A de-identification tool for users in medical operations and public health

Abstract

Multiple clinical, business and operational use cases from Electronic Health Records (EHR) systems has resulted in the availability of large quantities of longitudinal data. The secondary use of these data for research provides opportunities to generate insights that can help shape the design and delivery of health care services. Methods that allow the de-identification of these datasets facilitate their use for research while minimizing the loss of privacy. In addition, to optimize the use of the data, particularly for longitudinal datasets, the ability to generate the same unique identifier for each individual allows re-linking as more data on the same individual becomes available over time. This paper details an open source software tool named Ezy De-Identifier, developed to make the assignment of pseudo-identifiers simple to users in a medical operations and public health setting, and reports the user's perspectives of the tool through a survey. In addition, we view the tool from the perspective of research reproducibility and security, and explore its application to generate a dataset satisfying established dataset protection requirements. � 2016 IEEE.