Assessing the privacy risks of data sharing in genomics

The protection of identity of participants in medical research has traditionally been guaranteed by the maintenance of the confidentiality of health information through mechanisms such as only releasing data in an aggregated form or after identifying variables have been removed. This protection of privacy is regarded as a fundamental principle of research ethics, through which the support of research participants and the public is maintained. Whilst this traditional model was adopted for genetics and genomics research, and was generally considered broadly fit for purpose, we argue that this approach is increasingly untenable in genomics. Privacy risk assessments need to have regard to the whole data environment, not merely the quality of the dataset to be released in isolation. As sources of data proliferate, issues of privacy protection are increasingly problematic in relation to the release of genomic data. However, we conclude that, by paying careful attention to potential pitfalls, scientific funders and researchers can take an important part in attempts to safeguard the public and ensure the continuation of potentially important scientific research.