Fitness for Use of Anatomical Therapeutic Classification for Real World Data Research

Real-world data (RWD) is gaining importance in research. For instance, the European Medicines Agency (EMA) is currently in the process of establishing a cross-national research network that utilizes RWD for research. However, data harmonization across countries must be carefully considered to avoid misclassification and bias.

This paper aims to investigate the extent to which a correct assignment of RxNorm ingredients is possible for medication orders that include only ATC codes.

In this study, we analyzed 1,506,059 medication orders from the University Hospital Dresden (UKD) and merged them with the ATC vocabulary in the Observational Medical Outcomes Partnership (OMOP) including relevant relationship mappings to RxNorm.

We identified 70.25% of all medication orders were single ingredients with direct mapping to RxNorm. However, we also identified a significant complexity in mappings for the other medication orders that was visualized in an interactive scatterplot.

The majority of medication orders under observation (70.25%) are single ingredients and can be standardized to RxNorm, combination drugs pose a challenge due to the different approaches of ingredient assignments in ATC and RxNorm. The provided visualization can help research teams gain a better understanding of problematic data and further investigate identified issues.