Introduction: Characterization of regional variation in asthma treatment across Germany will improve
understanding of real-world therapy and inform physician education. We aimed to describe
regional variation in asthma prevalence and oral corticosteroid (OCS) use in Germany.
Methods: We developed a machine learning gradient boosted tree model, with 91% accuracy in
predicting the presence of asthma and chronic obstructive pulmonary disease, using
IMS® Disease Analyzer electronic medical records, which cover 3% of German patients. This
model was applied to the IMS Longitudinal Prescription database, with 75% national
coverage, to classify patients receiving airflow obstruction treatment from Oct 2017
to Sept 2018 in 63 regions.
Results: Of 2.4 million patients predicted to have asthma, 13.7%, 18.7%, 36.5%, 29.4%, and
1.7% were categorized as Global Initiative for Asthma (GINA) stages 1, 2, 3, 4, and
5, respectively. At least one short-term OCS prescription was received by 7 – 15%
of patients in GINA stages 1 – 4 and 35% of patients in stage 5. At least one maintenance
OCS prescription was received by 1 – 3% of patients in GINA stages 1 – 4 and 86% in
stage 5. Cumulative OCS dosage and percentage of OCS-treated patients differed substantially
across regions ([Fig. 1]). Regions with less maintenance OCS use had greater biologic use.
Fig. 1 Cumulative Annual OCS Dosage (mg Prednisolone per Patient with Prescription) for
OCS-Treated Patients with Asthma. OCS, oral corticosteroids; Rx, prescription.
Conclusions: Maintenance and short-term OCS use varied across Germany and were greater than recommended
by guidelines across the spectrum of asthma severity.
