DiAna

DiAna standardizes reported drug names from the FDA Adverse Event Reporting System (FAERS) by translating diverse nomenclature into active ingredients and mapping them to WHO Anatomical Therapeutic Chemical (ATC) codes to support more accurate pharmacovigilance analyses.

Key Features:

• DiAna dictionary: Translates reported drug names (including brand names, active ingredients, abbreviations, and misspellings) into standardized active ingredients.
• Integration with ATC classification: Links each standardized ingredient to the WHO Anatomical Therapeutic Chemical (ATC) system and assigns a primary ATC code.
• FAERS data coverage: Processes FAERS data from 2004 to December 2022, encompassing 18,151,842 reports and 74,143,411 drug entries.
• Automated translation using RxNorm and string editing: Employs existing dictionaries such as RxNorm and string editing techniques to automatically translate names; applied to the first 14,832 terms (terms occurring ≥200 times, representing 96.88% of total drug entries) yielding 6,282 unique active ingredients.
• Extended automatic standardization: Extended standardization to an additional 346,854 terms, achieving coverage of 98.94% of total drug entries via unchecked automatic translations.
• Manual revision for accuracy: Conducted manual review and correction of automatically translated terms to improve translation accuracy.

Scientific Applications:

• Enhanced data control: Cleanses inconsistent drug names in FAERS to improve dataset quality and facilitate more accurate adverse event signal detection.
• Improved interpretability: Produces transparent and consistent drug name mappings that enhance interpretability of pharmacovigilance findings and safety studies.
• Support for disproportionality analysis and reproducibility: Provides standardized drug-to-ingredient mappings that improve the accuracy and replicability of disproportionality analyses.

Methodology:

Processed FAERS drug-name entries from 2004 to December 2022 (18,151,842 reports, 74,143,411 drug entries); performed automated translation using existing dictionaries (e.g., RxNorm) and string-editing techniques on high-frequency terms (first 14,832 terms → 6,282 unique active ingredients) and extended automatic translations to an additional 346,854 terms; mapped standardized ingredients to WHO ATC codes assigning primary ATC classifications.