Last week, European Commission and the Ministry of Health from Germany, with the support of Microsoft, organized a two days workshop on semantic interoperability in EHRs. The initiative was triggered by the work performed by 12 countries in the project epSOS - the large scale pilot conceiving and testing a EU shared EHR and ePrescribing solution. The workshop involved experts in terminologies from Europe, United States and from international organizations such as WHO, European Commission and International Health Terminology Standard Development Organization (IHTSDO).
As a reminder, the epSOS project (Smart Open Services for European Patients) is Large Scale Pilot (2008-2011) funded by European Commission and 12 Member States (additional 6 Member States are joining now). It is a Public-Private collaboration of National Authorities, Competence Centers and a Industry-Team (35 international and European IT-vendors managed by IHE-Europe). Cooperation has been initiated with European eGovernment Pilot on electronic Identification and electronic Health Professional and Health Insurance Card projects.
We had all a lot of expectations from this workshop, knowing that the pilot epSOS is in a stage where a structure and terminology must be chosen for building the "Health Passport". The terminology must be simple, enough standardized and unequivocal, in order to allow easy understanding and translation across 12 languages. Moreover, because of the economic pressure, patients empowerment and the need for sustainable health, semantic interoperability is a key to exchange of knowledge, clinic decision support, intelligent alerts, business intelligence. In short, all the enabling tools for a knowledge based healthcare.
Here enclosed are the main messages delivered during the workshop. I think it was a very open, productive and intense session.
The meeting was opened by Ilias Iakovidis, Deputy Head of Unit from eHealth, DG INFSO and Media, EC who outlined that the overarching goal of this workshop is to advance the understanding of patient summary and EHR-related standards with a view to enhanced semantic interoperability. The issues to be addressed are important not only for reasons of efficient data exchange, but also for industry, in light of the possibility of arriving at commonly agreed-upon standards for certification and interoperability testing ("test once use everywhere").
The first presentation, made by Erwin Bertels, representative of the German Ministry of Health, outlined that the European Commission, Member States and Associated States of the European Union agreed during the eHealth Conference in April 2007 in Berlin on a structured and coordinated collaboration in eHealth. This European eHealth Initiative intends to encourage a European and international collaboration on the development and implementation of interoperable and open eHealth infrastructures and electronic health services mainly based on international standards.
This collaboration will be structured on several levels according to the principles outlined in the EC Recommendation on cross-border interoperability of electronic health record (COM(2008)3282):
At political level, an eHealth governance process will be established under the Swedish Presidency in the second half of 2009, an European eHealth roadmap is under development, including the implementation of a Electronic Health Record framework (European Commission recommendation: until 2015), ensuring "second use" of medical data for biomedical research, evidence based medicine, epidemiology, biostatistics, health policy and management.
At organizational Level the national eHealth Competence Centers will co-operate to facilitate implementation and operation of cross-border eHealth services. At semantic level, it is acknowledged that the European multi-lingual environment requires semantic interoperable electronic health services, the knowledge must be human and computer interpretable.
The epSOS project developed 2 use-cases:
"A European citizen from Country A requires unattended medical care in Country B as e.g. a tourist or travelling salesperson or attended care as e.g. a citizen living in Country A but crossing the border on the way to her or his working place in Country B daily. Receiving consent from the patient the physician in Country B can access the patient summary located in Country A and use it in his or her own native language."
"During a medical care event described above the physician can prescribe drugs urgently required by the patient based on his or her medication record as part of the patient summary or new drugs checking for interactions and contraindications. Dispensation of the prescription can take place either in Country A or B.
In these two situations, epSOS Patient Summary and Electronic Prescribing Services must support sharing of medical knowledge about a patient between two physicians in two different countries normally requiring translation (in at least one of 11 "epSOS" languages). The experts (physician and computer scientists - participating in the epSOS Work-Package: Semantic Services" - unanimously agreed that mapping of terms is required but is not sufficient to map the complete semantic of a patient summary. Therefore, Patient safety requires ontology-driven semantic interoperability (comprising content, context and interrelations).
Erwin expressed his expectations from the workshop regarding short term recommendations that can benefit projects such as epSOS and longer-term recommendations to be taken account for example by the European Commission eHealth Standardization Mandate 403.
Charles Parisot from GE, actually representing the IHE-Europe in this context, presented an overview of the US HIT-related provisions of the American Recovery and Reinvestment Act insisting on ARRA / HITECH calls for "meaningful use of EHRs" and the importance of semantic interoperability for future investment.
Barry Smith from US National Center for Biomedical Ontology summarized the Roadmap for Interoperability of eHealth Systems advanced by the EU's RIDE Coordination Action. He showed how elements of this Roadmap are being realized in a practical demonstrator project that is designed to yield practical experience of integration of useful medical data across linguistic boundaries within the framework of the epSOS Patient Summary Scenario. The EU RIDE project was entitled to design secure networks to exchange medical summaries (incl. Emergency Dataset), Business process interoperability across healthcare domains, Mechanisms to uniquely identify patients, healthcare professionals and institutions, semantic interoperability based on classification and coding scheme standards. Using the principles developed in RIDE: "Applying realist ontology to terminologies and EHR architectures means in the first place applying it to those entities in reality to which these artifacts of the human intellect refer, such as concrete patients, diseases and therapies", Barry is proposing: "Finally, it is time to solve the problems of semantics by using the theories and tools that have been developed so far, and that have been tested under laboratory conditions. This means using . an ontology that is able explicitly and unambiguously to relate coding systems, biomedical terminologies and electronic health care records (including their architecture) to the real world". However, this is a very difficult and costly project due to a lack of unambiguous mapping of languages in several terminologies such as UMLS, SNOMED-CT.
The epSOS demonstrator that Barry describes is based on a simple terminology (less then 200 terms) focusing on emergency dataset. In the eventuality that a patient is unconscious, we have urgent need for a small amount of information about the patient to be rapidly accessible to and reliably interpreted by the healthcare provider.
The item needed and tested would be:
1. Term lists from each project country
2. Shared reference ontology to support automatic translation and evolution over time
3. Summary snapshots / screenshots, one for each country (a template, to be filled in using terms taken from the term lists).
The terms will consist initially of the statistically most frequently used terms in all project languages, they will be organized into classes and subclasses under major headings such as allergies, medications, clinical problems. The goal is to find terms which, in total, cover some 90% of all relevant cases in each of the dimensions distinguished - focusing on those terms relating to features likely to be of relevance to cross-border healthcare. Thus, focus exclusively on those features on the side of the patient relevant to emergency care - not e.g. on healthcare transactions.
The snapshot will provide an emergency practitioner in country B with a quick overview of relevant features of the condition of the patient visiting from country A.
This template would be incrementally enlarged and with appropriate software will allow creation of patient snapshots via drop-down lists followed by an additional request: Name other allergies [etc.] from which this patient suffers and which you believe may be of relevance in case of need for urgent care. Entries under this heading will be collected and used as basis for extensions of the system in the reference ontology and in the separate term lists.
My presentation outlined the necessity for semantic interoperability: the exasperation of users due to the low usability of today's solutions, the wealth of studies showing that, without Clinical Decision Support systems, the EHRs are not really improving the clinical outcome. The new trends, disease management, care communities, Regional health information networks need semantic interoperability for increased efficiency and safety. Moreover, Microsoft made important efforts in the area of interoperability, developing standardized connectors for its HealthVault platform and Amalga Unified Intelligence System and Radiology Information System, allowing export and import of reports from and to legacy systems in standards such as Continuity of Care Record (ASTM E34), Continuity of Care Document (HL7), XML. Moreover, Microsoft and its partners contributed to IHE XDS b profile and HL7 ballots. 
Microsoft issued the Common User Interface guidelines as a result of a project funded by the UK NHS. As a formal definition, the Microsoft Health Common User Interface (CUI) provides User Interface Design Guidance and Toolkit controls that address a wide range of patient safety concerns for healthcare organizations worldwide, enabling a new generation of safer, more usable and compelling health applications to be quickly and easily created.
I also reminded the efforts of Microsoft in the area of ICT system architecture with the Connected Health Framework, a vendor-agnostic set of best practices and approach based on Services Oriented Architecture (SOA), for architecting eHealth solutions for health information networks ranging from within health organizations to across multiple government agencies.
In conclusion, in my view, as industry members, we would like to have practical guidelines to enable implementation of interoperable solutions in the area of chronic disease management, condition management including behavioral changes. We know that the semantic layer in industrial solutions is still underdeveloped and the ICT Industry would need consensus about a terminology to be used in EHRs, Clinical Decision support and Alert systems.
Thomas Beale from Ocean Informatics and OpenEHR described healthcare is an information-intensive business where healthcare data is captured piecemeal during clinical work processes but used by other processes. Clinical care of patients is shared among multiple provider enterprises (exacerbated by increasingly mobile citizens), requiring information sharing. Information needs to be aggregated per-patient to be computable - to allow personalized healthcare and decision support and then across populations, for public health analysis and medical research.
Thomas described the archetypes structure allowing semantic interoperability and set the requirements for true interoperability in his view:
Stefano Bertolo, from the European Commission, DG Information Society and Media made an overview of the projects funded in 6th and 7th Framework Programme. In FP6 the project supported collaborative work such as NeOn (advanced ontology editor) and OpenKnowledge (semantic P2P). In FP7 the project funded were designed to support work on massive scale: OKKAM (global identity management), LarKC (large scale inference), Focus K3D (semantically interpreted 3D), CALBC (massive collaborative annotation).
Ioana Singureanu (Eversolve, HL7) presented some considerations about the work performed by various vendors in the framework of HL7 organization.
Since its introduction in the mid 1980s, the HL7 messaging protocol has successfully displaced a series of vendor-proprietary messaging formats for Electronic Data Interchange. Not happy to stop there, the HL7 organization has worked diligently with its latest version to move the standard in a new direction that is object oriented and model driven. Recently, HL7 has been exploring a Service-Oriented Approach, but what is the verdict on model-driven design and development? Ioana's presentation explored the use of the HL7 Reference Information Model as a mid-level ontology for achieving semantic interoperability by aligning concepts between organizations or jurisdictions. While HL7 Version 2 is still the work-horse of clinical enterprise interoperability, how has HL7 Version 3 delivered on its promise of clear conformance and semantic interoperability? In answering this question, the presentation described practical implementation methods that harvest the findings of over 10 years of HL7 information model development and use in order to develop semantically interoperable solutions that cross jurisdictional and language boundaries. The presentation covered the work in progress on defining healthcare privacy requirements and specifying engineering solutions for privacy policy and data consent enforcement.
The day 2 of the workshop, Jennifer Zelmer from International Health Terminology Standard Development Organization (IHTSDO) stated that effective use of standardized terminologies such as SNOMED CT, along with other health information standards, is key to making this possible, whether information is being shared around the corner or across national borders. She explained that IHTSDO orchestrates a globally co-ordinated effort for agreement on a core terminology for recording and sharing of health information, seeking pooling of resources to share the costs and benefits relating to the sustainable development and maintenance of the terminology products, consistent promotion of the uptake and correct use of the terminology and active harmonization activity with other SDOs.
She gave a brief staus update:
She gave relevant examples about successful implementation of SNOMED CT:
Jennifer concluded with a call for action for improving SNOMED CT use and uptake by sharing experiences, becoming active in the Community of Practice, visiting their collaborative web site, affiliate Forum and contribute with suggestions for improving the standard (additions, changes, etc.).
Her presentation was followed by a through presentation of the new classification terminology from WHO - ICD 11. Robert Jakob presented the main components of ICD:
Then he detailed what are the expected changes in ICD 11:
The ambitious goals of ICD 11 is to serve as an international and multilingual reference standard for scientific comparability and communication purposes and to ensure that ICD-11 will seamlessly function in an electronic health records environment, to link ICD logically to underpinning terminologies and ontologies. The ICD Categories are "defined" by "logical operational rules" on their associations and details. ICD 11 will be an electronic database available on-line. The calendar of implementation can be found bellow.
Dr Sylvia Thun from DMDI (Germany) presented the latest news from epSOS and the decision regarding use of very simple models for the emergency summary. Starting from a health passport available on paper version, they have proposed the use of a simple model with less then 100 terms.
Dr Thun presented after some snapshots of a demonstrator using this simple terminology:
It appears that epSOS is ready to prove that we can transmit medical information from one language to another in a medical useful way (human and machine understandable). Though very simple, it can be an invaluable way of preventing errors of interpretation of symptoms and treatment for patients travelling cross border or found unconscious.
Dipak Kalra from the University College of London (UK) had a provocative speech about practical ways of achieving semantic interoperability. Achieving semantic interoperability, ensuring that the clinical meaning of entries in an electronic health record are fully and safely computable across systems and countries, is a significant challenge. The tools available to help achieve this, such as EHR information models, archetypes and SNOMED CT, have yet to be proved at scale but nevertheless are critical to this challenge. The presentation summarized the findings of the Semantic Health roadmap on near-term and practically-achievable semantic interoperability goals. The roadmap is an excellent visionary exercise on long term interoperability and I recommend you to take a look. *
Paolo Ciccarese from Harvard University presented Medicognos platform - an innovative EHR using embedded semantics. He excited us with the paradigm shift from the forty-year old problem-oriented medical record to a fully process-oriented knowledge management platform for continuity of care and disease management, possible only through adoption of explicit semantics. Electronic Health Record he described rely on a unified semantic model combining in one single framework biomedical data, clinical processes and a model of information. This comprehensive approach will make it possible to document and share not only patient data but also other fundamental aspects of the health care process. Many questions were raised about the interoperability to other existing solutions but it appears that the EHR platform he conceived with the help of enthusiastic GPs is so advanced that can interoperate with multiple systems if, of course, they use the same core terminology. The model looks very appealing and corresponds to the last generation of EHRs able to support collaborative healthcare process and disease management.
The brainstorming at the end of the workshop around semantic interoperability agreed that EHR vendors to be urged to create a common format for presentation of de-identified data. We need to collaborate across vendor organizations, epSOS project and invited experts to build a Venn diagram of semantic requirements, testing criteria and debate with the users on the possible business case and incentive programs. We hope that, at the political level, we will have an eHealth governance and in the next CIP (Competitiveness Innovation Programme) we will advance towards common health services pilots.
In conclusion, we assisted at a collection of views and solutions for semantic interoperability, we saw different roadmaps from RIDE, Semantic Health, WHO, SNOMED CT, HL7, we admired some implementations (Ocean Informatics, Medicognos) and we acknowledged the important steps that epSOS did on the path to create an interoperable Health Passport in Europe.
We agreed that a new meeting will be organized after the summer, in order to refine the terminology proposed by epSOS and get additional input from experts.
Dr Octavian Purcarea
Global Industry Manager
World Wide Health
Microsoft
