Date: June 2nd 2010 from 12pm to 1pm EST
Abstract:Standardized clinical features, such as diagnosis codes, derived from electronic medical records are increasingly combined with genomic sequences and shared to enable large-scale and low-cost Genome-Wide Association Studies (GWAS). However, disseminating these data “as is” may lead to patient re-identification, because they enable the linkage of genomic sequences to resources that contain the corresponding patients’ identity information. In this talk, I will illustrate the magnitude of this problem using patient records involved in a GWAS at the Vanderbilt University Medical Center. I will show that the vast majority of these records are uniquely re-identifiable and that popular protection methods, such as suppression and generalization, are inadequate to prevent this type of data linkage without harming data utility. Subsequently, I will present the first approach that can provably preserve privacy while still allowing data to be used in the context of GWAS and clinical case analysis tasks.
Speaker: Grigorios Loukides , Vanderbilt University
Biography: Grigorios Loukides received his Diploma from University of Crete and his PhD from Cardiff University, both in Computer Science. He is currently a Postdoctoral Research Fellow at the Health Information Privacy Laboratory at Vanderbilt University. His research interests are broadly in the area of data analysis, with an emphasis on privacy-preserving data publication and mining. His work has been published in top venues in computer science and biomedical informatics and has been reported on by various media outlets, including KDnuggets, MIT Technology Review, Nature News, Science News and Scientific American.