Trails Learning Project
Health data that appears anonymous, such as DNA
records, can be re-identified to named patients via
location visit patterns, or trails. This is a realistic
privacy concern which continues to exist because data
holders do not collaborate prior to making disclosures.
In this paper, we present STRANON, a novel
computational protocol that enables data holders to
work together to determine records that can be
disclosed and satisfy a formal privacy protection
model. STRANON incorporates a secure encrypted
environment, so no data holder reveals information
until the trails of disclosed records are provably
unlinkable. We evaluate STRANON on real-world
datasets with known susceptibilities and demonstrate
data holders can release significant quantities of data
with zero trail re-identifiability.
Bradley Malin and Latanya Sweeney. A Secure Protocol to Distribute Unlinkable Health Data. Proceedings, Journal of the American Medical Informatics Association (AMIA). Washington, DC. Oct 2005: 485-489. (PDF).