The shift from version to version 10 of the ICD-code ("International Statistical Classification of Diseases and Related Health Problems") causes enormous problems for the exploitation of medical data warehouses, since conventional data warehouses do not support the change of the structure of dimensions, i.e.the content and relationships of master data like the diagnostic codes, or other key values. This shortcoming results in a reduction of possible analysis, and unfortunately is the cause of many wrong statistics and analysis results. In this paper we analyze the problem and show how to superimpose conventional multidimensional data warehouses with temporal master data to allow queries spanning multiple periods to return correct answers.