Erik Flikkenschild

Profile

Erik Flikkenschild is educated as an electronic engineer. Since the 80-ties he is connected to the LUMC where he started  as a System Manager of the BAZIS/HIS with special attention to privacy and security (AIM / Seismed / CEN256 wg6). Later, in the role as manager of the central facilities group (Directorate ICT) Erik was building the LUMC (TCP/IP) PC network  and developing the serverpark, virtualizing servers with Citrix / VMware till 2012. In 2013 Erik took the role of Information Manager Research with the first task to develop a Research ICT Strategy with prof. Jeanine Houwing. Now Erik is a member of the core team of the Research ICT (implementation) program (2014 – 2017). Erik is co-writer of the NFU-Data4lifescience program and member of the operational board. Since 10 years Erik is active within the biobank world (BBMRI) and the Parelsnoer Institute (PSI) in the role of ICT coordinator and Security Officer. Since 2013 Erik is also member of the ZonMw GGG register commission. From 2017 also working on the national research network (www.lcrdm.nl, VSNU / NFU/SURFsara), developing a national reference architecture framework and glossary (workpackage leader). Since 2017 he is chairman of the national working group (SIG) secure linking personal data, which is tightly connected to the national programs LCRDM and Data4Lifesciences.

Personal data linkage in health and biobanking research: combining scientific robustness with privacy and security requirements

Being able to combine pseudonymised personal data from biobanks with registries, including data such as cause-of-death clinical, follow-up data and pathology reports is of great scientific value. Ambitions to facilitate such access and to re-use health research data, require methods and procedures for personal data linkage. These should be based on FAIR principles namely Findable, Accessible, Interoperable, Reusable. Experience with such types of data linkage that delivers scientifically robust secure results, with full respect for participants’ privacy, are of interest for the bio-banking community and should be shared. 

A fundamental design choice is needed: either a deterministic method, which relies on uniform (national) individual identifiers; or a probabilistic method that tries to link data on combinations of partially identifying information. A probabilistic linkage has higher privacy risks, but as long as the items are exclusively used to link datasets and are separately safely stored, privacy can be guaranteed.

As a use-case (EU) project TRANSFoRm has built an application involving pseudonymisation, by a Trusted Third Party, matching data  from general practice registries (Nivel) and bio-banking data for people with type 2 diabetes of the Dutch Parelsnoer Institute. The used combination of proven strong encryption technology and uniform deterministic identifiers allows for more secure linkage, resulting in robust datasets anonymized for the purposes of research. 

The Dutch Parelsnoer Institute  plays a role in working towards practical national guidelines, use-case descriptions and practical solutions. The Dutch expert group (“SIG secure linkage”) is working on a set of representative, easily understandable use cases.