Table 3 Recommendations to give effect to a “Big Data Ethics by Design” approach

Recommendation 1

To mitigate concerns of mistrust, and to garner sustained trust by publics at the design stage, researchers ought to identify the particular drivers of mistrust relevant to that particular project and use an ethically informed approach to address and overcome mistrust. This should involve an in-depth mapping exercise, iteratively conducted over the beginning stages of the research endeavour (i.e. from the design and early-implementation stages) to identify potential drivers of mistrust. This mapping exercise in turn ideally should involve both desk-based research and in-person engagement with the communities being studied (e.g. information discussions, focus groups, interviews) to ensure that all potential drivers are identified. This also ought to include engagement with HCWs’ employers (e.g. NHS trusts, NHS Health Boards) and their professional regulators (e.g. General Medical Council, Nursing and Midwifery Council) 

Recommendation 2

Research teams should be mindful of relevant vulnerabilities of individual participants in their study, as well as groups of participants (e.g. sectors of HCWs, ethnic minority communities) and the risk of group harms such as stigma, mitigating these where appropriate. This also includes the need for an “intersectional awareness”, i.e. that individuals’ social and political lives are shaped by multiple axes of social division (such as race, socioeconomic status, gender, etc.), which interact and influence each other to create modes of discrimination and privilege. This awareness can be enhanced through the mapping exercise (Recommendation 1) as well as through engagement with organisations representing the groups of participants/communities involved in the research (e.g. a nurses or doctors union for a particular ethnic minority community)

Recommendation 3

Research teams should ensure there is meaningful involvement of communities in research, in a manner that is sensitive to the particularity of the community or group being studied, and is attuned to the nature of communication or involvement that would be required to facilitate their proper involvement. We suggest that this involvement can take different forms, participatory action approaches, and models of co-production. The form will necessarily be contextual and depend on the specific nature and objective of the project, as well as the community setting in which the research is taking place. This can range from citizens juries to workshops to inclusion of a community panel or advisory board within the research team. Research teams should engage with and encourage participation of individuals with morally relevant interests as key stakeholders. The makeup of the stakeholder groups should be reviewed on an ongoing basis to ensure those with morally relevant interests are included at all stages and afforded a meaningful opportunity to influence decision-making and the direction of the study. This meaningful involvement can help assure a watching brief is kept on the relevant ethical values at play, in turn helping assure that the study’s activities continue to be proportional to the benefit sought, and to identify any new issues of ethical concern that may arise

Recommendation 4

Where data and datasets are deemed to be anonymised according to data protection law and commonly accepted standards (e.g. those promulgated by a national data protection authority), studies should continue to consider what legal and ethical obligations arise even after anonymisation has been achieved. This “dynamic” approach to anonymisation includes keeping a watching brief on what is required to ensure the data remain anonymous throughout the lifecycle of their use in a study and in any subsequent, downstream uses, as well as a watching brief on confidentiality of the data, risk of re-identification, human rights, equality/anti-discrimination protection, and other relevant interests

Recommendation 5

Research teams should ensure that in the collection, storage and dissemination of research findings, there is no discriminatory impact (directly or indirectly) on the participants and communities involved in their study. Involvement of a panel of the particular community or communities involved during the research findings stage of the research, and prior to any public release or dissemination, ought to be considered. The ethical values of reciprocity and benefit sharing suggest that researchers ought to be engage communities from the outset to identify the tangible and intangible benefits that research participation might bring, and how involvement of those communities may lead to positive (health) outcomes for the community in which they are situated