Scientists from academia and industry gathered in Utrecht, Netherlands at the latest ELIXIR Innovation and SME Forum last week to discuss the challenges at the forefront of data analysis in healthcare.
Re-using and combining datasets has become a common practice in scientific research. But the situation is different for healthcare data which grow rapidly as genomic sequencing is increasingly used in clinical practice. As healthcare data are collected by multiple actors - hospitals, GPs or even patients themselves - and contain sensitive personal information, they cannot be easily shared within the research and medical community.
The latest ELIXIR Innovation and SME Forum held on 10-11 October 2019 in Utrecht, Netherlands, explored this rapidly developing field and presented the different aspects and approaches to sharing, using and storing healthcare data.
The event brought together over 60 data scientists, managers and life science researchers from industry, academia and government to present their perspectives on using and re-using data to improve healthcare outcomes.
The first session set the scene by introducing the concept of distributed data analysis. Rather than bringing data to where the analysis is carried out, distributed data analysis brings algorithms to where data are collected and stored - to hospitals, GPs and other healthcare providers.
The first keynote by Andre Dekker (Maastricht University) presented the Personal Health Train, the Dutch system for distributed analysis of hospital data. Katrin Crameri’s (Swiss Institute of Bioinformatics) talk on the Swiss Personalized Health Network presented the methodologies developed in Switzerland, which reinforced the message of distributed data analysis as a key step in utilising healthcare data.
The following talks presented different approaches to building a federated data environment in healthcare and exploring the potential of data re-use for healthcare providers and clinical practice in general.
The flash talk session provided a sample of the emerging market for healthcare data services based on FAIR principles. The session featured a number of small companies and their products in health data management and analysis.
Daan Knoors who presented IKNL and their personal health train approach in the flash talks said: "t was very interesting to see how each organisation has their own ideas about making data FAIR and easier to access in a distributed setting. However, while our data is fragmented, our approach to benefit from it shouldn’t be. I hope we can work together to build the tools that tap into the potential of distributed data and limit the disclosure of sensitive information, whether it's through adopting FAIR principles or building an infrastructure for federated learning and secure computation. (...) This ELIXIR event promotes this type of collaboration, so for me it was a great experience."
The programme on the second day focused on exploring the value of FAIR data for industry and FAIR capacity building. Kees van Bochove from the Hyve stressed the need for interoperability and collaboration on all levels, from the IT infrastructure and analysis tools to creating common policy and regulatory frameworks at European level.
Alexandra Grebe de Barron from Bayer, gave a perspective from a large pharmaceutical company and talked about establishing a FAIR culture as a trusted and open environment.
The presentations and discussions throughout the two-day event showed there is a need for closer collaboration among companies, universities and research centres in managing and analysing data in the healthcare sector.
Nikolaos Molybdris from Swiss SME decentriq said: “The approach to data analysis in healthcare is certainly going towards distributed datasets. This ELIXIR event showcases the work of many big initiatives in this area, so it's a great place to share and discuss all aspects of integrating and analysing distributed datasets.”