying that it should be in an electronic form.
The editors rejected the paper but also requested the raw data. The authors have submitted the raw data on CD and these are currently being analysed by a statistician.