Preventing re-identification of anonymised study participants
A group of research funders today outlined the steps they will take to reduce the risk of anonymised individual research subjects in the UK being re-identified from genomic, epidemiological and social science data. These steps include possible sanctions, such as withdrawal of funding, for researchers who deliberately attempt to re-identify individuals without their consent.
The Medical Research Council along with Cancer Research UK, Economic and Social Research Council, and the Wellcome Trust have issued a joint response to a statement from the Expert Advisory Group on Data Access (EAGDA), a group of experts established by these four funders to provide strategic advice on emerging scientific, ethical and legal issues relating to data access.
The issue of re-identification was highlighted by a study in the journal Science last year, which showed that it was possible to identify participants in the 1000 Genomes Project through a complex technique that involved combining publicly available demographic information with anonymised genomic datasets.
In October 2013, EAGDA noted that "although the data in a genomic dataset may be fully anonymised in the conventional sense, cross-linking with general demographic data that are available from elsewhere makes it technically possible in some circumstances to triangulate the identities of individual research participants. Large datasets, particularly those including extensive genomic information, cannot be completely safe from inferential exploitation, including subject re-identification. Although the likelihood of such re-identification may currently be low for most types of study, it is likely to increase in the future."
The Expert Group has now set out a number of recommendations to reduce the impact of this risk and these have been accepted in full by the four organisations. They include: assessing and regularly reviewing the risk of re-identification through linking with other data; explaining the risk to participants when obtaining consent for studies; controlling access to data that could potentially identify individuals; and including sanctions that are proportionate to the nature of the offence, such as a withdrawal of funding, if researchers deliberately attempt to re-identify individuals from anonymised data.
Sir John Savill, Chief Executive of the Medical Research Council, says: "It's important that we protect the interests and anonymity of individuals while enabling research that benefits all society. As funders, we are committed to working together to reduce the risk of re-identification in a way that does not block valuable research to advance social and medical science and improve health."
Dr Jeremy Farrar, Director of the Wellcome Trust, says: "We accept the recommendations made by EAGDA to manage the potential risks and will set about implementing them into our funding policies and communicating this to the research community. Whilst it is impossible to eliminate entirely the risk of re-identification of individuals, it is possible to minimise this risk with proportionate safeguards. We believe that a deliberate attempt to re-identify individuals should be viewed as malpractice and be met with appropriate sanctions."
The four funders have committed to ensuring that both the EAGDA statement and their response are clearly communicated to their funded communities and are incorporated where appropriate into relevant guidance for studies. The principles and sanctions will be incorporated into their respective existing policies on research misconduct.