[Quality assurance of scientific information and databanks].

The situation in cases of fraud and scientific misconduct--either true or false--is reviewed and its reflection in databases illustrated by examples. Exemplary quality guidelines for these cases have been published by the US National Library of Medicine, one of the major medical database producers. However, no timely and consequent application was found in some cases studied. On the whole, results in the major worldwide medical/life sciences databases are still dissatifactory. The implicite meaning of database quality (as illustrated by cases of fraud) for the decision making process, and its results for patients too--is broached and has to be discussed in the future. Some proposals for quality improvement are submitted.