Reliable data, clear accountability and teams equipped to question AI: three key themes from Tuesday’s “Clinical AI in the ICU: how to prepare for the inevitable” session at ESICM LIVES.
Three highlights from colleagues involved in INDICATE:
• Christian Jung, INDICATE PI, and Michael Beil, member of INDICATE’s External Expert Advisory Board, presented ten dimensions of ICU data quality: completeness, accuracy, plausibility, temporal accuracy, timeliness, consistency, conformance, granularity, uniqueness, and context/provenance. Trust in AI starts with understanding and actively managing its underlying data.
• Paul Elbers, INDICATE consortium contributor, addressed clinical decision support and human–AI interaction: where model outputs meet clinical judgement and bedside decisions.
• Michel van Genderen, INDICATE PI, emphasised a crucial principle: “Accountability must be assigned before deployment, not after harm.” Clinical trust, workflow integration and evaluation of the impact on care are essential to responsible implementation.
• Stefan Schaller, INDICATE consortium contributor, together with Davy van de Sande, from the Datahub at Erasmus MC, explored how education can address “never-skilling”, “deskilling” and “misskilling”. As we increasingly rely on AI for knowledge and decisions, we must keep questioning and seeking to understand its outputs.
For INDICATE, these discussions reinforce the importance of bringing data infrastructure, clinical expertise and education together to support responsible AI in intensive care.
AI is advancing rapidly. Turning that progress into better intensive care depends on something much less visible: access to high-quality, representative data.
That was a central message from Michel van Genderen, INDICATE’s PI, on day I during ESICM Lives 2026.
The session “INDICATE: Connecting data in European Intensive Care” was moderated by co-PI Christian Jung and work package lead Maria Theodorakopoulou.
Michel started with the wider picture: increasingly capable AI models, growing pressure on healthcare systems, and the gap between promising research and tools that actually support clinicians at the bedside.
Closing that gap takes more than model performance. We need to address bias, help healthcare professionals use AI effectively, and ensure that models are developed and evaluated using data that reflects the patients they will serve.
Yet across Europe, ICU data remains fragmented across hospitals.
This is where INDICATE comes in.
We are building a federated infrastructure that connects European intensive care data through shared data standards and agreements on how it can be used responsibly.
The principle is simple: code travels to the data. Patient data stays at the hospital. Only aggregated outputs return.
Over the past two years, we have invested in the foundations: data standards, governance, legal and ethical frameworks, and the technical infrastructure needed to make this work across hospitals.
Now, the first data providers have been onboarded, and we are moving towards scaling the network.
That next step is also about growing our community. We welcome hospitals, researchers and organisations beyond the consortium who want to contribute to the future of data-driven intensive care.
The INDICATE Consortium Meeting, hosted by Fondazione Policlinico Universitario Agostino Gemelli IRCCS in Rome taking place from Monday September 14 till Wednesday September 16, marked an important milestone in the journey from developing the INDICATE infrastructure concept to putting it into practice and building a sustainable future beyond the project.
Over three days, consortium partners came together to take stock of progress and lessons learned, align on the implementation pathway, and translate the project’s ambitions into concrete actions. The meeting equipped partners to act as INDICATE ambassadors, with a shared understanding of the end-to-end journey for data providers, data users and research projects, as well as the actions needed to support the successful implementation and transition of the infrastructure.
The programme followed a clear three-day path:
Day 1 – Understand The first day focused on taking stock of progress, reflecting on lessons learned and building a shared understanding of the implementation challenges ahead.
Day 2 – Implement On the second day, partners aligned roadmaps, addressed key dependencies and translated plans into concrete actions for their organisations.
Day 3 – Sustain The final day looked beyond the project horizon, focusing on the future organisation, business model and consortium commitments needed to sustain INDICATE beyond May 2028.
The INDICATE Consortium Meeting, hosted by Fondazione Policlinico Universitario Agostino Gemelli IRCCS in Rome taking place from Monday September 14 till Wednesday September 16, marked an important milestone in the journey from developing the INDICATE infrastructure concept to putting it into practice and building a sustainable future beyond the project.
Over three days, consortium partners came together to take stock of progress and lessons learned, align on the implementation pathway, and translate the project’s ambitions into concrete actions. The meeting equipped partners to act as INDICATE ambassadors, with a shared understanding of the end-to-end journey for data providers, data users and research projects, as well as the actions needed to support the successful implementation and transition of the infrastructure.
Ready, set, access! Preparing for responsible data use
Led by Bert Cappelle (Ghent University Hospital), the second session of the Legal Framework training sessions, provided participants with a practical overview of the legal and organisational framework required before researchers can access health data.
Rather than focusing on legislation alone, the training demonstrated how legal, ethical and technical safeguards work together to enable secure and responsible data sharing across Europe.
Metadata standards, cataloguing workflows and examples of best practices
The fifth INDICATE training session brought together participants from across the consortium to understand the role of metadata in supporting interoperability, data discovery, and federated research across institutions.
INDICATE consortium reviews progress and prepares for next project phase
On 9 July, the INDICATE consortium came together for its monthly online update meeting to discuss project progress, upcoming milestones and the next steps towards expanding the INDICATE platform.
INDICATE abstract accepted for presentation at ESICM LIVES 2026
We are pleased to announce that the INDICATE study abstract, “A Multi-Dimensional Assessment Framework for Evaluating AI and Federated Data Training in Intensive Care: The INDICATE Study” has been accepted as an e-Poster at the ESICM LIVES 2026 Congress in Lisbon.
The abstract will be presented by Maria Theodorakopoulou (the Lead of Work Package (WP) 5) during the Data Science 3 session on Wednesday, 14 October 2026, in Lisbon. In addition to the presentation, the abstract will be published in the online supplementary issue of Intensive Care Medicine Experimental, the official journal of ESICM.
MIE 2026: “Standardizing ICU Data Across Europe: the INDICATE Minimal Data Dictionary”
On Monday 25 May 2026, Boris Delange (MD, Medical Informatics & Intensive Care, University Hospital of Rennes), presented the INDICATE Minimal Data Dictionary to the European medical informatics community during MIE 2026 (Medical Informatics Europe), held in Genova, Italy (25-28 May 2026).
How can we prepare people for federated health data research?
How do we ensure that clinicians, researchers, nurses, legal experts, and data engineers have the knowledge, skills, and confidence needed to participate in federated health data research? On June 10, the INDICATE Training and Education Workgroup met at ESICM in Brussels to discuss one of the most important ingredients for successful federated data research: people.
Advancing trustworthy data through ETL validation and data quality assessment
On 21 May, the fourth INDICATE training session brought together participants from across the consortium to explore one of the most critical aspects of data integration: ensuring data quality throughout the ETL (Extract, Transform, Load) process.
The help center is specifically set up for data providers who have technical questions related to the INDICATE portal. By submitting a ticket, they can easily reach the right support and receive focused and efficient assistance with the technical questions.
The next activities of the Data Provider Support Workgroup have been established to provide practical guidance to participating hospitals during the onboarding process.
Each data provider will be invited to a dedicated one-to-one onboarding interview to review progress, discuss technical or organisational challenges and define an onboarding roadmap. Institutions are encouraged to participate with representatives from different disciplines, including project management, IT, data engineering, legal and clinical teams.
Towards one clear onboarding journey
The Training & Education Working Group and Data Provider Support Working Group are collaborating to consolidate onboarding information into a single, structured onboarding journey.
The objective is to reduce fragmentation of documentation and make onboarding requirements easier to navigate. The onboarding flow will provide guidance on:
Study participation readiness.
Stakeholder identification and engagement.
Data governance and compliance requirements.
Technical preparation.
ETL processes and OMOP conversion.
Connection to the INDICATE infrastructure.
Romain Chor
AP-HP
I am based in Paris, France. I hold a M.Sc. in Statistics and a Ph.D. in Statistical learning. I previously worked as a research engineer in Gustave Eiffel University.
My background includes several areas of applied Mathematics such as Statistics, Statical learning, Optimization and Information Theory.
What do I do in INDICATE?
I joined INDICATE as a research engineer in Statistics, to contribute to WP6 use case 2, in Lariboisière hospital (APHP, France). In particular, my work focuses on early detection of organ failure based on time series data.
Find out more about Romain’s contributions to INDICATE in his bio.
At LIVES 2026 in Lisbon, the innovative formats of on-site sessions and activities will foster networking and knowledge exchange among colleagues and friends. The scientific programme will feature a diverse array of formats, including dynamic debates, case presentations, interactive lectures, symposia, and case discussions, among many others!
During the congress, the INDICATE study abstract, “A Multi-Dimensional Assessment Framework for Evaluating AI and Federated Data Training in Intensive Care: The INDICATE Study”will be presented by Maria Theodorakopoulou (the Lead of Work Package (WP) 5) during the Data Science 3 session on Wednesday, 14 October 2026, in Lisbon.
INDICATE will be represented during ESICM LIVES in the thematic session “INDICATE: Connecting Data in European Intensive Care.” The session will be moderated by Maria Theodorakopoulou (WP5 Lead) and Christian Jung (Co-PI INDICATE), with Michel van Genderen (Co-PI INDICATE).
Mobilizing Computable Biomedical Knowledge (MCBK) Global Meeting 2026
Location: Online
The 2026 Mobilising Computable Biomedical Knowledge (MCBK) Global Meeting will take place online on 13 and 14 October, bringing together the international community made up of representatives from academia, the scientific community and the public sector who work together to ensure that biomedical knowledge in computable format is easy to find, accessible, interoperable and reusable.
Do you have a special request? Would you like to share news or a publication? Would you like to be (digitally) connected to a certain person? Did you speak or went to an event related to INDICATE or INDICATE-subjects? Please feel very welcome to share your questions or input with: info@indicate-europe.eu
When sharing your news, please make sure to attach your photos/images or figures and always make sure when people are visible, you have their permission to use the pictures.
Below with the format you can use to share your news with us:
What is the purpose of your news?
What title can we use?
Who attended (names, roles, if related to INDICATE meetings/events)?
What were the main topics discussed or key findings?
Were any follow-up actions or appointments agreed upon?
All data elements for the use cases (WP6) are defined (SAS)
WP6 | MS18
Public release of the MIMIC-EU federated database (UDUS)
All consortium partners can access the official presentation templates and branding materials for INDICATE via the INDICATE Wiki. There you can find the PowerPoint template, Word template, Teams background, logo files, branding guidelines, and icons. Everything you need to ensure consistent and professional project communication.
After two days of reflecting, demonstrating and working hands-on with the INDICATE infrastructure, today we turned our attention to the future: How do we sustain INDICATE and build an organisation that can continue to grow beyond the project?
Prof. Dr. Christian Jung opened with a reflection on how far we have come. From writing the original grant (an intense process!) to today, where the project really starts to materialise in a very valuable cooperation: hospitals are preparing for onboarding, the federated data platform is taking shape, and potential new partners are already coming forward.
Dr. Michel van Genderen: “We are not only building a platform. We are building trust.”
Trust between hospitals, researchers and partners. Trust that data can be used responsibly. And trust within a community that is getting to know each other better and growing together. A community of people with expertise, commitment and trust in one another. No single party can build this alone.
🌍 So how do we take that community beyond the project?
Today Hylke Kingma from KPMG introduced the ambition for INDICATE beyond May 2028: build a sustainable entity supporting a growing ecosystem of data providers, data users and research projects. Hylke: “Hospitals need to move from local data capture to ecosystem-ready data reuse. ”
The work hospitals are doing today to prepare for INDICATE can therefore also help them prepare for the European Health Data Space and a future in which health data can be reused securely and responsibly across a wider European ecosystem.
Mark Driessen and Anouk Kruiswijk from KPMG then took us further into what that the future organisation could look like to support sustained operations after the end of the project phase.
Moving from a European project to sustainable and scalable operations means making choices about governance, organisation and financing. It requires a solid business model, clear roles and an organisation that can evolve as the number of participants, services and research projects grows.
It also requires a clear value proposition: why should hospitals, data users and partners join INDICATE and why should they remain part of the community?
Anouk also gave us an important reminder: “Building a sustainable organisation is not the goal in itself. It is the means to continue creating impact. The ultimate goal remains what it has been from the start: improving healthcare and outcomes for patients.”
That feels like a fitting perspective for our final day in Rome!
Over these three days, we moved from reflecting on two years of progress, to putting federated learning and onboarding into practice, and finally to asking how we can make what we are building last.
After yesterday’s reflections on two years of INDICATE, today was all about putting the vision into practice!
Prof. Dr. Christian Jung opened the second day of the INDICATE Consortium Meeting with a live demonstration of the INDICATE platform before moving on to the central question of the day: What is federated learning, and how to make it work across hospitals in Europe?
Jan van den Brand took us through the federated data platform, how it works and the roadmap towards training the very first federated machine learning model on INDICATE infrastructure. The principle sounds simple, but has major implications for how we can collaborate: Data stays. Insights travel.
Dr. Chong Shen Ng of Flower Labs introduced Flower and showed us how collaborative AI works in practice. From the difference between centralised and decentralised AI to the architecture behind federated learning, his live demo gave us a look under the hood of the technology enabling this new way of working.
Joost Huiskens from Microsoft Netherlands brought another important perspective: “Clinicians do not interact with models, they interact with solutions”. If we want AI to improve daily clinical practice, we need to think beyond developing good models. We need solutions that are robust, resilient, sustainable and able to fit into real clinical environments. Not only now, but also in the future. We need to think about achieving sovereignty together.
And that brings us to one of today’s second question: What does a hospital actually need to participate in INDICATE?
With Maarten Ligtenberg, Celia Alvarez Romero, Mark Driessen, Jan van den Brand and Bert Cappelle, we moved into a hands-on onboarding workshop. Participants worked through the steps required for both data providers and data users who will be using the tools, templates and guidance being developed within INDICATE. We collected experiences, identified obstacles and used feedback from our partners to make onboarding clearer and more scalable.
This afternoon, we went from onboarding an organisation to actually starting a research project on the INDICATE platform. Marcel Giemsa from University Hospital Düsseldorf guided us through the steps to onboard on the federated data platform of INDICATE.
First the organisation is onboarded, then its users can start creating and submitting research projects.The onboarding process was shared, tested and discussed with the consortium today. The feedback gathered will help us further improve the process before making the guidance available to future data providers and users.
We then explored key practical challenges through interactive challenge tables, covering topics from data dictionary validation and sandbox requirements to scalable legal and ethical approvals, educational needs and communication ecosystem engagement.
Building a European federated infrastructure is not just about making the technology work. It is also about making it clear and practical for organisations and researchers to actually use it.
More insights from our consortium meeting tomorrow on Day 3!