Today, hospitals and researchers hold enormous amounts of valuable health information, but sharing it safely is one of the biggest challenges in modern medicine. The European project TRUMPET set out to explore how we can use this data to improve healthcare without ever exposing patients’ private information. Here, we will present our main results as achievement of a 3 years journey. This has been realized thanks to the restless efforts of our consortium
TRUMPET Results #1. A new way to learn from Health Data without sharing it
At the heart of TRUMPET is a new approach called Federated Learning. Instead of moving patient data from hospital to hospital, Federated Learning allows each hospital to keep its data safely on-site. Only the “learning”, not the sensitive information, travels. TRUMPET built a privacy-preserving platform that lets hospitals and research centers work together to train AI models while staying fully compliant with European privacy laws.
For patients, this means more secure data.
For doctors and researchers, this means better tools for diagnosis and care.
For Europe, this means progress without compromising trust.
TRUMPET Results #2. Understanding the rules: clear legal guidance for Healthcare innovation
One major achievement of TRUMPET is the development of practical legal know-how for anyone working with medical data.
The project clarified:
- who is responsible for what when using privacy-preserving technologies
- what risks need to be considered
- how to comply with GDPR while still enabling innovation
- how organizations can design policies, terms of use, and data-protection plans correctly
This legal insight is reusable far beyond the project by hospitals, biobanks, companies, and research organizations. It helps them take the right steps from the beginning, making safer healthcare technology easier to build.
TRUMPET Results #3. Deep analysis on how to make privacy measurable
One of TRUMPET’s core goals was to show that privacy is not just a promise but it can be tested and measured.
A dedicated privacy-validation analysis was created. This analysis represents the foundation for the creation of future tools to implement in the platform and that will help organizations to check:
- whether their systems keep data safe
- whether they meet privacy regulations
- whether they are protected from common types of attacks
TRUMPET Results #3. Stronger protections behind the scenes
TRUMPET explored advanced privacy-enhancing technologies (PETs) that strengthen the protection of patient information, even when data never leaves its original site. Here is what this means for the public:
- Hospitals can collaborate without revealing any patient identities
- Sensitive information stays locked, even during analysis
- Systems remain protected against attempts to extract private details
- Privacy safeguards become part of the infrastructure itself
These tools help ensure that the promise of Federated Learning — “learning without sharing data” — holds true in practice.
TRUMPET Results #4. Guarding against privacy threats
TRUMPET also looked at how attackers might try to guess whether a specific patient’s data was used in an AI model, or attempt to uncover private attributes. By studying these risks, the project strengthened protection methods and developed ways to test systems against such threats. This is key for building public trust as AI becomes more widely used in healthcare.
TRUMPET Results #5. A solid foundation for Healthcare innovation for data-owners and researchers
Perhaps the biggest achievement of TRUMPET is that its results are ready to be reused across Europe.
The project produced:
- a secure and privacy-preserving platform
- a dashboard for users
- legal guidelines and templates
- ways to measure privacy
- methods to prevent data leaks
- technologies that allow hospitals to work together safely
These results help create a future where healthcare advances quickly, responsibly, and with respect for every patient’s privacy.