Semantic-Based Privacy Protection of Electronic Health Records for Collaborative Research

Abstract

Combined health information and web-based technologies can be used to support healthcare and research activities associated with electronic health records (EHRs).EHRs used for research purposes demand privacy, confidentiality and all information governance concerns are addressed. However, existing solutions are unable to meet the evolving research needs especially when supporting data access and linkage across organization boundaries. In this work, we show how semantic methods can aid in the specification and enforcement of policies for privacy protection. This is illustrated through a case study associated with the Australasian Diabetes Data Network (ADDN), the national paediatric type-1 diabetes data registry and the Australian Urban Research Infrastructure Network (AURIN) platform that supports Australia-wide access to urban and built environment data sets. Specifically we show that through extending the eXtensible Access Control Markup Language (XACML) with semantic capabilities, we are able to support fine-grained privacy-preserving policies leveraging semantic reasoning that is not directly available in XACML or other existing security policy specification languages.

Privacy in electronic health records: a systematic mapping study

“…A blockchain is used to effectively implement the proposed architecture because it maintains access control only for specific participants who will grant access to the data. In this work, (Lu and Sinnott 2016) propose a semantic methods XACML (eXtensible Access Control Markup Language) model provides access to personal information authorizing access to confidential enforcement of privacy protection policies.…”

From data flows to privacy‐benefit trade‐offs: A user‐centric semantic model

“…Web ontology language (OWL) and semantic web rule language (SWRL) have been widely utilized in specifying security and privacy policy constraints on data usage. [17][18][19][20] In this section, we use OWL and SWRL to formalize our model and show how reasoning can be done to detect privacy issues automatically. For the sake of simplicity, in this section we will focus on a subset of the entity types and relations.…”

Privacy-Preserving Access Control in Electronic Health Record Linkage

“…In this work, associations mined from "history publications" are extended and later enforced on the transformed linkage. More importantly, instead of focusing on simply protecting access to and use of sensitive attributes [17], this work effectively limits the chance of re-identification to an acceptable level and shows how semantic reasoning can deliver protection over real-time schemes and verify attributes based on the ever-growing external knowledge of collaborators and indeed potentially adversaries. To the best of our knowledge, no existing solutions have been designed for repeated linkage 2 which can be demanded in the long-term health studies [51].…”

“…Leveraging an existing solution [17] where XACML policy components are expressed and used for semantic reasoning in supporting access control policy compliance and privacy protection, we formalise anonymity schemes within obligation components so that the semantic-based anonymisation can be enforced when evaluating linkage requests and releases. Fig.…”