Article type
Year
Abstract
Background: It is essential that the processes used for literature searching and selecting studies for inclusion reliably identify retracted papers so they are not unwittingly included in systematic reviews. There is, however, evidence that retracted papers continue to be cited without reference to the retraction. In 2009 multiple papers authored by Dr Scott S Reuben were retracted due to falsification of data. We used these citations to investigate whether databases and journals recorded the information on retraction appropriately.
Objectives: To investigate whether three bibliographic databases recorded notices of retraction in a clear and timely way and whether electronic journals displayed information about retractions in a way that prevented inadvertent use of fraudulent research data.
Methods: We searched MEDLINE, EMBASE and CENTRAL databases for 18 papers by Reuben that were known to have been formally retracted, based on retraction notices. We noted whether the database recorded that each of the articles was retracted. The 18 papers were published across nine journals. We retrieved one paper from each of the nine electronic journals and assessed whether it was possible to identify the study as being retracted.
Results: In our case study all of the retracted papers identified on MEDLINE had been annotated appropriately as compared to 6% of EMBASE records and 80% of CENTRAL records. Similarly, there was a high level of variability across journals in how clearly they annotated that a paper had been retracted.
Conclusions:Lack of consistency in how databases and electronic journals record retractions and the time taken to record the retractions make it difficult for researchers and information specialists to identify retracted studies and extra care is needed to ensure that they are not included in systematic reviews.
Objectives: To investigate whether three bibliographic databases recorded notices of retraction in a clear and timely way and whether electronic journals displayed information about retractions in a way that prevented inadvertent use of fraudulent research data.
Methods: We searched MEDLINE, EMBASE and CENTRAL databases for 18 papers by Reuben that were known to have been formally retracted, based on retraction notices. We noted whether the database recorded that each of the articles was retracted. The 18 papers were published across nine journals. We retrieved one paper from each of the nine electronic journals and assessed whether it was possible to identify the study as being retracted.
Results: In our case study all of the retracted papers identified on MEDLINE had been annotated appropriately as compared to 6% of EMBASE records and 80% of CENTRAL records. Similarly, there was a high level of variability across journals in how clearly they annotated that a paper had been retracted.
Conclusions:Lack of consistency in how databases and electronic journals record retractions and the time taken to record the retractions make it difficult for researchers and information specialists to identify retracted studies and extra care is needed to ensure that they are not included in systematic reviews.