1. FAIR4Health project: Improving Health Research in EU through FAIR Data
Carlos Luis Parra-Calderón
Consortium Coordinator
23 Sept 2019
Santiago de Compostela
First of all, I’d like to thank the organisers for the opportunity to share with you our project. FAIR4Health started on last December and we are looking forward to establish collaborations and synergies at any level with all EU initiatives dealing with FAIR data in the health domain. So in case you are involved in other similar initiatives, please don’t hesitate to contact me at the end of the session and I will put you in contact with the coordinators to set out potential collaborations.

2. Outline Call: SwafS-04-2018 Use of FAIR data in Health
FAIR4Health Project:
Consortium
Objectives
Implementation
Pathfinder use cases
Technological platform
Open Community
In this talk I will introduce you the FAIR4Health project by including some information about the specific call under which this project has been funded and some use cases of FAIR data policy in the health domain as an introduction. Then we’ll meet the FAIR4Health team and the main objectives of the project following an overview of the project work plan, and afterwards I will show you in more detail which are the use cases and the technological tools that will be developed, and we’ll finalise with the open community that should sustain the outcomes beyond the project lifespan.

3. Call: SwafS-04-2018 TOPIC: Specific Challenge:
All research builds on former work and depends on scientists' possibilities to access and share scientific information. In the context of Open Science and Responsible Research and Innovation the European Commission therefore strongly supports the optimal open access to and re-use of research data (considering e.g. robust opt-outs). As a concrete action the EC has extended the Open Research Data Pilot to cover all areas of Horizon 2020 (as of the 2017 Work Programme). This will result in more data becoming available for re-use. However, it is necessary to adopt further actions to reach the Commission's overall objective of findable, accessible, interoperable and re-usable (FAIR) data by 2020.
SwafS : Encouraging the re-use of research data generated by publically funded research projects
The topic is called: Encouraging the re-use of research data generated by publically funded research projects.
In this topic, the challenge posed by the EC is the need to adopt actions to make FAIR all data and metadata derived from publicly funded research. This is make the data comply to be Findable, Accessible, Interoperable and Reusable.

4. Call: SwafS-04-2018 TOPIC: Scope:
To support the FAIRification of data, stressing on data quality (certification), their interoperability and reproducibility of research
To generate pathfinder use cases to demonstrate how data sharing and re-use can generate a groundbreaking innovative product, service, or treatment
To generate a prototype of such innovative product, service, or treatment
To include at least 10 different EU countries in the consortium
SwafS : Encouraging the re-use of research data generated by publically funded research projects
The scope of this call could be summarized in these 4 points:
To support FAIRification of data, addressing data quality, interoperability and reproducibility of research
To generate pathfinder use cases that could lead to the development of innovative services
To generate the prototype of such services
And to include participants of at least 10 different EU member states, which gives the idea of the geographic scope that the EC wants to address.

5. Call: SwafS-04-2018 TOPIC: Expected impacts:
To increase the visibility of the Commission's open FAIR data policy through dedicated communication activities, and networking of relevant actors including industry.
To generate a beneficial impact for science, the economy and society by means of:
Increasing the reproducibility of research.
Cross-fertilisation of interdisciplinary research.
Boosting citizen science.
Generating added value for innovative companies (including SMEs and start-ups) in the EU DSM.
Key Performance Indicators:
Increase in FAIR data in those domains identified by the beneficiaries for action.
Contribution of the pathfinder case studies to innovative data sharing and re-use.
This is the last slide about the details of the call, but I think that it is important to clearly define the framework in which the FAIR4Health project will evolve in the next 3 years.
This call has 2 main expected impacts:
To increase the visibility of the EC open FAIR data policy
And to generate a beneficial impact for science, economy and society by increasing reproducibility of reasearch, boosting interdisciplinary research and citizen science and generating added value for data-based innovative companies belonging to the Digital Single Market
This call also poses its own KPIs that should be monitored throughout the project: the increase in the use of FAIR data in the domains identified by the project and the contribution of the use cases to the development of services based on innovative data sharing and reuse.

6. Use of FAIR data in Health
Corpas et al. (2018)
Rare diseases (active IN)
PHT (preparatory IN)
Metabolomics (preparatory IN)
PubMed search of FAIR data in health ( )
Here you can see several examples of the use of FAIR data strategy in the health domain. We can find examples rising from the academic domain, such as the FAIR guide for data providers in human genomic data, in bottom-up initiatives such as the GO FAIR Implementation Networks for a practical implementation of the EOSC and, also, top-down initiatives like the “data for better health” which is being driven by the Belgian government.
FAIR data in health is gaining momentum since the publication in 2016 of the FAIR data guiding principles by the FORCE11 community, and it is foreseen to keep on growing in the coming years.

7. FAIR Guiding Principles
(F)INDABLE
F1. (Meta)data are assigned a globally unique and persistent identifier
F2. Data are described with rich metadata (defined by R1 below)
F3. Metadata clearly and explicitly include the identifier of the data they describe
F4. (Meta)data are registered or indexed in a searchable resource
(A)CCESSIBLE
A1. (Meta)data are retrievable by their identifier using a standardised communications protocol
A1.1 The protocol is open, free, and universally implementable
A1.2 The protocol allows for an authentication and authorisation procedure, where necessary
A2. Metadata are accessible, even when the data are no longer available
(I)NTEROPERABLE
I1. (Meta)data use a formal, accessible, shared, and broadly applicable language for knowledge representation.
I2. (Meta)data use vocabularies that follow FAIR principles
I3. (Meta)data include qualified references to other (meta)data
(R)EUSABLE
R1. Meta(data) are richly described with a plurality of accurate and relevant attributes
R1.1. (Meta)data are released with a clear and accessible data usage license
R1.2. (Meta)data are associated with detailed provenance
R1.3. (Meta)data meet domain-relevant community standards
Wilkinson, M. D., et al. (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific data, 3
Los principios FAIR son las características, normas y buenas prácticas que los recursos, herramientas e infraestructuras de datos deben cumplir para que sean Localizables, Accesibles, Interoperables y Reusables.

8. FAIR4Health

9. 17 partners from 11 EU and non-EU countries
Consortium
Coordinated by Virgen del Rocío University Hospital, Andalusian Health Service (SAS)
17 partners from 11 EU and non-EU countries
6 health research organisations
2 universities experts in data management
4 academic partners with strong background on medical informatics
5 business actors
FAIR4Health is coordinated by Prof. Carlos Parra, head of the Technological Innovation Unit at Virgen del Rocio University Hospital as part of the Andalusian Health Service in Spain. The consortium accounts for 17 partners from 11 EU and non-EU countries. There is a strong representation from both southern and central Europe, and it includes non-EU countries such as Switzerland, Serbia and Turkey (and maybe UK in the next months…)
FAIR4Health brings together expertise from different domains (health research, data managers, medical informatics, software developers, standards and lawyers) all of them key stakeholders to address the challenges posed by the project.

10. Objectives
To facilitate and encourage the EU Health Research community to FAIRify, share and reuse their datasets derived from publicly funded research initiatives through the demonstration of the potential impact that such strategy will have on health outcomes and health research.
SO 1. To design and implement an effective outreach strategy at EU level
SO 2. To produce a set of guidelines to set the foundations for a FAIR data certification roadmap
SO 3. To develop and validate an intuitive, user- centered FAIR4Health platform and FAIR4Health agents
SO 4. To demonstrate the potential impact in health research and health outcomes
The overall objective of FAIR4Health is to facilitate and encourage the EU health research community to FAIRify, share and reuse their datasets derived from publicly funded research initiatives.
Specific Objective 1. To design and implement an effective outreach strategy at EU level based on trust building and shared benefit
Specific Objective 2. To produce a set of guidelines to inform a number of Research Data Alliance recommendations in order to set the foundations for a FAIR data certification roadmap to guarantee high quality of data
Specific Objective 3. To develop and validate intuitive, user-centered technological tools to enable the translation from raw data to FAIR data and support the FAIRification workflow
Specific Objective 4. To demonstrate the potential impact that the implementation of such FAIR data strategy will have in terms of health outcomes and health research through the development and validation of 2 pathfinder case studies.

11. Implementation
Besides the coordination, the Andalusian Health Service will lead the dissemination and outreach of the project, which receives inputs from the other work packages to inform the development of this strategy. Work package 2 is being led by Eva Méndez from the University Carlos III of Madrid, and its role is to perform a comprehensive analysis of all the barriers and challenges that must be overcome to implement a FAIR data strategy in the health research domain. The work package 3 is being led by Catherine Chronaki from Health Level 7, and its objective is to define all the technical requirements needed to develop the FAIR4Health platform and agents, task that will be led by Blanca Jordán and her team from Atos in work package 4. Once the technological tools are ready, Christian Lovis from University of Geneve will lead the development of the use cases demonstrators in work package 5. Finally, the sustainability and economic assessment of the project will be led by Gokçe Banu from Software R&D Consultancy within the work package 6.

12. FAIRificación process
In this slide you can see a first vision of the FAIR4Health Open Community and how they will interact.
At the bottom of the image are the health researchers, who will be the main contributors in the FAIRification tasks. In turn, they will be able to browse and access to FAIRified datasets from their peers, always in compliance with the data owners’ policies in terms of licences, approval from local ethical board, regulatory framework, etc. For this purpose, FAIR4Health will define a clear workflow for the implementation of such FAIR data policy including all these details.
Within the project, this workflow will be needed to successfully develop the use cases proposed.
But the project goes beyond the health research domain and also addresses the beneficial impact that such FAIR data strategy may have in health outcomes. For this purpose, data scientists will be able to develop innovative eHealth services built upon Privacy Preserving Distributed Data Mining approaches that could be applied over the FAIRified datasets without the need of hosting them outside the dataset owner’s facilities. These services will be available for the healthcare providers that request them in the context of the EU Digital Single Market.
In summary, there is a triple win behind this community: health researchers could access larger datasets and accelerate the discovery of knowledge while avoiding bias due to local datasets; eHealth services providers could develop and exploit innovative services in the EU Digital Single Market; and healthcare providers could have access to this eHealth services portfolio that will allow them to improve their quality of care and, therefore, closing the cycle of a Learning Healthcare System based on FAIR.

13. Data sets I

14. Data sets II
En total, información disponible de 5 millones de pacientes.

15. Pathfinder Use Cases
Innovative eHealth services based on FAIR data reuse:
#1 To support the discovery of disease onset triggers and disease association patterns in comorbid patients and demonstrate the reproducibility of research.
#2 To develop and pilot a prediction service for 30-days readmission risk in complex chronic patients
FAIR4Health will design, develop and demonstrate 2 pathfinder use cases, in this case they will be innovative eHealth services based on FAIR data reuse.
The first one will address the feasibility for health researchers to work with FAIRified and federated datasets to support the discovery of disease onset triggers and disease association patterns in comorbid patients, and demonstrate the reproducibility of such research.
The second one will address the development of an innovative prediction service for 30-days readmission risk in complex chronic patients making use of a privacy preserving distributed data mining approach.

16. Technological Platform
FAIR4Health Agents
FAIR4Health
Platform
Communication Interface with FAIR4Health Platform
PPDDM Agent
Curation, Normalization and Mapping
ETL
User
Interface
FAIR4Health Agents
The technological solution that will support this project will be based on two main entities: The FAIR4Health platform and FAIR4Health agents.
The FAIR4Health agents, which will be located at the data owner’s premises, will enable the FAIRification of local datasets through its user-driven Extraction Transformation and Load (ETL) functionalities. At the end of this process, datasets will be normalised, curated and mapped to domain vocabularies and ontologies, acting like Data FAIRports. The agents will also host instances of the Privacy-Preserving Distributed Data Mining services, so they could be run locally without the need of hosting these datasets outside the owner’s premises.

17. Technological Platform
FAIR4Health Platform
FAIR4Health
Platform
Comm. Interface PPDDM services
PPDDM Manager
Federated Query Manager
Comm. Interface Agents
Security Layer for P2P Comm.
Actionable PPDDM repository
User
Interface
FAIR4Health Agents
The FAIR4Health platform will act as the facilitator for deploying and delivering innovative data-driven services to the FAIR4Health community by hosting the repository of actionable data mining models that could be executed on real time. Moreover, the platform will allow for P2P sharing of these datasets under specified licensing agreements between the data requester and the data owner that will also be put in place by the platform. This P2P sharing will be duly secured from end to end in order to minimise the risk of disclosing information. The FAIR4Health platform will act as a hub of the federated agents. It will also manage the innovative services developed upon PPDDM technologies, including security filters to avoid the execution of malicious services.

18. FAIR4Health Community
In this slide you can see a first vision of the FAIR4Health Open Community and how they will interact.
At the bottom of the image are the health researchers, who will be the main contributors in the FAIRification tasks. In turn, they will be able to browse and access to FAIRified datasets from their peers, always in compliance with the data owners’ policies in terms of licences, approval from local ethical board, regulatory framework, etc. For this purpose, FAIR4Health will define a clear workflow for the implementation of such FAIR data policy including all these details.
Within the project, this workflow will be needed to successfully develop the use cases proposed.
But the project goes beyond the health research domain and also addresses the beneficial impact that such FAIR data strategy may have in health outcomes. For this purpose, data scientists will be able to develop innovative eHealth services built upon Privacy Preserving Distributed Data Mining approaches that could be applied over the FAIRified datasets without the need of hosting them outside the dataset owner’s facilities. These services will be available for the healthcare providers that request them in the context of the EU Digital Single Market.
In summary, there is a triple win behind this community: health researchers could access larger datasets and accelerate the discovery of knowledge while avoiding bias due to local datasets; eHealth services providers could develop and exploit innovative services in the EU Digital Single Market; and healthcare providers could have access to this eHealth services portfolio that will allow them to improve their quality of care and, therefore, closing the cycle of a Learning Healthcare System based on FAIR.

19. Relevant links:

20. MUITO
OBRIGADO!
@FAIR4Health