HL7 Europe Common Cancer Model, published by HL7 Europe. This guide is not an authorized publication; it is the continuous build for version 0.1.0 built by the FHIR (HL7® FHIR® Standard) CI Build. This version is based on the current content of https://github.com/hl7-eu/cancer-common/ and changes regularly. See the Directory of published versions

Challenges

The development of the HL7 Europe Cancer Common Data Model (ECCDM) addresses a set of recurring challenges identified through experience in European and national initiatives working on cancer data.

One of the main challenges is the heterogeneity of data models currently adopted across projects, institutions, and countries. Different initiatives support different needs and use cases, leading to the definition of project-specific datasets and models that are difficult to compare, reuse, or align beyond their original context.

Another major challenge concerns data availability and usability in real-world EHR systems. The actual presence, structure, and quality of cancer-related information are often difficult to assess a priori, and vary significantly across care providers. This makes it challenging to define data models that are both theoretically complete and practically implementable.

Reconstructing the longitudinal cancer journey represents an additional challenge. Cancer patients are often treated across multiple care settings and over long periods of time, and their disease history is not confined to a single institution. Many existing models are patient-centric or event-centric, making it difficult to consistently represent disease progression, treatments, response evaluation, and follow-up over time.

The coexistence of different technical standards further complicates interoperability. In the European context, OMOP is widely adopted for research and secondary use of data, while HL7 FHIR is increasingly used for data exchange and interoperability, including primary data capture. Mapping between these standards is complex and often addressed on a project-by-project basis, resulting in ad hoc solutions that are difficult to maintain and reuse.

Supporting both primary and secondary use of cancer data is itself a challenge. Secondary use (e.g. research, analytics, AI model training) depends on the availability of structured and high-quality primary data. However, primary data capture is often not designed with reuse in mind, leading to loss of information or the need for extensive post-hoc transformation. Bridging this gap requires a shared understanding of which information must be captured during care to enable meaningful reuse.

Finally, existing international specifications, such as mCODE, provide valuable implementation experience but are tailored to specific contexts and requirements. Adopting such specifications directly in the European landscape is not always feasible due to differences in base profiles, governance, and interoperability needs, highlighting the need for a model-driven, context-aware European approach.

These challenges motivate the need for a shared, minimal, and extensible cancer data model that can support interoperability across standards, enable consistent mapping between HL7 FHIR and OMOP, and facilitate both primary and secondary use of cancer data in the European context.