Optimising cohort data in Europe

To meet the ethical requirements of beneficence and non-maleficence, the following has been proposed: y y On the basis of beneficence, justice and respect for the person, validated, verified and clinically useful research data should be communicated to participants, if they wish so (Bilkey et al., 2019). Communication of results should however be done carefully, considering the social and personal impact of these results, and taking into account potential feelings of stigmatisation and exclusion. y y During data collection, researchers should ensure that all communities of the target population are represented equally so that no group would be excluded from potential research benefits. This involves granting research access to those communities that either lack the technical means for participation (e.g. no Internet connection) or lack digital literacy. This can be done through health surveillance systems and wearable technologies as they are able to identify populations that are not in contact with traditional medical systems. New communication technologies have radically changed the parameters, procedures and modalities of participation, meaning that it is no longer possible to determine the potential benefits and harms of research participation in advance. To protect participants, technical capabilities for effortless and retroactive retraction of data should be explored. y y Researchers should be aware of the limitations of beneficence both for big data contexts in general and cohort research in particular. Beneficence principles predate modern social technologies and as such do not fully correspond to the participatory arrangements in cohort research (e.g. digital collection technologies, such as social media, give participants a more important role as they are able to report their data voluntarily). Hence, researchers should thrive towards beneficence principles while taking care to avoid paternalistic attitudes towards participants (e.g. not letting participants add information). Solutions to overcome potential obstacles related to preserving privacy and confidentiality include: y y Federated structures that allow data to be shared in a de-identified form. Growing out of international efforts to ensure data comparability in registry databases and biobanks, several mechanisms have been suggested to preserve confidentiality while benefiting from access to integrated data. FAIR-compliant platforms (i.e. Findable, Accessible, Interoperable, and Re-usable) and federated facilities are feasible options. The Personal Health Train (Beyan et al., 2020) is a federated, cooperative infrastructure in which the original data steward (fully implementing the FAIR principles) can maintain control over the data without sacrificing any confidentiality promises that have been arranged with the individual who provided the data. Similarly, the UK Data Service- Secure Lab (Jones et al., 2016), DataSHIELD