A stumble in data preparation earned a retraction for a paper on delirium tremens, a life-threatening side effect of alcohol withdrawal that spans a wide range of symptoms, including hallucinations and seizures.
Though the initial retraction notice was extremely unhelpful, the author stepped in to give us a better picture of the errors that led to the paper’s demise.
Here’s the notice from Alcohol and Alcoholism about “Biochemical Predictors of Delirium Tremens in Patients in Alcohol Withdrawal”:
The Editors-in-Chief and the Publishers have withdrawn this article after the authors re-visited the paper and found an error in the data that was originally submitted.We would like to apologize for any inconvenience this causes to readers of the journal.
And a more thorough explanation from author Devavrat Harshe of Padmashree Dr D Y Patil Hospital, Nerul, Navi Mumbai, India:
The paper in question was the first of 2 observational studies conducted to assess the course of alcohol withdrawal. During the initial data analysis and manuscript preparation, there was an error in the grouping of a set of patients under mutually exclusive study groups after the data entry was completed. This yielded erroneous mean scores on comparison. Since the initial grouping itself was wrong, further group distribution about each predictor and variable and specific predictors’ analysis also yielded wrong results. It was my responsibility to preserve the scientific accuracy of the manuscript and to re-run the analyses during the pre-publication review, and I take responsibility and express my apologies for this grave error. Although the scores obtained were different in 2 analyses, the conclusion of the paper remained unchanged. However that is definitely not an excuse. We informed the journal of this error along with the changes and the raw data and enquired whether the error can be corrected in a corrigendum or a retraction.
Once again, I express my apologies to the scientific community for the error. We have rectified the errors in the manuscript, verified it with the data and are planning to submit it for publication after our ongoing research projects are completed.