openEHR origins

Origins of openEHR
David Ingram, Emeritus Professor of Health Informatics at UCL
October 2002

[This keynote talk by David Ingram at Medinfo 2007 in Australia details the origins of openEHR.]

Overview of openEHR’s Development
Around a decade after the 1989 establishment of the GEHR project, its partners regrouped to review their experiences. It’s promising that an upcoming ISO standard will formally prioritise clinical requirements in electronic healthcare record standards. Numerous systems claim to be electronic healthcare records, yet often lack a unified concept of purpose and structure.

Research in this field has evolved chaotically, heavily influenced by commercial, political, and academic factors as well as the need to accommodate legacy systems. Debates over poorly defined clinical terminology and processes have hampered progress, as has a continuous reinvention of foundational concepts. A diverse and empirical study of electronic healthcare records remains crucial, informed by global consensus on key requirements.

The AIM Initiative in Europe
In 1988, the EU launched the Advanced Informatics in Medicine (AIM) initiative as part of the Framework Programme for Research and Technology Development, aiming to:

1. Strengthen economic and social cohesion across the EU
2. Drive collaboration among member states for widespread benefits
3. Support a unified scientific and technical market in the EU

The programme catalysed partnerships across sectors, including healthcare, where its 1988 objectives were to unify activities, enhance competitiveness, and improve healthcare quality. Under Dr Niels Rossing’s leadership, AIM was implemented in key phases:

• 1988-90: Exploratory phase, 43 projects, 20 million ECU
• 1990-94: 110 million ECU, 38 projects, 12 concerted actions (e.g., Medirec)
• 1994-98: ~60 projects, 135 million ECU

A major objective was developing an electronic healthcare record architecture, sparking intense competition among consortia seeking funding. Early AIM projects also focused on clinical terminology and protocols (e.g., Galen, Games, Dilemma).

The GEHR Project Proposal
In 1991, Dr Alain Maskens and Dr Sam Heard led a consortium to bid on electronic health record architecture. Dr Maskens, a Belgian oncologist, ran HDMP, a small software company specialising in GP electronic healthcare records, while Dr Heard, an Australian GP and lecturer at St Bartholomew’s, London, had developed a GP system. They collaborated to develop a generic EHR system.

The GEHR consortium included seven professional, industrial, and academic partners, with St Bartholomew’s Medical College as the lead. Professor David Ingram led the consortium, preparing the proposal, and later served as Project Director, with support from Lesley Southgate. The proposal, named “The Good European Health Record” (GEHR), was completed in three months and submitted after a final weekend push. Although unexpected, it was enthusiastically endorsed, despite significant competition. David Ingram led negotiations with the EU Commission, and the project officially began in January 1992.

Summary of the GEHR Project and Synapses

Participants and Key Contributions Key contributors to the ongoing GEHR project include Dr. Dipak Kalra, who led the Clinical Task Group, and David Lloyd from the technical team, both joining early at St. Bartholomew’s. Other collaborators like Dr. Jo Milan and Dr. Stanley Sheppard joined as subcontractors but maintained close connections even after leaving. Consultant Tom Beale joined in 1993 to assist in modeling, leading to the first GEHR object model. David Ingram became Professor of Health Informatics at UCL in 1995, moving the team to establish CHIME, with Marcia Jacks transitioning as Project Administrator. Many other professionals contributed to GEHR’s progress, with some participants later taking on significant roles in medical informatics.

GEHR Project Achievements The GEHR Project developed a pioneering electronic health record (EHR) architecture based on object modeling, which evolved iteratively through prototyping and empirical testing. Results were openly published to encourage widespread adoption. Though the project faced challenges, including opposition, its innovations informed the EU’s pre-standard health records and contributed to the CEN standard ENV 12265. The GEHR approach fostered collaboration and incorporated feedback to refine its public domain outputs, emphasising EHR standardisation.

GEHR and CEN Standardisation Efforts In parallel, the CEN’s TC/251 Committee, led by Prof. Georges de Moor, aimed to standardise medical informatics with limited resources, emphasising consensus-driven standards through expert task forces. The GEHR team collaborated with CEN’s Project Team to influence CEN’s health record standards, contributing early GEHR results and models.

Transition to Synapses In 1994, the GEHR project concluded, with some continuation under Synapses, a project focused on federated EHR integration. Key members like David Ingram and Jo Milan joined a new EU consortium, including major suppliers like Siemens, to implement Synapses. This project introduced a dual-model approach, separating EHR structure (Synom) from clinical content (Synod), using an object dictionary for consistency. Despite challenges in aligning with legacy systems and standards, Synapses succeeded in implementing pilot record servers across Europe, extending the concepts from GEHR.

Ongoing Influence and Challenges Although the GEHR model sparked resistance in the standards community, it persisted as a foundation for future health record standards. It promoted a record architecture paradigm, distinct from message-based approaches like HL7. The Synapses project further advanced these principles, addressing integration with legacy systems and refining the foundational GEHR Object Model (GOM) to accommodate broader implementation needs. The GEHR framework remains influential, supporting complex healthcare requirements and aligning with emerging needs in health informatics.

Australia awakens the world electronic health care records community to the approach of GEHR
In 1996, Australia reinvigorated global efforts in electronic healthcare records (EHR) through the development of the GEHR Object Model, led by Sam Heard and Tom Beale. They refined a two-level modeling approach, involving both record architecture and clinical standards, resulting in the “GEHR archetype system.” This approach paralleled SynOD, later harmonised into the common openEHR architecture.

Implementation, Implementation, implementation!
This innovation in GEHR spurred collaborations, such as David Ingram and Peter Schloeffel’s partnership, to advance the GEHR project’s outcomes. This work influenced international EHR discussions and led to Australia’s prominence in EHR development, particularly in projects with DSTC. A key focus was building a GEHR server with an archetype-based content model aligned with the original GEHR Object Model.

The work faced challenges, as standards communities often prioritised competition over empirical research, hampering practical applications of EHR standards. This fragmented approach lacked rigorous domain definitions and led to costly failures affecting patient care. The UCL team concluded that practical, implemented EHR systems were paramount, hence, prioritising “implementation, implementation, implementation.”

Under Dipak Kalra’s leadership, the team focused on EHR development across EU projects like Synapses, Synex, and Medicate. These projects tested GEHR-based concepts in diverse clinical fields and contexts. Through practical trials rather than committee standardisation, the team achieved working clinical EHR prototypes.

First ideas for the openEHR Foundation
The openEHR Foundation was conceptualised in 1998, aiming to provide an open-source framework for high-quality EHRs. UCL and Australian teams sought to harmonise their diverging models, coining “openEHR” as the Foundation’s name. Their principles emphasised inclusivity, empirical development, and a focus on high-quality, interoperable EHR systems. David Lloyd coordinated the initiative’s web presence, while Peter Schloeffel worked to establish the Australian chapter.

The Foundation set goals for defining rigorous clinical requirements, diverse model implementations, and open collaboration. Recognising the complexities of EHR standardisation, openEHR advocated for empirical, open-source solutions, distinct from commercially driven approaches. In 2001, an international meeting cemented collaboration between UCL and Australian teams, reinforcing the Foundation’s shared pathway.

The Australian team’s success in forming an openEHR presence in the Asia-Pacific region underscored the vision of a tri-continental Foundation spanning Australia, Europe, and the USA. A clear focus emerged on expanding partnerships and aligning international funding to support openEHR’s mission.

The openEHR Foundation’s efforts continued to gain momentum as they refined their strategy, balancing open-source standardisation with commercial opportunities. Recognising the need for sustainable funding, the team explored options for a Foundation-owned trading arm to generate revenue, supporting the Foundation’s non-profit goals. This dual structure would allow openEHR to maintain its open-source principles while creating financial streams to fuel research, development, and community support.

At the same time, the Foundation emphasised the importance of unified leadership and global collaboration, especially as openEHR grew its presence in Australia and the Asia-Pacific. This vision included establishing three regional branches in Australia, Europe, and the USA, fostering a continuous “three-time-zone” collaboration model. This setup aimed to enhance knowledge sharing and accelerate development by maintaining active, interconnected teams across global time zones.

A critical part of openEHR’s agenda was expanding partnerships with healthcare providers, technology companies, and academic institutions. This included collaborations with the NHS, Oracle, and key medical services to create demonstrable, clinical EHR prototypes. For example, the collaboration with Oracle in the NHS ERDIP project focused on creating a cardiovascular disease EHR, illustrating the potential of openEHR as a foundational model for practical healthcare solutions.

To solidify its influence, openEHR engaged with governmental and inter-governmental bodies, showcasing the importance of a clinically-driven approach over purely theoretical models. By promoting evidence-based, implementable frameworks, openEHR aimed to guide national EHR strategies and standardisation efforts, countering industry claims that solutions like HL7 or proprietary EHRs offered all-encompassing solutions.

In 2001, an intensive week-long meeting in London gathered the core openEHR team, including key figures like Sam Heard, Tom Beale, Peter Schloeffel, and Dipak Kalra. Together, they reviewed the progress and alignment of their Australian and UCL implementations. With both teams now freed from prior project constraints, their systems and methodologies proved largely compatible, and the benefits of aligning fully were clear. This meeting established a concrete convergence path within openEHR, with both teams committing to harmonised development of the two-level architecture and clinical archetypes.

Whehr now?
Looking ahead, the openEHR Foundation resolved to strengthen its presence across governments and health systems worldwide, actively seeking partnerships for funding, while also focusing on expanding its development community. With the progress in Australia and UCL, the Foundation’s goal was clear: to deliver interoperable, clinically viable EHR systems that could form the backbone of national and international healthcare infrastructure. The team’s commitment to delivering tangible, scalable results remained steadfast, laying a robust foundation for the future of EHR worldwide.