Drug Coding to standardized terminologies is a crucial data processing step to enable structuring drug information from various data sources such as Electronic Health Records (EHRs) systems, Electronic Data Capturing systems for clinical trials and spontaneous reporting databases for post-marketing surveillance of drugs. WHODrug Global is a resource for drug coding in international databases. It has global coverage of trade names from 167 different drug markets and is developed and maintained by the Uppsala Monitoring Centre (UMC), the WHO Collaborating Centre for International Drug Monitoring. WHODrug Global contains information about medicinal products and active substances intended for human and medicinal use, both of conventional and natural origin.
In postmarketing surveillance and clinical trials, drug coding involves coding the verbatim describing the active substance or product used by the patient to a drug terminology. Additional information, such as the strength, the dosage form or the indication of the product may be used to inform the coding decision. Since the verbatim might be entered manually, it can contain abbreviations, misspellings, and the trade name, with or without the corresponding active ingredients, and might not be formatted consistently, depending on data entry conventions. Drug coding can also be challenging when the trade names used are ambiguous. Aircort, for example, is a trade name that is marketed in both Morocco and Italy; however, while it represents a product containing Beclometasone dipropionate in Morocco, Aircort represents a product containing Budesonide in Italy.
Tools for drug coding through direct matches of the verbatim, with or without transformation, to drug terminologies, can be referred to as auto-encoders and the process itself auto-coding or auto-encoding. Direct matches are coded automatically to the drug terminology of interest and synonym lists are commonly created by organizations during manual coding to record and reapply coding decisions. However, manually coding terms that are not auto coded can take a significant amount of time.
More advanced, AI-based systems may help in dealing with ambiguities and cut down on time. WHODrug Koda is an automated coding engine custom-built by UMC. Koda is one of the first AI-based drug coding systems available for use.1 The purpose of Koda is to assist coders in interpreting free-text drug information and selecting the most appropriate drug name in WHODrug. Koda can scale-up coding capacity and support drug coders in their manual work. Eclipse coding integrates with WHODrug Koda from UMC.
Have questions or need support with Koda™? Contact us today to learn how Eclipse™ can enhance your clinical coding efficiency.