Session 2 | Structured High-Quality Data with AI and other means
The EHDS addresses the syntax of health systems transactions, though not yet the use of semantic resources. What are the necessary conditions to implement a meaningful - in particular semantic - interoperability strategy in a dedicated organisation? Which infrastructure, tools and competences are necessary? Can AI meet the clinician dream with coding happening in the background with no or very minimal human intervention?
A working paper on "Increasing the availability of high-quality and structured health data, the potential of AI and other means" will lie at the heart of this session.
It offers you the opportunity to:
- Influence the AI potential within the EHDS: Ensure real-world healthcare and innovation needs are reflected in practice.
- Bring practice to policy: Help make the EHDS regulation more realistic, sustainable, and workable.
- Strengthen EHTEL’s collective voice: Contribute to a united position that amplifies EHTEL's influence in European digital health policy.
Session 3 | EHR-Algorithm Communication
This session will discuss the need to have data that fuels algorithmic tools that are of high quality, meaningfully structured, and easily sharable. Achieving impact requires a comprehensive, multi-level strategies - from local hospital workflows up to European semantic frameworks - with humans and AI working hand-in-hand to progressively refine data ecosystems.
A working paper on "How can EHR system users make the best of algorithm-based tools?" will lie at the heart of this session.
It offers you the opportunity to:
- Influence the EHDS: Ensure real-world healthcare and innovation needs are reflected in practice.
- Bring practice to policy: Help make the EHDS regulation more realistic, sustainable, and workable.
- Strengthen EHTEL’s collective voice: Contribute to a united position that amplifies EHTEL's influence in European digital health policy.
- Oznake
- EHDS
Komentarji
This event raises a key and often underestimated point: high-quality, structured and interoperable data are necessary, but not sufficient, to ensure safe and effective AI use in healthcare.
EHDS rightly focuses on syntax, data availability and sharing. However, when AI systems and especially agent-based or semi-autonomous workflows are introduced, the main risks no longer come only from data quality, but from behaviour in context.
Even when AI relies on validated sources (e.g. semantic resources or standardized context protocols), failures can emerge from:
context shifts in real clinical workflows,
interaction between multiple algorithmic tools,
long-term drift after deployment,
pressure from users, time constraints or ambiguous cases.
Simple real-world examples illustrate this:
an AI tool using high-quality EHR data may behave safely in isolation, but produce unsafe recommendations when combined with another decision-support system;
automated clinical coding with minimal human intervention can work well for routine cases, yet silently degrade in rare diseases or edge cases;
interoperability enables scale, but can also amplify errors if behavioural effects are not continuously observed.
This is why EHDS implementation should combine:
semantic and structural interoperability,
clear governance and traceability,
behavioural testing and post-market monitoring, especially for high-risk and agent-enabled AI systems.
Moving from “data readiness” to “behavioural readiness” will be essential to make the automated workflows discussed in Session 2 safe in practice, and fully aligned with the AI Act requirements on post-market monitoring and real-world risk control.