Skip to content
Get in touch

EHDS and data value

Turn EHDS readiness into clinical insight and sustainable hospital revenue.

The European Health Data Space requires hospitals to make clinical data interoperable and accessible, without adding to the workload of clinical staff. Clinomic provides a turnkey, federated data infrastructure that keeps patient records safely in your custody while opening up internal analytics, EU research grants and sponsored clinical trials, even for regional and community hospitals.

Hospital 1satellite nodeHospital 2satellite nodeHospital 3satellite nodeHospital 4satellite nodeHospital 5satellite nodeHospital 6satellite nodeGovernance hubstudy intent · consent · approvalmetadata only, no patient records
  • MDR Class IIa
  • ISO 13485
  • ISO/IEC 27001
  • BSI C5 Type 2
Hospitals
45+
Medical device types
200+
European countries
7
ICUdata4EU

Proven in practice with ICUdata4EU

Across the EU-funded ICUdata4EU consortium, community and university hospitals alike already put this federated infrastructure to work, taking part in international clinical research with minimal internal IT overhead.

About ICUdata4EU

Built for data holders

From regulatory obligation to a sustainable hospital asset

  • Turnkey EHDS compliance

    Meet your obligations as a data holder without extra effort. Clinomic captures high-frequency bedside device data automatically, alongside hospital information system and laboratory records, standardises it into interoperable formats (HL7 FHIR, SNOMED CT) with ready-to-publish metadata catalogues, and keeps the raw data on site under your governance.

  • New revenue without heavy IT overhead

    Make research readiness a sustainable source of income. A low-barrier, full-service setup and a single contracting partner let regional and community hospitals join sponsored clinical trials, post-market surveillance studies and EU-funded research grants without building an in-house data engineering team.

  • Better clinical and operational insights

    Put your own data to work on the ward. Monitors, ventilators, diagnostics and the electronic health record come together in one clean, contextualised layer, so your teams get reliable analytics for process assessment, quality improvement and internal clinical decision support.

The ICUdata4EU pipeline: how it works

From the hospital information system to a secure research environment, in six steps

  1. 01 · Hospital IT · In the hospital

    Extraction

    Hospital information system, laboratory, PDMS and research databases are read inside the hospital, on its own network.

  2. 02 · Satellite node · In the hospital

    Standardisation

    Harmonised on site to the HL7 FHIR International Patient Summary with SNOMED CT, and pseudonymised a first time.

  3. 03 · Satellite node · At the hospital's edge (DMZ)

    Stored on site

    The standardised data stays in a warehouse under the hospital's custody, with consent managed alongside it.

  4. 04 · Governance hub · At the hospital's edge (DMZ)

    Consented cohorts

    A study's inclusion and exclusion criteria select a cohort, and only from patients whose consent covers the study.

  5. 05 · Key service · At the hospital's edge (DMZ)

    Second pseudonymisation

    Study-specific pseudonyms, with the keys held apart, before anything leaves the hospital.

  6. 06 · EHDS SPE or TRE · Research environment

    Secure processing

    Flattened cohorts reach a secure processing or trusted research environment for analysis and federated learning.

The path a value takes

From the bedside to a federated analysis

What has to be true of hospital data before a study can use it, and where each of those things happens.

  1. 01The problem

    Density

    At a monitored bed, the data never stops arriving.

    Continuous vitals, ventilator settings and measured mechanics, blood gases, infusion rates that change through the night. Captured by hand it becomes hourly snapshots; captured from the devices it stays at the resolution the patient actually had. That difference decides which questions can be asked afterwards.

  2. 02Standardisation

    Meaning

    Two hospitals do not mean the same thing by the same number.

    Units differ, local codes differ, the same analyte has three names. Standardisation onto HL7 FHIR, with diagnoses carrying ICD-10-GM and SNOMED CT codes, is the work that decides whether a multi-site cohort holds together.

  3. 03Governance

    The boundary

    Research does not get an open door to the operational record.

    Between clinical care and secondary use sits a de-identification stage, and beyond it the data lake and the Data Twin concept. Removing identifiers, controlling who may access what, and judging whether a dataset suits a particular question are three separate decisions with three different people answering them.

  4. 04Across sites

    Federation

    Federated analysis runs where the data lives.

    Federated approaches let a question run at each participating institution while the records stay where they were collected. That is what ICUdata4EU is for, and what the forthcoming SCAILED initiative builds on. It also asks something of every site: the same representation, the same governance, the same discipline about what a value means.

  5. 05VentAI

    A worked example

    Reinforcement learning for ventilation, evaluated on real ICU cohorts.

    VentAI developed and validated a reinforcement learning approach to PEEP, FiO2 and tidal volume, evaluated retrospectively on MIMIC-III and eICU and published in npj Digital Medicine. It is a worked example of what this data path is for, and of what a retrospective result can establish: a hypothesis worth testing prospectively, not a treatment recommendation.

    How to read retrospective results

Published research

Peer reviewed and published

Clinomic came out of university intensive care medicine. The work behind the platform has been peer reviewed and published in the journals below.

Published in

FAQ

Questions research teams ask first

What does monaOS provide for research today?

The full data path: capture during care into one representation, standardisation, a de-identification stage and federated sharing.

Does de-identification make a dataset safe to use?

It reduces identifiability. Whether a dataset is appropriate for a particular purpose is a separate judgement, made by the people responsible for governance rather than by the platform.

What does federation ask of a participating site?

The same representation of clinical information, agreed governance, and consistency about what a value means. That is the real work, and it is why federated infrastructure is built between institutions rather than bought.

What can a retrospective result establish?

That an approach is worth evaluating prospectively. It cannot establish clinical benefit on its own, and any assistance built on it stays under human review.

Where does the data physically live?

Research data stays on the satellite node in the hospital, under the hospital's custody. The clinical record runs on the Clinomic Cloud, operated around the clock by Clinomic, or on-premise in the hospital's own data centre.

Talk to the people who built the data path

Bring the question you want to answer and the sites you want to answer it across. We start with what your data contains.