Modeling of disease trajectories and influencing factors to create prediction models for clinical endpoints in patients with asthma and COPD – A multicenter consortium of the Medical Informatics Initiative (MII)

Authors: Hannah Lemper1, Sabine Feig1, Irina Lehmann2, Roland Eils2, Thomas Ganslandt3, Christian Taube4, Gernot Rohde5, Daiana Stolz6, Harald Binder7, Henning Schneider8, Frank Ückert9, Anna-Maria Dittrich10, Robert Bals11,Konstantin Strauch12, Hans Kestler13, Harald Renz1

Affiliation: 1. Institute of Laboratory Medicine and Pathobiochemistry, Molecular Diagnostics, Philipps University of Marburg, Marburg, Germany. 2. Digital Health Center, Berlin Institute of Health (BIH) at Charité – Universitätsmedizin Berlin, Berlin, Germany. 3. Chair of Medical Informatics, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany. 4. Department of Pulmonary Medicine, University Hospital – Ruhrlandklinik, Essen, Germany. 5. Medical Department I, Department of Respiratory Medicine, Goethe University Hospital Frankfurt/Main, Germany. 6. Department of Pneumology, University Medical Centre Freiburg, Germany. 7. Institute of Medical Biometry and Statistics, Faculty of Medicine and Medical Centre – University of Freiburg, Germany. 8. Institute of Medical Informatics, Justus-Liebig-University Giessen – Medicine Faculty, Giessen, Germany. 9. Applied Medical Informatics, University Medical Center Eppendorf, Hamburg, Germany. 10. Department for Pediatric Pneumology, Allergy and Neonatology, Hannover Medical School (MHH), Hannover, Germany. 11. Department of Internal Medicine V – Pulmonology, Allergology, Critical Care Medicine, Saarland University Medical Centre, Saarland University Hospital, Homburg/Saar, Germany. 12. Institute of Medical Biostatistics, Epidemiology and Informatics (IMBEI), University Medical Center, Johannes Gutenberg University, Mainz, Germany 13. Institute of Medical Systems Biology, Ulm University, Ulm, Germany.

Precision medicine approaches in the context of COPD and asthma are the objective of the Module 3 project „CALM QE“ of the Medical Informatics Initiative (MII). When investigating influencing factors and predictive parameters, it is important to include data sets from routine care and direct patient monitoring, so-called “real-world data”. Prediction models that include disease trajectories as well as influencing factors and relevant clinical outcomes can serve as support when making treatment decisions. The CALM QE project focuses on the development of such models for individual disease progression. In addition to processing data from inpatient care, data from outpatient care and smart wearables will also be analyzed. Important factors influencing the course of the disease as well as diseasetypical phenotypes and endotypes will be investigated. Risk factors for severe courses are also highlighted. Methodologically, various statistical models and machine learning methods are used in this context. In the future, the prediction models developed should help to improve risk assessment when making treatment decisions and to identify risk constellations at an early stage, thereby reducing the occurrence of exacerbations as far as possible as part of personalized treatment approaches. The CALM QE project involves interdisciplinary and cross-sectorial cooperation between clinics and institutes at university facilities at twelve locations in Germany as well as outpatient practices and patient organizations. In the project, inpatient patient data and outpatient data from the hospital’s own outpatient clinics are extracted and evaluated. Both a retrospective approach, in which data from recent years is processed, and a prospective longitudinal approach with an additional survey of further assessment criteria such as quality of life, are being pursued. The aim is to collect a meaningful and comprehensive clinical and diagnostic data set for evaluation and model development. In addition, the data under consideration will be compared with a large research data set from the German multicenter COPD cohort study COSYCONET. Environmental and weather data will also be included. In further sub-projects, data is collected from patients in private practice and from pediatric care in order to include not only severe courses of disease but also a broad spectrum of disease trajectories and to obtain a differentiated picture across the various age groups