Luis Marco-Ruiz^a,b, Birger Haarbrandt ^a, Silje Liosland Bakke^c , Dongsheng Zhao^d, Borut Fabjan^e

, Ian McNicoll^f

^a Peter L. Reichertz Institute for Medical Informatics, TU Braunschweig and Hannover Medical School, Germany

^b Norwegian Centre for e-Health Research, University Hospital of North Norway, Tromsø, Norway

^c  Nasjonal IKT HF, Bergen, Norway

^d Academy of Military Medical Science, Beijing, China  

^e Marand d.o.o., Ljubljana, Slovenia

^f openEHR Foundation, London, UK

^g Department of Medical Imaging, Institute Two, City Two, State Two, Country Two

Abstract

Clinical data is a strategic asset for overcoming some of current healthcare challenges and advance towards the Learning Healthcare System. To that end, data needs to be understandable for various stakeholders and applications. Currently, several projects in Germany, Norway, China, Brazil, and Australia are relying on openEHR as an open standard for building open platform EHRs and also as a specification for enabling secondary use of clinical data. The panel will discuss the different strategies and lessons learned while implementing openEHR in different countries. In particular, it will cover the large adoption of EHRs to build open platforms avoiding vendor locking and the recent advances in using openEHR technologies for performing phenotyping queries and data-driven analytics.

Panel description

Background

The paradigm of the Learning Healthcare System (LHS) sets data reuse as a cornerstone for the continuous improvement of our healthcare system and the rapid adoption of new evidence. This has led health data not only to be considered as a strategic asset, but also to become a public good that can improve healthcare in particular, and the society in general [1,2]. For realizing the LHS, health data need to be accessible (when all privacy requirements are met), and understandable across the different stakeholders that use them. This means that clinical data need to be seamlessly incorporated to various Electronic Health Records (EHRs) when it is transmitted from one organization to another. Luis Marco - I might say "when it is shared between organisations" - allows for different anon-exchange architectures. 

Only then is it possible to build longitudinal cross-institutional EHRs that can be shared and accessed by the different health providers involved in a patient´s care. From a data reuse perspective, the challenge is not only in incorporating data from one EHR into another so it is available to humans, but also to make these data machine-interpretable so data-driven analytics can be performed. Therefore, allowing semantic interoperability is needed for preserving the meaning of data. This has motivated health authorities in several countries to support and fund initiatives towards the standardization of clinical information for care, clinical research, epidemiology surveillance, population health, decision support etc. Examples are: Norway, where three out of four regions rely on openEHR for structuring the EHR; the US, where the Meaningful Use initiative relies on HL7 CDA for structuring the Continuity of Care Document; or Spain, where the central government uses ISO13606 for sharing the patient summary. @luis - suggest add "UK where several regions have chosen openEHR to underpin local shared care initiatives"

However, despite the efforts in standardizing key sections of the EHR, the long-standing problem of clinical data silos still persists to some extent [1]. The reason is that summary EHR extracts are minimal subsets of the EHR that still do not allow to support many of the information needs for building longitudinal EHRs, performing patient centered medicine, or allowing data reuse networks to perform deep phenotyping. Moreover, some standards allow for exporting some EHR sections into standard compliant extracts, but the structure of the EHR and its information model keeps on being proprietary. The consequence is that a high investment is required when migrating from one health information systems to another. This has led several national initiatives to adopt the open standard openEHR for structuring the complete internal information model of the EHR. That is, to use the openEHR specification not only as a canonical format to generate extracts containing  key sections of the EHR (e.g. allergies) when interoperability among health organizations is needed, but to define the internal information schema of the whole EHR. Although high costs are involved in the implementation of a complete openEHR architecture, there are also many benefits. The first is that all openEHR-based systems comply with the same API and reference model, thus avoiding vendor locking since the transition to another openEHR-based implementation is seamless. The second benefit is that openEHR facilitates the secondary use of clinical data since the combination of archetypes with terminologies provides a semantically rich specification of clinical information, thus allowing third parties to faithfully interpret clinical data.

International openEHR initiatives

Currently, various countries are fostering initiatives that rely on openEHR for the standardization and reuse of clinical data. In Germany, the HiGHmed project is using openEHR as a core specification for building a data reuse network involving eight university hospitals and 14 partners from industry and academia [3]. In Norway, the national infraestructure projects Learning Healthcare System Toolbox [4] and Praksis Net [5] are using openEHR to standardize data from GP offices for its secondary use in research, epidemiology, and quality measures. Both projects leverage the results of the Norwegian national archetype governance work conducted by Nasjonal IKT. Beyond European initiatives, China, and Brazil are relying on openEHR as open standard for the standardization of EHRs and data reuse [6,7]. Up to November 2018, the China Stroke Data Center has been used by more than 300 top hospitals, 3.700 primary care hospitals and community clinics from 31 provinces of China. Over 9 million peoples’ screening data and 0.7 million patients’ EMR and follow-up data were enrolled. Also in China, the National MEHR (Military EHR) Project has started in 2017 to use openEHR for the integration of patient records from over 200 military hospitals and 1,000 military primary care clinics.

The role of openEHR in open platforms development

OpenEHR defines a methodology for building interoperable EHRs which core is the definition of commonly agreed Clinical Information Models (CIMs) known as archetypes. This is a key feature that translates to: a) the ability to build Health Information Systems based on stable conceptual models provided that archetypes are validated by multidisciplinary teams at a national level [8-10]; and (b) the empowerment of health providers by allowing them to manage and govern their health data independently from vendors, thus allowing them to switch from one product to another with a minimal impact [8,9].

In addition, the openEHR specification contains its own query language known as the Archetype Query Language (AQL). The AQL references archetypes rather than a specific database schema, thus making queries independent from the underlying technology, and also making the query interoperable across openEHR-based health information systems. Another important feature is that AQL allows for embedding references to terminology servers, thus boosting their expressivity. For example, one could include SNOMED-CT Expression Constraint Language (ECL) expressions inside the AQL query. This allows executing the terminology query against a terminology server and using the result to calculate the final result set returned by the AQL query.

All these characteristics make openEHR a versatile and scalable open standard for EHRs and data reuse. However, it also makes the adoption of openEHR a complex process for several reasons. First, the organizational process for adopting openEHR is very complex since the vendors, clinics, and national health trusts need to carefully coordinate their developments for eliciting, validating, and implementing archetypes. Secondly, from a technical point of view, the openEHR methodology involves challenges related to the selection of technologies, and the integration of openEHR servers with other agents that are part of the EHR ecosystem such as terminology servers and middleware needed to interact with systems based on other standards (IHE, FHIR, CDA, etc.).

Participation and audience engagement strategy

Within the global health informatics community, there are ongoing discussions about the methodologies and architectures of national EHR projects and large scale data reuse. As experts with diverse backgrounds from other regional, national or international initiatives, that might have chosen a different approach or that are currently evaluating the specifications and standards, are likely to attend this panel, we expect high audience engagement. As every participant of the panel provides a unique perspective on the diverse requirements and challenges within their respective projects and countries, we expect detailed and multi-faceted discussions about the use of openEHR and complementing standards. 

Panel organizer and participants

Luis Marco-Ruiz, PhD is a data engineer and medical informatician. Luis holds a PhD in Health Science, and a Master in Science in Statistics and Operational research. Since 2007 he has participated as a developer, advisor, and researcher in private- and public-funded projects in Norway, Germany, the UK, and Spain. At the moment, he works as a Research Fellow at the Norwegian Centre for E-health Research (Norway) and as a Semantic Interoperability Specialist at the Peter L. Reichertz Institute for Medical Informatics for the HiGHmed consortium. Luis focuses his work on the interoperability of national data reuse networks and assesses the adoption of information standards and terminologies for cross-institutional interoperability.

Birger Haarbrandt holds a B.A. in Medical Information Management and an M.Sc. in Computer Science. Between 2013 and 2017, Birger has established the Hannover Medical School Translational Research Framework (HaMSTR), investigating the enhancement of traditional data warehousing approaches (including i2b2/tranSMART) with openEHR. He previously worked on the establishment of a regional health network in the state of Lower-Saxony based on IHE XDS and worked as a software developer for a German GP software vendor. Since 2015, he has worked on the technical concept of the HiGHmed consortium to apply for the Medical Informatics Initiative, a national research project to enable secondary use of health data across institutions. Since the start of the project in January 2018, he is working in HiGHmed as a software architect for the Peter L. Reichertz Institute for Medical Informatics, aiming at the establishment of an open platform based on IHE, openEHR and FHIR between eight German university hospitals.

Silje Ljosland Bakke is an informatician and a registered nurse, with a clinical background in surgical nursing as well as clinical research from the University Hospital of Northern Norway. She has worked in health IT in the Norwegian hospital sector since 2009, since 2015 as an information architect in the Nasjonal IKT health trust for strategic IT cooperation within the Norwegian public hospital system. She has been a leading figure in Norway's openEHR modelling, governance, and training effort since 2013, and is a Clinical Program Co-Lead and a board member of the openEHR Foundation. She will discuss the experiences in archetype modelling for EHR national implementation, the coordination with vendors for adopting the archetypes, and the work with data reuse networks to model their information structures.

Dongsheng Zhao is the director of information center of Academy of Military Medical Science, China. His main research areas include health information system modeling, biomedical big data analysis, and medical Artificial Intelligence. He was the past Vice President of CMIA, and served for several national/international conferences, such as EC member of Medinfo 2017, co-chair of World Chinese Health Informatics Symposium on Medinfo 2017, CMIA2015 SPC co-chairs, SPC member of 16th China-Japan-Korea Joint Symposium on Medical Informatics. At present, He is leading the development of several National openEHR-based health information systems, such as the EHR systems for Chinese Army, integrated clinical and follow-up data platform for China Stroke Screening and Intervention Project (CSSIP).

References 

[1]       I. of M. (US) R. on V.& S.-D.H. Care, Clinical Data as the Basic Staple of Health Learning: Creating and Protecting a Public Good: Workshop Summary., National Academies Press (US), 2010. https://www.ncbi.nlm.nih.gov/books/NBK54310/ (accessed November 23, 2018).

[2]       Institute of Medicine (US) Roundtable on Evidence-Based Medicine, The Learning Healthcare System: Workshop Summary, National Academies Press (US), Washington (DC), 2007. http://www.ncbi.nlm.nih.gov/books/NBK53494/ (accessed January 4, 2015).

[3]       B. Haarbrandt, B. Schreiweis, S. Rey, U. Sax, S. Scheithauer, O. Rienhoff, P. Knaup-Gregori, U. Bavendiek, C. Dieterich, B. Brors, I. Kraus, C.M. Thoms, D. Jäger, V. Ellenrieder, B. Bergh, R. Yahyapour, R. Eils, H. Consortium, and M. Marschollek, HiGHmed – An Open Platform Approach to Enhance Care and Research across Institutional Boundaries, Methods Inf. Med. 57 (2018) e66–e81. doi:10.3414/ME18-02-0002.

[4]       A. Budrionis, L. Marco-Ruiz, K.Y. Yigzaw, and J.G. Bellika, Building a Learning Healthcare System in North Norway, in: Proc. 14th Scand. Conf. Health Inform. 2016 Gothenbg. Swed. April 6-7 2016, Linköping University Electronic Press, 2016: pp. 1–5. http://www.ep.liu.se/ecp/article.asp?issue=122&article=001 (accessed July 25, 2016).

[5]       PraksisNett, Univ. Bergen. (n.d.). https://www.uib.no/nb/praksisnett (accessed November 24, 2018).

[6]       L. Min, L. Wang, X. Lu, and H. Duan, Case Study: Applying OpenEHR Archetypes to a Clinical Data Repository in a Chinese Hospital, Stud. Health Technol. Inform. 216 (2015) 207–211.

[7]       D. Teodoro, E. Sundvall, M. João Junior, P. Ruch, and S. Miranda Freire, ORBDA: An openEHR benchmark dataset for performance assessment of electronic health record servers, PloS One. 13 (2018) e0190028. doi:10.1371/journal.pone.0190028.

[8]       L. Marco-Ruiz, D. Moner, J.A. Maldonado, N. Kolstrup, and J.G. Bellika, Archetype-based data warehouse environment to enable the reuse of electronic health record data, Int. J. Med. Inf. 84 (2015) 702–714. doi:10.1016/j.ijmedinf.2015.05.016.

[9]       B. Haarbrandt, E. Tute, and M. Marschollek, Automated population of an i2b2 clinical data warehouse from an openEHR-based data repository, J. Biomed. Inform. 63 (2016) 277–294. doi:10.1016/j.jbi.2016.08.007.

[10]     L. Marco Ruiz, J.A. Maldonado, R. Karlsen, and J.G. Bellika, Multidisciplinary Modelling of Symptoms and Signs with Archetypes and SNOMEDCT for Clinical Decision Support, in: Stud Health Technol Inf., IOS press, Madrid, 2015.

==========================Organization=======================

1-Place and time:

29 Aug 2019 from 11:50 to 13:10, room: Tête d'or 1 

2-Distribution of time_

Since we are many presenters, I propose the distribution of time below based on the amount of topics to cover and prioritizing organizational stuff and lessons learned since these are topics less covered by literature. I will reduce my part 2.2 and merge it in slot 2.1. since it is more technical. If you want to propose any change please just do…

The proposed organization is as follows:

*Presentation and introduction: 5 min

*Part I – Challenges and success factors in openehr adoption -> 25 min

**(1.1. Coordination among archetype governance bodies & clinical models governance) (Silje + Ian)->10 minutes

**(1.3. Technology selection and transition planning & 1.4 Post-deployment follow-up) (Rune, D. Zhao, Fabio) ->15 minutes.

*Part II – Openehr and data reuse - 20 min

                        ** (2.1. Modeling the data reuse API & Standardization of clinical data and AQL for phenotyping) (X. Lu, Birger, Luis)->  10 min

**(2.2. Privacy preservation and GDPR compliance) (Gustav) -> 5 min.

*Part III – Coexistence with disparate standards (Gustav/Birger) - 5 min

*Open discussion – 30 min

3-Due date for slides: In order to optimize the use of our time we should avoid the need to change computer, sticks etc. during the panel. Therefore, please provide all the presentations in advance by the 27th so I can merge them in one single file and host them in the organization’s computer.

4-Other information:

Interactive program available at: https://medinfo-lyon.org/en/programme/preliminary/

Detailed program available at: https://medinfo-lyon.org/en/programme/preliminary/

If any of you have had a last minute change and will not be able to attend but wants to have presence, I will be happy to show one or two slides if you: a) add you name and affiliation, b) add a copyright mark (e.g. © [your_name you_family_name] 2019), c) add your organization logo, and c) provide a paragraph on the ideas you want to share as a part of it. This way you can still have some presence at the panel and the audience can grasp your main ideasJ