Treatment data from practices and specialization centers, especially in the increasingly specialized areas which university clinics do not cover, are very important for evaluating the effectiveness and efficiency of dental examination and treatment methods. In the case of paper-based documentation, the evaluation of these data usually fails because of the cost it entails. With the use of electronic medical records, this expense can be markedly lower, provided the data acquisition and storage is structured accordingly. Since access to sensitive person-related data is simplified considerably by this method, such health data are protected, especially on the European level.Other than generally assumed, this protection is not restricted solely to the confidentiality principle, but also comprises the power of disposition over the data (data protection). The result is that from a legal point of view, the treatment data cannot be readily used for scientific studies, not even by dentists and physicians who have collected the data legally during the course of their therapeutic work. The technical separation of treatment data from the personal data offers a legally acceptable solution to this problem. It must ensure that a later assignment to individual persons will not be feasible at a realistic expense ("effective anonymization"). This article describes the legal and information technology principles and their practical implementation, as illustrated by the concept of a respective compliant IT architecture for the dentaConcept CMDfact diagnostic software. Here, a special export function automatically separates the anonymized treatment data and thus facilitates multicentric studies within an institution and among dental practices. Keywords: person-related data, treatment data, virtual anonymization, IT architecture, relational databases