Premium
Similarity and consistency assessment of three major online drug–drug interaction resources
Author(s) -
Kontsioti Elpida,
Maskell Simon,
Bensalem Amina,
Dutta Bhaskar,
Pirmohamed Munir
Publication year - 2022
Publication title -
british journal of clinical pharmacology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.216
H-Index - 146
eISSN - 1365-2125
pISSN - 0306-5251
DOI - 10.1111/bcp.15341
Subject(s) - drug , drug drug interaction , similarity (geometry) , consistency (knowledge bases) , pharmacology , medicine , computer science , computational biology , biology , artificial intelligence , image (mathematics)
Aims The aim of this study was to explore the level of agreement on drug–drug interaction (DDI) information listed in three major online drug information resources (DIRs) in terms of: (1) interacting drug pairs; (2) severity rating; (3) evidence rating; and (4) clinical management recommendations. Methods We extracted information from the British National Formulary (BNF), Thesaurus and Micromedex . Following drug name normalisation, we estimated the overlap of the DIRs in terms of DDI. We annotated clinical management recommendations either manually, where possible, or through application of a machine learning algorithm. Results The DIRs contained 51 481 ( BNF ), 38 037 ( Thesaurus ) and 65 446 ( Micromedex ) drug pairs involved in DDIs. The number of common DDIs across the three DIRs was 6970 (13.54% of BNF , 18.32% of Thesaurus and 10.65% of Micromedex ). Micromedex and Thesaurus overall showed higher levels of similarity in their severity ratings, while the BNF agreed more with Micromedex on the critical severity ratings and with Thesaurus on the least significant ones. Evidence rating agreement between BNF and Micromedex was generally poor. Variation in clinical management recommendations was also identified, with some categories (i.e., Monitor and Adjust dose ) showing higher levels of agreement compared to others (i.e., Use with caution , Wash‐out , Modify administration ). Conclusions There is considerable variation in the DDIs included in the examined DIRs, together with variability in categorisation of severity and clinical advice given. DDIs labelled as critical were more likely to appear in multiple DIRs. Such variability in information could have deleterious consequences for patient safety, and there is a need for harmonisation and standardisation.