Openehr Open Data Platforms in Medical Informatics

Openehr Open Data Platforms in Medical Informatics

open data plaꢀorms in medical informaꢁcs

Greeꢁngs 3
Prof. Dr. Anneꢀe Grüters-Kieslich
Greeꢁngs 4
Prof. Dr. Roland Eils openEHR – What is it? 5openEHR in HiGHmed 7openEHR in pracꢁce openEHR in the EU 10
openEHR in the UK 12 openEHR in Slovenia 14 openEHR in Norway 16
German openEHR User Group 18
Imprint and contact data 19
Supported by

Prof. Dr. Anneꢀe Grüters-Kieslich
Chief Medical Director
Dear Reader,
Along with more advanced technologies in the healthcare system comes a constantly growing quanꢁty of health-related data for everyday clinical acꢁviꢁes. However, the lack of uniformity inherent in the available data, in terms of quality, type and format, has emphasized the criꢁcality of the establishment of a standardizaꢁon framework. and Chairwoman of the Board
Heidelberg University Hospital
The medical informaꢁcs consorꢁum HiGHmed funded by the German Federal Ministry of Educaꢁon and Research (BMBF) addresses this challenge and aims to develop and use innovaꢁve informaꢁon infrastructures to increase the efficiency of clinical research and to swiꢃly translate research results into tangible and validated improvements of paꢁent care. These aims are ꢁghtly connected to challenges to integrate and further develop soluꢁons of innovaꢁve, internaꢁonally interoperable data integraꢁon and methods, with the intending to demonstrate their added value for health research and also paꢁent care.
A chosen infrastructure for HiGHmed is the open source soluꢁon openEHR (Open Data Plaꢀorms in Medical Informaꢁcs), a creaꢁon of semanꢁc interoperable electronic health record architecture in a vendor- and technology-neutral format. openEHR standardizes heterogeneous medical IT systems and the resulꢁng dispersion of medical data in archetypes and templates. These archetypes are developed in collaboraꢁon with the internaꢁonal openEHR community, which consists of physicians, computer scienꢁsts and processing as well as standardizaꢁon experts. The results are systems and tools that add value to decision-making support in dayto-day clinical pracꢁce and support research quesꢁons and will be the basis for a data-centric “app” ecosystem. These tools will reduce workload and significantly improve semanꢁc traceability. Furthermore, model-generated code and user interfaces are an area of conꢁnuous innovaꢁon in openEHR and promise to revoluꢁonize health compuꢁng.
With this brochure we would like to introduce openEHR to you. A plaꢀorm where internaꢁonal CIOs, IT architects and medical informaꢁcs experts share their experiences and present their use cases for novel informaꢁon system architectures based on openEHR.
With this, I hope that you will enjoy reading about the diverse projects featured in this brochure.
Prof. Dr. Anneꢀe Grüters-Kieslich
Chief Medical Director and Chairwoman of the Board
Heidelberg University Hospital
Greeꢁngs Prof. Dr. Anneꢂe Grüters-Kieslich

Prof. Dr. Roland Eils
Coordinator of HiGHmed
Dear Reader,
A standardized management of vast amounts of data from disparate organizaꢁons is crucial for use of big data in healthcare for efficient knowledge discovery. Likewise, newly emerging paradigms as precision and systems medicine require a holisꢁc picture of the individual paꢁent across the health care system.
Medical Informaꢁcs Consorꢁum Germany
The highly complex and diverse data originaꢁng from care, omics, clinical and biomedical research for analyꢁcs, machine learning and clinical decision support systems needs to be findable, accessible, interoperable and reusable across insꢁtuꢁonal boundaries.
While the organizaꢁonal framework to enable data exchange and trusꢀul cooperaꢁon in full compliance with the paꢁents’ right of self-determinaꢁon and data protecꢁon legislaꢁon is a major challenge, the socio-technical aspects of data sharing must not be neglected. Today, we find that contemporary hospital informaꢁon system architectures are not well suited for data reuse within and across insꢁtuꢁons in an economical and sustainable way. Opening up exisꢁng and avoiding future ‘data silos’ by implemenꢁng interoperability standards is a prerequisite here.
Recently, open source solutions based on openEHR (Open Data Platforms in Medical Informatics), a semantic, interoperable electronic health record architecture, have gained tracꢁon in the United Kingdom, Norway, Slovenia, Australia and cross-boundary organizaꢁons like Eurotransplant. The primary focus of its endeavor is on electronic health records (EHR) and related systems. For
HiGHmed, openEHR plays a vital role by enabling collaboraꢁve informaꢁon management and by providing interoperable and vendorindependent data repositories. openEHR thus enables the use of standardized specificaꢁons across mulꢁple sites and promotes future service-oriented soꢃware architecture and innovaꢁve system soluꢁons in clinical care.
In this brochure we present the challenges and opportunities of openEHR for the healthcare system. We hope that you enjoy reading and that you will follow our enthusiasm for open data standards in clinical care and research.
Prof. Dr. Roland Eils
Coordinator of HiGHmed
Medical Informaꢁcs Consorꢁum Germany
Greeꢁngs Prof. Dr. Roland Eils

What is it?
In recent decades it has become clear that the value of informaꢁon technology in health
(oꢃen called e-health) has lagged far behind its value in other domains such as banking, process control and logisꢁcs.
Moꢁvaꢁon openEHR as a technology
People the world over rouꢁnely and jusꢁfiably wonder why their health records don’t work like their online banking does. Similarly, following: healthcare professionals ask why their Electronic Medical Record
(EMR) systems sꢁll don’t talk to each other, are so expensive, and why it is so difficult to maximize the value of the vast amounts of available health data. openEHR is designed to address this need, by providing the A multi-level framework that separates data models from
•domain models
An open platform architecture that can be used to represent
•patient-centric health data, which are accessed by institu-
There are many conꢁngent reasons for the poor progress of IT in health relaꢁng to poliꢁcs, commercial interests, and the inability to focus funding across the mixture of public and private economic sectors, but there are also fundamental causal factors at work, which if not addressed will conꢁnue to block progress.
Chief among these are the complexity of informaꢀon and processes (ulꢁmately due to the innate complexity both of human biology and society), and secondly the fact that the focus of healthcare – the paꢁent – rouꢁnely moves across enterprise and jurisdicꢀonal boundaries while expecꢁng seamless care. tions and products but not controlled by them
A modelling factory environment that continually produces
•computable domain models (known as archetypes and templates), developed by domain professionals, in any language
Tools that machine-convert domain models into technical
•forms useful for developers
We can visualise an openEHR technology ecosystem that implements the above as follows:
The model-driven openEHR technology ecosystem openEHR What is it?
5The value of using openEHR
An openEHR plaꢀorm soluꢁon may be deployed in a single hospital much as any EMR soluꢁon is, but also across a city, region or whole country. It is in the laꢂer deployments where the capability of persisꢁng data in a paꢁent-centric rather than insꢁtuꢁon- or product-centric fashion is realised, and vendor lock-in is avoided.
Under this architectural approach, the enꢁrety of the deployed soꢃware soluꢁon is based on (at design ꢁme) and driven by (at runꢁme) computable models of content and process created by domain professionals. Notably, the data representaꢁon depends only on the data model, which ensures that physical database contents are not affected by new domain models.
The separaꢁon of domain models from the technical layers qualitaꢁvely changes the soꢃware engineering economics of kinds of acꢁvity: soluꢁons, because it allows the plaꢀorm to be built and deployed independently, with domain models being injected at runꢀme, removing one of the major sources of cost at a stroke. It also allows domain professionals, who know their own data and work-
flows, to be in the driving seat when specifying the semanꢁcs of Healthcare Informaꢁon System soluꢁons.
As a community-based organisaꢁon, openEHR undertakes three Publishing the technical specifications that define both the •platform and the domain models
Developing software for modelling tools and repository

Publishing clinical models, which act as de facto standards for fragments of the domain
The technical advances of openEHR lead naturally to a plugand-play plaꢃorm economy, in which any vendor or developer can produce a soluꢁon component, as long as it conforms to the published data and Applicaꢁon Programming Interface (API) base standards of openEHR, and addiꢁonally, the domain content models created by the community of clinical professionals.
The technical specificaꢁons include informaꢁon models for healthcare data, including the EHR and demographics; a portable query language, formal languages for expressing domain content and data sets, and finally, an open API specificaꢁon.
The openEHR query language represents a major innovaꢁon, which enables the wriꢁng of model-based queries that are independent of physical database schemas, and thus portable across systems. This enables a sustainable approach to clinical decision support and business analyꢁcs, which otherwise are either ꢁed to a single database, or else have to be rewriꢂen for every target system.
The use of openEHR also entails new freedom with respect to health data: iniꢁally, it is liberated from products and vendors, to be owned by providers; eventually it can move to full paꢁent ownership, with healthcare professionals as guardians – the ultimate realisaꢁon of the paꢁent-centric EHR.
It is the goal of the openEHR Foundaꢁon and community to fundamentally change the quality of informaꢁon technology in the service of medicine, so as to improve outcomes in clinical healthcare, public health and the value of secondary data use.
Interoperability is solved in a way common outside of the healthcare domain, which is by machine-generaꢁon of schemas and soꢂware components from models, rather than hand-building of message or document definiꢁons. In a similar way, the difficulty of application development is greatly reduced via machinegeneraꢁon of applicaꢁon soꢃware and UI components.
Thomas Beale
Management Board, openEHR Foundaꢁon
6openEHR What is it? openEHR in HiGHmed
HiGHmed – an open plaꢀorm approach to enhance cross-enterprise care provision and research
Modern medicine is confronted with an unprecedented data
flood arising from recent technological breakthroughs in genome
Open Health Plaꢀorm sequencing, imaging and remote sensing amongst others, which are becoming part of rouꢁne clinical care at an ever increasing pace. Approach
Despite this massive increase in data sources, diversity and volume, only a surprisingly small fracꢁon of this data is currently integrated for rouꢁne processes in clinical pracꢁce. This integraꢁon is mainly hindered by the current state of hospital informaꢁon systems, which favor an applicaꢁon-centered approach over a collaboraꢁve and data-driven mindset, creaꢁng a vast amount of data silos.
Consequently, care providers are prevented from taking advantage of a holisꢁc view on all available data that would allow opꢁmal diagnosis and treatment. Moreover, those data sets, which are oꢃen unique and invaluable, are not available for research purposes, thereby acꢁvely impeding progress in clinical research and the translaꢁon of cuꢄng-edge research insights into clinical pracꢁce.
In HiGHmed all university hospitals will establish Medical Data
Integraꢁon Centers (MeDICs) based on a generic and scalable reference architecture. This architecture is based on IHE XDS, openEHR and HL7 FHIR for integraꢁng data from care, research, and external sources, which will facilitate the development of new soluꢁons for medical data analyꢁcs and benefit clinicians, paꢁents and researchers. A shared informaꢁon governance structure between all involved clinical sites in accordance with the FAIR (Findable, Accessible, Interoperable, Reusable) principles will provide the foundaꢁon for meaningful data exchange.
The Medical Informaꢁcs funding scheme of the German Federal
The geographical distribuꢁon of parꢁcipaꢁng
Ministry of Educaꢁon and Research aims at overcoming these ob- university hospitals and their connecꢁon via the stacles and thereby promoꢁng the opportuniꢁes of digitalizaꢁon in medicine. By establishing integrated IT soluꢁons, the exchange and use of data from healthcare, clinical and biomedical research across insꢁtuꢁons and locaꢁons will be greatly facilitated. During a four-year development and networking phase, which started in
January 2018, funding worth a total of about €150 million will be allocated to four consorꢁa.
HiGHmed plaꢃorm
The HiGHmed consorꢁum currently consists of the three university hospitals in Heidelberg, Göꢄngen and Hanover, and is complemented by more than 20 partners from academia and industry.
The HiGHmed consorꢁum will integrate five addiꢁonal university hospitals: the University Medical Center Schleswig-Holstein, the University Hospital Cologne, the University Hospital of Würzburg,
Charité – Universitätsmedizin Berlin and the Münster University
Hospital, which have formally applied to join HiGHmed by the end of 2018. The HiGHmed consorꢁum will thus significantly increase its impact in clinical care and research across Germany through the inclusion of 25ꢅ% of all German University hospitals. openEHR in HiGHmed
7To demonstrate the capability of the technical design and the organizaꢁonal structure, the MeDICs take the responsibility to prove the suitability of the HiGHmed approach by supporꢁng three medical use cases in the domains of oncology, cardiology and infecꢁon control.
By providing such an open plaꢀorm, HiGHmed avoids any mandatory procurement of proprietary soluꢁons that would cause vendor lock-in. Instead, parꢁcipants in HiGHmed are able to acquire relevant components from different vendors, open source iniꢁaꢁves or by self-development. This architecture will foster an ecosystem, based on open service interfaces and clinical models.
To achieve the aspired goals and provide a solid foundaꢁon to address the specific requirements of the clinical use cases,
HiGHmed’s approach includes the iteraꢁve definiꢁon of an interoperable, open health data plaꢀorm specificaꢁon. The following characterisꢁcs are essenꢁal to the HiGHmed plaꢀorm:
The role of openEHR
Within the HiGHmed plaꢀorm, the management of structured medical data is of utmost importance. The availability of structured data is the prerequisite for any meaningful computaꢁon as staꢁsꢁcal analysis, machine learning and the execuꢁon of decision support algorithms. To address the challenges regarding standardizaꢁon, management and computability of structured data, openEHR provides several capabiliꢁes needed to achieve the aspired objecꢁves.
1. Open Service Models: All specificaꢁons of the provided applicaꢁon programming interfaces (APIs) are openly accessible to everybody. Specificaꢁons include data security and privacy, electronic health record management and database queries.
2. Open Informaꢁon Models: All clinical models are well defined based on established open standards. Data based on these models can be reliably processed and computed in local and distributed environments. In addiꢁon, all models are openly available in HiGHmed.
Firstly, an informaꢁon model governance framework, derived from the one pioneered by Norwegian Nasjonal IKT (see from page 16), helps to establish a common understanding of the data between the parꢁcipaꢁng hospitals by allowing collaboraꢁve work on formal representaꢁons of clinical informaꢁon models, called archetypes. The creaꢁon and curaꢁon of archetypes is driven by dedicated data stewards, a new socio-technical role
3. Open System Specificaꢁons: All system components and protocols are openly specified using licenses feasible for commercial and non-commercial use. This assures that every component in the system can be replaced by soꢃware from mulꢁple vendors, including open source communiꢁes.
The core components of the HiGHmed platform are based on
IHE and openEHR specifications.
A service bus integrates the components and provides a secure and standardized data access layer. Based on the platform, an ecosystem of applications can be developed.
8openEHR in HiGHmed within hospitals, that takes responsibility of professional longterm care of data from design to processing and sharing data for the different purposes. In close cooperaꢁon with clinical stakeholders, they drive the collaboraꢁve modelling and maintenance acꢁviꢁes, which will establish a cross-enterprise informaꢁon management resulꢁng in a shared data dicꢁonary.
In the medium term, based on the scalable and solid architecture of openEHR, comprehensive and full-blown clinical applicaꢁon systems, like closed-loop medicaꢁon management systems or paꢁent data management systems, can be developed against the HiGHmed plaꢀorm and provide standard-compliant and interoperable data “by design” and without costly mappings. Likewise, lightweight paꢁent-facing apps can use the same data semanꢁcs to provide valuable data from wearables and paꢁent reported outcome.
This shared data dicꢁonary can be used to assist the manual implementaꢁon of the semanꢁcs agreed upon within databases, and data capture forms within soꢃware systems. However, this approach will sꢁll require addiꢁonal effort to integrate the data into a clinical data repository. Preferably, standardized data models can be directly incorporated within the HiGHmed platform to create new clinical and research applicaꢁon systems and databases. For example, interoperable clinical registries could be developed based on the capabiliꢁes of the HiGHmed plaꢀorm and shared semanꢁcs.
By providing a semanꢁcally-enabled data querying through the Archetype Query Language (AQL), openEHR offers a sound technical foundaꢁon to acquire data needed from disparate healthcare providing organizaꢁons. Queries can be directly expressed using the centrally managed models residing in the shared data dicꢁonary. For the HiGHmed plaꢀorm, this introduces a reliable and safe way to query data, deploy algorithms and develop clinical decision-support systems in a highly-distributed environment.
Another important aspect of openEHR is the separaꢁon of data and applicaꢁons. This means that applicaꢁons don’t use their own database layer, which then forms the typical data silo, but le- Outlook verage the plaꢀorm to store any structured paꢁent data instead.
Through this approach, all electronic health record data can potenꢁally be used immediately by other clinical and research applicaꢁons and for data analyꢁcs within the HiGHmed consorꢁum.
Addiꢁonally, locally developed, and highly specialized soluꢁons can be easily exchanged between sites that have implemented the HiGHmed plaꢀorm.
While the HiGHmed plaꢀorm will start with a narrow focus on the use cases, it provides a framework to extend its supported data models and funcꢁons in an open and transparent way. By connecꢁng with internaꢁonal programs and iniꢁaꢁves, and by aiming at providing an open source implementaꢁon of the main components of the plaꢀorm, we seek to facilitate the further rollout and the proliferaꢁon of the open plaꢀorm approach.
In contrast, building such systems in a tradiꢁonal way and creaꢁng mappings to a standardized messaging interface seems neither
financially sustainable nor realisꢁc given the sparse resources of hospital IT departments. Additionally, without a platform approach, integraꢁng newly developed applicaꢁons across different sites will be burdened with high costs.
Birger Haarbrandt
Hannover Medical School openEHR in HiGHmed
9openEHR in the EU
The Eurotransplant approach to interoperability
Eurotransplant has chosen openEHR as the basis for its clinical data repository as a means to enable the sharing of data across an internaꢁonal cooperaꢁon of hospitals, clinical registries and naꢁonal authoriꢁes.
This internaꢁonal collaboraꢁve framework includes all transplant hospitals, ꢁssue-typing laboratories and hospitals where organ donaꢁons take place. As mediator between donor and recipient,
Eurotransplant plays a key role in the allocaꢁon and distribuꢁon of donor organs for transplantaꢁon. The mission statement and goals of Eurotransplant express the foundaꢁon’s main target: to ensure an opꢁmal use of available donor organs.
The integrity and availability of data can be improved by sharing data within the transplant process chain automaꢁcally. The system renewal project, currently being executed, should prepare the system for future interoperability between hospital and laboratory systems and Eurotransplant systems. To be more flexible the current monolithic system will be split in several applicaꢁons: donor, allocaꢁon and waiꢁng list.