Optimising cohort data in Europe

the data-sharing centre has the ownership of the uploaded data (Cohen et al., 2020) although this practice obscures the meaning and application of ownership as a whole (Richterich, 2018). That is, just because institutions and data controllers store and hold patient data, this does not mean that they own these data per se (Hulsen et al., 2019). Instead, under such a perspective, data is the property of the patients themselves while institutions merely act as data custodians (Hodgson et al., 2020). This means that data should be destroyed at the end of the research project (Yuzbasioglu et al., 2018). However, the concept of patients as primary data owners is not necessarily straightforward for researchers and data custodians alike. This is because credit attribution shifts from the researchers to the participants themselves (Kostkova, 2018). If patients actively participate in the research, provide data and frame data governance, then they should get adequate credit for the work performed as well as intellectual property rights (Thorogood, 2020). The issue is however not clear in the literature because the matter of credit attribution is often mixed with issues of control and access (Molnár-Gábor and Korbel, 2020). A reason for this is that participants’ view about their potential for data access, acknowledgement and accreditation is not neutral. For instance, while they generally agree to give stewardship of the data to researchers, they also expect to be informed if their data are moved to another database for subsequent use (Kondo et al., 2019). In this context, in their attempts to adapt to general GDPR and EU Clinical Trial regulations, EU biobanks vary in their capacity to timely notify participants about data transfer (Salokannel et al., 2018). To some extent, this is a problem faced by many cohort sample collections as they include data that have been stored and used in different ways for diverse purposes (Lesko et al., 2018). Namely, biobanks generally do not only contain data samples that have been consented to for being transferred on legal grounds but also include legacy samples without any consent and even some samples with informed consent (Pastorino et al., 2019). A commonly used strategy in this context is to renew the ethical approval of the biobank each year (Coppola et al., 2019) or/and give independent custodians (such as universities) the possibility to link direct person-related data to the samples (Kiehntopf, 2019). As it stands, the current heterogeneity of purposes, uses, and consent schemes within biobank data impedes that participants' views on data governance can be fully integrated - especially when biobanks include a large amount of data obtained without consent (Organization for Economic Cooperation and Development, OECD, 2019). In this sense, patients' participation in the governance, design and data collection of cohort research projects is not a substitute for an ethical engagement with the challenges and ambiguities of data-intensive research. Instead, it is useful to evaluate how the increasing “datafication” of cohort research amplifies the tension between community control and scientific openness.