Semantic Interoperability and Artificial Intelligence in Healthcare
Carina Vorisek
Prof. Carina Vorisek, M.D., and her research group develop data infrastructures that make health data discoverable and reusable for research and artificial intelligence. As part of NFDI4Health, she is helping to shape the development of a nationwide research data infrastructure and serves as co-director of the Standards, Software, and Services division. A key focus is on standardized data and metadata models and their integration into the European Health Data Space. At the same time, the group is investigating how AI can support standardization and how the description and structuring of data influence its use and the validity of AI models. Clinical use cases from women’s health and research on rare diseases combine infrastructure and method development with specific research questions.
Research Focus Areas
Data Infrastructures for Health Research
We build data infrastructures that make health data discoverable and reusable across studies and institutions. In doing so, we align the needs of researchers with the development of common standards, technical services, and structured access pathways.
Within NFDI4Health, the group serves as co-leader of the Standards, Software, and Services division. This division encompasses, among other things, study and data portals, metadata standards, data harmonization, and distributed analysis methods. Applied projects on rare diseases and health insurance data serve to further develop the infrastructure in line with specific scientific requirements.
Semantic Interoperability and Metadata
For health data to be shared, its meaning and the context in which it was generated must be understandable. We develop data and metadata models based on international standards such as HL7 FHIR, SNOMED CT, and LOINC. Our focus is on the FAIR principles: data should be findable, accessible, interoperable, and reusable.
Within NFDI4Health, the group leads the further development of metadata standards and their implementation in HL7 FHIR. To facilitate integration with the European Health Data Space, we are investigating how different metadata descriptions can be mapped to the European standard HealthDCAT-AP. As part of this effort, we are also developing and testing automated mapping procedures.
Artificial Intelligence for and with Interoperable Data
We are investigating AI from two interconnected perspectives: How can it support the development and use of data infrastructures? And what data foundations does medical AI need to deliver transparent and reliable results?
One focus is the use of language models for encoding medical terms and mapping metadata. At the same time, we are examining how missing information, the representation of diversity characteristics, and the choice of medical terminology affect analyses and AI models. In the MOTHER project, we combine these questions with AI-supported research on the health of girls and women. In this way, we bring together infrastructure development, semantic interoperability, and responsible AI in concrete applications.
Projects
NFDI4Health II – National Research Data Infrastructure for Identifiable Health Data
NFDI4Health makes data from clinical trials, epidemiological cohorts, and public health studies discoverable and supports their reuse and joint analysis. More than 20 institutions are participating in the nationwide consortium.
The working group is responsible for the following tasks:
Area
Research Group Contribution
Standards, Software, and Services – Task Area 4 Co-led jointly with ZB MED and TMF; coordination of work on data portals, analytical methods, standards, harmonization, and data quality Metadata Standards – Measure 4.4 Head of the project to further develop the NFDI4Health metadata schema, its implementation in HL7 FHIR, and its integration with terminologies and HealthDCAT-AP Application Projects – Task Area 5 Head of two use cases at BIH, in collaboration with BIPS: rare diseases in European mega-cohorts, with Marfan syndrome as the pilot, and a joint FHIR data model for German health insurance data Helpdesk – Task Area 2 Advising Researchers on Standards and Interoperability
Funding: Federal-State Program for the National Research Data Infrastructure (DFG) · Second funding phase since 2025 · Role: Principal Investigator and Co-Lead of Task Area 4MOTHER – Maternal Health, Obesity Trajectories and Health Equity Research
MOTHER investigates how maternal health and factors surrounding childbirth are linked to the later development of overweight and metabolic disorders in girls. The study is based on the European IDEFICS/I.Family cohort, which includes over 22,000 children and adolescents from nine countries.
The project develops AI-supported trajectory models and specifically accounts for female developmental characteristics such as the onset of menstruation. It examines how missing data, incomplete mother-child linkages, and study dropouts influence the results. Planned outcomes include recommendations for gender-sensitive prevention, reusable bias-assessment procedures, and an FHIR-based minimum data set. Girls, young women, parents, and prevention specialists will be involved in consultations.
MOTHER builds on the cohort documented in NFDI4Health. The results are also intended to improve the description and reuse of research data.
Funding: BMFTR · Duration: 24 months starting in January 2027 · Role: Coordination; project leadership jointly with Dr. Vince Madai and Prof. Fabian Prasser (BIH) · Partners: BIPS and NFDI4Health
Midwife Roberta – Digital Support During Pregnancy
At the BIH Digital Health Accelerator, the team collaborated with pregnant women to develop a concept for digital support featuring clear, personalized information. A key component was the structured collection of pregnancy-related data and the design of a standardized process for sharing this data with midwives, gynecological practices, and clinics via HL7 FHIR, SNOMED CT, and LOINC. Interviews and co-creation sessions with patients guided the development process.
Participation in the Accelerator concluded in 2025. The project is currently on hold; further development is planned if additional funding is secured.
BIH Digital Health Accelerator · Cohort 2025 · Role: Team Lead
Selected publications
- Vorisek CN, Klopfenstein SAI, Löbe M, et al. Towards an Interoperability Landscape for a National Research Data Infrastructure for Personal Health Data.Scientific Data. 2024;11:772. DOI
- Abaza H, Shutsko A, Klopfenstein SAI, Vorisek CN, et al. Toward a Domain-Overarching Metadata Schema for Making Health Research Studies FAIR: Development of the NFDI4Health Metadata Schema.JMIR Medical Informatics. 2025;13. DOI
- Hölter T, Klopfenstein SAI, Mayer PJ, Sass J, Thun S, Vorisek CN. HL7 FHIR in Health Research: A FHIR Specification for Metadata in Clinical, Epidemiological, and Public Health Studies.Studies in Health Technology and Informatics. 2024;316:1960–1961. DOI
- Muzoora MR, Haffer N, Thun S, Vorisek CN. How Aligned are SNOMED and LOINC with the Diversity Minimal Item Set?Studies in Health Technology and Informatics. 2025;329:1530–1534. DOI
- Vorisek CN, Stellmach C, Mayer PJ, et al. Artificial Intelligence Bias in Health Care: Web-Based Survey.Journal of Medical Internet Research. 2023;25. DOI
- Vorisek CN, Lehne M, Klopfenstein SAI, et al. Fast Healthcare Interoperability Resources (FHIR) for Interoperability in Health Research: Systematic Review.JMIR Medical Informatics. 2022;10. DOI
| Name | Position | Focus Area |
|---|---|---|
| Prof. Dr. med. Carina Vorisek, M.Sc. | Group Leader | Semantic Interoperability, Data Infrastructures, FAIR AI, and Clinical Applications |
| Dr. Hanneke Leegwater | PostDoc | NFDI4Health, Rare Diseases, and Marfan Syndrome |
| Sophie Klopfenstein | Research Assistant | NFDI4Health, Semantics, and Metadata |
| Atakan Poyraz | PhD Student | Premature Birth and AI |
The research group is growing. Undergraduate students, doctoral candidates, and postdocs in medicine, computer science, data science, and digital health who are interested in our research topics are welcome to contact us.