iDrug: Text mining driven international adverse drug-drug interactions

Abstract

Abstract iDrug: Text mining driven international adverse drug-drug interactions

Saejung Kim Electrical and Computer Engineering The Graduate School Seoul National University

With the remarkable development of medical analysis techniques, many drug-drug interactions (DDI) have been identified. Especially, as drugs with adverse DDIs can bring severe side-effects when used together, this information should be given with priority to medical experts in an emergency situation such as in intensive care unit (ICU). However, drug compendia that provide drug-drug interaction severity often have inconsistencies of DDI severity among different compendia. Therefore, it is critical to identify the correct DDI severity and organize such information in a consistent and structured way.

We applied text mining techniques to the extraction of drug-drug interactions from prescribing information in six countries (the United States, the United Kingdom, Canada, France, Switzerland, and Japan), and constructed a DDI severity database that stores two levels of DDI severity, contraindications and precautions, for the six countries. Based on the information we extracted, we analyzed the characteristics of the DDI network for each nation. We computed the pairwise differences of adverse DDIs between nations to understand the inconsistency levels of the DDIs of each country compared to the other countries. To the best of our knowledge, our work is the first attempt to automatically extract PI from the national regulatory authorities of the six countries by using text-mining techniques. Since our data were populated from PI approved by national regulatory authorities, we believe that our database can serve as a more authentic and reliable guide for medical experts.

keywords : Adverse drug-drug interaction, Text mining Student Number : 2014-21735