Towards a socio-technical approach for privacy requirements analysis for next-generation trusted research environments

Increasingly, advanced analytics methods – artificial intelligence/machine learning – are being used to discover value in big datasets. These methods are driving new data processing patterns and forms of research collaborations underpinned by the federated sharing and processing of data. Such multi-stakeholder processing raises the need for a standard privacy risk assessment framework that can fully deal with privacy risks arising in this context. In this paper, we argue that a socio-technical approach to privacy requirements analysis provides a crucial starting point for developing such a framework – as a means to foster a shared understanding of privacy risk in a specific context for effective risk communication, modelling, simulation, and evaluation. By way of example, we concentrate on three main areas. First, to describe the scope and boundaries for privacy risk assessment, we provide an overview of trusted research environments and emerging data usage patterns in operational health networks. Second, for effective and meaningful risk communication in respect of privacy concerns, expectations, and protective measures, we focus on the Five Safes as well-known principles and dimensions used to structure discussions and decision-making about access to sensitive data. Third, to promote a shared understanding through a conceptual mapping of common types of risk factors, we compare the ISO/IEC 27005 methodology for information security risk management with other selected privacy risk assessment methodologies.