Real-world drug use in ATC and ICD-10: an expert-curated drug-diagnosis resource based on UK primary care and Danish hospitalization electronic health records
- 1University of Copenhagen
Abstract
As the use of electronic health records in drug repositioning research increases, so does the need for a well-curated resource describing real-world drug-diagnosis relationships. This need is particularly important in the context of polypharmacy. Although literature-based drug-disease maps exist, they are typically based on mechanistic disease ontologies, which are not widely used in clinical settings and do not align well with the ICD system, which is most often used in healthcare. Here, we used real-world primary and secondary healthcare data from approximately 1.5 million individuals to identify drug-diagnosis co-occurrences ( 736,000 pairs), significant associations (7,763 pairs), and assess direct drug usage through medical expert curation. The final resource comprises 7,763 associations with odds ratios > 3.5, manually annotated by six independent clinicians (3 in the UK and 3 in Denmark). Finally, the clinician annotations were scored using an Expectation-Maximization-based framework providing a confidence score for each pair. Our resource shows that the vast majority of significantly associated drugs and diagnoses in healthcare records are not due to direct treatment of the diagnosis. Additionally, through the annotation results, we demonstrate the importance of accounting for systematic differences among annotators when working with real-world data.
Related
- Diagnosis-code co-occurrence patterns and their associations with hospitalization costs among patients with lung cancer: a province-wide retrospective study in Sichuan, China
- How age and underlying conditions affect incidence of hospitalization and death after documented COVID-19 diagnosis among individuals aged >=20 years, a retrospective cohort study using PCORnet, April 2022 to March 2025
- Why do patients bypass primary care? Trust, resource concentration, and risk in China's tiered healthcare system: A comparative policy perspective
- AI-Based Screening Approach for Primary Aldosteronism Using Longitudinal Electronic Health Records
- Missed Opportunities for Chronic Kidney Disease Recognition in Irish Primary Care: A Nationwide Electronic Health Record Study
BibTeX
@misc{louloudis2026realworld,
title = {{Real-world drug use in ATC and ICD-10: an expert-curated drug-diagnosis resource based on UK primary care and Danish hospitalization electronic health records}},
author = {I. Louloudis and H. Currant and C. Ytsma and S. C. Lindgaard and A. D. Haue and W. Chen and S. J. Y. Thio and M. Pisliakova and S. Brunak and S. Denaxas},
howpublished = {medRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.06.26362128},
url = {https://www.medrxiv.org/content/10.64898/2026.09.06.26362128v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Real-world drug use in ATC and ICD-10: an expert-curated drug-diagnosis resource based on UK primary care and Danish hospitalization electronic health records
AU - I. Louloudis
AU - H. Currant
AU - C. Ytsma
AU - S. C. Lindgaard
AU - A. D. Haue
AU - W. Chen
AU - S. J. Y. Thio
AU - M. Pisliakova
AU - S. Brunak
AU - S. Denaxas
PY - 2026
DA - 2026/09/06
PB - medRxiv
DO - 10.64898/2026.09.06.26362128
UR - https://www.medrxiv.org/content/10.64898/2026.09.06.26362128v1
AB - As the use of electronic health records in drug repositioning research increases, so does the need for a well-curated resource describing real-world drug-diagnosis relationships. This need is particularly important in the context of polypharmacy. Although literature-based drug-disease maps exist, they are typically based on mechanistic disease ontologies, which are not widely used in clinical settings and do not align well with the ICD system, which is most often used in healthcare. Here, we used real-world primary and secondary healthcare data from approximately 1.5 million individuals to identify drug-diagnosis co-occurrences ( 736,000 pairs), significant associations (7,763 pairs), and assess direct drug usage through medical expert curation. The final resource comprises 7,763 associations with odds ratios > 3.5, manually annotated by six independent clinicians (3 in the UK and 3 in Denmark). Finally, the clinician annotations were scored using an Expectation-Maximization-based framework providing a confidence score for each pair. Our resource shows that the vast majority of significantly associated drugs and diagnoses in healthcare records are not due to direct treatment of the diagnosis. Additionally, through the annotation results, we demonstrate the importance of accounting for systematic differences among annotators when working with real-world data.
ER -