The scientific value of multicenter studies is still questionable, as there are no methods available that allow to determine whether in fact identical items are recorded in the same way in different centers. Subjective classifications (for example according to the New York Heart Association (NYHA)), definitions of items, documentation habits and procedures may not be comparable in different hospitals. For that reason the concept of a "virtual patient" for calibration of database systems was developed as a new methodological approach. The "virtual patients" describes a computer program that includes and uses multimedia-tools to represent an exactly defined patient. A sound card enables a patient to talk and to represent intracardial and intrapulmonary sounds. Additional information including echocardiographic sequences, perioperative video sequences, x-ray analysis, angiograms, etc. is represented in the program. Thus during the perioperative course of a patient the complete relevant pre-, intra- and postoperative information regarding the patient is provided. Any physician participating in a multicenter study is asked to record the data of this virtual patient and to store it in a database system using the sheets applied in a multicenter study. As the patient is exactly defined (virtual patient) incorrect classifications and incomplete recordings, etc. can easily be detected and discussed. The virtual patient is a tool that allows calibration of database systems by direct comparison of the documentation technique between different physicians and different hospitals participating in multicenter studies. It is thus a multimedia-instrument to support standardization and quality assurance of database systems.