NLM's practices for handling errata and retractions.

The keystone of the scientific method is solid experimental design and reproducible results. The publishing of findings advances knowledge and establishes the basis for further research. In recent years, the foundations of this principle have been shaken as a small, but significant portion of the scientific literature is being flawed by the appearance of fraudulently produced research. Potentially as damaging are errors that result from poor editing and proofreading. Fraudulent articles and errors lead, at best, to misunderstandings and, at worst, to dire consequences in the treatment of patients. Errata and retraction notices are generally carried in the published literature but usually are not linked to the original data. Database producers, such as NLM, have the means to establish this link and to inform users of incorrect information in the source documents as well. This paper reports NLM's experience in bringing published retraction and errata notices to the public's attention and relates this experience to the library's overall interest in quality assurance.