Article révisé par les pairs
Résumé : Databases should ideally evolve over time along with changes in their environment according to their uses. Taking these changes into account has a strategic impact on administrative data quality, and therefore on the statistical information systems using them. In order to help manage the transformations resulting from the observable reality affecting data, this article proposes an innovative, operational approach that can be generalized to any relational database management system.Thanks to research improvements in data quality, studying anomalies and their management gave rise to an original prototype, called ATMS (Anomalies & Transactions Management System). This service allows tracking and process anomalies, supporting the back tracking method: in a preventive approach of data quality, the method intends to structurally improve quality at the source, and its implementation provides a significant return on investment. The characteristics of the ATMS prototype are combined with the use of data quality tools in curative approaches, offering new perspectives for statistical information systems.