People with symptoms of a disease expect in our healthcare system to achieve treatment of their conditions. A correct diagnosis is a prerequisite for this. Rare diseases pose a particular challenge for those affected as well as the healthcare system, as diagnosis can often take years and only limited treatment options are available. Patient empowerment and process optimization for diagnosing rare diseases can help here.

The aim of the project is to design and implement a demonstrator "SEPE platform" for process optimization of diagnosing patients without a diagnosis in order to identify patients with rare diseases. The SEPE platform is intended to provide interactive support, mediated by human-machine interaction, for patients in the structured processing of existing medical records from physicians' letters (findings and letters), computer-based practice information systems, and the electronic health record, as well as the subsequent semi-automated NLP-based collection of additional parameters. This enables an optimized database for diagnosis by physicians and proposes the collection of additional data that supports a more reliable diagnosis. As part of the "SEPE study”, enriched, retrospective, and anonymized patient data will be selected, and appropriate machine learning models will be trained and evaluated for process optimisation. The results will be presented in the form of an aggregated demonstrator.