gms | German Medical Science

Background and state of research: The Swiss Personalized Health Network (SPHN) is a Swiss national initiative. It contributes to the development of coordinated data infrastructures to make health-relevant data interoperable for research. SPHN developed a framework based on a semantic layer of information and graph technologies for exchange. The SPHN Demonstrator projects demonstrate the added value of the data resources and infrastructure elements for data-driven personalized health research, public health research, and clinical use. While clinicians obtain a huge number of data in electronic health records (EHR) to identify and diagnose sepsis this amount of data is not sufficiently applied for clinical and research use. The Demonstrator project INFRA INFectionRadar is funded by the SPHN as an inter-institutional project to overcome this gap.

Research question and objectives, hypothesis: Aim of the study is to systematically identify the evolution of infection by integrating, analyzing and visualizing conventional EHR data. In alignment with the SPHN concepts a knowledge platform containing a specific ontology with different terminologies will be developed. Risk profiles and patterns will be visualized for the clinician to use in daily routine.

Method: To achieve the objectives, we will use the routinely collected data to build an innovative digital patient model. We will define the concepts, ensure the semantic meaning applying the SPHN Semantic Interoperability Framework and develop rule-based algorithms for scores. Sources of structured and unstructured data include discharge letters and findings; vital signs, medication (Anatomical Therapeutic Chemical classification (ATC)), microbiology (SNOMED CT and partly LOINC), laboratory (LOINC), medical coding data (ICD-10), Swiss procedure code nomenclature (CHOP), data from the intensive care unit (ICU) and others. Infrastructure readiness concerning SNOMED CT integration and the ability of the CDWH to produce the entire data pipeline from the source system to BioMedIT collaboration repository will be established. In the context of this project, patient representatives and clinicians will be involved to prioritize their needs.

Results: Current state of the project: The newly defined concepts (e.g. organ complications of sepsis, surgical site infections and clinical signs thereof, risk factors) are at least partly defined. Some are in the process of annotation. A semi-automated process of annotating to SNOMED CT has been established. The BioMedIT infrastructure is ensured and accessible.

Expected results: By October 2023 we expect the ontology to be developed, explorative analysis ongoing, visualization in process (Note: The results will be updated for the DKVF).

Discussion: We want to promote comprehensive accessibility of EHR data for clinicians and research. We assume that the innovative tool INFRA will enhance the visibility of the clinical patient profiles in sepsis as a proof of concept.

Implication for research: Enhancing the development of new concepts for national projects will promote personalized health research for sepsis. Sharing common concepts and data infrastructure make health-relevant data interoperable for research.

Funding: Other funding; SPHN DEM-2022-08