Optimising cohort data in Europe

In this sense, pseudonymisation is representative of the commitment to data protection by design. This logic extends not only to the collection of personal data but also to its storage and its modalities of access (Hansson, 2021). Allocation of responsibilities is thus grounded in the necessity and proportionality principles. This is implemented in various models of access such as controlled access and registered access for sharing data. In the former case, data controllers grant access to datasets only to approved data users under certain conditions (Shabani et al., 2021). As such, controlled access is the exact opposite of open access and is suitable for cases where privacy concerns are particularly acute (e.g. genomic research). By contrast, registered access concerns low risk personal data (e.g. non-stigmatising data from healthy individuals who have consented to data sharing). The model assumes that since the processing of data would not create risks of re-identification, it would be enough to simply restrict data access to trusted users or more precisely, bona fide researchers (Peloquin et al., 2020; Slokenberga et al., 2021). Reviewing ethical aspects of research proposals or performing subsequent access reviews is thus not necessary. (b) Obtaining data access In general, the process of obtaining access to samples or data in a cohort or biobank comprises three main steps: 1) requesting access to data or samples, 2) obtaining the necessary study approvals, and 3) executing the access. The absence of proper cataloguing of samples and data is the main obstacle of the first step, whereas heterogeneous access governance through the different cohorts across Europe is the main obstacle of the second step. Most cohorts have local scientific access committees or equivalent decision bodies for evaluating access requests and dealing with ethical, legal and administrative issues. In many cases, information about these cohort-specific committees and their requirements for granting access to data and samples are unclear or unavailable. In order to executing the access, establishing the necessary Material Transfer Agreements (MTA) that govern use of the cohort data and samples might prove to be a significant administrative obstacle and cause additional delays. 2.1.4. Data sharing (a) Integration of participants in the data sharing process Data sharing operates according to the legal frameworks, the biobank networks and the type of data architecture available. In ethical terms, trust is a central condition for data sharing, which can be integrated through a “learned intermediary” model: an independent panel reviews applications and grants access to data on the basis of applicants' expertise and quality of the proposed research (Steinbach et al., 2018). The learned intermediary model requires applicants to clearly specify the research timeline, the analytical processes to be used and potential specifications of data. In most cases,