Modeling and acquisition of drug-drug interaction knowledge.

Objectives: The effectiveness of computerized clinical decision support systems (CDSS) depends on the quality of the knowledge they refer to. In this article, we are interested in the acquisition, modeling and representation of the knowledge embedded in the "national reference framework of drug-drug interaction" published by the French Health Products Safety Agency.

Methods: A model of drug-drug interactions has been designed using bottom-up and top-down approaches. This model is the basis for the design of an XML format to represent and extract information on drug interactions from the reference framework.

Results: A specific tool has been developed to extract the information from a corpus of 1053 drug monographs using a methodology similar to the one used by the GEM-Cutter tool to extract information from clinical guidelines. Strategies to integrate the XML files produced into CDDSSs are discussed.

Discussion-conclusion: Modeling and acquisition of drug-drug interaction knowledge from a corpus of drug monographs is a potential approach to foster the development of CDSS and improve their specificity.