Big Health Data - P6 BOF session

24 September 2015- BREAKOUT 4 - 11:00

Background

This BOF session is rooted in a long series of European, international, and national projects in the area of biomedical informatics, in which the coordinators have been involved in the past decade. These include projects Health-e-Child (www.health-e-child.org), Sim-e-Child (www.sim-e-child.org), MD-Paedigree (www.md-paedigree.eu), p-medicine (www.p-medicine.eu), and others. The scientific and practitioner community developed during these projects is quite extensive and several members of them are interested in the discussions during the BOF session.

Context and Objectives

Big Data Healthcare characterizes a fundamental shift in the way biomedical data are collected and processed, as well as how biomedical research is performed and will allow us to capitalise on growing patient and health system data availability to generate healthcare innovation. However, to bring about this revolution in healthcare, there are legal, technical and cultural/societal barriers that must be overcome. The “Big Health Data” BOF seeks to start the discussion on several issues that arise when using Big Data analytics in a Healthcare setting by bringing together stakeholders from all relevant sides, particularly focusing on scalability, data analytics, privacy and security.

Bottom-up (evidence-oriented) analysis, seeking to extract useful knowledge by mining the daily routine's streaming data, is of fundamental interest in model-guided personalized medicine. In this context, Big Data techniques are applied aiming to identify latent factors (disease signatures) that can explain and predict variabilities in drug therapies and disease evolution, reveal similarities among patients stratifying patient groups and build patient specific simulation and prediction models. Such an approach goes beyond classical flat file data analysis, batch learning procedures, and simple data analysis techniques commonly focused only on a few variables of interest and a well specified dataset from a specific clinical trial. On the contrary, Knowledge Discovery and Data Mining (KDD) platforms in this area should be able to handle massive volumes of uncertain, streaming heterogeneous biomedical data, to curate, validate and analyze them in an incremental/on-line fashion from multiple points of view and under different assumptions, as well as to include or exclude dimensions, combine different modalities and incorporate existing knowledge and previous beliefs.

The still growing potential of Big Data is today fully acknowledged, but it may remain partly undeveloped or lead to undesirable outcomes or misuses of data if not carried in parallel with a deeper understanding of the regulatory and legal challenges it poses to patients’ privacy and data protection. At the same time, harming innovation and putting restrictions on research should be avoided. Indeed, as debates and proposals held in different countries show (such as on a “Magna Charta for Big Data”)[1], there’s a societal need for a more adequate legislative framework for ethical leveraging of Big Data applications, balancing the needs and rights of data providers and owners.

It must be remarked that Privacy and regulatory issues related to the process of “data-intensive scientific discovery” are a greater matter of attention for the EU, in particular in view of the new General Data Protection Regulation, which is expected to determine a more comprehensive legal framework to refer to, with the aim to strengthening individuals’ trust and confidence in the digital environment and enhancing legal certainty.  The EU debate on big data turns around three core themes: the need to ensure that citizens’ data are adequately protected; the need for Open Access to data for research purposes; the need to develop a vibrant Big Data industry in health, enabling EU to remain a major competitor in this field. As a consequence, it is necessary to strike the appropriate balance between individual privacy concerns in the healthcare setting and research purposes and innovation, which can greatly benefit patients.

Given the lack of legal harmonisation at the EU level and the different national implementations of data protection, different approaches and protocols are adopted (many of them in accordance to HIPAA, which still is, as yet, the strongest worldwide de-identifying constraint expression). Comparing and discussing these approaches is a fundamental need for the improvement of Big Data in Healthcare. 

The BOF will provide its members with a forum to discuss and highlight the legal, technological, ethical and societal challenges to the adoption of Big Data techniques in healthcare, to exchange opinions and compare experiences, and form working groups to address these challenges.

Agenda

·       Review recent advances, research and experience in Big Data Healthcare

·       Data access and protection

o   sharing best practice on pseudonymisation and anonymisation

o   developing models for consent that protect patients while enabling research

o   providing a forum for discussing, explaining and responding to data protection regulation

o   secure opening up of data to facilitate research

·       Big Data Healthcare for Personalised medicine

o   disease signatures identification

o   stratification of patient groups

o   patient-specific simulation and prediction

·       Big Data literacy in Healthcare

o   providing materials for education of healthcare professionals on use and misuse of big data

·       Patient data repositories

·       In silico drug development and clinical trials

·       Policy making

o   representing interests of the big data healthcare community to policy makers

o   Identify and discuss related challenges, interdisciplinary research needs and potential roadmap

Potential Participants

This BOF session is open to all RDA members to participate. It particularly welcomes individuals with the following expertise to actively participate in its discussion:

·       Clinicians wanting to use Big Data to improve practice

·       Biomedical researchers using data heavy analytical techniques

·       Healthcare Data Analytics with data mining, machine learning, physiological modelling and image processing expertise

·       HPC and distributed computing experts with experience in MapReduce paradigms

·       Policy-makers for Healthcare

·       Health bioinformatics legal experts

·       Healthcare administrators and Health Maintenance Organisations

·       Pharmaceutical industry researchers and manufacturers

·       Medical equipment researchers and manufacturers

·       In silico modelling, testing and clinical trial expertise

Related Documents

1.      Omiros Metaxas, Harry Dimitropoulos, Yannis Ioannidis, “AITION: A scalable KDD platform for Big Data Healthcare”, in Proc. of the IEEE Int’l Conference on Biomedical & Health Informatics, Valencia, Spain, June 2014. (http://emb.citengine.com/event/bhi-2014/paper-details?pdID=177)

 2.      Edwin Morley Fletcher, "Big Data Healthcare", in ICT 2014, Vilnius, Lithouania (http://ec.europa.eu/digital-agenda/en/news/big-data-what-it-and-why-it-important)

3.      Edwin Morley Fletcher, "Healthy data?", in Horizon 2020 Projects portal (www.horizon2020projects.com) and in the Digital Agenda News (http://ec.europa.eu/digital-agenda/en/news/discussion-big-data-and-healthcare-new-knowledge-era-world-healthcare)

Contact Person: Yannis Ioannidis (yannis@di.uoa.gr)