25gmds040 10.3205/25gmds040 urn:nbn:de:0183-25gmds0403 Meeting Abstract Neumann Neumann Daniel D

Institute for Medical Informatics, Statistics and Epidemiology (IMISE), Leipzig University, Leipzig, Germany

author Kesselmeier Kesselmeier Miriam M

Institut für Medizinische Statistik, Informatik und Datenwissenschaften (IMSID), Universitätsklinikum Jena, Jena, Germany

author Redeker Redeker Louisa L

Department of Clinical Pharmacology, School of Medicine, Faculty of Health, Witten/Herdecke University, Witten, Germany

author Schmidt Schmidt Florian F

Institute for Medical Informatics, Statistics and Epidemiology, Leipzig University, Leipzig, Germany

author Peschel Peschel Thomas T

Institute for Medical Informatics, Statistics and Epidemiology (IMISE), Leipzig University, Leipzig, Germany

author Haerterisch Haerterisch Steffen S

Hospital Pharmacy, University Medical Center Hamburg-Eppendorf, Hamburg, Germany

author Thalheim Thalheim Torsten T

Interdisciplinary Centre for Bioinformatics, University Leipzig, Leipzig, Germany

author Meineke Meineke Frank A. FA

Universität Leipzig, Leipzig, Germany

author Schmiedl Schmiedl Sven S

Helios University Hospital Wuppertal, Chair of Clinical Pharmacology, University Witten/Herdecke, Witten/Herdecke, Germany

author Thürmann Thürmann Petra A. PA

Helios University Hospital Wuppertal, Chair of Clinical Pharmacology, University Witten/Herdecke, Witten/Herdecke, Germany

author Scherag Scherag André A

Institut für Medizinische Statistik, Informatik und Datenwissenschaften, Universitätsklinikum Jena, Jena, Germany

author Löffler Löffler Markus M

Universität Leipzig, Med. Fakultät, Leipzig, Germany

author German Medical Science GMS Publishing House

Düsseldorf

610 Medical Informatics Initiative Germany (MII) distributed computing clinical routine data data quality plausibility checks health data research data validation 20251103 engl This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License. Dieser Artikel ist ein Open-Access-Artikel und steht unter den Lizenzbedingungen der Creative Commons Attribution 4.0 License (Namensnennung). M0631 040 Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie 70. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS) V: Gesundheitsdatennutzung – data quality Jena 20250907 20250911 Abstr. 373 TextIntroduction: The German Medical Informatics Initiative (MII) has established a nationwide FHIR-based core dataset and local Data Integration Centers (DICs) to enable secondary use of routine clinical data while preserving data sovereignty and privacy . Amitriptyline, a tricyclic antidepressant with strong anticholinergic and sedative effects, is classified as a potentially inappropriate medication (PIM) in patients ≥65 years due to elevated risks of delirium and falls . We aimed to quantify amitriptyline prescriptions in elderly inpatients across multiple university hospitals and to assess the reliability of the corresponding FHIR-extracted data via a decentralized validation pipeline.Methods: In this retrospective multicenter study, ten university hospitals participating in POLAR_MI deployed identical R scripts (using the fhircrackr package) to query FHIR servers at their DICs for all hospital encounters of patients aged ≥65 years and to count documented amitriptyline orders , . Only aggregate counts were shared centrally. To validate data quality, each site performed bidirectional plausibility checks on a random sample of 20 patients: manual chart review in the EMR versus FHIR-extracted results, covering inpatient (during stay) and outpatient (prior to or within 24 h of admission) use. Discrepancies were reconciled locally and fed back into the extraction logic. Central meta-analysis employed a random-effects model to pool prescription rates and quantified inter-site heterogeneity (I²). Data-quality metrics included sensitivity, specificity, and Krippendorff’s alpha for interrater agreement.Results: Among ~800 000 encounters, 414 587 involved patients ≥65 years; medication data were available for 97 917 cases. Pooled amitriptyline prevalence was 0.6% (95% CI: 0.3–1.3%; I² = 96.6%). Inpatient prescribing checks (n=158) achieved 100% sensitivity and specificity (Krippendorff’s α = 0.98). Outpatient checks showed 82.7% accuracy (sensitivity 46.8%; specificity 100%; α = 0.70). Misclassification analysis revealed normalized mutual information of 0.95 for inpatient and 0.50–0.72 for outpatient data, driven primarily by unstructured free-text entries.Conclusion: Our FHIR-based, decentralized extraction and bidirectional validation approach proved feasible and reliable for quantifying amitriptyline use in elderly inpatients. High accuracy for inpatient data demonstrates robust performance of the MII infrastructure, while lower outpatient sensitivity highlights the need for improved structured documentation. This scalable, privacy-preserving pipeline offers a template for large-scale pharmacovigilance and quality-of-care studies using routine EMR data.The authors declare that they have no competing interests.The authors declare that an ethics committee vote is not required.The contribution has already been presented: MIE'25, 2025 May 19-21, Glasgow, Scotland (STHI) Semler SC Wissing F Heyder R German Medical Informatics Initiative 2018 Methods Inf Med e50–e56 Semler SC, Wissing F, Heyder R. German Medical Informatics Initiative. Methods Inf Med. 2018;57(S 01):e50–e56. Davies LE Spiers G Kingston A Adverse outcomes of polypharmacy in older people: systematic review of reviews 2020 J Am Med Dir Assoc 181–187 Davies LE, Spiers G, Kingston A, et al. Adverse outcomes of polypharmacy in older people: systematic review of reviews. J Am Med Dir Assoc. 2020;21(2):181–187. Scherag A Andrikyan W Dreischulte T POLAR – “POLypharmazie, Arzneimittelwechselwirkungen und Risiken”: wie können Daten aus der stationären Krankenversorgung zur Beurteilung beitragen? Prävention Gesundheitsf 2022 Scherag A, Andrikyan W, Dreischulte T, et al. POLAR – “POLypharmazie, Arzneimittelwechselwirkungen und Risiken”: wie können Daten aus der stationären Krankenversorgung zur Beurteilung beitragen? Prävention Gesundheitsf. 2022;17(Suppl):1704–1708. Palm J Meineke FA Przybilla J Peschel T fhircrackr: An R package unlocking FHIR for statistical analysis 2023 Appl Clin Inform 54–64 Palm J, Meineke FA, Przybilla J, Peschel T. fhircrackr: An R package unlocking FHIR for statistical analysis. Appl Clin Inform. 2023;14(1):54–64. 0 0 0 0