Representativeness of a German AI-enabled data network for secondary epidemiological analysis based on electronic health records - Research article

Sep 29, 2026

Upload by jayni

Study evaluating the representativeness of a German AI-enabled federated machine learning data network built on electronic health records across many hospitals. Basic network data were compared with publicly available DESTATIS statistics to assess scientific validity for future epidemiological analyses. A retrospective secondary analysis examined 8,106,105 cases, comparing age, gender, length of stay, ICD-10 diagnoses and OPS codes, and further evaluating substance-abuse codes and regional socioeconomic confounding. Results showed high concordance for most clinical variables and no evidence of socioeconomic or regional deviations.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages37

Reading Time1 h 33 min

AccessFree to download

Abstract

Introduction

Background

Methods

Results

Discussion

OPEN ACCESS Citation: Eichhorn S, Mitze FN, Wagner F, Seuthe IMC, Brinks R, Zimmermann A, et al. (2026) Representativeness of a German AI-enabled data network for secondary epidemiological analysis based on electronic health records. PLoS One 21(1): e0339647 . [https://doi.org/10.1371/journal.pone.0339647](https://doi.org/10.1371/journal.pone.0339647) [Editor:](Editor: Leonard Ighodalo Uzairue)[ Leonard Ighodalo Uzairue](Editor: Leonard Ighodalo Uzairue), De Montfort University Faculty of Health and Life Sciences, UNITED KINGDOM OF GREAT BRITAIN AND NORTHERN IRELAND Received: June 6, 2025 Accepted: December 9, 2025 Published: January 6, 2026 Copyright: © 2026 Eichhorn et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data availability statement: Due to the presence of potentially identifying hospital information, especially in regional analyses, the dataset cannot be made publicly available RESEARCH ARTICLE Representativeness of a German AI-enabled data network for secondary epidemiological analysis based on electronic health records Sabine Eichhorn1☯*, Franz Niklas Mitze1☯, Fritz Wagner1, Inga Marte Charlott Seuthe1, Ralph Brinks2, André Zimmermann3, Mehdi Dastur3, Josef Ladenbauer3, Jonas Jae-Hyun Park1 1 University of Witten/Herdecke, Department of Otorhinolaryngology and Head and Neck Surgery, St. Josefs-Hospital, Hagen, Germany, 2 University of Witten/Herdecke, Medical Biometry and Epidemiology, Witten, Germany, 3 Tiplu GmbH, Hamburg/Berlin, Germany ☯ These authors contributed equally to this work. * [sabine.eichhorn@uni-wh.de](sabine.eichhorn@uni-wh.de) Abstract Introduction The ongoing digitalization of medicine, increased computing power and low-cost storage capacities enable the use of AI-based algorithms for epidemiological big data analysis of electronic patient records. The aim of this study was to evaluate the representativeness of a data network with infrastructure for federated machine learning (ML) across numerous German hospitals. This was done by comparing basic data from the ML data network with publicly available data from the Federal Statistical Office (DESTATIS) to test the scientific validity for future epidemiological analyses. Methods In a retrospective epidemiological secondary analysis, 8,106,105 case files from the ML network were examined and compared to DESTATIS data regarding age, gender, length of hospital stay, ICD-10 diagnoses, and OPS codes. In addition, ICD-10 codes for substance abuse and the regional distribution were compared to examine socioeconomic confounders. Results The variables age, gender and length of stay, as well as the most common general ICD-10 and OPS codes and ENT-specific OPS codes, showed a high level of concordance based on clinical relevance. For the ENT-specific ICD-10 codes, 2 out of 11 of the most frequent codes showed a maximum deviation of 3.71% . The analysis of socioeconomic factors and regional distribution showed no evidence for deviations. PLOS One | [https://doi.org/10.1371/journal.pone.0339647](https://doi.org/10.1371/journal.pone.0339647) January 6, 2026 1 / 37 in compliance with ethical standards and institutional privacy policies. De-identified datasets may be made available upon reasonable request, subject to approval by the responsible data protection body. Data access is regulated by Tiplu GmbH. Requests may be directed to: Stephanie Hoh, Information Security Officer at Tiplu GmbH, [s.hoh@tiplu.de](s.hoh@tiplu.de). Funding: This work was supported by the German Federal Ministry of Education and Research (BMBF) and the European Union via the program NextGenerationEU as part of the project geNuGes (Funding No. 16DTM309)(Figs 26 and 27) . M.D. was responsible for acquiring the funding. The funders had no role in the

FAQs

What was the study’s main goal?

To evaluate how representative a German AI-enabled federated ML data network is for secondary epidemiological analyses based on electronic health records.

Which data sources and comparisons were used?

The network’s basic data were compared with publicly available DESTATIS data, using age, gender, length of hospital stay, ICD-10 diagnoses, and OPS codes, plus substance-abuse codes and regional distribution.

What did the results show about the network’s representativeness?

Most examined variables showed high concordance with clinical relevance; socioeconomic factors and regional distribution showed no evidence of deviations.

Sep 29, 2026

Upload by jayni