We recently discussed the announcement from HL7 International (the organisation behind FHIR and more) and openEHR International that they are exploring whether certain standards and specifications can be aligned for the common good. Great news for those professionals who have been working in the field of open healthcare information standards for some time, but for those who are not yet as familiar with this subject matter, let's go back to basics: how do FHIR and openEHR differ and how can they strengthen each other?
Open?
At their core, both FHIR (Fast Healthcare Interoperability Resources) and openEHR are open standards for information and communication in healthcare. Even so, the 'openness' of such standards does not imply that information must be publically available - for example, Nictiz uses the following criteria to assess the openness of standards:
- The standard is approved and managed by a not-for-profit organisation. Decision-making is open and accessible to stakeholders.
- The standard has been published and the documents are freely available or at low cost and can be freely copied.
- The intellectual property is available royalty-free.
- There are no restrictions on reuse of the standard.
There are many more such standards - think of, for example, the ISO 13606, DICOM, Edifact, SNOMED-CT, LOINC... For the scope of this article, we will stick to (the basics of) openEHR and FHIR.
FHIR 'versus' openEHR
The question has been posed more often: if both FHIR and openEHR are open standards for healthcare information and communication, and using a proliferation of standards only complicates interoperability (the interoperability paradox), why not simply choose one standard and be done with it? Unfortunately, that is highly unlikely to work. As Gerda Meijboom says in her article ‘Is there still a future for zibs?’, "If you look closely at the various standards, you will see that each standard has its own goal and its own set-up to achieve that goal."
Where the focus used to be on data exchange, we see the conversation shift to include data storage and we are also seeing a growing interest in openEHR.
But in what ways do FHIR and openEHR differ, then? We propose the following definitions:
FHIR
FHIR is an HL7 standard for exchanging data digitally between healthcare providers and healthcare providers and patients. FHIR lends itself to all forms of communication in healthcare.
openEHR
openEHR is an open standard for modelling, storing and handling healthcare data. This set of documents allows you to develop information and interoperability solutions for healthcare, such as healthcare information systems (e.g. an EHR).
There is some overlap between the two standards: for instance, the FHIR specification provides some guidance for data storage, but in a much more limited form than openEHR. Besides the fact that openEHR can facilitate more detail for data storage, it is simply more ‘mature’ in terms of storage. For more perspectives on this, see also this interesting conversation on the openEHR forum.
On the other hand, openEHR also offers APIs for data exchange, but these are relatively new compared to the possibilities offered by FHIR.
In the words of our colleague Jorn Duwel, who has a lot of experience with FHIR: "You can cut your food with a fork reasonably well, but there is a better tool."
The bottom line is that choosing just one of these standards to solve all scenarios is not a good idea - they both have their strengths and can reinforce each other. So, not FHIR versus openEHR, but...
FHIR plus openEHR
At CODE24, we tend to emphasise openEHR because we focus on data availability through standardised data storage. We believe that this is where the future lies, but this does not mean that there is (or will be) no role for data exchange. We simply cannot do without it, especially in the coming years - if only to be able to exchange data from all kinds of ‘data silos’, such as systems that do not use openEHR. FHIR is a proven and widely supported option for this, which we have also been using ourselves since 2017 (including mapping between the two standards and in the role of MedMij DVA).
While openEHR has been around longer, FHIR has gained a faster foothold in the Netherlands. For instance, MedMij uses FHIR and in 2023 the Ministry of Health, Welfare and Sport indicated in a letter to parliament that it would choose FHIR for "improving data availability through a nationwide network of infrastructures". The Dutch government has not made any such statement about a standard for modelling and storing healthcare data yet, but things are definitely in motion: organisations such as Nictiz are investigating the role of the various standards to arrive at an ideal scenario for the Dutch healthcare information system. Read, for example, the ‘Research on future scenarios for the zibs’. The fact that both FHIR and openEHR already have a large support base in the sector through their international communities is also certainly being taken into account in the various analyses.
Convergence
That the organisations behind the standards themselves are now also exploring how openEHR and FHIR can reinforce each other is a good and important step. In addition to the announcement from HL7 and openEHR in May this year, Guy Tsafnat, Rachel Dunscombe, Davera Gabriel, Grahame Grieve and Christian Reich from both openEHR International and HL7 International also published the opinion piece ‘Converge or Collide? Making Sense of a Plethora of Open Data Standards in Health Care'. In this article, the authors propose choices for three specialised standards for three specific healthcare data domains:
-
Clinical care and administration (documentation of care data for e.g. continuity of care and billing)
-
Data exchange (communication between stakeholders, from health care providers to health care users and health insurers)
-
Longitudinal analysis (examining patterns, trends and predictability)
For these domains, the authors propose that openEHR will be used for clinical care and administration, FHIR for data exchange and the OMOP (Observational Medical Outcomes Partnership) Common Data Model for longitudinal analysis: “An interoperable health system would use openEHR to collect data, FHIR to transmit data between systems en organizations, and OMOP to find insights in the data.”
Harmonisation
The intention to harmonise FHIR and openEHR more is an important key to enable the two to be used side by side in a powerful way. This was also foreshadowed in a session organised by the openEHR Masterclass, where FHIR founder Grahame Grieve made a comparison of the two standards and suggested ways to work together. Meanwhile, both communities are actively investigating the possibilities for harmonisation and we also expect interesting perspectives on this at this year's openEHR International Conference in November.
Ultimately, conversion of data between standards will regularly lead to data loss - for example, information may be missing after conversion, or context or an interrelationship between information may be lost. Closer cooperation and harmonisation where possible can help ensure that we can harness the power of different standards alongside each other with minimal data loss. We will continue to follow the developments closely.