Drug-Drug Interaction Dataset Derived from FDA 2026 DailyMed Drug Labels

This dataset provides a comprehensive and up-to-date resource of drug–drug interactions (DDIs) extracted from FDA Structured Product Labeling (SPL) documents available through DailyMed. The data were obtained from the January 2026 release of human prescription drug labels and processed using a reproducible natural language processing (NLP) pipeline.

The dataset captures interactions by identifying co-occurring drug pairs and associated interaction terms within the same sentence of drug labeling sections, ensuring clinically relevant contextual relationships. Drug names were standardized using curated synonym mappings to active ingredients, and each drug was enriched with external identifiers including PubChem Compound Identifiers (CIDs) and Anatomical Therapeutic Chemical (ATC) classification codes to enhance interoperability with other biomedical databases.

In total, the dataset includes 2,038 unique drugs and more than 90,000 interaction pairs, substantially expanding coverage compared to legacy resources such as TWOSIDES (last updated in 2019). Unlike existing datasets, this resource reflects recent pharmaceutical developments, including drugs approved after the COVID-19 pandemic, and is directly aligned with U.S. FDA regulatory labeling.