Paediatric Personalized Research Network Switzerland (SwissPedHealth): A Joint Paediatric National Data Stream

Rebeca Mozún, Fabiën N. Belle, Andrea Agostini, Matthias R. Baumgartner, Jacques Fellay, Christopher B. Forrest, D. Sean Froese, Éric Giannoni, Sandra Goetze, K. Hofmann, Philipp Latzin, Roger Lauener, Aurélie Martin Necker, Kelly E. Ormond, Jana Pachlopnik Schmid, Patrick G. A. Pedrioli, Klara M. Posfay‐Barbe, Anita Rauch, Sven Schulzke, Martin Stocker, Ben D. Spycher, Effy Vayena, Tatjana Welzel, Nicola Zamboni, Julia E. Vogt, Luregn J. Schlapbach, Julia Bielicki, Claudia E. KuehnimedRxiv 2024 · DOI

TL;DR

SwissPedHealth is a national initiative to transform fragmented Swiss paediatric clinical data into a standardized, interoperable "data stream." By "FAIRifying" routine electronic health record (EHR) data and integrating it with high-density multi-omics, the project aims to create a learning health system that enables personalized care for children.

What Problem Does This Paper Address?

In the world of medicine, children are often treated as "small adults." Because paediatric clinical trials are difficult to conduct, much of the evidence for treating children is either inferred from adult data or based on small, isolated study cohorts. This creates a significant gap in evidence-based care.

In Switzerland, this problem is compounded by a fragmented healthcare system. Clinical data is trapped in "local silos"—isolated hospital repositories with different formats and governance rules. Furthermore, existing paediatric registries lack interoperability. The authors argue that to move toward personalized medicine, Switzerland needs a scalable infrastructure that can harmonize routine clinical data from across the country into a single, research-ready stream.

Methods

SwissPedHealth (launched in September 2022) is one of four National Data Streams (NDS) in Switzerland. Its approach is built on three pillars:

  1. Technical Standardization: The project uses the Resource Description Framework (RDF) to ensure data follows FAIR principles (Findable, Accessible, Interoperable, and Reusable). It maps routine hospital data to the SPHN ontology, which incorporates international standards like SNOMED CT (for clinical terms), ICD-10 (for diagnoses), LOINC (for lab results), and ATC (for medications).
  2. Automated Pipelines: Data is extracted from hospital EHRs into local Clinical Data Warehouses (CDWs). From there, it is pseudonymized and transferred via the SETT (Secure Encryption and Transfer Tool) to a central, secure "B-space" on the BioMedIT network for analysis.
  3. Governance: The authors developed a three-tiered regulatory framework to manage data flow, ranging from general service level agreements between hospitals to project-specific data transfer and use agreements.

Datasets

The project primarily utilizes routine clinical data from five Swiss university paediatric hospitals and two non-academic hospitals.

  • Historical Data: Extraction of demographic, treatment, and medical history data for the years 2017–2023.
  • Estimated Sample Sizes: For the growth-tracking project alone, the authors estimate a sample of +/- 28,000 inpatients and +/- 246,000 outpatients per year.
  • External Linkages: The stream is designed to link with the Swiss National Cohort, the Federal Statistical Office, and the Swiss Childhood Cancer Registry (ChCR).
  • Lighthouse Cohorts: For rare disease research, the project integrates data from the Swiss Paediatric Sepsis Study (SPSS) and the Genetic Study of Immunodeficiency (GSI).

Key Insights and Ideas

  • The "Learning Health System": The core philosophy is that every clinical encounter should contribute to a broader pool of knowledge. By automating the extraction of routine data, the system reduces the "lag time" between research findings and clinical implementation.
  • Multi-Omics Integration: The "Lighthouse" project is a standout feature. It doesn't just look at EHR data; it integrates Whole Genome Sequencing (WGS), RNA-seq, proteomics, and metabolomics to improve the diagnosis of rare diseases in children with life-threatening conditions.
  • Nested Use Cases: Rather than building a generic database, the team launched four "nested projects" (Growth, Cancer, Lung Function, and Antibiotics) to "stress test" the infrastructure and prove its utility in specific clinical domains.
  • Patient and Public Involvement (PPI): Unusually for a technical infrastructure paper, there is a heavy emphasis on PPI. Families and survivors are part of the Steering Committee, ensuring the research remains relevant to the people it serves.

Key Results

  • Infrastructure Approval: The Infrastructure Consortium Agreement has been successfully approved by all seven participating hospitals and three BioMedIT nodes.
  • Regulatory Milestone: The health database regulation for SwissPedHealth received ethics committee approval (AO_2022-00018), providing a legal basis for the national data stream.
  • Lighthouse Progress: Phase 1 (proof of concept) is complete. Phase 2 is currently processing a retrospective cohort of ~250 patients, and Phase 3 (prospective recruitment) began in late 2022.
  • Standardization: Successful implementation of the SPHN Connector tool across sites to transform local hospital data into a unified RDF format.

Limitations and Future Work

  • Regulatory Friction: Despite standardized templates, the authors acknowledge that legal agreements and multi-center ethics proposals remain "cumbersome" and time-consuming.
  • Resource Intensity: Data harmonization at local sites requires highly skilled technical and clinical staff, which are often in short supply.
  • Coverage Gaps: The current focus is on University Hospitals. To truly represent the Swiss child population, the network must eventually expand to primary care and smaller regional clinics.
  • International Interoperability: While currently using the SPHN/RDF framework, the authors note the need to explore future compatibility with the OMOP Common Data Model to facilitate international research collaborations.

Summary updated

Abstract

ABSTRACT Introduction Children represent a large and vulnerable patient group. However, the evidence-base for most paediatric diagnostic and therapeutic procedures remains limited or is often inferred from adults. There is urgency to improve paediatric health care provision based on real-world evidence generation. The digital transformation is a unique opportunity to shape a data-driven, agile, learning health care system and deliver more efficient and personalized care to children and their families. The goal of SwissPedHealth is to build a sustainable and scalable infrastructure to make routine clinical data from paediatric hospitals in Switzerland interoperable, standardized, quality-controlled, and ready for observational research, quality assurance, trials, and health-policy creation. This paper describes the design, aims, and current achievements of SwissPedHealth. Methods and analysis SwissPedHealth started in September 2022 as one of four National Data Streams co-funded by the Swiss Personalized Health Network (SPHN) and the Personalized Health and Related Technologies (PHRT). SwissPedHealth develops modular governance and regulatory strategies, and harnesses SPHN automatization procedures, in collaboration with clinical data warehouses, the Data Coordination Center, Biomedical Information Technology Network, and other SPHN institutions and funded projects. The SwissPedHealth consortium is led by a multi-site, multi-disciplinary Steering Committee incorporating patient and family representatives. The data stream contains work packages focusing on: 1) governance and implementation of standardized data collection, 2) nested projects to test the feasibility of the data stream, 3) a lighthouse project that enriches the data stream by integrating multi-omics data, aiming to improve diagnoses of rare diseases, and 4) engagement with families through patient and public involvement activities and bioethics interviews. Ethics and dissemination The health database regulation of SwissPedHealth was approved by the ethics committee (AO_2022-00018). Research findings will be disseminated through national and international conferences, publications in peer-reviewed journals and in lay language via online media and podcasts. ARTICLE SUMMARY ⍰ The paediatric national data stream SwissPedHealth focuses on routine clinical data from children at Swiss University Children’s Hospitals, spans disciplines and is built in a scalable and modular way in terms of governance, data infrastructure, and patient and public involvement, to enable a gradual increase in coverage of the Swiss child population. ⍰ SwissPedHealth strives to increase readiness for quality improvement, research, and personalized paediatric health care. ⍰ SwissPedHealth’s infrastructure aligns with the national frameworks, safeguarding data security and adhering to a standard interoperability framework based on the Resource Description Framework (RDF). ⍰ SwissPedHealth seeks to explore integration of data from external sources such as federal statistics, cohorts, and registries, which have their own governance and data formats. • SwissPedHealth investigates the use of multi-omics workflows to improve diagnosis of rare diseases in children with life-threatening phenotypes
BibTeX
@misc{mozun2024paediatric,
  author = {Rebeca Mozún and Fabiën N. Belle and Andrea Agostini and Matthias R. Baumgartner and Jacques Fellay and Christopher B. Forrest and D. Sean Froese and Éric Giannoni and Sandra Goetze and K. Hofmann and Philipp Latzin and Roger Lauener and Aurélie Martin Necker and Kelly E. Ormond and Jana Pachlopnik Schmid and Patrick G. A. Pedrioli and Klara M. Posfay‐Barbe and Anita Rauch and Sven Schulzke and Martin Stocker and Ben D. Spycher and Effy Vayena and Tatjana Welzel and Nicola Zamboni and Julia E. Vogt and Luregn J. Schlapbach and Julia Bielicki and Claudia E. Kuehni},
  title = {{Paediatric Personalized Research Network Switzerland (SwissPedHealth): A Joint Paediatric National Data Stream}},
  year = {2024},
  doi = {10.1101/2024.07.24.24310922},
  howpublished = {medRxiv},
}
Figure