The above-mentioned project aims to investigate the relationship between the cardiac biomarker NT-proBNP and atrial fibrillation in everyday clinical practice. To this end, a distributed computing architecture using the DataSHIELD system will be used to analyze retrospective data. The data will be managed by the respective DIZ and made available in a separate database specifically for the project. In addition to specially authorized persons at the site, designated employees of the data user, the University of Leipzig, Institute for Medical Informatics, Statistics, and Epidemiology, have access to the data. Access for both client groups is authenticated via the above-mentioned DataSHIELD software system. The data is pseudonymized in the database for this project and thus clearly separated from other data.
For analysis, the data is analyzed by the data user using special functions. The functions are available locally and return aggregates. Aggregates in the sense of the analysis comprise summaries of data for a group of patients. The data user has no direct access to the individual data, but can only perform statistical functions on it and obtain the results, i.e., the aggregates.
In addition to the above-mentioned content-related goal, the project will test the feasibility of distributed data analysis with the DataSHIELD infrastructure in order to draw conclusions for future projects.
