Anonymization of Phenotypic Data to Support Genotype Association Studies

• Activity Number: 106
• Type: Invited
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Committee on Privacy and Confidentiality
• Author(s): Bradley Malin*+
• Companies: Vanderbilt University

Abstract

This presentation will report on current data sharing practices for genomic data associated with clinical phenotypes in association studies. It will review the mechanism by which clinical data is collected, the extent to which it can be exploited for identification purposes, and anonymization strategies that can be invoked to minimize the risk of such information. With respect to the latter, this talk will illustrate how adversarial knowledge can be modeled, with a particular focus on how case-control cohorts are sampled from a certain population, and how it influences the degree to which information can be retained. The illustration will be based on an empirical investigation with data derived from over 100,000 patients' electronic medical records from a large academic medical center.