Interview Niels Bolding (Health-RI) about data sharing in the healthcare and research sector

Health-RI: creating a national health data infrastructure that is more effective, scalable, affordable, and secure

Available health and disease data are not optimally (re)used for the benefit of citizens and professionals. By combining data from different sources, we can gain a deeper understanding of the complexity of disease and health. This leads to better treatments, improved healthcare products, and ultimately, better health outcomes for citizens and patients. We spoke with Niels Bolding, Technical Lead at Health-RI, about the challenges and opportunities of data sharing in the healthcare and research sector, and about the efforts of Health-RI to improve data reusability.

Obstacles to data reuse in healthcare and research

Niels sees four key challenges in sharing healthcare data:

1. Right of say of citizens, thus giving data subjects the option to opt-in or opt-out. Currently, this is organised by data holders in different ways and very purpose-specific. If researchers want to use the data for another purpose, they must obtain consent again. This is very time consuming for patients, researchers and healthcare organisations and not transparent for citizens and patients. There is a need for a national opt-out registry with clear information and communication about health data, its use and how privacy and right of say are protected.
2. There is no unique identifier per person to combine data from different sources. Everyone has a citizen’s service number (BSN), but researchers are not allowed to use it to combine data from multiple sources. Alternative methods exist, like probabilistic encryption for pseudonymisation, but they are more error-prone which reduces research quality. There is a need for change of national legislation to be able to use the BSN for research purposes.
3. Consistent security assurances are not always in place. For instance, researchers sometimes receive data from Electronic Health Record (EHR) administrators on USB sticks, which poses security risks. This needs to be more secure, especially with the European Health Data Space (EHDS) coming up.
4. Data harmonisation. Healthcare data is stored in many different ways. No single electronic health record (EHR) stores data in the same way. And researchers look at data differently compared to healthcare professionals. This means it takes time to understand the method of a source or to interpret the data correctly. In addition, if you want to combine data from different sources, it helps if you compare related data.

Niels adds that, due to these obstacles, researchers frequently have to wait months to receive the requested data. “About 80% of a researcher’s budget is spent on receiving, interpreting, and preparing data before they can even start their research. This needs to change.”

Towards a national health data infrastructure

Currently, available health and disease data are not optimally (re)used for the benefit of citizens and professionals. Health-RI is a public-private partnership of organisations involved in health research and care. Its goal is to achieve better health for citizens and patients by reusing health data with an integrated national health data infrastructure for research, policy and innovation. To develop this infrastructure, Health-RI, the Dutch Ministries of Economic Affairs, Health, Welfare and Sport, and Education, Culture and Science, have committed to identify and remove obstacles. Niels says: “Health-RI works closely with the three ministries to design a generic health data infrastructure for research, policy and innovation. The goal is to make health data findable and accessible for reuse by healthcare professionals, researchers, and companies under the right conditions while guaranteeing privacy and security. This infrastructure will create a healthcare system that is more effective, scalable, affordable, and secure.”

Stages of implementation

To implement the infrastructure, Health-RI is working on different stages of implementation.

1. Finding data. Researchers need to be able to find data to conduct research. A catalogue is crucial to make data findable. Health-RI has already made this functionality operational.
2. Exploring data. Health-RI is developing the so-called ‘explorer function’ that allows researchers to assess whether a dataset matches their research questions before submitting a request.
3. Requesting data. Health-RI is developing a request tool that will document who is allowed to access data, where it will be stored, and how it will be used. The first version is expected to be operational by mid-May.
4. Making data available (the issuance process). It is possible to bring the research question to the data and only give back the aggregated result with Privacy-enhancing technologies (PETs). Health-RI is creating a design for an exchange platform where the data is prepared before it is issued. There is unity of technology and common agreements, which will improve data interoperability.
5. Analysing data. Once access is granted, data is provided in a pseudonymised, anonymised, or synthesised format, ensuring privacy while enabling research.

Two main challenges

To implement this national health data infrastructure, Niels sees two main challenges. The first is funding. Niels explains: “The data holder has to make an investment to prepare and transfer data for when a data user requests it. But the holder doesn’t actually gain anything – at least not instantly. The data user, on the other hand, receives the data much faster and in an easily interpretable, harmonised way. In the long run, the data holder also benefits, because the healthcare process becomes more efficient. But the holders don’t always see that now and have to invest in people and processes to be able to make the data available. We have to make sure that they see the value too. The Blueprint for Data Spaces by the Data Spaces Support Center (DSSC) also says you need to think carefully about the business model.”

The second challenge has to do with the harmonisation of health data, to not only focus on data availability for primary use in the care setting, but also on secondary use of data, which implies the usage of data for research purposes too. Niels says: “We would like the secondary use of data to be included and seen as just as important as primary care in the strategy and implementation of unity of language (semantic interoperability) and technology. Without research and innovation, healthcare cannot be improved.”

Collaboration is key

Health-RI is tackling these challenges by fostering collaboration. That’s why Health-RI, for example, participates in the core group of the CoE-DSC. Niels continues: “As the orchestration party, we bring data users (such as researchers), data holders (all parties with health data, such as a hospital with an Electronic Patient Record), and the Ministry of Health, Welfare and Sport together to discuss preparation functions to make data available. They can save money if they, in advance, make data that is often requested suitable for multiple usage.”

To promote secondary data use, Health-RI continuously engages with various stakeholders. Niels says: “Five years ago, Ministries barely considered secondary use, but now it is happening more often. Our obstacle removal trajectory is a useful tool, in which the three ministries also play an important role. Economic Affairs provided Growth Fund money, the Ministry of Health, Welfare and Sport organises interoperability for health data, and the Ministry of Education, Culture and Science facilitates research. As a participant in the HDAB-NL programme, we are also working with the Ministry of Health, Welfare and Sport, ICTU, CBS, and RIVM, among others, on a central metadata catalogue. By sharing our knowledge and expertise, we aim to inspire, coordinate, and orchestrate.”

Increase interoperability between data spaces

Research does not always limit itself to data from one domain. But in order to use all relevant data, data spaces have to be interoperable. Niels gives an example: “If you want to see if there is a causal relationship between a person’s socioeconomic status and for example obesity, you need access to both health and socioeconomic data. However, this data is siloed. Where Health-RI can provide health data, ODISSEI (Open Data Infrastructure for Social Science and Economic Innovations), the national research infrastructure for the Dutch social sciences, can provide social-economic data. By making these two data infrastructures interoperable, researchers can combine data from different domains.”

Niels concludes with another real-world example: “We are part of the European Cancer Imaging project, based in the EU. Suppose the Dutch Erasmus University has cancer images. These are for example photos from MRI scans that show the location of tumours. How can you serve both the Dutch national health data catalogue and the EU catalogue? Our goal is to enhance interoperability between separate data spaces. Not everyone follows the same standards, which is a challenge on a national level, let alone a European level. So it is crucial to find ways to collaborate and exchange data effectively. We already participate in European projects such as EuCAIM, GDI, and Lifescience comunity ELIXER. I look forward to meet with other data spaces to exchange experiences and best practices.”

Het bericht Interview Niels Bolding (Health-RI) about data sharing in the healthcare and research sector verscheen eerst op Centre of Excellence for Data Sharing & Cloud.