NUM RAPID — Registry of Adult and Pediatric Intensive Care Data — is a decentralized, federated registry initiative for routine intensive-care data from adult and pediatric intensive care units in Germany.
The project develops the technical, semantic, and organizational foundation for using high-resolution ICU routine data across sites while preserving local data sovereignty and supporting privacy-preserving analyses.
NUM RAPID aims to enable scalable, multicenter use of intensive-care routine data for:
- clinical and epidemiological research,
- health-services research,
- quality indicator analysis and benchmarking,
- outcome-oriented evaluation of intensive-care processes,
- pandemic and crisis preparedness.
Rather than transferring patient-level data into a central registry, RAPID follows a federated approach: data remain locally at the participating sites, while research queries are distributed to local data stores and evaluated there. Only the required aggregated results for a specific question are returned.
Participating sites retain local control over their data. Analyses are executed locally, and only query-specific outputs are shared according to the project’s governance, privacy, and data-minimization requirements.
RAPID builds on automated extraction of routine data from intensive-care information systems. This reduces manual documentation burden and enables the use of granular clinical data generated during routine care.
The project works toward a standardized intensive-care dataset, formalized data elements, and interoperable technical representations, including FHIR profiles where appropriate.
The RAPID dataset and analytical use cases are developed in close collaboration between clinicians, data specialists, medical informaticians, and participating university hospitals.
An initial focus is the implementation of a minimal viable product supporting the analysis of DIVI quality indicators and relevant outcome variables. The long-term goal is to support site-level learning and quality improvement through comparable, clinically interpretable results.
Repositories in this organization may include components for:
- data models and semantic specifications,
- FHIR profiles and implementation guidance,
- local data extraction and transformation workflows,
- federated query execution,
- quality indicator definitions,
- validation and benchmarking tooling,
- documentation for participating sites and developers.
NUM RAPID is part of the Netzwerk Universitätsmedizin (NUM) and builds on existing NUM infrastructures, including platforms for acute, intensive-care, and emergency medicine. The project contributes to the sustainable use of intensive-care routine data for research, healthcare improvement, and preparedness.
Official project page:
https://www.netzwerk-universitaetsmedizin.de/forschung/rapid
This GitHub organization is intended to host technical artifacts, documentation, and open-source components developed in the context of NUM RAPID.
Individual repositories may differ in maturity. Please refer to each repository’s README for status, installation instructions, licensing, and contribution guidance.
For general information about NUM RAPID, please refer to the official NUM project page.
For repository-specific questions, please open an issue in the corresponding repository.