NHS Canonical Data Model (CDM)

The NHS CDM is NHS England’s national logical canonical data model for health and care data and is part of the National Data Architecture.

It provides a common structure and language for data from disparate NHS systems, helping information from electronic patient records, GP systems, operational platforms and other sources to be understood, combined and reused consistently.

The model:

• defines common data concepts
• describes how those concepts relate to each other
• supports consistent use of data across the NHS

It is a governed national reference that is not tied to a specific system or technology.

It provides a shared logical structure that different platforms and implementations can align to in a clear and consistent way, while remaining independent of any single supplier or delivery route.

Health and care data is created and is made available through many different systems, including GP systems, EPR systems, and platforms that support social care and community settings. 

This means data can be:

• structured in different ways
• difficult to join-up
• hard to interpret consistently

Existing standards are important, but they are often designed for specific purposes.

The model helps bring these together and reduce legacy data debt by:

• reconciling what already exists, including where standards or local approaches conflict
• linking different standards and data sources
• creating meaningful relationships between data
• helping resolve historic variation and legacy data design debt over time

Ultimately, it supports a more consistent and complete understanding of data across the NHS, supports interoperability at the time of the design and supports the ability to provide the right clinical care for the patient and derive coherent national reporting and insight from dispersed datasets.

The NHS CDM allows the NHS to have control over the data which historically has been managed in silos. CDM breaks these by providing alignment on the semantic data layer, managing relationships between existing standards and models.  This means we can use the same data for multiple purposes, from supporting direct care and the Single Patient Record to simplified reporting.

It is organised around clearly defined concepts, definitions, properties and relationships so users can understand the canonical position and apply it consistently. The concepts are defined at the local level which means that the data specification can vary when the model is implemented.

The model is supported by metadata and managed reference data. This means that with each mapping of CDM to any other standard, we are developing a rich semantic layer of data for the NHS .

In practice, this means the CDM can be used to:

• bring together data from multiple systems
• reconcile data stored in different schema specified by a supplier
• support consistent interpretation of data across systems

The NHS CDM is NHS-owned and supported at the national level.

At national level, this support includes:

• defining and governing the NHS CDM as a national logical model
• aligning the model to relevant NHS standards and reference artefacts
• managing governed change, versioning and improvement
• helping implementation teams align to the canonical model
• using implementation feedback to identify gaps and clarify concepts
• convening CDM advisory, modelling and engagement groups
• setting strategic direction for adoption and evolution

It does not manage:

• the individual implementation decisions
• the detailed physical design of local systems
• full solution design for each consuming product or interface

This ensures the model provides clear national direction while remaining flexible for local implementation.

The NHS CDM helps systems work together by aligning with relevant health and care data standards, information models and reference artefacts, often uplifting them, referencing them or providing relationships between them.

It is designed to work alongside established standards and communities, including  Fast Healthcare Interoperability Resources (FHIR), Health Level Seven^® International (HL7), Observational Medical Outcomes Partnership (OMOP) and openEHR, which represent majority of the common standards used in the NHS landscape.

For example, so that CDM can:

• represent data extracted from a system using a FHIR compliant API endpoint, or
• translate data relevant to research into OMOP format which is the preferred standard use for these use-cases, or
• reflect OpenEHR clinical archetypes which reflect clinical elements inside the CDM, or
• link to managed reference data such as dm+d, SNOMED CT or ICD, or
• show how standards relate to each other

The model does not replace existing standards. Instead, it brings them together by providing a common structure for representing data consistently across different uses.

This means:

• standards such as FHIR and HL7 can support data exchange
• approaches such as openEHR can support detailed clinical modelling
• the model provides a unified way to represent data for wider use, including analytics and reporting

The model is standards aligned, not standards bound. It uses recognised standards to inform its design, while allowing flexibility to make clear and consistent decisions where needed.

The model is designed for practical use. Current focus areas include:

• Single Patient Record early use cases, including frailty and maternity
• NHS workforce modelling to support consistent staff data
• FDP product development through the NHS FDP CDM implementation
• medicines and devices modelling for local and national formularies
• uplift and redesign of national collections, including mental health, community and maternity dataset

Information, clinical and reporting needs are translated into clear data requirements and tested with stakeholders and implementation teams.

This keeps the model practical, usable and fit for purpose.

Implementation experience tests whether the model supports delivery without allowing any one route to control the canonical design.

The NHS CDM and NHS FDP CDM are closely related but have distinct roles.

The NHS CDM:

• is a national logical model
• is independent of any platform
• supports use across the NHS

NHS FDP CDM:

• is an implementation within NHS FDP (ontology)
• reflects a subset of the wider model
• supports NHS FDP use cases and products

NHS FDP is one route for expressing, testing and applying the NHS CDM. The CDM remains the national logical reference point; NHS FDP provides one implementation context for specific products and use cases.

Like any NHS CDM implementation, NHS FDP can reveal gaps and areas that need clarification, extension or refinement through governed change.

Implementation routes, including NHS FDP, FHIR-based messaging and analytical or operational uses, help test and improve the model but do not define it.

A core principle is one design, many implementations: one governed logical model that supports many physical implementations and use cases.

The model is reusable and platform-agnostic. Different systems can align to it while keeping data structured and understood consistently.

It aims to improve consistency, usability, clarity and national reporting across health and care data.

The CDM also helps address legacy data debt by identifying duplication, inconsistency and conflicting definitions, then resolving them through a shared governed model.

By providing a consistent and reusable way to structure data, it helps to:

• improve reuse of existing data
• reduce burden by minimising what data is collected
• support better national reporting and insight

It also supports:

• a shared understanding of data across organisations
• better interoperability between systems
• more effective use of data for clinical, operational and analytical purposes

By clarifying how data is collected and structured, the model makes reuse safer and easier. Over time, this can reduce burden while improving the quality and value of NHS data.

The NHS CDM is developed as an open health and care community model, supported by a national function that defines, maintains, socialises and governs it.

It is shaped through collaboration with:

• NHS England and DHSC teams
• clinicians and subject matter experts
• technical and non-technical stakeholders
• international standards communities

Feedback and challenge help ensure the model works in practice, reflects real needs, supports adoption and remains independent of any single implementation route.

The NHS CDM will continue to evolve as an open health and care model shaped by community input and real-world use.

It improves through:

• real-world use and implementation experience
• input from clinicians, professionals, data specialists and other health and care stakeholders
• alignment with relevant public NHS information standards and reference artefacts

Implementation feedback is used to identify gaps, clarify concepts and improve the model through governed change. This prevents design drift and keeps the CDM practical and relevant.

Changes are managed through governance, agreed decision routes and versioning so the model matures safely and consistently.

The model is maintained through governance that supports consistency, quality and trust.

Governance provides control, clarity and confidence without slowing progress unnecessarily.

Governance focuses on:

• controlled change through formal governance and version control
• clear decision-making supported by architectural review, stakeholder input and agreed decision routes
• ongoing improvement informed by real-world use and feedback
• adoption and alignment to help teams apply the canonical model consistently

The model is shaped through collaboration between:

• data architecture and modelling teams
• implementation and delivery teams
• clinical and subject matter experts
• governance and assurance groups

This approach ensures the model:

• remains consistent across the NHS
• continues to meet real user needs
• can evolve in a controlled and transparent way

The national function supports progress at pace while maintaining control, enabling model content to be released, tested and improved through governed change.

The model has been approved through national governance, including the Data Assurance Board (DAPB4121).

Success will be judged by practical use, improved consistency and clarity, smoother governed evolution, shared understanding and better national reporting from dispersed datasets.