Semantic Anonymization of Medical Records
Project Members
Tatiana Ringenberg, Julia M. Taylor, Victor Raskin
Tatiana Ringenberg, Julia M. Taylor, Victor Raskin
Abstract
With the availability of large amounts of data in the medical industry, it is becoming necessary, due to both regulatory and ethical concerns, to find unique ways of protecting patient identities. A name and social security number are no longer the only fields in a patient's record that can identify them. Data under HIPAA requires the removal of several Protected Health Information Identifiers. Symptoms themselves can also distinctly identify an individual in a large group. To prevent this, the Purdue OST Anonymization Project is using semantics to determine the degree to which any patient record is identifiable from others in a system. Our approach combines the conceptual mapping of Ontological Semantic Technology with the anonymity principles of K-Anonymity to semantically anonymize patient data for compliance with regulatory and research policies.